-
Notifications
You must be signed in to change notification settings - Fork 0
/
summarizer.py
281 lines (230 loc) · 10.6 KB
/
summarizer.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
import numpy as np
import nltk
import os
import re
from nltk.tokenize import sent_tokenize, word_tokenize
from nltk.corpus import stopwords as sw
from nltk.cluster.util import cosine_distance
from nltk.stem.porter import PorterStemmer
from operator import itemgetter
from Sastrawi.Stemmer.StemmerFactory import StemmerFactory
from scraper import Scraper
from scipy.sparse.linalg import svds
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
TEXT_RANK = 'text_rank'
INDONESIAN = 'indonesian'
ENGLISH = 'english'
class TextRankSummarizer(object):
def __init__(self, language):
dir_path = os.path.dirname(os.path.realpath(__file__)) + '/nltk_data/'
nltk.data.path = [dir_path]
self.stopwords = sw.words(language)
self.scraper = Scraper(language)
self.language = language
if language == INDONESIAN:
factory = StemmerFactory()
self.stemmer = factory.create_stemmer()
else:
self.stemmer = PorterStemmer()
def sentence_similarity(self, sentence1, sentence2):
if self.stopwords is None:
self.stopwords = []
sentence1 = [self.stemmer.stem(w.lower()) for w in word_tokenize(sentence1)]
sentence2 = [self.stemmer.stem(w.lower()) for w in word_tokenize(sentence2)]
all_words = list(set(sentence1 + sentence2))
vector1 = [0] * len(all_words)
vector2 = [0] * len(all_words)
for w in sentence1:
if w in self.stopwords:
continue
vector1[all_words.index(w)] += 1
for w in sentence2:
if w in self.stopwords:
continue
vector2[all_words.index(w)] += 1
return 1 - cosine_distance(vector1, vector2)
def build_similarity_matrix(self, sentences):
similarity_matrix = np.zeros((len(sentences), len(sentences)))
for idx1 in range(len(sentences)):
for idx2 in range(len(sentences)):
if idx1 == idx2:
continue
similarity_matrix[idx1][idx2] = self.sentence_similarity(sentences[idx1], sentences[idx2])
for idx in range(len(similarity_matrix)):
if similarity_matrix[idx].sum() != 0:
similarity_matrix[idx] /= similarity_matrix[idx].sum()
return similarity_matrix
def page_rank(self, similarity_matrix, eps=0.0001, d=0.85):
probs = np.ones(len(similarity_matrix)) / len(similarity_matrix)
while True:
new_probs = np.ones(len(similarity_matrix)) * (1 - d) / len(
similarity_matrix) + d * similarity_matrix.T.dot(probs)
delta = abs((new_probs - probs).sum())
if delta <= eps:
return new_probs
probs = new_probs
def summarize(self, query=None, size=1, text=None):
suggested_query = None
lang = None
status = 0
if query:
for ch in ['&',':','-','+','.',',']:
query = query.replace(ch,' ')
words = word_tokenize(query.lower())
filtered_words = [word for word in words if word not in self.stopwords and word.isalnum()]
new_query = " ".join(filtered_words)
suggested_query, status, lang = self.scraper.get_query(new_query)
if status == -1:
suggested_query, status, lang = self.scraper.get_query(new_query,isInverse=True)
if status == -1:
suggested_query, status, lang = self.scraper.get_query(new_query)
text = text if text else self.scraper.get_intro_lang(suggested_query, lang)
# remove formula notation and multiple spaces
text = re.sub('{.+}', '', text)
text = re.sub('\s+', ' ', text)
if not text:
if self.language == INDONESIAN:
return "mohon maaf {q} tidak ditemukan".format(q=query)
else:
return "{q} not found".format(q=query)
sentences = sent_tokenize(text)
similarity_matrix = self.build_similarity_matrix(sentences)
sentence_ranks = self.page_rank(similarity_matrix)
ranked_sentence_indexes = [item[0] for item in sorted(enumerate(sentence_ranks), key=lambda item: -item[1])]
selected_sentences = sorted(ranked_sentence_indexes[:size])
summary = itemgetter(*selected_sentences)(sentences)
if isinstance(summary, tuple):
if status == 0:
return ' '.join(summary)
elif lang == self.language:
res = ' '.join(summary)
if lang == INDONESIAN:
return "mungkin maksud anda adalah {sq}\n{s}".format(sq=suggested_query, s=res)
else:
return "maybe this is what you want {sq}\n{s}".format(sq=suggested_query, s=res)
else:
return summary
if status == 0:
return summary
elif lang == self.language:
if lang == INDONESIAN:
return "mungkin maksud anda adalah {sq}\n{s}".format(sq=suggested_query, s=summary)
else:
return "maybe this is what you want {sq}\n{s}".format(sq=suggested_query, s=summary)
else:
return summary
class LSASumarizer():
def __init__(self, language):
dir_path = os.path.dirname(os.path.realpath(__file__)) + '/nltk_data/'
nltk.data.path = [dir_path]
self.stopwords = sw.words(language)
if self.stopwords is None:
self.stopwords = []
self.scraper = Scraper(language)
self.language = language
if language == INDONESIAN:
factory = StemmerFactory()
self.stemmer = factory.create_stemmer()
else:
self.stemmer = PorterStemmer()
def _build_feature_matrix(self, documents, feature_type='frequency'):
def _tokenize(sentence):
return [self.stemmer.stem(w.lower()) for w in word_tokenize(sentence)]
feature_type = feature_type.lower().strip()
if feature_type == 'binary':
vectorizer = CountVectorizer(binary=True,
min_df=1,
stop_words=self.stopwords,
tokenizer=_tokenize,
ngram_range=(1, 2))
elif feature_type == 'frequency':
vectorizer = CountVectorizer(binary=False,
min_df=1,
analyzer= 'word',
stop_words=self.stopwords,
tokenizer=_tokenize,
ngram_range=(1, 2))
elif feature_type == 'tfidf':
vectorizer = TfidfVectorizer(min_df=1,
stop_words=self.stopwords,
tokenizer=_tokenize,
ngram_range=(1, 2))
else:
raise Exception("Wrong feature type entered. Possible values: 'binary', 'frequency', 'tfidf'")
feature_matrix = vectorizer.fit_transform(documents).astype(float)
return vectorizer, feature_matrix
def _low_rank_svd(self, matrix, singular_count=2):
u, s, vt = svds(matrix, k=singular_count)
return u, s, vt
def _summarize(self, document, num_sentences=2,
num_topics=1, feature_type='frequency',
sv_threshold=0.5):
sentences = sent_tokenize(document)
vec, dt_matrix = self._build_feature_matrix(sentences,
feature_type=feature_type)
td_matrix = dt_matrix.transpose()
td_matrix = td_matrix.multiply(td_matrix > 0)
u, s, vt = self._low_rank_svd(td_matrix, singular_count=num_topics)
min_sigma_value = max(s) * sv_threshold
s[s < min_sigma_value] = 0
salience_scores = np.sqrt(np.dot(np.square(s), np.square(vt)))
top_sentence_indices = salience_scores.argsort()[-num_sentences:][::-1]
top_sentence_indices.sort()
result = ""
for index in top_sentence_indices:
result += sentences[index]
return result
def summarize(self, query=None, size=2, text=None):
suggested_query = None
lang = None
status = 0
if query:
for ch in ['&',':','-','+','.',',']:
query = query.replace(ch,' ')
query = re.sub('[^ 0-9a-zA-Z]+', '', query)
words = word_tokenize(query.lower())
filtered_words = [word for word in words if word not in self.stopwords and word.isalnum()]
new_query = " ".join(filtered_words)
suggested_query, status, lang = self.scraper.get_query(new_query)
if status == -1:
suggested_query, status, lang = self.scraper.get_query(new_query,isInverse=True)
if status == -1:
suggested_query, status, lang = self.scraper.get_query(new_query)
text = text if text else self.scraper.get_intro_lang(suggested_query, lang)
# remove formula notation and multiple spaces
text = re.sub('{.+}', '', text)
text = re.sub('\s+', ' ', text)
if not text:
if self.language == INDONESIAN:
return "mohon maaf {q} tidak ditemukan".format(q=query)
else:
return "{q} not found".format(q=query)
summary = self._summarize(text, size)
if status == 0:
return summary
elif lang == self.language:
if lang == INDONESIAN:
return "mungkin maksud anda adalah {sq}\n{s}".format(sq=suggested_query, s=summary)
else:
return "maybe this is what you want {sq}\n{s}".format(sq=suggested_query, s=summary)
else:
return summary
class Summarizer():
def __init__(self, method=TEXT_RANK):
if method == TEXT_RANK:
self.english_summarizer = TextRankSummarizer(ENGLISH)
self.indonesian_summarizer = TextRankSummarizer(INDONESIAN)
else:
self.english_summarizer = LSASumarizer(ENGLISH)
self.indonesian_summarizer = LSASumarizer(INDONESIAN)
def summarize(self, language, size, query=None, text=None):
if language == INDONESIAN:
try:
return self.indonesian_summarizer.summarize(query, size, text)
except:
return "saya tidak paham maksud anda :("
else:
try:
return self.english_summarizer.summarize(query, size, text)
except:
return "I can't understand :("