Skip to content

JupyterHub

JupyterLab is a web-based interactive development environment for jupyter notebooks, code and data.

Jupyter Notebook is a web-based interactive computational environment. It allows to create notebooks containing code, text, visualizations. It is language agnostic and is compatible with more than 40 programming languages depending on kernels installed.

JupyterHub is a cloud plateform that can run instances of Jupyterlab or Jupyter notebooks.

Access

One jupyter per IPSL cluster is available.

The jupyterhub are only available inside IPSL network. If you are not using a VPN, you need to create an SSH tunnel to access it from outside. See here for more information.

Create a password

If you don't already have a password for IPSL mesocentre (this is different from your ssh key passphrase), you can create one by going to https://meso-pass.ipsl.fr/pwrst/.

Note

Your login is the one you use to connect to spirit and spiritx. It is the login in <your_login>@spirit1.ipsl.fr or <your_login>@spiritx1.ipsl.fr.

Resources and cluster

The jupyterhub allows you to reserve up to 4 cores and get 8BG of RAM per core. It means if you request 2 cores, you will get 16GB of RAM. The walltime can be chosen between 1 and 8 hours.

jupyterhub MOSS simple

At the bottom of the screen, you can see the number of cores available on the partition.

Advance interface

jupyterhub MOSS advance

This menu allows you to start the jupyterhub with your own environment but it needs to have the package below at least: - batchspawer 1.3.0 - jupyterhub 4.1.5

The batchspanwer is available both on pip and conda-forge repositories

To install it:

  • with pip

    pip install 'batchspawner>=1.3.0' 'jupyterhub<4.2'

  • with conda

    conda install -c conda-forge 'batchspawner>=1.3.0' 'jupyterhub<4.2'

For the path of the environment, you can also use python environments built within an apptainer container.

How much memory does my server use ?

The current memory usage is displyed at bottom left of your dashboard.

current_memory_usage.png

Access to data

You can access to the data served by ESPRI data and computiong center directly from your Jupyter session. Go to this page to know how to access your data.

Default Kernels

All the most recent python environments available with the module command are avaialble by default in jupyter

default kernels

How to install your own kernel

From scratch

You can make your own environment available as a kernel in jupyter.

  1. Activate the module pangeo-meso/2024.01.22
  2. Create you environment and install the package you want. You environement need to contain the package ipykernel.
  3. Install the module or the environement into JupyterLab.
  4. A new kernel will be available at the next start of JupyterLab.

Example for the module pangeo-meso/2024.01.22:

# load the module
module load pangeo-meso/2024.01.22

# Create and install your environment
mkdir -p /scratchx/<your_login>/python_envs
conda create -p /scratchx/<your_login>/python_envs/myenv -c conda-forge python ipykernel your_package ...

# load your environment
conda activate /scratchx/<your_login>/python_envs/myenv

# install the kernel
python -m ipykernel install --user --name "myenv" --display-name "env name in jupyter"

Existing environment

Example for a conda environement (named myenv), provided the package ipykernel is already installed:

conda activate myenv
python -m ipykernel install --user --name=myenv

Example for a Python virtual environement (named myenv), provided the package ipykernel is already installed:

source "/path_to_my_env/myenv/bin/activate"
python -m ipykernel install --user --name=myenv

Example for an extended mesocenter Python module:

The module, named module_name, is extended by a Python virtual environement, named myenv. For more informations, read the module extension procedure.

module load module_name
source "/path_to_my_env/myenv/bin/activate"
python -m ipykernel install --user --name=myenv

R environment

R/4.0.5 is available on JupyterLab. To launch R environment - select the following icon on your JupyterLab server.

r_icon.png

You can modify R environment by following instructions here.

JupyterLab extensions

Several JupyterLab extensions are installed. We do not allow user to install other JupyterLab.

List of extensions

jupyterlab-git

This extension allows you to use git with a gui but be aware that it is not possible to push or pull if your remote use ssh. Only https remote repositories are working but you will have to enter your password each time you do a fetch, pull or push.

jupyterlab-git.png

If you want to keep a ssh remote. You can do commit, branch with the jupyterlab-git extension and use a terminal for the fetch, pull and push commands.

How to logout properly

On the navigation top bar, go to "File" > "Hub control panel".

hub_control_panel.png

Then click on "Stop My Server" red button.


stop_server.png

You can then log out from the "Home" page. Click on "Logout" button on the top right corner of the screen.