holylovenia
commited on
Commit
•
4e22ee4
1
Parent(s):
7587130
Upload x_fact.py with huggingface_hub
Browse files
x_fact.py
ADDED
@@ -0,0 +1,164 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from pathlib import Path
|
2 |
+
from typing import Dict, List, Tuple
|
3 |
+
|
4 |
+
import datasets
|
5 |
+
|
6 |
+
from nusacrowd.nusa_datasets.x_fact.utils.x_fact_utils import \
|
7 |
+
load_x_fact_dataset
|
8 |
+
from nusacrowd.utils.configs import NusantaraConfig
|
9 |
+
from nusacrowd.utils.constants import Tasks
|
10 |
+
|
11 |
+
_CITATION = """\
|
12 |
+
@inproceedings{gupta2021xfact,
|
13 |
+
title={{X-FACT: A New Benchmark Dataset for Multilingual Fact Checking}},
|
14 |
+
author={Gupta, Ashim and Srikumar, Vivek},
|
15 |
+
booktitle = "Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics",
|
16 |
+
month = jul,
|
17 |
+
year = "2021",
|
18 |
+
address = "Online",
|
19 |
+
publisher = "Association for Computational Linguistics",
|
20 |
+
}
|
21 |
+
"""
|
22 |
+
_DATASETNAME = "x_fact"
|
23 |
+
|
24 |
+
_DESCRIPTION = """\
|
25 |
+
X-FACT: the largest publicly available multilingual dataset for factual verification of naturally existing realworld claims.
|
26 |
+
"""
|
27 |
+
|
28 |
+
_HOMEPAGE = "https://github.com/utahnlp/x-fact"
|
29 |
+
|
30 |
+
_LANGUAGES = [
|
31 |
+
'ara', 'aze', 'ben', 'deu', 'spa',
|
32 |
+
'fas', 'fra', 'guj', 'hin', 'ind',
|
33 |
+
'ita', 'kat', 'mar', 'nor', 'nld',
|
34 |
+
'pan', 'pol', 'por', 'ron', 'rus',
|
35 |
+
'sin', 'srp', 'sqi', 'tam', 'tur'
|
36 |
+
]
|
37 |
+
_LOCAL = False
|
38 |
+
|
39 |
+
_LICENSE = "MIT"
|
40 |
+
|
41 |
+
_URLS = {
|
42 |
+
"train": "https://raw.githubusercontent.com/utahnlp/x-fact/main/data/x-fact-including-en/train.all.tsv",
|
43 |
+
"validation": "https://raw.githubusercontent.com/utahnlp/x-fact/main/data/x-fact-including-en/dev.all.tsv",
|
44 |
+
"test": {
|
45 |
+
"in_domain": "https://raw.githubusercontent.com/utahnlp/x-fact/main/data/x-fact-including-en/test.all.tsv",
|
46 |
+
"out_domain": "https://raw.githubusercontent.com/utahnlp/x-fact/main/data/x-fact-including-en/ood.tsv",
|
47 |
+
},
|
48 |
+
}
|
49 |
+
|
50 |
+
_SUPPORTED_TASKS = [Tasks.FACT_CHECKING]
|
51 |
+
|
52 |
+
_SOURCE_VERSION = "1.0.0"
|
53 |
+
|
54 |
+
_NUSANTARA_VERSION = "1.0.0"
|
55 |
+
|
56 |
+
|
57 |
+
class XFact(datasets.GeneratorBasedBuilder):
|
58 |
+
"""X-FACT: the largest publicly available multilingual dataset for factual verification of naturally existing realworld claims."""
|
59 |
+
|
60 |
+
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
61 |
+
NUSANTARA_VERSION = datasets.Version(_NUSANTARA_VERSION)
|
62 |
+
|
63 |
+
BUILDER_CONFIGS = [
|
64 |
+
NusantaraConfig(
|
65 |
+
name="x_fact_source",
|
66 |
+
version=SOURCE_VERSION,
|
67 |
+
description="x_fact source schema",
|
68 |
+
schema="source",
|
69 |
+
subset_id="x_fact",
|
70 |
+
),
|
71 |
+
]
|
72 |
+
|
73 |
+
DEFAULT_CONFIG_NAME = "x_fact_source"
|
74 |
+
|
75 |
+
def _info(self) -> datasets.DatasetInfo:
|
76 |
+
|
77 |
+
if self.config.schema == "source":
|
78 |
+
features = datasets.Features(
|
79 |
+
{
|
80 |
+
"language": datasets.Value("string"),
|
81 |
+
"site": datasets.Value("string"),
|
82 |
+
"evidence_1": datasets.Value("string"),
|
83 |
+
"evidence_2": datasets.Value("string"),
|
84 |
+
"evidence_3": datasets.Value("string"),
|
85 |
+
"evidence_4": datasets.Value("string"),
|
86 |
+
"evidence_5": datasets.Value("string"),
|
87 |
+
"link_1": datasets.Value("string"),
|
88 |
+
"link_2": datasets.Value("string"),
|
89 |
+
"link_3": datasets.Value("string"),
|
90 |
+
"link_4": datasets.Value("string"),
|
91 |
+
"link_5": datasets.Value("string"),
|
92 |
+
"claimDate": datasets.Value("string"),
|
93 |
+
"reviewDate": datasets.Value("string"),
|
94 |
+
"claimant": datasets.Value("string"),
|
95 |
+
"claim": datasets.Value("string"),
|
96 |
+
"label": datasets.Value("string"),
|
97 |
+
}
|
98 |
+
)
|
99 |
+
|
100 |
+
return datasets.DatasetInfo(
|
101 |
+
description=_DESCRIPTION,
|
102 |
+
features=features,
|
103 |
+
homepage=_HOMEPAGE,
|
104 |
+
license=_LICENSE,
|
105 |
+
citation=_CITATION,
|
106 |
+
)
|
107 |
+
|
108 |
+
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
109 |
+
return [
|
110 |
+
datasets.SplitGenerator(
|
111 |
+
name=datasets.Split.TRAIN,
|
112 |
+
gen_kwargs={
|
113 |
+
"filepath": _URLS["train"],
|
114 |
+
"split": "train",
|
115 |
+
},
|
116 |
+
),
|
117 |
+
datasets.SplitGenerator(
|
118 |
+
name=datasets.Split.VALIDATION,
|
119 |
+
gen_kwargs={
|
120 |
+
"filepath": _URLS["validation"],
|
121 |
+
"split": "dev",
|
122 |
+
},
|
123 |
+
),
|
124 |
+
datasets.SplitGenerator(
|
125 |
+
name=datasets.splits.NamedSplit("TEST_IN_DOMAIN"),
|
126 |
+
gen_kwargs={
|
127 |
+
"filepath": _URLS["test"]["in_domain"],
|
128 |
+
"split": "test_in_domain",
|
129 |
+
},
|
130 |
+
),
|
131 |
+
datasets.SplitGenerator(
|
132 |
+
name=datasets.splits.NamedSplit("TEST_OUT_DOMAIN"),
|
133 |
+
gen_kwargs={
|
134 |
+
"filepath": _URLS["test"]["out_domain"],
|
135 |
+
"split": "test_out_domain",
|
136 |
+
},
|
137 |
+
),
|
138 |
+
]
|
139 |
+
|
140 |
+
def _generate_examples(self, filepath: Path, split: str) -> Tuple[int, Dict]:
|
141 |
+
|
142 |
+
df = load_x_fact_dataset(filepath)
|
143 |
+
if self.config.schema == "source":
|
144 |
+
for row in df.itertuples():
|
145 |
+
entry = {
|
146 |
+
"language": row.language,
|
147 |
+
"site": row.site,
|
148 |
+
"evidence_1": row.evidence_1,
|
149 |
+
"evidence_2": row.evidence_2,
|
150 |
+
"evidence_3": row.evidence_3,
|
151 |
+
"evidence_4": row.evidence_4,
|
152 |
+
"evidence_5": row.evidence_5,
|
153 |
+
"link_1": row.link_1,
|
154 |
+
"link_2": row.link_2,
|
155 |
+
"link_3": row.link_3,
|
156 |
+
"link_4": row.link_4,
|
157 |
+
"link_5": row.link_5,
|
158 |
+
"claimDate": row.claimDate,
|
159 |
+
"reviewDate": row.reviewDate,
|
160 |
+
"claimant": row.claimant,
|
161 |
+
"claim": row.claim,
|
162 |
+
"label": row.label,
|
163 |
+
}
|
164 |
+
yield row.index, entry
|