deep/shallow copy metadata

[akaszynski]

Problem

As pointed out in this comment, metadata is always shallow copied (or not at all) when copying datasets:

>>> import numpy as np
>>> import pyvista as pv
>>> sphere_a = pv.Sphere()
>>> sphere_a['data0'] = np.random.random((sphere_a.n_points, 3))

>>> sphere_b = sphere_a.copy(deep=True)
>>> sphere_a._active_scalars_info is sphere_b._active_scalars_info
True

This should be False.

On the other hand, the textures themselves are always shallow copied, but a new set is created, being neither a shallow or deep copy.

Solution

This PR cleans up copy_meta_from and cleans up the implementation to either perform a true "deep copy" or "shallow copy" depending on the deep parameter.

[codecov]

Codecov Report

Merging #2779 (bd79d13) into main (d5c89f1) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #2779      +/-   ##
==========================================
+ Coverage   93.86%   93.88%   +0.01%     
==========================================
  Files          76       76              
  Lines       16230    16265      +35     
==========================================
+ Hits        15235    15270      +35     
  Misses        995      995              

[adeak]

CI failure is unrelated, due to an ongoing "incident" with GitHub Actions Cache.

[adeak]

LGTM!