TST Add test to check if estimators reset model when fit is called

raghavrv · raghavrv · commit 2183d28613e1 · 2015-01-27T19:36:57.000+05:30
diff --git a/sklearn/tests/test_common.py b/sklearn/tests/test_common.py
@@ -32,6 +32,7 @@
     check_estimator_sparse_data,
     check_transformer,
     check_clustering,
+    check_fit_reset,
     check_clusterer_compute_labels_predict,
     check_regressors_int,
     check_regressors_train,
@@ -100,6 +101,8 @@ def test_non_meta_estimators():
             yield check_sparsify_coefficients, name, Estimator
 
         yield check_estimator_sparse_data, name, Estimator
+        # test if fit resets model
+        yield check_fit_reset, name, Alg
 
 
 def test_transformers():
diff --git a/sklearn/utils/estimator_checks.py b/sklearn/utils/estimator_checks.py
@@ -1004,3 +1004,35 @@ def check_transformer_n_iter(name, estimator):
             assert_greater(iter_, 1)
     else:
         assert_greater(estimator.n_iter_, 1)
+
+
+@ignore_warnings
+def check_fit_reset(name, Estimator):
+    X1, y1 = make_blobs(n_samples=50, n_features=2, center_box=(-200, -150),
+                        centers=2, random_state=0)
+    X2, y2 = make_blobs(n_samples=100, n_features=3, center_box=(-1, 1),
+                        centers=1, random_state=1)
+    X3, y3 = make_blobs(n_samples=200, n_features=4, center_box=(-100, -50),
+                        centers=5, random_state=2)
+    X4, y4 = make_blobs(n_samples=150, n_features=5, center_box=(50, 100),
+                        centers=10, random_state=3)
+
+    estimator_1 = Estimator()
+    estimator_2 = Estimator()
+
+    set_fast_parameters(estimator_1)
+    set_fast_parameters(estimator_2)
+
+    set_random_state(estimator_1)
+    set_random_state(estimator_2)
+
+    _fit(estimator_1, X1, y1)
+    _fit(estimator_2, X3, y3)
+    assert_not_same_model(estimator_1, estimator_2)
+
+    _fit(estimator_2, X4, y4)
+    assert_not_same_model(estimator_1, estimator_2)
+
+    _fit(estimator_1, X2, y2)
+    _fit(estimator_2, X2, y2)
+    assert_same_model(estimator_1, estimator_2)
