twitter_sentiment.py

"""Custom classification dataset."""

import csv

import datasets
from datasets.tasks import TextClassification

_DESCRIPTION = """\
"""

_CITATION = """\
"""

_TRAIN_DOWNLOAD_URL = "train.csv"
_TEST_DOWNLOAD_URL = "test.csv"

_DATASET_MAP = {"NEGATIVE": 0, "POSITIVE": 1}


class Custom(datasets.GeneratorBasedBuilder):
    """Custom classification dataset."""

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "text": datasets.Value("string"),
                    "label": datasets.features.ClassLabel(
                        names=["NEGATIVE", "POSITIVE"]
                    ),
                }
            ),
            homepage="",
            citation=_CITATION,
            task_templates=[
                TextClassification(text_column="text", label_column="label")
            ],
        )

    def _split_generators(self, dl_manager):
        train_path = dl_manager.download_and_extract(_TRAIN_DOWNLOAD_URL)
        test_path = dl_manager.download_and_extract(_TEST_DOWNLOAD_URL)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN, gen_kwargs={"filepath": train_path}
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION, gen_kwargs={"filepath": test_path}
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST, gen_kwargs={"filepath": test_path}
            ),
        ]

    def _generate_examples(self, filepath):
        """Generate Custom examples."""

        with open(filepath, encoding="utf-8") as csv_file:
            csv_reader = csv.reader(
                csv_file,
                quotechar='"',
                delimiter=",",
                quoting=csv.QUOTE_ALL,
                skipinitialspace=True,
            )
            for id_, row in enumerate(csv_reader):
                text, label = row
                label = _DATASET_MAP[label]
                yield id_, {"text": text, "label": label}