𓂀𓁢𓋹𝔸ℕ𝕌𝔹𝕀𝕊𓋹𓁢𓂀 - Intelligent Guidance for

# 🪟 Windows Docker Publication Guide ## 🎯 Complete Guide for Publishing Anubis to Docker Hub on Windows ### 🆕 What's New in This Version - **🌱 Integrated Database Seeding**: Workflow rules are automatically seeded during Docker build - **⚡ Instant Startup**: No manual `npm run rules:gen` required - everything is pre-configured - **🔧 Enhanced Build Process**: Improved multi-stage build with validation and testing - **📦 Production Ready**: Optimized for both development and production deployments ### 📋 Prerequisites 1. **Docker Desktop for Windows** - Installed and running 2. **Docker Hub Account** - Create at [hub.docker.com](https://hub.docker.com) 3. **PowerShell 5.1+** - Built into Windows 10/11 4. **Git for Windows** - For version control 5. **Node.js 22+** - For testing before publication ### 🔧 Pre-Publication Setup #### Step 1: Verify Docker Installation ```powershell # Check Docker is running docker --version docker info # Login to Docker Hub docker login # Enter your Docker Hub username and password ``` #### Step 3: Test Your Application ```powershell # Install dependencies npm install # Build the application npm run build # Run tests (if available) npm test # Test Docker build locally docker build -t test-mcp-workflow . docker run --rm test-mcp-workflow --help ``` ### 🚀 Publication Process #### Option A: Manual Step-by-Step 1. **Build Multi-Platform Image**: ```powershell # Create and use buildx builder docker buildx create --name mcp-builder --use --bootstrap # Build and push for multiple platforms docker buildx build --platform linux/amd64,linux/arm64 --tag hiveacademy/anubis:latest --tag hiveacademy/anubis:1.2.20 --push . ``` 2. **Verify Publication**: ```powershell # Check the image was published docker pull hiveacademy/anubis:latest # Test the published image docker run --rm hiveacademy/anubis:latest --help ``` ### 🔧 Configuration Options #### Environment Variables for Different Use Cases **1. Basic STDIO Usage (Default)**: ```powershell # For Claude Desktop integration docker run --rm -i ` -v mcp-workflow-data:/app/data ` hiveacademy/anubis:latest ``` **2. HTTP Server Mode**: ```powershell # For web-based clients docker run --rm ` -p 3000:3000 ` -e MCP_TRANSPORT_TYPE=SSE ` -e NODE_ENV=production ` -v mcp-workflow-data:/app/data ` hiveacademy/anubis:latest ``` **3. Production with PostgreSQL**: ```powershell # With external database docker run --rm ` -p 3000:3000 ` -e DATABASE_URL="postgresql://user:pass@postgres:5432/workflow_db" ` -e MCP_TRANSPORT_TYPE=SSE ` -e NODE_ENV=production ` hiveacademy/anubis:latest ``` ### 📝 Docker Hub Configuration #### Repository Settings on Docker Hub 1. **Repository Name**: `anubis` 2. **Description**: "A comprehensive Model Context Protocol server for AI workflow automation and task management" 3. **README**: Copy content from `DOCKER_HUB_README.md` 4. **Tags**: - `latest` (always points to newest stable) - `1.0.0`, `1.1.0`, etc. (specific versions) - `develop` (development builds) #### Automated Builds (Optional) Link your GitHub repository for automatic builds: 1. Go to Docker Hub → Account Settings → Linked Accounts 2. Link your GitHub account 3. Create automated build from `Hive-Academy/Workflow_Manager_MCP` ### 🔍 Testing Publication #### Local Testing ```powershell # Test different configurations # 1. Basic functionality docker run --rm -v mcp-test-data:/app/data hiveacademy/anubis:latest --help # 2. HTTP mode docker run --rm -p 3000:3000 -e MCP_TRANSPORT_TYPE=SSE hiveacademy/anubis:latest & Start-Sleep 5 Invoke-WebRequest http://localhost:3000/health -UseBasicParsing Stop-Process -Name node ``` #### Integration Testing ```powershell # Test with Claude Desktop configuration $claudeConfig = @{ mcpServers = @{ "anubis" = @{ command = "docker" args = @("run", "--rm", "-i", "-v", "mcp-workflow-data:/app/data", "hiveacademy/anubis:latest") } } } | ConvertTo-Json -Depth 3 Write-Host "Add this to your Claude Desktop config:" Write-Host $claudeConfig ``` ### 📊 Monitoring and Maintenance #### Health Checks ```powershell # Check container health docker run --rm --health-cmd "npm run health-check" hiveacademy/anubis:latest # Monitor running containers docker stats ``` #### Updates and Versioning ```powershell # When releasing new versions # 1. Update package.json version npm version patch # or minor/major # 2. Build and publish new version .\scripts\docker-publish.ps1 -Version $(node -p "require('./package.json').version") # 3. Test the new version docker pull hiveacademy/anubis:latest ``` ### 🛡️ Security Considerations #### Container Security - ✅ Image runs as non-root user - ✅ Minimal attack surface (Alpine Linux base) - ✅ No sensitive data in image layers - ✅ Environment variables for configuration - ✅ Health checks enabled #### Network Security ```powershell # For production deployment docker run --rm ` --network mcp-network ` --restart unless-stopped ` -v mcp-workflow-data:/app/data ` hiveacademy/anubis:latest ``` ### 🌟 Best Practices #### Version Management - Use semantic versioning (1.0.0, 1.1.0, etc.) - Tag releases in Git - Maintain changelog - Test before publishing #### Performance Optimization ```powershell # Optimize image size docker images hiveacademy/anubis # Check layer sizes docker history hiveacademy/anubis:latest ``` ### 🎉 Success Verification After successful publication, verify: 1. **Docker Hub Page**: https://hub.docker.com/r/hiveacademy/anubis 2. **Image Pull**: `docker pull hiveacademy/anubis` 3. **Multi-Platform**: Check AMD64 and ARM64 support 4. **Documentation**: Verify README displays correctly 5. **Tags**: Confirm version tags are available ### 🔗 Integration Examples #### Claude Desktop Configuration ```json { "mcpServers": { "anubis": { "command": "docker", "args": [ "run", "--rm", "-i", "-v", "mcp-workflow-data:/app/data", "hiveacademy/anubis" ] } } } ``` #### Docker Compose ```yaml version: '3.8' services: anubis: image: hiveacademy/anubis:latest container_name: mcp-workflow environment: - MCP_TRANSPORT_TYPE=SSE - NODE_ENV=production ports: - '3000:3000' volumes: - mcp-workflow-data:/app/data restart: unless-stopped volumes: mcp-workflow-data: ``` ### 📞 Support and Troubleshooting #### Common Issues 1. **Docker Build Fails**: ```powershell # Clear Docker cache docker system prune -a ``` 2. **Multi-Platform Build Issues**: ```powershell # Reset buildx docker buildx rm mcp-builder docker buildx create --name mcp-builder --use --bootstrap ``` 3. **Permission Issues**: ```powershell # Run PowerShell as Administrator Set-ExecutionPolicy -ExecutionPolicy RemoteSigned -Scope CurrentUser ``` #### Getting Help - 📖 Documentation: See README.md - 🐛 Issues: GitHub Issues page - 💬 Community: Docker Hub comments - 📧 Support: Repository maintainers --- **🎯 Quick Start Summary for Windows**: 1. `.\scripts\clean-for-publish.ps1` - Clean repository 2. `docker login` - Login to Docker Hub 3. `.\scripts\docker-publish.ps1` - Publish image 4. Verify at https://hub.docker.com/r/hiveacademy/anubis **Total time**: ~10-15 minutes for first publication! 🚀 ### 🌱 Database Seeding Integration #### How It Works The new Docker build process automatically handles database seeding: 1. **Build-Time Seeding**: During Docker image creation, the workflow rules are seeded into a build-time database 2. **Runtime Detection**: When the container starts, it detects if seeding is already complete 3. **Automatic Fallback**: If seeding is needed at runtime, it happens automatically using `npx prisma db seed` #### Key Benefits - **⚡ Instant Startup**: No waiting for database initialization - **🔄 Consistent State**: Every container starts with the same workflow rules - **🛡️ Error Resilient**: Automatic fallback if build-time seeding fails - **📦 Self-Contained**: No external dependencies or manual setup required #### Environment Variables ```powershell # These are automatically set in the Docker image BUILD_TIME_SEEDING_DEPLOYED=true # Indicates seeding completed during build MIGRATIONS_PRE_DEPLOYED=true # Indicates migrations completed during build ``` #### Troubleshooting Seeding If you encounter seeding issues: ```powershell # Force re-seeding in development docker run --rm -e FORCE_RESET=true -v mcp-data:/app/data hiveacademy/anubis # Check seeding status docker run --rm -v mcp-data:/app/data hiveacademy/anubis --verbose # Manual seeding (if needed) docker exec -it <container-id> npx prisma db seed ```