feat(python): Support zero-copy conversion for temporal types in DataFrame.to_numpy

[stinodego]

Closes #16415

Refactors the code to align more with how the Series view is made (though obviously there are some differences), and in the process allows zero copy support for Datetime and Duration types.

That's the last NumPy PR from me for now 😅

[codecov]

Codecov Report

Attention: Patch coverage is 98.40000% with 2 lines in your changes are missing coverage. Please review.

Project coverage is 81.43%. Comparing base (f0e38b5) to head (a714d42).
Report is 1 commits behind head on main.

Files	Patch %	Lines
py-polars/src/interop/numpy/to_numpy_df.rs	98.13%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #16429      +/-   ##
==========================================
+ Coverage   81.42%   81.43%   +0.01%     
==========================================
  Files        1409     1409              
  Lines      184594   184662      +68     
  Branches     2971     2971              
==========================================
+ Hits       150314   150388      +74     
+ Misses      33760    33754       -6     
  Partials      520      520              

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