Support RAG System

A powerful, modular RAG (Retrieval-Augmented Generation) system with semantic caching, hybrid search, and multi-hop reasoning.

📋 Table of Contents

• What is Support RAG?
• Key Features
• Quick Start
• Installation
• Usage
• Documentation
• Project Structure
• FAQ
• Contributing
• License
• Support

💡 What is Support RAG?

Support RAG is an advanced open-source RAG system designed for high-performance automated customer support and knowledge retrieval. It goes beyond simple semantic search by integrating semantic caching, hybrid search (lexical + vector), and a modular pipeline architecture with 29+ specialized nodes.

It is built to handle complex queries, manage conversation state, and provide accurate, context-aware responses with citations.

✨ Key Features

• 🚀 Modular Pipeline Architecture: 29+ specialized nodes for flexible query processing
• 🧠 Semantic Caching: Caches responses based on semantic similarity to reduce latency and costs
• 🔍 Hybrid Search: Combines Dense Retrieval (Vector) and Sparse Retrieval (BM25/Lexical) for best accuracy
• 🔄 Multi-hop Reasoning: Breaks down complex user queries into sub-questions
• 🛡️ Guardrails: Input and output validation to ensure safety and quality
• 💬 Conversation Management: Handles context and history for natural multi-turn dialogue
• 🔌 Webhooks & API: Real-time event notifications and comprehensive REST API

🚀 Quick Start

Get up and running in minutes using Docker Compose.

1. Clone the repository

   git clone https://github.com/letya999/support_rag.git
   cd support_rag

2. Configure Environment

   cp .env.example .env
   # Edit .env with your API keys (OpenAI, etc.)

3. Start Services

   docker-compose up -d

4. Ingest Data (Optional initial setup)

   docker-compose exec app python scripts/ingest.py --file datasets/qa_data.json

5. Test API

   curl -X POST http://localhost:8000/api/v1/chat/completions \
     -H "Content-Type: application/json" \
     -d '{"messages": [{"role": "user", "content": "How do I return an item?"}]}'

🛠️ Installation

Prerequisites

• Docker & Docker Compose
• Python 3.9+ (for local development)
• PostgreSQL
• Redis

Local Development Setup

For those who want to contribute or run without Docker:

1. Create a virtual environment:

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

2. Install dependencies:

   pip install -r requirements.txt

3. Run migrations and start the server:

   python scripts/run_migrations.py
   uvicorn app.main:app --reload

See DEVELOPMENT.md for detailed instructions.

📖 Usage

REST API

The system provides a comprehensive API for chat completions and RAG queries.

Example Request:

POST /api/v1/chat/completions
{
  "messages": [
    {"role": "user", "content": "Tell me about shipping policies."}
  ],
  "temperature": 0.7
}

Telegram Bot

A Telegram bot integration is included. Configure your TELEGRAM_BOT_TOKEN in .env and start the bot container.

📚 Documentation

• Development Guide: Setup, testing, and contribution guide.
• Architecture: Deep dive into the system design, nodes, and data flow.
• API Documentation: Detailed API reference.
• Database Schema: Database structure and tables.

📂 Project Structure

support_rag/
├── app/                 # Main application code
│   ├── api/             # API endpoints
│   ├── core/            # Config and core logic
│   ├── nodes/           # Pipeline nodes (processing logic)
│   └── services/        # Business logic services
├── datasets/            # Example datasets
├── docs/                # Documentation
├── scripts/             # Utility and setup scripts
├── tests/               # Test suite
├── docker-compose.yml   # Docker services config
└── requirements.txt     # Python dependencies

❓ FAQ

Q: Can I use my own LLM? A: Yes, the system handles OpenAI-compatible APIs. Configure the base URL and API key in .env.

Q: How does caching work? A: We use Redis for semantic caching. Similar queries are matched using vector embeddings to return cached responses instantly.

🤝 Contributing

We welcome contributions! Please see our CONTRIBUTING.md for details on how to submit pull requests, report issues, and setup your development environment.

📄 License

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

📞 Support

If you have any questions or run into issues, please open an issue on GitHub.