holylovenia
commited on
Commit
•
618286f
1
Parent(s):
5b7ebd8
Upload utils.py with huggingface_hub
Browse files
utils.py
ADDED
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import typing as T
|
2 |
+
|
3 |
+
from conllu.exceptions import ParseException
|
4 |
+
from conllu.models import Metadata, TokenList
|
5 |
+
from conllu.parser import (DEFAULT_FIELD_PARSERS, DEFAULT_FIELDS,
|
6 |
+
_FieldParserType, _MetadataParserType,
|
7 |
+
parse_comment_line, parse_line)
|
8 |
+
|
9 |
+
imputed_sent_id: int = 1
|
10 |
+
|
11 |
+
|
12 |
+
def parse_token_and_impute_metadata(data: str, fields: T.Optional[T.Sequence[str]] = None, field_parsers: T.Optional[T.Dict[str, _FieldParserType]] = None, metadata_parsers: T.Optional[T.Dict[str, _MetadataParserType]] = None) -> TokenList:
|
13 |
+
"""
|
14 |
+
Overrides conllu.parse_token_and_metadata via monkey patching.
|
15 |
+
This function imputes the following metadata, if these are not found in the .conllu file:
|
16 |
+
- sent_id (int): an integer identifier for each sentence.
|
17 |
+
- text (str): a concatenated string of token forms. This assumes that all token forms
|
18 |
+
are separated with an empty space ' ', and does not consider the `SpaceAfter` field.
|
19 |
+
"""
|
20 |
+
|
21 |
+
if not data:
|
22 |
+
raise ParseException("Can't create TokenList, no data sent to constructor.")
|
23 |
+
|
24 |
+
fields = fields or DEFAULT_FIELDS
|
25 |
+
global imputed_sent_id
|
26 |
+
|
27 |
+
if not field_parsers:
|
28 |
+
field_parsers = DEFAULT_FIELD_PARSERS.copy()
|
29 |
+
elif sorted(field_parsers.keys()) != sorted(fields):
|
30 |
+
new_field_parsers = DEFAULT_FIELD_PARSERS.copy()
|
31 |
+
new_field_parsers.update(field_parsers)
|
32 |
+
field_parsers = new_field_parsers
|
33 |
+
|
34 |
+
tokens = []
|
35 |
+
metadata = Metadata()
|
36 |
+
|
37 |
+
for line in data.split('\n'):
|
38 |
+
line = line.strip()
|
39 |
+
|
40 |
+
if not line:
|
41 |
+
continue
|
42 |
+
|
43 |
+
if line.startswith('#'):
|
44 |
+
pairs = parse_comment_line(line, metadata_parsers=metadata_parsers)
|
45 |
+
for key, value in pairs:
|
46 |
+
metadata[key] = value
|
47 |
+
|
48 |
+
else:
|
49 |
+
tokens.append(parse_line(line, fields, field_parsers))
|
50 |
+
|
51 |
+
if 'sent_id' not in metadata:
|
52 |
+
metadata['sent_id'] = str(imputed_sent_id)
|
53 |
+
imputed_sent_id += 1
|
54 |
+
|
55 |
+
if 'text' not in metadata:
|
56 |
+
imputed_text = ""
|
57 |
+
for token in tokens:
|
58 |
+
imputed_text += str(token['form']) + " "
|
59 |
+
metadata['text'] = imputed_text
|
60 |
+
|
61 |
+
return TokenList(tokens, metadata, default_fields=fields)
|