Using Extensions with EpiAware

[SamuelBrand1]

Since we added SciML deps to EpiAware the precompilation has been a bit more heavy. This is not surprising because even "low dep" usage of things like OrdinaryDiffEq loads a lot of (sub)packages which define individual solvers.

I'm not sure if going forwards:

• Each submodule of EpiAware ends up becoming its own pkg.
• We have a strategic usage of weakdeps and package extensions so that heavier duty bits of the code base are only loaded if the user wants to use ODE model (for example).

This is much more a discussion as opposed to an issue.

[seabbs]

Good point. I think the answer is both here