run_demo#

scikitplot.mlflow.run_demo(*, profile='local', project_root=None, open_ui_seconds=10.0, experiment_name=None, fmt='toml', overwrite=False)[source]#

Run a beginner-friendly end-to-end demo workflow.

The workflow demonstrates: 1) Export demo config shipped in the library into your project. 2) Optionally customize the config (experiment name). 3) Start a session and log a “train” run. 4) Export project config to YAML (backup / editable format). 5) Re-open session from YAML and keep UI open briefly. 6) Start a “predict” run.

Parameters:
profilestr, default=”local”

Profile name in the config file.

project_rootpathlib.Path or None, default=None

Project root. If None, auto-detected.

open_ui_secondsfloat, default=10.0

How long to keep the UI reachable (sleep) in the “UI check” step. Set to 0 to skip sleeping.

experiment_namestr or None, default=None

If provided, patch the exported config to use this experiment name.

fmt{“toml”, “yaml”}, default=”toml”

Which built-in config format to export initially.

overwritebool, default=False

Overwrite existing config files in the project.

Returns:
WorkflowPaths

Paths used in the workflow.

Raises:
ImportError

If MLflow is not installed when attempting to start a session.

Parameters:
  • profile (str)

  • project_root (Path | None)

  • open_ui_seconds (float)

  • experiment_name (str | None)

  • fmt (str)

  • overwrite (bool)

Return type:

WorkflowPaths