Merge pull request #420 from JuliaArrays/qrpivot

Allow for setting QR pivot types

Project.toml

@@ -1,6 +1,6 @@
 name = "ArrayInterface"
 uuid = "4fba245c-0d91-5ea0-9b3e-6abc04ee57a9"
-version = "7.4.10"
+version = "7.4.11"
 
 [deps]
 Adapt = "79e6a3ab-5dfb-504d-930d-738a2a938a0e"

src/ArrayInterface.jl

@@ -622,21 +622,25 @@ else
 end
 
 """
-  qr_instance(A) -> qr_factorization_instance
+  qr_instance(A, pivot = NoPivot()) -> qr_factorization_instance
 
 Returns an instance of the QR factorization object with the correct type
 cheaply.
 """
-function qr_instance(A::Matrix{T}) where {T}
-    LinearAlgebra.QRCompactWY(zeros(T,0,0),zeros(T,0,0))
+function qr_instance(A::Matrix{T},pivot = DEFAULT_CHOLESKY_PIVOT) where {T}
+    if pivot === DEFAULT_CHOLESKY_PIVOT
+        LinearAlgebra.QRCompactWY(zeros(T,0,0),zeros(T,0,0))
+    else
+        LinearAlgebra.QRPivoted(zeros(T,0,0),zeros(T,0),zeros(Int,0))
+    end
 end
 
-function qr_instance(A::Matrix{BigFloat})
+function qr_instance(A::Matrix{BigFloat},pivot = DEFAULT_CHOLESKY_PIVOT)
     LinearAlgebra.QR(zeros(BigFloat,0,0),zeros(BigFloat,0))
 end
 
 # Could be optimized but this should work for any real case.
-function qr_instance(jac_prototype::SparseMatrixCSC)
+function qr_instance(jac_prototype::SparseMatrixCSC, pivot = DEFAULT_CHOLESKY_PIVOT)
     qr(sparse(rand(1,1)))
 end
 
@@ -645,15 +649,15 @@ end
 
 Returns the number.
 """
-qr_instance(a::Number) = a
+qr_instance(a::Number, pivot = DEFAULT_CHOLESKY_PIVOT) = a
 
 """
     qr_instance(a::Any) -> qr(a)
 
 Slow fallback which gets the instance via factorization. Should get
 specialized for new matrix types.
 """
-qr_instance(a::Any) = qr(a)# check = false)
+qr_instance(a::Any, pivot = DEFAULT_CHOLESKY_PIVOT) = qr(a)# check = false)
 
 """
   svd_instance(A) -> qr_factorization_instance