refactor(api): replace json.loads with Pydantic validation in controllers and infra layers (#34277)

Co-authored-by: autofix-ci[bot] <114827586+autofix-ci[bot]@users.noreply.github.com>

api/extensions/logstore/repositories/logstore_workflow_node_execution_repository.py

@@ -20,6 +20,7 @@ from graphon.workflow_type_encoder import WorkflowRuntimeTypeConverter
 from sqlalchemy.engine import Engine
 from sqlalchemy.orm import sessionmaker
 
+from core.ops.utils import JSON_DICT_ADAPTER
 from core.repositories import SQLAlchemyWorkflowNodeExecutionRepository
 from core.repositories.factory import OrderConfig, WorkflowNodeExecutionRepository
 from extensions.logstore.aliyun_logstore import AliyunLogStore
@@ -48,10 +49,10 @@ def _dict_to_workflow_node_execution(data: dict[str, Any]) -> WorkflowNodeExecut
     """
     logger.debug("_dict_to_workflow_node_execution: data keys=%s", list(data.keys())[:5])
     # Parse JSON fields
-    inputs = json.loads(data.get("inputs", "{}"))
-    process_data = json.loads(data.get("process_data", "{}"))
-    outputs = json.loads(data.get("outputs", "{}"))
-    metadata = json.loads(data.get("execution_metadata", "{}"))
+    inputs = JSON_DICT_ADAPTER.validate_json(data.get("inputs") or "{}")
+    process_data = JSON_DICT_ADAPTER.validate_json(data.get("process_data") or "{}")
+    outputs = JSON_DICT_ADAPTER.validate_json(data.get("outputs") or "{}")
+    metadata = JSON_DICT_ADAPTER.validate_json(data.get("execution_metadata") or "{}")
 
     # Convert metadata to domain enum keys
     domain_metadata = {}

api/extensions/storage/clickzetta_volume/file_lifecycle.py

@@ -15,8 +15,12 @@ from datetime import datetime
 from enum import StrEnum, auto
 from typing import Any
 
+from pydantic import TypeAdapter
+
 logger = logging.getLogger(__name__)
 
+_metadata_adapter: TypeAdapter[dict[str, Any]] = TypeAdapter(dict[str, Any])
+
 
 class FileStatus(StrEnum):
     """File status enumeration"""
@@ -455,8 +459,8 @@ class FileLifecycleManager:
         try:
             if self._storage.exists(self._metadata_file):
                 metadata_content = self._storage.load_once(self._metadata_file)
-                result = json.loads(metadata_content.decode("utf-8"))
-                return dict(result) if result else {}
+                result = _metadata_adapter.validate_json(metadata_content)
+                return result or {}
             else:
                 return {}
         except Exception as e:

api/extensions/storage/google_cloud_storage.py

@@ -1,13 +1,16 @@
 import base64
 import io
-import json
 from collections.abc import Generator
+from typing import Any
 
 from google.cloud import storage as google_cloud_storage  # type: ignore
+from pydantic import TypeAdapter
 
 from configs import dify_config
 from extensions.storage.base_storage import BaseStorage
 
+_service_account_adapter: TypeAdapter[dict[str, Any]] = TypeAdapter(dict[str, Any])
+
 
 class GoogleCloudStorage(BaseStorage):
     """Implementation for Google Cloud storage."""
@@ -21,7 +24,7 @@ class GoogleCloudStorage(BaseStorage):
         if service_account_json_str:
             service_account_json = base64.b64decode(service_account_json_str).decode("utf-8")
             # convert str to object
-            service_account_obj = json.loads(service_account_json)
+            service_account_obj = _service_account_adapter.validate_json(service_account_json)
             self.client = google_cloud_storage.Client.from_service_account_info(service_account_obj)
         else:
             self.client = google_cloud_storage.Client()