Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.


The HPC and Data Processing Group periodically provides a pre-configured module with various scientific Python and R packages and libraries preinstalled. The advantages of this are: * You, as a user, don't need to install and maintain your own conda environment * Conda environments typically contain many files which is not good for the performance of distributed storage systems, such as the one
Markdown
-
from-url
https://gitlab.awi.de/hpc/user-docs/-/raw/master/using-the-analysis-toolbox.md Albedo. By offering a global environment
  to the users that covers most of the packages used we hope to reduce the number of 
  private conda environments existing in our system.

## Basic Usage

You can access this package using the standard `module load` command:

```console
$ module avail analysis_toolbox
$ module load analysis_toolbox/01.2024
```

You can inspect which packages are available via:

```console
$ mamba list -p /albedo/soft/sw/conda-sw/analysis-toolbox/01.2024  # Alternatively `conda list`
```

## Branching Off

It might be that our pre-configured toolbox does not contain all of the tools you might need. To
that end, it is possible to build on top and add your own packages:. There are two possibilities.

1. **Cloning the environment:** Using this method, you create your own local copy of the toolbox
   and use it as a starting point. [Here](https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#cloning-an-environment)
   is how it can be done:

   ```console
   $ mamba create --name <MY_ENV> --clone /albedo/soft/sw/conda-sw/analysis-toolbox/01.2024
   ```

2. **Stacking Environments:** In this case you still build on top of the original environment, but only
   install new binaries. You can keep the `$PATH` of the initial environment. Examples are provided in 
   [greater detail here](https://conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html#nested-activation),
   but effectively you need:

   ```console
   $ conda activate --stack <MY_ENV>
   ```