Fix notebooks so they run again

[teutoburg]

Doesn't add them in the tests, nor really brings the information in there up to date, but at least they run now.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (5674cf3) 75.82% compared to head (bfff055) 76.00%.
Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #16      +/-   ##
==========================================
+ Coverage   75.82%   76.00%   +0.18%     
==========================================
  Files           9        9              
  Lines         823      821       -2     
==========================================
  Hits          624      624              
+ Misses        199      197       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[hugobuddel]

So I don't really understand, but this breaks RTD:

• This commit: https://readthedocs.org/projects/spextrahb/builds/23166762/
• master, reran just now: https://readthedocs.org/projects/spextrahb/builds/23167021/

I think the problem is

jupyter_client.kernelspec.NoSuchKernel: No such kernel named python3

However, nothing changed in that respect, in fact, the notebooks were broken before. It seems that RTD only runs the notebooks when they are 'un-run' in the repository.

In fact, whether RTD runs the notebooks was exactly what I wanted to verify. You added the notebooks without output, that is without running them before committing, which I think was on purpose. I do like to have the notebooks as an example about 'what to expect', and their output must be available for that. I'm fine with having the notebooks without output in the repository, as long as they have output on RTD. But that is apparently broken.

I admit that having broken notebooks that do show up in the documentation is worse than having working notebooks with broken documentation, so I will approve the PR anyway.

There seems to be much more broken with RTD / our documentation though: also the log of the successful build is full of Python Exceptions. So we should fix the RTD workflow to ensure the notebooks can be ran, and that exceptions raised in the rst files will trigger a failure. This probably applies to the other repositories as well.

[teutoburg]

Hmm, I missed that e.g. docs/database.rst has some code like from spextra.database import Database, which gets executed by RTD (as it should), but doesn't work anymore because of the restructuring I made. But I think it's best to address that in a separate PR, because it's not directly related to the notebooks....

[teutoburg]

Oh, and I'm looking at https://readthedocs.org/projects/spextra/, not spextraHB. But that doesn't seem to build the branches...