DriverTrac/PROJECT_STRUCTURE.md

📁 Project Structure

Clean Organization

Driver_DSMS_ADAS/
│
├── 📂 src/                          # Source code
│   ├── poc_demo.py                 # ⭐ Main POC application (World-Class Demo)
│   └── check_dependencies.py       # Dependency checker
│
├── 📂 config/                       # Configuration files
│   └── poc_config.yaml             # POC configuration
│
├── 📂 models/                       # ML Models
│   ├── yolov8n.pt                  # YOLO PyTorch model
│   └── yolov8n.onnx                 # YOLO ONNX model (optimized)
│
├── 📂 logs/                         # Application logs
│   ├── poc_demo.log                # POC demo logs
│   └── predictions.log             # Prediction logs
│
├── 📂 docs/                         # Documentation
│   ├── ASSESSMENT_REPORT.md         # Comprehensive assessment
│   ├── BUG_FIX_SUMMARY.md           # Bug fixes documentation
│   ├── QUICK_START.md               # Quick start guide
│   ├── RASPBERRY_PI_GUIDE.md       # Raspberry Pi deployment
│   └── README.md                    # Original README
│
├── 📂 assets/                       # Media assets (images, videos)
│
├── 📂 venv/                         # Virtual environment (gitignored)
│
├── 📄 requirements.txt              # Python dependencies
├── 📄 run_poc.sh                    # Quick start script
├── 📄 POC_README.md                 # POC documentation
├── 📄 PROJECT_STRUCTURE.md          # This file
│
├── 📄 track_drive.py                # Original full version (legacy)
└── 📄 track_drive copy.py           # Backup copy

File Purposes

🎯 POC Demo (src/poc_demo.py)

Purpose: World-class, production-ready demo with only reliable features

Features:

• Drowsiness (PERCLOS)
• Distraction (Head Pose)
• Driver Absent
• Phone/Vehicle/Pedestrian Detection

Optimized for: Raspberry Pi, low-spec CPUs

⚙️ Configuration (config/poc_config.yaml)

Purpose: Centralized configuration for POC

Contains:

• Model paths
• Thresholds
• Performance settings
• Feature toggles

📊 Models (models/)

Purpose: Store ML models

Files:

• yolov8n.pt - PyTorch model (auto-downloaded)
• yolov8n.onnx - ONNX optimized model (auto-exported)

📝 Logs (logs/)

Purpose: Application logging

Files:

• poc_demo.log - POC application logs
• predictions.log - Prediction history

📚 Documentation (docs/)

Purpose: All project documentation

Files:

• Assessment reports
• Guides
• README files

Quick Navigation

To Run POC:

./run_poc.sh
# or
streamlit run src/poc_demo.py

To Configure:

Edit config/poc_config.yaml

To Check Logs:

tail -f logs/poc_demo.log

To View Documentation:

• POC: POC_README.md
• Full Guide: docs/README.md
• Raspberry Pi: docs/RASPBERRY_PI_GUIDE.md

Best Practices

1. Keep src/ clean - Only production code
2. Config in config/ - Never hardcode settings
3. Models in models/ - Centralized model storage
4. Logs in logs/ - All logging centralized
5. Docs in docs/ - All documentation organized

Git Ignore Recommendations

Add to .gitignore:

venv/
__pycache__/
*.pyc
*.log
models/*.pt
models/*.onnx
*.pt
*.onnx
.DS_Store

Migration from Old Structure

Old files are preserved:

• track_drive.py - Original full version
• track_drive copy.py - Backup

New structure is in:

• src/poc_demo.py - Streamlined POC

---

Clean. Organized. Professional. 🚀