Possible regression when simplifying system to 0 states

[ggkountouras]

MTK v8.47.0

Result

MethodError: no method matching cleansyms(::Vector{Any}) due to the system getting simplified to 0 states.

Expected result

Precompute everything and leave at least one state for plotting and analysis.

# M(not)WE
using ModelingToolkit, Sundials, DifferentialEquations, Plots
using ModelingToolkitStandardLibrary.Electrical
using ModelingToolkitStandardLibrary.Blocks

function ShockleyDiode(;name, Is = 1.0e-15, eta = 1.0, Vt = 26.0e-3)
    @named oneport = OnePort()
    @unpack v, i = oneport
    ps = @parameters Is=Is eta=eta Vt=Vt
    eqs = [
           i ~ Is*(exp(v/Vt) - 1)
          ]
    extend(ODESystem(eqs, t, [], ps; name=name), oneport)
end

@variables t
@named Vin = Voltage()
@named Vin_V = Ramp(height=0.4, duration=6.0)
@named D = ShockleyDiode(Is=2.5e-10)
@named GND = Ground()
test_eqs = [
       connect(Vin_V.output, Vin.V)
       connect(Vin.p, D.p)
       connect(Vin.n, D.n, GND.g)
       ]
@named test_model = ODESystem(test_eqs, t, systems=[Vin, Vin_V, D, GND])
sys = structural_simplify(test_model)
du0 = fill(0.0, length(states(sys)))
u0 = fill(0.0, length(states(sys)))
prob = DAEProblem(sys, du0, u0, (0, 6.0))
sol = solve(prob, IDA())
plot(sol, idxs=[D.v, D.i])

[uwbanjoman]

I experienced the same issue with MTK v8.51.0 on Julia 1.9.0-beta4

[ChrisRackauckas]

@oscardssmith I think all of these cases were handled? Can you check this one and see if we need another test, or close this?

[oscardssmith]

I believe this can be closed