import csv
import os
from datasets import DatasetBuilder, DatasetInfo, Features, Value, Audio, Split, SplitGenerator, DownloadManager, BuilderConfig

class TobiasChineseTTSConfig(BuilderConfig):
    """BuilderConfig for Tobias Chinese TTS Dataset."""
    
    def __init__(self, **kwargs):
        super(TobiasChineseTTSConfig, self).__init__(**kwargs)

class TobiasChineseTTS(DatasetBuilder):
    """Tobias Chinese TTS Dataset."""
    
    BUILDER_CONFIGS = [
        TobiasChineseTTSConfig(
            name="default",
            version="1.0.0",
            description="Tobias Chinese TTS Dataset"
        )
    ]
    
    def _info(self):
        return DatasetInfo(
            description="Chinese TTS dataset with audio-text pairs",
            features=Features({
                "text": Value("string"),
                "audio": Audio(sampling_rate=16000),
            }),
            supervised_keys=("text", "audio"),
        )
    
    def _split_generators(self, dl_manager: DownloadManager):
        return [
            SplitGenerator(
                name=Split.TRAIN,
                gen_kwargs={"filepath": "train.csv", "split": "train"},
            ),
            SplitGenerator(
                name=Split.VALIDATION,
                gen_kwargs={"filepath": "validation.csv", "split": "validation"},
            ),
        ]
    
    def _generate_examples(self, filepath, split):
        with open(filepath, encoding="utf-8") as f:
            reader = csv.DictReader(f)
            for idx, row in enumerate(reader):
                yield idx, {
                    "text": row["text"],
                    "audio": row["audio"],
                }