Enable symbolic indexing for PEtab results?

[TorkelE]

Just writing stuff here as I get ideas, feel free to close/consolidate to single issues if easier to handle that way.

It could be neat to permit accessing a parameters (optimal) value from a PEtab result like:

res = calibrate_model_multistart(petab_problem, ...)
res[p1]

to get its value.

Long-term, if rerouting through OptimizationProblems this would be handled automatically. However, it seems unlikely that SciML/Optimization will have full support for desired stuff (e.g. saving traces for plotting) any time soon.

[sebapersson]

I think this would be great. Will have it on the TODO for v3.0.

[sebapersson]

I wonder if we need this anymore. Now PEtabOptmizationResult as well as PEtabMultistartResult output the result as a ComponentArray. This means that fitted parameters can be accessed as:

xmin = res.xmin
k2 = xmin[:k2] # or xmin.k2

What do you think? Maybe it is better to rather add a get_xmin function (where the user can also choose which scale to convert the parameters to)