PsiAnimator-MCP

# PsiAnimator-MCP Deployment Guide ## Smithery Deployment This repository is configured for deployment on [Smithery](https://smithery.ai) with simplified, reliable configuration. ### Quick Deploy 1. **Push to GitHub**: Ensure your code is pushed to the repository 2. **Connect to Smithery**: Import your repository in Smithery 3. **Deploy**: Smithery will automatically detect the configuration and deploy ### Configuration Files - **`Dockerfile`**: Simplified build optimized for cloud deployment - **`Dockerfile.minimal`**: Ultra-lightweight version (core MCP only) - **`smithery.yaml`**: Simplified Smithery deployment configuration - **`config/default_config.json`**: Default server configuration ### Build Options (if you encounter timeouts) **Option 1: Default** (Recommended) ```yaml # Current smithery.yaml uses this dockerfile: "./Dockerfile" ``` **Option 2: Minimal Build** (If timeout issues persist) ```yaml # Update smithery.yaml to use: dockerfile: "./Dockerfile.minimal" ``` **Option 3: Conda-based** (Alternative approach) ```yaml # Update smithery.yaml to use: dockerfile: "./Dockerfile.light" ``` ### ✅ Current Status (Optimized for Smithery) **IMPORTANT: The repository has been optimized for Smithery deployment!** Changes made: - ✅ **Dockerfile**: Now uses only prebuilt wheels (`--only-binary=all`) - ✅ **smithery.yaml**: References standard `./Dockerfile` - ✅ **Simplified build**: No complex fallbacks or timeouts - ✅ **Committed locally**: Ready to push to GitHub ### 🚀 Next Steps for Deployment 1. **Push the optimized changes:** ```bash git push ``` 2. **Wait 2-3 minutes** for Smithery cache to refresh 3. **Try Smithery deployment again** - it should now find both files and build successfully ### Troubleshooting (if still having issues) **If Smithery still says "Could not find Dockerfile":** 1. **Check GitHub**: Verify files are visible at https://github.com/manasp21/PsiAnimator-MCP 2. **Wait longer**: Smithery cache can take up to 5 minutes to refresh 3. **Try minimal build**: Update `smithery.yaml` dockerfile to `"./Dockerfile.minimal"` **Build Taking Too Long:** - Default: Uses prebuilt wheels (faster) - Minimal: Only essential dependencies - Light: Uses conda-forge (may be slower but more reliable) ### Local Testing Test builds locally to debug: ```bash # Test default build docker build -t test-default . # Test minimal build (fastest) docker build -f Dockerfile.minimal -t test-minimal . # Test locally docker run --rm test-default python -c "import psianimator_mcp; print('Success')" ``` ### Environment Variables Configure in Smithery: - `PSIANIMATOR_TRANSPORT`: `stdio` (default) or `websocket` - `PYTHONUNBUFFERED`: `1` (for proper logging) ### Resource Requirements **Minimal:** - Memory: 256Mi - CPU: 100m **Default:** - Memory: 1Gi - CPU: 500m ### Production Notes - Start with minimal build for fastest deployment - Upgrade to full scientific stack once running - MCP server has conditional imports for graceful degradation - Heavy physics computations optional ### If All Else Fails Create a minimal `Dockerfile` with just: ```dockerfile FROM python:3.11-slim WORKDIR /app COPY . . RUN pip install mcp pydantic typer rich RUN pip install -e . --no-deps CMD ["python", "-m", "psianimator_mcp.cli", "serve"] ``` The MCP server is designed to work with or without heavy dependencies!