setup a CRON job for testing examples

[MridulS]

fixes #943

[sbenthall]

• Any idea why the tests failed here?
• Looks like this runs only one example. Is the idea to extend it and add the others as well?

[MridulS]

Any idea why the tests failed here?

The tests are timing out because of matplotlib and it is trying to build a font cache (erratic behaviour)

Looks like this runs only one example. Is the idea to extend it and add the others as well?

Yeah that's the idea, this PR is just a test for it.

[codecov-io]

Codecov Report

Merging #944 (4d18621) into master (799943e) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #944   +/-   ##
=======================================
  Coverage   71.85%   71.85%           
=======================================
  Files          62       62           
  Lines        9205     9205           
=======================================
  Hits         6614     6614           
  Misses       2591     2591           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 799943e...4d18621. Read the comment docs.

[sbenthall]

I've been trying to get this running all weekend, and now see this doesn't even test the notebook examples, but the python files cut from the ConsumptionSaving code like a year ago? Why are we maintaining them?

in any case, this is not ready for the 1.0 release. I'm removing the milestone.

[MridulS]

the python files are supposed to be jupytext synced with the notebook files.
It doesn't really matter if we test running python files or notebooks, but if we keep both of them then we should test both of them.

[sbenthall]

they do not seem to be jupytext synced to the notebooks.

[llorracc]

I wonder whether it might make more sense to just maintain the .py files, and purge the .ipynb files. If we move to exclusive use of JupyterLab, it can open .py files as notebooks. You can edit them as notebooks (sadly, without some of the nice jupyter notebook extensions -- like codefolding ...) but when you quit they are saved as both .py and .ipynb files. So we could basically treat the .ipynb files as ephemera -- the .py files are the "single source of truth."

…

On Sun, Feb 28, 2021 at 6:16 PM Sebastian Benthall ***@***.***> wrote: I've been trying to get this running all weekend, and now see this doesn't even test the notebook examples, but the python files cut from the ConsumptionSaving code like a year ago? Why are we maintaining them? in any case, this is not ready for the 1.0 release. I'm removing the milestone. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#944 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAKCK73Q3DMN33C74KAUI7LTBLFFZANCNFSM4XJFN25A> .

-- - Chris Carroll

[llorracc]

OK, I did what I described above and locally deleted all the .ipynb notebooks. I found a few minor problems and have now managed to fix most of them. There are only a couple remaining, and those might be fixed if I do the opposite (load from the ipynb file disregarding the .py). Bottom line is I think all DemARKs can be working by end of day.

[MridulS]

btw everything passes if we just use pytest --nbval-lax (this only checks if every cell executes), currently looking into making it work with pytest --nbval too (not a priority but makes the tests more robust).

[MridulS]

@sbenthall the CRON job is setup and it passes now. Let me know if you are okay with the changes to the example notebooks and you can merge it then.