my_stt_dataset.py

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

# BuilderConfig ni aniqlaymiz: bu yerda til qisqartmasi va asosiy data papkasi kiritiladi.
class STTConfig(BuilderConfig):
    def __init__(self, language_abbr, data_dir, **kwargs):
        """
        Args:
            language_abbr (str): Til qisqartmasi, masalan "uz".
            data_dir (str): Dataset joylashgan asosiy papka (misol: "Dataset_STT").
            **kwargs: Qolgan BuilderConfig parametrlar.
        """
        super().__init__(**kwargs)
        self.language_abbr = language_abbr
        self.data_dir = data_dir


class MySTTDataset(datasets.GeneratorBasedBuilder):
    """
    Common Voice uslubidagi minimal STT dataset yuklash skripti:
      - 3 ta tar fayl (train, test, validation) ichida .mp3 audio fayllar mavjud.
      - Har bir split uchun mos TSV fayl (train.tsv, test.tsv, validation.tsv) transkripsiyalarni o‘z ichiga oladi.
      - 'audio' ustuni Audio() tipida bo‘lib, Hugging Face Dataset Viewer’da "play" tugmasi orqali audio eshittirish imkoniyatini beradi.
    """
    VERSION = datasets.Version("1.0.0")

    # Agar bir nechta konfiguratsiya bo‘lmasa, oddiy qilib bitta config ishlatamiz.
    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",  # Bu yerga ma'lumotlar joylashgan asosiy papkani kiriting.
        )
    ]
    DEFAULT_CONFIG_NAME = "uz"

    def _info(self):
        """
        Datasetning xususiyatlari (features) aniqlanadi.
        Agar 'audio' ustuni Audio() tipida bo‘lsa, Dataset Viewer audio pleyerni ko‘rsatadi.
        """
        return datasets.DatasetInfo(
            description="Uzbek STT dataset: audio fayllar .tar arxivda saqlanadi, transcriptions esa .tsv faylda.",
            features=datasets.Features({
                "id": datasets.Value("string"),
                "audio": Audio(sampling_rate=None),  # sampling_rate=None degani audio fayldan olingan 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 va tsv fayllar yo‘llarini belgilab, dl_manager orqali ularni yuklab yoki extract qildik.
        """
        config = self.config  # STTConfig obyekti
        base_dir = config.data_dir  # Masalan: "Dataset_STT"
        lang = config.language_abbr   # Masalan: "uz"

        # Audio va transkript fayllarining yo‘llarini shakllantiramiz:
        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 fayllarni extract qilamiz (agar lokal bo‘lsa, dl_manager.extract mos yo‘lni qaytaradi)
        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,  # Tar fayl extract qilingan papka
                    "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‘qib, metadata lug‘atini yaratadi va
        extract qilingan archive papkasidan mos .mp3 faylni topadi.
        """
        # TSV faylini ochamiz va DictReader yordamida o‘qiymiz.
        with open(tsv_path, "r", encoding="utf-8") as f:
            reader = csv.DictReader(f, delimiter="\t")
            for idx, row in enumerate(reader):
                # TSV faylida kutilayotgan ustunlar: id, sentence, duration, age, gender, accents, locale
                audio_id = row["id"]
                mp3_file = audio_id + ".mp3"
                mp3_path = os.path.join(archive_dir, mp3_file)

                # Agar extract qilingan papkada audio fayl mavjud bo‘lsa:
                if os.path.isfile(mp3_path):
                    yield idx, {
                        "id": audio_id,
                        "audio": mp3_path,  # Audio() tipidagi ustun avtomatik ravishda faylni o‘qiydi va dekodlaydi.
                        "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:
                    # Agar audio fayl topilmasa, bu yozuvni o'tkazib yuboramiz yoki xatolik chiqarish mumkin.
                    continue