Modularize LearningSolver into components; implement branch-priority

miplearn/tests/test_benchmark.py

@@ -19,15 +19,16 @@ def test_benchmark():
     # Training phase...
     training_solver = LearningSolver()
     training_solver.parallel_solve(train_instances, n_jobs=10)
-    training_solver.save("data.bin")
+    training_solver.fit()
+    training_solver.save_state("data.bin")
 
     # Test phase...
     test_solvers = {
-        "Strategy A": LearningSolver(ws_predictor=None),
-        "Strategy B": LearningSolver(ws_predictor=None),
+        "Strategy A": LearningSolver(),
+        "Strategy B": LearningSolver(),
     }
     benchmark = BenchmarkRunner(test_solvers)
-    benchmark.load_fit("data.bin")
+    benchmark.load_state("data.bin")
     benchmark.parallel_solve(test_instances, n_jobs=2, n_trials=2)
     assert benchmark.raw_results().values.shape == (12,12)
     

miplearn/tests/test_solver.py

@@ -4,6 +4,8 @@
 
 from miplearn import LearningSolver
 from miplearn.problems.knapsack import KnapsackInstance2
+from miplearn.branching import BranchPriorityComponent
+from miplearn.warmstart import WarmStartComponent
 import numpy as np
 
 
@@ -16,21 +18,29 @@ def test_solver():
     solver.fit()
     solver.solve(instance)
 
-def test_solve_save_load():
+def test_solve_save_load_state():
     instance = KnapsackInstance2(weights=[23., 26., 20., 18.],
                                  prices=[505., 352., 458., 220.],
                                  capacity=67.)
-    solver = LearningSolver()
+    components_before = {
+        "warm-start": WarmStartComponent(),
+        "branch-priority": BranchPriorityComponent(),
+    }
+    solver = LearningSolver(components=components_before)
     solver.solve(instance)
     solver.fit()
-    solver.save("/tmp/knapsack_train.bin")
-    prev_x_train_len = len(solver.x_train)
-    prev_y_train_len = len(solver.y_train)
+    solver.save_state("/tmp/knapsack_train.bin")
+    prev_x_train_len = len(solver.components["warm-start"].x_train)
+    prev_y_train_len = len(solver.components["warm-start"].y_train)
     
-    solver = LearningSolver()
-    solver.load("/tmp/knapsack_train.bin")
-    assert len(solver.x_train) == prev_x_train_len
-    assert len(solver.y_train) == prev_y_train_len
+    components_after = {
+        "warm-start": WarmStartComponent(),
+    }
+    solver = LearningSolver(components=components_after)
+    solver.load_state("/tmp/knapsack_train.bin")
+    assert len(solver.components.keys()) == 1
+    assert len(solver.components["warm-start"].x_train) == prev_x_train_len
+    assert len(solver.components["warm-start"].y_train) == prev_y_train_len
 
 def test_parallel_solve():
     instances = [KnapsackInstance2(weights=np.random.rand(5),
@@ -38,13 +48,18 @@ def test_parallel_solve():
                                    capacity=3.0)
                  for _ in range(10)]
     solver = LearningSolver()
-    solver.parallel_solve(instances, n_jobs=3)
-    assert len(solver.x_train[0]) == 10
-    assert len(solver.y_train[0]) == 10
+    results = solver.parallel_solve(instances, n_jobs=3)
+    assert len(results) == 10
+    assert len(solver.components["warm-start"].x_train[0]) == 10
+    assert len(solver.components["warm-start"].y_train[0]) == 10
     
 def test_solver_random_branch_priority():
     instance = KnapsackInstance2(weights=[23., 26., 20., 18.],
                                  prices=[505., 352., 458., 220.],
                                  capacity=67.)
-    solver = LearningSolver(branch_priority=[1, 2, 3, 4])
-    solver.solve(instance, tee=True)
+    components = {
+        "warm-start": BranchPriorityComponent(priority=np.array([1, 2, 3, 4])),
+    }
+    solver = LearningSolver(components=components)
+    solver.solve(instance)
+    solver.fit()