parallel_execute_layer#

Defined in: voxatlas.pipeline.executor

voxatlas.pipeline.executor.parallel_execute_layer(layer, registry, feature_input, n_jobs, feature_params=None)[source]#

Execute a dependency layer sequentially or in parallel.

Parameters:

  • layer (list of str) – Feature names belonging to the same dependency layer.

  • registry (FeatureRegistry) – Registry used to resolve extractor classes.

  • feature_input (FeatureInput) – Shared input bundle for the current stream.

  • n_jobs (int) – Maximum number of worker processes to use.

  • feature_params (dict | None) – Optional mapping from feature name to resolved configuration.

Returns:

Mapping from feature name to computed output.

Return type:

dict

Notes

Features in the same layer are assumed to have no unresolved interdependencies.

Examples

Usage example:

outputs = parallel_execute_layer(layer, registry, feature_input, n_jobs=1)
print(outputs.keys())