Add tests, pylint fixes

Author: AndrejaKovacic

Orange/preprocess/preprocess.py

@@ -606,7 +606,11 @@ def __init__(self,
                  normalize_datetime=False,
                  center=Scale.Mean,
                  scale=Scale.Std):
-        self.normalize_pps = Normalize(zero_based, norm_type, transform_class, ~(center is not False), normalize_datetime)
+        self.normalize_pps = Normalize(zero_based,
+                                       norm_type,
+                                       transform_class,
+                                       ~(center is not False),
+                                       normalize_datetime)
         self.scale_pps = Scale(center=Scale.NoCentering, scale=scale)
 
     def __call__(self, data):

Orange/tests/test_preprocess.py

@@ -14,7 +14,7 @@
     SelectRandomFeatures, EqualFreq, RemoveNaNColumns, DropInstances, \
     EqualWidth, SelectBestFeatures, RemoveNaNRows, Preprocess, Scale, \
     Randomize, Continuize, Discretize, Impute, SklImpute, Normalize, \
-    ProjectCUR, ProjectPCA, RemoveConstant
+    ProjectCUR, ProjectPCA, RemoveConstant, SmartNormalize
 from Orange.util import OrangeDeprecationWarning
 
 
@@ -166,3 +166,25 @@ def test_scaling_pickling(self):
         c1 = pickle.loads(s)
         self.assertIs(c1.center, c.center)
         self.assertIs(c1.scale, c.scale)
+
+
+class TestSmartNormalize(unittest.TestCase):
+    """
+    Checks if output for sparse data is the same as for Scale
+    preprocessor. For dense data the output should match that
+     of Normalize preprocessor.
+    """
+
+    def test_dense_pps(self):
+        data = Table("iris")
+        true_out = Normalize()(data)
+        out = SmartNormalize()(data)
+        np.testing.assert_array_equal(out, true_out)
+
+    def test_sparse_pps(self):
+        data = Table("iris")
+        data.X = csr_matrix(data.X)
+        out = SmartNormalize()(data)
+        true_out = Scale(center=Scale.NoCentering)(data)
+        np.testing.assert_array_equal(out, true_out)
+

Orange/widgets/model/owsvm.py

@@ -32,7 +32,7 @@ class Outputs(OWBaseLearner.Outputs):
         support_vectors = Output("Support vectors", Table, explicit=True)
 
     class Warning(OWWidget.Warning):
-        sparse_data = Msg('Input data is sparse, default preprocessing is to scale it. If you want mean centering, use preprocessing widget.')
+        sparse_data = Msg('Input data is sparse, default preprocessing is to scale it.')
         outdated_learner = Msg("Press Apply to submit changes.")
 
     #: Different types of SVMs

Orange/widgets/model/tests/test_owsvm.py

@@ -1,12 +1,15 @@
 # Test methods with long descriptive names can omit docstrings
 # pylint: disable=missing-docstring
+from scipy.sparse import csr_matrix
+
 from Orange.widgets.model.owsvm import OWSVM
 from Orange.widgets.tests.base import (
     WidgetTest,
     DefaultParameterMapping,
     ParameterMapping,
     WidgetLearnerTestMixin
 )
+from Orange.data import Table
 
 
 class TestOWSVMClassification(WidgetTest, WidgetLearnerTestMixin):
@@ -87,3 +90,16 @@ def test_kernel_spins(self):
             self.widget.kernel_box.buttons[i].click()
             self.assertEqual([self.widget._kernel_params[j].box.isHidden()
                               for j in range(3)], hidden)
+
+    def test_sparse_warning(self):
+        """Check if the user is warned about sparse input"""
+        data = Table("iris")
+        data.X = csr_matrix(data.X)
+        self.send_signal("Data", data)
+        self.assertTrue(self.widget.Warning.sparse_data.is_shown())
+
+    def test_dense_input(self):
+        data = Table("iris")
+        self.send_signal("Data", data)
+        self.assertFalse(self.widget.Warning.sparse_data.is_shown())
+