RAG Applications with LangChain: Building AI Search

Retrieval Augmented Generation (RAG) combines document retrieval with LLM generation.

Setup

npm install langchain @langchain/openai @langchain/community

Document Loading and Splitting

import { TextLoader } from 'langchain/document_loaders/fs/text';
import { RecursiveCharacterTextSplitter } from 'langchain/text_splitter';

const loader = new TextLoader('documents/data.txt');
const docs = await loader.load();

const splitter = new RecursiveCharacterTextSplitter({
  chunkSize: 1000,
  chunkOverlap: 200,
});

const splitDocs = await splitter.splitDocuments(docs);

Vector Store

import { OpenAIEmbeddings } from '@langchain/openai';
import { MemoryVectorStore } from 'langchain/vectorstores/memory';

const vectorStore = await MemoryVectorStore.fromDocuments(
  splitDocs,
  new OpenAIEmbeddings()
);

Retrieval Chain

import { createRetrievalChain } from 'langchain/chains/retrieval';
import { createStuffDocumentsChain } from 'langchain/chains/combine_documents';
import { ChatOpenAI } from '@langchain/openai';
import { PromptTemplate } from '@langchain/core/prompts';

const llm = new ChatOpenAI({ model: 'gpt-4' });

const prompt = PromptTemplate.fromTemplate(`
Answer based on context:
{context}

Question: {input}
`);

const combineDocsChain = await createStuffDocumentsChain({
  llm,
  prompt,
});

const retriever = vectorStore.asRetriever();

const retrievalChain = await createRetrievalChain({
  retriever,
  combineDocsChain,
});

Query

const response = await retrievalChain.invoke({
  input: 'What is the main topic?',
});

console.log(response.answer);

Conclusion

RAG enables building intelligent Q&A systems that can search and reason over your documents.