`agrag.eval.answer.CitationScoreBreakdown`

Bases: TypedDict

Available precision, recall, and sentence-count fields for one score.

All fields are optional because abstentions and uncited answers have partial breakdowns.

Attributes:

citation_precision (float) – Fraction of cited sentences supported by evidence.
citation_recall (float) – Fraction of eligible sentences supported by evidence.
cited_sentences (int) – Number of sentences with citations.
supported_sentences (int) – Number of cited sentences supported by evidence.
sentences (int) – Number of eligible sentences in the answer.
sentence_rows (list[CitationSentenceRow]) – One row per cited sentence with its keys and verdict.

`citation_precision`

citation_precision: float

citation_recall: float

cited_sentences: int

sentence_rows: list[CitationSentenceRow]

sentences: int

supported_sentences: int