← Back to API reference

neuralset.base.Step

pydantic model neuralset.base.Step

Base class for composable pipeline nodes.

A Step represents a single operation in a data processing pipeline, such as loading a study (Study) or transforming events (EventsTransform). Steps can be executed individually via .run() or grouped into a Chain.

Inherits caching and execution infrastructure from exca.steps.Step (configured via the infra parameter).

Parameters:

infra (exca.steps.Backend, optional) – Caching and execution infrastructure. Supports caching to disk (e.g. Cached backend) and executing remotely on a cluster (e.g. Slurm). Pydantic will coerce a dict, e.g. {"backend": "Cached", "folder": "~/.cache/neuralset"}.

Fields:

requirements: tp.ClassVar[tuple[str, ...]] = ()

CACHE_TYPE: ClassVar[str | None] = None

← Back to API reference