JupyterHub

Access

JupyterHub provides a graphical interface to Alderaan from a web browser. You need to be on the CU Denver internal network, or VPN, or connect from VMWare Horizon Remote Desktop (Click on Complimentary after you log in). Start a web browser (Firefox recommended), and go to

https://math-alderaan

or, if the short name does not work, use the full name

https://math-alderaan.ucdenver.pvt

Click away the warnings, and log in with your University username and password. Then click on the blue New Server button. After a short time, you should see the JupyterLab screen. Your JupyterLab session will run somewhere on Alderaan on a dedicated CPU with at least 4GB of memoty.

To access files and directories outside of your home directory, you can make soft links to them using a terminal. For example

ln -s /data001/Projects/myuserid ~/MyPrpjects

and then click on the soft link MyProjects in your home directory in the file tree.

JupyterHub is a new, still experimental addition to Alderaan. The web interface may be unavailable and running sessions may be interrupted as we keep working on it, so please save your files frequently.

Please write Alderaan Help from CU Denver email if you have any suggestioms or requests.

Adding a Conda Environment as a Jupyter Kernel

Using the Terminal

  1. Activate your conda environment (replace myenv with your environment's name): bash conda activate myenv

  2. Install the ipykernel package (if not already installed): bash conda install ipykernel

  3. Register the environment as a Jupyter kernel: bash python -m ipykernel install --user --name myenv --display-name "Python (myenv)"

This command registers the environment as a kernel available in Jupyter Notebook under the display name “Python (myenv)”.

Using a Jupyter Notebook

You can run the same commands within a notebook cell by prefixing them with !. For example:

# Install ipykernel in your environment (if needed)
!conda install ipykernel -y

# Register the kernel (ensure your notebook is running in the desired environment)
!python -m ipykernel install --user --name myenv --display-name "Python (myenv)"

After running these commands, restart your Jupyter Notebook server. When creating a new notebook or changing the kernel in an existing one, the "Python (myenv)" option should be available.

Removing a Conda Environment Kernel

Using the Terminal

  1. List all installed kernels: bash jupyter kernelspec list

Example output: myenv /home/username/.local/share/jupyter/kernels/myenv python3 /usr/local/share/jupyter/kernels/python3

  1. Remove the kernel using its kernel spec name (for example, myenv): bash jupyter kernelspec remove myenv

Confirm the deletion when prompted.

Using a Jupyter Notebook

You can also run the removal command from a notebook cell:

!jupyter kernelspec list
# Identify the kernel spec name you want to remove, then:
!jupyter kernelspec remove myenv

Additional Notes

  • Display Name vs. Kernel Spec Name:
    The --display-name option sets the name displayed in the Jupyter Notebook launcher, while the kernel is managed by its spec name (e.g., myenv). When removing a kernel, always use the kernel spec name.

  • User Permissions:
    These commands are run under your user account, so no admin rights are required.

  • Refreshing the Environment:
    After adding or removing kernels, you may need to restart your Jupyter Notebook server to see the changes.