def main():
global TM
global LABELMAP
global CTMAP
global GENDER_TO_PRONOUN
global TOKEN_TO_GENDER
cfg = CONFIG[args.config]
catpeople = DbfilenameShelf(args.in_shelf, protocol=-1, flag='r')
TM = catpeople['__TOKEN_MAPPER__']
TM.finalize()
LABELMAP = util_catpeople.get_labelmap()
CTMAP = util_catpeople.get_coarse_tagmap()
GENDER_TO_PRONOUN = get_gender_to_pronoun(TM)
TOKEN_TO_GENDER = get_token_to_gender(TM)
if args.print_to_conll:
# Print CatPeople in Conll Format
partial_print_to_conll = functools.partial(print_to_conll,
catpeople=catpeople)
n_jobs = 4
Parallel(n_jobs=n_jobs)(
delayed(partial_print_to_conll)(out_fn=out_fn, urls=urls)
for (out_fn,
urls) in itertools.izip((args.out_fn + str(i) for i in range(
n_jobs)), split(catpeople['__URL_LIST__'], n_jobs)))
return
else:
name = cfg._name
if name.startswith(UNIGRAM):
return doc_to_unigrams(cfg, catpeople)
# doc_to_unigrams
# --> entity_list_to_ngram_csr_mat(n=0, width=None)
# --> get_ngrams_from_catpeople_entity
# --> yield_ngrams
# --> catpeople_sentence_iterator
elif name.startswith(BIGRAM):
return doc_to_bigrams(cfg, catpeople)
# doc_to_unigrams
# --> entity_list_to_ngram_csr_mat(n=0, width=None)
# --> get_width_for_bigrams
# --> entity_list_to_ngram_csr_mat(n=1, width=width)
elif name.startswith(UNIVEC):
return doc_to_univec(cfg, catpeople)
# doc_to_univec
# --> save_vec_file
# --> entity_list_to_ngram_csr_mat(n=0, width=None)
elif name.startswith(BIVEC):
return doc_to_bivec(cfg)
elif name.startswith(DSCTOK) or name.startswith(DSCSUF):
return doc_to_dscfeat(cfg, catpeople)
# --> entity_list_to_dscfeat_csr_mat
# --> get_dscfeat_from_catpeople_entity
# --> catpeople_sentence_iterator
# --> yield_dsctok
elif name.startswith(DSCTOKVEC):
return doc_to_dsctokvec(cfg)
elif name.startswith(UNISUF):
return doc_to_unisuf(cfg, catpeople)
else:
raise NotImplementedError(name)
def setup():
''' Load the catpeople data.
'''
url_mention = DbfilenameShelf(args.in_shelf, protocol=-1, flag='r')
TM = url_mention['__TOKEN_MAPPER__']
TM.finalize(catpeople_baseline_nb_config.MAX_TOK)
E = url_mention['__URL_LIST__']
DF = url_mention['__DF__']
cat_folds = pkl.load(open(args.fold_fn))
cat2url = util_catpeople.load_cat2url(args.cat2url_fn)
performance_aggregator = Performance_Aggregator(args=args)
return (url_mention, TM, E, cat_folds, cat2url, performance_aggregator, DF)
def __init__(self, datacfg, ppcfg, expcfg):
# Init Part 0
self.datacfg = datacfg
self.ppcfg = ppcfg
self.expcfg = expcfg
with rasengan.tictoc('Init Part 1 : The Datacfg'):
self.cp = DbfilenameShelf(
r'%s/%s'%(uc.get_pfx(),self.datacfg.cp_fn),
protocol=-1,
flag='r')
self.url_list = self.cp['__URL_LIST__']
self.TM = self.cp['__TOKEN_MAPPER__']
# self.TM.final must be patched to work with older
# versions of TokenMapper that are in the pickle.
if not hasattr(self.TM, 'final'):
self.TM.final = False
if self.is_malignull():
self.TM([self.expcfg.NULL_KEY])
self.bos_idx = self.TM.finalize()
self.pa = Aggregator(
datacfg=datacfg,
ppcfg=ppcfg,
expcfg=expcfg,
url_list=self.url_list,
TM=self.TM)
self.cat_folds = pkl.load(uc.proj_open(self.datacfg.fold_fn))
self.cat2url = uc.load_cat2url(uc.proj_open(self.datacfg.cat2url_fn))
self.url_to_idx = dict((b,a) for a,b in enumerate(self.url_list))
self.scratch = {}
pass
with rasengan.tictoc('Init Part 2 : The PP CFG'):
print 'Reading', 'catpeople_pp_%d'%args.ppcfg
self.smat = io.mmread(uc.proj_open('catpeople_pp_%d'%args.ppcfg))
assert scipy.sparse.isspmatrix_coo(self.smat)
if self.pp_prefix_is([UNIVEC, BIVEC, MALIGNER, DSCTOKVEC]):
self.vectors = np.load(uc.proj_open('catpeople_pp_%d.vec'%args.ppcfg))
pass
if self.is_malignull():
self.NULL_VEC = np.zeros((1,self.vectors.shape[1]))
if self.exp_prefix_is([NBKERNEL, KERMACH, MALIGNER]):
assert self.pp_prefix_is([UNIVEC, BIVEC, DSCTOKVEC])
if self.expcfg.rm_fn_word:
# Internally Manipulates smat
self.remove_fn_word()
if self.expcfg.weight_method.endswith('/df'):
self.populate_idf()
return
def update_shelf():
url_mention = DbfilenameShelf(args.in_shelf, protocol=-1)
TM = url_mention['__TOKEN_MAPPER__']
TM.finalize(catpeople_baseline_nb_config.MAX_TOK)
E = url_mention['__URL_LIST__']
n_doc = 10000
with rasengan.tictoc('Extracting Contexts'):
df_obj = TextualClueObject(E[:n_doc], url_mention, TM)
df = defaultdict(int)
for features in df_obj.features.itervalues():
for f in features:
df[f] += 1
for f in df.keys():
df[f] = df[f] / float(n_doc)
url_mention['__DF__'] = dict(df)
url_mention.close()
return
def setUpClass(cls):
super(TestEntityDescriptors, cls).setUpClass()
global TM
global LABELMAP
global CTMAP
cls.cpfn = (util_catpeople.get_pfx() + '/catpeople_clean_segmented_context.shelf')
cls.parsefn = (util_catpeople.get_pfx() + '/catpeople.parse.pkl')
cls.catpeople = DbfilenameShelf(cls.cpfn, protocol=-1, flag='r')
TM = cls.catpeople['__TOKEN_MAPPER__']
TM.finalize()
LABELMAP = util_catpeople.get_labelmap()
CTMAP = util_catpeople.get_coarse_tagmap()
# Inject global variables to module's namespace.
catpeople_preprocessor.TM = TM
catpeople_preprocessor.LABELMAP = LABELMAP
catpeople_preprocessor.CTMAP = CTMAP
catpeople_preprocessor.GENDER_TO_PRONOUN = catpeople_preprocessor.get_gender_to_pronoun(TM)
catpeople_preprocessor.TOKEN_TO_GENDER = catpeople_preprocessor.get_token_to_gender(TM)
catpeople_preprocessor.populate_dsctok_globals()
cls.testid = 1
print 'Calling setup'
import argparse
import sys, os
arg_parser = argparse.ArgumentParser(
description='Remove junk from catpeople wikimic')
arg_parser.add_argument('--seed', default=0, type=int, help='Default={0}')
arg_parser.add_argument('--MAX_CHAR_IN_SENT', default=1000, type=int)
PDIR = ('/export/b15/prastog3' if os.uname()[1] == 'b15' else 'data/')
arg_parser.add_argument('--in_shelf',
default='%s/catpeople_wikilink_mentions.shelf' % PDIR,
type=str)
arg_parser.add_argument('--out_shelf',
default='%s/catpeople_clean_segmented_context.shelf' %
PDIR,
type=str)
args = arg_parser.parse_args()
in_shelf = DbfilenameShelf(args.in_shelf, protocol=-1, flag='r')
out_shelf = DbfilenameShelf(args.out_shelf, protocol=-1)
urls = in_shelf['__URL_LIST__']
PAT_TOKENIZER = get_tokenizer()
TOKEN_MAPPER = TokenMapper()
MAX_CHAR_IN_SENT = args.MAX_CHAR_IN_SENT
import re
MIDDLE_NAME_REGEX = re.compile('[A-Z][^ ]*? [A-Z]\. [A-Z]')
for url_idx, url in enumerate(urls):
print >> sys.stderr, ('Done: %.3f \r' %
(float(url_idx) * 100 / len(urls))),
mentions = in_shelf[url]
out_mentions = []
for mention in mentions:
result = """
<section>
<p>You are not logged in.</p>
<p>
<a href="login.py">Login</a> | <a href="accounts/register.py">Register</a>
</p>
</section>"""
try:
cookie = SimpleCookie()
http_cookie_header = environ.get('HTTP_COOKIE')
if http_cookie_header:
cookie.load(http_cookie_header)
if 'sid' in cookie:
sid = cookie['sid'].value
session_store = DbfilenameShelf('sessions/sess_' + sid,
writeback=False)
if session_store.get('authenticated'):
message = ''
form_data = FieldStorage()
username = session_store.get('username')
form = """<p>
Hey, %s. Sorry to see you go.
</p>
<p>
<strong>Warning! This action is permenant.</strong> All of your scores will be lost.
</p>
<form action="delete_account.py" method="post">
<label for="pass1">Enter password: </label>
<input type="password" name="pass1" id="pass1" placeholder="Enter password" required />
<label for="pass2">Reenter password: </label>
<input type="password" name="pass2" id="pass2" placeholder="Reenter password" required />
import os
from pathlib import Path
import shelve
from shelve import Shelf, DbfilenameShelf
data = {'a': 0, 'b': 1, 'c': 'c-string'}
filename = str(Path.home() / "shelf")
#os.remove(filename + ".db")
db = DbfilenameShelf(filename, flag='c', protocol=3, writeback=True)
#db.update(data)
#db.sync()
print(f"shelf: {dict(db)}")
if len(form_data) != 0:
try:
cookie = SimpleCookie()
http_cookie_header = environ.get('HTTP_COOKIE')
if not http_cookie_header:
sid = sha256(repr(time()).encode()).hexdigest()
cookie['reset'] = sid
else:
cookie.load(http_cookie_header)
if 'reset' not in cookie:
sid = sha256(repr(time()).encode()).hexdigest()
cookie['reset'] = sid
else:
sid = cookie['reset'].value
session_store = DbfilenameShelf('../sessions/reset_' + sid,
writeback=True)
if session_store.get('code'):
code = escape(form_data.getfirst('code', '').strip())
if code:
form = """<form action="forgot.py" method="post">
<label for="code">Code: </label>
<input type="number" name="code" id="code" min="0" max="99999" value="%s" required />
<label for="pass1">Enter new password: </label>
<input type="password" name="pass1" id="pass1" required />
<label for="pass2">Reenter password: </label>
<input type="password" name="pass2" id="pass2" required />
<input type="submit" />
</form>""" % code
if session_store.get('code') == code:
pass1 = escape(form_data.getfirst('pass1', '').strip())
pass2 = escape(form_data.getfirst('pass2', '').strip())
import cPickle as pkl
from rasengan import groupby
PFX = get_pfx()
arg_parser = argparse.ArgumentParser(description='')
arg_parser.add_argument('--in_shelf',
default=PFX +
'/catpeople_clean_segmented_context.shelf',
type=str)
arg_parser.add_argument('--parsefn',
default=PFX + '/catpeople.parse.gz',
type=str)
arg_parser.add_argument('--parse_pkl',
default=PFX + '/catpeople.parse.pkl',
type=str)
args = arg_parser.parse_args()
catpeople = DbfilenameShelf(args.in_shelf, protocol=-1, flag='r')
TM = catpeople['__TOKEN_MAPPER__']
labelmap = get_labelmap()
ctmap = get_coarse_tagmap()
ftmap = get_fine_tagmap()
f = gzip.GzipFile(fileobj=proj_open(args.parsefn))
def get(e):
e = e.split('\t')
return [e[1], int(e[6]), e[7], e[3], e[4]]
PARSES = {}
for parse in groupby(f):
token, parent, labels, ctags, ftags = zip(*[get(r) for r in parse])
def _get_shelf_data(path):
with closing(DbfilenameShelf(path, flag='r')) as shelf:
return dict(shelf)
out_val.default_factory = None # FINALIZE out_val
return out_val
with rasengan.tictoc('Reading wikilinks'):
# With joblib this takes only 8 minutes !!
from joblib import Parallel, delayed
out_val_list = Parallel(n_jobs=10)(
delayed(get_mention_from_wikilink_thrift_file)(fn)
for fn in range(1, args.last_f2r))
# out_val_list = [get_mention_from_wikilink_thrift_file(fn)
# for fn in range(1, args.last_f2r)]
with rasengan.tictoc('Shelving'):
import shelve
from shelve import DbfilenameShelf
total_data = defaultdict(list)
for out_val in out_val_list:
for url in out_val:
total_data[url].extend(out_val[url])
total_data.default_factory = None # FINALIZE out_val
# Save the results of the processing.
shelf = DbfilenameShelf(args.out_fn, protocol=-1)
shelf['__URL_LIST__'] = total_data.keys()
for url in shelf['__URL_LIST__']:
shelf[url] = total_data[url]
shelf.close()
# Validation
for e in POOL:
assert e in total_data
