code stringlengths 1 1.72M | language stringclasses 1
value |
|---|---|
import logging
# Change default log level to debug in dev_appserver to match production
logging.getLogger().setLevel(logging.DEBUG)
# but leave the default for App Engine APIs
logging.getLogger('google.appengine').setLevel(logging.INFO)
| Python |
#!/usr/bin/env python2.7
import json
import logging
import os
import pprint as pp
import sys
import unittest
import urllib2
try:
from webtest import TestApp
except:
print """Please install webtest from http://webtest.pythonpaste.org/"""
sys.exit(1)
# Attempt to locate the App Engine SDK based on the system PATH
for d in sorted(os.environ['PATH'].split(os.path.pathsep)):
path = os.path.join(d, 'dev_appserver.py')
if not os.path.isfile(path):
continue
print 'Found the App Engine SDK directory: %s' % d
sys.path.insert(0, d)
# The App Engine SDK root is now expected to be in sys.path (possibly provided via PYTHONPATH)
try:
import dev_appserver
except ImportError:
error_msg = ('The path to the App Engine Python SDK must be in the '
'PYTHONPATH environment variable to run unittests.')
# The app engine SDK isn't in sys.path. If we're on Windows, we can try to
# guess where it is.
import platform
if platform.system() == 'Windows':
sys.path = sys.path + ['C:\\Program Files\\Google\\google_appengine']
try:
import dev_appserver # pylint: disable-msg=C6204
except ImportError:
print error_msg
raise
else:
print error_msg
raise
# add App Engine libraries
sys.path += dev_appserver.EXTRA_PATHS
from google.appengine.ext import testbed
from google.appengine.api import backends
def make_state(serverid, extras=None):
"""Create test game server state."""
# something pseudo predictable / stable
uptime = hash(serverid) / 1e16
result = { 'serverid': serverid, 'uptime': uptime }
if extras:
result.update(extras)
return result
def make_game(serverid, name, extras=None):
"""Create test game instance."""
result = {
'serverid': serverid,
'name': name,
'game_state': {'players': {}, 'min_players': 2, 'max_players': 4},
}
if extras:
result.update(extras)
return result
class MatchMakerTest(unittest.TestCase):
def setUp(self):
# see testbed docs https://developers.google.com/appengine/docs/python/tools/localunittesting
self.testbed = testbed.Testbed()
self.testbed.activate()
self.testbed.init_memcache_stub()
self.testbed.init_datastore_v3_stub()
self.testbed.init_urlfetch_stub()
# import matcher now that testbed has been activated
from client import matcher
# instatiate a new matcher
self.mm = matcher.MatchMaker()
# no maximum length for diffs
self.maxDiff = None
def tearDown(self):
self.testbed.deactivate()
def assertGameServers(self, expected_game_servers):
"""Assert the global game servers state."""
if expected_game_servers != self.mm._game_servers:
self.fail('Unexpected game servers state:\n----------\nEXPECTED:\n%s\n\nACTUAL:\n%s\n----------' % (expected_game_servers, self.mm._game_servers))
#self.assertEquals(expected_game_servers, self.mm._game_servers)
def assertPlayerGameState(self, expected_game, player_game):
"""Assert a specifc game state."""
self.assertIn('game', player_game)
self.assertIn('player_game_key', player_game)
self.assertEquals(expected_game, player_game['game'])
def assertGameServerWait(self, players_needed_for_next_game, pg):
"""Assert that the player is being asked to wait."""
self.assertEquals({'result': 'wait', 'players_needed_for_next_game': players_needed_for_next_game}, pg)
#######################################################################
# tests to verify server and game updates
#######################################################################
def test_initialized_state(self):
self.assertGameServers({})
def test_update_server_info(self):
self.assertGameServers({})
state_foo = make_state('foo')
expected = { 'foo':
{ 'server_info': state_foo,
'games': {},
}
}
self.mm.update_server_info(state_foo)
self.assertGameServers(expected)
def test_update_server_info_twice(self):
state_foo = make_state('foo')
state_bar = make_state('bar')
expected = { 'foo':
{ 'server_info': state_foo,
'games': {},
},
'bar':
{ 'server_info': state_bar,
'games': {},
},
}
self.assertGameServers({})
self.mm.update_server_info(state_foo)
self.mm.update_server_info(state_bar)
self.assertGameServers(expected)
def test_update_server_info_thrice(self):
state_foo1 = make_state('foo', {'update': 17} )
state_foo2 = make_state('foo', {'update': 42} )
state_bar = make_state('bar')
expected = { 'foo':
{ 'server_info': state_foo2,
'games': {},
},
'bar':
{ 'server_info': state_bar,
'games': {},
},
}
self.assertGameServers({})
self.mm.update_server_info(state_foo1)
self.mm.update_server_info(state_bar)
self.mm.update_server_info(state_foo2)
self.assertGameServers(expected)
def test_del_server_info(self):
state_foo = make_state('foo')
state_bar = make_state('bar')
expected = { 'bar':
{ 'server_info': state_bar,
'games': {},
},
}
self.assertGameServers({})
self.mm.update_server_info(state_foo)
self.mm.update_server_info(state_bar)
self.mm.del_game_server('foo')
self.assertGameServers(expected)
def test_update_game(self):
state_foo = make_state('foo')
game_abcd = make_game('foo', 'abcd')
expected = { 'foo':
{ 'server_info': state_foo,
'games': { 'abcd' : game_abcd,
}
}
}
self.assertGameServers({})
self.mm.update_server_info(state_foo)
self.mm.update_game(game_abcd)
self.assertGameServers(expected)
#######################################################################
# tests to verify match making
#######################################################################
def test_do_not_find_full_game(self):
state_foo = make_state('foo')
game_abcd = make_game('foo', 'abcd', {'game_state': {'players': {}, 'min_players': 0, 'max_players': 0}})
self.mm.update_server_info(state_foo)
# we have a game, but no players are needed
self.mm.update_game(game_abcd)
# do not find full game
pg = self.mm.find_player_game('fred')
self.assertGameServerWait(-1, pg)
def test_find_single_player_game(self):
state_foo = make_state('foo')
game_abcd = make_game('foo', 'abcd', {'game_state': {'players': {}, 'min_players': 1, 'max_players': 1}})
self.mm.update_server_info(state_foo)
# no available game yet
pg = self.mm.find_player_game('fred')
self.assertGameServerWait(-1, pg)
# game_abcd becomes available
self.mm.update_game(game_abcd)
# we should find it
pg = self.mm.find_player_game('fred')
self.assertPlayerGameState(game_abcd, pg)
# we should find it again
pg = self.mm.lookup_player_game('fred')
self.assertPlayerGameState(game_abcd, pg)
def test_wait_for_enough_players(self):
state_foo = make_state('foo')
# only 3 game slots remaining
game_abcd = make_game('foo', 'abcd', {'game_state': {'players': {}, 'min_players': 1, 'max_players': 1}})
game_efgh = make_game('foo', 'efgh', {'game_state': {'players': {}, 'min_players': 2, 'max_players': 2}})
self.mm.update_server_info(state_foo)
self.mm.update_game(game_abcd)
self.mm.update_game(game_efgh)
# player1 enters game_abcd
pg = self.mm.find_player_game('player1')
self.assertPlayerGameState(game_abcd, pg)
# player2 waits for a second player
pg = self.mm.find_player_game('player2')
self.assertGameServerWait(2, pg)
# player3 enters game_efgh
pg = self.mm.find_player_game('player3')
self.assertPlayerGameState(game_efgh, pg)
# player4 does not get in
pg = self.mm.find_player_game('player4')
self.assertGameServerWait(-1, pg)
# player2 finally enter game_efgh
pg = self.mm.lookup_player_game('player2')
self.assertPlayerGameState(game_efgh, pg)
def test_honor_max_players(self):
state_foo = make_state('foo')
# only 3 game slots remaining
game_abcd = make_game('foo', 'abcd', {'game_state': {'players': {}, 'min_players': 2, 'max_players': 4}})
self.mm.update_server_info(state_foo)
# player1 is told that a new game needs to be spun up
pg = self.mm.find_player_game('player1')
self.assertGameServerWait(-1, pg)
# a new game is made available
self.mm.update_game(game_abcd)
# player1 is told 2 players are required
pg = self.mm.find_player_game('player1')
self.assertGameServerWait(2, pg)
# player2 allows the game to start
pg = self.mm.find_player_game('player2')
self.assertPlayerGameState(game_abcd, pg)
# player3 drops in
pg = self.mm.find_player_game('player3')
self.assertPlayerGameState(game_abcd, pg)
# player4 drops in
pg = self.mm.find_player_game('player4')
self.assertPlayerGameState(game_abcd, pg)
# player5 does not get in
pg = self.mm.find_player_game('player5')
self.assertGameServerWait(-1, pg)
if __name__ == '__main__':
unittest.main()
| Python |
import os, zipfile
def zip(target, source, env):
f = zipfile.ZipFile(str(target[0]), 'w', zipfile.ZIP_DEFLATED)
chdir = None
try:
chdir = env['ZIPCHDIR'] + os.sep
except:
pass
for s in source:
s = str(s)
if chdir and s.find(chdir) == 0:
arcname = s[len(chdir):]
else:
arcname = s
f.write(s, arcname)
f.close()
zipbld = Builder(action = zip)
Export('zipbld')
objs = []
for subdir in ['components', 'chrome']:
r = SConscript(['%s/SConscript' % subdir])
objs.extend(r)
env = Environment()
r = env.Install('dist', 'chrome.manifest')
objs.append(r)
r = env.Install('dist', 'install.rdf')
objs.append(r)
r = env.Install('dist', 'License.txt')
objs.append(r)
r = env.Install('dist/defaults/preferences', 'defaults/preferences/prefs.js')
objs.append(r)
zipenv = Environment(BUILDERS = {'Zip' : zipbld}, ZIPCHDIR = 'dist')
zipenv.Zip('firetray.xpi', objs)
Alias('xpi', 'firetray.xpi')
| Python |
import sys, os, string, re
platform = sys.platform
try:
FLAGS = os.environ['CXXFLAGS'];
except:
FLAGS = '-O2';
linking_options=' --libs ';
try:
DYNAMIC_LINKING = os.environ['DYNAMIC_LINKING'];
except:
linking_options=' --static '
print "linking: " + linking_options
FLAGS += ' -include xpcom-config.h -include mozilla-config.h'
FLAGS += ' -fno-rtti -fno-exceptions -fshort-wchar'
FLAGS += ' -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wno-long-long'
FLAGS += ' -pedantic -pthread -pipe'
try:
lib_arch = os.environ['LIB_ARCH'];
except:
lib_arch = '';
try:
gecko_bin = os.environ['GECKO_SDK_BIN'];
if gecko_bin[-1] != os.sep: gecko_bin += os.sep
gecko_include = os.environ['GECKO_SDK_INCLUDE'];
if gecko_include[-1] != os.sep: gecko_include += os.sep
gecko_idl = os.environ['GECKO_SDK_IDL'];
if gecko_idl[-1] != os.sep: gecko_idl += os.sep
gecko_lib = os.environ['GECKO_SDK_LIB'];
if gecko_lib[-1] != os.sep: gecko_lib += os.sep
except:
try:
geckosdk = os.environ['GECKO_SDK']
print "Using GECKO_SDK=" + geckosdk
if geckosdk[-1] != os.sep: geckosdk += os.sep
gecko_bin = geckosdk + 'bin'
gecko_include = geckosdk + 'include'
gecko_idl = geckosdk + 'idl'
gecko_lib = geckosdk + 'lib'
except:
print "Please set environment variable GECKO_SDK first (or in alternative the variables GECKO_SDK_BIN, GECKO_SDK_INCLUDE, GECKO_SDK_IDL, GECKO_SDK_LIB)."
sys.exit(1)
# Hack to detect Mozilla version
version_re = re.compile('#define MOZILLA_VERSION "(.*?)"')
xpcom_libs = ['xpcomglue_s']
with open(gecko_include + os.sep + 'mozilla-config.h', 'r') as f:
for line in f:
version_match = version_re.match(line)
if version_match:
version = string.split(version_match.group(1), '.')
if int(version[0]) >= 2:
# OK, we're building with Mozilla 2.0
FLAGS += ' -DGECKO_2'
if 'GECKO_19_COMPAT' in os.environ:
FLAGS += ' -DMOZ_NO_MOZALLOC'
xpcom_libs = ['xpcomglue_s_nomozalloc']
else:
xpcom_libs.append('mozalloc')
break
# Create two builders to create xpt and header files from idl.
bxpt = Builder(action = 'xpidl -w -m typelib -Icomponents -I' + gecko_idl + ' -e $TARGET $SOURCE', suffix = '.xpt', src_suffix = '.idl')
bhdr = Builder(action = 'xpidl -w -m header -Icomponents -I' + gecko_idl + ' -e $TARGET $SOURCE', suffix = '.h', src_suffix = '.idl')
# Create environment object for build
env = Environment(
CPPPATH = [gecko_include],
LIBPATH = [gecko_lib],
LIBS = xpcom_libs,
ENV = os.environ)
env.AppendENVPath('PATH', gecko_bin)
env.Append(BUILDERS = {'MozXPT' : bxpt, 'MozHeader' : bhdr })
env.ParseConfig('pkg-config ' + linking_options + ' --cflags gtk+-2.0') # libnotify
env.ParseConfig('pkg-config --cflags nspr')
env.Append( CXXFLAGS = FLAGS )
# Create headers and xpt files from idl
xpts = [env.MozXPT('nsITray'), env.MozXPT('nsIFireTrayHandler')]
headers = [env.MozHeader('nsITray')]
parts = []
parts.extend(['nsTray.cpp', 'nsTrayModule.cpp'])
nptray_name = 'nptray' + lib_arch
print "NPTRAY: " + nptray_name
nptray = env.SharedLibrary(nptray_name, parts)
r = env.Install('#dist/components', [nptray, xpts, 'nsFireTrayHandler.js'])
Default([xpts, headers, nptray])
Return('r')
| Python |
import os
Import('zipbld')
contents = []
content = 'core.js browserOverlay.xul mailOverlay.xul songOverlay.xul navigatorOverlay.xul options.xul options.js icon.png'
for c in content.split(' '):
contents.append('content' + os.sep + c)
locales = 'bg-BG ca-AD de-DE en-US es-AR es-ES fa-IR fr-FR it-IT ja-JP mk-MK nl-NL pl-PL pt-BR ru-RU sk-SK sl-SI sv-SE tr-TR uk-UA zh-CN zh-TW'
locale_files = 'browserOverlay.dtd core.properties mailOverlay.dtd options.dtd'
for l in locales.split(' '):
for f in locale_files.split(' '):
contents.append('locale' + os.sep + l + os.sep + f)
env = Environment(ZIPCHDIR= 'chrome', BUILDERS = {'Zip' : zipbld})
r = env.Zip('#dist/chrome/firetray.jar', contents)
Return('r')
| Python |
import os, zipfile
def zip(target, source, env):
f = zipfile.ZipFile(str(target[0]), 'w', zipfile.ZIP_DEFLATED)
chdir = None
try:
chdir = env['ZIPCHDIR'] + os.sep
except:
pass
for s in source:
s = str(s)
if chdir and s.find(chdir) == 0:
arcname = s[len(chdir):]
else:
arcname = s
f.write(s, arcname)
f.close()
zipbld = Builder(action = zip)
Export('zipbld')
objs = []
for subdir in ['components', 'chrome']:
r = SConscript(['%s/SConscript' % subdir])
objs.extend(r)
env = Environment()
r = env.Install('dist', 'chrome.manifest')
objs.append(r)
r = env.Install('dist', 'install.rdf')
objs.append(r)
r = env.Install('dist', 'License.txt')
objs.append(r)
r = env.Install('dist/defaults/preferences', 'defaults/preferences/prefs.js')
objs.append(r)
zipenv = Environment(BUILDERS = {'Zip' : zipbld}, ZIPCHDIR = 'dist')
zipenv.Zip('firetray.xpi', objs)
Alias('xpi', 'firetray.xpi')
| Python |
import sys, os, string, re
platform = sys.platform
try:
FLAGS = os.environ['CXXFLAGS'];
except:
FLAGS = '-O2';
linking_options=' --libs ';
try:
DYNAMIC_LINKING = os.environ['DYNAMIC_LINKING'];
except:
linking_options=' --static '
print "linking: " + linking_options
FLAGS += ' -include xpcom-config.h -include mozilla-config.h'
FLAGS += ' -fno-rtti -fno-exceptions -fshort-wchar'
FLAGS += ' -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-non-virtual-dtor -Wno-long-long'
FLAGS += ' -pedantic -pthread -pipe'
try:
lib_arch = os.environ['LIB_ARCH'];
except:
lib_arch = '';
try:
gecko_bin = os.environ['GECKO_SDK_BIN'];
if gecko_bin[-1] != os.sep: gecko_bin += os.sep
gecko_include = os.environ['GECKO_SDK_INCLUDE'];
if gecko_include[-1] != os.sep: gecko_include += os.sep
gecko_idl = os.environ['GECKO_SDK_IDL'];
if gecko_idl[-1] != os.sep: gecko_idl += os.sep
gecko_lib = os.environ['GECKO_SDK_LIB'];
if gecko_lib[-1] != os.sep: gecko_lib += os.sep
except:
try:
geckosdk = os.environ['GECKO_SDK']
print "Using GECKO_SDK=" + geckosdk
if geckosdk[-1] != os.sep: geckosdk += os.sep
gecko_bin = geckosdk + 'bin'
gecko_sdk_bin = geckosdk + 'sdk/bin'
gecko_include = geckosdk + 'include'
gecko_idl = geckosdk + 'idl'
gecko_lib = geckosdk + 'lib'
except:
print "Please set environment variable GECKO_SDK first (or in alternative the variables GECKO_SDK_BIN, GECKO_SDK_INCLUDE, GECKO_SDK_IDL, GECKO_SDK_LIB)."
sys.exit(1)
# Hack to detect Mozilla version
version_re = re.compile('#define MOZILLA_VERSION "(.*?)"')
xpcom_libs = ['xpcomglue_s']
with open(gecko_include + os.sep + 'mozilla-config.h', 'r') as f:
for line in f:
version_match = version_re.match(line)
if version_match:
version = string.split(version_match.group(1), '.')
if int(version[0]) >= 2:
# OK, we're building with Mozilla 2.0
FLAGS += ' -DGECKO_2'
if 'GECKO_19_COMPAT' in os.environ:
FLAGS += ' -DMOZ_NO_MOZALLOC'
xpcom_libs = ['xpcomglue_s_nomozalloc']
else:
xpcom_libs.append('mozalloc')
break
# Create two builders to create xpt and header files from idl.
bxpt = Builder(action = 'typelib.py -Icomponents -I' + gecko_idl + ' -o $TARGET $SOURCE' + ' --cachedir="."', suffix = '.xpt', src_suffix = '.idl')
bhdr = Builder(action = 'header.py -Icomponents -I' + gecko_idl + ' -o $TARGET $SOURCE' + ' --cachedir="."', suffix = '.h', src_suffix = '.idl')
# Create environment object for build
env = Environment(
CPPPATH = [gecko_include],
LIBPATH = [gecko_lib],
LIBS = xpcom_libs,
ENV = os.environ)
env.AppendENVPath('PATH', gecko_bin)
env.AppendENVPath('PATH', gecko_sdk_bin)
env.Append(BUILDERS = {'MozXPT' : bxpt, 'MozHeader' : bhdr })
env.ParseConfig('pkg-config ' + linking_options + ' --cflags gtk+-2.0') # libnotify
env.ParseConfig('pkg-config --cflags nspr')
env.Append( CXXFLAGS = FLAGS )
# Create headers and xpt files from idl
xpts = [env.MozXPT('nsITray'), env.MozXPT('nsIFireTrayHandler')]
headers = [env.MozHeader('nsITray')]
parts = []
parts.extend(['nsTray.cpp', 'nsTrayModule.cpp'])
nptray_name = 'nptray' + lib_arch
print "NPTRAY: " + nptray_name
nptray = env.SharedLibrary(nptray_name, parts)
r = env.Install('#dist/components', [nptray, xpts, 'nsFireTrayHandler.js'])
Default([xpts, headers, nptray])
Return('r')
| Python |
import os
Import('zipbld')
contents = []
content = 'core.js browserOverlay.xul mailOverlay.xul songOverlay.xul navigatorOverlay.xul options.xul options.js icon.png'
for c in content.split(' '):
contents.append('content' + os.sep + c)
locales = 'bg-BG ca-AD de-DE en-US es-AR es-ES fa-IR fr-FR it-IT ja-JP mk-MK nl-NL pl-PL pt-BR ru-RU sk-SK sl-SI sv-SE tr-TR uk-UA zh-CN zh-TW'
locale_files = 'browserOverlay.dtd core.properties mailOverlay.dtd options.dtd'
for l in locales.split(' '):
for f in locale_files.split(' '):
contents.append('locale' + os.sep + l + os.sep + f)
env = Environment(ZIPCHDIR= 'chrome', BUILDERS = {'Zip' : zipbld})
r = env.Zip('#dist/chrome/firetray.jar', contents)
Return('r')
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
import sys,cgi,re,Image,Gnuplot,string,os
sys.path+=[os.getcwd()]
from types import *
from firesocket import *
try:
form=cgi.FieldStorage()
if form.has_key("server"):
server=form["server"].value
else:
server="localhost"
if form.has_key("port"):
port=int(form["port"].value)
else:
port=12960
if form.has_key("image"):
image=form["image"].value
else:
image="100.jpg"
if form.has_key("feature"):
feature=form["feature"].value
else:
feature="1"
s=FIRESocket()
s.connect(server,port)
s.sendcmd("feature "+image+" "+feature)
line=s.getline()
suffix=string.split(line)[1]
lines=[]
line=s.getline()
while line!="end":
lines+=[line]
line=s.getline()
s.sendcmd("bye")
if suffix.endswith(".histo") or suffix.endswith(".histo.gz") or suffix.endswith("oldhisto") or suffix.endswith("oldhisto.gz"):
for l in lines:
if l.startswith("data"):
tokens=string.split(l)
tokens.pop(0)
tokens=map(lambda t:
float(t),tokens)
bins=len(tokens)
maximum=max(tokens)
g=Gnuplot.Gnuplot()
g('set terminal png size 300,200')
g('unset xtics')
g('unset ytics')
g('set data style boxes')
print "Content-Type: image/png\n"
sys.stdout.flush()
g.plot(Gnuplot.Data(tokens))
elif suffix.endswith(".vec.gz") or suffix.endswith(".oldvec.gz") or suffix.endswith(".vec"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
if l.startswith("data"):
print l[5:]
print "</body></html>\n"
elif suffix.endswith(".textID"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body>\n</html>"
elif suffix.endswith("txt") or suffix.endswith("txt.gz"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body></html>\n"
elif suffix.endswith(".png"):
for l in lines:
if l.startswith("sizes"):
tok=string.split(l)
xsize=int(tok[1])
ysize=int(tok[2])
zsize=int(tok[3])
if l.startswith("data"):
tok=string.split(l)
tok.pop(0)
if zsize==1:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
g=int(float(tok.pop(0))*255)
i.putpixel((x,y),(g,g,g))
elif zsize==3:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
r=int(float(tok.pop(0))*255)
g=int(float(tok.pop(0))*255)
b=int(float(tok.pop(0))*255)
i.putpixel((x,y),(r,g,b))
print "Content-Type: image/png\n"
sys.stdout.flush()
i.save(sys.stdout,"png")
else:
print "Content-Type: text/html\n"
sys.stdout.flush()
print suffix
print "feature not supported: ",suffix
for l in lines:
print l+"<br>"
print "</body></html>\n"
except IOError:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Something failed"
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
import sys,cgi,re,Image,Gnuplot,string,os
sys.path+=[os.getcwd()]
from types import *
from firesocket import *
try:
form=cgi.FieldStorage()
if form.has_key("server"):
server=form["server"].value
else:
server="localhost"
if form.has_key("port"):
port=int(form["port"].value)
else:
port=12960
if form.has_key("image"):
image=form["image"].value
else:
image="100.jpg"
if form.has_key("feature"):
feature=form["feature"].value
else:
feature="1"
s=FIRESocket()
s.connect(server,port)
s.sendcmd("feature "+image+" "+feature)
line=s.getline()
suffix=string.split(line)[1]
lines=[]
line=s.getline()
while line!="end":
lines+=[line]
line=s.getline()
s.sendcmd("bye")
if suffix.endswith(".histo") or suffix.endswith(".histo.gz") or suffix.endswith("oldhisto") or suffix.endswith("oldhisto.gz"):
for l in lines:
if l.startswith("data"):
tokens=string.split(l)
tokens.pop(0)
tokens=map(lambda t:
float(t),tokens)
bins=len(tokens)
maximum=max(tokens)
g=Gnuplot.Gnuplot()
g('set terminal png size 300,200')
g('unset xtics')
g('unset ytics')
g('set data style boxes')
print "Content-Type: image/png\n"
sys.stdout.flush()
g.plot(Gnuplot.Data(tokens))
elif suffix.endswith(".vec.gz") or suffix.endswith(".oldvec.gz") or suffix.endswith(".vec"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
if l.startswith("data"):
print l[5:]
print "</body></html>\n"
elif suffix.endswith(".textID"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body>\n</html>"
elif suffix.endswith("txt") or suffix.endswith("txt.gz"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body></html>\n"
elif suffix.endswith(".png"):
for l in lines:
if l.startswith("sizes"):
tok=string.split(l)
xsize=int(tok[1])
ysize=int(tok[2])
zsize=int(tok[3])
if l.startswith("data"):
tok=string.split(l)
tok.pop(0)
if zsize==1:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
g=int(float(tok.pop(0))*255)
i.putpixel((x,y),(g,g,g))
elif zsize==3:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
r=int(float(tok.pop(0))*255)
g=int(float(tok.pop(0))*255)
b=int(float(tok.pop(0))*255)
i.putpixel((x,y),(r,g,b))
print "Content-Type: image/png\n"
sys.stdout.flush()
i.save(sys.stdout,"png")
else:
print "Content-Type: text/html\n"
sys.stdout.flush()
print suffix
print "feature not supported: ",suffix
for l in lines:
print l+"<br>"
print "</body></html>\n"
except IOError:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Something failed"
| Python |
#!/usr/bin/python
import sys
import cgi
import re
import Image
sys.stderr=sys.stdout
try:
form=cgi.FieldStorage()
imagefilename=form["image"].value
outWidth=-1
outHeight=-1
maxDim=-1
if(form.has_key("max")):
maxDim=int(form["max"].value)
if(form.has_key("width")):
outWidth=int(form["width"].value)
if(form.has_key("height")):
outHeight=int(form["height"].value)
inImg=Image.open(imagefilename)
inWidth=inImg.size[0]
inHeight=inImg.size[1]
if maxDim!=-1:
if inWidth>inHeight:
outWidth=maxDim
else:
outHeight=maxDim
if (outWidth==-1 and outHeight==-1):
outImg=inImg
elif (outWidth==-1):
scaleFactor=float(outHeight)/float(inHeight)
outWidth=int(scaleFactor*inWidth)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
elif (outHeight==-1):
scaleFactor=float(outWidth)/float(inWidth)
outHeight=int(scaleFactor*inHeight)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
else:
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
contenttype="image/jpg"
print "Content-Type: "+contenttype+"\n"
outImg=outImg.convert("RGB")
outImg.save(sys.stdout,"jpeg")
except:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Access not permitted"
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
#$Header$
# print cgi-header
print "Content-Type: text/html\n\n"
print
#print "<pre style=\"font-family:Fixedsys,Courier,1px;\">" # to have error messages as readable as possible, this
# is switched of right before displaying the main part of the page
#import lots of stuff
import cgi, re, sys, traceback, os, Image
#import cgitb; cgitb.enable()
from urllib import quote,unquote
from types import *
# redirect stderr to have error messages on the webpage and not in the
# webserver's log
sys.stderr=sys.stdout
# to be able to find the config
sys.path+=[os.getcwd()]
#import "my" socket implementation
from firesocket import *
# ----------------------------------------------------------------------
# import the config
# it is assumed there is a file called config in the same file as the cgi.
# an example for a valid config:
## settings for i6
#positiveImg="http://www-i6.iatik.rwth-aachen.de/~deselaers/images/positive.png"
#negativeImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/negativ.png"
#neutralImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/neutral.png"
#fireLogo="/u/deselaers/work/src/fire/cgi/fire-logo.png"
#i6Logo= "/u/deselaers/work/src/fire/cgi/i6.png"
#TemplateFile = "/u/deselaers/work/src/fire/cgi/fire-template.html"
#fireServer="deuterium"
#firePort=12960
#tempdir="/tmp"
#maxSize=1000, 1000
#minSize=200, 200
# ----------------------------------------------------------------------
import config
# settings for the linear (standard) scoring algorithm. using this
# algorithm the weights can be changed from the web interface
class LinearScoring:
def __init__(self):
self.size=0
self.weights=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
if(tmp=="weight"):
idx=status.pop(0)
w=status.pop(0)
self.weights+=[w]
else:
print "Cannot parse LinearScoring settings"
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for LinearScoring</th></tr>"
for i in range(len(self.weights)):
result+="<tr><td> Weight "+str(i)+"</td>"
result+="<td colspan=2><input name=\"weight"+str(i)+"\" type=\"text\" value=\""+self.weights[i]+"\"></td></tr>"
return result
# settings for maxent scoring algorithm
class MaxEntScoring:
def __init__(self):
self.size=0
self.factor=0
self.offset=0
self.lambdas=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
tmp=status.pop(0)
self.factor=status.pop(0)
tmp=status.pop(0)
self.offset=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
idx=status.pop(0)
l=status.pop(0)
self.lambdas+=[l]
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for MaxEntScoring</th></tr>"
for i in range(len(self.lambdas)/2):
result+="<tr><td> Lambda "+str(i)+"</td>"
result+="<td>"+self.lambdas[i]+"</td><td>"+self.lambdas[i+len(self.lambdas)/2]+"</td></tr>"
return result
# a class to store the settings of fire this classes uses the classes
# defined above to manage the settings of the active scoring algorithm
class FireSettings:
def __init__(self):
# variable initilization only. all member variables are defined here.
self.fireServer=config.fireServer
self.firePort=config.firePort
self.dbsize=0
self.scoringname=""
self.results=0
self.expansions=0
self.distances=[]
self.suffices=[]
self.path=""
self.scoring=""
def get(self,s):
# get the settings from the server, i.e. read the string s and parse it.
self.distances=[]
self.suffices=[]
s.sendcmd("info")
msg=s.getline()
status=re.split(" ",msg)
print "<!--"+str(status)+"-->"
while(len(status)>0):
keyword=status.pop(0)
if keyword=="filelist":
self.dbsize=status.pop(0)
elif keyword=="results":
self.results=status.pop(0)
elif keyword=="path":
self.path=status.pop(0)
elif keyword=="extensions":
self.expansions=status.pop(0)
elif keyword=="scoring":
self.scoringname=status.pop(0)
if self.scoringname=="linear":
self.scoring=LinearScoring()
elif self.scoringname=="maxent":
self.scoring=MaxEntScoring()
self.scoring.get(status)
elif keyword=="suffix":
no=status.pop(0)
suffix=status.pop(0)
self.suffices+=[suffix]
distname=status.pop(0)
self.distances+=[distname]
else:
print "<!-- Unknown keyword in stream : "+keyword+" -->"
def display(self):
# display the settings dialog
self.get(s)
result="<h4>Settings for fire</h4>"
result=result+"<form method=\"post\" name=\"settingsForm\">"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+str(self.fireServer)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(self.firePort)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"settings\" value=\"1\"/>\n"
result=result+"<table border=\"1\">\n"
result=result+"<tr><th colspan=3> General Settings</th></tr>"
result=result+"<tr><td>Number of images in database</td><td colspan=\"2\">"+str(self.dbsize)+"</td></tr>\n"
result=result+"<tr><td>Number of results shown</td><td colspan=\"2\"><input type=\"text\" value=\""+self.results+"\" name=\"results\"/></td></tr>\n"
result=result+"<tr><td>Scoring</td><td colspan=\"2\">"+self.getScoringChooser()+"\n"
result=result+"<tr><td>Number of Images for Query Expansion</td><td colspan=\"2\"><input type=\"text\" value=\""+self.expansions+"\" name=\"extensions\"/></td></tr>\n"
result=result+self.getDistForm()
result=result+self.getScoringForm()
result=result+"</table>\n"
result=result+"Password: <input type=\"password\" name=\"password\">\n"
result=result+"<input type=\"submit\" name=\"Save\" value=\"Save\"/>\n"
result=result+"</form>"
return result
def getScoringChooser(self):
# select the chooser for the server
result="<select name=\"scoring\">\n<option value=\"\">"
for scor in ['linear','maxent']:
if self.scoringname==scor:
result=result+"<option selected value=\""+scor+"\">"+scor+"</option>\n"
else:
result=result+"<option value=\""+scor+"\">"+scor+"</option>\n"
result+="</select>\n"
return result
def getScoringForm(self):
# display the form element to select the scoring algorihtm
return self.scoring.getForm()
def getDistForm(self):
result=""
result+="<tr><th colspan=3> Settings for Distance Functions</th></tr>"
for no in range(len(self.distances)):
result+="<tr><td> Distance "+str(no)+": "+self.suffices[no]+"</td><td>"+self.getDistChooser(no,self.distances[no])+"</td></tr>"
return result
def getDistChooser(self,no,distName):
result="<select name=\"dist"+str(no)+"\">\n<option value=\"\">"
s.sendcmd("setdist")
distsString=s.getline()
dists=re.split(" ",distsString)
for d in dists:
if d==distName:
result=result+"<option selected value=\""+d+"\">"+d+"</option>\n"
else:
result=result+"<option value=\""+d+"\">"+d+"</option>\n"
result=result+"</select>\n"
return result
def process(self,form):
result="<!-- processing settings -->\n"
if form.has_key("results"):
if form.has_key("password"):
password=form["password"].value
else:
password=""
s.sendcmd("password "+password)
l=s.getline()
result+="<!-- "+l+"-->"
if l!="ok":
result+="Changing settings denied: Not authorized!"
for i in form.keys():
if i.find("dist")==0:
no=i[4:]
d=form[i].value
if d!=self.distances[int(no)]: # only change when changed
s.sendcmd("setdist "+no+" "+d)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("weight")==0:
no=i[6:]
w=float(form[i].value)
if self.scoringname=="linear":
if(float(self.scoring.weights[int(no)])!=float(w)):
s.sendcmd("setweight "+no+" "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
else:
print "weights not supported"
elif i.find("results")==0:
if int(form[i].value)!=int(self.results):
s.sendcmd("setresults "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("extensions")==0:
if int(self.expansions)!=int(form[i].value):
s.sendcmd("setextensions "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("scoring")==0:
sn=form[i].value
if(sn!=self.scoringname): # only change when changed
s.sendcmd("setscoring "+sn)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
self.scoringname=sn
else:
result=result+"<!-- "+i+"-->\n"
result=result+"<!-- settings processed -->\n"
return result
# ----------------------------------------------------------------------
# Load the template file and display the content in this template file
# the string FIRELOGOSTRINGHERE, I6LOGOSTRINGHERE, and "INSERT CONTENT
# HERE" are replaced by the configured/generated strings
# ----------------------------------------------------------------------
def Display(Content):
TemplateHandle = open(config.TemplateFile, "r") # open in read only mode
# read the entire file as a string
TemplateInput = TemplateHandle.read()
TemplateHandle.close() # close the file
# this defines an exception string in case our
# template file is messed up
BadTemplateException = "There was a problem with the HTML template."
TemplateInput = re.sub("FIRELOGOSTRINGHERE",config.fireLogo,TemplateInput)
TemplateInput = re.sub("I6LOGOSTRINGHERE",config.i6Logo,TemplateInput)
SubResult = re.subn("INSERT CONTENT HERE",Content,TemplateInput)
if SubResult[1] == 0:
raise BadTemplateException
print SubResult[0]
def sendretrieve(querystring):
filterstring=""
# if(form.has_key("filter1")):
# filterstring+=" -porn2-sorted/009-www.ficken.bz-images-ficken.png"
# if(form.has_key("filter2")):
# filterstring+=" -porn2-sorted/037-www.free-porn-free-sex.net-free-porn-free-sex-free-sex-sites.png"
s.sendcmd("retrieve "+querystring+filterstring)
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by clicking a random
# image, that is:
# 1. process the form
# - extract the filename of the image which is used as query
# 2. send the query to the server
# 3. wait for the servers answer
# 4. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def retrieve(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["queryImage"].value
sendretrieve("+"+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimage,0)
return result
# ----------------------------------------------------------------------
# resizes an image while keeping aspect ratio
# ----------------------------------------------------------------------
def resizeImage(im, newSize):
newX, newY = newSize
if(im.size < newSize):
xSize, ySize = im.size
ratio=float(xSize)/float(ySize)
if(ySize < xSize):
newSize=(int(newY*ratio),newY)
else:
newSize=(newX,int(newX*(1/ratio)))
out = Image.new("RGB", newSize)
out = im.resize(newSize, Image.ANTIALIAS)
else:
im.thumbnail(newSize, Image.ANTIALIAS)
out=im
return out
# ----------------------------------------------------------------------
# verifies the image file and resizes the image according to
# minSize and maxSize parameters in config.py
# ----------------------------------------------------------------------
def checkImage(imagefile):
maxSize=config.maxSize
minSize=config.minSize
try:
im=Image.open(imagefile)
except:
return 0
if(im.size<minSize):
im=resizeImage(im, minSize)
elif(im.size>maxSize):
im=resizeImage(im, maxSize)
im.save(imagefile)
return 1
# ----------------------------------------------------------------------
# handle a serverfile-request
# ----------------------------------------------------------------------
def serverFile(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["serverfile"].value
checkImage(queryimage)
imagename=queryimage[queryimage.rfind("/")+1:len(queryimage)]
s.sendcmd("newfile 2 "+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+imagename,0)
return result
# ----------------------------------------------------------------------
# handle a newfile-request
# ----------------------------------------------------------------------
def newFile(form):
temporaryFiles=config.tempdir # standard: "/tmp"
result="<h3>Retrieval Result</h3>\n"
if(save_uploaded_file(form,"absolute_path",temporaryFiles)==0):
result=result+"No Retrieval Result available!"
return result
fileitem = form["absolute_path"]
# f=os.popen("ls -l "+ temporaryFiles+"/"+fileitem.filename)
# print f.readlines()
if(checkImage(temporaryFiles+"/"+fileitem.filename)==0):
result=result+"Delivered file was not a valid image!"
return result
s.sendcmd("newfile 2 "+temporaryFiles+"/"+fileitem.filename+".png")
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+fileitem.filename,0)
return result
# This saves a file uploaded by an HTML form.
# The form_field is the name of the file input field from the form.
# For example, the following form_field would be "file_1":
# <input name="file_1" type="file">
# The upload_dir is the directory where the file will be written.
# If no file was uploaded or if the field does not exist then
# this does nothing.
# written by Noah Spurrier, modified by Fabian Schwahn
def save_uploaded_file(form, form_field, upload_dir):
if not form.has_key(form_field): return 0
fileitem = form[form_field]
completepath = upload_dir + fileitem.filename
if os.path.isdir(completepath): return 0
if not fileitem.file: return 0
fout = open (os.path.join(upload_dir, fileitem.filename), 'wb')
while 1:
chunk = fileitem.file.read(100000)
if not chunk: break
fout.write (chunk)
fout.close()
return 1
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by relevance feedback.
# That is:
# 1. process the form
# - extract the filenames of positive and negative examples
# 2. generate the query string
# 3. send the query to the server
# 4. wait for the servers answer
# 5. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def feedbackretrieve(form):
print "<!-- feedback retrieve \n"
print form
print "-->\n"
result="<h3>Retrieval Result</h3>\n"
queryimages=""
for field in form.keys():
if field.startswith("relevance"):
imageno=re.sub("relevance","",field)
imagename=form["resultImage"+imageno].value
relevance=form[field].value
if relevance!="0":
queryimages=queryimages+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
result=result+"\n<!-- "+queryimages+" -->\n"
sendretrieve(queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimages,0)
return result
# ----------------------------------------------------------------------
# given the retrieval result string of the server process this answer
# and create the part of the html page displaying these results. this
# includes the buttons for relevance feedback and the hidden fields
# for server settings (e.g.\ which server, which port). also the
# buttons for "more results" and "relevance feedback" are generated.
# the button for saving relevances is deaktivated as it is not used
# ----------------------------------------------------------------------
def displayResults(tokens,querytype,querystring,resultsStep):
if(re.search("[0-9]+(\.[0-9]+)?", tokens[0]) == None):
result="Could not read retrieval result.\n"
s.flush()
return result
print "<!-- "+str(tokens)+"-->"
i=0
result="<form method=\"post\" name=\"resultForm\"><table>\n<tr>\n"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+querystring+"\"/>\n"
result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+quote(querystring)+"\" />\n"
result=result+"<input type=\"hidden\" name=\"resultsstep\" value=\""+str(resultsStep)+"\"/>\n"
first=True
while(len(tokens) > 0):
resNo=str(i)
img=tokens.pop(0)
score=float(tokens.pop(0))
result=result+"<td>\n"
result=result+"<a href=\""+config.fireurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&imageinfo="+img+"\" target=\"_blank\">"
# support for the select-rectangle tool
#-------------
# result=result+"<a href=\"save_coord.cgi?to_do=UpdateImage&upload_type=server&port="+str(settings.firePort)+"&image_name="+settings.path+"/"+img+"\" target=\"_top\">"
#-------------
# shows the image
#-------------
# result=result+"<a href=\"img.py?image="+settings.path+"/"+img+"\" target=\"_blank\">"
#------------
result=result+"<img src=\""+config.imgpyurl+"?max=150&image="+settings.path+"/"+img+"\" title=\""+str(score)+"-"+img+"\" name=\""+str(score)+"-"+img+"\">\n<br>\n"
result=result+"</a>"
result=result+"<input type=\"hidden\" name=\"resultImage"+resNo+"\" value=\""+img+"\">\n"
if first:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\" checked><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\"><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
first=False
else:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\"><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\" checked><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
result=result+"</td>\n\n"
i=i+1
if i%5==0:
result=result+"</tr>\n\n\n<tr>"
result=result+"</tr></table>\n"
# result+=filterbuttons()
result=result+"<button title=\"expandresults\" name=\"expandresults\" type=\"submit\" onClick=\"document.resultForm.feedback.value='expandresults';document.resultForm.submit()\">more results</button>\n"
result=result+"<button title=\"requery\" name=\"requery\" type=\"submit\" onClick=\"document.resultForm.feedback.value='requery';document.resultForm.submit()\">requery</button>\n"
# result=result+"<button title=\"saverelevances\" name=\"relevancessave\" onClick=\"document.resultForm.feedback.value='saverelevances';document.resultForm.submit()\" \"type=\"submit\">save relevances</button>\n"
result=result+"<input type=\"hidden\" name=\"feedback\" value=\"yes\"/>\n"
result=result+"</form>\n"
return result
def filterbuttons():
# display the buttons for the filter
result=""
result=result+"filter 1: <input type=\"checkbox\" name=\"filter1\" value=\"1\">\n"
result=result+"filter 2: <input type=\"checkbox\" name=\"filter2\" value=\"2\">\n<br>\n"
return result
# ----------------------------------------------------------------------
# ask the server for filenames of random images and create the part of
# the html page showing random images. Steps:
# 1. send "random" to the server
# 2. receive the servers answer
# 3. process the answer:
# - get the filenames and for each filename create a querybutton
# 4. put the "more random images" button, such that we can get a new set
# of random images
# ----------------------------------------------------------------------
def randomImages():
# display random images
s.sendcmd("random")
result="<hr><h3> Random Images - Click one to start a query </h3>"
result+="""<form method="post" name="queryForm" enctype="multipart/form-data">
<input type="hidden" name="queryImage" value="">
"""
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result+=filterbuttons()
res=s.getline()
for img in re.split(" ",res):
if img != "" and img != "\n":
result=result+"<button title=\""+img+"\" name=\"searchButton\" type=\"submit\" onClick=\"document.queryForm.queryImage.value='"+img+"';document.queryForm.submit()\" value=\""+img+"\">\n"
result=result+"<img src=\""+config.imgpyurl+"?max=100&image="+settings.path+"/"+img+"\" title=\""+img+"\">\n"
result=result+"</button>\n"
if form.has_key("demomode"):
seconds=form["demomode"].value
url="http://"
if(os.environ.has_key("HTTP_HOST") and os.environ.has_key("REQUEST_URI")):
part1=os.environ["HTTP_HOST"]
part2=re.sub("&queryImage=.*$","",os.environ["REQUEST_URI"])
url=url+part1+part2+"&queryImage="+img
print "<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result=result+"<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result+="<br><div align=\"right\">\n<button title=\"new random images\" name=\"morerandom\" type=\"submit\">more random images</button>\n</div>\n"
return result
# ----------------------------------------------------------------------
# displays the newfile-form
# ----------------------------------------------------------------------
def newFileForm():
result="<hr><h3> Upload - Use query image from your file system </h3>"
result+="""<input type="hidden" name="newFile" value="0">"""
result+="""<input name="absolute_path" type="file" size="30" onClick="document.queryForm.newFile.value=1;document.queryForm.submit()">"""
result+="\n"
return result
# ----------------------------------------------------------------------
# generate the text field and the query button for text
# queries
# ----------------------------------------------------------------------
def textQueryForm(settings):
result = "<hr><h3>Query by "
if("textfeature" in settings.distances):
result += "description"
if("metafeature" in settings.distances):
result += " or meta information"
elif("metafeature" in settings.distances):
result += "meta information"
result += "</h3>\n"
result += "<form method=\"post\" name=\"textQueryForm\">\n"
result += "<input type=\"hidden\" name=\"server\" value=\""+str(settings.fireServer)+"\"/>\n"
result += "<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
result += "<input type=\"hidden\" name=\"resultsstep\" value=\"0\"/>\n"
result += "<input type=\"text\" name=\"textquerystring\" size=\"30\" value=\"\"/>\n"
result += "<input type=\"submit\" value=\"query\"/>\n"
if("metafeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&metafeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on meta info</a>"
if("textfeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&textfeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on description</a>"
result += "</form>"
return result
def displayMetaFeatureInfo():
s.sendcmd("metafeatureinfo")
mfi = s.getline()
mfil = mfi.split(" ")
result = "<h4>Help on queries by meta info</h4>\n"
result += """If you want to find images with certain meta information attached,
type in a request of the form
<h5>key1:val1,key2:val2,...</h5>\n
The following meta keys are available in this corpus:<br><br>\n"""
#TODO: The table looks ugly!
result += "<table>\n"
result += "<tr>\n"
result += " <th>key</th>\n"
result += " <th>example value</th>\n"
result += "</tr>\n"
for mf in mfil:
mfl = mf.split(":")
result += "<tr>\n"
result += " <td><b>"+mfl[0]+"</b></td>\n" # I know <b> is deprecated, but it still works :-)
result += " <td>"+mfl[1]+"</td>\n"
result += "</tr>\n"
result += "</table>\n"
return result
def displayTextFeatureInfo():
result = "<h4>Help on queries by description</h4>\n"
result += """If you want to find images which have text information attached to them,
just enter the query into the textbox. Fire will then use an information
retrieval engine to find the images that best match your query.<br><br>
If you have text information in multiple languages for each image, you can give
a query for every language. For example, if you have german and french text
information and you want to search for "hand" in the respective languages, you enter
<h5>Ge:"hand" Fr:"mains"</h5>
Here, "Ge" has to be the suffix for the german textfiles and "Fr" has to be the suffix for the
fench textfiles. Fire will then use both queries and send them to separate information retrieval
engines."""
return result
# ----------------------------------------------------------------------
# this function makes the server save a relevances logfile
# but it is not used at the moment, because the saverelevances button
# is deaktivated
# ----------------------------------------------------------------------
def saveRelevances(form):
querystring=form["querystring"].value
relevancestring=""
for field in form.keys():
if field.startswith("relevance-"):
imagename=re.sub("relevance-","",field)
relevance=form[field].value
if relevance!="0":
relevancestring=relevancestring+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
s.sendcmd("saverelevances Q "+querystring+" R "+relevancestring)
res=s.getline();
message="<a href=\"javascript:back()\"> Go back to last query </a>"
return message
# ----------------------------------------------------------------------
# this is the "getMoreResults" function it basically does the same as
# retrieve, uses the same query string as the last query used (this is
# gotten from the form) and the calls displayResults
# ----------------------------------------------------------------------
def expandQuery(form):
resultsstep=int(form["resultsstep"].value)
resultsstep=resultsstep+1
if(form.has_key("metaquerystring")):
result="<h3>Metatag-based retrieval result</h3>\n"
queryfield = "metaquerystring"
cmd = "metaexpand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
elif(form.has_key("textquerystring")):
result="<h3>Text-based retrieval result</h3>\n"
queryfield = "textquerystring"
cmd = "textexpand"
queryimages = unquote(form["textquerystring"].value)
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
else:
result="<h3>Retrieval Result</h3>\n"
queryfield = "querystring"
cmd = "expand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
if(tokens[0] == "notextinformation"):
result=result+"No images matched your query: "+querystring
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,queryfield,queryimages,resultsstep)
return result
# ----------------------------------------------------------------------
# process a query using text- or metainformation from the text input
# field. The query type is set to metaquery when the search string has
# a leading "META "
# This works identical as the retrieve function
# ----------------------------------------------------------------------
def processTextQuery(form):
# Check if it's meta or text
# It's meta either if there's only the metafeature or it has a leading "META "
hasmeta = "metafeature" in settings.distances
hastext = "textfeature" in settings.distances
querystring = form["textquerystring"].value
if (hasmeta and (not hastext or querystring[:5] == "META ")):
if(querystring[:5] == "META "):
querystring = querystring[5:]
querytype = "metaquerystring"
result="<h3>Metatag-based retrieval result</h3>\n"
s.sendcmd("metaretrieve +"+querystring)
else:
querytype = "textquerystring"
result="<h3>Text-based retrieval result</h3>\n"
s.sendcmd("textretrieve +"+querystring)
msg=s.getline()
tokens=re.split(" ",msg)
if( (tokens[0] == "notextinformation") or (tokens[0] == "nometainformation") ):
result=result+"No images matched your query: "+querystring
if(querytype == "metaquerystring"):
if(hastext):
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'META key1:val1,key2:val2,...'<br><br>\n"
else:
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'key1:val1,key2:val2,...'<br><br>\n"
result=result+"Also have a look at the help to see which keys are available.<br>\n"
if(querytype == "textquerystring"):
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,querytype,querystring,0)
return result
# ----------------------------------------------------------------------
# create the link for the "settings" at the bottom of the page
# ----------------------------------------------------------------------
def adminLink(fireServer, firePort):
result=""
result=result+"<a name=\"#\"></a>"
result=result+"<a href=\"#\" "+\
"onClick=\"window.open('"+config.fireurl+"?server="+ \
fireServer+"&port="+str(firePort)+"&settings=1','newwindow',"+\
"'height=600, width=600')\">settings</a>"
return result
#<SCRIPT LANGUAGE="javascript">
#<!--
#window.open ('titlepage.html', 'newwindow', config='height=100,
#width=400, toolbar=no, menubar=no, scrollbars=no, resizable=no,
#location=no, directories=no, status=no')
#-->
#</SCRIPT>
# ----------------------------------------------------------------------
# the main program
# ----------------------------------------------------------------------
# get form information in easily accessible manner
form=cgi.FieldStorage()
# make socket
s = FIRESocket()
settings=FireSettings()
# see what server we have and if the webinterface specifed another one
# than the one in the config file
if form.has_key("server"):
settings.fireServer=form["server"].value
if form.has_key("port"):
settings.firePort=int(form["port"].value)
try:
# connect to the server
s.connect(settings.fireServer, settings.firePort)
except:
# if the server is down, give an error message
# print "</pre>"
message="""
<h2>FIRE Server down</h2>
<p>
Try again later, complain to
<a href="mailto:deselaers@informatik.rwth-aachen.de">deselaers@informatik.rwth-aachen.de</a>
or try other servers.
"""
Display(message)
else:
# everything is fine, connection to server is ok
# try:
message=""
settings.get(s)
# do we want to show the settings window?
if(form.has_key("settings")):
# see if there have been new settings defined
message=settings.process(form)
# generate the settings dialog
message+=settings.display()
# disonnect from server
s.sendcmd("bye")
print "</pre>"
# display the settings dialog
Display(message)
# do we want to show the metafeatureinfo window?
elif(form.has_key("metafeatureinfo")):
message+=displayMetaFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
elif(form.has_key("imageinfo")):
imagename=form["imageinfo"].value
s.sendcmd("image "+imagename)
line=""
while(line != "end"):
line=s.getline()
tokens=re.split(" ",line)
if tokens[0]=="imagefilename":
message+="<img src=\""+config.imgpyurl+"?image="+tokens[1]+"\" alt=\""+tokens[1]+"\"/><br>"+"\n"
elif tokens[0]=="class":
message+="class: "+tokens[1]+"<br>\n"
elif tokens[0]=="description":
message+="description: "+tokens[1:]+"<br>\n"
elif tokens[0]=="features":
features=[]
noffeatures=int(tokens[1])
for i in range(noffeatures):
line=s.getline()
tokens=re.split(" ",line)
message+="<h3>"+str(i)+": "+tokens[2]+"</h3>\n<br>\n"
if tokens[2].endswith(".histo.gz") \
or tokens[2].endswith(".histo") \
or tokens[2].endswith(".oldhisto.gz") \
or tokens[2].endswith(".png"):
message+="<img src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\" alt=\""+tokens[2]+"\"/><br><br>\n"
elif tokens[2].endswith(".vec.gz") \
or tokens[2].endswith(".oldvec.gz") \
or tokens[2].endswith(".vec") \
or tokens[2].endswith("txt") \
or tokens[2].endswith("txt.gz") \
or tokens[2].endswith("textID"):
message+="<iframe src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\"></iframe><br><br>\n"
else:
message+="not yet supported<br>\n"
line=s.getline() # get end
s.sendcmd("bye")
Display(message)
elif(form.has_key("textfeatureinfo")):
message+=displayTextFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
else:
# we do not want to show the settings window
if(form.has_key("newFile") and form["newFile"].value=="1"):
#message+=str(form["newFile"])
message+=newFile(form)
elif(form.has_key("serverfile")):
message+=serverFile(form)
elif(form.has_key("feedback")):
# result images have been selected and are part of this query
if form["feedback"].value=="requery":
# relevance feedback
message+=feedbackretrieve(form)
elif form["feedback"].value=="expandresults":
# more results
message+=expandQuery(form)
else:
# save relevances (this will not happen currently)
message=saveRelevances(form)
elif(form.has_key("queryImage") and form["queryImage"].value!=""):
# no relevance feedback, but just one of the random images was clicked
message+=retrieve(form)
elif form.has_key('textquerystring'):
# no relevance feedback, no query image, but a query string for text queries
message+=processTextQuery(form)
# now generate the remainder of the website: queryfield for meta information, random images, adminlink
if("textfeature" in settings.distances or "metafeature" in settings.distances):
message+=textQueryForm(settings)
message+=randomImages()
message+=newFileForm()
message+="<p align=\"right\">"+adminLink(settings.fireServer, settings.firePort)+"</p>"
message+="</form>"
# disconnect from server
s.sendcmd("bye")
print "</pre>"
# display the generated webpage
Display(message)
# except Exception, e:
# if anything went wrong:
# show the error message as readable as possible,
# disconnect from the server to keep it alive
# and end yourself
# print "<pre>"
# print "Exception: ",e
# s.sendcmd("bye")
# print "</pre>"
| Python |
#default settings
positiveImg="images/positive.png"
negativeImg="images/negativ.png"
neutralImg="images/neutral.png"
fireLogo="images/fire-logo.png"
i6Logo="images/i6.png"
TemplateFile = "fire-template.html"
fireServer="localhost"
firePort=12960
imgpyurl="img.py"
fireurl="fire.py"
#tempdir="/tmp"
#maxSize=1000, 1000
#minSize=200, 200
| Python |
#!/usr/bin/python
import sys
import cgi
import re
import Image
sys.stderr=sys.stdout
try:
form=cgi.FieldStorage()
imagefilename=form["image"].value
outWidth=-1
outHeight=-1
maxDim=-1
if(form.has_key("max")):
maxDim=int(form["max"].value)
if(form.has_key("width")):
outWidth=int(form["width"].value)
if(form.has_key("height")):
outHeight=int(form["height"].value)
inImg=Image.open(imagefilename)
inWidth=inImg.size[0]
inHeight=inImg.size[1]
if maxDim!=-1:
if inWidth>inHeight:
outWidth=maxDim
else:
outHeight=maxDim
if (outWidth==-1 and outHeight==-1):
outImg=inImg
elif (outWidth==-1):
scaleFactor=float(outHeight)/float(inHeight)
outWidth=int(scaleFactor*inWidth)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
elif (outHeight==-1):
scaleFactor=float(outWidth)/float(inWidth)
outHeight=int(scaleFactor*inHeight)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
else:
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
contenttype="image/jpg"
print "Content-Type: "+contenttype+"\n"
outImg=outImg.convert("RGB")
outImg.save(sys.stdout,"jpeg")
except:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Access not permitted"
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
#$Header$
# print cgi-header
print "Content-Type: text/html\n\n"
print
#print "<pre style=\"font-family:Fixedsys,Courier,1px;\">" # to have error messages as readable as possible, this
# is switched of right before displaying the main part of the page
#import lots of stuff
import cgi, re, sys, traceback, os, Image
#import cgitb; cgitb.enable()
from urllib import quote,unquote
from types import *
# redirect stderr to have error messages on the webpage and not in the
# webserver's log
sys.stderr=sys.stdout
# to be able to find the config
sys.path+=[os.getcwd()]
#import "my" socket implementation
from firesocket import *
# ----------------------------------------------------------------------
# import the config
# it is assumed there is a file called config in the same file as the cgi.
# an example for a valid config:
## settings for i6
#positiveImg="http://www-i6.iatik.rwth-aachen.de/~deselaers/images/positive.png"
#negativeImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/negativ.png"
#neutralImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/neutral.png"
#fireLogo="/u/deselaers/work/src/fire/cgi/fire-logo.png"
#i6Logo= "/u/deselaers/work/src/fire/cgi/i6.png"
#TemplateFile = "/u/deselaers/work/src/fire/cgi/fire-template.html"
#fireServer="deuterium"
#firePort=12960
#tempdir="/tmp"
#maxSize=1000, 1000
#minSize=200, 200
# ----------------------------------------------------------------------
import config
# settings for the linear (standard) scoring algorithm. using this
# algorithm the weights can be changed from the web interface
class LinearScoring:
def __init__(self):
self.size=0
self.weights=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
if(tmp=="weight"):
idx=status.pop(0)
w=status.pop(0)
self.weights+=[w]
else:
print "Cannot parse LinearScoring settings"
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for LinearScoring</th></tr>"
for i in range(len(self.weights)):
result+="<tr><td> Weight "+str(i)+"</td>"
result+="<td colspan=2><input name=\"weight"+str(i)+"\" type=\"text\" value=\""+self.weights[i]+"\"></td></tr>"
return result
# settings for maxent scoring algorithm
class MaxEntScoring:
def __init__(self):
self.size=0
self.factor=0
self.offset=0
self.lambdas=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
tmp=status.pop(0)
self.factor=status.pop(0)
tmp=status.pop(0)
self.offset=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
idx=status.pop(0)
l=status.pop(0)
self.lambdas+=[l]
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for MaxEntScoring</th></tr>"
for i in range(len(self.lambdas)/2):
result+="<tr><td> Lambda "+str(i)+"</td>"
result+="<td>"+self.lambdas[i]+"</td><td>"+self.lambdas[i+len(self.lambdas)/2]+"</td></tr>"
return result
# a class to store the settings of fire this classes uses the classes
# defined above to manage the settings of the active scoring algorithm
class FireSettings:
def __init__(self):
# variable initilization only. all member variables are defined here.
self.fireServer=config.fireServer
self.firePort=config.firePort
self.dbsize=0
self.scoringname=""
self.results=0
self.expansions=0
self.distances=[]
self.suffices=[]
self.path=""
self.scoring=""
def get(self,s):
# get the settings from the server, i.e. read the string s and parse it.
self.distances=[]
self.suffices=[]
s.sendcmd("info")
msg=s.getline()
status=re.split(" ",msg)
print "<!--"+str(status)+"-->"
while(len(status)>0):
keyword=status.pop(0)
if keyword=="filelist":
self.dbsize=status.pop(0)
elif keyword=="results":
self.results=status.pop(0)
elif keyword=="path":
self.path=status.pop(0)
elif keyword=="extensions":
self.expansions=status.pop(0)
elif keyword=="scoring":
self.scoringname=status.pop(0)
if self.scoringname=="linear":
self.scoring=LinearScoring()
elif self.scoringname=="maxent":
self.scoring=MaxEntScoring()
self.scoring.get(status)
elif keyword=="suffix":
no=status.pop(0)
suffix=status.pop(0)
self.suffices+=[suffix]
distname=status.pop(0)
self.distances+=[distname]
else:
print "<!-- Unknown keyword in stream : "+keyword+" -->"
def display(self):
# display the settings dialog
self.get(s)
result="<h4>Settings for fire</h4>"
result=result+"<form method=\"post\" name=\"settingsForm\">"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+str(self.fireServer)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(self.firePort)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"settings\" value=\"1\"/>\n"
result=result+"<table border=\"1\">\n"
result=result+"<tr><th colspan=3> General Settings</th></tr>"
result=result+"<tr><td>Number of images in database</td><td colspan=\"2\">"+str(self.dbsize)+"</td></tr>\n"
result=result+"<tr><td>Number of results shown</td><td colspan=\"2\"><input type=\"text\" value=\""+self.results+"\" name=\"results\"/></td></tr>\n"
result=result+"<tr><td>Scoring</td><td colspan=\"2\">"+self.getScoringChooser()+"\n"
result=result+"<tr><td>Number of Images for Query Expansion</td><td colspan=\"2\"><input type=\"text\" value=\""+self.expansions+"\" name=\"extensions\"/></td></tr>\n"
result=result+self.getDistForm()
result=result+self.getScoringForm()
result=result+"</table>\n"
result=result+"Password: <input type=\"password\" name=\"password\">\n"
result=result+"<input type=\"submit\" name=\"Save\" value=\"Save\"/>\n"
result=result+"</form>"
return result
def getScoringChooser(self):
# select the chooser for the server
result="<select name=\"scoring\">\n<option value=\"\">"
for scor in ['linear','maxent']:
if self.scoringname==scor:
result=result+"<option selected value=\""+scor+"\">"+scor+"</option>\n"
else:
result=result+"<option value=\""+scor+"\">"+scor+"</option>\n"
result+="</select>\n"
return result
def getScoringForm(self):
# display the form element to select the scoring algorihtm
return self.scoring.getForm()
def getDistForm(self):
result=""
result+="<tr><th colspan=3> Settings for Distance Functions</th></tr>"
for no in range(len(self.distances)):
result+="<tr><td> Distance "+str(no)+": "+self.suffices[no]+"</td><td>"+self.getDistChooser(no,self.distances[no])+"</td></tr>"
return result
def getDistChooser(self,no,distName):
result="<select name=\"dist"+str(no)+"\">\n<option value=\"\">"
s.sendcmd("setdist")
distsString=s.getline()
dists=re.split(" ",distsString)
for d in dists:
if d==distName:
result=result+"<option selected value=\""+d+"\">"+d+"</option>\n"
else:
result=result+"<option value=\""+d+"\">"+d+"</option>\n"
result=result+"</select>\n"
return result
def process(self,form):
result="<!-- processing settings -->\n"
if form.has_key("results"):
if form.has_key("password"):
password=form["password"].value
else:
password=""
s.sendcmd("password "+password)
l=s.getline()
result+="<!-- "+l+"-->"
if l!="ok":
result+="Changing settings denied: Not authorized!"
for i in form.keys():
if i.find("dist")==0:
no=i[4:]
d=form[i].value
if d!=self.distances[int(no)]: # only change when changed
s.sendcmd("setdist "+no+" "+d)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("weight")==0:
no=i[6:]
w=float(form[i].value)
if self.scoringname=="linear":
if(float(self.scoring.weights[int(no)])!=float(w)):
s.sendcmd("setweight "+no+" "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
else:
print "weights not supported"
elif i.find("results")==0:
if int(form[i].value)!=int(self.results):
s.sendcmd("setresults "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("extensions")==0:
if int(self.expansions)!=int(form[i].value):
s.sendcmd("setextensions "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("scoring")==0:
sn=form[i].value
if(sn!=self.scoringname): # only change when changed
s.sendcmd("setscoring "+sn)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
self.scoringname=sn
else:
result=result+"<!-- "+i+"-->\n"
result=result+"<!-- settings processed -->\n"
return result
# ----------------------------------------------------------------------
# Load the template file and display the content in this template file
# the string FIRELOGOSTRINGHERE, I6LOGOSTRINGHERE, and "INSERT CONTENT
# HERE" are replaced by the configured/generated strings
# ----------------------------------------------------------------------
def Display(Content):
TemplateHandle = open(config.TemplateFile, "r") # open in read only mode
# read the entire file as a string
TemplateInput = TemplateHandle.read()
TemplateHandle.close() # close the file
# this defines an exception string in case our
# template file is messed up
BadTemplateException = "There was a problem with the HTML template."
TemplateInput = re.sub("FIRELOGOSTRINGHERE",config.fireLogo,TemplateInput)
TemplateInput = re.sub("I6LOGOSTRINGHERE",config.i6Logo,TemplateInput)
SubResult = re.subn("INSERT CONTENT HERE",Content,TemplateInput)
if SubResult[1] == 0:
raise BadTemplateException
print SubResult[0]
def sendretrieve(querystring):
filterstring=""
# if(form.has_key("filter1")):
# filterstring+=" -porn2-sorted/009-www.ficken.bz-images-ficken.png"
# if(form.has_key("filter2")):
# filterstring+=" -porn2-sorted/037-www.free-porn-free-sex.net-free-porn-free-sex-free-sex-sites.png"
s.sendcmd("retrieve "+querystring+filterstring)
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by clicking a random
# image, that is:
# 1. process the form
# - extract the filename of the image which is used as query
# 2. send the query to the server
# 3. wait for the servers answer
# 4. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def retrieve(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["queryImage"].value
sendretrieve("+"+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimage,0)
return result
# ----------------------------------------------------------------------
# resizes an image while keeping aspect ratio
# ----------------------------------------------------------------------
def resizeImage(im, newSize):
newX, newY = newSize
if(im.size < newSize):
xSize, ySize = im.size
ratio=float(xSize)/float(ySize)
if(ySize < xSize):
newSize=(int(newY*ratio),newY)
else:
newSize=(newX,int(newX*(1/ratio)))
out = Image.new("RGB", newSize)
out = im.resize(newSize, Image.ANTIALIAS)
else:
im.thumbnail(newSize, Image.ANTIALIAS)
out=im
return out
# ----------------------------------------------------------------------
# verifies the image file and resizes the image according to
# minSize and maxSize parameters in config.py
# ----------------------------------------------------------------------
def checkImage(imagefile):
maxSize=config.maxSize
minSize=config.minSize
try:
im=Image.open(imagefile)
except:
return 0
if(im.size<minSize):
im=resizeImage(im, minSize)
elif(im.size>maxSize):
im=resizeImage(im, maxSize)
im.save(imagefile)
return 1
# ----------------------------------------------------------------------
# handle a serverfile-request
# ----------------------------------------------------------------------
def serverFile(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["serverfile"].value
checkImage(queryimage)
imagename=queryimage[queryimage.rfind("/")+1:len(queryimage)]
s.sendcmd("newfile 2 "+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+imagename,0)
return result
# ----------------------------------------------------------------------
# handle a newfile-request
# ----------------------------------------------------------------------
def newFile(form):
temporaryFiles=config.tempdir # standard: "/tmp"
result="<h3>Retrieval Result</h3>\n"
if(save_uploaded_file(form,"absolute_path",temporaryFiles)==0):
result=result+"No Retrieval Result available!"
return result
fileitem = form["absolute_path"]
# f=os.popen("ls -l "+ temporaryFiles+"/"+fileitem.filename)
# print f.readlines()
if(checkImage(temporaryFiles+"/"+fileitem.filename)==0):
result=result+"Delivered file was not a valid image!"
return result
s.sendcmd("newfile 2 "+temporaryFiles+"/"+fileitem.filename+".png")
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+fileitem.filename,0)
return result
# This saves a file uploaded by an HTML form.
# The form_field is the name of the file input field from the form.
# For example, the following form_field would be "file_1":
# <input name="file_1" type="file">
# The upload_dir is the directory where the file will be written.
# If no file was uploaded or if the field does not exist then
# this does nothing.
# written by Noah Spurrier, modified by Fabian Schwahn
def save_uploaded_file(form, form_field, upload_dir):
if not form.has_key(form_field): return 0
fileitem = form[form_field]
completepath = upload_dir + fileitem.filename
if os.path.isdir(completepath): return 0
if not fileitem.file: return 0
fout = open (os.path.join(upload_dir, fileitem.filename), 'wb')
while 1:
chunk = fileitem.file.read(100000)
if not chunk: break
fout.write (chunk)
fout.close()
return 1
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by relevance feedback.
# That is:
# 1. process the form
# - extract the filenames of positive and negative examples
# 2. generate the query string
# 3. send the query to the server
# 4. wait for the servers answer
# 5. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def feedbackretrieve(form):
print "<!-- feedback retrieve \n"
print form
print "-->\n"
result="<h3>Retrieval Result</h3>\n"
queryimages=""
for field in form.keys():
if field.startswith("relevance"):
imageno=re.sub("relevance","",field)
imagename=form["resultImage"+imageno].value
relevance=form[field].value
if relevance!="0":
queryimages=queryimages+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
result=result+"\n<!-- "+queryimages+" -->\n"
sendretrieve(queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimages,0)
return result
# ----------------------------------------------------------------------
# given the retrieval result string of the server process this answer
# and create the part of the html page displaying these results. this
# includes the buttons for relevance feedback and the hidden fields
# for server settings (e.g.\ which server, which port). also the
# buttons for "more results" and "relevance feedback" are generated.
# the button for saving relevances is deaktivated as it is not used
# ----------------------------------------------------------------------
def displayResults(tokens,querytype,querystring,resultsStep):
if(re.search("[0-9]+(\.[0-9]+)?", tokens[0]) == None):
result="Could not read retrieval result.\n"
s.flush()
return result
print "<!-- "+str(tokens)+"-->"
i=0
result="<form method=\"post\" name=\"resultForm\"><table>\n<tr>\n"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+querystring+"\"/>\n"
result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+quote(querystring)+"\" />\n"
result=result+"<input type=\"hidden\" name=\"resultsstep\" value=\""+str(resultsStep)+"\"/>\n"
first=True
while(len(tokens) > 0):
resNo=str(i)
img=tokens.pop(0)
score=float(tokens.pop(0))
result=result+"<td>\n"
result=result+"<a href=\""+config.fireurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&imageinfo="+img+"\" target=\"_blank\">"
# support for the select-rectangle tool
#-------------
# result=result+"<a href=\"save_coord.cgi?to_do=UpdateImage&upload_type=server&port="+str(settings.firePort)+"&image_name="+settings.path+"/"+img+"\" target=\"_top\">"
#-------------
# shows the image
#-------------
# result=result+"<a href=\"img.py?image="+settings.path+"/"+img+"\" target=\"_blank\">"
#------------
result=result+"<img src=\""+config.imgpyurl+"?max=150&image="+settings.path+"/"+img+"\" title=\""+str(score)+"-"+img+"\" name=\""+str(score)+"-"+img+"\">\n<br>\n"
result=result+"</a>"
result=result+"<input type=\"hidden\" name=\"resultImage"+resNo+"\" value=\""+img+"\">\n"
if first:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\" checked><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\"><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
first=False
else:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\"><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\" checked><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
result=result+"</td>\n\n"
i=i+1
if i%5==0:
result=result+"</tr>\n\n\n<tr>"
result=result+"</tr></table>\n"
# result+=filterbuttons()
result=result+"<button title=\"expandresults\" name=\"expandresults\" type=\"submit\" onClick=\"document.resultForm.feedback.value='expandresults';document.resultForm.submit()\">more results</button>\n"
result=result+"<button title=\"requery\" name=\"requery\" type=\"submit\" onClick=\"document.resultForm.feedback.value='requery';document.resultForm.submit()\">requery</button>\n"
# result=result+"<button title=\"saverelevances\" name=\"relevancessave\" onClick=\"document.resultForm.feedback.value='saverelevances';document.resultForm.submit()\" \"type=\"submit\">save relevances</button>\n"
result=result+"<input type=\"hidden\" name=\"feedback\" value=\"yes\"/>\n"
result=result+"</form>\n"
return result
def filterbuttons():
# display the buttons for the filter
result=""
result=result+"filter 1: <input type=\"checkbox\" name=\"filter1\" value=\"1\">\n"
result=result+"filter 2: <input type=\"checkbox\" name=\"filter2\" value=\"2\">\n<br>\n"
return result
# ----------------------------------------------------------------------
# ask the server for filenames of random images and create the part of
# the html page showing random images. Steps:
# 1. send "random" to the server
# 2. receive the servers answer
# 3. process the answer:
# - get the filenames and for each filename create a querybutton
# 4. put the "more random images" button, such that we can get a new set
# of random images
# ----------------------------------------------------------------------
def randomImages():
# display random images
s.sendcmd("random")
result="<hr><h3> Random Images - Click one to start a query </h3>"
result+="""<form method="post" name="queryForm" enctype="multipart/form-data">
<input type="hidden" name="queryImage" value="">
"""
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result+=filterbuttons()
res=s.getline()
for img in re.split(" ",res):
if img != "" and img != "\n":
result=result+"<button title=\""+img+"\" name=\"searchButton\" type=\"submit\" onClick=\"document.queryForm.queryImage.value='"+img+"';document.queryForm.submit()\" value=\""+img+"\">\n"
result=result+"<img src=\""+config.imgpyurl+"?max=100&image="+settings.path+"/"+img+"\" title=\""+img+"\">\n"
result=result+"</button>\n"
if form.has_key("demomode"):
seconds=form["demomode"].value
url="http://"
if(os.environ.has_key("HTTP_HOST") and os.environ.has_key("REQUEST_URI")):
part1=os.environ["HTTP_HOST"]
part2=re.sub("&queryImage=.*$","",os.environ["REQUEST_URI"])
url=url+part1+part2+"&queryImage="+img
print "<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result=result+"<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result+="<br><div align=\"right\">\n<button title=\"new random images\" name=\"morerandom\" type=\"submit\">more random images</button>\n</div>\n"
return result
# ----------------------------------------------------------------------
# displays the newfile-form
# ----------------------------------------------------------------------
def newFileForm():
result="<hr><h3> Upload - Use query image from your file system </h3>"
result+="""<input type="hidden" name="newFile" value="0">"""
result+="""<input name="absolute_path" type="file" size="30" onClick="document.queryForm.newFile.value=1;document.queryForm.submit()">"""
result+="\n"
return result
# ----------------------------------------------------------------------
# generate the text field and the query button for text
# queries
# ----------------------------------------------------------------------
def textQueryForm(settings):
result = "<hr><h3>Query by "
if("textfeature" in settings.distances):
result += "description"
if("metafeature" in settings.distances):
result += " or meta information"
elif("metafeature" in settings.distances):
result += "meta information"
result += "</h3>\n"
result += "<form method=\"post\" name=\"textQueryForm\">\n"
result += "<input type=\"hidden\" name=\"server\" value=\""+str(settings.fireServer)+"\"/>\n"
result += "<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
result += "<input type=\"hidden\" name=\"resultsstep\" value=\"0\"/>\n"
result += "<input type=\"text\" name=\"textquerystring\" size=\"30\" value=\"\"/>\n"
result += "<input type=\"submit\" value=\"query\"/>\n"
if("metafeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&metafeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on meta info</a>"
if("textfeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&textfeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on description</a>"
result += "</form>"
return result
def displayMetaFeatureInfo():
s.sendcmd("metafeatureinfo")
mfi = s.getline()
mfil = mfi.split(" ")
result = "<h4>Help on queries by meta info</h4>\n"
result += """If you want to find images with certain meta information attached,
type in a request of the form
<h5>key1:val1,key2:val2,...</h5>\n
The following meta keys are available in this corpus:<br><br>\n"""
#TODO: The table looks ugly!
result += "<table>\n"
result += "<tr>\n"
result += " <th>key</th>\n"
result += " <th>example value</th>\n"
result += "</tr>\n"
for mf in mfil:
mfl = mf.split(":")
result += "<tr>\n"
result += " <td><b>"+mfl[0]+"</b></td>\n" # I know <b> is deprecated, but it still works :-)
result += " <td>"+mfl[1]+"</td>\n"
result += "</tr>\n"
result += "</table>\n"
return result
def displayTextFeatureInfo():
result = "<h4>Help on queries by description</h4>\n"
result += """If you want to find images which have text information attached to them,
just enter the query into the textbox. Fire will then use an information
retrieval engine to find the images that best match your query.<br><br>
If you have text information in multiple languages for each image, you can give
a query for every language. For example, if you have german and french text
information and you want to search for "hand" in the respective languages, you enter
<h5>Ge:"hand" Fr:"mains"</h5>
Here, "Ge" has to be the suffix for the german textfiles and "Fr" has to be the suffix for the
fench textfiles. Fire will then use both queries and send them to separate information retrieval
engines."""
return result
# ----------------------------------------------------------------------
# this function makes the server save a relevances logfile
# but it is not used at the moment, because the saverelevances button
# is deaktivated
# ----------------------------------------------------------------------
def saveRelevances(form):
querystring=form["querystring"].value
relevancestring=""
for field in form.keys():
if field.startswith("relevance-"):
imagename=re.sub("relevance-","",field)
relevance=form[field].value
if relevance!="0":
relevancestring=relevancestring+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
s.sendcmd("saverelevances Q "+querystring+" R "+relevancestring)
res=s.getline();
message="<a href=\"javascript:back()\"> Go back to last query </a>"
return message
# ----------------------------------------------------------------------
# this is the "getMoreResults" function it basically does the same as
# retrieve, uses the same query string as the last query used (this is
# gotten from the form) and the calls displayResults
# ----------------------------------------------------------------------
def expandQuery(form):
resultsstep=int(form["resultsstep"].value)
resultsstep=resultsstep+1
if(form.has_key("metaquerystring")):
result="<h3>Metatag-based retrieval result</h3>\n"
queryfield = "metaquerystring"
cmd = "metaexpand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
elif(form.has_key("textquerystring")):
result="<h3>Text-based retrieval result</h3>\n"
queryfield = "textquerystring"
cmd = "textexpand"
queryimages = unquote(form["textquerystring"].value)
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
else:
result="<h3>Retrieval Result</h3>\n"
queryfield = "querystring"
cmd = "expand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
if(tokens[0] == "notextinformation"):
result=result+"No images matched your query: "+querystring
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,queryfield,queryimages,resultsstep)
return result
# ----------------------------------------------------------------------
# process a query using text- or metainformation from the text input
# field. The query type is set to metaquery when the search string has
# a leading "META "
# This works identical as the retrieve function
# ----------------------------------------------------------------------
def processTextQuery(form):
# Check if it's meta or text
# It's meta either if there's only the metafeature or it has a leading "META "
hasmeta = "metafeature" in settings.distances
hastext = "textfeature" in settings.distances
querystring = form["textquerystring"].value
if (hasmeta and (not hastext or querystring[:5] == "META ")):
if(querystring[:5] == "META "):
querystring = querystring[5:]
querytype = "metaquerystring"
result="<h3>Metatag-based retrieval result</h3>\n"
s.sendcmd("metaretrieve +"+querystring)
else:
querytype = "textquerystring"
result="<h3>Text-based retrieval result</h3>\n"
s.sendcmd("textretrieve +"+querystring)
msg=s.getline()
tokens=re.split(" ",msg)
if( (tokens[0] == "notextinformation") or (tokens[0] == "nometainformation") ):
result=result+"No images matched your query: "+querystring
if(querytype == "metaquerystring"):
if(hastext):
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'META key1:val1,key2:val2,...'<br><br>\n"
else:
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'key1:val1,key2:val2,...'<br><br>\n"
result=result+"Also have a look at the help to see which keys are available.<br>\n"
if(querytype == "textquerystring"):
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,querytype,querystring,0)
return result
# ----------------------------------------------------------------------
# create the link for the "settings" at the bottom of the page
# ----------------------------------------------------------------------
def adminLink(fireServer, firePort):
result=""
result=result+"<a name=\"#\"></a>"
result=result+"<a href=\"#\" "+\
"onClick=\"window.open('"+config.fireurl+"?server="+ \
fireServer+"&port="+str(firePort)+"&settings=1','newwindow',"+\
"'height=600, width=600')\">settings</a>"
return result
#<SCRIPT LANGUAGE="javascript">
#<!--
#window.open ('titlepage.html', 'newwindow', config='height=100,
#width=400, toolbar=no, menubar=no, scrollbars=no, resizable=no,
#location=no, directories=no, status=no')
#-->
#</SCRIPT>
# ----------------------------------------------------------------------
# the main program
# ----------------------------------------------------------------------
# get form information in easily accessible manner
form=cgi.FieldStorage()
# make socket
s = FIRESocket()
settings=FireSettings()
# see what server we have and if the webinterface specifed another one
# than the one in the config file
if form.has_key("server"):
settings.fireServer=form["server"].value
if form.has_key("port"):
settings.firePort=int(form["port"].value)
try:
# connect to the server
s.connect(settings.fireServer, settings.firePort)
except:
# if the server is down, give an error message
# print "</pre>"
message="""
<h2>FIRE Server down</h2>
<p>
Try again later, complain to
<a href="mailto:deselaers@informatik.rwth-aachen.de">deselaers@informatik.rwth-aachen.de</a>
or try other servers.
"""
Display(message)
else:
# everything is fine, connection to server is ok
# try:
message=""
settings.get(s)
# do we want to show the settings window?
if(form.has_key("settings")):
# see if there have been new settings defined
message=settings.process(form)
# generate the settings dialog
message+=settings.display()
# disonnect from server
s.sendcmd("bye")
print "</pre>"
# display the settings dialog
Display(message)
# do we want to show the metafeatureinfo window?
elif(form.has_key("metafeatureinfo")):
message+=displayMetaFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
elif(form.has_key("imageinfo")):
imagename=form["imageinfo"].value
s.sendcmd("image "+imagename)
line=""
while(line != "end"):
line=s.getline()
tokens=re.split(" ",line)
if tokens[0]=="imagefilename":
message+="<img src=\""+config.imgpyurl+"?image="+tokens[1]+"\" alt=\""+tokens[1]+"\"/><br>"+"\n"
elif tokens[0]=="class":
message+="class: "+tokens[1]+"<br>\n"
elif tokens[0]=="description":
message+="description: "+tokens[1:]+"<br>\n"
elif tokens[0]=="features":
features=[]
noffeatures=int(tokens[1])
for i in range(noffeatures):
line=s.getline()
tokens=re.split(" ",line)
message+="<h3>"+str(i)+": "+tokens[2]+"</h3>\n<br>\n"
if tokens[2].endswith(".histo.gz") \
or tokens[2].endswith(".histo") \
or tokens[2].endswith(".oldhisto.gz") \
or tokens[2].endswith(".png"):
message+="<img src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\" alt=\""+tokens[2]+"\"/><br><br>\n"
elif tokens[2].endswith(".vec.gz") \
or tokens[2].endswith(".oldvec.gz") \
or tokens[2].endswith(".vec") \
or tokens[2].endswith("txt") \
or tokens[2].endswith("txt.gz") \
or tokens[2].endswith("textID"):
message+="<iframe src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\"></iframe><br><br>\n"
else:
message+="not yet supported<br>\n"
line=s.getline() # get end
s.sendcmd("bye")
Display(message)
elif(form.has_key("textfeatureinfo")):
message+=displayTextFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
else:
# we do not want to show the settings window
if(form.has_key("newFile") and form["newFile"].value=="1"):
#message+=str(form["newFile"])
message+=newFile(form)
elif(form.has_key("serverfile")):
message+=serverFile(form)
elif(form.has_key("feedback")):
# result images have been selected and are part of this query
if form["feedback"].value=="requery":
# relevance feedback
message+=feedbackretrieve(form)
elif form["feedback"].value=="expandresults":
# more results
message+=expandQuery(form)
else:
# save relevances (this will not happen currently)
message=saveRelevances(form)
elif(form.has_key("queryImage") and form["queryImage"].value!=""):
# no relevance feedback, but just one of the random images was clicked
message+=retrieve(form)
elif form.has_key('textquerystring'):
# no relevance feedback, no query image, but a query string for text queries
message+=processTextQuery(form)
# now generate the remainder of the website: queryfield for meta information, random images, adminlink
if("textfeature" in settings.distances or "metafeature" in settings.distances):
message+=textQueryForm(settings)
message+=randomImages()
message+=newFileForm()
message+="<p align=\"right\">"+adminLink(settings.fireServer, settings.firePort)+"</p>"
message+="</form>"
# disconnect from server
s.sendcmd("bye")
print "</pre>"
# display the generated webpage
Display(message)
# except Exception, e:
# if anything went wrong:
# show the error message as readable as possible,
# disconnect from the server to keep it alive
# and end yourself
# print "<pre>"
# print "Exception: ",e
# s.sendcmd("bye")
# print "</pre>"
| Python |
#!/usr/bin/python
import sys,re
import filelist
filelist=filelist.FileList()
filelist.load(sys.argv[1])
for i in filelist:
cls=i[1]
fname=i[0]
print fname,
for j in filelist:
if j[1]==cls and j[0]!=fname:
print j[0],
print
#print filelist
| Python |
import re,sys,string,os
class FileList:
def __init__(self):
self.files=[]
self.classes=False
self.featuredirectories=False
self.descriptions=False
self.path=""
self.rawpath=""
self.suffices=[]
self.filename2idx={}
def __str__(self):
result=""
result+="Suffices: "+str(self.suffices)
result+="Files: "+str(self.files)
return result
def __getitem__(self,i):
return self.files[i]
def __getslice__(self,i):
return self.files[i]
def index(self,filename):
if self.filename2idx.has_key(filename):
return self.filename2idx[filename]
else:
print >>sys.stderr,filename,"not in database, returning -1"
#print >>sys.stderr,self.filename2idx
#print >>sys.stderr,self.files
return -1
#result=-1
#for i in range(len(self.files)):
# if self.files[i][0]==filename:
# result=i
#if result==-1:
# print "filelist.py: file",filename,"not in database"
#return result
def load(self, filename):
fl=open(filename,"r")
lines=fl.readlines()
for l in lines:
l=re.sub("\n","",l)
tokens=string.split(l)
if tokens[0] == "file":
f=tokens[1]
if self.classes==True:
cls=int(tokens[2])
if self.descriptions:
desc=tokens[3:]
self.files+=[[f,cls,desc]]
else:
self.files+=[[f,cls]]
elif self.descriptions:
desc=tokens[2:]
self.files+=[[f,0,desc]]
else:
self.files+=[[f,]]
elif tokens[0] == "classes":
if tokens[1]=="yes":
self.classes=True
elif tokens[0] == "descriptions" or tokens[0] == "description":
if tokens[1]=="yes":
self.descriptions=True
elif tokens[0] == "featuredirectories":
if tokens[1]=="yes":
self.featuredirectories=True
elif tokens[0] == "path":
self.rawpath=tokens[1]
self.path=self.rawpath
if self.path.startswith("this"):
self.path=self.path.replace("this+","")
offs=os.path.dirname(filename)
cwd=os.path.realpath(os.path.curdir)
self.path=offs+"/"+self.path+"/"
elif tokens[0] == "suffix":
self.suffices+=[tokens[1]]
self.initMap()
def add(self, filename, class_=0, description=""):
if description == "":
assert(self.descriptions == False)
self.files.append([filename, class_, description])
else:
assert(self.descriptions == True)
self.files.append([filename, class_])
def save(self, filename):
if filename=="-":
f=sys.stdout
else:
f=open(filename,"w")
print >>f,"FIRE_filelist"
if self.classes:
print >>f,"classes yes"
if self.descriptions:
print >>f,"descriptions yes"
if self.featuredirectories:
print >>f,"featuredirectories yes"
print >>f,"path",self.rawpath
for s in self.suffices:
print >>f,"suffix",s
for fi in self.files:
print >>f, "file %s" % " ".join(map(lambda x: str(x), fi))
def initMap(self):
self.filename2idx={}
for i in range(len(self.files)):
self.filename2idx[self.files[i][0]]=i
def basenameify(self):
for i in range(len(self.files)):
if len(self.files[i])==3:
self.files[i]=(os.path.basename(self.files[i][0]),self.files[i][1], self.files[i][2])
if len(self.files[i])==2:
self.files[i]=(os.path.basename(self.files[i][0]),self.files[i][1])
if len(self.files[i])==1:
self.files[i]=(os.path.basename(self.files[i][0]),)
def nameWithSuffix(self, item, suffix):
assert item < len(self.files)
assert suffix < len(self.suffices)
return self.files[item][0] + "." + self.suffices[suffix]
# generator for images with full path
def fullpaths(self):
for image in self.files:
yield os.path.join(self.rawpath, image[0])
def get_full_path_for_item(self, item, suffixidx=-1):
name = os.path.join(self.rawpath, self.files[item][0])
if suffixidx > -1 and len(self.suffices) > 0:
name += "." + self.suffices[suffixidx]
return name | Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
#-----------------------------------------------------------------
s=firesocket.FIRESocket()
try:
if len(sys.argv) < 5:
print """USAGE:
makedistfiles.py <querylist> <server> <port> <suffix> [-q]
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
suffix=sys.argv[4]
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
lines=ql.readlines()
lines=map(lambda line:
re.sub("\n","",line), lines)
ql.close()
i=0
for line in lines:
print i,"/",len(lines)
i+=1
tokens=re.split(" ",line)
file=tokens[0]
cmd="savedistances "+file+" "+file+"."+suffix
print "SENDING: ",cmd
s.sendcmd(cmd)
res=s.getline()
print res
sys.stdout.flush()
sys.stdout.flush()
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
filelist=filelist.FileList()
filelist.load(sys.argv[1])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in filelist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class SparseHistogramFeature:
def __init__(self):
self.bins={}
self.data={}
self.steps=[]
self.stepSize=[]
self.min=[]
self.max=[]
self.dim=0
self.numberofBins=0
self.counter=0
self.filename=""
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_sparse_histogram"
print >>f,"# sparse histogram file for FireV2"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps",reduce(lambda x,y: str(x)+" "+str(y),self.steps)
print >>f,"min",reduce(lambda x,y: str(x)+" "+str(y),self.min)
print >>f,"max",reduce(lambda x,y: str(x)+" "+str(y),self.max)
print >>f,"bins",len(self.bins)
for i in self.bins:
print >>f,"data",i,self.bins[i]
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_sparse_histogram":
fine=True
elif line[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="bins":
self.numberofBins=int(toks[1])
elif toks[0]=="data":
self.bins[toks[1]]=int(toks[2])
self.data[toks[1]]=float(toks[2])/float(self.counter)
else:
print "Unparsed line: '"+line+"'."
| Python |
#------------------------------------------------------------------------------
# proxyserver implementation for communication with the web frontend and fire
# server system
# server must be reachable via network sockets in order to be of any use
#
#------------------------------------------------------------------------------
__author__ = "Jens Forster <jens.forster@rwth-aachen.de>"
__version__= "0.1"
from SocketServer import *
from firesocket import *
# Needed for stringhandling
# re provides regular expressions and methods
import re
import string
# Needed for suitable choosing of random images
# hence each retrieval server will only have access to unique
# part of the database
import random
# Needed for logging
import logging
#from logging.handlers import TimedRotatingFileHandler
import sys
import os
import os.path
__all__ =["FIREProxyServer", "FIREProxyHandler"]
class FIREProxyServer(TCPServer):
request_queque_size = 1
allow_reuse_address = True
def __init__(self,server_address, retr_addrs, RequestHandlerClass,logDestination=None):
TCPServer.__init__(self,server_address,RequestHandlerClass)
if len(retr_addrs) == 0:
raise RuntimeError, "no retrieval servers spezified"
else:
self.retr_addrs = retr_addrs
self.retr_sock = []
# initializing loggin system
self.log = self._server_log(logDestination)
self.log.log(logging.INFO,'Initializing proxyserver')
self.log.log(logging.DEBUG,'Logging system and retrieval servers list ready')
# initializing boolean values for request and server termination
# and user authorization
self.notBye = True # true -> web interface didn't quit yet
self.notQuit = True # true -> proxy and retrieval servers still running
self.notProxyquit = True # false -> proxy going down, retrieval servers will remain online
self.auth = False
self.sendErrorsFrom = [] # a list which will contain retrieval servers which crashed or aren't reachable anymore
self.results = 0 # given results per retrieval server
self.log.log(logging.INFO,"Proxyserver running on "+str(server_address))
def sendToRetrievers(self,cmd,upperBound=None):
if upperBound != None:
if upperBound < 0:
raise RuntimeError, "no retrieval servers specified for sending, Proxyserver: method sendToRetrievers: upperBound "+str(upperBound)
relevantRetr = zip(self.retr_sock,range(upperBound))
for sock,idx in relevantRetr:
try:
sock.sendcmd(cmd)
self.log.log(logging.getLevelName("SEND "+str(self.retr_addrs[idx])),cmd)
except:
# iff no socket error occured until now
name,port = self.retr_addrs[idx]
if self.sendErrorsFrom == []:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str((name,port))+" did break" )
raise RuntimeError, "socket connection to retrieval server "+str((name,port))+" did break"
elif str(name) in self.sendErrorsFrom:
pass
else:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str((name,port))+" did break")
else:
for sock,addr in zip(self.retr_sock,self.retr_addrs):
try:
sock.sendcmd(cmd)
self.log.log(logging.getLevelName("SEND "+str(addr)),cmd)
except:
name,port = addr
self.log.log(logging.INFO,"sendErrorsFrom: "+str(self.sendErrorsFrom))
if self.sendErrorsFrom == []:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str(addr)+" did break")
raise RuntimeError, "socket connection to retrieval server "+str(addr)+" did break"
elif str(name) in self.sendErrorsFrom:
self.log.log(logging.DEBUG,"error in socket communication for retrieval server "+str(name)+" already known")
else:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str(addr)+" did break")
def getFromRetrievers(self):
answer = []
for sock,addr in zip(self.retr_sock,self.retr_addrs):
try:
ans = sock.getline()
except:
self.log.log(logging.CRITICAL,"retrieval server "+str(addr)+" doesn't respond")
name,port = addr
self.sendErrorsFrom.append(str(name))
raise RuntimeError, "retrieval server "+str(addr)+" didn't respond in method getFromRetrievers"
answer.append(ans)
self.log.log(logging.getLevelName("RECV "+str(addr)),ans)
return answer
def handle_request(self):
"""Handle one request in terms of one connection possibly containing
several actions, possibly blocking."""
errorCode = 0
try:
request, client_address = self.get_request()
self.log.log(logging.INFO,"client "+str(client_address)+" connected")
except:
self.notQuit = False
self.log.log(logging.ERROR,"an error occured during client's connection attempt")
self.log.log(logging.ERROR,"maybe the socket/pipe did break or there are network problems")
self.log.log(logging.ERROR,"proxyserver shutting down (retrieval server system may still be online)")
try:
request.close()
self.socket.close()
except:
pass
errorCode = 1
return errorCode
else:
try:
for serverNum in range(len(self.retr_addrs)):
serverNum_sock = FIRESocket()
self.retr_sock.append(serverNum_sock)
except:
self.notQuit = False
self.log.log(logging.ERROR,"couldn't create necessary network sockets for communication with retrieval servers")
self.log.log(logging.ERROR,"proxyserver shutting down (retrieval server system may still be online)")
request.close()
self.socket.close()
while self.retr_sock != []:
firesock = self.retr_sock.pop(0)
firesock.sock.close()
del firesock
errorCode = 2
return errorCode
else:
self.log.log(logging.DEBUG,"network sockets for communication with retrieval servers created")
if self.verify_request(request, client_address):
try:
self.process_request(request, client_address)
return errorCode
except RuntimeError, msg:
self.handle_error(request,msg)
errorCode = 3
return errorCode
except:
self.handle_error(request,"an error involving networksockets or an unknown error occured")
errorCode = 3
return errorCode
def process_request(self , request, client_addr):
self.finish_request(request, client_addr)
try:
self.close_request(request)
except:
pass
def close_request(self, client):
client.close()
self.log.log(logging.INFO,"client disconnected")
self.auth = False
# the network sockets designated for communication with retrieval servers
# will only be closed if bye, quit or an error are encountered
if self.notQuit:
self.sendToRetrievers("bye")
while self.retr_sock != []:
firesock = self.retr_sock.pop(0)
firesock.sock.close()
del firesock
def serve(self):
self.log.log(logging.INFO,"proxy handling requests")
# iff the error code is True no error occured while handling
errorCode = 0
while self.notQuit and self.notProxyquit :
errorCode = self.handle_request()
self.log.log(logging.DEBUG,"errorCode "+str(errorCode))
if errorCode == 0:
self.server_close()
else:
logging.shutdown()
def server_close(self):
# no auth necessary due to prior authentification
if self.notProxyquit:
self.log.log(logging.INFO,"sending shutdown signal to retrieval servers")
self.sendToRetrievers("quit")
self.socket.close()
# closing and deleting network sockets
while self.retr_sock != []:
sock = self.retr_sock.pop(0)
sock.sock.close()
del sock
if self.notProxyquit:
self.log.log(logging.INFO,"proxyserver and retrieval servers stopped")
else:
self.log.log(logging.INFO,"proxyserver stopped, retrieval servers running")
self.log.log(logging.DEBUG,"shutting logging system down")
logging.shutdown()
# printing the error message in to the logging file, closing connection to all retrieval servers
# and shutting down
def handle_error(self,request,msg):
self.notQuit = False
self.log.log(logging.ERROR,"an error occured while handling requests")
self.log.log(logging.ERROR,"for further information view the log file")
self.log.log(logging.DEBUG,msg)
self.log.log(logging.ERROR,"shutting proxyserver down")
request.close()
self.sendToRetrievers("bye")
self.socket.close()
while self.retr_sock != []:
sock = self.retr_sock.pop(0)
sock.sock.close()
del sock
del self.retr_sock
# logging.shutdown()
# Note that _server_log will always return a logger which will always log global statements to stdout
# and if intended will log everything to a file depending on the message's importance level
def _server_log(self,logDestination):
if logDestination == None:
location = None
else:
# exists is only needed if in future an other file logger should be used
# that doesn't check fileexistance
location, exists = self._logFilehandling(logDestination)
# Setting up the logging System
# The root logger shall log to stdout, all debug msg will be found in the files
# Adding new message levels to the logging system
# DEBUG (10) and NOTSET (0) needn't to be altered
logging.addLevelName(1020,"CRITICAL")
logging.addLevelName(1010,"ERROR")
logging.addLevelName(1005,"WARNING")
logging.addLevelName(1000,"INFO")
logging.addLevelName(11,"RECV (Web/Bash)")
logging.addLevelName(12,"SEND (Web/Bash)")
logging.INFO = 1000
logging.ERROR = 1010
logging.CRITICAL = 1020
logging.WARNING = 1005
# so the levels 13 up to 999 are free for retrieval servers
# because every server needs two numbers this logging system is capable of 493 retrieval servers
# now levels for the retrieval servers are added
# they are formated as SEND/RECV (IP-Address/Name)
level = 13
for addrs in self.retr_addrs:
logging.addLevelName(level,"SEND "+str(addrs))
logging.addLevelName(level+1,"RECV "+str(addrs))
level += 2
# now configurating the root logger which will log only to stdout
sys.stdout = sys.stderr
root = logging.getLogger('')
root.setLevel(logging.DEBUG)
console = logging.StreamHandler(sys.stdout)
console.setLevel(logging.INFO)
formatter = logging.Formatter('%(asctime)s %(levelname)-10s %(message)s','%a, %d-%m-%Y %H:%M:%S')
console.setFormatter(formatter)
root.addHandler(console)
# setting up the main file logger
if location != None:
# when = 'midnight'
# filesPerDay = 1
# backUpCount = 7
# file = TimedRotatingFileHandler(location,when,filesPerDay,backUpCount)
if exists:
file = logging.FileHandler(location,'a')
else:
file = logging.FileHandler(location,'w')
file.setLevel(logging.DEBUG)
formatter = logging.Formatter('%(asctime)s %(levelname)-6s %(message)s','%a, %d-%m-%Y %H:%M:%S')
file.setFormatter(formatter)
root.addHandler(file)
return root
def _logFilehandling(self,logDestination):
if logDestination == "default":
cwd = os.getcwd()
# determinating whether mac, unix/linux or windows
if os.name in ["mac","nt","posix"]:
if os.name == "nt":
# windows os
osPathTag = "\\"
else:
# unix/linux os or macintosch
osPathTag = "/"
try:
# checking whether dir exists
location = cwd+osPathTag+"fireproxylogs"
exist = os.path.isdir(location)
if not exist:
os.mkdir(location)
location += osPathTag+"proxylog"
del exist
del osPathTag
except:
raise RuntimeError, "error initalizing logging system (default logging location)"
else:
raise RuntimeError, "OS not supported for logging"
del cwd
# iff the user specified a storing location
# outer if-clause
else:
# if only a directory is given (logDestination must be terminated by the os's
# path separator (e.g / on linux \ on windows os)) default logfile name will be used
head,tail = os.path.split(logDestination)
osType = os.name
if osType in ["mac","posix"]:
osPathTag = "/"
elif osType == "nt":
osPathTag = "\\"
else:
raise RuntimeError, "OS not supported for logging"
# checking whether a path name is given
if head != "":
# checking whether the user set all path separators' right
head = os.path.normcase(head)
head = os.path.normpath(head)
# figuring out whether the supplied directories exist and
# if not which have to be created
if not os.path.isdir(head):
dirToCreate = [head]
newHead, newTail = os.path.split(head)
while newTail != "" and (not os.path.isdir(newHead)):
dirToCreate.append(newHead)
newHead, newTail = os.path.split(newHead)
while dirToCreate != []:
os.mkdir(dirToCreate.pop())
del newHead
del newTail
# path is valid now checking if a filename was supplied
if tail != "":
location = logDestination
else:
location = logDestination+osPathTag+"proxylog"
del dirToCreate
# no path given by the user
# tail must be != "" in this case otherwise it would be the default case
else:
# assuming current workdirectory was intended
cwd = os.getcwd()
location = cwd+osPathTag+tail
del osPathTag
del osType
del head
del tail
# Outer If Clause ends here
# testing if the file exist or not
exists = True
try:
file = open(location,'r+')
except IOError:
exists = False
else:
file.close()
return (location,exists)
# Overidden to be useless, because this proxy shall quit when all retrievers quit
def serve_forever(self):
pass
class FIREProxyHandler(StreamRequestHandler):
def __init__(self, request, client_addr, server):
StreamRequestHandler.__init__(self,request, client_addr, server)
def sendToWeb(self,msg):
self.wfile.write(msg+"\r\n")
self.server.log.log(logging.getLevelName("SEND (Web/Bash)"),msg)
def _validateSettings(self,settings):
self.server.log.log(logging.DEBUG,"valSet Argu:"+repr(settings))
leng = len(settings)
self.server.log.log(logging.DEBUG,"len: "+repr(leng))
if leng:
if settings[0][0] != "filelist":
retur = (False,"Settingsstream starts with wrong keyword")
else:
retur = (True, "")
if leng > 1:
lenSingle = map(lambda single: len(single),settings)
boolList = map(lambda single: single == lenSingle[0],lenSingle)
if not (reduce(lambda fst,snd: fst == snd,boolList)):
retur = (False,"At least one retrieval server didn't send complete settings")
# Fine all retrieval servers did send there complete settings
# or all made the same mistake :D
lenSetting = lenSingle[0]
# deleting helper variables
del boolList
# checking rest
for idx in range(lenSetting):
if (idx != 1):
for jdx in range(leng):
if (settings[0][idx] != settings[jdx][idx]):
# settings differ and singaling it
return (False,"settings of servers "+str(0)+" and "+str(jdx)+" differ in "+ settings[0][idx]+" "+settings[jdx][idx])
else:
retur = (True,"")
return retur
else:
return (False,"no info information at all received")
def setup(self):
StreamRequestHandler.setup(self)
error = False
try:
for retr, addr in zip(self.server.retr_sock,self.server.retr_addrs):
name, port = addr
try:
retr.connect(name,port)
self.server.log.log(logging.DEBUG,"connection to retrival server "+str(addr)+" established")
except:
# sending error message to web frontend
self.sendToWeb("failure "+str(name)+" down")
self.server.sendErrorsFrom.append(str(name))
# closing
error = True
self.server.log.log(logging.CRITICAL,"retrieval server "+str(addr)+" not reachable")
except:
pass
else:
if error:
self.finish()
raise RuntimeError, "some of the retrieval servers arent't reachable via network"
else:
self.server.notBye = True
# signal: connection to retrieval servers established
# currently unix specific linefeed
self.server.log.log(logging.INFO,"connection to retrieval server system established")
def handle(self):
retrieveAndMetaCommands = ["retrieve","expand","retrieveandsaveranks","metaretrieve","metaexpand","textretrieve","textexpand"]
# handling actions until "bye", "quit" or doomsday
while self.server.notBye and self.server.notQuit:
webMsg = self.rfile.readline()
self.server.log.log(logging.DEBUG,webMsg)
webMsg = re.sub("[\r\n\t\v\f]","",webMsg)
self.server.log.log(logging.getLevelName("RECV (Web/Bash)"),webMsg)
webKeyS = string.split(webMsg)
keyWord = webKeyS.pop(0) # grabbing the very first element
if keyWord == "info":
self.server.sendToRetrievers("info")
answer = self.server.getFromRetrievers()
# convert list of single strings in usable format
# e.g ["hello world","how do you do?"] becomes
# [['hello','world'],['how','do','you','do?']]
settings = map(lambda strin: string.split(strin),answer)
bool, comment = self._validateSettings(settings)
if bool:
sum = 0
# summing up total databank size
for idx in range(len(settings)):
sum += int(settings[idx][1])
settings[0][1] = sum
self.server.results = int(settings[0][3])
self.sendToWeb(re.sub("[[,'\]]","",str(settings[0])))
# deleting helper variables
del bool
del answer
del comment
del settings
else:
self.finish()
raise RuntimeError, "handle():"+comment
elif keyWord == "password":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
bool = True
idx = 0
while bool and idx < len(answer):
if tokens[idx][0] != "ok":
bool = False
idx += 1
if not bool:
self.sendToWeb(answer[idx-1])
self.server.auth = False
else:
self.sendToWeb("ok")
self.server.auth = True
# deleting helper variables
del answer
del tokens
del idx
del bool
elif keyWord == "bye":
self.server.notBye = False
elif keyWord == "quit":
self.server.notBye = False
if self.server.auth:
self.server.notQuit = False
else:
self.server.sendToRetrievers("bye")
elif keyWord in retrieveAndMetaCommands:
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
self.server.log.log(logging.DEBUG,"answer(handle()): "+repr(answer))
# converting retrieval servers' answers from strings to list of
# strings like above and afterwards merging all such lists into
# one
if keyWord in ["metaretrieve","metaexpand","textretrieve","textexpand"]:
listed = map(lambda string: re.split("[ :]",string),answer)
else:
listed = map(lambda strin: string.split(strin),answer)
listed = reduce(lambda fst,snd: fst+snd,listed)
if "nometainformation" in listed:
self.sendToWeb("nometainformation")
elif "notextinformation" in listed:
self.sendToWeb("notextinformation")
elif "" in listed:
self.server.log.log(logging.INFO,"at least one retrival server answered with an empty string")
self.server.log.log(logging.ERROR,"retrival aborted")
elif not self.server.results == 0:
# for sorting purposes changing data structure to a list of tupels
# consistent of distance as a float and filename as a string
toSort = [(float(listed[i+1]),listed[i]) for i in range(0,len(listed),2)]
toSort.sort()
toSort.reverse()
retrievalString = ""
for tup in toSort[:self.server.results]:
dist, name = tup
if keyWord in ["metaretrieve","metaexpand","textretrieve","textexpand"]:
retrievalString += name+":"+str(dist)
else:
retrievalString += name+" "+str(dist)
retrievalString += " "
# removing trailing blank
retrievalString = re.sub(" $","",retrievalString)
self.sendToWeb(retrievalString)
#deleting helper variables
del toSort
del retrievalString
else:
self.server.log.log(logging.INFO,"Proxyserver assumes that settings for results equals 0")
self.server.log.log(logging.INFO,"Please use setresults to set results to an usefull value")
self.server.log.log(logging.DEBUG,"possible mismath between retrieval servers settings and proxyserver")
self.server.log.log(logging.ERROR,"retrieval aborted")
# deleting helper variables
del answer
del listed
elif keyWord == "help":
self.server.sendToRetrievers(webMsg,1)
answer = self.server.retr_sock[0].getline()
ans = string.split(answer)
try:
ans.remove("filelist")
except:
pass
answer = "proxy only: quitproxy ||normal fire commands: "
for cmd in ans:
answer += cmd
answer += " "
answer += "||not supported in proxy mode: filelist"
self.sendToWeb(answer)
del answer
elif keyWord == "savedistances":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
fst = answer[0]
bool = True
for name in answer:
if name != fst:
bool = False
if not bool:
self.finish()
raise RuntimeError,"distance filenames differ"
self.sendToWeb(fst)
del answer
del bool
del fst
elif keyWord == "random":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
tokenList = reduce(lambda fst,snd:fst+snd,tokens)
# choosing randomly images from all random images
# because every retrieval server only accesses an unique database
# subset
randomImg = random.sample(tokenList,self.server.results)
self.sendToWeb(re.sub("[[',\]]","",str(randomImg)))
# deleting helper variables
del answer
del tokens
del tokenList
del randomImg
elif keyWord == "saverelevances":
self.server.sendToRetrievers(webMsg)
elif keyWord in ["setscoring","setresults","setextensions","setdist","setweight"]:
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
keyToken = re.sub("set","",keyWord)
# checking if all retrieval servers accepted new settings
# first checking keyword
boolList = map(lambda list: list[0] == keyToken,tokens)
boolAll = reduce(lambda fst,snd:fst == snd,boolList)
# checking iff snd argument must be checked
# in case it will be "=" testing is omitted
if keyToken in ["weight","dist"]:
sndArg = tokens[0][1]
boolList = map(lambda list: list[1] == sndArg,tokens)
boolSnd = reduce(lambda fst,snd: fst == snd,boolList)
boolAll = boolAll and boolSnd
del boolSnd
# checking third argument
trdArg = tokens[0][2]
boolList = map(lambda list: list[2] == trdArg,tokens)
boolTrd = reduce(lambda fst,snd:fst == snd,boolList)
boolAll = boolAll and boolTrd
if boolAll:
self.sendToWeb(answer[0])
if keyToken == "results" and tokens[0][0] != "Invalid":
self.server.results = int(tokens[0][2])
else:
self.server.sendToRetrievers(keyWord,1)
alterMsg = self.server.retr_sock[0].getline()
self.sendToWeb(alterMsg)
del alterMsg
del boolAll
del boolTrd
del boolList
del answer
del tokens
elif keyWord == "listfiles":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
result = ""
for idx in range(len(answer)):
result += answer[idx]
result += " "
# deleting trailing blank
result = re.sub(" $","",result)
self.sendToWeb(result)
del answer
del result
elif keyWord == "class":
self.server.sendToRetrievers(webMsg)
self.server.log.log(logging.INFO,webMsg+" send")
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
tokens = reduce(lambda fst,snd: fst+snd,tokens)
# since retrieval servers will only return a single value
# it is only neccessary to destinguish the three possible modes of the class command
# for error handling
# do all retrieval servers have classes or don't they
if "yes" in tokens:
for item in tokens:
if "yes" != item:
self.finish()
raise RuntimeError, "not all retrieval servers do have classes"
self.sendToWeb("yes")
elif "no" in tokens:
for item in tokens:
if "no" != item:
self.finish()
raise RuntimeError, "not all retrieval servers don't have classes"
self.sendToWeb("no")
else:
try:
intList = []
for item in tokens:
intItem = int(item)
if intItem >= 0:
intList.append(intItem)
except:
self.finish()
raise RuntimeError, "neiter yes/no nor numbers were returned for classes"
else:
boolList = map(lambda it: intList[0]==it,intList)
booL = reduce(lambda fst, snd: fst and snd,boolList)
if booL:
self.sendToWeb(str(intList[0]))
else:
self.finish()
raise RuntimeError, "retrieval servers classes differ"
del boolList
del booL
del tokens
del answer
elif keyWord == "metafeatureinfo":
self.server.sendToRetrievers(webMsg)
self.server.log.log(logging.INFO,webMsg+" send")
answer = self.server.getFromRetrievers()
answer = map(lambda string: re.sub(" $","",string),answer)
returnstring = ""
for strings in answer:
returnstring += strings
returnstring += " "
self.sendToWeb(re.sub(" $","",returnstring))
del answer
del returnstring
# Note that the commands filelist is dumped on purpose
# because every retrieval server accesses it's own databade subset
# The interactor command is dumped too.
elif keyWord in ["filelist","interactor"]:
self.sendToWeb(keyWord+" is not avaible since fireproxy is in use")
elif keyWord == "quitproxy":
self.server.notBye = False
self.server.notProxyquit = False
else:
self.sendToWeb("Unknown command: "+webMsg)
del retrieveAndMetaCommands
del webKeyS
del webMsg
| Python |
#!/usr/bin/python
from string import split, lower, strip
from random import shuffle
import sys
import os
from os.path import join, getsize
ARCH=os.environ["ARCH"]
# returns the string after a certain option string
def getStringAfter(name, array, default):
for i in range(len(array)):
if array[i] == name:
if (len(array) == i+1):
return default
else:
return array[i+1]
return default
def usage(commandline):
print "note: before you use this program make \"make all\" in the firedirectory. "
print commandline[0], "(-np|-npfromdb|-ndb|-ndbfromp|-showf|-e|-h) [options]"
print
print "mandatory options (choose exactly one of them):"
print " -h shows this help"
print " -showf show all possible features that can be extracted"
print " -e extracts features. options: -limages -ldb -ld -selectf "
print " -p -db -d -c -f -q -n"
print " -np creates a new purefiles file using the filenames of the "
print " files in the directory. options: -d -p"
print " -npfromdb creates a new purefiles file from the data of the dbfile."
print " options: -d -db -p"
print " -ndb creates a new database file using the filenames of the "
print " files in the directory. options: -d -db"
print " -ndbfromp creates a new database file from the data of the purefiles "
print " file. options: -d -p -db"
print
print "options:"
print " -d dir directory of where the files are. default: ."
print " -p path path of an [existing] purefiles file. default: purefiles"
print " -db path path of an [existing] database file. default: list"
print " -c path configfile that shall be used. "
print " default:", os.path.dirname(sys.argv[0]) + "/features.conf"
print " -f dir directory where fire resides. "
print " default:", os.path.dirname(sys.argv[0]) + "../../bin/"+ARCH+"/"
print " -q (day|week)"
print " put all jobs into the queuingsystem. default: no queuing"
print " -n name of the jobs in the queue. default: extractor"
print " -selectf feature1 [feature2 [... [featureN]]"
print " extracts only feature1 ... featureN of the config file"
print " -v verbose"
print " -limages image1 [image2 [... [imageN]]"
print " extracts features of image1 ... imageN only"
print " -ldb extracts the images from the dbfile"
print " (hint: it is smarter to create a db file first, using"
print " -ndb or -ndbfromp)"
print " -ld extracts the images of the directory"
print " (hint: it is smarter to create a purefiles file first,"
print " using -np or -npfromdb)"
class Extractor:
def __init__(self):
self.datadir = "."
self.purefiles = "purefiles"
self.dbfile = "list"
self.progdir = os.path.dirname(sys.argv[0]) # path of the tool
self.firedir = self.progdir + "/../../bin/"+ARCH+"/"
self.config = "features.conf"
self.extensionlist = ["jpg", "png", "gif", "jpeg", "tif", "tiff"]
self.images = []
self.features = {}
self.selectedFeatures = []
self.queue = None
self.name = "extractor"
# gets for the attributes
def setPurefiles(self, new):
self.purefiles = new
def setDBFile(self, new):
self.dbfile = new
def setDatadir(self, new):
self.datadir = new
def setFiredir(self, new):
self.firedir = new
def setConfigfile(self, new):
self.config = new
def setExtensionlist(self, newl):
self.extensionlist = newl
def setQueue(self, new):
self.queue = new
def setName(self, new):
self.name = new
# printing
def printInfo(self):
print "config info of extractor"
print "datadir", self.datadir
print "purefiles", self.purefiles
print "dbfile", self.dbfile
print "progdir", self.progdir
print "firedir", self.firedir
print "configfile", self.config
print "extensionlist", self.extensionlist
# prints available features
def printFeatures(self):
for feature in self.features.keys():
print feature, self.features[feature]
def printImages(self):
if self.images == []:
print "error: no images loaded."
else:
print "images:"
for image in self.images:
print image
# loading
# load images from the directory
def loadImagesFromDir(self):
print "loading images from directory."
self.images = []
for root, dirs, files in os.walk(self.datadir):
for file in files:
if self.checkExtension(file):
if (root == self.datadir):
self.images.append(file)
else:
self.images.append(root[len(self.datadir)+1:] + "/" + file)
if self.images == []:
print "error: no images found."
# loads images from database file
# (does not consider path information of the db file yet)
def loadImagesFromDB(self):
print "loading images from dbfile."
f = open(self.dbfile, "r")
lines = f.readlines()
f.close()
for line in lines:
if strip(split(line)[0]) == "file":
self.images.append(strip(split(line)[1]))
#print "file:", strip(split(line)[1])
if self.images == []:
print "error: no images found."
# loads features from the config file
def loadFeatures(self):
print "loading features file."
try:
f = open(self.config, "r")
except IOError:
print "warning: no specific config file found. I try the one in the tooldir."
f = open(self.progdir + "/features.conf")
lines = f.readlines()
f.close()
for line in lines:
if line[0] == "#":
#print "comment: ", line
pass
elif len(strip(line)) < 1:
pass
else:
key = strip(split(line)[0])
#print "key", key
feature = strip(line)[len(key)+1:]
#print "feature", feature
self.features[key] = feature
# loads an existing purefiles list
def loadImagesFromPurefiles(self):
print "loading images from purefiles."
self.images = []
file = open(self.purefiles, 'r')
lines = file.readlines()
file.close()
num = len(lines)
print "info: number of images =", num
for line in lines:
self.images.append(strip(line))
# loads images from commandline
def loadImagesFromCommandline(self, argv):
self.images = []
print "loading images from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-limages":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
self.images.append(command)
if self.images == []:
print "error: no images found"
def loadSelectedFeaturesFromCommandline(self, argv):
self.selectedFeatures = []
print "loading selected features from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-selectf":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
#print "added", command
self.selectedFeatures.append(command)
if self.selectedFeatures == []:
print "error: no features found"
# checking
# check if image file has a correct extension
def checkExtension(self, filename):
extension = split(filename, ".")[-1]
if lower(extension) in self.extensionlist:
return True
return False
def pureFileExists(self):
try:
f = open(self.purefiles, "r")
except IOError:
return False
else:
return True
# writing
# writes the purefilesfile
def writePurefiles(self):
print "writing purefiles file."
if self.images == []:
print "error: cannot write a purefiles file, because image list is empty."
else:
f = open(self.purefiles, "w")
for image in self.images:
f.write(image)
f.write("\n")
f.close()
# creates the database file
def writeDBFile(self):
# possible feature: dbfiles with feature directories?
# possible feature: non recursive traversion
print "writing dbfile."
f = open(self.dbfile, "w")
f.write("FIRE_filelist\n")
f.write("classes no\n")
f.write("descriptions no\n")
f.write("featuredirectories no\n")
f.write("path this\n")
for file in self.images:
f.write("file ")
f.write(file)
f.write("\n")
f.close()
# selection
# selects features
def selectFeatures(self):
if self.selectedFeatures == []:
print "error: there are no features to select"
print "hint: maybe wrong spelling?"
temp = {}
for key in self.features.keys():
#print "key", key
if key in self.selectedFeatures:
temp[key] = self.features[key]
self.features = temp
# extraction
# extracts all features for all images
def extractFeatures(self):
for key in self.features.keys():
self.extractFeature(key)
# extracts one feature for all images
def extractFeature(self, feature):
if self.images != []:
commandline = self.firedir + self.features[feature] + " --images"
for image in self.images:
commandline = commandline + " " + image
else:
if not self.pureFileExists():
print "error: purefiles file not existing"
print "hint: create with -np a new file or use -limage, -ldb or -ld to load images"
return False
else:
#self.printInfo()
#print "self.features[feature]", self.features[feature]
#self.printFeatures()
commandline = self.firedir + self.features[feature] + " --filelist " + self.purefiles
print "commandline:", commandline
if self.queue == "day":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 8:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
elif self.queue == "week":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 24:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
else:
os.system(commandline)
if __name__ == "__main__":
if "-h" in sys.argv:
usage(sys.argv)
else:
e = Extractor()
# adjusting file locations
if "-p" in sys.argv:
e.setPurefiles(getStringAfter("-p", sys.argv, None))
if "-db" in sys.argv:
e.setDBFile(getStringAfter("-db", sys.argv, None))
if "-d" in sys.argv:
e.setDatadir(getStringAfter("-d", sys.argv, None))
if "-c" in sys.argv:
e.setConfigfile(getStringAfter("-c", sys.argv, None))
if "-f" in sys.argv:
e.setFiredir(getStringAfter("-f", sys.argv, None))
if "-q" in sys.argv:
e.setQueue(getStringAfter("-q", sys.argv, None))
if "-n" in sys.argv:
e.setName(getStringAfter("-n", sys.argv, None))
## maybe later
# if "-el" in sys.argv
# e.setExtensionlist(...)
if "-v" in sys.argv:
e.printInfo()
if "-np" in sys.argv:
e.loadImagesFromDir()
e.writePurefiles()
elif "-npfromdb" in sys.argv:
e.loadImagesFromDB()
e.writePurefiles()
elif "-ndb" in sys.argv:
e.loadImagesFromDir()
e.writeDBFile()
elif "-ndbfromp" in sys.argv:
e.loadImagesFromPurefiles()
e.writeDBFile()
elif "-showf" in sys.argv:
e.loadFeatures()
e.printFeatures()
elif "-e" in sys.argv:
# loading the images (if no purefiles file is used)
if "-limages" in sys.argv:
e.loadImagesFromCommandline(sys.argv)
if "-v" in sys.argv:
e.printImages()
elif "-ldb" in sys.argv:
e.loadImagesFromDBFile()
elif "-ld" in sys.argv:
e.loadImagesFromDatadir()
else:
pass
e.loadFeatures()
if "-selectf" in sys.argv:
e.loadSelectedFeaturesFromCommandline(sys.argv)
e.selectFeatures()
if "-v" in sys.argv:
print "selected features:"
e.printFeatures()
e.extractFeatures()
else:
print "wrong usage. use -h for help."
| Python |
#!/usr/bin/python
from string import split, lower, strip
from random import shuffle
import sys
import os
from os.path import join, getsize
ARCH=os.environ["ARCH"]
# returns the string after a certain option string
def getStringAfter(name, array, default):
for i in range(len(array)):
if array[i] == name:
if (len(array) == i+1):
return default
else:
return array[i+1]
return default
def usage(commandline):
print "note: before you use this program make \"make all\" in the firedirectory. "
print commandline[0], "(-np|-npfromdb|-ndb|-ndbfromp|-showf|-e|-h) [options]"
print
print "mandatory options (choose exactly one of them):"
print " -h shows this help"
print " -showf show all possible features that can be extracted"
print " -e extracts features. options: -limages -ldb -ld -selectf "
print " -p -db -d -c -f -q -n"
print " -np creates a new purefiles file using the filenames of the "
print " files in the directory. options: -d -p"
print " -npfromdb creates a new purefiles file from the data of the dbfile."
print " options: -d -db -p"
print " -ndb creates a new database file using the filenames of the "
print " files in the directory. options: -d -db"
print " -ndbfromp creates a new database file from the data of the purefiles "
print " file. options: -d -p -db"
print
print "options:"
print " -d dir directory of where the files are. default: ."
print " -p path path of an [existing] purefiles file. default: purefiles"
print " -db path path of an [existing] database file. default: list"
print " -c path configfile that shall be used. "
print " default:", os.path.dirname(sys.argv[0]) + "/features.conf"
print " -f dir directory where fire resides. "
print " default:", os.path.dirname(sys.argv[0]) + "../../bin/"+ARCH+"/"
print " -q (day|week)"
print " put all jobs into the queuingsystem. default: no queuing"
print " -n name of the jobs in the queue. default: extractor"
print " -selectf feature1 [feature2 [... [featureN]]"
print " extracts only feature1 ... featureN of the config file"
print " -v verbose"
print " -limages image1 [image2 [... [imageN]]"
print " extracts features of image1 ... imageN only"
print " -ldb extracts the images from the dbfile"
print " (hint: it is smarter to create a db file first, using"
print " -ndb or -ndbfromp)"
print " -ld extracts the images of the directory"
print " (hint: it is smarter to create a purefiles file first,"
print " using -np or -npfromdb)"
class Extractor:
def __init__(self):
self.datadir = "."
self.purefiles = "purefiles"
self.dbfile = "list"
self.progdir = os.path.dirname(sys.argv[0]) # path of the tool
self.firedir = self.progdir + "/../../bin/"+ARCH+"/"
self.config = "features.conf"
self.extensionlist = ["jpg", "png", "gif", "jpeg", "tif", "tiff"]
self.images = []
self.features = {}
self.selectedFeatures = []
self.queue = None
self.name = "extractor"
# gets for the attributes
def setPurefiles(self, new):
self.purefiles = new
def setDBFile(self, new):
self.dbfile = new
def setDatadir(self, new):
self.datadir = new
def setFiredir(self, new):
self.firedir = new
def setConfigfile(self, new):
self.config = new
def setExtensionlist(self, newl):
self.extensionlist = newl
def setQueue(self, new):
self.queue = new
def setName(self, new):
self.name = new
# printing
def printInfo(self):
print "config info of extractor"
print "datadir", self.datadir
print "purefiles", self.purefiles
print "dbfile", self.dbfile
print "progdir", self.progdir
print "firedir", self.firedir
print "configfile", self.config
print "extensionlist", self.extensionlist
# prints available features
def printFeatures(self):
for feature in self.features.keys():
print feature, self.features[feature]
def printImages(self):
if self.images == []:
print "error: no images loaded."
else:
print "images:"
for image in self.images:
print image
# loading
# load images from the directory
def loadImagesFromDir(self):
print "loading images from directory."
self.images = []
for root, dirs, files in os.walk(self.datadir):
for file in files:
if self.checkExtension(file):
if (root == self.datadir):
self.images.append(file)
else:
self.images.append(root[len(self.datadir)+1:] + "/" + file)
if self.images == []:
print "error: no images found."
# loads images from database file
# (does not consider path information of the db file yet)
def loadImagesFromDB(self):
print "loading images from dbfile."
f = open(self.dbfile, "r")
lines = f.readlines()
f.close()
for line in lines:
if strip(split(line)[0]) == "file":
self.images.append(strip(split(line)[1]))
#print "file:", strip(split(line)[1])
if self.images == []:
print "error: no images found."
# loads features from the config file
def loadFeatures(self):
print "loading features file."
try:
f = open(self.config, "r")
except IOError:
print "warning: no specific config file found. I try the one in the tooldir."
f = open(self.progdir + "/features.conf")
lines = f.readlines()
f.close()
for line in lines:
if line[0] == "#":
#print "comment: ", line
pass
elif len(strip(line)) < 1:
pass
else:
key = strip(split(line)[0])
#print "key", key
feature = strip(line)[len(key)+1:]
#print "feature", feature
self.features[key] = feature
# loads an existing purefiles list
def loadImagesFromPurefiles(self):
print "loading images from purefiles."
self.images = []
file = open(self.purefiles, 'r')
lines = file.readlines()
file.close()
num = len(lines)
print "info: number of images =", num
for line in lines:
self.images.append(strip(line))
# loads images from commandline
def loadImagesFromCommandline(self, argv):
self.images = []
print "loading images from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-limages":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
self.images.append(command)
if self.images == []:
print "error: no images found"
def loadSelectedFeaturesFromCommandline(self, argv):
self.selectedFeatures = []
print "loading selected features from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-selectf":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
#print "added", command
self.selectedFeatures.append(command)
if self.selectedFeatures == []:
print "error: no features found"
# checking
# check if image file has a correct extension
def checkExtension(self, filename):
extension = split(filename, ".")[-1]
if lower(extension) in self.extensionlist:
return True
return False
def pureFileExists(self):
try:
f = open(self.purefiles, "r")
except IOError:
return False
else:
return True
# writing
# writes the purefilesfile
def writePurefiles(self):
print "writing purefiles file."
if self.images == []:
print "error: cannot write a purefiles file, because image list is empty."
else:
f = open(self.purefiles, "w")
for image in self.images:
f.write(image)
f.write("\n")
f.close()
# creates the database file
def writeDBFile(self):
# possible feature: dbfiles with feature directories?
# possible feature: non recursive traversion
print "writing dbfile."
f = open(self.dbfile, "w")
f.write("FIRE_filelist\n")
f.write("classes no\n")
f.write("descriptions no\n")
f.write("featuredirectories no\n")
f.write("path this\n")
for file in self.images:
f.write("file ")
f.write(file)
f.write("\n")
f.close()
# selection
# selects features
def selectFeatures(self):
if self.selectedFeatures == []:
print "error: there are no features to select"
print "hint: maybe wrong spelling?"
temp = {}
for key in self.features.keys():
#print "key", key
if key in self.selectedFeatures:
temp[key] = self.features[key]
self.features = temp
# extraction
# extracts all features for all images
def extractFeatures(self):
for key in self.features.keys():
self.extractFeature(key)
# extracts one feature for all images
def extractFeature(self, feature):
if self.images != []:
commandline = self.firedir + self.features[feature] + " --images"
for image in self.images:
commandline = commandline + " " + image
else:
if not self.pureFileExists():
print "error: purefiles file not existing"
print "hint: create with -np a new file or use -limage, -ldb or -ld to load images"
return False
else:
#self.printInfo()
#print "self.features[feature]", self.features[feature]
#self.printFeatures()
commandline = self.firedir + self.features[feature] + " --filelist " + self.purefiles
print "commandline:", commandline
if self.queue == "day":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 8:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
elif self.queue == "week":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 24:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
else:
os.system(commandline)
if __name__ == "__main__":
if "-h" in sys.argv:
usage(sys.argv)
else:
e = Extractor()
# adjusting file locations
if "-p" in sys.argv:
e.setPurefiles(getStringAfter("-p", sys.argv, None))
if "-db" in sys.argv:
e.setDBFile(getStringAfter("-db", sys.argv, None))
if "-d" in sys.argv:
e.setDatadir(getStringAfter("-d", sys.argv, None))
if "-c" in sys.argv:
e.setConfigfile(getStringAfter("-c", sys.argv, None))
if "-f" in sys.argv:
e.setFiredir(getStringAfter("-f", sys.argv, None))
if "-q" in sys.argv:
e.setQueue(getStringAfter("-q", sys.argv, None))
if "-n" in sys.argv:
e.setName(getStringAfter("-n", sys.argv, None))
## maybe later
# if "-el" in sys.argv
# e.setExtensionlist(...)
if "-v" in sys.argv:
e.printInfo()
if "-np" in sys.argv:
e.loadImagesFromDir()
e.writePurefiles()
elif "-npfromdb" in sys.argv:
e.loadImagesFromDB()
e.writePurefiles()
elif "-ndb" in sys.argv:
e.loadImagesFromDir()
e.writeDBFile()
elif "-ndbfromp" in sys.argv:
e.loadImagesFromPurefiles()
e.writeDBFile()
elif "-showf" in sys.argv:
e.loadFeatures()
e.printFeatures()
elif "-e" in sys.argv:
# loading the images (if no purefiles file is used)
if "-limages" in sys.argv:
e.loadImagesFromCommandline(sys.argv)
if "-v" in sys.argv:
e.printImages()
elif "-ldb" in sys.argv:
e.loadImagesFromDBFile()
elif "-ld" in sys.argv:
e.loadImagesFromDatadir()
else:
pass
e.loadFeatures()
if "-selectf" in sys.argv:
e.loadSelectedFeaturesFromCommandline(sys.argv)
e.selectFeatures()
if "-v" in sys.argv:
print "selected features:"
e.printFeatures()
e.extractFeatures()
else:
print "wrong usage. use -h for help."
| Python |
#!/usr/bin/env python
import sys,gzip
class Feature:
def __init__(self):
self.posx=0
self.posy=0
self.scale=0
self.vec=[]
class LocalFeatures:
def __init__(self):
self.winsize=0
self.dim=0
self.subsampling=0
self.padding=0
self.numberOfFeatures=0
self.varthreshold=0
self.zsize=0
self.filename=""
self.imagesize=(0,0)
self.features=[]
self.selffilename=""
def __getitem__(self,i):
return self.features[i]
def load(self,filename):
f=gzip.GzipFile(filename)
self.selffilename=filename
for l in f:
toks=l.split()
if toks[0]=="FIRE_localfeatures":
fine=True
elif toks[0]=="winsize":
self.winsize=int(toks[1])
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="subsampling":
self.subsampling=int(toks[1])
elif toks[0]=="padding":
self.padding=int(toks[1])
elif toks[0]=="numberOfFeatures":
self.numberOfFeatures=int(toks[1])
elif toks[0]=="varthreshold":
self.varthreshold=float(toks[1])
elif toks[0]=="zsize":
self.zsize=int(toks[1])
elif toks[0]=="filename":
self.filename=toks[1]
elif toks[0]=="imagesize":
self.imagesize=(int(toks[1]),int(toks[2]))
elif toks[0]=="features":
if self.numberOfFeatures!=int(toks[1]):
print "Weird... inconsistent number of Features and features expected in the file"
elif toks[0]=="feature":
f=Feature()
toks.pop(0)
f.posx=float(toks.pop(0))
f.posy=float(toks.pop(0))
f.scale=float(toks.pop(0))
f.vec=map(lambda x: float(x), toks)
self.features+=[f]
else:
print "'%s' is an unknown keyword... ignoring and continuiing nonetheless."%(toks[0])
def save(self,filename):
f=gzip.GzipFile(filename,"w")
f.write("FIRE_localfeatures\n")
f.write("winsize %d\ndim %d\nsubsampling %d\npadding %d\nnumberOfFeatures %d\nvarthreshold %d\nzsize %d\nfilename %s\nimagesize %d %d\nfeatures %d\n" % (self.winsize,self.dim,self.subsampling,self.padding,self.numberOfFeatures,self.varthreshold,self.zsize,self.filename,self.imagesize[0],self.imagesize[1],len(self.features)))
for lf in self.features:
f.write("feature %d %d %d " % (lf.posx,lf.posy,lf.scale))
for v in lf.vec:
f.write(str(v)+" ")
f.write("\n")
f.close()
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class HistogramFeature:
def __init__(self):
self.steps=[]
self.bins=[]
self.min=[]
self.max=[]
self.stepsize=[]
self.dim=0
self.counter=0
self.filename=""
self.data=[]
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_histogram"
print >>f,"# created with histogramfeature.py"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps"," ".join(map(lambda x: str(x),self.steps))
print >>f,"min"," ".join(map(lambda x: str(x), self.min))
print >>f,"max"," ".join(map(lambda x: str(x), self.max))
print >>f,"data"," ".join(map(lambda x: str(x), self.data))
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_histogram":
fine=True
elif toks[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="data":
self.bins=map(lambda x: int(x),toks[1:])
try:
self.data=map(lambda x: float(x)/float(self.counter), self.bins)
except:
print "Error reading file:", filename,"probably division by zero, continuing anyway."
else:
print "Unparsed line: '"+line+"'."
| Python |
#!/usr/bin/python
import sys
#this script merges fire distance files.
#
# as command line parameters, arbitrary many distance files are given and
# the merged distance file is printed.
infiles=[]
line=[]
for i in sys.argv[1:]:
infiles.append(open(i,"r"))
line.append("")
line[0]=infiles[0].readline()
count=0
while(line[0]):
for i in range(1,len(infiles)):
line[i]=infiles[i].readline()
if line[0].startswith("nofdistances"):
nDist=0
nImg=0
for l in line:
toks=l.split()
nDist+=int(toks[2])
nImg=int(toks[1])
print "nofdistances",nImg,nDist
elif not line[0].startswith("#"):
print count,
for l in line:
toks=l.split()
for i in range(1,len(toks)):
print float(toks[i]),
print
count+=1
line[0]=infiles[0].readline()
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
querylist=filelist.FileList()
filelist=filelist.FileList()
filelist.load(sys.argv[1])
querylist.load(sys.argv[2])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in querylist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
if len(sys.argv) != 4:
print """USAGE:
querylistwqueryexpansion.py <querylist> <server> <port>
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
files=ql.readlines()
ql.close()
s.sendcmd("setresults 10")
s.sendcmd("setextensions 2")
for file in files:
file=re.sub('\n','',file)
cmd="retrieveandsaveranks 1030 "+file+".ranks "+file
s.sendcmd(cmd)
res=s.getline()
print res
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback,numarray
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-f":
database=sys.argv.pop(0)
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-l1o":
l1o=True
elif argv=="-x":
quit=True
else:
print "Unknown option:",argv
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
if database=="" or querybase=="" and not l1o:
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
sys.stdout.flush()
f=filelist.FileList()
f.load(database)
q=filelist.FileList()
if not l1o:
q.load(querybase)
else:
q=f
if not f.classes:
print "Need classes in database file"
sys.exit(10)
if not l1o and not q.classes:
print "Need classes in querybase file"
sys.exit(10)
s=firesocket.FIRESocket()
s.connect(server,port)
try:
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
keyword=status.pop(0)
dbSize=int(status.pop(0))
if dbSize!=len(f.files):
print "database in retriever and database loaded are of different size:", dbSize,"!=",len(f.files)
s.sendcmd("bye")
time.sleep(1)
sys.exit(10)
print "database has",dbSize,"images."
if not l1o:
s.sendcmd("setresults 1"); res=s.getline()
else:
s.sendcmd("setresults 2"); res=s.getline()
classified=0; errors=0; correct=0
no_classes=0
for qf in q.files:
if qf[1]>no_classes:
no_classes=qf[1]
for df in f.files:
if df[1]>no_classes:
no_classes=df[1]
no_classes+=1
print no_classes,"classes."
confusionmatrix=numarray.zeros((no_classes,no_classes),numarray.Int)
counter=0
for qf in q.files:
counter+=1
qname=qf[0]
qcls=qf[1]
s.sendcmd("retrieve "+qname)
res=s.getline()
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
if not l1o:
if len(results)!=2:
print "expected 2 tokens, got",len(results)
sys.exit(10)
else:
if len(results)!=4:
print "expected 4 tokens, got",len(results)
sys.exit(10)
results.pop(0)
results.pop(0)
returnedimage=results.pop(0)
score=float(results.pop(0))
for df in f.files:
if not l1o:
if df[0]==returnedimage:
dcls=df[1]
break
else:
if df[0]==returnedimage and df[0]!=qname:
dcls=df[1]
break
classified+=1
if dcls==qcls:
correct+=1
else:
errors+=1
confusionmatrix[dcls][qcls]+=1
print "Query "+str(counter)+"/"+str(len(q.files))+":",qname,"("+str(qcls)+") NN:",returnedimage,"("+str(dcls)+")","ER:",float(errors)/float(classified)*100
sys.stdout.flush()
print "RESULT: ER:",float(errors)/float(classified)*100
print confusionmatrix
time.sleep(1)
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
useX="-x" in sys.argv
verbose="-v" in sys.argv
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
fileprefix=""
if len(sys.argv)>4:
if sys.argv[4]!="-q" and sys.argv[4]!="-x" and sys.argv[4]!="-v":
fileprefix=sys.argv[4]
lines=ql.readlines()
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
ql.close()
classified=0
###query images and evaluate results
for line in lines:
line=re.sub('\n','',line)
relevant=re.split(" ",line)
query=fileprefix+relevant[0] # this is the name of the query image
relevant=relevant[1:] # this are the names of the images relevant to the query image
cmd="retrieve +"+query
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimage=re.sub(fileprefix,"",returnedimage)
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
if verbose:
print "results:",returnedimages[0],"->0",
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
param["useX"] = useX
param["verbose"] = verbose
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
useX = param["useX"]
verbose = param["verbose"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
if useX:
g=Gnuplot.Gnuplot()
g.title('PR graph')
g('set data style linespoints')
g('set xrange [0.0:1.0]')
g('set yrange [0.0:1.0]')
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
if verbose:
print returnedimages[Ri],"->",Ri,
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
if useX:
g.plot(Gnuplot.Data(PRgraph,title="PRgraph current query"),Gnuplot.Data(tmpAPRgraph,title="PRgraph average"))
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [<fileprefix>] [-q] [-x]
-q: quit the server after finished eval
-x: show PR graph after each query on X display
-v: verbose
<fileprefix> should not be necessary
"""
else:
param = {}
retrieveData(param)
calculate(param)
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
host=sys.argv[1]
if len(sys.argv) >=3:
port=int(sys.argv[2])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
s.sendcmd("listfiles")
fileline=s.getline()
files=string.split(fileline," ")
s.sendcmd("setresults 2")
s.getline()
classified=0
correct=0
error=0
for file in files:
if len(file)>0:
cmd="retrieve +"+file
s.sendcmd(cmd)
res=s.getline()
parts=string.split(res," ")
result=parts[2]
print res
#get classes
s.sendcmd("class "+file)
qclass=s.getline()
s.sendcmd("class "+result)
rclass=s.getline()
qc=int(qclass)
rc=int(rclass)
if(rc==qc):
correct=correct+1
else:
error=error+1
classified=classified+1
er=float(error)/float(classified)*100.0
print file+" wrong="+str(error)+" correct="+str(correct)+" classified="+str(classified)+" ER="+str(er)
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
filelist=filelist.FileList()
filelist.load(sys.argv[1])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in filelist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class SparseHistogramFeature:
def __init__(self):
self.bins={}
self.data={}
self.steps=[]
self.stepSize=[]
self.min=[]
self.max=[]
self.dim=0
self.numberofBins=0
self.counter=0
self.filename=""
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_sparse_histogram"
print >>f,"# sparse histogram file for FireV2"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps",reduce(lambda x,y: str(x)+" "+str(y),self.steps)
print >>f,"min",reduce(lambda x,y: str(x)+" "+str(y),self.min)
print >>f,"max",reduce(lambda x,y: str(x)+" "+str(y),self.max)
print >>f,"bins",len(self.bins)
for i in self.bins:
print >>f,"data",i,self.bins[i]
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_sparse_histogram":
fine=True
elif line[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="bins":
self.numberofBins=int(toks[1])
elif toks[0]=="data":
self.bins[toks[1]]=int(toks[2])
self.data[toks[1]]=float(toks[2])/float(self.counter)
else:
print "Unparsed line: '"+line+"'."
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
#-----------------------------------------------------------------
s=firesocket.FIRESocket()
try:
if len(sys.argv) < 5:
print """USAGE:
makedistfiles.py <querylist> <server> <port> <suffix> [-q]
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
suffix=sys.argv[4]
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
lines=ql.readlines()
lines=map(lambda line:
re.sub("\n","",line), lines)
ql.close()
i=0
for line in lines:
print i,"/",len(lines)
i+=1
tokens=re.split(" ",line)
file=tokens[0]
cmd="savedistances "+file+" "+file+"."+suffix
print "SENDING: ",cmd
s.sendcmd(cmd)
res=s.getline()
print res
sys.stdout.flush()
sys.stdout.flush()
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
classified=0
###query images and evaluate results
for line in ql:
[query,relevant]=line.split('#')
relevant=relevant.strip()
query=query.strip()
cmd="retrieve "+query
relevant=relevant.split()
print "SEND:'"+cmd+"'."
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [-q]
-q: quit the server after finished eval
"""
else:
param = {}
# do the queries
retrieveData(param)
# querying finished, now evaluated the gathered data
calculate(param)
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback
#sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
start=-1
stop=-1
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-x":
quit=True
elif argv=="-S":
suffix=sys.argv.pop(0)
elif argv=="-start":
start=int(sys.argv.pop(0))
elif argv=="-stop":
stop=int(sys.argv.pop(0))
else:
print "Unknown option:",argv
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist, -1 for full database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"querybase=",querybase
if querybase=="":
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist
-x exit after having finished
"""
sys.exit(10)
sys.stdout.flush()
q=filelist.FileList()
q.load(querybase)
s=firesocket.FIRESocket()
s.connect(server,port)
if start==-1:
start=0
if stop==-1:
stop=len(q.files)
try:
for qf in q.files[start:stop]:
cmd="savedistances "+qf[0]+" "+qf[0]+"."+suffix
s.sendcmd(cmd)
res=s.getline()
print res
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/python
import sys,re
import filelist
filelist=filelist.FileList()
filelist.load(sys.argv[1])
for i in filelist:
cls=i[1]
fname=i[0]
print fname,
for j in filelist:
if j[1]==cls and j[0]!=fname:
print j[0],
print
#print filelist
| Python |
#!/usr/bin/python
import sys,re,filelist
if len(sys.argv)<3:
print """USAGE filelistquery2querylist.py <querylist> <dblist>"""
sys.exit(5)
qlist=filelist.FileList()
qlist.load(sys.argv[1])
dblist=filelist.FileList()
dblist.load(sys.argv[2])
for i in qlist:
cls=i[1]
fname=i[0]
print fname,
for j in dblist:
if j[1]==cls:
print j[0],
print
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
useX="-x" in sys.argv
verbose="-v" in sys.argv
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
fileprefix=""
if len(sys.argv)>4:
if sys.argv[4]!="-q" and sys.argv[4]!="-x" and sys.argv[4]!="-v":
fileprefix=sys.argv[4]
lines=ql.readlines()
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
ql.close()
classified=0
###query images and evaluate results
for line in lines:
line=re.sub('\n','',line)
relevant=re.split(" ",line)
query=fileprefix+relevant[0] # this is the name of the query image
relevant=relevant[1:] # this are the names of the images relevant to the query image
cmd="retrieve +"+query
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimage=re.sub(fileprefix,"",returnedimage)
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
if verbose:
print "results:",returnedimages[0],"->0",
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
param["useX"] = useX
param["verbose"] = verbose
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
useX = param["useX"]
verbose = param["verbose"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
if useX:
g=Gnuplot.Gnuplot()
g.title('PR graph')
g('set data style linespoints')
g('set xrange [0.0:1.0]')
g('set yrange [0.0:1.0]')
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
if verbose:
print returnedimages[Ri],"->",Ri,
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
if useX:
g.plot(Gnuplot.Data(PRgraph,title="PRgraph current query"),Gnuplot.Data(tmpAPRgraph,title="PRgraph average"))
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [<fileprefix>] [-q] [-x]
-q: quit the server after finished eval
-x: show PR graph after each query on X display
-v: verbose
<fileprefix> should not be necessary
"""
else:
param = {}
retrieveData(param)
calculate(param)
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
host=sys.argv[1]
if len(sys.argv) >=3:
port=int(sys.argv[2])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
s.sendcmd("listfiles")
fileline=s.getline()
files=string.split(fileline," ")
s.sendcmd("setresults 2")
s.getline()
classified=0
correct=0
error=0
for file in files:
if len(file)>0:
cmd="retrieve +"+file
s.sendcmd(cmd)
res=s.getline()
parts=string.split(res," ")
result=parts[2]
print res
#get classes
s.sendcmd("class "+file)
qclass=s.getline()
s.sendcmd("class "+result)
rclass=s.getline()
qc=int(qclass)
rc=int(rclass)
if(rc==qc):
correct=correct+1
else:
error=error+1
classified=classified+1
er=float(error)/float(classified)*100.0
print file+" wrong="+str(error)+" correct="+str(correct)+" classified="+str(classified)+" ER="+str(er)
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
if len(sys.argv) != 4:
print """USAGE:
querylistwqueryexpansion.py <querylist> <server> <port>
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
files=ql.readlines()
ql.close()
s.sendcmd("setresults 10")
s.sendcmd("setextensions 2")
for file in files:
file=re.sub('\n','',file)
cmd="retrieveandsaveranks 1030 "+file+".ranks "+file
s.sendcmd(cmd)
res=s.getline()
print res
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys,re,filelist
if len(sys.argv)<3:
print """USAGE filelistquery2querylist.py <querylist> <dblist>"""
sys.exit(5)
qlist=filelist.FileList()
qlist.load(sys.argv[1])
dblist=filelist.FileList()
dblist.load(sys.argv[2])
for i in qlist:
cls=i[1]
fname=i[0]
print fname,
for j in dblist:
if j[1]==cls:
print j[0],
print
| Python |
import socket, os, sys,re
# ----------------------------------------------------------------------
# socket implementation for comunication with the server system
# running on any computer which is reachable via network sockets
# here we need mainly the functions
# * connect
# * sendcmd
# * getline
# * close
# for the communication
# ----------------------------------------------------------------------
class FIRESocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
#self.sock.settimeout(10.0)
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
#print "PUT:",msg; sys.stdout.flush()
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
#print "GOTR:",result; sys.stdout.flush()
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk!= '\r' and chunk != '\n':
msg=msg+chunk
msg=re.sub("[ ]$","",msg)
#print "GOTL:",msg; sys.stdout.flush()
return msg
def flush(self):
chunk=self.sock.recv(5000)
return
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
querylist=filelist.FileList()
filelist=filelist.FileList()
filelist.load(sys.argv[1])
querylist.load(sys.argv[2])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in querylist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback
#sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
start=-1
stop=-1
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-x":
quit=True
elif argv=="-S":
suffix=sys.argv.pop(0)
elif argv=="-start":
start=int(sys.argv.pop(0))
elif argv=="-stop":
stop=int(sys.argv.pop(0))
else:
print "Unknown option:",argv
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist, -1 for full database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"querybase=",querybase
if querybase=="":
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist
-x exit after having finished
"""
sys.exit(10)
sys.stdout.flush()
q=filelist.FileList()
q.load(querybase)
s=firesocket.FIRESocket()
s.connect(server,port)
if start==-1:
start=0
if stop==-1:
stop=len(q.files)
try:
for qf in q.files[start:stop]:
cmd="savedistances "+qf[0]+" "+qf[0]+"."+suffix
s.sendcmd(cmd)
res=s.getline()
print res
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class HistogramFeature:
def __init__(self):
self.steps=[]
self.bins=[]
self.min=[]
self.max=[]
self.stepsize=[]
self.dim=0
self.counter=0
self.filename=""
self.data=[]
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_histogram"
print >>f,"# created with histogramfeature.py"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps"," ".join(map(lambda x: str(x),self.steps))
print >>f,"min"," ".join(map(lambda x: str(x), self.min))
print >>f,"max"," ".join(map(lambda x: str(x), self.max))
print >>f,"data"," ".join(map(lambda x: str(x), self.data))
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_histogram":
fine=True
elif toks[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="data":
self.bins=map(lambda x: int(x),toks[1:])
try:
self.data=map(lambda x: float(x)/float(self.counter), self.bins)
except:
print "Error reading file:", filename,"probably division by zero, continuing anyway."
else:
print "Unparsed line: '"+line+"'."
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
classified=0
###query images and evaluate results
for line in ql:
[query,relevant]=line.split('#')
relevant=relevant.strip()
query=query.strip()
cmd="retrieve "+query
relevant=relevant.split()
print "SEND:'"+cmd+"'."
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [-q]
-q: quit the server after finished eval
"""
else:
param = {}
# do the queries
retrieveData(param)
# querying finished, now evaluated the gathered data
calculate(param)
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback,numarray
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-f":
database=sys.argv.pop(0)
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-l1o":
l1o=True
elif argv=="-x":
quit=True
else:
print "Unknown option:",argv
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
if database=="" or querybase=="" and not l1o:
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
sys.stdout.flush()
f=filelist.FileList()
f.load(database)
q=filelist.FileList()
if not l1o:
q.load(querybase)
else:
q=f
if not f.classes:
print "Need classes in database file"
sys.exit(10)
if not l1o and not q.classes:
print "Need classes in querybase file"
sys.exit(10)
s=firesocket.FIRESocket()
s.connect(server,port)
try:
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
keyword=status.pop(0)
dbSize=int(status.pop(0))
if dbSize!=len(f.files):
print "database in retriever and database loaded are of different size:", dbSize,"!=",len(f.files)
s.sendcmd("bye")
time.sleep(1)
sys.exit(10)
print "database has",dbSize,"images."
if not l1o:
s.sendcmd("setresults 1"); res=s.getline()
else:
s.sendcmd("setresults 2"); res=s.getline()
classified=0; errors=0; correct=0
no_classes=0
for qf in q.files:
if qf[1]>no_classes:
no_classes=qf[1]
for df in f.files:
if df[1]>no_classes:
no_classes=df[1]
no_classes+=1
print no_classes,"classes."
confusionmatrix=numarray.zeros((no_classes,no_classes),numarray.Int)
counter=0
for qf in q.files:
counter+=1
qname=qf[0]
qcls=qf[1]
s.sendcmd("retrieve "+qname)
res=s.getline()
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
if not l1o:
if len(results)!=2:
print "expected 2 tokens, got",len(results)
sys.exit(10)
else:
if len(results)!=4:
print "expected 4 tokens, got",len(results)
sys.exit(10)
results.pop(0)
results.pop(0)
returnedimage=results.pop(0)
score=float(results.pop(0))
for df in f.files:
if not l1o:
if df[0]==returnedimage:
dcls=df[1]
break
else:
if df[0]==returnedimage and df[0]!=qname:
dcls=df[1]
break
classified+=1
if dcls==qcls:
correct+=1
else:
errors+=1
confusionmatrix[dcls][qcls]+=1
print "Query "+str(counter)+"/"+str(len(q.files))+":",qname,"("+str(qcls)+") NN:",returnedimage,"("+str(dcls)+")","ER:",float(errors)/float(classified)*100
sys.stdout.flush()
print "RESULT: ER:",float(errors)/float(classified)*100
print confusionmatrix
time.sleep(1)
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/env python
import sys,gzip
class Feature:
def __init__(self):
self.posx=0
self.posy=0
self.scale=0
self.vec=[]
class LocalFeatures:
def __init__(self):
self.winsize=0
self.dim=0
self.subsampling=0
self.padding=0
self.numberOfFeatures=0
self.varthreshold=0
self.zsize=0
self.filename=""
self.imagesize=(0,0)
self.features=[]
self.selffilename=""
def __getitem__(self,i):
return self.features[i]
def load(self,filename):
f=gzip.GzipFile(filename)
self.selffilename=filename
for l in f:
toks=l.split()
if toks[0]=="FIRE_localfeatures":
fine=True
elif toks[0]=="winsize":
self.winsize=int(toks[1])
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="subsampling":
self.subsampling=int(toks[1])
elif toks[0]=="padding":
self.padding=int(toks[1])
elif toks[0]=="numberOfFeatures":
self.numberOfFeatures=int(toks[1])
elif toks[0]=="varthreshold":
self.varthreshold=float(toks[1])
elif toks[0]=="zsize":
self.zsize=int(toks[1])
elif toks[0]=="filename":
self.filename=toks[1]
elif toks[0]=="imagesize":
self.imagesize=(int(toks[1]),int(toks[2]))
elif toks[0]=="features":
if self.numberOfFeatures!=int(toks[1]):
print "Weird... inconsistent number of Features and features expected in the file"
elif toks[0]=="feature":
f=Feature()
toks.pop(0)
f.posx=float(toks.pop(0))
f.posy=float(toks.pop(0))
f.scale=float(toks.pop(0))
f.vec=map(lambda x: float(x), toks)
self.features+=[f]
else:
print "'%s' is an unknown keyword... ignoring and continuiing nonetheless."%(toks[0])
def save(self,filename):
f=gzip.GzipFile(filename,"w")
f.write("FIRE_localfeatures\n")
f.write("winsize %d\ndim %d\nsubsampling %d\npadding %d\nnumberOfFeatures %d\nvarthreshold %d\nzsize %d\nfilename %s\nimagesize %d %d\nfeatures %d\n" % (self.winsize,self.dim,self.subsampling,self.padding,self.numberOfFeatures,self.varthreshold,self.zsize,self.filename,self.imagesize[0],self.imagesize[1],len(self.features)))
for lf in self.features:
f.write("feature %d %d %d " % (lf.posx,lf.posy,lf.scale))
for v in lf.vec:
f.write(str(v)+" ")
f.write("\n")
f.close()
| Python |
#!/usr/bin/python
import sys
#this script merges fire distance files.
#
# as command line parameters, arbitrary many distance files are given and
# the merged distance file is printed.
infiles=[]
line=[]
for i in sys.argv[1:]:
infiles.append(open(i,"r"))
line.append("")
line[0]=infiles[0].readline()
count=0
while(line[0]):
for i in range(1,len(infiles)):
line[i]=infiles[i].readline()
if line[0].startswith("nofdistances"):
nDist=0
nImg=0
for l in line:
toks=l.split()
nDist+=int(toks[2])
nImg=int(toks[1])
print "nofdistances",nImg,nDist
elif not line[0].startswith("#"):
print count,
for l in line:
toks=l.split()
for i in range(1,len(toks)):
print float(toks[i]),
print
count+=1
line[0]=infiles[0].readline()
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
import sys,cgi,re,Image,Gnuplot,string,os
sys.path+=[os.getcwd()]
from types import *
from firesocket import *
try:
form=cgi.FieldStorage()
if form.has_key("server"):
server=form["server"].value
else:
server="localhost"
if form.has_key("port"):
port=int(form["port"].value)
else:
port=12960
if form.has_key("image"):
image=form["image"].value
else:
image="100.jpg"
if form.has_key("feature"):
feature=form["feature"].value
else:
feature="1"
s=FIRESocket()
s.connect(server,port)
s.sendcmd("feature "+image+" "+feature)
line=s.getline()
suffix=string.split(line)[1]
lines=[]
line=s.getline()
while line!="end":
lines+=[line]
line=s.getline()
s.sendcmd("bye")
if suffix.endswith(".histo") or suffix.endswith(".histo.gz") or suffix.endswith("oldhisto") or suffix.endswith("oldhisto.gz"):
for l in lines:
if l.startswith("data"):
tokens=string.split(l)
tokens.pop(0)
tokens=map(lambda t:
float(t),tokens)
bins=len(tokens)
maximum=max(tokens)
g=Gnuplot.Gnuplot()
g('set terminal png size 300,200')
g('unset xtics')
g('unset ytics')
g('set data style boxes')
print "Content-Type: image/png\n"
sys.stdout.flush()
g.plot(Gnuplot.Data(tokens))
elif suffix.endswith(".vec.gz") or suffix.endswith(".oldvec.gz") or suffix.endswith(".vec"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
if l.startswith("data"):
print l[5:]
print "</body></html>\n"
elif suffix.endswith(".textID"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body>\n</html>"
elif suffix.endswith("txt") or suffix.endswith("txt.gz"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body></html>\n"
elif suffix.endswith(".png"):
for l in lines:
if l.startswith("sizes"):
tok=string.split(l)
xsize=int(tok[1])
ysize=int(tok[2])
zsize=int(tok[3])
if l.startswith("data"):
tok=string.split(l)
tok.pop(0)
if zsize==1:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
g=int(float(tok.pop(0))*255)
i.putpixel((x,y),(g,g,g))
elif zsize==3:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
r=int(float(tok.pop(0))*255)
g=int(float(tok.pop(0))*255)
b=int(float(tok.pop(0))*255)
i.putpixel((x,y),(r,g,b))
print "Content-Type: image/png\n"
sys.stdout.flush()
i.save(sys.stdout,"png")
else:
print "Content-Type: text/html\n"
sys.stdout.flush()
print suffix
print "feature not supported: ",suffix
for l in lines:
print l+"<br>"
print "</body></html>\n"
except IOError:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Something failed"
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
import sys,cgi,re,Image,Gnuplot,string,os
sys.path+=[os.getcwd()]
from types import *
from firesocket import *
try:
form=cgi.FieldStorage()
if form.has_key("server"):
server=form["server"].value
else:
server="localhost"
if form.has_key("port"):
port=int(form["port"].value)
else:
port=12960
if form.has_key("image"):
image=form["image"].value
else:
image="100.jpg"
if form.has_key("feature"):
feature=form["feature"].value
else:
feature="1"
s=FIRESocket()
s.connect(server,port)
s.sendcmd("feature "+image+" "+feature)
line=s.getline()
suffix=string.split(line)[1]
lines=[]
line=s.getline()
while line!="end":
lines+=[line]
line=s.getline()
s.sendcmd("bye")
if suffix.endswith(".histo") or suffix.endswith(".histo.gz") or suffix.endswith("oldhisto") or suffix.endswith("oldhisto.gz"):
for l in lines:
if l.startswith("data"):
tokens=string.split(l)
tokens.pop(0)
tokens=map(lambda t:
float(t),tokens)
bins=len(tokens)
maximum=max(tokens)
g=Gnuplot.Gnuplot()
g('set terminal png size 300,200')
g('unset xtics')
g('unset ytics')
g('set data style boxes')
print "Content-Type: image/png\n"
sys.stdout.flush()
g.plot(Gnuplot.Data(tokens))
elif suffix.endswith(".vec.gz") or suffix.endswith(".oldvec.gz") or suffix.endswith(".vec"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
if l.startswith("data"):
print l[5:]
print "</body></html>\n"
elif suffix.endswith(".textID"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body>\n</html>"
elif suffix.endswith("txt") or suffix.endswith("txt.gz"):
print "Content-Type: text/html\n"
sys.stdout.flush()
print "<html><body>\n"
for l in lines:
print l
print "</body></html>\n"
elif suffix.endswith(".png"):
for l in lines:
if l.startswith("sizes"):
tok=string.split(l)
xsize=int(tok[1])
ysize=int(tok[2])
zsize=int(tok[3])
if l.startswith("data"):
tok=string.split(l)
tok.pop(0)
if zsize==1:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
g=int(float(tok.pop(0))*255)
i.putpixel((x,y),(g,g,g))
elif zsize==3:
i=Image.new("RGB",(xsize,ysize))
for x in range(xsize):
for y in range(ysize):
r=int(float(tok.pop(0))*255)
g=int(float(tok.pop(0))*255)
b=int(float(tok.pop(0))*255)
i.putpixel((x,y),(r,g,b))
print "Content-Type: image/png\n"
sys.stdout.flush()
i.save(sys.stdout,"png")
else:
print "Content-Type: text/html\n"
sys.stdout.flush()
print suffix
print "feature not supported: ",suffix
for l in lines:
print l+"<br>"
print "</body></html>\n"
except IOError:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Something failed"
| Python |
#!/usr/bin/python
import sys
import cgi
import re
import Image
sys.stderr=sys.stdout
try:
form=cgi.FieldStorage()
imagefilename=form["image"].value
outWidth=-1
outHeight=-1
maxDim=-1
if(form.has_key("max")):
maxDim=int(form["max"].value)
if(form.has_key("width")):
outWidth=int(form["width"].value)
if(form.has_key("height")):
outHeight=int(form["height"].value)
inImg=Image.open(imagefilename)
inWidth=inImg.size[0]
inHeight=inImg.size[1]
if maxDim!=-1:
if inWidth>inHeight:
outWidth=maxDim
else:
outHeight=maxDim
if (outWidth==-1 and outHeight==-1):
outImg=inImg
elif (outWidth==-1):
scaleFactor=float(outHeight)/float(inHeight)
outWidth=int(scaleFactor*inWidth)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
elif (outHeight==-1):
scaleFactor=float(outWidth)/float(inWidth)
outHeight=int(scaleFactor*inHeight)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
else:
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
contenttype="image/jpg"
print "Content-Type: "+contenttype+"\n"
outImg=outImg.convert("RGB")
outImg.save(sys.stdout,"jpeg")
except:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Access not permitted"
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
#$Header$
# print cgi-header
print "Content-Type: text/html\n\n"
print
#print "<pre style=\"font-family:Fixedsys,Courier,1px;\">" # to have error messages as readable as possible, this
# is switched of right before displaying the main part of the page
#import lots of stuff
import cgi, re, sys, traceback, os, Image
#import cgitb; cgitb.enable()
from urllib import quote,unquote
from types import *
# redirect stderr to have error messages on the webpage and not in the
# webserver's log
sys.stderr=sys.stdout
# to be able to find the config
sys.path+=[os.getcwd()]
#import "my" socket implementation
from firesocket import *
# ----------------------------------------------------------------------
# import the config
# it is assumed there is a file called config in the same file as the cgi.
# an example for a valid config:
## settings for i6
#positiveImg="http://www-i6.iatik.rwth-aachen.de/~deselaers/images/positive.png"
#negativeImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/negativ.png"
#neutralImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/neutral.png"
#fireLogo="/u/deselaers/work/src/fire/cgi/fire-logo.png"
#i6Logo= "/u/deselaers/work/src/fire/cgi/i6.png"
#TemplateFile = "/u/deselaers/work/src/fire/cgi/fire-template.html"
#fireServer="deuterium"
#firePort=12960
#tempdir="/tmp"
#maxSize=1000, 1000
#minSize=200, 200
# ----------------------------------------------------------------------
import config
# settings for the linear (standard) scoring algorithm. using this
# algorithm the weights can be changed from the web interface
class LinearScoring:
def __init__(self):
self.size=0
self.weights=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
if(tmp=="weight"):
idx=status.pop(0)
w=status.pop(0)
self.weights+=[w]
else:
print "Cannot parse LinearScoring settings"
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for LinearScoring</th></tr>"
for i in range(len(self.weights)):
result+="<tr><td> Weight "+str(i)+"</td>"
result+="<td colspan=2><input name=\"weight"+str(i)+"\" type=\"text\" value=\""+self.weights[i]+"\"></td></tr>"
return result
# settings for maxent scoring algorithm
class MaxEntScoring:
def __init__(self):
self.size=0
self.factor=0
self.offset=0
self.lambdas=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
tmp=status.pop(0)
self.factor=status.pop(0)
tmp=status.pop(0)
self.offset=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
idx=status.pop(0)
l=status.pop(0)
self.lambdas+=[l]
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for MaxEntScoring</th></tr>"
for i in range(len(self.lambdas)/2):
result+="<tr><td> Lambda "+str(i)+"</td>"
result+="<td>"+self.lambdas[i]+"</td><td>"+self.lambdas[i+len(self.lambdas)/2]+"</td></tr>"
return result
# a class to store the settings of fire this classes uses the classes
# defined above to manage the settings of the active scoring algorithm
class FireSettings:
def __init__(self):
# variable initilization only. all member variables are defined here.
self.fireServer=config.fireServer
self.firePort=config.firePort
self.dbsize=0
self.scoringname=""
self.results=0
self.expansions=0
self.distances=[]
self.suffices=[]
self.path=""
self.scoring=""
def get(self,s):
# get the settings from the server, i.e. read the string s and parse it.
self.distances=[]
self.suffices=[]
s.sendcmd("info")
msg=s.getline()
status=re.split(" ",msg)
print "<!--"+str(status)+"-->"
while(len(status)>0):
keyword=status.pop(0)
if keyword=="filelist":
self.dbsize=status.pop(0)
elif keyword=="results":
self.results=status.pop(0)
elif keyword=="path":
self.path=status.pop(0)
elif keyword=="extensions":
self.expansions=status.pop(0)
elif keyword=="scoring":
self.scoringname=status.pop(0)
if self.scoringname=="linear":
self.scoring=LinearScoring()
elif self.scoringname=="maxent":
self.scoring=MaxEntScoring()
self.scoring.get(status)
elif keyword=="suffix":
no=status.pop(0)
suffix=status.pop(0)
self.suffices+=[suffix]
distname=status.pop(0)
self.distances+=[distname]
else:
print "<!-- Unknown keyword in stream : "+keyword+" -->"
def display(self):
# display the settings dialog
self.get(s)
result="<h4>Settings for fire</h4>"
result=result+"<form method=\"post\" name=\"settingsForm\">"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+str(self.fireServer)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(self.firePort)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"settings\" value=\"1\"/>\n"
result=result+"<table border=\"1\">\n"
result=result+"<tr><th colspan=3> General Settings</th></tr>"
result=result+"<tr><td>Number of images in database</td><td colspan=\"2\">"+str(self.dbsize)+"</td></tr>\n"
result=result+"<tr><td>Number of results shown</td><td colspan=\"2\"><input type=\"text\" value=\""+self.results+"\" name=\"results\"/></td></tr>\n"
result=result+"<tr><td>Scoring</td><td colspan=\"2\">"+self.getScoringChooser()+"\n"
result=result+"<tr><td>Number of Images for Query Expansion</td><td colspan=\"2\"><input type=\"text\" value=\""+self.expansions+"\" name=\"extensions\"/></td></tr>\n"
result=result+self.getDistForm()
result=result+self.getScoringForm()
result=result+"</table>\n"
result=result+"Password: <input type=\"password\" name=\"password\">\n"
result=result+"<input type=\"submit\" name=\"Save\" value=\"Save\"/>\n"
result=result+"</form>"
return result
def getScoringChooser(self):
# select the chooser for the server
result="<select name=\"scoring\">\n<option value=\"\">"
for scor in ['linear','maxent']:
if self.scoringname==scor:
result=result+"<option selected value=\""+scor+"\">"+scor+"</option>\n"
else:
result=result+"<option value=\""+scor+"\">"+scor+"</option>\n"
result+="</select>\n"
return result
def getScoringForm(self):
# display the form element to select the scoring algorihtm
return self.scoring.getForm()
def getDistForm(self):
result=""
result+="<tr><th colspan=3> Settings for Distance Functions</th></tr>"
for no in range(len(self.distances)):
result+="<tr><td> Distance "+str(no)+": "+self.suffices[no]+"</td><td>"+self.getDistChooser(no,self.distances[no])+"</td></tr>"
return result
def getDistChooser(self,no,distName):
result="<select name=\"dist"+str(no)+"\">\n<option value=\"\">"
s.sendcmd("setdist")
distsString=s.getline()
dists=re.split(" ",distsString)
for d in dists:
if d==distName:
result=result+"<option selected value=\""+d+"\">"+d+"</option>\n"
else:
result=result+"<option value=\""+d+"\">"+d+"</option>\n"
result=result+"</select>\n"
return result
def process(self,form):
result="<!-- processing settings -->\n"
if form.has_key("results"):
if form.has_key("password"):
password=form["password"].value
else:
password=""
s.sendcmd("password "+password)
l=s.getline()
result+="<!-- "+l+"-->"
if l!="ok":
result+="Changing settings denied: Not authorized!"
for i in form.keys():
if i.find("dist")==0:
no=i[4:]
d=form[i].value
if d!=self.distances[int(no)]: # only change when changed
s.sendcmd("setdist "+no+" "+d)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("weight")==0:
no=i[6:]
w=float(form[i].value)
if self.scoringname=="linear":
if(float(self.scoring.weights[int(no)])!=float(w)):
s.sendcmd("setweight "+no+" "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
else:
print "weights not supported"
elif i.find("results")==0:
if int(form[i].value)!=int(self.results):
s.sendcmd("setresults "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("extensions")==0:
if int(self.expansions)!=int(form[i].value):
s.sendcmd("setextensions "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("scoring")==0:
sn=form[i].value
if(sn!=self.scoringname): # only change when changed
s.sendcmd("setscoring "+sn)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
self.scoringname=sn
else:
result=result+"<!-- "+i+"-->\n"
result=result+"<!-- settings processed -->\n"
return result
# ----------------------------------------------------------------------
# Load the template file and display the content in this template file
# the string FIRELOGOSTRINGHERE, I6LOGOSTRINGHERE, and "INSERT CONTENT
# HERE" are replaced by the configured/generated strings
# ----------------------------------------------------------------------
def Display(Content):
TemplateHandle = open(config.TemplateFile, "r") # open in read only mode
# read the entire file as a string
TemplateInput = TemplateHandle.read()
TemplateHandle.close() # close the file
# this defines an exception string in case our
# template file is messed up
BadTemplateException = "There was a problem with the HTML template."
TemplateInput = re.sub("FIRELOGOSTRINGHERE",config.fireLogo,TemplateInput)
TemplateInput = re.sub("I6LOGOSTRINGHERE",config.i6Logo,TemplateInput)
SubResult = re.subn("INSERT CONTENT HERE",Content,TemplateInput)
if SubResult[1] == 0:
raise BadTemplateException
print SubResult[0]
def sendretrieve(querystring):
filterstring=""
# if(form.has_key("filter1")):
# filterstring+=" -porn2-sorted/009-www.ficken.bz-images-ficken.png"
# if(form.has_key("filter2")):
# filterstring+=" -porn2-sorted/037-www.free-porn-free-sex.net-free-porn-free-sex-free-sex-sites.png"
s.sendcmd("retrieve "+querystring+filterstring)
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by clicking a random
# image, that is:
# 1. process the form
# - extract the filename of the image which is used as query
# 2. send the query to the server
# 3. wait for the servers answer
# 4. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def retrieve(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["queryImage"].value
sendretrieve("+"+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimage,0)
return result
# ----------------------------------------------------------------------
# resizes an image while keeping aspect ratio
# ----------------------------------------------------------------------
def resizeImage(im, newSize):
newX, newY = newSize
if(im.size < newSize):
xSize, ySize = im.size
ratio=float(xSize)/float(ySize)
if(ySize < xSize):
newSize=(int(newY*ratio),newY)
else:
newSize=(newX,int(newX*(1/ratio)))
out = Image.new("RGB", newSize)
out = im.resize(newSize, Image.ANTIALIAS)
else:
im.thumbnail(newSize, Image.ANTIALIAS)
out=im
return out
# ----------------------------------------------------------------------
# verifies the image file and resizes the image according to
# minSize and maxSize parameters in config.py
# ----------------------------------------------------------------------
def checkImage(imagefile):
maxSize=config.maxSize
minSize=config.minSize
try:
im=Image.open(imagefile)
except:
return 0
if(im.size<minSize):
im=resizeImage(im, minSize)
elif(im.size>maxSize):
im=resizeImage(im, maxSize)
im.save(imagefile)
return 1
# ----------------------------------------------------------------------
# handle a serverfile-request
# ----------------------------------------------------------------------
def serverFile(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["serverfile"].value
checkImage(queryimage)
imagename=queryimage[queryimage.rfind("/")+1:len(queryimage)]
s.sendcmd("newfile 2 "+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+imagename,0)
return result
# ----------------------------------------------------------------------
# handle a newfile-request
# ----------------------------------------------------------------------
def newFile(form):
temporaryFiles=config.tempdir # standard: "/tmp"
result="<h3>Retrieval Result</h3>\n"
if(save_uploaded_file(form,"absolute_path",temporaryFiles)==0):
result=result+"No Retrieval Result available!"
return result
fileitem = form["absolute_path"]
# f=os.popen("ls -l "+ temporaryFiles+"/"+fileitem.filename)
# print f.readlines()
if(checkImage(temporaryFiles+"/"+fileitem.filename)==0):
result=result+"Delivered file was not a valid image!"
return result
s.sendcmd("newfile 2 "+temporaryFiles+"/"+fileitem.filename+".png")
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+fileitem.filename,0)
return result
# This saves a file uploaded by an HTML form.
# The form_field is the name of the file input field from the form.
# For example, the following form_field would be "file_1":
# <input name="file_1" type="file">
# The upload_dir is the directory where the file will be written.
# If no file was uploaded or if the field does not exist then
# this does nothing.
# written by Noah Spurrier, modified by Fabian Schwahn
def save_uploaded_file(form, form_field, upload_dir):
if not form.has_key(form_field): return 0
fileitem = form[form_field]
completepath = upload_dir + fileitem.filename
if os.path.isdir(completepath): return 0
if not fileitem.file: return 0
fout = open (os.path.join(upload_dir, fileitem.filename), 'wb')
while 1:
chunk = fileitem.file.read(100000)
if not chunk: break
fout.write (chunk)
fout.close()
return 1
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by relevance feedback.
# That is:
# 1. process the form
# - extract the filenames of positive and negative examples
# 2. generate the query string
# 3. send the query to the server
# 4. wait for the servers answer
# 5. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def feedbackretrieve(form):
print "<!-- feedback retrieve \n"
print form
print "-->\n"
result="<h3>Retrieval Result</h3>\n"
queryimages=""
for field in form.keys():
if field.startswith("relevance"):
imageno=re.sub("relevance","",field)
imagename=form["resultImage"+imageno].value
relevance=form[field].value
if relevance!="0":
queryimages=queryimages+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
result=result+"\n<!-- "+queryimages+" -->\n"
sendretrieve(queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimages,0)
return result
# ----------------------------------------------------------------------
# given the retrieval result string of the server process this answer
# and create the part of the html page displaying these results. this
# includes the buttons for relevance feedback and the hidden fields
# for server settings (e.g.\ which server, which port). also the
# buttons for "more results" and "relevance feedback" are generated.
# the button for saving relevances is deaktivated as it is not used
# ----------------------------------------------------------------------
def displayResults(tokens,querytype,querystring,resultsStep):
if(re.search("[0-9]+(\.[0-9]+)?", tokens[0]) == None):
result="Could not read retrieval result.\n"
s.flush()
return result
print "<!-- "+str(tokens)+"-->"
i=0
result="<form method=\"post\" name=\"resultForm\"><table>\n<tr>\n"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+querystring+"\"/>\n"
result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+quote(querystring)+"\" />\n"
result=result+"<input type=\"hidden\" name=\"resultsstep\" value=\""+str(resultsStep)+"\"/>\n"
first=True
while(len(tokens) > 0):
resNo=str(i)
img=tokens.pop(0)
score=float(tokens.pop(0))
result=result+"<td>\n"
result=result+"<a href=\""+config.fireurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&imageinfo="+img+"\" target=\"_blank\">"
# support for the select-rectangle tool
#-------------
# result=result+"<a href=\"save_coord.cgi?to_do=UpdateImage&upload_type=server&port="+str(settings.firePort)+"&image_name="+settings.path+"/"+img+"\" target=\"_top\">"
#-------------
# shows the image
#-------------
# result=result+"<a href=\"img.py?image="+settings.path+"/"+img+"\" target=\"_blank\">"
#------------
result=result+"<img src=\""+config.imgpyurl+"?max=150&image="+settings.path+"/"+img+"\" title=\""+str(score)+"-"+img+"\" name=\""+str(score)+"-"+img+"\">\n<br>\n"
result=result+"</a>"
result=result+"<input type=\"hidden\" name=\"resultImage"+resNo+"\" value=\""+img+"\">\n"
if first:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\" checked><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\"><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
first=False
else:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\"><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\" checked><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
result=result+"</td>\n\n"
i=i+1
if i%5==0:
result=result+"</tr>\n\n\n<tr>"
result=result+"</tr></table>\n"
# result+=filterbuttons()
result=result+"<button title=\"expandresults\" name=\"expandresults\" type=\"submit\" onClick=\"document.resultForm.feedback.value='expandresults';document.resultForm.submit()\">more results</button>\n"
result=result+"<button title=\"requery\" name=\"requery\" type=\"submit\" onClick=\"document.resultForm.feedback.value='requery';document.resultForm.submit()\">requery</button>\n"
# result=result+"<button title=\"saverelevances\" name=\"relevancessave\" onClick=\"document.resultForm.feedback.value='saverelevances';document.resultForm.submit()\" \"type=\"submit\">save relevances</button>\n"
result=result+"<input type=\"hidden\" name=\"feedback\" value=\"yes\"/>\n"
result=result+"</form>\n"
return result
def filterbuttons():
# display the buttons for the filter
result=""
result=result+"filter 1: <input type=\"checkbox\" name=\"filter1\" value=\"1\">\n"
result=result+"filter 2: <input type=\"checkbox\" name=\"filter2\" value=\"2\">\n<br>\n"
return result
# ----------------------------------------------------------------------
# ask the server for filenames of random images and create the part of
# the html page showing random images. Steps:
# 1. send "random" to the server
# 2. receive the servers answer
# 3. process the answer:
# - get the filenames and for each filename create a querybutton
# 4. put the "more random images" button, such that we can get a new set
# of random images
# ----------------------------------------------------------------------
def randomImages():
# display random images
s.sendcmd("random")
result="<hr><h3> Random Images - Click one to start a query </h3>"
result+="""<form method="post" name="queryForm" enctype="multipart/form-data">
<input type="hidden" name="queryImage" value="">
"""
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result+=filterbuttons()
res=s.getline()
for img in re.split(" ",res):
if img != "" and img != "\n":
result=result+"<button title=\""+img+"\" name=\"searchButton\" type=\"submit\" onClick=\"document.queryForm.queryImage.value='"+img+"';document.queryForm.submit()\" value=\""+img+"\">\n"
result=result+"<img src=\""+config.imgpyurl+"?max=100&image="+settings.path+"/"+img+"\" title=\""+img+"\">\n"
result=result+"</button>\n"
if form.has_key("demomode"):
seconds=form["demomode"].value
url="http://"
if(os.environ.has_key("HTTP_HOST") and os.environ.has_key("REQUEST_URI")):
part1=os.environ["HTTP_HOST"]
part2=re.sub("&queryImage=.*$","",os.environ["REQUEST_URI"])
url=url+part1+part2+"&queryImage="+img
print "<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result=result+"<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result+="<br><div align=\"right\">\n<button title=\"new random images\" name=\"morerandom\" type=\"submit\">more random images</button>\n</div>\n"
return result
# ----------------------------------------------------------------------
# displays the newfile-form
# ----------------------------------------------------------------------
def newFileForm():
result="<hr><h3> Upload - Use query image from your file system </h3>"
result+="""<input type="hidden" name="newFile" value="0">"""
result+="""<input name="absolute_path" type="file" size="30" onClick="document.queryForm.newFile.value=1;document.queryForm.submit()">"""
result+="\n"
return result
# ----------------------------------------------------------------------
# generate the text field and the query button for text
# queries
# ----------------------------------------------------------------------
def textQueryForm(settings):
result = "<hr><h3>Query by "
if("textfeature" in settings.distances):
result += "description"
if("metafeature" in settings.distances):
result += " or meta information"
elif("metafeature" in settings.distances):
result += "meta information"
result += "</h3>\n"
result += "<form method=\"post\" name=\"textQueryForm\">\n"
result += "<input type=\"hidden\" name=\"server\" value=\""+str(settings.fireServer)+"\"/>\n"
result += "<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
result += "<input type=\"hidden\" name=\"resultsstep\" value=\"0\"/>\n"
result += "<input type=\"text\" name=\"textquerystring\" size=\"30\" value=\"\"/>\n"
result += "<input type=\"submit\" value=\"query\"/>\n"
if("metafeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&metafeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on meta info</a>"
if("textfeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&textfeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on description</a>"
result += "</form>"
return result
def displayMetaFeatureInfo():
s.sendcmd("metafeatureinfo")
mfi = s.getline()
mfil = mfi.split(" ")
result = "<h4>Help on queries by meta info</h4>\n"
result += """If you want to find images with certain meta information attached,
type in a request of the form
<h5>key1:val1,key2:val2,...</h5>\n
The following meta keys are available in this corpus:<br><br>\n"""
#TODO: The table looks ugly!
result += "<table>\n"
result += "<tr>\n"
result += " <th>key</th>\n"
result += " <th>example value</th>\n"
result += "</tr>\n"
for mf in mfil:
mfl = mf.split(":")
result += "<tr>\n"
result += " <td><b>"+mfl[0]+"</b></td>\n" # I know <b> is deprecated, but it still works :-)
result += " <td>"+mfl[1]+"</td>\n"
result += "</tr>\n"
result += "</table>\n"
return result
def displayTextFeatureInfo():
result = "<h4>Help on queries by description</h4>\n"
result += """If you want to find images which have text information attached to them,
just enter the query into the textbox. Fire will then use an information
retrieval engine to find the images that best match your query.<br><br>
If you have text information in multiple languages for each image, you can give
a query for every language. For example, if you have german and french text
information and you want to search for "hand" in the respective languages, you enter
<h5>Ge:"hand" Fr:"mains"</h5>
Here, "Ge" has to be the suffix for the german textfiles and "Fr" has to be the suffix for the
fench textfiles. Fire will then use both queries and send them to separate information retrieval
engines."""
return result
# ----------------------------------------------------------------------
# this function makes the server save a relevances logfile
# but it is not used at the moment, because the saverelevances button
# is deaktivated
# ----------------------------------------------------------------------
def saveRelevances(form):
querystring=form["querystring"].value
relevancestring=""
for field in form.keys():
if field.startswith("relevance-"):
imagename=re.sub("relevance-","",field)
relevance=form[field].value
if relevance!="0":
relevancestring=relevancestring+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
s.sendcmd("saverelevances Q "+querystring+" R "+relevancestring)
res=s.getline();
message="<a href=\"javascript:back()\"> Go back to last query </a>"
return message
# ----------------------------------------------------------------------
# this is the "getMoreResults" function it basically does the same as
# retrieve, uses the same query string as the last query used (this is
# gotten from the form) and the calls displayResults
# ----------------------------------------------------------------------
def expandQuery(form):
resultsstep=int(form["resultsstep"].value)
resultsstep=resultsstep+1
if(form.has_key("metaquerystring")):
result="<h3>Metatag-based retrieval result</h3>\n"
queryfield = "metaquerystring"
cmd = "metaexpand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
elif(form.has_key("textquerystring")):
result="<h3>Text-based retrieval result</h3>\n"
queryfield = "textquerystring"
cmd = "textexpand"
queryimages = unquote(form["textquerystring"].value)
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
else:
result="<h3>Retrieval Result</h3>\n"
queryfield = "querystring"
cmd = "expand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
if(tokens[0] == "notextinformation"):
result=result+"No images matched your query: "+querystring
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,queryfield,queryimages,resultsstep)
return result
# ----------------------------------------------------------------------
# process a query using text- or metainformation from the text input
# field. The query type is set to metaquery when the search string has
# a leading "META "
# This works identical as the retrieve function
# ----------------------------------------------------------------------
def processTextQuery(form):
# Check if it's meta or text
# It's meta either if there's only the metafeature or it has a leading "META "
hasmeta = "metafeature" in settings.distances
hastext = "textfeature" in settings.distances
querystring = form["textquerystring"].value
if (hasmeta and (not hastext or querystring[:5] == "META ")):
if(querystring[:5] == "META "):
querystring = querystring[5:]
querytype = "metaquerystring"
result="<h3>Metatag-based retrieval result</h3>\n"
s.sendcmd("metaretrieve +"+querystring)
else:
querytype = "textquerystring"
result="<h3>Text-based retrieval result</h3>\n"
s.sendcmd("textretrieve +"+querystring)
msg=s.getline()
tokens=re.split(" ",msg)
if( (tokens[0] == "notextinformation") or (tokens[0] == "nometainformation") ):
result=result+"No images matched your query: "+querystring
if(querytype == "metaquerystring"):
if(hastext):
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'META key1:val1,key2:val2,...'<br><br>\n"
else:
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'key1:val1,key2:val2,...'<br><br>\n"
result=result+"Also have a look at the help to see which keys are available.<br>\n"
if(querytype == "textquerystring"):
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,querytype,querystring,0)
return result
# ----------------------------------------------------------------------
# create the link for the "settings" at the bottom of the page
# ----------------------------------------------------------------------
def adminLink(fireServer, firePort):
result=""
result=result+"<a name=\"#\"></a>"
result=result+"<a href=\"#\" "+\
"onClick=\"window.open('"+config.fireurl+"?server="+ \
fireServer+"&port="+str(firePort)+"&settings=1','newwindow',"+\
"'height=600, width=600')\">settings</a>"
return result
#<SCRIPT LANGUAGE="javascript">
#<!--
#window.open ('titlepage.html', 'newwindow', config='height=100,
#width=400, toolbar=no, menubar=no, scrollbars=no, resizable=no,
#location=no, directories=no, status=no')
#-->
#</SCRIPT>
# ----------------------------------------------------------------------
# the main program
# ----------------------------------------------------------------------
# get form information in easily accessible manner
form=cgi.FieldStorage()
# make socket
s = FIRESocket()
settings=FireSettings()
# see what server we have and if the webinterface specifed another one
# than the one in the config file
if form.has_key("server"):
settings.fireServer=form["server"].value
if form.has_key("port"):
settings.firePort=int(form["port"].value)
try:
# connect to the server
s.connect(settings.fireServer, settings.firePort)
except:
# if the server is down, give an error message
# print "</pre>"
message="""
<h2>FIRE Server down</h2>
<p>
Try again later, complain to
<a href="mailto:deselaers@informatik.rwth-aachen.de">deselaers@informatik.rwth-aachen.de</a>
or try other servers.
"""
Display(message)
else:
# everything is fine, connection to server is ok
# try:
message=""
settings.get(s)
# do we want to show the settings window?
if(form.has_key("settings")):
# see if there have been new settings defined
message=settings.process(form)
# generate the settings dialog
message+=settings.display()
# disonnect from server
s.sendcmd("bye")
print "</pre>"
# display the settings dialog
Display(message)
# do we want to show the metafeatureinfo window?
elif(form.has_key("metafeatureinfo")):
message+=displayMetaFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
elif(form.has_key("imageinfo")):
imagename=form["imageinfo"].value
s.sendcmd("image "+imagename)
line=""
while(line != "end"):
line=s.getline()
tokens=re.split(" ",line)
if tokens[0]=="imagefilename":
message+="<img src=\""+config.imgpyurl+"?image="+tokens[1]+"\" alt=\""+tokens[1]+"\"/><br>"+"\n"
elif tokens[0]=="class":
message+="class: "+tokens[1]+"<br>\n"
elif tokens[0]=="description":
message+="description: "+tokens[1:]+"<br>\n"
elif tokens[0]=="features":
features=[]
noffeatures=int(tokens[1])
for i in range(noffeatures):
line=s.getline()
tokens=re.split(" ",line)
message+="<h3>"+str(i)+": "+tokens[2]+"</h3>\n<br>\n"
if tokens[2].endswith(".histo.gz") \
or tokens[2].endswith(".histo") \
or tokens[2].endswith(".oldhisto.gz") \
or tokens[2].endswith(".png"):
message+="<img src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\" alt=\""+tokens[2]+"\"/><br><br>\n"
elif tokens[2].endswith(".vec.gz") \
or tokens[2].endswith(".oldvec.gz") \
or tokens[2].endswith(".vec") \
or tokens[2].endswith("txt") \
or tokens[2].endswith("txt.gz") \
or tokens[2].endswith("textID"):
message+="<iframe src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\"></iframe><br><br>\n"
else:
message+="not yet supported<br>\n"
line=s.getline() # get end
s.sendcmd("bye")
Display(message)
elif(form.has_key("textfeatureinfo")):
message+=displayTextFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
else:
# we do not want to show the settings window
if(form.has_key("newFile") and form["newFile"].value=="1"):
#message+=str(form["newFile"])
message+=newFile(form)
elif(form.has_key("serverfile")):
message+=serverFile(form)
elif(form.has_key("feedback")):
# result images have been selected and are part of this query
if form["feedback"].value=="requery":
# relevance feedback
message+=feedbackretrieve(form)
elif form["feedback"].value=="expandresults":
# more results
message+=expandQuery(form)
else:
# save relevances (this will not happen currently)
message=saveRelevances(form)
elif(form.has_key("queryImage") and form["queryImage"].value!=""):
# no relevance feedback, but just one of the random images was clicked
message+=retrieve(form)
elif form.has_key('textquerystring'):
# no relevance feedback, no query image, but a query string for text queries
message+=processTextQuery(form)
# now generate the remainder of the website: queryfield for meta information, random images, adminlink
if("textfeature" in settings.distances or "metafeature" in settings.distances):
message+=textQueryForm(settings)
message+=randomImages()
message+=newFileForm()
message+="<p align=\"right\">"+adminLink(settings.fireServer, settings.firePort)+"</p>"
message+="</form>"
# disconnect from server
s.sendcmd("bye")
print "</pre>"
# display the generated webpage
Display(message)
# except Exception, e:
# if anything went wrong:
# show the error message as readable as possible,
# disconnect from the server to keep it alive
# and end yourself
# print "<pre>"
# print "Exception: ",e
# s.sendcmd("bye")
# print "</pre>"
| Python |
#default settings
positiveImg="images/positive.png"
negativeImg="images/negativ.png"
neutralImg="images/neutral.png"
fireLogo="images/fire-logo.png"
i6Logo="images/i6.png"
TemplateFile = "fire-template.html"
fireServer="localhost"
firePort=12960
imgpyurl="img.py"
fireurl="fire.py"
#tempdir="/tmp"
#maxSize=1000, 1000
#minSize=200, 200
| Python |
#!/usr/bin/python
import sys
import cgi
import re
import Image
sys.stderr=sys.stdout
try:
form=cgi.FieldStorage()
imagefilename=form["image"].value
outWidth=-1
outHeight=-1
maxDim=-1
if(form.has_key("max")):
maxDim=int(form["max"].value)
if(form.has_key("width")):
outWidth=int(form["width"].value)
if(form.has_key("height")):
outHeight=int(form["height"].value)
inImg=Image.open(imagefilename)
inWidth=inImg.size[0]
inHeight=inImg.size[1]
if maxDim!=-1:
if inWidth>inHeight:
outWidth=maxDim
else:
outHeight=maxDim
if (outWidth==-1 and outHeight==-1):
outImg=inImg
elif (outWidth==-1):
scaleFactor=float(outHeight)/float(inHeight)
outWidth=int(scaleFactor*inWidth)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
elif (outHeight==-1):
scaleFactor=float(outWidth)/float(inWidth)
outHeight=int(scaleFactor*inHeight)
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
else:
outImg=inImg.resize((outWidth,outHeight),Image.BILINEAR)
contenttype="image/jpg"
print "Content-Type: "+contenttype+"\n"
outImg=outImg.convert("RGB")
outImg.save(sys.stdout,"jpeg")
except:
contenttype="text/html"
print "Content-Type: "+contenttype+"\n"
print "Access not permitted"
| Python |
#!/usr/bin/python
# -*- coding: iso-8859-1 -*-
#$Header$
# print cgi-header
print "Content-Type: text/html\n\n"
print
#print "<pre style=\"font-family:Fixedsys,Courier,1px;\">" # to have error messages as readable as possible, this
# is switched of right before displaying the main part of the page
#import lots of stuff
import cgi, re, sys, traceback, os, Image
#import cgitb; cgitb.enable()
from urllib import quote,unquote
from types import *
# redirect stderr to have error messages on the webpage and not in the
# webserver's log
sys.stderr=sys.stdout
# to be able to find the config
sys.path+=[os.getcwd()]
#import "my" socket implementation
from firesocket import *
# ----------------------------------------------------------------------
# import the config
# it is assumed there is a file called config in the same file as the cgi.
# an example for a valid config:
## settings for i6
#positiveImg="http://www-i6.iatik.rwth-aachen.de/~deselaers/images/positive.png"
#negativeImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/negativ.png"
#neutralImg="http://www-i6.informatik.rwth-aachen.de/~deselaers/images/neutral.png"
#fireLogo="/u/deselaers/work/src/fire/cgi/fire-logo.png"
#i6Logo= "/u/deselaers/work/src/fire/cgi/i6.png"
#TemplateFile = "/u/deselaers/work/src/fire/cgi/fire-template.html"
#fireServer="deuterium"
#firePort=12960
#tempdir="/tmp"
#maxSize=1000, 1000
#minSize=200, 200
# ----------------------------------------------------------------------
import config
# settings for the linear (standard) scoring algorithm. using this
# algorithm the weights can be changed from the web interface
class LinearScoring:
def __init__(self):
self.size=0
self.weights=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
if(tmp=="weight"):
idx=status.pop(0)
w=status.pop(0)
self.weights+=[w]
else:
print "Cannot parse LinearScoring settings"
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for LinearScoring</th></tr>"
for i in range(len(self.weights)):
result+="<tr><td> Weight "+str(i)+"</td>"
result+="<td colspan=2><input name=\"weight"+str(i)+"\" type=\"text\" value=\""+self.weights[i]+"\"></td></tr>"
return result
# settings for maxent scoring algorithm
class MaxEntScoring:
def __init__(self):
self.size=0
self.factor=0
self.offset=0
self.lambdas=[]
def get(self,status):
tmp=status.pop(0)
self.size=status.pop(0)
tmp=status.pop(0)
self.factor=status.pop(0)
tmp=status.pop(0)
self.offset=status.pop(0)
for i in range(int(self.size)):
tmp=status.pop(0)
idx=status.pop(0)
l=status.pop(0)
self.lambdas+=[l]
def getForm(self):
result=""
result+="<tr><th colspan=3> Settings for MaxEntScoring</th></tr>"
for i in range(len(self.lambdas)/2):
result+="<tr><td> Lambda "+str(i)+"</td>"
result+="<td>"+self.lambdas[i]+"</td><td>"+self.lambdas[i+len(self.lambdas)/2]+"</td></tr>"
return result
# a class to store the settings of fire this classes uses the classes
# defined above to manage the settings of the active scoring algorithm
class FireSettings:
def __init__(self):
# variable initilization only. all member variables are defined here.
self.fireServer=config.fireServer
self.firePort=config.firePort
self.dbsize=0
self.scoringname=""
self.results=0
self.expansions=0
self.distances=[]
self.suffices=[]
self.path=""
self.scoring=""
def get(self,s):
# get the settings from the server, i.e. read the string s and parse it.
self.distances=[]
self.suffices=[]
s.sendcmd("info")
msg=s.getline()
status=re.split(" ",msg)
print "<!--"+str(status)+"-->"
while(len(status)>0):
keyword=status.pop(0)
if keyword=="filelist":
self.dbsize=status.pop(0)
elif keyword=="results":
self.results=status.pop(0)
elif keyword=="path":
self.path=status.pop(0)
elif keyword=="extensions":
self.expansions=status.pop(0)
elif keyword=="scoring":
self.scoringname=status.pop(0)
if self.scoringname=="linear":
self.scoring=LinearScoring()
elif self.scoringname=="maxent":
self.scoring=MaxEntScoring()
self.scoring.get(status)
elif keyword=="suffix":
no=status.pop(0)
suffix=status.pop(0)
self.suffices+=[suffix]
distname=status.pop(0)
self.distances+=[distname]
else:
print "<!-- Unknown keyword in stream : "+keyword+" -->"
def display(self):
# display the settings dialog
self.get(s)
result="<h4>Settings for fire</h4>"
result=result+"<form method=\"post\" name=\"settingsForm\">"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+str(self.fireServer)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(self.firePort)+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"settings\" value=\"1\"/>\n"
result=result+"<table border=\"1\">\n"
result=result+"<tr><th colspan=3> General Settings</th></tr>"
result=result+"<tr><td>Number of images in database</td><td colspan=\"2\">"+str(self.dbsize)+"</td></tr>\n"
result=result+"<tr><td>Number of results shown</td><td colspan=\"2\"><input type=\"text\" value=\""+self.results+"\" name=\"results\"/></td></tr>\n"
result=result+"<tr><td>Scoring</td><td colspan=\"2\">"+self.getScoringChooser()+"\n"
result=result+"<tr><td>Number of Images for Query Expansion</td><td colspan=\"2\"><input type=\"text\" value=\""+self.expansions+"\" name=\"extensions\"/></td></tr>\n"
result=result+self.getDistForm()
result=result+self.getScoringForm()
result=result+"</table>\n"
result=result+"Password: <input type=\"password\" name=\"password\">\n"
result=result+"<input type=\"submit\" name=\"Save\" value=\"Save\"/>\n"
result=result+"</form>"
return result
def getScoringChooser(self):
# select the chooser for the server
result="<select name=\"scoring\">\n<option value=\"\">"
for scor in ['linear','maxent']:
if self.scoringname==scor:
result=result+"<option selected value=\""+scor+"\">"+scor+"</option>\n"
else:
result=result+"<option value=\""+scor+"\">"+scor+"</option>\n"
result+="</select>\n"
return result
def getScoringForm(self):
# display the form element to select the scoring algorihtm
return self.scoring.getForm()
def getDistForm(self):
result=""
result+="<tr><th colspan=3> Settings for Distance Functions</th></tr>"
for no in range(len(self.distances)):
result+="<tr><td> Distance "+str(no)+": "+self.suffices[no]+"</td><td>"+self.getDistChooser(no,self.distances[no])+"</td></tr>"
return result
def getDistChooser(self,no,distName):
result="<select name=\"dist"+str(no)+"\">\n<option value=\"\">"
s.sendcmd("setdist")
distsString=s.getline()
dists=re.split(" ",distsString)
for d in dists:
if d==distName:
result=result+"<option selected value=\""+d+"\">"+d+"</option>\n"
else:
result=result+"<option value=\""+d+"\">"+d+"</option>\n"
result=result+"</select>\n"
return result
def process(self,form):
result="<!-- processing settings -->\n"
if form.has_key("results"):
if form.has_key("password"):
password=form["password"].value
else:
password=""
s.sendcmd("password "+password)
l=s.getline()
result+="<!-- "+l+"-->"
if l!="ok":
result+="Changing settings denied: Not authorized!"
for i in form.keys():
if i.find("dist")==0:
no=i[4:]
d=form[i].value
if d!=self.distances[int(no)]: # only change when changed
s.sendcmd("setdist "+no+" "+d)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("weight")==0:
no=i[6:]
w=float(form[i].value)
if self.scoringname=="linear":
if(float(self.scoring.weights[int(no)])!=float(w)):
s.sendcmd("setweight "+no+" "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
else:
print "weights not supported"
elif i.find("results")==0:
if int(form[i].value)!=int(self.results):
s.sendcmd("setresults "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("extensions")==0:
if int(self.expansions)!=int(form[i].value):
s.sendcmd("setextensions "+form[i].value)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
elif i.find("scoring")==0:
sn=form[i].value
if(sn!=self.scoringname): # only change when changed
s.sendcmd("setscoring "+sn)
l=s.getline()
result=result+"<!-- "+l+" -->\n"
self.scoringname=sn
else:
result=result+"<!-- "+i+"-->\n"
result=result+"<!-- settings processed -->\n"
return result
# ----------------------------------------------------------------------
# Load the template file and display the content in this template file
# the string FIRELOGOSTRINGHERE, I6LOGOSTRINGHERE, and "INSERT CONTENT
# HERE" are replaced by the configured/generated strings
# ----------------------------------------------------------------------
def Display(Content):
TemplateHandle = open(config.TemplateFile, "r") # open in read only mode
# read the entire file as a string
TemplateInput = TemplateHandle.read()
TemplateHandle.close() # close the file
# this defines an exception string in case our
# template file is messed up
BadTemplateException = "There was a problem with the HTML template."
TemplateInput = re.sub("FIRELOGOSTRINGHERE",config.fireLogo,TemplateInput)
TemplateInput = re.sub("I6LOGOSTRINGHERE",config.i6Logo,TemplateInput)
SubResult = re.subn("INSERT CONTENT HERE",Content,TemplateInput)
if SubResult[1] == 0:
raise BadTemplateException
print SubResult[0]
def sendretrieve(querystring):
filterstring=""
# if(form.has_key("filter1")):
# filterstring+=" -porn2-sorted/009-www.ficken.bz-images-ficken.png"
# if(form.has_key("filter2")):
# filterstring+=" -porn2-sorted/037-www.free-porn-free-sex.net-free-porn-free-sex-free-sex-sites.png"
s.sendcmd("retrieve "+querystring+filterstring)
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by clicking a random
# image, that is:
# 1. process the form
# - extract the filename of the image which is used as query
# 2. send the query to the server
# 3. wait for the servers answer
# 4. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def retrieve(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["queryImage"].value
sendretrieve("+"+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimage,0)
return result
# ----------------------------------------------------------------------
# resizes an image while keeping aspect ratio
# ----------------------------------------------------------------------
def resizeImage(im, newSize):
newX, newY = newSize
if(im.size < newSize):
xSize, ySize = im.size
ratio=float(xSize)/float(ySize)
if(ySize < xSize):
newSize=(int(newY*ratio),newY)
else:
newSize=(newX,int(newX*(1/ratio)))
out = Image.new("RGB", newSize)
out = im.resize(newSize, Image.ANTIALIAS)
else:
im.thumbnail(newSize, Image.ANTIALIAS)
out=im
return out
# ----------------------------------------------------------------------
# verifies the image file and resizes the image according to
# minSize and maxSize parameters in config.py
# ----------------------------------------------------------------------
def checkImage(imagefile):
maxSize=config.maxSize
minSize=config.minSize
try:
im=Image.open(imagefile)
except:
return 0
if(im.size<minSize):
im=resizeImage(im, minSize)
elif(im.size>maxSize):
im=resizeImage(im, maxSize)
im.save(imagefile)
return 1
# ----------------------------------------------------------------------
# handle a serverfile-request
# ----------------------------------------------------------------------
def serverFile(form):
result="<h3>Retrieval Result</h3>\n"
queryimage=form["serverfile"].value
checkImage(queryimage)
imagename=queryimage[queryimage.rfind("/")+1:len(queryimage)]
s.sendcmd("newfile 2 "+queryimage)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+imagename,0)
return result
# ----------------------------------------------------------------------
# handle a newfile-request
# ----------------------------------------------------------------------
def newFile(form):
temporaryFiles=config.tempdir # standard: "/tmp"
result="<h3>Retrieval Result</h3>\n"
if(save_uploaded_file(form,"absolute_path",temporaryFiles)==0):
result=result+"No Retrieval Result available!"
return result
fileitem = form["absolute_path"]
# f=os.popen("ls -l "+ temporaryFiles+"/"+fileitem.filename)
# print f.readlines()
if(checkImage(temporaryFiles+"/"+fileitem.filename)==0):
result=result+"Delivered file was not a valid image!"
return result
s.sendcmd("newfile 2 "+temporaryFiles+"/"+fileitem.filename+".png")
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring","./"+fileitem.filename,0)
return result
# This saves a file uploaded by an HTML form.
# The form_field is the name of the file input field from the form.
# For example, the following form_field would be "file_1":
# <input name="file_1" type="file">
# The upload_dir is the directory where the file will be written.
# If no file was uploaded or if the field does not exist then
# this does nothing.
# written by Noah Spurrier, modified by Fabian Schwahn
def save_uploaded_file(form, form_field, upload_dir):
if not form.has_key(form_field): return 0
fileitem = form[form_field]
completepath = upload_dir + fileitem.filename
if os.path.isdir(completepath): return 0
if not fileitem.file: return 0
fout = open (os.path.join(upload_dir, fileitem.filename), 'wb')
while 1:
chunk = fileitem.file.read(100000)
if not chunk: break
fout.write (chunk)
fout.close()
return 1
# ----------------------------------------------------------------------
# handle a retrieval request which was generated by relevance feedback.
# That is:
# 1. process the form
# - extract the filenames of positive and negative examples
# 2. generate the query string
# 3. send the query to the server
# 4. wait for the servers answer
# 5. call displayResults with the servers answer to display the results
# ----------------------------------------------------------------------
def feedbackretrieve(form):
print "<!-- feedback retrieve \n"
print form
print "-->\n"
result="<h3>Retrieval Result</h3>\n"
queryimages=""
for field in form.keys():
if field.startswith("relevance"):
imageno=re.sub("relevance","",field)
imagename=form["resultImage"+imageno].value
relevance=form[field].value
if relevance!="0":
queryimages=queryimages+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
result=result+"\n<!-- "+queryimages+" -->\n"
sendretrieve(queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
result=result+displayResults(tokens,"querystring",queryimages,0)
return result
# ----------------------------------------------------------------------
# given the retrieval result string of the server process this answer
# and create the part of the html page displaying these results. this
# includes the buttons for relevance feedback and the hidden fields
# for server settings (e.g.\ which server, which port). also the
# buttons for "more results" and "relevance feedback" are generated.
# the button for saving relevances is deaktivated as it is not used
# ----------------------------------------------------------------------
def displayResults(tokens,querytype,querystring,resultsStep):
if(re.search("[0-9]+(\.[0-9]+)?", tokens[0]) == None):
result="Could not read retrieval result.\n"
s.flush()
return result
print "<!-- "+str(tokens)+"-->"
i=0
result="<form method=\"post\" name=\"resultForm\"><table>\n<tr>\n"
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+querystring+"\"/>\n"
result=result+"<input type=\"hidden\" name=\""+querytype+"\" value=\""+quote(querystring)+"\" />\n"
result=result+"<input type=\"hidden\" name=\"resultsstep\" value=\""+str(resultsStep)+"\"/>\n"
first=True
while(len(tokens) > 0):
resNo=str(i)
img=tokens.pop(0)
score=float(tokens.pop(0))
result=result+"<td>\n"
result=result+"<a href=\""+config.fireurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&imageinfo="+img+"\" target=\"_blank\">"
# support for the select-rectangle tool
#-------------
# result=result+"<a href=\"save_coord.cgi?to_do=UpdateImage&upload_type=server&port="+str(settings.firePort)+"&image_name="+settings.path+"/"+img+"\" target=\"_top\">"
#-------------
# shows the image
#-------------
# result=result+"<a href=\"img.py?image="+settings.path+"/"+img+"\" target=\"_blank\">"
#------------
result=result+"<img src=\""+config.imgpyurl+"?max=150&image="+settings.path+"/"+img+"\" title=\""+str(score)+"-"+img+"\" name=\""+str(score)+"-"+img+"\">\n<br>\n"
result=result+"</a>"
result=result+"<input type=\"hidden\" name=\"resultImage"+resNo+"\" value=\""+img+"\">\n"
if first:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\" checked><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\"><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
first=False
else:
result=result+"<center><input type=\"radio\" name=\"relevance"+resNo+"\" value=\"+\"><img src=\""+config.positiveImg+"\" alt=\"+\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[0].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"0\" checked><img src=\""+config.neutralImg+"\" alt=\"0\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[1].checked=true;\"> \n"
result=result+"<input type=\"radio\" name=\"relevance"+resNo+"\" value=\"-\"><img src=\""+config.negativeImg+"\" alt=\"-\" onClick=\"javascript:document.resultForm.relevance"+resNo+"[2].checked=true;\"></center>\n"
result=result+"</td>\n\n"
i=i+1
if i%5==0:
result=result+"</tr>\n\n\n<tr>"
result=result+"</tr></table>\n"
# result+=filterbuttons()
result=result+"<button title=\"expandresults\" name=\"expandresults\" type=\"submit\" onClick=\"document.resultForm.feedback.value='expandresults';document.resultForm.submit()\">more results</button>\n"
result=result+"<button title=\"requery\" name=\"requery\" type=\"submit\" onClick=\"document.resultForm.feedback.value='requery';document.resultForm.submit()\">requery</button>\n"
# result=result+"<button title=\"saverelevances\" name=\"relevancessave\" onClick=\"document.resultForm.feedback.value='saverelevances';document.resultForm.submit()\" \"type=\"submit\">save relevances</button>\n"
result=result+"<input type=\"hidden\" name=\"feedback\" value=\"yes\"/>\n"
result=result+"</form>\n"
return result
def filterbuttons():
# display the buttons for the filter
result=""
result=result+"filter 1: <input type=\"checkbox\" name=\"filter1\" value=\"1\">\n"
result=result+"filter 2: <input type=\"checkbox\" name=\"filter2\" value=\"2\">\n<br>\n"
return result
# ----------------------------------------------------------------------
# ask the server for filenames of random images and create the part of
# the html page showing random images. Steps:
# 1. send "random" to the server
# 2. receive the servers answer
# 3. process the answer:
# - get the filenames and for each filename create a querybutton
# 4. put the "more random images" button, such that we can get a new set
# of random images
# ----------------------------------------------------------------------
def randomImages():
# display random images
s.sendcmd("random")
result="<hr><h3> Random Images - Click one to start a query </h3>"
result+="""<form method="post" name="queryForm" enctype="multipart/form-data">
<input type="hidden" name="queryImage" value="">
"""
result=result+"<input type=\"hidden\" name=\"server\" value=\""+settings.fireServer+"\"/>\n"
result=result+"<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
#result+=filterbuttons()
res=s.getline()
for img in re.split(" ",res):
if img != "" and img != "\n":
result=result+"<button title=\""+img+"\" name=\"searchButton\" type=\"submit\" onClick=\"document.queryForm.queryImage.value='"+img+"';document.queryForm.submit()\" value=\""+img+"\">\n"
result=result+"<img src=\""+config.imgpyurl+"?max=100&image="+settings.path+"/"+img+"\" title=\""+img+"\">\n"
result=result+"</button>\n"
if form.has_key("demomode"):
seconds=form["demomode"].value
url="http://"
if(os.environ.has_key("HTTP_HOST") and os.environ.has_key("REQUEST_URI")):
part1=os.environ["HTTP_HOST"]
part2=re.sub("&queryImage=.*$","",os.environ["REQUEST_URI"])
url=url+part1+part2+"&queryImage="+img
print "<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result=result+"<meta http-equiv=\"refresh\" content=\""+seconds+"; URL="+url+"\">\n"
result+="<br><div align=\"right\">\n<button title=\"new random images\" name=\"morerandom\" type=\"submit\">more random images</button>\n</div>\n"
return result
# ----------------------------------------------------------------------
# displays the newfile-form
# ----------------------------------------------------------------------
def newFileForm():
result="<hr><h3> Upload - Use query image from your file system </h3>"
result+="""<input type="hidden" name="newFile" value="0">"""
result+="""<input name="absolute_path" type="file" size="30" onClick="document.queryForm.newFile.value=1;document.queryForm.submit()">"""
result+="\n"
return result
# ----------------------------------------------------------------------
# generate the text field and the query button for text
# queries
# ----------------------------------------------------------------------
def textQueryForm(settings):
result = "<hr><h3>Query by "
if("textfeature" in settings.distances):
result += "description"
if("metafeature" in settings.distances):
result += " or meta information"
elif("metafeature" in settings.distances):
result += "meta information"
result += "</h3>\n"
result += "<form method=\"post\" name=\"textQueryForm\">\n"
result += "<input type=\"hidden\" name=\"server\" value=\""+str(settings.fireServer)+"\"/>\n"
result += "<input type=\"hidden\" name=\"port\" value=\""+str(settings.firePort)+"\"/>\n"
result += "<input type=\"hidden\" name=\"resultsstep\" value=\"0\"/>\n"
result += "<input type=\"text\" name=\"textquerystring\" size=\"30\" value=\"\"/>\n"
result += "<input type=\"submit\" value=\"query\"/>\n"
if("metafeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&metafeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on meta info</a>"
if("textfeature" in settings.distances):
result += "<a href=\"#\" onClick=\"window.open('"+config.fireurl+"?server="+ \
settings.fireServer+"&port="+str(settings.firePort)+ \
"&textfeatureinfo=1','newwindow',"+ \
"'height=600, width=600')\">help on description</a>"
result += "</form>"
return result
def displayMetaFeatureInfo():
s.sendcmd("metafeatureinfo")
mfi = s.getline()
mfil = mfi.split(" ")
result = "<h4>Help on queries by meta info</h4>\n"
result += """If you want to find images with certain meta information attached,
type in a request of the form
<h5>key1:val1,key2:val2,...</h5>\n
The following meta keys are available in this corpus:<br><br>\n"""
#TODO: The table looks ugly!
result += "<table>\n"
result += "<tr>\n"
result += " <th>key</th>\n"
result += " <th>example value</th>\n"
result += "</tr>\n"
for mf in mfil:
mfl = mf.split(":")
result += "<tr>\n"
result += " <td><b>"+mfl[0]+"</b></td>\n" # I know <b> is deprecated, but it still works :-)
result += " <td>"+mfl[1]+"</td>\n"
result += "</tr>\n"
result += "</table>\n"
return result
def displayTextFeatureInfo():
result = "<h4>Help on queries by description</h4>\n"
result += """If you want to find images which have text information attached to them,
just enter the query into the textbox. Fire will then use an information
retrieval engine to find the images that best match your query.<br><br>
If you have text information in multiple languages for each image, you can give
a query for every language. For example, if you have german and french text
information and you want to search for "hand" in the respective languages, you enter
<h5>Ge:"hand" Fr:"mains"</h5>
Here, "Ge" has to be the suffix for the german textfiles and "Fr" has to be the suffix for the
fench textfiles. Fire will then use both queries and send them to separate information retrieval
engines."""
return result
# ----------------------------------------------------------------------
# this function makes the server save a relevances logfile
# but it is not used at the moment, because the saverelevances button
# is deaktivated
# ----------------------------------------------------------------------
def saveRelevances(form):
querystring=form["querystring"].value
relevancestring=""
for field in form.keys():
if field.startswith("relevance-"):
imagename=re.sub("relevance-","",field)
relevance=form[field].value
if relevance!="0":
relevancestring=relevancestring+" "+relevance+imagename
else:
print "<!-- field not processed: "+field+" -->"
s.sendcmd("saverelevances Q "+querystring+" R "+relevancestring)
res=s.getline();
message="<a href=\"javascript:back()\"> Go back to last query </a>"
return message
# ----------------------------------------------------------------------
# this is the "getMoreResults" function it basically does the same as
# retrieve, uses the same query string as the last query used (this is
# gotten from the form) and the calls displayResults
# ----------------------------------------------------------------------
def expandQuery(form):
resultsstep=int(form["resultsstep"].value)
resultsstep=resultsstep+1
if(form.has_key("metaquerystring")):
result="<h3>Metatag-based retrieval result</h3>\n"
queryfield = "metaquerystring"
cmd = "metaexpand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
elif(form.has_key("textquerystring")):
result="<h3>Text-based retrieval result</h3>\n"
queryfield = "textquerystring"
cmd = "textexpand"
queryimages = unquote(form["textquerystring"].value)
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
else:
result="<h3>Retrieval Result</h3>\n"
queryfield = "querystring"
cmd = "expand"
queryimages=form[queryfield].value
s.sendcmd(cmd+" "+str(resultsstep)+" "+queryimages)
msg=s.getline()
tokens=re.split(" ",msg)
if(tokens[0] == "notextinformation"):
result=result+"No images matched your query: "+querystring
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,queryfield,queryimages,resultsstep)
return result
# ----------------------------------------------------------------------
# process a query using text- or metainformation from the text input
# field. The query type is set to metaquery when the search string has
# a leading "META "
# This works identical as the retrieve function
# ----------------------------------------------------------------------
def processTextQuery(form):
# Check if it's meta or text
# It's meta either if there's only the metafeature or it has a leading "META "
hasmeta = "metafeature" in settings.distances
hastext = "textfeature" in settings.distances
querystring = form["textquerystring"].value
if (hasmeta and (not hastext or querystring[:5] == "META ")):
if(querystring[:5] == "META "):
querystring = querystring[5:]
querytype = "metaquerystring"
result="<h3>Metatag-based retrieval result</h3>\n"
s.sendcmd("metaretrieve +"+querystring)
else:
querytype = "textquerystring"
result="<h3>Text-based retrieval result</h3>\n"
s.sendcmd("textretrieve +"+querystring)
msg=s.getline()
tokens=re.split(" ",msg)
if( (tokens[0] == "notextinformation") or (tokens[0] == "nometainformation") ):
result=result+"No images matched your query: "+querystring
if(querytype == "metaquerystring"):
if(hastext):
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'META key1:val1,key2:val2,...'<br><br>\n"
else:
result=result+"<br><br>Check if the syntax is correct.<br><br>Example: 'key1:val1,key2:val2,...'<br><br>\n"
result=result+"Also have a look at the help to see which keys are available.<br>\n"
if(querytype == "textquerystring"):
result=result+"<br>This could be an error in your syntax. Check the help for more information.<br><br>\n"
else:
result=result+displayResults(tokens,querytype,querystring,0)
return result
# ----------------------------------------------------------------------
# create the link for the "settings" at the bottom of the page
# ----------------------------------------------------------------------
def adminLink(fireServer, firePort):
result=""
result=result+"<a name=\"#\"></a>"
result=result+"<a href=\"#\" "+\
"onClick=\"window.open('"+config.fireurl+"?server="+ \
fireServer+"&port="+str(firePort)+"&settings=1','newwindow',"+\
"'height=600, width=600')\">settings</a>"
return result
#<SCRIPT LANGUAGE="javascript">
#<!--
#window.open ('titlepage.html', 'newwindow', config='height=100,
#width=400, toolbar=no, menubar=no, scrollbars=no, resizable=no,
#location=no, directories=no, status=no')
#-->
#</SCRIPT>
# ----------------------------------------------------------------------
# the main program
# ----------------------------------------------------------------------
# get form information in easily accessible manner
form=cgi.FieldStorage()
# make socket
s = FIRESocket()
settings=FireSettings()
# see what server we have and if the webinterface specifed another one
# than the one in the config file
if form.has_key("server"):
settings.fireServer=form["server"].value
if form.has_key("port"):
settings.firePort=int(form["port"].value)
try:
# connect to the server
s.connect(settings.fireServer, settings.firePort)
except:
# if the server is down, give an error message
# print "</pre>"
message="""
<h2>FIRE Server down</h2>
<p>
Try again later, complain to
<a href="mailto:deselaers@informatik.rwth-aachen.de">deselaers@informatik.rwth-aachen.de</a>
or try other servers.
"""
Display(message)
else:
# everything is fine, connection to server is ok
# try:
message=""
settings.get(s)
# do we want to show the settings window?
if(form.has_key("settings")):
# see if there have been new settings defined
message=settings.process(form)
# generate the settings dialog
message+=settings.display()
# disonnect from server
s.sendcmd("bye")
print "</pre>"
# display the settings dialog
Display(message)
# do we want to show the metafeatureinfo window?
elif(form.has_key("metafeatureinfo")):
message+=displayMetaFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
elif(form.has_key("imageinfo")):
imagename=form["imageinfo"].value
s.sendcmd("image "+imagename)
line=""
while(line != "end"):
line=s.getline()
tokens=re.split(" ",line)
if tokens[0]=="imagefilename":
message+="<img src=\""+config.imgpyurl+"?image="+tokens[1]+"\" alt=\""+tokens[1]+"\"/><br>"+"\n"
elif tokens[0]=="class":
message+="class: "+tokens[1]+"<br>\n"
elif tokens[0]=="description":
message+="description: "+tokens[1:]+"<br>\n"
elif tokens[0]=="features":
features=[]
noffeatures=int(tokens[1])
for i in range(noffeatures):
line=s.getline()
tokens=re.split(" ",line)
message+="<h3>"+str(i)+": "+tokens[2]+"</h3>\n<br>\n"
if tokens[2].endswith(".histo.gz") \
or tokens[2].endswith(".histo") \
or tokens[2].endswith(".oldhisto.gz") \
or tokens[2].endswith(".png"):
message+="<img src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\" alt=\""+tokens[2]+"\"/><br><br>\n"
elif tokens[2].endswith(".vec.gz") \
or tokens[2].endswith(".oldvec.gz") \
or tokens[2].endswith(".vec") \
or tokens[2].endswith("txt") \
or tokens[2].endswith("txt.gz") \
or tokens[2].endswith("textID"):
message+="<iframe src=\""+config.featureurl+"?server="+settings.fireServer+"&port="+str(settings.firePort)+"&image="+imagename+"&feature="+str(i)+"\"></iframe><br><br>\n"
else:
message+="not yet supported<br>\n"
line=s.getline() # get end
s.sendcmd("bye")
Display(message)
elif(form.has_key("textfeatureinfo")):
message+=displayTextFeatureInfo()
s.sendcmd("bye")
print "</pre>"
Display(message)
else:
# we do not want to show the settings window
if(form.has_key("newFile") and form["newFile"].value=="1"):
#message+=str(form["newFile"])
message+=newFile(form)
elif(form.has_key("serverfile")):
message+=serverFile(form)
elif(form.has_key("feedback")):
# result images have been selected and are part of this query
if form["feedback"].value=="requery":
# relevance feedback
message+=feedbackretrieve(form)
elif form["feedback"].value=="expandresults":
# more results
message+=expandQuery(form)
else:
# save relevances (this will not happen currently)
message=saveRelevances(form)
elif(form.has_key("queryImage") and form["queryImage"].value!=""):
# no relevance feedback, but just one of the random images was clicked
message+=retrieve(form)
elif form.has_key('textquerystring'):
# no relevance feedback, no query image, but a query string for text queries
message+=processTextQuery(form)
# now generate the remainder of the website: queryfield for meta information, random images, adminlink
if("textfeature" in settings.distances or "metafeature" in settings.distances):
message+=textQueryForm(settings)
message+=randomImages()
message+=newFileForm()
message+="<p align=\"right\">"+adminLink(settings.fireServer, settings.firePort)+"</p>"
message+="</form>"
# disconnect from server
s.sendcmd("bye")
print "</pre>"
# display the generated webpage
Display(message)
# except Exception, e:
# if anything went wrong:
# show the error message as readable as possible,
# disconnect from the server to keep it alive
# and end yourself
# print "<pre>"
# print "Exception: ",e
# s.sendcmd("bye")
# print "</pre>"
| Python |
#!/usr/bin/python
import sys,re
import filelist
filelist=filelist.FileList()
filelist.load(sys.argv[1])
for i in filelist:
cls=i[1]
fname=i[0]
print fname,
for j in filelist:
if j[1]==cls and j[0]!=fname:
print j[0],
print
#print filelist
| Python |
import re,sys,string,os
class FileList:
def __init__(self):
self.files=[]
self.classes=False
self.featuredirectories=False
self.descriptions=False
self.path=""
self.rawpath=""
self.suffices=[]
self.filename2idx={}
def __str__(self):
result=""
result+="Suffices: "+str(self.suffices)
result+="Files: "+str(self.files)
return result
def __getitem__(self,i):
return self.files[i]
def __getslice__(self,i):
return self.files[i]
def index(self,filename):
if self.filename2idx.has_key(filename):
return self.filename2idx[filename]
else:
print >>sys.stderr,filename,"not in database, returning -1"
#print >>sys.stderr,self.filename2idx
#print >>sys.stderr,self.files
return -1
#result=-1
#for i in range(len(self.files)):
# if self.files[i][0]==filename:
# result=i
#if result==-1:
# print "filelist.py: file",filename,"not in database"
#return result
def load(self, filename):
fl=open(filename,"r")
lines=fl.readlines()
for l in lines:
l=re.sub("\n","",l)
tokens=string.split(l)
if tokens[0] == "file":
f=tokens[1]
if self.classes==True:
cls=int(tokens[2])
if self.descriptions:
desc=tokens[3:]
self.files+=[[f,cls,desc]]
else:
self.files+=[[f,cls]]
elif self.descriptions:
desc=tokens[2:]
self.files+=[[f,0,desc]]
else:
self.files+=[[f,]]
elif tokens[0] == "classes":
if tokens[1]=="yes":
self.classes=True
elif tokens[0] == "descriptions" or tokens[0] == "description":
if tokens[1]=="yes":
self.descriptions=True
elif tokens[0] == "featuredirectories":
if tokens[1]=="yes":
self.featuredirectories=True
elif tokens[0] == "path":
self.rawpath=tokens[1]
self.path=self.rawpath
if self.path.startswith("this"):
self.path=self.path.replace("this+","")
offs=os.path.dirname(filename)
cwd=os.path.realpath(os.path.curdir)
self.path=offs+"/"+self.path+"/"
elif tokens[0] == "suffix":
self.suffices+=[tokens[1]]
self.initMap()
def add(self, filename, class_=0, description=""):
if description == "":
assert(self.descriptions == False)
self.files.append([filename, class_, description])
else:
assert(self.descriptions == True)
self.files.append([filename, class_])
def save(self, filename):
if filename=="-":
f=sys.stdout
else:
f=open(filename,"w")
print >>f,"FIRE_filelist"
if self.classes:
print >>f,"classes yes"
if self.descriptions:
print >>f,"descriptions yes"
if self.featuredirectories:
print >>f,"featuredirectories yes"
print >>f,"path",self.rawpath
for s in self.suffices:
print >>f,"suffix",s
for fi in self.files:
print >>f, "file %s" % " ".join(map(lambda x: str(x), fi))
def initMap(self):
self.filename2idx={}
for i in range(len(self.files)):
self.filename2idx[self.files[i][0]]=i
def basenameify(self):
for i in range(len(self.files)):
if len(self.files[i])==3:
self.files[i]=(os.path.basename(self.files[i][0]),self.files[i][1], self.files[i][2])
if len(self.files[i])==2:
self.files[i]=(os.path.basename(self.files[i][0]),self.files[i][1])
if len(self.files[i])==1:
self.files[i]=(os.path.basename(self.files[i][0]),)
def nameWithSuffix(self, item, suffix):
assert item < len(self.files)
assert suffix < len(self.suffices)
return self.files[item][0] + "." + self.suffices[suffix]
# generator for images with full path
def fullpaths(self):
for image in self.files:
yield os.path.join(self.rawpath, image[0])
def get_full_path_for_item(self, item, suffixidx=-1):
name = os.path.join(self.rawpath, self.files[item][0])
if suffixidx > -1 and len(self.suffices) > 0:
name += "." + self.suffices[suffixidx]
return name | Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
#-----------------------------------------------------------------
s=firesocket.FIRESocket()
try:
if len(sys.argv) < 5:
print """USAGE:
makedistfiles.py <querylist> <server> <port> <suffix> [-q]
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
suffix=sys.argv[4]
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
lines=ql.readlines()
lines=map(lambda line:
re.sub("\n","",line), lines)
ql.close()
i=0
for line in lines:
print i,"/",len(lines)
i+=1
tokens=re.split(" ",line)
file=tokens[0]
cmd="savedistances "+file+" "+file+"."+suffix
print "SENDING: ",cmd
s.sendcmd(cmd)
res=s.getline()
print res
sys.stdout.flush()
sys.stdout.flush()
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
filelist=filelist.FileList()
filelist.load(sys.argv[1])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in filelist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class SparseHistogramFeature:
def __init__(self):
self.bins={}
self.data={}
self.steps=[]
self.stepSize=[]
self.min=[]
self.max=[]
self.dim=0
self.numberofBins=0
self.counter=0
self.filename=""
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_sparse_histogram"
print >>f,"# sparse histogram file for FireV2"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps",reduce(lambda x,y: str(x)+" "+str(y),self.steps)
print >>f,"min",reduce(lambda x,y: str(x)+" "+str(y),self.min)
print >>f,"max",reduce(lambda x,y: str(x)+" "+str(y),self.max)
print >>f,"bins",len(self.bins)
for i in self.bins:
print >>f,"data",i,self.bins[i]
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_sparse_histogram":
fine=True
elif line[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="bins":
self.numberofBins=int(toks[1])
elif toks[0]=="data":
self.bins[toks[1]]=int(toks[2])
self.data[toks[1]]=float(toks[2])/float(self.counter)
else:
print "Unparsed line: '"+line+"'."
| Python |
#------------------------------------------------------------------------------
# proxyserver implementation for communication with the web frontend and fire
# server system
# server must be reachable via network sockets in order to be of any use
#
#------------------------------------------------------------------------------
__author__ = "Jens Forster <jens.forster@rwth-aachen.de>"
__version__= "0.1"
from SocketServer import *
from firesocket import *
# Needed for stringhandling
# re provides regular expressions and methods
import re
import string
# Needed for suitable choosing of random images
# hence each retrieval server will only have access to unique
# part of the database
import random
# Needed for logging
import logging
#from logging.handlers import TimedRotatingFileHandler
import sys
import os
import os.path
__all__ =["FIREProxyServer", "FIREProxyHandler"]
class FIREProxyServer(TCPServer):
request_queque_size = 1
allow_reuse_address = True
def __init__(self,server_address, retr_addrs, RequestHandlerClass,logDestination=None):
TCPServer.__init__(self,server_address,RequestHandlerClass)
if len(retr_addrs) == 0:
raise RuntimeError, "no retrieval servers spezified"
else:
self.retr_addrs = retr_addrs
self.retr_sock = []
# initializing loggin system
self.log = self._server_log(logDestination)
self.log.log(logging.INFO,'Initializing proxyserver')
self.log.log(logging.DEBUG,'Logging system and retrieval servers list ready')
# initializing boolean values for request and server termination
# and user authorization
self.notBye = True # true -> web interface didn't quit yet
self.notQuit = True # true -> proxy and retrieval servers still running
self.notProxyquit = True # false -> proxy going down, retrieval servers will remain online
self.auth = False
self.sendErrorsFrom = [] # a list which will contain retrieval servers which crashed or aren't reachable anymore
self.results = 0 # given results per retrieval server
self.log.log(logging.INFO,"Proxyserver running on "+str(server_address))
def sendToRetrievers(self,cmd,upperBound=None):
if upperBound != None:
if upperBound < 0:
raise RuntimeError, "no retrieval servers specified for sending, Proxyserver: method sendToRetrievers: upperBound "+str(upperBound)
relevantRetr = zip(self.retr_sock,range(upperBound))
for sock,idx in relevantRetr:
try:
sock.sendcmd(cmd)
self.log.log(logging.getLevelName("SEND "+str(self.retr_addrs[idx])),cmd)
except:
# iff no socket error occured until now
name,port = self.retr_addrs[idx]
if self.sendErrorsFrom == []:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str((name,port))+" did break" )
raise RuntimeError, "socket connection to retrieval server "+str((name,port))+" did break"
elif str(name) in self.sendErrorsFrom:
pass
else:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str((name,port))+" did break")
else:
for sock,addr in zip(self.retr_sock,self.retr_addrs):
try:
sock.sendcmd(cmd)
self.log.log(logging.getLevelName("SEND "+str(addr)),cmd)
except:
name,port = addr
self.log.log(logging.INFO,"sendErrorsFrom: "+str(self.sendErrorsFrom))
if self.sendErrorsFrom == []:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str(addr)+" did break")
raise RuntimeError, "socket connection to retrieval server "+str(addr)+" did break"
elif str(name) in self.sendErrorsFrom:
self.log.log(logging.DEBUG,"error in socket communication for retrieval server "+str(name)+" already known")
else:
self.sendErrorsFrom.append(str(name))
self.log.log(logging.CRITICAL,"socket connection to retrieval server "+str(addr)+" did break")
def getFromRetrievers(self):
answer = []
for sock,addr in zip(self.retr_sock,self.retr_addrs):
try:
ans = sock.getline()
except:
self.log.log(logging.CRITICAL,"retrieval server "+str(addr)+" doesn't respond")
name,port = addr
self.sendErrorsFrom.append(str(name))
raise RuntimeError, "retrieval server "+str(addr)+" didn't respond in method getFromRetrievers"
answer.append(ans)
self.log.log(logging.getLevelName("RECV "+str(addr)),ans)
return answer
def handle_request(self):
"""Handle one request in terms of one connection possibly containing
several actions, possibly blocking."""
errorCode = 0
try:
request, client_address = self.get_request()
self.log.log(logging.INFO,"client "+str(client_address)+" connected")
except:
self.notQuit = False
self.log.log(logging.ERROR,"an error occured during client's connection attempt")
self.log.log(logging.ERROR,"maybe the socket/pipe did break or there are network problems")
self.log.log(logging.ERROR,"proxyserver shutting down (retrieval server system may still be online)")
try:
request.close()
self.socket.close()
except:
pass
errorCode = 1
return errorCode
else:
try:
for serverNum in range(len(self.retr_addrs)):
serverNum_sock = FIRESocket()
self.retr_sock.append(serverNum_sock)
except:
self.notQuit = False
self.log.log(logging.ERROR,"couldn't create necessary network sockets for communication with retrieval servers")
self.log.log(logging.ERROR,"proxyserver shutting down (retrieval server system may still be online)")
request.close()
self.socket.close()
while self.retr_sock != []:
firesock = self.retr_sock.pop(0)
firesock.sock.close()
del firesock
errorCode = 2
return errorCode
else:
self.log.log(logging.DEBUG,"network sockets for communication with retrieval servers created")
if self.verify_request(request, client_address):
try:
self.process_request(request, client_address)
return errorCode
except RuntimeError, msg:
self.handle_error(request,msg)
errorCode = 3
return errorCode
except:
self.handle_error(request,"an error involving networksockets or an unknown error occured")
errorCode = 3
return errorCode
def process_request(self , request, client_addr):
self.finish_request(request, client_addr)
try:
self.close_request(request)
except:
pass
def close_request(self, client):
client.close()
self.log.log(logging.INFO,"client disconnected")
self.auth = False
# the network sockets designated for communication with retrieval servers
# will only be closed if bye, quit or an error are encountered
if self.notQuit:
self.sendToRetrievers("bye")
while self.retr_sock != []:
firesock = self.retr_sock.pop(0)
firesock.sock.close()
del firesock
def serve(self):
self.log.log(logging.INFO,"proxy handling requests")
# iff the error code is True no error occured while handling
errorCode = 0
while self.notQuit and self.notProxyquit :
errorCode = self.handle_request()
self.log.log(logging.DEBUG,"errorCode "+str(errorCode))
if errorCode == 0:
self.server_close()
else:
logging.shutdown()
def server_close(self):
# no auth necessary due to prior authentification
if self.notProxyquit:
self.log.log(logging.INFO,"sending shutdown signal to retrieval servers")
self.sendToRetrievers("quit")
self.socket.close()
# closing and deleting network sockets
while self.retr_sock != []:
sock = self.retr_sock.pop(0)
sock.sock.close()
del sock
if self.notProxyquit:
self.log.log(logging.INFO,"proxyserver and retrieval servers stopped")
else:
self.log.log(logging.INFO,"proxyserver stopped, retrieval servers running")
self.log.log(logging.DEBUG,"shutting logging system down")
logging.shutdown()
# printing the error message in to the logging file, closing connection to all retrieval servers
# and shutting down
def handle_error(self,request,msg):
self.notQuit = False
self.log.log(logging.ERROR,"an error occured while handling requests")
self.log.log(logging.ERROR,"for further information view the log file")
self.log.log(logging.DEBUG,msg)
self.log.log(logging.ERROR,"shutting proxyserver down")
request.close()
self.sendToRetrievers("bye")
self.socket.close()
while self.retr_sock != []:
sock = self.retr_sock.pop(0)
sock.sock.close()
del sock
del self.retr_sock
# logging.shutdown()
# Note that _server_log will always return a logger which will always log global statements to stdout
# and if intended will log everything to a file depending on the message's importance level
def _server_log(self,logDestination):
if logDestination == None:
location = None
else:
# exists is only needed if in future an other file logger should be used
# that doesn't check fileexistance
location, exists = self._logFilehandling(logDestination)
# Setting up the logging System
# The root logger shall log to stdout, all debug msg will be found in the files
# Adding new message levels to the logging system
# DEBUG (10) and NOTSET (0) needn't to be altered
logging.addLevelName(1020,"CRITICAL")
logging.addLevelName(1010,"ERROR")
logging.addLevelName(1005,"WARNING")
logging.addLevelName(1000,"INFO")
logging.addLevelName(11,"RECV (Web/Bash)")
logging.addLevelName(12,"SEND (Web/Bash)")
logging.INFO = 1000
logging.ERROR = 1010
logging.CRITICAL = 1020
logging.WARNING = 1005
# so the levels 13 up to 999 are free for retrieval servers
# because every server needs two numbers this logging system is capable of 493 retrieval servers
# now levels for the retrieval servers are added
# they are formated as SEND/RECV (IP-Address/Name)
level = 13
for addrs in self.retr_addrs:
logging.addLevelName(level,"SEND "+str(addrs))
logging.addLevelName(level+1,"RECV "+str(addrs))
level += 2
# now configurating the root logger which will log only to stdout
sys.stdout = sys.stderr
root = logging.getLogger('')
root.setLevel(logging.DEBUG)
console = logging.StreamHandler(sys.stdout)
console.setLevel(logging.INFO)
formatter = logging.Formatter('%(asctime)s %(levelname)-10s %(message)s','%a, %d-%m-%Y %H:%M:%S')
console.setFormatter(formatter)
root.addHandler(console)
# setting up the main file logger
if location != None:
# when = 'midnight'
# filesPerDay = 1
# backUpCount = 7
# file = TimedRotatingFileHandler(location,when,filesPerDay,backUpCount)
if exists:
file = logging.FileHandler(location,'a')
else:
file = logging.FileHandler(location,'w')
file.setLevel(logging.DEBUG)
formatter = logging.Formatter('%(asctime)s %(levelname)-6s %(message)s','%a, %d-%m-%Y %H:%M:%S')
file.setFormatter(formatter)
root.addHandler(file)
return root
def _logFilehandling(self,logDestination):
if logDestination == "default":
cwd = os.getcwd()
# determinating whether mac, unix/linux or windows
if os.name in ["mac","nt","posix"]:
if os.name == "nt":
# windows os
osPathTag = "\\"
else:
# unix/linux os or macintosch
osPathTag = "/"
try:
# checking whether dir exists
location = cwd+osPathTag+"fireproxylogs"
exist = os.path.isdir(location)
if not exist:
os.mkdir(location)
location += osPathTag+"proxylog"
del exist
del osPathTag
except:
raise RuntimeError, "error initalizing logging system (default logging location)"
else:
raise RuntimeError, "OS not supported for logging"
del cwd
# iff the user specified a storing location
# outer if-clause
else:
# if only a directory is given (logDestination must be terminated by the os's
# path separator (e.g / on linux \ on windows os)) default logfile name will be used
head,tail = os.path.split(logDestination)
osType = os.name
if osType in ["mac","posix"]:
osPathTag = "/"
elif osType == "nt":
osPathTag = "\\"
else:
raise RuntimeError, "OS not supported for logging"
# checking whether a path name is given
if head != "":
# checking whether the user set all path separators' right
head = os.path.normcase(head)
head = os.path.normpath(head)
# figuring out whether the supplied directories exist and
# if not which have to be created
if not os.path.isdir(head):
dirToCreate = [head]
newHead, newTail = os.path.split(head)
while newTail != "" and (not os.path.isdir(newHead)):
dirToCreate.append(newHead)
newHead, newTail = os.path.split(newHead)
while dirToCreate != []:
os.mkdir(dirToCreate.pop())
del newHead
del newTail
# path is valid now checking if a filename was supplied
if tail != "":
location = logDestination
else:
location = logDestination+osPathTag+"proxylog"
del dirToCreate
# no path given by the user
# tail must be != "" in this case otherwise it would be the default case
else:
# assuming current workdirectory was intended
cwd = os.getcwd()
location = cwd+osPathTag+tail
del osPathTag
del osType
del head
del tail
# Outer If Clause ends here
# testing if the file exist or not
exists = True
try:
file = open(location,'r+')
except IOError:
exists = False
else:
file.close()
return (location,exists)
# Overidden to be useless, because this proxy shall quit when all retrievers quit
def serve_forever(self):
pass
class FIREProxyHandler(StreamRequestHandler):
def __init__(self, request, client_addr, server):
StreamRequestHandler.__init__(self,request, client_addr, server)
def sendToWeb(self,msg):
self.wfile.write(msg+"\r\n")
self.server.log.log(logging.getLevelName("SEND (Web/Bash)"),msg)
def _validateSettings(self,settings):
self.server.log.log(logging.DEBUG,"valSet Argu:"+repr(settings))
leng = len(settings)
self.server.log.log(logging.DEBUG,"len: "+repr(leng))
if leng:
if settings[0][0] != "filelist":
retur = (False,"Settingsstream starts with wrong keyword")
else:
retur = (True, "")
if leng > 1:
lenSingle = map(lambda single: len(single),settings)
boolList = map(lambda single: single == lenSingle[0],lenSingle)
if not (reduce(lambda fst,snd: fst == snd,boolList)):
retur = (False,"At least one retrieval server didn't send complete settings")
# Fine all retrieval servers did send there complete settings
# or all made the same mistake :D
lenSetting = lenSingle[0]
# deleting helper variables
del boolList
# checking rest
for idx in range(lenSetting):
if (idx != 1):
for jdx in range(leng):
if (settings[0][idx] != settings[jdx][idx]):
# settings differ and singaling it
return (False,"settings of servers "+str(0)+" and "+str(jdx)+" differ in "+ settings[0][idx]+" "+settings[jdx][idx])
else:
retur = (True,"")
return retur
else:
return (False,"no info information at all received")
def setup(self):
StreamRequestHandler.setup(self)
error = False
try:
for retr, addr in zip(self.server.retr_sock,self.server.retr_addrs):
name, port = addr
try:
retr.connect(name,port)
self.server.log.log(logging.DEBUG,"connection to retrival server "+str(addr)+" established")
except:
# sending error message to web frontend
self.sendToWeb("failure "+str(name)+" down")
self.server.sendErrorsFrom.append(str(name))
# closing
error = True
self.server.log.log(logging.CRITICAL,"retrieval server "+str(addr)+" not reachable")
except:
pass
else:
if error:
self.finish()
raise RuntimeError, "some of the retrieval servers arent't reachable via network"
else:
self.server.notBye = True
# signal: connection to retrieval servers established
# currently unix specific linefeed
self.server.log.log(logging.INFO,"connection to retrieval server system established")
def handle(self):
retrieveAndMetaCommands = ["retrieve","expand","retrieveandsaveranks","metaretrieve","metaexpand","textretrieve","textexpand"]
# handling actions until "bye", "quit" or doomsday
while self.server.notBye and self.server.notQuit:
webMsg = self.rfile.readline()
self.server.log.log(logging.DEBUG,webMsg)
webMsg = re.sub("[\r\n\t\v\f]","",webMsg)
self.server.log.log(logging.getLevelName("RECV (Web/Bash)"),webMsg)
webKeyS = string.split(webMsg)
keyWord = webKeyS.pop(0) # grabbing the very first element
if keyWord == "info":
self.server.sendToRetrievers("info")
answer = self.server.getFromRetrievers()
# convert list of single strings in usable format
# e.g ["hello world","how do you do?"] becomes
# [['hello','world'],['how','do','you','do?']]
settings = map(lambda strin: string.split(strin),answer)
bool, comment = self._validateSettings(settings)
if bool:
sum = 0
# summing up total databank size
for idx in range(len(settings)):
sum += int(settings[idx][1])
settings[0][1] = sum
self.server.results = int(settings[0][3])
self.sendToWeb(re.sub("[[,'\]]","",str(settings[0])))
# deleting helper variables
del bool
del answer
del comment
del settings
else:
self.finish()
raise RuntimeError, "handle():"+comment
elif keyWord == "password":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
bool = True
idx = 0
while bool and idx < len(answer):
if tokens[idx][0] != "ok":
bool = False
idx += 1
if not bool:
self.sendToWeb(answer[idx-1])
self.server.auth = False
else:
self.sendToWeb("ok")
self.server.auth = True
# deleting helper variables
del answer
del tokens
del idx
del bool
elif keyWord == "bye":
self.server.notBye = False
elif keyWord == "quit":
self.server.notBye = False
if self.server.auth:
self.server.notQuit = False
else:
self.server.sendToRetrievers("bye")
elif keyWord in retrieveAndMetaCommands:
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
self.server.log.log(logging.DEBUG,"answer(handle()): "+repr(answer))
# converting retrieval servers' answers from strings to list of
# strings like above and afterwards merging all such lists into
# one
if keyWord in ["metaretrieve","metaexpand","textretrieve","textexpand"]:
listed = map(lambda string: re.split("[ :]",string),answer)
else:
listed = map(lambda strin: string.split(strin),answer)
listed = reduce(lambda fst,snd: fst+snd,listed)
if "nometainformation" in listed:
self.sendToWeb("nometainformation")
elif "notextinformation" in listed:
self.sendToWeb("notextinformation")
elif "" in listed:
self.server.log.log(logging.INFO,"at least one retrival server answered with an empty string")
self.server.log.log(logging.ERROR,"retrival aborted")
elif not self.server.results == 0:
# for sorting purposes changing data structure to a list of tupels
# consistent of distance as a float and filename as a string
toSort = [(float(listed[i+1]),listed[i]) for i in range(0,len(listed),2)]
toSort.sort()
toSort.reverse()
retrievalString = ""
for tup in toSort[:self.server.results]:
dist, name = tup
if keyWord in ["metaretrieve","metaexpand","textretrieve","textexpand"]:
retrievalString += name+":"+str(dist)
else:
retrievalString += name+" "+str(dist)
retrievalString += " "
# removing trailing blank
retrievalString = re.sub(" $","",retrievalString)
self.sendToWeb(retrievalString)
#deleting helper variables
del toSort
del retrievalString
else:
self.server.log.log(logging.INFO,"Proxyserver assumes that settings for results equals 0")
self.server.log.log(logging.INFO,"Please use setresults to set results to an usefull value")
self.server.log.log(logging.DEBUG,"possible mismath between retrieval servers settings and proxyserver")
self.server.log.log(logging.ERROR,"retrieval aborted")
# deleting helper variables
del answer
del listed
elif keyWord == "help":
self.server.sendToRetrievers(webMsg,1)
answer = self.server.retr_sock[0].getline()
ans = string.split(answer)
try:
ans.remove("filelist")
except:
pass
answer = "proxy only: quitproxy ||normal fire commands: "
for cmd in ans:
answer += cmd
answer += " "
answer += "||not supported in proxy mode: filelist"
self.sendToWeb(answer)
del answer
elif keyWord == "savedistances":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
fst = answer[0]
bool = True
for name in answer:
if name != fst:
bool = False
if not bool:
self.finish()
raise RuntimeError,"distance filenames differ"
self.sendToWeb(fst)
del answer
del bool
del fst
elif keyWord == "random":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
tokenList = reduce(lambda fst,snd:fst+snd,tokens)
# choosing randomly images from all random images
# because every retrieval server only accesses an unique database
# subset
randomImg = random.sample(tokenList,self.server.results)
self.sendToWeb(re.sub("[[',\]]","",str(randomImg)))
# deleting helper variables
del answer
del tokens
del tokenList
del randomImg
elif keyWord == "saverelevances":
self.server.sendToRetrievers(webMsg)
elif keyWord in ["setscoring","setresults","setextensions","setdist","setweight"]:
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
keyToken = re.sub("set","",keyWord)
# checking if all retrieval servers accepted new settings
# first checking keyword
boolList = map(lambda list: list[0] == keyToken,tokens)
boolAll = reduce(lambda fst,snd:fst == snd,boolList)
# checking iff snd argument must be checked
# in case it will be "=" testing is omitted
if keyToken in ["weight","dist"]:
sndArg = tokens[0][1]
boolList = map(lambda list: list[1] == sndArg,tokens)
boolSnd = reduce(lambda fst,snd: fst == snd,boolList)
boolAll = boolAll and boolSnd
del boolSnd
# checking third argument
trdArg = tokens[0][2]
boolList = map(lambda list: list[2] == trdArg,tokens)
boolTrd = reduce(lambda fst,snd:fst == snd,boolList)
boolAll = boolAll and boolTrd
if boolAll:
self.sendToWeb(answer[0])
if keyToken == "results" and tokens[0][0] != "Invalid":
self.server.results = int(tokens[0][2])
else:
self.server.sendToRetrievers(keyWord,1)
alterMsg = self.server.retr_sock[0].getline()
self.sendToWeb(alterMsg)
del alterMsg
del boolAll
del boolTrd
del boolList
del answer
del tokens
elif keyWord == "listfiles":
self.server.sendToRetrievers(webMsg)
answer = self.server.getFromRetrievers()
result = ""
for idx in range(len(answer)):
result += answer[idx]
result += " "
# deleting trailing blank
result = re.sub(" $","",result)
self.sendToWeb(result)
del answer
del result
elif keyWord == "class":
self.server.sendToRetrievers(webMsg)
self.server.log.log(logging.INFO,webMsg+" send")
answer = self.server.getFromRetrievers()
tokens = map(lambda strin: string.split(strin),answer)
tokens = reduce(lambda fst,snd: fst+snd,tokens)
# since retrieval servers will only return a single value
# it is only neccessary to destinguish the three possible modes of the class command
# for error handling
# do all retrieval servers have classes or don't they
if "yes" in tokens:
for item in tokens:
if "yes" != item:
self.finish()
raise RuntimeError, "not all retrieval servers do have classes"
self.sendToWeb("yes")
elif "no" in tokens:
for item in tokens:
if "no" != item:
self.finish()
raise RuntimeError, "not all retrieval servers don't have classes"
self.sendToWeb("no")
else:
try:
intList = []
for item in tokens:
intItem = int(item)
if intItem >= 0:
intList.append(intItem)
except:
self.finish()
raise RuntimeError, "neiter yes/no nor numbers were returned for classes"
else:
boolList = map(lambda it: intList[0]==it,intList)
booL = reduce(lambda fst, snd: fst and snd,boolList)
if booL:
self.sendToWeb(str(intList[0]))
else:
self.finish()
raise RuntimeError, "retrieval servers classes differ"
del boolList
del booL
del tokens
del answer
elif keyWord == "metafeatureinfo":
self.server.sendToRetrievers(webMsg)
self.server.log.log(logging.INFO,webMsg+" send")
answer = self.server.getFromRetrievers()
answer = map(lambda string: re.sub(" $","",string),answer)
returnstring = ""
for strings in answer:
returnstring += strings
returnstring += " "
self.sendToWeb(re.sub(" $","",returnstring))
del answer
del returnstring
# Note that the commands filelist is dumped on purpose
# because every retrieval server accesses it's own databade subset
# The interactor command is dumped too.
elif keyWord in ["filelist","interactor"]:
self.sendToWeb(keyWord+" is not avaible since fireproxy is in use")
elif keyWord == "quitproxy":
self.server.notBye = False
self.server.notProxyquit = False
else:
self.sendToWeb("Unknown command: "+webMsg)
del retrieveAndMetaCommands
del webKeyS
del webMsg
| Python |
#!/usr/bin/python
from string import split, lower, strip
from random import shuffle
import sys
import os
from os.path import join, getsize
ARCH=os.environ["ARCH"]
# returns the string after a certain option string
def getStringAfter(name, array, default):
for i in range(len(array)):
if array[i] == name:
if (len(array) == i+1):
return default
else:
return array[i+1]
return default
def usage(commandline):
print "note: before you use this program make \"make all\" in the firedirectory. "
print commandline[0], "(-np|-npfromdb|-ndb|-ndbfromp|-showf|-e|-h) [options]"
print
print "mandatory options (choose exactly one of them):"
print " -h shows this help"
print " -showf show all possible features that can be extracted"
print " -e extracts features. options: -limages -ldb -ld -selectf "
print " -p -db -d -c -f -q -n"
print " -np creates a new purefiles file using the filenames of the "
print " files in the directory. options: -d -p"
print " -npfromdb creates a new purefiles file from the data of the dbfile."
print " options: -d -db -p"
print " -ndb creates a new database file using the filenames of the "
print " files in the directory. options: -d -db"
print " -ndbfromp creates a new database file from the data of the purefiles "
print " file. options: -d -p -db"
print
print "options:"
print " -d dir directory of where the files are. default: ."
print " -p path path of an [existing] purefiles file. default: purefiles"
print " -db path path of an [existing] database file. default: list"
print " -c path configfile that shall be used. "
print " default:", os.path.dirname(sys.argv[0]) + "/features.conf"
print " -f dir directory where fire resides. "
print " default:", os.path.dirname(sys.argv[0]) + "../../bin/"+ARCH+"/"
print " -q (day|week)"
print " put all jobs into the queuingsystem. default: no queuing"
print " -n name of the jobs in the queue. default: extractor"
print " -selectf feature1 [feature2 [... [featureN]]"
print " extracts only feature1 ... featureN of the config file"
print " -v verbose"
print " -limages image1 [image2 [... [imageN]]"
print " extracts features of image1 ... imageN only"
print " -ldb extracts the images from the dbfile"
print " (hint: it is smarter to create a db file first, using"
print " -ndb or -ndbfromp)"
print " -ld extracts the images of the directory"
print " (hint: it is smarter to create a purefiles file first,"
print " using -np or -npfromdb)"
class Extractor:
def __init__(self):
self.datadir = "."
self.purefiles = "purefiles"
self.dbfile = "list"
self.progdir = os.path.dirname(sys.argv[0]) # path of the tool
self.firedir = self.progdir + "/../../bin/"+ARCH+"/"
self.config = "features.conf"
self.extensionlist = ["jpg", "png", "gif", "jpeg", "tif", "tiff"]
self.images = []
self.features = {}
self.selectedFeatures = []
self.queue = None
self.name = "extractor"
# gets for the attributes
def setPurefiles(self, new):
self.purefiles = new
def setDBFile(self, new):
self.dbfile = new
def setDatadir(self, new):
self.datadir = new
def setFiredir(self, new):
self.firedir = new
def setConfigfile(self, new):
self.config = new
def setExtensionlist(self, newl):
self.extensionlist = newl
def setQueue(self, new):
self.queue = new
def setName(self, new):
self.name = new
# printing
def printInfo(self):
print "config info of extractor"
print "datadir", self.datadir
print "purefiles", self.purefiles
print "dbfile", self.dbfile
print "progdir", self.progdir
print "firedir", self.firedir
print "configfile", self.config
print "extensionlist", self.extensionlist
# prints available features
def printFeatures(self):
for feature in self.features.keys():
print feature, self.features[feature]
def printImages(self):
if self.images == []:
print "error: no images loaded."
else:
print "images:"
for image in self.images:
print image
# loading
# load images from the directory
def loadImagesFromDir(self):
print "loading images from directory."
self.images = []
for root, dirs, files in os.walk(self.datadir):
for file in files:
if self.checkExtension(file):
if (root == self.datadir):
self.images.append(file)
else:
self.images.append(root[len(self.datadir)+1:] + "/" + file)
if self.images == []:
print "error: no images found."
# loads images from database file
# (does not consider path information of the db file yet)
def loadImagesFromDB(self):
print "loading images from dbfile."
f = open(self.dbfile, "r")
lines = f.readlines()
f.close()
for line in lines:
if strip(split(line)[0]) == "file":
self.images.append(strip(split(line)[1]))
#print "file:", strip(split(line)[1])
if self.images == []:
print "error: no images found."
# loads features from the config file
def loadFeatures(self):
print "loading features file."
try:
f = open(self.config, "r")
except IOError:
print "warning: no specific config file found. I try the one in the tooldir."
f = open(self.progdir + "/features.conf")
lines = f.readlines()
f.close()
for line in lines:
if line[0] == "#":
#print "comment: ", line
pass
elif len(strip(line)) < 1:
pass
else:
key = strip(split(line)[0])
#print "key", key
feature = strip(line)[len(key)+1:]
#print "feature", feature
self.features[key] = feature
# loads an existing purefiles list
def loadImagesFromPurefiles(self):
print "loading images from purefiles."
self.images = []
file = open(self.purefiles, 'r')
lines = file.readlines()
file.close()
num = len(lines)
print "info: number of images =", num
for line in lines:
self.images.append(strip(line))
# loads images from commandline
def loadImagesFromCommandline(self, argv):
self.images = []
print "loading images from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-limages":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
self.images.append(command)
if self.images == []:
print "error: no images found"
def loadSelectedFeaturesFromCommandline(self, argv):
self.selectedFeatures = []
print "loading selected features from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-selectf":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
#print "added", command
self.selectedFeatures.append(command)
if self.selectedFeatures == []:
print "error: no features found"
# checking
# check if image file has a correct extension
def checkExtension(self, filename):
extension = split(filename, ".")[-1]
if lower(extension) in self.extensionlist:
return True
return False
def pureFileExists(self):
try:
f = open(self.purefiles, "r")
except IOError:
return False
else:
return True
# writing
# writes the purefilesfile
def writePurefiles(self):
print "writing purefiles file."
if self.images == []:
print "error: cannot write a purefiles file, because image list is empty."
else:
f = open(self.purefiles, "w")
for image in self.images:
f.write(image)
f.write("\n")
f.close()
# creates the database file
def writeDBFile(self):
# possible feature: dbfiles with feature directories?
# possible feature: non recursive traversion
print "writing dbfile."
f = open(self.dbfile, "w")
f.write("FIRE_filelist\n")
f.write("classes no\n")
f.write("descriptions no\n")
f.write("featuredirectories no\n")
f.write("path this\n")
for file in self.images:
f.write("file ")
f.write(file)
f.write("\n")
f.close()
# selection
# selects features
def selectFeatures(self):
if self.selectedFeatures == []:
print "error: there are no features to select"
print "hint: maybe wrong spelling?"
temp = {}
for key in self.features.keys():
#print "key", key
if key in self.selectedFeatures:
temp[key] = self.features[key]
self.features = temp
# extraction
# extracts all features for all images
def extractFeatures(self):
for key in self.features.keys():
self.extractFeature(key)
# extracts one feature for all images
def extractFeature(self, feature):
if self.images != []:
commandline = self.firedir + self.features[feature] + " --images"
for image in self.images:
commandline = commandline + " " + image
else:
if not self.pureFileExists():
print "error: purefiles file not existing"
print "hint: create with -np a new file or use -limage, -ldb or -ld to load images"
return False
else:
#self.printInfo()
#print "self.features[feature]", self.features[feature]
#self.printFeatures()
commandline = self.firedir + self.features[feature] + " --filelist " + self.purefiles
print "commandline:", commandline
if self.queue == "day":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 8:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
elif self.queue == "week":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 24:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
else:
os.system(commandline)
if __name__ == "__main__":
if "-h" in sys.argv:
usage(sys.argv)
else:
e = Extractor()
# adjusting file locations
if "-p" in sys.argv:
e.setPurefiles(getStringAfter("-p", sys.argv, None))
if "-db" in sys.argv:
e.setDBFile(getStringAfter("-db", sys.argv, None))
if "-d" in sys.argv:
e.setDatadir(getStringAfter("-d", sys.argv, None))
if "-c" in sys.argv:
e.setConfigfile(getStringAfter("-c", sys.argv, None))
if "-f" in sys.argv:
e.setFiredir(getStringAfter("-f", sys.argv, None))
if "-q" in sys.argv:
e.setQueue(getStringAfter("-q", sys.argv, None))
if "-n" in sys.argv:
e.setName(getStringAfter("-n", sys.argv, None))
## maybe later
# if "-el" in sys.argv
# e.setExtensionlist(...)
if "-v" in sys.argv:
e.printInfo()
if "-np" in sys.argv:
e.loadImagesFromDir()
e.writePurefiles()
elif "-npfromdb" in sys.argv:
e.loadImagesFromDB()
e.writePurefiles()
elif "-ndb" in sys.argv:
e.loadImagesFromDir()
e.writeDBFile()
elif "-ndbfromp" in sys.argv:
e.loadImagesFromPurefiles()
e.writeDBFile()
elif "-showf" in sys.argv:
e.loadFeatures()
e.printFeatures()
elif "-e" in sys.argv:
# loading the images (if no purefiles file is used)
if "-limages" in sys.argv:
e.loadImagesFromCommandline(sys.argv)
if "-v" in sys.argv:
e.printImages()
elif "-ldb" in sys.argv:
e.loadImagesFromDBFile()
elif "-ld" in sys.argv:
e.loadImagesFromDatadir()
else:
pass
e.loadFeatures()
if "-selectf" in sys.argv:
e.loadSelectedFeaturesFromCommandline(sys.argv)
e.selectFeatures()
if "-v" in sys.argv:
print "selected features:"
e.printFeatures()
e.extractFeatures()
else:
print "wrong usage. use -h for help."
| Python |
#!/usr/bin/python
from string import split, lower, strip
from random import shuffle
import sys
import os
from os.path import join, getsize
ARCH=os.environ["ARCH"]
# returns the string after a certain option string
def getStringAfter(name, array, default):
for i in range(len(array)):
if array[i] == name:
if (len(array) == i+1):
return default
else:
return array[i+1]
return default
def usage(commandline):
print "note: before you use this program make \"make all\" in the firedirectory. "
print commandline[0], "(-np|-npfromdb|-ndb|-ndbfromp|-showf|-e|-h) [options]"
print
print "mandatory options (choose exactly one of them):"
print " -h shows this help"
print " -showf show all possible features that can be extracted"
print " -e extracts features. options: -limages -ldb -ld -selectf "
print " -p -db -d -c -f -q -n"
print " -np creates a new purefiles file using the filenames of the "
print " files in the directory. options: -d -p"
print " -npfromdb creates a new purefiles file from the data of the dbfile."
print " options: -d -db -p"
print " -ndb creates a new database file using the filenames of the "
print " files in the directory. options: -d -db"
print " -ndbfromp creates a new database file from the data of the purefiles "
print " file. options: -d -p -db"
print
print "options:"
print " -d dir directory of where the files are. default: ."
print " -p path path of an [existing] purefiles file. default: purefiles"
print " -db path path of an [existing] database file. default: list"
print " -c path configfile that shall be used. "
print " default:", os.path.dirname(sys.argv[0]) + "/features.conf"
print " -f dir directory where fire resides. "
print " default:", os.path.dirname(sys.argv[0]) + "../../bin/"+ARCH+"/"
print " -q (day|week)"
print " put all jobs into the queuingsystem. default: no queuing"
print " -n name of the jobs in the queue. default: extractor"
print " -selectf feature1 [feature2 [... [featureN]]"
print " extracts only feature1 ... featureN of the config file"
print " -v verbose"
print " -limages image1 [image2 [... [imageN]]"
print " extracts features of image1 ... imageN only"
print " -ldb extracts the images from the dbfile"
print " (hint: it is smarter to create a db file first, using"
print " -ndb or -ndbfromp)"
print " -ld extracts the images of the directory"
print " (hint: it is smarter to create a purefiles file first,"
print " using -np or -npfromdb)"
class Extractor:
def __init__(self):
self.datadir = "."
self.purefiles = "purefiles"
self.dbfile = "list"
self.progdir = os.path.dirname(sys.argv[0]) # path of the tool
self.firedir = self.progdir + "/../../bin/"+ARCH+"/"
self.config = "features.conf"
self.extensionlist = ["jpg", "png", "gif", "jpeg", "tif", "tiff"]
self.images = []
self.features = {}
self.selectedFeatures = []
self.queue = None
self.name = "extractor"
# gets for the attributes
def setPurefiles(self, new):
self.purefiles = new
def setDBFile(self, new):
self.dbfile = new
def setDatadir(self, new):
self.datadir = new
def setFiredir(self, new):
self.firedir = new
def setConfigfile(self, new):
self.config = new
def setExtensionlist(self, newl):
self.extensionlist = newl
def setQueue(self, new):
self.queue = new
def setName(self, new):
self.name = new
# printing
def printInfo(self):
print "config info of extractor"
print "datadir", self.datadir
print "purefiles", self.purefiles
print "dbfile", self.dbfile
print "progdir", self.progdir
print "firedir", self.firedir
print "configfile", self.config
print "extensionlist", self.extensionlist
# prints available features
def printFeatures(self):
for feature in self.features.keys():
print feature, self.features[feature]
def printImages(self):
if self.images == []:
print "error: no images loaded."
else:
print "images:"
for image in self.images:
print image
# loading
# load images from the directory
def loadImagesFromDir(self):
print "loading images from directory."
self.images = []
for root, dirs, files in os.walk(self.datadir):
for file in files:
if self.checkExtension(file):
if (root == self.datadir):
self.images.append(file)
else:
self.images.append(root[len(self.datadir)+1:] + "/" + file)
if self.images == []:
print "error: no images found."
# loads images from database file
# (does not consider path information of the db file yet)
def loadImagesFromDB(self):
print "loading images from dbfile."
f = open(self.dbfile, "r")
lines = f.readlines()
f.close()
for line in lines:
if strip(split(line)[0]) == "file":
self.images.append(strip(split(line)[1]))
#print "file:", strip(split(line)[1])
if self.images == []:
print "error: no images found."
# loads features from the config file
def loadFeatures(self):
print "loading features file."
try:
f = open(self.config, "r")
except IOError:
print "warning: no specific config file found. I try the one in the tooldir."
f = open(self.progdir + "/features.conf")
lines = f.readlines()
f.close()
for line in lines:
if line[0] == "#":
#print "comment: ", line
pass
elif len(strip(line)) < 1:
pass
else:
key = strip(split(line)[0])
#print "key", key
feature = strip(line)[len(key)+1:]
#print "feature", feature
self.features[key] = feature
# loads an existing purefiles list
def loadImagesFromPurefiles(self):
print "loading images from purefiles."
self.images = []
file = open(self.purefiles, 'r')
lines = file.readlines()
file.close()
num = len(lines)
print "info: number of images =", num
for line in lines:
self.images.append(strip(line))
# loads images from commandline
def loadImagesFromCommandline(self, argv):
self.images = []
print "loading images from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-limages":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
self.images.append(command)
if self.images == []:
print "error: no images found"
def loadSelectedFeaturesFromCommandline(self, argv):
self.selectedFeatures = []
print "loading selected features from commandline."
catch = False
for command in argv:
#print "command:", command
if command == "-selectf":
#print "image tag found."
catch = True
elif command[0] == "-":
#print "end tag found."
catch = False
else:
if catch:
#print "added", command
self.selectedFeatures.append(command)
if self.selectedFeatures == []:
print "error: no features found"
# checking
# check if image file has a correct extension
def checkExtension(self, filename):
extension = split(filename, ".")[-1]
if lower(extension) in self.extensionlist:
return True
return False
def pureFileExists(self):
try:
f = open(self.purefiles, "r")
except IOError:
return False
else:
return True
# writing
# writes the purefilesfile
def writePurefiles(self):
print "writing purefiles file."
if self.images == []:
print "error: cannot write a purefiles file, because image list is empty."
else:
f = open(self.purefiles, "w")
for image in self.images:
f.write(image)
f.write("\n")
f.close()
# creates the database file
def writeDBFile(self):
# possible feature: dbfiles with feature directories?
# possible feature: non recursive traversion
print "writing dbfile."
f = open(self.dbfile, "w")
f.write("FIRE_filelist\n")
f.write("classes no\n")
f.write("descriptions no\n")
f.write("featuredirectories no\n")
f.write("path this\n")
for file in self.images:
f.write("file ")
f.write(file)
f.write("\n")
f.close()
# selection
# selects features
def selectFeatures(self):
if self.selectedFeatures == []:
print "error: there are no features to select"
print "hint: maybe wrong spelling?"
temp = {}
for key in self.features.keys():
#print "key", key
if key in self.selectedFeatures:
temp[key] = self.features[key]
self.features = temp
# extraction
# extracts all features for all images
def extractFeatures(self):
for key in self.features.keys():
self.extractFeature(key)
# extracts one feature for all images
def extractFeature(self, feature):
if self.images != []:
commandline = self.firedir + self.features[feature] + " --images"
for image in self.images:
commandline = commandline + " " + image
else:
if not self.pureFileExists():
print "error: purefiles file not existing"
print "hint: create with -np a new file or use -limage, -ldb or -ld to load images"
return False
else:
#self.printInfo()
#print "self.features[feature]", self.features[feature]
#self.printFeatures()
commandline = self.firedir + self.features[feature] + " --filelist " + self.purefiles
print "commandline:", commandline
if self.queue == "day":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 8:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
elif self.queue == "week":
queueline = "qsubmit -n " + self.name + "." + feature + " -m 0.4 -t 24:00:00 -bash "
print "job", queueline + commandline
os.system(queueline + commandline)
else:
os.system(commandline)
if __name__ == "__main__":
if "-h" in sys.argv:
usage(sys.argv)
else:
e = Extractor()
# adjusting file locations
if "-p" in sys.argv:
e.setPurefiles(getStringAfter("-p", sys.argv, None))
if "-db" in sys.argv:
e.setDBFile(getStringAfter("-db", sys.argv, None))
if "-d" in sys.argv:
e.setDatadir(getStringAfter("-d", sys.argv, None))
if "-c" in sys.argv:
e.setConfigfile(getStringAfter("-c", sys.argv, None))
if "-f" in sys.argv:
e.setFiredir(getStringAfter("-f", sys.argv, None))
if "-q" in sys.argv:
e.setQueue(getStringAfter("-q", sys.argv, None))
if "-n" in sys.argv:
e.setName(getStringAfter("-n", sys.argv, None))
## maybe later
# if "-el" in sys.argv
# e.setExtensionlist(...)
if "-v" in sys.argv:
e.printInfo()
if "-np" in sys.argv:
e.loadImagesFromDir()
e.writePurefiles()
elif "-npfromdb" in sys.argv:
e.loadImagesFromDB()
e.writePurefiles()
elif "-ndb" in sys.argv:
e.loadImagesFromDir()
e.writeDBFile()
elif "-ndbfromp" in sys.argv:
e.loadImagesFromPurefiles()
e.writeDBFile()
elif "-showf" in sys.argv:
e.loadFeatures()
e.printFeatures()
elif "-e" in sys.argv:
# loading the images (if no purefiles file is used)
if "-limages" in sys.argv:
e.loadImagesFromCommandline(sys.argv)
if "-v" in sys.argv:
e.printImages()
elif "-ldb" in sys.argv:
e.loadImagesFromDBFile()
elif "-ld" in sys.argv:
e.loadImagesFromDatadir()
else:
pass
e.loadFeatures()
if "-selectf" in sys.argv:
e.loadSelectedFeaturesFromCommandline(sys.argv)
e.selectFeatures()
if "-v" in sys.argv:
print "selected features:"
e.printFeatures()
e.extractFeatures()
else:
print "wrong usage. use -h for help."
| Python |
#!/usr/bin/env python
import sys,gzip
class Feature:
def __init__(self):
self.posx=0
self.posy=0
self.scale=0
self.vec=[]
class LocalFeatures:
def __init__(self):
self.winsize=0
self.dim=0
self.subsampling=0
self.padding=0
self.numberOfFeatures=0
self.varthreshold=0
self.zsize=0
self.filename=""
self.imagesize=(0,0)
self.features=[]
self.selffilename=""
def __getitem__(self,i):
return self.features[i]
def load(self,filename):
f=gzip.GzipFile(filename)
self.selffilename=filename
for l in f:
toks=l.split()
if toks[0]=="FIRE_localfeatures":
fine=True
elif toks[0]=="winsize":
self.winsize=int(toks[1])
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="subsampling":
self.subsampling=int(toks[1])
elif toks[0]=="padding":
self.padding=int(toks[1])
elif toks[0]=="numberOfFeatures":
self.numberOfFeatures=int(toks[1])
elif toks[0]=="varthreshold":
self.varthreshold=float(toks[1])
elif toks[0]=="zsize":
self.zsize=int(toks[1])
elif toks[0]=="filename":
self.filename=toks[1]
elif toks[0]=="imagesize":
self.imagesize=(int(toks[1]),int(toks[2]))
elif toks[0]=="features":
if self.numberOfFeatures!=int(toks[1]):
print "Weird... inconsistent number of Features and features expected in the file"
elif toks[0]=="feature":
f=Feature()
toks.pop(0)
f.posx=float(toks.pop(0))
f.posy=float(toks.pop(0))
f.scale=float(toks.pop(0))
f.vec=map(lambda x: float(x), toks)
self.features+=[f]
else:
print "'%s' is an unknown keyword... ignoring and continuiing nonetheless."%(toks[0])
def save(self,filename):
f=gzip.GzipFile(filename,"w")
f.write("FIRE_localfeatures\n")
f.write("winsize %d\ndim %d\nsubsampling %d\npadding %d\nnumberOfFeatures %d\nvarthreshold %d\nzsize %d\nfilename %s\nimagesize %d %d\nfeatures %d\n" % (self.winsize,self.dim,self.subsampling,self.padding,self.numberOfFeatures,self.varthreshold,self.zsize,self.filename,self.imagesize[0],self.imagesize[1],len(self.features)))
for lf in self.features:
f.write("feature %d %d %d " % (lf.posx,lf.posy,lf.scale))
for v in lf.vec:
f.write(str(v)+" ")
f.write("\n")
f.close()
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class HistogramFeature:
def __init__(self):
self.steps=[]
self.bins=[]
self.min=[]
self.max=[]
self.stepsize=[]
self.dim=0
self.counter=0
self.filename=""
self.data=[]
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_histogram"
print >>f,"# created with histogramfeature.py"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps"," ".join(map(lambda x: str(x),self.steps))
print >>f,"min"," ".join(map(lambda x: str(x), self.min))
print >>f,"max"," ".join(map(lambda x: str(x), self.max))
print >>f,"data"," ".join(map(lambda x: str(x), self.data))
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_histogram":
fine=True
elif toks[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="data":
self.bins=map(lambda x: int(x),toks[1:])
try:
self.data=map(lambda x: float(x)/float(self.counter), self.bins)
except:
print "Error reading file:", filename,"probably division by zero, continuing anyway."
else:
print "Unparsed line: '"+line+"'."
| Python |
#!/usr/bin/python
import sys
#this script merges fire distance files.
#
# as command line parameters, arbitrary many distance files are given and
# the merged distance file is printed.
infiles=[]
line=[]
for i in sys.argv[1:]:
infiles.append(open(i,"r"))
line.append("")
line[0]=infiles[0].readline()
count=0
while(line[0]):
for i in range(1,len(infiles)):
line[i]=infiles[i].readline()
if line[0].startswith("nofdistances"):
nDist=0
nImg=0
for l in line:
toks=l.split()
nDist+=int(toks[2])
nImg=int(toks[1])
print "nofdistances",nImg,nDist
elif not line[0].startswith("#"):
print count,
for l in line:
toks=l.split()
for i in range(1,len(toks)):
print float(toks[i]),
print
count+=1
line[0]=infiles[0].readline()
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
querylist=filelist.FileList()
filelist=filelist.FileList()
filelist.load(sys.argv[1])
querylist.load(sys.argv[2])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in querylist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
if len(sys.argv) != 4:
print """USAGE:
querylistwqueryexpansion.py <querylist> <server> <port>
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
files=ql.readlines()
ql.close()
s.sendcmd("setresults 10")
s.sendcmd("setextensions 2")
for file in files:
file=re.sub('\n','',file)
cmd="retrieveandsaveranks 1030 "+file+".ranks "+file
s.sendcmd(cmd)
res=s.getline()
print res
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback,numarray
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-f":
database=sys.argv.pop(0)
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-l1o":
l1o=True
elif argv=="-x":
quit=True
else:
print "Unknown option:",argv
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
if database=="" or querybase=="" and not l1o:
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
sys.stdout.flush()
f=filelist.FileList()
f.load(database)
q=filelist.FileList()
if not l1o:
q.load(querybase)
else:
q=f
if not f.classes:
print "Need classes in database file"
sys.exit(10)
if not l1o and not q.classes:
print "Need classes in querybase file"
sys.exit(10)
s=firesocket.FIRESocket()
s.connect(server,port)
try:
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
keyword=status.pop(0)
dbSize=int(status.pop(0))
if dbSize!=len(f.files):
print "database in retriever and database loaded are of different size:", dbSize,"!=",len(f.files)
s.sendcmd("bye")
time.sleep(1)
sys.exit(10)
print "database has",dbSize,"images."
if not l1o:
s.sendcmd("setresults 1"); res=s.getline()
else:
s.sendcmd("setresults 2"); res=s.getline()
classified=0; errors=0; correct=0
no_classes=0
for qf in q.files:
if qf[1]>no_classes:
no_classes=qf[1]
for df in f.files:
if df[1]>no_classes:
no_classes=df[1]
no_classes+=1
print no_classes,"classes."
confusionmatrix=numarray.zeros((no_classes,no_classes),numarray.Int)
counter=0
for qf in q.files:
counter+=1
qname=qf[0]
qcls=qf[1]
s.sendcmd("retrieve "+qname)
res=s.getline()
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
if not l1o:
if len(results)!=2:
print "expected 2 tokens, got",len(results)
sys.exit(10)
else:
if len(results)!=4:
print "expected 4 tokens, got",len(results)
sys.exit(10)
results.pop(0)
results.pop(0)
returnedimage=results.pop(0)
score=float(results.pop(0))
for df in f.files:
if not l1o:
if df[0]==returnedimage:
dcls=df[1]
break
else:
if df[0]==returnedimage and df[0]!=qname:
dcls=df[1]
break
classified+=1
if dcls==qcls:
correct+=1
else:
errors+=1
confusionmatrix[dcls][qcls]+=1
print "Query "+str(counter)+"/"+str(len(q.files))+":",qname,"("+str(qcls)+") NN:",returnedimage,"("+str(dcls)+")","ER:",float(errors)/float(classified)*100
sys.stdout.flush()
print "RESULT: ER:",float(errors)/float(classified)*100
print confusionmatrix
time.sleep(1)
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
useX="-x" in sys.argv
verbose="-v" in sys.argv
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
fileprefix=""
if len(sys.argv)>4:
if sys.argv[4]!="-q" and sys.argv[4]!="-x" and sys.argv[4]!="-v":
fileprefix=sys.argv[4]
lines=ql.readlines()
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
ql.close()
classified=0
###query images and evaluate results
for line in lines:
line=re.sub('\n','',line)
relevant=re.split(" ",line)
query=fileprefix+relevant[0] # this is the name of the query image
relevant=relevant[1:] # this are the names of the images relevant to the query image
cmd="retrieve +"+query
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimage=re.sub(fileprefix,"",returnedimage)
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
if verbose:
print "results:",returnedimages[0],"->0",
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
param["useX"] = useX
param["verbose"] = verbose
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
useX = param["useX"]
verbose = param["verbose"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
if useX:
g=Gnuplot.Gnuplot()
g.title('PR graph')
g('set data style linespoints')
g('set xrange [0.0:1.0]')
g('set yrange [0.0:1.0]')
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
if verbose:
print returnedimages[Ri],"->",Ri,
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
if useX:
g.plot(Gnuplot.Data(PRgraph,title="PRgraph current query"),Gnuplot.Data(tmpAPRgraph,title="PRgraph average"))
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [<fileprefix>] [-q] [-x]
-q: quit the server after finished eval
-x: show PR graph after each query on X display
-v: verbose
<fileprefix> should not be necessary
"""
else:
param = {}
retrieveData(param)
calculate(param)
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
host=sys.argv[1]
if len(sys.argv) >=3:
port=int(sys.argv[2])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
s.sendcmd("listfiles")
fileline=s.getline()
files=string.split(fileline," ")
s.sendcmd("setresults 2")
s.getline()
classified=0
correct=0
error=0
for file in files:
if len(file)>0:
cmd="retrieve +"+file
s.sendcmd(cmd)
res=s.getline()
parts=string.split(res," ")
result=parts[2]
print res
#get classes
s.sendcmd("class "+file)
qclass=s.getline()
s.sendcmd("class "+result)
rclass=s.getline()
qc=int(qclass)
rc=int(rclass)
if(rc==qc):
correct=correct+1
else:
error=error+1
classified=classified+1
er=float(error)/float(classified)*100.0
print file+" wrong="+str(error)+" correct="+str(correct)+" classified="+str(classified)+" ER="+str(er)
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
filelist=filelist.FileList()
filelist.load(sys.argv[1])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in filelist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class SparseHistogramFeature:
def __init__(self):
self.bins={}
self.data={}
self.steps=[]
self.stepSize=[]
self.min=[]
self.max=[]
self.dim=0
self.numberofBins=0
self.counter=0
self.filename=""
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_sparse_histogram"
print >>f,"# sparse histogram file for FireV2"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps",reduce(lambda x,y: str(x)+" "+str(y),self.steps)
print >>f,"min",reduce(lambda x,y: str(x)+" "+str(y),self.min)
print >>f,"max",reduce(lambda x,y: str(x)+" "+str(y),self.max)
print >>f,"bins",len(self.bins)
for i in self.bins:
print >>f,"data",i,self.bins[i]
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_sparse_histogram":
fine=True
elif line[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="bins":
self.numberofBins=int(toks[1])
elif toks[0]=="data":
self.bins[toks[1]]=int(toks[2])
self.data[toks[1]]=float(toks[2])/float(self.counter)
else:
print "Unparsed line: '"+line+"'."
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
#-----------------------------------------------------------------
s=firesocket.FIRESocket()
try:
if len(sys.argv) < 5:
print """USAGE:
makedistfiles.py <querylist> <server> <port> <suffix> [-q]
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
suffix=sys.argv[4]
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
lines=ql.readlines()
lines=map(lambda line:
re.sub("\n","",line), lines)
ql.close()
i=0
for line in lines:
print i,"/",len(lines)
i+=1
tokens=re.split(" ",line)
file=tokens[0]
cmd="savedistances "+file+" "+file+"."+suffix
print "SENDING: ",cmd
s.sendcmd(cmd)
res=s.getline()
print res
sys.stdout.flush()
sys.stdout.flush()
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
classified=0
###query images and evaluate results
for line in ql:
[query,relevant]=line.split('#')
relevant=relevant.strip()
query=query.strip()
cmd="retrieve "+query
relevant=relevant.split()
print "SEND:'"+cmd+"'."
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [-q]
-q: quit the server after finished eval
"""
else:
param = {}
# do the queries
retrieveData(param)
# querying finished, now evaluated the gathered data
calculate(param)
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback
#sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
start=-1
stop=-1
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-x":
quit=True
elif argv=="-S":
suffix=sys.argv.pop(0)
elif argv=="-start":
start=int(sys.argv.pop(0))
elif argv=="-stop":
stop=int(sys.argv.pop(0))
else:
print "Unknown option:",argv
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist, -1 for full database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"querybase=",querybase
if querybase=="":
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist
-x exit after having finished
"""
sys.exit(10)
sys.stdout.flush()
q=filelist.FileList()
q.load(querybase)
s=firesocket.FIRESocket()
s.connect(server,port)
if start==-1:
start=0
if stop==-1:
stop=len(q.files)
try:
for qf in q.files[start:stop]:
cmd="savedistances "+qf[0]+" "+qf[0]+"."+suffix
s.sendcmd(cmd)
res=s.getline()
print res
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/python
import sys,re
import filelist
filelist=filelist.FileList()
filelist.load(sys.argv[1])
for i in filelist:
cls=i[1]
fname=i[0]
print fname,
for j in filelist:
if j[1]==cls and j[0]!=fname:
print j[0],
print
#print filelist
| Python |
#!/usr/bin/python
import sys,re,filelist
if len(sys.argv)<3:
print """USAGE filelistquery2querylist.py <querylist> <dblist>"""
sys.exit(5)
qlist=filelist.FileList()
qlist.load(sys.argv[1])
dblist=filelist.FileList()
dblist.load(sys.argv[2])
for i in qlist:
cls=i[1]
fname=i[0]
print fname,
for j in dblist:
if j[1]==cls:
print j[0],
print
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
useX="-x" in sys.argv
verbose="-v" in sys.argv
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
fileprefix=""
if len(sys.argv)>4:
if sys.argv[4]!="-q" and sys.argv[4]!="-x" and sys.argv[4]!="-v":
fileprefix=sys.argv[4]
lines=ql.readlines()
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
ql.close()
classified=0
###query images and evaluate results
for line in lines:
line=re.sub('\n','',line)
relevant=re.split(" ",line)
query=fileprefix+relevant[0] # this is the name of the query image
relevant=relevant[1:] # this are the names of the images relevant to the query image
cmd="retrieve +"+query
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimage=re.sub(fileprefix,"",returnedimage)
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
if verbose:
print "results:",returnedimages[0],"->0",
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
param["useX"] = useX
param["verbose"] = verbose
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
useX = param["useX"]
verbose = param["verbose"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
if useX:
g=Gnuplot.Gnuplot()
g.title('PR graph')
g('set data style linespoints')
g('set xrange [0.0:1.0]')
g('set yrange [0.0:1.0]')
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
if verbose:
print returnedimages[Ri],"->",Ri,
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
if useX:
g.plot(Gnuplot.Data(PRgraph,title="PRgraph current query"),Gnuplot.Data(tmpAPRgraph,title="PRgraph average"))
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [<fileprefix>] [-q] [-x]
-q: quit the server after finished eval
-x: show PR graph after each query on X display
-v: verbose
<fileprefix> should not be necessary
"""
else:
param = {}
retrieveData(param)
calculate(param)
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
host=sys.argv[1]
if len(sys.argv) >=3:
port=int(sys.argv[2])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
s.sendcmd("listfiles")
fileline=s.getline()
files=string.split(fileline," ")
s.sendcmd("setresults 2")
s.getline()
classified=0
correct=0
error=0
for file in files:
if len(file)>0:
cmd="retrieve +"+file
s.sendcmd(cmd)
res=s.getline()
parts=string.split(res," ")
result=parts[2]
print res
#get classes
s.sendcmd("class "+file)
qclass=s.getline()
s.sendcmd("class "+result)
rclass=s.getline()
qc=int(qclass)
rc=int(rclass)
if(rc==qc):
correct=correct+1
else:
error=error+1
classified=classified+1
er=float(error)/float(classified)*100.0
print file+" wrong="+str(error)+" correct="+str(correct)+" classified="+str(classified)+" ER="+str(er)
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time
class mysocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk != '\n':
msg=msg+chunk
return msg
#-----------------------------------------------------------------
s=mysocket()
try:
if len(sys.argv) != 4:
print """USAGE:
querylistwqueryexpansion.py <querylist> <server> <port>
"""
else:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
files=ql.readlines()
ql.close()
s.sendcmd("setresults 10")
s.sendcmd("setextensions 2")
for file in files:
file=re.sub('\n','',file)
cmd="retrieveandsaveranks 1030 "+file+".ranks "+file
s.sendcmd(cmd)
res=s.getline()
print res
s.sendcmd("bye")
except Exception, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
| Python |
#!/usr/bin/python
import sys,re,filelist
if len(sys.argv)<3:
print """USAGE filelistquery2querylist.py <querylist> <dblist>"""
sys.exit(5)
qlist=filelist.FileList()
qlist.load(sys.argv[1])
dblist=filelist.FileList()
dblist.load(sys.argv[2])
for i in qlist:
cls=i[1]
fname=i[0]
print fname,
for j in dblist:
if j[1]==cls:
print j[0],
print
| Python |
import socket, os, sys,re
# ----------------------------------------------------------------------
# socket implementation for comunication with the server system
# running on any computer which is reachable via network sockets
# here we need mainly the functions
# * connect
# * sendcmd
# * getline
# * close
# for the communication
# ----------------------------------------------------------------------
class FIRESocket:
'''Socket stolen somewhere from the net'''
def __init__(self, sock=None):
if sock is None:
self.sock = socket.socket(
socket.AF_INET, socket.SOCK_STREAM)
else:
self.sock = sock
#self.sock.settimeout(10.0)
def connect(self,host, port):
self.sock.connect((host, port))
def send(self,msg):
#print "PUT:",msg; sys.stdout.flush()
totalsent = 0
MSGLEN=len(msg)
while totalsent < MSGLEN:
sent = self.sock.send(msg[totalsent:])
if sent == 0:
raise RuntimeError, "socket connection broken"
totalsent = totalsent + sent
def sendcmd(self,cmd):
cmd = cmd + "\r\n"
self.send(cmd)
def getResult(self):
result=""
line=""
while line[0:3] != "end":
line=self.getline()
if line[0:3] != "end":
result=result+line
#print "GOTR:",result; sys.stdout.flush()
return result
def receive(self):
msg = ''
while len(msg) < MSGLEN:
chunk = self.sock.recv(MSGLEN-len(msg))
if chunk == '':
raise RuntimeError, "socket connection broken"
msg = msg + chunk
return msg
def getline(self):
msg=''
chunk='a'
while chunk != '\n':
chunk=self.sock.recv(1)
if chunk == '':
raise RuntimeError, "socket connection broken"
if chunk!= '\r' and chunk != '\n':
msg=msg+chunk
msg=re.sub("[ ]$","",msg)
#print "GOTL:",msg; sys.stdout.flush()
return msg
def flush(self):
chunk=self.sock.recv(5000)
return
| Python |
#!/usr/bin/python
import sys,re
import filelist,porterstemmer
querylist=filelist.FileList()
filelist=filelist.FileList()
filelist.load(sys.argv[1])
querylist.load(sys.argv[2])
stemmer=porterstemmer.PorterStemmer()
if not filelist.descriptions:
print "no descriptions, this program is not appropriate"
sys.exit(10)
for i in querylist:
cls=i[1]
desc=[]
#print "Before stemming: ",i[2]
for w in i[2]:
w=stemmer.stem(w,0,len(w)-1)
desc+=[w]
#print "After stemming:",desc
rels={}
for j in filelist:
desc2=[]
for w in j[2]:
w=stemmer.stem(w,0,len(w)-1)
if w in desc:
rels[j[0]]=1
print i[0],
for f in rels:
print f,
print
#print filelist
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback
#sys.path.append("/u/deselaers/work/src/fire/python")
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
start=-1
stop=-1
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-x":
quit=True
elif argv=="-S":
suffix=sys.argv.pop(0)
elif argv=="-start":
start=int(sys.argv.pop(0))
elif argv=="-stop":
stop=int(sys.argv.pop(0))
else:
print "Unknown option:",argv
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist, -1 for full database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"querybase=",querybase
if querybase=="":
print """
USAGE: makedistfilesforfilelist.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-q database for query
-S suffix
-start <n>, -stop <n> to process a subset of the querylist
-x exit after having finished
"""
sys.exit(10)
sys.stdout.flush()
q=filelist.FileList()
q.load(querybase)
s=firesocket.FIRESocket()
s.connect(server,port)
if start==-1:
start=0
if stop==-1:
stop=len(q.files)
try:
for qf in q.files[start:stop]:
cmd="savedistances "+qf[0]+" "+qf[0]+"."+suffix
s.sendcmd(cmd)
res=s.getline()
print res
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/env python
import gzip
# class for FIRE histogramfeatures
class HistogramFeature:
def __init__(self):
self.steps=[]
self.bins=[]
self.min=[]
self.max=[]
self.stepsize=[]
self.dim=0
self.counter=0
self.filename=""
self.data=[]
def save(self,filename):
f=gzip.GzipFile(filename,"w")
print >>f,"FIRE_histogram"
print >>f,"# created with histogramfeature.py"
print >>f,"dim",self.dim
print >>f,"counter",self.counter
print >>f,"steps"," ".join(map(lambda x: str(x),self.steps))
print >>f,"min"," ".join(map(lambda x: str(x), self.min))
print >>f,"max"," ".join(map(lambda x: str(x), self.max))
print >>f,"data"," ".join(map(lambda x: str(x), self.data))
f.close()
def load(self,filename):
f=gzip.GzipFile(filename)
self.filename=filename
for line in f:
toks=line.split()
if toks[0]=="FIRE_histogram":
fine=True
elif toks[0]=="#":
fine=True
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="counter":
self.counter=int(toks[1])
elif toks[0]=="steps":
self.steps=map(lambda x: int(x), toks[1:])
elif toks[0]=="min":
self.min=map(lambda x: float(x),toks[1:])
elif toks[0]=="max":
self.max=map(lambda x: float(x),toks[1:])
elif toks[0]=="data":
self.bins=map(lambda x: int(x),toks[1:])
try:
self.data=map(lambda x: float(x)/float(self.counter), self.bins)
except:
print "Error reading file:", filename,"probably division by zero, continuing anyway."
else:
print "Unparsed line: '"+line+"'."
| Python |
#!/usr/bin/python
import sys, socket, re, os, string, time, traceback
#sys.path.append("../python")
import firesocket
def retrieveData(param):
data = {}
s=firesocket.FIRESocket()
try:
ql=open(sys.argv[1],"r")
host=sys.argv[2]
port=int(sys.argv[3])
try:
time.sleep(1) #don't kill the server by exiting tooo fast
s.connect(host,port)
except:
print "Connecting to server "+host+" on port "+str(port)+" failed."
# get number of files in retriever
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
status.reverse()
keyword=status.pop()
dbSize=int(status.pop())
print "database has",dbSize,"images."
s.sendcmd("setresults "+str(dbSize)); res=s.getline();
classified=0
###query images and evaluate results
for line in ql:
[query,relevant]=line.split('#')
relevant=relevant.strip()
query=query.strip()
cmd="retrieve "+query
relevant=relevant.split()
print "SEND:'"+cmd+"'."
s.sendcmd(cmd)
res=s.getline() # get the results from the retriever
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
results.reverse()
returnedimages=[]
while len(results)>0:
returnedimage=results.pop()
score=results.pop()
returnedimages=returnedimages+[returnedimage]
# --------------------------------------------------------------
# performance evaluation
# --------------------------------------------------------------
classified=classified+1
if returnedimages[0] != query:
# if the query image is not the first returned, we assume that it is
# not part of the database, and thus we have to start with the first
# returned image for query evaluation. This is a workaround, usually we start
# at position 1, to keep this valid we insert a new one on position 0
# this makes the old position 0 to position 1 and everything hereafter
# stays valid
returnedimages.insert(0,"DUMMYIMAGE")
key = classified
data[key] = [query, relevant, returnedimages]
#print "key", key, "data", data[key]
print "query:", query, "image no.:", classified, "of", dbSize, "relevant:", relevant[:5], "returnedimages:", returnedimages[:5]
sys.stdout.flush()
s.sendcmd("info")
res=s.getline()
print "SETTINGS:",res
sys.stdout.flush()
time.sleep(1)
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if "-q" in sys.argv:
s.sendcmd("quit")
else:
s.sendcmd("bye")
traceback.print_exc(file=sys.stdout)
# t=sys.last_traceback
# traceback.print_tb(t)
print e
time.sleep(1) #don't kill the server by exiting tooo fast
param["data"] = data
param["dbSize"] = dbSize
def calculate(param):
data = param["data"]
dbSize = param["dbSize"]
# init variables for quantitative comparison
classified=0; correct=0; error=0; AMP=0.0; ARE=0.0; ARP05=0.0;APRP=0.0; MAP=0.0
AP1=0; AP20=0; AP50=0; APNrel=0; AR100=0; ARank1=0; ARanktilde=0;
APRarea=0.0; APRgraph=[(0.0,0),(0.1,0),(0.2,0),(0.3,0),(0.4,0),(0.5,0),(0.6,0),(0.7,0),(0.8,0),(0.9,0),(1.0,0)]
data.keys().sort()
for j in data.keys():
#print "j", j
relevant = data[j][1]
returnedimages = data[j][2]
# calculate performance measures: Error rate
classified=classified+1
if returnedimages[1] in relevant:
correct=correct+1
else:
error=error+1
# calculate performance measures: ARE, AMP, Rank1, Ranktilde
MPQ=0.0; REQ=0.0; MPQ_div=0.0; REQ_div=0.0; i=1; RP05=-1.0; PRP=-1.0; AvPr=0.0;
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
AvPr+=(float(i)/float(Ri))
MPQ+=float(dbSize-Ri)/float(dbSize-(i))
REQ+=float(Ri); REQ_div+=float(i)
#print "i:", i
if i==1:
Rank1=Ri
i+=1
print
Ranktilde=float(REQ-((len(relevant)*(len(relevant)-1))/2))/float(dbSize*len(relevant))
MPQ*=float(100)/float(len(relevant))
if REQ_div!=0:
REQ/=REQ_div
AMP+=MPQ
ARE+=REQ
AvPr*=float(100)/float(len(relevant))
# calculate performance measures: P,R, PR-graph
Nrelret=0;
PR=[]
for Ri in range(1,len(returnedimages)):
if returnedimages[Ri] in relevant:
Nrelret+=1
P=float(Nrelret)/float(Ri)
R=float(Nrelret)/float(len(relevant))
PR+=[(P,R)]
# smoothing
Ris=range(1,len(returnedimages)-1)
Ris.reverse()
maxP=0
for Ri in Ris:
if PR[Ri][0]>maxP:
maxP=PR[Ri][0]
else:
PR[Ri]=(maxP,PR[Ri][1])
if PR[Ri][0]>0.5 and RP05<0:
RP05=PR[Ri][1]
if PR[Ri][0]>PR[Ri][1] and PRP<0:
PRP=PR[Ri][0]
if RP05<0:
RP05=0
if PRP<0:
PRP=0
# pick the eleven values R=0.0,0.1,...,1.0 from the PR values for the PRgraph
i=0
PRgraph=[]
for r in range(11):
R=float(r)/float(10)
while i<len(PR) and PR[i][1] < R:
i+=1
if i>len(PR)-1:
i=len(PR)-1
P=PR[i][0]
PRgraph+=[(R,P)]
if PRgraph[0][1] < PRgraph[1][1]:
PRgraph[0]=(PRgraph[0][0],PRgraph[1][1])
PRarea=0.0
for i in PRgraph[1:-1]:
PRarea+=i[1]
PRarea+=0.5*(PRgraph[0][1]+PRgraph[-1][1])
PRarea*=0.1
P1=PR[0][0]
P20=0
if len(PR)>20:
P20=PR[20][0]
P50=0
if len(PR)>50:
P50=PR[50][0]
PNrel=PR[len(relevant)][0]
if len(PR)>100:
R100=PR[100][1]
else:
R100=1
AP1 +=P1
AP20 +=P20
AP50 +=P50
APNrel+=PNrel
AR100 +=R100
ARank1+=Rank1
ARanktilde+=Ranktilde
APRarea+=PRarea
ARP05+=RP05
APRP+=PRP
MAP+=AvPr
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]+PRgraph[i][1])
tmpAPRgraph=[]
for i in range(len(APRgraph)):
tmpAPRgraph+=[(APRgraph[i][0],APRgraph[i][1]/float(classified))]
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
#print "PRgraph: ",PRgraph
print "classfied: %i" % (classified)
for i in range(len(APRgraph)):
APRgraph[i]=(APRgraph[i][0],APRgraph[i][1]/float(classified))
print "RESULT:",
print " AMP: %1.2f" % (float(AMP)/float(classified)),
print "ARE: %1.2f" % (float(ARE)/float(classified)),
print "MAP: %1.2f" % (float(MAP)/float(classified)),
print "ER: %1.2f:" % (float(error)*100/float(classified)),
print "P1: %1.2f" % (float(AP1)/float(classified)),
print "P20: %1.2f" % ( float(AP20)/float(classified)),
print "P50: %1.2f" % ( float(AP50)/float(classified)),
print "PNrel: %1.2f" % ( float(APNrel)/float(classified)),
print "R100: %1.2f" % (float(AR100)/float(classified)),
print "Rank1: %1.2f" % (float(ARank1)/float(classified)),
print "Ranktilde: %1.2f" % (float(ARanktilde)/float(classified)),
print "PRarea: %1.2f" % (float(APRarea)/float(classified)),
print "R(P=0.5): %1.2f" % (float(ARP05)/float(classified)),
print "P(R=P): %1.2f" % (float(APRP)/float(classified)),
print "classfied: %i" % (classified)
print "PRgraph: ",APRgraph
if __name__ == "__main__":
if len(sys.argv) < 4:
print """USAGE:
querylist.py <querylist> <server> <port> [-q]
-q: quit the server after finished eval
"""
else:
param = {}
# do the queries
retrieveData(param)
# querying finished, now evaluated the gathered data
calculate(param)
| Python |
#!/usr/bin/python
# easy script for automatic batches with fire
# takes database files for queries
# calculates error rate only
import sys, socket, re, os, string, time, traceback,numarray
import firesocket,filelist
# forget own name in commandline options
sys.argv.pop(0)
server="localhost"
port=12960
database=""
querybase=""
l1o=False
quit=False
while len(sys.argv)>0:
argv=sys.argv.pop(0)
if argv=="-s":
server=sys.argv.pop(0)
elif argv=="-p":
port=int(sys.argv.pop(0))
elif argv=="-f":
database=sys.argv.pop(0)
elif argv=="-q":
querybase=sys.argv.pop(0)
elif argv=="-l1o":
l1o=True
elif argv=="-x":
quit=True
else:
print "Unknown option:",argv
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
if database=="" or querybase=="" and not l1o:
print """
USAGE: queryfilelistwcls.py <options>
-h show this help
-s server (default: localhost)
-p port (default: 12960)
-f database
-q database for query
-l1o make leaving one out with database
-x exit after having finished
"""
sys.exit(10)
print "SETTINGS: server=",server,"port=",port,"database=",database,"querybase=",querybase,"l1o=",l1o
sys.stdout.flush()
f=filelist.FileList()
f.load(database)
q=filelist.FileList()
if not l1o:
q.load(querybase)
else:
q=f
if not f.classes:
print "Need classes in database file"
sys.exit(10)
if not l1o and not q.classes:
print "Need classes in querybase file"
sys.exit(10)
s=firesocket.FIRESocket()
s.connect(server,port)
try:
s.sendcmd("info")
res=s.getline()
status=re.split(" ",res)
keyword=status.pop(0)
dbSize=int(status.pop(0))
if dbSize!=len(f.files):
print "database in retriever and database loaded are of different size:", dbSize,"!=",len(f.files)
s.sendcmd("bye")
time.sleep(1)
sys.exit(10)
print "database has",dbSize,"images."
if not l1o:
s.sendcmd("setresults 1"); res=s.getline()
else:
s.sendcmd("setresults 2"); res=s.getline()
classified=0; errors=0; correct=0
no_classes=0
for qf in q.files:
if qf[1]>no_classes:
no_classes=qf[1]
for df in f.files:
if df[1]>no_classes:
no_classes=df[1]
no_classes+=1
print no_classes,"classes."
confusionmatrix=numarray.zeros((no_classes,no_classes),numarray.Int)
counter=0
for qf in q.files:
counter+=1
qname=qf[0]
qcls=qf[1]
s.sendcmd("retrieve "+qname)
res=s.getline()
res=re.sub('[ ]*$','',res) # and parse the results
results=re.split(" ",res)
if not l1o:
if len(results)!=2:
print "expected 2 tokens, got",len(results)
sys.exit(10)
else:
if len(results)!=4:
print "expected 4 tokens, got",len(results)
sys.exit(10)
results.pop(0)
results.pop(0)
returnedimage=results.pop(0)
score=float(results.pop(0))
for df in f.files:
if not l1o:
if df[0]==returnedimage:
dcls=df[1]
break
else:
if df[0]==returnedimage and df[0]!=qname:
dcls=df[1]
break
classified+=1
if dcls==qcls:
correct+=1
else:
errors+=1
confusionmatrix[dcls][qcls]+=1
print "Query "+str(counter)+"/"+str(len(q.files))+":",qname,"("+str(qcls)+") NN:",returnedimage,"("+str(dcls)+")","ER:",float(errors)/float(classified)*100
sys.stdout.flush()
print "RESULT: ER:",float(errors)/float(classified)*100
print confusionmatrix
time.sleep(1)
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
time.sleep(1) #don't kill the server by exiting tooo fast
except KeyboardInterrupt, e:
s.sendcmd("bye")
print e
time.sleep(1) #don't kill the server by exiting tooo fast
except Exception, e:
if quit:
s.sendcmd("quit")
else:
s.sendcmd("bye")
print e
time.sleep(1)
| Python |
#!/usr/bin/env python
import sys,gzip
class Feature:
def __init__(self):
self.posx=0
self.posy=0
self.scale=0
self.vec=[]
class LocalFeatures:
def __init__(self):
self.winsize=0
self.dim=0
self.subsampling=0
self.padding=0
self.numberOfFeatures=0
self.varthreshold=0
self.zsize=0
self.filename=""
self.imagesize=(0,0)
self.features=[]
self.selffilename=""
def __getitem__(self,i):
return self.features[i]
def load(self,filename):
f=gzip.GzipFile(filename)
self.selffilename=filename
for l in f:
toks=l.split()
if toks[0]=="FIRE_localfeatures":
fine=True
elif toks[0]=="winsize":
self.winsize=int(toks[1])
elif toks[0]=="dim":
self.dim=int(toks[1])
elif toks[0]=="subsampling":
self.subsampling=int(toks[1])
elif toks[0]=="padding":
self.padding=int(toks[1])
elif toks[0]=="numberOfFeatures":
self.numberOfFeatures=int(toks[1])
elif toks[0]=="varthreshold":
self.varthreshold=float(toks[1])
elif toks[0]=="zsize":
self.zsize=int(toks[1])
elif toks[0]=="filename":
self.filename=toks[1]
elif toks[0]=="imagesize":
self.imagesize=(int(toks[1]),int(toks[2]))
elif toks[0]=="features":
if self.numberOfFeatures!=int(toks[1]):
print "Weird... inconsistent number of Features and features expected in the file"
elif toks[0]=="feature":
f=Feature()
toks.pop(0)
f.posx=float(toks.pop(0))
f.posy=float(toks.pop(0))
f.scale=float(toks.pop(0))
f.vec=map(lambda x: float(x), toks)
self.features+=[f]
else:
print "'%s' is an unknown keyword... ignoring and continuiing nonetheless."%(toks[0])
def save(self,filename):
f=gzip.GzipFile(filename,"w")
f.write("FIRE_localfeatures\n")
f.write("winsize %d\ndim %d\nsubsampling %d\npadding %d\nnumberOfFeatures %d\nvarthreshold %d\nzsize %d\nfilename %s\nimagesize %d %d\nfeatures %d\n" % (self.winsize,self.dim,self.subsampling,self.padding,self.numberOfFeatures,self.varthreshold,self.zsize,self.filename,self.imagesize[0],self.imagesize[1],len(self.features)))
for lf in self.features:
f.write("feature %d %d %d " % (lf.posx,lf.posy,lf.scale))
for v in lf.vec:
f.write(str(v)+" ")
f.write("\n")
f.close()
| Python |
#!/usr/bin/python
import sys
#this script merges fire distance files.
#
# as command line parameters, arbitrary many distance files are given and
# the merged distance file is printed.
infiles=[]
line=[]
for i in sys.argv[1:]:
infiles.append(open(i,"r"))
line.append("")
line[0]=infiles[0].readline()
count=0
while(line[0]):
for i in range(1,len(infiles)):
line[i]=infiles[i].readline()
if line[0].startswith("nofdistances"):
nDist=0
nImg=0
for l in line:
toks=l.split()
nDist+=int(toks[2])
nImg=int(toks[1])
print "nofdistances",nImg,nDist
elif not line[0].startswith("#"):
print count,
for l in line:
toks=l.split()
for i in range(1,len(toks)):
print float(toks[i]),
print
count+=1
line[0]=infiles[0].readline()
| Python |
#
# jython examples for jas.
from jas import Module
from jas import SubModule
# module example
r = Module( "Rat(u,v,l) L", cols=4 );
print "Module: " + str(r);
print;
G = r.gens();
print "gens() = ", [str(e) for e in G];
L = [ e.elem.val for e in G ]
print "gens() = ", [str(e) for e in L];
M = r.submodul( list=L );
print "M = ", M;
P = M.mset.getPolynomialList();
print "P = ", P.toScript();
print "M.isGB(): ", M.isGB();
| Python |
#
# jython examples for jas.
# $Id$
#
from java.lang import System
from java.lang import Integer
from jas import Ring, PolyRing
from jas import terminate
from jas import startLog
from jas import QQ, DD
# polynomial examples: ideal primary decomposition
#r = Ring( "Rat(x) L" );
#r = Ring( "Q(x) L" );
r = PolyRing(QQ(),"x,y,z",PolyRing.lex);
print "Ring: " + str(r);
print;
[one,x,y,z] = r.gens();
f1 = (x**2 - 5)**2;
f2 = (y**2 - x)**2;
f3 = z**2 - x ;
#print "f1 = ", f1;
print "f2 = ", f2;
print "f3 = ", f3;
print;
F = r.ideal( list=[f2,f3] );
print "F = ", F;
print;
startLog();
t = System.currentTimeMillis();
Q = F.primaryDecomp();
t = System.currentTimeMillis() - t;
print "Q = ", Q;
print;
print "primary decomp time =", t, "milliseconds";
print;
print "F = ", F;
print;
#startLog();
terminate();
| Python |
#
# jython examples for jas.
# $Id$
#
import sys;
from jas import Ring, PolyRing, Ideal, QQ, ZZ, GF, ZM, CC
from jas import startLog, terminate
# trinks 6/7 example
# QQ = rational numbers, ZZ = integers, CC = complex rational numbers, GF = finite field
#QQ = QQ(); ZZ = ZZ(); CC = CC();
#r = PolyRing( GF(19),"B,S,T,Z,P,W", PolyRing.lex);
#r = PolyRing( GF(1152921504606846883),"B,S,T,Z,P,W", PolyRing.lex); # 2^60-93
#r = PolyRing( GF(2**60-93),"B,S,T,Z,P,W", PolyRing.lex);
#r = PolyRing( CC(),"B,S,T,Z,P,W", PolyRing.lex);
#r = PolyRing( ZZ(),"B,S,T,Z,P,W", PolyRing.lex); # not for parallel
r = PolyRing( QQ(),"B,S,T,Z,P,W", PolyRing.lex);
print "Ring: " + str(r);
print;
# sage like: with generators for the polynomial ring
#[one,I,B,S,T,Z,P,W] = r.gens(); # is automaticaly included
#[one,B,S,T,Z,P,W] = r.gens(); # is automaticaly included
f1 = 45 * P + 35 * S - 165 * B - 36;
f2 = 35 * P + 40 * Z + 25 * T - 27 * S;
f3 = 15 * W + 25 * S * P + 30 * Z - 18 * T - 165 * B**2;
f4 = - 9 * W + 15 * T * P + 20 * S * Z;
f5 = P * W + 2 * T * Z - 11 * B**3;
f6 = 99 * W - 11 *B * S + 3 * B**2;
f7 = 10000 * B**2 + 6600 * B + 2673;
F = [ f1, f2, f3, f4, f5, f6, f7 ]; # smaller, faster
#F = [ f1, f2, f3, f4, f5, f6 ]; # bigger, needs more time
#print "F = ", [ str(f) for f in F ];
#print;
I = r.ideal( "", list=F );
print "Ideal: " + str(I);
print;
rg = I.GB();
print "seq Output:", rg;
print;
#startLog();
rg = I.parGB(2);
#print "par Output:", rg;
#print;
#sys.exit(); # if using ZZ coefficients
I.distClient(); # starts in background
rg = I.distGB(2);
#print "dist Output:", rg;
#print;
I.distClientStop(); # stops them
terminate();
| Python |
#
# jython examples for jas.
# $Id$
#
import sys;
from jas import PolyRing, QQ, RF
from jas import Ideal
from jas import startLog
from jas import terminate
# Montes JSC 2002, 33, 183-208, example 11.1
# integral function coefficients
r = PolyRing( PolyRing(QQ(),"c, b, a",PolyRing.lex), "z,y,x", PolyRing.lex );
print "Ring: " + str(r);
print;
[one,c,b,a,z,y,x] = r.gens();
print "gens: ", [ str(f) for f in r.gens() ];
print;
f1 = x + c * y + b * z + a;
f2 = c * x + y + a * z + b;
f3 = b * x + a * y + z + c;
F = [f1,f2,f3];
print "F: ", [ str(f) for f in F ];
print;
#startLog();
If = r.paramideal( "", list = F );
print "ParamIdeal: " + str(If);
print;
## G = If.GB();
## print "GB: " + str(G);
## print;
## sys.exit();
GS = If.CGBsystem();
GS = If.CGBsystem();
GS = If.CGBsystem();
print "CGBsystem: " + str(GS);
print;
bg = GS.isCGBsystem();
if bg:
print "isCGBsystem: true";
else:
print "isCGBsystem: false";
print;
terminate();
sys.exit();
CG = If.CGB();
print "CGB: " + str(CG);
print;
bg = CG.isCGB();
if bg:
print "isCGB: true";
else:
print "isCGB: false";
print;
terminate();
#------------------------------------------
#sys.exit();
| Python |
#
# jython examples for jas.
# $Id$
#
from java.lang import System
from java.lang import Integer
from jas import SolvableRing, SolvPolyRing, PolyRing
from jas import QQ, startLog, SRC, SRF
# Ore extension solvable polynomial example, Gomez-Torrecillas, 2003
pcz = PolyRing(QQ(),"x,y,z");
#is automatic: [one,x,y,z] = pcz.gens();
zrelations = [z, y, y * z + x
];
print "zrelations: = " + str([ str(f) for f in zrelations ]);
print;
pz = SolvPolyRing(QQ(), "x,y,z", PolyRing.lex, zrelations);
print "SolvPolyRing: " + str(pz);
print;
pzq = SRF(pz);
print "SolvableQuotientRing: " + str(pzq.ring.toScript()); # + ", assoz: " + str(pzq.ring.isAssociative());
#print "gens =" + str([ str(f) for f in pzq.ring.generators() ]);
print;
pct = PolyRing(pzq,"t");
#is automatic: [one,x,y,z,t] = p.gens();
trelations = [t, y, y * t + y,
t, z, z * t - z
];
print "relations: = " + str([ str(f) for f in trelations ]);
print;
pt = SolvPolyRing(pzq, "t", PolyRing.lex, trelations);
print "SolvPolyRing: " + str(pt);
print;
print "gens =", [ str(f) for f in pt.gens() ];
#print "t = " + str(t);
#print "z = " + str(z);
zi = 1 / z;
yi = 1 / y;
xi = 1 / x;
#startLog();
a = ( t * zi ) * y;
b = t * ( zi * y );
c = a - b;
print "t * 1/z * y: ";
print "a = " +str(a);
print "b = " +str(b);
print "a-b = " +str(c);
print
#bm = b.monic()
#print "monic(b) = " +str(bm);
#print
#exit(0);
f1 = x**2 + y**2 + z**2 + t**2 + 1;
print "f1 = ", f1;
f2 = t * x * f1;
print "f2 = ", f2;
F = [ f1, f2 ];
print "F =", [ str(f) for f in F ];
print
I = pt.ideal( list=F );
print "SolvableIdeal: " + str(I);
print;
#exit(0);
rgl = I.leftGB();
print "seq left GB:" + str(rgl);
print "isLeftGB: " + str(rgl.isLeftGB());
print;
#rgr = I.rightGB();
#print "seq right GB:" + str(rgr);
#print "isRightGB: " + str(rgr.isRightGB());
#print;
#startLog();
rgt = I.twosidedGB();
print "seq twosided GB:" + str(rgt);
print "isTwosidedGB: " + str(rgt.isTwosidedGB());
print;
#exit(0);
#startLog();
rgi = rgl.intersect(rgt);
print "leftGB intersect twosidedGB:" + str(rgi);
print;
#terminate();
| Python |
# jython examples for jas.
# $Id$
#
import sys
from java.lang import System
from jas import Ring, RF, QQ, PolyRing
from jas import terminate
from jas import startLog
# elementary integration atan examples
r = PolyRing(QQ(),"x",PolyRing.lex);
print "r = " + str(r);
rf = RF(r);
print "rf = " + str(rf.factory());
[one,x] = rf.gens();
print "one = " + str(one);
print "x = " + str(x);
print;
#f = 1 / ( 1 + x**2 );
#f = 1 / ( x**2 - 2 );
#f = 1 / ( x**3 - 2 );
#f = ( x + 3 ) / ( x**2- 3 * x - 40 );
f = ( x**7 - 24 * x**4 - 4 * x**2 + 8 * x - 8 ) / ( x**8 + 6 * x**6 + 12 * x**4 + 8 * x**2 );
print "f = ", f;
print;
#sys.exit();
#startLog();
t = System.currentTimeMillis();
e1 = r.integrate(f);
t = System.currentTimeMillis() - t;
print "e1 = ", e1;
print "integration time =", t, "milliseconds";
print
t = System.currentTimeMillis();
e2 = f.integrate();
t = System.currentTimeMillis() - t;
print "e2 = ", e2;
print "integration time =", t, "milliseconds";
print
#startLog();
terminate();
#sys.exit();
| Python |
#
# jython examples for jas.
# $Id$
#
from java.lang import System
from java.lang import Integer
from jas import Ring, PolyRing
from jas import terminate
from jas import startLog
from jas import QQ, DD
# polynomial examples: real roots over Q
#r = Ring( "Rat(x) L" );
#r = Ring( "Q(x) L" );
r = PolyRing(QQ(),"x",PolyRing.lex);
print "Ring: " + str(r);
print;
[one,x] = r.gens();
f1 = x * ( x - 1 ) * ( x - 2 ) * ( x - 3 ) * ( x - 4 ) * ( x - 5 ) * ( x - 6 ) * ( x - 7 ) ;
f2 = ( x - (1,2) ) * ( x - (1,3) ) * ( x - (1,4) ) * ( x - (1,5) ) * ( x - (1,6) ) * ( x - (1,7) ) ;
f3 = ( x - (1,2**2) ) * ( x - (1,2**3) ) * ( x - (1,2**4) ) * ( x - (1,2**5) ) * ( x - (1,2**6) ) * ( x - (1,2**7) ) ;
#f = f1 * f2 * f3;
f = f1 * f2;
#f = f1 * f3;
#f = f2 * f3;
#f = f3;
#f = ( x**2 - 2 );
print "f = ", f;
print;
startLog();
t = System.currentTimeMillis();
R = r.realRoots(f);
t = System.currentTimeMillis() - t;
#print "R = ", R;
print "R = ", [ a.elem.ring.getRoot() for a in R ];
print "real roots time =", t, "milliseconds";
eps = QQ(1,10) ** (DD().elem.DEFAULT_PRECISION-3);
print "eps = ", eps;
t = System.currentTimeMillis();
R = r.realRoots(f,eps);
t = System.currentTimeMillis() - t;
#print "R = ", [ str(r) for r in R ];
print "R = ", [ a.elem.decimalMagnitude() for a in R ];
print "real roots time =", t, "milliseconds";
#startLog();
terminate();
| Python |
#
# jython examples for jas.
# $Id$
#
import sys;
from jas import Ring
from jas import Ideal
from jas import startLog
from jas import terminate
# ? example
#r = Ring( "Rat(t,x,y,z) L" );
#r = Ring( "Z(t,x,y,z) L" );
#r = Ring( "Mod 11(t,x,y,z) L" );
r = Ring( "Rat(t,x,y) L" );
print "Ring: " + str(r);
print;
ps = """
(
( t - x - 2 y ),
( x^4 + 4 ),
( y^4 + 4 )
)
""";
# ( y^4 + 4 x y^3 - 2 y^3 - 6 x y^2 + 1 y^2 + 6 x y + 2 y - 2 x + 3 ),
# ( x^2 + 1 ),
# ( y^3 - 1 )
# ( y^2 + 2 x y + 2 y + 2 x + 1 ),
# ( y^4 + 4 x - 2 y^3 - 6 x + 1 y^2 + 6 x + 2 y - 2 x + 3 ),
# ( t - x - 2 y ),
# ( 786361/784 y^2 + 557267/392 y + 432049/784 ),
# ( x^7 + 3 y x^4 + 27/8 x^4 + 2 y x^3 + 51/7 x^3 + 801/28 y + 1041/56 )
# ( x**2 + 1 ),
# ( y**2 + 1 )
f = r.ideal( ps );
print "Ideal: " + str(f);
print;
#startLog();
rg = f.GB();
print "seq Output:", rg;
print;
terminate();
| Python |
Subsets and Splits
SQL Console for ajibawa-2023/Python-Code-Large
Provides a useful breakdown of language distribution in the training data, showing which languages have the most samples and helping identify potential imbalances across different language groups.