Get rewrite up and running (#59)

* Do not track manifest

* Bump dependencies and use HiGHS as test solver

* Correct type in dictionary

* Add field for segment structure and constructors to match old interface

* Specialize piecewiselinear for typical 1D and 2D usage

* Include segment structure in triangle selection

* Update tests to reflect changes, use HiGHS and run all bivariate

* Fix triangulation

* Avoid explicit use of MOI

Author: trulsf

.gitignore

@@ -1 +1,3 @@
 .DS_Store
+Manifest.toml
+.vscode/settings.json

Manifest.toml

@@ -6,19 +6,16 @@ version = "0.3.0"
 [deps]
 JuMP = "4076af6c-e467-56ae-b986-b466b2749572"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
-MathOptInterface = "b8f27783-ece8-5eb3-8dc8-9495eed66fee"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
 
 [compat]
-Cbc = "≥ 0.6.0"
-JuMP = "~0.21"
-MathOptInterface = "~0.9"
+HiGHS = "1"
+JuMP = "1"
 julia = "1"
 
 [extras]
-Cbc = "9961bab8-2fa3-5c5a-9d89-47fab24efd76"
-Gurobi = "2e9cd046-0924-5485-92f1-d5272153d98b"
+HiGHS = "87dc4568-4c63-4d18-b0c0-bb2238e4078b"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test", "Cbc", "Gurobi"]
+test = ["Test", "HiGHS"]

Project.toml

@@ -3,8 +3,6 @@ __precompile__()
 module PiecewiseLinearOpt
 
 import JuMP
-import MathOptInterface
-const MOI = MathOptInterface
 using LinearAlgebra
 using Random
 
@@ -26,38 +24,45 @@ end
 
 const VarOrAff = Union{JuMP.VariableRef,JuMP.AffExpr}
 
-include(joinpath("methods", "util.jl"))
+include("methods/util.jl")
 
 export Incremental, LogarithmicEmbedding, LogarithmicIndependentBranching, NativeSOS2, ZigZagBinary, ZigZagInteger
-include(joinpath("methods", "univariate", "incremental.jl"))
-include(joinpath("methods", "univariate", "logarithmic_embedding.jl"))
-include(joinpath("methods", "univariate", "logarithmic_independent_branching.jl"))
-include(joinpath("methods", "univariate", "native_sos2.jl"))
-include(joinpath("methods", "univariate", "zig_zag_binary.jl"))
-include(joinpath("methods", "univariate", "zig_zag_integer.jl"))
+include("methods/univariate/incremental.jl")
+
+include("methods/univariate/logarithmic_embedding.jl")
+include("methods/univariate/logarithmic_independent_branching.jl")
+include("methods/univariate/native_sos2.jl")
+include("methods/univariate/zig_zag_binary.jl")
+include("methods/univariate/zig_zag_integer.jl")
 # ConvexCombination has an SOS2 formulation, so defer this until after the
 # multivariate formulations are defined
-include(joinpath("methods", "univariate", "sos2_formulation_base.jl"))
+include("methods/univariate/sos2_formulation_base.jl")
 
 # Consider the colloqial "log" to refer to the embedding formulation
 const Logarithmic = LogarithmicEmbedding
 export Logarithmic
 
-export K1, NineStencil, OptimalIndendentBranching, OptimalTriangleSelection, SixStencil, UnionJack
-include(joinpath("methods", "bivariate", "k1.jl"))
-include(joinpath("methods", "bivariate", "nine_stencil.jl"))
-include(joinpath("methods", "bivariate", "optimal_independent_branching.jl"))
-include(joinpath("methods", "bivariate", "optimal_triangle_selection.jl"))
-include(joinpath("methods", "bivariate", "six_stencil.jl"))
-include(joinpath("methods", "bivariate", "union_jack.jl"))
-include(joinpath("methods", "bivariate", "common.jl"))
-
-export ConvexCombination, DisaggregatedLogarithmic, MultipleChoice, OptimalIndependentBranching, OptimalTriangleSelection
-include(joinpath("methods", "multivariate", "convex_combination.jl"))
-include(joinpath("methods", "multivariate", "disaggregated_logarithmic.jl"))
-include(joinpath("methods", "multivariate", "multiple_choice.jl"))
-
-function formulate_pwl!(model::JuMP.Model, input_vals::Vector{NTuple{D,VarOrAff}}, output_vals::Vector{NTuple{F,VarOrAff}}, pwl::PWLFunction, method::Method, direction::DIRECTION) where {D,F}
+export K1, NineStencil, OptimalIndependentBranching, OptimalTriangleSelection, SixStencil, UnionJack
+include("methods/bivariate/k1.jl")
+include("methods/bivariate/nine_stencil.jl")
+include("methods/bivariate/optimal_independent_branching.jl")
+include("methods/bivariate/optimal_triangle_selection.jl")
+include("methods/bivariate/six_stencil.jl")
+include("methods/bivariate/union_jack.jl")
+include("methods/bivariate/common.jl")
+
+export ConvexCombination, DisaggregatedLogarithmic, MultipleChoice
+include("methods/multivariate/convex_combination.jl")
+include("methods/multivariate/disaggregated_logarithmic.jl")
+include("methods/multivariate/multiple_choice.jl")
+
+function formulate_pwl!(
+    model::JuMP.Model,
+    input_vals::Vector{NTuple{D,VarOrAff}},
+    output_vals::Vector{NTuple{F,VarOrAff}},
+    pwl::PWLFunction,
+    method::Method,
+    direction::DIRECTION) where {D,F}
     error("No support for a R^$D -> R^$F piecewise linear function using the $method method.")
 end
 
@@ -93,4 +98,42 @@ function piecewiselinear(model::JuMP.Model,
     return output_vars
 end
 
+function piecewiselinear(
+    model::JuMP.Model,
+    input_var::VarOrAff,
+    pwl::PWLFunction{1,1,SegmentPointRep{1,1}};
+    method::Method = _default_method(Val(1)),
+    direction::DIRECTION = Graph,
+    output_var::Union{Nothing, VarOrAff} = nothing
+)
+    return piecewiselinear(
+        model,
+        (input_var,),
+        pwl;
+        method = method,
+        direction = direction,
+        output_vars = isnothing(output_var) ? nothing : (output_var,)
+    )[1]
+end
+
+function piecewiselinear(
+    model::JuMP.Model,
+    input_var_x::VarOrAff,
+    input_var_y::VarOrAff,
+    pwl::PWLFunction{2,1,SegmentPointRep{2,1}};
+    method::Method = _default_method(Val(2)),
+    direction::DIRECTION = Graph,
+    output_var::Union{Nothing, VarOrAff} = nothing
+)
+    return piecewiselinear(
+        model,
+        (input_var_x, input_var_y),
+        pwl;
+        method = method,
+        direction = direction,
+        output_vars = isnothing(output_var) ? nothing : (output_var,)
+    )[1]
+end
+
+
 end # module

src/PiecewiseLinearOpt.jl

@@ -1,6 +1,6 @@
 const BivariateSOS2Method = Union{K1, OptimalTriangleSelection, NineStencil, SixStencil, UnionJack}
 
-function formulate_pwl!(model::JuMP.Model, input_vars::NTuple{2, VarOrAff}, output_vars::NTuple{F, VarOrAff}, pwl::BivariatePWLFunction{F}, method::BivariateSOS2Method, direction::DIRECTION) where {F}
+function formulate_pwl!(model::JuMP.Model, input_vars::NTuple{2, VarOrAff}, output_vars::NTuple{F, VarOrAff}, pwl::PWLFunctionPointRep{2, F}, method::BivariateSOS2Method, direction::DIRECTION) where {F}
     initPWL!(model)
     counter = model.ext[:PWL].counter
     counter += 1
@@ -58,5 +58,9 @@ function formulate_pwl!(model::JuMP.Model, input_vars::NTuple{2, VarOrAff}, outp
         end
     end
 
-    formulate_triangle_selection!(model, λ, triangle_direction, method)
+    formulate_triangle_selection!(model, λ, triangle_direction, method, pwl.structure)
+end
+
+function formulate_triangle_selection!(model::JuMP.Model, λ::Matrix{JuMP.VariableRef}, triangle_direction::Matrix{Bool}, method::BivariateSOS2Method, structure::GridTriangulation)
+    error("The triangulation structure $structure is not suppported for method $method")
 end

src/methods/bivariate/common.jl

@@ -5,7 +5,7 @@ K1() = K1(Logarithmic())
 
 axis_method(method::K1) = method.axis_method
 
-function formulate_triangle_selection!(model::JuMP.Model, λ::Matrix{JuMP.VariableRef}, triangle_direction::Matrix{Bool}, method::K1)
+function formulate_triangle_selection!(model::JuMP.Model, λ::Matrix{JuMP.VariableRef}, triangle_direction::Matrix{Bool}, method::K1, structure::K1Triangulation)
     n_1, n_2 = size(λ)
     @assert size(triangle_direction) == (n_1 - 1, n_2 - 1)
     counter = model.ext[:PWL].counter

src/methods/bivariate/k1.jl

@@ -6,7 +6,7 @@ NineStencil() = NineStencil(Logarithmic())
 axis_method(method::NineStencil) = method.axis_method
 
 # TODO: Unit tests for biclique cover
-function formulate_triangle_selection!(model::JuMP.Model, λ::Matrix{JuMP.VariableRef}, triangle_direction::Matrix{Bool}, method::NineStencil)
+function formulate_triangle_selection!(model::JuMP.Model, λ::Matrix{JuMP.VariableRef}, triangle_direction::Matrix{Bool}, method::NineStencil, _::GridTriangulation)
     n_1, n_2 = size(λ)
     @assert size(triangle_direction) == (n_1 - 1, n_2 - 1)
     counter = model.ext[:PWL].counter