def getAllBoW(self):
"""Return unigram, bigram words as bag of words."""
fname = 'cache/word_ngrams.%s.set%d.dom%d.pickle' % (
self.getFilename(), self.getEssaySet(), self.getDomain())
try:
f = open(fname, 'rb')
bow = pickle.load(f)
except:
bigram_measures = nltk.collocations.BigramAssocMeasures()
bow = list()
for line in self.getRawText():
cur = LanguageUtils.tokenize(line)
finder = BigramCollocationFinder.from_words(cur)
scored = finder.score_ngrams(bigram_measures.pmi)
for bigram, score in scored:
cur.append(bigram)
bow.append(cur)
pickle.dump(bow, open(fname, 'w'))
return bow
def extractFeatures(self, ds, corpus):
"""Extracts features from a DataSet ds"""
# load into memory the string from data/essay_set_desc_?.txt
f = open('data/essay_set_desc_%d.txt' % ds.getEssaySet(), 'r')
prompt = f.read()
# tokenize into unigrams & bigrams
tokens = LanguageUtils.tokenize(prompt)
bigram_measures = nltk.collocations.BigramAssocMeasures()
finder = BigramCollocationFinder.from_words(tokens)
scored = finder.score_ngrams(bigram_measures.raw_freq)
for bigram, score in scored:
tokens.append(bigram)
# Get feature bows for projection into LSI
dictionary = corpus.getWordDictionary()
# get lsi
lsi = corpus.getLSA()
tfidf = corpus.getTfidf()
mm_corpus = ds.getGensimCorpus()
# project into lsi space
vec_bow = dictionary.doc2bow(tokens)
vec_lsi = lsi[tfidf[vec_bow]]
index = gensim.similarities.MatrixSimilarity(lsi[tfidf[ds.getGensimCorpus()]])
sims = index[vec_lsi]
feats = list()
for sim in sims:
cur_feat = list()
cur_feat.append(sim)
feats.append(cur_feat)
self.features = np.asarray(feats)
return
