Deface2 Development Setup

Overview

This guide explains how to set up your development environment for Deface2.

Prerequisites

• Windows 10/11 (primary development platform)
• NVIDIA GPU (for GPU acceleration testing)
• 8GB+ RAM (recommended for ML processing)
• 20GB+ free disk space

---

1. Install Core Tools

1.1 Python 3.11+

# Download from https://www.python.org/downloads/
# OR using winget (Windows 10+)
winget install Python.Python.3.11

# Verify installation
python --version

1.2 Node.js 20+ (LTS)

# Using winget
winget install OpenJS.NodeJS.LTS

# OR download from https://nodejs.org/

1.3 Rust Toolchain

# Download and run rustup-init.exe from https://rustup.rs/
# OR using winget
winget install Rustlang.Rustup

# Verify installation
rustc --version
cargo --version

1.4 Git

winget install Git.Git

---

2. Install Python Dependencies

Create a virtual environment and install ML dependencies:

# Create virtual environment
python -m venv venv
.\venv\Scripts\activate

# Install core dependencies
pip install --upgrade pip
pip install numpy opencv-python onnxruntime onnx

# Install face detection libraries (we'll test which works best)
pip install mediapipe torch torchvision --index-url https://download.pytorch.org/whl/cu118
pip install ultralytics  # YOLOv8

# Install ORB-HD deface as reference
pip install deface

# Verify installations
python -c "import cv2; print(f'OpenCV: {cv2.__version__}')"
python -c "import mediapipe; print('MediaPipe OK')"
python -c "import onnxruntime; print(f'ONNX Runtime: {onnxruntime.__version__}')"

Note: The torch installation with cu118 is for CUDA 11.8. If you have a different CUDA version, check https://pytorch.org/get-started/locally/ for the correct command.

---

3. Install Tauri CLI

# Install Rust dependencies for Tauri
cargo install tauri-cli

# OR install via cargo-binstall (faster)
cargo install cargo-binstall
cargo-binstall tauri-cli

# Verify
cargo tauri --version

---

4. Set Up Node.js Dependencies for Tauri

# Navigate to project root
cd C:\Users\user\Downloads\Deface2

# Install dependencies (package.json will be created)
npm install

---

5. Project Structure

Deface2/
├── src/                    # Source code
│   ├── core/              # Python ML core
│   │   ├── __init__.py
│   │   ├── detector.py   # Face detection interface
│   │   ├── engines/       # Detection engine implementations
│   │   │   ├── mediapipe_engine.py
│   │   │   ├── yolov8_engine.py
│   │   │   └── deface_engine.py
│   │   └── anonymizer.py  # Anonymization logic
│   │
│   ├── ui/                # Tauri frontend (Svelte)
│   │   ├── src/
│   │   │   ├── App.svelte
│   │   │   ├── lib/
│   │   │   └── main.js
│   │   ├── index.html
│   │   └── package.json
│   │
│   └── main.rs            # Tauri entry point
│
├── tests/                 # Tests
│   ├── unit/
│   └── integration/
│
├── scripts/              # Build and utility scripts
│   ├── build.bat
│   └── package.bat
│
├── resources/            # Icons, assets
│
├── .github/              # GitHub Actions CI/CD
│   └── workflows/
│
├── Cargo.toml            # Rust dependencies
├── package.json          # Node dependencies
├── pyproject.toml        # Python dependencies
├── tauri.conf.json       # Tauri configuration
└── README.md

---

6. Run Development Server

Start Python ML Service (for development)

# Terminal 1: Start Python backend
cd Deface2
.\venv\Scripts\activate
python -m src.core.server

# Or run the face detection prototype
python tests/prototype/face_detection_test.py

Start Tauri Dev Server

# Terminal 2: Start Tauri UI
cd Deface2
cargo tauri dev

This will:

1. Start the Tauri development window
2. Connect to the Python backend via IPC
3. Enable hot reload for UI changes

---

7. Build for Testing

Build Desktop App (Windows)

# Build production bundle
cargo tauri build

# Output will be in: src-tauri/target/release/bundle/

Build with Debug Info

cargo tauri build --debug

---

8. Testing GPU Acceleration

Verify CUDA/ONNX Runtime

.\venv\Scripts\activate
python -c "
import onnxruntime as ort
providers = ort.get_available_providers()
print(f'Available providers: {providers}')
print(f'Current provider: {ort.get_default_execution_provider()}')
"

Expected output should include CUDAExecutionProvider if your NVIDIA GPU is detected.

Test Face Detection Engines

python tests/prototype/compare_engines.py --input tests/fixtures/sample.jpg

This will benchmark MediaPipe, YOLOv8, and ORB-HD/deface on the same image.

---

9. Troubleshooting

Python virtual environment not activating

# PowerShell execution policy may need adjustment
Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser

CUDA not detected

1. Verify NVIDIA driver: nvidia-smi
2. Check CUDA toolkit: nvcc --version
3. Reinstall PyTorch with correct CUDA version from pytorch.org

Tauri build fails

# Update Rust and Tauri
rustup update stable
cargo install tauri-cli --force

# Clean build
cargo clean
cargo tauri build

---

10. Useful Commands Summary

Task	Command
Activate venv	.\venv\Scripts\activate
Install Python deps	pip install -r requirements.txt
Run Tauri dev	cargo tauri dev
Build production	cargo tauri build
Run tests	pytest tests/
Lint Python	ruff check src/
Format Python	black src/

---

Next Steps

1. Complete environment setup (verify all tools work)
2. Run python tests/prototype/face_detection_test.py to test ML setup
3. Run cargo tauri dev to launch the app
4. Begin Phase 1.2: ML Core Prototype development