-
Notifications
You must be signed in to change notification settings - Fork 89
/
rank_bm25.py
257 lines (219 loc) · 9.21 KB
/
rank_bm25.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
#!/usr/bin/env python
import math
import numpy as np
from multiprocessing import Pool, cpu_count
"""
All of these algorithms have been taken from the paper:
Trotmam et al, Improvements to BM25 and Language Models Examined
Here we implement all the BM25 variations mentioned.
"""
class BM25:
def __init__(self, corpus, tokenizer=None):
self.corpus_size = 0
self.avgdl = 0
self.doc_freqs = []
self.idf = {}
self.doc_len = []
self.tokenizer = tokenizer
if tokenizer:
corpus = self._tokenize_corpus(corpus)
nd = self._initialize(corpus)
self._calc_idf(nd)
def _initialize(self, corpus):
nd = {} # word -> number of documents with word
num_doc = 0
for document in corpus:
self.doc_len.append(len(document))
num_doc += len(document)
frequencies = {}
for word in document:
if word not in frequencies:
frequencies[word] = 0
frequencies[word] += 1
self.doc_freqs.append(frequencies)
for word, freq in frequencies.items():
try:
nd[word]+=1
except KeyError:
nd[word] = 1
self.corpus_size += 1
self.avgdl = num_doc / self.corpus_size
return nd
def _tokenize_corpus(self, corpus):
pool = Pool(cpu_count())
tokenized_corpus = pool.map(self.tokenizer, corpus)
return tokenized_corpus
def _calc_idf(self, nd):
raise NotImplementedError()
def get_scores(self, query):
raise NotImplementedError()
def get_batch_scores(self, query, doc_ids):
raise NotImplementedError()
def get_top_n(self, query, documents, n=5):
assert self.corpus_size == len(documents), "The documents given don't match the index corpus!"
scores = self.get_scores(query)
top_n = np.argsort(scores)[::-1][:n]
return [documents[i] for i in top_n]
class BM25Okapi(BM25):
def __init__(self, corpus, tokenizer=None, k1=1.5, b=0.75, epsilon=0.25):
self.k1 = k1
self.b = b
self.epsilon = epsilon
super().__init__(corpus, tokenizer)
def _calc_idf(self, nd):
"""
Calculates frequencies of terms in documents and in corpus.
This algorithm sets a floor on the idf values to eps * average_idf
"""
# collect idf sum to calculate an average idf for epsilon value
idf_sum = 0
# collect words with negative idf to set them a special epsilon value.
# idf can be negative if word is contained in more than half of documents
negative_idfs = []
for word, freq in nd.items():
idf = math.log(self.corpus_size - freq + 0.5) - math.log(freq + 0.5)
self.idf[word] = idf
idf_sum += idf
if idf < 0:
negative_idfs.append(word)
self.average_idf = idf_sum / len(self.idf)
eps = self.epsilon * self.average_idf
for word in negative_idfs:
self.idf[word] = eps
def get_scores(self, query):
"""
The ATIRE BM25 variant uses an idf function which uses a log(idf) score. To prevent negative idf scores,
this algorithm also adds a floor to the idf value of epsilon.
See [Trotman, A., X. Jia, M. Crane, Towards an Efficient and Effective Search Engine] for more info
:param query:
:return:
"""
score = np.zeros(self.corpus_size)
doc_len = np.array(self.doc_len)
for q in query:
q_freq = np.array([(doc.get(q) or 0) for doc in self.doc_freqs])
score += (self.idf.get(q) or 0) * (q_freq * (self.k1 + 1) /
(q_freq + self.k1 * (1 - self.b + self.b * doc_len / self.avgdl)))
return score
def get_batch_scores(self, query, doc_ids):
"""
Calculate bm25 scores between query and subset of all docs
"""
assert all(di < len(self.doc_freqs) for di in doc_ids)
score = np.zeros(len(doc_ids))
doc_len = np.array(self.doc_len)[doc_ids]
for q in query:
q_freq = np.array([(self.doc_freqs[di].get(q) or 0) for di in doc_ids])
score += (self.idf.get(q) or 0) * (q_freq * (self.k1 + 1) /
(q_freq + self.k1 * (1 - self.b + self.b * doc_len / self.avgdl)))
return score.tolist()
class BM25L(BM25):
def __init__(self, corpus, tokenizer=None, k1=1.5, b=0.75, delta=0.5):
# Algorithm specific parameters
self.k1 = k1
self.b = b
self.delta = delta
super().__init__(corpus, tokenizer)
def _calc_idf(self, nd):
for word, freq in nd.items():
idf = math.log(self.corpus_size + 1) - math.log(freq + 0.5)
self.idf[word] = idf
def get_scores(self, query):
score = np.zeros(self.corpus_size)
doc_len = np.array(self.doc_len)
for q in query:
q_freq = np.array([(doc.get(q) or 0) for doc in self.doc_freqs])
ctd = q_freq / (1 - self.b + self.b * doc_len / self.avgdl)
score += (self.idf.get(q) or 0) * (self.k1 + 1) * (ctd + self.delta) / \
(self.k1 + ctd + self.delta)
return score
def get_batch_scores(self, query, doc_ids):
"""
Calculate bm25 scores between query and subset of all docs
"""
assert all(di < len(self.doc_freqs) for di in doc_ids)
score = np.zeros(len(doc_ids))
doc_len = np.array(self.doc_len)[doc_ids]
for q in query:
q_freq = np.array([(self.doc_freqs[di].get(q) or 0) for di in doc_ids])
ctd = q_freq / (1 - self.b + self.b * doc_len / self.avgdl)
score += (self.idf.get(q) or 0) * (self.k1 + 1) * (ctd + self.delta) / \
(self.k1 + ctd + self.delta)
return score.tolist()
class BM25Plus(BM25):
def __init__(self, corpus, tokenizer=None, k1=1.5, b=0.75, delta=1):
# Algorithm specific parameters
self.k1 = k1
self.b = b
self.delta = delta
super().__init__(corpus, tokenizer)
def _calc_idf(self, nd):
for word, freq in nd.items():
idf = math.log(self.corpus_size + 1) - math.log(freq)
self.idf[word] = idf
def get_scores(self, query):
score = np.zeros(self.corpus_size)
doc_len = np.array(self.doc_len)
for q in query:
q_freq = np.array([(doc.get(q) or 0) for doc in self.doc_freqs])
score += (self.idf.get(q) or 0) * (self.delta + (q_freq * (self.k1 + 1)) /
(self.k1 * (1 - self.b + self.b * doc_len / self.avgdl) + q_freq))
return score
def get_batch_scores(self, query, doc_ids):
"""
Calculate bm25 scores between query and subset of all docs
"""
assert all(di < len(self.doc_freqs) for di in doc_ids)
score = np.zeros(len(doc_ids))
doc_len = np.array(self.doc_len)[doc_ids]
for q in query:
q_freq = np.array([(self.doc_freqs[di].get(q) or 0) for di in doc_ids])
score += (self.idf.get(q) or 0) * (self.delta + (q_freq * (self.k1 + 1)) /
(self.k1 * (1 - self.b + self.b * doc_len / self.avgdl) + q_freq))
return score.tolist()
# BM25Adpt and BM25T are a bit more complicated than the previous algorithms here. Here a term-specific k1
# parameter is calculated before scoring is done
# class BM25Adpt(BM25):
# def __init__(self, corpus, k1=1.5, b=0.75, delta=1):
# # Algorithm specific parameters
# self.k1 = k1
# self.b = b
# self.delta = delta
# super().__init__(corpus)
#
# def _calc_idf(self, nd):
# for word, freq in nd.items():
# idf = math.log((self.corpus_size + 1) / freq)
# self.idf[word] = idf
#
# def get_scores(self, query):
# score = np.zeros(self.corpus_size)
# doc_len = np.array(self.doc_len)
# for q in query:
# q_freq = np.array([(doc.get(q) or 0) for doc in self.doc_freqs])
# score += (self.idf.get(q) or 0) * (self.delta + (q_freq * (self.k1 + 1)) /
# (self.k1 * (1 - self.b + self.b * doc_len / self.avgdl) + q_freq))
# return score
#
#
# class BM25T(BM25):
# def __init__(self, corpus, k1=1.5, b=0.75, delta=1):
# # Algorithm specific parameters
# self.k1 = k1
# self.b = b
# self.delta = delta
# super().__init__(corpus)
#
# def _calc_idf(self, nd):
# for word, freq in nd.items():
# idf = math.log((self.corpus_size + 1) / freq)
# self.idf[word] = idf
#
# def get_scores(self, query):
# score = np.zeros(self.corpus_size)
# doc_len = np.array(self.doc_len)
# for q in query:
# q_freq = np.array([(doc.get(q) or 0) for doc in self.doc_freqs])
# score += (self.idf.get(q) or 0) * (self.delta + (q_freq * (self.k1 + 1)) /
# (self.k1 * (1 - self.b + self.b * doc_len / self.avgdl) + q_freq))
# return score