add argument return_background for get_centroids and get_element_instances in the case of labels

[giovp]

closes #614

[giovp]

here the labels has values 0,100 but the 0 element is background, so is effectively filtered in the left/right joins when get_element_instances is called on labels. Hence the indices of the instances are different.

[giovp]

@melonora here I would also need your green light since, due to the change of get_element_instances, the results of the tests had to be modified. In short, because now the get_element_instances doesn't return the background as default, hence the joins with the table produce a different results. Please see inline comments above, let me know if anything is unclear and if you agree with this.
I guess one option would be to expose the return_background in join, but I really don't see it to be of any value at the moment. Wdyt?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.91%. Comparing base (95d69ff) to head (3b7026e).
Report is 45 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #621      +/-   ##
==========================================
- Coverage   91.93%   91.91%   -0.03%     
==========================================
  Files          44       44              
  Lines        6661     6666       +5     
==========================================
+ Hits         6124     6127       +3     
- Misses        537      539       +2     

Files with missing lines	Coverage Δ
src/spatialdata/_core/centroids.py	100.00% <100.00%> (ø)
src/spatialdata/_core/query/relational_query.py	90.65% <100.00%> (-0.17%)	⬇️

... and 2 files with indirect coverage changes

[melonora]

Hi @giovp, LGTM! However as you said I agree that it does not make sense to have it in the join, but I would add it in the extra explanation of the docstring for the function before merging, but for now I pre-approve. Did you already check whether any tests on the other repos fail?

[giovp]

@melonora indeed some tests in spatialdata-plot fails, so I will open a separate PR before merging this