def __init__(self):
logging.basicConfig()
self.loop_flag = True
self.bot_state = 0
self.QuestionObject = Q.Question()
trial1 = Pipeline([
('vectorizer',
TfidfVectorizer(tokenizer=stemming_tokenizer,
stop_words=stopwords.words('english'))),
('classifier', MultinomialNB()),
])
self.detector0 = train(trial1, xtrain, ytrain, xtest, ytest)
self.sentiment = sent.classifier()
def search_and_return(subreddit, search_string):
NBclassifier = classifier()
# I may need to set this as a global
# For now, I will call one every single time
# *NOT SCALABLE*
# r = praw.Reddit('search application')
now = datetime.datetime.now()
# search
search_results = r.search(search_string,
subreddit = subreddit,
sort = True)
for i in xrange(search_results.__sizeof__()):
comments = []
post = search_results.next()
for comment in post.comments:
try:
comment_str = str(comment.body.encode('utf-8'))
comments.append(comment_str)
comments.append(comment.score)
except AttributeError:
pass
score = 0
for i in range(0, len(comments), 2):
for comment in comments[i].split('\n'):
if any(i in str(comment) for i in search_string.lower().split(' ')):
#print comment
sentiment = NBclassifier.classify(word_feats(comment))
if sentiment == 'neg':
score += -1 * int(comments[i+1])
else:
score += 1 * int(comments[i+1])
if score > 10:
result = str(now) + " POSITIVE " + str(score)
elif score < -10:
result = str(now) + " NEGATIVE " + str(score)
else:
result = str(now) + " NEUTRAL " + str(score)
#return result, max_post_comment, pos_score, max_neg_comment, neg_score
return result
def search_and_return(subreddit, search_string):
NBclassifier = classifier()
# I may need to set this as a global
# For now, I will call one every single time
# *NOT SCALABLE*
# r = praw.Reddit('search application')
now = datetime.datetime.now()
# search
search_results = r.search(search_string, subreddit=subreddit, sort=True)
for i in xrange(search_results.__sizeof__()):
comments = []
post = search_results.next()
for comment in post.comments:
try:
comment_str = str(comment.body.encode('utf-8'))
comments.append(comment_str)
comments.append(comment.score)
except AttributeError:
pass
score = 0
for i in range(0, len(comments), 2):
for comment in comments[i].split('\n'):
if any(i in str(comment)
for i in search_string.lower().split(' ')):
#print comment
sentiment = NBclassifier.classify(word_feats(comment))
if sentiment == 'neg':
score += -1 * int(comments[i + 1])
else:
score += 1 * int(comments[i + 1])
if score > 10:
result = str(now) + " POSITIVE " + str(score)
elif score < -10:
result = str(now) + " NEGATIVE " + str(score)
else:
result = str(now) + " NEUTRAL " + str(score)
#return result, max_post_comment, pos_score, max_neg_comment, neg_score
return result
for i in range(0, len(comments), 2):
for comment in comments[i].split('\n'):
if any(i in str(comment)
for i in search_string.lower().split(' ')):
#print comment
sentiment = classifier.classify(word_feats(comment))
if sentiment == 'neg':
score += -1 * int(comments[i + 1])
else:
score += 1 * int(comments[i + 1])
if score > 10:
result = str(now) + " POSITIVE " + str(score)
elif score < -10:
result = str(now) + " NEGATIVE " + str(score)
else:
result = str(now) + " NEUTRAL " + str(score)
#return result, max_post_comment, pos_score, max_neg_comment, neg_score
return result
#return flask.render_template('result.html',result=result)
NBclassifier = classifier()
print search_and_return('lebron', 'nba', NBclassifier)
print search_and_return('trump', 'politics', NBclassifier)
print search_and_return('peyton manning', 'nfl', NBclassifier)
print search_and_return('monty williams', 'nba', NBclassifier)
print search_and_return('frank underwood', 'houseofcards', NBclassifier)
print search_and_return('bernie sanders', 'politics', NBclassifier)
from copy import deepcopy
import numpy as np
from sklearn.covariance import GraphLassoCV
from sklearn import manifold, cluster
from matplotlib.collections import LineCollection
import pylab as pl
##############################################################################
# Retrieve the data
begin = datetime.date( 2011, 1, 3 )
end = datetime.date( 2013, 11, 27 )
tickerList = keywords.getTickerList()
keywordsMap = keywords.getKeywordToIndexMap()
sentCounter = count.SentimentWordCounter( keywordsMap, sentiment.classifier() )
mentionCounter = count.WordCounter( keywordsMap )
empiricalDf = matrices.getEmpiricalDataFrame( tickerList, begin, end )[ tickerList ]
getTfIdf = lambda wordCounter, aggregator: normalize.TfIdf()( matrices.getCountDataFrame( tickerList, wordCounter, empiricalDf.index, aggregator = aggregator ) )[ tickerList ]
tfIdfSentArticle = getTfIdf( sentCounter, None )[ tickerList ]
tfIdfSentDay = getTfIdf( sentCounter, np.sum )[ tickerList ]
tfIdfMentionArticle = getTfIdf( mentionCounter, None )[ tickerList ]
tfIdfMentionDay = getTfIdf( mentionCounter, np.sum )[ tickerList ]
matrices = { 'Empirical' : {
'By Day' : { 'Data' : empiricalDf }
},
'Signed Mentions' : {
'By Day' : { 'Data' : tfIdfSentDay },
'By Article' : { 'Data' : tfIdfSentArticle }
pass
for i in range(0, len(comments), 2):
for comment in comments[i].split('\n'):
if any(i in str(comment) for i in search_string.lower().split(' ')):
#print comment
sentiment = classifier.classify(word_feats(comment))
if sentiment == 'neg':
score += -1 * int(comments[i+1])
else:
score += 1 * int(comments[i+1])
if score > 10:
result = str(now) + " POSITIVE " + str(score)
elif score < -10:
result = str(now) + " NEGATIVE " + str(score)
else:
result = str(now) + " NEUTRAL " + str(score)
#return result, max_post_comment, pos_score, max_neg_comment, neg_score
return result
#return flask.render_template('result.html',result=result)
NBclassifier = classifier()
print search_and_return('lebron','nba',NBclassifier)
print search_and_return('trump' , 'politics', NBclassifier)
print search_and_return('peyton manning', 'nfl', NBclassifier )
print search_and_return('monty williams', 'nba', NBclassifier)
print search_and_return('frank underwood', 'houseofcards', NBclassifier)
print search_and_return('bernie sanders', 'politics', NBclassifier)
