Tile sizes were previously taken directly from the input zarr's chunk sizes. Now, if the chunk size is a multiple of 16 it will be used directly, otherwise it will be rounded up to the next multiple of 16. This mostly affects the smallest resolution images (which are typically a single tile) and datasets generated by bioformats2raw with non-default tile sizes. When the tile size is calculated to be different from the chunk size, a warning is now printed.
This shouldn't actually affect pixel data that is returned when reading converted OME-TIFFs (with Bio-Formats, at least). It should eliminate warnings from libtiff as described in the test case in #88.
Fixes #88.
Tile sizes were previously taken directly from the input zarr's chunk sizes. Now, if the chunk size is a multiple of 16 it will be used directly, otherwise it will be rounded up to the next multiple of 16. This mostly affects the smallest resolution images (which are typically a single tile) and datasets generated by bioformats2raw with non-default tile sizes. When the tile size is calculated to be different from the chunk size, a warning is now printed.
This shouldn't actually affect pixel data that is returned when reading converted OME-TIFFs (with Bio-Formats, at least). It should eliminate warnings from libtiff as described in the test case in #88.