data_config_schema ¶

@field_validator("params")
def validate_dtype(cls, params: dict) -> dict:  # noqa: N805
    """Validate that the 'dtype' key is present in the encoder parameters."""
    if "dtype" not in params:
        raise ValueError("params must contain 'dtype' key")
    return params

@field_validator("columns")
@classmethod
def validate_param_lists_across_columns(
    cls,
    columns: list[TransformColumns],
) -> list[TransformColumns]:
    """Validate that parameter lists across columns have consistent lengths.

    Args:
        columns: List of transform columns to validate

    Returns:
        The validated columns list
    """
    # Get all parameter list lengths across all columns and transformations
    all_list_lengths: set[int] = set()

    for column in columns:
        for transformation in column.transformations:
            if transformation.params and any(
                isinstance(param_value, list) and len(param_value) > 0
                for param_value in transformation.params.values()
            ):
                all_list_lengths.update(
                    len(param_value)
                    for param_value in transformation.params.values()
                    if isinstance(param_value, list) and len(param_value) > 0
                )

    # Skip validation if no lists found
    if not all_list_lengths:
        return columns

    # Check if all lists either have length 1, or all have the same length
    all_list_lengths.discard(1)  # Remove length 1 as it's always valid
    if len(all_list_lengths) > 1:  # Multiple different lengths found
        raise ValueError(
            "All parameter lists across columns must either contain one element or have the same length",
        )

    return columns