A published container image carries the compiled solver, the Python layer and the input workbook, so that nothing has to be built locally:

docker pull ghcr.io/ghoniem/radcluster:v2.2.0
docker run --rm -it ghcr.io/ghoniem/radcluster:v2.2.0

Inside the container the solver is on the PATH, and the folder /work holds py_utils, input and codes, with PYTHONPATH already set.

The output is written inside the container, and it is lost when the container exits. Mount a host directory to keep it:

docker run --rm -v "$PWD/output:/work/output" ghcr.io/ghoniem/radcluster:v2.2.0 \
       python3 codes/Python_Testing/check_eurofer_rag.py

Note that the solver in the image is compiled for the x86-64-v2 architecture level and not for the architecture of the machine that builds it. It therefore runs on any x86-64 processor from about 2009 onwards, at a small cost in speed. For sustained production runs, build the solver from source. See Building the solver.

The image is built and published by the workflow .github/workflows/release.yml of the repository when a version tag is pushed.