eu_wikipedias.py

# This is a huggingface dataset script to load wikipedias for the eu languages using olm/wikipedia

# https://huggingface.co/datasets/olm/wikipedia/blob/main/wikipedia.py

import datasets
from datasets import DatasetDict

logger = datasets.logging.get_logger(__name__)


_CITATION = """\
@ONLINE {wikidump,
    author = {Wikimedia Foundation},
    title  = {Wikimedia Downloads},
    url    = {https://dumps.wikimedia.org}
}
"""

_DESCRIPTION = """\
Wikipedia dataset containing cleaned articles of all languages.
The datasets are built from the Wikipedia dump
(https://dumps.wikimedia.org/) with one split per language. Each example
contains the content of one full Wikipedia article with cleaning to strip
markdown and unwanted sections (references, etc.).
"""

_LICENSE = (
    "This work is licensed under the Creative Commons Attribution-ShareAlike "
    "3.0 Unported License. To view a copy of this license, visit "
    "http://creativecommons.org/licenses/by-sa/3.0/ or send a letter to "
    "Creative Commons, PO Box 1866, Mountain View, CA 94042, USA."
)

_VERSION = datasets.Version("1.0.0", "")


eu_languages = [
   "bg",
   "cs",
   "da",
   "de",
   "el",
   "en",
   "es",
   "et",
   "fi",
   "fr",
   "ga",
   "hr",
   "hu",
   "it",
   "lt",
   "lv",
   "mt",
   "nl",
   "pl",
   "pt",
   "ro",
   "sk",
   "sl",
   "sv",
]

class WikipediaConfig(datasets.BuilderConfig):
    """BuilderConfig for EuWikipedia."""

    def __init__(self, date=None, version=_VERSION, **kwargs):
        """BuilderConfig for Wikipedia.
        Args:
          date: string, date of the Wikipedia dump in YYYYMMDD format. A list of
            available dates can be found at https://dumps.wikimedia.org/enwiki/.
          **kwargs: keyword arguments forwarded to super.
        """
        super().__init__(
            name=f"{date}",
            description=f"Wikipedia dataset for EU languages, parsed from {date} dump.",
            version=version,
            **kwargs,
        )
        self.date = date


_DATE = "20221101"


class EuWikipedia(datasets.GeneratorBasedBuilder):
    """Wikipedia dataset."""

    # Use mirror (your.org) to avoid download caps.
    BUILDER_CONFIG_CLASS = WikipediaConfig
    BUILDER_CONFIGS = [WikipediaConfig(date=_DATE,)]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "id": datasets.Value("string"),
                    "url": datasets.Value("string"),
                    "title": datasets.Value("string"),
                    "text": datasets.Value("string"),
                }
            ),
            # No default supervised_keys.
            supervised_keys=None,
            homepage="https://dumps.wikimedia.org",
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN, gen_kwargs={"date": self.config.date}
            )
        ]

    def _generate_examples(self, date):
        # defer to olm/wikipedia
        for lang in eu_languages:
            for example in datasets.load_dataset("olm/wikipedia", language=lang, date=date)[datasets.Split.TRAIN]:
                yield f"{lang}-{example['id']}", example