VectorSearchProvider

avocado_core::storage

Trait VectorSearchProvider 

Source 
pub trait VectorSearchProvider: Send + Sync {
    // Required methods
    fn search<'life0, 'life1, 'async_trait>(
        &'life0 self,
        query_embedding: &'life1 [f32],
        k: usize,
    ) -> Pin<Box<dyn Future<Output = Result<Vec<VectorSearchResult>>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait;
    fn len(&self) -> usize;
    fn dimension(&self) -> usize;

    // Provided method
    fn is_empty(&self) -> bool { ... }
}
Expand description

Vector search provider trait

Implementations provide vector similarity search capabilities. This abstracts over HNSW (for SQLite) and pgvector (for PostgreSQL).

Required Methods§

Source

fn search<'life0, 'life1, 'async_trait>( &'life0 self, query_embedding: &'life1 [f32], k: usize, ) -> Pin<Box<dyn Future<Output = Result<Vec<VectorSearchResult>>> + Send + 'async_trait>>
where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Search for similar spans

§Arguments
  • query_embedding - Query vector
  • k - Number of results to return
§Returns

Vector of scored spans, sorted by relevance (highest score first)

Source

fn len(&self) -> usize

Get number of indexed spans

Source

fn dimension(&self) -> usize

Get embedding dimension

Provided Methods§

Source

fn is_empty(&self) -> bool

Check if index is empty

Implementors§