Agentic RAG Examples

Overview

This folder contains examples for building a Retrieval-Augmented Generation (RAG) agent using the LangChain library.

Features

• Construction of a RAG agent workflow using LangChain
• Integration with OpenAI models for language generation and retrieval
• Example usage of tools such as web search and response analysis, create rag response
• Auto-instrumentation with OpenInference decorators to fully instrument the agent
• End-to-end tracing with Phoenix to track agent performance

Requirements

• LangChain library
• OpenAI API key
• Langgraph
• Python 3.x
• Gradio (for UI)

Installation

1. Install the required libraries by running pip install -r requirements.txt
2. Run app.py and input the required Keys(OpenAI, Phoenix API Key)

Usage

1. Run the app.py script to start the RAG agent
2. Interact with the agent by providing input and receiving responses

Files

• app.py: The main script for starting the application, this will run the web server with default port(7860)
• agent.py: The main script for the RAG agent
• tools.py: Contains tools for web search and response analysis, create rag response
• rag.py: Contains functions for initializing and using the RAG vector store
• requirements.txt: Lists the required libraries for the project

Notes

• All the Key's must be inputted from the UI application.
• RAG will be loaded with default url in the UI, You can update the url and initialize the project with your own data source.
• This application will support the HTML based sources.