RAG AI assistant

This project provides an full stack application with a UI chatbot and RAG pipeline using google drive data to give contextual answers.

• Sample Deployment: http://20.117.106.138/

Screen.Recording.2025-06-29.at.3.mp4

Table of Contents

• Prerequisites
• Project Structure
• Backend Setup (Credentials & Environment Variables)
• How to Install Docker
• Running the Application with Docker Compose
• Troubleshooting

Prerequisites

• Docker Engine & Docker Compose: These will be installed via your Linux distribution's package manager.

🔧 Tech Stack Used

Layer	Technology	Why
Frontend	React (with Tailwind UI)	Responsive, easy integration with chatbot UI
Backend	FastAPI (Python)	Lightweight, async, perfect for ML workloads
Embedding Model	Google's models/embedding-001	Efficient and high-quality embeddings
LLM for RAG	Gemini via LangChain	Versatile, works well with RAG pipelines
Vector Store	FAISS	Fast similarity search for embeddings
File Parsing	unstructured, PyMuPDF, docx, pdfplumber	Handles various file formats
Google Drive Integration	Google Drive API + Google Auth	Authenticated, scoped access to user files
Orchestration	LangChain	Makes RAG pipeline setup easier
Deployment	Docker + Azure	Scalable, production-grade setup

Project Structure

.
├── backend/
│   ├── main.py
│   ├── requirements.txt
│   ├── credentials.json      # Google Service Account JSON (user added)
│   └── .env                  # GOOGLE_API_KEY (user added)
│   └── ...
├── ai-assistant-ui/
│   ├── package.json
│   ├── src/
│   ├── public/
│   ├── nginx.conf            # Nginx config for UI
│   └── ...
├── Dockerfile.backend        # Backend Dockerfile
├── Dockerfile.ai-assistant-ui  # UI Dockerfile
└── docker-compose.yml        # Docker Compose config

Backend Setup (Credentials & Environment Variables)

For security, credentials.json and .env are not committed to version control.

1. credentials.json:
   • Obtain: From Google Cloud Console > IAM & Admin > Service Accounts > Create new key (JSON).
   • Place: backend/credentials.json
2. .env file:
   • Create: backend/.env
   • Content: GOOGLE_API_KEY=YOUR_GOOGLE_API_KEY_HERE (Obtain from Google AI Studio).

How to Install Docker

Install Docker Engine and Docker Compose for your Linux distribution.

• For Debian/Ubuntu:
  sudo apt-get update
  sudo apt-get install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
  sudo usermod -aG docker $USER # Add user to docker group
  # Log out and log back in for changes to take effect

• For other Linux distributions: Refer to the official Docker documentation:

  • Install Docker Engine
  • Install Docker Compose

Verify installation and user permissions:

docker --version
docker compose version
docker run hello-world # Should run without permission errors

Running the Application with Docker Compose

1. Navigate: Go to the project root where docker-compose.yml is located.
   cd /path/to/your/project/root

2. Build & Run:
   docker compose up --build

   This builds images and starts both services. Logs will appear in the terminal.

3. Access:

   • AI Assistant UI: http://localhost:80
   • Backend: http://localhost:8000 (accessed by UI)

4. Stop:

   • Stop containers: docker compose stop
   • Stop & remove: docker compose down
   • Stop, remove containers & images: docker compose down --rmi all

Troubleshooting

• Container issues: Check logs with docker compose logs <service_name>.

• Permissions (Docker Daemon): If you get "permission denied", ensure your user is in the docker group and you've logged out/in.
  sudo usermod -aG docker $USER
  # Log out and back in

• AI Assistant UI: Verify npm run build completed, check browser console.

• Backend: Check main.py entry point, GOOGLE_API_KEY in .env.