Comments

Author: frances-h

rdt/transformers/numerical.py

@@ -670,6 +670,10 @@ def __init__(
         max_value=1.0,
         learn_rounding_scheme=False,
     ):
+        if min_value == max_value:
+            error_msg = 'The min_value and max_value for the logit function cannot be equal.'
+            raise TransformerInputError(error_msg)
+
         super().__init__(
             missing_value_replacement=missing_value_replacement,
             missing_value_generation=missing_value_generation,
@@ -707,18 +711,18 @@ def _transform(self, data):
         logit_vals = logit(transformed_vals, self.min_value, self.max_value)
         if transformed.ndim == 1:
             return logit_vals
-        else:
-            transformed[:, 0] = logit_vals
-            return transformed
+
+        transformed[:, 0] = logit_vals
+        return transformed
 
     def _reverse_transform(self, data):
         if not isinstance(data, np.ndarray):
             data = data.to_numpy()
 
         sampled_vals = data if data.ndim == 1 else data[:, 0]
-        reversed = sigmoid(sampled_vals, self.min_value, self.max_value)
+        reversed_values = sigmoid(sampled_vals, self.min_value, self.max_value)
         if data.ndim == 1:
-            return super()._reverse_transform(reversed)
-        else:
-            data[:, 0] = reversed
-            return super()._reverse_transform(data)
+            return super()._reverse_transform(reversed_values)
+
+        data[:, 0] = reversed_values
+        return super()._reverse_transform(data)

tests/integration/test_transformers.py

@@ -47,6 +47,28 @@
 }
 
 
+def _create_transformer_args_from_data(transformer_args, data):
+    """Helper to extract transformer arguments that are data-dependent.
+
+    Args:
+        transformer_args (dict):
+            The transformer arguments.
+        data (pd.Series):
+            The data for the transformer.
+
+    Returns:
+        dict:
+            The transformer arguments with data-specific arguments added.
+    """
+    if 'FROM_DATA' in transformer_args:
+        transformer_args = {**transformer_args}
+        args = transformer_args.pop('FROM_DATA')
+        for arg, arg_func in args.items():
+            transformer_args[arg] = arg_func(data)
+
+    return transformer_args
+
+
 def _validate_helper(validator_function, args, steps):
     """Wrap around validation functions to either return a boolean or assert.
 
@@ -157,11 +179,7 @@ def _test_transformer_with_dataset(transformer_class, input_data, steps):
     """
 
     transformer_args = TRANSFORMER_ARGS.get(transformer_class.__name__, {})
-    if 'FROM_DATA' in transformer_args:
-        transformer_args = {**transformer_args}
-        args = transformer_args.pop('FROM_DATA')
-        for arg, arg_func in args.items():
-            transformer_args[arg] = arg_func(input_data[TEST_COL])
+    transformer_args = _create_transformer_args_from_data(transformer_args, input_data[TEST_COL])
 
     transformer = transformer_class(**transformer_args)
     # Fit
@@ -217,12 +235,9 @@ def _test_transformer_with_hypertransformer(transformer_class, input_data, steps
     transformer_args = TRANSFORMER_ARGS.get(transformer_class.__name__, {})
     hypertransformer = HyperTransformer()
     if transformer_args:
-        if 'FROM_DATA' in transformer_args:
-            transformer_args = {**transformer_args}
-            args = transformer_args.pop('FROM_DATA')
-            for arg, arg_func in args.items():
-                transformer_args[arg] = arg_func(input_data[TEST_COL])
-
+        transformer_args = _create_transformer_args_from_data(
+            transformer_args, input_data[TEST_COL]
+        )
         field_transformers = {TEST_COL: transformer_class(**transformer_args)}
 
     else:

tests/unit/transformers/test_numerical.py

@@ -1889,7 +1889,18 @@ def test___init__(self):
         assert ls.max_value == 100.0
         assert ls.min_value == 2.0
 
-    def test__validate_logit_inputs(self):
+    def test___init___invalid_inputs(self):
+        """Test super() arguments are properly passed and set as attributes."""
+        # Setup
+        min_value = 10.0
+        max_value = 10.0
+
+        # Run / Assert
+        expected_msg = 'The min_value and max_value for the logit function cannot be equal.'
+        with pytest.raises(TransformerInputError, match=re.escape(expected_msg)):
+            LogitScaler(max_value=max_value, min_value=min_value)
+
+    def test__validate_logit_inputs_with_default_settings(self):
         """Test validating data against input arguments."""
         # Setup
         ls = LogitScaler()
@@ -1898,6 +1909,15 @@ def test__validate_logit_inputs(self):
         # Run and Assert
         ls._validate_logit_inputs(data)
 
+    def test__validate_logit_inputs_with_custom_inputs(self):
+        """Test validating data against input arguments."""
+        # Setup
+        ls = LogitScaler(min_value=0, max_value=100)
+        data = pd.Series([0.0, 10.1, 20.2, 30.3, 100])
+
+        # Run and Assert
+        ls._validate_logit_inputs(data)
+
     def test__validate_logit_inputs_errors_invalid_value(self):
         """Test error message contains invalid values."""
         # Setup
@@ -1944,7 +1964,7 @@ def test__fit(self):
     def test__transform(self, mock_logit):
         """Test the ``transform`` method."""
         # Setup
-        min_value = (1.0,)
+        min_value = 1.0
         max_value = 50.0
         ls = LogitScaler(min_value=min_value, max_value=max_value)
         ls._validate_logit_inputs = Mock()
@@ -1965,7 +1985,7 @@ def test__transform(self, mock_logit):
     def test__transform_multi_column(self, mock_logit):
         """Test the ``transform`` method with multiple columns."""
         # Setup
-        min_value = (1.0,)
+        min_value = 1.0
         max_value = 50.0
         ls = LogitScaler(min_value=min_value, max_value=max_value)
         ls._validate_logit_inputs = Mock()
@@ -1988,7 +2008,7 @@ def test__transform_multi_column(self, mock_logit):
     def test__reverse_transform(self, mock_sigmoid, ff_reverse_transform_mock):
         """Test the ``transform`` method."""
         # Setup
-        min_value = (1.0,)
+        min_value = 1.0
         max_value = 50.0
         ls = LogitScaler(min_value=min_value, max_value=max_value)
         data = pd.Series([1.0, 1.1, 1.2, 1.3, 2.0, 3.0, 4.0])
@@ -1997,40 +2017,40 @@ def test__reverse_transform(self, mock_sigmoid, ff_reverse_transform_mock):
         ls.null_transformer = null_transformer_mock
 
         # Run
-        reversed = ls._reverse_transform(data)
+        reversed_values = ls._reverse_transform(data)
 
         # Assert
         mock_sigmoid_args = mock_sigmoid.call_args[0]
         np.testing.assert_array_equal(mock_sigmoid_args[0], data.to_numpy())
         assert mock_sigmoid_args[1] == ls.min_value
         assert mock_sigmoid_args[2] == ls.max_value
         ff_reverse_transform_mock.assert_called_once_with(mock_sigmoid.return_value)
-        assert reversed == ff_reverse_transform_mock.return_value
+        assert reversed_values == ff_reverse_transform_mock.return_value
 
     @patch('rdt.transformers.numerical.FloatFormatter._reverse_transform')
     @patch('rdt.transformers.numerical.sigmoid')
     def test__reverse_transform_multi_column(self, mock_sigmoid, ff_reverse_transform_mock):
         """Test the ``transform`` method with multiple columns."""
         # Setup
-        min_value = (1.0,)
+        min_value = 1.0
         max_value = 50.0
         ls = LogitScaler(min_value=min_value, max_value=max_value)
         sampled_data = np.array([1.0, 1.1, 1.2, 1.3, 2.0, 3.0, 4.0])
         is_null = np.array([0.0, 0.0, 1.0, 1.0, 0.0, 1.0, 1.0])
         data = pd.DataFrame({'column': sampled_data, 'column.is_null': is_null})
         null_transformer_mock = Mock()
-        reversed = np.array([1.0, 1.1, np.nan, np.nan, 2.0, np.nan, np.nan])
-        null_transformer_mock.reverse_transform.return_value = reversed
+        reversed_values = np.array([1.0, 1.1, np.nan, np.nan, 2.0, np.nan, np.nan])
+        null_transformer_mock.reverse_transform.return_value = reversed_values
         ls.null_transformer = null_transformer_mock
         sigmoid_vals = np.array([3.0, 3.1, 3.3, 3.4, 2.1, 4.0, 4.6])
         mock_sigmoid.return_value = sigmoid_vals
 
         # Run
-        reversed = ls._reverse_transform(data)
+        reversed_values = ls._reverse_transform(data)
 
         # Assert
         ff_reverse_transform_args = ff_reverse_transform_mock.call_args[0]
         np.testing.assert_array_equal(
             ff_reverse_transform_args[0], np.array([sigmoid_vals, is_null]).T
         )
-        assert reversed == ff_reverse_transform_mock.return_value
+        assert reversed_values == ff_reverse_transform_mock.return_value