Merge pull request #695 from JuliaArrays/cjf/lu-api-updates

Some trait + LU api cleanup

src/StaticArrays.jl

@@ -86,7 +86,13 @@ const StaticMatrixLike{T} = Union{
     Adjoint{T, <:StaticVecOrMat{T}},
     Symmetric{T, <:StaticMatrix{<:Any, <:Any, T}},
     Hermitian{T, <:StaticMatrix{<:Any, <:Any, T}},
-    Diagonal{T, <:StaticVector{<:Any, T}}
+    Diagonal{T, <:StaticVector{<:Any, T}},
+    # We specifically list *Triangular here rather than using
+    # AbstractTriangular to avoid ambiguities in size() etc.
+    UpperTriangular{T, <:StaticMatrix{<:Any, <:Any, T}},
+    LowerTriangular{T, <:StaticMatrix{<:Any, <:Any, T}},
+    UnitUpperTriangular{T, <:StaticMatrix{<:Any, <:Any, T}},
+    UnitLowerTriangular{T, <:StaticMatrix{<:Any, <:Any, T}}
 }
 const StaticVecOrMatLike{T} = Union{StaticVector{<:Any, T}, StaticMatrixLike{T}}
 const StaticArrayLike{T} = Union{StaticVecOrMatLike{T}, StaticArray{<:Tuple, T}}

src/abstractarray.jl

@@ -1,13 +1,12 @@
-length(a::SA) where {SA <: StaticArrayLike} = length(SA)
+length(a::StaticArrayLike) = prod(Size(a))
 length(a::Type{SA}) where {SA <: StaticArrayLike} = prod(Size(SA))
 
-@pure size(::Type{SA}) where {SA <: StaticArrayLike} = get(Size(SA))
+@pure size(::Type{SA}) where {SA <: StaticArrayLike} = Tuple(Size(SA))
 @inline function size(t::Type{<:StaticArrayLike}, d::Int)
     S = size(t)
     d > length(S) ? 1 : S[d]
 end
-@inline size(a::StaticArrayLike) = size(typeof(a))
-@inline size(a::StaticArrayLike, d::Int) = size(typeof(a), d)
+@inline size(a::StaticArrayLike) = Tuple(Size(a))
 
 Base.axes(s::StaticArray) = _axes(Size(s))
 @pure function _axes(::Size{sizes}) where {sizes}

src/lu.jl

@@ -11,6 +11,24 @@ Base.iterate(S::LU, ::Val{:U}) = (S.U, Val(:p))
 Base.iterate(S::LU, ::Val{:p}) = (S.p, Val(:done))
 Base.iterate(S::LU, ::Val{:done}) = nothing
 
+@inline function Base.getproperty(F::LU, s::Symbol)
+    if s === :P
+        U = getfield(F, :U)
+        p = getfield(F, :p)
+        one(similar_type(p, Size(U)))[:,invperm(p)]
+    else
+        getfield(F, s)
+    end
+end
+
+function Base.show(io::IO, mime::MIME{Symbol("text/plain")}, F::LU)
+    println(io, LU) # Don't show full type - this will be in the factors
+    println(io, "L factor:")
+    show(io, mime, F.L)
+    println(io, "\nU factor:")
+    show(io, mime, F.U)
+end
+
 # LU decomposition
 function lu(A::StaticMatrix, pivot::Union{Val{false},Val{true}}=Val(true))
     L, U, p = _lu(A, pivot)
@@ -136,9 +154,6 @@ end
     :(SVector{$(M-1),Int}($(tuple(2:M...))))
 end
 
-# Base.lufact() interface is fairly inherently type unstable.  Punt on
-# implementing that, for now...
-
 \(F::LU, v::AbstractVector) = F.U \ (F.L \ v[F.p])
 \(F::LU, B::AbstractMatrix) = F.U \ (F.L \ B[F.p,:])
 

src/traits.jl

@@ -93,6 +93,7 @@ Size(::Type{Transpose{T, A}}) where {T, A <: AbstractVecOrMat{T}} = Size(Size(A)
 Size(::Type{Symmetric{T, A}}) where {T, A <: AbstractMatrix{T}} = Size(A)
 Size(::Type{Hermitian{T, A}}) where {T, A <: AbstractMatrix{T}} = Size(A)
 Size(::Type{Diagonal{T, A}}) where {T, A <: AbstractVector{T}} = Size(Size(A)[1], Size(A)[1])
+Size(::Type{<:LinearAlgebra.AbstractTriangular{T, A}}) where {T,A} = Size(A)
 
 @pure Size(::Type{<:AbstractArray{<:Any, N}}) where {N} = Size(ntuple(_ -> Dynamic(), N))
 
@@ -117,7 +118,7 @@ Length(::Size{S}) where {S} = _Length(S...)
 @inline _Length(S...) = Length{Dynamic()}()
 
 # Some @pure convenience functions for `Size`
-@pure get(::Size{S}) where {S} = S
+@pure (::Type{Tuple})(::Size{S}) where {S} = S
 
 @pure getindex(::Size{S}, i::Int) where {S} = i <= length(S) ? S[i] : 1
 
@@ -138,7 +139,7 @@ Base.LinearIndices(::Size{S}) where {S} = LinearIndices(S)
 @pure size_tuple(::Size{S}) where {S} = Tuple{S...}
 
 # Some @pure convenience functions for `Length`
-@pure get(::Length{L}) where {L} = L
+@pure (::Type{Int})(::Length{L}) where {L} = L
 
 @pure Base.:(==)(::Length{L}, l::Int) where {L} = L == l
 @pure Base.:(==)(l::Int, ::Length{L}) where {L} = l == L

src/triangular.jl

@@ -1,5 +1,3 @@
-@inline Size(A::LinearAlgebra.AbstractTriangular{<:Any,<:StaticMatrix}) = Size(A.data)
-
 @inline transpose(A::LinearAlgebra.LowerTriangular{<:Any,<:StaticMatrix}) =
     LinearAlgebra.UpperTriangular(transpose(A.data))
 @inline adjoint(A::LinearAlgebra.LowerTriangular{<:Any,<:StaticMatrix}) =

test/core.jl

@@ -155,6 +155,9 @@
         @test Size(Adjoint(zero(SMatrix{2, 3}))) == Size(3, 2)
         @test Size(Diagonal(SVector(1, 2, 3))) == Size(3, 3)
         @test Size(Transpose(Diagonal(SVector(1, 2, 3)))) == Size(3, 3)
+        @test Size(UpperTriangular(zero(SMatrix{2, 2}))) == Size(2, 2)
+        @test Size(LowerTriangular(zero(SMatrix{2, 2}))) == Size(2, 2)
+        @test Size(LowerTriangular(Symmetric(zero(SMatrix{2, 2})))) == Size(2,2)
     end
 
     @testset "dimmatch" begin

test/lu.jl

@@ -1,5 +1,14 @@
 using StaticArrays, Test, LinearAlgebra
 
+@testset "LU utils" begin
+    F = lu(SA[1 2; 3 4])
+
+    @test @inferred((F->F.p)(F)) === SA[2, 1]
+    @test @inferred((F->F.P)(F)) === SA[0 1; 1 0]
+
+    @test occursin(r"^StaticArrays.LU.*L factor.*U factor"s, sprint(show, MIME("text/plain"), F))
+end
+
 @testset "LU decomposition ($m×$n, pivot=$pivot)" for pivot in (true, false), m in [0:4..., 15], n in [0:4..., 15]
     a = SMatrix{m,n,Int}(1:(m*n))
     l, u, p = @inferred(lu(a, Val{pivot}()))