feat: Add application/x-parquet support to clarify.ModelConfig (Issue #3450)

sagemaker-core/src/sagemaker/core/clarify/__init__.py

@@ -380,12 +380,19 @@
                     "image/jpeg",
                     "image/png",
                     "application/x-npy",
+                    "application/x-parquet",
                 ),
             ),
             SchemaOptional("accept_type"): And(
                 str,
                 Use(str.lower),
-                lambda s: s in ("text/csv", "application/jsonlines", "application/json"),
+                lambda s: s
+                in (
+                    "text/csv",
+                    "application/jsonlines",
+                    "application/json",
+                    "application/x-parquet",
+                ),
             ),
             SchemaOptional("label"): Or(str, int),
             SchemaOptional("probability"): Or(str, int),
@@ -1021,11 +1028,13 @@ def __init__(
                 Must be set with ``instance_count``, ``model_name``
             accept_type (str): The model output format to be used for getting inferences with the
                 shadow endpoint. Valid values are ``"text/csv"`` for CSV,
-                ``"application/jsonlines"`` for JSON Lines, and ``"application/json"`` for JSON.
+                ``"application/jsonlines"`` for JSON Lines, ``"application/json"`` for JSON, and
+                ``"application/x-parquet"`` for Parquet.
                 Default is the same as ``content_type``.
             content_type (str): The model input format to be used for getting inferences with the
                 shadow endpoint. Valid values are ``"text/csv"`` for CSV,
-                ``"application/jsonlines"`` for JSON Lines, and ``"application/json"`` for JSON.
+                ``"application/jsonlines"`` for JSON Lines, ``"application/json"`` for JSON, and
+                ``"application/x-parquet"`` for Parquet.
                 Default is the same as ``dataset_format``.
             content_template (str): A template string to be used to construct the model input from
                 dataset instances. It is only used, and required, when ``model_content_type`` is
@@ -1161,10 +1170,15 @@ def __init__(
                 )
             self.predictor_config["endpoint_name_prefix"] = endpoint_name_prefix
         if accept_type is not None:
-            if accept_type not in ["text/csv", "application/jsonlines", "application/json"]:
+            if accept_type not in [
+                "text/csv",
+                "application/jsonlines",
+                "application/json",
+                "application/x-parquet",
+            ]:
                 raise ValueError(
                     f"Invalid accept_type {accept_type}."
-                    f" Please choose text/csv or application/jsonlines."
+                    f" Please choose text/csv, application/jsonlines, application/json, or application/x-parquet."
                 )
             if time_series_model_config and accept_type == "text/csv":
                 raise ValueError(
@@ -1180,10 +1194,11 @@ def __init__(
                 "image/jpg",
                 "image/png",
                 "application/x-npy",
+                "application/x-parquet",
             ]:
                 raise ValueError(
                     f"Invalid content_type {content_type}."
-                    f" Please choose text/csv or application/jsonlines."
+                    f" Please choose text/csv, application/jsonlines, application/json, image/jpeg, image/jpg, image/png, application/x-npy, or application/x-parquet."
                 )
             if content_type == "application/jsonlines":
                 if content_template is None:

sagemaker-core/src/sagemaker/core/workflow/job_config_document.py

@@ -28,6 +28,7 @@
   number or boolean reaches the service quoted. Parameterise string-valued
   fields only.
 """
+
 from __future__ import absolute_import
 
 import json
@@ -67,9 +68,7 @@ def _placeholder(obj):
         if isinstance(obj, PipelineVariable):
             variables.append(obj)
             return "%s%d__" % (token_prefix, len(variables) - 1)
-        raise TypeError(
-            "Object of type %s is not JSON serializable" % obj.__class__.__name__
-        )
+        raise TypeError("Object of type %s is not JSON serializable" % obj.__class__.__name__)
 
     document = json.dumps(job_config, default=_placeholder)
     if not variables:

sagemaker-core/src/sagemaker/core/workflow/pipeline_capture.py

@@ -11,6 +11,7 @@
 # ANY KIND, either express or implied. See the License for the specific
 # language governing permissions and limitations under the License.
 """Shared capture paths for producers running under a ``PipelineSession``."""
+
 from __future__ import absolute_import
 
 from typing import Any, Dict, List, Optional
@@ -87,9 +88,11 @@ def capture_training_request(pipeline_session, create_args: Dict[str, Any]) -> N
         # Dict-form tags arrive in both shapes: lowercase keys from the SDK's own
         # tag builders, PascalCase from callers passing wire-form dicts through.
         request["Tags"] = [
-            Tag(key=tag.get("key", tag.get("Key")), value=tag.get("value", tag.get("Value")))
-            if isinstance(tag, dict)
-            else tag
+            (
+                Tag(key=tag.get("key", tag.get("Key")), value=tag.get("value", tag.get("Value")))
+                if isinstance(tag, dict)
+                else tag
+            )
             for tag in request["Tags"]
         ]
     pipeline_session._intercept_create_request(serialize(request), None, "train")

sagemaker-core/tests/unit/test_clarify.py

@@ -566,6 +566,20 @@ def test_init_with_endpoint_name(self):
         config = ModelConfig(endpoint_name="my-endpoint")
         assert config.predictor_config["endpoint_name"] == "my-endpoint"
 
+    def test_init_with_parquet_content_and_accept_type(self):
+        """Test initialization with application/x-parquet content_type and accept_type."""
+        from sagemaker.core.clarify import ModelConfig
+
+        config = ModelConfig(
+            model_name="my-model",
+            instance_count=1,
+            instance_type="ml.m5.xlarge",
+            content_type="application/x-parquet",
+            accept_type="application/x-parquet",
+        )
+        assert config.predictor_config["content_type"] == "application/x-parquet"
+        assert config.predictor_config["accept_type"] == "application/x-parquet"
+
     def test_init_with_invalid_endpoint_prefix_raises_error(self):
         """Test that invalid endpoint_name_prefix raises ValueError."""
         from sagemaker.core.clarify import ModelConfig

sagemaker-core/tests/unit/workflow/test_job_config_document.py

@@ -11,6 +11,7 @@
 # ANY KIND, either express or implied. See the License for the specific
 # language governing permissions and limitations under the License.
 """Tests for the ``JobConfigDocument`` encoder."""
+
 from __future__ import absolute_import
 
 import json

sagemaker-core/tests/unit/workflow/test_pipeline_capture.py

@@ -11,6 +11,7 @@
 # ANY KIND, either express or implied. See the License for the specific
 # language governing permissions and limitations under the License.
 """Tests for the shared ``PipelineSession`` capture helpers."""
+
 from __future__ import absolute_import
 
 import json

sagemaker-core/tests/unit/workflow/test_retrieve_caller_name.py

@@ -19,6 +19,7 @@
 -- `ProcessingStep` expects `"run"` from here while interception is called with
 `"process"` -- so a step guard cannot be satisfied by the interception name alone.
 """
+
 from __future__ import absolute_import
 
 import pytest