NSW Development Compliance Engine

An enhanced development compliance checking system for NSW Inner West Council areas, featuring dual semantic processing with LangExtract and AutoSchemaKG for sophisticated regulatory rule understanding.

Features

Enhanced Semantic Processing

• Dual Semantic Pipeline: Combines LangExtract + AutoSchemaKG for complex rule understanding
• Conditional Logic Support: Handles rules like "0.9m minimum OR 0.5 times building height, whichever is greater"
• Source Text Grounding: Provides exact regulatory citations for legal defensibility
• Three-Tier Classification: Mandatory, Recommended, and Informational rule enforcement levels

API Integration

• Enhanced Compliance API: /api/compliance/check with semantic rule processing
• Setbacks API: /api/compliance/setbacks with fallback mechanisms
• Property Data API: NSW Planning Portal integration for official constraints
• Performance Caching: Intelligent caching for repeated queries

Frontend Interface

• Above-the-fold Design: Compact, professional interface
• Source Citation Display: Shows regulatory text with highlighting
• Processing Method Indicators: Visual badges for Semantic/Manual/Fallback
• Confidence Scoring: HIGH/MEDIUM/LOW reliability indicators
• Responsive Design: Works on mobile, tablet, and desktop

Architecture

compliance-engine/
├── app/
│ ├── api/compliance/ # Enhanced APIs with semantic processing
│ ├── property/enhanced/ # Advanced frontend with source citations
│ └── globals.css # Enhanced styling
├── lib/
│ ├── enhanced-compliance-engine.ts # Core semantic compliance logic
│ ├── semantic-compliance-bridge.ts # Bridge between semantic and API layers
│ ├── enhanced-setback-processor.ts # Semantic rule processor
│ └── property-data.ts # NSW Planning Portal integration
├── examples/regulatory-engine/
│ ├── dual_semantic_processor.py # Main semantic processing engine
│ ├── langextract_config.py # LangExtract configuration
│ └── compare_extraction_methods.py # Regex vs Semantic comparison
└── public/regulatory-data/ # Processed DCP data and caches

Quick Start

Development Setup

# Clone the repository
git clone https://github.com/lmcdo/nswcompliance.git
cd nswcompliance

# Install Node.js dependencies
npm install

# Set up Python virtual environment
python -m venv venv_linux
source venv_linux/bin/activate # Linux/Mac
# or venv_linux\Scripts\activate # Windows

# Install Python dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.local.example .env.local
# Edit .env.local with your API keys

# Start the development server
npm run dev

Database Configuration

PostgreSQL Connection Parameters:

• Host: localhost
• Port: 5432 (standard PostgreSQL port)
• Database: nsw_planning
• User: postgres
• Password: postgres

Test your database connection:

python db_config.py

Usage

1. Standard Compliance Check: Visit /property for basic compliance checking
2. Enhanced Compliance Check: Visit /property/enhanced for semantic processing with source citations

API Endpoints

# Get enhanced setback rules
curl "http://localhost:3000/api/compliance/setbacks?address=123%20Main%20St%20Ashfield%202131&semantic=true"

# Check development compliance
curl -X POST "http://localhost:3000/api/compliance/check" \
 -H "Content-Type: application/json" \
 -d '{
 "propertyData": {...},
 "proposal": {...},
 "formerCouncilArea": "Ashfield",
 "useSemanticRules": true
 }'

Technical Details

Semantic Processing Pipeline

1. LangExtract: Precise source text grounding with regulatory citations
2. AutoSchemaKG: Knowledge graph construction for complex relationships
3. Three-Tier Classification: Automatic enforcement level detection
4. Bridge Layer: Converts semantic outputs to compliance rule format

Supported Areas

• Ashfield: Former Ashfield Council area
• Leichhardt: Former Leichhardt Council area
• Marrickville: Former Marrickville Council area

Rule Types

• LEP Rules: Height and FSR from NSW Planning Portal (HIGH confidence)
• DCP Setbacks: Semantic extraction from council documents (MEDIUM confidence)
• Complex Conditions: Conditional logic like "minimum OR calculated, whichever is greater"

Performance

• Semantic Processing: ~2-3 seconds for initial extraction, <1 second for cached
• API Response Time: <500ms for cached data, <2 seconds for live processing
• Accuracy: 93% confidence for complex conditional rules vs 40% for regex
• Coverage: Handles 15+ different setback rule patterns vs 3 for basic regex

Testing

# Run Python tests
pytest examples/regulatory-engine/

# Run comparison analysis
python examples/regulatory-engine/compare_extraction_methods.py

# Test API endpoints
npm run test # (if test suite exists)

Deployment

The system is designed for production deployment with:

• Caching Strategy: Pre-processed rules cached in public/regulatory-data/
• Fallback Mechanisms: Graceful degradation from semantic to basic processing
• Error Handling: Comprehensive error catching with user-friendly messages
• Performance Monitoring: Processing time tracking and confidence scoring

Documentation

• Integration Plan: See AUTOSCHEMAKG_INTEGRATION_PLAN.md
• API Documentation: Swagger/OpenAPI specs available at /api/docs
• Processing Comparison: See public/regulatory-data/extraction_method_comparison.json

Contributing

This project uses Context Engineering principles with Claude Code:

1. Review CLAUDE.md for development guidelines
2. Add examples to examples/ folder for new patterns
3. Follow the three-tier rule classification system
4. Ensure all regulatory extractions include source grounding

License

This project is licensed under the MIT License - see the LICENSE file for details.

🆘 Support

For issues and feature requests:

• GitHub Issues: https://github.com/lmcdo/nswcompliance/issues
• Documentation: See /docs folder for detailed guides

---

Built with Context Engineering and Claude Code

Features dual semantic processing for legally defensible development compliance checking with source authority citations.