Move Statistics.jl to weak dependency extension

Project.toml

@@ -11,7 +11,6 @@ GPUArraysCore = "46192b85-c4d5-4398-a991-12ede77f4527"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 RecipesBase = "3cdcf5f2-1ef4-517c-9805-6587b60abb01"
 StaticArraysCore = "1e83bf80-4336-4d27-bf5d-d5a4f845583c"
-Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 SymbolicIndexingInterface = "2efcf032-c050-4f8e-a9bb-153293bab1f5"
 
 [weakdeps]
@@ -22,6 +21,7 @@ Measurements = "eff96d63-e80a-5855-80a2-b1b0885c5ab7"
 MonteCarloMeasurements = "0987c9cc-fe09-11e8-30f0-b96dd679fdca"
 ReverseDiff = "37e2e3b7-166d-5795-8a7a-e32c996b4267"
 SparseArrays = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
+Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 StructArrays = "09ab397b-f2b6-538f-b94a-2f83cf4a842a"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"
 Tracker = "9f7883ad-71c0-57eb-9f7f-b5c9e6d3789c"
@@ -35,6 +35,7 @@ RecursiveArrayToolsMeasurementsExt = "Measurements"
 RecursiveArrayToolsMonteCarloMeasurementsExt = "MonteCarloMeasurements"
 RecursiveArrayToolsReverseDiffExt = ["ReverseDiff", "Zygote"]
 RecursiveArrayToolsSparseArraysExt = ["SparseArrays"]
+RecursiveArrayToolsStatisticsExt = "Statistics"
 RecursiveArrayToolsStructArraysExt = "StructArrays"
 RecursiveArrayToolsTablesExt = ["Tables"]
 RecursiveArrayToolsTrackerExt = "Tracker"
@@ -88,11 +89,12 @@ SafeTestsets = "1bc83da4-3b8d-516f-aca4-4fe02f6d838f"
 SciMLBase = "0bca4576-84f4-4d90-8ffe-ffa030f20462"
 SparseArrays = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
 StaticArrays = "90137ffa-7385-5640-81b9-e52037218182"
+Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 StructArrays = "09ab397b-f2b6-538f-b94a-2f83cf4a842a"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 Tracker = "9f7883ad-71c0-57eb-9f7f-b5c9e6d3789c"
 Unitful = "1986cc42-f94f-5a68-af5c-568840ba703d"
 Zygote = "e88e6eb3-aa80-5325-afca-941959d7151f"
 
 [targets]
-test = ["Aqua", "FastBroadcast", "ForwardDiff", "JET", "KernelAbstractions", "Measurements", "NLsolve", "Pkg", "Random", "SafeTestsets", "SciMLBase", "SparseArrays", "StaticArrays", "StructArrays", "Tables", "Test", "Unitful", "Zygote"]
+test = ["Aqua", "FastBroadcast", "ForwardDiff", "JET", "KernelAbstractions", "Measurements", "NLsolve", "Pkg", "Random", "SafeTestsets", "SciMLBase", "SparseArrays", "StaticArrays", "Statistics", "StructArrays", "Tables", "Test", "Unitful", "Zygote"]

ext/RecursiveArrayToolsStatisticsExt.jl

@@ -0,0 +1,15 @@
+module RecursiveArrayToolsStatisticsExt
+
+using RecursiveArrayTools
+using Statistics
+
+@inline Statistics.mean(VA::AbstractVectorOfArray; kwargs...) = mean(Array(VA); kwargs...)
+@inline function Statistics.median(VA::AbstractVectorOfArray; kwargs...)
+    median(Array(VA); kwargs...)
+end
+@inline Statistics.std(VA::AbstractVectorOfArray; kwargs...) = std(Array(VA); kwargs...)
+@inline Statistics.var(VA::AbstractVectorOfArray; kwargs...) = var(Array(VA); kwargs...)
+@inline Statistics.cov(VA::AbstractVectorOfArray; kwargs...) = cov(Array(VA); kwargs...)
+@inline Statistics.cor(VA::AbstractVectorOfArray; kwargs...) = cor(Array(VA); kwargs...)
+
+end

src/RecursiveArrayTools.jl

@@ -5,7 +5,7 @@ $(DocStringExtensions.README)
 module RecursiveArrayTools
 
 using DocStringExtensions
-using RecipesBase, StaticArraysCore, Statistics,
+using RecipesBase, StaticArraysCore,
       ArrayInterface, LinearAlgebra
 using SymbolicIndexingInterface
 

src/array_partition.jl

@@ -342,7 +342,17 @@ function recursivecopy!(A::ArrayPartition{T, S},
     return A
 end
 
-recursive_mean(A::ArrayPartition) = mean((recursive_mean(x) for x in A.x))
+function recursive_mean(A::ArrayPartition)
+    n = npartitions(A)
+    if n == 0
+        return zero(eltype(A))
+    end
+    total = recursive_mean(A.x[1])
+    for i in 2:n
+        total += recursive_mean(A.x[i])
+    end
+    return total / n
+end
 
 # note: consider only first partition for recursive one and eltype
 recursive_one(A::ArrayPartition) = recursive_one(first(A.x))

src/utils.jl

@@ -293,7 +293,6 @@ end
 recursive_unitless_eltype(a::Type{T}) where {T <: Number} = typeof(one(eltype(a)))
 recursive_unitless_eltype(::Type{<:Enum{T}}) where {T} = T
 
-recursive_mean(x...) = mean(x...)
 function recursive_mean(vecvec::Vector{T}) where {T <: AbstractArray}
     out = zero(vecvec[1])
     for i in eachindex(vecvec)
@@ -302,6 +301,15 @@ function recursive_mean(vecvec::Vector{T}) where {T <: AbstractArray}
     out / length(vecvec)
 end
 
+# Fallback for scalars and general cases without Statistics
+function recursive_mean(x::AbstractArray{T}) where {T <: Number}
+    sum(x) / length(x)
+end
+
+function recursive_mean(x::Number)
+    x
+end
+
 # From Iterators.jl. Moved here since Iterators.jl is not precompile safe anymore.
 
 # Concatenate the output of n iterators

src/vector_of_array.jl

@@ -1153,14 +1153,6 @@ end
     mapreduce(f, Base.mul_prod, VA; kwargs...)
 end
 
-@inline Statistics.mean(VA::AbstractVectorOfArray; kwargs...) = mean(Array(VA); kwargs...)
-@inline function Statistics.median(VA::AbstractVectorOfArray; kwargs...)
-    median(Array(VA); kwargs...)
-end
-@inline Statistics.std(VA::AbstractVectorOfArray; kwargs...) = std(Array(VA); kwargs...)
-@inline Statistics.var(VA::AbstractVectorOfArray; kwargs...) = var(Array(VA); kwargs...)
-@inline Statistics.cov(VA::AbstractVectorOfArray; kwargs...) = cov(Array(VA); kwargs...)
-@inline Statistics.cor(VA::AbstractVectorOfArray; kwargs...) = cor(Array(VA); kwargs...)
 @inline Base.adjoint(VA::AbstractVectorOfArray) = Adjoint(VA)
 
 # linear algebra