Enum FusionAlgorithm 

pub enum FusionAlgorithm {
    RRF {
        k: f64,
    },
    WeightedSum {
        lexical_weight: f32,
        vector_weight: f32,
    },
}

Algorithm used to combine lexical and vector scores in hybrid search.

The default fusion algorithm (when none is specified in a SearchRequest) is RRF with k = 60.0.

Variants

RRF

Reciprocal Rank Fusion (RRF).

Combines results based on rank position rather than raw scores, making it effective when score magnitudes are not comparable (e.g. BM25 vs cosine similarity). The score for each document is sum(1 / (k + rank)) across the result lists.

Fields

k: f64

Smoothing constant k. Higher values reduce the influence of top-ranked documents. Typical default is 60.0.

WeightedSum

Weighted Sum with automatic min-max score normalization.

Before weighting, the engine independently normalizes lexical and vector scores to the [0.0, 1.0] range using min-max normalization over their respective result sets.

Fields

lexical_weight: f32

Weight for the normalized lexical score (clamped to 0.0..=1.0).

vector_weight: f32

Weight for the normalized vector score (clamped to 0.0..=1.0).

Trait Implementations

impl Clone for FusionAlgorithm

fn clone(&self) -> FusionAlgorithm

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for FusionAlgorithm

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Copy for FusionAlgorithm