BenchmarkRunner: Ensure outputs are discarded

test/instance/file_test.jl

@@ -18,7 +18,7 @@ using Gurobi
         solver = LearningSolver(Gurobi.Optimizer)
         solve!(solver, file_instance)
 
-        loaded = load_jump_instance(filename)
+        loaded = load_instance(filename)
         @test length(loaded.py.samples) == 1
     end    
 end

test/instance/jump_test.jl

@@ -20,7 +20,7 @@ using MIPLearn
         @test isfile(filename)
 
         # Read model from file
-        loaded = load_jump_instance(filename)
+        loaded = load_instance(filename)
         x = variable_by_name(loaded.model, "x")
         @test loaded.model.ext[:miplearn][:variable_features][x] == [1.0]
         @test loaded.model.ext[:miplearn][:variable_categories][x] == "cat1"

test/solvers/learning_test.jl

@@ -35,4 +35,12 @@ using Gurobi
         loaded = load_solver(filename)
         @test loaded.py.components == "Placeholder"
     end
+
+    @testset "Discard output" begin
+        instance = build_knapsack_file_instance()
+        solver = LearningSolver(Gurobi.Optimizer)
+        solve!(solver, instance, discard_output=true)
+        loaded = load_instance(instance.filename)
+        @test length(loaded.py.samples) == 0
+    end
 end

test/utils/benchmark_test.jl

@@ -8,24 +8,39 @@ using Gurobi
 
 
 @testset "BenchmarkRunner" begin
-    # Configure benchmark suite
-    benchmark = BenchmarkRunner(
-        solvers=Dict(
-            "Baseline" => LearningSolver(Gurobi.Optimizer, components=[]),
-            "Proposed" => LearningSolver(Gurobi.Optimizer),
-        ),
-    )
-
-    # Solve instances in parallel
+    # Initialie instances and generate training data
     instances = [
         build_knapsack_file_instance(),
         build_knapsack_file_instance(),
     ]
-    parallel_solve!(benchmark, instances)
+    parallel_solve!(
+        LearningSolver(Gurobi.Optimizer),
+        instances,
+    )
+
+    # Fit and benchmark
+    benchmark = BenchmarkRunner(
+        solvers=Dict(
+            "baseline" => LearningSolver(
+                Gurobi.Optimizer,
+                components=[],
+            ),
+            "ml-exact" => LearningSolver(
+                Gurobi.Optimizer,
+            ),
+            "ml-heur" => LearningSolver(
+                Gurobi.Optimizer,
+                mode="heuristic",
+            ),
+        ),
+    )
+    fit!(benchmark, instances)
+    parallel_solve!(benchmark, instances, n_trials=1)
 
     # Write CSV
     csv_filename = tempname()
     write_csv!(benchmark, csv_filename)
     @test isfile(csv_filename)
     csv = DataFrame(CSV.File(csv_filename))
+    @test size(csv)[1] == 6
 end