Interface: HnswlibVectorStoreConfig
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:33
Configuration for HnswlibVectorStore
Extends
Properties
customProps?
optionalcustomProps:Record<string,any>
Defined in: packages/agentos/src/rag/IVectorStore.ts:32
Any other custom properties or configurations specific to this provider instance not covered by standard fields.
Inherited from
VectorStoreProviderConfig.customProps
defaultEmbeddingDimension?
optionaldefaultEmbeddingDimension:number
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:38
Default embedding dimension for new collections
hnswEfConstruction?
optionalhnswEfConstruction:number
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:44
HNSW efConstruction parameter (index build quality, default 200)
hnswEfSearch?
optionalhnswEfSearch:number
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:46
HNSW efSearch parameter (search quality, default 100)
hnswM?
optionalhnswM:number
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:42
HNSW M parameter (number of connections per node, default 16)
id
id:
string
Defined in: packages/agentos/src/rag/IVectorStore.ts:30
A unique identifier for this specific provider instance (e.g., "pinecone-main-prod", "weaviate-dev-local"). This ID is used by the VectorStoreManager to retrieve this provider.
Inherited from
persistDirectory?
optionalpersistDirectory:string
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:36
Directory to persist index files. If not set, indexes are in-memory only.
similarityMetric?
optionalsimilarityMetric:"cosine"|"euclidean"|"dotproduct"
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:40
Default similarity metric
type
type:
"hnswlib"
Defined in: packages/agentos/src/rag/implementations/vector_stores/HnswlibVectorStore.ts:34
The type of the vector store provider (e.g., "pinecone", "weaviate", "in_memory", "lancedb"). This helps in selecting the correct implementation.