Deprecation warning about multiclass, adapt tests, remove liblinear w…

…ith multiple classes (because that is going to be unsupported)

Orange/classification/logistic_regression.py

@@ -1,10 +1,14 @@
+import warnings
+
 import numpy as np
 import sklearn.linear_model as skl_linear_model
 
 from Orange.classification import SklLearner, SklModel
 from Orange.preprocess import Normalize
 from Orange.preprocess.score import LearnerScorer
 from Orange.data import Variable, DiscreteVariable
+from Orange.util import OrangeDeprecationWarning
+
 
 __all__ = ["LogisticRegressionLearner"]
 
@@ -38,7 +42,15 @@ class LogisticRegressionLearner(SklLearner, _FeatureScorerMixin):
     def __init__(self, penalty="l2", dual=False, tol=0.0001, C=1.0,
                  fit_intercept=True, intercept_scaling=1, class_weight=None,
                  random_state=None, solver="auto", max_iter=100,
-                 verbose=0, n_jobs=1, preprocessors=None):
+                 multi_class="deprecated", verbose=0, n_jobs=1, preprocessors=None):
+        if multi_class == "deprecated":
+            del multi_class
+        else:
+            warnings.warn("The multi_class parameter was "
+                          "deprecated in scikit-learn 1.5. Using it with "
+                          "scikit-learn 1.7 will lead to a crash.",
+                          OrangeDeprecationWarning,
+                          stacklevel=2)
         super().__init__(preprocessors=preprocessors)
         self.params = vars()
 

Orange/evaluation/scoring.py

@@ -5,7 +5,7 @@
 --------
 >>> import Orange
 >>> data = Orange.data.Table('iris')
->>> learner = Orange.classification.LogisticRegressionLearner(solver="liblinear")
+>>> learner = Orange.classification.LogisticRegressionLearner()
 >>> results = Orange.evaluation.TestOnTrainingData(data, [learner])
 
 """

Orange/tests/test_logistic_regression.py

@@ -149,7 +149,7 @@ def test_auto_solver(self):
         # liblinear is default for l2 penalty
         lr = LogisticRegressionLearner(penalty="l1", solver="auto")
         skl_clf = lr._initialize_wrapped()
-        self.assertEqual(skl_clf.solver, "liblinear")
+        self.assertEqual(skl_clf.solver, "saga")
         self.assertEqual(skl_clf.penalty, "l1")
 
     def test_supports_weights(self):

Orange/widgets/evaluate/tests/test_owpredictions.py

@@ -1303,7 +1303,7 @@ def test_output_error_cls(self):
         log_reg = LogisticRegressionLearner()
         self.send_signal(self.widget.Inputs.predictors, log_reg(data), 0)
         self.send_signal(self.widget.Inputs.predictors,
-                         LogisticRegressionLearner(penalty="l1")(data), 1)
+                         LogisticRegressionLearner(penalty="l1", max_iter=1000)(data), 1)
         with data.unlocked(data.Y):
             data.Y[1] = np.nan
         self.send_signal(self.widget.Inputs.data, data)

Orange/widgets/visualize/tests/test_ownomogram.py

@@ -97,10 +97,8 @@ def test_nomogram_nb_multiclass(self):
     def test_nomogram_lr_multiclass(self):
         """Check probabilities for logistic regression classifier for various
         values of classes and radio buttons for multiclass data"""
-        cls = LogisticRegressionLearner(
-            solver="liblinear"
-        )(self.lenses)
-        self._test_helper(cls, [9, 45, 52])
+        cls = LogisticRegressionLearner(max_iter=100)(self.lenses)
+        self._test_helper(cls, [18, 56, 78])
 
     def test_nomogram_with_instance_nb(self):
         """Check initialized marker values and feature sorting for naive bayes