This service has been removed. This page is retained for archive purposes.
Anyone with a HEP account can log in to our Jupyterhub service and run or create Notebooks. The service is intended for teaching and testing purposes, large scale processing should be performed on interactive or batch nodes directly.
Use a supported browser to log in at https://hep.ph.liv.ac.uk/jhub/
with your HEP account. Supported browsers are currently Firefox and Chrome. Although Safari is supported the current version runs very slowly. Notebooks may run in Edge/IE but we can't guarantee it.
When you first log in you can choose a pre-defined server to spawn. The minimal setup is very minimal, probably only useful for teaching basic Python. The Datascience server has many standard scientific modules added eg numpy, matplotlib etc and is probably the best to choose for most purposes. The Liverpool HEP server is based on the Datascience server, with some extra Python modules added and may change regularly.
Hub or Lab
By default the interface will start in JupyterLab
mode, which is intended to replace the older JupyterHub
interface. If you still wish to use the older interface it can be accessed from Help>Launch Classic Notebook.
To logout use File>Log out.
You can stop your current server and start with a different one. Under File>Hub Control Panel you can choose to stop your server. After a while you can then start a new server. Note you should close the original browser tab before doing this otherwise a namespace conflict will occur.
Each user will have a kernel spawned on the cloud system, providing 20GB of storage for files and output, up to 2GB of RAM and up to 1 CPU. This kernel will persist after you have logged out. After some period of time the kernel may be stopped to free resources but the storage will be restored if you log in again. The storage is not backed up and may be cleared, any important files should be copied to permanent storage elsewhere.
There is a shared storage area available under /home/shared. This is available to all accounts so can be used to provide materials to multiple users. Note that all accounts can read/write/delete any files in this area so treat it as a scratch area not permanent storage.
The backend cloud service is currently running on refurbished hardware and should be considered less reliable than our core services.
The available CPUs unfortunately do not support AVX instructions so Tensorflow is currently not supported in Python notebooks.
The Lab interface has extensive online help available in the Help menu.
Lots of examples and courses using notebooks are linked from the Jupyterhub webpages https://github.com/jupyter/jupyter/wiki/A-gallery-of-interesting-Jupyter-Notebooks
. Many can be downloaded as .ipynb files and then edited or uploaded to the HEP Jupyterlab.