dataeval.scope¶

Evaluate data completeness and coverage of a dataset’s label and embedding space.

Scope evaluators assess whether a dataset adequately spans the space it is meant to cover, identifying gaps and prioritizing new data for labeling — across an ontology’s label space (Representation) and a dataset’s latent embedding space (Coverage, Prioritize).

Classes¶

`Coverage`	Evaluate a dataset’s embedding-space coverage and per-class variety.
`Prioritize`	Prioritize dataset samples based on their position in the embedding space.
`Representation`	Evaluate a dataset’s coverage of an ontology and prioritize what to collect.

Output Classes¶

`CoverageOutput`	A dataset’s per-class embedding-space coverage.
`PrioritizeOutput`	Ranking result with lazy index computation based on order and policy.
`RepresentationOutput`	A dataset’s collection worklist against an `Ontology`.