Trait VectorDB 

pub trait VectorDB: Send + Sync {
    // Required methods
    fn create_collection<'life0, 'life1, 'life2, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
        dimension: usize,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait;
    fn has_collection<'life0, 'life1, 'life2, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<bool>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait;
    fn index_points<'life0, 'life1, 'life2, 'life3, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
        points: &'life3 [VectorPoint],
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait,
             'life3: 'async_trait;
    fn search_similar<'life0, 'life1, 'life2, 'life3, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
        query_vector: &'life3 [f32],
        top_k: usize,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<Vec<SearchResult>>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait,
             'life3: 'async_trait;
    fn delete_collection<'life0, 'life1, 'life2, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait;
    fn collection_size<'life0, 'life1, 'life2, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<usize>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait;

    // Provided methods
    fn delete_points<'life0, 'life1, 'life2, 'life3, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
        point_ids: &'life3 [Uuid],
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait,
             'life3: 'async_trait { ... }
    fn list_collections<'life0, 'async_trait>(
        &'life0 self,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<Vec<(String, String)>>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait { ... }
    fn prune<'life0, 'async_trait>(
        &'life0 self,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait { ... }
    fn batch_search_similar<'life0, 'life1, 'life2, 'life3, 'async_trait>(
        &'life0 self,
        data_type: &'life1 str,
        field_name: &'life2 str,
        query_vectors: &'life3 [Vec<f32>],
        top_k: usize,
    ) -> Pin<Box<dyn Future<Output = VectorDBResult<Vec<Vec<SearchResult>>>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait,
             'life3: 'async_trait { ... }
}

Vector database trait

Required Methods

fn create_collection<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, dimension: usize, ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Create a collection for (data_type, field_name) pair

Arguments

• data_type - Type name (e.g., “DocumentChunk”, “Entity”)
• field_name - Field name (e.g., “text”, “name”)
• dimension - Vector dimension (e.g., 384 for MiniLM)

Example

vector_db.create_collection("DocumentChunk", "text", 384).await?;

fn has_collection<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, ) -> Pin<Box<dyn Future<Output = VectorDBResult<bool>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Check if collection exists

Arguments

• data_type - Type name
• field_name - Field name

fn index_points<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, points: &'life3 [VectorPoint], ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

Index data points (batch upsert with embeddings already generated)

Arguments

• data_type - Type name
• field_name - Field name
• points - Vector points with embeddings

Example

let points = vec![
    VectorPoint::new(chunk_id, embedding)
        .with_metadata("type", json!("DocumentChunk"))
        .with_metadata("field", json!("text")),
];
vector_db.index_points("DocumentChunk", "text", &points).await?;

fn search_similar<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, query_vector: &'life3 [f32], top_k: usize, ) -> Pin<Box<dyn Future<Output = VectorDBResult<Vec<SearchResult>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

Search for similar vectors

Arguments

• data_type - Type name
• field_name - Field name
• query_vector - Query embedding vector
• top_k - Number of results to return

Returns

Vector of search results sorted by similarity (descending)

fn delete_collection<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Delete collection

fn collection_size<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, ) -> Pin<Box<dyn Future<Output = VectorDBResult<usize>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Get collection statistics

Provided Methods

fn delete_points<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, point_ids: &'life3 [Uuid], ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

Delete points by IDs from an existing collection.

fn list_collections<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = VectorDBResult<Vec<(String, String)>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

List all existing vector collections as (data_type, field_name) pairs.

Default implementation returns an empty list. Backends should override to return the actual collections they hold.

fn prune<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = VectorDBResult<()>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Remove all vector collections.

Default implementation lists all collections and deletes each one. Backends may override with a more efficient bulk operation.

Equivalent to Python’s vector_engine.prune().

fn batch_search_similar<'life0, 'life1, 'life2, 'life3, 'async_trait>( &'life0 self, data_type: &'life1 str, field_name: &'life2 str, query_vectors: &'life3 [Vec<f32>], top_k: usize, ) -> Pin<Box<dyn Future<Output = VectorDBResult<Vec<Vec<SearchResult>>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, 'life3: 'async_trait,

Perform multiple vector similarity searches in sequence.

Default implementation loops over [search_similar]. Backends may override this with a native batch API for better performance.

Dyn Compatibility

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors

impl VectorDB for BruteForceVectorDB

impl VectorDB for LanceDbAdapter