libera_utils.scene_identification.scene_id.standard_scene_definitions

libera_utils.scene_identification.scene_id.standard_scene_definitions(scene_types: list[str] | None = None) → list[SceneDefinition]

Build the standard libera_utils scene definitions, optionally limited to specific types.

Parameters:

scene_types (list of str or None) – Scene types to include, drawn from the keys of STANDARD_SCENE_DEFINITION_CONFIG_KEYS ("trmm", "erbe", "unfiltering"). If None (default), all standard definitions are returned in that dictionary’s insertion order.

Returns:

The requested standard scene definitions, loaded from the CSV paths configured in config.json.

Return type:

list[SceneDefinition]

Raises:

ValueError – If a requested scene type is not a known standard definition.

Notes

These are only the definitions shipped with libera_utils. A user can supply a custom scene definition without editing library code: build a CSV with the same column layout as the shipped definitions in libera_utils/data/scene_definitions/ (scene_id plus <variable>_Min / <variable>_Max columns), wrap it in a SceneDefinition, and pass it through the additional_scene_definitions_files argument of FootprintData.identify_scenes() (or pass an explicit scene_definitions list). The CSV is the source of truth for the bins, which is what makes the classification reconfigurable in an analysis context. TODO[LIBSDC-857]: expand this into fuller end-user documentation of the custom-scene-definition workflow.