Merge pull request #6917 from ales-erjavec/fixes/owcalibratedlearner-…

…base-learner-modify [FIX] Calibrated Learner: Prevent in place modification of base learner
biolab · Oct 25, 2024 · c27a803 · c27a803
2 parents e42b9b1 + 4e7e541
commit c27a803
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 8 deletions.
diff --git a/Orange/widgets/model/owcalibratedlearner.py b/Orange/widgets/model/owcalibratedlearner.py
@@ -1,3 +1,5 @@
+import copy
+
 from Orange.classification import CalibratedLearner, ThresholdLearner, \
     NaiveBayesLearner
 from Orange.data import Table
@@ -65,7 +67,6 @@ def set_learner(self, learner):
         self.learner = self.model = None
 
     def _set_default_name(self):
-
         if self.base_learner is None:
             self.set_default_learner_name("")
         else:
@@ -80,10 +81,6 @@ def calibration_options_changed(self):
         self.apply()
 
     def create_learner(self):
-        class IdentityWrapper(Learner):
-            def fit_storage(self, data):
-                return self.base_learner.fit_storage(data)
-
         if self.base_learner is None:
             return None
         learner = self.base_learner
@@ -93,10 +90,11 @@ def fit_storage(self, data):
         if self.threshold != self.NoThresholdOptimization:
             learner = ThresholdLearner(learner,
                                        self.ThresholdMap[self.threshold])
+        if learner is self.base_learner:
+            learner = copy.deepcopy(learner)
         if self.preprocessors:
-            if learner is self.base_learner:
-                learner = IdentityWrapper()
             learner.preprocessors = (self.preprocessors, )
+        assert learner is not self.base_learner
         return learner
 
     def get_learner_parameters(self):

diff --git a/Orange/widgets/model/tests/test_owcalibratedlearner.py b/Orange/widgets/model/tests/test_owcalibratedlearner.py
@@ -95,7 +95,7 @@ def test_create_learner(self):
         widget.calibration = widget.NoCalibration
         widget.threshold = widget.NoThresholdOptimization
         learner = self.widget.create_learner()
-        self.assertIs(learner, self.widget.base_learner)
+        self.assertIsNot(learner, self.widget.base_learner)
 
         widget.calibration = widget.SigmoidCalibration
         widget.threshold = widget.OptimizeF1