Easily run the memote command line tools from a Docker container. You can find the main documentation on readthedocs. If you use this Docker image, please cite the corresponding Zenodo DOI: [placeholder]. For the original software, please cite Memote.

The container exposes the memote command line tool. If you run the container without any arguments you will see the base help message.

docker run ghcr.io/metabolicatlas/memote-docker:0.17

For now, the best way to have memote interact with files is to mount a local directory into the container. The only hurdle is that you may have to afterwards change the permissions of the output file to your own user.

docker run -v ~/local/path/to/models/directory:/opt ghcr.io/metabolicatlas/memote-docker:0.17 memote run /opt/my-model.xml

The following functionality has been added when comparing to the original repository:

• Memote v0.17 (versus v0.12)
• Python v3.11 (versus v3.6)
• addition of scipy and yamllint
• Zenodo DOIs
• Gurobi support
• arm64 support (Apple Silicon) in addition to amd64

The image bundles gurobipy; if a Gurobi license is provided the container switches to Gurobi, otherwise it falls back to GLPK.

• Preferred: set GUROBI_LICENSE_B64 to a base64-encoded gurobi.lic string (preserves newlines). The entrypoint decodes it to ~/.gurobi/gurobi.lic and sets COBRA_SOLVER=gurobi.
• Fallback: set GUROBI_LICENSE to the raw license text (written as-is to ~/.gurobi/gurobi.lic).
• If neither is set, COBRA_SOLVER defaults to glpk and the existing behaviour is unchanged.

GitHub Actions example using base64:

- name: Run memote with Gurobi
  env:
    GUROBI_LICENSE_B64: ${{ secrets.GUROBI_LICENSE_B64 }}
  run: |
    docker run -e GUROBI_LICENSE_B64="$GUROBI_LICENSE_B64" \
      ghcr.io/metabolicatlas/memote-docker:0.17 \
      memote run /opt/my-model.xml