Switch guessers to catching and raising NoDataErrors

[lilyminium]

Fixes #4749 (maybe?)

Changes made in this Pull Request:

• Switches errors to NoDataErrors

PR Checklist

• Tests?
• Docs?
• CHANGELOG updated?
• Issue raised/referenced?

Developers certificate of origin

• I certify that this contribution is covered by the LGPLv2.1+ license as defined in our LICENSE and adheres to the Developer Certificate of Origin.

---

📚 Documentation preview 📚: https://mdanalysis--4755.org.readthedocs.build/en/4755/

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.57%. Comparing base (d7153b9) to head (7c9e176).
Report is 2 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #4755      +/-   ##
===========================================
- Coverage    93.59%   93.57%   -0.03%     
===========================================
  Files          177      189      +12     
  Lines        21708    22775    +1067     
  Branches      3052     3052              
===========================================
+ Hits         20318    21312     +994     
- Misses         943     1016      +73     
  Partials       447      447              

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

[orbeckst]

From reading the issue and glancing at the code, changing the raised ValueError to NoDataError makes sense and seems more consistent.

I am not really familiar with the new guesser code so this is just a surface-level review. Other reviews would be great.

@lilyminium I'll leave it to you to gather more opinions if you need them.

[lilyminium]

@IAlibay could you please have a look to see if this resolves issue #4749?

[yuxuanzhuang]

I think what @IAlibay means is that there is a nested exception error message:

a = np.array([
    [0., 0., 150.], [0., 0., 150.], [200., 0., 150.],
    [0., 0., 150.], [100., 100., 150.], [200., 100., 150.],
    [0., 200., 150.], [100., 200., 150.], [200., 200., 150.]
])
u = mda.Universe(a, n_atoms=9, to_guess=['elements'])

> ...
> NoDataError: This Universe does not contain name information
> During handling of the above exception, another exception occurred:
> ...
> NoDataError: there is no reference attributes in this universeto guess types from

which is kinda ugly. you can explicitly suppress the chaining of exceptions using the from None syntax (see suggestions)

[IAlibay]

I think what @IAlibay means is that there is a nested exception error message:

which is kinda ugly. you can explicitly suppress the chaining of exceptions using the from None syntax (see suggestions)

Yes, thanks so much @yuxuanzhuang, that's exactly what I was looking at! Sorry I conflated that with the error type for some reason.

[IAlibay]

LGTM with @yuxuanzhuang's suggestions.

Changes accepted