def award_names(year):
best_re = re.compile('best ', re.IGNORECASE)
trans_words = re.compile('\s(at|for|dressed|award|and|i\s|golden|http)',
re.IGNORECASE)
awards = []
for tweet in tweets_i_care_about(year):
tweet_text = tweet['text'].encode('ascii', 'ignore')
tweet_pieces = re.compile('[^\s\w-]').split(tweet_text.replace(
'#', ''))
award_tweets = [
x.lower() for x in tweet_pieces
if 'wins best' in x or 'won best' in x
]
for award_piece in award_tweets:
award_name = 'best ' + best_re.split(award_piece)[-1].rstrip()
awards.append(
trans_words.split(award_name)[0].encode('ascii', 'ignore'))
thresh = max(AWARDS_THRESH, len(awards) / 400)
fake_awards = [
'best', 'best dressed', 'best speech', 'best act', 'best actor',
'best actress'
]
awardCounter = collections.Counter(awards).iteritems()
ret = [[k, v] for k, v in awardCounter
if v > thresh and k not in fake_awards]
# ret = delete_duplicate_names(ret)
ret = [k for k, v in ret]
return ret
def award_names(year):
best_re = re.compile('best ', re.IGNORECASE)
trans_words = re.compile('\s(at|for|dressed|award|and|i\s|golden|http)', re.IGNORECASE)
awards = []
for tweet in tweets_i_care_about(year):
tweet_text = tweet['text'].encode('ascii', 'ignore')
tweet_pieces = re.compile('[^\s\w-]').split(tweet_text.replace('#',''))
award_tweets = [x.lower() for x in tweet_pieces if 'wins best' in x or 'won best' in x]
for award_piece in award_tweets:
award_name = 'best ' + best_re.split(award_piece)[-1].rstrip()
awards.append(trans_words.split(award_name)[0].encode('ascii', 'ignore'))
thresh = max(AWARDS_THRESH, len(awards)/400)
fake_awards = ['best', 'best dressed', 'best speech', 'best act', 'best actor', 'best actress']
awardCounter = collections.Counter(awards).iteritems()
ret = [[k,v] for k, v in awardCounter if v > thresh and k not in fake_awards]
# ret = delete_duplicate_names(ret)
ret = [k for k, v in ret]
return ret
def find_noms(year):
current_award = None
simple_official_awards = get_simple_official_awards()
unverified_noms_for_current_award = []
verified_noms = {}
for k, v in simple_official_awards.iteritems():
verified_noms[k] = {}
for tweet in tweets_i_care_about(year):
# code to associate with a specific award ###################
autoverified = False
tweet_text = tweet['text'].encode('ascii', 'ignore')
lc_tw = tweet_text.lower()
puncless_tw = remove_punc(lc_tw)
if current_award and simple_official_awards[
current_award] in puncless_tw:
verified_noms = add_count_to_dict(
unverified_noms_for_current_award, verified_noms,
current_award)
unverified_noms_for_current_award = []
autoverified = True
else:
for (real, simple) in simple_official_awards.iteritems():
if simple in puncless_tw:
current_award = real
autoverified = True
unverified_noms_for_current_award = []
break
##############################################################
if 'http' in lc_tw:
continue
# salty people ###############################################
# re.compile('i wish\s (had)?won)
# if 'was better than ' + winner ....
# shoulda_regex = re.compile('should have won', re.IGNORECASE)
# if re.search(shoulda_regex, lc_tw):
# subject_clause = shoulda_regex.split(tweet_text)[0]
# proper_noun_regex = re.compile('(([A-Z)]\w*\s?)+(\w*\s?)*)+\Z')
# print '.'
# mtch = re.search(proper_noun_regex, subject_clause)
# print '.'
# if not mtch:
# unverified_noms_for_current_award.append(mtch.group(1))
# print mtch.group(1)
# nominees lists ##############################################
# method using lists after 'nominees'
presenter_regex = re.compile(
r'(@\w*|[A-Z]\w*\s[A-Z]\w*)\s(present|introduce)s?')
tw_without_presenters = re.sub(presenter_regex, '', tweet_text)
noms_regex = re.compile(r'(nominees:\s)|(nominees are\s)',
re.IGNORECASE)
if re.search(noms_regex, tw_without_presenters):
nominees_text = noms_regex.split(tw_without_presenters)[-1]
nominees_text = re.sub(re.compile(r'#\w*'), '', nominees_text)
nominees_text = nominees_text.split('.')[0]
noms = list(
set(
re.compile('(\s&\s|\sand\s|,\s|\n)').split(
nominees_text)))
unverified_noms_for_current_award += noms
# add if autoverified #########################################
if autoverified:
verified_noms = add_count_to_dict(
unverified_noms_for_current_award, verified_noms,
current_award)
unverified_noms_for_current_award = []
# post processing #################################################
verified_noms = trim_nom_dict(verified_noms)
for award, noms in verified_noms.iteritems():
print award
for nom in noms:
print '-> -> ' + nom
return verified_noms
def find_noms(year):
current_award = None
simple_official_awards = get_simple_official_awards()
unverified_noms_for_current_award = []
verified_noms = {}
for k, v in simple_official_awards.iteritems():
verified_noms[k] = {}
for tweet in tweets_i_care_about(year):
# code to associate with a specific award ###################
autoverified = False
tweet_text = tweet['text'].encode('ascii', 'ignore')
lc_tw = tweet_text.lower()
puncless_tw = remove_punc(lc_tw)
if current_award and simple_official_awards[current_award] in puncless_tw:
verified_noms = add_count_to_dict(unverified_noms_for_current_award, verified_noms, current_award)
unverified_noms_for_current_award = []
autoverified = True
else:
for (real, simple) in simple_official_awards.iteritems():
if simple in puncless_tw:
current_award = real
autoverified = True
unverified_noms_for_current_award = []
break
##############################################################
if 'http' in lc_tw:
continue
# salty people ###############################################
# re.compile('i wish\s (had)?won)
# if 'was better than ' + winner ....
# shoulda_regex = re.compile('should have won', re.IGNORECASE)
# if re.search(shoulda_regex, lc_tw):
# subject_clause = shoulda_regex.split(tweet_text)[0]
# proper_noun_regex = re.compile('(([A-Z)]\w*\s?)+(\w*\s?)*)+\Z')
# print '.'
# mtch = re.search(proper_noun_regex, subject_clause)
# print '.'
# if not mtch:
# unverified_noms_for_current_award.append(mtch.group(1))
# print mtch.group(1)
# nominees lists ##############################################
# method using lists after 'nominees'
presenter_regex = re.compile(r'(@\w*|[A-Z]\w*\s[A-Z]\w*)\s(present|introduce)s?')
tw_without_presenters = re.sub(presenter_regex, '', tweet_text)
noms_regex = re.compile(r'(nominees:\s)|(nominees are\s)', re.IGNORECASE)
if re.search(noms_regex, tw_without_presenters):
nominees_text = noms_regex.split(tw_without_presenters)[-1]
nominees_text = re.sub(re.compile(r'#\w*'),'', nominees_text)
nominees_text = nominees_text.split('.')[0]
noms = list(set(re.compile('(\s&\s|\sand\s|,\s|\n)').split(nominees_text)))
unverified_noms_for_current_award += noms
# add if autoverified #########################################
if autoverified:
verified_noms = add_count_to_dict(unverified_noms_for_current_award, verified_noms, current_award)
unverified_noms_for_current_award = []
# post processing #################################################
verified_noms = trim_nom_dict(verified_noms)
for award, noms in verified_noms.iteritems():
print award
for nom in noms:
print '-> -> ' + nom
return verified_noms
def find_winners(year):
current_award = None
simple_official_awards = get_simple_official_awards()
unverified_wins_current = []
verified_wins = {}
for k, v in simple_official_awards.iteritems():
verified_wins[k] = defaultdict(lambda: 0)
for tweet in tweets_i_care_about(year):
autoverified = False
lc_tw = tweet['text'].lower()
puncless_tw = remove_punc(lc_tw)
if current_award and simple_official_awards[current_award] in puncless_tw:
verified_wins = add_count_to_dict(unverified_wins_current, verified_wins, current_award)
unverified_wins_current = []
autoverified = True
else:
for (real, simple) in simple_official_awards.iteritems():
#print simple
for x in award_name_gen(simple):
if x in puncless_tw:
#print x
current_award = real
verified_wins = add_count_to_dict(unverified_wins_current, verified_wins, current_award)
#print current_award
autoverified = True
unverified_wins_current = []
if 'wins' not in lc_tw and 'won' not in lc_tw and 'goes to' not in lc_tw or 'http' in lc_tw:
#print "no"
continue
presenter_regex = re.compile(r'(@\w*|[A-Z]\w*\s[A-Z]\w*)\s(present|introduce)')
tw_without_presenters = re.sub(presenter_regex, ' ', tweet['text'])
if 'goes to' in lc_tw or 'wins' in lc_tw or 'won' in lc_tw:
#print current_award
if 'wins' in lc_tw:
#print "yes"
winner_text = re.compile(r'wins', re.IGNORECASE).split(tw_without_presenters)[0]
winner_text = winner_text.encode('ascii', 'ignore')
winner_text = re.sub(re.compile(r'#\w*'),'', winner_text)
if 'won' in lc_tw:
#print "yes"
winner_text = re.compile(r'won', re.IGNORECASE).split(tw_without_presenters)[0]
winner_text = winner_text.encode('ascii', 'ignore')
winner_text = re.sub(re.compile(r'#\w*'),'', winner_text)
# print winner_text
if 'goes to' in lc_tw:
winner_text = re.compile(r'goes to', re.IGNORECASE).split(tw_without_presenters)[-1]
winner_text = re.compile(r'for', re.IGNORECASE).split(winner_text)[0]
winner_text = winner_text.encode('ascii', 'ignore')
winner_text = re.sub(re.compile(r'#\w*'),'', winner_text)
# print winner_text
# if 'winner' in lc_tw:
# #winner_text = re.compile(r'winner', re.IGNORECASE).split(tw_without_presenters)[-1]
# winner_text = tw_without_presenters.encode('ascii', 'ignore')
# winner_text = re.sub(re.compile(r'#\w*'),'', winner_text)
prop = re.compile(r'([A-Z]{1}[a-z]{1,}(\s[A-Z]{1}[a-z]{1,})?)')
matches = prop.findall(winner_text)
#print matches, current_award
matches = [i[0] for i in matches]
matches = filter_false_positive(matches)
#if autoverified:
#if autoverified:
#print current_award, matches
wins = list(matches)
#print "matches:", wins, current_award
unverified_wins_current += wins
#print wins
#print wins
if autoverified:
verified_wins = add_count_to_dict(unverified_wins_current, verified_wins, current_award)
unverified_wins_current = []
#verified_wins = trim_nom_dict(verified_wins)
#print verified_wins
final_dict = {}
for k, v in verified_wins.iteritems():
#print k
highWin=''
highCount= 0
for win, count in v.iteritems():
if count > highCount and win != '':
highCount = count
highWin = win
final_dict[k] = highWin
return final_dict
# find_noms(2015)