create_forecasting_workflow#

openstef_foundation_models.presets.create_forecasting_workflow(config: ForecastingWorkflowConfig) → CustomForecastingWorkflow[source]#

Build a foundation-model forecasting workflow from a declarative config.

Resolves the checkpoint (lazily importing the inference runtime), composes the forecaster for the selected model family on the configured backend, and wraps it in a workflow with feature-selection preprocessing and quantile-sorting postprocessing. There is no training step: the model is zero-shot, so CustomForecastingWorkflow.fit() only fits the feature selector.

Parameters:
Returns:

A ready-to-use workflow composing the configured backend.

Return type:

CustomForecastingWorkflow