def Insert_Donnees_AVI() :
con = psycopg2.connect(database="BASE_CURATED_ZONE", user="******", password="******", host="127.0.0.1", port="5433")
cur = con.cursor()
myFilePathName = myPathRoot_CURRATEDZONE + "AVI.txt"
myFilePtr = open(myFilePathName, "r", encoding="utf-8", errors="ignore")
myFileContents = myFilePtr.readlines()
del myFileContents[0]
for myLineRead in myFileContents:
line = myLineRead.split(";")
cle_unique = int(line[0])
emplacement_source=line[1]
datetime_ingestion=line[2]
privacy_level=line[3]
entreprise=line[4]
if line[5] == 'NULL' :
date = 'May 24, 2020'
else :
date = line[5]
review_titre=TextBlob(line[6], pos_tagger = PatternTagger(), analyzer= PatternAnalyzer()).sentiment[0]
status_employe=line[7]
lieu=line[8]
recommande=line[9]
commentaire=TextBlob(line[10], pos_tagger = PatternTagger(), analyzer= PatternAnalyzer()).sentiment[0]
avantage= line[11].lower()
incovenient= line[12].lower()
cur.execute("INSERT INTO AVIS VALUES (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s)",
(cle_unique,emplacement_source,datetime_ingestion,privacy_level,entreprise,date,review_titre,status_employe,lieu,recommande,commentaire,avantage,incovenient))
myFilePtr.close()
con.commit()
con.close()
return (True)
def textblob_sentiment_fr(tweets):
""" Get Textblob Sentiment scores (fr) """
tb_fr = []
for tweet in tweets:
blob_tweet_fr = TextBlob(tweet, pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
tb_fr.append(blob_tweet_fr.sentiment[0])
return tb_fr
def sentiment_analysis_basic(tweets, lang):
positive_tweets = 0
neutral_tweets = 0
negative_tweets = 0
for tweet in tweets:
if lang == 'english':
analysis = TextBlob(tweet)
sentiment = analysis.sentiment.polarity
else: # french
analysis = TextBlob(tweet, pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
sentiment = analysis.sentiment[0]
if sentiment > 0:
positive_tweets += 1
elif sentiment == 0:
neutral_tweets += 1
else:
negative_tweets += 1
total_tweets_analysed = positive_tweets + neutral_tweets + negative_tweets
positive_tweets_percentage = positive_tweets / total_tweets_analysed * 100
neutral_tweets_percentage = neutral_tweets / total_tweets_analysed * 100
print("\nNo. of positive tweets = {} Percentage = {}".format(
positive_tweets, positive_tweets_percentage))
print("No. of neutral tweets = {} Percentage = {}".format(
neutral_tweets, neutral_tweets_percentage))
print("No. of negative tweets = {} Percentage = {}".format(
negative_tweets, 100 - (positive_tweets_percentage + neutral_tweets_percentage)))
def match_syntagm_text_blob_multi(syntagms, text):
from textblob import TextBlob
from textblob_fr import PatternTagger, PatternAnalyzer
blob = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
return match_sequences(syntagms, list(blob.tokenize()))
def querytextanswers(self) -> List[TextAnswer]:
res = list()
form_ids = [f.id for f in self.forms_list]
student_ids = [s.id for s in self.students_list]
questions = [
q for q in self.questions_list if q.isint == Const.DBFALSE
]
question_ids = [q.id for q in questions]
answers = db.session.query(Answer).filter(
Answer.form_id.in_(form_ids), Answer.student_id.in_(student_ids),
Answer.question_id.in_(question_ids)).order_by(Answer.question_id)
#
# see example https://github.com/sloria/textblob-fr
for question in questions:
polvalues = list()
subvalues = list()
texts = [
answer.text
for answer in answers.filter(Answer.question_id == question.id)
]
for text in texts:
blob = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
qsentiment = blob.sentiment
polvalues.append(qsentiment[0])
subvalues.append(qsentiment[1])
res.append(
TextAnswer(questiontext=question.text,
polarities=polvalues,
subjectivities=subvalues))
return res
def analyse(comments):
allcomments = []
polarity = []
for comment in comments:
try:
allcomments.append(comment)
try:
if detect(comment) == 'de':
text = TextBlobDE(comment)
x = text.sentiment.polarity
polarity.append(x)
elif detect(comment) == 'fr':
blob = TextBlob(comment,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
x = blob.sentiment[0]
polarity.append(x)
else:
text = TextBlob(comment)
x = text.sentiment.polarity
polarity.append(x)
except:
text = TextBlob(comment)
x = text.sentiment.polarity
polarity.append(x)
except:
pass
return allcomments, polarity
def extract_keywords(txt):
""" Extract keywords from FR text"""
blob = TextBlob(any2utf8(txt),
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
tags = blob.tags
return [tag for tag in tags]
def saveSentiments():
"""
opinion minning average for one day
save data in CSV file with date
"""
data = pd.read_csv("input/" + str(datetime.utcnow().date().today()) +
".csv")
f = open("output/sentimentAnalysis.csv", "a", newline='', encoding='utf-8')
writer = csv.writer(f, delimiter=',')
polarityAverage = []
subjectivityAverage = []
for i in range(len(data.index)):
blob = TextBlob(cleanText(data.iloc[i, 1]),
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
polarityAverage.append(blob.sentiment[0])
subjectivityAverage.append(blob.sentiment[1])
sum = 0
for ele in polarityAverage:
sum += ele
polarityAverage = sum / len(polarityAverage)
for ele in subjectivityAverage:
sum += ele
subjectivityAverage = sum / len(subjectivityAverage)
writer.writerow([
str(datetime.utcnow().date().today()),
len(data.index), polarityAverage, subjectivityAverage
])
f.close()
def extract_dictionary(txt):
""" Extract from FR text"""
blob = TextBlob(any2utf8(txt),
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
seg_list = blob.words
return list(seg_list)
def get_polarity(topic):
#Step 2 - Prepare query features
#List of candidates to French Republicans Primary Elections
candidates_names = ['Clinton', 'Trump']
#Hashtag related to the debate
name_of_debate = str(topic)
#Date of the debate : October 13th
since_date = "2016-10-01"
until_date = "2016-11-06"
#Step 3 - Retrieve Tweets and Save Them
all_polarities = dict()
for candidate in candidates_names:
this_candidate_polarities = []
#Get the tweets about the debate and the candidate between the dates
this_candidate_tweets = api.search(q=[name_of_debate, candidate], count=100, since = since_date, until=until_date)
#Save the tweets in csv
with open('%s_tweets.csv' % candidate, 'wb') as this_candidate_file:
this_candidate_file.write('tweet,sentiment_label\n')
for tweet in this_candidate_tweets:
analysis = TextBlob(tweet.text, pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
#Get the label corresponding to the sentiment analysis
this_candidate_polarities.append(analysis.sentiment[0])
this_candidate_file.write('%s,%s\n' % (tweet.text.encode('utf8'), get_label(analysis)))
#Save the mean for final results
all_polarities[candidate] = np.mean(this_candidate_polarities)
#Step bonus - Print a Result
sorted_analysis = sorted(all_polarities.items(), key=operator.itemgetter(1), reverse=True)
print 'Mean Sentiment Polarity in descending order :'
for candidate, polarity in sorted_analysis:
print '%s : %0.3f' % (candidate, polarity)
class TestPatternAnalyzer(unittest.TestCase):
def setUp(self):
self.analyzer = FrAnalyzer()
self.neg = u"C'est une voiture terribles."
self.pos = u"Quelle belle matinée!"
def test_analyze(self):
pos_sentiment = self.analyzer.analyze(self.pos)
assert_true(pos_sentiment[0] > 0.0)
neg_sentiment = self.analyzer.analyze(self.neg)
assert_true(neg_sentiment[0] < 0.0)
def test_blob_analyze(self):
pos_blob = TextBlob(self.pos, analyzer=self.analyzer)
assert_true(pos_blob.sentiment[0] > 0.0)
neg_blob = TextBlob(self.neg, analyzer=self.analyzer)
assert_true(neg_blob.sentiment[0] < 0.0)
def analize_sentiment_fr(tweet):
analysis = TextBlob(clean_tweet(tweet),pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
#print(analysis)
if float(analysis.sentiment[0]) > 0:
return 1
elif float(analysis.sentiment[0]) == 0:
return 0
else:
return -1
def get_usefull_words(text): blob = TextBlob(text, pos_tagger=PatternTagger(), analyzer=PatternAnalyzer()) blob = blob.correct() print(blob) print(blob.noun_phrases) for tag in blob.tags: if tag[1] in ['NN', 'JJ', 'VB']: print(tag[0],tag[1])
class TestPatternAnalyzer(unittest.TestCase):
def setUp(self):
self.analyzer = FrAnalyzer()
self.neg = u"C'est une voiture terribles."
self.pos = u"Quelle belle matinée!"
def test_analyze(self):
pos_sentiment = self.analyzer.analyze(self.pos)
assert_true(pos_sentiment[0] > 0.0)
neg_sentiment = self.analyzer.analyze(self.neg)
assert_true(neg_sentiment[0] < 0.0)
def test_blob_analyze(self):
pos_blob = TextBlob(self.pos, analyzer=self.analyzer)
assert_true(pos_blob.sentiment[0] > 0.0)
neg_blob = TextBlob(self.neg, analyzer=self.analyzer)
assert_true(neg_blob.sentiment[0] < 0.0)
def sentiment_analyser(text):
blob = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
sentiments = blob.sentiment
# on renvoie le résultat
return sentiments
def getSentimentForReview(review):
#print(review)
blob = TextBlob(review,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
sentiment = blob.sentiment
#print(sentiment)
#print()
return sentiment
def run(self):
Myanalyzer = SentimentalAnalysis
print("Analyzer begin")
#
#consumer = KafkaConsumer(bootstrap_servers='localhost:9092','37.163.95.205',
consumer = KafkaConsumer(bootstrap_servers='localhost:9092',
auto_offset_reset='earliest',
consumer_timeout_ms=10000)
consumer.subscribe(['my-topic'])
print("Subscription analyzer: OK")
moviesList = []
for message in consumer:
if allocine:
jsoned = json.loads(message.value)
#note = Myanalyzer.analysis(jsoned['spectators_reviews'])
from textblob import TextBlob
from textblob_fr import PatternTagger, PatternAnalyzer
print(jsoned)
if (jsoned['spectators_reviews']):
note = 0
nbNote = 1
for elem in jsoned['spectators_reviews']:
blob = TextBlob(elem,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
note += blob.sentiment[0]
nbNote += 1
note /= nbNote
# print("Note obtenu ", note)
jsoned['ownRating'] = note
moviesList.append(jsoned)
else:
jsoned = json.loads(message.value)
#moviesList.append(jsoned['review'])
note = Myanalyzer.analysis(jsoned['review'])
if (note):
jsoned['ownRating'] = note
else:
jsoned['ownRating'] = 0
moviesList.append(jsoned)
producer = KafkaProducer(bootstrap_servers='localhost:9092')
for data in moviesList:
if (debug):
print("dataAnalyzer to send ", data)
data[
'spectators_reviews'] = "" # Supprime les reviews pour pas s'embeter
producer.send("my-ratings", json.dumps(data))
print("Sent")
producer.flush()
producer.close()
print("DataAnalyse produced")
def __init__(self):
self.french_stop_words = get_stop_words('french')
# TODO Acev SpaCy, quelle différence entre "md" et "sm" ??
self.nlp = spacy.load(
'fr_core_news_md'
) # Utilisé par SapCy pour la Lemmatisation et Stemmatisation
self.blob = Blobber(
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer()) # Analyse de sentiments
def textblob_sentiment(tweets):
""" Get Textblob Sentiment scores (en/fr) """
tb = []
tb_fr = []
for tweet in tweets:
blob_tweet = TextBlob(tweet)
tb.append(blob_tweet.sentiment.polarity)
blob_tweet_fr = TextBlob(
tweet, pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
tb_fr.append(blob_tweet_fr.sentiment[0])
return tb, tb_fr
def is_neutral(self, text):
btext = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
if btext.sentiment[0] == 0:
if (self.bayes_clf.predict([text])[0] == 0) or (self.svm_clf.predict([text])[0] == 0):
return True
else:
if (self.bayes_clf.predict([text])[0] == 0) and (self.svm_clf.predict([text])[0] == 0):
return True
return False
def sentiment_analysis(name, dictionary):
"""
This function takes a file and creates a dictionary of each line's sentiment analysis.
>>> sentiment_analysis('EmmanuelMacron', {})
{'EmmanuelMacron': [0.1466666666666667, 0.0, -0.1, 0.0, 0.42000000000000004, 0.0, 0.115, 0.0, 0.1325, 0.0, 0.03333333333333333, 0.0, 0.27, -0.12, 0.0, 0.22, 0.27, 0.1, 0.15, 0.075, 0.0, 0.0, 0.0, 0.17, 0.0, 0.07666666666666666, 0.2, 0.0, 0.0, 0.2, 0.2525, -0.35, 0.0, 0.0, 0.1, 0.0, 0.15, 0.0, 0.0, 0.56, 0.0, 0.25, 0.22, 0.0, 0.0, 0.45, 0.0, 0.0, 0.023333333333333334, 0.025000000000000022, 0.0, 0.0, -0.125, 0.0, 0.0, 0.0, 0.15, 0.13666666666666666, 0.1, 0.11, 0.0, 0.0, -0.4, 0.0, 0.0, 0.2, 0.625, 0.0, 0.0, 0.0, 0.09999999999999999, 0.0, 0.05, 0.25, 0.0, 0.0, 0.0, 0.22, 0.0, 0.22, 0.22, 0.53, -0.15, 0.0, 0.0, 0.4, 0.0, 0.0, 0.009999999999999995, 0.0, 0.0, -0.016666666666666663, 0.1, 0.0, 0.15, 0.0, 0.1, 0.0, -0.25, 0.0, -0.25166666666666665, 0.22, 0.17, 0.0, 0.0, -0.7, 0.0, 0.22, 0.22, 0.0, 0.2, 0.0, 0.0, 0.0, 0.13, 0.17, 0.0, 0.1275, 0.0, 0.0, 0.1, 0.15, -0.16249999999999998, 0.1, 0.8, 0.14, 0.0, 0.0, -0.1, 0.0, 0.0, 0.0, 0.30833333333333335, 0.0, 0.185, 0.0, 0.0, 0.0, -0.09000000000000001, 0.0, 0.08, -0.75, 0.22, 0.0, -0.3, 0.21000000000000002, 0.010000000000000009, -0.03125, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.17500000000000002, 0.3499999999999999, 0.09833333333333334, 0.135, 0.0, 0.0, 0.08, 0.2, 0.0, -0.2, 0.0, 0.2233333333333333, 0.0, 0.29, 0.0, 0.0, 0.0, 0.0, 0.6625000000000001, 0.29, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.32, 0.4, -0.24, 0.0, -0.125, 0.15, 0.0, 0.7, 0.0, 0.22, 0.0, 0.0, 0.5, 0.0, 0.2, -0.21875, 0.25, 0.26, 0.185, 0.08333333333333333, 0.23]}
"""
l = open(name + '.txt')
lines = l.readlines()
dictionary[name] = []
for i in lines:
blob = TextBlob(i, pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
dictionary[name].append(blob.sentiment[0])
return(dictionary)
def labellize_sentence(self, sentence, model):
"""Use the CRF to labellize some string"""
# NOW, TAG THAT
blob = TextBlob(sentence,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
encoded_sentence = self.sent2features(blob.tags)
print()
LABELS = model.predict([encoded_sentence])
for token, label in zip(blob.tags, LABELS[0]):
print(token, label)
def get_tweet_sentiment(self, row):
analysis = TextBlob(str(row),
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
strAnalysis = str(analysis.sentiment)
sentiment = strAnalysis[strAnalysis.find('(') +
1:strAnalysis.find(',')]
if float(sentiment) > 0:
return 'positive'
elif float(sentiment) == 0:
return 'neutral'
else:
return 'negative'
def sentiments_cava(msg_user):
happy_bot = ["Super!", "Content que tout aille bien!"]
sad_bot = [
"Désolé d'entendre ça",
"Je voudrais te consoler mais je suis trop basique pour ça."
]
ok_bot = ["Ok...", "Bon..."]
blob = TextBlob(msg_user,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
if blob.sentiment[0] >= 0.2:
return random.choice(happy_bot)
if blob.sentiment[0] <= -0.2:
return random.choice(sad_bot)
else:
return random.choice(ok_bot)
def detect_sentiment(text, lang):
if lang == 'fr':
sentiment = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer()).sentiment
polarity = sentiment[0]
else:
sentiment = TextBlob(text).sentiment
polarity = sentiment.polarity
if polarity > 0.3:
text_sentiment = 'Positive'
elif polarity < -0.3:
text_sentiment = 'Negative'
else:
text_sentiment = 'Neutral or Undefined'
return text_sentiment
def convert_training_set(self):
"""From STandfor format to scikit-learn crfsuite one"""
training_set_folder = '/home/emeric/1_Github/RecipeAnalyzor/models/training_set_crf'
crfsuite_annotation = []
file_list = os.listdir(training_set_folder)
for filename in file_list:
if filename.endswith('.ann'):
root_name = re.findall('([0-9]{1,10}).ann', filename)[0]
with open('{}/{}.txt'.format(training_set_folder, root_name),
'r') as handler:
text = handler.read()
with open('{}/{}.ann'.format(training_set_folder, root_name),
'r') as handler:
annotations = [
ann.strip('\n').split('\t')
for ann in handler.readlines()
]
blob = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
file_annotation = []
tagged_tokens = [
x[2] for x in annotations if x[0].startswith('T')
]
#~ print(tagged_tokens)
for token in blob.tags:
# NOW TAG THE SHIT
if token[0] in tagged_tokens:
tag = list(
set([
x[1].split(' ')[0] for x in annotations
if x[0].startswith('T') and x[2] == token[0]
]))[0]
else:
tag = '0'
file_annotation.append((token[0], token[1], tag))
crfsuite_annotation.append(file_annotation)
return crfsuite_annotation
def get_sentiment_from_sentence(sentence):
tb = Blobber(pos_tagger=PatternTagger(), analyzer=PatternAnalyzer())
input_text = sentence.line
blob = tb(input_text)
pola, subj = blob.sentiment
perc = f"{100*abs(pola):.0f}"
if pola > 0:
sent = "positive"
elif pola < 0:
sent = "negative"
else:
sent = "neutral"
sentiment = {'pola': pola, 'perc': perc, 'sent': sent}
return sentiment
def blob_sentiment(text, Polarity=True, Subjectivity=True):
# Utilise un fichier xml ou il y a 5300 mot ayant une polarité et une subjectivité
# Le score de polarité est un float entre [-1,1] quand c'est égale à -1 l'opinion est négative et 1 l'opinion est positive
# Le score de subjectivité est un float entre [0,1] quand c'est égale à 0 le texte est objectif et quand c'est égale 1 le texte est subjectif
blob = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
if Polarity == True and Subjectivity == True:
return blob.sentiment[0], blob.sentiment[1]
elif Polarity == True and Subjectivity == False:
return blob.sentiment[0]
elif Polarity == False and Subjectivity == True:
return blob.sentiment[1]
else:
return None
def polarity(list_phrase, data):
polarite = []
for i in range(len(list_phrase)):
if list_phrase[i] == '':
polarite.append('vide')
else:
text = list_phrase[i]
blob = TextBlob(text,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
if (blob.sentiment[0] < 0.15 and blob.sentiment[0] > -0.15):
polarite.append(['neutre', data[i]['date']])
elif (blob.sentiment[0] > 0.15):
polarite.append(['positif', data[i]['date']])
else:
polarite.append(['negatif', data[i]['date']])
return (polarite)
def polarizer_text_blob_french(df_Verbatim):
"""
Method which generate a list of categories of sentiments (neutral, positive, negative) given a dataframe of textual comments in french
Input : a dataframe with only one column of strings
Output : a list of strings
"""
sentiment_scores_tb = [
round(
TextBlob(article,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer()).sentiment[0], 3)
for article in df_Verbatim
]
sentiment_category_tb = [
'positive' if score > 0 else 'negative' if score < 0 else 'neutral'
for score in sentiment_scores_tb
]
return (sentiment_category_tb)
def data(file):
title = file.iloc[:, 0].dropna().tolist()
comment = file.iloc[:, 0].dropna().tolist()
data = [title, comment]
pos = []
sub = []
for item in data:
for phrase in item:
blob = TextBlob(phrase,
pos_tagger=PatternTagger(),
analyzer=PatternAnalyzer())
sen = blob.sentiment
pos.append(sen[0])
sub.append(sen[1])
print(phrase)
print(sen)
def setUp(self):
self.analyzer = FrAnalyzer()
self.neg = u"C'est une voiture terribles."
self.pos = u"Quelle belle matinée!"
