feat(api): simplify `execution_metadata` handling for `WorkflowNodeExecution`

Return an empty dictionary instead of `None` in
`execution_metadata_dict` when metadata is absent.

api/core/repositories/sqlalchemy_workflow_node_execution_repository.py

@@ -103,10 +103,7 @@ class SQLAlchemyWorkflowNodeExecutionRepository(WorkflowNodeExecutionRepository)
         inputs = db_model.inputs_dict
         process_data = db_model.process_data_dict
         outputs = db_model.outputs_dict
-        if db_model.execution_metadata_dict:
+        metadata = {NodeRunMetadataKey(k): v for k, v in db_model.execution_metadata_dict.items()}
-            metadata = {NodeRunMetadataKey(k): v for k, v in db_model.execution_metadata_dict.items()}
-        else:
-            metadata = {}
 
         # Convert status to domain enum
         status = NodeExecutionStatus(db_model.status)

api/models/workflow.py

@@ -666,8 +666,11 @@ class WorkflowNodeExecution(Base):
         return json.loads(self.process_data) if self.process_data else None
 
     @property
-    def execution_metadata_dict(self) -> dict[str, Any] | None:
+    def execution_metadata_dict(self) -> dict[str, Any]:
-        return json.loads(self.execution_metadata) if self.execution_metadata else None
+        # When the metadata is unset, we return an empty dictionary instead of `None`.
+        # This approach streamlines the logic for the caller, making it easier to handle
+        # cases where metadata is absent.
+        return json.loads(self.execution_metadata) if self.execution_metadata else {}
 
     @property
     def extras(self):