interface VectorStoreWeaviateNodeParameters {
embeddingBatchSize?: number;
id?: string;
includeDocumentMetadata?: boolean;
mode?:
| "load"
| "insert"
| "retrieve"
| "retrieve-as-tool";
options?: | {
clearStore?: boolean;
proxy_grpc?: string;
skip_init_checks?: boolean;
tenant?: string;
textKey?: string;
timeout_init?: number;
timeout_insert?: number;
timeout_query?: number;
}
| {
alpha?: number;
autoCutLimit?: number;
fusionType?: "RelativeScore"
| "Ranked";
hybridExplainScore?: boolean;
hybridQuery?: string;
maxVectorDistance?: number;
metadataKeys?: string;
proxy_grpc?: string;
queryProperties?: string;
searchFilterJson?: string;
skip_init_checks?: boolean;
tenant?: string;
textKey?: string;
timeout_init?: number;
timeout_insert?: number;
timeout_query?: number;
};
prompt?: string;
toolDescription?: string;
toolName?: string;
topK?: number;
useReranker?: boolean;
weaviateCollection?: {
mode: "id"
| "list";
value: string;
};
}Properties§
§§§§§§§§§§
readonly embedding Batch Size?: number§
readonly id?: stringID of an embedding entry
readonly include Document Metadata?: booleanWhether or not to include document metadata Default: true
readonly mode?: "load" | "insert" | "retrieve" | "retrieve-as-tool"Default: "retrieve"
readonly options?:
| {
clearStore?: boolean;
proxy_grpc?: string;
skip_init_checks?: boolean;
tenant?: string;
textKey?: string;
timeout_init?: number;
timeout_insert?: number;
timeout_query?: number;
}
| {
alpha?: number;
autoCutLimit?: number;
fusionType?: "RelativeScore"
| "Ranked";
hybridExplainScore?: boolean;
hybridQuery?: string;
maxVectorDistance?: number;
metadataKeys?: string;
proxy_grpc?: string;
queryProperties?: string;
searchFilterJson?: string;
skip_init_checks?: boolean;
tenant?: string;
textKey?: string;
timeout_init?: number;
timeout_insert?: number;
timeout_query?: number;
}Default: {}
readonly prompt?: stringSearch prompt to retrieve matching documents from the vector store using similarity-based ranking
readonly tool Description?: stringExplain to the LLM what this tool does, a good, specific description would allow LLMs to produce expected results much more often Type options: {"rows":2}
readonly tool Name?: stringName of the vector store
readonly top K?: numberNumber of top results to fetch from vector store Default: 4
readonly use Reranker?: booleanWhether or not to rerank results
readonly weaviate Collection?: { ... }Default: {"mode":"list","value":""}
Number of documents to embed in a single batch Default: 200