preallocate stats

[dpo]

The objective of this PR is to preallocate the contents of the stats so they can be passed in to a solver for a re-solve, and so we can say something like stats.solution .= x to update the solution.

cf: #3
cc @sshin23

[codecov]

Codecov Report

Merging #25 (080fa9f) into main (b93373d) will decrease coverage by 0.21%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      #25      +/-   ##
==========================================
- Coverage   88.67%   88.46%   -0.22%     
==========================================
  Files           3        3              
  Lines         106      104       -2     
==========================================
- Hits           94       92       -2     
  Misses         12       12              

Impacted Files	Coverage Δ
src/stats.jl	88.23% <100.00%> (-0.28%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b93373d...080fa9f. Read the comment docs.

[github-actions]

Package name	latest	stable
CaNNOLeS.jl
DCI.jl
DerivativeFreeSolvers.jl
JSOSolvers.jl
NLPModelsIpopt.jl
Percival.jl
RipQP.jl
SolverTest.jl

[github-actions]

Package name	latest	stable
CaNNOLeS.jl
DCI.jl
DerivativeFreeSolvers.jl
JSOSolvers.jl
NLPModelsIpopt.jl
Percival.jl
RipQP.jl
SolverTest.jl

[tmigot]

Looks good to me, thanks!

[abelsiqueira]

Thanks for the PR.
Should we have a flag to let the user know if the solution and multipliers were touched? Along with a reset! method.

[github-actions]

Package name	latest	stable
CaNNOLeS.jl
DCI.jl
DerivativeFreeSolvers.jl
JSOSolvers.jl
NLPModelsIpopt.jl
Percival.jl
RipQP.jl
SolverTest.jl

[dpo]

Should we have a flag to let the user know if the solution and multipliers were touched? Along with a reset! method.

What would reset! reset?

[geoffroyleconte]

Should we have a flag to let the user know if the solution and multipliers were touched? Along with a reset! method.

What would reset! reset?

The reset! method is a good idea! Maybe it could reset the flag, elapsed_time, iter and counters? And maybe empty solver_specific?

[dpo]

I'm not sure I see the point of resetting the stats structure. Either it comes out a solve, and in that case it holds the solve info, or it's passed to solve!(), and in that case the solver will overwrite all of it. The counters are in the model, and those can be reset.

I don't see how to reset solver specific stuff, since only the solver knows what to do with that info.

[dpo]

Thanks guys!

[abelsiqueira]

The flags are useful in particular for the contained solvers that don't touch the multipliers. That means that the solver needs to let the user know that the output.multipliers is not reliable.
The reset! would set the flags to their initial state.

[dpo]

In what situation would that occur? Presumably, the user knows if they're solving a constrained or unconstrained problem.

[abelsiqueira]

The solver is for constrained optimization but does not give an estimate of the multipliers because it uses something different.

[dpo]

Ok, like TRON or the nonsmooth solvers.I'll add that.