API Reference

This page provides detailed documentation for the Ragify.js API.

Core Classes

Ragify

The main class for interacting with the RAG system.


class Ragify {
  constructor(config: RagifyConfig);
  initialize(): Promise<void>;
  ingest(doc: Document | string, options?: IngestOptions): Promise<void>;
  query(query: string, options?: QueryOptions): Promise<QueryResult[]>;
}

Constructor


interface RagifyConfig {
  embeddingProvider: EmbeddingProvider;
  collectionName?: string;
  qdrantUrl?: string;
  qdrantApiKey?: string;
  chunkingConfig?: ChunkingConfig;
}

Methods

`initialize()`

Initializes the vector store collection.


await ragify.initialize();

`ingest()`

Ingests a document into the vector store.


interface Document {
  text: string;
  metadata?: Record<string, unknown>;
}
 
interface IngestOptions {
  chunkSize?: number;
  chunkOverlap?: number;
}
 
await ragify.ingest({
  text: "Your document text...",
  metadata: { source: "example" }
});

`query()`

Queries the vector store for similar documents.


interface QueryOptions {
  topK?: number;
  threshold?: number;
  metadata?: Record<string, unknown>;
}
 
interface QueryResult {
  text: string;
  metadata?: Record<string, unknown>;
  score: number;
}
 
const results = await ragify.query("Your query here", {
  topK: 5,
  threshold: 0.7
});

Embedding Providers

OpenAIEmbeddings


interface OpenAIEmbeddingsConfig {
  apiKey: string;
  model?: "text-embedding-3-small" | "text-embedding-3-large" | "text-embedding-ada-002";
  batchSize?: number;
}
 
class OpenAIEmbeddings implements EmbeddingProvider {
  constructor(config: OpenAIEmbeddingsConfig);
  generateEmbeddings(texts: string[]): Promise<number[][]>;
  getDimension(): Promise<number>;
}

Vector Store

QdrantVectorStore


interface QdrantConfig {
  url?: string;
  apiKey?: string;
  collectionName?: string;
  distance?: "Cosine" | "Euclid" | "Dot" | "Manhattan";
}
 
class QdrantVectorStore implements VectorStore {
  constructor(config: QdrantConfig);
  initialize(): Promise<void>;
  upsert(vectors: Vector[]): Promise<void>;
  search(query: number[], options: SearchOptions): Promise<SearchResult[]>;
}

Types

Document


interface Document {
  text: string;
  metadata?: Record<string, unknown>;
}

Chunk


interface Chunk {
  text: string;
  metadata?: Record<string, unknown>;
  embedding?: number[];
}

Vector


interface Vector {
  id: string;
  vector: number[];
  payload?: Record<string, unknown>;
}

SearchOptions


interface SearchOptions {
  topK?: number;
  threshold?: number;
  filter?: Record<string, unknown>;
}

SearchResult


interface SearchResult {
  id: string;
  score: number;
  payload?: Record<string, unknown>;
}

Error Handling

Ragify.js uses custom error classes for better error handling:


class RagifyError extends Error {
  constructor(message: string);
}
 
class EmbeddingError extends RagifyError {
  constructor(message: string);
}
 
class VectorStoreError extends RagifyError {
  constructor(message: string);
}

Events

The Ragify class emits events for various operations:


interface RagifyEvents {
  "initialize": () => void;
  "ingest": (doc: Document) => void;
  "query": (query: string, results: QueryResult[]) => void;
  "error": (error: Error) => void;
}
 
ragify.on("ingest", (doc) => {
  console.log("Document ingested:", doc);
});