From 577f3a3e3a39e4e7eea1f816facdf6ea86f6df1a Mon Sep 17 00:00:00 2001
From: Y-oHr-N <kon.y.ohr.n@gmail.com>
Date: Wed, 5 Feb 2020 17:00:14 +0900
Subject: [PATCH] Fix AttributeError

---
 optgbm/sklearn.py     | 38 ++++++++++++++++++--------------------
 tests/test_sklearn.py | 20 +++++++++++++++-----
 2 files changed, 33 insertions(+), 25 deletions(-)

diff --git a/optgbm/sklearn.py b/optgbm/sklearn.py
index 8114bde..5f437ae 100644
--- a/optgbm/sklearn.py
+++ b/optgbm/sklearn.py
@@ -233,24 +233,6 @@ def feature_importance(self, **kwargs: Any) -> np.ndarray:
 
 
 class _BaseOGBMModel(lgb.LGBMModel):
-    @property
-    def _param_distributions(
-        self,
-    ) -> Dict[str, optuna.distributions.BaseDistribution]:
-        if self.param_distributions is None:
-            return DEFAULT_PARAM_DISTRIBUTIONS
-
-        return self.param_distributions
-
-    @property
-    def _random_state(self) -> Optional[int]:
-        if self.random_state is None or isinstance(self.random_state, int):
-            return self.random_state
-
-        random_state = check_random_state(self.random_state)
-
-        return random_state.randint(0, MAX_INT)
-
     def __init__(
         self,
         boosting_type: str = "gbdt",
@@ -315,6 +297,22 @@ def __init__(
     def _check_is_fitted(self) -> None:
         check_is_fitted(self, "n_features_")
 
+    def _get_param_distributions(
+        self,
+    ) -> Dict[str, optuna.distributions.BaseDistribution]:
+        if self.param_distributions is None:
+            return DEFAULT_PARAM_DISTRIBUTIONS
+
+        return self.param_distributions
+
+    def _get_random_state(self) -> Optional[int]:
+        if self.random_state is None or isinstance(self.random_state, int):
+            return self.random_state
+
+        random_state = check_random_state(self.random_state)
+
+        return random_state.randint(0, MAX_INT)
+
     def fit(
         self,
         X: TWO_DIM_ARRAYLIKE_TYPE,
@@ -374,7 +372,7 @@ def fit(
         is_classifier = self._estimator_type == "classifier"
         cv = check_cv(self.cv, y, is_classifier)
 
-        seed = self._random_state
+        seed = self._get_random_state()
 
         params = self.get_params()
 
@@ -442,7 +440,7 @@ def fit(
         objective = _Objective(
             params,
             dataset,
-            self._param_distributions,
+            self._get_param_distributions(),
             eval_name,
             is_higher_better,
             callbacks=callbacks,
diff --git a/tests/test_sklearn.py b/tests/test_sklearn.py
index 52b01bd..8aaed73 100644
--- a/tests/test_sklearn.py
+++ b/tests/test_sklearn.py
@@ -15,7 +15,9 @@
 from sklearn.datasets import load_iris
 from sklearn.datasets import load_wine
 from sklearn.model_selection import train_test_split
-from sklearn.utils.estimator_checks import check_estimator
+
+# from sklearn.utils.estimator_checks import check_estimator
+from sklearn.utils.estimator_checks import check_set_params
 
 from optgbm.sklearn import OGBMClassifier
 from optgbm.sklearn import OGBMRegressor
@@ -31,14 +33,22 @@ def zero_one_loss(
     return "zero_one_loss", np.mean(y_true != y_pred), False
 
 
-@pytest.mark.skip
 def test_ogbm_classifier() -> None:
-    check_estimator(OGBMClassifier)
+    clf = OGBMClassifier()
+    name = clf.__class__.__name__
+
+    # check_estimator(clf)
+
+    check_set_params(name, clf)
 
 
-@pytest.mark.skip
 def test_ogbm_regressor() -> None:
-    check_estimator(OGBMRegressor)
+    reg = OGBMRegressor()
+    name = reg.__class__.__name__
+
+    # check_estimator(reg)
+
+    check_set_params(name, reg)
 
 
 @pytest.mark.parametrize("reg_sqrt", [False, True])