Append uuid to endpoint name (#90)

* Add unit test and fix HyperPod Manager

1. Default namespace can be set by HyperpodManager.set_context()

2. Added unit tests for inference

* Remove debug print

* Append uuid to model name and endpoint name

* minor fix in create method

Author: zhaoqizqwang

src/sagemaker/hyperpod/common/utils.py

@@ -3,6 +3,7 @@
 from pydantic import ValidationError
 from kubernetes.client.exceptions import ApiException
 from kubernetes import config
+import uuid
 
 
 def validate_cluster_connection():
@@ -37,13 +38,13 @@ def handle_exception(e: Exception, name: str, namespace: str):
             raise Exception(f"Credentials unauthorized.") from e
         elif e.status == 403:
             raise Exception(
-                f"Access denied to resource '{name}' in '{namespace}'."
+                f"Access denied to resource '{name}' in namespace '{namespace}'."
             ) from e
         if e.status == 404:
-            raise Exception(f"Resource '{name}' not found in '{namespace}'.") from e
+            raise Exception(f"Resource '{name}' not found in namespace '{namespace}'.") from e
         elif e.status == 409:
             raise Exception(
-                f"Resource '{name}' already exists in '{namespace}'."
+                f"Resource '{name}' already exists in namespace '{namespace}'."
             ) from e
         elif 500 <= e.status < 600:
             raise Exception("Kubernetes API internal server error.") from e
@@ -54,3 +55,6 @@ def handle_exception(e: Exception, name: str, namespace: str):
         raise Exception("Response did not match expected schema.") from e
 
     raise e
+
+def append_uuid(name: str) -> str:
+    return f"{name}-{str(uuid.uuid4())[:4]}"

src/sagemaker/hyperpod/inference/config/common.py

@@ -1,5 +1,6 @@
 from pydantic import BaseModel, ConfigDict, Field
 from typing import Optional, List, Dict, Union, Literal
+from sagemaker.hyperpod.common.config import *
 
 
 class Dimensions(BaseModel):

src/sagemaker/hyperpod/inference/config/hp_endpoint_config.py

@@ -1,5 +1,6 @@
 from pydantic import BaseModel, ConfigDict, Field
 from typing import Optional, List, Dict, Union, Literal
+from sagemaker.hyperpod.common.config import *
 
 
 class Dimensions(BaseModel):

src/sagemaker/hyperpod/inference/config/hp_jumpstart_endpoint_config.py

@@ -1,5 +1,6 @@
 from sagemaker.hyperpod.common.config.metadata import Metadata
 from sagemaker.hyperpod.inference.config.constants import *
+from sagemaker.hyperpod.common.utils import append_uuid
 from sagemaker.hyperpod.inference.config.hp_endpoint_config import (
     InferenceEndpointConfigStatus,
     _HPEndpoint,
@@ -27,14 +28,18 @@ def create(
             self.get_logger().setLevel(logging.DEBUG)
         else:
             self.get_logger().setLevel(logging.INFO)
+
         spec = _HPEndpoint(**self.model_dump(by_alias=True, exclude_none=True))
 
         if not name:
-            name = spec.modelName
+            name = append_uuid(spec.modelName)
 
         if not namespace:
             namespace = get_default_namespace()
 
+        if spec.endpointName:
+            spec.endpointName = append_uuid(spec.endpointName)
+
         self.call_create_api(
             name=name,  # use model name as metadata name
             kind=INFERENCE_ENDPOINT_CONFIG_KIND,
@@ -60,11 +65,14 @@ def create_from_dict(
         spec = _HPEndpoint.model_validate(input, by_name=True)
 
         if not name:
-            name = spec.modelName
+            name = append_uuid(spec.modelName)
 
         if not namespace:
             namespace = get_default_namespace()
 
+        if spec.endpointName:
+            spec.endpointName = append_uuid(spec.endpointName)
+
         self.call_create_api(
             name=name,  # use model name as metadata name
             kind=INFERENCE_ENDPOINT_CONFIG_KIND,

src/sagemaker/hyperpod/inference/hp_endpoint.py

@@ -1,6 +1,7 @@
 from sagemaker.hyperpod.inference.config.constants import *
 from sagemaker.hyperpod.inference.hp_endpoint_base import HPEndpointBase
 from sagemaker.hyperpod.common.config.metadata import Metadata
+from sagemaker.hyperpod.common.utils import append_uuid
 from sagemaker.hyperpod.inference.config.hp_jumpstart_endpoint_config import (
     _HPJumpStartEndpoint,
     JumpStartModelStatus,
@@ -31,7 +32,10 @@ def create(
         spec = _HPJumpStartEndpoint(**self.model_dump(by_alias=True, exclude_none=True))
 
         if not name:
-            name = spec.model.modelId
+            name = append_uuid(spec.model.modelId)
+
+        if spec.sageMakerEndpoint and spec.sageMakerEndpoint.name:
+            spec.sageMakerEndpoint.name = append_uuid(spec.sageMakerEndpoint.name)
 
         if not namespace:
             namespace = get_default_namespace()
@@ -61,7 +65,10 @@ def create_from_dict(
         spec = _HPJumpStartEndpoint.model_validate(input, by_name=True)
 
         if not name:
-            name = spec.model.modelId
+            name = append_uuid(spec.model.modelId)
+
+        if spec.sageMakerEndpoint and spec.sageMakerEndpoint.name:
+            spec.sageMakerEndpoint.name = append_uuid(spec.sageMakerEndpoint.name)
 
         if not namespace:
             namespace = get_default_namespace()

src/sagemaker/hyperpod/inference/hp_jumpstart_endpoint.py

@@ -0,0 +1 @@
+from sagemaker.hyperpod.common.config import *

src/sagemaker/hyperpod/training/config/__init__.py

@@ -29,7 +29,7 @@ def test_handle_api_exception_404(self):
         with self.assertRaises(Exception) as context:
             handle_exception(exception, "test-job", "default")
         self.assertIn(
-            "Resource 'test-job' not found in 'default'", str(context.exception)
+            "Resource 'test-job' not found in namespace 'default'", str(context.exception)
         )
 
     def test_handle_api_exception_409(self):