def TextBlobCleanEmoji():
'''
TextBlob model with Emoticon scoring.
'''
tweet_counter = 0
with open("results_textblob_emoji.txt", "w",
encoding="utf-8") as preresults:
newWriter = csv.writer(preresults,
delimiter='\t',
quotechar='|',
quoting=csv.QUOTE_MINIMAL)
with open("raw_twitter.txt", "r", encoding="utf-8") as preproccessed:
for line in preproccessed.readlines():
tweet_counter += 1
try:
print("Processing tweet: {}".format(tweet_counter))
tweet = tweetCleaner.lowercase(line)
tweet = tweetCleaner.StopWordRemover(tweet)
tweet = tweetCleaner.removeSpecialChars(tweet)
tweet, score = tweetProcesser.emoticon_score(tweet)
tweet = tweetCleaner.removeAllNonAlpha(tweet)
tweet = tweetCleaner.lemmatizer(tweet)
wiki = TextBlob(tweet)
normalized_score, sentiment_label = tweetProcesser.sentimentClassifier(
wiki, score)
newWriter.writerow(
[normalized_score, sentiment_label, tweet])
except:
newWriter.writerow(["0", "neutral", "ERROR"])
print("ERROR processing tweet: {}".format(tweet_counter))
def TextBlobCleanAbbrev():
'''
TextBlob model with abbreviations extended.
'''
tweet_counter = 0
tweetProcesser.abbreviation_extender()
with open("results_textblob_abbrev.txt", "w",
encoding="utf-8") as postresults:
newWriter = csv.writer(postresults,
delimiter='\t',
quotechar='|',
quoting=csv.QUOTE_MINIMAL)
with open("abbreviations_twitter.txt", "r",
encoding="utf-8") as postprocessed:
for line in postprocessed.readlines():
tweet_counter += 1
try:
print("Processing tweet: {}".format(tweet_counter))
tweet = tweetCleaner.StopWordRemover(line)
tweet = tweetCleaner.removeSpecialChars(tweet)
tweet = tweetCleaner.removeAllNonAlpha(tweet)
tweet = tweetCleaner.lemmatizer(tweet)
wiki = TextBlob(tweet)
normalized_score, sentiment_label = tweetProcesser.sentimentClassifier(
wiki, 0)
newWriter.writerow(
[normalized_score, sentiment_label, tweet])
except:
newWriter.writerow(["0", "neutral", "BLANK"])
print("ERROR processing tweet: {}".format(tweet_counter))
def NLTKCleanAbbrevEmoji():
"""
NLTK model with extended abbreviations AND emoticon scoring
"""
tweet_counter = 0
tweetProcesser.abbreviation_extender()
with open("results_nltk_abbrev_emoji.txt", "w",
encoding="utf-8") as postresults:
newWriter = csv.writer(postresults,
delimiter='\t',
quotechar='|',
quoting=csv.QUOTE_MINIMAL)
with open("abbreviations_twitter.txt", "r",
encoding="utf-8") as postprocessed:
for line in postprocessed.readlines():
total_score = 0
tweet_counter += 1
try:
print("Processing tweet: {}".format(tweet_counter))
tweet = tweetCleaner.lowercase(line)
tweet = tweetCleaner.StopWordRemover(tweet)
tweet = tweetCleaner.removeSpecialChars(tweet)
tweet, total_score = tweetProcesser.emoticon_score(tweet)
tweet = tweetCleaner.removeAllNonAlpha(tweet)
tweet = tweetCleaner.lemmatizer(tweet)
lines_list = tokenize.sent_tokenize(tweet)
for line in lines_list:
ss = sentiment.polarity_scores(line)
total_score -= ss["neg"]
total_score += ss["pos"]
total_score = round(total_score, 3)
if total_score == 0:
newWriter.writerow([0, "neutral"])
elif total_score > 0:
newWriter.writerow([total_score, "positive"])
else:
newWriter.writerow([total_score, "negative"])
except:
newWriter.writerow([0, "neutral"])
print("ERROR processing tweet: {}".format(tweet_counter))
def NLTKCleanRaw():
'''
Raw NLTK model
'''
tweet_counter = 0
with open("results_nltk_raw.txt", "w", encoding="utf-8") as postresults:
newWriter = csv.writer(postresults,
delimiter='\t',
quotechar='|',
quoting=csv.QUOTE_MINIMAL)
with open("raw_twitter.txt", "r", encoding="utf-8") as postprocessed:
for line in postprocessed.readlines():
total_score = 0
tweet_counter += 1
try:
print("Processing tweet: {}".format(tweet_counter))
tweet = tweetCleaner.lowercase(line)
tweet = tweetCleaner.StopWordRemover(tweet)
tweet = tweetCleaner.removeSpecialChars(tweet)
tweet = tweetCleaner.removeAllNonAlpha(tweet)
tweet = tweetCleaner.lemmatizer(tweet)
lines_list = tokenize.sent_tokenize(tweet)
for sentence in lines_list:
ss = sentiment.polarity_scores(sentence)
total_score -= ss["neg"]
total_score += ss["pos"]
total_score = round(total_score, 3)
if total_score == 0:
newWriter.writerow([0, "neutral"])
elif total_score > 0:
newWriter.writerow([total_score, "positive"])
else:
newWriter.writerow([total_score, "negative"])
except:
newWriter.writerow([0, "neutral"])
print("ERROR processing tweet: {}".format(tweet_counter))
