Struct RrfFusionScoring 

pub struct RrfFusionScoring {
    pub ego_depth: usize,
    pub k: f64,
}

Reciprocal-rank fusion of the structural (EGO) and lexical (BM25) signals.

The two are complementary rather than redundant: on genuine retrieval the lexical component alone outranks the deployed graph+lexical mixture, and their score-free union raises the reachable recall well above either — a miscalibrated-mixture signature. RRF fuses on ranks only, so neither component’s score scale can dominate the other, which is exactly the failure the weighted mixture had.

Fields

ego_depth: usize

k: f64

Implementations

impl RrfFusionScoring

pub fn new(ego_depth: usize) -> Self

Trait Implementations

impl ScoringStrategy for RrfFusionScoring

fn score_and_filter( &self, all_fragments: &[Fragment], core_ids: &FxHashSet<FragmentId>, hunks: &[DiffHunk], repo_root: Option<&Path>, seed_weights: Option<&FxHashMap<FragmentId, f64>>, discovered_paths: Option<&FxHashSet<Arc<str>>>, deadline: Deadline, ) -> ScoringResult