Merge branch 'master' into bugfix/update-batch-monitoring-schedule

Author: keshav-chandak

src/sagemaker/feature_store/feature_group.py

@@ -60,6 +60,7 @@
     FeatureValue,
     FeatureParameter,
     TableFormatEnum,
+    DeletionModeEnum,
 )
 from sagemaker.utils import resolve_value_from_config
 
@@ -785,6 +786,7 @@ def delete_record(
         self,
         record_identifier_value_as_string: str,
         event_time: str,
+        deletion_mode: DeletionModeEnum = DeletionModeEnum.SOFT_DELETE,
     ):
         """Delete a single record from a FeatureGroup.
 
@@ -793,11 +795,15 @@ def delete_record(
                 a String representing the value of the record identifier.
             event_time (String):
                 a timestamp format String indicating when the deletion event occurred.
+            deletion_mode (DeletionModeEnum):
+                deletion mode for deleting record. (default: DetectionModeEnum.SOFT_DELETE)
         """
+
         return self.sagemaker_session.delete_record(
             feature_group_name=self.name,
             record_identifier_value_as_string=record_identifier_value_as_string,
             event_time=event_time,
+            deletion_mode=deletion_mode.value,
         )
 
     def ingest(

src/sagemaker/feature_store/inputs.py

@@ -369,3 +369,13 @@ def to_dict(self) -> Dict[str, Any]:
             RecordIdentifiersValueAsString=self.record_identifiers_value_as_string,
             FeatureNames=None if not self.feature_names else self.feature_names,
         )
+
+
+class DeletionModeEnum(Enum):
+    """Enum of deletion modes.
+
+    The deletion mode for deleting records can be SoftDelete or HardDelete.
+    """
+
+    SOFT_DELETE = "SoftDelete"
+    HARD_DELETE = "HardDelete"

src/sagemaker/session.py

@@ -5196,18 +5196,21 @@ def delete_record(
         feature_group_name: str,
         record_identifier_value_as_string: str,
         event_time: str,
+        deletion_mode: str = None,
     ):
         """Deletes a single record from the FeatureGroup.
 
         Args:
             feature_group_name (str): name of the FeatureGroup.
             record_identifier_value_as_string (str): name of the record identifier.
             event_time (str): a timestamp indicating when the deletion event occurred.
+            deletion_mode: (str): deletion mode for deleting record.
         """
         return self.sagemaker_featurestore_runtime_client.delete_record(
             FeatureGroupName=feature_group_name,
             RecordIdentifierValueAsString=record_identifier_value_as_string,
             EventTime=event_time,
+            DeletionMode=deletion_mode,
         )
 
     def get_record(

tests/integ/test_feature_store.py

@@ -33,6 +33,7 @@
     Filter,
     ResourceEnum,
     Identifier,
+    DeletionModeEnum,
 )
 from sagemaker.feature_store.dataset_builder import (
     JoinTypeEnum,
@@ -162,6 +163,15 @@ def pandas_data_frame_without_string():
     return df
 
 
+@pytest.fixture
+def historic_record():
+    return [
+        FeatureValue(feature_name="feature1", value_as_string="10.0"),
+        FeatureValue(feature_name="feature2", value_as_string="7"),
+        FeatureValue(feature_name="feature3", value_as_string="2020-10-29T03:43:21Z"),
+    ]
+
+
 @pytest.fixture
 def record():
     return [
@@ -398,7 +408,7 @@ def test_get_and_batch_get_record(
             assert feature["FeatureName"] is not removed_feature_name
 
 
-def test_delete_record(
+def test_soft_delete_record(
     feature_store_session,
     role,
     feature_group_name,
@@ -437,6 +447,55 @@ def test_delete_record(
         assert retrieved_record is None
 
 
+def test_hard_delete_record(
+    feature_store_session,
+    role,
+    feature_group_name,
+    pandas_data_frame,
+    historic_record,
+    record,
+):
+    feature_group = FeatureGroup(name=feature_group_name, sagemaker_session=feature_store_session)
+    feature_group.load_feature_definitions(data_frame=pandas_data_frame)
+
+    record_identifier_value_as_string = record[0].value_as_string
+    historic_record_identifier_value_as_string = historic_record[0].value_as_string
+    with cleanup_feature_group(feature_group):
+        feature_group.create(
+            s3_uri=False,
+            record_identifier_name="feature1",
+            event_time_feature_name="feature3",
+            role_arn=role,
+            enable_online_store=True,
+        )
+        _wait_for_feature_group_create(feature_group)
+        # Ingest data
+        feature_group.put_record(record=record)
+        # Retrieve data
+        retrieved_record = feature_group.get_record(
+            record_identifier_value_as_string=record_identifier_value_as_string,
+        )
+        assert retrieved_record is not None
+        # Delete data
+        feature_group.delete_record(
+            record_identifier_value_as_string=record_identifier_value_as_string,
+            event_time=datetime.datetime.now().replace(microsecond=0).isoformat() + "Z",
+            deletion_mode=DeletionModeEnum.HARD_DELETE,
+        )
+        # Retrieve data
+        retrieved_record = feature_group.get_record(
+            record_identifier_value_as_string=record_identifier_value_as_string,
+        )
+        assert retrieved_record is None
+        # Ingest data
+        feature_group.put_record(historic_record)
+        # Retrieve data
+        retrieved_record = feature_group.get_record(
+            record_identifier_value_as_string=historic_record_identifier_value_as_string,
+        )
+        assert retrieved_record is not None
+
+
 def test_update_feature_group(
     feature_store_session,
     role,

tests/unit/sagemaker/feature_store/test_feature_group.py

@@ -31,7 +31,7 @@
     AthenaQuery,
     IngestionError,
 )
-from sagemaker.feature_store.inputs import FeatureParameter
+from sagemaker.feature_store.inputs import FeatureParameter, DeletionModeEnum
 
 from tests.unit import SAGEMAKER_CONFIG_FEATURE_GROUP
 
@@ -296,6 +296,41 @@ def test_delete_record(sagemaker_session_mock):
         feature_group_name="MyFeatureGroup",
         record_identifier_value_as_string=record_identifier_value_as_string,
         event_time=event_time,
+        deletion_mode=DeletionModeEnum.SOFT_DELETE.value,
+    )
+
+
+def test_soft_delete_record(sagemaker_session_mock):
+    feature_group = FeatureGroup(name="MyFeatureGroup", sagemaker_session=sagemaker_session_mock)
+    record_identifier_value_as_string = "1.0"
+    event_time = "2022-09-14"
+    feature_group.delete_record(
+        record_identifier_value_as_string=record_identifier_value_as_string,
+        event_time=event_time,
+        deletion_mode=DeletionModeEnum.SOFT_DELETE,
+    )
+    sagemaker_session_mock.delete_record.assert_called_with(
+        feature_group_name="MyFeatureGroup",
+        record_identifier_value_as_string=record_identifier_value_as_string,
+        event_time=event_time,
+        deletion_mode=DeletionModeEnum.SOFT_DELETE.value,
+    )
+
+
+def test_hard_delete_record(sagemaker_session_mock):
+    feature_group = FeatureGroup(name="MyFeatureGroup", sagemaker_session=sagemaker_session_mock)
+    record_identifier_value_as_string = "1.0"
+    event_time = "2022-09-14"
+    feature_group.delete_record(
+        record_identifier_value_as_string=record_identifier_value_as_string,
+        event_time=event_time,
+        deletion_mode=DeletionModeEnum.HARD_DELETE,
+    )
+    sagemaker_session_mock.delete_record.assert_called_with(
+        feature_group_name="MyFeatureGroup",
+        record_identifier_value_as_string=record_identifier_value_as_string,
+        event_time=event_time,
+        deletion_mode=DeletionModeEnum.HARD_DELETE.value,
     )