Add metadata for internal parameters

[TorkelE]

So right now we have parameters that the user is aware of, and then stuff that MTK adds on its own accord. Basically the difference between what get_ps and `parameters return.

Would it make sense to have a metadata entry for "internal parameters" (with a value set to true). That was one could easily check if a parameter is relevant to the user or not. That way it might be easier to handle functions like parameters. I.e. right now if MTK adds a new internal parameter, unless I am told an updated Catalyst code internally we will get a problem (and wherever we take this into account). However, this way, when a new type of internal parameter is introduced the metadata could just be checked as true and, at least some stuff in SciML, might automatically take this into account.

[ChrisRackauckas]

The idea of an "internal" parameter isn't well defined? D(x) ~ 0 is that internal?

[TorkelE]

sorry, what is the parameter in that? In this case internal parameters to me are parameters stored in the get_ps field but which is not returned by parameters (but are returned by get_ps).

[ChrisRackauckas]

It's a state that a user defines that can become a parameter because it doesn't need to be solved by the ODE solver.

[TorkelE]

Agreed, that makes sense. Then I guess it depends on what MTK considers it in in parameters.