Struct EmbeddingRetriever 

pub struct EmbeddingRetriever<P: EmbeddingProvider> { /* private fields */ }

In-memory vector retriever using cosine similarity search.

Stores document embeddings and retrieves the top-k most similar documents for a query, optionally filtered by a minimum similarity threshold.

Implementations

impl<P: EmbeddingProvider> EmbeddingRetriever<P>

pub fn new(provider: P) -> Self

Create a new retriever backed by the given EmbeddingProvider.

pub fn with_similarity_threshold(self, threshold: f64) -> Self

Set the minimum cosine similarity required to include a result.

Results with similarity below this threshold are excluded.

Example

let retriever = EmbeddingRetriever::new(Dummy).with_similarity_threshold(0.7);

pub async fn add_documents(&mut self, documents: Vec<Document>) -> Result<()>

Embed and store documents in the in-memory vector store.

Calls embed() exactly once with all document texts.

Errors

Returns an error if the embedding provider fails or returns the wrong number of embeddings.

pub fn len(&self) -> usize

Number of stored documents.

pub fn is_empty(&self) -> bool

Whether the vector store is empty.

Trait Implementations

impl<P: EmbeddingProvider> Retriever for EmbeddingRetriever<P>

fn retrieve<'life0, 'life1, 'async_trait>( &'life0 self, query: &'life1 str, limit: usize, ) -> Pin<Box<dyn Future<Output = Result<Vec<Document>>> + Send + 'async_trait>>

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

Embed query, compute cosine similarity with all stored docs, return top-k.