changes in tests

Author: Tomasz Latkowski

config/main.ini

@@ -1,6 +1,21 @@
 import tensorflow as tf
 
 
+def selection_wrapper(data, num_instances, selection_method=None, num_features=None):
+    if data is None:
+        raise ValueError('Provide data to make selection.')
+
+    if selection_method is None:
+        raise ValueError('Provide selection method.')
+
+    if num_features is None:
+        data = tf.convert_to_tensor(data)
+        num_features = data.get_shape().as_list()[-1]
+
+    values, indices = selection_method(data, num_instances, num_features)
+    return values, tf.gather(data, indices, axis=1)
+
+
 def fisher(data, num_instances: list, top_k_features=2):
     """
     Performs Fisher feature selection method according to the following formula:
@@ -20,9 +35,7 @@ def fisher(data, num_instances: list, top_k_features=2):
     mean1, std1 = tf.nn.moments(class1, axes=0)
     mean2, std2 = tf.nn.moments(class2, axes=0)
     fisher_coeffs = tf.abs(mean1 - mean2) / (std1 + std2)
-    values, indices = tf.nn.top_k(fisher_coeffs, k=top_k_features)
-    most_sig_f = tf.gather(data, indices, axis=1)
-    return most_sig_f
+    return tf.nn.top_k(fisher_coeffs, k=top_k_features)
 
 
 def feature_correlation_with_class(data, num_instances: list, top_k_features=10):

methods/selection.py

@@ -1,7 +1,7 @@
 import numpy as np
 import tensorflow as tf
 
-from methods.selection import fisher
+from methods.selection import fisher, selection_wrapper
 
 
 class TestFisherSelection(tf.test.TestCase):
@@ -14,29 +14,52 @@ def testFisherCorrectScore(self):
                              [5, 6]])
             num_instances = [2, 2]
             top_k = 2
-            actual_most_significant_features = test_session.run(fisher(data, num_instances, top_k))
+            actual_most_significant_features, _ = test_session.run(fisher(data, num_instances, top_k))
             correct_most_significant_features = [3., .5]
 
-            self.assertAllEqual(actual_most_significant_features.values, correct_most_significant_features)
+            self.assertAllEqual(actual_most_significant_features, correct_most_significant_features)
 
     def testFisherPickFirstSignificantFeature(self):
         with self.test_session() as test_session:
             data = np.array([[2, 2],
                              [4, 4],
                              [3, 6],
                              [5, 6]])
+
             num_instances = [2, 2]
             top_k = 1
-            actual_most_significant_features = test_session.run(fisher(data, num_instances, top_k))
+            _, actual_most_significant_features = test_session.run(selection_wrapper(data,
+                                                                                     num_instances,
+                                                                                     fisher,
+                                                                                     num_features=top_k))
             correct_most_significant_features = [[2.], [4.], [6.], [6.]]
 
             self.assertAllEqual(actual_most_significant_features, correct_most_significant_features)
 
     def testFisherCorrectOrderOfFeatures(self):
-        raise NotImplementedError
+        with self.test_session() as test_session:
+            data = np.array([[2, 2],
+                             [4, 4],
+                             [3, 6],
+                             [5, 6]])
+            num_instances = [2, 2]
+            top_k = 2
+            _, actual_most_significant_features = test_session.run(fisher(data, num_instances, top_k))
+            correct_most_significant_features = [1., 0.]
+
+            self.assertAllEqual(actual_most_significant_features, correct_most_significant_features)
 
     def testMoreThan2ClassesIsNotAllowed(self):
-        raise NotImplementedError
+        with self.test_session() as test_session:
+            data = np.array([[2, 2],
+                             [4, 4],
+                             [3, 6],
+                             [5, 6]])
+            num_instances = [2, 2, 2]
+            top_k = 2
+            with self.assertRaises(AssertionError):
+                _, actual_most_significant_features = test_session.run(fisher(data, num_instances, top_k))
+
 
 if __name__ == '__main__':
     tf.test.main()

tests/test_fisher.py

@@ -9,25 +9,32 @@ def testPearsonCoefficientValueForTwoVectors(self):
         with self.test_session() as test_session:
             x1 = np.array([2., 3., 4.])
             x2 = np.array([3., 1., 5.])
+
             actual_pearson_coefficient = test_session.run(pearson_correlation(x1, x2))
-            correct_pearson_coefficient = tf.constant([.5])
-            self.assertEqual(actual_pearson_coefficient, correct_pearson_coefficient.eval())
+            correct_pearson_coefficient = [.5]
+
+            self.assertEqual(actual_pearson_coefficient, correct_pearson_coefficient)
 
     def testNegativePearsonCoefficientValueForTwoVectors(self):
         with self.test_session() as test_session:
             x1 = np.array([1., 2., 3.])
             x2 = np.array([-1., -2., -3.])
+
             actual_pearson_coefficient = test_session.run(pearson_correlation(x1, x2))
-            correct_pearson_coefficient = tf.constant([-1.])
-            self.assertEqual(actual_pearson_coefficient, correct_pearson_coefficient.eval())
+            correct_pearson_coefficient = [-1.]
+
+            self.assertEqual(actual_pearson_coefficient, correct_pearson_coefficient)
 
     def testPositivePearsonCoefficientValueForTwoVectors(self):
         with self.test_session() as test_session:
             x1 = np.array([1., 2., 3.])
             x2 = np.array([1., 2., 3.])
+
             actual_pearson_coefficient = test_session.run(pearson_correlation(x1, x2))
-            correct_pearson_coefficient = tf.constant([1.])
-            self.assertEqual(actual_pearson_coefficient, correct_pearson_coefficient.eval())
+            correct_pearson_coefficient = [1.]
+
+            self.assertEqual(actual_pearson_coefficient, correct_pearson_coefficient)
+
 
 if __name__ == '__main__':
     tf.test.main()