ModelsFacade

class scikitplot.mlflow.ModelsFacade(mlflow_module=None, client=None)

Model helper facade bound to a session-bound MLflow client.

Parameters:

mlflow_modulemodule

Imported mlflow module.

clientMlflowClient

Client bound to the session tracking/registry URIs.

Parameters:

• mlflow_module (Any)

• client (Any)

Notes

This facade intentionally stays thin and uses MLflow public APIs.

client: Any = None

load_model(model_uri, *, flavor=None)

Load a model by URI.

Parameters:

model_uristr

Model URI (e.g., “runs:/<run_id>/model” or “models:/Name/Stage”).

flavorstr or None, default=None

Optional flavor to load. If provided, attempts to load via mlflow.<flavor>.load_model. Otherwise, uses mlflow.pyfunc.load_model.

Returns:

Any

Loaded model object.

Raises:

AttributeError

If requested flavor is not available.

Parameters:

• model_uri (str)

• flavor (str | None)

Return type:

Any

mlflow_module: Any = None

register_model(model_uri, name)

Register a model version.

Parameters:

model_uristr

Model source URI.

namestr

Registered model name.

Returns:

Any

Model version object (type depends on MLflow version).

Parameters:

• model_uri (str)

• name (str)

Return type:

Any