Add vector of scalar variables with in syntax

[joaquimg]

Partially solves: #2148

this enables both:

model = Model()
@variable(model, x[1:2] in MOI.Integer())

and

model = Model()
@variable(model, x[1:2] in [MOI.Integer(), MOI.ZeroOne()])

.in does not parse in julia, so @variable(model, x[1:2] .in MOI.GreaterThan.([3, 2])) is not an option.

Also added protection to other cases.

This would be a great replacement for jump-dev/ParametricOptInterface.jl#40

cc @guilhermebodin

[codecov]

Codecov Report

Merging #2657 (0cfa41c) into master (28a57b2) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2657      +/-   ##
==========================================
+ Coverage   93.84%   93.85%   +0.01%     
==========================================
  Files          44       44              
  Lines        5524     5534      +10     
==========================================
+ Hits         5184     5194      +10     
  Misses        340      340              

Impacted Files	Coverage Δ
src/Containers/DenseAxisArray.jl	88.77% <100.00%> (ø)
src/macros.jl	94.94% <100.00%> (+0.06%)	⬆️
src/variables.jl	98.62% <100.00%> (+<0.01%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 28a57b2...0cfa41c. Read the comment docs.

[odow]

Just my style pet-peeves. JuliaFormatter doesn't have an option for this yet.

[joaquimg]

Latest version allows SparseAxisArray

@variable(model, b[i=1:2, j=1:2; i+j==3] in MOI.GreaterThan(3.0))
# and
@variable(model, a[i=1:2, j=1:2; i+j==3] in
    JuMP.Containers.SparseAxisArray(
        Dict(
            (1,2) => MOI.GreaterThan(3.0),
            (2,1) => MOI.GreaterThan(3.0),
)))

But DenseAxisArray like:

@variable(model, z1[1:2, 2:4] in MOI.GreaterThan(3.0))

fails with:

Add vector of variables: Error During Test at C:\Users\joaquimgarcia\.julia\dev\JuMP\test\macros.jl:920
  Got exception outside of a @test
  MethodError: no method matching similar(::Type{Array{VariableRef, N} where N}, ::Tuple{Base.OneTo{Int64}, UnitRange{Int64}})
  Closest candidates are:
    similar(::AbstractArray{T, N} where N, ::Tuple) where T at abstractarray.jl:740
    similar(::Type{A}, ::Type{T}, ::StaticArrays.Size{S}) where {A<:Array, T, S} at C:\Users\joaquimgarcia\.julia\packages\StaticArrays\lpEdQ\src\abstractarray.jl:163
    similar(::Type{T}, ::Union{Integer, AbstractUnitRange}...) where T<:AbstractArray at abstractarray.jl:783
    ...
  Stacktrace:
    [1] similar(#unused#::Base.Broadcast.Broadcasted{Base.Broadcast.ArrayConflict, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, typeof(add_variable), Tuple{Base.RefValue{Model}, JuMP.Containers.DenseAxisArray{VariableConstrainedOnCreation{MathOptInterface.GreaterThan{Float64}, ScalarVariable{Float64, Float64, Float64, Float64}}, 2, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, Tuple{JuMP.Containers._AxisLookup{Base.OneTo{Int64}}, JuMP.Containers._AxisLookup{Tuple{Int64, Int64}}}}, JuMP.Containers.DenseAxisArray{String, 2, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, Tuple{JuMP.Containers._AxisLookup{Base.OneTo{Int64}}, JuMP.Containers._AxisLookup{Tuple{Int64, Int64}}}}}}, #unused#::Type{VariableRef}, dims::Tuple{Base.OneTo{Int64}, UnitRange{Int64}})
      @ Base.Broadcast .\broadcast.jl:202
    [2] similar(bc::Base.Broadcast.Broadcasted{Base.Broadcast.ArrayConflict, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, typeof(add_variable), Tuple{Base.RefValue{Model}, JuMP.Containers.DenseAxisArray{VariableConstrainedOnCreation{MathOptInterface.GreaterThan{Float64}, ScalarVariable{Float64, Float64, Float64, Float64}}, 2, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, Tuple{JuMP.Containers._AxisLookup{Base.OneTo{Int64}}, JuMP.Containers._AxisLookup{Tuple{Int64, Int64}}}}, JuMP.Containers.DenseAxisArray{String, 2, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, Tuple{JuMP.Containers._AxisLookup{Base.OneTo{Int64}}, JuMP.Containers._AxisLookup{Tuple{Int64, Int64}}}}}}, #unused#::Type{VariableRef})
      @ Base.Broadcast .\broadcast.jl:196
    [3] copy
      @ .\broadcast.jl:908 [inlined]
    [4] materialize
      @ .\broadcast.jl:883 [inlined]
    [5] add_variable(model::Model, variables::JuMP.Containers.DenseAxisArray{VariableConstrainedOnCreation{MathOptInterface.GreaterThan{Float64}, ScalarVariable{Float64, Float64, Float64, Float64}}, 2, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, Tuple{JuMP.Containers._AxisLookup{Base.OneTo{Int64}}, JuMP.Containers._AxisLookup{Tuple{Int64, Int64}}}}, names::JuMP.Containers.DenseAxisArray{String, 2, Tuple{Base.OneTo{Int64}, UnitRange{Int64}}, Tuple{JuMP.Containers._AxisLookup{Base.OneTo{Int64}}, JuMP.Containers._AxisLookup{Tuple{Int64, Int64}}}})
      @ JuMP C:\Users\joaquimgarcia\.julia\dev\JuMP\src\variables.jl:1082

Any ideas?

[odow]

Hmm. I'm guessing I did something wrong here: #2533. Broadcasting over multiple DenseAxisArrays is hard.

[odow]

We probably need to modify this somehow

JuMP.jl/src/Containers/DenseAxisArray.jl

Lines 190 to 206 in 2a87c29

	# We specify `Ax` for the type of `axes` to avoid conflict where `axes` has type `Tuple{Vararg{Int,N}}`.
	function Base.similar(
	A::DenseAxisArray{T,N,Ax},
	::Type{S},
	axes::Ax,
	) where {T,N,Ax,S}
	return construct_undef_array(S, axes)
	end
	# Avoid conflict with method defined in Julia Base when the axes of the
	# `DenseAxisArray` are all `Base.OneTo`:
	function Base.similar(
	::DenseAxisArray{T,N,Ax},
	::Type{S},
	axes::Ax,
	) where {T,N,Ax<:Tuple{Base.OneTo,Vararg{Base.OneTo}},S}
	return construct_undef_array(S, axes)
	end

[odow]

@joaquimg I pushed a fix for the broadcasting

[joaquimg]

this was subtle