Agentic RAG Implementation with LangGraph 🦜🔍

Implementation of Adaptive RAG for visual impaired users learning LangGraph🦜🕸️.

This code was guided by Gemini-CLI.

This repository contains a refactored version of the original LangChain's Cookbook,

See Original YouTube video:Advance RAG control flow with Mistral and LangChain

of Sophia Young from Mistral & Lance Martin from LangChain

Features

An accessible RAG CLI tool to summarize or query files from (default) input directory and store results in (default) output directory.

% poetry run python ragagent2.py --help     
usage: ragagent2.py [-h] [--query [QUERY]] [--summarize] [--input INPUT] [--output OUTPUT] [--rich]

options:
  -h, --help       show this help message and exit
  --query [QUERY]  The question to ask the RAG agent.
  --summarize      Summarize all documents in the input directory.
  --input INPUT    Directory for source documents (default 'input').
  --output OUTPUT  Directory to save results (default 'output').
  --rich           Enable rich text output for non-screen reader users.


% poetry run python ragagent2.py --summarize
% poetry run python ragagent2.py --query 'Was is a multi-agent environment?' 

Both cases are using default 'input' and 'output' directory

Environment Variables

To run this project, need to add the following environment variables to your .env file:

OPENAI_API_KEY=your_openai_api_key_here
TAVILY_API_KEY=your_tavily_api_key_here  # For web search capabilities
LANGCHAIN_API_KEY=your_langchain_api_key_here  # Optional, for tracing
LANGCHAIN_TRACING_V2=true                      # Optional
LANGCHAIN_PROJECT=your-agentic-rag             # Optional

Note: For query tracing via a LangSmith, to inspect queries and their results between LLM calls, the setting LANGCHAIN_TRACING_V2=true must be defined. Also, a valid LangSmith API key must be defined in LANGCHAIN_API_KEY. Without a valid API key, the application will throw an error.

Getting Started

Clone the repository:

git clone https://github.com/kds215/RagAgent2.git
cd RagAgent2
git checkout project/RagAgent2

Install dependencies:

pip install -r requirements.txt
# or if using Poetry:
poetry install

Run:

% poetry run python ragagent2.py --help 

Acknowledgements

(1) this RagAgent2 code was GGBK published: “Gemini-Guided-By-Klaus”.

(2) Original LangChain repository: LangChain Cookbook By Sophia Young from Mistral & Lance Martin from LangChain