Add fallback option for loading data

[mdoucet]

Occasionally, and for older data, the current way of filtering data fails.
In this case, we case use another PV to do the filtering.

This PR also fixed a small bug where the reduction would fail if there was only one cross-section because it was assuming that the return value of the reduction Mantid algo was a iterable.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 15.00000% with 17 lines in your changes missing coverage. Please review.

Project coverage is 35.16%. Comparing base (d11eaee) to head (5855a66).

Files	Patch %	Lines
...flectivity_ui/interfaces/data_handling/data_set.py	0.00%	8 Missing ⚠️
...ectivity_ui/interfaces/data_handling/instrument.py	0.00%	6 Missing ⚠️
reflectivity_ui/interfaces/data_manager.py	50.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             next      #94      +/-   ##
==========================================
- Coverage   35.26%   35.16%   -0.11%     
==========================================
  Files          30       30              
  Lines        6085     6100      +15     
==========================================
- Hits         2146     2145       -1     
- Misses       3939     3955      +16     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[backmari]

This change allows loading a run where the sample log required by MRFilterCrossSections is missing due to a problem in the data acquisition. Before this change, loading the run number 42006 fails. After the change, loading run 42006 succeeds by instead filtering by the log BL4A:SF:ICP:getDI. There is a log message and a status bar message to warn the user about the different filtering used.
Other than my question/comment about the return when qmin is None, this looks good to me.