Fix benchmark scripts; add more input checks

benchmark/benchmark.py

@@ -61,7 +61,7 @@ def read_pickle_gz(filename):
         return pickle.load(file)
 
 
-def write_multiple(objs, dirname):
+def write_pickle_gz_multiple(objs, dirname):
     for (i, obj) in enumerate(objs):
         write_pickle_gz(obj, f"{dirname}/{i:05d}.pkl.gz")
 
@@ -73,8 +73,8 @@ def train(args):
     challenge = getattr(pkg, challenge_name)()
 
     if not os.path.isdir(f"{basepath}/train"):
-        write_multiple(challenge.training_instances, f"{basepath}/train")
-        write_multiple(challenge.test_instances, f"{basepath}/test")
+        write_pickle_gz_multiple(challenge.training_instances, f"{basepath}/train")
+        write_pickle_gz_multiple(challenge.test_instances, f"{basepath}/test")
 
     done_filename = f"{basepath}/train/done"
     if not os.path.isfile(done_filename):
@@ -114,7 +114,7 @@ def test_baseline(args):
             test_instances,
             n_jobs=int(args["--test-jobs"]),
         )
-        benchmark.save_results(csv_filename)
+        benchmark.write_csv(csv_filename)
 
 
 def test_ml(args):
@@ -148,7 +148,7 @@ def test_ml(args):
             test_instances,
             n_jobs=int(args["--test-jobs"]),
         )
-        benchmark.save_results(csv_filename)
+        benchmark.write_csv(csv_filename)
 
 
 def charts(args):
@@ -171,11 +171,11 @@ def charts(args):
     if (sense == "min").any():
         primal_column = "Relative upper bound"
         obj_column = "Upper bound"
-        predicted_obj_column = "Predicted UB"
+        predicted_obj_column = "Objective: Predicted UB"
     else:
         primal_column = "Relative lower bound"
         obj_column = "Lower bound"
-        predicted_obj_column = "Predicted LB"
+        predicted_obj_column = "Objective: Predicted LB"
 
     palette = {"baseline": "#9b59b6", "ml-exact": "#3498db", "ml-heuristic": "#95a5a6"}
     fig, (ax1, ax2, ax3, ax4) = plt.subplots(

miplearn/classifiers/__init__.py

@@ -112,8 +112,14 @@ class Regressor(ABC):
         assert isinstance(y_train, np.ndarray)
         assert x_train.dtype in [np.float16, np.float32, np.float64]
         assert y_train.dtype in [np.float16, np.float32, np.float64]
-        assert len(x_train.shape) == 2
-        assert len(y_train.shape) == 2
+        assert len(x_train.shape) == 2, (
+            f"Parameter x_train should be a square matrix. "
+            f"Found {x_train.shape} ndarray instead."
+        )
+        assert len(y_train.shape) == 2, (
+            f"Parameter y_train should be a square matrix. "
+            f"Found {y_train.shape} ndarray instead."
+        )
         (n_samples_x, n_inputs) = x_train.shape
         (n_samples_y, n_outputs) = y_train.shape
         assert n_samples_y == n_samples_x

miplearn/features.py

@@ -41,12 +41,13 @@ class FeaturesExtractor:
                     assert isinstance(user_features, list), (
                         f"Variable features must be a list. "
                         f"Found {type(user_features).__name__} instead for "
-                        f"var={var_name}."
+                        f"var={var_name}[{idx}]."
                     )
-                    assert isinstance(user_features[0], numbers.Real), (
+                    for v in user_features:
+                        assert isinstance(v, numbers.Real), (
                             f"Variable features must be a list of numbers. "
-                        f"Found {type(user_features[0]).__name__} instead "
-                        f"for var={var_name}."
+                            f"Found {type(v).__name__} instead "
+                            f"for var={var_name}[{idx}]."
                         )
                 var_dict[idx] = {
                     "Category": category,
@@ -92,4 +93,15 @@ class FeaturesExtractor:
 
     @staticmethod
     def _extract_instance(instance: "Instance") -> InstanceFeatures:
-        return {"User features": instance.get_instance_features()}
+        user_features = instance.get_instance_features()
+        assert isinstance(user_features, list), (
+            f"Instance features must be a list. "
+            f"Found {type(user_features).__name__} instead for "
+            f"var={var_name}[{idx}]."
+        )
+        for v in user_features:
+            assert isinstance(v, numbers.Real), (
+                f"Instance features must be a list of numbers. "
+                f"Found {type(v).__name__} instead."
+            )
+        return {"User features": user_features}

miplearn/log.py

@@ -5,6 +5,10 @@
 import logging
 import sys
 import time
+import warnings
+import traceback
+
+_formatwarning = warnings.formatwarning
 
 
 class TimeFormatter(logging.Formatter):
@@ -28,6 +32,13 @@ class TimeFormatter(logging.Formatter):
         )
 
 
+def formatwarning_tb(*args, **kwargs):
+    s = _formatwarning(*args, **kwargs)
+    tb = traceback.format_stack()
+    s += "".join(tb[:-1])
+    return s
+
+
 def setup_logger(start_time=None, force_color=False):
     if start_time is None:
         start_time = time.time()
@@ -49,3 +60,5 @@ def setup_logger(start_time=None, force_color=False):
     handler.setFormatter(TimeFormatter(start_time, log_colors))
     logging.getLogger().addHandler(handler)
     logging.getLogger("miplearn").setLevel(logging.INFO)
+    warnings.formatwarning = formatwarning_tb
+    logging.captureWarnings(True)

miplearn/problems/knapsack.py

@@ -85,24 +85,10 @@ class MultiKnapsackInstance(Instance):
         return model
 
     def get_instance_features(self):
-        return np.hstack(
-            [
-                np.mean(self.prices),
-                self.capacities,
-            ]
-        )
+        return [np.mean(self.prices)] + list(self.capacities)
 
     def get_variable_features(self, var, index):
-        return np.hstack(
-            [
-                self.prices[index],
-                self.weights[:, index],
-            ]
-        )
-
-
-#     def get_variable_category(self, var, index):
-#         return index
+        return [self.prices[index]] + list(self.weights[:, index])
 
 
 class MultiKnapsackGenerator: