add a test for testing all _most_frequent options

DavidKatz-il · DavidKatz-il · commit 20b5c87b3bb5 · 2020-12-11T01:09:29.000+02:00
diff --git a/sklearn/impute/tests/test_impute.py b/sklearn/impute/tests/test_impute.py
@@ -27,6 +27,7 @@
 from sklearn import tree
 from sklearn.random_projection import _sparse_random_matrix
 from sklearn.exceptions import ConvergenceWarning
+from .._base import _most_frequent
 
 
 def _check_statistics(X, X_true,
@@ -1474,3 +1475,29 @@ def test_simple_imputation_inverse_transform_exceptions(missing_value):
     with pytest.raises(ValueError,
                        match=f"Got 'add_indicator={imputer.add_indicator}'"):
         imputer.inverse_transform(X_1_trans)
+
+
+def test_most_frequent():
+    # collections.Counter
+    assert np.isnan(_most_frequent(np.array([]), 'extra_value', 0))
+    assert 'extra_value' == _most_frequent(
+        np.array(['a', 'b', 'c'], dtype=object), 'extra_value', 2
+    )
+    assert 'most_frequent' == _most_frequent(
+        np.array(
+            ['most_frequent_value', 'most_frequent_value', 'value'],
+            dtype=object), 'extra_value', 1
+    )
+    assert 'a' == _most_frequent(np.array(
+        ['min_value', 'min_value' 'value'], dtype=object), 'a', 2
+    )
+    assert 'min_value' == _most_frequent(
+        np.array(['min_value', 'min_value', 'value'], dtype=object), 'z', 2
+    )
+
+    # scipy.stats.mode
+    assert np.isnan(_most_frequent(np.array([]), 10, 0))
+    assert 10 == _most_frequent(np.array([1, 2, 3]), 10, 2)
+    assert 1 == _most_frequent(np.array([1, 1, 2]), 10, 1)
+    assert 10 == _most_frequent(np.array([20, 20, 1]), 10, 2)
+    assert 1 == _most_frequent(np.array([1, 1, 20], dtype=object), 10, 2)
