kernelbits/TnLicencePlateReader

A tunisian Licence Plate reader # Tunisian License Plate Reader A full-stack application for detecting and reading Tunisian license plates using Roboflow for detection, PaddleOCR for character recognition, and Supabase for driver information lookup. ## Project Structure - `backend/`: FastAPI server for plate detection and OCR. - `frontend/`: Gradio-based user interface. - `data/`: Contains training images, test images, and database CSVs. - `plateReader.py`: Script for standalone detection testing. - `upload_data.py`: Script to upload images and annotations to Roboflow. - `requirements.txt`: Project dependencies. ## Setup 1. Install dependencies: ```bash pip install -r requirements.txt pip install -r frontend/requirements.txt ``` 2. Configure `.env` file with your API keys: - `RF_API_KEY`: Roboflow API key - `SUPABASE_URL`: Supabase project URL - `SUPABASE_ANON_KEY`: Supabase anonymous key - `OCR_API_URL`: PaddleOCR endpoint URL - `OCR_TOKEN`: PaddleOCR access token 3. (Optional) Install and start **Ollama** for the AI Assistant: - Download Ollama from ollama.com. - Pull the required model: `ollama pull llama3.2:3b`. - Ensure Ollama is running in the background. ## Usage 1. Start the backend: ```bash python -m uvicorn backend.backend:app --reload --host 0.0.0.0 --port 8000 ``` 2. Start the frontend: ```bash BACKEND_URL=127.0.0.1 python frontend/main.py ``` ## Docker (Local) 1. Ensure your `.env` is present at the project root (contains RF_API_KEY, SUPABASE_URL, SUPABASE_ANON_KEY, OCR_API_URL, OCR_TOKEN, SUPABASE_SERVICE_ROLE_KEY, etc.). 2. Build and run with Docker Compose: ```bash docker compose up --build ``` - Backend: localhost - Frontend (Gradio): localhost - Ollama is started as a service. Pull the model in a separate terminal once (first run only): ```bash docker exec -it _ollama_1 ollama pull llama3.2:3b ``` - Alternatively, comment out the `ollama` service in `docker-compose.yml` and set `OLLAMA_API_URL` to your external Ollama host. ## Feat …