Struct OpenSearchVectorStore 

pub struct OpenSearchVectorStore { /* private fields */ }

A VectorStore backed by OpenSearch with the k-NN plugin.

Use OpenSearchVectorStoreBuilder to construct.

Implementations

impl OpenSearchVectorStore

pub fn client(&self) -> &OpenSearch

Returns a reference to the underlying OpenSearch client.

pub fn index(&self) -> &str

Returns the index name used by this store.

pub fn dimensions(&self) -> usize

Returns the configured vector dimensions.

pub fn space_type(&self) -> SpaceType

Returns the configured k-NN space type (distance metric).

This determines how OpenSearch scores query hits; see query for why those scores are only comparable within a single space type.

Trait Implementations

impl VectorStore for OpenSearchVectorStore

async fn upsert( &self, id: &str, embedding: Vec<f32>, document: Document, ) -> Result<()>

Inserts or updates a document with a pre-computed embedding vector.

async fn upsert_many( &self, items: Vec<(String, Vec<f32>, Document)>, ) -> Result<()>

Inserts or updates a batch of documents with pre-computed embeddings.

async fn query( &self, embedding: Vec<f32>, top_k: usize, ) -> Result<Vec<ScoredDocument>>

Queries the store for the top_k most similar documents to the given embedding vector. Returns results sorted by descending similarity.

async fn delete(&self, id: &str) -> Result<bool>

Deletes a document by ID. Returns true if the document was found and deleted.

async fn count(&self) -> Result<usize>

Returns the number of documents in the store.