You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
So in the branch feature/get_extent there is the new get_extent function which should be outsourced from spatialdata-plot.
I just compared it to the _get_extent function from spatialdata-plot and it looks like both compute the same numbers but the spatialdata output has the x and y coordinates switched which seems to be a bug. I added code to hopefully reproduce the issue.
Would be cool if you could have a look at it :)
importspatialdataassdimportspatialdata_plot# This example doesn't really work with blobs (bc it has identical x and y extents), so I used the example data from https://s3.embl.de/spatialdata/spatialdata-sandbox/visium_associated_xenium_io.zipvisium_sdata=sd.read_zarr("visium_associated_xenium_io.zarr")
print("Old function (from spatialdata-plot):")
print(spatialdata_plot.pl.utils._get_extent(visium_sdata))
# output format is [xmin, xmax, ymin, ymax], so for "global" I get xmax = 19505
print("New function (from spatialdata):")
print(sd._core.data_extent.get_extent(visium_sdata, coordinate_system="global"))
# This would imply ymax = 19505
print("Ground truth:")
print(visium_sdata.images)
# (c: 3, y: 21571, x: 19505) would mean the spatialdata output is incorrect because x and y are swapped
Activity
LucaMarconato commentedon Aug 10, 2023
Great catch Sonja, thanks for reporting!