SciMLLogging Integration

[jClugstor]

Checklist

• Appropriate tests were added
• Any code changes were done in a way that does not break public API
• All documentation related to code changes were updated
• The new code follows the
  contributor guidelines, in particular the SciML Style Guide and
  COLPRAC.
• Any new documentation only uses public API

Additional context

Add any other context about the problem here.

[SebastianM-C]

Would progress information also go through this?

[jClugstor]

It definitely could, yeah.

[jClugstor]

@SebastianM-C can you make sure this looks okay? Also, let me know if there are any other log messages you would like to add that I haven't yet. Thanks!

[SebastianM-C]

The looks good overall. Can you take a look at OptimzationIpopt? It looks like that was not update. That has an integer verbosity that Ipopt interprets, but I'm not sure how do we translate that, since that's specific to Ipopt. I saw that you translate the verbose for MadNLP and I was wondering how do we encode the custom log levels of external packages in general.

[jClugstor]

Thanks, I'll take a look at OptimizationIpopt.

For packages with integer / bool interfaces for verbosity, in general I've been adding a verbosity toggle for those in to to the verbosity specifier, and then using a utility function to translate them in to a Bool / Int. So for example, in LinearSolve.jl, it can use HYPRE, which uses integers to specify the verbosity levels. So I added a verbosity toggle "HYPRE_verbosity", that is then translated in to an Int that HYPRE uses: https://github.com/SciML/LinearSolve.jl/blob/b88069a4f245cf2bc421c174b6fbc8948f4a4b6f/ext/LinearSolveHYPREExt.jl#L193 . The SciMLLogging docs have the mapping between the message levels and integers, and if you use a CustomLevel, the utility function automatically uses the integer inside of the CustomLevel.

[jClugstor]

@SebastianM-C I added a setting for Ipopt verbosity, let me know if that looks good.

[ChrisRackauckas]

OptimizationBase needs a major on this as it's breaking?

[jClugstor]

Yes I believe so, since this adds a field to OptimizationCache

[ChrisRackauckas]

Okay so make the bump and bump the lower bounds and we'll release