allocine.py

"""Allocine Dataset: A Large-Scale French Movie Reviews Dataset."""


import json

import datasets
from datasets.tasks import TextClassification


_CITATION = """\
@misc{blard2019allocine,
  author = {Blard, Theophile},
  title = {french-sentiment-analysis-with-bert},
  year = {2020},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished={\\url{https://github.com/TheophileBlard/french-sentiment-analysis-with-bert}},
}
"""

_DESCRIPTION = """\
 Allocine Dataset: A Large-Scale French Movie Reviews Dataset.
 This is a dataset for binary sentiment classification, made of user reviews scraped from Allocine.fr.
 It contains 100k positive and 100k negative reviews divided into 3 balanced splits: train (160k reviews), val (20k) and test (20k).
"""


class AllocineConfig(datasets.BuilderConfig):
    """BuilderConfig for Allocine."""

    def __init__(self, **kwargs):
        """BuilderConfig for Allocine.

        Args:
          **kwargs: keyword arguments forwarded to super.
        """
        super(AllocineConfig, self).__init__(**kwargs)


class AllocineDataset(datasets.GeneratorBasedBuilder):
    """Allocine Dataset: A Large-Scale French Movie Reviews Dataset."""

    _DOWNLOAD_URL = "https://github.com/TheophileBlard/french-sentiment-analysis-with-bert/raw/master/allocine_dataset/data.tar.bz2"
    _TRAIN_FILE = "train.jsonl"
    _VAL_FILE = "val.jsonl"
    _TEST_FILE = "test.jsonl"

    BUILDER_CONFIGS = [
        AllocineConfig(
            name="allocine",
            version=datasets.Version("1.0.0"),
            description="Allocine Dataset: A Large-Scale French Movie Reviews Dataset",
        ),
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "review": datasets.Value("string"),
                    "label": datasets.features.ClassLabel(names=["neg", "pos"]),
                }
            ),
            supervised_keys=None,
            homepage="https://github.com/TheophileBlard/french-sentiment-analysis-with-bert",
            citation=_CITATION,
            task_templates=[TextClassification(text_column="review", label_column="label")],
        )

    def _split_generators(self, dl_manager):
        archive_path = dl_manager.download(self._DOWNLOAD_URL)
        data_dir = "data"
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "filepath": f"{data_dir}/{self._TRAIN_FILE}",
                    "files": dl_manager.iter_archive(archive_path),
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "filepath": f"{data_dir}/{self._VAL_FILE}",
                    "files": dl_manager.iter_archive(archive_path),
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "filepath": f"{data_dir}/{self._TEST_FILE}",
                    "files": dl_manager.iter_archive(archive_path),
                },
            ),
        ]

    def _generate_examples(self, filepath, files):
        """Generate Allocine examples."""
        for path, file in files:
            if path == filepath:
                for id_, row in enumerate(file):
                    data = json.loads(row.decode("utf-8"))
                    review = data["review"]
                    label = "neg" if data["polarity"] == 0 else "pos"
                    yield id_, {"review": review, "label": label}