Writing 3D tiffs throws deprecation warning

[K-Meech]

Writing a 3D tiff via the image_io sub-module throws a DeprecationWarning e.g.:

image = np.ones((4, 4, 4))
file_path = Path(tmp_dir) / "test_array.tif"
save.to_tiff(image, file_path)

DeprecationWarning: <tifffile.TiffWriter.write> data with shape (4, 4, 4) and dtype 'float64' are stored as RGB with contiguous samples. Future versions will store such data as MINISBLACK in separate pages by default, unless the 'photometric' parameter is specified.

Should the photometric / planarconfig be set explicitly to keep the old settings? Or are you happy to use the new tifffile defaults and just add this warning to the list of those ignored by pytest?

[adamltyson]

It would be good to know what's going on here, so we can explicitly set the parameter. We rarely (if ever) need to write either float64 or RGB though, so it would be good to know what this should be set to for our "normal" usage (typically 16bit greyscale, 2/3D).

[K-Meech]

The same error occurs with int32 data, so it's more to do with the image's shape. It appears to be recognising any image with a length of 3 or 4 in the last axis as rgb. In their examples they recommend using photometric='minisblack' for grayscale images. The explanation of the different photometric options is linked here

[adamltyson]

In that case I would definitely set it to MINISBLACK (ideally as a default parameter in the BrainGlobe function rather than hardcoding) unless @alessandrofelder thinks differently?

[alessandrofelder]

I would definitely set it to MINISBLACK

Yep, I agree with this

[K-Meech]

Ok - will do! I'll make a PR to add this