Simple RAG

A simple RAG (Retrieval-Augmented Generation) application built with LangChain, ChromaDB, and Gradio.

Features

• Document ingestion from PDF and text files
• Webpage ingestion with URL support
• Efficient document chunking and embedding
• Interactive chat interface
• Persistent vector store
• Modular architecture with clear separation of concerns

Setup

1. Create a virtual environment:

python -m venv .venv
source .venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file:

cp .env.example .env

4. Run the application:

python main.py

Project Structure

simple-rag/
├── app/
│   ├── core/               # Core functionality (config, vectorstore)
│   ├── document/           # Document processing (loader, processor)
│   ├── rag/               # RAG implementation (chain)
│   └── ui/                # Gradio interface
├── data/
│   └── vectorstore/       # ChromaDB storage

Usage

1. Launch the application
2. Upload Documents: Select PDF, TXT, or MD files to process
3. Add Webpages: Enter URLs (one per line or comma-separated) to ingest web content
4. Start chatting with your documents and webpages

Architecture

The application follows a modular architecture:

• UI Layer (app/ui/): Gradio interface components
• Document Layer (app/document/): Document loading and processing logic
• RAG Layer (app/rag/): Retrieval-augmented generation implementation
• Core Layer (app/core/): Configuration and vector store management

Reference

To understand RAG in the bigger picture checkout my LinkedIn Post

License

MIT