holylovenia
commited on
Upload limesoda.py with huggingface_hub
Browse files- limesoda.py +172 -0
limesoda.py
ADDED
@@ -0,0 +1,172 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
from pathlib import Path
|
3 |
+
from typing import Dict, List, Tuple
|
4 |
+
|
5 |
+
import datasets
|
6 |
+
|
7 |
+
from seacrowd.utils import schemas
|
8 |
+
from seacrowd.utils.configs import SEACrowdConfig
|
9 |
+
from seacrowd.utils.constants import Licenses, Tasks
|
10 |
+
|
11 |
+
_CITATION = """\
|
12 |
+
@INPROCEEDINGS{9678187,
|
13 |
+
author={Payoungkhamdee, Patomporn and Porkaew, Peerachet and Sinthunyathum, Atthasith and Songphum, Phattharaphon and Kawidam, Witsarut and Loha-Udom, Wichayut and Boonkwan, Prachya and Sutantayawalee, Vipas},
|
14 |
+
booktitle={2021 16th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP)},
|
15 |
+
title={LimeSoda: Dataset for Fake News Detection in Healthcare Domain},
|
16 |
+
year={2021},
|
17 |
+
volume={},
|
18 |
+
number={},
|
19 |
+
pages={1-6},
|
20 |
+
doi={10.1109/iSAI-NLP54397.2021.9678187}}
|
21 |
+
"""
|
22 |
+
|
23 |
+
_DATASETNAME = "limesoda"
|
24 |
+
|
25 |
+
_DESCRIPTION = """\
|
26 |
+
Thai fake news dataset in the healthcare domain consisting of curate and manually annotated 7,191 documents
|
27 |
+
(only 4,141 documents contain token labels and are used as a test set of the baseline models).
|
28 |
+
Each document in the dataset is classified as fact, fake, or undefined.
|
29 |
+
"""
|
30 |
+
|
31 |
+
_HOMEPAGE = "https://github.com/byinth/LimeSoda"
|
32 |
+
|
33 |
+
_LICENSE = Licenses.CC_BY_4_0.value
|
34 |
+
|
35 |
+
_LANGUAGES = ["tha"]
|
36 |
+
_LOCAL = False
|
37 |
+
|
38 |
+
_URLS = {
|
39 |
+
"split": {
|
40 |
+
"train": "https://raw.githubusercontent.com/byinth/LimeSoda/main/dataset_train_wo_tokentags_v1/train_v1.jsonl",
|
41 |
+
"val": "https://raw.githubusercontent.com/byinth/LimeSoda/main/dataset_train_wo_tokentags_v1/val_v1.jsonl",
|
42 |
+
"test": "https://raw.githubusercontent.com/byinth/LimeSoda/main/dataset_train_wo_tokentags_v1/test_v1.jsonl",
|
43 |
+
},
|
44 |
+
"raw": "https://raw.githubusercontent.com/byinth/LimeSoda/main/LimeSoda/Limesoda.jsonl",
|
45 |
+
}
|
46 |
+
|
47 |
+
_SUPPORTED_TASKS = [Tasks.HOAX_NEWS_CLASSIFICATION]
|
48 |
+
|
49 |
+
_SOURCE_VERSION = "1.0.0"
|
50 |
+
|
51 |
+
_SEACROWD_VERSION = "2024.06.20"
|
52 |
+
|
53 |
+
|
54 |
+
class LimeSodaDataset(datasets.GeneratorBasedBuilder):
|
55 |
+
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
56 |
+
SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
|
57 |
+
|
58 |
+
BUILDER_CONFIGS = [
|
59 |
+
SEACrowdConfig(
|
60 |
+
name=f"{_DATASETNAME}_source",
|
61 |
+
version=SOURCE_VERSION,
|
62 |
+
description="limesoda source schema",
|
63 |
+
schema="source",
|
64 |
+
subset_id=_DATASETNAME,
|
65 |
+
),
|
66 |
+
SEACrowdConfig(
|
67 |
+
name=f"{_DATASETNAME}_split_source",
|
68 |
+
version=SOURCE_VERSION,
|
69 |
+
description="limesoda source schema",
|
70 |
+
schema="source",
|
71 |
+
subset_id=f"{_DATASETNAME}_split",
|
72 |
+
),
|
73 |
+
SEACrowdConfig(
|
74 |
+
name=f"{_DATASETNAME}_seacrowd_text",
|
75 |
+
version=SEACROWD_VERSION,
|
76 |
+
description="limesoda SEACrowd schema",
|
77 |
+
schema="seacrowd_text",
|
78 |
+
subset_id=_DATASETNAME,
|
79 |
+
),
|
80 |
+
SEACrowdConfig(
|
81 |
+
name=f"{_DATASETNAME}_split_seacrowd_text",
|
82 |
+
version=SEACROWD_VERSION,
|
83 |
+
description="limesoda: split SEACrowd schema",
|
84 |
+
schema="seacrowd_text",
|
85 |
+
subset_id=f"{_DATASETNAME}_split",
|
86 |
+
),
|
87 |
+
]
|
88 |
+
|
89 |
+
DEFAULT_CONFIG_NAME = f"{_DATASETNAME}_source"
|
90 |
+
|
91 |
+
def _info(self) -> datasets.DatasetInfo:
|
92 |
+
if self.config.schema == "source":
|
93 |
+
if self.config.subset_id == "limesoda":
|
94 |
+
features = datasets.Features(
|
95 |
+
{
|
96 |
+
"id": datasets.Value("string"),
|
97 |
+
"title": datasets.Value("string"),
|
98 |
+
"detail": datasets.Sequence(datasets.Value("string")),
|
99 |
+
"title_token_tags": datasets.Value("string"),
|
100 |
+
"detail_token_tags": datasets.Sequence(datasets.Value("string")),
|
101 |
+
"document_tag": datasets.Value("string"),
|
102 |
+
}
|
103 |
+
)
|
104 |
+
else:
|
105 |
+
features = datasets.Features({"id": datasets.Value("string"), "text": datasets.Value("string"), "document_tag": datasets.Value("string")})
|
106 |
+
elif self.config.schema == "seacrowd_text":
|
107 |
+
features = schemas.text_features(["Fact News", "Fake News", "Undefined"])
|
108 |
+
|
109 |
+
return datasets.DatasetInfo(
|
110 |
+
description=_DESCRIPTION,
|
111 |
+
features=features,
|
112 |
+
homepage=_HOMEPAGE,
|
113 |
+
license=_LICENSE,
|
114 |
+
citation=_CITATION,
|
115 |
+
)
|
116 |
+
|
117 |
+
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
118 |
+
"""Returns SplitGenerators."""
|
119 |
+
path_dict = dl_manager.download_and_extract(_URLS)
|
120 |
+
if self.config.subset_id == "limesoda":
|
121 |
+
raw_path = path_dict["raw"]
|
122 |
+
return [
|
123 |
+
datasets.SplitGenerator(
|
124 |
+
name=datasets.Split.TRAIN,
|
125 |
+
gen_kwargs={
|
126 |
+
"filepath": raw_path,
|
127 |
+
},
|
128 |
+
),
|
129 |
+
]
|
130 |
+
elif self.config.subset_id == "limesoda_split":
|
131 |
+
train_path, val_path, test_path = path_dict["split"]["train"], path_dict["split"]["val"], path_dict["split"]["test"]
|
132 |
+
return [
|
133 |
+
datasets.SplitGenerator(
|
134 |
+
name=datasets.Split.TRAIN,
|
135 |
+
gen_kwargs={
|
136 |
+
"filepath": train_path,
|
137 |
+
},
|
138 |
+
),
|
139 |
+
datasets.SplitGenerator(
|
140 |
+
name=datasets.Split.TEST,
|
141 |
+
gen_kwargs={
|
142 |
+
"filepath": test_path,
|
143 |
+
},
|
144 |
+
),
|
145 |
+
datasets.SplitGenerator(
|
146 |
+
name=datasets.Split.VALIDATION,
|
147 |
+
gen_kwargs={
|
148 |
+
"filepath": val_path,
|
149 |
+
},
|
150 |
+
),
|
151 |
+
]
|
152 |
+
|
153 |
+
def _generate_examples(self, filepath: Path) -> Tuple[int, Dict]:
|
154 |
+
with open(filepath, "r") as f:
|
155 |
+
entries = [json.loads(line) for line in f.readlines()]
|
156 |
+
if self.config.schema == "source":
|
157 |
+
if self.config.subset_id == "limesoda":
|
158 |
+
for i, row in enumerate(entries):
|
159 |
+
ex = {"id": str(i), "title": row["Title"], "detail": row["Detail"], "title_token_tags": row["Title Token Tags"], "detail_token_tags": row["Detail Token Tags"], "document_tag": row["Document Tag"]}
|
160 |
+
yield i, ex
|
161 |
+
else:
|
162 |
+
for i, row in enumerate(entries):
|
163 |
+
ex = {"id": str(i), "text": row["Text"], "document_tag": row["Document Tag"]}
|
164 |
+
yield i, ex
|
165 |
+
elif self.config.schema == "seacrowd_text":
|
166 |
+
for i, row in enumerate(entries):
|
167 |
+
ex = {
|
168 |
+
"id": str(i),
|
169 |
+
"text": row["Detail"] if self.config.subset_id == "limesoda" else row["Text"],
|
170 |
+
"label": row["Document Tag"],
|
171 |
+
}
|
172 |
+
yield i, ex
|