Struct VectorDatabase 

pub struct VectorDatabase { /* private fields */ }

👎Deprecated since 0.1.0:

Use SemanticSearch over storage-backed VectorEntry slices for runtime retrieval. VectorDatabase is an internal/test abstraction.

Vector database for storing and searching embeddings

Deprecated: VectorDatabase is an in-memory test/development abstraction. The live retrieval path uses SemanticSearch over storage-backed VectorEntry slices (see crate::search::SemanticSearch). SemanticSearch is the actual runtime retrieval path used by RepresentationService.

Implementations

impl VectorDatabase

pub fn new() -> Self

Create a new vector database

pub fn with_dimension(dimension: usize) -> Self

Create with custom dimension

pub async fn in_memory() -> Result<Self>

Create a new in-memory database (async for API compatibility)

pub fn insert_with_priority( &mut self, entry: VectorEntry, priority: Option<u8>, ) -> Result<()>

Insert a vector with priority for graph tree

pub fn insert(&mut self, entry: VectorEntry) -> Result<()>

Insert a vector (backward compatible)

pub fn get(&self, id: i64) -> Option<&VectorEntry>

Get a vector by ID

pub fn remove(&mut self, id: i64) -> Option<VectorEntry>

Remove a vector by ID

pub fn ids(&self) -> Vec<i64>

Get all vector IDs

pub fn len(&self) -> usize

Get vector count

pub fn is_empty(&self) -> bool

Check if empty

pub fn by_namespace(&self, namespace_id: i64) -> Vec<&VectorEntry>

Get vectors by namespace

pub fn by_category(&self, category: &str) -> Vec<&VectorEntry>

Get vectors by category

pub fn dimension(&self) -> usize

Get the embedding dimension

pub fn tree(&self) -> &GraphTree

Get reference to the graph tree

pub fn tree_mut(&mut self) -> &mut GraphTree

Get mutable reference to the graph tree

pub fn search( &self, query: &[f32], namespace_id: i64, limit: usize, threshold: f32, ) -> Result<(Vec<VectorSearchResult>, SearchLatency)>

Search for similar vectors

Returns results sorted by boosted score (descending). Target latency: <10ms

pub fn search_by_category( &self, query: &[f32], namespace_id: i64, category: &str, limit: usize, threshold: f32, ) -> Result<(Vec<VectorSearchResult>, SearchLatency)>

Search with category filter

pub fn insert_batch(&mut self, entries: Vec<VectorEntry>) -> Result<usize>

Batch insert multiple vectors

More efficient than individual inserts for bulk loading.

pub fn insert_batch_with_priorities( &mut self, entries: Vec<(VectorEntry, Option<u8>)>, ) -> Result<usize>

Batch insert with priorities

pub fn remove_batch(&mut self, ids: &[i64]) -> Vec<Option<VectorEntry>>

Batch remove multiple vectors

pub fn search_batch( &self, queries: &[Vec<f32>], namespace_id: i64, limit: usize, threshold: f32, ) -> Result<Vec<(Vec<VectorSearchResult>, SearchLatency)>>

Search with multiple queries (for batch operations)

pub fn find_similar( &self, memory_id: i64, limit: usize, threshold: f32, ) -> Result<(Vec<VectorSearchResult>, SearchLatency)>

Find similar vectors to a stored vector by ID

pub fn stats(&self) -> VectorDatabaseStats

Get statistics about the vector database

pub fn clear(&mut self)

Clear all vectors from the database

pub fn contains(&self, id: i64) -> bool

Check if a vector exists

pub fn all_vectors(&self) -> Vec<&VectorEntry>

Get all vectors as a slice (read-only access)

pub fn update_embedding( &mut self, id: i64, new_embedding: Vec<f32>, ) -> Result<()>

Update an existing vector’s embedding

Trait Implementations

impl Debug for VectorDatabase

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

Formats the value using the given formatter.

impl Default for VectorDatabase

fn default() -> VectorDatabase

Returns the “default value” for a type.