Deployment Configs - Follow-ups (aws#4626)

* Init Deployment configs outside Model init.

* Testing with NB

* Testing with NB-V2

* Refactoring, NB testing

* NB Testing and Refactoring

* Testing

* Refactoring

* Testing with NB

* Debug

* Debug display API

* Debug with NB

* Testing with NB

* Refactoring

* Refactoring

* Refactoring and NB testing

* Testing with NB

* Refactoring

* Prefix instance type with ml

* Fix unit tests

---------

Co-authored-by: Jonathan Makunga <[email protected]>

Author: 2 people

src/sagemaker/jumpstart/model.py

@@ -41,18 +41,17 @@
 from sagemaker.jumpstart.types import (
     JumpStartSerializablePayload,
     DeploymentConfigMetadata,
-    JumpStartBenchmarkStat,
-    JumpStartMetadataConfig,
 )
 from sagemaker.jumpstart.utils import (
     validate_model_id_and_get_type,
     verify_model_region_and_return_specs,
     get_jumpstart_configs,
     get_metrics_from_deployment_configs,
+    add_instance_rate_stats_to_benchmark_metrics,
 )
 from sagemaker.jumpstart.constants import JUMPSTART_LOGGER
 from sagemaker.jumpstart.enums import JumpStartModelType
-from sagemaker.utils import stringify_object, format_tags, Tags, get_instance_rate_per_hour
+from sagemaker.utils import stringify_object, format_tags, Tags
 from sagemaker.model import (
     Model,
     ModelPackage,
@@ -361,17 +360,13 @@ def _validate_model_id_and_type():
         self.model_package_arn = model_init_kwargs.model_package_arn
         self.init_kwargs = model_init_kwargs.to_kwargs_dict(False)
 
-        metadata_configs = get_jumpstart_configs(
+        self._metadata_configs = get_jumpstart_configs(
             region=self.region,
             model_id=self.model_id,
             model_version=self.model_version,
             sagemaker_session=self.sagemaker_session,
             model_type=self.model_type,
         )
-        self._deployment_configs = [
-            self._convert_to_deployment_config_metadata(config_name, config)
-            for config_name, config in metadata_configs.items()
-        ]
 
     def log_subscription_warning(self) -> None:
         """Log message prompting the customer to subscribe to the proprietary model."""
@@ -449,33 +444,46 @@ def set_deployment_config(self, config_name: str, instance_type: str) -> None:
 
     @property
     def deployment_config(self) -> Optional[Dict[str, Any]]:
-        """The deployment config that will be applied to the model.
+        """The deployment config that will be applied to ``This`` model.
 
         Returns:
-            Optional[Dict[str, Any]]: Deployment config that will be applied to the model.
+            Optional[Dict[str, Any]]: Deployment config.
         """
-        return self._retrieve_selected_deployment_config(self.config_name)
+        deployment_config = self._retrieve_selected_deployment_config(
+            self.config_name, self.instance_type
+        )
+        return deployment_config.to_json() if deployment_config is not None else None
 
     @property
     def benchmark_metrics(self) -> pd.DataFrame:
-        """Benchmark Metrics for deployment configs
+        """Benchmark Metrics for deployment configs.
 
         Returns:
-            Metrics: Pandas DataFrame object.
+            Benchmark Metrics: Pandas DataFrame object.
         """
-        return pd.DataFrame(self._get_benchmarks_data(self.config_name))
+        benchmark_metrics_data = self._get_deployment_configs_benchmarks_data(
+            self.config_name, self.instance_type
+        )
+        keys = list(benchmark_metrics_data.keys())
+        df = pd.DataFrame(benchmark_metrics_data).sort_values(by=[keys[0], keys[1]])
+        return df
 
     def display_benchmark_metrics(self) -> None:
-        """Display Benchmark Metrics for deployment configs."""
-        print(self.benchmark_metrics.to_markdown())
+        """Display deployment configs benchmark metrics."""
+        print(self.benchmark_metrics.to_markdown(index=False))
 
     def list_deployment_configs(self) -> List[Dict[str, Any]]:
         """List deployment configs for ``This`` model.
 
         Returns:
             List[Dict[str, Any]]: A list of deployment configs.
         """
-        return self._deployment_configs
+        return [
+            deployment_config.to_json()
+            for deployment_config in self._get_deployment_configs(
+                self.config_name, self.instance_type
+            )
+        ]
 
     def _create_sagemaker_model(
         self,
@@ -866,92 +874,94 @@ def register_deploy_wrapper(*args, **kwargs):
         return model_package
 
     @lru_cache
-    def _get_benchmarks_data(self, config_name: str) -> Dict[str, List[str]]:
+    def _get_deployment_configs_benchmarks_data(
+        self, config_name: str, instance_type: str
+    ) -> Dict[str, Any]:
         """Deployment configs benchmark metrics.
 
         Args:
-            config_name (str): The name of the selected deployment config.
+            config_name (str): Name of selected deployment config.
+            instance_type (str): The selected Instance type.
         Returns:
             Dict[str, List[str]]: Deployment config benchmark data.
         """
         return get_metrics_from_deployment_configs(
-            self._deployment_configs,
-            config_name,
+            self._get_deployment_configs(config_name, instance_type)
         )
 
     @lru_cache
-    def _retrieve_selected_deployment_config(self, config_name: str) -> Optional[Dict[str, Any]]:
-        """Retrieve the deployment config to apply to the model.
+    def _retrieve_selected_deployment_config(
+        self, config_name: str, instance_type: str
+    ) -> Optional[DeploymentConfigMetadata]:
+        """Retrieve the deployment config to apply to `This` model.
 
         Args:
             config_name (str): The name of the deployment config to retrieve.
+            instance_type (str): The instance type of the deployment config to retrieve.
         Returns:
             Optional[Dict[str, Any]]: The retrieved deployment config.
         """
         if config_name is None:
             return None
 
-        for deployment_config in self._deployment_configs:
-            if deployment_config.get("DeploymentConfigName") == config_name:
+        for deployment_config in self._get_deployment_configs(config_name, instance_type):
+            if deployment_config.deployment_config_name == config_name:
                 return deployment_config
         return None
 
-    def _convert_to_deployment_config_metadata(
-        self, config_name: str, metadata_config: JumpStartMetadataConfig
-    ) -> Dict[str, Any]:
-        """Retrieve deployment config for config name.
+    @lru_cache
+    def _get_deployment_configs(
+        self, selected_config_name: str, selected_instance_type: str
+    ) -> List[DeploymentConfigMetadata]:
+        """Retrieve deployment configs metadata.
 
         Args:
-            config_name (str): Name of deployment config.
-            metadata_config (JumpStartMetadataConfig): Metadata config for deployment config.
-        Returns:
-            A deployment metadata config for config name (dict[str, Any]).
+            selected_config_name (str): The name of the selected deployment config.
+            selected_instance_type (str): The selected instance type.
         """
-        default_inference_instance_type = metadata_config.resolved_config.get(
-            "default_inference_instance_type"
-        )
-
-        benchmark_metrics = (
-            metadata_config.benchmark_metrics.get(default_inference_instance_type)
-            if metadata_config.benchmark_metrics is not None
-            else None
-        )
-
-        should_fetch_instance_rate_metric = True
-        if benchmark_metrics is not None:
-            for benchmark_metric in benchmark_metrics:
-                if benchmark_metric.name.lower() == "instance rate":
-                    should_fetch_instance_rate_metric = False
-                    break
-
-        if should_fetch_instance_rate_metric:
-            instance_rate = get_instance_rate_per_hour(
-                instance_type=default_inference_instance_type, region=self.region
+        deployment_configs = []
+        if self._metadata_configs is None:
+            return deployment_configs
+
+        err = None
+        for config_name, metadata_config in self._metadata_configs.items():
+            if err is None or "is not authorized to perform: pricing:GetProducts" not in err:
+                err, metadata_config.benchmark_metrics = (
+                    add_instance_rate_stats_to_benchmark_metrics(
+                        self.region, metadata_config.benchmark_metrics
+                    )
+                )
+
+            resolved_config = metadata_config.resolved_config
+            if selected_config_name == config_name:
+                instance_type_to_use = selected_instance_type
+            else:
+                instance_type_to_use = resolved_config.get("default_inference_instance_type")
+
+            init_kwargs = get_init_kwargs(
+                model_id=self.model_id,
+                instance_type=instance_type_to_use,
+                sagemaker_session=self.sagemaker_session,
             )
-            if instance_rate is not None:
-                instance_rate_metric = JumpStartBenchmarkStat(instance_rate)
-
-                if benchmark_metrics is None:
-                    benchmark_metrics = [instance_rate_metric]
-                else:
-                    benchmark_metrics.append(instance_rate_metric)
-
-        init_kwargs = get_init_kwargs(
-            model_id=self.model_id,
-            instance_type=default_inference_instance_type,
-            sagemaker_session=self.sagemaker_session,
-        )
-        deploy_kwargs = get_deploy_kwargs(
-            model_id=self.model_id,
-            instance_type=default_inference_instance_type,
-            sagemaker_session=self.sagemaker_session,
-        )
+            deploy_kwargs = get_deploy_kwargs(
+                model_id=self.model_id,
+                instance_type=instance_type_to_use,
+                sagemaker_session=self.sagemaker_session,
+            )
+            deployment_config_metadata = DeploymentConfigMetadata(
+                config_name,
+                metadata_config.benchmark_metrics,
+                resolved_config,
+                init_kwargs,
+                deploy_kwargs,
+            )
+            deployment_configs.append(deployment_config_metadata)
 
-        deployment_config_metadata = DeploymentConfigMetadata(
-            config_name, benchmark_metrics, init_kwargs, deploy_kwargs
-        )
+        if err is not None and "is not authorized to perform: pricing:GetProducts" in err:
+            error_message = "Instance rate metrics will be omitted. Reason: %s"
+            JUMPSTART_LOGGER.warning(error_message, err)
 
-        return deployment_config_metadata.to_json()
+        return deployment_configs
 
     def __str__(self) -> str:
         """Overriding str(*) method to make more human-readable."""

src/sagemaker/jumpstart/types.py

@@ -2238,29 +2238,37 @@ def to_json(self) -> Dict[str, Any]:
             if hasattr(self, att):
                 cur_val = getattr(self, att)
                 att = self._convert_to_pascal_case(att)
-                if issubclass(type(cur_val), JumpStartDataHolderType):
-                    json_obj[att] = cur_val.to_json()
-                elif isinstance(cur_val, list):
-                    json_obj[att] = []
-                    for obj in cur_val:
-                        if issubclass(type(obj), JumpStartDataHolderType):
-                            json_obj[att].append(obj.to_json())
-                        else:
-                            json_obj[att].append(obj)
-                elif isinstance(cur_val, dict):
-                    json_obj[att] = {}
-                    for key, val in cur_val.items():
-                        if issubclass(type(val), JumpStartDataHolderType):
-                            json_obj[att][self._convert_to_pascal_case(key)] = val.to_json()
-                        else:
-                            json_obj[att][key] = val
-                else:
-                    json_obj[att] = cur_val
+                json_obj[att] = self._val_to_json(cur_val)
         return json_obj
 
+    def _val_to_json(self, val: Any) -> Any:
+        """Converts the given value to JSON.
+
+        Args:
+            val (Any): The value to convert.
+        Returns:
+            Any: The converted json value.
+        """
+        if issubclass(type(val), JumpStartDataHolderType):
+            return val.to_json()
+        if isinstance(val, list):
+            list_obj = []
+            for obj in val:
+                list_obj.append(self._val_to_json(obj))
+            return list_obj
+        if isinstance(val, dict):
+            dict_obj = {}
+            for k, v in val.items():
+                if isinstance(v, JumpStartDataHolderType):
+                    dict_obj[self._convert_to_pascal_case(k)] = self._val_to_json(v)
+                else:
+                    dict_obj[k] = self._val_to_json(v)
+            return dict_obj
+        return val
+
 
 class DeploymentArgs(BaseDeploymentConfigDataHolder):
-    """Dataclass representing a Deployment Config."""
+    """Dataclass representing a Deployment Args."""
 
     __slots__ = [
         "image_uri",
@@ -2273,9 +2281,12 @@ class DeploymentArgs(BaseDeploymentConfigDataHolder):
     ]
 
     def __init__(
-        self, init_kwargs: JumpStartModelInitKwargs, deploy_kwargs: JumpStartModelDeployKwargs
+        self,
+        init_kwargs: Optional[JumpStartModelInitKwargs] = None,
+        deploy_kwargs: Optional[JumpStartModelDeployKwargs] = None,
+        resolved_config: Optional[Dict[str, Any]] = None,
     ):
-        """Instantiates DeploymentConfig object."""
+        """Instantiates DeploymentArgs object."""
         if init_kwargs is not None:
             self.image_uri = init_kwargs.image_uri
             self.model_data = init_kwargs.model_data
@@ -2290,6 +2301,11 @@ def __init__(
             self.container_startup_health_check_timeout = (
                 deploy_kwargs.container_startup_health_check_timeout
             )
+        if resolved_config is not None:
+            self.default_instance_type = resolved_config.get("default_inference_instance_type")
+            self.supported_instance_types = resolved_config.get(
+                "supported_inference_instance_types"
+            )
 
 
 class DeploymentConfigMetadata(BaseDeploymentConfigDataHolder):
@@ -2304,13 +2320,15 @@ class DeploymentConfigMetadata(BaseDeploymentConfigDataHolder):
 
     def __init__(
         self,
-        config_name: str,
-        benchmark_metrics: List[JumpStartBenchmarkStat],
-        init_kwargs: JumpStartModelInitKwargs,
-        deploy_kwargs: JumpStartModelDeployKwargs,
+        config_name: Optional[str] = None,
+        benchmark_metrics: Optional[Dict[str, List[JumpStartBenchmarkStat]]] = None,
+        resolved_config: Optional[Dict[str, Any]] = None,
+        init_kwargs: Optional[JumpStartModelInitKwargs] = None,
+        deploy_kwargs: Optional[JumpStartModelDeployKwargs] = None,
     ):
         """Instantiates DeploymentConfigMetadata object."""
         self.deployment_config_name = config_name
-        self.deployment_args = DeploymentArgs(init_kwargs, deploy_kwargs)
-        self.acceleration_configs = None
+        self.deployment_args = DeploymentArgs(init_kwargs, deploy_kwargs, resolved_config)
         self.benchmark_metrics = benchmark_metrics
+        if resolved_config is not None:
+            self.acceleration_configs = resolved_config.get("acceleration_configs")