Improvements/fixes: merscope reader

[quentinblampey]

Hello,

I made a few improvements and fixes for the merscope reader:

• previously, the table was not kept when subsetting a SpatialData object. This is something I discussed with @LucaMarconato to better understand what was called "region" in the TableModel.parse. Now, region=f"{dataset_id}_polygons", and I create another variable (vizgen_region) to denote the region name used by vizgen, and not mix the two. It now works as intended.
• The Merscope cell shapes are MultiPolygons with only one polygon. Even though this is not throwing any error, shapes will not be displayed by spatialdata-plot. Now, I use the first (and only) polygon from the MultiPolygon for each cell.
• Except for the toy dataset, I couldn't save any image to .zarr, because I was not making chunks. I used the same logic as in the xenium reader with the imread_kwargs and image_models_kwargs variables.
• I now use the micron space by default (instead of the pixels space), which feels more natural because we use to deal with distances in microns.

NB: sorry the 19 first commits shouldn't have been shown (commits from the first PR). Only the last two commits are new

Let me know!

[codecov-commenter]

Codecov Report

Merging #73 (137fe8a) into main (755d475) will decrease coverage by 0.38%.
The diff coverage is 15.38%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #73      +/-   ##
==========================================
- Coverage   41.92%   41.54%   -0.38%     
==========================================
  Files          16       16              
  Lines         854      869      +15     
==========================================
+ Hits          358      361       +3     
- Misses        496      508      +12     

Files Changed	Coverage Δ
src/spatialdata_io/readers/merscope.py	25.00% <8.33%> (-2.03%)	⬇️
src/spatialdata_io/_constants/_constants.py	100.00% <100.00%> (ø)

[LucaMarconato]

Looks great to me, thanks for the refinement!