Selection Metadata Helper
GitHub Link to Code.
Metadata operations helper for feature selection system.
Provides metadata collection, processing, and validation operations for the feature selection system in PipelineData.
- class mdxplain.pipeline.helper.selection_metadata_helper.SelectionMetadataHelper
Helper class for metadata operations in feature selection system.
Provides static methods for collecting metadata for selections, processing individual feature metadata, and finalizing metadata results.
- static collect_metadata_for_selection(pipeline_data: PipelineData, name: str) List[Dict[str, Any]]
Collect metadata for all features in the selection using stored reference trajectory.
Parameters
- pipeline_dataPipelineData
Pipeline data object containing feature data
- namestr
Name of the selection to collect metadata for
Returns
- list
List of metadata for all features in the selection
Raises
- ValueError
If the selection does not exist or reference trajectory not set/valid
- static finalize_metadata_result(selected_metadata: list, name: str) ndarray
Finalize and validate the metadata result.
Parameters
- selected_metadatalist
List of metadata for all features in the selection
- namestr
Name of the selection to get metadata for
Returns
- numpy.ndarray
Array of dictionaries, one for each column in the selected matrix.
Raises
- ValueError
If no valid metadata is found for the selection