test_hf_dataset.py

#!/usr/bin/env python3

from datasets import Dataset
import json


def test_local_dataset():
    """Test loading the dataset locally with Hugging Face datasets"""

    # Read the JSONL metadata file
    records = []
    try:
        with open("data/metadata.jsonl", "r", encoding="utf-8") as f:
            for line in f:
                records.append(json.loads(line))
    except FileNotFoundError:
        print("Error: data/metadata.jsonl not found")
        return

    print(f"Loaded {len(records)} records from metadata.jsonl")

    # Create a dataset from the records
    dataset = Dataset.from_list(records)

    # Cast the audio column to Audio type (without automatic decoding)
    from datasets import Audio
    dataset = dataset.cast_column("audio", Audio(decode=False))

    print("\nDataset info:")
    print(f"- Number of samples: {len(dataset)}")
    print(f"- Features: {dataset.features}")

    # Test accessing first few examples
    print("\nFirst 3 examples:")
    for i, example in enumerate(dataset.select(range(3))):
        print(f"\n{i+1}. Word: {example['word']}")
        print(f"   IPA: {example['ipa']}")
        audio = example['audio']
        if audio:
            print(f"   Audio path: {audio.get('path', 'N/A')}")
            if audio.get('bytes'):
                print(f"   Audio bytes: {len(audio['bytes'])} bytes")
            else:
                print("   Audio loaded successfully")

    return dataset


if __name__ == "__main__":
    print("Testing Hugging Face dataset compatibility...")
    print("=" * 50)

    dataset = test_local_dataset()

    if dataset:
        print(f"\n✓ Dataset successfully loaded with {len(dataset)} examples")
        print("✓ Audio files are properly loaded as Audio features")
    else:
        print("✗ Failed to load dataset")