[python] Append-mode enum bit-width mods

[johnkerl]

Issue and/or context: Presented in detail on #2456.

Changes: As on #2456.

Notes for Reviewer: See also updates below.

[johnkerl]

Update: @nguyenv 's recent cppification has fixed the case when a larger-bit-width enum exists on disk, with a smaller-bit-width append -- this now works correctly.

The only remaining issue is: when a smaller bit width is done for the initial ingest, and then we overflow.

• For thetiledbsoma API: this is intentional: if a user supplies say int8 index type in their user-provided Arrow schema and user-provided data exceed that capacity, then we must fail -- and we do
• For the tiledbsoma.io API: users do not pass in Arrow schema -- they simply pass in AnnData objects and they are entitled to expect appendability. The necessary step here is for tiledbsoma.io to broaden our inference for index types to int32 -- regardless of the initial input data -- since any enumerated column on any dataset could be appended to.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.19%. Comparing base (94bfd1c) to head (83fd5c4).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2455      +/-   ##
==========================================
+ Coverage   89.87%   90.19%   +0.31%     
==========================================
  Files          37       37              
  Lines        3941     3988      +47     
==========================================
+ Hits         3542     3597      +55     
+ Misses        399      391       -8     

Flag	Coverage Δ
python	90.19% <100.00%> (+0.31%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
python_api	90.19% <100.00%> (+0.31%)	⬆️
libtiledbsoma	∅ <ø> (∅)

[johnkerl]

@nguyenv 83fd5c4