enhanced-cobald-dataset.py

import datasets
from datasets import GeneratorBasedBuilder, BuilderConfig, Sequence, Value

from .parsing import parse_incr


class LanguageSpecificConfig(BuilderConfig):
    def __init__(self, language, data_file, **kwargs):
        super().__init__(**kwargs)
        self.language = language
        self.data_file = data_file


class EnhancedCobaldDataset(GeneratorBasedBuilder):
    BUILDER_CONFIGS = [
        LanguageSpecificConfig(
            name="en",
            language="en",
            data_file="https://raw.githubusercontent.com/CobaldAnnotation/CobaldEng/refs/heads/main/enhanced/train.conllu",
            description="English dataset."
        ),
        LanguageSpecificConfig(
            name="ru",
            language="ru",
            data_file="https://raw.githubusercontent.com/CobaldAnnotation/CobaldRus/refs/heads/main/enhanced/train.conllu",
            description="Russian dataset."
        ),
        # Other languages here
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description="A CoBaLD dataset in CoNLL-U plus format.",
            features=datasets.Features({
                "ids": Sequence(Value("string")),
                "words": Sequence(Value("string")),
                "lemmas": Sequence(Value("string")),
                "upos": Sequence(Value("string")),
                "xpos": Sequence(Value("string")),
                # huggingface datasets can't handle dicts with dynamic keys, so represent feats as string
                "feats": Sequence(Value("string")),
                "heads": Sequence(Value("int32")),
                "deprels": Sequence(Value("string")),
                "deps": Sequence(Value("string")),
                "miscs": Sequence(Value("string")),
                "deepslots": Sequence(Value("string")),
                "semclasses": Sequence(Value("string")),
                "sent_id": Value("string"),
                "text": Value("string"),
            })
        )

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

    def _generate_examples(self, filepath: str):
        """
        Generator function that reads a CoNLL-U file and yields one sentence at a time.
        Each sentence is represented as a dictionary where each field is a list.
        """
        yield from enumerate(parse_incr(filepath))