-
Notifications
You must be signed in to change notification settings - Fork 35
/
base.py
167 lines (140 loc) · 5.31 KB
/
base.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
import re
from typing import List, Union
from spacy.lang.en import English
class RCSentence:
def __init__(self, sentence_text, is_blank, start=None, end=None) -> None:
self.text = sentence_text
self.is_blank = is_blank
self.start = start
self.end = end
def __repr__(self) -> str:
return self.text
def to_dict(self):
return {'text': self.text, 'is_blank': self.is_blank, 'start': self.start, 'end': self.end}
@classmethod
def from_dict(cls, sent_dict: dict):
return cls(text=sent_dict['text'], is_blank=sent_dict['is_blank'], start=sent_dict['start'], end=sent_dict['end'])
class RCText:
def __init__(self, response_text) -> None:
self.orig_text = response_text
self.nlp = English()
self.nlp.add_pipe("sentencizer")
self.sentences = None
self._sent_id_to_index = dict()
self.sentencize()
self.indexed_response = None
def sentencize(self):
blanks = [' ', '\n', '\r']
sents = []
for s in self.nlp(self.orig_text).sents:
s_text = s.text
prefix = ''
start_idx = 0
while start_idx < len(s_text):
if s_text[start_idx] in blanks:
prefix += s_text[start_idx]
start_idx += 1
else:
start_idx = start_idx
break
if len(prefix):
sents.append(RCSentence(prefix, is_blank=True))
surfix = ''
end_idx = len(s_text) - 1
while end_idx > start_idx:
if s_text[end_idx] in blanks:
surfix = s_text[end_idx] + surfix
end_idx -= 1
else:
break
if len(s_text[start_idx: end_idx+1]):
sents.append(RCSentence(s_text[start_idx: end_idx+1], is_blank=False, start=s.start, end=s.end))
if len(surfix):
sents.append(RCSentence(surfix, is_blank=True))
self.sentences = sents
sent_id = 1
for index, sent in enumerate(self.sentences):
if not sent.is_blank:
self._sent_id_to_index[str(sent_id)] = index
sent_id += 1
def get_indexed_response(self, condense_newlines: bool):
if self.indexed_response is None:
sent_id = 1
res = ''
for i, s in enumerate(self.sentences):
sent_text = s.text
if condense_newlines:
sent_text = re.sub(r'(\n\s*)+\n', '\n', sent_text)
sent_text = re.sub(r'(\r\s*)+\r', '\r', sent_text)
if s.is_blank:
res += sent_text
else:
res += f'[{sent_id}] {sent_text}'
sent_id += 1
if i < len(self.sentences) - 1:
res += ' '
self.indexed_response = res
return self.indexed_response
def get_sentence_by_id(self, sent_id: str):
assert sent_id in self._sent_id_to_index, "Invalid sentence ID"
assert self._sent_id_to_index[sent_id] < len(self.sentences)
return self.sentences[self._sent_id_to_index[sent_id]]
def get_sentence_ids(self):
return list(self._sent_id_to_index.keys())
def to_dict(self):
return {
'sents': [s.to_dict() for s in self.sentences],
'sent_id_to_index': self._sent_id_to_index
}
class RCClaim:
def __init__(
self,
format: str,
content: Union[str, list],
attributed_sent_ids: List[str]
) -> None:
self.format = format
self.content = content
self.attributed_sent_ids = attributed_sent_ids
def __repr__(self) -> str:
if self.format == 'triplet':
return f'("{self.content[0]}", "{self.content[1]}", "{self.content[2]}")'
elif self.format == 'subsentence':
ret = self.content + ' '
for sid in self.attributed_sent_ids:
ret += f'[{sid}]'
return ret
else:
raise ValueError(f'Unknown Claim Format: {self.format}')
def get_content(self, preserve_triplet_form=False):
if self.format == 'triplet':
if preserve_triplet_form:
return f'("{self.content[0]}", "{self.content[1]}", "{self.content[2]}")'
else:
return f'{self.content[0]} {self.content[1]} {self.content[2]}'
else:
return self.content
def to_dict(self):
ret = {
'format': self.format,
'content': self.content,
'attributed_sent_ids': self.attributed_sent_ids
}
return ret
@classmethod
def from_dict(cls, claim_dict: dict):
return cls(
format=claim_dict['format'],
content=claim_dict['content'],
attributed_sent_ids=claim_dict['attributed_sent_ids']
)
class ExtractionResult:
def __init__(
self,
claims: List[RCClaim],
response: Union[str, RCText],
extractor_response: str = None,
) -> None:
self.claims = claims
self.response = response
self.extractor_response = extractor_response