RAG Support Client

⚠️ Proof of Concept (POC) - AI-Generated Application This application was entirely developed through prompting ChatGPT, Gemini, Claude and local Open Source LLM. Demonstrating that in 2024, anyone can create a local RAG AI application through simple conversations with AI, without coding expertise. The code was generated by AI, tested, and assembled through copy-paste. This project is provided "as-is" for educational and demonstration purposes only.

A powerful Retrieval-Augmented Generation (RAG) support client that provides context-aware responses from Markdown documentation. Built with LangChain, Ollama, and ChromaDB, featuring both a Streamlit UI and FastAPI backend.

Project Goals

• Demonstrate AI application development accessibility without coding expertise
• Show the feasibility of creating local RAG systems without cloud dependencies
• Illustrate how modern tools (LangChain, Ollama) simplify AI application development

Features

• 🔍 Smart document retrieval using RAG technology
• 🤖 Local LLM integration with Ollama
• 💾 Efficient vector storage with ChromaDB
• 🌐 Dual interface: Streamlit UI and FastAPI backend
• 📝 Markdown documentation processing
• 🔄 Real-time conversation context management
• 🎯 High-precision response generation
• 📊 Administrative interface for document management

Architecture

The application follows a modular architecture with these key components:

• RAG Chain: Central component handling document retrieval and response generation
• Document Processing: Intelligent Markdown processing with metadata extraction
• Vector Store: ChromaDB-based similarity search
• LLM Integration: Local deployment using Ollama
• Dual Interface: FastAPI backend for remote API client as a ChatBot and a Streamlit UI for local testing

Deployment Options

Prerequisites

For Production Deployment

• Docker and Docker Compose
• 8GB RAM minimum (16GB recommended)
• 4 CPU cores minimum
• 20GB free disk space (more depending on your data volume)
• Linux server (recommended) or Windows/macOS with Docker support

For Development

• Python 3.11 or higher
• Git
• Visual Studio Code (recommended)
• Ollama installed locally
• ChromaDB prerequisites

Option 1: Docker Deployment (with Ollama)

This option deploys both the RAG Support Client and Ollama in Docker containers:

1. Clone the repository:

git clone https://github.com/espritdunet/rag-support-client.git
cd rag-support-client

2. Configure your environment:

   cp .env.example .env
   # Edit .env file with your settings

3. Start the services:

   docker compose up -d

4. Access the applications: FastAPI Documentation: http://your-server:8000/docs Streamlit UI: http://your-server:8501

Option 2: Docker with External Ollama

Use this option if you have Ollama running on a separate server:

1. Follow steps 1-2 from Option 1

2. Edit .env and set OLLAMA_BASE_URL to your Ollama server

3. Start the service:

   docker compose -f docker-compose.external-ollama.yml up -d

Option 3: Development Setup

1. Create virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # or `.venv\Scripts\activate` on Windows

2. Install dependencies:

   pip install -e '.[dev]'

3. Configure environment:

   cp .env.example .env
   # Edit .env with your settings

4. Run the applications:

   # Start FastAPI server
   python main.py
   
   # Start Streamlit UI (new terminal)
   streamlit run streamlit_app.py

Data Persistence

Docker Volumes

/var/lib/rag-support/data: Source documents and processed files /var/lib/rag-support/vector_store: ChromaDB vector database /var/log/rag-support: Application logs

Backup and Restore

```bash
# Backup
docker run --rm -v rag-support-vector-store:/source -v $(pwd)/backups:/backup \
ubuntu tar czf /backup/vector-store-$(date +%Y%m%d).tar.gz /source

# Restore
docker run --rm -v rag-support-vector-store:/target -v $(pwd)/backups:/backup \
ubuntu tar xzf /backup/vector-store-20240101.tar.gz -C /target --strip-components=1
```

Maintenance

Updating the Application

```bash
# Pull latest changes
    git pull

# Rebuild and restart
docker compose down
docker compose build --no-cache
docker compose up -d
```

Monitoring

```bash
# View logs
docker compose logs -f

# Check status
docker compose ps

# Monitor resources
docker stats
```

Troubleshooting

Common Issues

1. Ollama Connection Failed

   Check Ollama is running: curl http://ollama:11434/api/tags Verify network connectivity Check Ollama logs: docker compose logs ollama

2. Memory Issues

   Verify server meets minimum requirements Check Docker resource limits Monitor memory usage: docker stats

3. Data Persistence

   Check volume permissions Verify mount points Ensure sufficient disk space

Getting Help

Check the Issues page
Submit detailed bug reports with logs
Join our Discussions

Contributing

Contributions are welcome! Please read our Contributing Guidelines before submitting pull requests.

License

This project is licensed under the MIT License - see the LICENSE file for details.