The published Docker image includes ARC and RMG along with convenience entrypoints.
Bind-mount your working directory and pass an input file path that exists inside
the container. For best write access on bind mounts, pass your host UID/GID as
PUID/PGID (the entrypoint remaps the mambauser account).
Run ARC non-interactively:
docker run --rm \
-v "$PWD:/work" -w /work \
-e PUID=$(id -u) -e PGID=$(id -g) \
laxzal/arc:latest arc my_case/input.yml
Run RMG non-interactively:
docker run --rm \
-v "$PWD:/work" -w /work \
-e PUID=$(id -u) -e PGID=$(id -g) \
laxzal/arc:latest rmg my_case/input.py
Manual RMG invocation:
docker run --rm \
-v "$PWD:/work" -w /work \
-e PUID=$(id -u) -e PGID=$(id -g) \
laxzal/arc:latest \
micromamba run -n rmg_env python /home/mambauser/Code/RMG-Py/rmg.py my_case/input.py
Manual ARC invocation:
docker run --rm \
-v "$PWD:/work" -w /work \
-e PUID=$(id -u) -e PGID=$(id -g) \
laxzal/arc:latest \
micromamba run -n arc_env python /home/mambauser/Code/ARC/ARC.py my_case/input.yml
Open an interactive shell:
docker run --rm -it \
-v "$PWD:/work" -w /work \
-e PUID=$(id -u) -e PGID=$(id -g) \
laxzal/arc:latest
For job submission, the scheduler client tools must be available in the container or accessed via SSH on a remote host.
When you open an interactive shell, the image provides these aliases:
rc # reload ~/.bashrc rce, erc # edit ~/.bashrc with nano mamba, conda # micromamba deact # micromamba deactivate rmge, arce # activate rmg_env / arc_env rmgcode # cd to /home/mambauser/Code/RMG-Py dbcode # cd to /home/mambauser/Code/RMG-database arcode # cd to /home/mambauser/Code/ARC rmg # run RMG with input.py, tee logs arkane # run Arkane with input.py, tee logs arc # run ARC with input.yml, tee logs arcrestart # run ARC with restart.yml, tee logs