my_stt_dataset.py

import os
import csv
import datasets
from datasets import Audio, BuilderConfig

# Konfiguratsiya sinfi: til qisqartmasi va ma'lumotlar joylashgan papkani belgilaydi.
class STTConfig(BuilderConfig):
    def __init__(self, language_abbr, data_dir, **kwargs):
        """
        Args:
            language_abbr (str): Masalan, "uz".
            data_dir (str): Dataset joylashgan asosiy papka, masalan "Dataset_STT".
            **kwargs: Qolgan parametrlar.
        """
        super().__init__(**kwargs)
        self.language_abbr = language_abbr
        self.data_dir = data_dir

# Dataset yuklash skripti
class MySTTDataset(datasets.GeneratorBasedBuilder):
    """
    Uzbek STT dataset yuklash skripti:
      - Audio fayllar .tar arxiv ichida joylashgan.
      - Transkripsiya ma'lumotlari mos TSV faylda.
      - "audio" ustuni Audio() tipida aniqlangan, shuning uchun Hub Viewer "play" tugmasini ko'rsatadi.
    """
    VERSION = datasets.Version("1.0.0")

    BUILDER_CONFIGS = [
        STTConfig(
            name="uz",
            version=datasets.Version("1.0.0"),
            description="Uzbek subset of the STT dataset",
            language_abbr="uz",
            data_dir="Dataset_STT",  # Asosiy papka nomi
        )
    ]
    DEFAULT_CONFIG_NAME = "uz"

    def _info(self):
        """
        Dataset ustunlarini aniqlaydi.
        "audio" ustuni Audio() tipida belgilangan – bu orqali Viewer audio faylni avtomatik dekodlaydi.
        """
        return datasets.DatasetInfo(
            description="Uzbek STT dataset: audio fayllar .tar arxivda, transcriptions esa TSV faylda saqlanadi.",
            features=datasets.Features({
                "id": datasets.Value("string"),
                "audio": Audio(sampling_rate=None),  # Asl sampling rate saqlanadi
                "sentence": datasets.Value("string"),
                "duration": datasets.Value("float"),
                "age": datasets.Value("string"),
                "gender": datasets.Value("string"),
                "accents": datasets.Value("string"),
                "locale": datasets.Value("string"),
            }),
            supervised_keys=None,
            version=self.VERSION,
        )

    def _split_generators(self, dl_manager):
        """
        Har bir split uchun: tar arxiv va mos TSV fayllarining yo'llari aniqlanadi,
        va tar fayllar dl_manager.extract() orqali ochiladi.
        """
        config = self.config
        base_dir = config.data_dir  # Misol: "Dataset_STT"
        lang = config.language_abbr   # Misol: "uz"

        train_tar = os.path.join(base_dir, "audio", lang, "train.tar")
        train_tsv = os.path.join(base_dir, "transcript", lang, "train.tsv")

        test_tar = os.path.join(base_dir, "audio", lang, "test.tar")
        test_tsv = os.path.join(base_dir, "transcript", lang, "test.tsv")

        val_tar = os.path.join(base_dir, "audio", lang, "validation.tar")
        val_tsv = os.path.join(base_dir, "transcript", lang, "validation.tsv")

        # Tar arxiv extract qilinadi:
        train_tar_extracted = dl_manager.extract(train_tar)
        test_tar_extracted = dl_manager.extract(test_tar)
        val_tar_extracted = dl_manager.extract(val_tar)

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={
                    "archive_dir": train_tar_extracted,
                    "tsv_path": train_tsv,
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.TEST,
                gen_kwargs={
                    "archive_dir": test_tar_extracted,
                    "tsv_path": test_tsv,
                },
            ),
            datasets.SplitGenerator(
                name=datasets.Split.VALIDATION,
                gen_kwargs={
                    "archive_dir": val_tar_extracted,
                    "tsv_path": val_tsv,
                },
            ),
        ]

    def _generate_examples(self, archive_dir, tsv_path):
        """
        TSV faylini qatorma-qator o'qiydi va metadata lug'atini tuzadi.
        Keyin, archive papkasidan mos .mp3 faylni topib,
        audio ustunini quyidagicha shakllantiradi:
            {"path": <relative file name>, "bytes": <audio file baytlari>}
        Bu shakl Dataset Viewer tomonidan Audio() sifatida aniqlanib, "play" tugmasini ko'rsatishga imkon beradi.
        """
        with open(tsv_path, "r", encoding="utf-8") as f:
            reader = csv.DictReader(f, delimiter="\t")
            for idx, row in enumerate(reader):
                audio_id = row["id"]
                mp3_file = audio_id + ".mp3"
                full_path = os.path.join(archive_dir, mp3_file)

                if os.path.isfile(full_path):
                    with open(full_path, "rb") as audio_file:
                        audio_bytes = audio_file.read()
                    # MUHIM: "path" qiymatini lokal extract qilingan papka o'rniga, faqat fayl nomi sifatida uzatamiz
                    yield idx, {
                        "id": audio_id,
                        "audio": {"path": mp3_file, "bytes": audio_bytes},
                        "sentence": row.get("sentence", ""),
                        "duration": float(row.get("duration", 0.0)),
                        "age": row.get("age", ""),
                        "gender": row.get("gender", ""),
                        "accents": row.get("accents", ""),
                        "locale": row.get("locale", ""),
                    }
                else:
                    continue