Handle uninitialized variable in propagate_solution of scaling transformation

[Robbybp]

Fixes #3273, Fixes #2817

Changes proposed in this PR:

• If a variable has a value of None in the scaled model, we don't attempt to set its value in the original model

Legal Acknowledgement

By contributing to this software project, I have read the contribution guide and agree to the following terms and conditions for my contribution:

1. I agree my contributions are submitted under the BSD license.
2. I represent I am authorized to make the contributions and grant the license. If my employer has rights to intellectual property that includes these contributions, I represent that I have received permission to make contributions and grant the required license on behalf of that employer.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.56%. Comparing base (5206eee) to head (beb31f6).
Report is 286 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3275   +/-   ##
=======================================
  Coverage   88.56%   88.56%           
=======================================
  Files         877      877           
  Lines       99276    99282    +6     
=======================================
+ Hits        87920    87927    +7     
+ Misses      11356    11355    -1     

Flag	Coverage Δ
linux	86.10% <100.00%> (+<0.01%)	⬆️
osx	75.76% <100.00%> (+<0.01%)	⬆️
other	86.59% <100.00%> (+<0.01%)	⬆️
win	83.92% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[jsiirola]

Is the correct check here to test None, or to not transfer the values of any stale variables? I am wondering if the correct answer is to only propagate variable values when .stale==False

[Robbybp]

I guess I can see three options:

1. Propagate values in the model, unless they are None (what this PR currently does)
2. Propagate values in the model, regardless of what they are
3. Propagate non-stale values in the model

My instinct is that relying on the stale flag can lead to non-intuitive behavior, e.g. if a user sets a value after a solve (or solves via solve_strongly_connected_components), only the last values updated will be propagated. If this is what we decide on, the docstring should be updated to indicate that only the most recent solution is propagated.

[Robbybp]

Based on @dallan-keylogic's comment on #2817, I've updated this PR to propagate values from the scaled model even if they are None, but to log a warning if None overrides a non-None value.

[jsiirola]

I think I am OK with this behavior, but if we are always going to propagate the entire model state, I wonder if a better name for this method is propagate_state() (as we are propagating the current state of the model and not necessarily the last thing that came back from the solver).

[Robbybp]

I agree that propagate_state, or something like propagate_variable_values, would be a better name, although I'm not convinced it's worth changing.

[Robbybp]

This should fix #2817 as well.

[jsiirola]

I have a minor comment that would improve testing, but that shouldn't necessarily prevent merging.