Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

DNMY: rely on promote_op for many operations #604

Closed
wants to merge 1 commit into from
Closed

DNMY: rely on promote_op for many operations #604

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion src/JuMP.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,11 +8,12 @@
# See http://github.com/JuliaOpt/JuMP.jl
#############################################################################

isdefined(Base, :__precompile__) && __precompile__()
__precompile__()

module JuMP

importall Base.Operators
using Base: MulFun, AddFun, SubFun, LDivFun, RDivFun

import MathProgBase

Expand Down
117 changes: 49 additions & 68 deletions src/operators.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -309,6 +309,54 @@ end
# A bunch of operator junk to make matrix multiplication and friends act
# reasonably sane with JuMP types

divlike = (LDivFun, RDivFun)
Base.promote_op{R<:Real}(::AddFun, ::Type{Variable}, ::Type{R} ) = AffExpr
Base.promote_op{R<:Real}(::AddFun, ::Type{R}, ::Type{Variable} ) = AffExpr
Base.promote_op{R<:Real}(::SubFun, ::Type{Variable}, ::Type{R} ) = AffExpr
Base.promote_op{R<:Real}(::SubFun, ::Type{R}, ::Type{Variable} ) = AffExpr
Base.promote_op{R<:Real}(::MulFun, ::Type{Variable}, ::Type{R} ) = AffExpr
Base.promote_op{R<:Real}(::MulFun, ::Type{R}, ::Type{Variable} ) = AffExpr
for F in divlike
@eval Base.promote_op{R<:Real}(::$F, ::Type{Variable}, ::Type{R} ) = AffExpr
end

Base.promote_op{R<:Real}(::AddFun, ::Type{AffExpr}, ::Type{R} ) = AffExpr
Base.promote_op{R<:Real}(::AddFun, ::Type{R}, ::Type{AffExpr} ) = AffExpr
Base.promote_op{R<:Real}(::SubFun, ::Type{AffExpr}, ::Type{R} ) = AffExpr
Base.promote_op{R<:Real}(::SubFun, ::Type{R}, ::Type{AffExpr} ) = AffExpr
Base.promote_op{R<:Real}(::MulFun, ::Type{AffExpr}, ::Type{R} ) = AffExpr
Base.promote_op{R<:Real}(::MulFun, ::Type{R}, ::Type{AffExpr} ) = AffExpr
for F in divlike
@eval Base.promote_op{R<:Real}(::$F, ::Type{AffExpr}, ::Type{R} ) = AffExpr
end

Base.promote_op{R<:Real}(::AddFun, ::Type{QuadExpr}, ::Type{R} ) = QuadExpr
Base.promote_op{R<:Real}(::AddFun, ::Type{R}, ::Type{QuadExpr} ) = QuadExpr
Base.promote_op{R<:Real}(::SubFun, ::Type{QuadExpr}, ::Type{R} ) = QuadExpr
Base.promote_op{R<:Real}(::SubFun, ::Type{R}, ::Type{QuadExpr} ) = QuadExpr
Base.promote_op{R<:Real}(::MulFun, ::Type{QuadExpr}, ::Type{R} ) = QuadExpr
Base.promote_op{R<:Real}(::MulFun, ::Type{R}, ::Type{QuadExpr} ) = QuadExpr
for F in divlike
@eval Base.promote_op{R<:Real}(::$F, ::Type{QuadExpr}, ::Type{R} ) = QuadExpr
end

Base.promote_op( ::AddFun, ::Type{Variable}, ::Type{Variable} ) = AffExpr
Base.promote_op( ::SubFun, ::Type{Variable}, ::Type{Variable} ) = AffExpr
Base.promote_op( ::MulFun, ::Type{Variable}, ::Type{Variable} ) = QuadExpr

Base.promote_op( ::AddFun, ::Type{Variable}, ::Type{AffExpr} ) = AffExpr
Base.promote_op( ::AddFun, ::Type{AffExpr}, ::Type{Variable}) = AffExpr
Base.promote_op( ::SubFun, ::Type{Variable}, ::Type{AffExpr} ) = AffExpr
Base.promote_op( ::SubFun, ::Type{AffExpr}, ::Type{Variable}) = AffExpr
Base.promote_op( ::MulFun, ::Type{Variable}, ::Type{AffExpr} ) = QuadExpr
Base.promote_op( ::MulFun, ::Type{AffExpr}, ::Type{Variable}) = QuadExpr

Base.promote_op( ::AddFun, ::Type{Variable}, ::Type{QuadExpr}) = QuadExpr
Base.promote_op( ::AddFun, ::Type{QuadExpr}, ::Type{Variable}) = QuadExpr
Base.promote_op( ::SubFun, ::Type{Variable}, ::Type{QuadExpr}) = QuadExpr
Base.promote_op( ::SubFun, ::Type{QuadExpr}, ::Type{Variable}) = QuadExpr

# Make concatenation work properly
Base.promote_rule{R<:Real}(::Type{Variable},::Type{R} ) = AffExpr
Base.promote_rule( ::Type{Variable},::Type{AffExpr} ) = AffExpr
Base.promote_rule( ::Type{Variable},::Type{QuadExpr}) = QuadExpr
Expand Down Expand Up @@ -417,20 +465,6 @@ end
typealias ArrayOrSparseMat{T} Union{Array{T}, SparseMatrixCSC{T}}

for op in [:+, :-]; @eval begin
function $op{T<:JuMPTypes}(lhs::Number,rhs::ArrayOrSparseMat{T})
ret = Array(typeof($op(lhs, zero(T))), size(rhs))
for I in eachindex(ret)
ret[I] = $op(lhs, rhs[I])
end
ret
end
function $op{T<:JuMPTypes}(lhs::ArrayOrSparseMat{T},rhs::Number)
ret = Array(typeof($op(zero(T), rhs)), size(lhs))
for I in eachindex(ret)
ret[I] = $op(lhs[I], rhs)
end
ret
end
function $op{T<:JuMPTypes,S}(lhs::T,rhs::ArrayOrSparseMat{S})
ret = Array(typeof($op(lhs, zero(S))), size(rhs))
for I in eachindex(ret)
Expand All @@ -447,21 +481,7 @@ for op in [:+, :-]; @eval begin
end
end; end

for op in [:*, :/]; @eval begin
function $op{T<:JuMPTypes}(lhs::Number,rhs::Array{T})
ret = Array(typeof($op(lhs, zero(T))), size(rhs))
for I in eachindex(ret)
ret[I] = $op(lhs, rhs[I])
end
ret
end
function $op{T<:JuMPTypes}(lhs::Array{T},rhs::Number)
ret = Array(typeof($op(zero(T), rhs)), size(lhs))
for I in eachindex(ret)
ret[I] = $op(lhs[I], rhs)
end
ret
end
for op in [:*, :/, :.*, :./]; @eval begin
function $op{T<:JuMPTypes,S}(lhs::T,rhs::Array{S})
ret = Array(typeof($op(lhs, zero(S))), size(rhs))
for I in eachindex(ret)
Expand All @@ -479,56 +499,17 @@ for op in [:*, :/]; @eval begin
end; end

# Special-case sparse matrix scalar multiplication/division
(*){T<:JuMPTypes}(lhs::Number, rhs::SparseMatrixCSC{T}) =
SparseMatrixCSC(rhs.m, rhs.n, copy(rhs.colptr), copy(rhs.rowval), lhs .* rhs.nzval)
(*)(lhs::JuMPTypes, rhs::SparseMatrixCSC) =
SparseMatrixCSC(rhs.m, rhs.n, copy(rhs.colptr), copy(rhs.rowval), lhs .* rhs.nzval)
(*){T<:JuMPTypes}(lhs::SparseMatrixCSC{T}, rhs::Number) =
SparseMatrixCSC(lhs.m, lhs.n, copy(lhs.colptr), copy(lhs.rowval), lhs.nzval .* rhs)
(*)(lhs::SparseMatrixCSC, rhs::JuMPTypes) =
SparseMatrixCSC(lhs.m, lhs.n, copy(lhs.colptr), copy(lhs.rowval), lhs.nzval .* rhs)
(/){T<:JuMPTypes}(lhs::SparseMatrixCSC{T}, rhs::Number) =
SparseMatrixCSC(lhs.m, lhs.n, copy(lhs.colptr), copy(lhs.rowval), lhs.nzval ./ rhs)

# The following are primarily there for internal use in the macro code for @addConstraint
for op in [:(+), :(-)]; @eval begin
function $op(lhs::Array{Variable},rhs::Array{Variable})
(sz = size(lhs)) == size(rhs) || error("Incompatible sizes for $op: $sz $op $(size(rhs))")
ret = Array(AffExpr, sz)
for I in eachindex(ret)
ret[I] = $op(lhs[I], rhs[I])
end
ret
end
end; end

for (dotop,op) in [(:.+,:+), (:.-,:-), (:.*,:*), (:./,:/)]
@eval begin
$dotop(lhs::Number,rhs::JuMPTypes) = $op(lhs,rhs)
$dotop(lhs::JuMPTypes,rhs::Number) = $op(lhs,rhs)
end
for (T1,T2) in [(:JuMPTypes,:Number),(:JuMPTypes,:JuMPTypes),(:Number,:JuMPTypes)]
# Need these looks over S1,S2 for v0.3 because Union{Array,SparseMatrix}
# gives ambiguity warnings
for S1 in (:Array,:SparseMatrixCSC)
@eval $dotop{S<:$T1}(lhs::$S1{S},rhs::$T2) = $op(lhs,rhs)
end
for (S1,S2) in [(:Array,:Array),(:Array,:SparseMatrixCSC),(:SparseMatrixCSC,:Array)]
@eval begin
function $dotop{S<:$T1,T<:$T2}(lhs::$S1{S},rhs::$S2{T})
size(lhs) == size(rhs) || error("Incompatible dimensions")
arr = Array(typeof($op(zero(S),zero(T))), size(rhs))
@inbounds for i in eachindex(lhs)
arr[i] = $op(lhs[i], rhs[i])
end
arr
end
end
end
for S2 in (:Array,:SparseMatrixCSC)
@eval $dotop{T<:$T2}(lhs::$T1,rhs::$S2{T}) = $op(lhs,rhs)
end
end
end


Expand Down
10 changes: 9 additions & 1 deletion test/operator.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -465,7 +465,7 @@ module TestHelper # weird scoping behavior with FactCheck...
return x
end

vec_eq(x,y) = vec_eq([x;], [y;])
vec_eq(x,y) = vec_eq(convert(Array, [x;]), convert(Array, [y;]))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is this change needed?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

isa(x, SparseMatrixCSC) == true. If you leave it as a sparse matrix, you get a stack overflow because all other methods are implemented only for Array and so this method gets called again.


function vec_eq(x::Array, y::Array)
size(x) == size(y) || return false
Expand Down Expand Up @@ -575,6 +575,14 @@ context("Vectorized arithmetic") do
x[2] + 3x[3]]) --> true

@fact TestHelper.vec_eq(@defExpr(A*x/2), A*x/2) --> true

x2 = x[1:2]
Xcov = [x[1]^2 x[1]*x[2];
x[1]*x[2] x[2]^2]
@fact TestHelper.vec_eq(x2 .* x2', Xcov) --> true

X = reshape(x2, 2, 1)
@fact TestHelper.vec_eq(X * X', Xcov) --> true
end

context("Dot-ops") do
Expand Down