Enum Normalization 

pub enum Normalization {
    MinMax,
    ZScore,
    Sum,
    Rank,
    None,
}

Score normalization methods.

Different retrievers produce scores on different scales. Normalization puts them on a common scale before combining.

Variants

MinMax

Min-max normalization: (score - min) / (max - min) → [0, 1]

Best when score distributions are similar. Sensitive to outliers.

ZScore

Z-score normalization: (score - mean) / std, clipped to [-3, 3]

More robust to outliers. Better when distributions differ.

Sum

Sum normalization: score / sum(scores)

Preserves relative magnitudes. Useful when scores represent probabilities.

Rank

Rank-based: convert scores to ranks, then normalize

Sorts input by score (descending), assigns ranks 0..n-1, then normalizes to [0, 1] range where rank 0 (best) → 1.0, rank n-1 (worst) → 1/n. Ignores score magnitudes entirely. Most robust but loses information.

None

No normalization: use raw scores

Only use when all retrievers use the same scale.

Trait Implementations

impl Clone for Normalization

fn clone(&self) -> Normalization

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Normalization

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

Formats the value using the given formatter.

impl Default for Normalization

fn default() -> Normalization

Returns the “default value” for a type.

impl PartialEq for Normalization

fn eq(&self, other: &Normalization) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for Normalization

impl Eq for Normalization

impl StructuralPartialEq for Normalization