Format code with JuliaFormatter v2.1.5

Applied JuliaFormatter to ensure consistent code formatting across the codebase.
This includes formatting improvements to source files, extensions, and test files.
Note: One file (src/vector_of_array.jl) was skipped due to parsing error with special comment syntax.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <[email protected]>

Author: claude

README.md

@@ -45,9 +45,9 @@ pB = ArrayPartition(b)
 pA .* pB # Now all standard array stuff works!
 
 # or do:
-x0 = rand(3,3)
-v0 = rand(3,3)
-a0 = rand(3,3)
+x0 = rand(3, 3)
+v0 = rand(3, 3)
+a0 = rand(3, 3)
 u0 = ArrayPartition(x0, v0, a0)
 u0.x[1] == x0 # true
 

ext/RecursiveArrayToolsKernelAbstractionsExt.jl

@@ -3,7 +3,6 @@ module RecursiveArrayToolsKernelAbstractionsExt
 import RecursiveArrayTools: VectorOfArray
 import KernelAbstractions
 
-
 function KernelAbstractions.get_backend(x::VectorOfArray)
     u = parent(x)
     if length(u) == 0

ext/RecursiveArrayToolsZygoteExt.jl

@@ -5,7 +5,6 @@ using RecursiveArrayTools
 using Zygote
 using Zygote: FillArrays, ChainRulesCore, literal_getproperty, @adjoint
 
-
 # Define a new species of projection operator for this type:
 # ChainRulesCore.ProjectTo(x::VectorOfArray) = ChainRulesCore.ProjectTo{VectorOfArray}()
 
@@ -216,10 +215,12 @@ z̄ -> (nothing, conj.(z̄))
 @adjoint Broadcast.broadcasted(::typeof(real), x::AbstractVectorOfArray) = real.(x),
 z̄ -> (nothing, real.(z̄))
 
-@adjoint Broadcast.broadcasted(::typeof(imag), x::AbstractVectorOfArray) = imag.(x),
+@adjoint Broadcast.broadcasted(
+    ::typeof(imag), x::AbstractVectorOfArray) = imag.(x),
 z̄ -> (nothing, im .* real.(z̄))
 
-@adjoint Broadcast.broadcasted(::typeof(abs2), x::AbstractVectorOfArray) = abs2.(x),
+@adjoint Broadcast.broadcasted(::typeof(abs2),
+    x::AbstractVectorOfArray) = abs2.(x),
 z̄ -> (nothing, 2 .* real.(z̄) .* x)
 
 @adjoint function Broadcast.broadcasted(
@@ -260,7 +261,9 @@ end
     end
 end
 
-@adjoint Broadcast.broadcasted(::Type{T}, x::AbstractVectorOfArray) where {T <: Number} = T.(x),
+@adjoint Broadcast.broadcasted(::Type{T},
+    x::AbstractVectorOfArray) where {T <:
+                                     Number} = T.(x),
 ȳ -> (nothing, Zygote._project(x, ȳ))
 
 function Zygote.unbroadcast(x::AbstractVectorOfArray, x̄)
@@ -273,11 +276,16 @@ function Zygote.unbroadcast(x::AbstractVectorOfArray, x̄)
     end
 end
 
-@adjoint Broadcast.broadcasted(::Broadcast.AbstractArrayStyle, f::F, a::AbstractVectorOfArray, b) where {F} = _broadcast_generic(
+@adjoint Broadcast.broadcasted(
+    ::Broadcast.AbstractArrayStyle, f::F, a::AbstractVectorOfArray,
+    b) where {F} = _broadcast_generic(
     __context__, f, a, b)
-@adjoint Broadcast.broadcasted(::Broadcast.AbstractArrayStyle, f::F, a, b::AbstractVectorOfArray) where {F} = _broadcast_generic(
+@adjoint Broadcast.broadcasted(::Broadcast.AbstractArrayStyle, f::F, a,
+    b::AbstractVectorOfArray) where {F} = _broadcast_generic(
     __context__, f, a, b)
-@adjoint Broadcast.broadcasted(::Broadcast.AbstractArrayStyle, f::F, a::AbstractVectorOfArray, b::AbstractVectorOfArray) where {F} = _broadcast_generic(
+@adjoint Broadcast.broadcasted(
+    ::Broadcast.AbstractArrayStyle, f::F, a::AbstractVectorOfArray,
+    b::AbstractVectorOfArray) where {F} = _broadcast_generic(
     __context__, f, a, b)
 
 @inline function _broadcast_generic(__context__, f::F, args...) where {F}

src/RecursiveArrayTools.jl

@@ -28,22 +28,15 @@ An AbstractVectorOfArray subtype should match the following behaviors.
 
 !!! note
 
-    In 2023 the linear indexing `A[i]`` was deprecated. It previously had the behavior that
-    `A[i] = A.u[i]`. However, this is incompatible with standard `AbstractArray` interfaces,
-    Since if `A = VectorOfArray([[1,2],[3,4]])` and `A` is supposed to act like `[1 3; 2 4]`,
-    then there is a difference `A[1] = [1,2]` for the VectorOfArray while `A[1] = 1` for the
-    matrix. This causes many issues if `AbstractVectorOfArray <: AbstractArray`. Thus we
-    plan in 2026 to complete the deprecation and thus have a breaking update where `A[i]`
-    matches the linear indexing of an `AbstractArray`, and then making
-    `AbstractVectorOfArray <: AbstractArray`. Until then, `AbstractVectorOfArray` due to
+    In 2023 the linear indexing `A[i]`` was deprecated. It previously had the behavior that `A[i] = A.u[i]`. However, this is incompatible with standard `AbstractArray`interfaces, Since if`A = VectorOfArray([[1,2],[3,4]])`and`A`is supposed to act like`[1 3; 2 4]`, then there is a difference `A[1] = [1,2]`for the VectorOfArray while`A[1] = 1`for the matrix. This causes many issues if`AbstractVectorOfArray <: AbstractArray`. Thus we plan in 2026 to complete the deprecation and thus have a breaking update where `A[i]`matches the linear indexing of an`AbstractArray`, and then making `AbstractVectorOfArray <: AbstractArray`. Until then, `AbstractVectorOfArray` due to
     this interface break but manually implements an AbstractArray-like interface for
     future compatibility.
 
 ## Fields
 
 An AbstractVectorOfArray has the following fields:
 
-* `u` which holds the Vector of values at each timestep
+  - `u` which holds the Vector of values at each timestep
 
 ## Array Interface
 
@@ -125,7 +118,7 @@ additional properties:
 
 An AbstractDiffEqArray adds the following fields:
 
-* `t` which holds the times of each timestep.
+  - `t` which holds the times of each timestep.
 """
 abstract type AbstractDiffEqArray{T, N, A} <: AbstractVectorOfArray{T, N, A} end
 

src/named_array_partition.jl

@@ -115,15 +115,20 @@ end
 
 # hook into ArrayPartition broadcasting routines
 @inline RecursiveArrayTools.npartitions(x::NamedArrayPartition) = npartitions(ArrayPartition(x))
-@inline RecursiveArrayTools.unpack(bc::Broadcast.Broadcasted{Broadcast.ArrayStyle{NamedArrayPartition}}, i) = Broadcast.Broadcasted(
+@inline RecursiveArrayTools.unpack(
+    bc::Broadcast.Broadcasted{Broadcast.ArrayStyle{NamedArrayPartition}},
+    i) = Broadcast.Broadcasted(
     bc.f, RecursiveArrayTools.unpack_args(i, bc.args))
 @inline RecursiveArrayTools.unpack(x::NamedArrayPartition, i) = unpack(ArrayPartition(x), i)
 
 function Base.copy(A::NamedArrayPartition{T, S, NT}) where {T, S, NT}
     NamedArrayPartition{T, S, NT}(copy(ArrayPartition(A)), getfield(A, :names_to_indices))
 end
 
-@inline NamedArrayPartition(f::F, N, names_to_indices) where {F <: Function} = NamedArrayPartition(
+@inline NamedArrayPartition(f::F,
+    N,
+    names_to_indices) where {F <:
+                             Function} = NamedArrayPartition(
     ArrayPartition(ntuple(f, Val(N))), names_to_indices)
 
 @inline function Base.copy(bc::Broadcast.Broadcasted{Broadcast.ArrayStyle{NamedArrayPartition}})

src/utils.jl

@@ -110,6 +110,7 @@ function recursivefill!(bs::AbstractVectorOfArray{T, N},
         a::T2) where {T <: StaticArraysCore.StaticArray,
         T2 <: StaticArraysCore.StaticArray, N}
     @inbounds for b in bs, i in eachindex(b)
+
         b[i] = copy(a)
     end
 end
@@ -126,6 +127,7 @@ function recursivefill!(bs::AbstractVectorOfArray{T, N},
         a::T2) where {T <: StaticArraysCore.SArray,
         T2 <: Union{Number, Bool}, N}
     @inbounds for b in bs, i in eachindex(b)
+
         b[i] = fill(a, typeof(b[i]))
     end
 end

test/adjoints.jl

@@ -13,6 +13,7 @@ function loss3(x)
     y = VectorOfArray([x .* i for i in 1:5])
     tmp = 0.0
     for i in 1:5, j in 1:5
+
         tmp += y[i, j]
     end
     tmp
@@ -22,6 +23,7 @@ function loss4(x)
     y = DiffEqArray([x .* i for i in 1:5], 1:5)
     tmp = 0.0
     for i in 1:5, j in 1:5
+
         tmp += y[i, j]
     end
     tmp

test/gpu/vectorofarray_gpu.jl

@@ -43,12 +43,12 @@ va_cu = convert(AbstractArray, va)
 @test size(va_cu) == size(x)
 
 a = VectorOfArray([ones(2) for i in 1:3])
-_a = Adapt.adapt(CuArray,a)
+_a = Adapt.adapt(CuArray, a)
 @test _a isa VectorOfArray
 @test _a.u isa Vector{<:CuArray}
 
-b = DiffEqArray([ones(2) for i in 1:3],ones(2))
-_b = Adapt.adapt(CuArray,b)
+b = DiffEqArray([ones(2) for i in 1:3], ones(2))
+_b = Adapt.adapt(CuArray, b)
 @test _b isa DiffEqArray
 @test _b.u isa Vector{<:CuArray}
 @test _b.t isa CuArray

test/partitions_test.jl

@@ -286,7 +286,8 @@ end
 @testset "Copy and zero with type changing array" begin
     # Motivating use case for this is ArrayPartitions of Arrow arrays which are mmap:ed and change type when copied
     struct TypeChangingArray{T, N} <: AbstractArray{T, N} end
-    Base.copy(::TypeChangingArray{T, N}) where {T, N} = Array{T, N}(undef,
+    Base.copy(::TypeChangingArray{
+        T, N}) where {T, N} = Array{T, N}(undef,
         ntuple(_ -> 0, N))
     Base.zero(::TypeChangingArray{T, N}) where {T, N} = zeros(T, ntuple(_ -> 0, N))