utils.py
1.84 KB
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
from collections import Counter
from scipy.sparse import csr_matrix
import numpy as np
def scan_vocabulary(sents, tokenize=None, min_count=2):
"""
Arguments
---------
sents : list of str
Sentence list
tokenize : callable
tokenize(str) returns list of str
min_count : int
Minumum term frequency
Returns
-------
idx_to_vocab : list of str
Vocabulary list
vocab_to_idx : dict
Vocabulary to index mapper.
"""
counter = Counter(w for sent in sents for w in tokenize(sent))
counter = {w:c for w,c in counter.items() if c >= min_count}
idx_to_vocab = [w for w, _ in sorted(counter.items(), key=lambda x:-x[1])]
vocab_to_idx = {vocab:idx for idx, vocab in enumerate(idx_to_vocab)}
return idx_to_vocab, vocab_to_idx
def tokenize_sents(sents, tokenize):
"""
Arguments
---------
sents : list of str
Sentence list
tokenize : callable
tokenize(sent) returns list of str (word sequence)
Returns
-------
tokenized sentence list : list of list of str
"""
return [tokenize(sent) for sent in sents]
def vectorize(tokens, vocab_to_idx):
"""
Arguments
---------
tokens : list of list of str
Tokenzed sentence list
vocab_to_idx : dict
Vocabulary to index mapper
Returns
-------
sentence bow : scipy.sparse.csr_matrix
shape = (n_sents, n_terms)
"""
rows, cols, data = [], [], []
for i, tokens_i in enumerate(tokens):
for t, c in Counter(tokens_i).items():
j = vocab_to_idx.get(t, -1)
if j == -1:
continue
rows.append(i)
cols.append(j)
data.append(c)
n_sents = len(tokens)
n_terms = len(vocab_to_idx)
x = csr_matrix((data, (rows, cols)), shape=(n_sents, n_terms))
return x