Fix: mosek_direct updated to use putqconk instead of putqcon

[Utkarsh-Detha]

Fixes

This fix concerns QCQP models when solved using mosek. In MOSEK's Optimizer API, the putqcon method resets the Q matrix entries for all constraints to zero, while putqconk does so only for k-th constraint. The _add_constraints method in mosek_direct would call putqcon, but this would lead to loss of Q info with every subsequent call to the _add_constraints (if new Q info was given). This is now fixed, because the Q matrix in each constraint is updated in its own call to putqconk.

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.

[Utkarsh-Detha]

Could someone help me figure out how to get this merged? Maybe @jsiirola? Thank you.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 88.39%. Comparing base (c849905) to head (9376b6b).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3199      +/-   ##
==========================================
- Coverage   88.40%   88.39%   -0.01%     
==========================================
  Files         846      846              
  Lines       95156    95153       -3     
==========================================
- Hits        84118    84107      -11     
- Misses      11038    11046       +8     

Flag	Coverage Δ
linux	86.33% <0.00%> (+<0.01%)	⬆️
osx	76.17% <0.00%> (+<0.01%)	⬆️
other	86.53% <0.00%> (-0.01%)	⬇️
win	83.82% <0.00%> (-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.

[mrmundt]

@Utkarsh-Detha - This looks fine to me. Our guidelines say that we need two approvals, and most of the Pyomo team is unfortunately on travel this week, so I think the earliest I can get someone else to look over it is Thursday. Apologies for the delay!

[jsiirola]

This looks fine - One question: is there a significant performance difference between putqcon and putqconk? If so, is it worth tracking if this is the first time add_constraints() was called for a model (allowing the interface to use putqcon the first time and putqconk for all subsequent calls)?

[Utkarsh-Detha]

Good question/suggestion, @jsiirola! My expectation was that for large problems putqcon should be faster than using a for-loop with putqconk. However, surprisingly enough, passing an n-by-n (dense) Q matrix in each constraint in an n-constraint problem is faster when using a for-loop with putqconk!

We are looking at as to why this is the case in Optimizer API. In the meantime, as far as mosek_direct is concerned, I think the incoming changes already implement the fastest way to do things!