Instructions for instructors

The DTU databar

The summer school exercises will be running in the DTU Databar, not on Niflheim. You can log in with your DTU username and password. Like Niflheim, it has a login node and a number of compute node, but unlike Niflheim it also has interactive compute nodes where the students and you can run e.g. your Jupyter notebooks.

Logging in

ssh -X login.gbar.dtu.dk

Then proceed from the login node to an interactive compute node with the command:

linuxsh -X

Setting up the Virtual Environment etc

To prepare your account, run the script (you only need to do this once):

bash ~jasc/CAMD2022/setup2022_teacher

(Note that this is an extended version of the students’ setup script which will give you an editable version of GPAW and its documentation).

This will create a folder CAMD2022, a “slim” virtual environment in CAMD2022/venv, and a MyQueue configuration. The slim virtual environment is a fake venv folder where most of the contents is links to a master environment. If this is causing trouble, please contact Jakob to get it fixed.

Remember to activate the virtual environment - this may be done automatically for the students (XXX fix later):

source ~/CAMD2022/venv/bin/activate

Summerschool Notebooks

The summerschool notebooks exist in two versions, the student version and the teacher version. The student version is somewhat censored, the teacher version also contain a working solution.

The student version is autogenerated from the teacher version and is available from the CAMd Summer School 2022 web page under the individual projects. Remember also to download associated .png files or they will not display properly.

The teacher version is stored in the GPAW source code, but as a .py file with special comments to help automatic conversion to Notebook format. In this way, the teacher versions can be run as part of the GPAW test suite, this is the only reason that we can expect the notebooks to still work after four years of code development!

The .py files are found in the project folders under venv/gpaw/doc/summerschools/summerschool22.

Student versions of all notebooks are extracted from the source Python files when you build the GPAW documentation (make in the doc folder). Edit your notebooks in another folder, as building the documentation mayoverwrite all notebooks without warning.

Making small modifications to the notebooks

It is probably easiest to first make the modification in the downloaded Notebook file, and then make the same modification in the .py file. Then submit a merge request.

Teacher versus student version

In the .py file, a cell can be marked to be left out in the student version like this:

# %%
# teacher:
print('N2 bond length:', slabN2.get_distance(8, 9))

The first line marks that a new cell should be made, the second that the cell should be left out of the student version.

You can also make a single line appear different in the teacher and student version:

vib = Vibrations(slab,
                 name='vib',
                 indices=[8, 9],  # student: indices=[?, ?],
                 nfree=4)

Here, one line reads indices=[8, 9], but indices=[?, ?], in the student version.

Making new notebooks or large modifications

When making a new notebook, make it as a normal notebook, use the teacher / student markup in the comments as needed, and then get help converting it to a .py file once it is ready.

To make major changes to an existing notebook, first extract the “teacher version” from the .py file with the command:

cd doc/summerschools/summerschool22
python convert.py --teacher

move it somewhere else to update it, and then get help converting it back once it is ready. Note that rebuilding the documentation or rerunning the command above will cause all notebooks to be overwritten.

The extracted notebook will start with a cell:

# teacher
import ase.visualize as viz
viz.view = lambda atoms, repeat=None: None

It disables the ASE gui when the source script is run as part of the GPAW test suite, but should probably be deleted while you develop the notebook.

Running Notebooks in the Databar

Notebooks can developed on Niflheim or elsewhere, but all notebooks should be testet in the DTU Databar.

Instructions on how to run notebooks in the databar can be found here:

• Starting and accessing a Jupyter Notebook (Windows)

• Starting and accessing a Jupyter Notebook (Linux / macOS)

Updating the text of the project pages

Remember to also update the web pages associated with the projects. This is done by editing the associated .rst files in doc/summerschools/summeschool22, and then submitting a merge request.

Notes on how to set this up

These notes are intended as a starting point for setting this up in 2024 !

• Build a GPAW venv called venv-master

• Upgrade sphinx, otherwise it will not work:

  pip install --upgrade  sphinx sphinx-rtd-theme