mc4-es-sampled.py

"""Perplexity Sampled mC4 dataset based on Common Crawl."""


import gzip
import json

import datasets
import numpy as np
from numpy.random import default_rng


logger = datasets.logging.get_logger(__name__)


_DESCRIPTION = """
50 million documents in Spanish extracted from mC4 applying perplexity sampling via mc4-sampling: "https://huggingface.co/datasets/bertin-project/mc4-sampling". Please, refer to BERTIN Project. The original dataset is the Multlingual Colossal, Cleaned version of Common Crawl's web crawl corpus (mC4), based on the Common Crawl dataset: "https://commoncrawl.org", and processed by AllenAI.
"""

_CITATION = """
@article{2019t5,
    author = {Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu},
    title = {Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer},
    journal = {arXiv e-prints},
    year = {2019},
    archivePrefix = {arXiv},
    eprint = {1910.10683},
}
"""

_URL = "https://github.com/allenai/allennlp/discussions/5056"

_DATA_URL_VALIDATION = "https://huggingface.co/datasets/allenai/c4/resolve/1ddc917116b730e1859edef32896ec5c16be51d0/multilingual/c4-es-validation.tfrecord-{index:05d}-of-{n_shards:05d}.json.gz"
_DATA_URL_TRAIN = "https://huggingface.co/datasets/bertin-project/mc4-es-sampled/resolve/main/mc4-es-train-50M-{config}-shard-{index:04d}-of-{n_shards:04d}.json.gz"

_CONFIGS = [
    "random",
    "stepwise",
    "gaussian",
]

_N_SHARDS_PER_SPLIT = {
    "random": {"train": 1024, "validation": 16},
    "stepwise": {"train": 1024, "validation": 16},
    "gaussian": {"train": 1024, "validation": 16},
}


class Mc4EsSampledConfig(datasets.BuilderConfig):
    """BuilderConfig for mC4."""

    def __init__(self, *args, configs, **kwargs):
        """BuilderConfig for mC4.
        Args:
            configs (:obj:`List[str]`): list of configs to load
            **kwargs: keyword arguments forwarded to super.
        """
        super().__init__(
            *args,
            name="+".join(configs),
            **kwargs,
        )
        self.configs = configs


class Mc4EsSampled(datasets.GeneratorBasedBuilder):
    """mC4, a colossal, cleaned version of Common Crawl's web crawl corpus."""

    BUILDER_CONFIGS = [Mc4EsSampledConfig(configs=[config]) for config in _CONFIGS]
    BUILDER_CONFIG_CLASS = Mc4EsSampledConfig

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "text": datasets.Value("string"),
                    "timestamp": datasets.Value("string"),
                    "url": datasets.Value("string"),
                }
            ),
            supervised_keys=None,
            homepage=_URL,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        data_urls = {}
        data_urls["train"] = [
            _DATA_URL_TRAIN.format(
                config=config,
                index=index + 1,
                n_shards=_N_SHARDS_PER_SPLIT[config]["train"],
            )
            for config in self.config.configs
            for index in range(_N_SHARDS_PER_SPLIT[config]["train"])
        ]
        data_urls["validation"] = [
            _DATA_URL_VALIDATION.format(
                index=index + 1,
                n_shards=_N_SHARDS_PER_SPLIT[config]["validation"],
            )
            for config in self.config.configs
            for index in range(_N_SHARDS_PER_SPLIT[config]["validation"] - 1)  # the last validation shard is not working in the original allen/c4 repo
        ]
        train_downloaded_files = dl_manager.download(data_urls["train"])
        validation_downloaded_files = dl_manager.download(data_urls["validation"])
        return [
            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"filepaths": train_downloaded_files}),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION, gen_kwargs={"filepaths": validation_downloaded_files}
            ),
        ]

    def _generate_examples(self, filepaths):
        """This function returns the examples in the raw (text) form by iterating on all the files."""
        id_ = 0
        for filepath in filepaths:
            with gzip.open(open(filepath, "rb"), "rt", encoding="utf-8") as f:
                for line in f:
                    if line:
                        example = json.loads(line)
                        yield id_, example
                        id_ += 1