Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

format #1114

Merged
merged 1 commit into from
Feb 12, 2025
Merged

format #1114

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions ext/DiffEqBaseForwardDiffExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,8 +2,9 @@ module DiffEqBaseForwardDiffExt

using DiffEqBase, ForwardDiff
using DiffEqBase.ArrayInterface
using DiffEqBase: Void, FunctionWrappersWrappers, OrdinaryDiffEqTag, AbstractTimeseriesSolution,
RecursiveArrayTools, reduce_tup, _promote_tspan, has_continuous_callback
using DiffEqBase: Void, FunctionWrappersWrappers, OrdinaryDiffEqTag,
AbstractTimeseriesSolution,
RecursiveArrayTools, reduce_tup, _promote_tspan, has_continuous_callback
import DiffEqBase: hasdualpromote, wrapfun_oop, wrapfun_iip, prob2dtmin,
promote_tspan, anyeltypedual, isdualtype, value, ODE_DEFAULT_NORM,
InternalITP, nextfloat_tdir, DualEltypeChecker, sse
Expand Down Expand Up @@ -502,7 +503,8 @@ unitfulvalue(x::ForwardDiff.Dual) = unitfulvalue(ForwardDiff.unitfulvalue(x))

sse(x::ForwardDiff.Dual) = sse(ForwardDiff.value(x)) + sum(sse, ForwardDiff.partials(x))
function DiffEqBase.totallength(x::ForwardDiff.Dual)
return DiffEqBase.totallength(ForwardDiff.value(x)) + sum(DiffEqBase.totallength, ForwardDiff.partials(x))
return DiffEqBase.totallength(ForwardDiff.value(x)) +
sum(DiffEqBase.totallength, ForwardDiff.partials(x))
end

@inline ODE_DEFAULT_NORM(u::ForwardDiff.Dual, ::Any) = sqrt(sse(u))
Expand Down
2 changes: 1 addition & 1 deletion ext/DiffEqBaseGTPSAExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,4 +35,4 @@ function ODE_DEFAULT_NORM(f::F, u::AbstractArray{<:TPS}, t) where {F}
Base.FastMath.sqrt_fast(x / max(length(u), 1))
end

end
end
5 changes: 4 additions & 1 deletion ext/DiffEqBaseMonteCarloMeasurementsExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,10 @@ end

DiffEqBase.value(x::Type{MonteCarloMeasurements.AbstractParticles{T, N}}) where {T, N} = T
DiffEqBase.value(x::MonteCarloMeasurements.AbstractParticles) = mean(x.particles)
DiffEqBase.unitfulvalue(x::Type{MonteCarloMeasurements.AbstractParticles{T, N}}) where {T, N} = T
function DiffEqBase.unitfulvalue(x::Type{MonteCarloMeasurements.AbstractParticles{
T, N}}) where {T, N}
T
end
DiffEqBase.unitfulvalue(x::MonteCarloMeasurements.AbstractParticles) = mean(x.particles)

# Support adaptive steps should be errorless
Expand Down
1 change: 0 additions & 1 deletion ext/DiffEqBaseReverseDiffExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,7 +22,6 @@ end
DiffEqBase.value(x::ReverseDiff.TrackedReal) = x.value
DiffEqBase.value(x::ReverseDiff.TrackedArray) = x.value


DiffEqBase.unitfulvalue(x::Type{ReverseDiff.TrackedReal{V, D, O}}) where {V, D, O} = V
function DiffEqBase.unitfulvalue(x::Type{
ReverseDiff.TrackedArray{V, D, N, VA, DA},
Expand Down
4 changes: 3 additions & 1 deletion ext/DiffEqBaseTrackerExt.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,9 @@ DiffEqBase.value(x::Tracker.TrackedReal) = x.data
DiffEqBase.value(x::Tracker.TrackedArray) = x.data

DiffEqBase.unitfulvalue(x::Type{Tracker.TrackedReal{T}}) where {T} = T
DiffEqBase.unitfulvalue(x::Type{Tracker.TrackedArray{T, N, A}}) where {T, N, A} = Array{T, N}
function DiffEqBase.unitfulvalue(x::Type{Tracker.TrackedArray{T, N, A}}) where {T, N, A}
Array{T, N}
end
DiffEqBase.unitfulvalue(x::Tracker.TrackedReal) = x.data
DiffEqBase.unitfulvalue(x::Tracker.TrackedArray) = x.data

Expand Down
8 changes: 5 additions & 3 deletions src/norecompile.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,9 +23,11 @@ end

# Default dispatch assumes no ForwardDiff, gets added in the new dispatch
function wrapfun_iip(ff, inputs)
FunctionWrappersWrappers.FunctionWrappersWrapper(Void(ff), (typeof(inputs),), (Nothing,))
FunctionWrappersWrappers.FunctionWrappersWrapper(
Void(ff), (typeof(inputs),), (Nothing,))
end

function wrapfun_oop(ff, inputs)
FunctionWrappersWrappers.FunctionWrappersWrapper(ff, (typeof(inputs),), (typeof(inputs[1]),))
end
FunctionWrappersWrappers.FunctionWrappersWrapper(
ff, (typeof(inputs),), (typeof(inputs[1]),))
end
39 changes: 20 additions & 19 deletions test/downstream/gtpsa.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,68 +9,69 @@ x = [1.0, 2.0, 3.0]
p = [4.0, 5.0, 6.0]

prob = ODEProblem(f!, x, (0.0, 1.0), p)
sol = solve(prob, Tsit5(), reltol=1e-16, abstol=1e-16)
sol = solve(prob, Tsit5(), reltol = 1e-16, abstol = 1e-16)

# Parametric GTPSA map
desc = Descriptor(3, 2, 3, 2) # 3 variables 3 parameters, both to 2nd order
dx = @vars(desc)
dp = @params(desc)
prob_GTPSA = ODEProblem(f!, x .+ dx, (0.0, 1.0), p .+ dp)
sol_GTPSA = solve(prob_GTPSA, Tsit5(), reltol=1e-16, abstol=1e-16)
sol_GTPSA = solve(prob_GTPSA, Tsit5(), reltol = 1e-16, abstol = 1e-16)

@test sol.u[end] ≈ scalar.(sol_GTPSA.u[end]) # scalar gets 0th order part

# Compare Jacobian against ForwardDiff
J_FD = ForwardDiff.jacobian([x..., p...]) do t
prob = ODEProblem(f!, t[1:3], (0.0, 1.0), t[4:6])
sol = solve(prob, Tsit5(), reltol=1e-16, abstol=1e-16)
sol = solve(prob, Tsit5(), reltol = 1e-16, abstol = 1e-16)
sol.u[end]
end

@test J_FD ≈ GTPSA.jacobian(sol_GTPSA.u[end], include_params=true)
@test J_FD ≈ GTPSA.jacobian(sol_GTPSA.u[end], include_params = true)

# Compare Hessians against ForwardDiff
for i in 1:3
Hi_FD = ForwardDiff.hessian([x..., p...]) do t
prob = ODEProblem(f!, t[1:3], (0.0, 1.0), t[4:6])
sol = solve(prob, Tsit5(), reltol=1e-16, abstol=1e-16)
sol = solve(prob, Tsit5(), reltol = 1e-16, abstol = 1e-16)
sol.u[end][i]
end
@test Hi_FD ≈ GTPSA.hessian(sol_GTPSA.u[end][i], include_params=true)
@test Hi_FD ≈ GTPSA.hessian(sol_GTPSA.u[end][i], include_params = true)
end


# ODEProblem 2 =======================
pdot!(dq, p, q, params, t) = dq .= [0.0, 0.0, 0.0]
qdot!(dp, p, q, params, t) = dp .= [p[1] / sqrt((1 + p[3])^2 - p[1]^2 - p[2]^2),
p[2] / sqrt((1 + p[3])^2 - p[1]^2 - p[2]^2),
p[3] / sqrt(1 + p[3]^2) - (p[3] + 1)/sqrt((1 + p[3])^2 - p[1]^2 - p[2]^2)]
pdot!(dq, p, q, params, t) = dq .= [0.0, 0.0, 0.0]
function qdot!(dp, p, q, params, t)
dp .= [p[1] / sqrt((1 + p[3])^2 - p[1]^2 - p[2]^2),
p[2] / sqrt((1 + p[3])^2 - p[1]^2 - p[2]^2),
p[3] / sqrt(1 + p[3]^2) - (p[3] + 1) / sqrt((1 + p[3])^2 - p[1]^2 - p[2]^2)]
end

prob = DynamicalODEProblem(pdot!, qdot!, [0.0, 0.0, 0.0], [0.0, 0.0, 0.0], (0.0, 25.0))
sol = solve(prob, Yoshida6(), dt = 1.0, reltol=1e-16, abstol=1e-16)
sol = solve(prob, Yoshida6(), dt = 1.0, reltol = 1e-16, abstol = 1e-16)

desc = Descriptor(6, 2) # 6 variables to 2nd order
dx = @vars(desc) # identity map
dx = @vars(desc) # identity map
prob_GTPSA = DynamicalODEProblem(pdot!, qdot!, dx[1:3], dx[4:6], (0.0, 25.0))
sol_GTPSA = solve(prob_GTPSA, Yoshida6(), dt = 1.0, reltol=1e-16, abstol=1e-16)
sol_GTPSA = solve(prob_GTPSA, Yoshida6(), dt = 1.0, reltol = 1e-16, abstol = 1e-16)

@test sol.u[end] ≈ scalar.(sol_GTPSA.u[end]) # scalar gets 0th order part

# Compare Jacobian against ForwardDiff
J_FD = ForwardDiff.jacobian(zeros(6)) do t
prob = DynamicalODEProblem(pdot!, qdot!, t[1:3], t[4:6], (0.0, 25.0))
sol = solve(prob, Yoshida6(), dt = 1.0, reltol=1e-16, abstol=1e-16)
sol = solve(prob, Yoshida6(), dt = 1.0, reltol = 1e-16, abstol = 1e-16)
sol.u[end]
end

@test J_FD ≈ GTPSA.jacobian(sol_GTPSA.u[end], include_params=true)
@test J_FD ≈ GTPSA.jacobian(sol_GTPSA.u[end], include_params = true)

# Compare Hessians against ForwardDiff
for i in 1:6
Hi_FD = ForwardDiff.hessian(zeros(6)) do t
prob = DynamicalODEProblem(pdot!, qdot!, t[1:3], t[4:6], (0.0, 25.0))
sol = solve(prob, Yoshida6(), dt = 1.0, reltol=1e-16, abstol=1e-16)
prob = DynamicalODEProblem(pdot!, qdot!, t[1:3], t[4:6], (0.0, 25.0))
sol = solve(prob, Yoshida6(), dt = 1.0, reltol = 1e-16, abstol = 1e-16)
sol.u[end][i]
end
@test Hi_FD ≈ GTPSA.hessian(sol_GTPSA.u[end][i], include_params=true)
@test Hi_FD ≈ GTPSA.hessian(sol_GTPSA.u[end][i], include_params = true)
end
3 changes: 2 additions & 1 deletion test/forwarddiff_dual_detection.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -387,5 +387,6 @@ u = Dual.(val, par)
@test DiffEqBase.totallength(val[1]) == 1
@test DiffEqBase.totallength(val) == length(val)
@test DiffEqBase.totallength(par) == length(par)
@test DiffEqBase.totallength(u[1]) == DiffEqBase.totallength(val[1]) + DiffEqBase.totallength(par[1])
@test DiffEqBase.totallength(u[1]) ==
DiffEqBase.totallength(val[1]) + DiffEqBase.totallength(par[1])
@test DiffEqBase.totallength(u) == sum(DiffEqBase.totallength, u)
Loading