Interface GraphRAGConfig

interface GraphRAGConfig {
    engineId: string;
    entityTypes?: string[];
    maxCommunityLevels?: number;
    minCommunitySize?: number;
    communityResolution?: number;
    generateEntityEmbeddings?: boolean;
    embeddingModelId?: string;
    embeddingDimension?: number;
    maxSummaryTokens?: number;
    vectorStoreProviderId?: string;
    entityCollectionName?: string;
    communityCollectionName?: string;
    tablePrefix?: string;
}

Properties

engineId: string

Unique ID for this GraphRAG engine instance

entityTypes?: string[]

Entity types to extract (e.g., ['person', 'organization', 'concept'])

maxCommunityLevels?: number

Maximum community hierarchy depth

minCommunitySize?: number

Minimum community size (entities) before splitting stops

communityResolution?: number

Louvain resolution parameter (higher = more communities)

generateEntityEmbeddings?: boolean

Whether to generate embeddings for entities

embeddingModelId?: string

Embedding model ID to use

embeddingDimension?: number

Embedding dimension for the selected embedding model.

Optional: when omitted and an embeddingManager is available, the engine will probe the embedding dimension at runtime by generating a tiny embedding once.

maxSummaryTokens?: number

Maximum tokens for community summaries

vectorStoreProviderId?: string

Vector store provider ID for entity embeddings

entityCollectionName?: string

Collection name for entity embeddings

communityCollectionName?: string

Collection name for community summary embeddings

tablePrefix?: string

SQL table prefix for persistence