490 fix notebooks

[Lukas113]

closes #490

[Lukas113]

Converted to a draft because it works locally but not on github runners and they're just triggered after an opened PR.

When everything works, then

if [[ ${{ github.event_name }} == pull_request ]]; then
    export RUN_SLOW_TESTS=true

needs to be changed back to false

[Lukas113]

I've disabled the notebook-tests entirely because they don't run on the github runners (see action). I'm not entirely sure about the reason, maybe because of limited RAM of 7GB?

However, I don't see how this can be fixed easily. In the meantime, I suggest running the notebooks on your local machine by setting RUN_SLOW_TESTS=true should be good enough.

[codecov[bot]]

Codecov Report

Patch and project coverage have no change.

Comparison is base (adfe033) 68.67% compared to head (ad80bb8) 68.67%. Report is 3 commits behind head on main.

:exclamation: Current head ad80bb8 differs from pull request most recent head cd7d7f4. Consider uploading reports for the commit cd7d7f4 to get more accurate results

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #491 +/- ## ======================================= Coverage 68.67% 68.67% ======================================= Files 34 34 Lines 3761 3761 Branches 475 475 ======================================= Hits 2583 2583 Misses 1001 1001 Partials 177 177 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[kenfus]

Hmm... That is wierd. Maybe check which one fails? Is that possible? In the notebook HIIM_Img_Recovery there is also leftovers of some dask tests, I guess? That is very annoying if our notebooks are not checked regularly

[Lukas113]

Ok, just the source_detection.ipynb is causing the issue. Therefore, the other notebooks are run. I also removed RUN_SLOW_TESTS option since the running two notebooks don't take a long for execution.

[Lukas113]

I've just removed the other two notebooks. They did run on individually on the runners, however they didn't in the whole test-framework (see action).

[Lukas113]

So, after a google-searched I figured out that the issue is probably that the maximum disk-space is at max (10GB, see bug 26351). This is probably caused because our messy behavior of FileHandle and it's usage. As far as I know there are cases where multiple copies of the same file are created on different locations. In addition, there is no easy way to clean-up these automatically intermediate files. Therefore, the disk-usage just grows and comes at it's limit as some point. This has to change if we want to keep running a lot of tests on the GitHub runners.

[Lukas113]

In the meantime, two Notebooks which did run on the runners, but not together with the other tests are skipped until we fix this issue.