gigl.experimental.knowledge_graph_embedding.lib.config.hydra_utils#

Functions#

`build_hydra_dict_from_object`(obj)	Builds a Hydra-compatible dictionary from a dataclass or namedtuple.
`get_fully_qualified_name`(cls_or_obj)	Returns the fully qualified name (module.ClassName) for a class or object.

Module Contents#

gigl.experimental.knowledge_graph_embedding.lib.config.hydra_utils.build_hydra_dict_from_object(obj)[source]#

Builds a Hydra-compatible dictionary from a dataclass or namedtuple.

Recursively processes object attributes and adds ‘_target_’ fields for Hydra instantiation. Handles nested structures, lists, and various Python types. Autonomously converts NewType instances to their base string representation.

Parameters:

obj (Any) – The object to convert. Can be a dataclass, namedtuple, list, enum, or primitive type (str, int, float, bool, None).

Returns:

A dictionary (for complex objects) or the original value (for primitives): that can be used in Hydra configuration files. Complex objects include a ‘_target_’ field pointing to their fully qualified class name.

Return type:

Any

gigl.experimental.knowledge_graph_embedding.lib.config.hydra_utils.get_fully_qualified_name(cls_or_obj)[source]#

Returns the fully qualified name (module.ClassName) for a class or object.

Handles both classes and instances by getting the type of the instance. This is useful for generating Hydra _target_ paths for dynamic instantiation.

Parameters:: cls_or_obj (Union[Type, Any]) – Either a class type or an instance of a class.
Returns:: The fully qualified name in the format “module.ClassName”.
Return type:: str