README.md

---
title: REMB - Land Redistribution API
emoji: 🏘️
colorFrom: blue
colorTo: green
sdk: docker
app_file: algorithms/backend/main.py
app_port: 7860
pinned: false
---

# REMB - Industrial Estate Master Planning API

AI-powered industrial park land subdivision and redistribution API using advanced optimization algorithms.

## 🚀 Features

- **Multi-stage Optimization**: NSGA-II genetic algorithm + OR-Tools constraint programming
- **DXF Import/Export**: Upload site boundaries from CAD files
- **Automated Layout**: Grid optimization, block subdivision, and infrastructure planning
- **100% Compliant**: Follows Vietnamese industrial planning regulations
- **Export Results**: Download results as GeoJSON or DXF

## 📚 API Documentation

Once the Space is running, visit:
- **Interactive API Docs**: `/docs`
- **Alternative Docs**: `/redoc`
- **Health Check**: `/health`

## 🔧 API Endpoints

### Health Check
```bash
GET /health
```

### Full Optimization Pipeline
```bash
POST /api/optimize
```

Runs the complete 3-stage optimization:
1. **Grid Optimization** (NSGA-II) - Find optimal grid orientation and spacing
2. **Block Subdivision** (OR-Tools) - Subdivide blocks into individual lots
3. **Infrastructure Planning** - Generate technical networks and drainage

**Request Body**:
```json
{
  "config": {
    "spacing_min": 20.0,
    "spacing_max": 30.0,
    "angle_min": 0.0,
    "angle_max": 90.0,
    "min_lot_width": 20.0,
    "max_lot_width": 80.0,
    "target_lot_width": 40.0,
    "population_size": 50,
    "generations": 100
  },
  "land_plots": [{
    "type": "Polygon",
    "coordinates": [[[0, 0], [100, 0], [100, 100], [0, 100], [0, 0]]]
  }]
}
```

### DXF Upload
```bash
POST /api/upload-dxf
Content-Type: multipart/form-data

file: <DXF file>
```

Upload DXF file and extract boundary polygon.

## 💻 Usage Example

```python
import requests

# API endpoint
url = "https://cuong2004-remb.hf.space/api/optimize"

# Configuration
payload = {
    "config": {
        "spacing_min": 20.0,
        "spacing_max": 30.0,
        "population_size": 50,
        "generations": 100
    },
    "land_plots": [{
        "type": "Polygon",
        "coordinates": [[[0, 0], [100, 0], [100, 100], [0, 100], [0, 0]]]
    }]
}

# Run optimization
response = requests.post(url, json=payload)
result = response.json()

print(f"Total blocks: {result['statistics']['total_blocks']}")
print(f"Total lots: {result['statistics']['total_lots']}")
```

## 🎨 Frontend

For a complete user interface with visualization, use the Streamlit frontend:
[Deploy Frontend to Streamlit Cloud →](https://github.com/nxc1802/REMB/tree/main/algorithms/frontend)

## 🛠️ Technology Stack

- **FastAPI**: High-performance Python web framework
- **DEAP**: Genetic algorithms (NSGA-II)
- **OR-Tools**: Google's constraint programming solver
- **Shapely**: Geometric operations and spatial analysis
- **ezdxf**: DXF file parsing and generation
- **Docker**: Containerized deployment

## 📖 Documentation

- [Full Documentation](https://github.com/nxc1802/REMB)
- [API Reference](https://github.com/nxc1802/REMB/blob/main/algorithms/README.md)
- [Deployment Guide](https://github.com/nxc1802/REMB/blob/main/algorithms/DEPLOYMENT.md)

## 🏗️ Architecture

### Module A - "The Architect" (NSGA-II)
- Multi-objective genetic algorithm
- Generate thousands of feasible layouts
- Optimize: sellable area, green space, road access

### Module B - "The Engineer" (OR-Tools)
- Constraint programming solver
- Ensure mathematical validity
- Check non-overlapping, road connectivity

### Module C - "The Inspector" (Regulation Checker)
- Expert system for regulatory compliance
- 100% adherence to Vietnamese regulations
- Validate: setbacks, fire safety, FAR, green space

## 📊 Performance

- **Planning Time**: 3 weeks → 2 hours (99% faster)
- **Compliance**: 100% regulatory adherence
- **Optimization Quality**: Pareto-optimal solutions

## 📝 License

MIT License - see [LICENSE](LICENSE) for details

## 👥 Team

Made with ❤️ by **PiXerse.AI Team**

---

**Note**: This Space contains the optimization API backend. For the complete user interface, deploy the frontend separately to Streamlit Cloud.