feat(core): add HubContent.get_all and keep OriginalCreationTime on listed hub content

sagemaker-core/src/sagemaker/core/resources.py

@@ -13830,6 +13830,7 @@ class HubContent(Base):
         hub_content_dependencies: The location of any dependencies that the hub content has, such as scripts, model artifacts, datasets, or notebooks.
         failure_reason: The failure reason if importing hub content failed.
         last_modified_time: The last modified time of the hub content.
+        original_creation_time: The date and time when the hub content was originally created, before any updates or revisions.
 
     """
 
@@ -13854,6 +13855,7 @@ class HubContent(Base):
     creation_time: Optional[datetime.datetime] = Unassigned()
     last_modified_time: Optional[datetime.datetime] = Unassigned()
     hub_name: Optional[str] = Unassigned()
+    original_creation_time: Optional[datetime.datetime] = Unassigned()
 
     def get_name(self) -> str:
         attributes = vars(self)
@@ -14210,6 +14212,85 @@ def load(
             region=region,
         )
 
+    @classmethod
+    @Base.add_validate_call
+    def get_all(
+        cls,
+        hub_name: StrPipeVar,
+        hub_content_type: StrPipeVar,
+        name_contains: Optional[StrPipeVar] = Unassigned(),
+        max_schema_version: Optional[StrPipeVar] = Unassigned(),
+        creation_time_before: Optional[datetime.datetime] = Unassigned(),
+        creation_time_after: Optional[datetime.datetime] = Unassigned(),
+        sort_by: Optional[StrPipeVar] = Unassigned(),
+        sort_order: Optional[StrPipeVar] = Unassigned(),
+        session: Optional[Session] = None,
+        region: Optional[StrPipeVar] = None,
+    ) -> ResourceIterator["HubContent"]:
+        """
+        Get all HubContent resources
+
+        Parameters:
+            hub_name: The name of the hub to list the contents of.
+            hub_content_type: The type of hub content to list.
+            name_contains: Only list hub content if the name contains the specified string.
+            max_schema_version: The upper bound of the hub content schema verion.
+            creation_time_before: Only list hub content that was created before the time specified.
+            creation_time_after: Only list hub content that was created after the time specified.
+            sort_by: Sort hub content versions by either name or creation time.
+            sort_order: Sort hubs by ascending or descending order.
+            max_results: The maximum amount of hub content to list.
+            next_token: If the response to a previous ListHubContents request was truncated, the response includes a NextToken. To retrieve the next set of hub content, use the token in the next request.
+            session: Boto3 session.
+            region: Region name.
+
+        Returns:
+            Iterator for listed HubContent resources.
+
+        Raises:
+            botocore.exceptions.ClientError: This exception is raised for AWS service related errors.
+                The error message and error code can be parsed from the exception as follows:
+                ```
+                try:
+                    # AWS service call here
+                except botocore.exceptions.ClientError as e:
+                    error_message = e.response['Error']['Message']
+                    error_code = e.response['Error']['Code']
+                ```
+            ResourceNotFound: Resource being access is not found.
+        """
+
+        client = Base.get_sagemaker_client(
+            session=session, region_name=region, service_name="sagemaker"
+        )
+
+        operation_input_args = {
+            "HubName": hub_name,
+            "HubContentType": hub_content_type,
+            "NameContains": name_contains,
+            "MaxSchemaVersion": max_schema_version,
+            "CreationTimeBefore": creation_time_before,
+            "CreationTimeAfter": creation_time_after,
+            "SortBy": sort_by,
+            "SortOrder": sort_order,
+        }
+
+        # serialize the input request
+        operation_input_args = serialize(operation_input_args)
+        logger.debug(f"Serialized input request: {operation_input_args}")
+
+        return ResourceIterator(
+            client=client,
+            session=session,
+            region=region,
+            list_method="list_hub_contents",
+            summaries_key="HubContentSummaries",
+            summary_name="HubContentInfo",
+            resource_cls=HubContent,
+            extra_init_data={"hub_name": hub_name},
+            list_method_kwargs=operation_input_args,
+        )
+
     @Base.add_validate_call
     def get_all_versions(
         self,
@@ -14279,6 +14360,7 @@ def get_all_versions(
             summary_name="HubContentInfo",
             resource_cls=HubContent,
             list_method_kwargs=operation_input_args,
+            extra_init_data={"hub_name": self.hub_name},
         )
 
 

sagemaker-core/src/sagemaker/core/tools/resources_codegen.py

@@ -106,6 +106,25 @@
     "ModelPackage": {"ModelPackageName": "model_package_arn"},
 }
 
+# Some List APIs return a summary member that the Describe API does not. A resource class
+# is generated from the Describe output, so ResourceIterator drops such members when it
+# builds the resource from the summary (`extra="forbid"` leaves no other option). This
+# maps a resource to the summary members it keeps as optional attributes. ListHubContents
+# returns OriginalCreationTime, the model-level catalog date, while DescribeHubContent
+# returns only the version-level CreationTime.
+LIST_SUMMARY_EXTRA_ATTRIBUTES = {
+    "HubContent": {"OriginalCreationTime": "Optional[datetime.datetime] = Unassigned()"},
+}
+
+# Some List APIs take an input member that the resource's Describe input requires but the
+# list summary does not return (ListHubContents takes HubName, HubContentInfo omits it).
+# get_all() cannot build a resource from such a summary unless the list argument is
+# copied onto every resource. This maps a resource to the list-input members to copy, as
+# {summary-less member: resource attribute}.
+GET_ALL_INPUT_ATTRIBUTES = {
+    "HubContent": {"HubName": "hub_name"},
+}
+
 
 class ResourcesCodeGen:
     """A class for generating resources based on a service JSON file.
@@ -535,8 +554,15 @@ def _get_class_attributes(self, resource_name: str, class_methods: list) -> tupl
                 class_attributes["hub_name"] = "Optional[str] = Unassigned()"
                 class_attributes_string = class_attributes_string.replace("hub_name: str", "")
                 class_attributes_string = (
-                    class_attributes_string + "hub_name: Optional[str] = Unassigned()"
+                    class_attributes_string + "hub_name: Optional[str] = Unassigned()\n"
                 )
+            for member, annotation in LIST_SUMMARY_EXTRA_ATTRIBUTES.get(resource_name, {}).items():
+                attribute = convert_to_snake_case(member)
+                class_attributes[attribute] = annotation
+                class_attributes_string += f"{attribute}: {annotation}\n"
+                attributes_and_documentation[member] = self.shapes_extractor.combined_shapes[
+                    self._list_summary_shape(resource_name)
+                ]["members"][member].get("documentation")
             if resource_name == "Endpoint":
                 class_attributes["serializer"] = "Optional[BaseSerializer] = None"
                 class_attributes_string = class_attributes_string.replace(
@@ -1714,6 +1740,21 @@ def generate_additional_get_all_method(self, method: Method, resource_attributes
             "list_method_kwargs=operation_input_args",
         ]
 
+        # An object method lists resources of its own type under its own identifiers
+        # (ListHubContentVersions lists the versions of one HubContent). The summaries
+        # omit the identifiers the list call was made with, so copy them onto every
+        # resource or the first refresh() fails with a missing required parameter.
+        if (
+            method.method_type == MethodType.OBJECT.value
+            and method.return_type == method.resource_name
+        ):
+            input_attributes = GET_ALL_INPUT_ATTRIBUTES.get(method.resource_name, {})
+            if input_attributes:
+                extra_init = ", ".join(
+                    f"'{attribute}': self.{attribute}" for attribute in input_attributes.values()
+                )
+                resource_iterator_args_list.append(f"extra_init_data={{{extra_init}}}")
+
         resource_iterator_args = ",\n".join(resource_iterator_args_list)
         resource_iterator_args = add_indent(resource_iterator_args, 8)
         serialize_operation_input = SERIALIZE_INPUT_TEMPLATE.format(
@@ -1934,6 +1975,14 @@ def generate_wait_for_delete_method(self, resource_name: str) -> str:
         )
         return formatted_method
 
+    def _list_summary_shape(self, resource_name: str) -> str:
+        """Return the summary shape name of the resource's List API output."""
+        operation_name = "List" + resource_name + "s"
+        output_shape = self.operations[operation_name]["output"]["shape"]
+        output_members = self.shapes[output_shape]["members"]
+        summaries_key = next(key for key in output_members if key != "NextToken")
+        return self.shapes[output_members[summaries_key]["shape"]]["member"]["shape"]
+
     def generate_get_all_method(self, resource_name: str) -> str:
         """Auto-Generate 'get_all' class Method [list API] for a resource.
 
@@ -1974,7 +2023,13 @@ def generate_get_all_method(self, resource_name: str) -> str:
             get_operation_required_input = []
 
         custom_key_mapping_str = ""
-        if any(member not in summary_members for member in get_operation_required_input):
+        input_attributes = GET_ALL_INPUT_ATTRIBUTES.get(resource_name, {})
+        missing_members = [
+            member
+            for member in get_operation_required_input
+            if member not in summary_members and member not in input_attributes
+        ]
+        if missing_members:
             if "MonitoringJobDefinitionSummary" == summary_name:
                 custom_key_mapping = {
                     "monitoring_job_definition_name": "job_definition_name",
@@ -2002,6 +2057,13 @@ def generate_get_all_method(self, resource_name: str) -> str:
         if custom_key_mapping_str:
             resource_iterator_args_list.append("custom_key_mapping=custom_key_mapping")
 
+        if input_attributes:
+            extra_init = ", ".join(
+                f"'{attribute}': {convert_to_snake_case(member)}"
+                for member, attribute in input_attributes.items()
+            )
+            resource_iterator_args_list.append(f"extra_init_data={{{extra_init}}}")
+
         exclude_list = ["next_token", "max_results"]
         get_all_args = self._generate_method_args(operation_input_shape_name, exclude_list)
 

sagemaker-core/src/sagemaker/core/utils/utils.py

@@ -546,6 +546,7 @@ def __init__(
         custom_key_mapping: dict = None,
         session=None,
         region: Optional[str] = None,
+        extra_init_data: dict = None,
     ):
         """Initialize a ResourceIterator object
 
@@ -563,6 +564,9 @@ def __init__(
                 Each returned resource is bound to it, so its refresh and other object methods
                 use the same account. Defaults to None (the process default client).
             region (str, optional): The region the list call was made in. Defaults to None.
+            extra_init_data (dict, optional): Attributes set on every resource in addition to
+                the summary, for a list argument the summary does not echo back (for example
+                the ``hub_name`` a ``ListHubContents`` call was made with). Defaults to None.
         """
         self.summaries_key = summaries_key
         self.summary_name = summary_name
@@ -572,6 +576,7 @@ def __init__(
         self.custom_key_mapping = custom_key_mapping
         self.session = session
         self.region = region
+        self.extra_init_data = extra_init_data or {}
 
         self.resource_cls = resource_cls
         self.index = 0
@@ -605,6 +610,8 @@ def __next__(self) -> T:
                 # Filter out the fields that are not in the resource class
                 fields = self.resource_cls.__annotations__
                 init_data = {k: v for k, v in init_data.items() if k in fields}
+                for key, value in self.extra_init_data.items():
+                    init_data.setdefault(key, value)
 
                 resource_object = self.resource_cls(**init_data)