Merge pull request #640 from SciML/downstream

Fix downstream tests for MTK v9

Author: ChrisRackauckas

test/downstream/ensemble_multi_prob.jl

@@ -1,14 +1,13 @@
 using ModelingToolkit, OrdinaryDiffEq, Test
+using ModelingToolkit: t_nounits as t, D_nounits as D
+@variables x(t), y(t)
 
-@variables t, x(t), y(t)
-D = Differential(t)
-
-@named sys1 = ODESystem([D(x) ~ x,
-    D(y) ~ -y])
-@named sys2 = ODESystem([D(x) ~ 2x,
-    D(y) ~ -2y])
-@named sys3 = ODESystem([D(x) ~ 3x,
-    D(y) ~ -3y])
+@mtkbuild sys1 = ODESystem([D(x) ~ x,
+        D(y) ~ -y], t)
+@mtkbuild sys2 = ODESystem([D(x) ~ 2x,
+        D(y) ~ -2y], t)
+@mtkbuild sys3 = ODESystem([D(x) ~ 3x,
+        D(y) ~ -3y], t)
 
 prob1 = ODEProblem(sys1, [1.0, 1.0], (0.0, 1.0))
 prob2 = ODEProblem(sys2, [2.0, 2.0], (0.0, 1.0))

test/downstream/integrator_indexing.jl

@@ -1,16 +1,15 @@
 using ModelingToolkit, OrdinaryDiffEq, RecursiveArrayTools, StochasticDiffEq,
       SymbolicIndexingInterface, Test
-
+using ModelingToolkit: t_nounits as t, D_nounits as D
 ### Tests on non-layered model (everything should work). ###
 
-@parameters t a b c d
+@parameters a b c d
 @variables s1(t) s2(t)
-D = Differential(t)
 
 eqs = [D(s1) ~ a * s1 / (1 + s1 + s2) - b * s1,
     D(s2) ~ +c * s2 / (1 + s1 + s2) - d * s2]
 
-@named population_model = ODESystem(eqs)
+@named population_model = ODESystem(eqs, t)
 
 # Tests on ODEProblem.
 u0 = [s1 => 2.0, s2 => 1.0]
@@ -119,23 +118,24 @@ integrator[noisy_population_model.s2] = 10.0
 integrator[:s1] = 1.0
 @test integrator[s1] == integrator[noisy_population_model.s1] == integrator[:s1] == 1.0
 
-@parameters t σ ρ β
+@parameters σ ρ β
 @variables x(t) y(t) z(t)
 D = Differential(t)
 
 eqs = [D(x) ~ σ * (y - x),
     D(y) ~ x * (ρ - z) - y,
     D(z) ~ x * y - β * z]
 
-@named lorenz1 = ODESystem(eqs)
-@named lorenz2 = ODESystem(eqs)
+@named lorenz1 = ODESystem(eqs, t)
+@named lorenz2 = ODESystem(eqs, t)
 
 @parameters γ
 @variables a(t) α(t)
 connections = [0 ~ lorenz1.x + lorenz2.y + a * γ,
     α ~ 2lorenz1.x + a * γ]
-@named sys = ODESystem(connections, t, [a, α], [γ], systems = [lorenz1, lorenz2])
-sys_simplified = structural_simplify(sys)
+@mtkbuild sys_simplified = ODESystem(
+    connections, t, [a, α], [γ], systems = [lorenz1, lorenz2])
+sys_simplified = complete(structural_simplify(sys))
 
 u0 = [lorenz1.x => 1.0,
     lorenz1.y => 0.0,
@@ -185,7 +185,7 @@ step!(integrator, 100.0, true)
 eqs = [D(q[1]) ~ 2q[1]
        D(q[2]) ~ 2.0]
 @named sys2 = ODESystem(eqs, t, [q...], [])
-sys2_simplified = structural_simplify(sys2)
+sys2_simplified = complete(structural_simplify(sys2))
 prob2 = ODEProblem(sys2, [], (0.0, 5.0))
 integrator2 = init(prob2, Tsit5())
 
@@ -198,7 +198,7 @@ integrator2 = init(prob2, Tsit5())
     @variables u(t)
     eqs = [D(u) ~ u]
 
-    @named sys2 = ODESystem(eqs)
+    @mtkbuild sys2 = ODESystem(eqs, t)
 
     tspan = (0.0, 5.0)
 
@@ -327,13 +327,12 @@ plot(sol,idxs=(t,α))
 =#
 
 using LinearAlgebra
-@variables t
 sts = @variables x(t)[1:3]=[1, 2, 3.0] y(t)=1.0
 ps = @parameters p[1:3] = [1, 2, 3]
 D = Differential(t)
 eqs = [collect(D.(x) .~ x)
        D(y) ~ norm(x) * y - x[1]]
-@named sys = ODESystem(eqs, t, [sts...;], [ps...;])
+@mtkbuild sys = ODESystem(eqs, t, [sts...;], [ps...;])
 prob = ODEProblem(sys, [], (0, 1.0))
 integrator = init(prob, Tsit5(), save_everystep = false)
 @test integrator[x] isa Vector{Float64}

test/downstream/problem_interface.jl

@@ -1,17 +1,15 @@
 using ModelingToolkit, OrdinaryDiffEq, Test
+using ModelingToolkit: t_nounits as t, D_nounits as D
 using SymbolicIndexingInterface
 
 @parameters σ ρ β
-@variables t x(t) y(t) z(t)
-D = Differential(t)
+@variables x(t) y(t) z(t)
 
 eqs = [D(D(x)) ~ σ * (y - x),
     D(y) ~ x * (ρ - z) - y,
     D(z) ~ x * y - β * z]
 
-@named sys = ODESystem(eqs)
-
-sys = structural_simplify(sys)
+@mtkbuild sys = ODESystem(eqs, t)
 
 u0 = [D(x) => 2.0,
     x => 1.0,
@@ -181,13 +179,12 @@ set_tuple!(sprob, [10.0, 10.0])
 @test get_tuple(sprob) == (10.0, 10.0)
 
 using LinearAlgebra
-@variables t
 sts = @variables x(t)[1:3]=[1, 2, 3.0] y(t)=1.0
 ps = @parameters p[1:3] = [1, 2, 3]
 D = Differential(t)
 eqs = [collect(D.(x) .~ x)
        D(y) ~ norm(x) * y - x[1]]
-@named sys = ODESystem(eqs, t, [sts...;], [ps...;])
+@mtkbuild sys = ODESystem(eqs, t, [sts...;], [ps...;])
 prob = ODEProblem(sys, [], (0, 1.0))
 @test getp(sys, p)(prob) == prob.ps[p] == [1, 2, 3]
 setp(sys, p)(prob, [4, 5, 6])

test/downstream/remake_autodiff.jl

@@ -1,6 +1,7 @@
 using OrdinaryDiffEq, ModelingToolkit, Zygote, SciMLSensitivity
+using ModelingToolkit: t_nounits as t, D_nounits as D
 
-@variables t x(t) o(t)
+@variables x(t) o(t)
 D = Differential(t)
 function lotka_volterra(; name = name)
     unknowns = @variables x(t)=1.0 y(t)=1.0 o(t)

test/downstream/solution_interface.jl

@@ -1,18 +1,15 @@
 using ModelingToolkit, OrdinaryDiffEq, RecursiveArrayTools, StochasticDiffEq, Test
-# compat for MTKv8 and v9
-unknowns = isdefined(ModelingToolkit, :states) ? ModelingToolkit.states :
-           ModelingToolkit.unknowns
+using ModelingToolkit: t_nounits as t, D_nounits as D
 
 ### Tests on non-layered model (everything should work). ###
 
-@parameters t a b c d
+@parameters a b c d
 @variables s1(t) s2(t)
-D = Differential(t)
 
 eqs = [D(s1) ~ a * s1 / (1 + s1 + s2) - b * s1,
     D(s2) ~ +c * s2 / (1 + s1 + s2) - d * s2]
 
-@named population_model = ODESystem(eqs)
+@mtkbuild population_model = ODESystem(eqs, t)
 
 # Tests on ODEProblem.
 u0 = [s1 => 2.0, s2 => 1.0]
@@ -32,7 +29,7 @@ sol = solve(oprob, Rodas4())
 noiseeqs = [0.1 * s1,
     0.1 * s2]
 @named noisy_population_model = SDESystem(population_model, noiseeqs)
-sprob = SDEProblem(noisy_population_model, u0, (0.0, 100.0), p)
+sprob = SDEProblem(complete(noisy_population_model), u0, (0.0, 100.0), p)
 sol = solve(sprob, ImplicitEM())
 
 @test sol[s1] == sol[noisy_population_model.s1] == sol[:s1]
@@ -42,23 +39,22 @@ sol = solve(sprob, ImplicitEM())
 @test_throws Exception sol[:a]
 ### Tests on layered model (some things should not work). ###
 
-@parameters t σ ρ β
+@parameters σ ρ β
 @variables x(t) y(t) z(t)
-D = Differential(t)
 
 eqs = [D(x) ~ σ * (y - x),
     D(y) ~ x * (ρ - z) - y,
     D(z) ~ x * y - β * z]
 
-@named lorenz1 = ODESystem(eqs)
-@named lorenz2 = ODESystem(eqs)
+@named lorenz1 = ODESystem(eqs,t)
+@named lorenz2 = ODESystem(eqs,t)
 
 @parameters γ
 @variables a(t) α(t)
 connections = [0 ~ lorenz1.x + lorenz2.y + a * γ,
     α ~ 2lorenz1.x + a * γ]
 @named sys = ODESystem(connections, t, [a, α], [γ], systems = [lorenz1, lorenz2])
-sys_simplified = structural_simplify(sys)
+sys_simplified = complete(structural_simplify(sys))
 
 u0 = [lorenz1.x => 1.0,
     lorenz1.y => 0.0,

test/downstream/symbol_indexing.jl

@@ -1,10 +1,8 @@
 using ModelingToolkit, OrdinaryDiffEq, RecursiveArrayTools, SymbolicIndexingInterface, Test
 using Optimization, OptimizationOptimJL
-# compat for MTKv8 and v9
-unknowns = isdefined(ModelingToolkit, :states) ? ModelingToolkit.states :
-           ModelingToolkit.unknowns
+using ModelingToolkit: t_nounits as t, D_nounits as D
 
-@parameters t σ ρ β
+@parameters σ ρ β
 @variables x(t) y(t) z(t)
 D = Differential(t)
 
@@ -19,8 +17,7 @@ eqs = [D(x) ~ σ * (y - x),
 @variables a(t) α(t)
 connections = [0 ~ lorenz1.x + lorenz2.y + a * γ,
     α ~ 2lorenz1.x + a * γ]
-@named sys = ODESystem(connections, t, [a, α], [γ], systems = [lorenz1, lorenz2])
-sys_simplified = structural_simplify(sys)
+@mtkbuild sys = ODESystem(connections, t, [a, α], [γ], systems = [lorenz1, lorenz2])
 
 u0 = [lorenz1.x => 1.0,
     lorenz1.y => 0.0,
@@ -39,7 +36,7 @@ p = [lorenz1.σ => 10.0,
     γ => 2.0]
 
 tspan = (0.0, 100.0)
-prob = ODEProblem(sys_simplified, u0, tspan, p)
+prob = ODEProblem(sys, u0, tspan, p)
 integ = init(prob, Rodas4())
 sol = solve(prob, Rodas4())
 
@@ -135,7 +132,7 @@ sol1 = sol(0.0:1.0:10.0)
 
 sol2 = sol(0.1)
 @test sol2 isa Vector
-@test length(sol2) == length(unknowns(sys_simplified))
+@test length(sol2) == length(unknowns(sys))
 @test first(sol2) isa Real
 
 sol3 = sol(0.0:1.0:10.0, idxs = [lorenz1.x, lorenz2.x])
@@ -191,9 +188,9 @@ sol10 = sol(0.1, idxs = 2)
 @test sol10 isa Real
 
 @test is_timeseries(sol) == Timeseries()
-getx = getu(sys_simplified, lorenz1.x)
-get_arr = getu(sys_simplified, [lorenz1.x, lorenz2.x])
-get_tuple = getu(sys_simplified, (lorenz1.x, lorenz2.x))
+getx = getu(sys, lorenz1.x)
+get_arr = getu(sys, [lorenz1.x, lorenz2.x])
+get_tuple = getu(sys, (lorenz1.x, lorenz2.x))
 get_obs = getu(sol, lorenz1.x + lorenz2.x) # can't use sys for observed
 get_obs_arr = getu(sol, [lorenz1.x + lorenz2.x, lorenz1.y + lorenz2.y])
 l1x_idx = variable_index(sol, lorenz1.x)
@@ -218,7 +215,6 @@ plot(sol,idxs=(t,α))
 =#
 
 using LinearAlgebra
-@variables t
 sts = @variables x(t)[1:3]=[1, 2, 3.0] y(t)=1.0
 ps = @parameters p[1:3] = [1, 2, 3]
 D = Differential(t)
@@ -342,7 +338,7 @@ for (sym, oldval, newval, check_inference) in [
 end
 
 # accessing parameters
-@variables t x(t)
+@variables x(t)
 @parameters tau
 D = Differential(t)