Enable functional/time dependent parameters

[TorkelE]

Adds docs and tests. Also modifies the main code to enable these to be used within Catalyst.

Example:

using Catalyst, DataInterpolations, OrdinaryDiffEqDefault, Plots

# Create input function/data
tend = 120.0
ts = collect(0.0:1.0:tend)
light = sin.(ts/6) .+ 1
light = [max(0.0, l - rand()) for l in light]
interpolated_light = LinearInterpolation(light, ts)
plot(interpolated_light)

# Create model
t = default_t()
in_type = typeof(interpolated_light)
@parameters kA kD (light_in::in_type)(..)
@species Pₐ(t) Pᵢ(t)
rxs = [
    Reaction(kA*light_in(t), [Pᵢ], [Pₐ]),
    Reaction(kD, [Pₐ], [Pᵢ])
]
@named rs = ReactionSystem(rxs, t)
rs = complete(rs)

# Simulate the model.
u0 = [Pᵢ => 1.0, Pₐ => 0.0]
ps = [kA => 1.5, kD => 1.0, light_in => interpolated_light]
oprob = ODEProblem(rs, u0, tend, ps)
sol = solve(oprob)
plot(sol)

[TorkelE]

This change is actually done in #1192 (but this PR is based off that one).

[TorkelE]

Builds on the change in #1192, however, but otherwise this is now ready.