Removed warnings only for warnings with category : UserWarning

[spyke7]

Hi @PicoCentauri ,
See firstly I thought of using warnings, but saw that in the find_cls_members function, the logic of ignoring warnings was backward. So I fixed it just by removing not keyword and it removed warning of category : UserWarning.

I have previously run the warnings.simplefilter("ignore") from warnings, and got that the warning -

• This module was moved to MDAnalysis.analysis.hydrogenbonds.hbond_autocorrel; hbonds.hbond_autocorrel will be removed in 3.0.0. is DeprecationWarning
• And others are UserWarning
  So even if I use filterwarnings and gave category of UserWarning, they got removed as well.

But the problem was the first one which is still there.
I have used to filterwarning and gave category DeprecationWarning and tried in both __init__.py and __main__.py. But it remained.

I found that in MDAnalysis/analysis/hbonds/hbond_autocorrel.py -
warnings.catch_warnings() context manager is used with warnings.simplefilter("always", DeprecationWarning) which always shows it even if I try to filterit.

So in the moment I think either we fix the hbond_autocorrel.py or try to monkey patch (I don't know if it would work as well 🥲)

So please check the things and update me, I will try to help.

---

📚 Documentation preview 📚: https://mdacli--129.org.readthedocs.build/en/129/

[spyke7]

I will update the CHANGELOG later for the docs to be passing. And should I generate a PR in Mdanalysis regarding this a probably changing hbonds a little?

[PicoCentauri]

And should I generate a PR in Mdanalysis regarding this a probably changing hbonds a little?

Yes, a PR in MDA that allows setting an own context filter should be good. Not allowing to disable warnings is a bit agains the zen of Python.

[PicoCentauri]

hahaha this was easy.

Maybe we add a test for the ignore_warnings switch.

[spyke7]

?
should I add a test regarding this - about ignoring warnings?

[PicoCentauri]

yeah basically, running this function once with ignore_warnings=True (checking that no warning will be emitted) and once with ignore_warnings=False (checking that a warning with pytest.warns is emitted)

[spyke7]

And should I generate a PR in Mdanalysis regarding this a probably changing hbonds a little?

Yes, a PR in MDA that allows setting an own context filter should be good. Not allowing to disable warnings is a bit agains the zen of Python.

Can you please guide me about the probable title or the description of the issue...
And I want to change this -
(hbond_autocorrel.py)

with warnings.catch_warnings():
    warnings.simplefilter("always", DeprecationWarning)
    wmsg = (
        "This module was moved to "
        "MDAnalysis.analysis.hydrogenbonds.hbond_autocorrel; "
        "hbonds.hbond_autocorrel will be removed in 3.0.0."
    )
    warnings.warn(wmsg, category=DeprecationWarning)

Here just changing the always in warnings.simplefilter

[PicoCentauri]

And should I generate a PR in Mdanalysis regarding this a probably changing hbonds a little?

Yes, a PR in MDA that allows setting an own context filter should be good. Not allowing to disable warnings is a bit agains the zen of Python.

Can you please guide me about the probable title or the description of the issue... And I want to change this - (hbond_autocorrel.py)

with warnings.catch_warnings():
    warnings.simplefilter("always", DeprecationWarning)
    wmsg = (
        "This module was moved to "
        "MDAnalysis.analysis.hydrogenbonds.hbond_autocorrel; "
        "hbonds.hbond_autocorrel will be removed in 3.0.0."
    )
    warnings.warn(wmsg, category=DeprecationWarning)

Here just changing the always in warnings.simplefilter

I would say something along the lines Don't always warn in hbond_autocorrel. In the description you can also link this PR.

[spyke7]

Generated PR in mdanalysis. Only the fix in find_cls_memebers is enough for suppressing the warnings. When it will be accepted I can work on tests.

[PicoCentauri]

Yes nice I saw it. I think one can already now work on a test. Probably the test should not depend on mda itself because they might remove all warnings at some point and this will lead to a failing test.

[spyke7]

Hi @PicoCentauri
I added the tests please check it and tell me if I need to add or edit anyother things
And inside mdanalysis I talked with the maintainers and proposed them for removing the always word but they suggested me to replace find_cls_members with an explicit list of classes, removing hbonds.hbond_autocorrel from the list.
Please check the issue there.

[PicoCentauri]

Skipping this class should be rather easy. There is already a list here:

mdacli/src/mdacli/__main__.py

Lines 32 to 39 in 6333fdf

	skip_mods = [
	"AnalysisFromFunction",
	"HydrogenBondAnalysis",
	"WaterBridgeAnalysis",
	"Contacts",
	"PersistenceLength",
	"InterRDF_s",
	]

we could just add the hbond code to it.

[spyke7]

I checked for class names and even tried to include HydrogenBondAutoCorrel inside the skip_mods but it didn't skip it.
I ran this as a test file -

from MDAnalysis.analysis import __all__
print([m for m in __all__ if 'hbond' in m.lower()])

and got ['hbonds']
So I skipped it from the module_list, by adding m!=hbonds

[PicoCentauri]

Okay, weird. But I think we had a similar issue before.

Cool that it is working. I will check the code next week.

[PicoCentauri]

Very nice! Thanks again @spyke7. After merge I will do a release.

[spyke7]

Welcome @PicoCentauri
I would like to solve more issues and also implement the new features