Merge pull request #483 from tdspora/EPMCTDM-7179_target_validation_f…

…or_utility_metric

Epmctdm 7179 target validation for utility metric

src/syngen/VERSION

@@ -1 +1 @@
-0.10.0
+0.10.1

src/syngen/ml/metrics/metrics_classes/metrics.py

@@ -1356,12 +1356,8 @@ def __model_process(self, model_object, targets, task_type):
                     original, model_y, self.sample_size
                 )
 
-            if len(set(model_y)) < 2:
-                logger.info(
-                    f"Column {col} has less than 2 classes as target. "
-                    f"It will not be used in metric "
-                    f"that measures regression results."
-                )
+            # validate number of unique values and instances in each class
+            if not self._valid_target(col, model_y, task_type):
                 continue
 
             (
@@ -1463,3 +1459,39 @@ def __create_regression_models(self, cont_targets):
             cont_targets, "regression"
         )
         return best_target, score, synthetic_score
+
+    @staticmethod
+    def _valid_target(col, model_y, task_type):
+        """
+        Validate the target column in utility metric calculation.
+
+        Args:
+        col (str): The name of the column being checked.
+        model_y (array-like): The target values.
+        task_type (str): type task
+
+        Returns:
+        bool: True if the column is valid
+        """
+        unique_values, unique_counts = np.unique(model_y, return_counts=True)
+
+        # check if there is only one unique value in the column
+        if len(unique_values) < 2:
+            logger.info(
+                f"Column '{col}' has only one unique value. "
+                "It will not be used as target column "
+                f"in utility metric calculation for {task_type}."
+            )
+            return False
+
+        # check if there is more than 1 sample in each class
+        if task_type in ["binary classification", "multiclass classification"]:
+            if np.min(unique_counts) < 2:
+                logger.info(
+                    f"Column '{col}' has a class with less than 2 samples. "
+                    "It will not be used as target column "
+                    f"in utility metric calculation for {task_type}."
+                )
+                return False
+
+        return True

src/tests/unit/metrics/test_metrics.py

@@ -1,6 +1,11 @@
 import pandas as pd
+import numpy as np
+import pytest
+
+from unittest.mock import patch
+
+from syngen.ml.metrics.metrics_classes.metrics import Clustering, Utility
 
-from syngen.ml.metrics.metrics_classes.metrics import Clustering
 from tests.conftest import SUCCESSFUL_MESSAGE, DIR_NAME
 
 
@@ -30,3 +35,42 @@ def test_clustering_calculate_all(rp_logger):
     assert mean_score >= threshold, f"Mean score shouldn't be less than {threshold}"
 
     rp_logger.info(SUCCESSFUL_MESSAGE)
+
+
+@pytest.mark.parametrize(
+    "col, model_y, task_type, expected_result, expected_log_message",
+    [
+        ("column_1", np.array([1, 1, 1, 1]), "binary classification", False, True),
+        ("column_2", np.array([0, 0, 0, 1]), "binary classification", False, True),
+        ("column_3", np.array([0, 1, 0, 1]), "binary classification", True, False),
+        ("column_4", np.array([0, 0, 1, 1, 2, 2]), "multiclass classification", True, False),
+        ("column_5", np.array([0, 0, 0, 1, 2, 2]), "multiclass classification", False, True),
+        ("column_6", np.array([1, 1, 1]), "multiclass classification", False, True),
+        ("column_7", np.array([1, 1, 1, 1]), "regression", False, True),
+        ("column_8", np.array([1, 2, 3, 4, 4]), "regression", True, False),
+    ]
+)
+def test_utility_valid_target(
+        rp_logger, col, model_y, task_type,
+        expected_result, expected_log_message):
+    """
+    Testing the _valid_target function in the Utility class
+    """
+    rp_logger.info(
+        "Testing the _valid_target function in the Utility class"
+    )
+
+    with patch(
+        'syngen.ml.metrics.metrics_classes.metrics.logger'
+    ) as mock_logger:
+        result = Utility._valid_target(col, model_y, task_type)
+
+    assert result == expected_result, \
+        f"Expected result is {expected_result}, got {result}"
+
+    if expected_log_message:
+        mock_logger.info.assert_called_once()
+    else:
+        mock_logger.info.assert_not_called()
+
+    rp_logger.info(SUCCESSFUL_MESSAGE)