bio-mcp-evo2

Bootstrap: docker From: nvcr.io/nvidia/pytorch:24.01-py3 %post # Update and install system dependencies apt-get update && apt-get install -y \ git \ wget \ curl \ python3-pip \ && rm -rf /var/lib/apt/lists/* # Set up Python environment pip install --upgrade pip # Install MCP and dependencies mkdir -p /app cd /app # Clone and install evo2 git clone --recurse-submodules https://github.com/ArcInstitute/evo2.git /opt/evo2 cd /opt/evo2 pip install --no-cache-dir . # Install MCP server dependencies cd /app pip install --no-cache-dir mcp>=1.1.0 pydantic>=2.0.0 pydantic-settings>=2.0.0 httpx>=0.24.0 nvidia-ml-py>=12.0.0 # Create working directory mkdir -p /tmp/mcp-work chmod 777 /tmp/mcp-work %files src /app/src pyproject.toml /app/ %environment export LC_ALL=C export PATH=/opt/conda/bin:$PATH export PYTHONPATH=/app:$PYTHONPATH export BIO_MCP_EVO2_TEMP_DIR=/tmp/mcp-work export BIO_MCP_EVO2_EXECUTION_MODE=local export PYTHONUNBUFFERED=1 %runscript cd /app exec python -m src.server "$@" %help This container provides an MCP server for the evo2 DNA language model. Usage: singularity run --nv evo2.sif The --nv flag is required to enable GPU support. Environment variables: BIO_MCP_EVO2_EXECUTION_MODE: Execution mode (local, sbatch, api) BIO_MCP_EVO2_MODEL_SIZE: Model size (7b or 40b) BIO_MCP_EVO2_CUDA_DEVICE: CUDA device index BIO_MCP_EVO2_NIM_API_KEY: Nvidia NIM API key for API mode %labels Author Bio-MCP Team Version 0.1.0 Description MCP server for evo2 DNA language model with GPU support