Registry

dataphy.dataset.registry

Dataset loader registry for Dataphy SDK.

Classes

Functions

register_dataset_loader(format_type: DatasetFormat, loader_class: Type[BaseDatasetLoader])

Register a new dataset loader.

Parameters:

Name	Type	Description	Default
format_type	DatasetFormat	The dataset format to register	required
loader_class	Type[BaseDatasetLoader]	The loader class to register for this format	required

Source code in src/dataphy/dataset/registry.py

def register_dataset_loader(format_type: DatasetFormat, loader_class: Type[BaseDatasetLoader]):
    """Register a new dataset loader.

    Args:
        format_type: The dataset format to register
        loader_class: The loader class to register for this format
    """
    _DATASET_LOADERS[format_type] = loader_class

get_dataset_loader(format_type: DatasetFormat) -> Type[BaseDatasetLoader]

Get a dataset loader by format.

Parameters:

Name	Type	Description	Default
format_type	DatasetFormat	The dataset format to get loader for	required

Returns:

Type	Description
Type[BaseDatasetLoader]	The loader class for the specified format

Raises:

Type	Description
ValueError	If the format is not supported

Source code in src/dataphy/dataset/registry.py

def get_dataset_loader(format_type: DatasetFormat) -> Type[BaseDatasetLoader]:
    """Get a dataset loader by format.

    Args:
        format_type: The dataset format to get loader for

    Returns:
        The loader class for the specified format

    Raises:
        ValueError: If the format is not supported
    """
    if format_type not in _DATASET_LOADERS:
        raise ValueError(f"Unsupported dataset format: {format_type}")
    return _DATASET_LOADERS[format_type]

create_dataset_loader(dataset_path: Union[str, Path], format_type: Optional[DatasetFormat] = None, **kwargs: Any) -> BaseDatasetLoader

Create a dataset loader instance with automatic format detection.

This is the main entry point for loading datasets. It automatically detects the dataset format based on directory structure and files, or uses the explicitly provided format.

Supported formats: - LeRobot: Robotics datasets with episodes, videos, and parquet data

Parameters:

Name	Type	Description	Default
dataset_path	Union[str, Path]	Path to the dataset directory	required
format_type	Optional[DatasetFormat]	Dataset format. If None, format will be auto-detected. Use DatasetFormat enum values (e.g., DatasetFormat.LEROBOT)	None
**kwargs	Any	Additional arguments passed to the specific loader	{}

Returns:

Name	Type	Description
BaseDatasetLoader	BaseDatasetLoader	Dataset loader instance for the detected/specified format

Raises:

Type	Description
ValueError	If dataset format cannot be detected or is unsupported
FileNotFoundError	If dataset path does not exist

Example

from dataphy.dataset.registry import create_dataset_loader, DatasetFormat

# Auto-detect format (recommended)
loader = create_dataset_loader("./my_dataset")

# Explicit format specification
loader = create_dataset_loader(
    "./my_dataset", 
    format_type=DatasetFormat.LEROBOT
)

# Get dataset information
info = loader.get_dataset_info()
episodes = loader.get_episode_ids()

Source code in src/dataphy/dataset/registry.py

def create_dataset_loader(
    dataset_path: Union[str, Path], 
    format_type: Optional[DatasetFormat] = None,
    **kwargs: Any
) -> BaseDatasetLoader:
    """Create a dataset loader instance with automatic format detection.

    This is the main entry point for loading datasets. It automatically detects
    the dataset format based on directory structure and files, or uses the
    explicitly provided format.

    Supported formats:
    - **LeRobot**: Robotics datasets with episodes, videos, and parquet data

    Args:
        dataset_path: Path to the dataset directory
        format_type: Dataset format. If None, format will be auto-detected.
            Use DatasetFormat enum values (e.g., DatasetFormat.LEROBOT)
        **kwargs: Additional arguments passed to the specific loader

    Returns:
        BaseDatasetLoader: Dataset loader instance for the detected/specified format

    Raises:
        ValueError: If dataset format cannot be detected or is unsupported
        FileNotFoundError: If dataset path does not exist

    Example:
        ```python
        from dataphy.dataset.registry import create_dataset_loader, DatasetFormat

        # Auto-detect format (recommended)
        loader = create_dataset_loader("./my_dataset")

        # Explicit format specification
        loader = create_dataset_loader(
            "./my_dataset", 
            format_type=DatasetFormat.LEROBOT
        )

        # Get dataset information
        info = loader.get_dataset_info()
        episodes = loader.get_episode_ids()
        ```
    """
    dataset_path = Path(dataset_path)

    if format_type is None:
        # Auto-detect format
        format_type = _detect_dataset_format(dataset_path)

    loader_class = get_dataset_loader(format_type)
    return loader_class(dataset_path, **kwargs)

list_supported_formats() -> List[DatasetFormat]

List all supported dataset formats.

Returns:

Type	Description
List[DatasetFormat]	List of all registered dataset formats

Source code in src/dataphy/dataset/registry.py

def list_supported_formats() -> List[DatasetFormat]:
    """List all supported dataset formats.

    Returns:
        List of all registered dataset formats
    """
    return list(_DATASET_LOADERS.keys())