JupyterHub

JupyterHub is accessible from UK’s network (or VPN) at https://jupyterhub.ccs.uky.edu.

Use your linkblue ID/password to log into the system.

Initial login will take a few minutes as it builds the default conda environment.

Create a custom conda environment and install it to JupyterHub

1. Log into JupyterHub. You should initialize to the “Lab” screen.

2. Click File > New > Terminal.

3. A new tab should be created with a terminal prompt. Create a conda environment with your specific package requirements, for example:

   conda create --name CustomEnv1 "python==3.9" "numpy>=1.10"

   This creates an environment named “CustomEnv1” that will install Python v3.9 and numpy>=1.10. You can customize the name and packages.

4. Once the environment is created, you need to activate the environment with the following commands:

   source /opt/miniconda/bin/activate conda activate CustomEnv1

5. Additional packages can be installed to customize the environment using the following:

   conda install matplotlib

6. Once the environment is finalized, we would like to install it as an ipython kernel to be accessible from JupyterHub:

7. Now you can restart your JupyterHub server (File > Hub Control Panel > Stop My Server THEN Start My Server), and the “Custom Environment” kernel will be available.

Remove a custom conda environment and kernel from JupyterHub

1. Log into JupyterHub. You should initialize to the “Lab” screen.

2. Click File > New > Terminal.

3. A new tab should be created with a shell. List the kernels that are currently installed:

4. Remove your chosen kernel:

5. Remove the conda environment:

6. Restart the JupyterHub server (File > Hub Control Panel > Stop My Server THEN Start My Server) and the kernel is no longer available.