Struct VectorSpaceService 

pub struct VectorSpaceService { /* private fields */ }

The main service for vector space operations.

This service provides a thread-safe interface for:

• Adding and removing embeddings
• Nearest neighbor search
• Building similarity graphs

Implementations

impl VectorSpaceService

pub fn new(config: HnswConfig) -> Self

Create a new vector space service with the given configuration.

pub fn from_index(index: HnswIndex, config: HnswConfig) -> Self

Create a service from an existing index.

pub fn dimensions(&self) -> usize

Get the index dimensions.

pub fn len(&self) -> usize

Get the current number of vectors.

pub fn is_empty(&self) -> bool

Check if the index is empty.

pub fn config(&self) -> &HnswConfig

Get a reference to the configuration.

pub async fn add_embedding( &self, id: EmbeddingId, vector: Vec<f32>, ) -> Result<(), VectorError>

Add a single embedding to the index.

The vector will be normalized if the configuration specifies normalization.

pub async fn add_embeddings_batch( &self, items: Vec<(EmbeddingId, Vec<f32>)>, ) -> Result<usize, VectorError>

Add multiple embeddings in a batch.

This is more efficient than multiple single adds due to amortized locking overhead.

pub async fn find_neighbors( &self, query: &[f32], k: usize, ) -> Result<Vec<Neighbor>, VectorError>

Find the k nearest neighbors to a query vector.

pub async fn find_neighbors_with_options( &self, query: &[f32], options: SearchOptions, ) -> Result<Vec<Neighbor>, VectorError>

Find neighbors with custom search options.

pub async fn find_neighbors_with_filter<F>( &self, query: &[f32], k: usize, filter: F, ) -> Result<Vec<Neighbor>, VectorError>

where F: Fn(&EmbeddingId) -> bool + Send + Sync,

Find neighbors using a filter predicate.

The filter function receives an EmbeddingId and returns true if the embedding should be included in results.

pub async fn remove_embedding( &self, id: &EmbeddingId, ) -> Result<(), VectorError>

Remove an embedding from the index.

pub fn contains(&self, id: &EmbeddingId) -> bool

Check if an embedding exists in the index.

pub fn get_vector(&self, id: &EmbeddingId) -> Option<Vec<f32>>

Get a vector by its ID.

pub async fn build_similarity_edges( &self, id: EmbeddingId, vector: &[f32], k: usize, min_similarity: f32, ) -> Result<Vec<SimilarityEdge>, VectorError>

Build similarity edges for an embedding.

This finds the k nearest neighbors and creates edges to them.

pub async fn compute_pairwise_similarities( &self, vectors: &[(EmbeddingId, Vec<f32>)], ) -> Result<Vec<(EmbeddingId, EmbeddingId, f32)>, VectorError>

Compute pairwise similarities for a set of embeddings.

pub async fn clear(&self) -> Result<(), VectorError>

Clear all embeddings from the index.

pub async fn save(&self, path: &Path) -> Result<(), VectorError>

Save the index to a file.

pub async fn load(path: &Path, config: HnswConfig) -> Result<Self, VectorError>

Load an index from a file.

pub fn stats(&self) -> IndexStatistics

Get index statistics.

Trait Implementations

impl Clone for VectorSpaceService

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.