OWTreeLearner: report error instead of crashing when can't binarize

Author: janezd

Orange/widgets/classify/owclassificationtree.py

@@ -9,6 +9,7 @@
 from Orange.widgets import gui
 from Orange.widgets.settings import Setting
 from Orange.widgets.utils.owlearnerwidget import OWBaseLearner
+from Orange.widgets.widget import Msg
 
 
 class OWTreeLearner(OWBaseLearner):
@@ -29,6 +30,12 @@ class OWTreeLearner(OWBaseLearner):
         ("Limit the maximal tree depth to: ",
          "limit_depth", "max_depth", 1, 1000))
 
+    class Error(OWBaseLearner.Error):
+        cannot_binarize = Msg("Binarization cannot handle '{}'\n"
+                              "because it has {} values. "
+                              "Binarization can handle up to {}.\n"
+                              "Disable 'Induce binary tree' to proceed.")
+
     def add_main_layout(self):
         box = gui.vBox(self.controlArea, True)
         # the checkbox is put into vBox for alignemnt with other checkboxes
@@ -49,6 +56,20 @@ def learner_kwargs(self):
             binarize=self.binary_trees,
             preprocessors=self.preprocessors)
 
+    def check_data(self):
+        self.Error.cannot_binarize.clear()
+        if not super().check_data():
+            return False
+        max_values, max_attr = max(
+            ((len(attr.values), attr)
+             for attr in self.data.domain.attributes if attr.is_discrete),
+            default=(0, None))
+        if max_values > self.LEARNER.MAX_BINARIZATION:
+            self.Error.cannot_binarize(
+                max_attr.name, max_values, self.LEARNER.MAX_BINARIZATION)
+            return False
+        return True
+
     def create_learner(self):
         # pylint: disable=not-callable
         return self.LEARNER(**self.learner_kwargs())

Orange/widgets/classify/tests/test_owclassificationtree.py

@@ -1,12 +1,18 @@
 # Test methods with long descriptive names can omit docstrings
 # pylint: disable=missing-docstring
+from Orange.data import Table, Domain, DiscreteVariable
 from Orange.base import Model
 from Orange.widgets.classify.owclassificationtree import OWClassificationTree
 from Orange.widgets.tests.base import (WidgetTest, DefaultParameterMapping,
                                        ParameterMapping, WidgetLearnerTestMixin)
 
 
 class TestOWClassificationTree(WidgetTest, WidgetLearnerTestMixin):
+    @classmethod
+    def setUpClass(cls):
+        super().setUpClass()
+        cls.iris = Table("iris")
+
     def setUp(self):
         self.widget = self.create_widget(OWClassificationTree,
                                          stored_settings={"auto_apply": False})
@@ -34,3 +40,25 @@ def test_parameters_unchecked(self):
                            for par, val in zip(self.parameters, (None, 2, 1))]
         self.test_parameters()
 
+    def test_cannot_binarize(self):
+        widget = self.widget
+        error_shown = self.widget.Error.cannot_binarize.is_shown
+        self.assertFalse(error_shown())
+        self.send_signal("Data", self.iris)
+        self.assertFalse(error_shown())
+        domain = Domain([
+            DiscreteVariable(
+                values=[str(x)
+                        for x in range(widget.LEARNER.MAX_BINARIZATION + 1)])],
+            DiscreteVariable(values="01"))
+        self.send_signal("Data", Table(domain, [[0, 0], [1, 1]]))
+        self.assertTrue(error_shown())
+        self.send_signal("Data", self.iris)
+        self.assertFalse(error_shown())
+        domain = Domain([
+            DiscreteVariable(
+                values=[str(x)
+                        for x in range(widget.LEARNER.MAX_BINARIZATION + 1)])],
+            DiscreteVariable(values="01"))
+        self.send_signal("Data", Table(domain))
+        self.assertFalse(error_shown())