delete cluster functionality (#247)

Co-authored-by: Mohamed Zeidan <[email protected]>

Author: 2 people

src/sagemaker/hyperpod/cli/common_utils.py

@@ -75,10 +75,10 @@ def parse_comma_separated_list(value: str) -> List[str]:
     """
     Parse a comma-separated string into a list of strings.
     Generic utility that can be reused across commands.
-    
+
     Args:
         value: Comma-separated string like "item1,item2,item3"
-        
+
     Returns:
         List of trimmed strings
     """
@@ -87,35 +87,35 @@ def parse_comma_separated_list(value: str) -> List[str]:
     return [item.strip() for item in value.split(",") if item.strip()]
 
 
-def categorize_resources_by_type(resources: List[Dict[str, Any]], 
+def categorize_resources_by_type(resources: List[Dict[str, Any]],
                                 type_mappings: Dict[str, List[str]]) -> Dict[str, List[str]]:
     """
     Generic function to categorize resources by type.
-    
+
     Args:
         resources: List of resource dictionaries with 'ResourceType' and 'LogicalResourceId'
         type_mappings: Dictionary mapping category names to lists of resource types
-        
+
     Returns:
         Dictionary of category -> list of resource names
     """
     categorized = {category: [] for category in type_mappings.keys()}
     categorized["Other"] = []
-    
+
     for resource in resources:
         resource_type = resource.get("ResourceType", "")
         logical_id = resource.get("LogicalResourceId", "")
-        
+
         # Find which category this resource type belongs to
         category_found = False
         for category, types in type_mappings.items():
             if any(resource_type.startswith(rt) for rt in types):
                 categorized[category].append(logical_id)
                 category_found = True
                 break
-        
+
         if not category_found:
             categorized["Other"].append(logical_id)
-    
+
     # Remove empty categories
     return {k: v for k, v in categorized.items() if v}

src/sagemaker/hyperpod/cluster_management/hp_cluster_stack.py

@@ -510,6 +510,116 @@ def check_status(stack_name: str, region: Optional[str] = None):
               >>> status = HpClusterStack.check_status("my-stack", region="us-west-2")
         """
         return HpClusterStack._get_stack_status_helper(stack_name, region)
+    
+    @staticmethod
+    def delete(stack_name: str, region: Optional[str] = None, retain_resources: Optional[List[str]] = None, 
+                logger: Optional[logging.Logger] = None) -> None:
+        """Deletes a HyperPod cluster CloudFormation stack.
+
+        Removes the specified CloudFormation stack and all associated AWS resources.
+        This operation cannot be undone and proceeds automatically without confirmation.
+
+        **Parameters:**
+
+        .. list-table::
+            :header-rows: 1
+            :widths: 20 20 60
+
+            * - Parameter
+                - Type
+                - Description
+            * - stack_name
+                - str
+                - Name of the CloudFormation stack to delete
+            * - region
+                - str, optional
+                - AWS region where the stack exists
+            * - retain_resources
+                - List[str], optional
+                - List of logical resource IDs to retain during deletion (only works on DELETE_FAILED stacks)
+            * - logger
+                - logging.Logger, optional
+                - Logger instance for output messages. Uses default logger if not provided
+
+        **Raises:**
+
+        ValueError: When stack doesn't exist or retain_resources limitation is encountered
+        RuntimeError: When CloudFormation deletion fails
+        Exception: For other deletion errors
+
+        .. dropdown:: Usage Examples
+            :open:
+
+            .. code-block:: python
+
+                >>> # Delete a stack (automatically proceeds without confirmation)
+                >>> HpClusterStack.delete("my-stack-name")
+                >>>
+                >>> # Delete in specific region
+                >>> HpClusterStack.delete("my-stack-name", region="us-west-2")
+                >>>
+                >>> # Delete with retained resources (only works on DELETE_FAILED stacks)
+                >>> HpClusterStack.delete("my-stack-name", retain_resources=["S3Bucket", "EFSFileSystem"])
+                >>>
+                >>> # Delete with custom logger
+                >>> import logging
+                >>> logger = logging.getLogger(__name__)
+                >>> HpClusterStack.delete("my-stack-name", logger=logger)
+        """
+        from sagemaker.hyperpod.cli.cluster_stack_utils import (
+            delete_stack_with_confirmation, 
+            StackNotFoundError
+        )
+        
+        if logger is None:
+            logger = logging.getLogger(__name__)
+        
+        # Convert retain_resources list to comma-separated string for the utility function
+        retain_resources_str = ",".join(retain_resources) if retain_resources else ""
+
+        def sdk_confirm_callback(message: str) -> bool:
+            """SDK-specific confirmation callback - always auto-confirms."""
+            logger.info(f"Auto-confirming: {message}")
+            return True
+        
+        try:
+            delete_stack_with_confirmation(
+                stack_name=stack_name,
+                region=region or boto3.session.Session().region_name,
+                retain_resources_str=retain_resources_str,
+                message_callback=logger.info,
+                confirm_callback=sdk_confirm_callback,
+                success_callback=logger.info
+            )
+        except StackNotFoundError:
+            error_msg = f"Stack '{stack_name}' not found"
+            logger.error(error_msg)
+            raise ValueError(error_msg)
+        except Exception as e:
+            error_str = str(e)
+            
+            # Handle CloudFormation retain-resources limitation with clear exception for SDK
+            if retain_resources and "specify which resources to retain only when the stack is in the DELETE_FAILED state" in error_str:
+                error_msg = (
+                    f"CloudFormation limitation: retain_resources can only be used on stacks in DELETE_FAILED state. "
+                    f"Current stack state allows normal deletion. Try deleting without retain_resources first, "
+                    f"then retry with retain_resources if deletion fails."
+                )
+                logger.error(error_msg)
+                raise ValueError(error_msg)
+            
+            # Handle termination protection
+            if "TerminationProtection is enabled" in error_str:
+                error_msg = (
+                    f"Stack deletion blocked: Termination Protection is enabled. "
+                    f"Disable termination protection first using AWS CLI or Console."
+                )
+                logger.error(error_msg)
+                raise RuntimeError(error_msg)
+            
+            # Handle other errors
+            logger.error(f"Failed to delete stack: {error_str}")
+            raise RuntimeError(f"Stack deletion failed: {error_str}")
 
     @staticmethod
     def delete(stack_name: str, region: Optional[str] = None, retain_resources: Optional[List[str]] = None, 
@@ -622,8 +732,8 @@ def sdk_confirm_callback(message: str) -> bool:
             raise RuntimeError(f"Stack deletion failed: {error_str}")
 
 
-def _yaml_to_json_string(yaml_path) -> str:
-    """Convert YAML file to JSON string"""
-    with open(yaml_path, 'r') as file:
-        yaml_data = yaml.safe_load(file)
-    return json.dumps(yaml_data, indent=2, ensure_ascii=False)
+    def _yaml_to_json_string(yaml_path) -> str:
+        """Convert YAML file to JSON string"""
+        with open(yaml_path, 'r') as file:
+            yaml_data = yaml.safe_load(file)
+        return json.dumps(yaml_data, indent=2, ensure_ascii=False)