bugnet.py

"""TODO: Add a description here."""


import json

import datasets

# TODO: Add BibTeX citation
_CITATION = """\
"""

# TODO: Add description of the dataset here
_DESCRIPTION = """\
"""

# TODO: Add a link to an official homepage for the dataset here
_HOMEPAGE = ""

# TODO: Add the licence for the dataset here if you can find it
_LICENSE = ""

_URL = "https://huggingface.co/datasets/alexjercan/bugnet/resolve/main/"


def _mk_urls(language):
    return {
        "train": _URL + language + "_train.jsonl",
        "validation": _URL + language + "_validation.jsonl",
        "test": _URL + language + "_test.jsonl",
        "descriptions": _URL + "problem_descriptions.json",
    }


class Bugnet(datasets.GeneratorBasedBuilder):
    """TODO: Short description of my dataset."""

    VERSION = datasets.Version("1.4.0")

    BUILDER_CONFIGS = [
        datasets.BuilderConfig(name="Python", version=VERSION, description="This part of bugnet contains Python bugs"),
        datasets.BuilderConfig(name="C++", version=VERSION, description="This part of bugnet contains C++ bugs"),
    ]

    DEFAULT_CONFIG_NAME = "Python"

    def _info(self):
        features = datasets.Features(
            {
                "problem_id": datasets.Value("string"),
                "original_status": datasets.Value("string"),
                "original_src": datasets.Value("string"),
                "changed_src": datasets.Value("string"),
                "change": datasets.Value("string"),
                "i1": datasets.Value("uint32"),
                "i2": datasets.Value("uint32"),
                "j1": datasets.Value("uint32"),
                "j2": datasets.Value("uint32"),
                "error": datasets.Value("string"),
                "stderr": datasets.Value("string"),
                "description": datasets.Value("string"),
            }
        )

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            homepage=_HOMEPAGE,
            license=_LICENSE,
            citation=_CITATION,
        )

    def _split_generators(self, dl_manager):
        urls = _mk_urls(self.config.name)
        data_dir = dl_manager.download_and_extract(urls)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "filepath": data_dir["train"],
                    "descriptions": data_dir["descriptions"],
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "filepath": data_dir["validation"],
                    "descriptions": data_dir["descriptions"],
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "filepath": data_dir["test"],
                    "descriptions": data_dir["descriptions"],
                },
            ),
        ]

    def _generate_examples(self, filepath, descriptions):
        with open(descriptions, encoding="utf-8") as file:
            data = json.load(file)

        descriptions = {}
        for item in data:
            key = item["problem_id"]
            value = item["description"]
            descriptions[key] = value

        with open(filepath, encoding="utf-8") as file:
            for key, row in enumerate(file):
                data = json.loads(row)

                yield key, {
                    "problem_id": data["problem_id"],
                    "original_status": data["original_status"],
                    "original_src": data["original_src"],
                    "changed_src": data["changed_src"],
                    "change": data["change"],
                    "i1": data["i1"],
                    "i2": data["i2"],
                    "j1": data["j1"],
                    "j2": data["j2"],
                    "error": data["error"],
                    "stderr": data["stderr"],
                    "description": descriptions[data["problem_id"]],
                }