dataeval.scope.RepresentationOutput¶

class dataeval.scope.RepresentationOutput(data, *, leaf_coverage, total_deficit, violations, dark_branches)¶

A dataset’s collection worklist against an Ontology.

The wrapped DataFrame is the worklist itself — one row per leaf species short of its expected share, sorted by deficit (largest first) — with columns concept, label, parent, action ("acquire" for unrepresented species, "augment" for under-represented ones), count, target, and deficit. The summary scalars and supporting frames hang off it as attributes.

leaf_coverage¶

Fraction of the ontology’s leaf species with any examples (carried through from label_coverage(); observation, not policy).

Type:¶

float

total_deficit¶

Sum of all positive deficits — an estimate of how many labels the dataset is short of its expected distribution. The single budgeting number.

Type:¶

int

violations¶

One row per asserted class (from expected) whose observed share falls below its floor: concept, label, floor, actual, shortfall. Empty when no assertions were made or all held.

Type:¶

polars.DataFrame

dark_branches¶

Maximal wholly-unpopulated internal branches, largest first: concept, label, leaves (leaf species under that branch). The branch-level headline above the per-species worklist.

Type:¶

polars.DataFrame

data()¶

Return the output data as a polars DataFrame.

meta()¶

Metadata about the execution of the function or method for the Output class.

Return type:¶

ExecutionMetadata