Struct CachedNorm 

pub struct CachedNorm { /* private fields */ }

Pre-computed L2 norm for efficient repeated cosine similarity calculations.

When comparing a query vector against many candidates, computing the query’s norm once and reusing it is more efficient than recomputing it for each comparison.

Example

use manifoldb_vector::distance::CachedNorm;

let query = [1.0, 2.0, 3.0];
let cached_norm = CachedNorm::new(&query);

// Use cached norm for many comparisons
for candidate in candidates {
    let candidate_norm = CachedNorm::new(&candidate);
    let similarity = cached_norm.cosine_similarity_to(&query, &candidate, &candidate_norm);
}

Implementations

impl CachedNorm

pub fn new(v: &[f32]) -> Self

Compute and cache the L2 norm of a vector.

pub const fn norm(&self) -> f32

Get the cached L2 norm.

pub const fn norm_squared(&self) -> f32

Get the cached squared L2 norm.

pub fn is_zero(&self) -> bool

Check if the vector has zero magnitude.

Trait Implementations

impl Clone for CachedNorm

fn clone(&self) -> CachedNorm

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CachedNorm

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

Formats the value using the given formatter.

impl Copy for CachedNorm