Functional

Sorry, the person who is responsible for the description was eaten by hydras last week.

hydra_slayer.functional.get_factory(name_or_object: Union[str, hydra_slayer.functional.T]) → Union[Type, Callable[[...], Any], hydra_slayer.functional.T]

Retrieves factory, without creating any objects with it.

Parameters

name_or_object – factory name or any valid python object

Returns

factory

Raises

LookupError – if no factory with provided name was registered

Examples

>>> to_int = get_factory("int")
>>> to_int("42")
42

hydra_slayer.functional.get_from_params(*, shared_params: Optional[Dict[str, Any]] = None, **kwargs) → Any

Creates instance based in configuration dict with instantiation_fn.

Note

The name of the factory to use should be provided by '_target_' keyword.

Parameters

• shared_params – params to pass on all levels in case of recursive creation

• **kwargs – named parameters for factory

Returns

result of calling instantiate_fn(factory, **sub_kwargs)

Examples

>>> get_from_params(_target_="torch.nn.Linear", in_features=20, out_features=30)
Linear(in_features=20, out_features=30, bias=True)

hydra_slayer.functional.get_instance(*args, meta_factory: Optional[Callable[[Union[Type, Callable[[...], Any]], Tuple, Mapping], Any]] = None, **kwargs) → Any

Creates instance by calling specified factory with instantiate_fn.

Note

The name of the factory to use should be provided as the first argument or directly by '_target_' keyword.

Parameters

• *args – positional arguments to pass to the factory

• meta_factory – function that calls factory in the right way. Default: factory.default_meta_factory()

• **kwargs – named parameters to pass to the factory

Returns

created instance

Examples

>>> get_instance(int, "42", base=10)
42