ultimate-mcp-server/dist/rag/types.d.ts

The definitive all-in-one Model Context Protocol server for AI-assisted coding across 30+ platforms

export interface Document {
    id: string;
    content: string;
    metadata: {
        source: string;
        title?: string;
        type?: string;
        timestamp?: Date;
        tags?: string[];
        [key: string]: any;
    };
    embedding?: number[];
    chunks?: DocumentChunk[];
}
export interface DocumentChunk {
    id: string;
    documentId: string;
    content: string;
    startIndex: number;
    endIndex: number;
    embedding?: number[];
    metadata?: Record<string, any>;
}
export interface VectorStore {
    name: string;
    dimension: number;
    addDocuments(documents: Document[]): Promise<void>;
    addChunks(chunks: DocumentChunk[]): Promise<void>;
    search(query: string | number[], options?: SearchOptions): Promise<SearchResult[]>;
    delete(ids: string[]): Promise<void>;
    clear(): Promise<void>;
}
export interface SearchOptions {
    topK?: number;
    threshold?: number;
    filter?: Record<string, any>;
    includeMetadata?: boolean;
    includeEmbeddings?: boolean;
}
export interface SearchResult {
    id: string;
    score: number;
    content: string;
    metadata?: Record<string, any>;
    embedding?: number[];
}
export interface EmbeddingProvider {
    name: string;
    model: string;
    dimension: number;
    embed(text: string): Promise<number[]>;
    embedBatch(texts: string[]): Promise<number[][]>;
}
export interface ChunkingStrategy {
    type: 'fixed' | 'sentence' | 'paragraph' | 'semantic';
    chunkSize?: number;
    chunkOverlap?: number;
    separators?: string[];
}
export interface RAGConfig {
    vectorStore: VectorStoreConfig;
    embedding: EmbeddingConfig;
    chunking: ChunkingStrategy;
    retrieval: RetrievalConfig;
    generation?: GenerationConfig;
}
export interface VectorStoreConfig {
    type: 'memory' | 'postgres' | 'pinecone' | 'weaviate' | 'chroma';
    connectionString?: string;
    apiKey?: string;
    index?: string;
    namespace?: string;
}
export interface EmbeddingConfig {
    provider: 'openai' | 'cohere' | 'local' | 'custom';
    model: string;
    apiKey?: string;
    dimension?: number;
}
export interface RetrievalConfig {
    topK: number;
    threshold: number;
    reranking?: boolean;
    hybridSearch?: boolean;
}
export interface GenerationConfig {
    provider: string;
    model: string;
    temperature?: number;
    maxTokens?: number;
    systemPrompt?: string;
}
//# sourceMappingURL=types.d.ts.map