Python sentiment.SentimentAnalyzer类代码示例

    def sentiment_analysis(self, testing_data, training_data=None):
        if training_data is None:
            training_data = self.training_data
            ## Apply sentiment analysis to data to extract new "features"

            # Initialize sentiment analyzer object
        sentiment_analyzer = SentimentAnalyzer()

        # Mark all negative words in training data, using existing list of negative words
        all_negative_words = sentiment_analyzer.all_words([mark_negation(data) for data in training_data])

        unigram_features = sentiment_analyzer.unigram_word_feats(all_negative_words, min_freq=4)
        len(unigram_features)
        sentiment_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_features)

        training_final = sentiment_analyzer.apply_features(training_data)
        testing_final = sentiment_analyzer.apply_features(testing_data)

        ## Traing model and test

        model = NaiveBayesClassifier.train
        classifer = sentiment_analyzer.train(model, training_final)

        for key, value in sorted(sentiment_analyzer.evaluate(testing_final).items()):
            print ("{0}: {1}".format(key, value))

def demo_subjectivity(trainer, save_analyzer=False, n_instances=None, output=None):
    """
    Train and test a classifier on instances of the Subjective Dataset by Pang and
    Lee. The dataset is made of 5000 subjective and 5000 objective sentences.
    All tokens (words and punctuation marks) are separated by a whitespace, so
    we use the basic WhitespaceTokenizer to parse the data.

    :param trainer: `train` method of a classifier.
    :param save_analyzer: if `True`, store the SentimentAnalyzer in a pickle file.
    :param n_instances: the number of total sentences that have to be used for
        training and testing. Sentences will be equally split between positive
        and negative.
    :param output: the output file where results have to be reported.
    """
    from nltk.sentiment import SentimentAnalyzer
    from nltk.corpus import subjectivity

    if n_instances is not None:
        n_instances = int(n_instances/2)

    subj_docs = [(sent, 'subj') for sent in subjectivity.sents(categories='subj')[:n_instances]]
    obj_docs = [(sent, 'obj') for sent in subjectivity.sents(categories='obj')[:n_instances]]

    # We separately split subjective and objective instances to keep a balanced
    # uniform class distribution in both train and test sets.
    train_subj_docs, test_subj_docs = split_train_test(subj_docs)
    train_obj_docs, test_obj_docs = split_train_test(obj_docs)

    training_docs = train_subj_docs+train_obj_docs
    testing_docs = test_subj_docs+test_obj_docs

    sentim_analyzer = SentimentAnalyzer()
    all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in training_docs])

    # Add simple unigram word features handling negation
    unigram_feats = sentim_analyzer.unigram_word_feats(all_words_neg, min_freq=4)
    sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)

    # Apply features to obtain a feature-value representation of our datasets
    training_set = sentim_analyzer.apply_features(training_docs)
    test_set = sentim_analyzer.apply_features(testing_docs)

    classifier = sentim_analyzer.train(trainer, training_set)
    try:
        classifier.show_most_informative_features()
    except AttributeError:
        print('Your classifier does not provide a show_most_informative_features() method.')
    results = sentim_analyzer.evaluate(test_set)

    if save_analyzer == True:
        save_file(sentim_analyzer, 'sa_subjectivity.pickle')

    if output:
        extr = [f.__name__ for f in sentim_analyzer.feat_extractors]
        output_markdown(output, Dataset='subjectivity', Classifier=type(classifier).__name__,
                        Tokenizer='WhitespaceTokenizer', Feats=extr,
                        Instances=n_instances, Results=results)

    return sentim_analyzer

def train():
  positive_tweets = read_tweets('positive.txt', 'positive')
  negative_tweets = read_tweets('negative.txt', 'negative')
  print len(positive_tweets)
  print len(negative_tweets)

  pos_train = positive_tweets[:len(positive_tweets)]
  neg_train = negative_tweets[:len(negative_tweets)]
  # pos_test = positive_tweets[len(positive_tweets)*80/100+1:]
  # neg_test = negative_tweets[len(positive_tweets)*80/100+1:]

  training_data = pos_train + neg_train
  # test_data = pos_test + neg_test

  sentim_analyzer = SentimentAnalyzer()
  all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in training_data])
  unigram_feats = sentim_analyzer.unigram_word_feats(all_words_neg, min_freq=4)
  print len(unigram_feats)

  sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)
  training_set = sentim_analyzer.apply_features(training_data)

  # test_set = sentim_analyzer.apply_features(test_data)
  # print test_set

  trainer = NaiveBayesClassifier.train
  sentim_analyzer.train(trainer, training_set)

  # for key,value in sorted(sentim_analyzer.evaluate(test_set).items()):
  #   print('{0}: {1}'.format(key, value))
  # print sentim_analyzer.classify(tokenize_sentence('I hate driving car at night'))

  return sentim_analyzer

def demo_movie_reviews(trainer, n_instances=None, output=None):
    """
    Train classifier on all instances of the Movie Reviews dataset.
    The corpus has been preprocessed using the default sentence tokenizer and
    WordPunctTokenizer.
    Features are composed of:
        - most frequent unigrams

    :param trainer: `train` method of a classifier.
    :param n_instances: the number of total reviews that have to be used for
        training and testing. Reviews will be equally split between positive and
        negative.
    :param output: the output file where results have to be reported.
    """
    from nltk.corpus import movie_reviews
    from nltk.sentiment import SentimentAnalyzer

    if n_instances is not None:
        n_instances = int(n_instances/2)

    pos_docs = [(list(movie_reviews.words(pos_id)), 'pos') for pos_id in movie_reviews.fileids('pos')[:n_instances]]
    neg_docs = [(list(movie_reviews.words(neg_id)), 'neg') for neg_id in movie_reviews.fileids('neg')[:n_instances]]
    # We separately split positive and negative instances to keep a balanced
    # uniform class distribution in both train and test sets.
    train_pos_docs, test_pos_docs = split_train_test(pos_docs)
    train_neg_docs, test_neg_docs = split_train_test(neg_docs)

    training_docs = train_pos_docs+train_neg_docs
    testing_docs = test_pos_docs+test_neg_docs

    sentim_analyzer = SentimentAnalyzer()
    all_words = sentim_analyzer.all_words(training_docs)

    # Add simple unigram word features
    unigram_feats = sentim_analyzer.unigram_word_feats(all_words, min_freq=4)
    sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)
    # Apply features to obtain a feature-value representation of our datasets
    training_set = sentim_analyzer.apply_features(training_docs)
    test_set = sentim_analyzer.apply_features(testing_docs)

    classifier = sentim_analyzer.train(trainer, training_set)
    try:
        classifier.show_most_informative_features()
    except AttributeError:
        print('Your classifier does not provide a show_most_informative_features() method.')
    results = sentim_analyzer.evaluate(test_set)

    if output:
        extr = [f.__name__ for f in sentim_analyzer.feat_extractors]
        output_markdown(output, Dataset='Movie_reviews', Classifier=type(classifier).__name__,
                        Tokenizer='WordPunctTokenizer', Feats=extr, Results=results,
                        Instances=n_instances)

    def __init__(self, sentiment_only, num_phrases_to_track=20):
        # neg_phrases = filter_negative_phrases(load_csv_sentences('thoughtsandfeelings.csv'))
        # pos_phrases = filter_positive_phrases(load_csv_sentences('spiritualforums.csv'))
        # file_pos = open("pos_phrases.txt", 'w')
        # file_neg = open("neg_phrases.txt", 'w')

        # for item in pos_phrases:
        #     print>>file_pos, item
        # for item in neg_phrases:
        #     print>>file_neg, item
        self.recent_sentiment_scores = []

        neg_file = open("ALL_neg_phrases_filtered.txt", "r")
        pos_file = open("webtext_phrases_with_lots_of_words.txt", "r")
        neg_phrases = neg_file.readlines()
        pos_phrases = pos_file.readlines()

        neg_docs = []
        pos_docs = []
        for phrase in neg_phrases:
            neg_docs.append((phrase.split(), 'suicidal'))
        for phrase in pos_phrases[:len(neg_phrases)]:
            pos_docs.append((phrase.split(), 'alright'))

        print len(neg_docs)
        print len(pos_docs)
        # negcutoff = len(neg_docs) * 3 / 4
        # poscutoff = len(pos_docs) * 3 / 4
        negcutoff = -200
        poscutoff = -200

        train_pos_docs = pos_docs[:poscutoff]
        test_pos_docs = pos_docs[poscutoff:]
        train_neg_docs = neg_docs[:negcutoff]
        test_neg_docs = neg_docs[negcutoff:]
        training_docs = train_pos_docs + train_neg_docs
        testing_docs = test_pos_docs + test_neg_docs

        self.sentim_analyzer = SentimentAnalyzer()

        if not sentiment_only:
            all_words = self.sentim_analyzer.all_words([doc for doc in training_docs])
            unigram_feats = self.sentim_analyzer.unigram_word_feats(all_words, min_freq=1)
            self.sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)

        self.sentim_analyzer.add_feat_extractor(vader_sentiment_feat)

        # bigram_feats = self.sentim_analyzer.bigram_collocation_feats(all_words, min_freq=1)
        # self.sentim_analyzer.add_feat_extractor(extract_bigram_feats, bigrams=bigram_feats)

        training_set = self.sentim_analyzer.apply_features(training_docs)
        test_set = self.sentim_analyzer.apply_features(testing_docs)
        trainer = NaiveBayesClassifier.train
        self.classifier = self.sentim_analyzer.train(trainer, training_set)
        for key, value in sorted(self.sentim_analyzer.evaluate(test_set).items()):
            print('{0}: {1}'.format(key, value))
        self.classifier.show_most_informative_features(20)

def train_model(training):
	## Apply sentiment analysis to data to extract new "features"

	# Initialize sentiment analyzer object
	sentiment_analyzer = SentimentAnalyzer()

	# Mark all negative words in training data, using existing list of negative words
	all_negative_words = sentiment_analyzer.all_words([mark_negation(data) for data in training])

	unigram_features = sentiment_analyzer.unigram_word_feats(all_negative_words, min_freq=4)
	len(unigram_features)
	sentiment_analyzer.add_feat_extractor(extract_unigram_feats,unigrams=unigram_features)

	training_final = sentiment_analyzer.apply_features(training)

	return [training_final]

def get_objectivity_analyzer():
    n_instances = 100
    subj_docs = [(sent, 'subj') for sent in subjectivity.sents(categories='subj')[:n_instances]]
    obj_docs = [(sent, 'obj') for sent in subjectivity.sents(categories='obj')[:n_instances]]
    
    train_subj_docs = subj_docs
    train_obj_docs = obj_docs
    training_docs = train_subj_docs+train_obj_docs
    sentim_analyzer = SentimentAnalyzer()
    all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in training_docs])
    
    unigram_feats = sentim_analyzer.unigram_word_feats(all_words_neg, min_freq=4)
    
    sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)
    
    training_set = sentim_analyzer.apply_features(training_docs)

    trainer = NaiveBayesClassifier.train
    sentiment_classifier = sentim_analyzer.train(trainer, training_set)
    
    return sentim_analyzer

    def __init__(self):
        self.sentim_analyzer = SentimentAnalyzer()
        self.genre_dict = read_file("jsons/movie_genre_quote_dict_2.json")
        context_file = "jsons/final_context.json"
        movie_file = "jsons/final_movies.json"
        quote_file = "jsons/final_quotes.json"
        year_rating_file = "jsons/final_year_rating.json"

        self.context = read_file(context_file)
        self.movies = read_file(movie_file)
        self.quotes = read_file(quote_file)
        self.year_rating_dict = read_file(year_rating_file)

        # Reincode to unicode
        for i in range(len(self.context)):
            self.context[i] = self.context[i].encode("utf-8").decode("utf-8")
            self.movies[i] = self.movies[i].encode("utf-8").decode("utf-8")
            self.quotes[i] = self.quotes[i].encode("utf-8").decode("utf-8")

        self.context, self.quotes, self.movies = quote_pruner(self.context, self.quotes, self.movies)

        self.inverted_index = read_file("jsons/f_inverted_index.json")
        self.idf = read_file("jsons/f_idf.json")

        # Initialize query tokenizer
        self.tokenizer = TreebankWordTokenizer()
        # Compute document norms
        self.norms = compute_doc_norms(self.inverted_index, self.idf, len(self.context))

        word_co_filename = "jsons/word_co.json"
        word_count_filename = "jsons/word_count_dict.json"
        pmi_dict_filename = "jsons/pmi_dict.json"
        # Read files
        self.word_co = read_file(word_co_filename)
        self.word_count_dict = read_file(word_count_filename)
        self.pmi_dict = read_file(pmi_dict_filename)

def demo_tweets(trainer, n_instances=None, output=None):
    """
    Train and test Naive Bayes classifier on 10000 tweets, tokenized using
    TweetTokenizer.
    Features are composed of:
        - 1000 most frequent unigrams
        - 100 top bigrams (using BigramAssocMeasures.pmi)

    :param trainer: `train` method of a classifier.
    :param n_instances: the number of total tweets that have to be used for
        training and testing. Tweets will be equally split between positive and
        negative.
    :param output: the output file where results have to be reported.
    """
    from nltk.tokenize import TweetTokenizer
    from nltk.sentiment import SentimentAnalyzer
    from nltk.corpus import twitter_samples, stopwords

    # Different customizations for the TweetTokenizer
    tokenizer = TweetTokenizer(preserve_case=False)
    # tokenizer = TweetTokenizer(preserve_case=True, strip_handles=True)
    # tokenizer = TweetTokenizer(reduce_len=True, strip_handles=True)

    if n_instances is not None:
        n_instances = int(n_instances/2)

    fields = ['id', 'text']
    positive_json = twitter_samples.abspath("positive_tweets.json")
    positive_csv = 'positive_tweets.csv'
    json2csv_preprocess(positive_json, positive_csv, fields, limit=n_instances)

    negative_json = twitter_samples.abspath("negative_tweets.json")
    negative_csv = 'negative_tweets.csv'
    json2csv_preprocess(negative_json, negative_csv, fields, limit=n_instances)

    neg_docs = parse_tweets_set(negative_csv, label='neg', word_tokenizer=tokenizer)
    pos_docs = parse_tweets_set(positive_csv, label='pos', word_tokenizer=tokenizer)

    # We separately split subjective and objective instances to keep a balanced
    # uniform class distribution in both train and test sets.
    train_pos_docs, test_pos_docs = split_train_test(pos_docs)
    train_neg_docs, test_neg_docs = split_train_test(neg_docs)

    training_tweets = train_pos_docs+train_neg_docs
    testing_tweets = test_pos_docs+test_neg_docs

    sentim_analyzer = SentimentAnalyzer()
    # stopwords = stopwords.words('english')
    # all_words = [word for word in sentim_analyzer.all_words(training_tweets) if word.lower() not in stopwords]
    all_words = [word for word in sentim_analyzer.all_words(training_tweets)]

    # Add simple unigram word features
    unigram_feats = sentim_analyzer.unigram_word_feats(all_words, top_n=1000)
    sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)

    # Add bigram collocation features
    bigram_collocs_feats = sentim_analyzer.bigram_collocation_feats([tweet[0] for tweet in training_tweets],
        top_n=100, min_freq=12)
    sentim_analyzer.add_feat_extractor(extract_bigram_feats, bigrams=bigram_collocs_feats)

    training_set = sentim_analyzer.apply_features(training_tweets)
    test_set = sentim_analyzer.apply_features(testing_tweets)

    classifier = sentim_analyzer.train(trainer, training_set)
    # classifier = sentim_analyzer.train(trainer, training_set, max_iter=4)
    try:
        classifier.show_most_informative_features()
    except AttributeError:
        print('Your classifier does not provide a show_most_informative_features() method.')
    results = sentim_analyzer.evaluate(test_set)

    if output:
        extr = [f.__name__ for f in sentim_analyzer.feat_extractors]
        output_markdown(output, Dataset='labeled_tweets', Classifier=type(classifier).__name__,
                        Tokenizer=tokenizer.__class__.__name__, Feats=extr,
                        Results=results, Instances=n_instances)

from nltk.classify import NaiveBayesClassifier
from nltk.corpus import subjectivity
from nltk.sentiment import SentimentAnalyzer
from nltk.sentiment.util import *

n_instances = 100
subj_docs = [(sent, 'subj') for sent in subjectivity.sents(categories='subj')[:n_instances]]
obj_docs = [(sent, 'obj') for sent in subjectivity.sents(categories='obj')[:n_instances]]

train_subj_docs = subj_docs[:80]
test_subj_docs = subj_docs[80:100]
train_obj_docs = obj_docs[:80]
test_obj_docs = obj_docs[80:100]
training_docs = train_subj_docs+train_obj_docs
testing_docs = test_subj_docs+test_obj_docs

sentim_analyzer = SentimentAnalyzer()
all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in training_docs])

print "creating data set"
i = 0
s1 = ""
s2 = ""
tup = (s1, s2)
for line in f:
    if i > 6718:
        break
    if i % 2 == 0:
        s1 = line.split()
    else:
        s2 = line
        tup = (s1, s2)
        train.append(tup)
    i += 1

print train
sentim_analyzer = SentimentAnalyzer()
all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in train])
print all_words_neg
unigram_feats = sentim_analyzer.unigram_word_feats(all_words_neg)
print unigram_feats
sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)
training_set = sentim_analyzer.apply_features(train)
trainer = MaxentClassifier.train
classifier = sentim_analyzer.train(trainer, training_set)

f = open('maxent_trained_with_80_percent.pickle', 'wb')
pickle.dump(classifier, f)
f.close()

training_subjective = subjective[: int(0.8 * n)]
test_subjective = subjective[int(0.8 * n) : n]
training_objective = objective[: int(0.8 * n)]
test_objective = objective[int(0.8 * n) : n]

# Now aggregate the training and test sets

training = training_subjective + training_objective
test = test_subjective + test_objective

## Apply sentiment analysis to data to extract new "features"

# Initialize sentiment analyzer object

sentiment_analyzer = SentimentAnalyzer()

# Mark all negative words in training data, using existing list of negative words

all_negative_words = sentiment_analyzer.all_words([mark_negation(data) for data in training])

unigram_features = sentiment_analyzer.unigram_word_feats(all_negative_words, min_freq=4)
len(unigram_features)
sentiment_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_features)

training_final = sentiment_analyzer.apply_features(training)
test_final = sentiment_analyzer.apply_features(test)

## Traing model and test

model = NaiveBayesClassifier.train

					text = tokenizer.tokenize(line[5].decode("utf-8"))
					text = [token for token in text if token != u'\ufffd']
					test.append((text, sent))
			

		return test, train



# Read in annotated data
NUM_TRAIN = 10000
NUM_TEST = 2500
test, train = read_input("train.csv",NUM_TRAIN,NUM_TEST)


sentiment_analyzer = SentimentAnalyzer()
#all_words = sentiment_analyzer.all_words([mark_negation(doc[0]) for doc in train])
all_words = sentiment_analyzer.all_words([doc[0] for doc in train])
unigrams = sentiment_analyzer.unigram_word_feats(all_words, min_freq=4)
# print unigrams
sentiment_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigrams)

training_set=sentiment_analyzer.apply_features(train)
test_set=sentiment_analyzer.apply_features(test)

trainer = NaiveBayesClassifier.train
classifier = sentiment_analyzer.train(trainer, training_set)
save_file(sentiment_analyzer, "sentiment_classifier.pkl")
for key,value in sorted(sentiment_analyzer.evaluate(test_set).items()):
	print("{0}: {1}".format(key,value))

  row = line.split(',')
  sentiment = row[1]
  tweet = row[3].strip()
  translator = str.maketrans({key: None for key in string.punctuation})
  tweet = tweet.translate(translator)
  tweet = tweet.split(' ')
  tweet_lower = []
  for word in tweet:
    tweet_lower.append(word.lower())
  return (tweet_lower, sentiment)

#call the function on each row in the dataset
train_data = train_data_raw.map(lambda line: get_row(line))

#create a SentimentAnalyzer object
sentim_analyzer = SentimentAnalyzer()

#get list of stopwords (with _NEG) to use as a filter
stopwords_all = []
for word in stopwords.words('english'):
  stopwords_all.append(word)
  stopwords_all.append(word + '_NEG')

#take 10,000 Tweets from this training dataset for this example and get all the words
#that are not stop words
train_data_sample = train_data.take(10000)
all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in train_data_sample])
all_words_neg_nostops = [x for x in all_words_neg if x not in stopwords_all]

#create unigram features and extract features
unigram_feats = sentim_analyzer.unigram_word_feats(all_words_neg_nostops, top_n=200)

len(subj_docs), len(obj_docs)
(100, 100)
subj_docs[0]


(['smart', 'and', 'alert', ',', 'thirteen', 'conversations', 'about', 'one',
'thing', 'is', 'a', 'small', 'gem', '.'], 'subj')


train_subj_docs = subj_docs[:80]
test_subj_docs = subj_docs[80:100]
train_obj_docs = obj_docs[:80]
test_obj_docs = obj_docs[80:100]
training_docs = train_subj_docs+train_obj_docs
testing_docs = test_subj_docs+test_obj_docs
sentim_analyzer = SentimentAnalyzer()
all_words_neg = sentim_analyzer.all_words([mark_negation(doc) for doc in training_docs])
unigram_feats = sentim_analyzer.unigram_word_feats(all_words_neg, min_freq=1)
len(unigram_feats)

sentim_analyzer.add_feat_extractor(extract_unigram_feats, unigrams=unigram_feats)
training_set = sentim_analyzer.apply_features(training_docs)
test_set = sentim_analyzer.apply_features(testing_docs)
trainer = NaiveBayesClassifier.train
classifier = sentim_analyzer.train(trainer, training_set)

for key,value in sorted(sentim_analyzer.evaluate(test_set).items()):
    print('{0}: {1}'.format(key, value))

    from nltk.sentiment.vader import SentimentIntensityAnalyzer
sentences = ["VADER is smart, handsome, and funny.", # positive sentence example