add exception if outcome is not binary when classifier is used

Author: SvenKlaassen

doubleml/plm/plr.py

@@ -97,7 +97,7 @@ def __init__(
         valid_scores = ["IV-type", "partialling out"]
         _check_score(self.score, valid_scores, allow_callable=True)
 
-        _ = self._check_learner(ml_l, "ml_l", regressor=True, classifier=False)
+        ml_l_is_classifier = self._check_learner(ml_l, "ml_l", regressor=True, classifier=True)
         ml_m_is_classifier = self._check_learner(ml_m, "ml_m", regressor=True, classifier=True)
         self._learner = {"ml_l": ml_l, "ml_m": ml_m}
 
@@ -117,7 +117,17 @@ def __init__(
             warnings.warn(("For score = 'IV-type', learners ml_l and ml_g should be specified. Set ml_g = clone(ml_l)."))
             self._learner["ml_g"] = clone(ml_l)
 
-        self._predict_method = {"ml_l": "predict"}
+        if ml_l_is_classifier:
+            if obj_dml_data.binary_outcome:
+                self._predict_method = {"ml_g": "predict_proba"}
+            else:
+                raise ValueError(
+                    f"The ml_l learner {str(ml_l)} was identified as classifier "
+                    "but the outcome variable is not binary with values 0 and 1."
+                )
+        else:
+            self._predict_method = {"ml_l": "predict"}
+
         if "ml_g" in self._learner:
             self._predict_method["ml_g"] = "predict"
         if ml_m_is_classifier:

doubleml/plm/tests/test_plr_binary_outcome.py

@@ -54,12 +54,20 @@ def generate_binary_data():
 def test_dml_plr_binary_warnings(generate_binary_data, learner_binary, score):
     data = generate_binary_data
     obj_dml_data = dml.DoubleMLData(data, "y", ["d"])
-    msg = "The ml_m learner {str(ml_m)} was identified as classifier ' \
-                          'but at least one treatment variable is not binary with values 0 and 1.'"
+    msg = "The ml_l learner {str(ml_l)} was identified as classifier"
     with pytest.warns(UserWarning, match=msg):
         _ = dml.DoubleMLPLR(obj_dml_data, clone(learner_binary), clone(learner_binary), score=score)
 
 
+@pytest.mark.ci
+def test_dml_plr_binary_exceptions(generate_binary_data, learner_binary, score):
+    data = generate_binary_data
+    obj_dml_data = dml.DoubleMLData(data, "X1", ["d"])
+    msg = "The ml_l learner .+ was identified as classifier but the outcome variable is not binary with values 0 and 1."
+    with pytest.raises(ValueError, match=msg):
+        _ = dml.DoubleMLPLR(obj_dml_data, clone(learner_binary), clone(learner_binary), score=score)
+
+
 @pytest.fixture(scope="module")
 def dml_plr_binary_fixture(generate_binary_data, learner_binary, score):
     boot_methods = ["normal"]