import pickle
from pathlib import Path

# Simple check of the latest dataset
dataset_path = Path("phoneme_segmented_data/ctc_results_20251009_000024.pkl")
print(f"Checking: {dataset_path}")
print(f"File exists: {dataset_path.exists()}")

if dataset_path.exists():
    print(f"File size: {dataset_path.stat().st_size / (1024*1024):.1f} MB")

    try:
        with open(dataset_path, 'rb') as f:
            data = pickle.load(f)

        print(f"Data type: {type(data)}")
        print(f"Keys: {len(data) if isinstance(data, dict) else 'N/A'}")

        # Quick sample
        if isinstance(data, dict):
            sample_keys = list(data.keys())[:3]
            for key in sample_keys:
                if 'trial_' in str(key):
                    trial = data[key]
                    if 'alignment_info' in trial:
                        print(f"Trial {key}: {len(trial['alignment_info'])} segments")

                        # Check first few segments for temporal ordering
                        for i, (phoneme, start, end, conf) in enumerate(trial['alignment_info'][:3]):
                            status = "OK" if end >= start else "BUG"
                            print(f"  {phoneme}: {start}-{end} [{status}]")
                        break

        print("✅ Dataset loaded successfully")

    except Exception as e:
        print(f"❌ Error: {e}")
else:
    print("❌ Dataset file not found")