added _check__ambiguous()

smarttree/_check.py

@@ -58,6 +58,18 @@ def check__params(
     if rank_features is not None:
         _check__rank_features(rank_features)
 
+    if num_features is not None and cat_features is not None:
+        param_name1, param_name2 = "num_features", "cat_features"
+        _check__ambiguous(num_features, cat_features, param_name1, param_name2)
+
+    if num_features is not None and rank_features is not None:
+        param_name1, param_name2 = "num_features", "rank_features"
+        _check__ambiguous(num_features, list(rank_features.keys()), param_name1, param_name2)
+
+    if cat_features is not None and rank_features is not None:
+        param_name1, param_name2 = "cat_features", "rank_features"
+        _check__ambiguous(cat_features, list(rank_features.keys()), param_name1, param_name2)
+
     if hierarchy is not None:
         _check__hierarchy(hierarchy)
 
@@ -221,6 +233,17 @@ def _check__features_contain_duplicates(param_name, features):
         raise ValueError(f"`{param_name}` contains duplicates.")
 
 
+def _check__ambiguous(features1, features2, param_name1, param_name2):
+    set_features1 = {features1} if isinstance(features1, str) else set(features1)
+    set_features2 = {features2} if isinstance(features2, str) else set(features2)
+    intersection = set_features1 & set_features2
+    if intersection:
+        raise ValueError(
+            "Following feature names are ambiguous, they are defined in both"
+            f" '{param_name1}' and '{param_name2}': {intersection}."
+        )
+
+
 def _check__hierarchy(hierarchy):
     common_message = (
         "`hierarchy` must be a dictionary"

tests/decision_tree/base/test__check_params.py

@@ -442,6 +442,57 @@ def test__check_params__features_contain_duplicates(params_to_set, expected_cont
         SmartDecisionTreeClassifier(**params_to_set)
 
 
+@pytest.mark.parametrize(
+    ("n_features", "c_features", "r_features", "expected_context"),
+    [
+        ("f1", "f2", None, does_not_raise()),
+        (
+            "f1",
+            "f1",
+            None,
+            pytest.raises(
+                ValueError,
+                match=(
+                    "Following feature names are ambiguous, they are defined in"
+                    " both 'num_features' and 'cat_features': {'f1'}."
+                ),
+            ),
+        ),
+        ("f1", None, {"f2": ["v2"]}, does_not_raise()),
+        (
+            "f1",
+            None,
+            {"f1": ["v1"]},
+            pytest.raises(
+                ValueError,
+                match=(
+                    "Following feature names are ambiguous, they are defined in"
+                    " both 'num_features' and 'rank_features': {'f1'}."
+                ),
+            ),
+        ),
+        (None, "f1", {"f2": ["v2"]}, does_not_raise()),
+        (
+            None,
+            "f1",
+            {"f1": ["v1"]},
+            pytest.raises(
+                ValueError,
+                match=(
+                    "Following feature names are ambiguous, they are defined in"
+                    " both 'cat_features' and 'rank_features': {'f1'}."
+                ),
+            ),
+        ),
+    ],
+)
+def test__check_params__ambiguous(n_features, c_features, r_features, expected_context):
+    with expected_context:
+        SmartDecisionTreeClassifier(
+            num_features=n_features, cat_features=c_features, rank_features=r_features
+        )
+
+
 @pytest.mark.parametrize(
     ("hierarchy", "expected_context"),
     [