In the current writing logic, the user has to calculate the coordinate_transformations themselves for each pyramidal layers before passing them on to the write_image function. I think it's much easier to simply pass a scale argument as a per-axis dict (i.e., {"z": 0.5, "y": 0.2, "x": 0.2}) and then just auto-calculate all the coordinate transformations for every scale level.
The default behavior for unpassed coordinatetransformations can stay as it is.
Working plan:
- add
scale argument to kwargs of write_image, write_labels functions
- deprecate
coordinate_transformations argument of these functions
cc @will-moore
cc @melonora : I think this is also orthogonal to your efforts regarding improving passage of arguments to zarr
In the current writing logic, the user has to calculate the
coordinate_transformationsthemselves for each pyramidal layers before passing them on to thewrite_imagefunction. I think it's much easier to simply pass ascaleargument as a per-axis dict (i.e.,{"z": 0.5, "y": 0.2, "x": 0.2}) and then just auto-calculate all the coordinate transformations for every scale level.The default behavior for unpassed coordinatetransformations can stay as it is.
Working plan:
scaleargument to kwargs ofwrite_image,write_labelsfunctionscoordinate_transformationsargument of these functionscc @will-moore
cc @melonora : I think this is also orthogonal to your efforts regarding improving passage of arguments to zarr