hub.oggm.org is our own JupyterHub deployment of OGGM on our servers in Bremen. It works similarly to MyBinder (see Try OGGM online) but it is bound to a username (you’ll need an account) and is therefore persistent (your files are saved between sessions). It also gives you access to more computing resources than MyBinder.
In order to be able to log in, you will need to have a (free) user account. It is super easy, just Get in touch if you want to try it out!
If you are new to the Jupyter Notebooks and JupyterLab, you will probably find this introduction to interactive notebooks quite useful.
hub.oggm.org is still experimental and we cannot guarantee that your work will always be safe here. We will do our best, but, you know, we are scientists after all. Please, make a copy of your files from time to time!
Hub or Binder?¶
We provide two solutions to try OGGM online, without local installation. Which is best for you? Here is a quick comparison.
|Files not saved between sessions||Files saved between sessions|
|Limited computational resources||Dedicated processing and space on OGGM servers|
|Use cases: quick tests and demos||Use cases: deeper explorations and teaching|
|No registration required||Contact us to register|
How does this work?¶
We use a single compute node located in Bremen to welcome the hub users.
Currently (we are still trying things out) each user gets enough CPUs (4) and
enough RAM (8Gb) to run OGGM on several glaciers at once.
This is not enough to do heavy work, but will get you through the exploratory
phase or even small regional runs. Each user also gets a persistent 16Gb disk
to save output data, notebooks and scripts. The OGGM-specific input data
(i.e. everything that is downloaded automatically between users, see
system settings) is shared among
shared folder in your
$HOME). The first time you run a
new glacier, OGGM will first check if the data is available in the
and if not it will download it for you and the other users.
When logging in, you can choose between two environments:
oggm_latest, with the latest OGGM installed from master (updated every few weeks)
oggm_vXXX(starting from OGGM v1.4), which are environments made with a pinned OGGM version
These environments are restarted each time you log-out and log-in again
(don’t worry, your
HOME and all it contains won’t be erased!). This means
that while you can install things in the root tree (e.g. with
it won’t be there the next time you open your hub. If you have special
requirements, please let us now so that we can add them, or install them in
pip install --user.
Accessing the tutorials (and other content) with nbgitpuller¶
You can execute this command in your JupyterHub terminal (Launcher -> Start a terminal):
$ gitpuller https://github.com/OGGM/tutorials master tutorials
That will copy the notebooks into the
in your home directory (on OGGM-Hub, not your local machine).
You can use a similar command to pull content from other repositories as well (e.g. the
Another way to pull content into your hub is to use a special weblink. Say, for example, that you would like to download the content of Lizz’s glacier course (Spanish notebooks) into your lab as well. You can use the nbgitpuller link generator to create the following links which, once clicked, will open your workspace with the new notebooks in it. Here are some useful links to add notebooks to your hub:
- Add OGGM’s tutorial notebooks to your hub
- Add Lizz’s Clubes de Ciencia Perú 2019 notebooks to your hub
nbgitpuller will never overwrite changes that the user made to the files in the pulled folder. This is very important to remember: sometimes you would like to get an updated version of the notebooks for example, and this will not work if you made changes to the file there. Therefore, it is always a good idea to make a working copy of the original file/folder before working on it (right-click -> rename).
The full set of rules used by nbgitpuller while pulling is explained here.
Top tip: copy-pasting text with the mouse in JupyterLab¶
Copying to and from JupyterLab can be annoying at times (context). This is one of the most frequent issue hitting users of JupyterLab when working in a terminal or when selecting text from notebook cells.
In these cases, press
shift + right click to experience a standard
“copy/paste” mouse menu.