Error when initializing iterator for SteadyStateProblem with DynamicSS and Dual numbers

[harisorgn]

MWE :

using ModelingToolkit
using DifferentialEquations
using Turing
using ForwardDiff
using ForwardDiff: Dual

@variables begin
    t
    y1(t) = 1.0
    y2(t) = 0.0
    y3(t) = 0.0
    y4(t) = 0.0
    y5(t) = 0.0
    y6(t) = 0.0
    y7(t) = 0.0
    y8(t) = 0.0057
end

@parameters begin
    k1 = 1.71
    k2 = 280.0
    k3 = 8.32
    k4 = 0.69
    k5 = 0.43
    k6 = 1.81
end

D = Differential(t)
eqs = [
    D(y1) ~ (-k1*y1 + k5*y2 + k3*y3 + 0.0007),
    D(y2) ~ (k1*y1 - 8.75*y2),
    D(y3) ~ (-10.03*y3 + k5*y4 + 0.035*y5),
    D(y4) ~ (k3*y2 + k1*y3 - 1.12*y4),
    D(y5) ~ (-1.745*y5 + k5*y6 + k5*y7),
    D(y6) ~ (-k2*y6*y8 + k4*y4 + k1*y5 - k5*y6 + k4*y7),
    D(y7) ~ (k2*y6*y8 - k6*y7),
    D(y8) ~ (-k2*y6*y8 + k6*y7)
]

@named model = ODESystem(eqs)

u0 = [
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(1.8983788068509213,0.6970981087506788,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.0057,0.0,0.0,0.0,0.0,0.0,0.0,0.0)
]

p = [
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(1.5174072564237708,0.0,0.38355212192378396,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.43,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(8.32,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(268.1197063467946,0.0,0.0,44.91823438292696,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(0.69,0.0,0.0,0.0,0.0,0.0,0.0,0.0),
    Dual{ForwardDiff.Tag{Turing.TuringTag, Float64}}(20.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0)
]

prob = SteadyStateProblem(model, u0, p)
alg = DynamicSS(QNDF())
int = init(prob, alg)

gives

ERROR: The chosen algorithm, SSRootfind{SteadyStateDiffEq.var"#2#4"}(SteadyStateDiffEq.var"#2#4"()), does not exist. Please verify that the appropriate solver package has been installed.

Top of stack trace is a call to

DifferentialEquations.jl/src/default_solve.jl

Lines 12 to 14 in c9632cc

	function DiffEqBase.__init(prob::DiffEqBase.DEProblem,
	alg::Union{Nothing, DiffEqBase.DEAlgorithm},
	args...; default_set = false, kwargs...)

This dispatch on alg::DiffEqBase.DEAlgorithm seems too wide and DynamicSS is treated like the recently deprecated SSRootfind?

Same behavior no matter what ODESolver I use inside DynamicSS (tried Tsit5 and TRBDF2).

cc @SebastianM-C

[ChrisRackauckas]

SSRootfind{SteadyStateDiffEq.var"#2#4"} Odd, SSRootfind shouldn't be involved at all. But I don't think init is defined on DynamicSS so that's probably the root cause here. It shouldn't be falling back to the default algorithm handling though, that's a bit odd so indeed it needs to be tightened.