Add docstring and oop array construction method for build_explicit_observed_function

[BenChung]

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.

[ChrisRackauckas]

Can we translate this to the standard? https://github.com/SciML/SciMLStyle?tab=readme-ov-file#documentation we should update that to use ## for the sections but still, the point being that it documents it by arg is a bit easier to read. I don't quite understand how it's written here.

[BenChung]

This one is tricky because it's substantially specifying the signature of a method it returns, rather than the method itself. I'm not sure where in the style guide it talks about documenting the argument and return types of returned function values? This is also an issue for SII, too, so it would be worthwhile coming up with a standard.

[ChrisRackauckas]

why would it be deprecated? How else would you control the rgf return serialization?

[BenChung]

I don't know; it's not actually used for anything in the function so far as I can tell.

That statement comes from Fredrik:

drop_expr is no longer needed, it was used to control whether or not the Expr was included in the generated function or not, for debugging purposes

[ChrisRackauckas]

It's fine to just always require RuntimeGeneratedFunctions.get_expression yes, but then if this argument is never documented or used, why mention it in the docstring at all? Just remove it?

[BenChung]

I think that when I wrote this I was just trying to document each argument... yeah, it should probably just be deleted. It looks like the only other use is

ModelingToolkit.jl/test/input_output_handling.jl

Line 144 in 9cf859f

drop_expr = identity)

, which is doing exactly nothing.

[BenChung]

@ChrisRackauckas What would help a lot is if there's an example/a standard for how to document first-class functions, since that's ultimately what's being described here (albeit sort of in a matrix of the argument configuration to the signature of the output function). I couldn't find it in the style guide you linked, but I might just be missing something.

The proximal problem is that the function (and the functions it returns) behaves very differently depending on the arguments and the system that it's generating an observed function for. It can return one or two functions, and the signatures of the returned functions can have between 0 and 5 arguments depending on a bunch of different properties.

[ChrisRackauckas]

Rebase