Struct OpenSearchVectorStoreBuilder 

pub struct OpenSearchVectorStoreBuilder { /* private fields */ }

Builds an OpenSearchVectorStore with optional auto-index-creation.

Example

use daimon_plugin_opensearch::{OpenSearchVectorStoreBuilder, SpaceType, Engine};

let store = OpenSearchVectorStoreBuilder::new("http://localhost:9200", 1536)
    .index("embeddings")
    .space_type(SpaceType::CosineSimilarity)
    .engine(Engine::Lucene)
    .hnsw_m(16)
    .hnsw_ef_construction(256)
    .auto_create_index(true)
    .build()
    .await?;

Implementations

impl OpenSearchVectorStoreBuilder

pub fn new(url: impl Into<String>, dimensions: usize) -> Self

Creates a new builder.

• url: OpenSearch cluster URL (e.g. "http://localhost:9200")
• dimensions: the fixed vector dimension count (must match your embedding model)

pub fn index(self, index: impl Into<String>) -> Self

Sets the index name. Default: "daimon_vectors".

pub fn space_type(self, space_type: SpaceType) -> Self

Sets the k-NN space type (distance metric). Default: SpaceType::CosineSimilarity.

pub fn engine(self, engine: Engine) -> Self

Sets the k-NN engine. Default: Engine::Lucene.

pub fn auto_create_index(self, enabled: bool) -> Self

Enables or disables automatic index creation on first use. Default: true.

When disabled, use the JSON from crate::index_settings to create the index manually.

pub fn hnsw_m(self, m: usize) -> Self

Sets the HNSW m parameter (max connections per layer). None uses the engine default.

pub fn hnsw_ef_construction(self, ef: usize) -> Self

Sets the HNSW ef_construction parameter (build-time search width). None uses the engine default.

pub async fn build_with_client( self, client: OpenSearch, ) -> Result<OpenSearchVectorStore>

Builds an OpenSearchVectorStore from a pre-existing OpenSearch client.

Use this when you need custom transport configuration (e.g. AWS SigV4, custom certificates, connection pool tuning).

pub async fn build(self) -> Result<OpenSearchVectorStore>

Builds the OpenSearchVectorStore, optionally creating the index.