BaseEvaluator

evaluator.BaseEvaluator(feedback=None)

Abstract base class for evaluation strategies.

An evaluator compares predictions with the source observation context and selects the most relevant candidate geometries.

Methods

Name	Description
evaluate	Execute evaluation strategy for the given aligner context.
get_brdr_observation_from_properties	Extract a BRDR observation from thematic feature properties.
get_observation_comparison_properties	Compare current observation against base observation.
get_observation_properties	Build observation properties for a process result.
update_evaluation_with_original	Update evaluated output by returning the original geometry.

evaluate

evaluator.BaseEvaluator.evaluate(
    aligner,
    relevant_distances=None,
    thematic_ids=None,
    metadata_field=METADATA_FIELD_NAME,
    full_reference_strategy=FullReferenceStrategy.NO_FULL_REFERENCE,
    max_predictions=-1,
    multi_to_best_prediction=True,
)

Execute evaluation strategy for the given aligner context.

get_brdr_observation_from_properties

evaluator.BaseEvaluator.get_brdr_observation_from_properties(
    aligner,
    id_theme,
    base_metadata_field,
)

Extract a BRDR observation from thematic feature properties.

get_observation_comparison_properties

evaluator.BaseEvaluator.get_observation_comparison_properties(
    aligner,
    process_result,
    base_brdr_observation=None,
)

Compare current observation against base observation.

get_observation_properties

evaluator.BaseEvaluator.get_observation_properties(aligner, process_result)

Build observation properties for a process result.

update_evaluation_with_original

evaluator.BaseEvaluator.update_evaluation_with_original(
    aligner,
    metadata_field,
    original_geometry,
    process_results_evaluated,
    theme_id,
    evaluation,
)

Update evaluated output by returning the original geometry.