#13 Add test for info leakage

Author: pockerman

src/tests/test_dataset_info_leakage.py

@@ -0,0 +1,110 @@
+import unittest
+import pytest
+
+from src.spaces.environment import Environment, EnvConfig
+from src.spaces.action_space import ActionSpace
+from src.spaces.actions import ActionSuppress, ActionIdentity, ActionGeneralize
+from src.utils.serial_hierarchy import SerialHierarchy
+from src.datasets.datasets_loaders import MockSubjectsLoader
+from src.utils.reward_manager import RewardManager
+from src.utils.string_distance_calculator import DistanceType
+from src.datasets.dataset_information_leakage import info_leakage
+
+
+class TestDatasetInfoLeakage(unittest.TestCase):
+
+    def setUp(self) -> None:
+        """
+        Setup the PandasDSWrapper to be used in the tests
+        :return: None
+        """
+
+        # load the dataset
+        self.ds = MockSubjectsLoader()
+
+        # specify the action space. We need to establish how these actions
+        # are performed
+        self.action_space = ActionSpace(n=4)
+
+        self.generalization_table = {"Mixed White/Asian": SerialHierarchy(values=["Mixed", ]),
+                                     "Chinese": SerialHierarchy(values=["Asian", ]),
+                                     "Indian": SerialHierarchy(values=["Asian", ]),
+                                     "Mixed White/Black African": SerialHierarchy(values=["Mixed", ]),
+                                     "Black African": SerialHierarchy(values=["Black", ]),
+                                     "Asian other": SerialHierarchy(values=["Asian", ]),
+                                     "Black other": SerialHierarchy(values=["Black", ]),
+                                     "Mixed White/Black Caribbean": SerialHierarchy(values=["Mixed", ]),
+                                     "Mixed other": SerialHierarchy(values=["Mixed", ]),
+                                     "Arab": SerialHierarchy(values=["Asian", ]),
+                                     "White Irish": SerialHierarchy(values=["White", ]),
+                                     "Not stated": SerialHierarchy(values=["Not stated"]),
+                                     "White Gypsy/Traveller": SerialHierarchy(values=["White", ]),
+                                     "White British": SerialHierarchy(values=["White", ]),
+                                     "Bangladeshi": SerialHierarchy(values=["Asian", ]),
+                                     "White other": SerialHierarchy(values=["White", ]),
+                                     "Black Caribbean": SerialHierarchy(values=["Black", ]),
+                                     "Pakistani": SerialHierarchy(values=["Asian", ])}
+
+
+        self.action_space.add_many(ActionSuppress(column_name="gender", suppress_table={"F": SerialHierarchy(values=['*', ]),
+                                                                                        'M': SerialHierarchy(values=['*', ])}),
+                              ActionIdentity(column_name="salary"), ActionIdentity(column_name="education"),
+                              ActionGeneralize(column_name="ethnicity", generalization_table=self.generalization_table))
+        self.reward_manager = RewardManager()
+
+    def test_info_leakage_1(self):
+        """
+        No distortion is applied on the data set so total distortion
+        should be zero
+        """
+
+        env_config = EnvConfig()
+        env_config.action_space = self.action_space
+        env_config.reward_manager = self.reward_manager
+        env_config.data_set = self.ds
+        env_config.start_column = "gender"
+        env_config.gamma = 0.99
+
+        # create the environment
+        env = Environment(env_config=env_config)
+
+        # initialize text distances
+        env.initialize_text_distances(distance_type=DistanceType.COSINE)
+
+        distances, sum_distances = info_leakage(ds1=env.data_set, ds2=env.start_ds, column_distances=env.column_distances)
+
+        # no leakage should exist as no trasformation is applied
+        self.assertEqual(0.0, sum_distances)
+
+    #@pytest.mark.skip(reason="no way of currently testing this")
+    def test_info_leakage_2(self):
+        """
+        We apply distortion on column gender
+        """
+
+        env_config = EnvConfig()
+        env_config.action_space = self.action_space
+        env_config.reward_manager = self.reward_manager
+        env_config.data_set = self.ds
+        env_config.start_column = "gender"
+        env_config.gamma = 0.99
+
+        # create the environment
+        env = Environment(env_config=env_config)
+
+        # initialize text distances
+        env.initialize_text_distances(distance_type=DistanceType.COSINE)
+
+        action = env.action_space.get_action_by_column_name(column_name="gender")
+
+        env.step(action=action)
+
+        distances, sum_distances = info_leakage(ds1=env.data_set, ds2=env.start_ds,
+                                                column_distances=env.column_distances)
+
+        # leakage should exist as we suppress the gender column
+        self.assertNotEqual(0.0, sum_distances)
+
+
+if __name__ == '__main__':
+    unittest.main()