Fixing tests for rdt 1.13.2 (#2320)

Co-authored-by: Felipe <[email protected]>

Author: amontanez24

pyproject.toml

@@ -35,7 +35,7 @@ dependencies = [
     'copulas>=0.12.0',
     'ctgan>=0.10.2',
     'deepecho>=0.6.1',
-    'rdt>=1.13.1',
+    'rdt>=1.13.2.dev0',
     'sdmetrics>=0.17.0',
     'platformdirs>=4.0',
     'pyyaml>=6.0.1',

sdv/constraints/tabular.py

@@ -299,7 +299,7 @@ def _fit(self, table_data):
         self._uuids_to_combinations = {}
         for combination in self._combinations.itertuples(index=False, name=None):
             mappable_combination = get_mappable_combination(combination)
-            uuid_str = str(uuid.uuid4())
+            uuid_str = str(uuid.uuid5(uuid.NAMESPACE_DNS, str(mappable_combination)))
             self._combinations_to_uuids[mappable_combination] = uuid_str
             self._uuids_to_combinations[uuid_str] = mappable_combination
 

tests/integration/constraints/test_tabular.py

@@ -69,7 +69,10 @@ def test_fixed_combinations_with_nans():
     assert len(synthetic_data) == 1000
     pd.testing.assert_frame_equal(
         synthetic_data.drop_duplicates(ignore_index=True),
-        data.drop_duplicates(ignore_index=True),
+        pd.DataFrame({
+            'A': [1, np.nan, 2],
+            'B': [10, 30, 20],
+        }).drop_duplicates(ignore_index=True),
         check_like=True,
     )
 

tests/integration/single_table/test_base.py

@@ -212,7 +212,7 @@ def test_sample_keys_are_scrambled():
 
     # Assert
     ids = sampled['guest_email'].head()
-    expected_keys = pd.Series(['ARH', 'BBH', 'BAP', 'AIF', 'AQP'], name='guest_email')
+    expected_keys = pd.Series(['AQQ', 'BBI', 'AET', 'AZM', 'AHZ'], name='guest_email')
     pd.testing.assert_series_equal(ids, expected_keys)
 
 

tests/integration/single_table/test_constraints.py

@@ -468,28 +468,29 @@ def test_inequality_constraint_with_datetimes_and_nones():
     sampled = synth.sample(10)
 
     # Assert
+    synth.validate(sampled)
     expected_sampled = pd.DataFrame({
         'A': [
             '2020-01-02',
             '2020-01-02',
+            np.nan,
+            np.nan,
             '2020-01-02',
+            np.nan,
             '2020-01-02',
-            '2020-01-02',
-            '2020-01-02',
-            '2020-01-02',
-            '2020-01-02',
+            np.nan,
             '2020-01-02',
             np.nan,
         ],
         'B': [
-            np.nan,
             '2021-12-30',
             '2021-12-30',
             '2021-12-30',
-            np.nan,
             '2021-12-30',
             np.nan,
             '2021-12-30',
+            '2021-12-30',
+            '2021-12-30',
             np.nan,
             '2021-12-30',
         ],
@@ -529,19 +530,20 @@ def test_scalar_inequality_constraint_with_datetimes_and_nones():
     sampled = synth.sample(5)
 
     # Assert
+    synth.validate(sampled)
     expected_sampled = pd.DataFrame({
         'A': {
             0: np.nan,
             1: '2020-01-19',
             2: np.nan,
-            3: '2020-01-29',
+            3: np.nan,
             4: '2020-01-31',
         },
         'B': {
-            0: '2021-07-28',
-            1: '2021-07-14',
-            2: '2021-07-26',
-            3: '2021-07-02',
+            0: np.nan,
+            1: np.nan,
+            2: np.nan,
+            3: np.nan,
             4: '2021-06-06',
         },
     })
@@ -585,27 +587,28 @@ def test_scalar_range_constraint_with_datetimes_and_nones():
     sampled = synth.sample(10)
 
     # Assert
+    synth.validate(sampled)
     expected_sampled = pd.DataFrame({
         'A': {
-            0: '2020-02-04',
+            0: np.nan,
             1: np.nan,
             2: '2020-02-07',
             3: np.nan,
-            4: np.nan,
+            4: '2020-02-29',
             5: '2020-02-29',
             6: np.nan,
             7: np.nan,
-            8: np.nan,
+            8: '2020-01-26',
             9: '2020-02-02',
         },
         'B': {
             0: np.nan,
             1: np.nan,
             2: np.nan,
             3: np.nan,
-            4: np.nan,
-            5: '2021-11-22',
-            6: np.nan,
+            4: '2021-06-21',
+            5: np.nan,
+            6: '2021-09-28',
             7: '2021-06-19',
             8: np.nan,
             9: np.nan,
@@ -657,14 +660,14 @@ def test_range_constraint_with_datetimes_and_nones():
         'A': [
             '2020-01-02',
             '2020-01-02',
+            np.nan,
             '2020-01-02',
             '2020-01-02',
+            np.nan,
             '2020-01-02',
             '2020-01-02',
             '2020-01-02',
-            '2020-01-02',
-            '2020-01-02',
-            '2020-01-02',
+            np.nan,
         ],
         'B': [
             np.nan,
@@ -674,7 +677,7 @@ def test_range_constraint_with_datetimes_and_nones():
             np.nan,
             '2021-12-30',
             np.nan,
-            np.nan,
+            '2021-12-30',
             np.nan,
             '2021-12-30',
         ],
@@ -686,12 +689,13 @@ def test_range_constraint_with_datetimes_and_nones():
             np.nan,
             '2022-12-30',
             np.nan,
-            np.nan,
+            '2022-12-30',
             np.nan,
             '2022-12-30',
         ],
     })
     pd.testing.assert_frame_equal(expected_sampled, sampled)
+    synth.validate(sampled)
 
 
 def test_inequality_constraint_all_possible_nans_configurations():