Open PallHaraldsson opened 11 months ago
It's not supported yet, but I do have the basic code in place. If I can muster up a days work for managing some slightly breaking changes and more comprehensive tests it should get done. I wish I could give a time estimate but I'm stuck in the middle of two papers and a thesis so my OSS contributions are pretty slow.
No pressure, I was just curious. I use neither. I suggest documenting only NIfTI-1 is supported (for now; in the README? this issue sort of does too), and that NIfTI-2 also exists and its support is on the TODO list.
What about both endian formats?
It detects whether the endian format a file is saved in and byte swaps accordingly. Users shouldn't really have to worry about that for NIfTI.
Some of the apprehension behind jumping into breaking changes for NIfTI-2 is that any software that supports it for standard MRIs also supports NIfTI-1 but the inverse is not true. v2 is also used for CIFTI in some cases. So it becomes a bit of a user interface nightmare if not done with a lot of thought and care.
I see from the source code you support NIfTI-1, but it's unclear if you support NIfTI-2 too, and both endian formats.
https://brainder.org/2015/04/03/the-nifti-2-file-format/
I think it should be documented either way. Do you know which is the most common format, and do you plan to support both?