Orientation Viewer doesn't work from nigthly build (on Mac)

[wpotrzebowski]

Describe the bug
Discovered while testing Mac installer but it may also occur on other platforms
Gives an error when trying to run:
07:41:09 - ERROR: Traceback (most recent call last):
File "sas/qtgui/MainWindow/GuiManager.py", line 1029, in actionOrientation_Viewer
File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 293, in show_orientation_viewer
File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 159, in init
File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 210, in on_angle_changed
File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 173, in _set_image_data
File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 261, in scatering_data
File "sasmodels/direct_model.py", line 375, in call
return self._calc_theory(pars, cutoff=self.cutoff)
File "sasmodels/direct_model.py", line 340, in _calc_theory
Iq_calc = call_kernel(self._kernel, pars, cutoff=cutoff)
File "sasmodels/direct_model.py", line 59, in call_kernel
mesh = get_mesh(calculator.info, pars, dim=calculator.dim, mono=mono)
File "sasmodels/direct_model.py", line 125, in get_mesh
raise TypeError(f"Unused parameters in call: {', '.join(values.keys())}")
TypeError: Unused parameters in call: a, b, c

To Reproduce
Steps to reproduce the behavior:

1. Download nightly build
2. Try to run "Orientation Viewer"

SasView version (please complete the following information):

• Version: nightly build

Operating system (please complete the following information):

• OS: Mac OSX

[lucas-wilkins]

hmmmm, OK, I'll check it on windows.

We still don't have a mac, but hopefully its just another bug associated with PySide

[lucas-wilkins]

although, it might be something to do with a change in the paracrystal model, if that has happened

[lucas-wilkins]

Yeah, it's caused by a change in sasmodels.

[lucas-wilkins]

This is the cause: SasView/sasmodels#563

It breaks things pretty hard. Not just an issue with parameter names.

[pkienzle]

The orientation viewer is using incorrect parameter names for parallelepiped. The new sasmodels kernel call correctly reports this.

The fix should be trivial: change orientation viewer calculator to use length_a, length_b, and length_c. There is also a typo in background in the original jitter.py code.

[pkienzle]

After fixing and merging the corrections to the orientation viewer the sasmodels PR SasView/sasmodels#575 can be merged.

[wpotrzebowski]

Orientation viewer doesn't work again (tested in nightly build)

07:40:52 - INFO:  --- SasView session started, version 6.0.0a1, 2023 ---
07:40:52 - INFO: Python: 3.8.17 (default, Jun  7 2023, 12:31:35) 
[Clang 13.0.0 (clang-1300.0.29.30)]
07:40:57 - ERROR: Traceback (most recent call last):
  File "sas/qtgui/MainWindow/GuiManager.py", line 1032, in actionOrientation_Viewer
  File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 293, in show_orientation_viewer
  File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 159, in __init__
  File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 210, in on_angle_changed
  File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 173, in _set_image_data
  File "sas/qtgui/Utilities/OrientationViewer/OrientationViewer.py", line 261, in scatering_data
  File "sasmodels/direct_model.py", line 377, in __call__
    return self._calc_theory(pars, cutoff=self.cutoff)
  File "sasmodels/direct_model.py", line 342, in _calc_theory
    Iq_calc = call_kernel(self._kernel, pars, cutoff=cutoff)
  File "sasmodels/direct_model.py", line 59, in call_kernel
    mesh = get_mesh(calculator.info, pars, dim=calculator.dim, mono=mono)
  File "sasmodels/direct_model.py", line 126, in get_mesh
    raise TypeError(f"Unused parameters in call: {', '.join(values.keys())}")
TypeError: Unused parameters in call: a, b, c

[lucas-wilkins]

Weird, I though I'd tested the change. I might have tested it with the parameter names changed. Will try it from the linked branch in a bit.

[lucas-wilkins]

Yeah, it's broken again