Merge branch 'master' of https://github.com/aws/sagemaker-python-sdk into smdmp-1.11.0-doc

Author: mchoi8739

CHANGELOG.md

@@ -1,5 +1,21 @@
 # Changelog
 
+## v2.105.0 (2022-08-19)
+
+### Features
+
+ * Added endpoint_name to clarify.ModelConfig
+ * adding workgroup functionality to athena query
+
+### Bug Fixes and Other Changes
+
+ * disable debugger/profiler in cgk region
+ * using unique name for lineage test to unblock PR checks
+
+### Documentation Changes
+
+ * update first-party algorithms and structural updates
+
 ## v2.104.0 (2022-08-17)
 
 ### Features

VERSION

@@ -1 +1 @@
-2.104.1.dev0
+2.105.1.dev0

src/sagemaker/clarify.py

@@ -823,6 +823,29 @@ def _prepare_debugger_for_training(self):
             self.debugger_hook_config.s3_output_path = self.output_path
         self.debugger_rule_configs = self._prepare_debugger_rules()
         self._prepare_collection_configs()
+        self._validate_and_set_debugger_configs()
+        if not self.debugger_hook_config:
+            if self.environment is None:
+                self.environment = {}
+            self.environment[DEBUGGER_FLAG] = "0"
+
+    def _validate_and_set_debugger_configs(self):
+        """Set defaults for debugging."""
+        region_supports_debugger = _region_supports_debugger(
+            self.sagemaker_session.boto_region_name
+        )
+
+        if region_supports_debugger:
+            if self.debugger_hook_config in [None, {}]:
+                self.debugger_hook_config = DebuggerHookConfig(s3_output_path=self.output_path)
+        else:
+            if self.debugger_hook_config is not False and self.debugger_hook_config:
+                # when user set debugger config in a unsupported region
+                raise ValueError(
+                    "Current region does not support debugger but debugger hook config is set!"
+                )
+            # disable debugger in unsupported regions
+            self.debugger_hook_config = False
 
     def _prepare_debugger_rules(self):
         """Set any necessary values in debugger rules, if they are provided."""
@@ -1766,6 +1789,8 @@ def enable_default_profiling(self):
         Debugger monitoring is disabled.
         """
         self._ensure_latest_training_job()
+        if not _region_supports_debugger(self.sagemaker_session.boto_region_name):
+            raise ValueError("Current region does not support profiler / debugger!")
 
         training_job_details = self.latest_training_job.describe()
 
@@ -1799,6 +1824,8 @@ def disable_profiling(self):
 
         """
         self._ensure_latest_training_job()
+        if not _region_supports_debugger(self.sagemaker_session.boto_region_name):
+            raise ValueError("Current region does not support profiler / debugger!")
 
         training_job_details = self.latest_training_job.describe()
 
@@ -1852,6 +1879,8 @@ def update_profiler(
 
         """
         self._ensure_latest_training_job()
+        if not _region_supports_debugger(self.sagemaker_session.boto_region_name):
+            raise ValueError("Current region does not support profiler / debugger!")
 
         if (
             not rules
@@ -2872,13 +2901,7 @@ def _script_mode_hyperparam_update(self, code_dir: str, script: str) -> None:
 
     def _validate_and_set_debugger_configs(self):
         """Set defaults for debugging."""
-        if self.debugger_hook_config is None and _region_supports_debugger(
-            self.sagemaker_session.boto_region_name
-        ):
-            self.debugger_hook_config = DebuggerHookConfig(s3_output_path=self.output_path)
-        elif not self.debugger_hook_config:
-            # set hook config to False if _region_supports_debugger is False
-            self.debugger_hook_config = False
+        super(Framework, self)._validate_and_set_debugger_configs()
 
         # Disable debugger if checkpointing is enabled by the customer
         if self.checkpoint_s3_uri and self.checkpoint_local_path and self.debugger_hook_config:
@@ -2901,11 +2924,6 @@ def _validate_and_set_debugger_configs(self):
                     )
                     self.debugger_hook_config = False
 
-        if self.debugger_hook_config is False:
-            if self.environment is None:
-                self.environment = {}
-            self.environment[DEBUGGER_FLAG] = "0"
-
     def _model_source_dir(self):
         """Get the appropriate value to pass as ``source_dir`` to a model constructor.
 

src/sagemaker/estimator.py

@@ -81,7 +81,9 @@ class AthenaQuery:
     _result_bucket: str = attr.ib(init=False, default=None)
     _result_file_prefix: str = attr.ib(init=False, default=None)
 
-    def run(self, query_string: str, output_location: str, kms_key: str = None) -> str:
+    def run(
+        self, query_string: str, output_location: str, kms_key: str = None, workgroup: str = None
+    ) -> str:
         """Execute a SQL query given a query string, output location and kms key.
 
         This method executes the SQL query using Athena and outputs the results to output_location
@@ -91,6 +93,7 @@ def run(self, query_string: str, output_location: str, kms_key: str = None) -> s
             query_string: SQL query string.
             output_location: S3 URI of the query result.
             kms_key: KMS key id. If set, will be used to encrypt the query result file.
+            workgroup (str): The name of the workgroup in which the query is being started.
 
         Returns:
             Execution id of the query.
@@ -101,6 +104,7 @@ def run(self, query_string: str, output_location: str, kms_key: str = None) -> s
             query_string=query_string,
             output_location=output_location,
             kms_key=kms_key,
+            workgroup=workgroup,
         )
         self._current_query_execution_id = response["QueryExecutionId"]
         parse_result = urlparse(output_location, allow_fragments=False)

src/sagemaker/feature_store/feature_group.py

@@ -53,8 +53,26 @@
     "only one worker per host regardless of the number of GPUs."
 )
 
-DEBUGGER_UNSUPPORTED_REGIONS = ("us-iso-east-1",)
-PROFILER_UNSUPPORTED_REGIONS = ("us-iso-east-1",)
+DEBUGGER_UNSUPPORTED_REGIONS = (
+    "us-iso-east-1",
+    "ap-southeast-3",
+    "ap-southeast-4",
+    "eu-south-2",
+    "me-central-1",
+    "ap-south-2",
+    "eu-central-2",
+    "us-gov-east-1",
+)
+PROFILER_UNSUPPORTED_REGIONS = (
+    "us-iso-east-1",
+    "ap-southeast-3",
+    "ap-southeast-4",
+    "eu-south-2",
+    "me-central-1",
+    "ap-south-2",
+    "eu-central-2",
+    "us-gov-east-1",
+)
 
 SINGLE_GPU_INSTANCE_TYPES = ("ml.p2.xlarge", "ml.p3.2xlarge")
 SM_DATAPARALLEL_SUPPORTED_INSTANCE_TYPES = (

src/sagemaker/fw_utils.py

@@ -4187,6 +4187,7 @@ def start_query_execution(
         query_string: str,
         output_location: str,
         kms_key: str = None,
+        workgroup: str = None,
     ) -> Dict[str, str]:
         """Start Athena query execution.
 
@@ -4196,6 +4197,8 @@ def start_query_execution(
             query_string (str): SQL expression.
             output_location (str): S3 location of the output file.
             kms_key (str): KMS key id will be used to encrypt the result if given.
+            workgroup (str): The name of the workgroup in which the query is being started.
+            If the workgroup is not specified, the default workgroup is used.
 
         Returns:
             Response dict from the service.
@@ -4210,6 +4213,9 @@ def start_query_execution(
             )
         kwargs.update(ResultConfiguration=result_config)
 
+        if workgroup:
+            kwargs.update(WorkGroup=workgroup)
+
         athena_client = self.boto_session.client("athena", region_name=self.boto_region_name)
         return athena_client.start_query_execution(**kwargs)
 

src/sagemaker/session.py

@@ -141,7 +141,7 @@ def action_objs(sagemaker_session):
 @pytest.fixture
 def artifact_obj(sagemaker_session):
     obj = artifact.Artifact.create(
-        artifact_name="SDKIntegrationTest",
+        artifact_name=name(),
         artifact_type="SDKIntegrationTest",
         source_uri=name(),
         properties={"k1": "v1"},

tests/integ/sagemaker/lineage/conftest.py

@@ -138,7 +138,9 @@ def data_path_no_label_index(training_set_no_label):
 def data_path_label_index(training_set_label_index):
     features, label, index = training_set_label_index
     data = pd.concat(
-        [pd.DataFrame(label), pd.DataFrame(features), pd.DataFrame(index)], axis=1, sort=False
+        [pd.DataFrame(label), pd.DataFrame(features), pd.DataFrame(index)],
+        axis=1,
+        sort=False,
     )
     with tempfile.TemporaryDirectory() as tmpdirname:
         filename = os.path.join(tmpdirname, "train_label_index.csv")
@@ -151,7 +153,12 @@ def data_path_label_index(training_set_label_index):
 def data_path_label_index_6col(training_set_label_index):
     features, label, index = training_set_label_index
     data = pd.concat(
-        [pd.DataFrame(label), pd.DataFrame(features), pd.DataFrame(features), pd.DataFrame(index)],
+        [
+            pd.DataFrame(label),
+            pd.DataFrame(features),
+            pd.DataFrame(features),
+            pd.DataFrame(index),
+        ],
         axis=1,
         sort=False,
     )
@@ -551,7 +558,10 @@ def test_pre_training_bias(clarify_processor, data_config, data_bias_config, sag
 
 
 def test_pre_training_bias_facets_not_included(
-    clarify_processor, data_config_facets_not_included, data_bias_config, sagemaker_session
+    clarify_processor,
+    data_config_facets_not_included,
+    data_bias_config,
+    sagemaker_session,
 ):
     with timeout.timeout(minutes=CLARIFY_DEFAULT_TIMEOUT_MINUTES):
         clarify_processor.run_pre_training_bias(
@@ -643,7 +653,9 @@ def test_post_training_bias_facets_not_included_excluded_columns(
             <= 1.0
         )
         check_analysis_config(
-            data_config_facets_not_included_multiple_files, sagemaker_session, "post_training_bias"
+            data_config_facets_not_included_multiple_files,
+            sagemaker_session,
+            "post_training_bias",
         )
 
 
@@ -704,6 +716,50 @@ def test_shap(clarify_processor, data_config, model_config, shap_config, sagemak
         check_analysis_config(data_config, sagemaker_session, "shap")
 
 
+def test_bias_and_explainability(
+    clarify_processor,
+    data_config,
+    model_config,
+    shap_config,
+    data_bias_config,
+    sagemaker_session,
+):
+    with timeout.timeout(minutes=CLARIFY_DEFAULT_TIMEOUT_MINUTES):
+        clarify_processor.run_bias_and_explainability(
+            data_config,
+            model_config,
+            shap_config,
+            data_bias_config,
+            pre_training_methods="all",
+            post_training_methods="all",
+            model_predicted_label_config="score",
+            job_name=utils.unique_name_from_base("clarify-bias-and-explainability"),
+            wait=True,
+        )
+        analysis_result_json = s3.S3Downloader.read_file(
+            data_config.s3_output_path + "/analysis.json",
+            sagemaker_session,
+        )
+        analysis_result = json.loads(analysis_result_json)
+        assert (
+            math.fabs(
+                analysis_result["explanations"]["kernel_shap"]["label0"]["global_shap_values"]["F2"]
+            )
+            <= 1
+        )
+        check_analysis_config(data_config, sagemaker_session, "shap")
+
+        assert (
+            math.fabs(
+                analysis_result["post_training_bias_metrics"]["facets"]["F1"][0]["metrics"][0][
+                    "value"
+                ]
+            )
+            <= 1.0
+        )
+        check_analysis_config(data_config, sagemaker_session, "post_training_bias")
+
+
 def check_analysis_config(data_config, sagemaker_session, method):
     analysis_config_json = s3.S3Downloader.read_file(
         data_config.s3_output_path + "/analysis_config.json",

tests/integ/test_clarify.py

@@ -468,14 +468,18 @@ def query(sagemaker_session_mock):
 
 
 def test_athena_query_run(sagemaker_session_mock, query):
+    WORKGROUP = "workgroup"
     sagemaker_session_mock.start_query_execution.return_value = {"QueryExecutionId": "query_id"}
-    query.run(query_string="query", output_location="s3://some-bucket/some-path")
+    query.run(
+        query_string="query", output_location="s3://some-bucket/some-path", workgroup=WORKGROUP
+    )
     sagemaker_session_mock.start_query_execution.assert_called_with(
         catalog="catalog",
         database="database",
         query_string="query",
         output_location="s3://some-bucket/some-path",
         kms_key=None,
+        workgroup=WORKGROUP,
     )
     assert "some-bucket" == query._result_bucket
     assert "some-path" == query._result_file_prefix

tests/unit/sagemaker/feature_store/test_feature_store.py

@@ -483,6 +483,7 @@ def test_fit_ps(time, strftime, sagemaker_session):
     expected_train_args = _create_train_job("1.11", ps=True, py_version="py2")
     expected_train_args["input_config"][0]["DataSource"]["S3DataSource"]["S3Uri"] = inputs
     expected_train_args["hyperparameters"][TensorFlow.LAUNCH_PS_ENV_NAME] = json.dumps(True)
+    expected_train_args["environment"] = {"USE_SMDEBUG": "0"}
 
     actual_train_args = sagemaker_session.method_calls[0][2]
     assert actual_train_args == expected_train_args

tests/unit/sagemaker/tensorflow/test_estimator.py

@@ -370,6 +370,10 @@ def test_training_step_base_estimator(sagemaker_session):
             },
             "RoleArn": ROLE,
             "StoppingCondition": {"MaxRuntimeInSeconds": 86400},
+            "DebugHookConfig": {
+                "S3OutputPath": {"Std:Join": {"On": "/", "Values": ["s3:/", "a", "b"]}},
+                "CollectionConfigurations": [],
+            },
             "ProfilerConfig": {
                 "ProfilingIntervalInMilliseconds": 500,
                 "S3OutputPath": {"Std:Join": {"On": "/", "Values": ["s3:/", "a", "b"]}},