Skip to main content

Software Environment Options

galyleo enables you to launch Jupyter notebook sessions within any software environment. To specify these software environments, it currently supports the use of environment modules, conda environments, and Singularity (or Docker) containers. The only requirement is Jupyter must be available within the software environment when the job starts. Otherwise, no Jupyter session can be created.

  • -e, --env-modules: comma-separated list of environment modules to load
  • --conda-yml: path to a conda environment.yml file
  • --cache: caches your conda environment; reused if the environment.yml file does not change between sessions
  • -s, --sif: URI to a Singularity (or Docker) container stored in a registry, or path to a Singularity container image file
  • -B, --bind: comma-separated list of user-defined bind paths to be mounted within a Singularity container
  • --nv: enable NVIDIA GPU support when running a Singularity container
tip

--conda-yml environments provide the most flexible option for most users. It dynamically builds your conda environment at runtime from the environment.yml file. You can even perform additional conda install commands during the live session to modify the environment as needed.

warning

Do not mix environment module, conda environment, and/or container options together in any one launch command. In general, you should choose only one way to specify the software environment for your Jupyter notebook session. However, there are exceptions.