mirror of
https://github.com/gsi-upm/senpy
synced 2024-11-13 04:02:29 +00:00
Added emotion-anew and sentiment-vader
This commit is contained in:
parent
23a5595d18
commit
23d15e9274
35
emotion-anew/README.md
Executable file
35
emotion-anew/README.md
Executable file
@ -0,0 +1,35 @@
|
||||
|
||||
#Plugin emotion-anew
|
||||
|
||||
This plugin consists on an **emotion classifier** that detects six possible emotions:
|
||||
- Anger : general-dislike.
|
||||
- Fear : negative-fear.
|
||||
- Disgust : shame.
|
||||
- Joy : gratitude, affective, enthusiasm, love, joy, liking.
|
||||
- Sadness : ingrattitude, daze, humlity, compassion, despair, anxiety, sadness.
|
||||
- Neutral: not detected a particulary emotion.
|
||||
|
||||
The plugin uses **ANEW lexicon** dictionary to calculate VAD (valence-arousal-dominance) of the sentence and determinate which emotion is closer to this value. To do this comparision, it is defined that each emotion has a centroid, calculated according to this article: http://www.aclweb.org/anthology/W10-0208.
|
||||
|
||||
The plugin is going to look for the words in the sentence that appear in the ANEW dictionary and calculate the average VAD score for the sentence. Once this score is calculated, it is going to seek the emotion that is closest to this value.
|
||||
|
||||
The response of this plugin uses [Onyx ontology](https://www.gsi.dit.upm.es/ontologies/onyx/) developed at GSI UPM, to express the information.
|
||||
|
||||
##Usage
|
||||
|
||||
Params accepted:
|
||||
- Language: English (en) and Spanish (es).
|
||||
- Input: input text to analyse.
|
||||
|
||||
|
||||
Example request:
|
||||
```
|
||||
http://senpy.cluster.gsi.dit.upm.es/api/?algo=emotion-anew&language=en&input=I%20love%20Madrid
|
||||
```
|
||||
|
||||
Example respond: This plugin follows the standard for the senpy plugin response. For more information, please visit [senpy documentation](http://senpy.readthedocs.io). Specifically, NIF API section.
|
||||
|
||||
![alt GSI Logo][logoGSI]
|
||||
|
||||
[logoES]: https://www.gsi.dit.upm.es/ontologies/onyx/img/eurosentiment_logo.png "EuroSentiment logo"
|
||||
[logoGSI]: http://www.gsi.dit.upm.es/images/stories/logos/gsi.png "GSI Logo"
|
155
emotion-anew/emotion-anew.py
Normal file
155
emotion-anew/emotion-anew.py
Normal file
@ -0,0 +1,155 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
|
||||
import re
|
||||
import nltk
|
||||
import csv
|
||||
import sys
|
||||
import os
|
||||
import unicodedata
|
||||
import string
|
||||
import xml.etree.ElementTree as ET
|
||||
import math
|
||||
|
||||
from sklearn.svm import LinearSVC
|
||||
from sklearn.feature_extraction import DictVectorizer
|
||||
|
||||
from nltk import bigrams
|
||||
from nltk import trigrams
|
||||
from nltk.corpus import stopwords
|
||||
|
||||
from pattern.en import parse as parse_en
|
||||
from pattern.es import parse as parse_es
|
||||
from senpy.plugins import SentimentPlugin, SenpyPlugin
|
||||
from senpy.models import Results, EmotionSet, Entry, Emotion
|
||||
|
||||
|
||||
class EmotionTextPlugin(SentimentPlugin):
|
||||
|
||||
def activate(self, *args, **kwargs):
|
||||
self._stopwords = stopwords.words('english')
|
||||
self._local_path=os.path.dirname(os.path.abspath(__file__))
|
||||
|
||||
def _my_preprocessor(self, text):
|
||||
|
||||
regHttp = re.compile('(http://)[a-zA-Z0-9]*.[a-zA-Z0-9/]*(.[a-zA-Z0-9]*)?')
|
||||
regHttps = re.compile('(https://)[a-zA-Z0-9]*.[a-zA-Z0-9/]*(.[a-zA-Z0-9]*)?')
|
||||
regAt = re.compile('@([a-zA-Z0-9]*[*_/&%#@$]*)*[a-zA-Z0-9]*')
|
||||
text = re.sub(regHttp, '', text)
|
||||
text = re.sub(regAt, '', text)
|
||||
text = re.sub('RT : ', '', text)
|
||||
text = re.sub(regHttps, '', text)
|
||||
text = re.sub('[0-9]', '', text)
|
||||
text = self._delete_punctuation(text)
|
||||
return text
|
||||
|
||||
def _delete_punctuation(self, text):
|
||||
|
||||
exclude = set(string.punctuation)
|
||||
s = ''.join(ch for ch in text if ch not in exclude)
|
||||
return s
|
||||
|
||||
def _extract_ngrams(self, text, lang):
|
||||
|
||||
unigrams_lemmas = []
|
||||
unigrams_words = []
|
||||
pos_tagged = []
|
||||
if lang == 'es':
|
||||
sentences = parse_es(text,lemmata=True).split()
|
||||
else:
|
||||
sentences = parse_en(text,lemmata=True).split()
|
||||
|
||||
for sentence in sentences:
|
||||
for token in sentence:
|
||||
if token[0].lower() not in self._stopwords:
|
||||
unigrams_words.append(token[0].lower())
|
||||
unigrams_lemmas.append(token[4])
|
||||
pos_tagged.append(token[1])
|
||||
|
||||
return unigrams_lemmas,unigrams_words,pos_tagged
|
||||
|
||||
def _find_ngrams(self, input_list, n):
|
||||
return zip(*[input_list[i:] for i in range(n)])
|
||||
|
||||
def _emotion_calculate(self, VAD):
|
||||
emotion=''
|
||||
value=10000000000000000000000.0
|
||||
for state in self.centroids:
|
||||
valence=VAD[0]-self.centroids[state]['V']
|
||||
arousal=VAD[1]-self.centroids[state]['A']
|
||||
dominance=VAD[2]-self.centroids[state]['D']
|
||||
new_value=math.sqrt((valence*valence)+(arousal*arousal)+(dominance*dominance))
|
||||
if new_value < value:
|
||||
value=new_value
|
||||
emotion=state
|
||||
return emotion
|
||||
|
||||
def _extract_features(self, tweet,dictionary,lang):
|
||||
feature_set={}
|
||||
ngrams_lemmas,ngrams_words,pos_tagged = self._extract_ngrams(tweet,lang)
|
||||
pos_tags={'NN':'NN', 'NNS':'NN', 'JJ':'JJ', 'JJR':'JJ', 'JJS':'JJ', 'RB':'RB', 'RBR':'RB',
|
||||
'RBS':'RB', 'VB':'VB', 'VBD':'VB', 'VGB':'VB', 'VBN':'VB', 'VBP':'VB', 'VBZ':'VB'}
|
||||
totalVAD=[0,0,0]
|
||||
matches=0
|
||||
for word in range(len(ngrams_lemmas)):
|
||||
VAD=[]
|
||||
if ngrams_lemmas[word] in dictionary:
|
||||
matches+=1
|
||||
totalVAD = [totalVAD[0]+float(dictionary[ngrams_lemmas[word]]['V']),
|
||||
totalVAD[1]+float(dictionary[ngrams_lemmas[word]]['A']),
|
||||
totalVAD[2]+float(dictionary[ngrams_lemmas[word]]['D'])]
|
||||
elif ngrams_words[word] in dictionary:
|
||||
matches+=1
|
||||
totalVAD = [totalVAD[0]+float(dictionary[ngrams_words[word]]['V']),
|
||||
totalVAD[1]+float(dictionary[ngrams_words[word]]['A']),
|
||||
totalVAD[2]+float(dictionary[ngrams_words[word]]['D'])]
|
||||
if matches==0:
|
||||
emotion='neutral'
|
||||
else:
|
||||
totalVAD=[totalVAD[0]/matches,totalVAD[1]/matches,totalVAD[2]/matches]
|
||||
emotion=self._emotion_calculate(totalVAD)
|
||||
feature_set['emotion']=emotion
|
||||
feature_set['V']=totalVAD[0]
|
||||
feature_set['A']=totalVAD[1]
|
||||
feature_set['D']=totalVAD[2]
|
||||
return feature_set
|
||||
|
||||
def analyse_entry(self, entry, params):
|
||||
|
||||
text_input = entry.get("text", None)
|
||||
|
||||
text= self._my_preprocessor(text_input)
|
||||
dictionary={}
|
||||
lang = params.get("language", "auto")
|
||||
if lang == 'es':
|
||||
with open(self._local_path+self.anew_path_es,'rb') as tabfile:
|
||||
reader = csv.reader(tabfile, delimiter='\t')
|
||||
for row in reader:
|
||||
dictionary[row[2]]={}
|
||||
dictionary[row[2]]['V']=row[3]
|
||||
dictionary[row[2]]['A']=row[5]
|
||||
dictionary[row[2]]['D']=row[7]
|
||||
else:
|
||||
with open(self._local_path+self.anew_path_en,'rb') as tabfile:
|
||||
reader = csv.reader(tabfile, delimiter='\t')
|
||||
for row in reader:
|
||||
dictionary[row[0]]={}
|
||||
dictionary[row[0]]['V']=row[2]
|
||||
dictionary[row[0]]['A']=row[4]
|
||||
dictionary[row[0]]['D']=row[6]
|
||||
|
||||
feature_set=self._extract_features(text,dictionary,lang)
|
||||
|
||||
emotions = EmotionSet()
|
||||
emotions.id = "Emotions0"
|
||||
|
||||
emotion1 = Emotion(id="Emotion0")
|
||||
|
||||
emotion1["onyx:hasEmotionCategory"] = self.emotions_ontology[feature_set['emotion']]
|
||||
emotion1["http://www.gsi.dit.upm.es/ontologies/onyx/vocabularies/anew/ns#valence"] = feature_set['V']
|
||||
emotion1["http://www.gsi.dit.upm.es/ontologies/onyx/vocabularies/anew/ns#arousal"] = feature_set['A']
|
||||
emotion1["http://www.gsi.dit.upm.es/ontologies/onyx/vocabularies/anew/ns#dominance"] = feature_set['D']
|
||||
|
||||
emotions.onyx__hasEmotion.append(emotion1)
|
||||
entry.emotions = [emotions,]
|
||||
|
||||
yield entry
|
64
emotion-anew/emotion-anew.senpy
Normal file
64
emotion-anew/emotion-anew.senpy
Normal file
@ -0,0 +1,64 @@
|
||||
{
|
||||
"name": "emotion-anew",
|
||||
"module": "emotion-anew",
|
||||
"description": "This plugin consists on an emotion classifier using ANEW lexicon dictionary to calculate VAD (valence-arousal-dominance) of the sentence and determinate which emotion is closer to this value. Each emotion has a centroid, calculated according to this article: http://www.aclweb.org/anthology/W10-0208. The plugin is going to look for the words in the sentence that appear in the ANEW dictionary and calculate the average VAD score for the sentence. Once this score is calculated, it is going to seek the emotion that is closest to this value.",
|
||||
"author": "@icorcuera",
|
||||
"version": "0.5",
|
||||
"extra_params": {
|
||||
"language": {
|
||||
"aliases": ["language", "l"],
|
||||
"required": true,
|
||||
"options": ["es","en"],
|
||||
"default": "en"
|
||||
}
|
||||
},
|
||||
"requirements": {},
|
||||
"anew_path_es": "/data/emotion-anew/Dictionary/Redondo(2007).csv",
|
||||
"anew_path_en": "/data/emotion-anew/Dictionary/ANEW2010All.txt",
|
||||
"centroids": {
|
||||
"anger": {
|
||||
"A": 6.95,
|
||||
"D": 5.1,
|
||||
"V": 2.7
|
||||
},
|
||||
"disgust": {
|
||||
"A": 5.3,
|
||||
"D": 8.05,
|
||||
"V": 2.7
|
||||
},
|
||||
"fear": {
|
||||
"A": 6.5,
|
||||
"D": 3.6,
|
||||
"V": 3.2
|
||||
},
|
||||
"joy": {
|
||||
"A": 7.22,
|
||||
"D": 6.28,
|
||||
"V": 8.6
|
||||
},
|
||||
"sadness": {
|
||||
"A": 5.21,
|
||||
"D": 2.82,
|
||||
"V": 2.21
|
||||
}
|
||||
},
|
||||
"emotions_ontology": {
|
||||
"anger": "http://gsi.dit.upm.es/ontologies/wnaffect/ns#anger",
|
||||
"disgust": "http://gsi.dit.upm.es/ontologies/wnaffect/ns#disgust",
|
||||
"fear": "http://gsi.dit.upm.es/ontologies/wnaffect/ns#negative-fear",
|
||||
"joy": "http://gsi.dit.upm.es/ontologies/wnaffect/ns#joy",
|
||||
"neutral": "http://gsi.dit.upm.es/ontologies/wnaffect/ns#neutral-emotion",
|
||||
"sadness": "http://gsi.dit.upm.es/ontologies/wnaffect/ns#sadness"
|
||||
},
|
||||
"requirements": [
|
||||
"numpy",
|
||||
"pandas",
|
||||
"nltk",
|
||||
"scipy",
|
||||
"scikit-learn",
|
||||
"textblob",
|
||||
"pattern",
|
||||
"lxml"
|
||||
],
|
||||
"onyx:usesEmotionModel": "emoml:big6",
|
||||
}
|
45
emotion-anew/test.py
Normal file
45
emotion-anew/test.py
Normal file
@ -0,0 +1,45 @@
|
||||
import os
|
||||
import logging
|
||||
logging.basicConfig()
|
||||
try:
|
||||
import unittest.mock as mock
|
||||
except ImportError:
|
||||
import mock
|
||||
from senpy.extensions import Senpy
|
||||
from flask import Flask
|
||||
import unittest
|
||||
import re
|
||||
|
||||
class emoTextANEWTest(unittest.TestCase):
|
||||
|
||||
def setUp(self):
|
||||
self.app = Flask("test_plugin")
|
||||
self.dir = os.path.join(os.path.dirname(__file__))
|
||||
self.senpy = Senpy(plugin_folder=self.dir, default_plugins=False)
|
||||
self.senpy.init_app(self.app)
|
||||
|
||||
def tearDown(self):
|
||||
self.senpy.deactivate_plugin("EmoTextANEW", sync=True)
|
||||
|
||||
def test_analyse(self):
|
||||
plugin = self.senpy.plugins["EmoTextANEW"]
|
||||
plugin.activate()
|
||||
|
||||
ontology = "http://gsi.dit.upm.es/ontologies/wnaffect/ns#"
|
||||
texts = {'I hate you': 'anger',
|
||||
'i am sad': 'sadness',
|
||||
'i am happy with my marks': 'joy',
|
||||
'This movie is scary': 'negative-fear',
|
||||
'this cake is disgusting' : 'negative-fear'}
|
||||
|
||||
for text in texts:
|
||||
response = plugin.analyse(input=text)
|
||||
expected = texts[text]
|
||||
emotionSet = response.entries[0].emotions[0]
|
||||
|
||||
assert emotionSet['onyx:hasEmotion'][0]['onyx:hasEmotionCategory'] == ontology+expected
|
||||
|
||||
plugin.deactivate()
|
||||
|
||||
if __name__ == '__main__':
|
||||
unittest.main()
|
39
sentiment-vader/README.md
Normal file
39
sentiment-vader/README.md
Normal file
@ -0,0 +1,39 @@
|
||||
# Sentimet-vader plugin
|
||||
|
||||
=========
|
||||
|
||||
Vader is a plugin developed at GSI UPM for sentiment analysis.
|
||||
|
||||
For developing this plugin, it has been used the module vaderSentiment, which is described in the paper:
|
||||
VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text
|
||||
C.J. Hutto and Eric Gilbert
|
||||
Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.
|
||||
|
||||
If you use this plugin in your research, please cite the above paper.
|
||||
|
||||
For more information about the functionality, check the official repository
|
||||
|
||||
https://github.com/cjhutto/vaderSentiment
|
||||
|
||||
The response of this plugin uses [Marl ontology](https://www.gsi.dit.upm.es/ontologies/marl/) developed at GSI UPM for semantic web.
|
||||
|
||||
## Usage
|
||||
|
||||
Params accepted:
|
||||
- Language: es (Spanish), en(English).
|
||||
- Input: Text to analyse.
|
||||
|
||||
|
||||
Example request:
|
||||
```
|
||||
http://senpy.cluster.gsi.dit.upm.es/api/?algo=sentiment-vader&language=en&input=I%20love%20Madrid
|
||||
```
|
||||
|
||||
Example respond: This plugin follows the standard for the senpy plugin response. For more information, please visit [senpy documentation](http://senpy.readthedocs.io). Specifically, NIF API section.
|
||||
|
||||
|
||||
![alt GSI Logo][logoGSI]
|
||||
|
||||
[logoGSI]: http://www.gsi.dit.upm.es/images/stories/logos/gsi.png "GSI Logo"
|
||||
|
||||
========
|
16
sentiment-vader/README.txt
Normal file
16
sentiment-vader/README.txt
Normal file
@ -0,0 +1,16 @@
|
||||
==========
|
||||
|
||||
This README file describes the plugin vaderSentiment.
|
||||
|
||||
For developing this plugin, it has been used the module vaderSentiment, which is described in the paper:
|
||||
VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text
|
||||
C.J. Hutto and Eric Gilbert
|
||||
Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.
|
||||
|
||||
If you use this plugin in your research, please cite the above paper
|
||||
|
||||
For more information about the functionality, check the official repository
|
||||
|
||||
https://github.com/cjhutto/vaderSentiment
|
||||
|
||||
========
|
49
sentiment-vader/sentiment-vader.py
Normal file
49
sentiment-vader/sentiment-vader.py
Normal file
@ -0,0 +1,49 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
|
||||
from vaderSentiment import sentiment
|
||||
from senpy.plugins import SentimentPlugin, SenpyPlugin
|
||||
from senpy.models import Results, Sentiment, Entry
|
||||
import logging
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
class vaderSentimentPlugin(SentimentPlugin):
|
||||
|
||||
def analyse_entry(self,entry,params):
|
||||
|
||||
logger.debug("Analysing with params {}".format(params))
|
||||
|
||||
text_input = entry.get("text", None)
|
||||
aggregate = params['aggregate']
|
||||
|
||||
score = sentiment(text_input)
|
||||
|
||||
opinion0 = Sentiment(id= "Opinion_positive",
|
||||
marl__hasPolarity= "marl:Positive",
|
||||
marl__algorithmConfidence= score['pos']
|
||||
)
|
||||
opinion1 = Sentiment(id= "Opinion_negative",
|
||||
marl__hasPolarity= "marl:Negative",
|
||||
marl__algorithmConfidence= score['neg']
|
||||
)
|
||||
opinion2 = Sentiment(id= "Opinion_neutral",
|
||||
marl__hasPolarity = "marl:Neutral",
|
||||
marl__algorithmConfidence = score['neu']
|
||||
)
|
||||
|
||||
if aggregate == 'true':
|
||||
res = None
|
||||
confident = max(score['neg'],score['neu'],score['pos'])
|
||||
if opinion0.marl__algorithmConfidence == confident:
|
||||
res = opinion0
|
||||
elif opinion1.marl__algorithmConfidence == confident:
|
||||
res = opinion1
|
||||
elif opinion2.marl__algorithmConfidence == confident:
|
||||
res = opinion2
|
||||
entry.sentiments.append(res)
|
||||
else:
|
||||
entry.sentiments.append(opinion0)
|
||||
entry.sentiments.append(opinion1)
|
||||
entry.sentiments.append(opinion2)
|
||||
|
||||
yield entry
|
25
sentiment-vader/sentiment-vader.senpy
Normal file
25
sentiment-vader/sentiment-vader.senpy
Normal file
@ -0,0 +1,25 @@
|
||||
{
|
||||
"name": "sentiment-vader",
|
||||
"module": "sentiment-vader",
|
||||
"description": "Sentiment classifier using vaderSentiment module. Params accepted: Language: {en, es}. The output uses Marl ontology developed at GSI UPM for semantic web.",
|
||||
"author": "@icorcuera",
|
||||
"version": "0.1",
|
||||
"extra_params": {
|
||||
"language": {
|
||||
"@id": "lang_rand",
|
||||
"aliases": ["language", "l"],
|
||||
"required": false,
|
||||
"options": ["es", "en", "auto"]
|
||||
},
|
||||
|
||||
"aggregate": {
|
||||
"aliases": ["aggregate","agg"],
|
||||
"options": ["true", "false"],
|
||||
"required": false,
|
||||
"default": false
|
||||
|
||||
}
|
||||
|
||||
},
|
||||
"requirements": {}
|
||||
}
|
44
sentiment-vader/test.py
Normal file
44
sentiment-vader/test.py
Normal file
@ -0,0 +1,44 @@
|
||||
import os
|
||||
import logging
|
||||
logging.basicConfig()
|
||||
try:
|
||||
import unittest.mock as mock
|
||||
except ImportError:
|
||||
import mock
|
||||
from senpy.extensions import Senpy
|
||||
from flask import Flask
|
||||
from flask.ext.testing import TestCase
|
||||
import unittest
|
||||
|
||||
class vaderTest(unittest.TestCase):
|
||||
|
||||
def setUp(self):
|
||||
self.app = Flask("test_plugin")
|
||||
self.dir = os.path.join(os.path.dirname(__file__))
|
||||
self.senpy = Senpy(plugin_folder=self.dir, default_plugins=False)
|
||||
self.senpy.init_app(self.app)
|
||||
|
||||
def tearDown(self):
|
||||
self.senpy.deactivate_plugin("vaderSentiment", sync=True)
|
||||
|
||||
def test_analyse(self):
|
||||
plugin = self.senpy.plugins["vaderSentiment"]
|
||||
plugin.activate()
|
||||
|
||||
texts = {'I am tired :(' : 'marl:Negative',
|
||||
'I love pizza' : 'marl:Positive',
|
||||
'I like going to the cinema :)' : 'marl:Positive',
|
||||
'This cake is disgusting' : 'marl:Negative'}
|
||||
|
||||
for text in texts:
|
||||
response = plugin.analyse(input=text)
|
||||
expected = texts[text]
|
||||
sentimentSet = response.entries[0].sentiments
|
||||
|
||||
max_sentiment = max(sentimentSet, key=lambda x: x['marl:polarityValue'])
|
||||
assert max_sentiment['marl:hasPolarity'] == expected
|
||||
|
||||
plugin.deactivate()
|
||||
|
||||
if __name__ == '__main__':
|
||||
unittest.main()
|
363
sentiment-vader/vaderSentiment.py
Normal file
363
sentiment-vader/vaderSentiment.py
Normal file
@ -0,0 +1,363 @@
|
||||
#!/usr/bin/python
|
||||
# coding: utf-8
|
||||
'''
|
||||
Created on July 04, 2013
|
||||
@author: C.J. Hutto
|
||||
|
||||
Citation Information
|
||||
|
||||
If you use any of the VADER sentiment analysis tools
|
||||
(VADER sentiment lexicon or Python code for rule-based sentiment
|
||||
analysis engine) in your work or research, please cite the paper.
|
||||
For example:
|
||||
|
||||
Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for
|
||||
Sentiment Analysis of Social Media Text. Eighth International Conference on
|
||||
Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.
|
||||
'''
|
||||
|
||||
import os, math, re, sys, fnmatch, string
|
||||
reload(sys)
|
||||
|
||||
def make_lex_dict(f):
|
||||
return dict(map(lambda (w, m): (w, float(m)), [wmsr.strip().split('\t')[0:2] for wmsr in open(f) ]))
|
||||
|
||||
f = 'vader_sentiment_lexicon.txt' # empirically derived valence ratings for words, emoticons, slang, swear words, acronyms/initialisms
|
||||
try:
|
||||
word_valence_dict = make_lex_dict(f)
|
||||
except:
|
||||
f = os.path.join(os.path.dirname(__file__),'vader_sentiment_lexicon.txt')
|
||||
word_valence_dict = make_lex_dict(f)
|
||||
|
||||
# for removing punctuation
|
||||
regex_remove_punctuation = re.compile('[%s]' % re.escape(string.punctuation))
|
||||
|
||||
def sentiment(text):
|
||||
"""
|
||||
Returns a float for sentiment strength based on the input text.
|
||||
Positive values are positive valence, negative value are negative valence.
|
||||
"""
|
||||
wordsAndEmoticons = str(text).split() #doesn't separate words from adjacent punctuation (keeps emoticons & contractions)
|
||||
text_mod = regex_remove_punctuation.sub('', text) # removes punctuation (but loses emoticons & contractions)
|
||||
wordsOnly = str(text_mod).split()
|
||||
# get rid of empty items or single letter "words" like 'a' and 'I' from wordsOnly
|
||||
for word in wordsOnly:
|
||||
if len(word) <= 1:
|
||||
wordsOnly.remove(word)
|
||||
# now remove adjacent & redundant punctuation from [wordsAndEmoticons] while keeping emoticons and contractions
|
||||
puncList = [".", "!", "?", ",", ";", ":", "-", "'", "\"",
|
||||
"!!", "!!!", "??", "???", "?!?", "!?!", "?!?!", "!?!?"]
|
||||
for word in wordsOnly:
|
||||
for p in puncList:
|
||||
pword = p + word
|
||||
x1 = wordsAndEmoticons.count(pword)
|
||||
while x1 > 0:
|
||||
i = wordsAndEmoticons.index(pword)
|
||||
wordsAndEmoticons.remove(pword)
|
||||
wordsAndEmoticons.insert(i, word)
|
||||
x1 = wordsAndEmoticons.count(pword)
|
||||
|
||||
wordp = word + p
|
||||
x2 = wordsAndEmoticons.count(wordp)
|
||||
while x2 > 0:
|
||||
i = wordsAndEmoticons.index(wordp)
|
||||
wordsAndEmoticons.remove(wordp)
|
||||
wordsAndEmoticons.insert(i, word)
|
||||
x2 = wordsAndEmoticons.count(wordp)
|
||||
# get rid of residual empty items or single letter "words" like 'a' and 'I' from wordsAndEmoticons
|
||||
for word in wordsAndEmoticons:
|
||||
if len(word) <= 1:
|
||||
wordsAndEmoticons.remove(word)
|
||||
|
||||
# remove stopwords from [wordsAndEmoticons]
|
||||
#stopwords = [str(word).strip() for word in open('stopwords.txt')]
|
||||
#for word in wordsAndEmoticons:
|
||||
# if word in stopwords:
|
||||
# wordsAndEmoticons.remove(word)
|
||||
|
||||
# check for negation
|
||||
negate = ["aint", "arent", "cannot", "cant", "couldnt", "darent", "didnt", "doesnt",
|
||||
"ain't", "aren't", "can't", "couldn't", "daren't", "didn't", "doesn't",
|
||||
"dont", "hadnt", "hasnt", "havent", "isnt", "mightnt", "mustnt", "neither",
|
||||
"don't", "hadn't", "hasn't", "haven't", "isn't", "mightn't", "mustn't",
|
||||
"neednt", "needn't", "never", "none", "nope", "nor", "not", "nothing", "nowhere",
|
||||
"oughtnt", "shant", "shouldnt", "uhuh", "wasnt", "werent",
|
||||
"oughtn't", "shan't", "shouldn't", "uh-uh", "wasn't", "weren't",
|
||||
"without", "wont", "wouldnt", "won't", "wouldn't", "rarely", "seldom", "despite"]
|
||||
def negated(list, nWords=[], includeNT=True):
|
||||
nWords.extend(negate)
|
||||
for word in nWords:
|
||||
if word in list:
|
||||
return True
|
||||
if includeNT:
|
||||
for word in list:
|
||||
if "n't" in word:
|
||||
return True
|
||||
if "least" in list:
|
||||
i = list.index("least")
|
||||
if i > 0 and list[i-1] != "at":
|
||||
return True
|
||||
return False
|
||||
|
||||
def normalize(score, alpha=15):
|
||||
# normalize the score to be between -1 and 1 using an alpha that approximates the max expected value
|
||||
normScore = score/math.sqrt( ((score*score) + alpha) )
|
||||
return normScore
|
||||
|
||||
def wildCardMatch(patternWithWildcard, listOfStringsToMatchAgainst):
|
||||
listOfMatches = fnmatch.filter(listOfStringsToMatchAgainst, patternWithWildcard)
|
||||
return listOfMatches
|
||||
|
||||
|
||||
def isALLCAP_differential(wordList):
|
||||
countALLCAPS= 0
|
||||
for w in wordList:
|
||||
if str(w).isupper():
|
||||
countALLCAPS += 1
|
||||
cap_differential = len(wordList) - countALLCAPS
|
||||
if cap_differential > 0 and cap_differential < len(wordList):
|
||||
isDiff = True
|
||||
else: isDiff = False
|
||||
return isDiff
|
||||
isCap_diff = isALLCAP_differential(wordsAndEmoticons)
|
||||
|
||||
b_incr = 0.293 #(empirically derived mean sentiment intensity rating increase for booster words)
|
||||
b_decr = -0.293
|
||||
# booster/dampener 'intensifiers' or 'degree adverbs' http://en.wiktionary.org/wiki/Category:English_degree_adverbs
|
||||
booster_dict = {"absolutely": b_incr, "amazingly": b_incr, "awfully": b_incr, "completely": b_incr, "considerably": b_incr,
|
||||
"decidedly": b_incr, "deeply": b_incr, "effing": b_incr, "enormously": b_incr,
|
||||
"entirely": b_incr, "especially": b_incr, "exceptionally": b_incr, "extremely": b_incr,
|
||||
"fabulously": b_incr, "flipping": b_incr, "flippin": b_incr,
|
||||
"fricking": b_incr, "frickin": b_incr, "frigging": b_incr, "friggin": b_incr, "fully": b_incr, "fucking": b_incr,
|
||||
"greatly": b_incr, "hella": b_incr, "highly": b_incr, "hugely": b_incr, "incredibly": b_incr,
|
||||
"intensely": b_incr, "majorly": b_incr, "more": b_incr, "most": b_incr, "particularly": b_incr,
|
||||
"purely": b_incr, "quite": b_incr, "really": b_incr, "remarkably": b_incr,
|
||||
"so": b_incr, "substantially": b_incr,
|
||||
"thoroughly": b_incr, "totally": b_incr, "tremendously": b_incr,
|
||||
"uber": b_incr, "unbelievably": b_incr, "unusually": b_incr, "utterly": b_incr,
|
||||
"very": b_incr,
|
||||
|
||||
"almost": b_decr, "barely": b_decr, "hardly": b_decr, "just enough": b_decr,
|
||||
"kind of": b_decr, "kinda": b_decr, "kindof": b_decr, "kind-of": b_decr,
|
||||
"less": b_decr, "little": b_decr, "marginally": b_decr, "occasionally": b_decr, "partly": b_decr,
|
||||
"scarcely": b_decr, "slightly": b_decr, "somewhat": b_decr,
|
||||
"sort of": b_decr, "sorta": b_decr, "sortof": b_decr, "sort-of": b_decr}
|
||||
sentiments = []
|
||||
for item in wordsAndEmoticons:
|
||||
v = 0
|
||||
i = wordsAndEmoticons.index(item)
|
||||
if (i < len(wordsAndEmoticons)-1 and str(item).lower() == "kind" and \
|
||||
str(wordsAndEmoticons[i+1]).lower() == "of") or str(item).lower() in booster_dict:
|
||||
sentiments.append(v)
|
||||
continue
|
||||
item_lowercase = str(item).lower()
|
||||
if item_lowercase in word_valence_dict:
|
||||
#get the sentiment valence
|
||||
v = float(word_valence_dict[item_lowercase])
|
||||
|
||||
#check if sentiment laden word is in ALLCAPS (while others aren't)
|
||||
c_incr = 0.733 #(empirically derived mean sentiment intensity rating increase for using ALLCAPs to emphasize a word)
|
||||
if str(item).isupper() and isCap_diff:
|
||||
if v > 0: v += c_incr
|
||||
else: v -= c_incr
|
||||
|
||||
#check if the preceding words increase, decrease, or negate/nullify the valence
|
||||
def scalar_inc_dec(word, valence):
|
||||
scalar = 0.0
|
||||
word_lower = str(word).lower()
|
||||
if word_lower in booster_dict:
|
||||
scalar = booster_dict[word_lower]
|
||||
if valence < 0: scalar *= -1
|
||||
#check if booster/dampener word is in ALLCAPS (while others aren't)
|
||||
if str(word).isupper() and isCap_diff:
|
||||
if valence > 0: scalar += c_incr
|
||||
else: scalar -= c_incr
|
||||
return scalar
|
||||
n_scalar = -0.74
|
||||
if i > 0 and str(wordsAndEmoticons[i-1]).lower() not in word_valence_dict:
|
||||
s1 = scalar_inc_dec(wordsAndEmoticons[i-1], v)
|
||||
v = v+s1
|
||||
if negated([wordsAndEmoticons[i-1]]): v = v*n_scalar
|
||||
if i > 1 and str(wordsAndEmoticons[i-2]).lower() not in word_valence_dict:
|
||||
s2 = scalar_inc_dec(wordsAndEmoticons[i-2], v)
|
||||
if s2 != 0: s2 = s2*0.95
|
||||
v = v+s2
|
||||
# check for special use of 'never' as valence modifier instead of negation
|
||||
if wordsAndEmoticons[i-2] == "never" and (wordsAndEmoticons[i-1] == "so" or wordsAndEmoticons[i-1] == "this"):
|
||||
v = v*1.5
|
||||
# otherwise, check for negation/nullification
|
||||
elif negated([wordsAndEmoticons[i-2]]): v = v*n_scalar
|
||||
if i > 2 and str(wordsAndEmoticons[i-3]).lower() not in word_valence_dict:
|
||||
s3 = scalar_inc_dec(wordsAndEmoticons[i-3], v)
|
||||
if s3 != 0: s3 = s3*0.9
|
||||
v = v+s3
|
||||
# check for special use of 'never' as valence modifier instead of negation
|
||||
if wordsAndEmoticons[i-3] == "never" and \
|
||||
(wordsAndEmoticons[i-2] == "so" or wordsAndEmoticons[i-2] == "this") or \
|
||||
(wordsAndEmoticons[i-1] == "so" or wordsAndEmoticons[i-1] == "this"):
|
||||
v = v*1.25
|
||||
# otherwise, check for negation/nullification
|
||||
elif negated([wordsAndEmoticons[i-3]]): v = v*n_scalar
|
||||
|
||||
# check for special case idioms using a sentiment-laden keyword known to SAGE
|
||||
special_case_idioms = {"the shit": 3, "the bomb": 3, "bad ass": 1.5, "yeah right": -2,
|
||||
"cut the mustard": 2, "kiss of death": -1.5, "hand to mouth": -2}
|
||||
# future work: consider other sentiment-laden idioms
|
||||
#other_idioms = {"back handed": -2, "blow smoke": -2, "blowing smoke": -2, "upper hand": 1, "break a leg": 2,
|
||||
# "cooking with gas": 2, "in the black": 2, "in the red": -2, "on the ball": 2,"under the weather": -2}
|
||||
onezero = "{} {}".format(str(wordsAndEmoticons[i-1]), str(wordsAndEmoticons[i]))
|
||||
twoonezero = "{} {} {}".format(str(wordsAndEmoticons[i-2]), str(wordsAndEmoticons[i-1]), str(wordsAndEmoticons[i]))
|
||||
twoone = "{} {}".format(str(wordsAndEmoticons[i-2]), str(wordsAndEmoticons[i-1]))
|
||||
threetwoone = "{} {} {}".format(str(wordsAndEmoticons[i-3]), str(wordsAndEmoticons[i-2]), str(wordsAndEmoticons[i-1]))
|
||||
threetwo = "{} {}".format(str(wordsAndEmoticons[i-3]), str(wordsAndEmoticons[i-2]))
|
||||
if onezero in special_case_idioms: v = special_case_idioms[onezero]
|
||||
elif twoonezero in special_case_idioms: v = special_case_idioms[twoonezero]
|
||||
elif twoone in special_case_idioms: v = special_case_idioms[twoone]
|
||||
elif threetwoone in special_case_idioms: v = special_case_idioms[threetwoone]
|
||||
elif threetwo in special_case_idioms: v = special_case_idioms[threetwo]
|
||||
if len(wordsAndEmoticons)-1 > i:
|
||||
zeroone = "{} {}".format(str(wordsAndEmoticons[i]), str(wordsAndEmoticons[i+1]))
|
||||
if zeroone in special_case_idioms: v = special_case_idioms[zeroone]
|
||||
if len(wordsAndEmoticons)-1 > i+1:
|
||||
zeroonetwo = "{} {}".format(str(wordsAndEmoticons[i]), str(wordsAndEmoticons[i+1]), str(wordsAndEmoticons[i+2]))
|
||||
if zeroonetwo in special_case_idioms: v = special_case_idioms[zeroonetwo]
|
||||
|
||||
# check for booster/dampener bi-grams such as 'sort of' or 'kind of'
|
||||
if threetwo in booster_dict or twoone in booster_dict:
|
||||
v = v+b_decr
|
||||
|
||||
# check for negation case using "least"
|
||||
if i > 1 and str(wordsAndEmoticons[i-1]).lower() not in word_valence_dict \
|
||||
and str(wordsAndEmoticons[i-1]).lower() == "least":
|
||||
if (str(wordsAndEmoticons[i-2]).lower() != "at" and str(wordsAndEmoticons[i-2]).lower() != "very"):
|
||||
v = v*n_scalar
|
||||
elif i > 0 and str(wordsAndEmoticons[i-1]).lower() not in word_valence_dict \
|
||||
and str(wordsAndEmoticons[i-1]).lower() == "least":
|
||||
v = v*n_scalar
|
||||
sentiments.append(v)
|
||||
|
||||
# check for modification in sentiment due to contrastive conjunction 'but'
|
||||
if 'but' in wordsAndEmoticons or 'BUT' in wordsAndEmoticons:
|
||||
try: bi = wordsAndEmoticons.index('but')
|
||||
except: bi = wordsAndEmoticons.index('BUT')
|
||||
for s in sentiments:
|
||||
si = sentiments.index(s)
|
||||
if si < bi:
|
||||
sentiments.pop(si)
|
||||
sentiments.insert(si, s*0.5)
|
||||
elif si > bi:
|
||||
sentiments.pop(si)
|
||||
sentiments.insert(si, s*1.5)
|
||||
|
||||
if sentiments:
|
||||
sum_s = float(sum(sentiments))
|
||||
#print sentiments, sum_s
|
||||
|
||||
# check for added emphasis resulting from exclamation points (up to 4 of them)
|
||||
ep_count = str(text).count("!")
|
||||
if ep_count > 4: ep_count = 4
|
||||
ep_amplifier = ep_count*0.292 #(empirically derived mean sentiment intensity rating increase for exclamation points)
|
||||
if sum_s > 0: sum_s += ep_amplifier
|
||||
elif sum_s < 0: sum_s -= ep_amplifier
|
||||
|
||||
# check for added emphasis resulting from question marks (2 or 3+)
|
||||
qm_count = str(text).count("?")
|
||||
qm_amplifier = 0
|
||||
if qm_count > 1:
|
||||
if qm_count <= 3: qm_amplifier = qm_count*0.18
|
||||
else: qm_amplifier = 0.96
|
||||
if sum_s > 0: sum_s += qm_amplifier
|
||||
elif sum_s < 0: sum_s -= qm_amplifier
|
||||
|
||||
compound = normalize(sum_s)
|
||||
|
||||
# want separate positive versus negative sentiment scores
|
||||
pos_sum = 0.0
|
||||
neg_sum = 0.0
|
||||
neu_count = 0
|
||||
for sentiment_score in sentiments:
|
||||
if sentiment_score > 0:
|
||||
pos_sum += (float(sentiment_score) +1) # compensates for neutral words that are counted as 1
|
||||
if sentiment_score < 0:
|
||||
neg_sum += (float(sentiment_score) -1) # when used with math.fabs(), compensates for neutrals
|
||||
if sentiment_score == 0:
|
||||
neu_count += 1
|
||||
|
||||
if pos_sum > math.fabs(neg_sum): pos_sum += (ep_amplifier+qm_amplifier)
|
||||
elif pos_sum < math.fabs(neg_sum): neg_sum -= (ep_amplifier+qm_amplifier)
|
||||
|
||||
total = pos_sum + math.fabs(neg_sum) + neu_count
|
||||
pos = math.fabs(pos_sum / total)
|
||||
neg = math.fabs(neg_sum / total)
|
||||
neu = math.fabs(neu_count / total)
|
||||
|
||||
else:
|
||||
compound = 0.0; pos = 0.0; neg = 0.0; neu = 0.0
|
||||
|
||||
s = {"neg" : round(neg, 3),
|
||||
"neu" : round(neu, 3),
|
||||
"pos" : round(pos, 3),
|
||||
"compound" : round(compound, 4)}
|
||||
return s
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
# --- examples -------
|
||||
sentences = [
|
||||
"VADER is smart, handsome, and funny.", # positive sentence example
|
||||
"VADER is smart, handsome, and funny!", # punctuation emphasis handled correctly (sentiment intensity adjusted)
|
||||
"VADER is very smart, handsome, and funny.", # booster words handled correctly (sentiment intensity adjusted)
|
||||
"VADER is VERY SMART, handsome, and FUNNY.", # emphasis for ALLCAPS handled
|
||||
"VADER is VERY SMART, handsome, and FUNNY!!!",# combination of signals - VADER appropriately adjusts intensity
|
||||
"VADER is VERY SMART, really handsome, and INCREDIBLY FUNNY!!!",# booster words & punctuation make this close to ceiling for score
|
||||
"The book was good.", # positive sentence
|
||||
"The book was kind of good.", # qualified positive sentence is handled correctly (intensity adjusted)
|
||||
"The plot was good, but the characters are uncompelling and the dialog is not great.", # mixed negation sentence
|
||||
"A really bad, horrible book.", # negative sentence with booster words
|
||||
"At least it isn't a horrible book.", # negated negative sentence with contraction
|
||||
":) and :D", # emoticons handled
|
||||
"", # an empty string is correctly handled
|
||||
"Today sux", # negative slang handled
|
||||
"Today sux!", # negative slang with punctuation emphasis handled
|
||||
"Today SUX!", # negative slang with capitalization emphasis
|
||||
"Today kinda sux! But I'll get by, lol" # mixed sentiment example with slang and constrastive conjunction "but"
|
||||
]
|
||||
paragraph = "It was one of the worst movies I've seen, despite good reviews. \
|
||||
Unbelievably bad acting!! Poor direction. VERY poor production. \
|
||||
The movie was bad. Very bad movie. VERY bad movie. VERY BAD movie. VERY BAD movie!"
|
||||
|
||||
from nltk import tokenize
|
||||
lines_list = tokenize.sent_tokenize(paragraph)
|
||||
sentences.extend(lines_list)
|
||||
|
||||
tricky_sentences = [
|
||||
"Most automated sentiment analysis tools are shit.",
|
||||
"VADER sentiment analysis is the shit.",
|
||||
"Sentiment analysis has never been good.",
|
||||
"Sentiment analysis with VADER has never been this good.",
|
||||
"Warren Beatty has never been so entertaining.",
|
||||
"I won't say that the movie is astounding and I wouldn't claim that the movie is too banal either.",
|
||||
"I like to hate Michael Bay films, but I couldn't fault this one",
|
||||
"It's one thing to watch an Uwe Boll film, but another thing entirely to pay for it",
|
||||
"The movie was too good",
|
||||
"This movie was actually neither that funny, nor super witty.",
|
||||
"This movie doesn't care about cleverness, wit or any other kind of intelligent humor.",
|
||||
"Those who find ugly meanings in beautiful things are corrupt without being charming.",
|
||||
"There are slow and repetitive parts, BUT it has just enough spice to keep it interesting.",
|
||||
"The script is not fantastic, but the acting is decent and the cinematography is EXCELLENT!",
|
||||
"Roger Dodger is one of the most compelling variations on this theme.",
|
||||
"Roger Dodger is one of the least compelling variations on this theme.",
|
||||
"Roger Dodger is at least compelling as a variation on the theme.",
|
||||
"they fall in love with the product",
|
||||
"but then it breaks",
|
||||
"usually around the time the 90 day warranty expires",
|
||||
"the twin towers collapsed today",
|
||||
"However, Mr. Carter solemnly argues, his client carried out the kidnapping under orders and in the ''least offensive way possible.''"
|
||||
]
|
||||
sentences.extend(tricky_sentences)
|
||||
for sentence in sentences:
|
||||
print sentence,
|
||||
ss = sentiment(sentence)
|
||||
print "\t" + str(ss)
|
||||
|
||||
print "\n\n Done!"
|
7517
sentiment-vader/vader_sentiment_lexicon.txt
Normal file
7517
sentiment-vader/vader_sentiment_lexicon.txt
Normal file
File diff suppressed because it is too large
Load Diff
Loading…
Reference in New Issue
Block a user