https://doi.org/10.5281/zenodo.18718420

Conda environments

Conda environments#

A tutorial for setting up your conda environments on Neurodesk.

This tutorial was created by Fernanda L. Ribeiro.

Email: fernanda.ribeiro@uq.edu.au

Github: @felenitaribeiro

Twitter: @NandaRibeiro93

Getting Set Up with Neurodesk

This tutorial documents how to create conda environments on Neurodesk.

Conda/Mamba environment#

The default conda environment is not persistent across sessions, so this means any packages you install in the standard environment will disappear after you restart the Jupyterlab instance. However, you can create your own conda environment, which will be stored in your homedirectory, by following the steps on this page. This method can also be used to install additional kernels, such as an R kernel.

  1. In a Terminal window, type in:

1_terminal

For Python:

mamba create -n test ipykernel pip
#OR
conda create -n test ipykernel pip

or for R:

mamba create -n r_env r-irkernel
#OR
conda create -n r_env r-irkernel

Important: Include both ipykernel and pip when you create a Python environment. JupyterLab uses ipykernel to run notebooks, and pip installs Python modules in the environment. If you omitted either package, activate the environment and install both packages:

conda activate test
conda install ipykernel pip

  1. To check the list of environments you have created, run the following:

mamba env list
#OR
conda env list

  1. Activate the environment:

For Python:

conda activate test

To load Neurodesk software modules from a notebook, install jupyterlmod in the environment:

python -m pip install jupyterlmod

jupyterlmod provides the Python package used by import module. A user-created conda environment does not inherit this package from the base environment. You do not need to install jupyterlmod to use module or ml in a terminal.

Install a Python module with python -m pip. For example, install NumPy:

python -m pip install numpy

To install the modules listed in a requirements file, run:

python -m pip install -r requirements.txt

For R:

conda activate r_env

  1. Open a new Launcher tab. The new environment can take a short time to appear on the JupyterLab start page after the installation finishes. When it appears, select the new Notebook option to start a Jupyter Notebook in that environment.

2_env

Switching the environment on a Jupyter Notebook is also possible on the top right corner dropdown menu.

3_notebook