Architecture Diagram Generator

AI-powered architecture diagram generator that converts natural language descriptions into visual architecture diagrams.

Features

• ✅ Natural Language Processing: Convert text descriptions to diagrams using Strands Agents
• ✅ Multi-Cloud Support: AWS, Azure, and GCP with provider consistency enforcement
• ✅ Multiple Output Formats: PNG, SVG, PDF, DOT
• ✅ Advanced Code Mode: Direct Python code editing with autocomplete
• ✅ Examples Panel: Pre-built architecture examples for each provider
• ✅ Session Management: Automatic expiration and cleanup
• ✅ Request Tracking: Request ID tracking for debugging
• ✅ Production Ready: Security hardening, API documentation, deployment scripts

Project Structure

diagrams/
├── backend/              # Python FastAPI backend
│   ├── src/
│   │   ├── api/         # API routes with request tracking
│   │   ├── agents/      # Strands agents (diagram, modification, classifier)
│   │   ├── generators/  # Diagram generators (universal, diagrams engine)
│   │   ├── models/      # Pydantic models (ArchitectureSpec, etc.)
│   │   ├── resolvers/   # Component resolvers (intelligent, library discovery)
│   │   ├── advisors/    # Architectural advisors (AWS advisor)
│   │   └── validators/  # Input validation
│   ├── config/          # Node registry YAML files
│   ├── tests/           # Test suite
│   └── main.py          # FastAPI application entry point
├── frontend/            # React TypeScript frontend
│   └── src/
│       ├── components/  # React components (DiagramGenerator, AdvancedCodeMode, etc.)
│       ├── services/    # API client services
│       ├── data/        # Example prompts data
│       └── pages/       # Page components (HelpPage)
├── deployment/          # EC2 deployment scripts and systemd services
├── docs/               # Comprehensive documentation
│   ├── API.md          # API endpoint documentation
│   ├── ARCHITECTURE.md # System architecture
│   ├── DECISIONS.md    # Architectural decisions (ADR format)
│   ├── EC2_DEPLOYMENT_GUIDE.md # EC2 deployment guide
│   └── EC2_QUICK_REFERENCE.md   # EC2 quick reference
└── config/             # Feature flags and environment config

Quick Start

Backend

cd backend
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

Frontend

cd frontend
npm install
npm run dev

Development Phases

• ✅ Phase 1: MVP - Basic AWS diagram generation
• ✅ Phase 2: Provider selection (AWS/Azure/GCP)
• ⏳ Phase 3: Chat-based modifications (not yet implemented)
• ✅ Phase 4: Universal generator
• ✅ Phase 5: Polish & production - Security, optimization, documentation

Documentation

Essential documentation:

• Architecture - System architecture overview
• API Documentation - API endpoints and quick reference
• Architectural Decisions - Key design decisions

Interactive API Docs:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Testing

Quick test run:

cd backend
pytest tests/ -v

For comprehensive test reports and detailed instructions, see Test Instructions.

Quick reference:

# Full test suite with reports
python tests/run_tests.py --all-reports --verbose

# Run specific test file
pytest tests/test_api.py -v

# Run with coverage
python tests/run_tests.py --coverage

Deployment

Prerequisites

• AWS EC2 instance (Amazon Linux 2023)
• Security group: ports 22, 3000, 8000
• IAM role attached to EC2 instance with Bedrock access permissions

Quick Deploy

# SSH into EC2
ssh -i your-key.pem ec2-user@your-ec2-ip

# Pull and deploy
cd /opt/diagram-generator
git pull origin main
bash deployment/deploy-git.sh

Initial Setup

1. Install dependencies: sudo yum install -y python3 python3-pip nodejs graphviz git
2. Clone repository to /opt/diagram-generator
3. Setup backend: cd backend && python3 -m venv venv && source venv/bin/activate && pip install -r requirements.txt
4. Setup frontend: cd frontend && npm install && npm run build
5. Configure .env in backend/ with AWS region and Bedrock model ID (IAM role provides credentials)
6. Create systemd services (see deployment/systemd/ directory)
7. Start services: sudo systemctl start diagram-api diagram-frontend

Updating

cd /opt/diagram-generator
git pull origin main
bash deployment/deploy-git.sh
# Or manually:
# cd backend && source venv/bin/activate && pip install -r requirements.txt && sudo systemctl restart diagram-api
# cd ../frontend && npm install && npm run build && sudo systemctl restart diagram-frontend

Troubleshooting

Service Issues:

• Check logs: sudo journalctl -u diagram-api.service -f
• Verify services: sudo systemctl status diagram-api diagram-frontend
• Restart services: sudo systemctl restart diagram-api diagram-frontend

Backend Errors:

• ModuleNotFoundError: No module named 'strands'

  cd backend
  source venv/bin/activate
  pip install git+https://github.com/strands-agents/sdk-python.git
  deactivate

• Dependency conflicts: Upgrade packages:

  pip install --upgrade uvicorn[standard] pydantic python-multipart anyio
  pip install -r requirements.txt --upgrade

• Python version: Ensure Python 3.10+ is installed

  python3 --version  # Should be 3.10+
  # If not available: sudo yum install -y python3.11 python3.11-pip python3.11-devel

Security Features

• ✅ Path traversal protection for file serving
• ✅ Input validation and sanitization
• ✅ Session expiration (1 hour TTL)
• ✅ CORS configuration for production
• ✅ Request ID tracking for debugging
• ✅ Automatic file cleanup (24-hour retention)

Recent Improvements

• ✅ Output format selection (PNG, SVG, PDF, DOT)
• ✅ Download format dropdown with regeneration
• ✅ Instance caching for performance optimization
• ✅ Comprehensive API documentation with Swagger UI
• ✅ Request ID tracking middleware
• ✅ Session management with automatic cleanup
• ✅ File cleanup for old diagrams
• ✅ TypeScript type safety improvements