Commit
·
68d33f4
1
Parent(s):
308ce55
Delete CSMD.py
Browse files
CSMD.py
DELETED
@@ -1,179 +0,0 @@
|
|
1 |
-
# coding=utf-8
|
2 |
-
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
3 |
-
#
|
4 |
-
# Licensed under the Apache License, Version 2.0 (the "License");
|
5 |
-
# you may not use this file except in compliance with the License.
|
6 |
-
# You may obtain a copy of the License at
|
7 |
-
#
|
8 |
-
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
-
#
|
10 |
-
# Unless required by applicable law or agreed to in writing, software
|
11 |
-
# distributed under the License is distributed on an "AS IS" BASIS,
|
12 |
-
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
-
# See the License for the specific language governing permissions and
|
14 |
-
# limitations under the License.
|
15 |
-
"""CSMD: a dataset for assessing meaning preservation between sentences"""
|
16 |
-
|
17 |
-
import csv
|
18 |
-
|
19 |
-
import datasets
|
20 |
-
from datasets import load_dataset
|
21 |
-
|
22 |
-
_CITATION = """\
|
23 |
-
@ARTICLE{10.3389/frai.2023.1223924,
|
24 |
-
AUTHOR={Beauchemin, David and Saggion, Horacio and Khoury, Richard},
|
25 |
-
TITLE={{MeaningBERT: Assessing Meaning Preservation Between Sentences}},
|
26 |
-
JOURNAL={Frontiers in Artificial Intelligence},
|
27 |
-
VOLUME={6},
|
28 |
-
YEAR={2023},
|
29 |
-
URL={https://www.frontiersin.org/articles/10.3389/frai.2023.1223924},
|
30 |
-
DOI={10.3389/frai.2023.1223924},
|
31 |
-
ISSN={2624-8212},
|
32 |
-
}
|
33 |
-
"""
|
34 |
-
|
35 |
-
_DESCRIPTION = """\
|
36 |
-
Continuous Scale Meaning Dataset (CSMD) is a dataset for assessing meaning preservation between sentences.
|
37 |
-
"""
|
38 |
-
|
39 |
-
_HOMEPAGE = "https://github.com/GRAAL-Research/csmd"
|
40 |
-
|
41 |
-
_LICENSE = "Attribution 4.0 International (CC BY 4.0)"
|
42 |
-
|
43 |
-
_URL_LIST = [
|
44 |
-
(
|
45 |
-
"meaning.train",
|
46 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning/train.tsv",
|
47 |
-
),
|
48 |
-
(
|
49 |
-
"meaning.dev",
|
50 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning/dev.tsv",
|
51 |
-
),
|
52 |
-
(
|
53 |
-
"meaning.test",
|
54 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning/test.tsv",
|
55 |
-
),
|
56 |
-
(
|
57 |
-
"meaning_with_data_augmentation.train",
|
58 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning_with_data_augmentation/train.tsv",
|
59 |
-
),
|
60 |
-
(
|
61 |
-
"meaning_with_data_augmentation.dev",
|
62 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning_with_data_augmentation/dev.tsv",
|
63 |
-
),
|
64 |
-
(
|
65 |
-
"meaning_with_data_augmentation.test",
|
66 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning_with_data_augmentation/test.tsv",
|
67 |
-
),
|
68 |
-
(
|
69 |
-
"identical",
|
70 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/holdout/identical.tsv",
|
71 |
-
),
|
72 |
-
(
|
73 |
-
"unrelated",
|
74 |
-
"https://github.com/GRAAL-Research/csmd/blob/main/dataset/holdout/unrelated.tsv",
|
75 |
-
),
|
76 |
-
]
|
77 |
-
|
78 |
-
_URLs = dict(_URL_LIST)
|
79 |
-
|
80 |
-
|
81 |
-
class CSMD(datasets.GeneratorBasedBuilder):
|
82 |
-
VERSION = datasets.Version("2.0.0")
|
83 |
-
|
84 |
-
BUILDER_CONFIGS = [
|
85 |
-
datasets.BuilderConfig(
|
86 |
-
name="meaning",
|
87 |
-
version=VERSION,
|
88 |
-
description="An instance consists of 1,355 meaning preservation triplets (Document, simplification, "
|
89 |
-
"label).",
|
90 |
-
),
|
91 |
-
datasets.BuilderConfig(
|
92 |
-
name="meaning_with_data_augmentation",
|
93 |
-
version=VERSION,
|
94 |
-
description="An instance consists of 1,355 meaning preservation triplets (Document, simplification, label) "
|
95 |
-
"along with 1,355 data augmentation triplets (Document, Document, 1) and 1,355 data "
|
96 |
-
"augmentation triplets (Document, Unrelated Document, 0) (See the sanity checks in our "
|
97 |
-
"article).",
|
98 |
-
),
|
99 |
-
datasets.BuilderConfig(
|
100 |
-
name="meaning_holdout_identical",
|
101 |
-
version=VERSION,
|
102 |
-
description="An instance consists of 359 meaning holdout preservation identical triplets (Document, "
|
103 |
-
"Document, 1) based on the ASSET Simplification dataset.",
|
104 |
-
),
|
105 |
-
datasets.BuilderConfig(
|
106 |
-
name="meaning_holdout_unrelated",
|
107 |
-
version=VERSION,
|
108 |
-
description="An instance consists of 359 meaning holdout preservation unrelated triplets (Document, "
|
109 |
-
"Unrelated Document, 0) based on the ASSET Simplification dataset.",
|
110 |
-
),
|
111 |
-
]
|
112 |
-
|
113 |
-
DEFAULT_CONFIG_NAME = "meaning"
|
114 |
-
|
115 |
-
def _info(self):
|
116 |
-
features = datasets.Features(
|
117 |
-
{
|
118 |
-
"document": datasets.Value(dtype="string"),
|
119 |
-
"simplification": datasets.Value(dtype="string"),
|
120 |
-
"labels": datasets.Value(dtype="string"),
|
121 |
-
}
|
122 |
-
)
|
123 |
-
return datasets.DatasetInfo(
|
124 |
-
description=_DESCRIPTION,
|
125 |
-
features=features,
|
126 |
-
supervised_keys=None,
|
127 |
-
homepage=_HOMEPAGE,
|
128 |
-
license=_LICENSE,
|
129 |
-
citation=_CITATION,
|
130 |
-
)
|
131 |
-
|
132 |
-
def _split_generators(self, dl_manager):
|
133 |
-
data_dir = dl_manager.download_and_extract(_URLs)
|
134 |
-
if self.config.name in ("meaning", "meaning_with_data_augmentation"):
|
135 |
-
return [
|
136 |
-
datasets.SplitGenerator(
|
137 |
-
name=datasets.Split.TRAIN,
|
138 |
-
gen_kwargs={
|
139 |
-
"filepaths": data_dir,
|
140 |
-
"split": "train",
|
141 |
-
},
|
142 |
-
),
|
143 |
-
datasets.SplitGenerator(
|
144 |
-
name=datasets.Split.VALIDATION,
|
145 |
-
gen_kwargs={
|
146 |
-
"filepaths": data_dir,
|
147 |
-
"split": "valid",
|
148 |
-
},
|
149 |
-
),
|
150 |
-
datasets.SplitGenerator(
|
151 |
-
name=datasets.Split.TEST,
|
152 |
-
gen_kwargs={"filepaths": data_dir, "split": "test"},
|
153 |
-
),
|
154 |
-
]
|
155 |
-
elif self.config.name in ("identical", "unrelated"):
|
156 |
-
return [
|
157 |
-
datasets.SplitGenerator(
|
158 |
-
name=f"{self.config.name}_{datasets.Split.TEST}",
|
159 |
-
gen_kwargs={
|
160 |
-
"filepaths": data_dir,
|
161 |
-
"split": "test",
|
162 |
-
},
|
163 |
-
),
|
164 |
-
]
|
165 |
-
|
166 |
-
def _generate_examples(self, filepaths, split):
|
167 |
-
with open(filepaths[split], encoding="utf-8") as f:
|
168 |
-
reader = csv.reader(f, delimiter="\t")
|
169 |
-
for id_, row in enumerate(reader):
|
170 |
-
if id_ == 0:
|
171 |
-
# Columns header
|
172 |
-
keys = row[:]
|
173 |
-
else:
|
174 |
-
res = dict([(k, v) for k, v in zip(keys, row)])
|
175 |
-
for k in ["document", "simplification", "labels"]:
|
176 |
-
res[k] = int(res[k])
|
177 |
-
yield (
|
178 |
-
id_ - 1
|
179 |
-
), res # Minus 1, since first idx is the columns header
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|