Update datetime transformer (#230)

* Update datetime transformer

* Update datetime transformer to match the new baseclass transformer

* Fix bugs

* Disable duplicate code lint error + fix readme.md error

* Fix lint

README.md

@@ -104,7 +104,7 @@ Before being able to transform the data, we need the transformer to learn from i
 We will do this by calling its `fit` method passing the column that we want to transform.
 
 ```python3
-transformer.fit(data['3_datetime'])
+transformer.fit(data, columns=['3_datetime'])
 ```
 
 ### 4. Transform the data
@@ -113,7 +113,7 @@ Once the transformer is fitted, we can pass the data again to its `transform` me
 to get the transformed version of the data.
 
 ```python3
-transformed = transformer.transform(data['3_datetime'])
+transformed = transformer.transform(data)
 ```
 
 The output will be a `numpy.ndarray` with two columns, one with the datetimes transformed

rdt/transformers/datetime.py

@@ -34,6 +34,11 @@ class DatetimeTransformer(BaseTransformer):
             are zero on the lower time units.
     """
 
+    INPUT_TYPE = 'datetime'
+    DETERMINISTIC_TRANSFORM = True
+    DETERMINISTIC_REVERSE = True
+    COMPOSITION_IS_IDENTITY = True
+
     null_transformer = None
     divider = None
 
@@ -42,6 +47,33 @@ def __init__(self, nan='mean', null_column=None, strip_constant=False):
         self.null_column = null_column
         self.strip_constant = strip_constant
 
+    def is_composition_identity(self):
+        """Return whether composition of transform and reverse transform produces the input data.
+
+        Returns:
+            bool:
+                Whether or not transforming and then reverse transforming returns the input data.
+        """
+        if self.null_transformer and not self.null_transformer.creates_null_column():
+            return False
+
+        return self.COMPOSITION_IS_IDENTITY
+
+    def get_output_types(self):
+        """Return the output types supported by the transformer.
+
+        Returns:
+            dict:
+                Mapping from the transformed column names to supported data types.
+        """
+        output_types = {
+            'value': 'float',
+        }
+        if self.null_transformer and self.null_transformer.creates_null_column():
+            output_types['is_null'] = 'float'
+
+        return self._add_prefix(output_types)
+
     def _find_divider(self, transformed):
         self.divider = 1
         multipliers = [10] * 9 + [60, 60, 24]
@@ -52,7 +84,7 @@ def _find_divider(self, transformed):
 
             self.divider = candidate
 
-    def _transform(self, datetimes):
+    def _transform_helper(self, datetimes):
         """Transform datetime values to integer."""
         nulls = datetimes.isnull()
         integers = pd.to_numeric(datetimes, errors='coerce').values.astype(np.float64)
@@ -65,7 +97,7 @@ def _transform(self, datetimes):
 
         return transformed
 
-    def fit(self, data):
+    def _fit(self, data):
         """Fit the transformer to the data.
 
         Args:
@@ -75,11 +107,11 @@ def fit(self, data):
         if isinstance(data, np.ndarray):
             data = pd.Series(data)
 
-        transformed = self._transform(data)
+        transformed = self._transform_helper(data)
         self.null_transformer = NullTransformer(self.nan, self.null_column, copy=True)
         self.null_transformer.fit(transformed)
 
-    def transform(self, data):
+    def _transform(self, data):
         """Transform datetime values to float values.
 
         Args:
@@ -92,11 +124,11 @@ def transform(self, data):
         if isinstance(data, np.ndarray):
             data = pd.Series(data)
 
-        data = self._transform(data)
+        data = self._transform_helper(data)
 
         return self.null_transformer.transform(data)
 
-    def reverse_transform(self, data):
+    def _reverse_transform(self, data):
         """Convert float values back to datetimes.
 
         Args:
@@ -106,6 +138,9 @@ def reverse_transform(self, data):
         Returns:
             pandas.Series
         """
+        if not isinstance(data, np.ndarray):
+            data = data.to_numpy()
+
         if self.nan is not None:
             data = self.null_transformer.reverse_transform(data)
 

setup.cfg

@@ -60,4 +60,4 @@ max-attributes = 11
 ignore-comments = yes
 ignore-docstrings = yes
 ignore-imports = yes
-disable = R0903, R0913, R0914, C0209, W0223, W0221, W0237
+disable = R0801, R0903, R0913, R0914, C0209, W0223, W0221, W0237

tests/unit/transformers/test_datetime.py

@@ -11,9 +11,9 @@ def test_no_strip(self):
         data = pd.to_datetime(pd.Series([None, '1996-10-17', '1965-05-23']))
 
         # Run
-        dtt.fit(data.copy().to_numpy())
-        transformed = dtt.transform(data.copy().to_numpy())
-        reverted = dtt.reverse_transform(transformed)
+        dtt._fit(data.copy().to_numpy())
+        transformed = dtt._transform(data.copy().to_numpy())
+        reverted = dtt._reverse_transform(transformed)
 
         # Asserts
         expect_trans = np.array([
@@ -29,9 +29,9 @@ def test_strip(self):
         data = pd.to_datetime(pd.Series([None, '1996-10-17', '1965-05-23']))
 
         # Run
-        dtt.fit(data.copy().to_numpy())
-        transformed = dtt.transform(data.copy().to_numpy())
-        reverted = dtt.reverse_transform(transformed)
+        dtt._fit(data.copy().to_numpy())
+        transformed = dtt._transform(data.copy().to_numpy())
+        reverted = dtt._reverse_transform(transformed)
 
         # Asserts
         expect_trans = np.array([
@@ -45,20 +45,20 @@ def test_strip(self):
     def test_reverse_transform_all_none(self):
         dt = pd.to_datetime(['2020-01-01'])
         dtt = DatetimeTransformer(strip_constant=True)
-        dtt.fit(dt)
+        dtt._fit(dt)
 
-        output = dtt.reverse_transform(np.array([None]))
+        output = dtt._reverse_transform(np.array([None]))
 
         expected = pd.to_datetime(['NaT'])
         pd.testing.assert_series_equal(output.to_series(), expected.to_series())
 
     def test_reverse_transform_2d_ndarray(self):
         dt = pd.to_datetime(['2020-01-01', '2020-02-01', '2020-03-01'])
         dtt = DatetimeTransformer(strip_constant=True)
-        dtt.fit(dt)
+        dtt._fit(dt)
 
         transformed = np.array([[18262.], [18293.], [18322.]])
-        output = dtt.reverse_transform(transformed)
+        output = dtt._reverse_transform(transformed)
 
         expected = pd.to_datetime(['2020-01-01', '2020-02-01', '2020-03-01'])
         pd.testing.assert_series_equal(output.to_series(), expected.to_series())