dataset.py

import csv
import os

import datasets

logger = datasets.logging.get_logger(__name__)

_DESCRIPTION = "Custom dataset for extracting audio files and matching sentences."

_DATA_URL = "https://huggingface.co/datasets/ugshanyu/dataset/resolve/main"  # Replace with the URL of your data

class CustomDataset(datasets.GeneratorBasedBuilder):

    VERSION = datasets.Version("1.0.0")

    def _info(self):
        features = datasets.Features(
            {
                "audio": datasets.Audio(sampling_rate=48_000),
                "sentence": datasets.Value("string"),
            }
        )

        return datasets.DatasetInfo(
            description=_DESCRIPTION,
            features=features,
            supervised_keys=("audio", "sentence"),
            homepage=None,
            citation=None,
        )

    def _split_generators(self, dl_manager):
        audio_path = dl_manager.download_and_extract(_DATA_URL+"/audo.zip")
        csv_path = dl_manager.download_and_extract(_DATA_URL+"/comma.csv")
        
        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={"audio_path": audio_path, "csv_path": csv_path},
            )
        ]

    def _generate_examples(self, audio_path, csv_path):
        print(audio_path)
        print(csv_path)
        key = 0
        print(os.listdir(audio_path))
        
        with open(csv_path, encoding="utf-8") as csv_file:
            csv_reader = csv.DictReader(csv_file)
            for row in csv_reader:
                original_sentence_id, sentence, locale = row.values()
                audio_file = f"{original_sentence_id}.mp3"
                audio_file_path = os.path.join(audio_path, audio_file)
                yield key, {
                    "audio": audio_file_path,
                    "sentence": sentence,
                }
                key += 1