to_nibabel and from_nibabel don't cancel out

[dangom]

Describe the bug
I believe this may be related to #52. I've just built AntsPy from master, though, so the fix for #52 doesn't solve the issue.
For an ANTsImage (RSA), converting to and from nibabel changes their orientation (or at least their display). Also values for spacing and direction change.

To Reproduce
Steps to reproduce the behavior:

import ants
data = ants.image_read("some_nifti_rsa.nii")
ants.plot_ortho(data, flat=True)
ants.plot_ortho(ants.from_nibabel(ants.to_nibabel(data)), flat=True)

Expected behavior
The to_nibabel and from_nibabel conversions cancel each other out?

Screenshots

Desktop (please complete the following information):

• OS: mac high sierra
• ANTsPy: master
• nibabel: 2.3.3

[stnava]

I've tested this in the past and it "works for me"

this probably tests the same thing:

ANTsPy/tests/test_core_ants_image_io.py

Lines 270 to 277 in 5ed4d11

	def test_nibabel(self):
	fn = ants.get_ants_data( 'mni' )
	ants_img = ants.image_read( fn )
	nii_mni = nib.load( fn )
	ants_mni = ants_img.to_nibabel()
	self.assertTrue( ( ants_mni.get_qform() == nii_mni.get_qform() ).all() )
	temp = ants.from_nibabel( nii_mni )
	self.assertTrue(ants.image_physical_space_consistency(ants_img,temp))

is there something different about these images? if you look at the code, I don't think there is any way that spacing could change as you describe ....... would need a reproducible example to debug.

[stnava]

also ITK coords are LPS.

[dangom]

Yes, get_ants_data('mni') doesn't showcase the issue. I tried that too as I thought that'd be a more reproducible example.
Try with the following data (zipped because github)
clone.nii.zip

import ants
data = ants.image_read("~/Downloads/clone.nii")
ants.plot_ortho(data, flat=True)
ants.plot_ortho(ants.from_nibabel(ants.to_nibabel(data)), flat=True)

[stnava]

probably resolved with 30b0038

please reopen if not