minimal fix to resolve #707

causalml/metrics/visualize.py

@@ -84,6 +84,10 @@ def get_cumlift(
         or treatment_effect_col in df.columns
     )
 
+    assert not (
+        (df[[outcome_col, treatment_col, treatment_effect_col]].isnull().values.any())
+    )
+
     df = df.copy()
     np.random.seed(random_seed)
     random_cols = []
@@ -219,6 +223,10 @@ def get_qini(
         or treatment_effect_col in df.columns
     )
 
+    assert not (
+        (df[[outcome_col, treatment_col, treatment_effect_col]].isnull().values.any())
+    )
+
     df = df.copy()
     np.random.seed(random_seed)
     random_cols = []
@@ -315,6 +323,8 @@ def get_tmlegain(
         or p_col in df.columns
     )
 
+    assert not ((df[[outcome_col, treatment_col, p_col]].isnull().values.any()))
+
     inference_col = [x for x in inference_col if x in df.columns]
 
     # Initialize TMLE
@@ -421,6 +431,8 @@ def get_tmleqini(
         or p_col in df.columns
     )
 
+    assert not ((df[[outcome_col, treatment_col, p_col]].isnull().values.any()))
+
     inference_col = [x for x in inference_col if x in df.columns]
 
     # Initialize TMLE

tests/test_cevae.py

@@ -38,9 +38,7 @@ def test_CEVAE():
     # check the accuracy of the ite accuracy
     ite = cevae.predict(X).flatten()
 
-    auuc_metrics = pd.DataFrame(
-        {"ite": ite, "W": treatment, "y": y, "treatment_effect_col": tau}
-    )
+    auuc_metrics = pd.DataFrame({"ite": ite, "W": treatment, "y": y, "tau": tau})
 
     cumgain = get_cumgain(
         auuc_metrics, outcome_col="y", treatment_col="W", treatment_effect_col="tau"

tests/test_ivlearner.py

@@ -71,7 +71,7 @@ def test_drivlearner():
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 

tests/test_meta_learners.py

@@ -123,7 +123,7 @@ def test_BaseSRegressor(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -179,7 +179,7 @@ def test_BaseTLearner(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -229,7 +229,7 @@ def test_BaseTRegressor(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -268,7 +268,7 @@ def test_MLPTRegressor(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -307,7 +307,7 @@ def test_XGBTRegressor(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -346,7 +346,7 @@ def test_BaseXLearner(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -398,7 +398,7 @@ def test_BaseXRegressor(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -439,7 +439,7 @@ def test_BaseXLearner_without_p(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -478,7 +478,7 @@ def test_BaseXRegressor_without_p(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -517,7 +517,7 @@ def test_BaseRLearner(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -568,7 +568,7 @@ def test_BaseRRegressor(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -607,7 +607,7 @@ def test_BaseRLearner_without_p(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -646,7 +646,7 @@ def test_BaseRRegressor_without_p(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 
@@ -698,15 +698,15 @@ def test_BaseSClassifier(generate_classification_data):
             "tau_pred": tau_pred.flatten(),
             "W": df_test["treatment_group_key"].values,
             CONVERSION: df_test[CONVERSION].values,
-            "treatment_effect_col": df_test["treatment_effect"].values,
+            "tau": df_test["treatment_effect"].values,
         }
     )
 
     cumgain = get_cumgain(
         auuc_metrics,
         outcome_col=CONVERSION,
         treatment_col="W",
-        treatment_effect_col="treatment_effect_col",
+        treatment_effect_col="tau",
     )
 
     # Check if the cumulative gain when using the model's prediction is
@@ -742,15 +742,15 @@ def test_BaseTClassifier(generate_classification_data):
             "tau_pred": tau_pred.flatten(),
             "W": df_test["treatment_group_key"].values,
             CONVERSION: df_test[CONVERSION].values,
-            "treatment_effect_col": df_test["treatment_effect"].values,
+            "tau": df_test["treatment_effect"].values,
         }
     )
 
     cumgain = get_cumgain(
         auuc_metrics,
         outcome_col=CONVERSION,
         treatment_col="W",
-        treatment_effect_col="treatment_effect_col",
+        treatment_effect_col="tau",
     )
 
     # Check if the cumulative gain when using the model's prediction is
@@ -812,15 +812,15 @@ def test_BaseXClassifier(generate_classification_data):
             "tau_pred": tau_pred.flatten(),
             "W": df_test["treatment_group_key"].values,
             CONVERSION: df_test[CONVERSION].values,
-            "treatment_effect_col": df_test["treatment_effect"].values,
+            "tau": df_test["treatment_effect"].values,
         }
     )
 
     cumgain = get_cumgain(
         auuc_metrics,
         outcome_col=CONVERSION,
         treatment_col="W",
-        treatment_effect_col="treatment_effect_col",
+        treatment_effect_col="tau",
     )
 
     # Check if the cumulative gain when using the model's prediction is
@@ -861,15 +861,15 @@ def test_BaseRClassifier(generate_classification_data):
             "tau_pred": tau_pred.flatten(),
             "W": df_test["treatment_group_key"].values,
             CONVERSION: df_test[CONVERSION].values,
-            "treatment_effect_col": df_test["treatment_effect"].values,
+            "tau": df_test["treatment_effect"].values,
         }
     )
 
     cumgain = get_cumgain(
         auuc_metrics,
         outcome_col=CONVERSION,
         treatment_col="W",
-        treatment_effect_col="treatment_effect_col",
+        treatment_effect_col="tau",
     )
 
     # Check if the cumulative gain when using the model's prediction is
@@ -912,15 +912,15 @@ def test_BaseRClassifier_with_sample_weights(generate_classification_data):
             "tau_pred": tau_pred.flatten(),
             "W": df_test["treatment_group_key"].values,
             CONVERSION: df_test[CONVERSION].values,
-            "treatment_effect_col": df_test["treatment_effect"].values,
+            "tau": df_test["treatment_effect"].values,
         }
     )
 
     cumgain = get_cumgain(
         auuc_metrics,
         outcome_col=CONVERSION,
         treatment_col="W",
-        treatment_effect_col="treatment_effect_col",
+        treatment_effect_col="tau",
     )
 
     # Check if the cumulative gain when using the model's prediction is
@@ -1005,7 +1005,7 @@ def test_BaseDRLearner(generate_regression_data):
             "cate_p": cate_p.flatten(),
             "W": treatment,
             "y": y,
-            "treatment_effect_col": tau,
+            "tau": tau,
         }
     )
 

tests/test_visualize.py

@@ -0,0 +1,14 @@
+import pandas as pd
+import numpy as np
+import pytest
+from causalml.metrics.visualize import get_cumlift
+
+
+def test_visualize_get_cumlift_errors_on_nan():
+    df = pd.DataFrame(
+        [[0, np.nan, 0.5], [1, np.nan, 0.1], [1, 1, 0.4], [0, 1, 0.3], [1, 1, 0.2]],
+        columns=["w", "y", "pred"],
+    )
+
+    with pytest.raises(Exception):
+        get_cumlift(df)