wip: 修改错误

docs/problem.md

@@ -6,14 +6,16 @@
 - [ ] /pretor/worker_individual待完善复合子个体和基础子个体
 
 #### 🛡️ 安全与合规 (Security & Auth)
-
+- [ ] 优化安全架构防止模型注入
+- [ ] 设计workflowEngine的自动扩缩容设计
+- [ ] 完善错误捕获和日志系统
 
 #### ⚡ 性能与资源优化 (Performance & Scalability)
 - [ ] 增加对应全workflow的情况追踪，使得在任务运行中人机交互更加自然方便
 - [ ] 优化import
 
 #### 🏗️ 架构演进 (Architecture & Refactoring)
-- ~~[ ] 使用fastapi-users完善用户系统~~(2026/4/19 fastapi-users会严重摧毁代码的优雅性)
+- 】~~使用fastapi-users完善用户系统~~(2026/4/19 fastapi-users会严重摧毁代码的优雅性)
 - [ ] 升级auth功能
 - [x] /pretor/api的接口函数进行重构
 - [ ] /dockerfile待完善
@@ -41,3 +43,8 @@
 - [ ] 对接更多的provider
 - [ ] 优化import
 - [ ] 升级auth功能
+
+#### 2026/4/20
+- [ ] 优化安全架构防止模型注入
+- [ ] 设计workflowEngine的自动扩缩容设计
+- [ ] 完善错误捕获和日志系统

pretor/api/platform/event.py

@@ -21,7 +21,7 @@ import asyncio
 
 class PretorEvent(BaseModel):
     model_config = ConfigDict(arbitrary_types_allowed=True)
-    event_id: str = Field(default_factory=lambda: str(ULID()), description="事件的唯一标识符")
+    trace_id: str = Field(default_factory=lambda: str(ULID()), description="事件的唯一标识符")
     platform: str = Field(description="消息来源的平台")
     user_id: str = Field(description="用户id")
     user_name: str = Field(description="用户名")

pretor/api/platform/frontend.py

@@ -36,8 +36,6 @@ async def create_message(message: Message,
     supervisory_node = ray_actor_hook("supervisor_node")
     message = await supervisory_node.working.remote(event)
     if message == "任务已创建":
-        global_state_machine = ray_actor_hook("global_state_machine")
-        global_state_machine.add_event.remote(event)
         return {"message": event.event_id}
     elif message == "未知相应类型":
         raise HTTPException(

pretor/core/global_state_machine/global_state_machine.py

@@ -18,6 +18,7 @@ from pretor.core.global_state_machine.provider_manager import ProviderManager
 from pretor.core.global_state_machine.tool_manager import GlobalToolManager
 from pretor.core.global_state_machine.model_provider import Provider, ProviderArgs
 import httpx
+import pathlib
 import json
 from loguru import logger
 from typing import Dict, Literal, List
@@ -46,6 +47,7 @@ class GlobalStateMachine:
     async def init_state_machine(self):
         await self.global_provider_manager.init_provider_register(self.postgres_database)
 
+
     ###以下方法为event_dict方法
     def add_event(self, event: PretorEvent) -> None:
         event.pending_queue = asyncio.Queue()
@@ -167,6 +169,7 @@ class GlobalStateMachine:
     def get_workflow_template_list(self) -> List[Dict[str, str]]:
         return self.global_workflow_template_manager.workflow_templates_registry
 
+
     ###以下为skill_manager方法
     def add_skill(self, skill_name: str):
         skill_plugin_dir = pathlib.Path(__file__).parent.parent.parent / "plugin" / "skill_plugin" / skill_name

pretor/core/workflow/workflow.py

@@ -23,7 +23,7 @@ NodeType = Literal[
 
 class EventInfo(BaseModel):
     platform: str
-    username: str
+    user_name: str
 
 class LogicGate(BaseModel):
     if_fail: str = Field(..., description="失败跳转目标，如 'jump_to_step_1'")

pretor/core/workflow/workflow_runner.py

@@ -15,7 +15,7 @@
 from pretor.utils.ray_hook import ray_actor_hook
 import ray
 import asyncio
-from pretor.core.workflow.workflow import PretorWorkflow, WorkStep
+from pretor.core.workflow.workflow import PretorWorkflow, WorkStep, EventInfo
 from loguru import logger
 from typing import Optional, Dict, Union, Any, List
 from pretor.utils.error import WorkflowError, WorkflowExit
@@ -31,6 +31,14 @@ from pretor.core.individual.consciousness_node.template import (
     ForWorkflowEngine
 )
 from pretor.core.individual.supervisory_node.template import TerminationMessage
+import pathlib
+
+
+def get_workflow_template(workflow_name: str) -> str:
+    workflow_template = pathlib.Path(__file__).parent.parent.parent / "workflow_template" / (workflow_name + "_workflow_template.json")
+    with open(workflow_template, "r", encoding="utf-8") as workflow_template_file:
+        workflow_template = workflow_template_file.read()
+        return workflow_template
 
 
 class WorkflowEngine:
@@ -237,6 +245,7 @@ class WorkflowRunningEngine:
         self.consciousness_node = consciousness_node
         self.control_node = control_node
         self.supervisory_node = supervisory_node
+        self.global_state_machine = ray_actor_hook("global_state_machine")
 
     async def run(self):
         self.runner_engine = {
@@ -264,6 +273,7 @@ class WorkflowRunningEngine:
                     raise WorkflowError("未配置 consciousness_node，无法生成工作流")
 
                 workflow_template = event.context.get("workflow_template", "")
+                workflow_template = get_workflow_template(workflow_template)
 
                 payload = ForWorkflowEngineInput(
                     original_command=event.message,
@@ -274,12 +284,16 @@ class WorkflowRunningEngine:
 
                 if isinstance(result_obj, ForWorkflowEngine):
                     workflow = result_obj.workflow
+
+                    workflow.trace_id = event.event_id
+                    workflow.command = event.message
+                    workflow.event_info = EventInfo(platform=event.platform,
+                                                    user_name=event.user_name,)
+
                     logger.info(
                         f"WorkflowRunningEngine: runner_{i} 成功生成工作流 {workflow.trace_id}:{workflow.title}")
-                    workflow.event_info = event
 
-                    global_state_machine = ray_actor_hook("global_state_machine")
+                    await self.global_state_machine.update_workflow.remote(event.event_id, workflow)
-                    await global_state_machine.update_workflow.remote(event.event_id, workflow)
 
                     workflow_engine = WorkflowEngine(workflow,
                                                      self.consciousness_node,

pretor/core/workflow/workflow_template_manager.py

@@ -42,3 +42,4 @@ class WorkflowManager:
             self.workflow_templates_registry[workflow_template.name] = workflow_template.desc
         except Exception as e:
             logger.exception("Failed to generate workflow template")
+

tests/core/global_state_machine/provider_manager_test.py

@@ -1,32 +1,25 @@
 import pytest
-from unittest.mock import MagicMock
+from unittest.mock import MagicMock, AsyncMock
 from pretor.core.global_state_machine.provider_manager import ProviderManager
 
 
-def test_provider_manager_init():
+@pytest.mark.asyncio
+async def test_provider_manager_init():
+    from pretor.core.global_state_machine.provider_manager import ProviderManager
     mock_postgres = MagicMock()
-
     mock_provider1 = MagicMock()
-    mock_provider1.title = "title1"
+    mock_provider1.provider_title = "title1"
 
     mock_provider2 = MagicMock()
-    mock_provider2.title = "title2"
+    mock_provider2.provider_title = "title2"
 
-    # In _load_provider_register, it calls `postgres.provider_database.get_provider.remote()`
+    mock_postgres.get_providers.remote = AsyncMock(return_value=[mock_provider1, mock_provider2])
-    # which returns a list of providers synchronously?
-    # Yes, it assumes `.remote()` returns an iterable in this context. Wait!
-    # `.remote()` in Ray actually returns an ObjectRef which is NOT iterable directly,
-    # it must be `ray.get()`.
-    # But let's mock it to return a list anyway because the code does `for provider in providers:`.
-
-    mock_postgres.provider_database.get_provider.remote.return_value = [mock_provider1, mock_provider2]
 
     manager = ProviderManager(mock_postgres)
+    await manager.init_provider_register(mock_postgres)
 
     assert "openai" in manager.provider_mapper
     assert "gemini" in manager.provider_mapper
     assert "claude" in manager.provider_mapper
 
     assert manager.provider_register["title1"] == mock_provider1
-    assert manager.provider_register["title2"] == mock_provider2
-    mock_postgres.provider_database.get_provider.remote.assert_called_once()

tests/core/workflow/workflow_runner_test.py

@@ -115,13 +115,39 @@ async def test_workflow_running_engine_submit():
 
 @pytest.mark.asyncio
 async def test_workflow_running_engine_runner():
-    engine = WorkflowRunningEngine("conscious", "control", "supervisor")
+    from pretor.api.platform.event import PretorEvent
-    engine.workflow_queue = asyncio.Queue()
+    from pretor.core.individual.consciousness_node.template import ForWorkflowEngine
+
+    mock_consciousness = MagicMock()
 
     mock_wf = MagicMock()
-    await engine.workflow_queue.put(mock_wf)
+    mock_wf.trace_id = "test_trace"
+    mock_wf.title = "test_title"
+
+    mock_result = MagicMock(spec=ForWorkflowEngine)
+    mock_result.workflow = mock_wf
+
+    mock_consciousness.working.remote = AsyncMock(return_value=mock_result)
+
+    engine = WorkflowRunningEngine(mock_consciousness, "control", "supervisor")
+    engine.workflow_queue = asyncio.Queue()
+
+    # Use real PretorEvent to avoid Pydantic validation errors on MagicMock properties
+    mock_event = PretorEvent(
+        platform="test_platform",
+        user_id="test_user",
+        user_name="test_user",
+        message="test_message",
+        context={"workflow_template": "test_template"}
+    )
+    await engine.workflow_queue.put(mock_event)
+
+    with patch("pretor.core.workflow.workflow_runner.WorkflowEngine") as mock_wf_engine_cls, \
+            patch("pretor.core.workflow.workflow_runner.ray_actor_hook") as mock_hook:
+        mock_gsm = MagicMock()
+        mock_gsm.update_workflow.remote = AsyncMock()
+        mock_hook.return_value = mock_gsm
 
-    with patch("pretor.core.workflow.workflow_runner.WorkflowEngine") as mock_wf_engine_cls:
         mock_engine_instance = MagicMock()
         mock_engine_instance.run = AsyncMock()
         mock_wf_engine_cls.return_value = mock_engine_instance
@@ -130,5 +156,4 @@ async def test_workflow_running_engine_runner():
         await asyncio.sleep(0.05)  # Give runner time to process one item
         task.cancel()  # Stop the infinite loop
 
-        mock_wf_engine_cls.assert_called_with(mock_wf, "conscious", "control", "supervisor")
+        mock_wf_engine_cls.assert_called_with(mock_wf, mock_consciousness, "control", "supervisor")
-        mock_engine_instance.run.assert_called_once()