create_complete_metadata_table¶

def create_complete_metadata_table(pdb_directory: Optional[str] = None) -> pd.DataFrame:
    """
    Create complete metadata table for all ProteinGym datasets.

    Args:
        pdb_directory: Optional directory containing PDB files

    Returns:
        DataFrame with all metadata entries
    """
    metadata_entries = []

    # Core datasets
    metadata_entries.append(create_alphamissense_metadata())
    metadata_entries.append(create_dms_substitutions_metadata())
    metadata_entries.append(create_dms_reference_metadata())
    metadata_entries.append(create_zeroshot_scores_metadata())
    metadata_entries.append(create_zeroshot_summary_v12_metadata())
    metadata_entries.append(create_zeroshot_model_scores_metadata())

    # Supervised datasets
    for fold_type in ["contiguous", "modulo", "random"]:
        metadata_entries.append(create_supervised_metadata(fold_type))

    metadata_entries.append(create_supervised_summary_metadata())

    # PDB files if directory provided
    if pdb_directory:
        pdb_metadata = generate_pdb_metadata(pdb_directory)
        metadata_entries.extend(pdb_metadata)

    # Convert to DataFrame
    rows = []
    for metadata in metadata_entries:
        row = {
            'Title': metadata.title,
            'Description': metadata.description,
            'BiocVersion': metadata.bioc_version,
            'Genome': metadata.genome,
            'SourceType': metadata.source_type,
            'SourceUrl': metadata.source_url,
            'SourceVersion': metadata.source_version,
            'Species': metadata.species,
            'TaxonomyId': metadata.taxonomy_id,
            'Coordinate_1_based': metadata.coordinate_1_based,
            'DataProvider': metadata.data_provider,
            'Maintainer': metadata.maintainer,
            'DataClass': metadata.data_class,
            'DispatchClass': metadata.dispatch_class,
            'DataPath': metadata.data_path
        }
        rows.append(row)

    return pd.DataFrame(rows)