ramonachristen
commited on
Commit
•
959866c
1
Parent(s):
0b78b08
Create MultiLegalNeg.py
Browse files- MultiLegalNeg.py +120 -0
MultiLegalNeg.py
ADDED
@@ -0,0 +1,120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
|
3 |
+
import datasets
|
4 |
+
import pandas as pd
|
5 |
+
from huggingface_hub.file_download import hf_hub_url
|
6 |
+
|
7 |
+
try:
|
8 |
+
import lzma as xz
|
9 |
+
except ImportError:
|
10 |
+
import pylzma as xz
|
11 |
+
|
12 |
+
datasets.logging.set_verbosity_info()
|
13 |
+
logger = datasets.logging.get_logger(__name__)
|
14 |
+
|
15 |
+
_DESCRIPTION ="""\
|
16 |
+
|
17 |
+
"""
|
18 |
+
|
19 |
+
_HOMEPAGE = ""
|
20 |
+
|
21 |
+
_LICENSE = ""
|
22 |
+
|
23 |
+
_CITATION = ""
|
24 |
+
|
25 |
+
_URL = {
|
26 |
+
'data/'
|
27 |
+
}
|
28 |
+
_LANGUAGES = [
|
29 |
+
"german", "french"
|
30 |
+
]
|
31 |
+
|
32 |
+
|
33 |
+
class MultiLegalNegConfig(datasets.BuilderConfig):
|
34 |
+
|
35 |
+
def __init__(self, name:str, **kwargs):
|
36 |
+
super( MultiLegalNegConfig, self).__init__(**kwargs)
|
37 |
+
self.name = name
|
38 |
+
self.language = name.split("_")[0]
|
39 |
+
|
40 |
+
class MultiLegalNeg(datasets.GeneratorBasedBuilder):
|
41 |
+
|
42 |
+
BUILDER_CONFIG_CLASS = MultiLegalNegConfig
|
43 |
+
|
44 |
+
BUILDER_CONFIGS = [
|
45 |
+
MultiLegalNegConfig(f"{language}")
|
46 |
+
for language in _LANGUAGES + ['all']
|
47 |
+
]
|
48 |
+
DEFAULT_CONFIG_NAME = 'all_all'
|
49 |
+
def _info(self):
|
50 |
+
features = datasets.Features(
|
51 |
+
{
|
52 |
+
"text": datasets.Value("string"),
|
53 |
+
"spans": [
|
54 |
+
{
|
55 |
+
"start": datasets.Value("int64"),
|
56 |
+
"end": datasets.Value("int64"),
|
57 |
+
"token_start": datasets.Value("int64"),
|
58 |
+
"token_end": datasets.Value("int64"),
|
59 |
+
"label": datasets.Value("string"),
|
60 |
+
"text": datasets.Value("string"),
|
61 |
+
"pattern": datasets.Value("int64")
|
62 |
+
}
|
63 |
+
],
|
64 |
+
"tokens": [
|
65 |
+
{
|
66 |
+
"text": datasets.Value("string"),
|
67 |
+
"start": datasets.Value("int64"),
|
68 |
+
"end": datasets.Value("int64"),
|
69 |
+
"id": datasets.Value("int64"),
|
70 |
+
"ws": datasets.Value("bool")
|
71 |
+
}
|
72 |
+
],
|
73 |
+
"multineg": datasets.Value("bool")
|
74 |
+
}
|
75 |
+
)
|
76 |
+
return datasets.DatasetInfo(
|
77 |
+
description=_DESCRIPTION,
|
78 |
+
features = features,
|
79 |
+
homepage = _HOMEPAGE,
|
80 |
+
citation=_CITATION
|
81 |
+
)
|
82 |
+
|
83 |
+
def _split_generators(self, dl_manager):
|
84 |
+
|
85 |
+
languages = _LANGUAGES if self.config.language == "all" else [self.config.language]
|
86 |
+
|
87 |
+
split_generators = []
|
88 |
+
for split in [datasets.Split.TRAIN]:
|
89 |
+
filepaths = []
|
90 |
+
for language in languages:
|
91 |
+
try:
|
92 |
+
filepaths.append(dl_manager.download((f'data/{language}.jsonl.xz')))
|
93 |
+
except:
|
94 |
+
break
|
95 |
+
|
96 |
+
split_generators.append(
|
97 |
+
datasets.SplitGenerator(name=split, gen_kwargs={'filepaths': filepaths})
|
98 |
+
)
|
99 |
+
|
100 |
+
return split_generators
|
101 |
+
|
102 |
+
def _generate_examples(self,filepaths):
|
103 |
+
id_ = 0
|
104 |
+
for filepath in filepaths:
|
105 |
+
if filepath:
|
106 |
+
logger.info("Generating examples from = %s", filepath)
|
107 |
+
try:
|
108 |
+
with xz.open(open(filepath,'rb'), 'rt', encoding='utf-8') as f:
|
109 |
+
json_list = list(f)
|
110 |
+
|
111 |
+
for json_str in json_list:
|
112 |
+
example = json.loads(json_str)
|
113 |
+
if example is not None and isinstance(example, dict):
|
114 |
+
yield id_, example
|
115 |
+
id_ +=1
|
116 |
+
|
117 |
+
except Exception:
|
118 |
+
logger.exception("Error while processing file %s", filepath)
|
119 |
+
|
120 |
+
|