Finally get right Orientation from raw data

[chaithyagr]

This is continuation of saga in #157 which was based on twixtools, which was not working well to read data.
Finally we have this working.

Some notes for how I got this working:

The base notes of how quaternion is ordered in data: https://github.com/ismrmrd/siemens_to_ismrmrd/blob/993ff5ffa74b4f719af36264e98698b05683498b/main.cpp#L1131-L1142

Which calls :
https://github.com/ismrmrd/ismrmrd/blob/802b52d69dcc31935888d3c69cc2a179f13a5608/libsrc/ismrmrd.c#L668-L683

However notice that call is Phase Read Slice, while the function expects Read Phase Slice. This mis-match is as data is expected to be Y X Z co-ordinates, cause I dont know why (@paquiteau has some links xD)

Based on this I almost got it working except sometimes the brain was inverted, and I found: https://github.com/ismrmrd/ismrmrd/blob/802b52d69dcc31935888d3c69cc2a179f13a5608/libsrc/ismrmrd.c#L578-L610

Where you try to change orientation based on if determinant of rotation matrix is negative (to keep it 1)

[paquiteau]

Nice job !

How do you want to use the orientation attribute later on ?

also, it may be nice to use:

https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.transform.Rotation.html

for the computation (that way I won't ask for a unit test :P )

[chaithyagr]

also, it may be nice to use:

https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.transform.Rotation.html

for the computation (that way I won't ask for a unit test :P )

I really wish I could, I tried a lot and could not get the 2 to match ...

[chaithyagr]

I think its because of all the intricacies like having it as Y X Z coordinates etc. Lemme see if I can fix it though