ForEach¶

class laminar.configurations.layers.ForEach(parameters: ~typing.Iterable[~laminar.configurations.layers.Parameter] = <factory>)¶

Bases: object

Configures a layer to perform a grid-foreach over the configured properties.

Notes

In order to foreach over each element of a layer, the layer’s artifact must be sharded. See Layer.shard().

If Parameter(index=None), ForEach will include artifacts from all Layer splits.

Usage:

@Flow.register(foreach=ForEach(...))

Methods

`grid`	Generate a grid of all combinations of foreach inputs.
`join`	Join together multiple artifact splits of a layer into a single Archive.
`set`	Set a foreach layer's parameters given the inputs from the foreach grid.
`splits`	Get the splits of the ForEach grid.

Attributes

`parameters`	Parameters to configure the foreach with.

__init__(parameters: ~typing.Iterable[~laminar.configurations.layers.Parameter] = <factory>)¶

grid(*, layer: Layer) → list[dict[Layer, dict[str, int]]]¶

Generate a grid of all combinations of foreach inputs.

Parameters¶

Returns¶

Index ordered inputs of layers mapped to attributes mapped to Accessor: index.

Return type¶

List[Dict[Layer, Dict[str, int]]]

join(*, layer: Layer, name: str) → Archive¶

Join together multiple artifact splits of a layer into a single Archive.

Parameters¶

Returns¶

Archive created from multiple layer archives.

Return type¶

datastores.Archive

parameters : Iterable[Parameter]¶: Parameters to configure the foreach with.

set(*, layer: Layer, parameters: tuple[Layer, ...]) → tuple[Layer, ...]¶

Set a foreach layer’s parameters given the inputs from the foreach grid.

Parameters¶

layer : Layer¶: Layer the ForEach is configured for.
parameters : Tuple[Layer, ...]¶: Parameters in the order they will be passed to the layer.

Returns¶

Parameters with modified values for the foreach evaluation.

Return type¶

Tuple[Layer, …]