Datasets:
Revert "Delete csfever_v2.py"
Browse filesThis reverts commit 4b9ba4ca445c19aa6e8037daeb7bb505d672d70c.
- csfever_v2.py +286 -0
csfever_v2.py
ADDED
@@ -0,0 +1,286 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
2 |
+
#
|
3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
4 |
+
# you may not use this file except in compliance with the License.
|
5 |
+
# You may obtain a copy of the License at
|
6 |
+
#
|
7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
8 |
+
#
|
9 |
+
# Unless required by applicable law or agreed to in writing, software
|
10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
12 |
+
# See the License for the specific language governing permissions and
|
13 |
+
# limitations under the License.
|
14 |
+
# TODO: Address all TODOs and remove all explanatory comments
|
15 |
+
"""CsFEVERv2 dataset"""
|
16 |
+
|
17 |
+
|
18 |
+
import csv
|
19 |
+
import json
|
20 |
+
import os
|
21 |
+
|
22 |
+
import datasets
|
23 |
+
|
24 |
+
|
25 |
+
# TODO: Add description of the dataset here
|
26 |
+
# You can copy an official description
|
27 |
+
_DESCRIPTION = """\
|
28 |
+
This new dataset is aimed on Czech fact-checking task.
|
29 |
+
"""
|
30 |
+
#TODO
|
31 |
+
_CITATION = ""
|
32 |
+
# TODO: Add a link to an official homepage for the dataset here
|
33 |
+
_HOMEPAGE = ""
|
34 |
+
|
35 |
+
# TODO: Add the licence for the dataset here if you can find it
|
36 |
+
_LICENSE = ""
|
37 |
+
|
38 |
+
# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
|
39 |
+
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
|
40 |
+
_URLS = {
|
41 |
+
"original": {"train": "./original/train.jsonl",
|
42 |
+
"dev" : "./original/dev.jsonl",
|
43 |
+
"test": "./original/test.jsonl"},
|
44 |
+
"f1": {"train": "./f1/train.jsonl",
|
45 |
+
"dev" : "./f1/dev.jsonl",
|
46 |
+
"test": "./f1/test.jsonl"},
|
47 |
+
"precision": {"train": "./precision/train.jsonl",
|
48 |
+
"dev" : "./precision/dev.jsonl",
|
49 |
+
"test": "./precision/test.jsonl"},
|
50 |
+
"07": {"train": "./07/train.jsonl",
|
51 |
+
"dev" : "./07/dev.jsonl",
|
52 |
+
"test": "./07/test.jsonl"},
|
53 |
+
"wiki_pages": "./wiki_pages/wiki_pages.jsonl",
|
54 |
+
"original_nli": {"train": "./original_nli/train.jsonl",
|
55 |
+
"dev" : "./original_nli/dev.jsonl",
|
56 |
+
"test": "./original_nli/test.jsonl"},
|
57 |
+
"f1_nli": {"train": "./f1_nli/train.jsonl",
|
58 |
+
"dev" : "./f1_nli/dev.jsonl",
|
59 |
+
"test": "./f1_nli/test.jsonl"},
|
60 |
+
"07_nli": {"train": "./07_nli/train.jsonl",
|
61 |
+
"dev" : "./07_nli/dev.jsonl",
|
62 |
+
"test": "./07_nli/test.jsonl"},
|
63 |
+
"precision_nli": {"train": "./precision_nli/train.jsonl",
|
64 |
+
"dev" : "./precision_nli/dev.jsonl",
|
65 |
+
"test": "./precision_nli/test.jsonl"},
|
66 |
+
}
|
67 |
+
|
68 |
+
_ORIGINAL_DESCRIPTION = ""
|
69 |
+
|
70 |
+
_NLI_NAMES = ["original_nli", "07_nli", "precision_nli", "f1_nli"]
|
71 |
+
|
72 |
+
|
73 |
+
#Name of the dataset usually matches the script name with CamelCase instead of snake_case
|
74 |
+
class CsFEVERv2(datasets.GeneratorBasedBuilder):
|
75 |
+
"""CsFEVERv2"""
|
76 |
+
|
77 |
+
VERSION = datasets.Version("1.1.0")
|
78 |
+
|
79 |
+
# This is an example of a dataset with multiple configurations.
|
80 |
+
# If you don't want/need to define several sub-sets in your dataset,
|
81 |
+
# just remove the BUILDER_CONFIG_CLASS and the BUILDER_CONFIGS attributes.
|
82 |
+
|
83 |
+
# If you need to make complex sub-parts in the datasets with configurable options
|
84 |
+
# You can create your own builder configuration class to store attribute, inheriting from datasets.BuilderConfig
|
85 |
+
# BUILDER_CONFIG_CLASS = MyBuilderConfig
|
86 |
+
|
87 |
+
# You will be able to load one or the other configurations in the following list with
|
88 |
+
# data = datasets.load_dataset('my_dataset', 'first_domain')
|
89 |
+
# data = datasets.load_dataset('my_dataset', 'second_domain')
|
90 |
+
|
91 |
+
BUILDER_CONFIGS = [
|
92 |
+
datasets.BuilderConfig(
|
93 |
+
name="original",
|
94 |
+
version=VERSION,
|
95 |
+
description=_ORIGINAL_DESCRIPTION,
|
96 |
+
),
|
97 |
+
datasets.BuilderConfig(
|
98 |
+
name="f1",
|
99 |
+
version=VERSION,
|
100 |
+
description=_ORIGINAL_DESCRIPTION,
|
101 |
+
),
|
102 |
+
datasets.BuilderConfig(
|
103 |
+
name="precision",
|
104 |
+
version=VERSION,
|
105 |
+
description=_ORIGINAL_DESCRIPTION
|
106 |
+
),
|
107 |
+
datasets.BuilderConfig(
|
108 |
+
name="07",
|
109 |
+
version=VERSION,
|
110 |
+
description=_ORIGINAL_DESCRIPTION
|
111 |
+
),
|
112 |
+
datasets.BuilderConfig(
|
113 |
+
name="wiki_pages",
|
114 |
+
version=VERSION,
|
115 |
+
description=_ORIGINAL_DESCRIPTION
|
116 |
+
),
|
117 |
+
datasets.BuilderConfig(
|
118 |
+
name="original_nli",
|
119 |
+
version=VERSION,
|
120 |
+
description=_ORIGINAL_DESCRIPTION
|
121 |
+
),
|
122 |
+
datasets.BuilderConfig(
|
123 |
+
name="07_nli",
|
124 |
+
version=VERSION,
|
125 |
+
description=_ORIGINAL_DESCRIPTION
|
126 |
+
),
|
127 |
+
datasets.BuilderConfig(
|
128 |
+
name="f1_nli",
|
129 |
+
version=VERSION,
|
130 |
+
description=_ORIGINAL_DESCRIPTION
|
131 |
+
),
|
132 |
+
datasets.BuilderConfig(
|
133 |
+
name="precision_nli",
|
134 |
+
version=VERSION,
|
135 |
+
description=_ORIGINAL_DESCRIPTION
|
136 |
+
),
|
137 |
+
]
|
138 |
+
|
139 |
+
DEFAULT_CONFIG_NAME = "original" # It's not mandatory to have a default configuration. Just use one if it make sense.
|
140 |
+
|
141 |
+
def _info(self):
|
142 |
+
#This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
|
143 |
+
if self.config.name == "original": # This is the name of the configuration selected in BUILDER_CONFIGS above
|
144 |
+
features = datasets.Features(
|
145 |
+
{
|
146 |
+
"id": datasets.Value("int32"),
|
147 |
+
"label": datasets.Value("string"),
|
148 |
+
"predicted_label": datasets.Value("string"),
|
149 |
+
"predicted_score": datasets.Value("float"),
|
150 |
+
"claim": datasets.Value("string"),
|
151 |
+
"evidence": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
|
152 |
+
# These are the features of your dataset like images, labels ...
|
153 |
+
}
|
154 |
+
)
|
155 |
+
elif self.config.name in _NLI_NAMES: # This is the name of the configuration selected in BUILDER_CONFIGS above
|
156 |
+
features = datasets.Features(
|
157 |
+
{
|
158 |
+
"id": datasets.Value("int32"),
|
159 |
+
"label": datasets.ClassLabel(num_classes=3, names=["SUPPORTS", "REFUTES", "NOT ENOUGH INFO"]),
|
160 |
+
"claim": datasets.Value("string"),
|
161 |
+
"evidence": datasets.Value("string"),
|
162 |
+
# These are the features of your dataset like images, labels ...
|
163 |
+
}
|
164 |
+
)
|
165 |
+
elif self.config.name == "wiki_pages":
|
166 |
+
features = datasets.Features(
|
167 |
+
{
|
168 |
+
"id": datasets.Value("int32"),
|
169 |
+
"revid": datasets.Value("int32"),
|
170 |
+
"url": datasets.Value("string"),
|
171 |
+
"title": datasets.Value("string"),
|
172 |
+
"text": datasets.Value("string"),
|
173 |
+
# These are the features of your dataset like images, labels ...
|
174 |
+
}
|
175 |
+
)
|
176 |
+
else: # This is an example to show how to have different features for "first_domain" and "second_domain"
|
177 |
+
features = datasets.Features(
|
178 |
+
{
|
179 |
+
"id": datasets.Value("int32"),
|
180 |
+
"label": datasets.Value("string"),
|
181 |
+
"claim": datasets.Value("string"),
|
182 |
+
"evidence": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
|
183 |
+
# These are the features of your dataset like images, labels ...
|
184 |
+
}
|
185 |
+
)
|
186 |
+
return datasets.DatasetInfo(
|
187 |
+
# This is the description that will appear on the datasets page.
|
188 |
+
description=_DESCRIPTION,
|
189 |
+
# This defines the different columns of the dataset and their types
|
190 |
+
features=features, # Here we define them above because they are different between the two configurations
|
191 |
+
# If there's a common (input, target) tuple from the features, uncomment supervised_keys line below and
|
192 |
+
# specify them. They'll be used if as_supervised=True in builder.as_dataset.
|
193 |
+
# supervised_keys=("sentence", "label"),
|
194 |
+
# Homepage of the dataset for documentation
|
195 |
+
homepage=_HOMEPAGE,
|
196 |
+
# License for the dataset if available
|
197 |
+
license=_LICENSE,
|
198 |
+
# Citation for the dataset
|
199 |
+
citation=_CITATION,
|
200 |
+
)
|
201 |
+
|
202 |
+
def _split_generators(self, dl_manager):
|
203 |
+
# This method is tasked with downloading/extracting the data and defining the splits depending on the configuration
|
204 |
+
# If several configurations are possible (listed in BUILDER_CONFIGS), the configuration selected by the user is in self.config.name
|
205 |
+
|
206 |
+
# dl_manager is a datasets.download.DownloadManager that can be used to download and extract URLS
|
207 |
+
# It can accept any type or nested list/dict and will give back the same structure with the url replaced with path to local files.
|
208 |
+
# By default the archives will be extracted and a path to a cached folder where they are extracted is returned instead of the archive
|
209 |
+
urls = _URLS[self.config.name]
|
210 |
+
data_dir = dl_manager.download_and_extract(urls)
|
211 |
+
if self.config.name == "wiki_pages":
|
212 |
+
return [datasets.SplitGenerator(
|
213 |
+
name="wiki_pages",
|
214 |
+
# These kwargs will be passed to _generate_examples
|
215 |
+
gen_kwargs={
|
216 |
+
"filepath": data_dir,
|
217 |
+
"split": "wiki_pages",
|
218 |
+
},
|
219 |
+
)]
|
220 |
+
else:
|
221 |
+
return [
|
222 |
+
datasets.SplitGenerator(
|
223 |
+
name=datasets.Split.TRAIN,
|
224 |
+
# These kwargs will be passed to _generate_examples
|
225 |
+
gen_kwargs={
|
226 |
+
"filepath": data_dir["train"],
|
227 |
+
"split": "train",
|
228 |
+
},
|
229 |
+
),
|
230 |
+
datasets.SplitGenerator(
|
231 |
+
name=datasets.Split.VALIDATION,
|
232 |
+
# These kwargs will be passed to _generate_examples
|
233 |
+
gen_kwargs={
|
234 |
+
"filepath": data_dir["dev"],
|
235 |
+
"split": "dev",
|
236 |
+
},
|
237 |
+
),
|
238 |
+
datasets.SplitGenerator(
|
239 |
+
name=datasets.Split.TEST,
|
240 |
+
# These kwargs will be passed to _generate_examples
|
241 |
+
gen_kwargs={
|
242 |
+
"filepath": data_dir["test"],
|
243 |
+
"split": "test"
|
244 |
+
},
|
245 |
+
),
|
246 |
+
]
|
247 |
+
|
248 |
+
# method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
249 |
+
def _generate_examples(self, filepath, split):
|
250 |
+
# This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
|
251 |
+
# The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
|
252 |
+
with open(filepath, encoding="utf-8") as f:
|
253 |
+
for key, row in enumerate(f):
|
254 |
+
data_point = json.loads(row)
|
255 |
+
if self.config.name == "original":
|
256 |
+
# Yields examples as (key, example) tuples
|
257 |
+
yield key, {
|
258 |
+
"id": data_point["id"],
|
259 |
+
"label": data_point["label"],
|
260 |
+
"predicted_label": data_point["predicted_label"],
|
261 |
+
"predicted_score": data_point["predicted_score"],
|
262 |
+
"claim": data_point["claim"],
|
263 |
+
"evidence": data_point["evidence"],
|
264 |
+
}
|
265 |
+
elif self.config.name in _NLI_NAMES:
|
266 |
+
yield key, {
|
267 |
+
"id": data_point["id"],
|
268 |
+
"label": data_point["label"],
|
269 |
+
"claim": data_point["claim"],
|
270 |
+
"evidence": data_point["evidence"],
|
271 |
+
}
|
272 |
+
elif self.config.name == "wiki_pages":
|
273 |
+
yield key, {
|
274 |
+
"id": data_point["id"],
|
275 |
+
"revid": data_point["revid"],
|
276 |
+
"url": data_point["url"],
|
277 |
+
"title": data_point["title"],
|
278 |
+
"text": data_point["text"],
|
279 |
+
}
|
280 |
+
else:
|
281 |
+
yield key, {
|
282 |
+
"id": data_point["id"],
|
283 |
+
"label": data_point["label"],
|
284 |
+
"claim": data_point["claim"],
|
285 |
+
"evidence": data_point["evidence"],
|
286 |
+
}
|