Support vector-valued constraints in normalized_coefficient

docs/src/manual/constraints.md

@@ -843,7 +843,7 @@ julia> normalized_coefficient(con, x[1], x[2])
 ### Vector constraints
 
 To modify the coefficients of a vector-valued constraint, use
-[`set_normalized_coefficients`](@ref).
+[`set_normalized_coefficient`](@ref).
 ```jldoctest
 julia> model = Model();
 
@@ -853,12 +853,12 @@ x
 julia> @constraint(model, con, [2x + 3x, 4x] in MOI.Nonnegatives(2))
 con : [5 x, 4 x] ∈ MathOptInterface.Nonnegatives(2)
 
-julia> set_normalized_coefficients(con, x, [(1, 3.0)])
+julia> set_normalized_coefficient(con, x, [(1, 3.0)])
 
 julia> con
 con : [3 x, 4 x] ∈ MathOptInterface.Nonnegatives(2)
 
-julia> set_normalized_coefficients(con, x, [(1, 2.0), (2, 5.0)])
+julia> set_normalized_coefficient(con, x, [(1, 2.0), (2, 5.0)])
 
 julia> con
 con : [2 x, 5 x] ∈ MathOptInterface.Nonnegatives(2)

src/variables.jl

@@ -2624,7 +2624,7 @@ function set_normalized_coefficient(
 end
 
 """
-    set_normalized_coefficients(
+    set_normalized_coefficient(
         con_ref::ConstraintRef,
         variable::AbstractVariableRef,
         new_coefficients::Vector{Tuple{Int64,T}},
@@ -2648,13 +2648,13 @@ x
 julia> @constraint(model, con, [2x + 3x, 4x] in MOI.Nonnegatives(2))
 con : [5 x, 4 x] ∈ MathOptInterface.Nonnegatives(2)
 
-julia> set_normalized_coefficients(con, x, [(1, 2.0), (2, 5.0)])
+julia> set_normalized_coefficient(con, x, [(1, 2.0), (2, 5.0)])
 
 julia> con
 con : [2 x, 5 x] ∈ MathOptInterface.Nonnegatives(2)
 ```
 """
-function set_normalized_coefficients(
+function set_normalized_coefficient(
     constraint::ConstraintRef{<:AbstractModel,<:MOI.ConstraintIndex{F}},
     variable::AbstractVariableRef,
     new_coefficients::Vector{Tuple{Int64,T}},
@@ -2669,6 +2669,22 @@ function set_normalized_coefficients(
     return
 end
 
+"""
+    set_normalized_coefficients(
+        constraint::ConstraintRef{<:AbstractModel,<:MOI.ConstraintIndex{F}},
+        variable::AbstractVariableRef,
+        new_coefficients::Vector{Tuple{Int64,T}},
+    ) where {T,F<:Union{MOI.VectorAffineFunction{T},MOI.VectorQuadraticFunction{T}}}
+
+A deprecated method that now redirects to [`set_normalized_coefficient`](@ref).
+"""
+function set_normalized_coefficients(
+    constraint::ConstraintRef{<:AbstractModel,<:MOI.ConstraintIndex{F}},
+    variable::AbstractVariableRef,
+    new_coefficients::Vector{Tuple{Int64,T}},
+) where {T,F<:Union{MOI.VectorAffineFunction{T},MOI.VectorQuadraticFunction{T}}}
+    return set_normalized_coefficient(constraint, variable, new_coefficients)
+end
 """
     set_normalized_coefficient(
         constraint::ConstraintRef,
@@ -2816,6 +2832,14 @@ con : 5 x ≤ 2
 
 julia> normalized_coefficient(con, x)
 5.0
+
+julia> @constraint(model, con_vec, [x, 2x + 1, 3] >= 0)
+con_vec : [x, 2 x + 1, 3] ∈ MathOptInterface.Nonnegatives(3)
+
+julia> normalized_coefficient(con_vec, x)
+2-element Vector{Tuple{Int64, Float64}}:
+ (1, 1.0)
+ (2, 2.0)
 ```
 """
 function normalized_coefficient(
@@ -2825,6 +2849,14 @@ function normalized_coefficient(
     return coefficient(constraint_object(constraint).func, variable)
 end
 
+function normalized_coefficient(
+    constraint::ConstraintRef{<:AbstractModel,<:MOI.ConstraintIndex{F}},
+    variable::AbstractVariableRef,
+) where {T,F<:Union{MOI.VectorAffineFunction{T},MOI.VectorQuadraticFunction{T}}}
+    c = coefficient.(constraint_object(constraint).func, variable)
+    return filter!(!iszero ∘ last, collect(enumerate(c)))
+end
+
 """
     normalized_coefficient(
         constraint::ConstraintRef,
@@ -2852,6 +2884,16 @@ julia> normalized_coefficient(con, x[1], x[1])
 
 julia> normalized_coefficient(con, x[1], x[2])
 3.0
+
+julia> @constraint(model, con_vec, x.^2 <= [1, 2])
+con_vec : [x[1]² - 1, x[2]² - 2] ∈ MathOptInterface.Nonpositives(2)
+
+julia> normalized_coefficient(con_vec, x[1], x[1])
+1-element Vector{Tuple{Int64, Float64}}:
+ (1, 1.0)
+
+julia> normalized_coefficient(con_vec, x[1], x[2])
+Tuple{Int64, Float64}[]
 ```
 """
 function normalized_coefficient(
@@ -2863,6 +2905,16 @@ function normalized_coefficient(
     return coefficient(con.func, variable_1, variable_2)
 end
 
+function normalized_coefficient(
+    constraint::ConstraintRef{<:AbstractModel,<:MOI.ConstraintIndex{F}},
+    variable_1::AbstractVariableRef,
+    variable_2::AbstractVariableRef,
+) where {T,F<:MOI.VectorQuadraticFunction{T}}
+    f = constraint_object(constraint).func
+    c = coefficient.(f, variable_1, variable_2)
+    return filter!(!iszero ∘ last, collect(enumerate(c)))
+end
+
 ###
 ### Error messages for common incorrect usages
 ###

test/test_constraint.jl

@@ -1009,6 +1009,18 @@ function test_change_coefficient_batch()
 end
 
 function test_change_coefficients_vector_function()
+    model = Model()
+    @variable(model, x)
+    @constraint(model, con, [2x + 3x, 4x] in MOI.Nonnegatives(2))
+    @test isequal_canonical(constraint_object(con).func, [5.0x, 4.0x])
+    set_normalized_coefficient(con, x, [(Int64(1), 3.0)])
+    @test isequal_canonical(constraint_object(con).func, [3.0x, 4.0x])
+    set_normalized_coefficient(con, x, [(Int64(1), 2.0), (Int64(2), 5.0)])
+    @test isequal_canonical(constraint_object(con).func, [2.0x, 5.0x])
+    return
+end
+
+function test_change_coefficients_vector_function_DEPRECATED()
     model = Model()
     @variable(model, x)
     @constraint(model, con, [2x + 3x, 4x] in MOI.Nonnegatives(2))

test/test_variable.jl

@@ -1626,4 +1626,29 @@ function test_variable_one()
     return
 end
 
+function test_variable_normalized_coefficient_vector()
+    model = Model()
+    @variable(model, x[1:3])
+    A = Float64[4 5 0; 0 6 7; 8 0 0; 9 10 11]
+    sparse(y) = filter!(!iszero ∘ last, collect(enumerate(y)))
+    @constraint(model, c, A * x >= 0)
+    for i in 1:3
+        @test normalized_coefficient(c, x[i]) == sparse(A[:, i])
+    end
+    return
+end
+
+function test_variable_normalized_coefficient_vector_quadratic()
+    model = Model()
+    @variable(model, x[1:3])
+    A = Float64[4 5 0; 0 6 7; 8 0 0; 9 10 11]
+    sparse(y) = filter!(!iszero ∘ last, collect(enumerate(y)))
+    @constraint(model, c, A * (x .^ 2) >= 0)
+    for i in 1:3, j in 1:3
+        b = ifelse(i == j, A[:, i], zeros(4))
+        @test normalized_coefficient(c, x[i], x[j]) == sparse(b)
+    end
+    return
+end
+
 end  # module TestVariable