feat: add XYZ scale to ArrayDisplayModel

[tlambert03]

something we've needed here at the QI course...

exposes the ability to set the scale part of the transform (and only the scale for now) on the image being displayed. mostly needed for anisotropic z sampling

may introduce some additional changes needed for the camera centering

[codecov]

Codecov Report

Attention: Patch coverage is 64.70588% with 12 lines in your changes missing coverage. Please review.

Project coverage is 45.53%. Comparing base (cdc4f5d) to head (d5da376).
Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
src/ndv/models/_array_display_model.py	76.19%	5 Missing ⚠️
src/ndv/views/_pygfx/_array_canvas.py	0.00%	4 Missing ⚠️
src/ndv/controllers/_array_viewer.py	50.00%	3 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (cdc4f5d) and HEAD (d5da376). Click for more details.

HEAD has 1 upload less than BASE

Flag	BASE (cdc4f5d)	HEAD (d5da376)
	2	1

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #165       +/-   ##
===========================================
- Coverage   80.80%   45.53%   -35.28%     
===========================================
  Files          45       45               
  Lines        4271     4553      +282     
===========================================
- Hits         3451     2073     -1378     
- Misses        820     2480     +1660     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[tlambert03]

one challenge this PR generally exposes: how does a user pass data-related metadata (like channel names, scales, etc...). Yes, we support named arrays and we try to get all this info from xarray etc... but i don't want someone to have to create an xarray just to add some metadata. The DataWrapper is the natural place for this to live... but someone shouldn't have to subclass it to customize it (if it's not in the array info itself already).

this PR could just make the visualization possible, without weighing in on the public API