How to Set Up Python Environment on the MDX Workflow Client

This guide explains how to set up a Python virtual environment for workflow execution on the mdx platform.

Prerequisites

Before you begin, ensure the following:

• You have an MDX workflow client account.
• SSH authentication to the workflow client is configured.

Instructions

[!IMPORTANT]
- Replace z12345 with your actual account name. - If you encounter an error indicating insufficient space in the ~ (home) directory on mdx, use the /large/z12345 directory instead.

Step 1: Log in to MDX Workflow Client

Connect to MDX workflow client:

ssh -A z12345@mdx-workflow.example.org

Step 2: Install uv Package Manager

Run the following command to install uv:

curl -LsSf https://astral.sh/uv/install.sh | sh

Step 3: Create a Python Virtual Environment

Use uv to create a virtual environment with Python 3.12:

uv venv ~/venv/prefect -p 3.12

This will install Python 3.12 and set up a new environment named prefect.

Step 4: Activate the Environment and Install Prefect.

Activate the virtual environment:

source ~/venv/prefect/bin/activate
uv pip install prefect

Your Python environment is now ready for workflow development.

Step 5: Add github.com to SSH known_hosts

If this is the first time to connect with github.com on this machine, make sure you added GitHub's SSH host key:

ssh -T git@github.com

You should see the output like:

Hi ...! You've successfully authenticated, but GitHub does not provide shell access.

---

END OF GUIDE