rename parameter up_angle to up_theta for consistency and check conversion from old models

[pkienzle]

Orientation angles are θ, φ, ψ most everywhere but magnetic orientation uses up_angle and up_phi for historical reasons. Rename up_angle to up_theta and make sure that older models get converted.

Verify that θ, φ use the same orientation system as the particle axis, both for applied field and for orientation of the magnetism within the particle.

Also, make sure that models saved without up_phi are loaded correctly in versions that support arbitrary applied field.

[dehoni]

In realspace.py, the rotation matrixes and the definition of orientation coordinates pol2rec are consistent with the coordinates for the polarisation vector p_hat in the function magnetic_sld .
I would need to guess the origin of the expressions for view and jitter matrices in kernel_iq.c, such that I can not state if they are in line. Similarly, the dedicated magnetic models in the model market place need an extra look, while doing the refactoring mentioned in ticket #450.

[pkienzle]

Related: sasview generic scattering calculator SasView/sasview#1971