nemo_curator.pipeline.workflow

Module Contents

Classes

Name	Description
`WorkflowBase`	-
`WorkflowRunResult`	Container returned by high-level workflows to expose pipeline outputs.

API

class nemo_curator.pipeline.workflow.WorkflowBase()

Abstract

nemo_curator.pipeline.workflow.WorkflowBase.run(
    args = (),
    kwargs = {}
) -> nemo_curator.pipeline.workflow.WorkflowRunResult

abstract

class nemo_curator.pipeline.workflow.WorkflowRunResult(
    workflow_name: str,
    pipeline_tasks: dict[str, list[nemo_curator.tasks.Task]] = dict(),
    metadata: dict[str, typing.Any] = dict()
)

Dataclass

Container returned by high-level workflows to expose pipeline outputs.

metadata

dict[str, Any] = field(default_factory=dict)

pipeline_tasks

dict[str, list[Task]] = field(default_factory=dict)

workflow_name

str

nemo_curator.pipeline.workflow.WorkflowRunResult.add_metadata(
    key: str,
    value: typing.Any
) -> None

Add a metadata key-value pair.

nemo_curator.pipeline.workflow.WorkflowRunResult.add_pipeline_tasks(
    pipeline_name: str,
    tasks: list[nemo_curator.tasks.Task] | None
) -> None

Record the tasks emitted by a pipeline run (empty list if None).

nemo_curator.pipeline.workflow.WorkflowRunResult.extend_metadata(
    updates: dict[str, typing.Any] | None = None
) -> None

Update metadata dictionary in-place.

nemo_curator.pipeline.workflow.WorkflowRunResult.get_metadata(
    key: str
) -> typing.Any

Get a metadata value.