#!/usr/bin/env python3

import json


def convert_to_hf_audio_format(jsonl_file="data/metadata.jsonl", output_file="data/metadata_hf.jsonl"):
    """Convert JSONL dataset to Hugging Face audio dataset format"""

    records = []

    # Read existing JSONL file
    with open(jsonl_file, "r", encoding="utf-8") as f:
        for line in f:
            old_record = json.loads(line)

            # Create new record following HF audio dataset format
            record = {
                "word": old_record["word"],
                "ipa": old_record["ipa"],
                "audio": old_record["audio"]["path"]  # HF expects string path, not dict
            }
            records.append(record)

    # Write new JSONL file
    with open(output_file, "w", encoding="utf-8") as f:
        for record in records:
            f.write(json.dumps(record, ensure_ascii=False) + "\n")

    print(f"Converted {len(records)} records to Hugging Face audio format")
    print(f"Output saved to: {output_file}")

    # Display first 3 records
    print("\nFirst 3 records in HF format:")
    for i, record in enumerate(records[:3], 1):
        print(f"{i}. {json.dumps(record, ensure_ascii=False)}")

    return records


if __name__ == "__main__":
    convert_to_hf_audio_format()