def extract_textfeatures_flavor2(comment):
words = ['f**k', 'bitch', 'shut', 'hate', 'suck', 'gay', 'ugli', 'work', 'beauti', 'sick']
analyzer = vaderSentiment.SentimentIntensityAnalyzer()
vader = analyzer.polarity_scores(comment).get('compound')
f = pd.read_csv('Terms-to-Block.csv')['0'].tolist()
ps = PorterStemmer()
wordnet_lemmatizer = WordNetLemmatizer()
words = [ps.stem(wordnet_lemmatizer.lemmatize(w)) for w in words]
comment = unicode(comment, errors='ignore')
comment = comment.decode('utf-8').encode('ascii', errors='ignore')
punc = sum([comment.count(i) for i in ["!", "?"]])
url =len(re.findall('https?://(?:[-\w.]|(?:%[\da-fA-F]{2}))+', comment))
user = len(re.findall(r'(?<=^|(?<=[^a-zA-Z0-9-_\.]))@([A-Za-z]+[A-Za-z0-9-_]+)', comment))
hashtag = len(re.findall(r"#(\w+)", comment))
comment = ' '.join(re.sub("(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)"," ",comment).split())
textlen = len(comment.split())
if textlen ==0:
textlen =1
punc = punc/textlen
uppercase = len(filter(lambda x: x in string.uppercase, comment)) / textlen
comment = [x.lower() for x in comment.split()]
comment = [ps.stem(wordnet_lemmatizer.lemmatize(w)) for w in comment]
comment = " ".join(comment)
badwords = sum([comment.count(i) for i in f])/textlen
# for uni in [comment.count(x) for x in words]:
# return uni
unigrams = [comment.count(x) for x in words]
all_features = tuple([textlen,punc,uppercase,badwords,vader,hashtag,user,url]+unigrams)
# return textlen,punc,uppercase,badwords,vader,hashtag,user,url,tuple(unigrams)
return all_features
def index():
form = NameForm()
if form.validate_on_submit():
score = vaderSentiment.SentimentIntensityAnalyzer().polarity_scores(
form.name.data)['compound']
if score > 0.05:
sentiment = 'Positive'
category = 'success'
elif score < -0.05:
sentiment = 'Negative'
category = 'danger'
else:
sentiment = 'Neutral'
category = 'info'
flash('Score = ' + str(score) + ' (' + sentiment + ')', category)
session['name'] = form.name.data
return redirect(url_for('index'))
return render_template('index.html', form=form, name=session.get('name'))
import time
from alpha_vantage.timeseries import TimeSeries
import json
import re
import vaderSentiment
ALPHA_API_KEY = ''
CONSUMER_KEY = ''
CONSUMER_SECRET_KEY = ''
ACCESS_TOKEN = ''
ACCESS_SECRET_TOKEN = ''
STOCK_NAME = 'TSLA'
TWEET_SAMPLING_TIME = 100
analyzer = vaderSentiment.SentimentIntensityAnalyzer()
def generate_stocks_csv(ts):
stock_data, meta_data = ts.get_daily(STOCK_NAME, outputsize='full')
#stock_data.drop('2. high', 1)
del stock_data['2. high']
del stock_data['1. open']
del stock_data['3. low']
stock_data.to_csv(STOCK_NAME + '.csv', sep=',')
def cleanse_tweet(tweet):
# process the tweets
#Convert to lower case