repository_name
stringclasses
316 values
func_path_in_repository
stringlengths
6
223
func_name
stringlengths
1
134
language
stringclasses
1 value
func_code_string
stringlengths
57
65.5k
func_documentation_string
stringlengths
1
46.3k
split_name
stringclasses
1 value
func_code_url
stringlengths
91
315
called_functions
listlengths
1
156
enclosing_scope
stringlengths
2
1.48M
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.prepare_params
python
def prepare_params(self): if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on]
Prepare the parameters passed to the templatetag
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L234-L256
[ "def get_expire_time(self):\n \"\"\"Update the expiry time with the multiplicator.\"\"\"\n expiry_time = super(TestCacheTag, self).get_expire_time()\n return self.multiplicator * expiry_time\n", "def get_expire_time(self):\n \"\"\"\n Return the expire time passed to the templatetag.\n Must be None or an integer.\n \"\"\"\n try:\n expire_time = self.node.expire_time.resolve(self.context)\n except template.VariableDoesNotExist:\n raise template.TemplateSyntaxError('\"%s\" tag got an unknown variable: %r' %\n (self.node.nodename, self.node.expire_time.var))\n try:\n if expire_time is not None:\n expire_time = str(expire_time)\n if not expire_time.isdigit():\n raise TypeError\n expire_time = int(expire_time)\n except (ValueError, TypeError):\n raise template.TemplateSyntaxError(\n '\"%s\" tag got a non-integer (or None) timeout value: %r' % (\n self.node.nodename, expire_time\n )\n )\n\n return expire_time\n", "def get_version(self):\n \"\"\"\n Return the stringified version passed to the templatetag.\n \"\"\"\n if not self.node.version:\n return None\n try:\n version = smart_str('%s' % self.node.version.resolve(self.context))\n except template.VariableDoesNotExist:\n raise template.TemplateSyntaxError('\"%s\" tag got an unknown variable: %r' %\n (self.node.nodename, self.node.version.var))\n\n return '%s' % version\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.get_expire_time
python
def get_expire_time(self): try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time
Return the expire time passed to the templatetag. Must be None or an integer.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L258-L281
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.get_version
python
def get_version(self): if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version
Return the stringified version passed to the templatetag.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L283-L295
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.hash_args
python
def hash_args(self): return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest()
Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L297-L302
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.get_cache_key_args
python
def get_cache_key_args(self): cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args
Return the arguments to be passed to the base cache key returned by `get_base_cache_key`.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L326-L338
[ "def hash_args(self):\n \"\"\"\n Take all the arguments passed after the fragment name and return a\n hashed version which will be used in the cache key\n \"\"\"\n return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest()\n", "def get_pk(self):\n \"\"\"\n Return the pk to use in the cache key. It's the first version of the\n templatetag arguments after the fragment name\n \"\"\"\n return self.vary_on[0]\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.cache_set
python
def cache_set(self, to_cache): self.cache.set(self.cache_key, to_cache, self.expire_time)
Set content into the cache
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L362-L366
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.join_content_version
python
def join_content_version(self, to_cache): parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts)
Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on)
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L368-L381
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.split_content_version
python
def split_content_version(self): try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None
Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L383-L406
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.decode_content
python
def decode_content(self): self.content = pickle.loads(zlib.decompress(self.content))
Decode (decompress...) the content got from the cache, to the final html
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L408-L413
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.render_node
python
def render_node(self): self.content = self.node.nodelist.render(self.context)
Render the template and save the generated content
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L421-L425
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.create_content
python
def create_content(self): self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment')
Render the template, apply options on it, and save it to the cache.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L427-L448
[ "def is_template_debug_activated():\n if django_version < (1, 8):\n return settings.TEMPLATE_DEBUG\n\n # not so simple now, it's an option of a template backend\n for template_settings in settings.TEMPLATES:\n if template_settings['BACKEND'] == 'django.template.backends.django.DjangoTemplates':\n return bool(template_settings.get('OPTIONS', {}).get('debug', False))\n\n return False\n", "def cache_set(self, to_cache):\n \"\"\"\n Set content into the cache\n \"\"\"\n self.cache.set(self.cache_key, to_cache, self.expire_time)\n", "def join_content_version(self, to_cache):\n \"\"\"\n Add the version(s) to the content to cache : internal version at first\n and then the template version if versioning is activated.\n Each version, and the content, are separated with `VERSION_SEPARATOR`.\n This method is called after the encoding (if \"compress\" or\n \"compress_spaces\" options are on)\n \"\"\"\n parts = [self.INTERNAL_VERSION]\n if self.options.versioning:\n parts.append(force_bytes(self.version))\n parts.append(force_bytes(to_cache))\n\n return self.VERSION_SEPARATOR.join(parts)\n", "def encode_content(self):\n \"\"\"\n Encode (compress...) the html to the data to be cached\n \"\"\"\n return zlib.compress(pickle.dumps(self.content))\n", "def render_node(self):\n \"\"\"\n Render the template and save the generated content\n \"\"\"\n self.content = self.node.nodelist.render(self.context)\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.load_content
python
def load_content(self): self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content)
It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L450-L490
[ "def is_template_debug_activated():\n if django_version < (1, 8):\n return settings.TEMPLATE_DEBUG\n\n # not so simple now, it's an option of a template backend\n for template_settings in settings.TEMPLATES:\n if template_settings['BACKEND'] == 'django.template.backends.django.DjangoTemplates':\n return bool(template_settings.get('OPTIONS', {}).get('debug', False))\n\n return False\n", "def cache_get(self):\n \"\"\"\n Get content from the cache\n \"\"\"\n return self.cache.get(self.cache_key)\n", "def split_content_version(self):\n \"\"\"\n Remove and return the version(s) from the cached content. First the\n internal version, and if versioning is activated, the template one.\n And finally save the content, but only if all versions match.\n The content saved is the encoded one (if \"compress\" or\n \"compress_spaces\" options are on). By doing so, we avoid decoding if\n the versions didn't match, to save some cpu cycles.\n \"\"\"\n try:\n nb_parts = 2\n if self.options.versioning:\n nb_parts = 3\n\n parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1)\n assert len(parts) == nb_parts\n\n self.content_internal_version = parts[0]\n if self.options.versioning:\n self.content_version = parts[1]\n\n self.content = parts[-1]\n except Exception:\n self.content = None\n", "def decode_content(self):\n \"\"\"\n Decode (decompress...) the content got from the cache, to the final\n html\n \"\"\"\n self.content = pickle.loads(zlib.decompress(self.content))\n", "def create_content(self):\n \"\"\"\n Render the template, apply options on it, and save it to the cache.\n \"\"\"\n self.render_node()\n\n if self.options.compress_spaces:\n self.content = self.RE_SPACELESS.sub(' ', self.content)\n\n if self.options.compress:\n to_cache = self.encode_content()\n else:\n to_cache = self.content\n\n to_cache = self.join_content_version(to_cache)\n\n try:\n self.cache_set(to_cache)\n except Exception:\n if is_template_debug_activated():\n raise\n logger.exception('Error when saving the cached template fragment')\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.render
python
def render(self): try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache()
Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L493-L516
[ "def is_template_debug_activated():\n if django_version < (1, 8):\n return settings.TEMPLATE_DEBUG\n\n # not so simple now, it's an option of a template backend\n for template_settings in settings.TEMPLATES:\n if template_settings['BACKEND'] == 'django.template.backends.django.DjangoTemplates':\n return bool(template_settings.get('OPTIONS', {}).get('debug', False))\n\n return False\n", "def load_content(self):\n \"\"\"\n It's the main method of the class.\n Try to load the template from cache, get the versions and decode the\n content.\n If something was wrong during this process (or if we had a\n `__regenerate__` value to True in the context), create new content and\n save it in cache.\n \"\"\"\n\n self.content = None\n\n if not self.regenerate:\n try:\n self.content = self.cache_get()\n except Exception:\n if is_template_debug_activated():\n raise\n logger.exception('Error when getting the cached template fragment')\n\n try:\n\n assert self.content\n\n self.split_content_version()\n\n assert self.content\n\n if self.content_internal_version != self.INTERNAL_VERSION or (\n self.options.versioning and self.content_version != self.version):\n self.content = None\n\n assert self.content\n\n if self.options.compress:\n self.decode_content()\n\n except Exception:\n self.create_content()\n\n self.content = smart_str(self.content)\n", "def render_nocache(self):\n \"\"\"\n Render the `nocache` blocks of the content and return the whole\n html\n \"\"\"\n tmpl = template.Template(''.join([\n # start by loading the cache library\n template.BLOCK_TAG_START,\n 'load %s' % self.get_templatetag_module(),\n template.BLOCK_TAG_END,\n # and surround the cached template by \"raw\" tags\n self.RAW_TOKEN_START,\n self.content,\n self.RAW_TOKEN_END,\n ]))\n return tmpl.render(self.context)\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.get_all_tags_and_filters_by_function
python
def get_all_tags_and_filters_by_function(): libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache
Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call.
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L519-L559
[ "def get_template_libraries():\n try:\n from django.template.base import libraries\n except ImportError:\n # Django >= 1.9\n from django.template import engines\n libraries = engines['django'].engine.template_libraries\n\n return libraries\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.get_templatetag_module
python
def get_templatetag_module(cls): if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls]
Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L562-L571
[ "def get_all_tags_and_filters_by_function():\n \"\"\"\n Return a dict with all the template tags (in the `tags` entry) and filters (in the\n `filters` entry) that are available.\n Both entries are a dict with the function as key, and a tuple with (library name, function\n name) as value.\n This is cached after the first call.\n \"\"\"\n\n libraries = get_template_libraries()\n\n force = False\n\n # We'll force the update of the cache if new libraries where added\n if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'):\n if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries:\n force = True\n\n if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'):\n CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries)\n available_tags = {}\n available_filters = {}\n\n for lib_name, lib in libraries.items():\n available_tags.update(\n (function, (lib_name, tag_name))\n for tag_name, function\n in lib.tags.items()\n )\n available_filters.update(\n (function, (lib_name, filter_name))\n for filter_name, function\n in lib.filters.items()\n )\n\n CacheTag.get_all_tags_and_filters_by_function._cache = {\n 'tags': available_tags,\n 'filters': available_filters\n }\n\n return CacheTag.get_all_tags_and_filters_by_function._cache\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.render_nocache
python
def render_nocache(self): tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context)
Render the `nocache` blocks of the content and return the whole html
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L573-L588
[ "def get_templatetag_module(cls):\n \"\"\"\n Return the templatetags module name for which the current class is used.\n It's used to render the nocache blocks by loading the correct module\n \"\"\"\n if cls not in CacheTag._templatetags_modules:\n # find the library including the main templatetag of the current class\n all_tags = cls.get_all_tags_and_filters_by_function()['tags']\n CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0]\n return CacheTag._templatetags_modules[cls]\n" ]
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.get_template_node_arguments
python
def get_template_node_arguments(cls, tokens): if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:]
Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing...
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L591-L602
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod @classmethod def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): """ Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag """ if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
twidi/django-adv-cache-tag
adv_cache_tag/tag.py
CacheTag.register
python
def register(cls, library_register, nodename='cache', nocache_nodename='nocache'): if cls in CacheTag._templatetags: raise RuntimeError('The adv-cache-tag class %s is already registered' % cls) CacheTag._templatetags[cls] = {} def templatetag_cache(parser, token): """ Return a new Node object for the main cache templatetag """ nodelist = parser.parse(('end%s' % nodename,)) parser.delete_first_token() args = cls.get_template_node_arguments(token.contents.split()) return cls.Node(nodename, nodelist, *args) library_register.tag(nodename, templatetag_cache) CacheTag._templatetags[cls]['cache'] = templatetag_cache def templatetag_raw(parser, token): """ Return a TextNode with all html not parsed, used for templatetags that need to not be parsed : the `nocache` one and the `RAW` one, used to surround cached html (to be not parsed again) Based on http://www.holovaty.com/writing/django-two-phased-rendering/ """ # Whatever is between {% nocache %} and {% endnocache %} will be preserved as # raw, un-rendered template code. text = [] parse_until = 'end%s' % token.contents tag_mapping = { TOKEN_TEXT: ('', ''), TOKEN_VAR: ('{{', '}}'), TOKEN_BLOCK: ('{%', '%}'), TOKEN_COMMENT: ('{#', '#}'), } # By the time this template tag is called, the template system has already # lexed the template into tokens. Here, we loop over the tokens until # {% endraw %} and parse them to TextNodes. We have to add the start and # end bits (e.g. "{{" for variables) because those have already been # stripped off in a previous part of the template-parsing process. while parser.tokens: token = parser.next_token() if token.token_type == TOKEN_BLOCK and token.contents == parse_until: return template.TextNode(''.join(text)) start, end = tag_mapping[token.token_type] text.append('%s%s%s' % (start, token.contents, end)) parser.unclosed_block_tag(parse_until) library_register.tag(cls.RAW_TOKEN, templatetag_raw) CacheTag._templatetags[cls]['raw'] = templatetag_raw def templatetag_nocache(parser, token): """ Return a TextNode with raw html from the `nocache` templatetag, and surround it with `endRAW` and `RAW` (precisely `cls.RAW_TOKEN_END` and `cls.RAW_TOKEN_START`). So for {% nocache %}foo{% endnocache %} we get {% endRAW... %}foo{% RAW... %} When the main cache templatetag content will be loaded from cache, it will be surrounded by the same templatetags, reversed. So if at first we had {% cache %}bar{% nocache %}foo{% endnocache %}baz{% endcache %} The cached version will be bar{% endRAW... %}foo{% RAW... %}baz And the final html to be rendered will be {% RAW... %}bar{% endRAW... %}foo{% RAW... %}baz{% endRAW... %} And the html within `RAW` and `endRAW` will not be parsed, as wanted """ # We'll load in the no-cache part all template tags and filters loaded in the main # template, to be able to use it when the no-cache will be rendered all_tags_and_filters = cls.get_all_tags_and_filters_by_function() available_tags = all_tags_and_filters['tags'] available_filters = all_tags_and_filters['filters'] needed = {} current_module = cls.get_templatetag_module() for function in parser.tags.values(): if function in available_tags: lib, name = available_tags[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) for function in parser.filters.values(): if function in available_filters: lib, name = available_filters[function] if lib == current_module: continue needed.setdefault(lib, set()).add(name) load_string = ''.join( '%sload %s from %s%s' % ( template.BLOCK_TAG_START, ' '.join(names), lib, template.BLOCK_TAG_END, ) for lib, names in needed.items() ) node = templatetag_raw(parser, token) node.s = cls.RAW_TOKEN_END + load_string + node.s + cls.RAW_TOKEN_START return node library_register.tag(nocache_nodename, templatetag_nocache) CacheTag._templatetags['nocache'] = templatetag_nocache
Register all needed templatetags, with these parameters : * library_register : the `register` object (result of `template.Library()`) in your templatetag module * nodename : the node to use for the cache templatetag (the default is "cache") * nocache_nodename : the node to use for the nocache templatetag
train
https://github.com/twidi/django-adv-cache-tag/blob/811f8db4dac73667c7d2fe0ea97a24969593eb8a/adv_cache_tag/tag.py#L605-L726
null
class CacheTag(object, metaclass=CacheTagMetaClass): """ The main class of `django-adv-cache-tag` which does all the work. To change its behaviour, simply change one or more of these settings (see in the `Meta` class for details) : * ADV_CACHE_VERSIONING * ADV_CACHE_COMPRESS * ADV_CACHE_COMPRESS_SPACES * ADV_CACHE_INCLUDE_PK * ADV_CACHE_BACKEND * ADV_CACHE_VERSION * ADV_CACHE_RESOLVE_NAME Or inherit from this class and don't forget to register your tag : from adv_cache_tag.tag import CacheTag register = template.Library() class MyCacheTag(CacheTag): # change something MyCacheTag.register(register, 'my_cache') By inheriting you can change many things as CacheTag implements a lot of small methods """ # Will change if the algorithm changes INTERNAL_VERSION = '1' # Used to separate internal version, template version, and the content VERSION_SEPARATOR = '::' # Regex used to reduce spaces/blanks (many spaces into one) RE_SPACELESS = re.compile(r'\s\s+') # generate a token for this site, based on the secret_key RAW_TOKEN = 'RAW_' + hashlib.sha1( b'RAW_TOKEN_SALT1' + force_bytes(hashlib.sha1( b'RAW_TOKEN_SALT2' + force_bytes(settings.SECRET_KEY) ).hexdigest()) ).hexdigest() # tokens to use around the already parsed parts of the cached template RAW_TOKEN_START = template.BLOCK_TAG_START + RAW_TOKEN + template.BLOCK_TAG_END RAW_TOKEN_END = template.BLOCK_TAG_START + 'end' + RAW_TOKEN + template.BLOCK_TAG_END # internal use only: keep reference to templatetags functions _templatetags = {} # internal use only: name of the templatetags module to load for this class and subclasses _templatetags_modules = {} options = None Node = Node class Meta: """ Options of this class. Accessible via cls.options or self.options. To force (and/or add) options in your own class, simply redefine a `Meta` class in your own main cache class with updated/add values """ # If versioning is activated (internal versioning is always on) versioning = getattr(settings, 'ADV_CACHE_VERSIONING', False) # If the content will be compressed before caching compress = getattr(settings, 'ADV_CACHE_COMPRESS', False) # If many spaces/blanks will be converted into one compress_spaces = getattr(settings, 'ADV_CACHE_COMPRESS_SPACES', False) # If a "pk" (you can pass what you want) will be added to the cache key include_pk = getattr(settings, 'ADV_CACHE_INCLUDE_PK', False) # The cache backend to use (or use the "default" one) cache_backend = getattr(settings, 'ADV_CACHE_BACKEND', 'default') # Part of the INTERNAL_VERSION configurable via settings internal_version = getattr(settings, 'ADV_CACHE_VERSION', '') # If the fragment name should be resolved or taken as is resolve_fragment = getattr(settings, 'ADV_CACHE_RESOLVE_NAME', False) # Use a metaclass to use the right class in the Node class, and assign Meta to options def __init__(self, node, context): """ Constructor of the Cache class: * preparing fields to be used later, * prepare the templatetag parameters * create the cache key """ super(CacheTag, self).__init__() # the actual Node object self.node = node # the context used for the rendering self.context = context # indicate that we force regenerating the cache, even if it exists self.regenerate = bool(self.context.get('__regenerate__', False)) # indicate if we only want html without parsing the nocache parts self.partial = bool(self.context.get('__partial__', False)) # the content of the template, will be used through the whole process self.content = '' # the version used in the cached templatetag self.content_version = None # Final "INTERNAL_VERSION" if self.options.internal_version: self.INTERNAL_VERSION = b'%s|%s' % (self.__class__.INTERNAL_VERSION, self.options.internal_version) else: self.INTERNAL_VERSION = force_bytes(self.__class__.INTERNAL_VERSION) self.VERSION_SEPARATOR = force_bytes(self.__class__.VERSION_SEPARATOR) # prepare all parameters passed to the templatetag self.expire_time = None self.version = None self.prepare_params() # get the cache and cache key self.cache = self.get_cache_object() self.cache_key = self.get_cache_key() def prepare_params(self): """ Prepare the parameters passed to the templatetag """ if self.options.resolve_fragment: self.fragment_name = self.node.fragment_name.resolve(self.context) else: self.fragment_name = str(self.node.fragment_name) # Remove quotes that surround the name for char in '\'\"': if self.fragment_name.startswith(char) or self.fragment_name.endswith(char): if self.fragment_name.startswith(char) and self.fragment_name.endswith(char): self.fragment_name = self.fragment_name[1:-1] break else: raise ValueError('Number of quotes around the fragment name is incoherent') self.expire_time = self.get_expire_time() if self.options.versioning: self.version = force_bytes(self.get_version()) self.vary_on = [template.Variable(var).resolve(self.context) for var in self.node.vary_on] def get_expire_time(self): """ Return the expire time passed to the templatetag. Must be None or an integer. """ try: expire_time = self.node.expire_time.resolve(self.context) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.expire_time.var)) try: if expire_time is not None: expire_time = str(expire_time) if not expire_time.isdigit(): raise TypeError expire_time = int(expire_time) except (ValueError, TypeError): raise template.TemplateSyntaxError( '"%s" tag got a non-integer (or None) timeout value: %r' % ( self.node.nodename, expire_time ) ) return expire_time def get_version(self): """ Return the stringified version passed to the templatetag. """ if not self.node.version: return None try: version = smart_str('%s' % self.node.version.resolve(self.context)) except template.VariableDoesNotExist: raise template.TemplateSyntaxError('"%s" tag got an unknown variable: %r' % (self.node.nodename, self.node.version.var)) return '%s' % version def hash_args(self): """ Take all the arguments passed after the fragment name and return a hashed version which will be used in the cache key """ return hashlib.md5(force_bytes(':'.join([urlquote(force_bytes(var)) for var in self.vary_on]))).hexdigest() def get_pk(self): """ Return the pk to use in the cache key. It's the first version of the templatetag arguments after the fragment name """ return self.vary_on[0] def get_base_cache_key(self): """ Return a string with format placeholder used as a source to compute the final cache key. Placeholders are : * %(nodename)s : the name of the templatetag * %(name)s : the fragment name passed to the templatetag * %(pk)s : the return of the `get_pk` method, passed only if `include_pk` is True * %(hash)s : the return of the `hash_args` method """ if self.options.include_pk: return 'template.%(nodename)s.%(name)s.%(pk)s.%(hash)s' else: return 'template.%(nodename)s.%(name)s.%(hash)s' def get_cache_key_args(self): """ Return the arguments to be passed to the base cache key returned by `get_base_cache_key`. """ cache_key_args = dict( nodename=self.node.nodename, name=self.fragment_name, hash=self.hash_args(), ) if self.options.include_pk: cache_key_args['pk'] = self.get_pk() return cache_key_args def get_cache_key(self): """ Compute and return the final cache key, using return values of `get_base_cache_key` and `get_cache_key_args`. """ return self.get_base_cache_key() % self.get_cache_key_args() def get_cache_object(self): """ Return the cache object to be used to set and get the values in cache. By default it's the default cache defined by django, but it can be every object with a `get` and a `set` method (or not, if `cache_get` and `cache_set` methods are overridden) """ return get_cache(self.node.cache_backend or self.options.cache_backend) def cache_get(self): """ Get content from the cache """ return self.cache.get(self.cache_key) def cache_set(self, to_cache): """ Set content into the cache """ self.cache.set(self.cache_key, to_cache, self.expire_time) def join_content_version(self, to_cache): """ Add the version(s) to the content to cache : internal version at first and then the template version if versioning is activated. Each version, and the content, are separated with `VERSION_SEPARATOR`. This method is called after the encoding (if "compress" or "compress_spaces" options are on) """ parts = [self.INTERNAL_VERSION] if self.options.versioning: parts.append(force_bytes(self.version)) parts.append(force_bytes(to_cache)) return self.VERSION_SEPARATOR.join(parts) def split_content_version(self): """ Remove and return the version(s) from the cached content. First the internal version, and if versioning is activated, the template one. And finally save the content, but only if all versions match. The content saved is the encoded one (if "compress" or "compress_spaces" options are on). By doing so, we avoid decoding if the versions didn't match, to save some cpu cycles. """ try: nb_parts = 2 if self.options.versioning: nb_parts = 3 parts = self.content.split(self.VERSION_SEPARATOR, nb_parts - 1) assert len(parts) == nb_parts self.content_internal_version = parts[0] if self.options.versioning: self.content_version = parts[1] self.content = parts[-1] except Exception: self.content = None def decode_content(self): """ Decode (decompress...) the content got from the cache, to the final html """ self.content = pickle.loads(zlib.decompress(self.content)) def encode_content(self): """ Encode (compress...) the html to the data to be cached """ return zlib.compress(pickle.dumps(self.content)) def render_node(self): """ Render the template and save the generated content """ self.content = self.node.nodelist.render(self.context) def create_content(self): """ Render the template, apply options on it, and save it to the cache. """ self.render_node() if self.options.compress_spaces: self.content = self.RE_SPACELESS.sub(' ', self.content) if self.options.compress: to_cache = self.encode_content() else: to_cache = self.content to_cache = self.join_content_version(to_cache) try: self.cache_set(to_cache) except Exception: if is_template_debug_activated(): raise logger.exception('Error when saving the cached template fragment') def load_content(self): """ It's the main method of the class. Try to load the template from cache, get the versions and decode the content. If something was wrong during this process (or if we had a `__regenerate__` value to True in the context), create new content and save it in cache. """ self.content = None if not self.regenerate: try: self.content = self.cache_get() except Exception: if is_template_debug_activated(): raise logger.exception('Error when getting the cached template fragment') try: assert self.content self.split_content_version() assert self.content if self.content_internal_version != self.INTERNAL_VERSION or ( self.options.versioning and self.content_version != self.version): self.content = None assert self.content if self.options.compress: self.decode_content() except Exception: self.create_content() self.content = smart_str(self.content) def render(self): """ Try to load content (from cache or by rendering the template). If it fails, return an empty string or raise the exception if it's a TemplateSyntaxError. With this, we can no parse and render the content included in the {% nocache %} blocks, but only if we have have this tag and if we don't have `__partial__` to True in the context (in this case we simple return the html with the {% nocache %} block not parsed. """ try: self.load_content() except template.TemplateSyntaxError: raise except Exception: if is_template_debug_activated(): raise logger.exception('Error when rendering template fragment') return '' if self.partial or self.RAW_TOKEN_START not in self.content: return self.content return self.render_nocache() @staticmethod def get_all_tags_and_filters_by_function(): """ Return a dict with all the template tags (in the `tags` entry) and filters (in the `filters` entry) that are available. Both entries are a dict with the function as key, and a tuple with (library name, function name) as value. This is cached after the first call. """ libraries = get_template_libraries() force = False # We'll force the update of the cache if new libraries where added if hasattr(CacheTag.get_all_tags_and_filters_by_function, '_len_libraries'): if len(libraries) != CacheTag.get_all_tags_and_filters_by_function._len_libraries: force = True if force or not hasattr(CacheTag.get_all_tags_and_filters_by_function, '_cache'): CacheTag.get_all_tags_and_filters_by_function._len_libraries = len(libraries) available_tags = {} available_filters = {} for lib_name, lib in libraries.items(): available_tags.update( (function, (lib_name, tag_name)) for tag_name, function in lib.tags.items() ) available_filters.update( (function, (lib_name, filter_name)) for filter_name, function in lib.filters.items() ) CacheTag.get_all_tags_and_filters_by_function._cache = { 'tags': available_tags, 'filters': available_filters } return CacheTag.get_all_tags_and_filters_by_function._cache @classmethod def get_templatetag_module(cls): """ Return the templatetags module name for which the current class is used. It's used to render the nocache blocks by loading the correct module """ if cls not in CacheTag._templatetags_modules: # find the library including the main templatetag of the current class all_tags = cls.get_all_tags_and_filters_by_function()['tags'] CacheTag._templatetags_modules[cls] = all_tags[CacheTag._templatetags[cls]['cache']][0] return CacheTag._templatetags_modules[cls] def render_nocache(self): """ Render the `nocache` blocks of the content and return the whole html """ tmpl = template.Template(''.join([ # start by loading the cache library template.BLOCK_TAG_START, 'load %s' % self.get_templatetag_module(), template.BLOCK_TAG_END, # and surround the cached template by "raw" tags self.RAW_TOKEN_START, self.content, self.RAW_TOKEN_END, ])) return tmpl.render(self.context) @classmethod def get_template_node_arguments(cls, tokens): """ Return the arguments taken from the templatetag that will be used to the Node class. Take a list of all tokens and return a list of real tokens. Here should be done some validations (number of tokens...) and eventually some parsing... """ if len(tokens) < 3: raise template.TemplateSyntaxError( "'%r' tag requires at least 2 arguments." % tokens[0]) return tokens[1], tokens[2], tokens[3:] @classmethod
pyfca/pyfca
pyfca/implications.py
L
python
def L(g,i): g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2)
recursively constructs L line for g; i = len(g)-1
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L33-L47
[ "Lwidth = Hwidth = lambda n: 3**n\n", "def L(g,i):\n \"\"\"recursively constructs L line for g; i = len(g)-1\"\"\"\n g1 = g&(2**i)\n if i:\n n = Lwidth(i)\n Ln = L(g,i-1)\n if g1:\n return Ln<<(2*n) | Ln<<n | Ln\n else:\n return int('1'*n,2)<<(2*n) | Ln<<n | Ln\n else:\n if g1:\n return int('000',2)\n else:\n return int('100',2)\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
H
python
def H(g,i): g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2)
recursively constructs H line for g; i = len(g)-1
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L48-L63
[ "Lwidth = Hwidth = lambda n: 3**n\n", "def L(g,i):\n \"\"\"recursively constructs L line for g; i = len(g)-1\"\"\"\n g1 = g&(2**i)\n if i:\n n = Lwidth(i)\n Ln = L(g,i-1)\n if g1:\n return Ln<<(2*n) | Ln<<n | Ln\n else:\n return int('1'*n,2)<<(2*n) | Ln<<n | Ln\n else:\n if g1:\n return int('000',2)\n else:\n return int('100',2)\n", "def H(g,i):\n \"\"\"recursively constructs H line for g; i = len(g)-1\"\"\"\n g1 = g&(2**i)\n if i:\n n = Hwidth(i)\n i=i-1\n Hn = H(g,i)\n if g1:\n return Hn<<(2*n) | Hn<<n | Hn\n else:\n return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn\n else:\n if g1:\n return int('111',2)\n else:\n return int('101',2)\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
UV_H
python
def UV_H(Hg,gw): lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K)
Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L65-L90
[ "Lwidth = Hwidth = lambda n: 3**n\n", "def istr(i,b,w,c=\"0123456789abcdefghijklmnopqrstuvwxyz\"):\n return ((w<=0 and i==0) and \" \") or (istr(i//b, b, w-1, c).lstrip() + c[i%b])\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
A
python
def A(g,i): g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2)
recursively constructs A line for g; i = len(g)-1
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L93-L107
[ "Awidth = lambda n: 2**n\n", "def A(g,i):\n \"\"\"recursively constructs A line for g; i = len(g)-1\"\"\"\n g1 = g&(2**i)\n if i:\n n = Awidth(i)\n An = A(g,i-1)\n if g1:\n return An<<n | An\n else:\n return int('1'*n,2)<<n | An\n else:\n if g1:\n return int('00',2)\n else:\n return int('10',2)\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
B
python
def B(g,i): g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0
recursively constructs B line for g; i = len(g)-1
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L109-L125
[ "Awidth = lambda n: 2**n\n", "Bwidth = lambda n:n*2**(n-1)\n", "def A(g,i):\n \"\"\"recursively constructs A line for g; i = len(g)-1\"\"\"\n g1 = g&(2**i)\n if i:\n n = Awidth(i)\n An = A(g,i-1)\n if g1:\n return An<<n | An\n else:\n return int('1'*n,2)<<n | An\n else:\n if g1:\n return int('00',2)\n else:\n return int('10',2)\n", "def B(g,i):\n \"\"\"recursively constructs B line for g; i = len(g)-1\"\"\"\n g1 = g&(2**i)\n if i:\n nA = Awidth(i)\n nB = Bwidth(i)\n i=i-1\n Bn = B(g,i)\n if g1:\n return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn\n else:\n return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn\n else:\n if g1:\n return 1\n else:\n return 0\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
B012
python
def B012(t,i): if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1)
Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L135-L151
[ "Awidth = lambda n: 2**n\n", "Bwidth = lambda n:n*2**(n-1)\n", "def A012(t,i):\n if i<0:\n return \"\"\n nA = Awidth(i)\n if t < nA:\n return \"0\"+A012(t,i-1)\n else:\n return \"2\"+A012(t-nA,i-1)\n", "def B012(t,i):\n \"\"\"\n Constructs ternary implication coding (0=not there, 2=U, 1=V)\n t is B column position\n i = |M|-1 to 0\n \"\"\"\n if not i:\n return \"1\"\n nA = Awidth(i)\n nB = Bwidth(i)\n nBB = nB + nA\n if t < nB:\n return \"0\"+B012(t,i-1)\n elif t < nBB:\n return \"1\"+A012(t-nB,i-1)\n else:\n return \"2\"+B012(t-nBB,i-1)\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
UV_B
python
def UV_B(Bg,gw): UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV
returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L153-L168
[ "Bwidth = lambda n:n*2**(n-1)\n", "def B012(t,i):\n \"\"\"\n Constructs ternary implication coding (0=not there, 2=U, 1=V)\n t is B column position\n i = |M|-1 to 0\n \"\"\"\n if not i:\n return \"1\"\n nA = Awidth(i)\n nB = Bwidth(i)\n nBB = nB + nA\n if t < nB:\n return \"0\"+B012(t,i-1)\n elif t < nBB:\n return \"1\"+A012(t-nB,i-1)\n else:\n return \"2\"+B012(t-nBB,i-1)\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
omega
python
def omega(imps): if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res
Calculates a measure for the size of the implication basis: \sum |U||V|
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L170-L191
null
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
respects
python
def respects(g,imp): if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV
g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L387-L401
[ "def istr(i,b,w,c=\"0123456789abcdefghijklmnopqrstuvwxyz\"):\n return ((w<=0 and i==0) and \" \") or (istr(i//b, b, w-1, c).lstrip() + c[i%b])\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
LL
python
def LL(n): if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2
constructs the LL context
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L532-L539
[ "def C1(w,h):\n return Context('\\n'.join(['1'*w]*h))\n", "def LL(n):\n \"\"\"constructs the LL context\"\"\"\n if (n<=0):return Context('0')\n else:\n LL1=LL(n-1)\n r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1\n r2 = LL1 - LL1 - LL1\n return r1 + r2\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
HH
python
def HH(n): if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2
constructs the HH context
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L540-L548
[ "def C1(w,h):\n return Context('\\n'.join(['1'*w]*h))\n", "def LL(n):\n \"\"\"constructs the LL context\"\"\"\n if (n<=0):return Context('0')\n else:\n LL1=LL(n-1)\n r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1\n r2 = LL1 - LL1 - LL1\n return r1 + r2\n", "def HH(n):\n \"\"\"constructs the HH context\"\"\"\n if (n<=0):return Context('1')\n else:\n LL1=LL(n-1)\n HH1=HH(n-1)\n r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1\n r2 = HH1 - HH1 - HH1\n return r1 + r2\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
AA
python
def AA(n): if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2
constructs the AA context
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L550-L557
[ "def C1(w,h):\n return Context('\\n'.join(['1'*w]*h))\n", "def AA(n):\n \"\"\"constructs the AA context\"\"\"\n if (n<=1):return Context('10\\n00')\n else:\n AA1=AA(n-1)\n r1 = C1(2**(n-1),2**(n-1)) - AA1\n r2 = AA1 - AA1\n return r1 + r2\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def BB(n): """constructs the BB context""" if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
BB
python
def BB(n): if (n<=1):return Context('0\n1') else: BB1=BB(n-1) AA1=AA(n-1) r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1 r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1; return r1 + r2
constructs the BB context
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L558-L566
[ "def C1(w,h):\n return Context('\\n'.join(['1'*w]*h))\n", "def AA(n):\n \"\"\"constructs the AA context\"\"\"\n if (n<=1):return Context('10\\n00')\n else:\n AA1=AA(n-1)\n r1 = C1(2**(n-1),2**(n-1)) - AA1\n r2 = AA1 - AA1\n return r1 + r2\n", "def BB(n):\n \"\"\"constructs the BB context\"\"\"\n if (n<=1):return Context('0\\n1')\n else:\n BB1=BB(n-1)\n AA1=AA(n-1)\n r1 = C1((n-1)*2**(n-2),2**(n-1)) - AA1 - BB1\n r2 = BB1 - C1(2**(n-1),2**(n-1)) - BB1;\n return r1 + r2\n" ]
#!/usr/bin/env python3 # encoding: utf-8 """ Implications ------------ This uses the python int as a bit field to store the FCA context. See this `blog`_ for more. .. _`blog`: http://rolandpuntaier.blogspot.com/2015/07/implications.html """ from math import trunc, log2 from functools import reduce from itertools import tee from collections import defaultdict def istr(i,b,w,c="0123456789abcdefghijklmnopqrstuvwxyz"): return ((w<=0 and i==0) and " ") or (istr(i//b, b, w-1, c).lstrip() + c[i%b]) digitat = lambda i,a,b: int(istr(i,b,a+1)[-a],b) digitat2 = lambda i,a: (i>>a)&1 #concatenate... horizontally = lambda K1,K2,b,w1,w2: [int(s,b) for s in [istr(k1,b,w1)+istr(k2,b,w2) for k1,k2 in zip(K1,K2)]] horizontally2 = lambda K1,K2,w1,w2: [(k1<<w2)|k2 for k1,k2 in zip(K1,K2)] vertically2 = vertically = lambda K1,K2: K1+K2 Lwidth = Hwidth = lambda n: 3**n def L(g,i): """recursively constructs L line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Lwidth(i) Ln = L(g,i-1) if g1: return Ln<<(2*n) | Ln<<n | Ln else: return int('1'*n,2)<<(2*n) | Ln<<n | Ln else: if g1: return int('000',2) else: return int('100',2) def H(g,i): """recursively constructs H line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Hwidth(i) i=i-1 Hn = H(g,i) if g1: return Hn<<(2*n) | Hn<<n | Hn else: return int('1'*n,2)<<(2*n) | L(g,i)<<n | Hn else: if g1: return int('111',2) else: return int('101',2) def UV_H(Hg,gw): """ Constructs implications and intents based on H gw = g width Hg = H(g), g is the binary coding of the attribute set UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets """ lefts = set() K = [] UV = [] p = Hwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Hg&pp: y = istr(p,3,gw) yy = y.replace('1','0') if yy not in lefts: if y.find('1') == -1:#y∈{0,2}^n K.append(y) else: UV.append(y) lefts.add(yy) return (UV,K) Awidth = lambda n: 2**n def A(g,i): """recursively constructs A line for g; i = len(g)-1""" g1 = g&(2**i) if i: n = Awidth(i) An = A(g,i-1) if g1: return An<<n | An else: return int('1'*n,2)<<n | An else: if g1: return int('00',2) else: return int('10',2) Bwidth = lambda n:n*2**(n-1) def B(g,i): """recursively constructs B line for g; i = len(g)-1""" g1 = g&(2**i) if i: nA = Awidth(i) nB = Bwidth(i) i=i-1 Bn = B(g,i) if g1: return Bn << (nA+nB) | int('1'*nA,2) << nB | Bn else: return int('1'*nB,2) << (nA+nB) | A(g,i) << nB | Bn else: if g1: return 1 else: return 0 def A012(t,i): if i<0: return "" nA = Awidth(i) if t < nA: return "0"+A012(t,i-1) else: return "2"+A012(t-nA,i-1) def B012(t,i): """ Constructs ternary implication coding (0=not there, 2=U, 1=V) t is B column position i = |M|-1 to 0 """ if not i: return "1" nA = Awidth(i) nB = Bwidth(i) nBB = nB + nA if t < nB: return "0"+B012(t,i-1) elif t < nBB: return "1"+A012(t-nB,i-1) else: return "2"+B012(t-nBB,i-1) def UV_B(Bg,gw): """ returns the implications UV based on B Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ UV = [] p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) UV.append(uv) return UV def omega(imps): """ Calculates a measure for the size of the implication basis: \sum |U||V| """ if isinstance(imps,v_Us_dict): return sum([omega(V) for U,V in imps.items()])#|V|=1 if isinstance(imps,list): return sum([omega(x) for x in imps]) if isinstance(imps,str): #imps = due[-1] try: U,V = imps.split("->") Us = U.split(",") if "," in U else U.split() Vs = V.split(",") if "," in V else V.split() res = len(Us)*len(Vs) return res except: return 0 if isinstance(imps,int): b=bin(imps)[2:] res = len([x for x in b if x=='1']) return res class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def koenig(self): """ This needs to be L = contextg.v_Us_B() """ L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z #Yn+1 #Lost = Ybar - (Y + ~Y) #assert len(Lost) == 0 def respects(g,imp): """ g is an int, where each bit is an attribute implication UV is ternary coded 1 = ∈V, 2 = ∈U, 0 otherwise g and UV have the same number of digits """ if isinstance(g,str): g = int(g,2) if isinstance(imp,int): imp = istr(imp,3,g.bit_length()) V = int(imp.replace('1','2').replace('2','1'),2) U = int(imp.replace('1','0').replace('2','1'),2) ginU = U&g == U ginV = V&g == V return not ginU or ginV class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right) C = Context def C1(w,h): return Context('\n'.join(['1'*w]*h)) def C0(w,h): return Context('\n'.join(['0'*w]*h)) #HH, LL, BB, AA are `\mathbb{H}`, `\mathbb{L}`, `\mathbb{B}`, `\mathbb{A}` from [1]_. #They are not needed to construct the implication basis. def LL(n): """constructs the LL context""" if (n<=0):return Context('0') else: LL1=LL(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - LL1 r2 = LL1 - LL1 - LL1 return r1 + r2 def HH(n): """constructs the HH context""" if (n<=0):return Context('1') else: LL1=LL(n-1) HH1=HH(n-1) r1 = C1(3**(n-1),2**(n-1)) - LL1 - HH1 r2 = HH1 - HH1 - HH1 return r1 + r2 def AA(n): """constructs the AA context""" if (n<=1):return Context('10\n00') else: AA1=AA(n-1) r1 = C1(2**(n-1),2**(n-1)) - AA1 r2 = AA1 - AA1 return r1 + r2 #.. _[1]: # # `Endliche Hüllensysteme und ihre Implikationenbasen <http://www.emis.de/journals/SLC/wpapers/s49koenig.pdf>`_ by Roman König.
pyfca/pyfca
pyfca/implications.py
v_Us_dict.koenig
python
def koenig(self): L = self Y = L - (L*L) while True: Ybar = Y + ~Y take = L - Ybar if not len(take): return Y else: ZZ = list(set(take)-set(Y))#use significant which is not in Y if len(ZZ) > 0: v = ZZ[0] z=(v,take[v][0]) else: z = next(take.flatten()) Yzgen = Y**z Y = (Y - Yzgen) + z
This needs to be L = contextg.v_Us_B()
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L364-L383
null
class v_Us_dict(defaultdict): """ In an implication U→u, u is the significant component. U is coded as int. u is the bit column of the implication's conclusion. {u:[U1,U2,...]} """ def __init__(self,Bg,gw): """ returns the implications {v:Us} based on B v is the significant component Bg = B(g), g∈2^M gw = |M|, M is the set of all attributes """ self.width = gw if isinstance(Bg,int): defaultdict.__init__(self,list) p = Bwidth(gw) pp = 2**p while p: pp = pp>>1 p = p-1 if Bg&pp: uv = B012(p,gw-1) #let's find minima regarding product order #{v:[Umin1,Umin2,...]} v = uv.find('1')#v=significant u = uv[:v]+'0'+uv[v+1:] u = int(u.replace('2','1'),2) Umin_s = self[gw-v-1]#bit position from right it = [i for i,U in enumerate(Umin_s) if U&u==u] for i in reversed(it): del Umin_s[i] else: Umin_s.append(u) elif isinstance(Bg,list): defaultdict.__init__(self,list) for k,v in Bg: assert isinstance(v,list) self[k] += v else: defaultdict.__init__(self,list,Bg) def __eq__(self, other): if len(self) != len(other): return False for v,U in self.items(): if v not in other: return False Uo = other[v] if not set(Uo)==set(U): return False return True def Code012(self): for v,Us in self.items(): vleft = self.width - v - 1 for u in Us: b = bin(u)[2:] w0 = self.width-len(b) c01 = '0'*w0+b c01 = c01.replace('1','2') c01 = c01[:vleft]+'1'+c01[vleft+1:] yield c01 def __str__(self): return defaultdict.__str__(self).replace('defaultdict','v_Us_dict') def __len__(self): return sum((len(x) for x in self.values())) def flatten(self): for v,Us in self.items(): for u in Us: yield (v,u) def __add__(self, other): res = v_Us_dict([],self.width) if isinstance(other,tuple): other = {other[0]:[other[1]]} keys = set(self)|set(other) for v in keys: t = set() if v in self: t |= set(self[v]) if v in other: t |= set(other[v]) if t: res[v] = list(t) return res def __sub__(self, other): res = v_Us_dict([],self.width) for v,U in self.items(): r = list(set(U) - set(other[v])) if r: res[v] = r return res def __mul__(self, other): """ This is the o operation in [1]_, that represents the 3rd Armstrong rule. It returns combinations for i‡j: (i,u1|u2) or (j,u1|u2), """ res = v_Us_dict([],self.width) if id(self)==id(other): s = iter(self.items()) try: while True: v1, us1 = next(s) vv1 = 2**v1 s, ss = tee(s)#remember s and iterate with copy ss try: while True: v2, us2 = next(ss) vv2 = 2**v2 for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) except StopIteration: pass except StopIteration: pass else: for v1,us1 in self.items(): vv1 = 2**v1 for v2,us2 in other.items(): vv2 = 2**v2 if v1 != v2: for u1 in us1: for u2 in us2: if vv2&u1 and not vv1&u2: res[v1].append((u1|u2)&~vv2) elif vv1&u2 and not vv2&u1: res[v2].append((u1|u2)&~vv1) for v,U in res.items(): res[v] = list(set(U))#remove duplicates return res def __invert__(self): """ U->v generated from L=∪ min L_i via the 3rd Armstrong rule Note, that this can become bigger than L. """ Y = self Yn = Y*Y while True: YnplusY = Yn+Y Yg = Yn*YnplusY #YgenNotInL = Yg - L #YgenInL = Yg - YgenNotInL #Yn1 = Yn + YgenInL Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn def __pow__(self, other): """ 'other' is a (v,u) couple generates U->v involving 'other' #other = (0,64) """ Y = self Z = v_Us_dict({other[0]:[other[1]]},self.width) Yn = Y*Z while True: YnplusY = Yn+Y Yg = Z*YnplusY #this does not work for test_basis1 #YnplusZ = Yn+Z #Yg = YnplusZ*YnplusY Yn1 = Yn + Yg if Yn1 == Yn: break Yn = Yn1 return Yn #Yn+1
pyfca/pyfca
pyfca/implications.py
Context.column
python
def column(self, i): return ''.join([str(digitat2(r,i)) for r in self])
from right
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L432-L434
null
class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right)
pyfca/pyfca
pyfca/implications.py
Context.UV_H
python
def UV_H(self): h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width)
UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets.
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L454-L463
[ "def UV_H(Hg,gw):\n \"\"\"\n Constructs implications and intents based on H\n gw = g width\n Hg = H(g), g is the binary coding of the attribute set\n UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U)\n K = all closed sets\n \"\"\"\n lefts = set()\n K = []\n UV = []\n p = Hwidth(gw)\n pp = 2**p\n while p:\n pp = pp>>1\n p = p-1\n if Hg&pp:\n y = istr(p,3,gw)\n yy = y.replace('1','0')\n if yy not in lefts: \n if y.find('1') == -1:#y∈{0,2}^n\n K.append(y)\n else:\n UV.append(y)\n lefts.add(yy)\n return (UV,K)\n" ]
class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right)
pyfca/pyfca
pyfca/implications.py
Context.UV_B
python
def UV_B(self): h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width)
returns UV = all respected U->Ux in ternary coding (1=V,2=U)
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L464-L469
[ "def UV_B(Bg,gw):\n \"\"\"\n returns the implications UV based on B\n Bg = B(g), g∈2^M\n gw = |M|, M is the set of all attributes\n \"\"\"\n UV = []\n p = Bwidth(gw)\n pp = 2**p\n while p:\n pp = pp>>1\n p = p-1\n if Bg&pp:\n uv = B012(p,gw-1)\n UV.append(uv)\n return UV\n" ]
class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def v_Us_B(self): """ returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_ """ Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right)
pyfca/pyfca
pyfca/implications.py
Context.v_Us_B
python
def v_Us_B(self): Bg = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) gw = self.width return v_Us_dict(Bg, gw)
returns the implications {v:Us} based on B This is L=∪ min L_i in [1]_
train
https://github.com/pyfca/pyfca/blob/cf8cea9e76076dbf4bb3f38996dcb5491b0eb0b0/pyfca/implications.py#L470-L477
null
class Context(list): def __init__(self, *args, **kwargs): """Context can be initialized with - a rectangular text block of 0s and 1s - a list of ints and a "width" keyword argument. A "mapping" keyword argument as list associates the bits with objects of any kind. """ if isinstance(args[0],str): lines = [s.strip() for s in args[0].splitlines() if s.strip()] linelens = [len(tt) for tt in lines] self.width = linelens[0] samelen = linelens.count(linelens[0])==len(linelens) assert samelen, "Context needs all lines to be of same number of 0s and 1s" super().__init__([int(s,2) for s in lines]) else: super().__init__(*args) self.width = kwargs['width'] try: self.mapping = kwargs['mapping'] except: self.mapping = [i for i in range(self.width)] def __add__(self, other): c = Context(list.__add__(self,other),width=self.width) return c def __sub__(self, other): c = Context(horizontally2(self,other,self.width,other.width),width=self.width+other.width) return c def column(self, i): """from right""" return ''.join([str(digitat2(r,i)) for r in self]) def row(self, i): try: r = istr(self[i],2,self.width) except IndexError: r = '0'*self.width return r def __getitem__(self,xy): if isinstance(xy,tuple): return digitat2(list.__getitem__(self,xy[0]),xy[1]) else: return list.__getitem__(self,xy) def transpose(self): cs='\n'.join([self.column(i) for i in reversed(range(self.width))]) return Context(cs) def __str__(self): rs='\n'.join([self.row(i) for i in range(len(self))]) return rs def size(self): return self.width, len(self) def UV_H(self): """ UV = all non-trivial (!V⊂U) implications U->V with UuV closed; in ternary coding (1=V,2=U) K = all closed sets This is UV_H function, but the returned implications are respected by all attribute sets of this context. This corresponds to a multiplication or & operation of the Hg sets. """ h = reduce(lambda x,y:x&y,(H(g,self.width-1) for g in self)) return UV_H(h, self.width) def UV_B(self): """ returns UV = all respected U->Ux in ternary coding (1=V,2=U) """ h = reduce(lambda x,y:x&y,(B(g,self.width-1) for g in self)) return UV_B(h, self.width) def respects(self, implications): if isinstance(implications,v_Us_dict): implications = implications.Code012() for g in self: for i in implications: if not respects(g,i): return False return True def __call__(self, intOrCode012, right = None): """ mapping from bits to attributes using mapping (which defaults to ints) - right, if available, is the conclusion of the implication; used if intOrCode012 is int """ if isinstance(intOrCode012,v_Us_dict): return frozenset(self(x,right=i) for i,x in intOrCode012.items()) if isinstance(intOrCode012,list): return frozenset(self(x,right=right) for x in intOrCode012) if isinstance(intOrCode012,int): res = [] pp = 1 for pos in range(self.width): if intOrCode012&pp: res.append(self.mapping[-pos-1]) pp = pp*2 if right != None: return (frozenset(res),frozenset([self.mapping[-right-1]])) else: return frozenset(res) if isinstance(intOrCode012,str): left = [] right = [] for pos in range(self.width): if intOrCode012[pos] == '2': left.append(self.mapping[pos]) elif intOrCode012[pos] == '1': right.append(self.mapping[pos]) if left: if right: return (frozenset(left),frozenset(right)) else: return frozenset(left) else: return frozenset(right)
Metatab/geoid
geoid/core.py
parse_to_gvid
python
def parse_to_gvid(v): from geoid.civick import GVid from geoid.acs import AcsGeoid m1 = '' try: return GVid.parse(v) except ValueError as e: m1 = str(e) try: return AcsGeoid.parse(v).convert(GVid) except ValueError as e: raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e)))
Parse an ACS Geoid or a GVID to a GVID
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L342-L357
[ "def parse(cls, gvid, exception=True):\n \"\"\"\n Parse a string value into the geoid of this class.\n\n :param gvid: String value to parse.\n :param exception: If true ( default) raise an eception on parse erorrs. If False, return a\n 'null' geoid.\n :return:\n \"\"\"\n\n if gvid == 'invalid':\n return cls.get_class('null')(0)\n\n if not bool(gvid):\n return None\n\n if not isinstance(gvid, six.string_types):\n raise TypeError(\"Can't parse; not a string. Got a '{}' \".format(type(gvid)))\n\n try:\n if not cls.sl:\n # Civick and ACS include the SL, so can call from base type.\n if six.PY3:\n fn = cls.decode\n else:\n fn = cls.decode.__func__\n\n sl = fn(gvid[0:cls.sl_width])\n else:\n sl = cls.sl # Otherwise must use derived class.\n\n except ValueError as e:\n if exception:\n raise ValueError(\"Failed to parse gvid '{}': {}\".format(gvid, str(e)))\n else:\n return cls.get_class('null')(0)\n\n try:\n cls = cls.sl_map[sl]\n except KeyError:\n if exception:\n raise ValueError(\"Failed to parse gvid '{}': Unknown summary level '{}' \".format(gvid, sl))\n else:\n return cls.get_class('null')(0)\n\n m = cls.regex.match(gvid)\n\n if not m:\n raise ValueError(\"Failed to match '{}' to '{}' \".format(gvid, cls.regex_str))\n\n d = m.groupdict()\n\n if not d:\n return None\n\n if six.PY3:\n fn = cls.decode\n else:\n fn = cls.decode.__func__\n\n d = {k: fn(v) for k, v in d.items()}\n\n try:\n del d['sl']\n except KeyError:\n pass\n\n return cls(**d)\n" ]
""" CLasses for working with census Geoids """ import inspect import re import sys import six names = { # (summary level value, base 10 chars, Base 62 chars, prefix fields) 'null': 1, 'us': 10, 'region': 20, 'division': 30, 'state': 40, 'county': 50, 'cosub': 60, 'place': 160, 'ua': 400, 'tract': 140, 'blockgroup': 150, 'block': 101, 'sdelm': 950, 'sdsec': 960, 'sduni': 970, 'zcta': 860, 'zip': 1200, 'sldl': 620, 'sldu': 610, 'cdcurr': 500, # Other Levels that don't have proper names yet. For these .allval() and # simplify() don't work properly. 'state_aianhh': 260, 'necta_nectadiv_state_county_cousub': 358, 'state_aianhh_place': 269, 'aianhh_state_county': 270, 'state_cbsa_metdiv': 323, 'state_aianhh280': 280, 'state_place_county': 155, 'aianhh_aitsce_state': 290, 'state_aianhh_aihhtli': 283, 'state_cdcurr_aianhh': 550, 'state_concit': 170, 'state_concit_place': 172, 'state_aianhh_aihhtli286': 286, 'cbsa': 310, 'cbsa_state': 311, 'cbsa_state_place': 312, 'cbsa_state_county': 313, 'cbsa_metdiv': 314, 'cbsa_metdiv_state': 315, 'state_cbsa': 320, 'state_cbsa_place': 321, 'state_cbsa_county': 322, 'state_county_cousub_submcd': 67, 'state_cbsa_metdiv_county': 324, 'state_county_cousub_place': 70, 'necta_state_county': 353, 'state_puma5': 795, 'csa': 330, 'csa_state': 331, 'csa_cbsa': 332, 'csa_cbsa_state': 333, 'cnecta': 335, 'state_county_cousub_place_tract': 80, 'cnecta_necta': 337, 'cnecta_necta_state': 338, 'state_csa': 340, 'state_csa_cbsa': 341, 'state_cnecta': 345, 'state_cnecta_necta': 346, 'necta': 350, 'necta_state': 351, 'necta_state_place': 352, 'cnecta_state': 336, 'necta_state_county_cousub': 354, 'necta_nectadiv': 355, 'necta_nectadiv_state': 356, 'state_anrc': 230, 'necta_nectadiv_state_county': 357, 'state_necta': 360, 'cbsa_metdiv_state_county': 316, 'state_necta_county': 362, 'state_necta_county_cousub': 363, 'state_necta_nectadiv': 364, 'state_necta_nectadiv_county': 365, 'state_necta_nectadiv_county_cousub': 366, 'ua_state': 410, 'ua_state_county': 430, 'state_sldu_county': 612, 'state_sldu': 610, 'state_sldl_county': 622, 'state_sldl': 620, 'state_cdcurr_county': 510, 'state_necta_place': 361, 'aianhh': 250, 'aianhh_aitsce': 251, 'aianhh_aihhtli': 252, 'state_sldl_county': 622, 'aianhh_aihhtli254': 254 } lengths = { 'null': 1, 'aianhh': 4, # American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 'aihhtli': '1', # American Indian Trust Land/ Hawaiian Home Land Indicator. A str b/c Census val is a str 'aitsce': 3, # American Indian Tribal Subdivision (Census) 'anrc': 5, # Alaska Native Regional Corporation (FIPS) 'blkgrp': 1, # Block Group 'blockgroup': 1, # Block Group 'block': 4, # Block 'cbsa': 5, # Metropolitan and Micropolitan Statistical Area 'cdcurr': 2, # Current Congressional District *** 'cnecta': 3, # New England City and Town Combined Statistical Area 'concit': 5, # Consolidated City 'county': 3, # County of current residence 'cousub': 5, # County Subdivision (FIPS) 'cosub': 5, # County Subdivision (FIPS) 'csa': 3, # Combined Statistical Area 'division': 1, # Census Division 'metdiv': 5, # Metropolitan Statistical Area- Metropolitan Division 'necta': 5, # New England City and Town Area 'nectadiv': 5, # New England City and Town Area Division 'place': 5, # Place (FIPS Code) 'puma5': 5, # Public Use Microdata Area 5% File 'region': 1, # Census Region 'sdelm': 5, # State-School District (Elementary) 'sdsec': 5, # State-School District (Secondary) 'sduni': 5, # State-School District (Unified) 'sldl': '3', # State Legislative District Lower. A String to signal that the census value is a string 'sldu': '3', # State Legislative District Upper. A String to signal that the census value is a string 'state': 2, # State (FIPS Code) 'submcd': 5, # Subminor Civil Division (FIPS) 'tract': 6, # Census Tract 'ua': 5, # Urban Area 'ur': 1, # Urban/Rural 'us': 0, 'zcta': 5, # Nonstandard 'zip': 5, } segments = { 1: ['null'], # United States 10: ['us'], # United States 20: ['region'], # Region 30: ['division'], # Division 40: ['state'], # State 50: ['state', 'county'], # County 60: ['state', 'county', 'cousub'], # County Subdivision 67: ['state', 'county', 'cousub', 'submcd'], # State (Puerto Rico Only)-County-County Subdivision-Subbarrio 70: ['state', 'county', 'cousub', 'place'], # County Subdivision-Place/Remainder 80: ['state', 'county', 'cousub', 'place', 'tract'], # County Subdivision-Place/Remainder-Census Tract 101: ['state', 'county', 'tract', 'block'], 140: ['state', 'county', 'tract'], # Census Tract 150: ['state', 'county', 'tract', 'blockgroup'], # Census Tract-Block Group 155: ['state', 'place', 'county'], # Place-County 160: ['state', 'place'], # Place 170: ['state', 'concit'], # Consolidated City 172: ['state', 'concit', 'place'], # Consolidated City-Place Within Consolidated City 230: ['state', 'anrc'], # State-Alaska Native Regional Corporation 250: ['aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land 251: ['aianhh', 'aitsce'], # American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder 252: ['aianhh', 'aihhtli'], # American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)4 254: ['aianhh', 'aihhtli'], # American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 260: ['state', 'aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State 269: ['state', 'aianhh', 'place'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder 270: ['aianhh', 'state', 'county'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County 280: ['state', 'aianhh'], # State-American Indian Area/Alaska Native Area/Hawaiian Home Land 283: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only) 286: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 290: ['aianhh', 'aitsce', 'state'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Tribal Subdivision/Remainder-State 310: ['cbsa'], # CBSA 311: ['cbsa', 'state'], # CBSA-State-County 312: ['cbsa', 'state', 'place'], # CBSA-State-Principal City 313: ['cbsa', 'state', 'county'], # CBSA-State-County 314: ['cbsa', 'metdiv'], # Metropolitan Statistical Area/Metropolitan Division 315: ['cbsa', 'metdiv', 'state'], # Metropolitan Statistical Area/Metropolitan Division-State 316: ['cbsa', 'metdiv', 'state', 'county'], # Metropolitan Statistical Area/Metropolitan Division-State-County 320: ['state', 'cbsa'], # State- CBSA 321: ['state', 'cbsa', 'place'], # State- CBSA -Principal City 322: ['state', 'cbsa', 'county'], # State- CBSA -County 323: ['state', 'cbsa', 'metdiv'], # State- Metropolitan Statistical Area/Metropolitan Division 324: ['state', 'cbsa', 'metdiv', 'county'], # State- Metropolitan Statistical Area/Metropolitan Division-County 330: ['csa'], # Combined Statistical Area 331: ['csa', 'state'], # Combined Statistical Area-State 332: ['csa', 'cbsa'], # Combined Statistical Area-CBSA 333: ['csa', 'cbsa', 'state'], # Combined Statistical Area-CBSA-State 335: ['cnecta'], # Combined New England City and Town Area 336: ['cnecta', 'state'], # Combined New England City and Town Area -State 337: ['cnecta', 'necta'], # Combined New England City and Town Area -New England City and Town Area 338: ['cnecta', 'necta', 'state'], # Combined New England City and Town Area -New England City and Town Area-State 340: ['state', 'csa'], # State-Combined Statistical Area 341: ['state', 'csa', 'cbsa'], # State-Combined Statistical Area-CBSA 345: ['state', 'cnecta'], # State-Combined New England City and Town Area 346: ['state', 'cnecta', 'necta'], # State-Combined New England City and Town Area-New England City and Town Area 350: ['necta'], # New England City and Town Area 351: ['necta', 'state'], # New England City and Town Area-State 352: ['necta', 'state', 'place'], # New England City and Town Area-State-Principal City 353: ['necta', 'state', 'county'], # New England City and Town Area-State-County 354: ['necta', 'state', 'county', 'cousub'], # New England City and Town Area-State-County-County Subdivision 355: ['necta', 'nectadiv'], # New England City and Town Area (NECTA)-NECTA Division 356: ['necta', 'nectadiv', 'state'], # New England City and Town Area (NECTA)-NECTA Division-State 357: ['necta', 'nectadiv', 'state', 'county'], # New England City and Town Area (NECTA)-NECTA Division-State-County 358: ['necta', 'nectadiv', 'state', 'county', 'cousub'], # New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision 360: ['state', 'necta'], # State-New England City and Town Area 361: ['state', 'necta', 'place'], # State-New England City and Town Area-Principal City 362: ['state', 'necta', 'county'], # State-New England City and Town Area-County 363: ['state', 'necta', 'county', 'cousub'], # State-New England City and Town Area-County-County Subdivision 364: ['state', 'necta', 'nectadiv'], # State-New England City and Town Area (NECTA)-NECTA Division 365: ['state', 'necta', 'nectadiv', 'county'], # State-New England City and Town Area (NECTA)-NECTA Division-County 366: ['state', 'necta', 'nectadiv', 'county', 'cousub'], # State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision 400: ['ua'], # Urban Area, 410: ['ua', 'state'], # Urban Area, State, 430: ['ua','state','county'], # Urban Area, State, County, 500: ['state', 'cdcurr'], # Congressional District 510: ['state', 'cdcurr', 'county'], # 550: ['state', 'cdcurr', 'aianhh'], # Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land 610: ['state', 'sldu'], # State Senate District 612: ['state', 'sldu', 'county'], # State Senate District-County 620: ['state', 'sldl'], # State House District 622: ['state', 'sldl', 'county'], # State House District-County 795: ['state', 'puma5'], # State-Public Use MicroSample Area 5% 860: ['zcta'], 950: ['state', 'sdelm'], # State-Elementary School District 960: ['state', 'sdsec'], # State-High School District 970: ['state', 'sduni'], # State-Unified School District # Nonstandard 1200: ['zip'] } descriptions = { 1: 'United States', 10: 'United States', 20: 'Region', 30: 'Division', 40: 'State', 50: 'County', 60: 'County Subdivision', 67: 'State (Puerto Rico Only)-County-County Subdivision-Subbarrio', 70: 'County Subdivision-Place/Remainder', 80: 'County Subdivision-Place/Remainder-Census Tract', 101: 'block', 140: 'Census Tract', 150: 'Census Tract-Block Group', 155: 'Place-County', 160: 'Place', 170: 'Consolidated City', 172: 'Consolidated City-Place Within Consolidated City', 230: 'State-Alaska Native Regional Corporation', 250: 'American Indian Area/Alaska Native Area/Hawaiian Home Land', 251: 'American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder', 252: 'American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)', 254: 'American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land', 260: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State', 269: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder', 270: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County', 280: 'State-American Indian Area/Alaska Native Area/Hawaiian Home Land', 283: 'aihhtli', 286: 'aihhtli', 290: 'state', 310: 'CBSA', 311: 'CBSA-State-County', 312: 'CBSA-State-Principal City', 313: 'CBSA-State-County', 314: 'Metropolitan Statistical Area/Metropolitan Division', 315: 'Metropolitan Statistical Area/Metropolitan Division-State', 316: 'Metropolitan Statistical Area/Metropolitan Division-State-County', 320: 'State- CBSA', 321: 'State- CBSA -Principal City', 322: 'State- CBSA -County', 323: 'State- Metropolitan Statistical Area/Metropolitan Division', 324: 'State- Metropolitan Statistical Area/Metropolitan Division-County', 330: 'Combined Statistical Area', 331: 'Combined Statistical Area-State', 332: 'Combined Statistical Area-CBSA', 333: 'Combined Statistical Area-CBSA-State', 335: 'Combined New England City and Town Area', 336: 'Combined New England City and Town Area -State', 337: 'Combined New England City and Town Area -New England City and Town Area', 338: 'Combined New England City and Town Area -New England City and Town Area-State', 340: 'State-Combined Statistical Area', 341: 'State-Combined Statistical Area-CBSA', 345: 'State-Combined New England City and Town Area', 346: 'State-Combined New England City and Town Area-New England City and Town Area', 350: 'New England City and Town Area', 351: 'New England City and Town Area-State', 352: 'New England City and Town Area-State-Principal City', 353: 'New England City and Town Area-State-County', 354: 'New England City and Town Area-State-County-County Subdivision', 355: 'New England City and Town Area (NECTA)-NECTA Division', 356: 'New England City and Town Area (NECTA)-NECTA Division-State', 357: 'New England City and Town Area (NECTA)-NECTA Division-State-County', 358: 'New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision', 360: 'State-New England City and Town Area', 361: 'State-New England City and Town Area-Principal City', 362: 'State-New England City and Town Area-County', 363: 'State-New England City and Town Area-County-County Subdivision', 364: 'State-New England City and Town Area (NECTA)-NECTA Division', 365: 'State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision', 400: 'Urban Area,', 410: 'Urban Area, State,', 430: 'Urban Area, State, County,', 500: 'Congressional District', 510: 'Congressional District, County', 550: 'Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land', 610: 'State Senate District', 612: 'State Senate District-County', 620: 'State House District', 622: 'State House District-County', 795: 'State-Public Use MicroSample Area 5%', 860: 'ZIP Code Tabulation Area', 950: 'State-Elementary School District', 960: 'State-High School District', 970: 'State-Unified School District', } plurals = { 'county': 'counties', 'place': 'places', 'Sdlu': 'State ' } class NotASummaryName(Exception): """An argument was not one of the valid summary names""" class ParseError(Exception): """Error parsing a geoid""" def base62_encode(num): """Encode a number in Base X. WIth the built-in alphabet, its base 62 `num`: The number to encode `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ num = int(num) alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' if (num == 0): return alphabet[0] arr = [] base = len(alphabet) while num: rem = num % base num = num // base arr.append(alphabet[rem]) arr.reverse() return ''.join(arr) def base62_decode(string): """Decode a Base X encoded string into the number Arguments: - `string`: The encoded string - `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' base = len(alphabet) strlen = len(string) num = 0 idx = 0 for char in string: power = (strlen - (idx + 1)) num += alphabet.index(char) * (base ** power) idx += 1 return int(num) def augment(module_name, base_class): """Call the augment() method for all of the derived classes in the module """ for name, cls in inspect.getmembers(sys.modules[module_name], lambda x : inspect.isclass(x) and issubclass(x, base_class) ): if cls == base_class: continue cls.augment() def get_class(module, sl): for name, named_sl in names.items(): if named_sl == sl or sl == name: return getattr(module, name.capitalize()) raise NotASummaryName("No class for summary_level {}".format(sl)) def make_classes(base_class, module): """Create derived classes and put them into the same module as the base class. This function is called at the end of each of the derived class modules, acs, census, civik and tiger. It will create a set of new derived class in the module, one for each of the enries in the `summary_levels` dict. """ from functools import partial for k in names: cls = base_class.class_factory(k.capitalize()) cls.augment() setattr(module, k.capitalize(), cls) setattr(module, 'get_class', partial(get_class, module)) class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return '' county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s") def generate_all(sumlevel, d): """Generate a dict that includes all of the available geoid values, with keys for the most common names for those values. """ from geoid.civick import GVid from geoid.tiger import TigerGeoid from geoid.acs import AcsGeoid sumlevel = int(sumlevel) d = dict(d.items()) # Map common name variants if 'cousub' in d: d['cosub'] = d['cousub'] del d['cousub'] if 'blkgrp' in d: d['blockgroup'] = d['blkgrp'] del d['blkgrp'] if 'zcta5' in d: d['zcta'] = d['zcta5'] del d['zcta5'] gvid_class = GVid.resolve_summary_level(sumlevel) if not gvid_class: return {} geoidt_class = TigerGeoid.resolve_summary_level(sumlevel) geoid_class = AcsGeoid.resolve_summary_level(sumlevel) try: return dict( gvid=str(gvid_class(**d)), geoid=str(geoid_class(**d)), geoidt=str(geoidt_class(**d)) ) except: raise def _generate_names(): """ Code to generate the state and county names >>> python -c 'import geoid; geoid._generate_names()' """ from ambry import get_library l = get_library() counties = l.partition('census.gov-acs-geofile-2009-geofile50-20095-50') states = l.partition('census.gov-acs-geofile-2009-geofile40-20095-40') names = {} for row in counties.remote_datafile.reader: names[(row.state, row.county)] = row.name for row in states.remote_datafile.reader: if row.component == '00': names[(row.state, 0)] = row.name pprint.pprint(names)
Metatab/geoid
geoid/core.py
base62_decode
python
def base62_decode(string): alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' base = len(alphabet) strlen = len(string) num = 0 idx = 0 for char in string: power = (strlen - (idx + 1)) num += alphabet.index(char) * (base ** power) idx += 1 return int(num)
Decode a Base X encoded string into the number Arguments: - `string`: The encoded string - `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L384-L405
null
""" CLasses for working with census Geoids """ import inspect import re import sys import six names = { # (summary level value, base 10 chars, Base 62 chars, prefix fields) 'null': 1, 'us': 10, 'region': 20, 'division': 30, 'state': 40, 'county': 50, 'cosub': 60, 'place': 160, 'ua': 400, 'tract': 140, 'blockgroup': 150, 'block': 101, 'sdelm': 950, 'sdsec': 960, 'sduni': 970, 'zcta': 860, 'zip': 1200, 'sldl': 620, 'sldu': 610, 'cdcurr': 500, # Other Levels that don't have proper names yet. For these .allval() and # simplify() don't work properly. 'state_aianhh': 260, 'necta_nectadiv_state_county_cousub': 358, 'state_aianhh_place': 269, 'aianhh_state_county': 270, 'state_cbsa_metdiv': 323, 'state_aianhh280': 280, 'state_place_county': 155, 'aianhh_aitsce_state': 290, 'state_aianhh_aihhtli': 283, 'state_cdcurr_aianhh': 550, 'state_concit': 170, 'state_concit_place': 172, 'state_aianhh_aihhtli286': 286, 'cbsa': 310, 'cbsa_state': 311, 'cbsa_state_place': 312, 'cbsa_state_county': 313, 'cbsa_metdiv': 314, 'cbsa_metdiv_state': 315, 'state_cbsa': 320, 'state_cbsa_place': 321, 'state_cbsa_county': 322, 'state_county_cousub_submcd': 67, 'state_cbsa_metdiv_county': 324, 'state_county_cousub_place': 70, 'necta_state_county': 353, 'state_puma5': 795, 'csa': 330, 'csa_state': 331, 'csa_cbsa': 332, 'csa_cbsa_state': 333, 'cnecta': 335, 'state_county_cousub_place_tract': 80, 'cnecta_necta': 337, 'cnecta_necta_state': 338, 'state_csa': 340, 'state_csa_cbsa': 341, 'state_cnecta': 345, 'state_cnecta_necta': 346, 'necta': 350, 'necta_state': 351, 'necta_state_place': 352, 'cnecta_state': 336, 'necta_state_county_cousub': 354, 'necta_nectadiv': 355, 'necta_nectadiv_state': 356, 'state_anrc': 230, 'necta_nectadiv_state_county': 357, 'state_necta': 360, 'cbsa_metdiv_state_county': 316, 'state_necta_county': 362, 'state_necta_county_cousub': 363, 'state_necta_nectadiv': 364, 'state_necta_nectadiv_county': 365, 'state_necta_nectadiv_county_cousub': 366, 'ua_state': 410, 'ua_state_county': 430, 'state_sldu_county': 612, 'state_sldu': 610, 'state_sldl_county': 622, 'state_sldl': 620, 'state_cdcurr_county': 510, 'state_necta_place': 361, 'aianhh': 250, 'aianhh_aitsce': 251, 'aianhh_aihhtli': 252, 'state_sldl_county': 622, 'aianhh_aihhtli254': 254 } lengths = { 'null': 1, 'aianhh': 4, # American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 'aihhtli': '1', # American Indian Trust Land/ Hawaiian Home Land Indicator. A str b/c Census val is a str 'aitsce': 3, # American Indian Tribal Subdivision (Census) 'anrc': 5, # Alaska Native Regional Corporation (FIPS) 'blkgrp': 1, # Block Group 'blockgroup': 1, # Block Group 'block': 4, # Block 'cbsa': 5, # Metropolitan and Micropolitan Statistical Area 'cdcurr': 2, # Current Congressional District *** 'cnecta': 3, # New England City and Town Combined Statistical Area 'concit': 5, # Consolidated City 'county': 3, # County of current residence 'cousub': 5, # County Subdivision (FIPS) 'cosub': 5, # County Subdivision (FIPS) 'csa': 3, # Combined Statistical Area 'division': 1, # Census Division 'metdiv': 5, # Metropolitan Statistical Area- Metropolitan Division 'necta': 5, # New England City and Town Area 'nectadiv': 5, # New England City and Town Area Division 'place': 5, # Place (FIPS Code) 'puma5': 5, # Public Use Microdata Area 5% File 'region': 1, # Census Region 'sdelm': 5, # State-School District (Elementary) 'sdsec': 5, # State-School District (Secondary) 'sduni': 5, # State-School District (Unified) 'sldl': '3', # State Legislative District Lower. A String to signal that the census value is a string 'sldu': '3', # State Legislative District Upper. A String to signal that the census value is a string 'state': 2, # State (FIPS Code) 'submcd': 5, # Subminor Civil Division (FIPS) 'tract': 6, # Census Tract 'ua': 5, # Urban Area 'ur': 1, # Urban/Rural 'us': 0, 'zcta': 5, # Nonstandard 'zip': 5, } segments = { 1: ['null'], # United States 10: ['us'], # United States 20: ['region'], # Region 30: ['division'], # Division 40: ['state'], # State 50: ['state', 'county'], # County 60: ['state', 'county', 'cousub'], # County Subdivision 67: ['state', 'county', 'cousub', 'submcd'], # State (Puerto Rico Only)-County-County Subdivision-Subbarrio 70: ['state', 'county', 'cousub', 'place'], # County Subdivision-Place/Remainder 80: ['state', 'county', 'cousub', 'place', 'tract'], # County Subdivision-Place/Remainder-Census Tract 101: ['state', 'county', 'tract', 'block'], 140: ['state', 'county', 'tract'], # Census Tract 150: ['state', 'county', 'tract', 'blockgroup'], # Census Tract-Block Group 155: ['state', 'place', 'county'], # Place-County 160: ['state', 'place'], # Place 170: ['state', 'concit'], # Consolidated City 172: ['state', 'concit', 'place'], # Consolidated City-Place Within Consolidated City 230: ['state', 'anrc'], # State-Alaska Native Regional Corporation 250: ['aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land 251: ['aianhh', 'aitsce'], # American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder 252: ['aianhh', 'aihhtli'], # American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)4 254: ['aianhh', 'aihhtli'], # American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 260: ['state', 'aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State 269: ['state', 'aianhh', 'place'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder 270: ['aianhh', 'state', 'county'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County 280: ['state', 'aianhh'], # State-American Indian Area/Alaska Native Area/Hawaiian Home Land 283: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only) 286: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 290: ['aianhh', 'aitsce', 'state'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Tribal Subdivision/Remainder-State 310: ['cbsa'], # CBSA 311: ['cbsa', 'state'], # CBSA-State-County 312: ['cbsa', 'state', 'place'], # CBSA-State-Principal City 313: ['cbsa', 'state', 'county'], # CBSA-State-County 314: ['cbsa', 'metdiv'], # Metropolitan Statistical Area/Metropolitan Division 315: ['cbsa', 'metdiv', 'state'], # Metropolitan Statistical Area/Metropolitan Division-State 316: ['cbsa', 'metdiv', 'state', 'county'], # Metropolitan Statistical Area/Metropolitan Division-State-County 320: ['state', 'cbsa'], # State- CBSA 321: ['state', 'cbsa', 'place'], # State- CBSA -Principal City 322: ['state', 'cbsa', 'county'], # State- CBSA -County 323: ['state', 'cbsa', 'metdiv'], # State- Metropolitan Statistical Area/Metropolitan Division 324: ['state', 'cbsa', 'metdiv', 'county'], # State- Metropolitan Statistical Area/Metropolitan Division-County 330: ['csa'], # Combined Statistical Area 331: ['csa', 'state'], # Combined Statistical Area-State 332: ['csa', 'cbsa'], # Combined Statistical Area-CBSA 333: ['csa', 'cbsa', 'state'], # Combined Statistical Area-CBSA-State 335: ['cnecta'], # Combined New England City and Town Area 336: ['cnecta', 'state'], # Combined New England City and Town Area -State 337: ['cnecta', 'necta'], # Combined New England City and Town Area -New England City and Town Area 338: ['cnecta', 'necta', 'state'], # Combined New England City and Town Area -New England City and Town Area-State 340: ['state', 'csa'], # State-Combined Statistical Area 341: ['state', 'csa', 'cbsa'], # State-Combined Statistical Area-CBSA 345: ['state', 'cnecta'], # State-Combined New England City and Town Area 346: ['state', 'cnecta', 'necta'], # State-Combined New England City and Town Area-New England City and Town Area 350: ['necta'], # New England City and Town Area 351: ['necta', 'state'], # New England City and Town Area-State 352: ['necta', 'state', 'place'], # New England City and Town Area-State-Principal City 353: ['necta', 'state', 'county'], # New England City and Town Area-State-County 354: ['necta', 'state', 'county', 'cousub'], # New England City and Town Area-State-County-County Subdivision 355: ['necta', 'nectadiv'], # New England City and Town Area (NECTA)-NECTA Division 356: ['necta', 'nectadiv', 'state'], # New England City and Town Area (NECTA)-NECTA Division-State 357: ['necta', 'nectadiv', 'state', 'county'], # New England City and Town Area (NECTA)-NECTA Division-State-County 358: ['necta', 'nectadiv', 'state', 'county', 'cousub'], # New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision 360: ['state', 'necta'], # State-New England City and Town Area 361: ['state', 'necta', 'place'], # State-New England City and Town Area-Principal City 362: ['state', 'necta', 'county'], # State-New England City and Town Area-County 363: ['state', 'necta', 'county', 'cousub'], # State-New England City and Town Area-County-County Subdivision 364: ['state', 'necta', 'nectadiv'], # State-New England City and Town Area (NECTA)-NECTA Division 365: ['state', 'necta', 'nectadiv', 'county'], # State-New England City and Town Area (NECTA)-NECTA Division-County 366: ['state', 'necta', 'nectadiv', 'county', 'cousub'], # State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision 400: ['ua'], # Urban Area, 410: ['ua', 'state'], # Urban Area, State, 430: ['ua','state','county'], # Urban Area, State, County, 500: ['state', 'cdcurr'], # Congressional District 510: ['state', 'cdcurr', 'county'], # 550: ['state', 'cdcurr', 'aianhh'], # Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land 610: ['state', 'sldu'], # State Senate District 612: ['state', 'sldu', 'county'], # State Senate District-County 620: ['state', 'sldl'], # State House District 622: ['state', 'sldl', 'county'], # State House District-County 795: ['state', 'puma5'], # State-Public Use MicroSample Area 5% 860: ['zcta'], 950: ['state', 'sdelm'], # State-Elementary School District 960: ['state', 'sdsec'], # State-High School District 970: ['state', 'sduni'], # State-Unified School District # Nonstandard 1200: ['zip'] } descriptions = { 1: 'United States', 10: 'United States', 20: 'Region', 30: 'Division', 40: 'State', 50: 'County', 60: 'County Subdivision', 67: 'State (Puerto Rico Only)-County-County Subdivision-Subbarrio', 70: 'County Subdivision-Place/Remainder', 80: 'County Subdivision-Place/Remainder-Census Tract', 101: 'block', 140: 'Census Tract', 150: 'Census Tract-Block Group', 155: 'Place-County', 160: 'Place', 170: 'Consolidated City', 172: 'Consolidated City-Place Within Consolidated City', 230: 'State-Alaska Native Regional Corporation', 250: 'American Indian Area/Alaska Native Area/Hawaiian Home Land', 251: 'American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder', 252: 'American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)', 254: 'American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land', 260: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State', 269: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder', 270: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County', 280: 'State-American Indian Area/Alaska Native Area/Hawaiian Home Land', 283: 'aihhtli', 286: 'aihhtli', 290: 'state', 310: 'CBSA', 311: 'CBSA-State-County', 312: 'CBSA-State-Principal City', 313: 'CBSA-State-County', 314: 'Metropolitan Statistical Area/Metropolitan Division', 315: 'Metropolitan Statistical Area/Metropolitan Division-State', 316: 'Metropolitan Statistical Area/Metropolitan Division-State-County', 320: 'State- CBSA', 321: 'State- CBSA -Principal City', 322: 'State- CBSA -County', 323: 'State- Metropolitan Statistical Area/Metropolitan Division', 324: 'State- Metropolitan Statistical Area/Metropolitan Division-County', 330: 'Combined Statistical Area', 331: 'Combined Statistical Area-State', 332: 'Combined Statistical Area-CBSA', 333: 'Combined Statistical Area-CBSA-State', 335: 'Combined New England City and Town Area', 336: 'Combined New England City and Town Area -State', 337: 'Combined New England City and Town Area -New England City and Town Area', 338: 'Combined New England City and Town Area -New England City and Town Area-State', 340: 'State-Combined Statistical Area', 341: 'State-Combined Statistical Area-CBSA', 345: 'State-Combined New England City and Town Area', 346: 'State-Combined New England City and Town Area-New England City and Town Area', 350: 'New England City and Town Area', 351: 'New England City and Town Area-State', 352: 'New England City and Town Area-State-Principal City', 353: 'New England City and Town Area-State-County', 354: 'New England City and Town Area-State-County-County Subdivision', 355: 'New England City and Town Area (NECTA)-NECTA Division', 356: 'New England City and Town Area (NECTA)-NECTA Division-State', 357: 'New England City and Town Area (NECTA)-NECTA Division-State-County', 358: 'New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision', 360: 'State-New England City and Town Area', 361: 'State-New England City and Town Area-Principal City', 362: 'State-New England City and Town Area-County', 363: 'State-New England City and Town Area-County-County Subdivision', 364: 'State-New England City and Town Area (NECTA)-NECTA Division', 365: 'State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision', 400: 'Urban Area,', 410: 'Urban Area, State,', 430: 'Urban Area, State, County,', 500: 'Congressional District', 510: 'Congressional District, County', 550: 'Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land', 610: 'State Senate District', 612: 'State Senate District-County', 620: 'State House District', 622: 'State House District-County', 795: 'State-Public Use MicroSample Area 5%', 860: 'ZIP Code Tabulation Area', 950: 'State-Elementary School District', 960: 'State-High School District', 970: 'State-Unified School District', } plurals = { 'county': 'counties', 'place': 'places', 'Sdlu': 'State ' } class NotASummaryName(Exception): """An argument was not one of the valid summary names""" class ParseError(Exception): """Error parsing a geoid""" def parse_to_gvid(v): """Parse an ACS Geoid or a GVID to a GVID""" from geoid.civick import GVid from geoid.acs import AcsGeoid m1 = '' try: return GVid.parse(v) except ValueError as e: m1 = str(e) try: return AcsGeoid.parse(v).convert(GVid) except ValueError as e: raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e))) def base62_encode(num): """Encode a number in Base X. WIth the built-in alphabet, its base 62 `num`: The number to encode `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ num = int(num) alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' if (num == 0): return alphabet[0] arr = [] base = len(alphabet) while num: rem = num % base num = num // base arr.append(alphabet[rem]) arr.reverse() return ''.join(arr) def augment(module_name, base_class): """Call the augment() method for all of the derived classes in the module """ for name, cls in inspect.getmembers(sys.modules[module_name], lambda x : inspect.isclass(x) and issubclass(x, base_class) ): if cls == base_class: continue cls.augment() def get_class(module, sl): for name, named_sl in names.items(): if named_sl == sl or sl == name: return getattr(module, name.capitalize()) raise NotASummaryName("No class for summary_level {}".format(sl)) def make_classes(base_class, module): """Create derived classes and put them into the same module as the base class. This function is called at the end of each of the derived class modules, acs, census, civik and tiger. It will create a set of new derived class in the module, one for each of the enries in the `summary_levels` dict. """ from functools import partial for k in names: cls = base_class.class_factory(k.capitalize()) cls.augment() setattr(module, k.capitalize(), cls) setattr(module, 'get_class', partial(get_class, module)) class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return '' county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s") def generate_all(sumlevel, d): """Generate a dict that includes all of the available geoid values, with keys for the most common names for those values. """ from geoid.civick import GVid from geoid.tiger import TigerGeoid from geoid.acs import AcsGeoid sumlevel = int(sumlevel) d = dict(d.items()) # Map common name variants if 'cousub' in d: d['cosub'] = d['cousub'] del d['cousub'] if 'blkgrp' in d: d['blockgroup'] = d['blkgrp'] del d['blkgrp'] if 'zcta5' in d: d['zcta'] = d['zcta5'] del d['zcta5'] gvid_class = GVid.resolve_summary_level(sumlevel) if not gvid_class: return {} geoidt_class = TigerGeoid.resolve_summary_level(sumlevel) geoid_class = AcsGeoid.resolve_summary_level(sumlevel) try: return dict( gvid=str(gvid_class(**d)), geoid=str(geoid_class(**d)), geoidt=str(geoidt_class(**d)) ) except: raise def _generate_names(): """ Code to generate the state and county names >>> python -c 'import geoid; geoid._generate_names()' """ from ambry import get_library l = get_library() counties = l.partition('census.gov-acs-geofile-2009-geofile50-20095-50') states = l.partition('census.gov-acs-geofile-2009-geofile40-20095-40') names = {} for row in counties.remote_datafile.reader: names[(row.state, row.county)] = row.name for row in states.remote_datafile.reader: if row.component == '00': names[(row.state, 0)] = row.name pprint.pprint(names)
Metatab/geoid
geoid/core.py
augment
python
def augment(module_name, base_class): for name, cls in inspect.getmembers(sys.modules[module_name], lambda x : inspect.isclass(x) and issubclass(x, base_class) ): if cls == base_class: continue cls.augment()
Call the augment() method for all of the derived classes in the module
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L408-L416
null
""" CLasses for working with census Geoids """ import inspect import re import sys import six names = { # (summary level value, base 10 chars, Base 62 chars, prefix fields) 'null': 1, 'us': 10, 'region': 20, 'division': 30, 'state': 40, 'county': 50, 'cosub': 60, 'place': 160, 'ua': 400, 'tract': 140, 'blockgroup': 150, 'block': 101, 'sdelm': 950, 'sdsec': 960, 'sduni': 970, 'zcta': 860, 'zip': 1200, 'sldl': 620, 'sldu': 610, 'cdcurr': 500, # Other Levels that don't have proper names yet. For these .allval() and # simplify() don't work properly. 'state_aianhh': 260, 'necta_nectadiv_state_county_cousub': 358, 'state_aianhh_place': 269, 'aianhh_state_county': 270, 'state_cbsa_metdiv': 323, 'state_aianhh280': 280, 'state_place_county': 155, 'aianhh_aitsce_state': 290, 'state_aianhh_aihhtli': 283, 'state_cdcurr_aianhh': 550, 'state_concit': 170, 'state_concit_place': 172, 'state_aianhh_aihhtli286': 286, 'cbsa': 310, 'cbsa_state': 311, 'cbsa_state_place': 312, 'cbsa_state_county': 313, 'cbsa_metdiv': 314, 'cbsa_metdiv_state': 315, 'state_cbsa': 320, 'state_cbsa_place': 321, 'state_cbsa_county': 322, 'state_county_cousub_submcd': 67, 'state_cbsa_metdiv_county': 324, 'state_county_cousub_place': 70, 'necta_state_county': 353, 'state_puma5': 795, 'csa': 330, 'csa_state': 331, 'csa_cbsa': 332, 'csa_cbsa_state': 333, 'cnecta': 335, 'state_county_cousub_place_tract': 80, 'cnecta_necta': 337, 'cnecta_necta_state': 338, 'state_csa': 340, 'state_csa_cbsa': 341, 'state_cnecta': 345, 'state_cnecta_necta': 346, 'necta': 350, 'necta_state': 351, 'necta_state_place': 352, 'cnecta_state': 336, 'necta_state_county_cousub': 354, 'necta_nectadiv': 355, 'necta_nectadiv_state': 356, 'state_anrc': 230, 'necta_nectadiv_state_county': 357, 'state_necta': 360, 'cbsa_metdiv_state_county': 316, 'state_necta_county': 362, 'state_necta_county_cousub': 363, 'state_necta_nectadiv': 364, 'state_necta_nectadiv_county': 365, 'state_necta_nectadiv_county_cousub': 366, 'ua_state': 410, 'ua_state_county': 430, 'state_sldu_county': 612, 'state_sldu': 610, 'state_sldl_county': 622, 'state_sldl': 620, 'state_cdcurr_county': 510, 'state_necta_place': 361, 'aianhh': 250, 'aianhh_aitsce': 251, 'aianhh_aihhtli': 252, 'state_sldl_county': 622, 'aianhh_aihhtli254': 254 } lengths = { 'null': 1, 'aianhh': 4, # American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 'aihhtli': '1', # American Indian Trust Land/ Hawaiian Home Land Indicator. A str b/c Census val is a str 'aitsce': 3, # American Indian Tribal Subdivision (Census) 'anrc': 5, # Alaska Native Regional Corporation (FIPS) 'blkgrp': 1, # Block Group 'blockgroup': 1, # Block Group 'block': 4, # Block 'cbsa': 5, # Metropolitan and Micropolitan Statistical Area 'cdcurr': 2, # Current Congressional District *** 'cnecta': 3, # New England City and Town Combined Statistical Area 'concit': 5, # Consolidated City 'county': 3, # County of current residence 'cousub': 5, # County Subdivision (FIPS) 'cosub': 5, # County Subdivision (FIPS) 'csa': 3, # Combined Statistical Area 'division': 1, # Census Division 'metdiv': 5, # Metropolitan Statistical Area- Metropolitan Division 'necta': 5, # New England City and Town Area 'nectadiv': 5, # New England City and Town Area Division 'place': 5, # Place (FIPS Code) 'puma5': 5, # Public Use Microdata Area 5% File 'region': 1, # Census Region 'sdelm': 5, # State-School District (Elementary) 'sdsec': 5, # State-School District (Secondary) 'sduni': 5, # State-School District (Unified) 'sldl': '3', # State Legislative District Lower. A String to signal that the census value is a string 'sldu': '3', # State Legislative District Upper. A String to signal that the census value is a string 'state': 2, # State (FIPS Code) 'submcd': 5, # Subminor Civil Division (FIPS) 'tract': 6, # Census Tract 'ua': 5, # Urban Area 'ur': 1, # Urban/Rural 'us': 0, 'zcta': 5, # Nonstandard 'zip': 5, } segments = { 1: ['null'], # United States 10: ['us'], # United States 20: ['region'], # Region 30: ['division'], # Division 40: ['state'], # State 50: ['state', 'county'], # County 60: ['state', 'county', 'cousub'], # County Subdivision 67: ['state', 'county', 'cousub', 'submcd'], # State (Puerto Rico Only)-County-County Subdivision-Subbarrio 70: ['state', 'county', 'cousub', 'place'], # County Subdivision-Place/Remainder 80: ['state', 'county', 'cousub', 'place', 'tract'], # County Subdivision-Place/Remainder-Census Tract 101: ['state', 'county', 'tract', 'block'], 140: ['state', 'county', 'tract'], # Census Tract 150: ['state', 'county', 'tract', 'blockgroup'], # Census Tract-Block Group 155: ['state', 'place', 'county'], # Place-County 160: ['state', 'place'], # Place 170: ['state', 'concit'], # Consolidated City 172: ['state', 'concit', 'place'], # Consolidated City-Place Within Consolidated City 230: ['state', 'anrc'], # State-Alaska Native Regional Corporation 250: ['aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land 251: ['aianhh', 'aitsce'], # American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder 252: ['aianhh', 'aihhtli'], # American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)4 254: ['aianhh', 'aihhtli'], # American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 260: ['state', 'aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State 269: ['state', 'aianhh', 'place'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder 270: ['aianhh', 'state', 'county'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County 280: ['state', 'aianhh'], # State-American Indian Area/Alaska Native Area/Hawaiian Home Land 283: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only) 286: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 290: ['aianhh', 'aitsce', 'state'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Tribal Subdivision/Remainder-State 310: ['cbsa'], # CBSA 311: ['cbsa', 'state'], # CBSA-State-County 312: ['cbsa', 'state', 'place'], # CBSA-State-Principal City 313: ['cbsa', 'state', 'county'], # CBSA-State-County 314: ['cbsa', 'metdiv'], # Metropolitan Statistical Area/Metropolitan Division 315: ['cbsa', 'metdiv', 'state'], # Metropolitan Statistical Area/Metropolitan Division-State 316: ['cbsa', 'metdiv', 'state', 'county'], # Metropolitan Statistical Area/Metropolitan Division-State-County 320: ['state', 'cbsa'], # State- CBSA 321: ['state', 'cbsa', 'place'], # State- CBSA -Principal City 322: ['state', 'cbsa', 'county'], # State- CBSA -County 323: ['state', 'cbsa', 'metdiv'], # State- Metropolitan Statistical Area/Metropolitan Division 324: ['state', 'cbsa', 'metdiv', 'county'], # State- Metropolitan Statistical Area/Metropolitan Division-County 330: ['csa'], # Combined Statistical Area 331: ['csa', 'state'], # Combined Statistical Area-State 332: ['csa', 'cbsa'], # Combined Statistical Area-CBSA 333: ['csa', 'cbsa', 'state'], # Combined Statistical Area-CBSA-State 335: ['cnecta'], # Combined New England City and Town Area 336: ['cnecta', 'state'], # Combined New England City and Town Area -State 337: ['cnecta', 'necta'], # Combined New England City and Town Area -New England City and Town Area 338: ['cnecta', 'necta', 'state'], # Combined New England City and Town Area -New England City and Town Area-State 340: ['state', 'csa'], # State-Combined Statistical Area 341: ['state', 'csa', 'cbsa'], # State-Combined Statistical Area-CBSA 345: ['state', 'cnecta'], # State-Combined New England City and Town Area 346: ['state', 'cnecta', 'necta'], # State-Combined New England City and Town Area-New England City and Town Area 350: ['necta'], # New England City and Town Area 351: ['necta', 'state'], # New England City and Town Area-State 352: ['necta', 'state', 'place'], # New England City and Town Area-State-Principal City 353: ['necta', 'state', 'county'], # New England City and Town Area-State-County 354: ['necta', 'state', 'county', 'cousub'], # New England City and Town Area-State-County-County Subdivision 355: ['necta', 'nectadiv'], # New England City and Town Area (NECTA)-NECTA Division 356: ['necta', 'nectadiv', 'state'], # New England City and Town Area (NECTA)-NECTA Division-State 357: ['necta', 'nectadiv', 'state', 'county'], # New England City and Town Area (NECTA)-NECTA Division-State-County 358: ['necta', 'nectadiv', 'state', 'county', 'cousub'], # New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision 360: ['state', 'necta'], # State-New England City and Town Area 361: ['state', 'necta', 'place'], # State-New England City and Town Area-Principal City 362: ['state', 'necta', 'county'], # State-New England City and Town Area-County 363: ['state', 'necta', 'county', 'cousub'], # State-New England City and Town Area-County-County Subdivision 364: ['state', 'necta', 'nectadiv'], # State-New England City and Town Area (NECTA)-NECTA Division 365: ['state', 'necta', 'nectadiv', 'county'], # State-New England City and Town Area (NECTA)-NECTA Division-County 366: ['state', 'necta', 'nectadiv', 'county', 'cousub'], # State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision 400: ['ua'], # Urban Area, 410: ['ua', 'state'], # Urban Area, State, 430: ['ua','state','county'], # Urban Area, State, County, 500: ['state', 'cdcurr'], # Congressional District 510: ['state', 'cdcurr', 'county'], # 550: ['state', 'cdcurr', 'aianhh'], # Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land 610: ['state', 'sldu'], # State Senate District 612: ['state', 'sldu', 'county'], # State Senate District-County 620: ['state', 'sldl'], # State House District 622: ['state', 'sldl', 'county'], # State House District-County 795: ['state', 'puma5'], # State-Public Use MicroSample Area 5% 860: ['zcta'], 950: ['state', 'sdelm'], # State-Elementary School District 960: ['state', 'sdsec'], # State-High School District 970: ['state', 'sduni'], # State-Unified School District # Nonstandard 1200: ['zip'] } descriptions = { 1: 'United States', 10: 'United States', 20: 'Region', 30: 'Division', 40: 'State', 50: 'County', 60: 'County Subdivision', 67: 'State (Puerto Rico Only)-County-County Subdivision-Subbarrio', 70: 'County Subdivision-Place/Remainder', 80: 'County Subdivision-Place/Remainder-Census Tract', 101: 'block', 140: 'Census Tract', 150: 'Census Tract-Block Group', 155: 'Place-County', 160: 'Place', 170: 'Consolidated City', 172: 'Consolidated City-Place Within Consolidated City', 230: 'State-Alaska Native Regional Corporation', 250: 'American Indian Area/Alaska Native Area/Hawaiian Home Land', 251: 'American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder', 252: 'American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)', 254: 'American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land', 260: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State', 269: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder', 270: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County', 280: 'State-American Indian Area/Alaska Native Area/Hawaiian Home Land', 283: 'aihhtli', 286: 'aihhtli', 290: 'state', 310: 'CBSA', 311: 'CBSA-State-County', 312: 'CBSA-State-Principal City', 313: 'CBSA-State-County', 314: 'Metropolitan Statistical Area/Metropolitan Division', 315: 'Metropolitan Statistical Area/Metropolitan Division-State', 316: 'Metropolitan Statistical Area/Metropolitan Division-State-County', 320: 'State- CBSA', 321: 'State- CBSA -Principal City', 322: 'State- CBSA -County', 323: 'State- Metropolitan Statistical Area/Metropolitan Division', 324: 'State- Metropolitan Statistical Area/Metropolitan Division-County', 330: 'Combined Statistical Area', 331: 'Combined Statistical Area-State', 332: 'Combined Statistical Area-CBSA', 333: 'Combined Statistical Area-CBSA-State', 335: 'Combined New England City and Town Area', 336: 'Combined New England City and Town Area -State', 337: 'Combined New England City and Town Area -New England City and Town Area', 338: 'Combined New England City and Town Area -New England City and Town Area-State', 340: 'State-Combined Statistical Area', 341: 'State-Combined Statistical Area-CBSA', 345: 'State-Combined New England City and Town Area', 346: 'State-Combined New England City and Town Area-New England City and Town Area', 350: 'New England City and Town Area', 351: 'New England City and Town Area-State', 352: 'New England City and Town Area-State-Principal City', 353: 'New England City and Town Area-State-County', 354: 'New England City and Town Area-State-County-County Subdivision', 355: 'New England City and Town Area (NECTA)-NECTA Division', 356: 'New England City and Town Area (NECTA)-NECTA Division-State', 357: 'New England City and Town Area (NECTA)-NECTA Division-State-County', 358: 'New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision', 360: 'State-New England City and Town Area', 361: 'State-New England City and Town Area-Principal City', 362: 'State-New England City and Town Area-County', 363: 'State-New England City and Town Area-County-County Subdivision', 364: 'State-New England City and Town Area (NECTA)-NECTA Division', 365: 'State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision', 400: 'Urban Area,', 410: 'Urban Area, State,', 430: 'Urban Area, State, County,', 500: 'Congressional District', 510: 'Congressional District, County', 550: 'Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land', 610: 'State Senate District', 612: 'State Senate District-County', 620: 'State House District', 622: 'State House District-County', 795: 'State-Public Use MicroSample Area 5%', 860: 'ZIP Code Tabulation Area', 950: 'State-Elementary School District', 960: 'State-High School District', 970: 'State-Unified School District', } plurals = { 'county': 'counties', 'place': 'places', 'Sdlu': 'State ' } class NotASummaryName(Exception): """An argument was not one of the valid summary names""" class ParseError(Exception): """Error parsing a geoid""" def parse_to_gvid(v): """Parse an ACS Geoid or a GVID to a GVID""" from geoid.civick import GVid from geoid.acs import AcsGeoid m1 = '' try: return GVid.parse(v) except ValueError as e: m1 = str(e) try: return AcsGeoid.parse(v).convert(GVid) except ValueError as e: raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e))) def base62_encode(num): """Encode a number in Base X. WIth the built-in alphabet, its base 62 `num`: The number to encode `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ num = int(num) alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' if (num == 0): return alphabet[0] arr = [] base = len(alphabet) while num: rem = num % base num = num // base arr.append(alphabet[rem]) arr.reverse() return ''.join(arr) def base62_decode(string): """Decode a Base X encoded string into the number Arguments: - `string`: The encoded string - `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' base = len(alphabet) strlen = len(string) num = 0 idx = 0 for char in string: power = (strlen - (idx + 1)) num += alphabet.index(char) * (base ** power) idx += 1 return int(num) def get_class(module, sl): for name, named_sl in names.items(): if named_sl == sl or sl == name: return getattr(module, name.capitalize()) raise NotASummaryName("No class for summary_level {}".format(sl)) def make_classes(base_class, module): """Create derived classes and put them into the same module as the base class. This function is called at the end of each of the derived class modules, acs, census, civik and tiger. It will create a set of new derived class in the module, one for each of the enries in the `summary_levels` dict. """ from functools import partial for k in names: cls = base_class.class_factory(k.capitalize()) cls.augment() setattr(module, k.capitalize(), cls) setattr(module, 'get_class', partial(get_class, module)) class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return '' county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s") def generate_all(sumlevel, d): """Generate a dict that includes all of the available geoid values, with keys for the most common names for those values. """ from geoid.civick import GVid from geoid.tiger import TigerGeoid from geoid.acs import AcsGeoid sumlevel = int(sumlevel) d = dict(d.items()) # Map common name variants if 'cousub' in d: d['cosub'] = d['cousub'] del d['cousub'] if 'blkgrp' in d: d['blockgroup'] = d['blkgrp'] del d['blkgrp'] if 'zcta5' in d: d['zcta'] = d['zcta5'] del d['zcta5'] gvid_class = GVid.resolve_summary_level(sumlevel) if not gvid_class: return {} geoidt_class = TigerGeoid.resolve_summary_level(sumlevel) geoid_class = AcsGeoid.resolve_summary_level(sumlevel) try: return dict( gvid=str(gvid_class(**d)), geoid=str(geoid_class(**d)), geoidt=str(geoidt_class(**d)) ) except: raise def _generate_names(): """ Code to generate the state and county names >>> python -c 'import geoid; geoid._generate_names()' """ from ambry import get_library l = get_library() counties = l.partition('census.gov-acs-geofile-2009-geofile50-20095-50') states = l.partition('census.gov-acs-geofile-2009-geofile40-20095-40') names = {} for row in counties.remote_datafile.reader: names[(row.state, row.county)] = row.name for row in states.remote_datafile.reader: if row.component == '00': names[(row.state, 0)] = row.name pprint.pprint(names)
Metatab/geoid
geoid/core.py
make_classes
python
def make_classes(base_class, module): from functools import partial for k in names: cls = base_class.class_factory(k.capitalize()) cls.augment() setattr(module, k.capitalize(), cls) setattr(module, 'get_class', partial(get_class, module))
Create derived classes and put them into the same module as the base class. This function is called at the end of each of the derived class modules, acs, census, civik and tiger. It will create a set of new derived class in the module, one for each of the enries in the `summary_levels` dict.
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L428-L446
[ "def class_factory(cls, name):\n def __init__(self, *args, **kwargs):\n cls.__init__(self, *args, **kwargs)\n\n return type(name, (cls,), {\"__init__\": __init__})\n", "def class_factory(cls, name):\n def __init__(self, *args, **kwargs):\n cls.__init__(self, *args, **kwargs)\n\n return type(name, (cls,), {\"__init__\": __init__})\n", "def class_factory(cls, name):\n\n def __init__(self, *args, **kwargs):\n cls.__init__(self, *args, **kwargs)\n\n return type(name, (cls,), {\"__init__\": __init__})\n", "def class_factory(cls, name):\n\n def __init__(self, *args, **kwargs):\n cls.__init__(self, *args, **kwargs)\n\n return type(name, (cls,), {\"__init__\": __init__})\n" ]
""" CLasses for working with census Geoids """ import inspect import re import sys import six names = { # (summary level value, base 10 chars, Base 62 chars, prefix fields) 'null': 1, 'us': 10, 'region': 20, 'division': 30, 'state': 40, 'county': 50, 'cosub': 60, 'place': 160, 'ua': 400, 'tract': 140, 'blockgroup': 150, 'block': 101, 'sdelm': 950, 'sdsec': 960, 'sduni': 970, 'zcta': 860, 'zip': 1200, 'sldl': 620, 'sldu': 610, 'cdcurr': 500, # Other Levels that don't have proper names yet. For these .allval() and # simplify() don't work properly. 'state_aianhh': 260, 'necta_nectadiv_state_county_cousub': 358, 'state_aianhh_place': 269, 'aianhh_state_county': 270, 'state_cbsa_metdiv': 323, 'state_aianhh280': 280, 'state_place_county': 155, 'aianhh_aitsce_state': 290, 'state_aianhh_aihhtli': 283, 'state_cdcurr_aianhh': 550, 'state_concit': 170, 'state_concit_place': 172, 'state_aianhh_aihhtli286': 286, 'cbsa': 310, 'cbsa_state': 311, 'cbsa_state_place': 312, 'cbsa_state_county': 313, 'cbsa_metdiv': 314, 'cbsa_metdiv_state': 315, 'state_cbsa': 320, 'state_cbsa_place': 321, 'state_cbsa_county': 322, 'state_county_cousub_submcd': 67, 'state_cbsa_metdiv_county': 324, 'state_county_cousub_place': 70, 'necta_state_county': 353, 'state_puma5': 795, 'csa': 330, 'csa_state': 331, 'csa_cbsa': 332, 'csa_cbsa_state': 333, 'cnecta': 335, 'state_county_cousub_place_tract': 80, 'cnecta_necta': 337, 'cnecta_necta_state': 338, 'state_csa': 340, 'state_csa_cbsa': 341, 'state_cnecta': 345, 'state_cnecta_necta': 346, 'necta': 350, 'necta_state': 351, 'necta_state_place': 352, 'cnecta_state': 336, 'necta_state_county_cousub': 354, 'necta_nectadiv': 355, 'necta_nectadiv_state': 356, 'state_anrc': 230, 'necta_nectadiv_state_county': 357, 'state_necta': 360, 'cbsa_metdiv_state_county': 316, 'state_necta_county': 362, 'state_necta_county_cousub': 363, 'state_necta_nectadiv': 364, 'state_necta_nectadiv_county': 365, 'state_necta_nectadiv_county_cousub': 366, 'ua_state': 410, 'ua_state_county': 430, 'state_sldu_county': 612, 'state_sldu': 610, 'state_sldl_county': 622, 'state_sldl': 620, 'state_cdcurr_county': 510, 'state_necta_place': 361, 'aianhh': 250, 'aianhh_aitsce': 251, 'aianhh_aihhtli': 252, 'state_sldl_county': 622, 'aianhh_aihhtli254': 254 } lengths = { 'null': 1, 'aianhh': 4, # American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 'aihhtli': '1', # American Indian Trust Land/ Hawaiian Home Land Indicator. A str b/c Census val is a str 'aitsce': 3, # American Indian Tribal Subdivision (Census) 'anrc': 5, # Alaska Native Regional Corporation (FIPS) 'blkgrp': 1, # Block Group 'blockgroup': 1, # Block Group 'block': 4, # Block 'cbsa': 5, # Metropolitan and Micropolitan Statistical Area 'cdcurr': 2, # Current Congressional District *** 'cnecta': 3, # New England City and Town Combined Statistical Area 'concit': 5, # Consolidated City 'county': 3, # County of current residence 'cousub': 5, # County Subdivision (FIPS) 'cosub': 5, # County Subdivision (FIPS) 'csa': 3, # Combined Statistical Area 'division': 1, # Census Division 'metdiv': 5, # Metropolitan Statistical Area- Metropolitan Division 'necta': 5, # New England City and Town Area 'nectadiv': 5, # New England City and Town Area Division 'place': 5, # Place (FIPS Code) 'puma5': 5, # Public Use Microdata Area 5% File 'region': 1, # Census Region 'sdelm': 5, # State-School District (Elementary) 'sdsec': 5, # State-School District (Secondary) 'sduni': 5, # State-School District (Unified) 'sldl': '3', # State Legislative District Lower. A String to signal that the census value is a string 'sldu': '3', # State Legislative District Upper. A String to signal that the census value is a string 'state': 2, # State (FIPS Code) 'submcd': 5, # Subminor Civil Division (FIPS) 'tract': 6, # Census Tract 'ua': 5, # Urban Area 'ur': 1, # Urban/Rural 'us': 0, 'zcta': 5, # Nonstandard 'zip': 5, } segments = { 1: ['null'], # United States 10: ['us'], # United States 20: ['region'], # Region 30: ['division'], # Division 40: ['state'], # State 50: ['state', 'county'], # County 60: ['state', 'county', 'cousub'], # County Subdivision 67: ['state', 'county', 'cousub', 'submcd'], # State (Puerto Rico Only)-County-County Subdivision-Subbarrio 70: ['state', 'county', 'cousub', 'place'], # County Subdivision-Place/Remainder 80: ['state', 'county', 'cousub', 'place', 'tract'], # County Subdivision-Place/Remainder-Census Tract 101: ['state', 'county', 'tract', 'block'], 140: ['state', 'county', 'tract'], # Census Tract 150: ['state', 'county', 'tract', 'blockgroup'], # Census Tract-Block Group 155: ['state', 'place', 'county'], # Place-County 160: ['state', 'place'], # Place 170: ['state', 'concit'], # Consolidated City 172: ['state', 'concit', 'place'], # Consolidated City-Place Within Consolidated City 230: ['state', 'anrc'], # State-Alaska Native Regional Corporation 250: ['aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land 251: ['aianhh', 'aitsce'], # American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder 252: ['aianhh', 'aihhtli'], # American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)4 254: ['aianhh', 'aihhtli'], # American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 260: ['state', 'aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State 269: ['state', 'aianhh', 'place'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder 270: ['aianhh', 'state', 'county'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County 280: ['state', 'aianhh'], # State-American Indian Area/Alaska Native Area/Hawaiian Home Land 283: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only) 286: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 290: ['aianhh', 'aitsce', 'state'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Tribal Subdivision/Remainder-State 310: ['cbsa'], # CBSA 311: ['cbsa', 'state'], # CBSA-State-County 312: ['cbsa', 'state', 'place'], # CBSA-State-Principal City 313: ['cbsa', 'state', 'county'], # CBSA-State-County 314: ['cbsa', 'metdiv'], # Metropolitan Statistical Area/Metropolitan Division 315: ['cbsa', 'metdiv', 'state'], # Metropolitan Statistical Area/Metropolitan Division-State 316: ['cbsa', 'metdiv', 'state', 'county'], # Metropolitan Statistical Area/Metropolitan Division-State-County 320: ['state', 'cbsa'], # State- CBSA 321: ['state', 'cbsa', 'place'], # State- CBSA -Principal City 322: ['state', 'cbsa', 'county'], # State- CBSA -County 323: ['state', 'cbsa', 'metdiv'], # State- Metropolitan Statistical Area/Metropolitan Division 324: ['state', 'cbsa', 'metdiv', 'county'], # State- Metropolitan Statistical Area/Metropolitan Division-County 330: ['csa'], # Combined Statistical Area 331: ['csa', 'state'], # Combined Statistical Area-State 332: ['csa', 'cbsa'], # Combined Statistical Area-CBSA 333: ['csa', 'cbsa', 'state'], # Combined Statistical Area-CBSA-State 335: ['cnecta'], # Combined New England City and Town Area 336: ['cnecta', 'state'], # Combined New England City and Town Area -State 337: ['cnecta', 'necta'], # Combined New England City and Town Area -New England City and Town Area 338: ['cnecta', 'necta', 'state'], # Combined New England City and Town Area -New England City and Town Area-State 340: ['state', 'csa'], # State-Combined Statistical Area 341: ['state', 'csa', 'cbsa'], # State-Combined Statistical Area-CBSA 345: ['state', 'cnecta'], # State-Combined New England City and Town Area 346: ['state', 'cnecta', 'necta'], # State-Combined New England City and Town Area-New England City and Town Area 350: ['necta'], # New England City and Town Area 351: ['necta', 'state'], # New England City and Town Area-State 352: ['necta', 'state', 'place'], # New England City and Town Area-State-Principal City 353: ['necta', 'state', 'county'], # New England City and Town Area-State-County 354: ['necta', 'state', 'county', 'cousub'], # New England City and Town Area-State-County-County Subdivision 355: ['necta', 'nectadiv'], # New England City and Town Area (NECTA)-NECTA Division 356: ['necta', 'nectadiv', 'state'], # New England City and Town Area (NECTA)-NECTA Division-State 357: ['necta', 'nectadiv', 'state', 'county'], # New England City and Town Area (NECTA)-NECTA Division-State-County 358: ['necta', 'nectadiv', 'state', 'county', 'cousub'], # New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision 360: ['state', 'necta'], # State-New England City and Town Area 361: ['state', 'necta', 'place'], # State-New England City and Town Area-Principal City 362: ['state', 'necta', 'county'], # State-New England City and Town Area-County 363: ['state', 'necta', 'county', 'cousub'], # State-New England City and Town Area-County-County Subdivision 364: ['state', 'necta', 'nectadiv'], # State-New England City and Town Area (NECTA)-NECTA Division 365: ['state', 'necta', 'nectadiv', 'county'], # State-New England City and Town Area (NECTA)-NECTA Division-County 366: ['state', 'necta', 'nectadiv', 'county', 'cousub'], # State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision 400: ['ua'], # Urban Area, 410: ['ua', 'state'], # Urban Area, State, 430: ['ua','state','county'], # Urban Area, State, County, 500: ['state', 'cdcurr'], # Congressional District 510: ['state', 'cdcurr', 'county'], # 550: ['state', 'cdcurr', 'aianhh'], # Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land 610: ['state', 'sldu'], # State Senate District 612: ['state', 'sldu', 'county'], # State Senate District-County 620: ['state', 'sldl'], # State House District 622: ['state', 'sldl', 'county'], # State House District-County 795: ['state', 'puma5'], # State-Public Use MicroSample Area 5% 860: ['zcta'], 950: ['state', 'sdelm'], # State-Elementary School District 960: ['state', 'sdsec'], # State-High School District 970: ['state', 'sduni'], # State-Unified School District # Nonstandard 1200: ['zip'] } descriptions = { 1: 'United States', 10: 'United States', 20: 'Region', 30: 'Division', 40: 'State', 50: 'County', 60: 'County Subdivision', 67: 'State (Puerto Rico Only)-County-County Subdivision-Subbarrio', 70: 'County Subdivision-Place/Remainder', 80: 'County Subdivision-Place/Remainder-Census Tract', 101: 'block', 140: 'Census Tract', 150: 'Census Tract-Block Group', 155: 'Place-County', 160: 'Place', 170: 'Consolidated City', 172: 'Consolidated City-Place Within Consolidated City', 230: 'State-Alaska Native Regional Corporation', 250: 'American Indian Area/Alaska Native Area/Hawaiian Home Land', 251: 'American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder', 252: 'American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)', 254: 'American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land', 260: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State', 269: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder', 270: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County', 280: 'State-American Indian Area/Alaska Native Area/Hawaiian Home Land', 283: 'aihhtli', 286: 'aihhtli', 290: 'state', 310: 'CBSA', 311: 'CBSA-State-County', 312: 'CBSA-State-Principal City', 313: 'CBSA-State-County', 314: 'Metropolitan Statistical Area/Metropolitan Division', 315: 'Metropolitan Statistical Area/Metropolitan Division-State', 316: 'Metropolitan Statistical Area/Metropolitan Division-State-County', 320: 'State- CBSA', 321: 'State- CBSA -Principal City', 322: 'State- CBSA -County', 323: 'State- Metropolitan Statistical Area/Metropolitan Division', 324: 'State- Metropolitan Statistical Area/Metropolitan Division-County', 330: 'Combined Statistical Area', 331: 'Combined Statistical Area-State', 332: 'Combined Statistical Area-CBSA', 333: 'Combined Statistical Area-CBSA-State', 335: 'Combined New England City and Town Area', 336: 'Combined New England City and Town Area -State', 337: 'Combined New England City and Town Area -New England City and Town Area', 338: 'Combined New England City and Town Area -New England City and Town Area-State', 340: 'State-Combined Statistical Area', 341: 'State-Combined Statistical Area-CBSA', 345: 'State-Combined New England City and Town Area', 346: 'State-Combined New England City and Town Area-New England City and Town Area', 350: 'New England City and Town Area', 351: 'New England City and Town Area-State', 352: 'New England City and Town Area-State-Principal City', 353: 'New England City and Town Area-State-County', 354: 'New England City and Town Area-State-County-County Subdivision', 355: 'New England City and Town Area (NECTA)-NECTA Division', 356: 'New England City and Town Area (NECTA)-NECTA Division-State', 357: 'New England City and Town Area (NECTA)-NECTA Division-State-County', 358: 'New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision', 360: 'State-New England City and Town Area', 361: 'State-New England City and Town Area-Principal City', 362: 'State-New England City and Town Area-County', 363: 'State-New England City and Town Area-County-County Subdivision', 364: 'State-New England City and Town Area (NECTA)-NECTA Division', 365: 'State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision', 400: 'Urban Area,', 410: 'Urban Area, State,', 430: 'Urban Area, State, County,', 500: 'Congressional District', 510: 'Congressional District, County', 550: 'Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land', 610: 'State Senate District', 612: 'State Senate District-County', 620: 'State House District', 622: 'State House District-County', 795: 'State-Public Use MicroSample Area 5%', 860: 'ZIP Code Tabulation Area', 950: 'State-Elementary School District', 960: 'State-High School District', 970: 'State-Unified School District', } plurals = { 'county': 'counties', 'place': 'places', 'Sdlu': 'State ' } class NotASummaryName(Exception): """An argument was not one of the valid summary names""" class ParseError(Exception): """Error parsing a geoid""" def parse_to_gvid(v): """Parse an ACS Geoid or a GVID to a GVID""" from geoid.civick import GVid from geoid.acs import AcsGeoid m1 = '' try: return GVid.parse(v) except ValueError as e: m1 = str(e) try: return AcsGeoid.parse(v).convert(GVid) except ValueError as e: raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e))) def base62_encode(num): """Encode a number in Base X. WIth the built-in alphabet, its base 62 `num`: The number to encode `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ num = int(num) alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' if (num == 0): return alphabet[0] arr = [] base = len(alphabet) while num: rem = num % base num = num // base arr.append(alphabet[rem]) arr.reverse() return ''.join(arr) def base62_decode(string): """Decode a Base X encoded string into the number Arguments: - `string`: The encoded string - `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' base = len(alphabet) strlen = len(string) num = 0 idx = 0 for char in string: power = (strlen - (idx + 1)) num += alphabet.index(char) * (base ** power) idx += 1 return int(num) def augment(module_name, base_class): """Call the augment() method for all of the derived classes in the module """ for name, cls in inspect.getmembers(sys.modules[module_name], lambda x : inspect.isclass(x) and issubclass(x, base_class) ): if cls == base_class: continue cls.augment() def get_class(module, sl): for name, named_sl in names.items(): if named_sl == sl or sl == name: return getattr(module, name.capitalize()) raise NotASummaryName("No class for summary_level {}".format(sl)) class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return '' county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s") def generate_all(sumlevel, d): """Generate a dict that includes all of the available geoid values, with keys for the most common names for those values. """ from geoid.civick import GVid from geoid.tiger import TigerGeoid from geoid.acs import AcsGeoid sumlevel = int(sumlevel) d = dict(d.items()) # Map common name variants if 'cousub' in d: d['cosub'] = d['cousub'] del d['cousub'] if 'blkgrp' in d: d['blockgroup'] = d['blkgrp'] del d['blkgrp'] if 'zcta5' in d: d['zcta'] = d['zcta5'] del d['zcta5'] gvid_class = GVid.resolve_summary_level(sumlevel) if not gvid_class: return {} geoidt_class = TigerGeoid.resolve_summary_level(sumlevel) geoid_class = AcsGeoid.resolve_summary_level(sumlevel) try: return dict( gvid=str(gvid_class(**d)), geoid=str(geoid_class(**d)), geoidt=str(geoidt_class(**d)) ) except: raise def _generate_names(): """ Code to generate the state and county names >>> python -c 'import geoid; geoid._generate_names()' """ from ambry import get_library l = get_library() counties = l.partition('census.gov-acs-geofile-2009-geofile50-20095-50') states = l.partition('census.gov-acs-geofile-2009-geofile40-20095-40') names = {} for row in counties.remote_datafile.reader: names[(row.state, row.county)] = row.name for row in states.remote_datafile.reader: if row.component == '00': names[(row.state, 0)] = row.name pprint.pprint(names)
Metatab/geoid
geoid/core.py
generate_all
python
def generate_all(sumlevel, d): from geoid.civick import GVid from geoid.tiger import TigerGeoid from geoid.acs import AcsGeoid sumlevel = int(sumlevel) d = dict(d.items()) # Map common name variants if 'cousub' in d: d['cosub'] = d['cousub'] del d['cousub'] if 'blkgrp' in d: d['blockgroup'] = d['blkgrp'] del d['blkgrp'] if 'zcta5' in d: d['zcta'] = d['zcta5'] del d['zcta5'] gvid_class = GVid.resolve_summary_level(sumlevel) if not gvid_class: return {} geoidt_class = TigerGeoid.resolve_summary_level(sumlevel) geoid_class = AcsGeoid.resolve_summary_level(sumlevel) try: return dict( gvid=str(gvid_class(**d)), geoid=str(geoid_class(**d)), geoidt=str(geoidt_class(**d)) ) except: raise
Generate a dict that includes all of the available geoid values, with keys for the most common names for those values.
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L887-L927
[ "def resolve_summary_level(cls, sl):\n try:\n return cls.sl_map[sl]\n except KeyError:\n return None\n" ]
""" CLasses for working with census Geoids """ import inspect import re import sys import six names = { # (summary level value, base 10 chars, Base 62 chars, prefix fields) 'null': 1, 'us': 10, 'region': 20, 'division': 30, 'state': 40, 'county': 50, 'cosub': 60, 'place': 160, 'ua': 400, 'tract': 140, 'blockgroup': 150, 'block': 101, 'sdelm': 950, 'sdsec': 960, 'sduni': 970, 'zcta': 860, 'zip': 1200, 'sldl': 620, 'sldu': 610, 'cdcurr': 500, # Other Levels that don't have proper names yet. For these .allval() and # simplify() don't work properly. 'state_aianhh': 260, 'necta_nectadiv_state_county_cousub': 358, 'state_aianhh_place': 269, 'aianhh_state_county': 270, 'state_cbsa_metdiv': 323, 'state_aianhh280': 280, 'state_place_county': 155, 'aianhh_aitsce_state': 290, 'state_aianhh_aihhtli': 283, 'state_cdcurr_aianhh': 550, 'state_concit': 170, 'state_concit_place': 172, 'state_aianhh_aihhtli286': 286, 'cbsa': 310, 'cbsa_state': 311, 'cbsa_state_place': 312, 'cbsa_state_county': 313, 'cbsa_metdiv': 314, 'cbsa_metdiv_state': 315, 'state_cbsa': 320, 'state_cbsa_place': 321, 'state_cbsa_county': 322, 'state_county_cousub_submcd': 67, 'state_cbsa_metdiv_county': 324, 'state_county_cousub_place': 70, 'necta_state_county': 353, 'state_puma5': 795, 'csa': 330, 'csa_state': 331, 'csa_cbsa': 332, 'csa_cbsa_state': 333, 'cnecta': 335, 'state_county_cousub_place_tract': 80, 'cnecta_necta': 337, 'cnecta_necta_state': 338, 'state_csa': 340, 'state_csa_cbsa': 341, 'state_cnecta': 345, 'state_cnecta_necta': 346, 'necta': 350, 'necta_state': 351, 'necta_state_place': 352, 'cnecta_state': 336, 'necta_state_county_cousub': 354, 'necta_nectadiv': 355, 'necta_nectadiv_state': 356, 'state_anrc': 230, 'necta_nectadiv_state_county': 357, 'state_necta': 360, 'cbsa_metdiv_state_county': 316, 'state_necta_county': 362, 'state_necta_county_cousub': 363, 'state_necta_nectadiv': 364, 'state_necta_nectadiv_county': 365, 'state_necta_nectadiv_county_cousub': 366, 'ua_state': 410, 'ua_state_county': 430, 'state_sldu_county': 612, 'state_sldu': 610, 'state_sldl_county': 622, 'state_sldl': 620, 'state_cdcurr_county': 510, 'state_necta_place': 361, 'aianhh': 250, 'aianhh_aitsce': 251, 'aianhh_aihhtli': 252, 'state_sldl_county': 622, 'aianhh_aihhtli254': 254 } lengths = { 'null': 1, 'aianhh': 4, # American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 'aihhtli': '1', # American Indian Trust Land/ Hawaiian Home Land Indicator. A str b/c Census val is a str 'aitsce': 3, # American Indian Tribal Subdivision (Census) 'anrc': 5, # Alaska Native Regional Corporation (FIPS) 'blkgrp': 1, # Block Group 'blockgroup': 1, # Block Group 'block': 4, # Block 'cbsa': 5, # Metropolitan and Micropolitan Statistical Area 'cdcurr': 2, # Current Congressional District *** 'cnecta': 3, # New England City and Town Combined Statistical Area 'concit': 5, # Consolidated City 'county': 3, # County of current residence 'cousub': 5, # County Subdivision (FIPS) 'cosub': 5, # County Subdivision (FIPS) 'csa': 3, # Combined Statistical Area 'division': 1, # Census Division 'metdiv': 5, # Metropolitan Statistical Area- Metropolitan Division 'necta': 5, # New England City and Town Area 'nectadiv': 5, # New England City and Town Area Division 'place': 5, # Place (FIPS Code) 'puma5': 5, # Public Use Microdata Area 5% File 'region': 1, # Census Region 'sdelm': 5, # State-School District (Elementary) 'sdsec': 5, # State-School District (Secondary) 'sduni': 5, # State-School District (Unified) 'sldl': '3', # State Legislative District Lower. A String to signal that the census value is a string 'sldu': '3', # State Legislative District Upper. A String to signal that the census value is a string 'state': 2, # State (FIPS Code) 'submcd': 5, # Subminor Civil Division (FIPS) 'tract': 6, # Census Tract 'ua': 5, # Urban Area 'ur': 1, # Urban/Rural 'us': 0, 'zcta': 5, # Nonstandard 'zip': 5, } segments = { 1: ['null'], # United States 10: ['us'], # United States 20: ['region'], # Region 30: ['division'], # Division 40: ['state'], # State 50: ['state', 'county'], # County 60: ['state', 'county', 'cousub'], # County Subdivision 67: ['state', 'county', 'cousub', 'submcd'], # State (Puerto Rico Only)-County-County Subdivision-Subbarrio 70: ['state', 'county', 'cousub', 'place'], # County Subdivision-Place/Remainder 80: ['state', 'county', 'cousub', 'place', 'tract'], # County Subdivision-Place/Remainder-Census Tract 101: ['state', 'county', 'tract', 'block'], 140: ['state', 'county', 'tract'], # Census Tract 150: ['state', 'county', 'tract', 'blockgroup'], # Census Tract-Block Group 155: ['state', 'place', 'county'], # Place-County 160: ['state', 'place'], # Place 170: ['state', 'concit'], # Consolidated City 172: ['state', 'concit', 'place'], # Consolidated City-Place Within Consolidated City 230: ['state', 'anrc'], # State-Alaska Native Regional Corporation 250: ['aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land 251: ['aianhh', 'aitsce'], # American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder 252: ['aianhh', 'aihhtli'], # American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)4 254: ['aianhh', 'aihhtli'], # American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 260: ['state', 'aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State 269: ['state', 'aianhh', 'place'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder 270: ['aianhh', 'state', 'county'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County 280: ['state', 'aianhh'], # State-American Indian Area/Alaska Native Area/Hawaiian Home Land 283: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only) 286: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 290: ['aianhh', 'aitsce', 'state'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Tribal Subdivision/Remainder-State 310: ['cbsa'], # CBSA 311: ['cbsa', 'state'], # CBSA-State-County 312: ['cbsa', 'state', 'place'], # CBSA-State-Principal City 313: ['cbsa', 'state', 'county'], # CBSA-State-County 314: ['cbsa', 'metdiv'], # Metropolitan Statistical Area/Metropolitan Division 315: ['cbsa', 'metdiv', 'state'], # Metropolitan Statistical Area/Metropolitan Division-State 316: ['cbsa', 'metdiv', 'state', 'county'], # Metropolitan Statistical Area/Metropolitan Division-State-County 320: ['state', 'cbsa'], # State- CBSA 321: ['state', 'cbsa', 'place'], # State- CBSA -Principal City 322: ['state', 'cbsa', 'county'], # State- CBSA -County 323: ['state', 'cbsa', 'metdiv'], # State- Metropolitan Statistical Area/Metropolitan Division 324: ['state', 'cbsa', 'metdiv', 'county'], # State- Metropolitan Statistical Area/Metropolitan Division-County 330: ['csa'], # Combined Statistical Area 331: ['csa', 'state'], # Combined Statistical Area-State 332: ['csa', 'cbsa'], # Combined Statistical Area-CBSA 333: ['csa', 'cbsa', 'state'], # Combined Statistical Area-CBSA-State 335: ['cnecta'], # Combined New England City and Town Area 336: ['cnecta', 'state'], # Combined New England City and Town Area -State 337: ['cnecta', 'necta'], # Combined New England City and Town Area -New England City and Town Area 338: ['cnecta', 'necta', 'state'], # Combined New England City and Town Area -New England City and Town Area-State 340: ['state', 'csa'], # State-Combined Statistical Area 341: ['state', 'csa', 'cbsa'], # State-Combined Statistical Area-CBSA 345: ['state', 'cnecta'], # State-Combined New England City and Town Area 346: ['state', 'cnecta', 'necta'], # State-Combined New England City and Town Area-New England City and Town Area 350: ['necta'], # New England City and Town Area 351: ['necta', 'state'], # New England City and Town Area-State 352: ['necta', 'state', 'place'], # New England City and Town Area-State-Principal City 353: ['necta', 'state', 'county'], # New England City and Town Area-State-County 354: ['necta', 'state', 'county', 'cousub'], # New England City and Town Area-State-County-County Subdivision 355: ['necta', 'nectadiv'], # New England City and Town Area (NECTA)-NECTA Division 356: ['necta', 'nectadiv', 'state'], # New England City and Town Area (NECTA)-NECTA Division-State 357: ['necta', 'nectadiv', 'state', 'county'], # New England City and Town Area (NECTA)-NECTA Division-State-County 358: ['necta', 'nectadiv', 'state', 'county', 'cousub'], # New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision 360: ['state', 'necta'], # State-New England City and Town Area 361: ['state', 'necta', 'place'], # State-New England City and Town Area-Principal City 362: ['state', 'necta', 'county'], # State-New England City and Town Area-County 363: ['state', 'necta', 'county', 'cousub'], # State-New England City and Town Area-County-County Subdivision 364: ['state', 'necta', 'nectadiv'], # State-New England City and Town Area (NECTA)-NECTA Division 365: ['state', 'necta', 'nectadiv', 'county'], # State-New England City and Town Area (NECTA)-NECTA Division-County 366: ['state', 'necta', 'nectadiv', 'county', 'cousub'], # State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision 400: ['ua'], # Urban Area, 410: ['ua', 'state'], # Urban Area, State, 430: ['ua','state','county'], # Urban Area, State, County, 500: ['state', 'cdcurr'], # Congressional District 510: ['state', 'cdcurr', 'county'], # 550: ['state', 'cdcurr', 'aianhh'], # Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land 610: ['state', 'sldu'], # State Senate District 612: ['state', 'sldu', 'county'], # State Senate District-County 620: ['state', 'sldl'], # State House District 622: ['state', 'sldl', 'county'], # State House District-County 795: ['state', 'puma5'], # State-Public Use MicroSample Area 5% 860: ['zcta'], 950: ['state', 'sdelm'], # State-Elementary School District 960: ['state', 'sdsec'], # State-High School District 970: ['state', 'sduni'], # State-Unified School District # Nonstandard 1200: ['zip'] } descriptions = { 1: 'United States', 10: 'United States', 20: 'Region', 30: 'Division', 40: 'State', 50: 'County', 60: 'County Subdivision', 67: 'State (Puerto Rico Only)-County-County Subdivision-Subbarrio', 70: 'County Subdivision-Place/Remainder', 80: 'County Subdivision-Place/Remainder-Census Tract', 101: 'block', 140: 'Census Tract', 150: 'Census Tract-Block Group', 155: 'Place-County', 160: 'Place', 170: 'Consolidated City', 172: 'Consolidated City-Place Within Consolidated City', 230: 'State-Alaska Native Regional Corporation', 250: 'American Indian Area/Alaska Native Area/Hawaiian Home Land', 251: 'American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder', 252: 'American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)', 254: 'American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land', 260: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State', 269: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder', 270: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County', 280: 'State-American Indian Area/Alaska Native Area/Hawaiian Home Land', 283: 'aihhtli', 286: 'aihhtli', 290: 'state', 310: 'CBSA', 311: 'CBSA-State-County', 312: 'CBSA-State-Principal City', 313: 'CBSA-State-County', 314: 'Metropolitan Statistical Area/Metropolitan Division', 315: 'Metropolitan Statistical Area/Metropolitan Division-State', 316: 'Metropolitan Statistical Area/Metropolitan Division-State-County', 320: 'State- CBSA', 321: 'State- CBSA -Principal City', 322: 'State- CBSA -County', 323: 'State- Metropolitan Statistical Area/Metropolitan Division', 324: 'State- Metropolitan Statistical Area/Metropolitan Division-County', 330: 'Combined Statistical Area', 331: 'Combined Statistical Area-State', 332: 'Combined Statistical Area-CBSA', 333: 'Combined Statistical Area-CBSA-State', 335: 'Combined New England City and Town Area', 336: 'Combined New England City and Town Area -State', 337: 'Combined New England City and Town Area -New England City and Town Area', 338: 'Combined New England City and Town Area -New England City and Town Area-State', 340: 'State-Combined Statistical Area', 341: 'State-Combined Statistical Area-CBSA', 345: 'State-Combined New England City and Town Area', 346: 'State-Combined New England City and Town Area-New England City and Town Area', 350: 'New England City and Town Area', 351: 'New England City and Town Area-State', 352: 'New England City and Town Area-State-Principal City', 353: 'New England City and Town Area-State-County', 354: 'New England City and Town Area-State-County-County Subdivision', 355: 'New England City and Town Area (NECTA)-NECTA Division', 356: 'New England City and Town Area (NECTA)-NECTA Division-State', 357: 'New England City and Town Area (NECTA)-NECTA Division-State-County', 358: 'New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision', 360: 'State-New England City and Town Area', 361: 'State-New England City and Town Area-Principal City', 362: 'State-New England City and Town Area-County', 363: 'State-New England City and Town Area-County-County Subdivision', 364: 'State-New England City and Town Area (NECTA)-NECTA Division', 365: 'State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision', 400: 'Urban Area,', 410: 'Urban Area, State,', 430: 'Urban Area, State, County,', 500: 'Congressional District', 510: 'Congressional District, County', 550: 'Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land', 610: 'State Senate District', 612: 'State Senate District-County', 620: 'State House District', 622: 'State House District-County', 795: 'State-Public Use MicroSample Area 5%', 860: 'ZIP Code Tabulation Area', 950: 'State-Elementary School District', 960: 'State-High School District', 970: 'State-Unified School District', } plurals = { 'county': 'counties', 'place': 'places', 'Sdlu': 'State ' } class NotASummaryName(Exception): """An argument was not one of the valid summary names""" class ParseError(Exception): """Error parsing a geoid""" def parse_to_gvid(v): """Parse an ACS Geoid or a GVID to a GVID""" from geoid.civick import GVid from geoid.acs import AcsGeoid m1 = '' try: return GVid.parse(v) except ValueError as e: m1 = str(e) try: return AcsGeoid.parse(v).convert(GVid) except ValueError as e: raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e))) def base62_encode(num): """Encode a number in Base X. WIth the built-in alphabet, its base 62 `num`: The number to encode `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ num = int(num) alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' if (num == 0): return alphabet[0] arr = [] base = len(alphabet) while num: rem = num % base num = num // base arr.append(alphabet[rem]) arr.reverse() return ''.join(arr) def base62_decode(string): """Decode a Base X encoded string into the number Arguments: - `string`: The encoded string - `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' base = len(alphabet) strlen = len(string) num = 0 idx = 0 for char in string: power = (strlen - (idx + 1)) num += alphabet.index(char) * (base ** power) idx += 1 return int(num) def augment(module_name, base_class): """Call the augment() method for all of the derived classes in the module """ for name, cls in inspect.getmembers(sys.modules[module_name], lambda x : inspect.isclass(x) and issubclass(x, base_class) ): if cls == base_class: continue cls.augment() def get_class(module, sl): for name, named_sl in names.items(): if named_sl == sl or sl == name: return getattr(module, name.capitalize()) raise NotASummaryName("No class for summary_level {}".format(sl)) def make_classes(base_class, module): """Create derived classes and put them into the same module as the base class. This function is called at the end of each of the derived class modules, acs, census, civik and tiger. It will create a set of new derived class in the module, one for each of the enries in the `summary_levels` dict. """ from functools import partial for k in names: cls = base_class.class_factory(k.capitalize()) cls.augment() setattr(module, k.capitalize(), cls) setattr(module, 'get_class', partial(get_class, module)) class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return '' county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s") def _generate_names(): """ Code to generate the state and county names >>> python -c 'import geoid; geoid._generate_names()' """ from ambry import get_library l = get_library() counties = l.partition('census.gov-acs-geofile-2009-geofile50-20095-50') states = l.partition('census.gov-acs-geofile-2009-geofile40-20095-40') names = {} for row in counties.remote_datafile.reader: names[(row.state, row.county)] = row.name for row in states.remote_datafile.reader: if row.component == '00': names[(row.state, 0)] = row.name pprint.pprint(names)
Metatab/geoid
geoid/core.py
_generate_names
python
def _generate_names(): from ambry import get_library l = get_library() counties = l.partition('census.gov-acs-geofile-2009-geofile50-20095-50') states = l.partition('census.gov-acs-geofile-2009-geofile40-20095-40') names = {} for row in counties.remote_datafile.reader: names[(row.state, row.county)] = row.name for row in states.remote_datafile.reader: if row.component == '00': names[(row.state, 0)] = row.name pprint.pprint(names)
Code to generate the state and county names >>> python -c 'import geoid; geoid._generate_names()'
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L930-L952
null
""" CLasses for working with census Geoids """ import inspect import re import sys import six names = { # (summary level value, base 10 chars, Base 62 chars, prefix fields) 'null': 1, 'us': 10, 'region': 20, 'division': 30, 'state': 40, 'county': 50, 'cosub': 60, 'place': 160, 'ua': 400, 'tract': 140, 'blockgroup': 150, 'block': 101, 'sdelm': 950, 'sdsec': 960, 'sduni': 970, 'zcta': 860, 'zip': 1200, 'sldl': 620, 'sldu': 610, 'cdcurr': 500, # Other Levels that don't have proper names yet. For these .allval() and # simplify() don't work properly. 'state_aianhh': 260, 'necta_nectadiv_state_county_cousub': 358, 'state_aianhh_place': 269, 'aianhh_state_county': 270, 'state_cbsa_metdiv': 323, 'state_aianhh280': 280, 'state_place_county': 155, 'aianhh_aitsce_state': 290, 'state_aianhh_aihhtli': 283, 'state_cdcurr_aianhh': 550, 'state_concit': 170, 'state_concit_place': 172, 'state_aianhh_aihhtli286': 286, 'cbsa': 310, 'cbsa_state': 311, 'cbsa_state_place': 312, 'cbsa_state_county': 313, 'cbsa_metdiv': 314, 'cbsa_metdiv_state': 315, 'state_cbsa': 320, 'state_cbsa_place': 321, 'state_cbsa_county': 322, 'state_county_cousub_submcd': 67, 'state_cbsa_metdiv_county': 324, 'state_county_cousub_place': 70, 'necta_state_county': 353, 'state_puma5': 795, 'csa': 330, 'csa_state': 331, 'csa_cbsa': 332, 'csa_cbsa_state': 333, 'cnecta': 335, 'state_county_cousub_place_tract': 80, 'cnecta_necta': 337, 'cnecta_necta_state': 338, 'state_csa': 340, 'state_csa_cbsa': 341, 'state_cnecta': 345, 'state_cnecta_necta': 346, 'necta': 350, 'necta_state': 351, 'necta_state_place': 352, 'cnecta_state': 336, 'necta_state_county_cousub': 354, 'necta_nectadiv': 355, 'necta_nectadiv_state': 356, 'state_anrc': 230, 'necta_nectadiv_state_county': 357, 'state_necta': 360, 'cbsa_metdiv_state_county': 316, 'state_necta_county': 362, 'state_necta_county_cousub': 363, 'state_necta_nectadiv': 364, 'state_necta_nectadiv_county': 365, 'state_necta_nectadiv_county_cousub': 366, 'ua_state': 410, 'ua_state_county': 430, 'state_sldu_county': 612, 'state_sldu': 610, 'state_sldl_county': 622, 'state_sldl': 620, 'state_cdcurr_county': 510, 'state_necta_place': 361, 'aianhh': 250, 'aianhh_aitsce': 251, 'aianhh_aihhtli': 252, 'state_sldl_county': 622, 'aianhh_aihhtli254': 254 } lengths = { 'null': 1, 'aianhh': 4, # American Indian Area/Alaska Native Area/ Hawaiian Home Land (Census) 'aihhtli': '1', # American Indian Trust Land/ Hawaiian Home Land Indicator. A str b/c Census val is a str 'aitsce': 3, # American Indian Tribal Subdivision (Census) 'anrc': 5, # Alaska Native Regional Corporation (FIPS) 'blkgrp': 1, # Block Group 'blockgroup': 1, # Block Group 'block': 4, # Block 'cbsa': 5, # Metropolitan and Micropolitan Statistical Area 'cdcurr': 2, # Current Congressional District *** 'cnecta': 3, # New England City and Town Combined Statistical Area 'concit': 5, # Consolidated City 'county': 3, # County of current residence 'cousub': 5, # County Subdivision (FIPS) 'cosub': 5, # County Subdivision (FIPS) 'csa': 3, # Combined Statistical Area 'division': 1, # Census Division 'metdiv': 5, # Metropolitan Statistical Area- Metropolitan Division 'necta': 5, # New England City and Town Area 'nectadiv': 5, # New England City and Town Area Division 'place': 5, # Place (FIPS Code) 'puma5': 5, # Public Use Microdata Area 5% File 'region': 1, # Census Region 'sdelm': 5, # State-School District (Elementary) 'sdsec': 5, # State-School District (Secondary) 'sduni': 5, # State-School District (Unified) 'sldl': '3', # State Legislative District Lower. A String to signal that the census value is a string 'sldu': '3', # State Legislative District Upper. A String to signal that the census value is a string 'state': 2, # State (FIPS Code) 'submcd': 5, # Subminor Civil Division (FIPS) 'tract': 6, # Census Tract 'ua': 5, # Urban Area 'ur': 1, # Urban/Rural 'us': 0, 'zcta': 5, # Nonstandard 'zip': 5, } segments = { 1: ['null'], # United States 10: ['us'], # United States 20: ['region'], # Region 30: ['division'], # Division 40: ['state'], # State 50: ['state', 'county'], # County 60: ['state', 'county', 'cousub'], # County Subdivision 67: ['state', 'county', 'cousub', 'submcd'], # State (Puerto Rico Only)-County-County Subdivision-Subbarrio 70: ['state', 'county', 'cousub', 'place'], # County Subdivision-Place/Remainder 80: ['state', 'county', 'cousub', 'place', 'tract'], # County Subdivision-Place/Remainder-Census Tract 101: ['state', 'county', 'tract', 'block'], 140: ['state', 'county', 'tract'], # Census Tract 150: ['state', 'county', 'tract', 'blockgroup'], # Census Tract-Block Group 155: ['state', 'place', 'county'], # Place-County 160: ['state', 'place'], # Place 170: ['state', 'concit'], # Consolidated City 172: ['state', 'concit', 'place'], # Consolidated City-Place Within Consolidated City 230: ['state', 'anrc'], # State-Alaska Native Regional Corporation 250: ['aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land 251: ['aianhh', 'aitsce'], # American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder 252: ['aianhh', 'aihhtli'], # American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)4 254: ['aianhh', 'aihhtli'], # American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 260: ['state', 'aianhh'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State 269: ['state', 'aianhh', 'place'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder 270: ['aianhh', 'state', 'county'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County 280: ['state', 'aianhh'], # State-American Indian Area/Alaska Native Area/Hawaiian Home Land 283: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only) 286: ['state', 'aianhh', 'aihhtli'], # State-American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land 290: ['aianhh', 'aitsce', 'state'], # American Indian Area/Alaska Native Area/Hawaiian Home Land-Tribal Subdivision/Remainder-State 310: ['cbsa'], # CBSA 311: ['cbsa', 'state'], # CBSA-State-County 312: ['cbsa', 'state', 'place'], # CBSA-State-Principal City 313: ['cbsa', 'state', 'county'], # CBSA-State-County 314: ['cbsa', 'metdiv'], # Metropolitan Statistical Area/Metropolitan Division 315: ['cbsa', 'metdiv', 'state'], # Metropolitan Statistical Area/Metropolitan Division-State 316: ['cbsa', 'metdiv', 'state', 'county'], # Metropolitan Statistical Area/Metropolitan Division-State-County 320: ['state', 'cbsa'], # State- CBSA 321: ['state', 'cbsa', 'place'], # State- CBSA -Principal City 322: ['state', 'cbsa', 'county'], # State- CBSA -County 323: ['state', 'cbsa', 'metdiv'], # State- Metropolitan Statistical Area/Metropolitan Division 324: ['state', 'cbsa', 'metdiv', 'county'], # State- Metropolitan Statistical Area/Metropolitan Division-County 330: ['csa'], # Combined Statistical Area 331: ['csa', 'state'], # Combined Statistical Area-State 332: ['csa', 'cbsa'], # Combined Statistical Area-CBSA 333: ['csa', 'cbsa', 'state'], # Combined Statistical Area-CBSA-State 335: ['cnecta'], # Combined New England City and Town Area 336: ['cnecta', 'state'], # Combined New England City and Town Area -State 337: ['cnecta', 'necta'], # Combined New England City and Town Area -New England City and Town Area 338: ['cnecta', 'necta', 'state'], # Combined New England City and Town Area -New England City and Town Area-State 340: ['state', 'csa'], # State-Combined Statistical Area 341: ['state', 'csa', 'cbsa'], # State-Combined Statistical Area-CBSA 345: ['state', 'cnecta'], # State-Combined New England City and Town Area 346: ['state', 'cnecta', 'necta'], # State-Combined New England City and Town Area-New England City and Town Area 350: ['necta'], # New England City and Town Area 351: ['necta', 'state'], # New England City and Town Area-State 352: ['necta', 'state', 'place'], # New England City and Town Area-State-Principal City 353: ['necta', 'state', 'county'], # New England City and Town Area-State-County 354: ['necta', 'state', 'county', 'cousub'], # New England City and Town Area-State-County-County Subdivision 355: ['necta', 'nectadiv'], # New England City and Town Area (NECTA)-NECTA Division 356: ['necta', 'nectadiv', 'state'], # New England City and Town Area (NECTA)-NECTA Division-State 357: ['necta', 'nectadiv', 'state', 'county'], # New England City and Town Area (NECTA)-NECTA Division-State-County 358: ['necta', 'nectadiv', 'state', 'county', 'cousub'], # New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision 360: ['state', 'necta'], # State-New England City and Town Area 361: ['state', 'necta', 'place'], # State-New England City and Town Area-Principal City 362: ['state', 'necta', 'county'], # State-New England City and Town Area-County 363: ['state', 'necta', 'county', 'cousub'], # State-New England City and Town Area-County-County Subdivision 364: ['state', 'necta', 'nectadiv'], # State-New England City and Town Area (NECTA)-NECTA Division 365: ['state', 'necta', 'nectadiv', 'county'], # State-New England City and Town Area (NECTA)-NECTA Division-County 366: ['state', 'necta', 'nectadiv', 'county', 'cousub'], # State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision 400: ['ua'], # Urban Area, 410: ['ua', 'state'], # Urban Area, State, 430: ['ua','state','county'], # Urban Area, State, County, 500: ['state', 'cdcurr'], # Congressional District 510: ['state', 'cdcurr', 'county'], # 550: ['state', 'cdcurr', 'aianhh'], # Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land 610: ['state', 'sldu'], # State Senate District 612: ['state', 'sldu', 'county'], # State Senate District-County 620: ['state', 'sldl'], # State House District 622: ['state', 'sldl', 'county'], # State House District-County 795: ['state', 'puma5'], # State-Public Use MicroSample Area 5% 860: ['zcta'], 950: ['state', 'sdelm'], # State-Elementary School District 960: ['state', 'sdsec'], # State-High School District 970: ['state', 'sduni'], # State-Unified School District # Nonstandard 1200: ['zip'] } descriptions = { 1: 'United States', 10: 'United States', 20: 'Region', 30: 'Division', 40: 'State', 50: 'County', 60: 'County Subdivision', 67: 'State (Puerto Rico Only)-County-County Subdivision-Subbarrio', 70: 'County Subdivision-Place/Remainder', 80: 'County Subdivision-Place/Remainder-Census Tract', 101: 'block', 140: 'Census Tract', 150: 'Census Tract-Block Group', 155: 'Place-County', 160: 'Place', 170: 'Consolidated City', 172: 'Consolidated City-Place Within Consolidated City', 230: 'State-Alaska Native Regional Corporation', 250: 'American Indian Area/Alaska Native Area/Hawaiian Home Land', 251: 'American Indian Area/Alaska NativeArea/HawaiianHomeLand-Tribal Subdivision/Remainder', 252: 'American Indian Area/Alaska Native Area (Reservation or Statistical Entity Only)', 254: 'American Indian Area (Off-Reservation Trust Land Only)/Hawaiian Home Land', 260: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State', 269: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-Place-Remainder', 270: 'American Indian Area/Alaska Native Area/Hawaiian Home Land-State-County', 280: 'State-American Indian Area/Alaska Native Area/Hawaiian Home Land', 283: 'aihhtli', 286: 'aihhtli', 290: 'state', 310: 'CBSA', 311: 'CBSA-State-County', 312: 'CBSA-State-Principal City', 313: 'CBSA-State-County', 314: 'Metropolitan Statistical Area/Metropolitan Division', 315: 'Metropolitan Statistical Area/Metropolitan Division-State', 316: 'Metropolitan Statistical Area/Metropolitan Division-State-County', 320: 'State- CBSA', 321: 'State- CBSA -Principal City', 322: 'State- CBSA -County', 323: 'State- Metropolitan Statistical Area/Metropolitan Division', 324: 'State- Metropolitan Statistical Area/Metropolitan Division-County', 330: 'Combined Statistical Area', 331: 'Combined Statistical Area-State', 332: 'Combined Statistical Area-CBSA', 333: 'Combined Statistical Area-CBSA-State', 335: 'Combined New England City and Town Area', 336: 'Combined New England City and Town Area -State', 337: 'Combined New England City and Town Area -New England City and Town Area', 338: 'Combined New England City and Town Area -New England City and Town Area-State', 340: 'State-Combined Statistical Area', 341: 'State-Combined Statistical Area-CBSA', 345: 'State-Combined New England City and Town Area', 346: 'State-Combined New England City and Town Area-New England City and Town Area', 350: 'New England City and Town Area', 351: 'New England City and Town Area-State', 352: 'New England City and Town Area-State-Principal City', 353: 'New England City and Town Area-State-County', 354: 'New England City and Town Area-State-County-County Subdivision', 355: 'New England City and Town Area (NECTA)-NECTA Division', 356: 'New England City and Town Area (NECTA)-NECTA Division-State', 357: 'New England City and Town Area (NECTA)-NECTA Division-State-County', 358: 'New England City and Town Area (NECTA)-NECTA Division-State-County-County Subdivision', 360: 'State-New England City and Town Area', 361: 'State-New England City and Town Area-Principal City', 362: 'State-New England City and Town Area-County', 363: 'State-New England City and Town Area-County-County Subdivision', 364: 'State-New England City and Town Area (NECTA)-NECTA Division', 365: 'State-New England City and Town Area (NECTA)-NECTA Division-County-County Subdivision', 400: 'Urban Area,', 410: 'Urban Area, State,', 430: 'Urban Area, State, County,', 500: 'Congressional District', 510: 'Congressional District, County', 550: 'Congressional District-American IndianArea/Alaska NativeArea/Hawaiian Home Land', 610: 'State Senate District', 612: 'State Senate District-County', 620: 'State House District', 622: 'State House District-County', 795: 'State-Public Use MicroSample Area 5%', 860: 'ZIP Code Tabulation Area', 950: 'State-Elementary School District', 960: 'State-High School District', 970: 'State-Unified School District', } plurals = { 'county': 'counties', 'place': 'places', 'Sdlu': 'State ' } class NotASummaryName(Exception): """An argument was not one of the valid summary names""" class ParseError(Exception): """Error parsing a geoid""" def parse_to_gvid(v): """Parse an ACS Geoid or a GVID to a GVID""" from geoid.civick import GVid from geoid.acs import AcsGeoid m1 = '' try: return GVid.parse(v) except ValueError as e: m1 = str(e) try: return AcsGeoid.parse(v).convert(GVid) except ValueError as e: raise ValueError("Failed to parse to either ACS or GVid: {}; {}".format(m1, str(e))) def base62_encode(num): """Encode a number in Base X. WIth the built-in alphabet, its base 62 `num`: The number to encode `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ num = int(num) alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' if (num == 0): return alphabet[0] arr = [] base = len(alphabet) while num: rem = num % base num = num // base arr.append(alphabet[rem]) arr.reverse() return ''.join(arr) def base62_decode(string): """Decode a Base X encoded string into the number Arguments: - `string`: The encoded string - `alphabet`: The alphabet to use for encoding Stolen from: http://stackoverflow.com/a/1119769/1144479 """ alphabet = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ' base = len(alphabet) strlen = len(string) num = 0 idx = 0 for char in string: power = (strlen - (idx + 1)) num += alphabet.index(char) * (base ** power) idx += 1 return int(num) def augment(module_name, base_class): """Call the augment() method for all of the derived classes in the module """ for name, cls in inspect.getmembers(sys.modules[module_name], lambda x : inspect.isclass(x) and issubclass(x, base_class) ): if cls == base_class: continue cls.augment() def get_class(module, sl): for name, named_sl in names.items(): if named_sl == sl or sl == name: return getattr(module, name.capitalize()) raise NotASummaryName("No class for summary_level {}".format(sl)) def make_classes(base_class, module): """Create derived classes and put them into the same module as the base class. This function is called at the end of each of the derived class modules, acs, census, civik and tiger. It will create a set of new derived class in the module, one for each of the enries in the `summary_levels` dict. """ from functools import partial for k in names: cls = base_class.class_factory(k.capitalize()) cls.augment() setattr(module, k.capitalize(), cls) setattr(module, 'get_class', partial(get_class, module)) class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return '' county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s") def generate_all(sumlevel, d): """Generate a dict that includes all of the available geoid values, with keys for the most common names for those values. """ from geoid.civick import GVid from geoid.tiger import TigerGeoid from geoid.acs import AcsGeoid sumlevel = int(sumlevel) d = dict(d.items()) # Map common name variants if 'cousub' in d: d['cosub'] = d['cousub'] del d['cousub'] if 'blkgrp' in d: d['blockgroup'] = d['blkgrp'] del d['blkgrp'] if 'zcta5' in d: d['zcta'] = d['zcta5'] del d['zcta5'] gvid_class = GVid.resolve_summary_level(sumlevel) if not gvid_class: return {} geoidt_class = TigerGeoid.resolve_summary_level(sumlevel) geoid_class = AcsGeoid.resolve_summary_level(sumlevel) try: return dict( gvid=str(gvid_class(**d)), geoid=str(geoid_class(**d)), geoidt=str(geoidt_class(**d)) ) except: raise
Metatab/geoid
geoid/core.py
CountyName.division_name
python
def division_name(self): """The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough'""" try: return next(e for e in self.type_names_re.search(self.name).groups() if e is not None) except AttributeError: # The search will fail for 'District of Columbia' return ''
The type designation for the county or county equivalent, such as 'County','Parish' or 'Borough
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L489-L495
null
class CountyName(object): """A Census county name, with methods to create shorter versions and return the types of division, which may be county, parish, borough, etc. """ # Strip the county and state name. THis doesn't work for some locations # where the county is actually called a parish or a bario. state_name_pattern = r', (.*)$' state_name_re = re.compile(state_name_pattern) def __init__(self, name): self.name = name def intuit_name(self, name): """Return a numeric value in the range [-1,1), indicating the likelyhood that the name is for a valuable of of this type. -1 indicates a strong non-match, 1 indicates a strong match, and 0 indicates uncertainty. """ raise NotImplementedError @property def state(self): try: county, state = self.name.split(',') return state except ValueError: # The search will fail for 'District of Columbia' return '' @property def medium_name(self): """The census name without the state""" return self.state_name_re.sub('', self.name) type_names = ( 'County', 'Municipio', 'Parish', 'Census Area', 'Borough', 'Municipality', 'city', 'City and Borough') type_name_pattern = '|'.join('({})'.format(e) for e in type_names) type_names_re = re.compile(type_name_pattern) @property county_name_pattern = r'(.+) {}, (.+)'.format(type_name_pattern) county_name_re = re.compile(county_name_pattern) @property def short_name(self): try: county, state = self.name.split(',') except ValueError: return self.name # 'District of Colombia' return self.type_names_re.sub('', county) def __str__(self): return self.name
Metatab/geoid
geoid/core.py
Geoid.augment
python
def augment(cls): import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl]
Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance.
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L562-L583
null
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.get_class
python
def get_class(cls, name_or_sl): try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl))
Return a derived class based on the class name or the summary_level
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L587-L598
null
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.geo_name
python
def geo_name(self): if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level)
Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return:
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L668-L689
null
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.parse
python
def parse(cls, gvid, exception=True): if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d)
Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return:
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L714-L781
[ "def get_class(cls, name_or_sl):\n \"\"\"Return a derived class based on the class name or the summary_level\"\"\"\n try:\n return cls.sl_map[int(name_or_sl)]\n\n except TypeError as e:\n raise TypeError(\"Bad name or sl: {} : {}\".format(name_or_sl, e))\n except ValueError:\n try:\n return cls.class_map[name_or_sl.lower()]\n except (KeyError, ValueError):\n raise NotASummaryName(\"Value '{}' is not a valid summary level\".format(name_or_sl))\n" ]
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.convert
python
def convert(self, root_cls): d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d)
Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc.
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L783-L796
[ "def get_class(cls, name_or_sl):\n \"\"\"Return a derived class based on the class name or the summary_level\"\"\"\n try:\n return cls.sl_map[int(name_or_sl)]\n\n except TypeError as e:\n raise TypeError(\"Bad name or sl: {} : {}\".format(name_or_sl, e))\n except ValueError:\n try:\n return cls.class_map[name_or_sl.lower()]\n except (KeyError, ValueError):\n raise NotASummaryName(\"Value '{}' is not a valid summary level\".format(name_or_sl))\n" ]
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.promote
python
def promote(self, level=None): if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d)
Convert to the next higher level summary level
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L810-L828
[ "def get_class(cls, name_or_sl):\n \"\"\"Return a derived class based on the class name or the summary_level\"\"\"\n try:\n return cls.sl_map[int(name_or_sl)]\n\n except TypeError as e:\n raise TypeError(\"Bad name or sl: {} : {}\".format(name_or_sl, e))\n except ValueError:\n try:\n return cls.class_map[name_or_sl.lower()]\n except (KeyError, ValueError):\n raise NotASummaryName(\"Value '{}' is not a valid summary level\".format(name_or_sl))\n" ]
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.allval
python
def allval(self): d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d)
Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state'
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L835-L846
[ "def get_class(cls, name_or_sl):\n \"\"\"Return a derived class based on the class name or the summary_level\"\"\"\n try:\n return cls.sl_map[int(name_or_sl)]\n\n except TypeError as e:\n raise TypeError(\"Bad name or sl: {} : {}\".format(name_or_sl, e))\n except ValueError:\n try:\n return cls.class_map[name_or_sl.lower()]\n except (KeyError, ValueError):\n raise NotASummaryName(\"Value '{}' is not a valid summary level\".format(name_or_sl))\n" ]
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError @classmethod def nullval(cls): """Create a new instance where all of the values are 0""" d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d) @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/core.py
Geoid.nullval
python
def nullval(cls): d = dict(cls.__dict__.items()) for k in d: d[k] = 0 d['sl'] = cls.sl d[cls.level] = 0 return cls(**d)
Create a new instance where all of the values are 0
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/core.py#L849-L860
null
class Geoid(object): @classmethod def resolve_summary_level(cls, sl): try: return cls.sl_map[sl] except KeyError: return None @classmethod def make_format_string(cls, level): sl_num = names[level] segs = segments[sl_num] formats = [] formats.append(cls.sl_format) for seg in segs: # Lengths dict may have strings to indicate string format usage. if int(lengths[seg]) <= 0: continue if isinstance(lengths[seg], int): fmt = cls.elem_format else: fmt = cls.elem_str_format formats.append(fmt.format(seg, cls.part_width(lengths[seg]))) return ''.join(formats) @classmethod def make_regex(cls, level): sl_num = names[level] segs = segments[sl_num] # Lengths dict may have strings to indicate string format usage. regexes = [cls.sl_regex] + [cls.elem_regex.format(seg, cls.part_width(lengths[seg])) for seg in segs if int(lengths[seg]) > 0] re_str = '^' + ''.join(regexes) + '$' return re_str @classmethod def augment(cls): """Augment the class with computed formats, regexes, and other things. This caches these values so they don't have to be created for every instance. """ import re level_name = cls.__name__.lower() cls.sl = names[level_name] cls.class_map[cls.__name__.lower()] = cls cls.sl_map[cls.sl] = cls cls.fmt = cls.make_format_string(cls.__name__.lower()) cls.regex_str = cls.make_regex(cls.__name__.lower()) cls.regex = re.compile(cls.regex_str) # List of field names cls.level = level_name cls.fields = segments[cls.sl] @classmethod def get_class(cls, name_or_sl): """Return a derived class based on the class name or the summary_level""" try: return cls.sl_map[int(name_or_sl)] except TypeError as e: raise TypeError("Bad name or sl: {} : {}".format(name_or_sl, e)) except ValueError: try: return cls.class_map[name_or_sl.lower()] except (KeyError, ValueError): raise NotASummaryName("Value '{}' is not a valid summary level".format(name_or_sl)) def __init__(self, *args, **kwargs): # This is a bit unusual, because it means, that , unlike normal # python args, a kwarg can overwrite a position arg. d = dict(zip(self.fields, args + ((0,) * 10))) # Add enough zeros to set all fields to zero d.update(kwargs) for k, v in d.items(): if k in self.fields: try: setattr(self, k, v) except TypeError as e: raise TypeError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) except ValueError as e: raise ValueError("Failed to convert '{}' ({}) for field '{}' in {}: {}" .format(v, type(v), k, type(self), e)) def __str__(self): d = self.__dict__ d['sl'] = self.sl # Hacks for special string cases if 'sldu' in d: d['sldu'] = str(d['sldu']).zfill(3) if 'sldl' in d: d['sldl'] = str(d['sldl']).zfill(3) try: fn = six.get_method_function(self.encode) kwargs = {k: fn(v) for k, v in d.items()} return self.fmt.format(**kwargs) except (ValueError, KeyError) as e: raise ValueError("Bad value in {}, data {} for format {}: {}".format(type(self), d, self.fmt, e)) @property def state_name(self): from geoid.censusnames import geo_names return geo_names[(self.state, 0)] @property def stusab(self): from geoid.censusnames import stusab try: return stusab[int(self.state)] except (AttributeError, ValueError): # Assume this is a Us object, or some other national object return 'US' @property def county_name(self): from geoid.censusnames import geo_names try: try: return CountyName(geo_names[(self.state, self.county)]) except KeyError: try: return CountyName("County #{}, {}".format(self.county,geo_names[(self.state, 0)])) except KeyError: return CountyName("County #{}, State#{}".format(self.county, self.state)) except Exception: return CountyName('') @property def geo_name(self): """ Return a name of the state or county, or, for other lowever levels, the name of the level type in the county. :return: """ if self.level == 'county': return str(self.county_name) elif self.level == 'state': return self.state_name else: if hasattr(self, 'county'): return "{} in {}".format(self.level,str(self.county_name)) elif hasattr(self, 'state'): return "{} in {}".format(self.level, self.state_name) else: return "a {}".format(self.level) def __hash__(self): return hash(str(self)) def __eq__(self, other): return str(self) == str(other) def __ne__(self, other): return str(self) != str(other) def __lt__(self, other): return str(self) < str(other) def __le__(self, other): return str(self) <= str(other) def __gt__(self, other): return str(self) > str(other) def __ge__(self, other): return str(self) >= str(other) @classmethod def parse(cls, gvid, exception=True): """ Parse a string value into the geoid of this class. :param gvid: String value to parse. :param exception: If true ( default) raise an eception on parse erorrs. If False, return a 'null' geoid. :return: """ if gvid == 'invalid': return cls.get_class('null')(0) if not bool(gvid): return None if not isinstance(gvid, six.string_types): raise TypeError("Can't parse; not a string. Got a '{}' ".format(type(gvid))) try: if not cls.sl: # Civick and ACS include the SL, so can call from base type. if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ sl = fn(gvid[0:cls.sl_width]) else: sl = cls.sl # Otherwise must use derived class. except ValueError as e: if exception: raise ValueError("Failed to parse gvid '{}': {}".format(gvid, str(e))) else: return cls.get_class('null')(0) try: cls = cls.sl_map[sl] except KeyError: if exception: raise ValueError("Failed to parse gvid '{}': Unknown summary level '{}' ".format(gvid, sl)) else: return cls.get_class('null')(0) m = cls.regex.match(gvid) if not m: raise ValueError("Failed to match '{}' to '{}' ".format(gvid, cls.regex_str)) d = m.groupdict() if not d: return None if six.PY3: fn = cls.decode else: fn = cls.decode.__func__ d = {k: fn(v) for k, v in d.items()} try: del d['sl'] except KeyError: pass return cls(**d) def convert(self, root_cls): """Convert to another derived class. cls is the base class for the derived type, ie AcsGeoid, TigerGeoid, etc. """ d = self.__dict__ d['sl'] = self.sl try: cls = root_cls.get_class(root_cls.sl) except (AttributeError, TypeError): # Hopefully because root_cls is a module cls = root_cls.get_class(self.sl) return cls(**d) def as_census(self): from geoid.census import CensusGeoid return self.convert(CensusGeoid) def as_acs(self): from geoid.acs import AcsGeoid return self.convert(AcsGeoid) def as_tiger(self): from geoid.tiger import TigerGeoid return self.convert(TigerGeoid) def promote(self, level=None): """Convert to the next higher level summary level""" if level is None: if len(self.fields) < 2: if self.level in ('region', 'division', 'state', 'ua'): cls = self.get_class('us') else: return None else: cls = self.get_class(self.fields[-2]) else: cls = self.get_class(level) d = dict(self.__dict__.items()) d['sl'] = self.sl return cls(**d) def summarize(self): """Convert all of the values to their max values. This form is used to represent the summary level""" raise NotImplementedError def allval(self): """Convert the last value to zero. This form represents the entire higher summary level at the granularity of the lower summary level. For example, for a county, it means 'All counties in the state' """ d = dict(self.__dict__.items()) d['sl'] = self.sl d[self.level] = 0 cls = self.get_class(self.sl) return cls(**d) @classmethod @property def tuples(self): """Return tuples of field, value, in the order of the levels as they are defined """ return [(field, getattr(self, field, None)) for field in self.fields] @property def is_summary(self): """Return True if this geoid is an summary -- all of the fields are 0""" return str(self) == str(self.summarize()) @property def is_allval(self): """Return True if this geoid is an allval -- the last field is zero, but the first is not""" tups = self.tuples return tups[-1][1] == 0 and tups[0][1] != 0 @property def level_plural(self): """Return the name of the level as a plural""" return plurals.get(self.level, self.level + "s")
Metatab/geoid
geoid/civick.py
GVid.summarize
python
def summarize(self): s = str(self.allval()) return self.parse(s[:2]+ ''.join(['Z']*len(s[2:])))
Convert all of the values to their max values. This form is used to represent the summary level
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/civick.py#L41-L46
[ "def parse(cls, gvid, exception=True):\n \"\"\"\n Parse a string value into the geoid of this class.\n\n :param gvid: String value to parse.\n :param exception: If true ( default) raise an eception on parse erorrs. If False, return a\n 'null' geoid.\n :return:\n \"\"\"\n\n if gvid == 'invalid':\n return cls.get_class('null')(0)\n\n if not bool(gvid):\n return None\n\n if not isinstance(gvid, six.string_types):\n raise TypeError(\"Can't parse; not a string. Got a '{}' \".format(type(gvid)))\n\n try:\n if not cls.sl:\n # Civick and ACS include the SL, so can call from base type.\n if six.PY3:\n fn = cls.decode\n else:\n fn = cls.decode.__func__\n\n sl = fn(gvid[0:cls.sl_width])\n else:\n sl = cls.sl # Otherwise must use derived class.\n\n except ValueError as e:\n if exception:\n raise ValueError(\"Failed to parse gvid '{}': {}\".format(gvid, str(e)))\n else:\n return cls.get_class('null')(0)\n\n try:\n cls = cls.sl_map[sl]\n except KeyError:\n if exception:\n raise ValueError(\"Failed to parse gvid '{}': Unknown summary level '{}' \".format(gvid, sl))\n else:\n return cls.get_class('null')(0)\n\n m = cls.regex.match(gvid)\n\n if not m:\n raise ValueError(\"Failed to match '{}' to '{}' \".format(gvid, cls.regex_str))\n\n d = m.groupdict()\n\n if not d:\n return None\n\n if six.PY3:\n fn = cls.decode\n else:\n fn = cls.decode.__func__\n\n d = {k: fn(v) for k, v in d.items()}\n\n try:\n del d['sl']\n except KeyError:\n pass\n\n return cls(**d)\n", "def allval(self):\n \"\"\"Convert the last value to zero. This form represents the entire higher summary level at the granularity\n of the lower summary level. For example, for a county, it means 'All counties in the state' \"\"\"\n\n d = dict(self.__dict__.items())\n d['sl'] = self.sl\n\n d[self.level] = 0\n\n cls = self.get_class(self.sl)\n\n return cls(**d)\n" ]
class GVid(Geoid): sl = None fmt = None class_map = {} sl_map = {} sl_width = 2 sl_format = '{sl:0>2s}' elem_format = '{{{}:0>{}s}}' elem_str_format = '{{{}:0>{}s}}' sl_regex = '(?P<sl>.{2})' elem_regex = '(?P<{}>.{{{}}})' encode = base62_encode decode = base62_decode @classmethod def part_width(cls, dec_width): # Convert a decimal number of digits to a base 62 number of digits, via strings. # Maybe would be faster to use log()? return len(base62_encode(int('9'*int(dec_width)))) @classmethod def class_factory(cls, name): def __init__(self, *args, **kwargs): cls.__init__(self, *args, **kwargs) return type(name, (cls,), {"__init__": __init__}) def __str__(self): try: r = super(GVid, self).__str__() assert r != '0' and r != 0 return r except ValueError: # There are a few types of geoids that can have strings in their values instead of numbers: # aihhtli and sdlu # FIXME: Do more analysis to determine if these values can be converted to numbers. return 'invalid'
Metatab/geoid
geoid/util.py
simplify
python
def simplify(geoids): from collections import defaultdict aggregated = defaultdict(set) d = {} for g in geoids: if not bool(g): continue av = g.allval() d[av] = None aggregated[av].add(g) compiled = set() for k, v in aggregated.items(): if len(v) >= 5: compiled.add(k) compiled.add(k.promote()) else: compiled |= v return compiled
Given a list of geoids, reduce it to a simpler set. If there are five or more geoids at one summary level convert them to a single geoid at the higher level. :param geoids: :return:
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/util.py#L3-L38
null
"""Utilities""" def isimplify(geoids): """Iteratively simplify until the set stops getting smaller. """ s0 = list(geoids) for i in range(10): s1 = simplify(s0) if len(s1) == len(s0): return s1 s0 = s1 def iallval(t): """Recursively promote and compute allvals """ if t: return [t.allval()] + iallval(t.promote()) else: return []
Metatab/geoid
geoid/util.py
isimplify
python
def isimplify(geoids): s0 = list(geoids) for i in range(10): s1 = simplify(s0) if len(s1) == len(s0): return s1 s0 = s1
Iteratively simplify until the set stops getting smaller.
train
https://github.com/Metatab/geoid/blob/4b7769406b00e59376fb6046b42a2f8ed706b33b/geoid/util.py#L40-L51
[ "def simplify(geoids):\n \"\"\"\n Given a list of geoids, reduce it to a simpler set. If there are five or more geoids at one summary level\n convert them to a single geoid at the higher level.\n\n :param geoids:\n :return:\n \"\"\"\n\n from collections import defaultdict\n\n aggregated = defaultdict(set)\n\n d = {}\n\n for g in geoids:\n\n if not bool(g):\n continue\n\n av = g.allval()\n\n d[av] = None\n\n aggregated[av].add(g)\n\n compiled = set()\n\n for k, v in aggregated.items():\n if len(v) >= 5:\n compiled.add(k)\n compiled.add(k.promote())\n else:\n compiled |= v\n\n return compiled\n" ]
"""Utilities""" def simplify(geoids): """ Given a list of geoids, reduce it to a simpler set. If there are five or more geoids at one summary level convert them to a single geoid at the higher level. :param geoids: :return: """ from collections import defaultdict aggregated = defaultdict(set) d = {} for g in geoids: if not bool(g): continue av = g.allval() d[av] = None aggregated[av].add(g) compiled = set() for k, v in aggregated.items(): if len(v) >= 5: compiled.add(k) compiled.add(k.promote()) else: compiled |= v return compiled def iallval(t): """Recursively promote and compute allvals """ if t: return [t.allval()] + iallval(t.promote()) else: return []
jpscaletti/authcode
authcode/auth.py
Auth.set_hasher
python
def set_hasher(self, hash, rounds=None): hash = hash.replace('-', '_') if hash not in VALID_HASHERS: raise WrongHashAlgorithm(WRONG_HASH_MESSAGE) hasher = getattr(ph, hash) utils.test_hasher(hasher) default_rounds = getattr(hasher, 'default_rounds', 1) min_rounds = getattr(hasher, 'min_rounds', 1) max_rounds = getattr(hasher, 'max_rounds', float("inf")) rounds = min(max(rounds or default_rounds, min_rounds), max_rounds) op = { 'schemes': VALID_HASHERS + DEPRECATED_HASHERS, 'deprecated': DEPRECATED_HASHERS, 'default': hash, hash + '__default_rounds': rounds } self.hasher = CryptContext(**op) self.hash = hash.replace('_', '-') # For testing self.rounds = rounds
Updates the has algorithm and, optionally, the number of rounds to use. Raises: `~WrongHashAlgorithm` if new algorithm isn't one of the three recomended options.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/auth.py#L140-L167
[ "def test_hasher(hasher):\n hasher.encrypt('test', rounds=hasher.min_rounds)\n" ]
class Auth(AuthenticationMixin, AuthorizationMixin, ViewsMixin): default_settings = { 'session_key': '_uhmac', 'user_name': 'user', 'csrf_key': '_csrf_token', 'csrf_header': 'X-CSRFToken', 'csrf_header_alt': 'X-CSRF-TOKEN', 'redirect_key': 'next', 'sign_in_redirect': '/', 'sign_out_redirect': '/', 'views': 'sign_in sign_out reset_password change_password'.split(' '), 'url_sign_in': '/sign-in/', 'url_sign_out': '/sign-out/', 'url_reset_password': '/reset-password/', 'url_change_password': '/change-password/', 'template_sign_in': None, 'template_sign_out': None, 'template_reset': None, 'template_reset_email': None, 'template_change_password': None, 'reset_email_subject': u'Reset your password', # Should logins be case insensitive? 'case_insensitive': True, # Prevent session fixation attacks, but # block having multiple logins at the same time. # If you set this to False, make sure to delete on logout all user's # information stored in the session. 'clear_session_on_logout': True, # Trying to save any password shorter than this will raise a ValueError. 'password_minlen': 5, # To help preventing denial-of-service via large passwords # See: https://www.djangoproject.com/weblog/2013/sep/15/security/ # Authenticaction with any password longer than this will automatically fail. # Trying to save any password longer than this will raise a ValueError. 'password_maxlen': 1024, 'token_life': 3 * 60, # minutes 'update_hash': True, 'wsgi': wsgi.werkzeug, 'pepper': u'', # considering deprecating it } def __init__(self, secret_key, db=None, hash=DEFAULT_HASHER, rounds=None, UserMixin=None, RoleMixin=None, roles=False, prefix=None, views_prefix=None, users_model_name=None, roles_model_name=None, **settings): self.secret_key = str(secret_key) assert len(self.secret_key) >= MIN_SECRET_LENGTH, \ "`secret_key` must be at least {} chars long".format(MIN_SECRET_LENGTH) self.set_hasher(hash, rounds) if prefix: prefix = prefix.lower().replace(' ', '') if not users_model_name: users_model_name = '{}User'.format(prefix.title()) if not roles_model_name: roles_model_name = '{}Role'.format(prefix.title()) if not views_prefix: views_prefix = '{}_'.format(prefix) settings.setdefault( 'url_sign_in', '/{prefix}{url}'.format( prefix=prefix, url=self.default_settings['url_sign_in'] ) ) settings.setdefault( 'url_sign_out', '/{prefix}{url}'.format( prefix=prefix, url=self.default_settings['url_sign_out'] ) ) settings.setdefault( 'url_reset_password', '/{prefix}{url}'.format( prefix=prefix, url=self.default_settings['url_reset_password'] ) ) settings.setdefault( 'url_change_password', '/{prefix}{url}'.format( prefix=prefix, url=self.default_settings['url_change_password'] ) ) self.db = db if db: self.users_model_name = users_model_name or 'User' roles = roles or RoleMixin self.User = extend_user_model(self, UserMixin, roles=roles) if roles: self.roles_model_name = roles_model_name or 'Role' self.Role = extend_role_model(self, self.User, RoleMixin) self.backends = [ self.auth_password, self.auth_token, ] self.session = {} self.views_prefix = views_prefix or u'' for name in self.default_settings: setattr(self, name, settings.get(name, self.default_settings[name]))
jpscaletti/authcode
authcode/auth_views_mixin.py
ViewsMixin.render_template
python
def render_template(self, name, **kwargs): custom_template = getattr(self, 'template_' + name) if custom_template: return self.render(custom_template, **kwargs) template = TEMPLATES.get(name) return self.default_render(template, **kwargs)
Search for a setting named ``template_<name>`` and renders it. If one is not defined it uses the default template of the library at ``autchode/templates/<name>,html``. To render the template uses the ``render`` function, a property that has been probably overwritten in a ``auth.setup_for_something`` function (eg. ``setup_for_flask``).
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/auth_views_mixin.py#L44-L57
[ "def default_render(self, template, **kwargs):\n tmpl = def_env.get_template(template)\n return tmpl.render(kwargs)\n", "def render(self, template, **kwargs):\n \"\"\"Should be overwritten in the setup\"\"\"\n return self.default_render(template, **kwargs) # pragma: no cover\n" ]
class ViewsMixin(object): ERROR_BAD_CSRF = 'BAD CSRF TOKEN' ERROR_SUSPENDED = 'ACCOUNT SUSPENDED' ERROR_CREDENTIALS = 'BAD CREDENTIALS' ERROR_BAD_TOKEN = 'WRONG TOKEN' ERROR_WRONG_TOKEN_USER = 'WRONG USER' ERROR_PASSW_TOO_SHORT = 'TOO SHORT' ERROR_PASSW_TOO_LONG = 'TOO LONG' ERROR_PASSW_MISMATCH = 'MISMATCH' ERROR_PASSW_CURRENT = 'FAIL' def auth_sign_in(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.sign_in(self, request, self.session, *args, **kwargs) def auth_sign_out(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.sign_out(self, request, **kwargs) def auth_reset_password(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.reset_password(self, request, **kwargs) def auth_change_password(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.change_password(self, request, **kwargs) def default_render(self, template, **kwargs): tmpl = def_env.get_template(template) return tmpl.render(kwargs) def render(self, template, **kwargs): """Should be overwritten in the setup""" return self.default_render(template, **kwargs) # pragma: no cover def send_email(self, user, subject, msg): """Should be overwritten in the setup""" print('To:', user) print('Subject:', subject) print(msg)
jpscaletti/authcode
authcode/auth_views_mixin.py
ViewsMixin.send_email
python
def send_email(self, user, subject, msg): print('To:', user) print('Subject:', subject) print(msg)
Should be overwritten in the setup
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/auth_views_mixin.py#L67-L71
null
class ViewsMixin(object): ERROR_BAD_CSRF = 'BAD CSRF TOKEN' ERROR_SUSPENDED = 'ACCOUNT SUSPENDED' ERROR_CREDENTIALS = 'BAD CREDENTIALS' ERROR_BAD_TOKEN = 'WRONG TOKEN' ERROR_WRONG_TOKEN_USER = 'WRONG USER' ERROR_PASSW_TOO_SHORT = 'TOO SHORT' ERROR_PASSW_TOO_LONG = 'TOO LONG' ERROR_PASSW_MISMATCH = 'MISMATCH' ERROR_PASSW_CURRENT = 'FAIL' def auth_sign_in(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.sign_in(self, request, self.session, *args, **kwargs) def auth_sign_out(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.sign_out(self, request, **kwargs) def auth_reset_password(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.reset_password(self, request, **kwargs) def auth_change_password(self, *args, **kwargs): request = self.request or kwargs.get('request') or args and args[0] return views.change_password(self, request, **kwargs) def render_template(self, name, **kwargs): """Search for a setting named ``template_<name>`` and renders it. If one is not defined it uses the default template of the library at ``autchode/templates/<name>,html``. To render the template uses the ``render`` function, a property that has been probably overwritten in a ``auth.setup_for_something`` function (eg. ``setup_for_flask``). """ custom_template = getattr(self, 'template_' + name) if custom_template: return self.render(custom_template, **kwargs) template = TEMPLATES.get(name) return self.default_render(template, **kwargs) def default_render(self, template, **kwargs): tmpl = def_env.get_template(template) return tmpl.render(kwargs) def render(self, template, **kwargs): """Should be overwritten in the setup""" return self.default_render(template, **kwargs) # pragma: no cover
jpscaletti/authcode
authcode/setups/setup_for_bottle.py
setup_for_bottle
python
def setup_for_bottle( auth, app, send_email=None, render=None, session=None, request=None, urloptions=None): import bottle auth.request = request or bottle.request if session is not None: auth.session = session if send_email: auth.send_email = send_email auth.render = render or bottle.template bottle.BaseTemplate.defaults['csrf_token'] = auth.get_csrf_token bottle.BaseTemplate.defaults['auth'] = auth @bottle.hook('before_request') def after_request(): auth.session = session or getattr(bottle.request, 'session') \ or bottle.request.environ.get('beaker.session') assert auth.session, 'Session not found' # By doing this, ``bottle.request`` now has a ``user`` attribute # that it's replaced by the real user object the first time is used. LazyUser(auth, bottle.request, user_name=auth.user_name) if auth.views: assert auth.render setup_for_bottle_views(auth, app, urloptions)
Set the session **before** calling ``setup_for_bottle`` like this: @hook('before_request') def setup_request(): request.session = request.environ['beaker.session']
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/setups/setup_for_bottle.py#L5-L40
[ "def setup_for_bottle_views(auth, app, urloptions):\n urloptions = urloptions or {}\n\n if 'sign_in' in auth.views:\n url_sign_in = eval_url(auth.url_sign_in)\n app.route(\n url_sign_in,\n method=['GET', 'POST'],\n name='{prefix}{name}'.format(\n prefix=auth.views_prefix,\n name='auth_sign_in'\n ),\n callback=auth.auth_sign_in,\n **urloptions\n )\n\n if 'sign_out' in auth.views:\n url_sign_out = eval_url(auth.url_sign_out)\n app.route(\n url_sign_out,\n method=['GET', 'POST'],\n name='{prefix}{name}'.format(\n prefix=auth.views_prefix,\n name='auth_sign_out'\n ),\n callback=auth.auth_sign_out,\n **urloptions\n )\n\n if 'change_password' in auth.views:\n url_change_password = eval_url(auth.url_change_password)\n app.route(\n url_change_password,\n method=['GET', 'POST'],\n name='{prefix}{name}'.format(\n prefix=auth.views_prefix,\n name='auth_change_password'\n ),\n callback=auth.auth_change_password,\n **urloptions\n )\n\n if 'reset_password' in auth.views:\n url_reset_password = eval_url(auth.url_reset_password)\n app.route(\n url_reset_password,\n method=['GET', 'POST'],\n name='{prefix}{name}'.format(\n prefix=auth.views_prefix,\n name='auth_reset_password'\n ),\n callback=auth.auth_reset_password,\n **urloptions\n )\n app.route(\n url_reset_password.rstrip('/') + '/<token>/',\n method=['GET', 'POST'],\n name='{prefix}{name}'.format(\n prefix=auth.views_prefix,\n name='auth_reset_password'\n ),\n callback=auth.auth_reset_password,\n **urloptions\n )\n" ]
# coding=utf-8 from ..utils import LazyUser, eval_url def setup_for_bottle_views(auth, app, urloptions): urloptions = urloptions or {} if 'sign_in' in auth.views: url_sign_in = eval_url(auth.url_sign_in) app.route( url_sign_in, method=['GET', 'POST'], name='{prefix}{name}'.format( prefix=auth.views_prefix, name='auth_sign_in' ), callback=auth.auth_sign_in, **urloptions ) if 'sign_out' in auth.views: url_sign_out = eval_url(auth.url_sign_out) app.route( url_sign_out, method=['GET', 'POST'], name='{prefix}{name}'.format( prefix=auth.views_prefix, name='auth_sign_out' ), callback=auth.auth_sign_out, **urloptions ) if 'change_password' in auth.views: url_change_password = eval_url(auth.url_change_password) app.route( url_change_password, method=['GET', 'POST'], name='{prefix}{name}'.format( prefix=auth.views_prefix, name='auth_change_password' ), callback=auth.auth_change_password, **urloptions ) if 'reset_password' in auth.views: url_reset_password = eval_url(auth.url_reset_password) app.route( url_reset_password, method=['GET', 'POST'], name='{prefix}{name}'.format( prefix=auth.views_prefix, name='auth_reset_password' ), callback=auth.auth_reset_password, **urloptions ) app.route( url_reset_password.rstrip('/') + '/<token>/', method=['GET', 'POST'], name='{prefix}{name}'.format( prefix=auth.views_prefix, name='auth_reset_password' ), callback=auth.auth_reset_password, **urloptions )
jpscaletti/authcode
authcode/wsgi/bottle.py
get_site_name
python
def get_site_name(request): urlparts = request.urlparts return ':'.join([urlparts.hostname, str(urlparts.port)])
Return the domain:port part of the URL without scheme. Eg: facebook.com, 127.0.0.1:8080, etc.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/bottle.py#L10-L15
null
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_FORBIDDEN = 403 def get_full_path(request): """Return the current relative path including the query string. Eg: “/foo/bar/?page=1” """ path = request.fullpath query_string = request.environ.get('QUERY_STRING') if query_string: path += '?' + to_native(query_string) return path def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ urlparts = request.urlparts return '{scheme}://{site}/{url}'.format( scheme=urlparts.scheme, site=get_site_name(request), url=url.lstrip('/'), ) def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from bottle import redirect redirect(url) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from bottle import abort abort(HTTP_FORBIDDEN, msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ return request.query.get(key) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ return request.headers.get(key) def get_post_data(request): """Return all the POST data from the request. """ return request.forms def make_response(body, mimetype='text/html'): """Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``. """ from bottle import response response.content_type = mimetype return body or u''
jpscaletti/authcode
authcode/wsgi/bottle.py
get_full_path
python
def get_full_path(request): path = request.fullpath query_string = request.environ.get('QUERY_STRING') if query_string: path += '?' + to_native(query_string) return path
Return the current relative path including the query string. Eg: “/foo/bar/?page=1”
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/bottle.py#L18-L26
[ "def to_native(x, charset='utf8', errors='ignore'):\n bb = to_bytes(x, charset=charset, errors=errors)\n if not bb:\n return bb\n return bb.decode('utf8')\n" ]
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_FORBIDDEN = 403 def get_site_name(request): """Return the domain:port part of the URL without scheme. Eg: facebook.com, 127.0.0.1:8080, etc. """ urlparts = request.urlparts return ':'.join([urlparts.hostname, str(urlparts.port)]) def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ urlparts = request.urlparts return '{scheme}://{site}/{url}'.format( scheme=urlparts.scheme, site=get_site_name(request), url=url.lstrip('/'), ) def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from bottle import redirect redirect(url) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from bottle import abort abort(HTTP_FORBIDDEN, msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ return request.query.get(key) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ return request.headers.get(key) def get_post_data(request): """Return all the POST data from the request. """ return request.forms def make_response(body, mimetype='text/html'): """Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``. """ from bottle import response response.content_type = mimetype return body or u''
jpscaletti/authcode
authcode/wsgi/bottle.py
make_full_url
python
def make_full_url(request, url): urlparts = request.urlparts return '{scheme}://{site}/{url}'.format( scheme=urlparts.scheme, site=get_site_name(request), url=url.lstrip('/'), )
Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open”
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/bottle.py#L29-L38
[ "def get_site_name(request):\n \"\"\"Return the domain:port part of the URL without scheme.\n Eg: facebook.com, 127.0.0.1:8080, etc.\n \"\"\"\n urlparts = request.urlparts\n return ':'.join([urlparts.hostname, str(urlparts.port)])\n" ]
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_FORBIDDEN = 403 def get_site_name(request): """Return the domain:port part of the URL without scheme. Eg: facebook.com, 127.0.0.1:8080, etc. """ urlparts = request.urlparts return ':'.join([urlparts.hostname, str(urlparts.port)]) def get_full_path(request): """Return the current relative path including the query string. Eg: “/foo/bar/?page=1” """ path = request.fullpath query_string = request.environ.get('QUERY_STRING') if query_string: path += '?' + to_native(query_string) return path def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from bottle import redirect redirect(url) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from bottle import abort abort(HTTP_FORBIDDEN, msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ return request.query.get(key) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ return request.headers.get(key) def get_post_data(request): """Return all the POST data from the request. """ return request.forms def make_response(body, mimetype='text/html'): """Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``. """ from bottle import response response.content_type = mimetype return body or u''
jpscaletti/authcode
authcode/wsgi/bottle.py
make_response
python
def make_response(body, mimetype='text/html'): from bottle import response response.content_type = mimetype return body or u''
Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/bottle.py#L87-L93
null
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_FORBIDDEN = 403 def get_site_name(request): """Return the domain:port part of the URL without scheme. Eg: facebook.com, 127.0.0.1:8080, etc. """ urlparts = request.urlparts return ':'.join([urlparts.hostname, str(urlparts.port)]) def get_full_path(request): """Return the current relative path including the query string. Eg: “/foo/bar/?page=1” """ path = request.fullpath query_string = request.environ.get('QUERY_STRING') if query_string: path += '?' + to_native(query_string) return path def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ urlparts = request.urlparts return '{scheme}://{site}/{url}'.format( scheme=urlparts.scheme, site=get_site_name(request), url=url.lstrip('/'), ) def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from bottle import redirect redirect(url) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from bottle import abort abort(HTTP_FORBIDDEN, msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ return request.query.get(key) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ return request.headers.get(key) def get_post_data(request): """Return all the POST data from the request. """ return request.forms
jpscaletti/authcode
authcode/auth_authentication_mixin.py
AuthenticationMixin.login
python
def login(self, user, remember=True, session=None): logger = logging.getLogger(__name__) logger.debug(u'User `{0}` logged in'.format(user.login)) if session is None: session = self.session session['permanent'] = remember session[self.session_key] = user.get_uhmac() if callable(getattr(session, 'save', None)): session.save()
Sets the current user UID in the session. Instead of just storing the user's id, it generates a hash from the password *salt*. That way, an admin or the user herself can invalidate the login in other computers just by changing (or re-saving) her password.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/auth_authentication_mixin.py#L130-L146
null
class AuthenticationMixin(object): def prepare_password(self, secret): return u'{pepper}{secret}'.format( pepper=to_unicode(self.pepper), secret=to_unicode(secret) ) def hash_password(self, secret): if secret is None: return None len_secret = len(secret) if len_secret < self.password_minlen: raise ValueError( 'Password is too short. Must have at least {} chars long'.format( self.password_minlen)) if len_secret > self.password_maxlen: raise ValueError( 'Password is too long. Must have at most {} chars long'.format( self.password_maxlen)) secret = self.prepare_password(secret) hashed = self.hasher.encrypt(secret) return hashed def password_is_valid(self, secret, hashed): if secret is None or hashed is None: return False # To help preventing denial-of-service via large passwords # See: https://www.djangoproject.com/weblog/2013/sep/15/security/ if len(secret) > self.password_maxlen: return False secret = self.prepare_password(secret) try: return self.hasher.verify(secret, hashed) except ValueError: return False def authenticate(self, credentials): for backend in self.backends: user = backend(credentials) if user: return user return None def auth_password(self, credentials): logger = logging.getLogger(__name__) login = credentials.get('login') secret = credentials.get('password') if login is None or secret is None: return None user = self.User.by_login(login) if not user: logger.debug(u'User `{0}` not found'.format(login)) return None if not user.password: logger.debug(u'User `{0}` has no password'.format(login)) return None if not self.password_is_valid(secret, user.password): logger.debug(u'Invalid password for user `{0}`'.format(login)) return None self._update_password_hash(secret, user) return user def _update_password_hash(self, secret, user): if not self.update_hash: return new_hash = self.hash_password(secret) if new_hash.split('$')[:3] == user.password.split('$')[:3]: return user.set_raw_password(new_hash) def auth_token(self, credentials, token_life=None): logger = logging.getLogger(__name__) token = credentials.get('token') if token is None: return None try: timestamp, uid = utils.split_token(str(token)) except ValueError: logger.info(u'Invalid auth token format') return None token_life = token_life or self.token_life user = self.User.by_id(uid) if not user: logger.info(u'Tampered auth token? uid `{0} not found'.format(uid[:20])) return None valid = user.get_token(timestamp) == token not_expired = timestamp + token_life >= int(time()) if valid and not_expired: return user logger.info(u'Invalid auth token') return None def get_user(self, session=None): if session is None: session = self.session user = None uhmac = session.get(self.session_key) if uhmac: try: uid = utils.split_uhmac(uhmac) user = self.User.by_id(uid) if not user or uhmac != user.get_uhmac() or not user.login: raise ValueError except ValueError: logger = logging.getLogger(__name__) logger.warn(u'Tampered uhmac?') user = None self.logout(session) return user def logout(self, session=None): if session is None: session = self.session if self.session_key in session: del session[self.session_key] if self.clear_session_on_logout: session.clear() if callable(getattr(session, 'save', None)): session.save()
jpscaletti/authcode
authcode/auth_authorization_mixin.py
AuthorizationMixin.protected
python
def protected(self, *tests, **kwargs): _role = kwargs.pop('role', None) _roles = kwargs.pop('roles', None) or [] _csrf = kwargs.pop('csrf', None) _url_sign_in = kwargs.pop('url_sign_in', None) _request = kwargs.pop('request', None) if _role: _roles.append(_role) _roles = [to_unicode(r) for r in _roles] _tests = tests _user_tests = kwargs def decorator(f): @functools.wraps(f) def wrapper(*args, **kwargs): logger = logging.getLogger(__name__) request = _request or self.request or args and args[0] url_sign_in = self._get_url_sign_in(request, _url_sign_in) user = self.get_user() if not user: return self._login_required(request, url_sign_in) if hasattr(user, 'has_role') and _roles: if not user.has_role(*_roles): logger.debug(u'User `{0}`: has_role fail'.format(user.login)) logger.debug(u'User roles: {0}'.format([r.name for r in user.roles])) return self.wsgi.raise_forbidden() for test in _tests: test_pass = test(user, *args, **kwargs) if not test_pass: logger.debug(u'User `{0}`: test fail'.format(user.login)) return self.wsgi.raise_forbidden() for name, value in _user_tests.items(): user_test = getattr(user, name) test_pass = user_test(value, *args, **kwargs) if not test_pass: logger.debug(u'User `{0}`: test fail'.format(user.login)) return self.wsgi.raise_forbidden() disable_csrf = _csrf == False # noqa if (not self.wsgi.is_idempotent(request) and not disable_csrf) or _csrf: if not self.csrf_token_is_valid(request): logger.debug(u'User `{0}`: invalid CSFR token'.format(user.login)) return self.wsgi.raise_forbidden("CSFR token isn't valid") return f(*args, **kwargs) return wrapper return decorator
Factory of decorators for limit the access to views. :tests: *function, optional One or more functions that takes the args and kwargs of the view and returns either `True` or `False`. All test must return True to show the view. Options: :role: str, optional Test for the user having a role with this name. :roles: list, optional Test for the user having **any** role in this list of names. :csrf: bool, None, optional If ``None`` (the default), the decorator will check the value of the CSFR token for POST, PUT or DELETE requests. If ``True`` it will do the same also for all requests. If ``False``, the value of the CSFR token will not be checked. :url_sign_in: str, function, optional If any required condition fail, redirect to this place. Override the default URL. This can also be a callable. :request: obj, optional Overwrite the request for testing. The rest of the ``key=value`` pairs in ``kwargs`` are interpreted as tests. The user must have a property `key` with a value equals to `value`. If the user has a method named `key`, that method is called with `value` as a single argument and must return True to show the view.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/auth_authorization_mixin.py#L31-L117
null
class AuthorizationMixin(object): # Useful for setting a cookie only if the CSRF token has changed. csrf_token_has_changed = False def get_csrf_token(self, session=None): logger = logging.getLogger(__name__) if session is None: session = self.session csrf_token = session.get(self.csrf_key) if not csrf_token: logger.debug(u'New CSFR token') csrf_token = self.make_csrf_token() session[self.csrf_key] = csrf_token if callable(getattr(session, 'save', None)): session.save() return csrf_token def make_csrf_token(self): self.csrf_token_has_changed = True return str(uuid4()).replace('-', '') def replace_flask_route(self, bp, *args, **kwargs): """Replace the Flask `app.route` or `blueprint.route` with a version that first apply the protected decorator to the view, so all views are automatically protected.""" protected = self.protected def protected_route(rule, **options): """Like :meth:`Flask.route` but for a blueprint. The endpoint for the :func:`url_for` function is prefixed with the name of the blueprint. """ def decorator(f): endpoint = options.pop("endpoint", f.__name__) protected_f = protected(*args, **kwargs)(f) bp.add_url_rule(rule, endpoint, protected_f, **options) return f return decorator bp.route = protected_route def csrf_token_is_valid(self, request, session=None): token = self._get_csrf_token_from_request(request) return token and self._csrf_token_is_valid(token, session) def _csrf_token_is_valid(self, token, session=None): new_token = self.get_csrf_token(session=session) return new_token == token def _login_required(self, request, url_sign_in): self.session[self.redirect_key] = self.wsgi.get_full_path(request) if callable(getattr(self.session, 'save', None)): self.session.save() return self.wsgi.redirect(url_sign_in) def _get_url_sign_in(self, request, url_sign_in): url_sign_in = url_sign_in or self.url_sign_in if callable(url_sign_in): url_sign_in = url_sign_in(request) return url_sign_in or '/' def _get_csrf_token_from_request(self, request): token = self.wsgi.get_from_params(request, self.csrf_key) or \ self.wsgi.get_from_headers(request, self.csrf_header) or \ self.wsgi.get_from_headers(request, self.csrf_header_alt) return token
jpscaletti/authcode
authcode/auth_authorization_mixin.py
AuthorizationMixin.replace_flask_route
python
def replace_flask_route(self, bp, *args, **kwargs): protected = self.protected def protected_route(rule, **options): """Like :meth:`Flask.route` but for a blueprint. The endpoint for the :func:`url_for` function is prefixed with the name of the blueprint. """ def decorator(f): endpoint = options.pop("endpoint", f.__name__) protected_f = protected(*args, **kwargs)(f) bp.add_url_rule(rule, endpoint, protected_f, **options) return f return decorator bp.route = protected_route
Replace the Flask `app.route` or `blueprint.route` with a version that first apply the protected decorator to the view, so all views are automatically protected.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/auth_authorization_mixin.py#L119-L136
null
class AuthorizationMixin(object): # Useful for setting a cookie only if the CSRF token has changed. csrf_token_has_changed = False def get_csrf_token(self, session=None): logger = logging.getLogger(__name__) if session is None: session = self.session csrf_token = session.get(self.csrf_key) if not csrf_token: logger.debug(u'New CSFR token') csrf_token = self.make_csrf_token() session[self.csrf_key] = csrf_token if callable(getattr(session, 'save', None)): session.save() return csrf_token def make_csrf_token(self): self.csrf_token_has_changed = True return str(uuid4()).replace('-', '') def protected(self, *tests, **kwargs): """Factory of decorators for limit the access to views. :tests: *function, optional One or more functions that takes the args and kwargs of the view and returns either `True` or `False`. All test must return True to show the view. Options: :role: str, optional Test for the user having a role with this name. :roles: list, optional Test for the user having **any** role in this list of names. :csrf: bool, None, optional If ``None`` (the default), the decorator will check the value of the CSFR token for POST, PUT or DELETE requests. If ``True`` it will do the same also for all requests. If ``False``, the value of the CSFR token will not be checked. :url_sign_in: str, function, optional If any required condition fail, redirect to this place. Override the default URL. This can also be a callable. :request: obj, optional Overwrite the request for testing. The rest of the ``key=value`` pairs in ``kwargs`` are interpreted as tests. The user must have a property `key` with a value equals to `value`. If the user has a method named `key`, that method is called with `value` as a single argument and must return True to show the view. """ _role = kwargs.pop('role', None) _roles = kwargs.pop('roles', None) or [] _csrf = kwargs.pop('csrf', None) _url_sign_in = kwargs.pop('url_sign_in', None) _request = kwargs.pop('request', None) if _role: _roles.append(_role) _roles = [to_unicode(r) for r in _roles] _tests = tests _user_tests = kwargs def decorator(f): @functools.wraps(f) def wrapper(*args, **kwargs): logger = logging.getLogger(__name__) request = _request or self.request or args and args[0] url_sign_in = self._get_url_sign_in(request, _url_sign_in) user = self.get_user() if not user: return self._login_required(request, url_sign_in) if hasattr(user, 'has_role') and _roles: if not user.has_role(*_roles): logger.debug(u'User `{0}`: has_role fail'.format(user.login)) logger.debug(u'User roles: {0}'.format([r.name for r in user.roles])) return self.wsgi.raise_forbidden() for test in _tests: test_pass = test(user, *args, **kwargs) if not test_pass: logger.debug(u'User `{0}`: test fail'.format(user.login)) return self.wsgi.raise_forbidden() for name, value in _user_tests.items(): user_test = getattr(user, name) test_pass = user_test(value, *args, **kwargs) if not test_pass: logger.debug(u'User `{0}`: test fail'.format(user.login)) return self.wsgi.raise_forbidden() disable_csrf = _csrf == False # noqa if (not self.wsgi.is_idempotent(request) and not disable_csrf) or _csrf: if not self.csrf_token_is_valid(request): logger.debug(u'User `{0}`: invalid CSFR token'.format(user.login)) return self.wsgi.raise_forbidden("CSFR token isn't valid") return f(*args, **kwargs) return wrapper return decorator def csrf_token_is_valid(self, request, session=None): token = self._get_csrf_token_from_request(request) return token and self._csrf_token_is_valid(token, session) def _csrf_token_is_valid(self, token, session=None): new_token = self.get_csrf_token(session=session) return new_token == token def _login_required(self, request, url_sign_in): self.session[self.redirect_key] = self.wsgi.get_full_path(request) if callable(getattr(self.session, 'save', None)): self.session.save() return self.wsgi.redirect(url_sign_in) def _get_url_sign_in(self, request, url_sign_in): url_sign_in = url_sign_in or self.url_sign_in if callable(url_sign_in): url_sign_in = url_sign_in(request) return url_sign_in or '/' def _get_csrf_token_from_request(self, request): token = self.wsgi.get_from_params(request, self.csrf_key) or \ self.wsgi.get_from_headers(request, self.csrf_header) or \ self.wsgi.get_from_headers(request, self.csrf_header_alt) return token
jpscaletti/authcode
authcode/wsgi/werkzeug.py
get_full_path
python
def get_full_path(request): path = request.path if request.query_string: path += '?' + to_native(request.query_string) return path
Return the current relative path including the query string. Eg: “/foo/bar/?page=1”
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/werkzeug.py#L10-L17
[ "def to_native(x, charset='utf8', errors='ignore'):\n bb = to_bytes(x, charset=charset, errors=errors)\n if not bb:\n return bb\n return bb.decode('utf8')\n" ]
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_SEE_OTHER = 303 def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ return request.url_root + url.lstrip('/') def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from werkzeug.utils import redirect return redirect(url, code=HTTP_SEE_OTHER) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from werkzeug.exceptions import Forbidden raise Forbidden(msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ data = getattr(request, 'json', None) or request.values value = data.get(key) return to_native(value) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ value = request.headers.get(key) return to_native(value) def get_post_data(request): """Return all the POST data from the request. """ return getattr(request, 'json', None) or request.form or {} def make_response(body, mimetype='text/html'): """Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``. """ from werkzeug.wrappers import Response if isinstance(body, Response): body.mimetype = mimetype return body return Response(body, mimetype=mimetype)
jpscaletti/authcode
authcode/wsgi/werkzeug.py
get_from_params
python
def get_from_params(request, key): data = getattr(request, 'json', None) or request.values value = data.get(key) return to_native(value)
Try to read a value named ``key`` from the GET parameters.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/werkzeug.py#L55-L60
[ "def to_native(x, charset='utf8', errors='ignore'):\n bb = to_bytes(x, charset=charset, errors=errors)\n if not bb:\n return bb\n return bb.decode('utf8')\n" ]
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_SEE_OTHER = 303 def get_full_path(request): """Return the current relative path including the query string. Eg: “/foo/bar/?page=1” """ path = request.path if request.query_string: path += '?' + to_native(request.query_string) return path def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ return request.url_root + url.lstrip('/') def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from werkzeug.utils import redirect return redirect(url, code=HTTP_SEE_OTHER) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from werkzeug.exceptions import Forbidden raise Forbidden(msg) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ value = request.headers.get(key) return to_native(value) def get_post_data(request): """Return all the POST data from the request. """ return getattr(request, 'json', None) or request.form or {} def make_response(body, mimetype='text/html'): """Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``. """ from werkzeug.wrappers import Response if isinstance(body, Response): body.mimetype = mimetype return body return Response(body, mimetype=mimetype)
jpscaletti/authcode
authcode/wsgi/werkzeug.py
get_from_headers
python
def get_from_headers(request, key): value = request.headers.get(key) return to_native(value)
Try to read a value named ``key`` from the headers.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/werkzeug.py#L63-L67
[ "def to_native(x, charset='utf8', errors='ignore'):\n bb = to_bytes(x, charset=charset, errors=errors)\n if not bb:\n return bb\n return bb.decode('utf8')\n" ]
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_SEE_OTHER = 303 def get_full_path(request): """Return the current relative path including the query string. Eg: “/foo/bar/?page=1” """ path = request.path if request.query_string: path += '?' + to_native(request.query_string) return path def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ return request.url_root + url.lstrip('/') def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from werkzeug.utils import redirect return redirect(url, code=HTTP_SEE_OTHER) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from werkzeug.exceptions import Forbidden raise Forbidden(msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ data = getattr(request, 'json', None) or request.values value = data.get(key) return to_native(value) def get_post_data(request): """Return all the POST data from the request. """ return getattr(request, 'json', None) or request.form or {} def make_response(body, mimetype='text/html'): """Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``. """ from werkzeug.wrappers import Response if isinstance(body, Response): body.mimetype = mimetype return body return Response(body, mimetype=mimetype)
jpscaletti/authcode
authcode/wsgi/werkzeug.py
make_response
python
def make_response(body, mimetype='text/html'): from werkzeug.wrappers import Response if isinstance(body, Response): body.mimetype = mimetype return body return Response(body, mimetype=mimetype)
Build a framework specific HTPP response, containing ``body`` and marked as the type ``mimetype``.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/wsgi/werkzeug.py#L76-L84
null
# coding=utf-8 from __future__ import absolute_import from .._compat import to_native HTTP_SEE_OTHER = 303 def get_full_path(request): """Return the current relative path including the query string. Eg: “/foo/bar/?page=1” """ path = request.path if request.query_string: path += '?' + to_native(request.query_string) return path def make_full_url(request, url): """Get a relative URL and returns the absolute version. Eg: “/foo/bar?q=is-open” ==> “http://example.com/foo/bar?q=is-open” """ return request.url_root + url.lstrip('/') def is_post(request): """Return ``True`` if the method of the request is ``POST``. """ return request.method.upper() == 'POST' def is_idempotent(request): """Return ``True`` if the method of the request is ``GET`` or ``HEAD``. """ return request.method.upper() in ('GET', 'HEAD') def redirect(url): """Return an HTTP 303 See Other response for this url, in the idiom of the framework. """ from werkzeug.utils import redirect return redirect(url, code=HTTP_SEE_OTHER) def raise_forbidden(msg='You are not allowed to access this.'): """Return an HTTP 403 Forbidden response (with the passed message), in the idiom of the framework. """ from werkzeug.exceptions import Forbidden raise Forbidden(msg) def get_from_params(request, key): """Try to read a value named ``key`` from the GET parameters. """ data = getattr(request, 'json', None) or request.values value = data.get(key) return to_native(value) def get_from_headers(request, key): """Try to read a value named ``key`` from the headers. """ value = request.headers.get(key) return to_native(value) def get_post_data(request): """Return all the POST data from the request. """ return getattr(request, 'json', None) or request.form or {}
jpscaletti/authcode
authcode/utils.py
get_uhmac
python
def get_uhmac(user, secret): secret = to_bytes(secret) key = '|'.join([ hashlib.sha1(secret).hexdigest(), str(user.id), get_hash_extract(user.password), ]) key = key.encode('utf8', 'ignore') mac = hmac.new(key, msg=None, digestmod=hashlib.sha512) mac = mac.hexdigest()[:50] uhmac = '{0}${1}'.format(user.id, mac) return uhmac
Make an unique identifier for the user (stored in the session), so it can stay logged between requests. By hashing a snippet of the current password hash salt, it makes possible to automatically logout from all other devices just by changing (or refreshing) the password.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/utils.py#L42-L60
[ "def to_bytes(x, charset='utf8', errors='ignore'):\n if x is None:\n return None\n if isinstance(x, (bytes, bytearray, memoryview)):\n return bytes(x)\n if isinstance(x, str):\n return x.encode(charset, errors)\n raise TypeError('Expected bytes')\n", "def get_hash_extract(hash):\n if not hash:\n return u''\n half = hash.rsplit('$', 1)[0]\n return half[-10:]\n" ]
# coding=utf-8 import hashlib import hmac from time import time from ._compat import to_bytes, to_unicode def eval_url(url): if callable(url): url = url() return url def test_hasher(hasher): hasher.encrypt('test', rounds=hasher.min_rounds) def to36(number): assert int(number) >= 0, 'Must be a positive integer' alphabet = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ' base36 = '' while number: number, i = divmod(number, 36) base36 = alphabet[i] + base36 return base36 or alphabet[0] def from36(snumber): snumber = snumber.upper() return int(snumber, 36) def get_hash_extract(hash): if not hash: return u'' half = hash.rsplit('$', 1)[0] return half[-10:] def get_token(user, secret, timestamp=None): """Make a timestamped one-time-use token that can be used to identifying the user. By hashing the `last_sign_in` attribute and a snippet of the current password hash salt, it produces a token that will be invalidated as soon as the user log in again or the is changed. A hash of the user ID is used, so the HMAC part of the token is always unique for each user. It also hash a secret key, so without access to the source code, fake tokens cannot be generated even if the database is compromised. """ timestamp = int(timestamp or time()) secret = to_bytes(secret) key = '|'.join([ hashlib.sha1(secret).hexdigest(), str(user.id), get_hash_extract(user.password), str(getattr(user, 'last_sign_in', 0)), str(timestamp), ]) key = key.encode('utf8', 'ignore') mac = hmac.new(key, msg=None, digestmod=hashlib.sha512) mac = mac.hexdigest()[:50] token = '{0}${1}${2}'.format(user.id, to36(timestamp), mac) return token def split_uhmac(uhmac): uid, mac = uhmac.split('$', 1) return uid def split_token(token): uid, t36, mac = token.split('$', 2) return from36(t36), uid class LazyUser(object): """Acts as a proxy for the current user. Forwards all operations to the proxied user. The only operations not supported for forwarding are right handed operands and any kind of assignment. """ __slots__ = ('__auth', '__storage', '__dict__') def __init__(self, auth, storage, user_name='user'): object.__setattr__(self, '_LazyUser__auth', auth) object.__setattr__(self, '_LazyUser__storage', storage) object.__setattr__(self, '_LazyUser__user_name', user_name) setattr(storage, user_name, self) def __get_user(self): """Return the real user object. """ storage = object.__getattribute__(self, '_LazyUser__storage') user = getattr(self.__auth, 'get_user')() setattr(storage, self.__user_name, user) return user @property def __dict__(self): return self.__get_user().__dict__ @property def __doc__(self): return self.__get_user().__doc__ def __repr__(self): return repr(self.__get_user()) def __bool__(self): user = self.__get_user() return user is not None __nonzero__ = __bool__ def __str__(self): return str(self.__get_user()) def __unicode__(self): return to_unicode(self.__get_user()) def __dir__(self): return dir(self.__get_user()) def __getattr__(self, name): return getattr(self.__get_user(), name) def __setattr__(self, name, value): setattr(self.__get_user(), name, value) def __delattr__(self, name): delattr(self.__get_user(), name) def __hash__(self): return hash(self.__get_user()) def __call__(self, *args, **kwargs): return self.__get_user()(*args, **kwargs) def __eq__(self, other): return self.__get_user() == other def __ne__(self, other): return self.__get_user() != other def __setitem__(self, key, value): self.__get_user()[key] = value def __delitem__(self, key): del self.__get_user()[key] def __lt__(self, other): return self.__get_user() < other def __le__(self, other): return self.__get_user() <= other def __gt__(self, other): return self.__get_user() > other def __ge__(self, other): return self.__get_user() >= other def __getitem__(self, i): return self.__get_user()[i] def __iter__(self): return iter(self.__get_user()) def __contains__(self, i): return i in self.__get_user() def __add__(self, other): return self.__get_user() + other def __sub__(self, other): return self.__get_user() - other def __mul__(self, other): return self.__get_user() * other def __floordiv__(self, other): return self.__get_user() // other def __mod__(self, other): return self.__get_user() % other def __divmod__(self, other): return self.__get_user().__divmod__(other) def __pow__(self, other): return self.__get_user() ** other def __lshift__(self, other): return self.__get_user() << other def __rshift__(self, other): return self.__get_user() >> other def __and__(self, other): return self.__get_user() & other def __xor__(self, other): return self.__get_user() ^ other def __or__(self, other): return self.__get_user() | other def __div__(self, other): return self.__get_user().__div__(other) def __truediv__(self, other): return self.__get_user().__truediv__(other) def __neg__(self): return -(self.__get_user()) def __pos__(self): return +(self.__get_user()) def __abs__(self): return abs(self.__get_user()) def __len__(self): return len(self.__get_user()) def __invert__(self): return ~(self.__get_user()) def __complex__(self): return complex(self.__get_user()) def __int__(self): return int(self.__get_user()) def __float__(self): return float(self.__get_user()) def __enter__(self): return self.__get_user().__enter__() def __exit__(self, *args, **kwargs): return self.__get_user().__exit__(*args, **kwargs)
jpscaletti/authcode
authcode/utils.py
get_token
python
def get_token(user, secret, timestamp=None): timestamp = int(timestamp or time()) secret = to_bytes(secret) key = '|'.join([ hashlib.sha1(secret).hexdigest(), str(user.id), get_hash_extract(user.password), str(getattr(user, 'last_sign_in', 0)), str(timestamp), ]) key = key.encode('utf8', 'ignore') mac = hmac.new(key, msg=None, digestmod=hashlib.sha512) mac = mac.hexdigest()[:50] token = '{0}${1}${2}'.format(user.id, to36(timestamp), mac) return token
Make a timestamped one-time-use token that can be used to identifying the user. By hashing the `last_sign_in` attribute and a snippet of the current password hash salt, it produces a token that will be invalidated as soon as the user log in again or the is changed. A hash of the user ID is used, so the HMAC part of the token is always unique for each user. It also hash a secret key, so without access to the source code, fake tokens cannot be generated even if the database is compromised.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/utils.py#L63-L90
[ "def to_bytes(x, charset='utf8', errors='ignore'):\n if x is None:\n return None\n if isinstance(x, (bytes, bytearray, memoryview)):\n return bytes(x)\n if isinstance(x, str):\n return x.encode(charset, errors)\n raise TypeError('Expected bytes')\n", "def get_hash_extract(hash):\n if not hash:\n return u''\n half = hash.rsplit('$', 1)[0]\n return half[-10:]\n", "def to36(number):\n assert int(number) >= 0, 'Must be a positive integer'\n alphabet = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ'\n base36 = ''\n while number:\n number, i = divmod(number, 36)\n base36 = alphabet[i] + base36\n\n return base36 or alphabet[0]\n" ]
# coding=utf-8 import hashlib import hmac from time import time from ._compat import to_bytes, to_unicode def eval_url(url): if callable(url): url = url() return url def test_hasher(hasher): hasher.encrypt('test', rounds=hasher.min_rounds) def to36(number): assert int(number) >= 0, 'Must be a positive integer' alphabet = '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ' base36 = '' while number: number, i = divmod(number, 36) base36 = alphabet[i] + base36 return base36 or alphabet[0] def from36(snumber): snumber = snumber.upper() return int(snumber, 36) def get_hash_extract(hash): if not hash: return u'' half = hash.rsplit('$', 1)[0] return half[-10:] def get_uhmac(user, secret): """Make an unique identifier for the user (stored in the session), so it can stay logged between requests. By hashing a snippet of the current password hash salt, it makes possible to automatically logout from all other devices just by changing (or refreshing) the password. """ secret = to_bytes(secret) key = '|'.join([ hashlib.sha1(secret).hexdigest(), str(user.id), get_hash_extract(user.password), ]) key = key.encode('utf8', 'ignore') mac = hmac.new(key, msg=None, digestmod=hashlib.sha512) mac = mac.hexdigest()[:50] uhmac = '{0}${1}'.format(user.id, mac) return uhmac def split_uhmac(uhmac): uid, mac = uhmac.split('$', 1) return uid def split_token(token): uid, t36, mac = token.split('$', 2) return from36(t36), uid class LazyUser(object): """Acts as a proxy for the current user. Forwards all operations to the proxied user. The only operations not supported for forwarding are right handed operands and any kind of assignment. """ __slots__ = ('__auth', '__storage', '__dict__') def __init__(self, auth, storage, user_name='user'): object.__setattr__(self, '_LazyUser__auth', auth) object.__setattr__(self, '_LazyUser__storage', storage) object.__setattr__(self, '_LazyUser__user_name', user_name) setattr(storage, user_name, self) def __get_user(self): """Return the real user object. """ storage = object.__getattribute__(self, '_LazyUser__storage') user = getattr(self.__auth, 'get_user')() setattr(storage, self.__user_name, user) return user @property def __dict__(self): return self.__get_user().__dict__ @property def __doc__(self): return self.__get_user().__doc__ def __repr__(self): return repr(self.__get_user()) def __bool__(self): user = self.__get_user() return user is not None __nonzero__ = __bool__ def __str__(self): return str(self.__get_user()) def __unicode__(self): return to_unicode(self.__get_user()) def __dir__(self): return dir(self.__get_user()) def __getattr__(self, name): return getattr(self.__get_user(), name) def __setattr__(self, name, value): setattr(self.__get_user(), name, value) def __delattr__(self, name): delattr(self.__get_user(), name) def __hash__(self): return hash(self.__get_user()) def __call__(self, *args, **kwargs): return self.__get_user()(*args, **kwargs) def __eq__(self, other): return self.__get_user() == other def __ne__(self, other): return self.__get_user() != other def __setitem__(self, key, value): self.__get_user()[key] = value def __delitem__(self, key): del self.__get_user()[key] def __lt__(self, other): return self.__get_user() < other def __le__(self, other): return self.__get_user() <= other def __gt__(self, other): return self.__get_user() > other def __ge__(self, other): return self.__get_user() >= other def __getitem__(self, i): return self.__get_user()[i] def __iter__(self): return iter(self.__get_user()) def __contains__(self, i): return i in self.__get_user() def __add__(self, other): return self.__get_user() + other def __sub__(self, other): return self.__get_user() - other def __mul__(self, other): return self.__get_user() * other def __floordiv__(self, other): return self.__get_user() // other def __mod__(self, other): return self.__get_user() % other def __divmod__(self, other): return self.__get_user().__divmod__(other) def __pow__(self, other): return self.__get_user() ** other def __lshift__(self, other): return self.__get_user() << other def __rshift__(self, other): return self.__get_user() >> other def __and__(self, other): return self.__get_user() & other def __xor__(self, other): return self.__get_user() ^ other def __or__(self, other): return self.__get_user() | other def __div__(self, other): return self.__get_user().__div__(other) def __truediv__(self, other): return self.__get_user().__truediv__(other) def __neg__(self): return -(self.__get_user()) def __pos__(self): return +(self.__get_user()) def __abs__(self): return abs(self.__get_user()) def __len__(self): return len(self.__get_user()) def __invert__(self): return ~(self.__get_user()) def __complex__(self): return complex(self.__get_user()) def __int__(self): return int(self.__get_user()) def __float__(self): return float(self.__get_user()) def __enter__(self): return self.__get_user().__enter__() def __exit__(self, *args, **kwargs): return self.__get_user().__exit__(*args, **kwargs)
jpscaletti/authcode
authcode/utils.py
LazyUser.__get_user
python
def __get_user(self): storage = object.__getattribute__(self, '_LazyUser__storage') user = getattr(self.__auth, 'get_user')() setattr(storage, self.__user_name, user) return user
Return the real user object.
train
https://github.com/jpscaletti/authcode/blob/91529b6d0caec07d1452758d937e1e0745826139/authcode/utils.py#L116-L122
null
class LazyUser(object): """Acts as a proxy for the current user. Forwards all operations to the proxied user. The only operations not supported for forwarding are right handed operands and any kind of assignment. """ __slots__ = ('__auth', '__storage', '__dict__') def __init__(self, auth, storage, user_name='user'): object.__setattr__(self, '_LazyUser__auth', auth) object.__setattr__(self, '_LazyUser__storage', storage) object.__setattr__(self, '_LazyUser__user_name', user_name) setattr(storage, user_name, self) @property def __dict__(self): return self.__get_user().__dict__ @property def __doc__(self): return self.__get_user().__doc__ def __repr__(self): return repr(self.__get_user()) def __bool__(self): user = self.__get_user() return user is not None __nonzero__ = __bool__ def __str__(self): return str(self.__get_user()) def __unicode__(self): return to_unicode(self.__get_user()) def __dir__(self): return dir(self.__get_user()) def __getattr__(self, name): return getattr(self.__get_user(), name) def __setattr__(self, name, value): setattr(self.__get_user(), name, value) def __delattr__(self, name): delattr(self.__get_user(), name) def __hash__(self): return hash(self.__get_user()) def __call__(self, *args, **kwargs): return self.__get_user()(*args, **kwargs) def __eq__(self, other): return self.__get_user() == other def __ne__(self, other): return self.__get_user() != other def __setitem__(self, key, value): self.__get_user()[key] = value def __delitem__(self, key): del self.__get_user()[key] def __lt__(self, other): return self.__get_user() < other def __le__(self, other): return self.__get_user() <= other def __gt__(self, other): return self.__get_user() > other def __ge__(self, other): return self.__get_user() >= other def __getitem__(self, i): return self.__get_user()[i] def __iter__(self): return iter(self.__get_user()) def __contains__(self, i): return i in self.__get_user() def __add__(self, other): return self.__get_user() + other def __sub__(self, other): return self.__get_user() - other def __mul__(self, other): return self.__get_user() * other def __floordiv__(self, other): return self.__get_user() // other def __mod__(self, other): return self.__get_user() % other def __divmod__(self, other): return self.__get_user().__divmod__(other) def __pow__(self, other): return self.__get_user() ** other def __lshift__(self, other): return self.__get_user() << other def __rshift__(self, other): return self.__get_user() >> other def __and__(self, other): return self.__get_user() & other def __xor__(self, other): return self.__get_user() ^ other def __or__(self, other): return self.__get_user() | other def __div__(self, other): return self.__get_user().__div__(other) def __truediv__(self, other): return self.__get_user().__truediv__(other) def __neg__(self): return -(self.__get_user()) def __pos__(self): return +(self.__get_user()) def __abs__(self): return abs(self.__get_user()) def __len__(self): return len(self.__get_user()) def __invert__(self): return ~(self.__get_user()) def __complex__(self): return complex(self.__get_user()) def __int__(self): return int(self.__get_user()) def __float__(self): return float(self.__get_user()) def __enter__(self): return self.__get_user().__enter__() def __exit__(self, *args, **kwargs): return self.__get_user().__exit__(*args, **kwargs)
bearyinnovative/bearychat.py
bearychat/openapi/client.py
Requester.request
python
def request(self, request_method, api_method, *args, **kwargs): url = self._build_url(api_method) resp = requests.request(request_method, url, *args, **kwargs) try: rv = resp.json() except ValueError: raise RequestFailedError(resp, 'not a json body') if not resp.ok: raise RequestFailedError(resp, rv.get('error')) return rv
Perform a request. Args: request_method: HTTP method for this request. api_method: API method name for this request. *args: Extra arguments to pass to the request. **kwargs: Extra keyword arguments to pass to the request. Returns: A dict contains the request response data. Raises: RequestFailedError: Raises when BearyChat's OpenAPI responses with status code != 2xx
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/openapi/client.py#L82-L109
[ "def _build_url(self, api_method):\n path = '{}/{}'.format(self.base_url.path, api_method.lstrip('/'))\n url = self.base_url._replace(path=path)\n return urlparse.urlunparse(url).rstrip('/')\n" ]
class Requester(object): def __init__(self, base_url): self.base_url = urlparse.urlparse(base_url.rstrip('/')) def _build_url(self, api_method): path = '{}/{}'.format(self.base_url.path, api_method.lstrip('/')) url = self.base_url._replace(path=path) return urlparse.urlunparse(url).rstrip('/')
bearyinnovative/bearychat.py
bearychat/incoming.py
validate
python
def validate(data): text = data.get('text') if not isinstance(text, _string_types) or len(text) == 0: raise ValueError('text field is required and should not be empty') if 'markdown' in data and not type(data['markdown']) is bool: raise ValueError('markdown field should be bool') if 'attachments' in data: if not isinstance(data['attachments'], (list, tuple)): raise ValueError('attachments field should be list or tuple') for attachment in data['attachments']: if 'text' not in attachment and 'title' not in attachment: raise ValueError('text or title is required in attachment') return True
Validates incoming data Args: data(dict): the incoming data Returns: True if the data is valid Raises: ValueError: the data is not valid
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/incoming.py#L13-L40
null
import sys PY3 = sys.version_info[0] == 3 import requests if PY3: _string_types = str, else: _string_types = basestring, # noqa def send(url, data): """Sends an incoming message Args: url(str): the incoming hook url data(dict): the sending data Returns: requests.Response """ validate(data) return requests.post(url, json=data)
bearyinnovative/bearychat.py
bearychat/incoming.py
send
python
def send(url, data): validate(data) return requests.post(url, json=data)
Sends an incoming message Args: url(str): the incoming hook url data(dict): the sending data Returns: requests.Response
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/incoming.py#L43-L55
[ "def validate(data):\n \"\"\"Validates incoming data\n\n Args:\n data(dict): the incoming data\n\n Returns:\n True if the data is valid\n\n Raises:\n ValueError: the data is not valid\n \"\"\"\n text = data.get('text')\n if not isinstance(text, _string_types) or len(text) == 0:\n raise ValueError('text field is required and should not be empty')\n\n if 'markdown' in data and not type(data['markdown']) is bool:\n raise ValueError('markdown field should be bool')\n\n if 'attachments' in data:\n if not isinstance(data['attachments'], (list, tuple)):\n raise ValueError('attachments field should be list or tuple')\n\n for attachment in data['attachments']:\n if 'text' not in attachment and 'title' not in attachment:\n raise ValueError('text or title is required in attachment')\n\n return True\n" ]
import sys PY3 = sys.version_info[0] == 3 import requests if PY3: _string_types = str, else: _string_types = basestring, # noqa def validate(data): """Validates incoming data Args: data(dict): the incoming data Returns: True if the data is valid Raises: ValueError: the data is not valid """ text = data.get('text') if not isinstance(text, _string_types) or len(text) == 0: raise ValueError('text field is required and should not be empty') if 'markdown' in data and not type(data['markdown']) is bool: raise ValueError('markdown field should be bool') if 'attachments' in data: if not isinstance(data['attachments'], (list, tuple)): raise ValueError('attachments field should be list or tuple') for attachment in data['attachments']: if 'text' not in attachment and 'title' not in attachment: raise ValueError('text or title is required in attachment') return True
bearyinnovative/bearychat.py
examples/rtm_loop.py
RTMLoop.on_open
python
def on_open(self, ws): def keep_alive(interval): while True: time.sleep(interval) self.ping() start_new_thread(keep_alive, (self.keep_alive_interval, ))
Websocket on_open event handler
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/examples/rtm_loop.py#L50-L57
null
class RTMLoop(object): """Real Time Message loop _errors(Queue): contains error message(dict("result", "msg")), looks self._set_error() _inbox(Queue): contains RTMMessage _worker(threading.Thread): a thread for running the loop Args: ws_host(str): websocket host """ def __init__(self, ws_host): self._call_id = 0 self._inbox = Queue() self._errors = Queue() self._ws = websocket.WebSocketApp( ws_host, on_open=self.on_open, on_message=self.on_message, on_close=self.on_close, on_error=self.on_error) self._worker = threading.Thread( target=self._ws.run_forever, kwargs={'sslopt': sslopt_with_ca_certs}) def on_message(self, ws, message): """Websocket on_message event handler Saves message as RTMMessage in self._inbox """ try: data = json.loads(message) except Exception: self._set_error(message, "decode message failed") else: self._inbox.put(RTMMessage(data)) def on_error(self, ws, error): """Websocket on_error event handler Saves error message in self._errors """ self._set_error(error, "read socket failed") def on_close(self, ws): """Websocket on_close event handler""" self._set_error("closed", "websocket closed") def _set_error(self, result, msg): """Puts a error to self._errors Args: result(mix): received data msg(str): message """ self._errors.put({"result": result, "msg": msg}) def start(self, keep_alive_interval=2): """Starts the main loop Args: keep_alive_interval(int): the interval(second) of sending keep alive message """ self.keep_alive_interval = keep_alive_interval self._worker.start() def stop(self): """Stops the main loop """ self._ws.close() def ping(self): """Sends ping message """ self.send(RTMMessage({"type": RTMMessageType.Ping})) def gen_call_id(self): """Generates a call_id Returns: int: the call_id """ self._call_id += 1 return self._call_id def send(self, message): """Sends a RTMMessage Should be called after starting the loop Args: message(RTMMessage): the sending message Raises: WebSocketConnectionClosedException: if the loop is closed """ if "call_id" not in message: message["call_id"] = self.gen_call_id() self._ws.send(message.to_json()) def get_message(self, block=False, timeout=None): """Removes and returns a RTMMessage from self._inbox Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: RTMMessage if self._inbox is not empty, else None """ try: message = self._inbox.get(block=block, timeout=timeout) return message except Exception: return None def get_error(self, block=False, timeout=None): """Removes and returns an error from self._errors Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: error if inbox is not empty, else None """ try: error = self._errors.get(block=block, timeout=timeout) return error except Exception: return None
bearyinnovative/bearychat.py
examples/rtm_loop.py
RTMLoop.on_message
python
def on_message(self, ws, message): try: data = json.loads(message) except Exception: self._set_error(message, "decode message failed") else: self._inbox.put(RTMMessage(data))
Websocket on_message event handler Saves message as RTMMessage in self._inbox
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/examples/rtm_loop.py#L59-L69
[ "def _set_error(self, result, msg):\n \"\"\"Puts a error to self._errors\n\n Args:\n result(mix): received data\n msg(str): message\n \"\"\"\n self._errors.put({\"result\": result, \"msg\": msg})\n" ]
class RTMLoop(object): """Real Time Message loop _errors(Queue): contains error message(dict("result", "msg")), looks self._set_error() _inbox(Queue): contains RTMMessage _worker(threading.Thread): a thread for running the loop Args: ws_host(str): websocket host """ def __init__(self, ws_host): self._call_id = 0 self._inbox = Queue() self._errors = Queue() self._ws = websocket.WebSocketApp( ws_host, on_open=self.on_open, on_message=self.on_message, on_close=self.on_close, on_error=self.on_error) self._worker = threading.Thread( target=self._ws.run_forever, kwargs={'sslopt': sslopt_with_ca_certs}) def on_open(self, ws): """Websocket on_open event handler""" def keep_alive(interval): while True: time.sleep(interval) self.ping() start_new_thread(keep_alive, (self.keep_alive_interval, )) def on_error(self, ws, error): """Websocket on_error event handler Saves error message in self._errors """ self._set_error(error, "read socket failed") def on_close(self, ws): """Websocket on_close event handler""" self._set_error("closed", "websocket closed") def _set_error(self, result, msg): """Puts a error to self._errors Args: result(mix): received data msg(str): message """ self._errors.put({"result": result, "msg": msg}) def start(self, keep_alive_interval=2): """Starts the main loop Args: keep_alive_interval(int): the interval(second) of sending keep alive message """ self.keep_alive_interval = keep_alive_interval self._worker.start() def stop(self): """Stops the main loop """ self._ws.close() def ping(self): """Sends ping message """ self.send(RTMMessage({"type": RTMMessageType.Ping})) def gen_call_id(self): """Generates a call_id Returns: int: the call_id """ self._call_id += 1 return self._call_id def send(self, message): """Sends a RTMMessage Should be called after starting the loop Args: message(RTMMessage): the sending message Raises: WebSocketConnectionClosedException: if the loop is closed """ if "call_id" not in message: message["call_id"] = self.gen_call_id() self._ws.send(message.to_json()) def get_message(self, block=False, timeout=None): """Removes and returns a RTMMessage from self._inbox Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: RTMMessage if self._inbox is not empty, else None """ try: message = self._inbox.get(block=block, timeout=timeout) return message except Exception: return None def get_error(self, block=False, timeout=None): """Removes and returns an error from self._errors Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: error if inbox is not empty, else None """ try: error = self._errors.get(block=block, timeout=timeout) return error except Exception: return None
bearyinnovative/bearychat.py
examples/rtm_loop.py
RTMLoop.send
python
def send(self, message): if "call_id" not in message: message["call_id"] = self.gen_call_id() self._ws.send(message.to_json())
Sends a RTMMessage Should be called after starting the loop Args: message(RTMMessage): the sending message Raises: WebSocketConnectionClosedException: if the loop is closed
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/examples/rtm_loop.py#L120-L133
[ "def gen_call_id(self):\n \"\"\"Generates a call_id\n\n Returns:\n int: the call_id\n \"\"\"\n self._call_id += 1\n return self._call_id\n", "def to_json(self):\n \"\"\"Transfers current message to json\n\n Returns:\n json\n \"\"\"\n return json.dumps(self._data)\n" ]
class RTMLoop(object): """Real Time Message loop _errors(Queue): contains error message(dict("result", "msg")), looks self._set_error() _inbox(Queue): contains RTMMessage _worker(threading.Thread): a thread for running the loop Args: ws_host(str): websocket host """ def __init__(self, ws_host): self._call_id = 0 self._inbox = Queue() self._errors = Queue() self._ws = websocket.WebSocketApp( ws_host, on_open=self.on_open, on_message=self.on_message, on_close=self.on_close, on_error=self.on_error) self._worker = threading.Thread( target=self._ws.run_forever, kwargs={'sslopt': sslopt_with_ca_certs}) def on_open(self, ws): """Websocket on_open event handler""" def keep_alive(interval): while True: time.sleep(interval) self.ping() start_new_thread(keep_alive, (self.keep_alive_interval, )) def on_message(self, ws, message): """Websocket on_message event handler Saves message as RTMMessage in self._inbox """ try: data = json.loads(message) except Exception: self._set_error(message, "decode message failed") else: self._inbox.put(RTMMessage(data)) def on_error(self, ws, error): """Websocket on_error event handler Saves error message in self._errors """ self._set_error(error, "read socket failed") def on_close(self, ws): """Websocket on_close event handler""" self._set_error("closed", "websocket closed") def _set_error(self, result, msg): """Puts a error to self._errors Args: result(mix): received data msg(str): message """ self._errors.put({"result": result, "msg": msg}) def start(self, keep_alive_interval=2): """Starts the main loop Args: keep_alive_interval(int): the interval(second) of sending keep alive message """ self.keep_alive_interval = keep_alive_interval self._worker.start() def stop(self): """Stops the main loop """ self._ws.close() def ping(self): """Sends ping message """ self.send(RTMMessage({"type": RTMMessageType.Ping})) def gen_call_id(self): """Generates a call_id Returns: int: the call_id """ self._call_id += 1 return self._call_id def get_message(self, block=False, timeout=None): """Removes and returns a RTMMessage from self._inbox Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: RTMMessage if self._inbox is not empty, else None """ try: message = self._inbox.get(block=block, timeout=timeout) return message except Exception: return None def get_error(self, block=False, timeout=None): """Removes and returns an error from self._errors Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: error if inbox is not empty, else None """ try: error = self._errors.get(block=block, timeout=timeout) return error except Exception: return None
bearyinnovative/bearychat.py
examples/rtm_loop.py
RTMLoop.get_message
python
def get_message(self, block=False, timeout=None): try: message = self._inbox.get(block=block, timeout=timeout) return message except Exception: return None
Removes and returns a RTMMessage from self._inbox Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: RTMMessage if self._inbox is not empty, else None
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/examples/rtm_loop.py#L135-L150
null
class RTMLoop(object): """Real Time Message loop _errors(Queue): contains error message(dict("result", "msg")), looks self._set_error() _inbox(Queue): contains RTMMessage _worker(threading.Thread): a thread for running the loop Args: ws_host(str): websocket host """ def __init__(self, ws_host): self._call_id = 0 self._inbox = Queue() self._errors = Queue() self._ws = websocket.WebSocketApp( ws_host, on_open=self.on_open, on_message=self.on_message, on_close=self.on_close, on_error=self.on_error) self._worker = threading.Thread( target=self._ws.run_forever, kwargs={'sslopt': sslopt_with_ca_certs}) def on_open(self, ws): """Websocket on_open event handler""" def keep_alive(interval): while True: time.sleep(interval) self.ping() start_new_thread(keep_alive, (self.keep_alive_interval, )) def on_message(self, ws, message): """Websocket on_message event handler Saves message as RTMMessage in self._inbox """ try: data = json.loads(message) except Exception: self._set_error(message, "decode message failed") else: self._inbox.put(RTMMessage(data)) def on_error(self, ws, error): """Websocket on_error event handler Saves error message in self._errors """ self._set_error(error, "read socket failed") def on_close(self, ws): """Websocket on_close event handler""" self._set_error("closed", "websocket closed") def _set_error(self, result, msg): """Puts a error to self._errors Args: result(mix): received data msg(str): message """ self._errors.put({"result": result, "msg": msg}) def start(self, keep_alive_interval=2): """Starts the main loop Args: keep_alive_interval(int): the interval(second) of sending keep alive message """ self.keep_alive_interval = keep_alive_interval self._worker.start() def stop(self): """Stops the main loop """ self._ws.close() def ping(self): """Sends ping message """ self.send(RTMMessage({"type": RTMMessageType.Ping})) def gen_call_id(self): """Generates a call_id Returns: int: the call_id """ self._call_id += 1 return self._call_id def send(self, message): """Sends a RTMMessage Should be called after starting the loop Args: message(RTMMessage): the sending message Raises: WebSocketConnectionClosedException: if the loop is closed """ if "call_id" not in message: message["call_id"] = self.gen_call_id() self._ws.send(message.to_json()) def get_error(self, block=False, timeout=None): """Removes and returns an error from self._errors Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: error if inbox is not empty, else None """ try: error = self._errors.get(block=block, timeout=timeout) return error except Exception: return None
bearyinnovative/bearychat.py
examples/rtm_loop.py
RTMLoop.get_error
python
def get_error(self, block=False, timeout=None): try: error = self._errors.get(block=block, timeout=timeout) return error except Exception: return None
Removes and returns an error from self._errors Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: error if inbox is not empty, else None
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/examples/rtm_loop.py#L152-L167
null
class RTMLoop(object): """Real Time Message loop _errors(Queue): contains error message(dict("result", "msg")), looks self._set_error() _inbox(Queue): contains RTMMessage _worker(threading.Thread): a thread for running the loop Args: ws_host(str): websocket host """ def __init__(self, ws_host): self._call_id = 0 self._inbox = Queue() self._errors = Queue() self._ws = websocket.WebSocketApp( ws_host, on_open=self.on_open, on_message=self.on_message, on_close=self.on_close, on_error=self.on_error) self._worker = threading.Thread( target=self._ws.run_forever, kwargs={'sslopt': sslopt_with_ca_certs}) def on_open(self, ws): """Websocket on_open event handler""" def keep_alive(interval): while True: time.sleep(interval) self.ping() start_new_thread(keep_alive, (self.keep_alive_interval, )) def on_message(self, ws, message): """Websocket on_message event handler Saves message as RTMMessage in self._inbox """ try: data = json.loads(message) except Exception: self._set_error(message, "decode message failed") else: self._inbox.put(RTMMessage(data)) def on_error(self, ws, error): """Websocket on_error event handler Saves error message in self._errors """ self._set_error(error, "read socket failed") def on_close(self, ws): """Websocket on_close event handler""" self._set_error("closed", "websocket closed") def _set_error(self, result, msg): """Puts a error to self._errors Args: result(mix): received data msg(str): message """ self._errors.put({"result": result, "msg": msg}) def start(self, keep_alive_interval=2): """Starts the main loop Args: keep_alive_interval(int): the interval(second) of sending keep alive message """ self.keep_alive_interval = keep_alive_interval self._worker.start() def stop(self): """Stops the main loop """ self._ws.close() def ping(self): """Sends ping message """ self.send(RTMMessage({"type": RTMMessageType.Ping})) def gen_call_id(self): """Generates a call_id Returns: int: the call_id """ self._call_id += 1 return self._call_id def send(self, message): """Sends a RTMMessage Should be called after starting the loop Args: message(RTMMessage): the sending message Raises: WebSocketConnectionClosedException: if the loop is closed """ if "call_id" not in message: message["call_id"] = self.gen_call_id() self._ws.send(message.to_json()) def get_message(self, block=False, timeout=None): """Removes and returns a RTMMessage from self._inbox Args: block(bool): if True block until a RTMMessage is available, else it will return None when self._inbox is empty timeout(int): it blocks at most timeout seconds Returns: RTMMessage if self._inbox is not empty, else None """ try: message = self._inbox.get(block=block, timeout=timeout) return message except Exception: return None
bearyinnovative/bearychat.py
bearychat/rtm_client_service.py
RTMCurrentTeam.members
python
def members(self): resp = self._rtm_client.get('v1/current_team.members?all=true') if resp.is_fail(): raise RTMServiceError( 'Failed to get members of current team', resp ) return resp.data['result']
Gets members of current team Returns: list of User Throws: RTMServiceError when request failed
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client_service.py#L33-L48
null
class RTMCurrentTeam(RTMService): def info(self): """Gets current team infomation Returns: Team if success Throws: RTMServiceError when request failed """ resp = self._rtm_client.get('v1/current_team.info') if resp.is_fail(): raise RTMServiceError( 'Failed to get current team infomation', resp ) return resp.data['result'] def channels(self): """Gets channels of current team Returns: list of Channel Throws: RTMServiceError when request failed """ resp = self._rtm_client.get('v1/current_team.channels') if resp.is_fail(): raise RTMServiceError( 'Failed to get channels of current team', resp ) return resp.data['result']
bearyinnovative/bearychat.py
bearychat/rtm_client_service.py
RTMCurrentTeam.channels
python
def channels(self): resp = self._rtm_client.get('v1/current_team.channels') if resp.is_fail(): raise RTMServiceError( 'Failed to get channels of current team', resp ) return resp.data['result']
Gets channels of current team Returns: list of Channel Throws: RTMServiceError when request failed
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client_service.py#L50-L65
null
class RTMCurrentTeam(RTMService): def info(self): """Gets current team infomation Returns: Team if success Throws: RTMServiceError when request failed """ resp = self._rtm_client.get('v1/current_team.info') if resp.is_fail(): raise RTMServiceError( 'Failed to get current team infomation', resp ) return resp.data['result'] def members(self): """Gets members of current team Returns: list of User Throws: RTMServiceError when request failed """ resp = self._rtm_client.get('v1/current_team.members?all=true') if resp.is_fail(): raise RTMServiceError( 'Failed to get members of current team', resp ) return resp.data['result']
bearyinnovative/bearychat.py
bearychat/rtm_client_service.py
RTMUser.info
python
def info(self, user_id): resp = self._rtm_client.get('v1/user.info?user_id={}'.format(user_id)) if resp.is_fail(): raise RTMServiceError('Failed to get user information', resp) return resp.data['result']
Gets user information by user id Args: user_id(int): the id of user Returns: User Throws: RTMServiceError when request failed
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client_service.py#L69-L85
null
class RTMUser(RTMService):
bearyinnovative/bearychat.py
bearychat/rtm_client_service.py
RTMChannel.info
python
def info(self, channel_id): resource = 'v1/channel.info?channel_id={}'.format(channel_id) resp = self._rtm_client.get(resource) if resp.is_fail(): raise RTMServiceError("Failed to get channel information", resp) return resp.data['result']
Gets channel information by channel id Args: channel_id(int): the id of channel Returns: Channel Throws: RTMServiceError when request failed
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client_service.py#L89-L106
null
class RTMChannel(RTMService):
bearyinnovative/bearychat.py
bearychat/rtm_message.py
RTMMessage.reply
python
def reply(self, text): data = {'text': text, 'vchannel_id': self['vchannel_id']} if self.is_p2p(): data['type'] = RTMMessageType.P2PMessage data['to_uid'] = self['uid'] else: data['type'] = RTMMessageType.ChannelMessage data['channel_id'] = self['channel_id'] return RTMMessage(data)
Replys a text message Args: text(str): message content Returns: RTMMessage
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_message.py#L53-L69
[ "def is_p2p(self):\n \"\"\"\n Returns:\n True if current message is p2p message\n \"\"\"\n return self['type'] in (RTMMessageType.P2PMessage,\n RTMMessageType.P2PTyping)\n" ]
class RTMMessage(object): """RTM Message Args: data(dict): message dict """ def __init__(self, data): self._data = data self.mention_user_ids = set() self.parse_mention_user_ids() def __setitem__(self, name, value): self._data[name] = value def __getitem__(self, name): return self._data[name] def __contains__(self, name): return name in self._data def parse_mention_user_ids(self): if not self.is_chat_message(): return if len(self['text']) == 0: return self.mention_user_ids = set(mention_regex.findall(self['text'])) def refer(self, text): """Refers current message and replys a new message Args: text(str): message content Returns: RTMMessage """ data = self.reply(text) data['refer_key'] = self['key'] return data def is_p2p(self): """ Returns: True if current message is p2p message """ return self['type'] in (RTMMessageType.P2PMessage, RTMMessageType.P2PTyping) def is_chat_message(self): """ Returns: True if current message is chatting message """ return self['type'] in (RTMMessageType.P2PMessage, RTMMessageType.ChannelMessage) def is_mention_user(self, user): """Check if current message mentions user Args: user(User) Returns: True if message mentions user """ return user.get('id') in self.mention_user_ids def is_from(self, user): """Checks if current message is sent by user Args: user(User) Returns: True if current message is sent by the user """ return self['uid'] == user.get('id') def to_json(self): """Transfers current message to json Returns: json """ return json.dumps(self._data)
bearyinnovative/bearychat.py
bearychat/rtm_message.py
RTMMessage.refer
python
def refer(self, text): data = self.reply(text) data['refer_key'] = self['key'] return data
Refers current message and replys a new message Args: text(str): message content Returns: RTMMessage
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_message.py#L71-L82
[ "def reply(self, text):\n \"\"\"Replys a text message\n\n Args:\n text(str): message content\n\n Returns:\n RTMMessage\n \"\"\"\n data = {'text': text, 'vchannel_id': self['vchannel_id']}\n if self.is_p2p():\n data['type'] = RTMMessageType.P2PMessage\n data['to_uid'] = self['uid']\n else:\n data['type'] = RTMMessageType.ChannelMessage\n data['channel_id'] = self['channel_id']\n return RTMMessage(data)\n" ]
class RTMMessage(object): """RTM Message Args: data(dict): message dict """ def __init__(self, data): self._data = data self.mention_user_ids = set() self.parse_mention_user_ids() def __setitem__(self, name, value): self._data[name] = value def __getitem__(self, name): return self._data[name] def __contains__(self, name): return name in self._data def parse_mention_user_ids(self): if not self.is_chat_message(): return if len(self['text']) == 0: return self.mention_user_ids = set(mention_regex.findall(self['text'])) def reply(self, text): """Replys a text message Args: text(str): message content Returns: RTMMessage """ data = {'text': text, 'vchannel_id': self['vchannel_id']} if self.is_p2p(): data['type'] = RTMMessageType.P2PMessage data['to_uid'] = self['uid'] else: data['type'] = RTMMessageType.ChannelMessage data['channel_id'] = self['channel_id'] return RTMMessage(data) def is_p2p(self): """ Returns: True if current message is p2p message """ return self['type'] in (RTMMessageType.P2PMessage, RTMMessageType.P2PTyping) def is_chat_message(self): """ Returns: True if current message is chatting message """ return self['type'] in (RTMMessageType.P2PMessage, RTMMessageType.ChannelMessage) def is_mention_user(self, user): """Check if current message mentions user Args: user(User) Returns: True if message mentions user """ return user.get('id') in self.mention_user_ids def is_from(self, user): """Checks if current message is sent by user Args: user(User) Returns: True if current message is sent by the user """ return self['uid'] == user.get('id') def to_json(self): """Transfers current message to json Returns: json """ return json.dumps(self._data)
bearyinnovative/bearychat.py
bearychat/rtm_client.py
RTMClient.start
python
def start(self): resp = self.post('start') if resp.is_fail(): return None if 'result' not in resp.data: return None result = resp.data['result'] return { 'user': result['user'], 'ws_host': result['ws_host'], }
Gets the rtm ws_host and user information Returns: None if request failed, else a dict containing "user"(User) and "ws_host"
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client.py#L52-L70
[ "def is_fail(self):\n return not self.is_ok()\n", "def post(self, resource, data=None, json=None):\n \"\"\"Sends a POST request\n\n Returns:\n RTMResponse\n \"\"\"\n return self.do(resource, 'POST', data=data, json=json)\n" ]
class RTMClient(object): """Real Time Message client Attributes: current_team(RTMCurrentTeam): service of current team user(RTMUser): service of current user channel(RTMChannel): service of current channel """ def __init__(self, token, api_base="https://rtm.bearychat.com"): """ Args: token(str): rtm token api_base(str): api url base """ self._token = token self._api_base = api_base self.current_team = RTMCurrentTeam(self) self.user = RTMUser(self) self.channel = RTMChannel(self) def do(self, resource, method, params=None, data=None, json=None, headers=None): """Does the request job Args: resource(str): resource uri(relative path) method(str): HTTP method params(dict): uri queries data(dict): HTTP body(form) json(dict): HTTP body(json) headers(dict): HTTP headers Returns: RTMResponse """ uri = "{0}/{1}".format(self._api_base, resource) if not params: params = {} params.update({'token': self._token}) req = Request( method=method, url=uri, params=params, headers=headers, data=data, json=json) s = Session() prepped = s.prepare_request(req) resp = s.send(prepped) return RTMResponse(resp) def get(self, resource, params=None, headers=None): """Sends a GET request Returns: RTMResponse """ return self.do(resource, 'GET', params=params, headers=headers) def post(self, resource, data=None, json=None): """Sends a POST request Returns: RTMResponse """ return self.do(resource, 'POST', data=data, json=json)
bearyinnovative/bearychat.py
bearychat/rtm_client.py
RTMClient.do
python
def do(self, resource, method, params=None, data=None, json=None, headers=None): uri = "{0}/{1}".format(self._api_base, resource) if not params: params = {} params.update({'token': self._token}) req = Request( method=method, url=uri, params=params, headers=headers, data=data, json=json) s = Session() prepped = s.prepare_request(req) resp = s.send(prepped) return RTMResponse(resp)
Does the request job Args: resource(str): resource uri(relative path) method(str): HTTP method params(dict): uri queries data(dict): HTTP body(form) json(dict): HTTP body(json) headers(dict): HTTP headers Returns: RTMResponse
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client.py#L72-L108
null
class RTMClient(object): """Real Time Message client Attributes: current_team(RTMCurrentTeam): service of current team user(RTMUser): service of current user channel(RTMChannel): service of current channel """ def __init__(self, token, api_base="https://rtm.bearychat.com"): """ Args: token(str): rtm token api_base(str): api url base """ self._token = token self._api_base = api_base self.current_team = RTMCurrentTeam(self) self.user = RTMUser(self) self.channel = RTMChannel(self) def start(self): """Gets the rtm ws_host and user information Returns: None if request failed, else a dict containing "user"(User) and "ws_host" """ resp = self.post('start') if resp.is_fail(): return None if 'result' not in resp.data: return None result = resp.data['result'] return { 'user': result['user'], 'ws_host': result['ws_host'], } def get(self, resource, params=None, headers=None): """Sends a GET request Returns: RTMResponse """ return self.do(resource, 'GET', params=params, headers=headers) def post(self, resource, data=None, json=None): """Sends a POST request Returns: RTMResponse """ return self.do(resource, 'POST', data=data, json=json)
bearyinnovative/bearychat.py
bearychat/rtm_client.py
RTMClient.get
python
def get(self, resource, params=None, headers=None): return self.do(resource, 'GET', params=params, headers=headers)
Sends a GET request Returns: RTMResponse
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client.py#L110-L116
[ "def do(self,\n resource,\n method,\n params=None,\n data=None,\n json=None,\n headers=None):\n \"\"\"Does the request job\n\n Args:\n resource(str): resource uri(relative path)\n method(str): HTTP method\n params(dict): uri queries\n data(dict): HTTP body(form)\n json(dict): HTTP body(json)\n headers(dict): HTTP headers\n\n Returns:\n RTMResponse\n \"\"\"\n uri = \"{0}/{1}\".format(self._api_base, resource)\n if not params:\n params = {}\n params.update({'token': self._token})\n\n req = Request(\n method=method,\n url=uri,\n params=params,\n headers=headers,\n data=data,\n json=json)\n s = Session()\n prepped = s.prepare_request(req)\n resp = s.send(prepped)\n\n return RTMResponse(resp)\n" ]
class RTMClient(object): """Real Time Message client Attributes: current_team(RTMCurrentTeam): service of current team user(RTMUser): service of current user channel(RTMChannel): service of current channel """ def __init__(self, token, api_base="https://rtm.bearychat.com"): """ Args: token(str): rtm token api_base(str): api url base """ self._token = token self._api_base = api_base self.current_team = RTMCurrentTeam(self) self.user = RTMUser(self) self.channel = RTMChannel(self) def start(self): """Gets the rtm ws_host and user information Returns: None if request failed, else a dict containing "user"(User) and "ws_host" """ resp = self.post('start') if resp.is_fail(): return None if 'result' not in resp.data: return None result = resp.data['result'] return { 'user': result['user'], 'ws_host': result['ws_host'], } def do(self, resource, method, params=None, data=None, json=None, headers=None): """Does the request job Args: resource(str): resource uri(relative path) method(str): HTTP method params(dict): uri queries data(dict): HTTP body(form) json(dict): HTTP body(json) headers(dict): HTTP headers Returns: RTMResponse """ uri = "{0}/{1}".format(self._api_base, resource) if not params: params = {} params.update({'token': self._token}) req = Request( method=method, url=uri, params=params, headers=headers, data=data, json=json) s = Session() prepped = s.prepare_request(req) resp = s.send(prepped) return RTMResponse(resp) def post(self, resource, data=None, json=None): """Sends a POST request Returns: RTMResponse """ return self.do(resource, 'POST', data=data, json=json)
bearyinnovative/bearychat.py
bearychat/rtm_client.py
RTMClient.post
python
def post(self, resource, data=None, json=None): return self.do(resource, 'POST', data=data, json=json)
Sends a POST request Returns: RTMResponse
train
https://github.com/bearyinnovative/bearychat.py/blob/6c7af2d215c2ff7135bb5af66ca333d0ea1089fd/bearychat/rtm_client.py#L118-L124
[ "def do(self,\n resource,\n method,\n params=None,\n data=None,\n json=None,\n headers=None):\n \"\"\"Does the request job\n\n Args:\n resource(str): resource uri(relative path)\n method(str): HTTP method\n params(dict): uri queries\n data(dict): HTTP body(form)\n json(dict): HTTP body(json)\n headers(dict): HTTP headers\n\n Returns:\n RTMResponse\n \"\"\"\n uri = \"{0}/{1}\".format(self._api_base, resource)\n if not params:\n params = {}\n params.update({'token': self._token})\n\n req = Request(\n method=method,\n url=uri,\n params=params,\n headers=headers,\n data=data,\n json=json)\n s = Session()\n prepped = s.prepare_request(req)\n resp = s.send(prepped)\n\n return RTMResponse(resp)\n" ]
class RTMClient(object): """Real Time Message client Attributes: current_team(RTMCurrentTeam): service of current team user(RTMUser): service of current user channel(RTMChannel): service of current channel """ def __init__(self, token, api_base="https://rtm.bearychat.com"): """ Args: token(str): rtm token api_base(str): api url base """ self._token = token self._api_base = api_base self.current_team = RTMCurrentTeam(self) self.user = RTMUser(self) self.channel = RTMChannel(self) def start(self): """Gets the rtm ws_host and user information Returns: None if request failed, else a dict containing "user"(User) and "ws_host" """ resp = self.post('start') if resp.is_fail(): return None if 'result' not in resp.data: return None result = resp.data['result'] return { 'user': result['user'], 'ws_host': result['ws_host'], } def do(self, resource, method, params=None, data=None, json=None, headers=None): """Does the request job Args: resource(str): resource uri(relative path) method(str): HTTP method params(dict): uri queries data(dict): HTTP body(form) json(dict): HTTP body(json) headers(dict): HTTP headers Returns: RTMResponse """ uri = "{0}/{1}".format(self._api_base, resource) if not params: params = {} params.update({'token': self._token}) req = Request( method=method, url=uri, params=params, headers=headers, data=data, json=json) s = Session() prepped = s.prepare_request(req) resp = s.send(prepped) return RTMResponse(resp) def get(self, resource, params=None, headers=None): """Sends a GET request Returns: RTMResponse """ return self.do(resource, 'GET', params=params, headers=headers)
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/base_request.py
BaseRequest.set_logger
python
def set_logger(self): """Method to build the base logging system. By default, logging level is set to INFO.""" logger = logging.getLogger(__name__) logger.setLevel(level=logging.INFO) logger_file = os.path.join(self.logs_path, 'dingtalk_sdk.logs') logger_handler = logging.FileHandler(logger_file) logger_handler.setLevel(logging.INFO) logger_formatter = logging.Formatter( '[%(asctime)s | %(name)s | %(levelname)s] %(message)s' ) logger_handler.setFormatter(logger_formatter) logger.addHandler(logger_handler) return logger
Method to build the base logging system. By default, logging level is set to INFO.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/base_request.py#L22-L35
null
class BaseRequest(object): """The base request for dingtalk""" logs_path = os.path.dirname(os.path.abspath(__file__)) request_url = None request_methods_valid = [ "get", "post", "put", "delete", "head", "options", "patch" ] def __init__(self, **kwargs): self.kwargs = kwargs self.logger = self.set_logger() self.response = None self.json_response = None self.call_status = False self._request_method = "get" def set_logger(self): """Method to build the base logging system. By default, logging level is set to INFO.""" logger = logging.getLogger(__name__) logger.setLevel(level=logging.INFO) logger_file = os.path.join(self.logs_path, 'dingtalk_sdk.logs') logger_handler = logging.FileHandler(logger_file) logger_handler.setLevel(logging.INFO) logger_formatter = logging.Formatter( '[%(asctime)s | %(name)s | %(levelname)s] %(message)s' ) logger_handler.setFormatter(logger_formatter) logger.addHandler(logger_handler) return logger @property def request_method(self): """Mostly, the get method is used to request wanted json data, as a result, the property of request_method is set to get by default.""" return self._request_method @request_method.setter def request_method(self, method_str): request_method_lower = method_str.lower() if request_method_lower in self.request_methods_valid: self._request_method = request_method_lower else: raise ValueError( "%s is not a valid HTTP request method, please choose one" "of %s to perform a normal http request, correct it now." "" % (method_str, ",".join(self.request_methods_valid)) ) def get_response(self): """Get the original response of requests""" request = getattr(requests, self.request_method, None) if request is None and self._request_method is None: raise ValueError("A effective http request method must be set") if self.request_url is None: raise ValueError( "Fatal error occurred, the class property \"request_url\" is" "set to None, reset it with an effective url of dingtalk api." ) response = request(self.request_url, **self.kwargs) self.response = response return response def get_json_response(self): """This method aims at catching the exception of ValueError, detail: http://docs.python-requests.org/zh_CN/latest/user/quickstart.html#json """ self.json_response = self.get_response().json() if self.json_response is not None: error_code = self.json_response.get("errcode", None) self.call_status = True if error_code == 0 else False return self.json_response def get_call_status(self): """The global status of api calling.""" return self.call_status
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/base_request.py
BaseRequest.get_response
python
def get_response(self): """Get the original response of requests""" request = getattr(requests, self.request_method, None) if request is None and self._request_method is None: raise ValueError("A effective http request method must be set") if self.request_url is None: raise ValueError( "Fatal error occurred, the class property \"request_url\" is" "set to None, reset it with an effective url of dingtalk api." ) response = request(self.request_url, **self.kwargs) self.response = response return response
Get the original response of requests
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/base_request.py#L55-L67
null
class BaseRequest(object): """The base request for dingtalk""" logs_path = os.path.dirname(os.path.abspath(__file__)) request_url = None request_methods_valid = [ "get", "post", "put", "delete", "head", "options", "patch" ] def __init__(self, **kwargs): self.kwargs = kwargs self.logger = self.set_logger() self.response = None self.json_response = None self.call_status = False self._request_method = "get" def set_logger(self): """Method to build the base logging system. By default, logging level is set to INFO.""" logger = logging.getLogger(__name__) logger.setLevel(level=logging.INFO) logger_file = os.path.join(self.logs_path, 'dingtalk_sdk.logs') logger_handler = logging.FileHandler(logger_file) logger_handler.setLevel(logging.INFO) logger_formatter = logging.Formatter( '[%(asctime)s | %(name)s | %(levelname)s] %(message)s' ) logger_handler.setFormatter(logger_formatter) logger.addHandler(logger_handler) return logger @property def request_method(self): """Mostly, the get method is used to request wanted json data, as a result, the property of request_method is set to get by default.""" return self._request_method @request_method.setter def request_method(self, method_str): request_method_lower = method_str.lower() if request_method_lower in self.request_methods_valid: self._request_method = request_method_lower else: raise ValueError( "%s is not a valid HTTP request method, please choose one" "of %s to perform a normal http request, correct it now." "" % (method_str, ",".join(self.request_methods_valid)) ) def get_response(self): """Get the original response of requests""" request = getattr(requests, self.request_method, None) if request is None and self._request_method is None: raise ValueError("A effective http request method must be set") if self.request_url is None: raise ValueError( "Fatal error occurred, the class property \"request_url\" is" "set to None, reset it with an effective url of dingtalk api." ) response = request(self.request_url, **self.kwargs) self.response = response return response def get_json_response(self): """This method aims at catching the exception of ValueError, detail: http://docs.python-requests.org/zh_CN/latest/user/quickstart.html#json """ self.json_response = self.get_response().json() if self.json_response is not None: error_code = self.json_response.get("errcode", None) self.call_status = True if error_code == 0 else False return self.json_response def get_call_status(self): """The global status of api calling.""" return self.call_status
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/base_request.py
BaseRequest.get_json_response
python
def get_json_response(self): """This method aims at catching the exception of ValueError, detail: http://docs.python-requests.org/zh_CN/latest/user/quickstart.html#json """ self.json_response = self.get_response().json() if self.json_response is not None: error_code = self.json_response.get("errcode", None) self.call_status = True if error_code == 0 else False return self.json_response
This method aims at catching the exception of ValueError, detail: http://docs.python-requests.org/zh_CN/latest/user/quickstart.html#json
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/base_request.py#L69-L77
[ "def get_response(self):\n \"\"\"Get the original response of requests\"\"\"\n request = getattr(requests, self.request_method, None)\n if request is None and self._request_method is None:\n raise ValueError(\"A effective http request method must be set\")\n if self.request_url is None:\n raise ValueError(\n \"Fatal error occurred, the class property \\\"request_url\\\" is\"\n \"set to None, reset it with an effective url of dingtalk api.\"\n )\n response = request(self.request_url, **self.kwargs)\n self.response = response\n return response\n" ]
class BaseRequest(object): """The base request for dingtalk""" logs_path = os.path.dirname(os.path.abspath(__file__)) request_url = None request_methods_valid = [ "get", "post", "put", "delete", "head", "options", "patch" ] def __init__(self, **kwargs): self.kwargs = kwargs self.logger = self.set_logger() self.response = None self.json_response = None self.call_status = False self._request_method = "get" def set_logger(self): """Method to build the base logging system. By default, logging level is set to INFO.""" logger = logging.getLogger(__name__) logger.setLevel(level=logging.INFO) logger_file = os.path.join(self.logs_path, 'dingtalk_sdk.logs') logger_handler = logging.FileHandler(logger_file) logger_handler.setLevel(logging.INFO) logger_formatter = logging.Formatter( '[%(asctime)s | %(name)s | %(levelname)s] %(message)s' ) logger_handler.setFormatter(logger_formatter) logger.addHandler(logger_handler) return logger @property def request_method(self): """Mostly, the get method is used to request wanted json data, as a result, the property of request_method is set to get by default.""" return self._request_method @request_method.setter def request_method(self, method_str): request_method_lower = method_str.lower() if request_method_lower in self.request_methods_valid: self._request_method = request_method_lower else: raise ValueError( "%s is not a valid HTTP request method, please choose one" "of %s to perform a normal http request, correct it now." "" % (method_str, ",".join(self.request_methods_valid)) ) def get_response(self): """Get the original response of requests""" request = getattr(requests, self.request_method, None) if request is None and self._request_method is None: raise ValueError("A effective http request method must be set") if self.request_url is None: raise ValueError( "Fatal error occurred, the class property \"request_url\" is" "set to None, reset it with an effective url of dingtalk api." ) response = request(self.request_url, **self.kwargs) self.response = response return response def get_json_response(self): """This method aims at catching the exception of ValueError, detail: http://docs.python-requests.org/zh_CN/latest/user/quickstart.html#json """ self.json_response = self.get_response().json() if self.json_response is not None: error_code = self.json_response.get("errcode", None) self.call_status = True if error_code == 0 else False return self.json_response def get_call_status(self): """The global status of api calling.""" return self.call_status
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/authority_request.py
PersistentCodeRequest.get_ticket_for_sns_token
python
def get_ticket_for_sns_token(self): """This is a shortcut for getting the sns_token, as a post data of request body.""" self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return { "openid": self.get_openid(), "persistent_code": self.get_persistent_code(), }
This is a shortcut for getting the sns_token, as a post data of request body.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/authority_request.py#L86-L93
[ "def get_openid(self):\n openid = self.json_response.get(\"openid\", None)\n self.logger.info(\"%s\\t%s\" % (self.request_method, self.request_url))\n return openid\n", "def get_persistent_code(self):\n persistent_code = self.json_response.get(\"persistent_code\", None)\n self.logger.info(\"%s\\t%s\" % (self.request_method, self.request_url))\n return persistent_code\n" ]
class PersistentCodeRequest(BaseRequest): """ Description: The persistent_code, openid and unionid are contained in the json response of this api, the parameter <access_token> and post data <tmp_auth_code> is required. parameter_R: <access_token> parameter_O: None post_data_R: <tmp_auth_code> post_data_O: None Return: the persistent_code, openid and unionid of dingtalk api. doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/athq8o """ request_url = settings.GET_PERSISTENT_CODE def get_openid(self): openid = self.json_response.get("openid", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return openid def get_unionid(self): unionid = self.json_response.get("unionid", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return unionid def get_persistent_code(self): persistent_code = self.json_response.get("persistent_code", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return persistent_code def get_ticket_for_sns_token(self): """This is a shortcut for getting the sns_token, as a post data of request body.""" self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return { "openid": self.get_openid(), "persistent_code": self.get_persistent_code(), }
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/message_request.py
WorkNoticeRequest.get_task_id
python
def get_task_id(self): """Method to get all department members.""" task_id = self.json_response.get("task_id", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return task_id
Method to get all department members.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/message_request.py#L24-L28
null
class WorkNoticeRequest(BaseRequest): """ Description: The WorkNoticeRequest send work notice to specified user (userid), parameters of <userid_list> and <dept_id_list> should NOT keep null simultaneously, or request would fail parameter_R: <access_token> parameter_O: None post_data_R: <agent_id>, <msg> post_data_O: <userid_list>, <dept_id_list>, <to_all_user> Return: send result json response doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/pgoxpy """ request_url = settings.SEND_WORK_NOTICE task_id = None def get_task_id(self): """Method to get all department members.""" task_id = self.json_response.get("task_id", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return task_id
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/message_request.py
GetWorkNoticeSendProgressRequest.get_progress
python
def get_progress(self): """Method to get the progress of work notice sending.""" progress = self.json_response.get("progress", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return progress
Method to get the progress of work notice sending.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/message_request.py#L48-L52
null
class GetWorkNoticeSendProgressRequest(BaseRequest): """ Description: The response of GetWorkNoticeSendProgressRequest shows the progress of sending work notice parameter_R: <access_token> parameter_O: None post_data_R: <agent_id>, <msg> post_data_O: <userid_list>, <dept_id_list>, <to_all_user> Return: the progress of sending work notice doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/pgoxpy """ request_url = settings.GET_WORK_NOTICE_SEND_PROGRESS def get_progress(self): """Method to get the progress of work notice sending.""" progress = self.json_response.get("progress", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return progress
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/message_request.py
GetWorkNoticeSendResultRequest.get_send_result
python
def get_send_result(self): """Method to get the progress of work notice sending.""" send_result = self.json_response.get("send_result", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return send_result
Method to get the progress of work notice sending.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/message_request.py#L72-L76
null
class GetWorkNoticeSendResultRequest(BaseRequest): """ Description: The response of GetWorkNoticeSendResultRequest shows the result of sending work notice parameter_R: <access_token> parameter_O: None post_data_R: None post_data_O: <agent_id>, <task_id> Return: send result json response doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/pgoxpy """ request_url = settings.GET_WORK_NOTICE_SEND_RESULT def get_send_result(self): """Method to get the progress of work notice sending.""" send_result = self.json_response.get("send_result", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return send_result
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/message_request.py
CreateGroupChatRequest.get_chat_id
python
def get_chat_id(self): """Method to get chatid of group created.""" chat_id = self.json_response.get("chatid", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return chat_id
Method to get chatid of group created.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/message_request.py#L96-L100
null
class CreateGroupChatRequest(BaseRequest): """ Description: The CreateGroupChatRequest aims to create a group chat for some user(userid) parameter_R: <access_token> parameter_O: None post_data_R: <name>, <owner>, <useridlist> post_data_O: <showHistoryType> Return: the chatid of group create json response doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/isu6nk """ request_url = settings.CREATE_GROUP_CHAT def get_chat_id(self): """Method to get chatid of group created.""" chat_id = self.json_response.get("chatid", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return chat_id
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/message_request.py
GetGroupChatRequest.get_chat_info
python
def get_chat_info(self): """Method to get chatid of group created.""" chat_info = self.json_response.get("chat_info", None) self.chat_info = chat_info self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return chat_info
Method to get chatid of group created.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/message_request.py#L139-L144
null
class GetGroupChatRequest(BaseRequest): """ Description: The GetGroupChatRequest aims to get information of a existed group chat by chatid parameter_R: <access_token> parameter_O: None post_data_R: <chatid> post_data_O: None Return: the chatid of group create json response doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/isu6nk """ request_url = settings.UPDATE_GROUP_CHAT chat_info = None def get_chat_info(self): """Method to get chatid of group created.""" chat_info = self.json_response.get("chat_info", None) self.chat_info = chat_info self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return chat_info def get_specified_group_user_ids(self): user_ids = self.chat_info.get("useridlist", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return user_ids
gmdzy2010/dingtalk_sdk_gmdzy2010
dingtalk_sdk_gmdzy2010/message_request.py
SendGroupChatRequest.get_message_id
python
def get_message_id(self): """Method to get messageId of group created.""" message_id = self.json_response.get("messageId", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return message_id
Method to get messageId of group created.
train
https://github.com/gmdzy2010/dingtalk_sdk_gmdzy2010/blob/b06cb1f78f89be9554dcb6101af8bc72718a9ecd/dingtalk_sdk_gmdzy2010/message_request.py#L168-L172
null
class SendGroupChatRequest(BaseRequest): """ Description: The SendGroupChatRequest aims to send group chat message parameter_R: <access_token> parameter_O: None post_data_R: <chatid>, <msg> post_data_O: None Return: the message id of group chat doc_links: https://open-doc.dingtalk.com/microapp/serverapi2/isu6nk """ request_url = settings.SEND_GROUP_CHAT def get_message_id(self): """Method to get messageId of group created.""" message_id = self.json_response.get("messageId", None) self.logger.info("%s\t%s" % (self.request_method, self.request_url)) return message_id