ms_marco.py

import datasets
import os
import json


_CITATION = """

"""

_DESCRIPTION = """

"""

class Builder(datasets.GeneratorBasedBuilder):
    VERSION = datasets.Version("1.0.0")

    BUILDER_CONFIGS = [
        datasets.BuilderConfig(name=name, version=datasets.Version("1.0.0"), description=_DESCRIPTION)
            for name in ["regular", "trec"]
        ]

    def _info(self):
        features = datasets.Features(
            {
                "text": datasets.Value("string"),
                "references": datasets.Sequence({
                    "id": datasets.Value("string"),
                    "text": datasets.Value("string"),
                    "is_gold": datasets.Value("bool"),
                    "relevance": datasets.Value("int64"),
                    "rank": datasets.Value("int64"),

                }),
                "golds":  datasets.Sequence({
                    "id": datasets.Value("string"),
                    "text": datasets.Value("string"),
                    "is_gold": datasets.Value("bool"),
                    "relevance": datasets.Value("int64"),
                    "rank": datasets.Value("int64"),
                }),
            }
        )
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage="",
            license="",
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        train_json = dl_manager.download(os.path.join(self.config.name, "train.jsonl"))
        valid_json = dl_manager.download(os.path.join(self.config.name, "valid.jsonl"))

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={"path": train_json},
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={"path": valid_json},
            ),
        ]

    # method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
    def _generate_examples(self, path):
        with open(path, encoding="utf-8") as f:
            for key, row in enumerate(f):
                yield key, json.loads(row)