def testTweets():
for i in range(1, 8):
statuses = api.GetSearch(term=settings_local.USERTRACK, lang="es", page=i, per_page=100)
for s in statuses:
t = Normalizator.normalize(s.text)
if not (t.isspace()):
if classifier.classify(extract_features(t.split())) == "reclamo":
print "\033[91m" + t
else:
print "\033[94m" + t
def saveTweetsOf(screen_name, category):
for i in range(1,4):
statuses = api.GetSearch(term=settings_local.USERTRACK, lang='es', page=i,per_page=100)
for s in statuses:
if(s.text.find('RT',0,2)==-1):
screen_name = s._user._screen_name
t = Normalizator.normalize(s.text)
em = Normalizator.usedPattern(Patterns.DUPLICATED_LETTERS)
print Normalizator.usedPattern(Patterns.SPECIALCHARS_AND_MENTIONS)
print smart_str(s.text.lower())
if (not(t.isspace())):
selected = raw_input('Es un reclamo?')
tweets = db[settings_local.CATEGORIES[int(selected)]]
tweet = {'normalized_tweet': t,
'tweet': s.text,
'screen_name': screen_name,
'emphatized': em
}
tweets.insert(tweet)
for s in statuses:
t = Normalizator.normalize(s.text)
if not (t.isspace()):
if classifier.classify(extract_features(t.split())) == "reclamo":
print "\033[91m" + t
else:
print "\033[94m" + t
allTweets = []
for c in settings_local.CATEGORIES:
statuses = db[c].find()
for s in statuses:
t = s["normalized_tweet"]
t = Normalizator.normalize(t)
if not (t.isspace()):
allTweets.append((t, c))
tweets = []
for (words, sentiment) in allTweets:
words_filtered = [e.lower() for e in words.split() if len(e) >= 3]
tweets.append((words_filtered, sentiment))
def get_words_in_tweets(tweets):
all_words = []
for (words, sentiment) in tweets:
all_words.extend(words)
return all_words