vertebrae.scoring.overlap

Internal OverlapIndex adapters.

Classes

OverlapScoreResult	Structured result from OverlapIndex or ContinuousOverlapIndex scoring.
OverlapIndexScorer	Internal adapter for MiniBatchKMeans-backed overlap scoring.

Functions

auto_k_for_class(n_class[, min_k, max_k, ...])	Resolve automatic k for a single class.
resolve_kmeans_k(y, config[, return_warnings, label_names])	Resolve MiniBatchKMeans k values per class.

Module Contents

class vertebrae.scoring.overlap.OverlapScoreResult

Structured result from OverlapIndex or ContinuousOverlapIndex scoring.

to_dict()

Serialize the score result to a JSON-safe dictionary.

Return type:

Dict[str, Any]

vertebrae.scoring.overlap.auto_k_for_class(n_class, min_k=10, max_k=50, min_samples_per_cluster=5)

Resolve automatic k for a single class.

Parameters:

• n_class (int)

• min_k (int)

• max_k (int)

• min_samples_per_cluster (int)

Return type:

int

vertebrae.scoring.overlap.resolve_kmeans_k(y, config, return_warnings=False, label_names=None)

Resolve MiniBatchKMeans k values per class.

Parameters:

• y (Any)

• config (vertebrae.config.OverlapScoringConfig)

• return_warnings (bool)

• label_names (Optional[Any])

Return type:

Union[Dict[Any, int], Tuple[Dict[Any, int], List[str]]]

class vertebrae.scoring.overlap.OverlapIndexScorer(config=None)

Internal adapter for MiniBatchKMeans-backed overlap scoring.

Parameters:

config (Optional[Union[vertebrae.config.OverlapScoringConfig, vertebrae.config.ContinuousOverlapScoringConfig]])

score(Z, y, seed=None, label_names=None, target_type='auto', target_names=None, label_catalog=None)

Score dense or sparse embeddings with OverlapIndex-family backends.

Parameters:

• Z (Any)

• y (Any)

• seed (Optional[int])

• label_names (Optional[Any])

• target_type (str)

• target_names (Optional[Any])

• label_catalog (Optional[Any])

Return type:

OverlapScoreResult