📗 Table of Contents

• 📖 About the Project
  • 🛠 Built With
    • Tech Stack
    • Key Features
  • 🚀 Live Demo
• 💻 Getting Started
  • Prerequisites
  • Setup
  • Install
  • Configuration
  • Usage
  • API Endpoints
• 👥 Authors
• 🔭 Future Features
• 🤝 Contributing
• ⭐️ Show your support
• 📝 License

📖 IT Support RAG Assistant

IT Support RAG Assistant is a powerful API designed to automate IT support queries using Retrieval-Augmented Generation (RAG). It combines the power of Large Language Models (Google Gemini) with a local knowledge base (ChromaDB) to provide accurate, context-aware answers to user questions.

The system features robust authentication, query history tracking, automatic query clustering using machine learning, and comprehensive experiment tracking with MLflow.

🛠 Built With

Tech Stack

• FastAPI - Modern, high-performance web framework for building APIs.
• PostgreSQL - Advanced open-source relational database.
• SQLAlchemy - The Python SQL Toolkit and Object Relational Mapper.
• LangChain - Framework for developing applications powered by LLMs.
• ChromaDB - AI-native open-source vector database.
• MLflow - Open source platform for the machine learning lifecycle.
• Google Gemini - Generative AI models.
• Hugging Face - Platform for ML models (Embeddings).

Key Features

• 🔐 Secure Authentication: User registration and login with JWT-based authentication stored in HTTP-only cookies.
• 🤖 RAG-Powered QA: Retrieval-Augmented Generation pipeline to answer support queries using your knowledge base.
• 📦 Query Clustering: Automatic clustering of user queries to categorize support topics using a trained ML model.
• 📊 Experiment Tracking: Full integration with MLflow to track RAG metrics (latency, number of chunks) and query results.
• 📜 History Management: Retrieve past user queries and generated answers.
• 🐳 Dockerized: Container-ready application with Dockerfile included.

(back to top)

🚀 Live Demo

• [Live Demo Link](link to deployed project)

(back to top)

💻 Getting Started

To get a local copy up and running, follow these steps.

Prerequisites

• uv (Recommended for managing dependencies and environments)
• Python 3.12+ (managed by uv)
• PostgreSQL
• Docker (optional)

Setup

Clone this repository to your desired folder:

git clone https://github.com/codehass/it-support-rag-assistant.git
cd it-support-rag-assistant

Install

Install dependencies and create the virtual environment automatically:

uv sync

Configuration

Create a .env file in the root directory and add your environment variables. You can copy .env.example as a template:

cp .env.example .env

Required .env Variables:

USER_DB=postgres
PASSWORD=your_password
DATABASE_HOST=localhost
PORT=5432
DATABASE=your_database_name

# HuggingFace & Google API
HF_TOKEN=your_huggingface_token_here
GOOGLE_API_KEY=your_google_api_key_here

# Authentication
SECRET_KEY=your_secure_secret
ALGORITHM=HS256
ACCESS_TOKEN_EXPIRE_MINUTES=30

FRONTEND_URL=http://localhost:3000

Usage

1. Start the Database: Ensure your PostgreSQL service is running and the database is created.

2. Run the API:

   uv run fastapi dev app/main.py

   The API will be available at http://localhost:8000.

3. Explore Documentation: Go to http://localhost:8000/docs for the interactive Swagger UI.

4. Start MLflow UI (Optional, for tracking):

   uv run mlflow ui

   Access MLflow dashboard at http://localhost:5000.

Testing & Linting

1. Run Linting (Ruff):

   uv run ruff check .

2. Run Tests (Pytest):

   uv run pytest

API Endpoints

Authentication

• POST /api/v1/auth/register - Register a new user
• POST /api/v1/auth/login - Login to get access token (cookie)
• POST /api/v1/auth/logout - Logout user
• GET /api/v1/auth/users/me - Get current user info

RAG Support

• POST /api/v1/rag/query - Ask a question to the IT Support Assistant
• GET /api/v1/rag/history - Get your query history
• GET /api/v1/rag/health - Check backend health status

(back to top)

👥 Authors

👤 Hassan El Ouardy

• GitHub: @codehass
• Twitter: @hassanelourdy
• LinkedIn: @hassanelourdy

(back to top)

🔭 Future Features

• Voice Interface: Enable voice-to-text for querying.
• Admin Dashboard: Analytics view for IT support managers.
• Feedback Loop: User feedback mechanism to improve RAG accuracy.

(back to top)

🤝 Contributing

Contributions, issues, and feature requests are welcome!

Feel free to check the issues page.

(back to top)

⭐️ Show your support

If this project helps you, give it a ⭐️!

(back to top)

📝 License

This project is MIT licensed.

(back to top)