dataeval.utils.data.requires_maite_dataset¶

dataeval.utils.data.requires_maite_dataset(arg_name='dataset', *, expected='any_target')¶

Validate a named dataset argument before the wrapped call runs (decorator).

Resolves the dataset argument by name (works for positional or keyword passing) and calls validate_dataset(). Compatible with regular functions, methods (including __init__), and classmethod-wrapped constructors like Embeddings.new().

Parameters:¶

arg_name='dataset'¶

Name of the dataset parameter on the wrapped callable.

expected='any_target'¶

Forwarded to validate_dataset().

Examples

>>> @requires_maite_dataset("dataset", expected="image_only")
... def fit(self, dataset): ...

>>> @requires_maite_dataset(expected="object_detection")
... def evaluate(self, dataset): ...