CSMD.py

# coding=utf-8
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""CSMD: a dataset for assessing meaning preservation between sentences"""

import csv

import datasets
from datasets import load_dataset

_CITATION = """\
@ARTICLE{10.3389/frai.2023.1223924,
AUTHOR={Beauchemin, David and Saggion, Horacio and Khoury, Richard},   
TITLE={{MeaningBERT: Assessing Meaning Preservation Between Sentences}},      
JOURNAL={Frontiers in Artificial Intelligence},      
VOLUME={6},           
YEAR={2023},      
URL={https://www.frontiersin.org/articles/10.3389/frai.2023.1223924},       
DOI={10.3389/frai.2023.1223924},      	
ISSN={2624-8212},   
}
"""

_DESCRIPTION = """\
Continuous Scale Meaning Dataset (CSMD) is a dataset for assessing meaning preservation between sentences.
"""

_HOMEPAGE = "https://github.com/GRAAL-Research/csmd"

_LICENSE = "Attribution 4.0 International (CC BY 4.0)"

_URL_LIST = [
    (
        "meaning.train",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning/train.tsv",
    ),
    (
        "meaning.dev",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning/dev.tsv",
    ),
    (
        "meaning.test",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning/test.tsv",
    ),
    (
        "meaning_with_data_augmentation.train",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning_with_data_augmentation/train.tsv",
    ),
    (
        "meaning_with_data_augmentation.dev",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning_with_data_augmentation/dev.tsv",
    ),
    (
        "meaning_with_data_augmentation.test",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/meaning_with_data_augmentation/test.tsv",
    ),
    (
        "identical",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/holdout/identical.tsv",
    ),
    (
        "unrelated",
        "https://github.com/GRAAL-Research/csmd/blob/main/dataset/holdout/unrelated.tsv",
    ),
]

_URLs = dict(_URL_LIST)


class CSMD(datasets.GeneratorBasedBuilder):
    VERSION = datasets.Version("2.0.0")

    BUILDER_CONFIGS = [
        datasets.BuilderConfig(
            name="meaning",
            version=VERSION,
            description="An instance consists of 1,355 meaning preservation triplets (Document, simplification, "
            "label).",
        ),
        datasets.BuilderConfig(
            name="meaning_with_data_augmentation",
            version=VERSION,
            description="An instance consists of 1,355 meaning preservation triplets (Document, simplification, label) "
            "along with 1,355 data augmentation triplets (Document, Document, 1) and 1,355 data "
            "augmentation triplets (Document, Unrelated Document, 0) (See the sanity checks in our "
            "article).",
        ),
        datasets.BuilderConfig(
            name="meaning_holdout_identical",
            version=VERSION,
            description="An  instance consists of 359 meaning holdout preservation identical triplets (Document, "
            "Document, 1) based on the ASSET Simplification dataset.",
        ),
        datasets.BuilderConfig(
            name="meaning_holdout_unrelated",
            version=VERSION,
            description="An instance consists of 359 meaning holdout preservation unrelated triplets (Document, "
            "Unrelated Document, 0) based on the ASSET Simplification dataset.",
        ),
    ]

    DEFAULT_CONFIG_NAME = "meaning"

    def _info(self):
        features = datasets.Features(
            {
                "document": datasets.Value(dtype="string"),
                "simplification": datasets.Value(dtype="string"),
                "labels": datasets.Value(dtype="string"),
            }
        )
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            supervised_keys=None,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        data_dir = dl_manager.download_and_extract(_URLs)
        if self.config.name in ("meaning", "meaning_with_data_augmentation"):
            return [
                datasets.SplitGenerator(
                    name=datasets.Split.TRAIN,
                    gen_kwargs={
                        "filepaths": data_dir,
                        "split": "train",
                    },
                ),
                datasets.SplitGenerator(
                    name=datasets.Split.VALIDATION,
                    gen_kwargs={
                        "filepaths": data_dir,
                        "split": "valid",
                    },
                ),
                datasets.SplitGenerator(
                    name=datasets.Split.TEST,
                    gen_kwargs={"filepaths": data_dir, "split": "test"},
                ),
            ]
        elif self.config.name in ("identical", "unrelated"):
            return [
                datasets.SplitGenerator(
                    name=f"{self.config.name}_{datasets.Split.TEST}",
                    gen_kwargs={
                        "filepaths": data_dir,
                        "split": "test",
                    },
                ),
            ]

    def _generate_examples(self, filepaths, split):
        with open(filepaths[split], encoding="utf-8") as f:
            reader = csv.reader(f, delimiter="\t")
            for id_, row in enumerate(reader):
                if id_ == 0:
                    # Columns header
                    keys = row[:]
                else:
                    res = dict([(k, v) for k, v in zip(keys, row)])
                    for k in ["document", "simplification", "labels"]:
                        res[k] = int(res[k])
                    yield (
                        id_ - 1
                    ), res  # Minus 1, since first idx is the columns header