RelayMD

RelayMD is a distributed orchestration system for long-running molecular dynamics workloads across HPC and cloud GPU capacity; project design and implementation decisions are split across several focused documents in the docs/ directory.

Start reading with:

• Core Architecture
• Job Lifecycle
• Tech Stack & Development Guidelines
• Deployment Guide

Live Documentation

The full documentation is deployed to GitHub Pages at https://ballaneypranav.github.io/RelayMD/.

Containers

RelayMD publishes two images to GitHub Container Registry (GHCR):

• ghcr.io/<org>/relaymd-worker:<tag>
• ghcr.io/<org>/relaymd-orchestrator:<tag>

Build and push commands:

make docker-build-worker ORG=<org>
make docker-push-worker ORG=<org>
make docker-build-orchestrator ORG=<org>
make docker-push-orchestrator ORG=<org>

Worker entrypoint: python -m relaymd.worker. Orchestrator entrypoint: relaymd orchestrator up. Runtime requirement: set INFISICAL_TOKEN=<client_id>:<client_secret> for both worker bootstrap and orchestrator startup.

Development

Frontend

The operator UI is built in frontend/ and served by the orchestrator on port 36158.

cd frontend
npm --cache ./.npm install
npm --cache ./.npm run build

Git Hooks

This repository uses Ruff and Pyright for code quality. We use git hooks to ensure these checks are run before every commit.

To set up the hooks locally, run:

make setup-hooks

The hooks are stored in .githooks/ and include:

• pre-commit: Runs ruff format, ruff check --fix, and pyright.
• pre-push: (Empty placeholder) Useful for longer-running checks like pytest.