Datasets:
thisisanshgupta
/
Pycode

Dataset card Data Studio Files
xet
 Community
text
stringlengths
0
828
return wrapped"
393,"def reducer(*, name = None, metric = call_default):
""""""Decorator to measure a function that consumes many items.
The wrapped ``func`` should take either a single ``iterable`` argument or
``*args`` (plus keyword arguments).
:arg function metric: f(name, count, total_time)
:arg str name: name for the metric
""""""
class instrument_reducer_decorator(object):
def __init__(self, func):
self.orig_func = func
self.wrapping = wraps(func)
self.metric_name = name if name is not None else func.__module__ + '.' +func.__name__
self.varargs = inspect.getargspec(func).varargs is not None
if self.varargs:
self.method = _varargs_to_iterable_method(func)
self.func = _varargs_to_iterable_func(func)
self.callme = _iterable_to_varargs_func(self._call)
else:
self.method = func
self.func = func
self.callme = self._call
# we need _call/callme b/c CPython short-circurits CALL_FUNCTION to
# directly access __call__, bypassing our varargs decorator
def __call__(self, *args, **kwargs):
return self.callme(*args, **kwargs)
def _call(self, iterable, **kwargs):
it = counted_iterable(iterable)
t = time.time()
try:
return self.func(it, **kwargs)
finally:
metric(self.metric_name, it.count, time.time() - t)
def __get__(self, instance, class_):
metric_name = name if name is not None else\
""."".join((class_.__module__, class_.__name__, self.orig_func.__name__))
def wrapped_method(iterable, **kwargs):
it = counted_iterable(iterable)
t = time.time()
try:
return self.method(instance, it, **kwargs)
finally:
metric(metric_name, it.count, time.time() - t)
# wrap in func version b/c self is handled for us by descriptor (ie, `instance`)
if self.varargs: wrapped_method = _iterable_to_varargs_func(wrapped_method)
wrapped_method = self.wrapping(wrapped_method)
return wrapped_method
return instrument_reducer_decorator"
394,"def producer(*, name = None, metric = call_default):
""""""Decorator to measure a function that produces many items.
The function should return an object that supports ``__len__`` (ie, a
list). If the function returns an iterator, use :func:`all` instead.
:arg function metric: f(name, count, total_time)
:arg str name: name for the metric
""""""
def wrapper(func):
def instrumenter(name_, *args, **kwargs):
t = time.time()
try:
ret = func(*args, **kwargs)
except Exception:
# record a metric for other exceptions, than raise
metric(name_, 0, time.time() - t)
raise
else:
# normal path, record metric & return
metric(name_, len(ret), time.time() - t)
return ret
name_ = name if name is not None else func.__module__ + '.' +func.__name__
class instrument_decorator(object): # must be a class for descriptor magic to work
@wraps(func)
def __call__(self, *args, **kwargs):
return instrumenter(name_, *args, **kwargs)
def __get__(self, instance, class_):
name_ = name if name is not None else\
""."".join((class_.__module__, class_.__name__, func.__name__))
@wraps(func)
def wrapped_method(*args, **kwargs):
return instrumenter(name_, instance, *args, **kwargs)
return wrapped_method
return instrument_decorator()
return wrapper"
395,"def block(*, name = None, metric = call_default, count = 1):
""""""Context manager to measure execution time of a block
:arg function metric: f(name, 1, time)