boost_camp.py

import csv
import datasets
from datasets.download.download_manager import DownloadManager

_CITATION = """"""
_DESCRIPTION = "hugging face dataset 공유해보기 연습"
_LICENSE = "CC-BY-SA-4.0"
_URL = "https://github.com/KyungHyunLim/hf_dataset_test"
_DATA_URLS = {
    "train": "https://huggingface.co/datasets/KHLim/boost_camp/resolve/main/train.csv",
    "valid": "https://huggingface.co/datasets/KHLim/boost_camp/resolve/main/dev.csv"
}

_VERSION = "0.0.0"

class MyDataConfig(datasets.BuilderConfig):
    def __init__(self, data_url, **kwargs):
        super().__init__(version=datasets.Version(_VERSION), **kwargs)
        self.data_url = data_url

class MYData(datasets.GeneratorBasedBuilder):
    DEFAULT_CONFIG_NAME = "mydata"
    BUILDER_CONFIGS = [
        MyDataConfig(
            name="mydata",
            data_url=_DATA_URLS,
            description=_DESCRIPTION
        )
    ]

    def _info(self):
        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=datasets.Features(
                {
                    "id": datasets.Value("string"),
                    "source": datasets.Value("string"),
                    "sentence_1": datasets.Value("string"),
                    "sentence_2": datasets.Value("string"),
                    "label": datasets.Value("float"),
                    "binary-label": datasets.Value("float")
                }
            ),
            homepage=_URL,
            license=_LICENSE,
            citation=_CITATION,
            supervised_keys=None,
        )
    
    def _split_generators(self, dl_manager: DownloadManager):
        data_file = dl_manager.download_and_extract(self.config.data_url)
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "data_file": data_file['train'],
                    "split": "train"
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "data_file": data_file['valid'],
                    "split": "valid"
                }
            )
        ]
    
    def _generate_examples(self, data_file:str, split:str):
        with open(data_file, newline='', encoding='utf-8') as f:
            reader = csv.reader(f, delimiter=',')
            features_names = next(reader)
            idx = 0
            for row in reader:
                features = {
                    "id": row[0],
                    "source": row[1],
                    "sentence_1": row[2],
                    "sentence_2": row[3],
                    "label": row[4],
                    "binary-label": row[5]
                }
                yield idx, features
                idx += 1