Fix all_scalars usage in threshold

[MatthewFlamm]

Overview

In the tests, it is shown that all_scalars=True does not use all the scalars but rather all the points in a cell for the threshold filter. This can also be seen when using cell data where there is only 1 value per cell therefore all_scalars=True does not have an effect even when having multiple cell data scalars.

Details

There was a discussion of this in #1488 (comment), where @adeak wished to have all_scalars to mean 'all points' match rather than 'all scalars' match. In my testing, this is indeed what this argument does.

I'm not sure if this would suggest a name change from all_scalars to all_points.

[adeak]

I can say with certainty that I had no clear idea back then that thresholding is a cell-only operation, so I was working with multiple levels of confusion 😄 These parts of the docs were certainly unclear to me when I looked at them the other day.

[MatthewFlamm]

My comment was not meant to disparage any understanding, but rather that your intuition of what would be useful was a good pointer that the pyvista implementation and description was off.

[codecov]

Codecov Report

Merging #2670 (0e29475) into main (608c592) will decrease coverage by 0.00%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main    #2670      +/-   ##
==========================================
- Coverage   93.70%   93.70%   -0.01%     
==========================================
  Files          75       75              
  Lines       16180    16178       -2     
==========================================
- Hits        15162    15160       -2     
  Misses       1018     1018              

[akaszynski]

Thanks for clearing this up @MatthewFlamm.