Survey of upstream package usage in prep for SciML

[jlperla]

To lower maintenance requirements and upstream dependencies, we should verify:

• QuantEcon.jl See separate issue for details
• VegaLite and Query - reference ecosystem instead?
• Do we want to avoid having Zygote directly in the docs and instead just refer to it until reverse-mode AD consolidates a little? Point to Enzyme as well?
• DataInterpolations.jl and Interpolations.jl vs. whatever sciml has in the spring?

Key sciml packages as they currently are specified:

• https://github.com/SciML/Integrals.jl
• https://github.com/SciML/LinearSolve.jl
• https://github.com/SciML/Optimization.jl
• https://github.com/SciML/NonlinearSolve.jl
• Decide if DifferenceEquations.jl is appropriate

Also, in anticipation of #56 and others we should review the optimization and related sciml packages to see what is required for direct use, can be consolidated, etc.

• BlackBoxOptim.jl
• Optim.jl
• NLopt
• LeastSquaresOptim
• Ipopt
• NLSolve and Roots replaced with sciml's nonlinear solver
• IterativeSolvers/LinerOperators/LinearMaps/etc. and preconditioners with sciml's linear solver interface which might consolidate options?
• DiffEqOperators is being replaced with scimloperators/MOL/etc.
• Expectations.jl vs FastGaussQuadrature etc. usage. What about sciml's integrals.jl? Is gaussian quadrature in scope?