Struct RerankerObjective 

pub struct RerankerObjective {
    pub reranker_name: String,
}

Scores a NoteCandidate using a named reranker’s pre-computed score.

Looks up candidate.rerank_scores[reranker_name]. Returns 0.0 when the reranker was not run (key absent) — callers should gate on RecallConfig.reranker_weights[name] > 0.0 before including this objective in a WeightedObjective composition.

Fields

reranker_name: String

Name of the reranker to look up in candidate.rerank_scores.

Implementations

impl RerankerObjective

pub fn new(name: impl Into<String>) -> Self

Create a new objective for the named reranker.

Trait Implementations

impl Objective<NoteCandidate> for RerankerObjective

fn score(&self, candidate: &NoteCandidate, _context: &ObjectiveContext) -> f64

Evaluate a single candidate.

fn name(&self) -> &str

Get the name of this objective.

fn precision(&self, _candidate: &T, _context: &ObjectiveContext) -> f64

Precision (inverse variance) of the score estimate; default 1.0 (fully trusted).

fn passes_score(&self, score: f64, context: &ObjectiveContext) -> bool

Check if a score passes the threshold; non-finite scores never pass.

fn passes(&self, candidate: &T, context: &ObjectiveContext) -> bool

Check if a candidate passes the threshold.

fn batch_score( &self, candidates: &[T], context: &ObjectiveContext, ) -> Vec<(usize, f64)>

Score a batch of candidates and return passing (index, score) pairs.

fn select<'a>( &self, candidates: &'a [T], context: &ObjectiveContext, ) -> Vec<Selection<&'a T>>

Select all passing candidates in score-descending order.

fn select_top<'a>( &self, candidates: &'a [T], n: usize, context: &ObjectiveContext, ) -> Vec<Selection<&'a T>>

Select the top N candidates by precision-weighted score.