lmflow.utils.deprecated#
Utilities for handling deprecated APIs and maintaining backwards compatibility.
Functions#
|
Decorator to handle deprecated function arguments. |
Module Contents#
- lmflow.utils.deprecated.deprecated_args(**deprecated_params: dict[str, Any])[source]#
Decorator to handle deprecated function arguments.
- Args:
- **deprecated_params: Mapping of deprecated argument names to their configuration.
Each value should be a dict with: - ‘replacement’: Name of the new argument (optional) - ‘mapper’: Function to map old value to new value (optional) - ‘message’: Custom deprecation message (optional)
- Example:
- @deprecated_args(
- use_vllm={
‘replacement’: ‘inference_engine’, ‘mapper’: lambda x: ‘vllm’ if x else ‘huggingface’, ‘message’: “use_vllm is deprecated. Use inference_engine=’vllm’ instead.”
}
) def my_function(inference_engine=’huggingface’, **kwargs):
pass