NGFF v06 transformations

[Tomaz-Vieira]

Mirrors Ngff* classes in the graph module, parses ome-zarr-models-py classes.

CoordSystem (based on NgffCoordinateSystem)

• Made immutable, so we can __hash__ (needed for usage in graph, but we might still have the identity issue when searching for verts)
• Can be marked "virtual" so that we know not to serialize it (that is, it has been created for the sake of putting it in a graph and maintaining the invariant of always having an input and output, but it should not show up in the zarr metadata)

*Edge (based on NgffTransformation)

• All transformations always have input and output, which diverge from the raw NGFF
• Don't do serialization to raw JSON anymore, but rather convert to and from omz-models-py classes
• Stronger invariants; checked in __init__ and nowhere else
• A few corner cases fixed

io_raster.py

• Adds function to read multiscales from v06, applying the scale/translation transformations as a RangeIndex to the resulting xarrays

Next steps

Roundtripping

This PR only does input dfor now; The full roundtrip is showing some issues that would delay this further, so I've removed all output for now and will submit it in another PR

Using unorderd axes in coordinate systems

Since our arrays are labeled, we probably shouldn't be using a Sequence of axes in our coordinate systems. This will probably simplify runtime quit a lot, but might create some more work during IO

[codecov]

Codecov Report

❌ Patch coverage is 85.18519% with 88 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (transformation_manager@b0ff71c). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
src/spatialdata/transformations/graph/edge.py	87.29%	53 Missing ⚠️
src/spatialdata/transformations/graph/vert.py	77.10%	19 Missing ⚠️
src/spatialdata/_io/io_raster.py	68.62%	16 Missing ⚠️

Additional details and impacted files

@@                    Coverage Diff                    @@
##             transformation_manager    #1182   +/-   ##
=========================================================
  Coverage                          ?   92.22%           
=========================================================
  Files                             ?       55           
  Lines                             ?     8654           
  Branches                          ?        0           
=========================================================
  Hits                              ?     7981           
  Misses                            ?      673           
  Partials                          ?        0           

Files with missing lines	Coverage Δ
...ialdata/_core/transformation_manager/exceptions.py	100.00% <100.00%> (ø)
src/spatialdata/transformations/__init__.py	100.00% <100.00%> (ø)
src/spatialdata/_io/io_raster.py	85.44% <68.62%> (ø)
src/spatialdata/transformations/graph/vert.py	77.10% <77.10%> (ø)
src/spatialdata/transformations/graph/edge.py	87.29% <87.29%> (ø)

🚀 New features to boost your workflow:

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

[LucaMarconato]

Rough edge: MapAxis is basically an affine (made of zeros and a few ones to permute the data). When we do repr on the affine we print the matrix and not the axes (which will be handled by the coordinate systems). Here we do the opposite, we print the axes and don't care about the affine.

It's not incorrect, but I see that this goes more towards the philosophy of spatialdata transformations, while the affine stays "true" to the NGFF transformations.

For clarify we could actually always print both: the "index-based/matrix representation" (in this case index 0 goes to 1, index 3 goes to 4 etc; in the affine case the matrix); and then the axes annotation (x -> y, y -> z here; in the case of affine the input and output axes, as we do when we print spatialdata Affine transformations).

[Tomaz-Vieira]

I think I'd rather avoid the indexed output since we have everything labeled (axes have names, arrays have dims), and hopefully we eventually completely hide the fact that indices have any meaning for us