[SC 16935] Fix IQROutliersBarPlot failure on boolean numeric

tests/unit_tests/data_validation/test_IQROutliersBarPlot.py

@@ -97,4 +97,4 @@ def test_boolean_dtype_excluded_from_raw_data(self):
         results = IQROutliersBarPlot(vm_dataset)
         raw_data = results[-1]
 
-        self.assertNotIn("flag", raw_data.outlier_counts_by_feature.index)
+        self.assertNotIn("flag", raw_data.outlier_counts_by_feature.columns)

tests/unit_tests/data_validation/test_IQROutliersTable.py

@@ -83,3 +83,25 @@ def test_binary_exclusion(self):
         # Verify binary column is not in results
         for summary in outliers_summary:
             self.assertNotIn("binary", summary["Variable"])
+
+    def test_boolean_dtype_excluded_from_raw_data(self):
+        n_samples = 100
+        normal_data = np.array([1.0] * 25 + [2.0] * 50 + [3.0] * 25)
+        data_with_outliers = normal_data.copy()
+        data_with_outliers[0:4] = [-15, -10, 10, 15]
+        df = pd.DataFrame(
+            {
+                "with_outliers": data_with_outliers,
+                "flag": np.random.choice([True, False], n_samples),
+            }
+        )
+        vm_dataset = vm.init_dataset(
+            input_id="test_boolean_dataset", dataset=df, __log=False
+        )
+
+        result, raw_data = IQROutliersTable(vm_dataset)
+        outliers_summary = result["Summary of Outliers Detected by IQR Method"]
+        summary_variables = [summary["Variable"] for summary in outliers_summary]
+
+        self.assertNotIn("flag", summary_variables)
+        self.assertNotIn("flag", raw_data.all_outliers)

validmind/tests/data_validation/IQROutliersBarPlot.py

@@ -5,6 +5,7 @@
 
 from typing import Tuple
 
+import pandas as pd
 import plotly.graph_objects as go
 
 from validmind import RawData, tags, tasks
@@ -76,9 +77,12 @@ def IQROutliersBarPlot(
     """
     df = dataset.df
 
-    # Exclude binary/boolean features (IQR is not meaningful and quantile fails on bool)
+    # Exclude boolean and binary features. The IQR is not meaningful for them and
+    # `quantile` raises "numpy boolean subtract" on boolean dtype columns.
     eligible_columns = [
-        col for col in dataset.feature_columns_numeric if len(df[col].unique()) > 2
+        col
+        for col in dataset.feature_columns_numeric
+        if not pd.api.types.is_bool_dtype(df[col]) and df[col].nunique() > 2
     ]
 
     figures = []

validmind/tests/data_validation/IQROutliersTable.py

@@ -5,6 +5,8 @@
 
 from typing import Any, Dict, Tuple
 
+import pandas as pd
+
 from validmind import RawData, tags, tasks
 from validmind.vm_models import VMDataset
 
@@ -72,8 +74,9 @@ def IQROutliersTable(
     all_outliers = {}
 
     for col in dataset.feature_columns_numeric:
-        # Skip binary features
-        if len(df[col].unique()) <= 2:
+        # Skip boolean and binary features. The IQR is not meaningful for them and
+        # `quantile` raises "numpy boolean subtract" on boolean dtype columns.
+        if pd.api.types.is_bool_dtype(df[col]) or df[col].nunique() <= 2:
             continue
 
         outliers = compute_outliers(df[col], threshold)