Overhaul solver member access via functions

[apozharski]

The goal of this PR is to attempt to minimize the number of structure dependent solver.* calls in core MadNLP.
The idea here is it easier in principle to create alternate subclasses of AbstractMadNLPSolver that only need to specialize some or all of the API found in IPM/api.jl and in doing so can make maximum use of the existing code in MadNLP with minimal required duplication.

I suspect there is more to be done in this direction, e.g. wrapping more "calculated" quantities over the current solver state, into functions _quantity(solver::AbstractMadNLPSolver) = <expression>, but I believe this is a good first pass.

(Note, that I am not married to the naming scheme etc., of this PR, if you would prefer set_<field>(solver::AbstractMadNLPSolver) and get_<field>(solver::AbstractMadNLPSolver), I would be happy with that as well, however due to shadowing pain I think just <field>(solver::AbstractMadNLPSolver) is not the best idea.)

[codecov]

Codecov Report

❌ Patch coverage is 90.08000% with 62 lines in your changes missing coverage. Please review.
✅ Project coverage is 84.56%. Comparing base (8a2fee6) to head (287dd4d).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
src/IPM/solver.jl	86.03%	37 Missing ⚠️
src/IPM/line_search.jl	78.20%	17 Missing ⚠️
src/IPM/types.jl	68.75%	5 Missing ⚠️
src/IPM/api.jl	62.50%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #497      +/-   ##
==========================================
- Coverage   84.63%   84.56%   -0.07%     
==========================================
  Files          49       51       +2     
  Lines        4400     4413      +13     
==========================================
+ Hits         3724     3732       +8     
- Misses        676      681       +5     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[apozharski]

Just a note: I cannot reproduce the test-madnlp failure locally on my machine, which is odd.

[sshin23]

@aspozharski could be due to the new release of julia v1.12

[frapac]

I like the direction where this is going. Having getters would also help with the C interface.

I would suggest generating all the getters automatically using metaprogramming. E.g. using the following code:

for (k, attribute) in enumerate(fieldnames(MadNLPSolver))
    fname = "get_$(attribute)"
    @eval begin
        function $(Symbol(fname))(solver::MadNLPSolver)
            return getfield(solver, $k)
        end
    end
end

[apozharski]

While I agree that the metaprogramming is nice from a maintenance perspective I always have found it to be a pain for future developers because things are not explicitly e.g. greppable.

[apozharski]

@aspozharski could be due to the new release of julia v1.12

Yep indeed there were some issues from the v1.12 release. #498 should fix these.