def use_indico(Xtest, ytest):
'''
Batch reviews in groups of 100 and send them to the Indico high quality API to get
sentiment results. Return the list of results, as well as the test features
and test targets, to be used in testing the results. Process each document as a
whole, rather than processing each sentence individually.
Isolate the fetching of the sentiment results from Indico from the use of
those results, so that if something goes wrong, we don't need to fetch again.
No need to clean and vectorize training reviews, or train a random forest
on them, because Indico has done all of that already. Just strip out html.
'''
print "Cleaning html from the test set of movie reviews..."
clean_test_reviews = remove_html(Xtest)
print "Running Indico queries..."
print "This will take a while..."
# process the reviews in batches of 1000, then finish with the leftovers, if any
# Indico is not splitting on sentences, returns one sentiment per review
sentiment_lists = []
for i in range(100,len(Xtest),100):
print "Processing reviews {0} to {1}...".format(i-100, i-1)
batch = clean_test_reviews[i-100:i]
results = indicoio.sentiment_hq(batch)
sentiment_lists += results
if len(sentiment_lists)<len(Xtest):
print "Processing final reviews {0} to {1}...".format(len(sentiment_lists),len(Xtest))
batch = clean_test_reviews[len(sentiment_lists):]
results = indicoio.sentiment_hq(batch)
sentiment_lists += results
print "{0} Indico sentiments returned".format(len(sentiment_lists))
return sentiment_lists
def analyze_image(self, gif_url):
"""
Checks to see if there are any recognizable objects in the gifs, and if there are performs sentiment analysis on the objects. Then it outputs the average of the sentiment analysis of all the objects
"""
gif_output = []
#print(gif_url)
while True:
try:
gif_output = indicoio.image_recognition(gif_url)
break
except ValueError:
#print('oops')
return None
#print(gif_output)
self.list_objects = [
name for name, prob in gif_output.items() if prob > .1
]
if len(self.list_objects) == 1:
self.sentiment_output = indicoio.sentiment_hq(self.list_objects[0])
elif len(self.list_objects) == 0:
return self.text_sentiment
else:
self.sentiment_output = indicoio.sentiment_hq(self.list_objects)
self.sentiment_output = np.mean(self.sentiment_output)
return self.sentiment_output
def sentiment(link, title):
# r = requests.post('http://text-processing.com/api/sentiment/', data = {'text': title})
# posnegneutral = json.loads(r.text)['label']
try:
link_pos = indicoio.sentiment_hq(link)
if link_pos > 0.9 or link_pos < 0.2:
# save an API call
# - don't need to waste one if page contents already > 0.9 or < 0.2 (pos)
title_pos = link_pos
else:
title_pos = indicoio.sentiment_hq(title)
avg_pos = (link_pos + title_pos) / 2
if link_pos > 0.9 or title_pos > 0.9:
posnegneutral = "pos - {} (max)".format(max(link_pos, title_pos))
elif avg_pos > 0.8:
posnegneutral = "pos - {} (avg)".format(avg_pos)
elif link_pos < 0.2 or title_pos < 0.2:
posnegneutral = "neg - {} (min)".format(min(link_pos, title_pos))
elif avg_pos < 0.3:
posnegneutral = "neg - {} (avg)".format(avg_pos)
else:
posnegneutral = "neutral (link_pos={} title_pos={} avg={})".format(
link_pos, title_pos, avg_pos)
except:
e = sys.exc_info()
posnegneutral = "ERROR Checking Sentiment: {}".format(e)
link_pos = 0
title_pos = 0
avg_pos = 0
return posnegneutral, link_pos, title_pos, avg_pos
def test_sentiment_hq(self):
test_string = "Worst song ever."
response = sentiment_hq(test_string)
self.assertTrue(isinstance(response, float))
self.assertTrue(response < 0.5)
test_string = "Best song ever."
response = sentiment_hq(test_string)
self.assertTrue(isinstance(response, float))
self.assertTrue(response > 0.5)
def test_sentiment_hq(self):
test_string = "Worst song ever."
response = sentiment_hq(test_string, job_options={"job": True})
self.assertTrue(isinstance(response, (float, np.float32)))
self.assertTrue(response < 0.5)
test_string = "Best song ever."
response = sentiment_hq(test_string, job_options={"job": True})
self.assertTrue(isinstance(response, (float, np.float32)))
self.assertTrue(response > 0.5)
def get_sentiment(statement, lookup_table, save_results=False):
if save_results:
if statement in lookup_table.index:
return lookup_table.loc[statement, 'sentiment_hq'], False
else:
print('Result not found in table')
return indicoio.sentiment_hq(statement), True
else:
if statement in lookup_table.index:
return lookup_table.loc[statement, 'sentiment_hq']
else:
print('Result not found in table')
return indicoio.sentiment_hq(statement)
def on_status(self, status):
global sentiment_value
body = status.text
s = indicoio.sentiment_hq(body)
s = Decimal(s).quantize(Decimal('.0001'), rounding=ROUND_DOWN)
sentiments.append(s)
sentiment_value = (sum(sentiments) / len(sentiments))
def main():
indicoio.config.api_key = '123273ff84fe220626891873d499ea07'
indicoio.config.language = 'russian'
# results:
#0.94399955814
#print indicoio.sentiment('хороший кот', language='russian')
#0.777086528524
#print indicoio.sentiment('постановление правительство', language='russian')
print indicoio.sentiment('хороший', language='russian')
print indicoio.sentiment('правительство', language='russian')
print indicoio.sentiment('кот', language='russian')
return
res = indicoio.sentiment_hq([
'хороший кот',
'постановление правительство',
'состоятельный оказаться',
'коррупционный правонарушение',
'конфликт интерес',
'первое квартал'
])
for r in res:
print r
def main():
if len(sys.argv) != 3:
return
inname = sys.argv[1]
outname = sys.argv[2]
with open(inname, mode='r') as inFile:
tweets = json.load(inFile)
count = 0
for tweet in tweets:
if tweet['lang'] == 'en':
tweet['positiveness'] = float(
indicoio.sentiment_hq(tweet['text']))
else:
tweet['positiveness'] = None
count += 1
if count % 100 == 0:
print(count)
with open(outname, 'w') as outfile:
json.dump(tweets, outfile)
with open(outname, 'w') as outfile:
json.dump(tweets, outfile)
def processData(self):
textNoSym = str.replace(str.replace((str)(self.text), '#', ''), '@',
'')
textNoSym = re.sub(r'https:.*$', '', textNoSym)
if textNoSym == '': # If without emojis and URLs the tweet is an empty string return neutral sentiment
return 50
return (int)(indicoio.sentiment_hq(textNoSym) * 100)
def get_weighted_average_sentiments(self):
if (len(self.yaks) > 0):
sentiment_sum = 0.0
yak_info = [[yak.message, yak.likes] for yak in self.yaks
if yak.likes != 10000]
upvotes = [yak[1] for yak in yak_info]
sentiments = indicoio.sentiment_hq([yak[0] for yak in yak_info])
entries = len(upvotes) + sum(
[math.fabs(upvote) for upvote in upvotes])
for index in range(len(sentiments)):
sentiment_sum += sentiments[index] * upvotes[index]
percent = (sentiment_sum / entries - .4) * 200
message = "Positive" if percent >= 0 else "Negative"
style = "color:green;" if percent >= 0 else "color:red;"
return {
"percent": "{0:.2f}".format(math.fabs(percent)),
"message": message,
"style": style
}
return {
"percent": 100.0,
"message": "Not Available",
"style": "color:orange;"
}
def indico_batch_sentiment():
"""a ONE-OFF method to call the indico.io API to HQ batch sentiment 18192 texts.
Kinda shows how badly designed OOP code this is right now for the hackathon, oops.
Should refactor when there is time.
"""
with open('sentiments.csv', 'wb') as f:
texts = []
writer = csv.writer(f)
with open('texts/filenames.txt', 'r') as filenames:
fn_list = map(str.strip, [filename for filename in filenames])
fn_list = map(lambda x: 'texts/texts/' + x, fn_list)
for fn in fn_list:
texts.append(get_texts(fn)) # returns TextMessage object
texts = [item for sublist in texts for item in sublist]
with open('indico_sentiment_hq_errors.txt', 'w') as error_log:
for text in texts:
sentiment_result = None
try:
sentiment_result = sentiment_hq(text.body.encode(),
api_key=INDICO_API_KEY)
except BaseException as e:
error_log.write(str(e))
finally:
writer.writerow([
unicode(s).encode('utf-8') for s in [
text.msg_id, text.posix,
repr(text.sent), text.body,
repr(text.mentions), sentiment_result
]
])
def transcribe_file(speech_file):
"""Transcribe the given audio file."""
from google.cloud import speech
from google.cloud.speech import enums
from google.cloud.speech import types
client = speech.SpeechClient()
with io.open(speech_file, 'rb') as audio_file:
content = audio_file.read()
audio = types.RecognitionAudio(content=content)
config = types.RecognitionConfig(
encoding=enums.RecognitionConfig.AudioEncoding.LINEAR16,
sample_rate_hertz=16000,
language_code='en-US')
response = client.recognize(config, audio)
# Each result is for a consecutive portion of the audio. Iterate through
# them to get the transcripts for the entire audio file.
y = ""
for result in response.results:
# The first alternative is the most likely one for this portion.
x = format(result.alternatives[0].transcript)
y = y + x
#get_text_sentiment()
sentiment = indicoio.sentiment_hq(y)
return "Transcript: " + y + " \n Sentiment:" + str(sentiment)
def indico_batch_sentiment():
"""a ONE-OFF method to call the indico.io API to HQ batch sentiment 18192 texts.
Kinda shows how badly designed OOP code this is right now for the hackathon, oops.
Should refactor when there is time.
"""
with open('sentiments.csv', 'wb') as f:
texts = []
writer = csv.writer(f)
with open('texts/filenames.txt', 'r') as filenames:
fn_list = map(str.strip, [filename for filename in filenames])
fn_list = map(lambda x: 'texts/texts/' + x, fn_list)
for fn in fn_list:
texts.append(get_texts(fn)) # returns TextMessage object
texts = [item for sublist in texts for item in sublist]
with open('indico_sentiment_hq_errors.txt', 'w') as error_log:
for text in texts:
sentiment_result = None
try:
sentiment_result = sentiment_hq(text.body.encode(), api_key=INDICO_API_KEY)
except BaseException as e:
error_log.write(str(e))
finally:
writer.writerow([unicode(s).encode('utf-8') for s in
[text.msg_id, text.posix, repr(text.sent),
text.body, repr(text.mentions), sentiment_result]])
def analysis(data):
sentiment = ind.sentiment_hq(data)
tags = sort(ind.text_tags(data))
languages = sort(ind.language(data))
politics = sort(ind.political(data))
keywords = sort(ind.keywords(data))
names = sort(ind.named_entities(data))
print "Sentiment", sentiment
print "\n\n\nTags"
for t in tags:
print t[0], float(t[1]) * 100
print "\n\n\nLanguages"
for l in languages:
print l[0], float(l[1]) * 100
print "\n\n\nPolitical"
for p in politics:
print p[0], float(p[1]) * 100
print "\n\nkeywords"
for k in keywords:
print k[0], float(k[1]) * 100
def sentiments(twitchchat): """ Prints the sentiment of the whole duration the program was open. """ sentimental = indicoio.sentiment_hq(twitchchat) #uses pattern's sentiment analysis to calculate how troll or nice the chat is print "This streamer has a sentiment polarity of " + str(sentimental)
def results():
keyword = str(request.args.get("keyword", None))
conn = sqlite3.connect(":memory:")
cursor = conn.cursor()
cursor.execute("""
CREATE TABLE IF NOT EXISTS main(
Keyword TEXT,
Title TEXT,
Description TEXT,
URL TEXT,
IMGLink TEXT,
Score INTEGER,
sqltime TIMESTAMP DEFAULT CURRENT_TIMESTAMP NOT NULL
);
""")
resultslist = []
scorelist = []
index = 0
limit = 20
data = newsapi.get_everything(
q=keyword,
sources='crypto-coins-news',
language='en',
sort_by='publishedAt',
)
load = data['articles']
for index, post in zip(range(limit), load):
title = post['title']
description = post['description']
url = post['url']
imglink = post['urlToImage']
scoredesc = indicoio.sentiment_hq(title)
calc = (scoredesc * 100)
score = ("%.2f" % calc)
scorelist.append(float(score))
nline = keyword, title, description, url, imglink, score
cursor.execute(
'INSERT INTO main (Keyword, Title, Description, URL, IMGLink, Score) VALUES (?, ?, ?, ?, ?, ?)',
(nline))
conn.commit()
for row in cursor.execute(
"SELECT Title, Description, URL, IMGLink, Score, sqltime FROM main ORDER BY Score DESC LIMIT 16"
):
resultslist.append(row)
sumlist = (sum(scorelist))
lenght_list = (len(scorelist))
before = (sumlist / lenght_list)
average = float("%.0f" % before) + 10
return render_template('analysis.html',
resultslist=resultslist,
keyword=keyword,
average=average)
def test_batched_error_handling(self):
test_data = ["Terribly interesting test data."] * 100
test_data[98] = ""
with self.assertRaises(IndicoError):
sentiment_hq(test_data, batch_size=20)
files = glob.glob('indico-sentimenthq-*.json')
assert len(files)
for filename in files:
data = json.load(open(filename, 'r'))
# first four batches should have returned
assert len(data) == 80
# clean up after ourselves
os.remove(filename)
def retrieve_indico(tokz):
indi = indicoio.sentiment_hq(tokz)
if indi>0.5:
isent = 'POS'
elif indi<0.5:
isent = 'NEG'
else: isent = 'NEU'
return isent
def test_batched_error_handling(self):
test_data = ["Terribly interesting test data."] * 100
test_data[98] = ""
with self.assertRaises(IndicoError):
sentiment_hq(test_data, batch_size=20)
files = glob.glob('indico-sentimenthq-*.json')
assert len(files)
for filename in files:
data = json.load(open(filename, 'r'))
# first four batches should have returned
assert len(data) == 80
# clean up after ourselves
os.remove(filename)
def handle(self, *args, **options):
tweet = Tweet.objects.get(id=options['tweet_id'][0])
score = indicoio.sentiment_hq(tweet.text)
tweet.sentiment_score = score
tweet.save()
self.stdout.write(str(score))
return
def tweety_search(searchName):
tweet_list = []
search = api.GetSearch(term=searchName, lang='en', result_type='recent', count=NUMBER_TWEETS, max_id='')
for t in search:
tweet_list.append(t.text.encode('utf-8'))
score = indicoio.sentiment_hq(tweet_list)
top_and_bottom(tweet_list, score)
score = statistics.mean(score)
return score
def sentiments(twitchchat):
"""
Prints the sentiment of the whole duration the program was open.
"""
sentimental = indicoio.sentiment_hq(
twitchchat
) #uses pattern's sentiment analysis to calculate how troll or nice the chat is
print "This streamer has a sentiment polarity of " + str(sentimental)
def main():
configure()
examples = filter(lambda sen: sen != '', get_examples('input.txt'))
# single examples
sentiments = indicoio.sentiment_hq(examples)
poli = indicoio.political(examples)
for i in range(len(examples)):
print('============')
print('{}\n\n{}\n\n{}\n'.format(examples[i], sentiments[i], poli[i]))
print('============')
def wordsentiment(words):
"""
returns the sentiment of the inputted phrase
"""
if words == '': #sometimes after filtering the individual phrase has nothing left, this is to account for that
sentiments = 0
else:
sentiments = indicoio.sentiment_hq(
words) #using indico's API because pattern's is bad
return sentiments
def wordsentiment(words): """ returns the sentiment of the inputted phrase """ if words == '': #sometimes after filtering the individual phrase has nothing left, this is to account for that sentiments = 0 else: sentiments = indicoio.sentiment_hq(words) #using indico's API because pattern's is bad return sentiments
def sms():
response = twiml.Response()
body = request.form['Body']
print(body)
value = indicoio.sentiment_hq(body)
print(value)
response.message("You sent: {0}\nSentiment analysis: {1}".format(body, value))
data = {
"text": body,
"value": value
}
messages.insert_one(data)
return str(response)
def loadRecentPosts(self,recent_tags, api):
for tag in recent_tags:
#split the string returned to get users id
temp, new_id = tag.id.split('_')
user = api.user(new_id)
#gets amount of posts user has made
postCount = user.counts['media']
#gets the amount of followers the user has
followers = user.counts['followed_by']
#gets the amount of people the user is following
following = user.counts['follows']
#gets the number of likes of the post
likes = tag.like_count
print 'Post Number:', self.numPosts
print likes, 'likes'
print "Users Number of Posts:", postCount
print "Followers:", followers
print "Following:", following
# Checks each word in caption to see if it is positive, neutral or negative and
# puts it into a list then calculates its radius based on number of followers
if tag.caption is not None:
print(tag.caption.text)
sentiment = indicoio.sentiment_hq(tag.caption.text)
if sentiment >= 0.66:
self.positivePosts+=1
self.positiveY.append(sentiment*100)
self.positiveX.append(self.numPosts%(MAX_COUNT/3))
self.positiveRadius = self.calculateRadius(self.positiveRadius,followers)
elif sentiment >= 0.33:
self.neutralPosts+=1
self.neutralY.append(sentiment*100)
self.neutralX.append(self.numPosts%(MAX_COUNT/3))
self.neutralRadius = self.calculateRadius(self.neutralRadius,followers)
else:
self.negativePosts+=1
self.negativeY.append(sentiment*100)
self.negativeX.append(self.numPosts%(MAX_COUNT/3))
self.negativeRadius = self.calculateRadius(self.negativeRadius,followers)
#Use Indico API to calculate image sentiment
imageUrl = tag.images['low_resolution'].url
self.imageSentiment.append(indicoio.fer(imageUrl))
print # separate each post with a new line
self.numPosts+=1
def spam_filter(msg=input("Enter message = ")):
msg = TextBlob(msg)
current_lang = msg.detect_language()
print("Language of this message is = ", current_lang)
if (current_lang != 'en'):
msg.translate(to='en')
else:
msg.correct()
X_dtm = vect.fit_transform(X)
test_dtm = vect.transform([str(msg)])
model.fit(X_dtm, y)
result = model.predict(test_dtm)
prob = model.predict_proba(test_dtm)
if result == [1]:
print("SPAM ALERT!")
else:
print("HAM")
predsa = clf.predict(vectsa.transform([str(msg)]))
if predsa == [1]:
print("Positive Feeling")
elif predsa == [0]:
print("Negative Feeling")
else:
print("Can't analyze ur Felling...Try API ? ....")
senti = indicoio.sentiment_hq(str(msg))
print("Online Help , Positivity of Incoming Message = ", senti)
p = indicoio.personality(str(msg))
d = []
d.append([
p['agreeableness'], p['conscientiousness'], p['extraversion'],
p['openness'], msg.sentiment.polarity, msg.sentiment.subjectivity
])
traits = pd.DataFrame(d,
columns=[
'agreeableness', 'conscientiousness',
'extraversion', 'openness', 'polarity',
'subjectivity'
])
print(profanity.contains_profanity(str(msg)), " Profanity")
print(profanity.censor(str(msg)))
print("Summarizing this message =", msg.noun_phrases)
percent = pd.DataFrame(prob, columns=["% HAM", "%SPAM"])
print(traits)
print(percent)
def group_msgs(msgs, ts_grouped): track_idx = 0 msgs_grouped = [] sentiment_mapping = [] for ts_list in ts_grouped: move_up = track_idx + len(ts_list) grouping = msgs[track_idx:move_up] # hits a problem right here... single_str = ' '.join(str(elem) for elem in grouping) sentiment_rating = indicoio.sentiment_hq(single_str) timemap = (ts_list[0].date(), sentiment_rating) sentiment_mapping.append(timemap) msgs_grouped.append(grouping) track_idx = move_up return [sentiment_mapping, msgs_grouped]
def reddit_search(searchName):
searchName = request.form['searchName']
r = praw.Reddit(user_agent='sentimenter')
submissions = r.search(searchName)
score = []
commentList = []
for searchSubmissionInd, submission in enumerate(submissions):
submission.replace_more_comments(limit=1, threshold=1)
flat_comments = praw.helpers.flatten_tree(submission.comments)
for comment in flat_comments:
commentList.append(comment.body)
print("Grabbing comments from reddit thread: " + submission.short_link)
if (searchSubmissionInd +1 == SUBMISSION_SEARCH_LIMIT):
break
score = indicoio.sentiment_hq(commentList)
top_and_bottom(commentList, score)
score = statistics.mean(score)
return score
def search():
searchName = request.form['searchName']
reddit = request.form.get('reddit', default=False, type=bool)
twitter = request.form.get('twitter', default=False, type=bool)
r = praw.Reddit(user_agent='sentimenter')
submissions = r.search(searchName)
score = []
commentList = []
for searchSubmissionInd, submission in enumerate(submissions):
submission.replace_more_comments(limit=1, threshold=1)
flat_comments = praw.helpers.flatten_tree(submission.comments)
for comment in flat_comments:
commentList.append(comment.body)
print("Grabbing comments from reddit thread: " + submission.short_link)
if (searchSubmissionInd +1 == SUBMISSION_SEARCH_LIMIT):
break
score = statistics.mean(indicoio.sentiment_hq(commentList))
return jsonify(searchName = searchName, score = score)
def correr():
indicoio.config.api_key = settings.KEY_INDICOIO
while (1):
# En un principio fue necesario usar objetos Q para usar or, pero ya no lo es pues se modifico el codigo para que se haga el analisis de sentimiento
# y de localizacion a la vez
# tweets = Tweet.objects.filter(Q(sentimiento=None) | Q(mina=None))
tweets = Tweet.objects.filter(sentimiento=None)
for tweet in tqdm(tweets):
sent = indicoio.sentiment_hq(tweet.text)
tweet.sentimiento = Decimal(sent)
lugares = Lugar.objects.all()
for lugar in lugares:
# Autocritica: Se puede evitar el bucle for aprovechando los QuerySets de Django. Mucho mas veloces.
if lugar.nombre in tweet.text:
# Obviamente si hay mas de un lugar, se va a quedar con el ultimo. Este caso no se considerara por ser prototipo de Hackaton
tweet.mina = lugar
tweet.save()
time.sleep(10)
def correr():
indicoio.config.api_key = settings.KEY_INDICOIO
while(1):
# En un principio fue necesario usar objetos Q para usar or, pero ya no lo es pues se modifico el codigo para que se haga el analisis de sentimiento
# y de localizacion a la vez
# tweets = Tweet.objects.filter(Q(sentimiento=None) | Q(mina=None))
tweets = Tweet.objects.filter(sentimiento=None)
for tweet in tqdm(tweets):
sent = indicoio.sentiment_hq(tweet.text)
tweet.sentimiento = Decimal(sent)
lugares = Lugar.objects.all()
for lugar in lugares:
# Autocritica: Se puede evitar el bucle for aprovechando los QuerySets de Django. Mucho mas veloces.
if lugar.nombre in tweet.text:
# Obviamente si hay mas de un lugar, se va a quedar con el ultimo. Este caso no se considerara por ser prototipo de Hackaton
tweet.mina = lugar
tweet.save()
time.sleep(10)
def sentiment():
"""
Uses indico.io's Sentiment Analysis API to analyse each text in tih-data.json and
assign it a sentiment value.
Writes a csv file of the form:
Item,Category,Year,Sentiment
X: YYYY/MM/DD,$category,YYYY,sentiment_score
Where: sentiment_score is an integer between -100 and 100.
and category has been previously manually assigned in tih-data.json.
"""
data = read()
with open('sentiment.csv', 'a') as outfile:
outfile.write('Item,Category,Year,Sentiment\n')
for i in range(len(data)):
# Item in the format X: YYYY/MM/DD
s1 = "{}{}{}".format(i + 1, ": ", data[str(i)]["date"])
# Category e.g. Opinion, Advert, etc.
s2 = data[str(i)]["cat"]
# Year
s3 = data[str(i)]["date"][0:4]
# Sentiment
s4 = indicoio.sentiment_hq(data[str(i)]["article"]["text"])
s4 = s4 * 100
if s4 > 50:
pass
elif s4 < 50:
s4 = s4 - 100
s4 = int(round(s4, 0))
s = "{},{},{},{}".format(s1, s2, s3, s4)
outfile.write("{}{}".format(s, '\n'))
print("{} {}".format(s1, s4))
def sentiment(text_input):
#song lists created using radio billboard list and indicoio algorithm on song lyrics
list1 = ["Sorry", "What do you mean", "Same Old Love", "Here", "Hit The Quan", "Good For You"]
list2 = ["Focus", "Where Ya At", "Hello", "Locked Away", "Lean On", "Confident"]
list3 = ["Jumpman", "Trap Queen", "Renegades", "White Iverson", "Shut Up and Dance", "Hotline Bling"]
list4 = ["Drag Me Down", "Downtown", "Perfect", "Again", "Stitches", "679"]
list5 = ["My Way", "Ex's & Oh's", "Antidote", "Like I'm gonna lose you", "Tennessee Whiskey", "See You Again", "Watch Me", "On My Mind", "I'll Show You",
"Cheerleader", "The Hills", "Uptown Funk!", "Die a Happy Man", "How Deep Is Your Love", "Photograph", "Can't feel my face"]
indicoio.config.api_key = os.environ.get('INDICO')
#text_input = raw_input('Tell me about your day: ') #raw_input for python 2.7
sentiment = indicoio.sentiment_hq(text_input)
songTitle = ""
dayDescription = ""
print(sentiment)
#0<SUPERSAD<.25<SAD<.45<NEUTRAL<.6<HAPPY<.8<SUPERHAPPY<1
if sentiment < .15:
dayDescription = ("awful")
songTitle = (list1[(int)(random.random()*len(list1))])
else:
if sentiment < .35:
dayDescription = ("bad")
songTitle = (list2[(int)(random.random()*len(list2))])
else:
if sentiment < .65:
dayDescription = ("ok")
songTitle = (list3[(int)(random.random()*len(list3))])
else:
if sentiment <.9:
dayDescription = ("good")
songTitle = (list4[(int)(random.random()*len(list4))])
else:
dayDescription = ("awesome")
songTitle = (list5[(int)(random.random()*len(list5))])
values = [dayDescription, songTitle]
return values
def get_tweet_sentiment(self, tweet):
'''
Utility function to classify sentiment of passed tweet
using indicoio
'''
indicoio.config.api_key = 'e7cf4e703a29b6cdcecec19c5898185e'
# calculate sentiment score
myString = self.clean_tweet(tweet)
if not bool(myString and myString.strip()):
return 0.5
try:
sentiment_score = indicoio.sentiment_hq(myString)
except:
print('connection issue with indicoio')
return 0.5
if sentiment_score > 0.5:
sentiment = 'positive'
elif sentiment_score == 0.5:
sentiment = 'neutral'
else:
sentiment = 'negative'
print(myString, ",", sentiment)
return sentiment_score
def send_to_indico():
'''
This route handles the server's response when
you post data to localhost:5000/crunch through
the form on index.html
'''
tweets_csv_string = request.form.get('tweets')
csv_list = tweets_csv_string.replace('\r', '').splitlines()
if len(csv_list) > 40:
csv_list = csv_list[:40]
print csv_list
tweet_list = []
for csv_tweet in csv_list:
tweet_only = csv_tweet.split(',')[2:]
tweet_list.append(','.join(tweet_only))
tweet_list = tweet_list[::-1]
#tweet_scores = indicoio.batch_sentiment(tweet_list, api_key="428b1c1039ed8d8eaa886ee88044debd")
tweet_scores = indicoio.sentiment_hq(tweet_list, api_key="428b1c1039ed8d8eaa886ee88044debd")
return json.dumps({'scores': tweet_scores, 'tweets': tweet_list}) # dumps converts res to a JSON object
def main():
indicoio.config.api_key = '123273ff84fe220626891873d499ea07'
indicoio.config.language = 'russian'
# results:
#0.94399955814
#print indicoio.sentiment('хороший кот', language='russian')
#0.777086528524
#print indicoio.sentiment('постановление правительство', language='russian')
print indicoio.sentiment('хороший', language='russian')
print indicoio.sentiment('правительство', language='russian')
print indicoio.sentiment('кот', language='russian')
return
res = indicoio.sentiment_hq([
'хороший кот', 'постановление правительство',
'состоятельный оказаться', 'коррупционный правонарушение',
'конфликт интерес', 'первое квартал'
])
for r in res:
print r
def determine_comment_sentiment(submissions, submission_urls, comments_dict):
sentiment_comments = dict()
Scounter = 0
for submission in submissions:
# reading the lines
print("Finding sentiment value of " + submission_urls[submission])
multiLine = comments_dict[str(submission)]
total = 0
Scounter = Scounter + 1
counter = 0
for line in multiLine:
if len(line) > 125:
line = line[0:125]
total = total + indicoio.sentiment_hq(line)
counter = counter + 1
if counter != 0:
commentSentiment = total / counter
sentiment_comments[str(submission)] = commentSentiment
else:
sentiment_comments[str(submission)] = 0.5
for key in sentiment_comments:
print(key, sentiment_comments[key])
return sentiment_comments
import pandas as pd
import indicoio
import cProfile
from pprint import pprint
import csv
indicoio.config.api_key = "bbc6aca3c1564961ba2f0ee5dc81f32f"
df = pd.read_csv("data/hotel_reviews.csv")
indicoio.sentiment_hq(["indico is so easy to use!", "everything is awesome!"])
register = template.Library()
#song lists created using radio billboard list and indicoio algorithm on song lyrics
list1 = ["Sorry", "What do you mean", "Same Old Love", "Here", "Hit The Quan", "Good For You"]
list2 = ["Focus", "Where Ya At", "Hello", "Locked Away", "Lean On", "Confident"]
list3 = ["Jumpman", "Trap Queen", "Renegades", "White Iverson", "Shut Up and Dance", "Hotline Bling"]
list4 = ["Drag Me Down", "Downtown", "Perfect", "Again", "Stitches", "679"]
list5 = ["My Way", "Ex's & Oh's", "Antidote", "Like I'm gonna lose you", "Tennessee Whiskey", "See You Again", "Watch Me", "On My Mind", "I'll Show You",
"Cheerleader", "The Hills", "Uptown Funk!", "Die a Happy Man", "How Deep Is Your Love", "Photograph", "Can't feel my face"]
indicoio.config.api_key = 'a894b47f344116b3d32d12de039bf690'
#text_input = input('Tell me about your day: ')
sentiment = indicoio.sentiment_hq(text_input)
songTitle = ""
dayDescription = ""
print(sentiment)
#0<SUPERSAD<.25<SAD<.45<NEUTRAL<.6<HAPPY<.8<SUPERHAPPY<1
if sentiment < .15:
dayDescription = ("Your day was: AWFUL")
songTitle = (list1[(int)(random.random()*len(list1))])
else:
if sentiment < .35:
dayDescription = ("Your day was: BAD")
songTitle = (list2[(int)(random.random()*len(list2))])
else:
if sentiment < .65:
dayDescription = ("Your day was: OK")
songTitle = (list3[(int)(random.random()*len(list3))])
# print(linedict[submission])
# Reading the lines
sentimentComments = dict()
Scounter = 0
for submission in submissions:
# reading the lines
print("Finding sentiment value of " + submissionUrls[submission])
multiLine = linedict[str(submission)]
total = 0
Scounter = Scounter + 1
counter = 0
for line in multiLine:
if len(line) > 125:
line = line[0:125]
total = total + indicoio.sentiment_hq(line)
counter = counter + 1
if counter != 0:
commentSentiment = total / counter
sentimentComments[str(submission)] = commentSentiment
else:
sentimentComments[str(submission)] = 0.5
for key in sentimentComments:
print(key, sentimentComments[key])
f.close()
objectSent = dict()
objectOccur = dict()
objectIDs = dict()
def analyze_Sentiment_indico(self):
art = self.cleanText()
s = ''.join(ch for ch in art)
sentiment = indicoio.sentiment_hq(s)
return sentiment
def test_batch_sentiment_hq(self):
test_data = ['Worst song ever', 'Best song ever']
response = sentiment_hq(test_data, api_key=self.api_key)
self.assertTrue(isinstance(response, list))
self.assertTrue(response[0] < 0.5)
#import indicoio api configuration
import indicoio
indicoio.config.api_key = '580dc401830e72a98d4834bafe5b7d7c'
#use api to calculate sentiment analysis. sentiment is number 0-1
text_input = raw_input('enter feels here: ') #take user input
sentiment = (indicoio.sentiment_hq(text_input)) #use indico api to calculate input. save value as sentiment
file = open("sentimentnumber*.csv", "w")
file.write(str(sentiment))
file.close()
def crawlTwits(term):
# get authentication
api = TwitterAPI('1KxHa1hlPbRdsggvL5yfBgHPY', 'afQVw38uLX3Q1YdILflyG4FjWhjkMzXgSP9ypLee4LM4QIMOea',
'2786140432-npYkBKUXdUj3ulYj5f2N7LLN7dVJD6L6KdoyyLi',
'qwOdzFwyNfBKcmO6tq2TbOElrDHfd0ooiXNhMy4a7kUMd')
indicoio.config.api_key = 'e2637d8d80afb64412b3dda3dda64bdd'
# keep a counter to sum the sentiment score
scoreSum = 0
# keep a counter to sum the number of twits
twitsNum = 0
# keep a list of keywords
listKeyWords = ""
# search twits
r = api.request('search/tweets', {'q': term})
for item in r:
# filter out patterns
patterns = re.compile(', u\'text\': u\'(.*?)\', u\'is_quote_status\':')
if patterns is None:
patterns = re.compile(', u\'text\': u\"(.*?), u\'is_quote_status\':')
# search for patterns from twits
text = patterns.search(str(item))
# if found
if text:
# group into a text
twit = text.group(1)
# send twit to indico to get sentiment analyzed
sentimentNum = indicoio.sentiment_hq(twit)
# sent twit to indico to get keywords
json_keyWords = indicoio.keywords(twit)
# go through dict object
for key, value in json_keyWords.items():
# if the key is relevant enough
if value >= 0.2:
# add keywords to the list
listKeyWords += key + ", "
# add up score sum
scoreSum += sentimentNum
# increment number of twits
twitsNum += 1
# Uncomment lines below to debug
# print(twit)
# print(sentimentNum)
# if sentimentNum < 0.3:
# print("Negative")
# elif sentimentNum > 0.7:
# print("Positive")
# else:
# print("Neutral")
# print('\n')
# compute the average sentiment score
average = scoreSum / twitsNum
# get the evaluation
if average <= 0.2:
rate = "very negative"
elif average <= 0.4:
rate = "slightly negative"
elif average >= 0.8:
rate = "very positive"
elif average >= 0.6:
rate = "slightly positive"
else:
rate = "neutral"
# string to return
string = "an average score of " + str(average) + "\nOverall, it is " + str(rate) + "\nKeywords are " + listKeyWords
return string
import indicoio
access_key = "734176927825682433-UPc1vYiJjO0B4Ux6kj3GwRFXgIA22sK"
access_secret = "PpMlpJ412wx87g34msMRdir2yqfaiDTB6P0WqOdrlMxUw"
consumer_key = "jRsDidUota8wzQOin3AmKS312"
consumer_secret = "fRWBT5mK9EXWaZiIJCLdAO9FpbfU71eT4IHuv41OxazTJCeUdC"
indicoio.config.api_key = "070ef4588cd091fb23e3a6c9727097ce"
twitter = Twitter(auth = OAuth(access_key, access_secret, consumer_key, consumer_secret))
query = twitter.search.tweets(q = sys.argv[1], count = 15)
count = 0
total = len(query["statuses"])
ratingsum = 0;
strings = []
for result in query["statuses"]:
tweet = result["text"].encode(encoding='UTF-8',errors='strict')
weight = indicoio.sentiment_hq(tweet)
print(weight)
ratingsum += weight
string = "@%s | https://twitter.com/%s/status/%s | %s\n" % (result["user"]["screen_name"], result["user"]["screen_name"], result["id_str"], tweet)
if weight > 0.5:
strings.append(string);
count = count + 1;
print("The average favorability of tweets, with 1 being the most favorable, is %2f" % (ratingsum/total))
print("%s tweets with positive sentiment found that mention %s\n" % (count, sys.argv[1]))
print("twitter user | tweet link | tweet message")
for index in range(0, count):
print(strings[index])
import indicoio as ind from sys import stdin ind.config.api_key = 'b88a14d4a97b56a6ed8f65efee05f9c4' str = stdin.read() print(ind.sentiment_hq(str))
import indicoio
indicoio.config.api_key = ''
import numpy
import matplotlib.pyplot as plt
#read the file line by line
with open(filename) as f:
lines = f.readlines()
#send request to Indico SA API and store the scores in a list
sent_scores=[]
for l in lines:
sent_scores.append(indicoio.sentiment_hq(l))
#print "All scores %s:" %sent_scores
#map score ranges to categories(arbitrarily assigned here)
pos_total=0
neutral_total=0
negative_total=0
for l in lines:
if indicoio.sentiment_hq(l)<0.3:
negative_total+=1
elif indicoio.sentiment_hq(l)>=0.3 and indicoio.sentiment_hq(l)<0.7:
neutral_total+=1
else:
def test_batch_size(self):
test_data = ["Terribly interesting test data."] * 100
response = sentiment_hq(test_data, batch_size=20)
self.assertTrue(isinstance(response, list))
self.assertTrue(all([isinstance(el, (float, np.float32)) for el in response]))
import indicoio as ind from sys import stdin ind.config.api_key = 'b88a14d4a97b56a6ed8f65efee05f9c4' happy_thresh = 0.75 sad_thresh = 0.25 str = stdin.read() val = ind.sentiment_hq(str) print(val)
import indicoio
indicoio.config.api_key = '580dc401830e72a98d4834bafe5b7d7c'
text_input = raw_input('enter feels here: ')
print text_input
print(indicoio.sentiment_hq(text_input))
