Beta approximation not working correctly

[butlerpd]

Describe the bug
The beta approximation no longer produces the beta or S_eff subplots. This suggests beta may not in fact be being applied. A few quick tests support that hypothesis, though better test data would be required to prove it. The priority depends on which is the case. I will set the priority as critical under the assumption that it is in fact not actually calculating beta even though the GUI suggests to the user it is doing so. If it is just the plots not being added to the explorer I would rate the severity as moderate.

To Reproduce
Steps to reproduce the behavior:

1. Load some data (2% hSDS for example) and send to fitting
2. choose sphere and HMSA structure factor and set parameters as given in tutorial (https://www.sasview.org/downloads/basic_1d_fitting_in_sasview_v5.pdf)
3. hit fit then check the subplots under the data set in the explorer
4. There will only be S, P and residuals. S_eff and beta will be absent

Expected behavior
In 5.0.6 once fit, the data file will have 5 subplots rather than just two. The two missing from 6.0.0 are the beta and S_eff plots.

SasView version (please complete the following information):

• Version: 6.0.0b2

Operating system (please complete the following information):

• OS: both windows and mac

Additional context
Reported as part of the NIST CNR summer school bugs

[lucas-wilkins]

Hmmm, I think diagnosing it would be much easier if the plots worked!

[lucas-wilkins]

@butlerpd you might be right that its not calculating it. I need to do more testing

[lucas-wilkins]

It definitely seems like its not being calculated. There's no difference in plotted output or in intermediate_results coming from the plots between PS and P(1 + beta(S-1)) for either polydisperse or asymmetric tests.

[lucas-wilkins]

Somewhere at the FittingThread level or below, plotting should automatically work if the data is there.

[smk78]

Retested using https://github.com/SasView/sasview/actions/runs/10638002500 on W10/x64.
The issue described above appears to be fixed and the beta(Q) and S_eff(Q) theories and plots are now being produced again when structure_factor_mode=P*(1+beta*(S-1)). Also checked that S(Q) and S_eff(Q) are in fact different using the cylinder model. They are!