-
Notifications
You must be signed in to change notification settings - Fork 1
/
preprocess.py
66 lines (47 loc) · 2.02 KB
/
preprocess.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
from nltk.corpus import stopwords
import spacy
import re
from tools import logging
class Tokenizer():
def __init__(self, tokenizer_type='normal', remove_stop_words=False):
self.is_remove_stop_words = remove_stop_words
if tokenizer_type == 'normal':
self.tokenizer = self.normal_token
elif tokenizer_type == 'spacy':
self.nlp = spacy.load('en_core_web_sm')
self.tokenizer = self.spacy_token
else:
raise RuntimeError(f'Tokenizer type is error, do not have type {tokenizer_type}')
self.token_type = tokenizer_type
self.stop_words = set(stopwords.words('english'))
for w in ["<br />", '!', ',', '.', '?', '-s', '-ly', '</s>', 's', '</', '>', '/>', 'br', '<']:
self.stop_words.add(w)
logging(f'using tokenizer {tokenizer_type}, is_remove_stop_words={remove_stop_words}')
def pre_process(self, text: str)->str:
text = text.lower().strip()
text = re.sub(r"<br />", "", text)
text = re.sub(r'(\W)(?=\1)', '', text)
text = re.sub(r"([.!?,])", r" \1", text)
text = re.sub(r"[^a-zA-Z.!?]+", r" ", text)
return text.strip()
def normal_token(self, text: str, is_word=True)->[str]:
if is_word:
return [tok for tok in text.split() if not tok.isspace()]
else:
return [tok for tok in text]
def spacy_token(self, text: str, is_word=True)->[str]:
if is_word:
text = self.nlp(text)
return [token.text for token in text if not token.text.isspace()]
else:
return [tok for tok in text]
def stop_words_filter(self, words: [str]):
return [word for word in words if word not in self.stop_words]
def __call__(self, text: str, is_word=True)->[str]:
text = self.pre_process(text)
words = self.tokenizer(text, is_word=is_word)
if self.is_remove_stop_words:
return self.stop_words_filter(words)
return words
if __name__ == '__main__':
pass