Python🔗

We install many versions of Python and all the common packages (many bundled with Python, many more as seprate modules)

The software we build for the cluster is optimized for the hardware. Pre-compiled versions are often only built generically, which will give up a lot of performance and features. Prefer using the packages from the the module tree.

Avoid using pip install directly. It will place user wide packages directly into your ~/.local/, not only using up most if not all or your disk quota, but won't be isolated and will leak into all other containers and environments likely breaking compatibility. Please only follow the examples below instead.

Don't mix modules with containers. They can never assumed to be compatible and will in the best case just crash.

Virtual environments🔗

To use virtualenv, we need load its module first. At the same time, we can load other preferable modules built from the same toolchain. Some of them may already have virtualenv as its dependency, but you can always explicitly load the module.

$ module load virtualenv/20.23.1-GCCcore-12.3.0 matplotlib/3.7.2-gfbf-2023a SciPy-bundle/2023.07-gfbf-2023a h5py/3.9.0-foss-2023a
$ module load JupyterLab/4.0.5-GCCcore-12.3.0  # Load JupyterLab if you are using the virtual environment on OnDemand portal

• Note: For toolchains earlier than 2023a, virtualenv is included in the Python modules and you don't have to load addition virtualenv module.

Once we loaded all needed modules, we create a new virtual environment (only done once), e.g.

$ virtualenv --system-site-packages my_venv

You will see a directory appear in your currenct directory. To use this environment, we must activate it and load the same modules we have loaded before creating the virtual environment (every time you log in)

$ module load virtualenv/20.23.1-GCCcore-12.3.0 matplotlib/3.7.2-gfbf-2023a SciPy-bundle/2023.07-gfbf-2023a h5py/3.9.0-foss-2023a
$ source my_venv/bin/activate

Once the virtual environment is activated, you should see the name of the virtual environment appears in front of you console. In this example, it will be

(my_venv) ... $ 

and you can use pip to install additional packages after activating the environment

$ module load virtualenv/20.23.1-GCCcore-12.3.0 matplotlib/3.7.2-gfbf-2023a SciPy-bundle/2023.07-gfbf-2023a h5py/3.9.0-foss-2023a
$ source my_venv/bin/activate
(my_venv) ... $ pip install <some_modules>
(my_venv) ... $ pip install <some_other_modules>

If you find that your packages are install into your ~/.local directory. It is most likely you forget to activate the virtual environment.

Note that there is a global pip configuration active on Vera and Alvis (see pip config -v list). That by default sets the following options for pip install:

• --no-user option is very important to avoid accidentally doing user installations that in turn would break a lot of things.
• --no-cache-dir option is required to avoid it from reusing earlier installations from the same user in a different environment.
• --no-build-isolation is to make sure that it uses the loaded modules from the module system when building any Cython libraries.

For more information see this external guide

Accessing virtual environments in Jupyter Notebook🔗

If you're using virtual-environments in connection with Jupyter Notebook you might have problems that the kernel used in Jupyter doesn't recognize the correct site-packages. To resolve this do the following after completing the above steps

$ module load virtualenv/20.23.1-GCCcore-12.3.0 matplotlib/3.7.2-gfbf-2023a SciPy-bundle/2023.07-gfbf-2023a h5py/3.9.0-foss-2023a
$ module load JupyterLab/4.0.5-GCCcore-12.3.0
$ source my_venv/bin/activate
$ python -m ipykernel install --user --name=my_venv --display-name="My Python"

and then when running your notebooks you should be able to select this as your kernel

• Note: about using your virtual environment on OnDemand portal, please refer to OpenOndemand portals

Conda environments🔗

The best way to create conda environments is instead do it with a container. There are example recipes at

/apps/containers/Conda/conda-example1.def
/apps/containers/Conda/conda-example2.def
/apps/containers/Conda/conda-example3.def

which after you've built them can be used as e.g.

apptainer exec conda-example.sif python my_script.py

This drastically reduces the installation size, and the number of files (down to 1), making it much better for network storage. Many conda packages might not be compatible with the operating system we use on the cluster. Using a container also solves this issue as well as allowing you to install additional OS packages.

Please see our container page for more details.

There are Anaconda3 modules available, which can be directly used (but might not be as optimized as the other software in the module tree). If you need to set up a conda environment, you should instead always use a container.

Note: The containers themselves serve as environments, you don't need to set up conda environments inside the containers (see example recipes) as it only complicates using the container.

Matplotlib🔗

Please note that when running Matplotlib, you might want to run

matplotlib.use('Agg')

after importing, to avoid Matplotlib trying to use the X Windows backend, which will fail if you didn't log in with X forwarding (which won't work in batch jobs).

NumPy and SciPy🔗

As it is impractical to have each individual extension as a module, so you can find the optimized builds of numpy and scipy as part of the SciPy-bundle. In general, you can use module keyword xxx to search for a particular extension that might be part of a module bundle.