load_mlflow_gateway_config#

scikitplot.llm_provider.load_mlflow_gateway_config(path)[source]#

Load MLflow Gateway model configuration from a YAML file.

Parses a configuration file defining multiple model endpoints. Supports environment variable expansion and validates expected structure.

Parameters:
pathstr

Path to the MLflow Gateway YAML file. May contain ~ or environment variables.

Returns:
dict of str to list of dict

A dictionary where each key is a provider name and each value is a list of dictionaries with keys: - ‘model_id’: str - ‘api_key’: str (may be empty if not resolved)

Raises:
ScikitplotException

If the YAML is malformed, missing fields, or the file doesn’t exist.

Parameters:

path (str)

Return type:

dict[str, list[dict[str, any]]]

Notes

  • API keys can be hardcoded or passed via environment variables using $VARNAME or ${VARNAME}.

  • This loader is tolerant to partial failures but logs all issues.

Examples

>>> cfg = load_mlflow_gateway_config("~/gateway.yaml")
>>> cfg["openai"][0]["model_id"]
'gpt-4'