zrald1/dist/core/vector-store.d.ts

Advanced Graph RAG MCP Server with file location identification, graph processing, and result summarization capabilities

import { Node, Chunk, File } from '../types/graph.js';
export interface VectorSearchResult {
    id: string;
    score: number;
    node?: Node;
    chunk?: Chunk;
    file?: File;
}
export declare class VectorStore {
    private dimension;
    private maxElements;
    private nodeMap;
    private chunkMap;
    private fileMap;
    private embeddings;
    private currentIndex;
    constructor(dimension?: number, maxElements?: number);
    initialize(): Promise<void>;
    addNode(node: Node): Promise<void>;
    addChunk(chunk: Chunk): Promise<void>;
    addFile(file: File, embedding?: number[]): Promise<void>;
    searchNodes(queryEmbedding: number[], topK?: number, threshold?: number): Promise<VectorSearchResult[]>;
    searchByNodeTypes(queryEmbedding: number[], nodeTypes: string[], topK?: number, threshold?: number): Promise<VectorSearchResult[]>;
    searchChunks(queryEmbedding: number[], topK?: number, threshold?: number): Promise<VectorSearchResult[]>;
    searchFiles(queryEmbedding: number[], topK?: number, threshold?: number): Promise<VectorSearchResult[]>;
    getStats(): {
        totalNodes: number;
        totalChunks: number;
        totalFiles: number;
        totalEmbeddings: number;
        dimension: number;
        maxElements: number;
    };
    static cosineSimilarity(a: number[], b: number[]): number;
    static euclideanDistance(a: number[], b: number[]): number;
    static dotProduct(a: number[], b: number[]): number;
    static generateRandomEmbedding(dimension: number): number[];
    clear(): void;
    getAllNodes(): Node[];
    getAllChunks(): Chunk[];
    getAllFiles(): File[];
    removeNode(nodeId: string): boolean;
    removeChunk(chunkId: string): boolean;
    removeFile(fileId: string): boolean;
}
//# sourceMappingURL=vector-store.d.ts.map