Rename translation function alias

[sngyo]

Corresponds to #424

In this project, _ is used as an alias for trans.gettext although _ is commonly used as a placeholder for ignored variables or loop variables in Python.
In this PR, I'm trying to rename _ to _i18n as discussed in #424, it would be clearer and more useful to avoid conflicts between these two uses.

P.S.
This is the first time for me to make a contribution, so I would appreciate it if you could kindly point out any inadequacies/advice 😉

TODO

• update CONTRIBUTING.rst since the link is outdated.
• there are bunch of strings which are not given to _i18n in serverapp.py

[kevin-bates]

Hi @sngyo - thanks for doing this. My primary comment is regarding constraining this to only the existing _() functions.

[kevin-bates]

I don't think we should be introducing new localized strings in this exercise. Only those statements wrapped in _() should be replaced with _i18n()

[kevin-bates]

Sorry, this comment is on the wrong line now and it looks like things were reverted, but I think the point is still the same - let's not introduce new strings for localization bundles in this PR. I think there are a few other cases in serverapp.py where this occurred.

[sngyo]

Hi @kevin-bates,
thanks for the advice, I'll revert the last two commits later! 😄
--> DONE!

[kevin-bates]

@goanpeca, @Zsailer - do you know if server extensions use this _ alias to wrap localizable strings? If so, we may not want to fully replace it, but define two aliases for now. Since we may not know, it might be best to keep both.

@sngyo - I suspect we'll want to add some kind of commented directive such that the "unused global variable" warning is not emitted. (This is probably another reason why _ was used.)

[sngyo]

@kevin-bates

1. two alias

How do you think introducing a deprecation warning for _() as below?

import warnings

def _trans_gettext_helper(*args, **kwargs):
    warn_msg = "The alias `_()` will be deprecated. Use `_i18n()` instead."
    warnings.warn(warn_msg, UserWarning)
    return trans.gettext(*args, **kwargs)


_ = _trans_gettext_helper
_i18n = trans.gettext

2. skip unused global variable

I'll look for the method to ignore (skip) the warning for Github CodeScanning. There should be something like this (maybe it will work 😆 ):

_i18n = trans.gettext  # noqa: E640

[kevin-bates]

1. two alias

I think it makes sense to add a deprecation warning for this. I see you're using UserWarning - I'm guessing due to the limitations for DeprecationWarning. I would suggest using FutureWarning for this instead since it more closely identifies with the intention. (We also used this warning level recently in jupyter_client.)

2. skip unused global variable

I had linting on my mind when I made this suggestion. I'm not immediately finding linting-style directives for codeQL, so this probably isn't a big deal since codeQL is still succeeding. Sorry for the misdirection.

[codecov-io]

Codecov Report

Merging #428 (5210fa2) into master (9b9b311) will increase coverage by 0.02%.
The diff coverage is 44.23%.

@@            Coverage Diff             @@
##           master     #428      +/-   ##
==========================================
+ Coverage   77.46%   77.48%   +0.02%     
==========================================
  Files         105      105              
  Lines        8936     8942       +6     
  Branches      956      956              
==========================================
+ Hits         6922     6929       +7     
+ Misses       1677     1676       -1     
  Partials      337      337              

Impacted Files	Coverage Δ
jupyter_server/serverapp.py	65.85% <33.33%> (ø)
jupyter_server/transutils.py	72.72% <57.14%> (-27.28%)	⬇️
jupyter_server/extension/application.py	72.44% <100.00%> (ø)
jupyter_server/services/contents/filemanager.py	66.95% <100.00%> (ø)
jupyter_server/services/contents/manager.py	88.32% <100.00%> (ø)
jupyter_server/services/kernels/handlers.py	58.24% <0.00%> (+1.06%)	⬆️

---

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 9b9b311...5210fa2. Read the comment docs.

[kevin-bates]

This looks good - thank you @sngyo. (I'm assuming this is no longer WIP, but please let us know if that's not the case.)

[sngyo]

@kevin-bates Thank you for the review and advice! It's ready to be merged 😄

[Zsailer]

LGTM! Thanks, @sngyo.

[kevin-bates]

Thanks @sngyo - this is helpful!