test renaming

tests/test_integration.py

@@ -29,25 +29,8 @@
 np.random.seed(42)
 
 
-def test_interactions_selection(prepare_data, feature_properties, features, is_plot):
-    X, y = prepare_data
-
-    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
-
-    best_interaction_term_features = select_interaction_terms_anova(X_train, y_train, feature_properties, 3, 5)
-
-    expected = [
-        ("PG_ID_3", "L_ID"),
-        ("PG_ID_3", "PROMOTION_TYPE"),
-        ("L_ID", "PROMOTION_TYPE"),
-        ("PG_ID_3", "L_ID", "dayofweek"),
-        ("PG_ID_3", "L_ID", "PROMOTION_TYPE"),
-    ]
-    assert best_interaction_term_features == expected
-
-    features_ext = features.copy()
-    features_ext += best_interaction_term_features
-
+@pytest.fixture(scope="function")
+def cb_poisson_regressor_model(features, feature_properties):
     explicit_smoothers = {
         ("dayofyear",): SeasonalSmoother(order=3),
         ("price_ratio",): IsotonicRegressor(increasing=False),
@@ -58,15 +41,25 @@ def test_interactions_selection(prepare_data, feature_properties, features, is_p
         observers.PlottingObserver(iteration=-1),
     ]
 
-    CB_est = pipeline_CBPoissonRegressor(
+    CB_pipeline = pipeline_CBPoissonRegressor(
         feature_properties=feature_properties,
-        feature_groups=features_ext,
+        feature_groups=features,
         observers=plobs,
         maximal_iterations=50,
         smoother_choice=common_smoothers.SmootherChoiceGroupBy(
             use_regression_type=True, use_normalization=False, explicit_smoothers=explicit_smoothers
         ),
     )
+
+    return CB_pipeline
+
+
+def test_poisson_regression(is_plot, prepare_data, cb_poisson_regressor_model):
+    X, y = prepare_data
+
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
+
+    CB_est = cb_poisson_regressor_model
     CB_est.fit(X_train, y_train)
 
     if is_plot:
@@ -76,11 +69,28 @@ def test_interactions_selection(prepare_data, feature_properties, features, is_p
     yhat = CB_est.predict(X_test)
 
     mad = np.nanmean(np.abs(y_test - yhat))
-    np.testing.assert_almost_equal(mad, 1.641, 3)
+    np.testing.assert_almost_equal(mad, 1.688, 3)
 
 
-@pytest.fixture(scope="function")
-def cb_poisson_regressor_model(features, feature_properties):
+def test_poisson_regression_interactions_selection(prepare_data, feature_properties, features, is_plot):
+    X, y = prepare_data
+
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
+
+    best_interaction_term_features = select_interaction_terms_anova(X_train, y_train, feature_properties, 3, 5)
+
+    expected = [
+        ("PG_ID_3", "L_ID"),
+        ("PG_ID_3", "PROMOTION_TYPE"),
+        ("L_ID", "PROMOTION_TYPE"),
+        ("PG_ID_3", "L_ID", "dayofweek"),
+        ("PG_ID_3", "L_ID", "PROMOTION_TYPE"),
+    ]
+    assert best_interaction_term_features == expected
+
+    features_ext = features.copy()
+    features_ext += best_interaction_term_features
+
     explicit_smoothers = {
         ("dayofyear",): SeasonalSmoother(order=3),
         ("price_ratio",): IsotonicRegressor(increasing=False),
@@ -91,25 +101,15 @@ def cb_poisson_regressor_model(features, feature_properties):
         observers.PlottingObserver(iteration=-1),
     ]
 
-    CB_pipeline = pipeline_CBPoissonRegressor(
+    CB_est = pipeline_CBPoissonRegressor(
         feature_properties=feature_properties,
-        feature_groups=features,
+        feature_groups=features_ext,
         observers=plobs,
         maximal_iterations=50,
         smoother_choice=common_smoothers.SmootherChoiceGroupBy(
             use_regression_type=True, use_normalization=False, explicit_smoothers=explicit_smoothers
         ),
     )
-
-    return CB_pipeline
-
-
-def test_poisson_regression(is_plot, prepare_data, cb_poisson_regressor_model):
-    X, y = prepare_data
-
-    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
-
-    CB_est = cb_poisson_regressor_model
     CB_est.fit(X_train, y_train)
 
     if is_plot:
@@ -119,7 +119,7 @@ def test_poisson_regression(is_plot, prepare_data, cb_poisson_regressor_model):
     yhat = CB_est.predict(X_test)
 
     mad = np.nanmean(np.abs(y_test - yhat))
-    np.testing.assert_almost_equal(mad, 1.688, 3)
+    np.testing.assert_almost_equal(mad, 1.641, 3)
 
 
 @pytest.fixture(scope="function")