LangChain Documentation RAG

A simple RAG (Retrieval Augmented Generation) system for querying LangChain documentation using LangGraph.

Overview

This project provides a simple way to query LangChain documentation using a retrieval-based system. It uses:

• LangGraph for orchestrating the retrieval and response generation
• Vector database for storing and retrieving documentation content
• LLMs for generating natural language responses

Features

• Document indexing for LangChain documentation
• Natural language querying of documentation content
• Contextual responses based on retrieved documentation

Getting Started

1. Copy .env.example to .env

cp .env.example .env 

2. Add your API keys to .env:

OPENAI_API_KEY=<your-key>
ELASTICSEARCH_URL=<your-url>
ELASTICSEARCH_API_KEY=<your-key>

3. Index the documentation:

python index.py

4. Start querying the documentation:

python query.py "How do I use LangChain agents?"

Customize

You can customize the:

• Vector store (Elasticsearch, MongoDB, Pinecone)
• Embedding model
• Language model for responses
• System prompts and retrieval parameters

Check the configuration files for available options.

Development

See the LangGraph documentation for more details on extending functionality.