plotting errors with time-dependent parameters

[isaacsas]

For this example from the docs:

t = ModelingToolkit.t_nounits
D = ModelingToolkit.D_nounits
@variables x(t)
@parameters c(t)
@mtkbuild sys = ODESystem(
    D(x) ~ c * cos(x), t, [x], [c]; discrete_events = [1.0 => [c ~ c + 1]])
prob = ODEProblem(sys, [x => 0.0], (0.0, 2pi), [c => 1.0])
sol = solve(prob, Tsit5())

Plotting the parameter's time-series kind of works

plot(sol; idxs = sys.c) 

but note the x-axis starts at t = 1 instead of t = 0.

Trying to plot the variable errors

plot(sol; idxs = sys.x) 

gives

ERROR: Cannot access discrete variable at time 0.0 before initial save 1.0
Stacktrace:
  [1] error(s::String)
    @ Base ./error.jl:35
  [2] _hold_discrete
    @ ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:186 [inlined]
  [3] _broadcast_getindex_evalf
    @ ./broadcast.jl:709 [inlined]
  [4] _broadcast_getindex
    @ ./broadcast.jl:682 [inlined]
  [5] getindex
    @ ./broadcast.jl:636 [inlined]
  [6] macro expansion
    @ ./broadcast.jl:1004 [inlined]
  [7] macro expansion
    @ ./simdloop.jl:77 [inlined]
  [8] copyto!
    @ ./broadcast.jl:1003 [inlined]
  [9] copyto!
    @ ./broadcast.jl:956 [inlined]
 [10] copy
    @ ./broadcast.jl:928 [inlined]
 [11] materialize
    @ ./broadcast.jl:903 [inlined]
 [12] hold_discrete(disc_u::Vector{ModelingToolkit.NestedGetIndex{Tuple{…}}}, disc_t::Vector{Float64}, t::Vector{Float64})
    @ SciMLBase ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:197
 [13] #480
    @ ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:205 [inlined]
 [14] iterate
    @ ./generator.jl:47 [inlined]
 [15] collect(itr::Base.Generator{ParameterTimeseriesCollection{…}, SciMLBase.var"#480#481"{…}})
    @ Base ./array.jl:834
 [16] map
    @ ./abstractarray.jl:3313 [inlined]
 [17] get_interpolated_discretes(sol::ODESolution{…}, t::Vector{…}, deriv::Type, continuity::Symbol)
    @ SciMLBase ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:204
 [18] (::ODESolution{…})(t::Vector{…}, ::Type{…}, idxs::Vector{…}, continuity::Symbol)
    @ SciMLBase ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:340
 [19] #_#482
    @ ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:215 [inlined]
 [20] AbstractODESolution
    @ ~/.julia/packages/SciMLBase/xlQdY/src/solutions/ode_solutions.jl:210 [inlined]
 [21] solplot_vecs_and_labels(dims::Int64, vars::Vector{…}, plott::Vector{…}, sol::ODESolution{…}, plot_analytic::Bool, plot_analytic_timeseries::Nothing)
    @ SciMLBase ~/.julia/packages/SciMLBase/xlQdY/src/solutions/solution_interface.jl:506
 [22] diffeq_to_arrays(sol::ODESolution{…}, plot_analytic::Bool, denseplot::Bool, plotdensity::Int64, tspan::Nothing, vars::Vector{…}, tscale::Symbol, plotat::Nothing)
    @ SciMLBase ~/.julia/packages/SciMLBase/xlQdY/src/solutions/solution_interface.jl:400
 [23] macro expansion
    @ ~/.julia/packages/SciMLBase/xlQdY/src/solutions/solution_interface.jl:210 [inlined]
 [24] macro expansion
    @ ~/.julia/packages/RecipesBase/BRe07/src/RecipesBase.jl:340 [inlined]
 [25] macro expansion
    @ ~/.julia/packages/SciMLBase/xlQdY/src/solutions/solution_interface.jl:204 [inlined]
 [26] apply_recipe(plotattributes::AbstractDict{Symbol, Any}, sol::SciMLBase.AbstractTimeseriesSolution)
    @ SciMLBase ~/.julia/packages/RecipesBase/BRe07/src/RecipesBase.jl:300
 [27] _process_userrecipes!(plt::Any, plotattributes::Any, args::Any)
    @ RecipesPipeline ~/.julia/packages/RecipesPipeline/BGM3l/src/user_recipe.jl:38
 [28] recipe_pipeline!(plt::Any, plotattributes::Any, args::Any)
    @ RecipesPipeline ~/.julia/packages/RecipesPipeline/BGM3l/src/RecipesPipeline.jl:72
 [29] _plot!(plt::Plots.Plot, plotattributes::Any, args::Any)
    @ Plots ~/.julia/packages/Plots/kLeqV/src/plot.jl:223
 [30] plot(args::Any; kw...)
    @ Plots ~/.julia/packages/Plots/kLeqV/src/plot.jl:102
 [31] top-level scope
    @ REPL[63]:1
Some type information was truncated. Use `show(err)` to see complete types.

Pkg.status is:

  [479239e8] Catalyst v14.4.0 `~/.julia/dev/Catalyst`
  [ccbc3e58] JumpProcesses v9.13.5
  [961ee093] ModelingToolkit v9.35.1
  [1dea7af3] OrdinaryDiffEq v6.89.0
  [91a5bcdd] Plots v1.40.8
  [2efcf032] SymbolicIndexingInterface v0.3.28

[AayushSabharwal]

There are two parts to this issue:

• MTK's events don't save at t=0, so the initial value of the discrete is not known
• Interpolation (rightly) refuses to interpolate the value of a discrete variable before the initial time it was saved. However, it should not do this if the value being interpolated isn't discrete.

[isaacsas]

Thanks!