Refactor GT.opt_all_caps()

[jrycw]

I'm curious if GT.opt_all_caps() can be refactored as suggested in this PR.

[codecov]

Codecov Report

Attention: Patch coverage is 89.47368% with 2 lines in your changes missing coverage. Please review.

Project coverage is 86.88%. Comparing base (bacda7e) to head (f0ae1a7).
Report is 6 commits behind head on main.

Files with missing lines	Patch %	Lines
great_tables/_options.py	88.88%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #436      +/-   ##
==========================================
+ Coverage   86.37%   86.88%   +0.50%     
==========================================
  Files          42       42              
  Lines        4683     4674       -9     
==========================================
+ Hits         4045     4061      +16     
+ Misses        638      613      -25     

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

[jrycw]

@rich-iannone and @machow , sorry for the messy commits.

The reason for this is that while trying to modify the locations parameter to favor Loc objects, a circular import issue occurred.

Currently, the type annotation and initial value of locations are incorrect, causing the documentation build to fail in CI. Since this is a breaking change, I would like to get your feedback on this proposed update.

[machow]

Hey, thanks for working on this -- this was a good nudge that we should probably circle back to #341 and round out the support of loc pieces in Great Tables. I can pick up #341 next week, so we can merge these changes in around the same time :).

[jrycw]

Thanks for bringing up the related issue. Initially, I was thinking of using a Loc object like loc.column_labels for the locations parameter. However, using a Loc instance like loc.column_labels() seems like a better idea.

[machow]

Sorry for the wait, now that the new location stuff is merged in, I can take a look at this next week