No cachedrule for DynamicQuantities types

[mikeingold]

I just started experimenting with the recent DynamicQuantities.jl package, which aims to provide the same functionality as Unitful.jl but with a statically-typed Quantity to improve performance. It looks like QuadGK currently doesn't support these DynamicQuantities Quantity types, but I'm not familiar enough with the codebase to quite understand why.

I'm interested in contributing by generating a PR for this, but would really appreciate a nudge in the right direction. It looks like the issue spawns from approximately here, where either a @generated cachedrule isn't being produced at compile-time or maybe the generic cachedrule below it isn't working.

Any ideas that might help me understand what I can do to help integrate this functionality?

Thanks,
Mike

Minimal reproduction (Julia v1.9.3):

julia> using DynamicQuantities  # [06fc5a27] DynamicQuantities v0.7.0

julia> using QuadGK  # [1fd47b50] QuadGK v2.8.2

julia> quadgk(t -> 5u"m/s"*t, 0u"s", 10u"s")
ERROR: MethodError: no method matching cachedrule(::Type{Quantity{Float64, Dimensions{DynamicQuantities.FixedRational{Int32, 25200}}}}, ::Int64)

Closest candidates are:
  cachedrule(::Union{Type{Complex{BigFloat}}, Type{BigFloat}}, ::Integer)
   @ QuadGK C:\Users\*\.julia\packages\QuadGK\XqIlh\src\gausskronrod.jl:253
  cachedrule(::Type{T}, ::Integer) where T<:Number
   @ QuadGK C:\Users\*\.julia\packages\QuadGK\XqIlh\src\gausskronrod.jl:264

[stevengj]

You need to provide a way to strip the units from the type. Currently, we use:

QuadGK.jl/src/gausskronrod.jl

Lines 571 to 572 in cb745d4

	cachedrule(::Type{T}, n::Integer) where {T<:Number} =
	_cachedrule(typeof(float(real(one(T)))), Int(n))

which assumes that one(T) returns the underlying dimensionless numeric type.

See e.g. JuliaPhysics/Measurements.jl#134

[mikeingold]

Thanks! I'll see what I can do to get these packages working together.

[stevengj]

Closing in favor of SymbolicML/DynamicQuantities.jl#40

[mikeingold]

Concur. Looks like this is due to DynamicQuantities’ current interpretation of the Base.one spec.