Giter Club home page Giter Club logo

Comments (3)

ChrisRackauckas avatar ChrisRackauckas commented on June 10, 2024

I pushed it along and got pretty far:

using DynamicQuantities, OrdinaryDiffEq, RecursiveArrayTools

function RecursiveArrayTools.recursive_unitless_bottom_eltype(a::Type{
                                                                      <:DynamicQuantities.Quantity{T}
                                                                      }) where T
    T
end

function RecursiveArrayTools.recursive_unitless_eltype(a::Type{<:DynamicQuantities.Quantity{T}}) where T
    T
end

DiffEqBase.value(x::DynamicQuantities.Quantity) = x.value
@inline function DiffEqBase.UNITLESS_ABS2(x::AbstractArray)
    mapreduce(DiffEqBase.UNITLESS_ABS2, DiffEqBase.abs2_and_sum, x, init = zero(real(first(DiffEqBase.value(x)))))
end
@inline function DiffEqBase.UNITLESS_ABS2(x::DynamicQuantities.Quantity)
    abs(DiffEqBase.value(x))
end

function DiffEqBase.abs2_and_sum(x::DynamicQuantities.Quantity, y::Float64)
    reduce(Base.add_sum, DiffEqBase.value(x), init = zero(real(DiffEqBase.value(x)))) +
    reduce(Base.add_sum, y, init = zero(real(DiffEqBase.value(eltype(y)))))
end

DiffEqBase.recursive_length(u::Array) = length(u)
Base.sign(x::DynamicQuantities.Quantity) = Base.sign(DiffEqBase.value(x))

function DiffEqBase.prob2dtmin(prob; use_end_time = true)
    DiffEqBase.prob2dtmin(prob.tspan, oneunit(first(prob.tspan)), use_end_time)
end

DiffEqBase.NAN_CHECK(x::DynamicQuantities.Quantity) = isnan(x)
Base.zero(x::Array{T}) where {T<:DynamicQuantities.Quantity} = zero.(x)

@inline function DiffEqBase.calculate_residuals(ũ, u₀, u₁, α, ρ, internalnorm, t)
    @. DiffEqBase.calculate_residuals(ũ, u₀, u₁, α, ρ, internalnorm, t)
end

f(u, p, t) = u / t;
problem = ODEProblem(f, [1.0u"km/s"], (0.0u"s", 1.0u"s"));
sol = solve(problem, Tsit5(), dt = 0.1u"s")

with just one internal modification. Two interface breaks are weird though:

First one:

julia> typeof(one(0.0u"s"))
Quantity{Float64, Dimensions{DynamicQuantities.FixedRational{Int32, 25200}}}

that should just be Float64?

Second there's something odd in brodcasting I haven't isolated yet.

from differentialequations.jl.

MilesCranmer avatar MilesCranmer commented on June 10, 2024

Thanks, nice work!

Regarding one, see the discussion here: SymbolicML/DynamicQuantities.jl#40. This resulted in the package BaseType.jl for specifically getting the base numeric type. But maybe an interim is to allow Float64 return value, I’m not sure.

Also one alternative to this sort of modification is some of the ideas in SymbolicML/DynamicQuantities.jl#76

from differentialequations.jl.

MilesCranmer avatar MilesCranmer commented on June 10, 2024

Here is the PR to implement these changes: SymbolicML/DynamicQuantities.jl#74

So I think the missing part is switching to oneunit(::T) and one(::T) in OrdinaryDiffEq.jl?

julia> sol = solve(problem, Tsit5(), dt = 0.1u"s")
ERROR: Cannot create a dimensionful 1 for a `UnionAbstractQuantity` type without knowing the dimensions. Please use `oneunit(::UnionAbstractQuantity)` instead.
Stacktrace:
 [1] error(s::String)
   @ Base ./error.jl:35
 [2] oneunit(::Type{Quantity{Float64, Dimensions{DynamicQuantities.FixedRational{Int32, 25200}}}})
   @ DynamicQuantities ~/Documents/DynamicQuantities.jl/src/utils.jl:191
 [3] __init(prob::ODEProblem{…}, alg::Tsit5{…}, timeseries_init::Tuple{}, ts_init::Tuple{}, ks_init::Tuple{}, recompile::Type{…}; saveat::Tuple{}, tstops::Tuple{}, d_discontinuities::Tuple{}, save_idxs::Nothing, save_everystep::Bool, save_on::Bool, save_start::Bool, save_end::Nothing, callback::Nothing, dense::Bool, calck::Bool, dt::Quantity{…}, dtmin::Nothing, dtmax::Quantity{…}, force_dtmin::Bool, adaptive::Bool, gamma::Rational{…}, abstol::Nothing, reltol::Nothing, qmin::Rational{…}, qmax::Int64, qsteady_min::Int64, qsteady_max::Int64, beta1::Nothing, beta2::Nothing, qoldinit::Rational{…}, controller::Nothing, fullnormalize::Bool, failfactor::Int64, maxiters::Int64, internalnorm::typeof(DiffEqBase.ODE_DEFAULT_NORM), internalopnorm::typeof(LinearAlgebra.opnorm), isoutofdomain::typeof(DiffEqBase.ODE_DEFAULT_ISOUTOFDOMAIN), unstable_check::typeof(DiffEqBase.ODE_DEFAULT_UNSTABLE_CHECK), verbose::Bool, timeseries_errors::Bool, dense_errors::Bool, advance_to_tstop::Bool, stop_at_next_tstop::Bool, initialize_save::Bool, progress::Bool, progress_steps::Int64, progress_name::String, progress_message::typeof(DiffEqBase.ODE_DEFAULT_PROG_MESSAGE), progress_id::Symbol, userdata::Nothing, allow_extrapolation::Bool, initialize_integrator::Bool, alias_u0::Bool, alias_du0::Bool, initializealg::OrdinaryDiffEq.DefaultInit, kwargs::@Kwargs{})
   @ OrdinaryDiffEq ~/.julia/packages/OrdinaryDiffEq/qxpST/src/solve.jl:220
 [4] __solve(::ODEProblem{…}, ::Tsit5{…}; kwargs::@Kwargs{})
   @ OrdinaryDiffEq ~/.julia/packages/OrdinaryDiffEq/qxpST/src/solve.jl:5
 [5] solve_call(_prob::ODEProblem{…}, args::Tsit5{…}; merge_callbacks::Bool, kwargshandle::Nothing, kwargs::@Kwargs{})
   @ DiffEqBase ~/.julia/packages/DiffEqBase/NYLhl/src/solve.jl:557
 [6] solve_up(prob::ODEProblem{…}, sensealg::Nothing, u0::Vector{…}, p::SciMLBase.NullParameters, args::Tsit5{…}; kwargs::@Kwargs{})
   @ DiffEqBase ~/.julia/packages/DiffEqBase/NYLhl/src/solve.jl:1006
 [7] solve(prob::ODEProblem{…}, args::Tsit5{…}; sensealg::Nothing, u0::Nothing, p::Nothing, wrap::Val{…}, kwargs::@Kwargs{})
   @ DiffEqBase ~/.julia/packages/DiffEqBase/NYLhl/src/solve.jl:929
 [8] top-level scope
   @ REPL[21]:1
Some type information was truncated. Use `show(err)` to see complete types.

from differentialequations.jl.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.