Trait ANNIndexExternal

pub trait ANNIndexExternal<'a, const N: usize> {
    type Index;

    // Required methods
    fn build<R: Rng>(
        num_trees: usize,
        max_leaf_size: usize,
        vectors: &'a [Vector<N>],
        rng: &mut R,
    ) -> Result<Self::Index, &'static str>;
    fn search(&self, query: &Vector<N>, top_k: usize) -> Vec<(usize, f32)>;
    fn memory_usage(&self) -> usize;
}

A trait for ANNIndex that references external data. The index is read-only once built.

Required Associated Types

type Index

Required Methods

fn build<R: Rng>( num_trees: usize, max_leaf_size: usize, vectors: &'a [Vector<N>], rng: &mut R, ) -> Result<Self::Index, &'static str>

Build the index for the given vectors and ids.

Arguments

• num_trees - The number of trees to build, higher means more accurate but slower and larger memory usage.
• max_leaf_size - The maximum number of vectors in a leaf node, lower means higher accuracy but slower search.
• vectors - The vectors to build the index from.

fn search(&self, query: &Vector<N>, top_k: usize) -> Vec<(usize, f32)>

Search for the top_k nearest neighbors of the query vector.

Arguments

• query - The query vector to search for.
• top_k - The number of nearest neighbors to return.

Returns

• Vec<(usize, f32)> - The top_k nearest (index, distance) of the query vector.

fn memory_usage(&self) -> usize

Get the memory usage of the index.

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety", so this trait is not object safe.

Implementors

impl<'a, const N: usize> ANNIndexExternal<'a, N> for LinearSearchExternal<'a, N>

type Index = LinearSearchExternal<'a, N>

impl<'a, const N: usize> ANNIndexExternal<'a, N> for LshExternal<'a, N>

type Index = LshExternal<'a, N>