Struct FlatEmbeddings 

pub struct FlatEmbeddings {
    pub data: Arc<[f32]>,
    pub dim: usize,
}

Flat row-major embedding matrix: data is n_vectors × dim floats in one contiguous heap allocation. Used everywhere in the dense search pipeline so the same allocation backs both the HNSW graph cache and per-query scoring — zero copies between layers.

Fields

data: Arc<[f32]>

dim: usize

Implementations

impl FlatEmbeddings

pub fn n_vectors(&self) -> usize

Number of vectors in the matrix.

pub fn get(&self, i: usize) -> &[f32]

View the i-th vector as a &[f32].

pub fn get_vec(&self, i: usize) -> Vec<f32>

Extract the i-th vector into an owned Vec<f32>.

pub fn from_vecs(vecs: Vec<Vec<f32>>) -> Self

Build from a Vec<Vec<f32>> (for tests / migration).

Trait Implementations

impl Clone for FlatEmbeddings

fn clone(&self) -> FlatEmbeddings

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for FlatEmbeddings

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

Formats the value using the given formatter.