Royal Enfield AI Diagnostic Assistant (Local)

What it does

Local RAG assistant for Royal Enfield motorcycles (v1):

• User chats with problem description
• Retrieval from local dataset stored in ChromaDB
• Gemini generates structured diagnosis/checks/solution strictly from retrieved context
• Chat history saved in SQLite

Requirements

• Python 3.11+
• Windows 11
• Environment variable: GEMINI_API_KEY

Setup

1. Create venv (project-local)

python -m venv venv
venv\Scripts\activate

2. Install dependencies

pip install -r requirements.txt

3. Set Gemini key

set GEMINI_API_KEY=your_key_here

Run Backend

uvicorn backend.main:app --reload --port 8000

Run Frontend (Streamlit)

streamlit run frontend/app.py --server.port 8501

How to Use

1. Start the backend server (see "Run Backend" above)
2. In a new terminal, start the Streamlit frontend
3. Open browser to http://localhost:8501
4. Describe your Royal Enfield motorcycle issue
5. Get AI-generated diagnosis and solutions based on your knowledge base

Project Structure

├── backend/
│   └── main.py          # FastAPI backend
├── frontend/
│   └── app.py           # Streamlit UI
├── knowledge_base/
│   └── royal_enfield_troubleshooting.csv  # RAG dataset
├── vector_db/           # ChromaDB persisted vectors
├── data/
│   └── royal_enfield.db # SQLite chat history
├── requirements.txt
└── README.md

Notes

• Dataset: knowledge_base/royal_enfield_troubleshooting.csv
• Vector DB: persisted under vector_db/
• Chat history: stored in data/royal_enfield.db