humaneval_infilling.py

import json
import datasets

logger = datasets.logging.get_logger(__name__)

_CITATION = """\
@article{bavarian2022efficient,
  title={Efficient Training of Language Models to Fill in the Middle},
  author={Bavarian, Mohammad and Jun, Heewoo and Tezak, Nikolas and Schulman, John and McLeavey, Christine and Tworek, Jerry and Chen, Mark},
  journal={arXiv preprint arXiv:2207.14255},
  year={2022}
}
"""

_DESCRIPTION = """\
An evaluation benchamrk for infilling tasks on HumanEval dataset for code generation.
"""

_SUBSETS = [ "MultiLineInfilling", "SingleLineInfilling", "RandomSpanInfilling", "RandomSpanInfillingLight" ]


class HumanevalConfig(datasets.BuilderConfig):
    """BuilderConfig for HumanevalConfig."""

    def __init__(
        self,
        subset,
        **kwargs,
    ):
        self.subset = subset
        name = f"HumanEval-{subset}"
        kwargs["name"] = name
        super(HumanevalConfig, self).__init__(**kwargs)


class MultiPLE(datasets.GeneratorBasedBuilder):
    BUILDER_CONFIG_CLASS = HumanevalConfig

    BUILDER_CONFIGS = [
        HumanevalConfig(
            subset=subset,
            version=datasets.Version("1.0.0"))
        for subset in _SUBSETS 
    ]

    DEFAULT_CONFIG_NAME = "HumanEval-SingleLineInfilling"

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            license="MIT",
            features = datasets.Features({'task_id': datasets.Value(dtype='string'),
                                        'entry_point': datasets.Value(dtype='string'),
                                        'prompt': datasets.Value(dtype='string'),
                                        'suffix': datasets.Value(dtype='string'),
                                        'canonical_solution': datasets.Value(dtype='string'),
                                        'test': datasets.Value(dtype='string')}),
            supervised_keys=None,
            homepage="https://github.com/openai/human-eval-infilling",
            citation=_CITATION
        )

    def _split_generators(self, dl_manager: datasets.DownloadManager):
        files = dl_manager.download(
            f"data/{self.config.name}.jsonl"
        )
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "filepath": files,
                }
            )
        ]

    def _generate_examples(self, filepath):
        with open(filepath) as f:
            for id, line in enumerate(f):
                row = json.loads(line)
                yield id, row