Add helper functions to tile image for inference

[timtreis]

import squidpy as sq

sdata = sq.datasets.visium_hne_sdata()
sq.experimental.im.make_tiles(sdata, image_key="hne", preview=True)

import squidpy as sq

sdata = sq.datasets.visium_hne_sdata()
sq.experimental.im.make_tiles_from_spots(
    sdata,
    spots_key="spots",
    image_key="hne",
    tissue_mask_key="hne_tissue",
	preview=True,
)

[codecov]

Codecov Report

❌ Patch coverage is 67.64045% with 144 lines in your changes missing coverage. Please review.
✅ Project coverage is 65.88%. Comparing base (cacd943) to head (623ab3e).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/squidpy/experimental/im/_make_tiles.py	69.26%	72 Missing and 50 partials ⚠️
src/squidpy/experimental/im/_utils.py	50.00%	8 Missing and 3 partials ⚠️
src/squidpy/experimental/im/_detect_tissue.py	58.82%	6 Missing and 1 partial ⚠️
src/squidpy/_utils.py	55.55%	2 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1053      +/-   ##
==========================================
+ Coverage   65.52%   65.88%   +0.35%     
==========================================
  Files          43       44       +1     
  Lines        6364     6762     +398     
  Branches     1064     1137      +73     
==========================================
+ Hits         4170     4455     +285     
- Misses       1810     1874      +64     
- Partials      384      433      +49     

Files with missing lines	Coverage Δ
src/squidpy/_utils.py	56.25% <55.55%> (+0.89%)	⬆️
src/squidpy/experimental/im/_detect_tissue.py	52.94% <58.82%> (+2.31%)	⬆️
src/squidpy/experimental/im/_utils.py	38.29% <50.00%> (+3.00%)	⬆️
src/squidpy/experimental/im/_make_tiles.py	69.26% <69.26%> (ø)

🚀 New features to boost your workflow:

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

[Zethson]

A few random comments as requested

[LLehner]

Is it necessary to restrict this to Visium spots? This could also work for arbitrary segmentations as long as cell coordinates + e.g radius are given. It definitely works on custom cell circles.

[selmanozleyen]

I think you need to import spatialdata-plot to fix that

[timtreis]

Is it necessary to restrict this to Visium spots? This could also work for arbitrary segmentations

I think yes, because it generates non-overlapping tiles. Yes, you can generate them centered around cell segmentation masks but then the tiles would overlap which would probably create interpretability issues. So I'd semantically restrict it to the clean usecase - what users then actually do with it is their issue imho.

I think you need to import spatialdata-plot to fix that

wdym @selmanozleyen ?

[LLehner]

Looks good! Just wondering about the plot warning, but maybe thats on my side.

[selmanozleyen]

Besides my comments I checked the code coverage and it suggested that some cases weren't tested.

•  if H_mask != H or W_mask != W: # (line 803)
  

I am not sure if its accurate but overall I'd check the coverage report to ensure robustness for some cases.

[selmanozleyen]

looks good