PipelineStep¶
- class verta.pipeline.PipelineStep(name: str, registered_model_version: RegisteredModelVersion, predecessors: Optional[Union[List[PipelineStep], Set[PipelineStep], Tuple[PipelineStep]]] = None)¶
Object representing a single step to be run within an inference pipeline.
- Parameters:
name (str) – Name of the step, for use within the scope of the pipeline only.
registered_model_version (
RegisteredModelVersion
) – Registered model version to run for this step.predecessors (list, set, tuple, optional) – List, set, or tuple of unique PipelineStep objects whose outputs will be treated as inputs to this step. If not included, the step is assumed to be an initial step.
- Variables:
name (str) – Name of the step within the scope of the pipeline.
registered_model_version (
RegisteredModelVersion
) – Registered model version run by this step.predecessors (set) – Set of PipelineSteps whose outputs will be treated as inputs to this step.
- set_predecessors(steps: Optional[Union[List[PipelineStep], Set[PipelineStep], Tuple[PipelineStep]]] = None) Set[PipelineStep] ¶
Set the predecessors associated with this step.
- Parameters:
steps (list, set, or tuple, optional) – List, set, or tuple of PipelineStep objects whose outputs will be treated as inputs to this step. All options are converted to a set, so order is irrelevant and duplicates are removed. An empty set is used if no input is provided.
- Returns:
set of PipelineStep – The new set of predecessors now set for this step.
- Raises:
TypeError – If the provided value for
steps
is not a set of PipelineStep objects.
- set_registered_model_version(registered_model_version: RegisteredModelVersion) RegisteredModelVersion ¶
Set the registered model version associated with this step.
- Parameters:
registered_model_version (
RegisteredModelVersion
) – Registered model version to use for the step.- Returns:
RegisteredModelVersion – The new registered model version now set for this step.
- Raises:
TypeError – If the provided value is not type RegisteredModelVersion.