Make LearningSolver working with JuMP

src/julia/src/MIPLearn.jl

@@ -32,6 +32,7 @@ end
     function set_warm_start(self, solution)
         for (basename, subsolution) in solution
             for (idx, value) in subsolution
+                value != nothing || continue
                 var = self.basename_idx_to_var[basename, idx]
                 JuMP.set_start_value(var, value)
             end
@@ -45,6 +46,7 @@ end
     function fix(self, solution)
         for (basename, subsolution) in solution
             for (idx, value) in subsolution
+                value != nothing || continue
                 var = self.basename_idx_to_var[basename, idx]
                 JuMP.fix(var, value, force=true)
             end

src/julia/test/learning_solver.jl

@@ -13,38 +13,21 @@ using Gurobi
                                     [505., 352., 458., 220.],
                                     67.0)
         model = instance.to_model()
-
+        solver = LearningSolver(solver=JuMPSolver(optimizer=optimizer),
-        solver = LearningSolver(solver=JuMPSolver(optimizer=optimizer))
+                                mode="heuristic")
         stats = solver.solve(instance, model)
-
+        @test instance.solution["x"]["1"] == 1.0
-        @test stats["Lower bound"] == 1183.0
+        @test instance.solution["x"]["2"] == 0.0
-        @test stats["Upper bound"] == 1183.0
+        @test instance.solution["x"]["3"] == 1.0
-        @test stats["Sense"] == "max"
+        @test instance.solution["x"]["4"] == 1.0
-        @test stats["Wallclock time"] > 0
+        @test instance.lower_bound == 1183.0
-
+        @test instance.upper_bound == 1183.0
-#         solution = solver.get_solution()
+        @test round(instance.lp_solution["x"]["1"], digits=3) == 1.000
-#         @test solution["x[1]"] == 1.0
+        @test round(instance.lp_solution["x"]["2"], digits=3) == 0.923
-#         @test solution["x[2]"] == 0.0
+        @test round(instance.lp_solution["x"]["3"], digits=3) == 1.000
-#         @test solution["x[3]"] == 1.0
+        @test round(instance.lp_solution["x"]["4"], digits=3) == 0.000
-#         @test solution["x[4]"] == 1.0
+        @test round(instance.lp_value, digits=3) == 1287.923
-#
+        solver.fit([instance])
-#         stats = solver.solve_lp()
+        solver.solve(instance)
-#         @test round(stats["Optimal value"], digits=3) == 1287.923
-#
-#         solution = solver.get_solution()
-#         @test round(solution["x[1]"], digits=3) == 1.000
-#         @test round(solution["x[2]"], digits=3) == 0.923
-#         @test round(solution["x[3]"], digits=3) == 1.000
-#         @test round(solution["x[4]"], digits=3) == 0.000
-#
-#         solver.fix(Dict(
-#             "x[1]" => 1.0,
-#             "x[2]" => 0.0,
-#             "x[3]" => 0.0,
-#             "x[4]" => 1.0,
-#         ))
-#         stats = solver.solve()
-#         @test stats["Lower bound"] == 725.0
-#         @test stats["Upper bound"] == 725.0
     end
 end

src/julia/test/runtests.jl

@@ -3,8 +3,12 @@
 #  Released under the modified BSD license. See COPYING.md for more details.
 
 using Test
+using PyCall
+
+logging = pyimport("logging")
+logging.basicConfig(format="%(levelname)10s %(message)s", level=logging.DEBUG)
 
 @testset "MIPLearn" begin
     include("jump_solver.jl")
-    #include("learning_solver.jl")
+    include("learning_solver.jl")
 end