Fix methods in SVD Toolbox

[dallan-keylogic]

Summary/Motivation:

Use PyNumero methods to get constraint/variable indices rather than list.index because (I think) list.index uses == to compare elements, which is not what should be done for Pyomo objects. Fixes #1285 .

Also add in string formatting for magnitude of Jacobian elements. Now it uses scientific notation with three digits past the decimal place. It's a lot better than having to deal with 16 digits.

Legal Acknowledgement

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

1. I agree my contributions are submitted under the license terms described in the LICENSE.txt file at the top level of this directory.
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 ✅

Comparison is base (214f32a) 77.41% compared to head (bb8f3ba) 77.40%.
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1288      +/-   ##
==========================================
- Coverage   77.41%   77.40%   -0.01%     
==========================================
  Files         390      390              
  Lines       63404    63399       -5     
  Branches    11658    11658              
==========================================
- Hits        49082    49074       -8     
- Misses      11793    11797       +4     
+ Partials     2529     2528       -1     

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

[andrewlee94]

I think you should make these private. There is no reason the user should be changing these.

[dallan-keylogic]

I don't think they should be changing them, but they might have the desire to look at them. We can still make them semiprivate, though.

[andrewlee94]

If they want to see them then they can access to the nlp object themselves, so I would tend toward hiding them until w have a good case otherwise (and then consider making them properties so they can't be changed). "Private" in Python is just convention anyways.

[andrewlee94]

Only one minor request.

[Robbybp]

Good fix. list.index is indeed not safe here.

[Robbybp]

What is the type of self.jacobian here? I.e. dense, COO, CSC, etc. I guess for a single variable, performance isn't an issue here.

[dallan-keylogic]

It's some sparse type, I forget whether it's row or column based. Performance probably isn't an issue, so this is probably unnecessary, but I'm familiar with the methods whereas it's unclear (at least to me) how indexing sparse matrices works.

[Robbybp]

getcol and indexing do the same thing AFAIK. So one of row/col is slow and one is fast depending on whether the matrix CSC or CSR. If the matrix is COO, both are slow.