完整实现 Swarm 多智能体协作系统
- 新增 CLIPluginAdapter 统一接口 (backend/app/core/agent_adapter.py) - 新增 LLM 服务层，支持 Anthropic/OpenAI/DeepSeek/Ollama (backend/app/services/llm_service.py) - 新增 Agent 执行引擎，支持文件锁自动管理 (backend/app/services/agent_executor.py) - 新增 NativeLLMAgent 原生 LLM 适配器 (backend/app/adapters/native_llm_agent.py) - 新增进程管理器 (backend/app/services/process_manager.py) - 新增 Agent 控制 API (backend/app/routers/agents_control.py) - 新增 WebSocket 实时通信 (backend/app/routers/websocket.py) - 更新前端 AgentsPage，支持启动/停止 Agent - 测试通过：Agent 启动、批量操作、栅栏同步 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Claude Code
473
backend/app/services/workflow_engine.py
Normal file
473
backend/app/services/workflow_engine.py
Normal file
@@ -0,0 +1,473 @@
|
||||
"""
|
||||
工作流引擎 - 管理和执行工作流
|
||||
支持从 YAML 文件加载工作流定义，并跟踪进度
|
||||
"""
|
||||
|
||||
import asyncio
|
||||
from pathlib import Path
|
||||
from typing import Dict, List, Optional
|
||||
from dataclasses import dataclass, field, asdict
|
||||
from datetime import datetime
|
||||
import yaml
|
||||
|
||||
from .storage import get_storage
|
||||
from .meeting_recorder import get_meeting_recorder
|
||||
|
||||
|
||||
@dataclass
|
||||
class WorkflowMeeting:
|
||||
"""工作流中的节点"""
|
||||
meeting_id: str
|
||||
title: str
|
||||
attendees: List[str]
|
||||
depends_on: List[str] = field(default_factory=list)
|
||||
completed: bool = False
|
||||
node_type: str = "meeting" # meeting | execution
|
||||
min_required: int = None # 最少完成人数（execution 节点用）
|
||||
on_failure: str = None # 失败时跳转的节点 ID
|
||||
|
||||
# 执行状态追踪（execution 节点专用）
|
||||
completed_attendees: List[str] = field(default_factory=list)
|
||||
|
||||
@property
|
||||
def is_ready(self) -> bool:
|
||||
"""检查节点是否可以开始（依赖已完成）"""
|
||||
return all(dep in self.depends_on for dep in self.depends_on)
|
||||
|
||||
@property
|
||||
def is_execution_ready(self) -> bool:
|
||||
"""检查执行节点是否所有人都完成了"""
|
||||
if self.node_type != "execution":
|
||||
return False
|
||||
required = self.min_required or len(self.attendees)
|
||||
return len(self.completed_attendees) >= required
|
||||
|
||||
@property
|
||||
def missing_attendees(self) -> List[str]:
|
||||
"""获取未完成的人员列表"""
|
||||
if self.node_type != "execution":
|
||||
return []
|
||||
return [a for a in self.attendees if a not in self.completed_attendees]
|
||||
|
||||
@property
|
||||
def progress(self) -> str:
|
||||
"""执行进度"""
|
||||
if self.node_type != "execution":
|
||||
return "N/A"
|
||||
return f"{len(self.completed_attendees)}/{len(self.attendees)}"
|
||||
|
||||
|
||||
@dataclass
|
||||
class Workflow:
|
||||
"""工作流定义"""
|
||||
workflow_id: str
|
||||
name: str
|
||||
description: str
|
||||
meetings: List[WorkflowMeeting]
|
||||
created_at: str = ""
|
||||
status: str = "pending" # pending, in_progress, completed
|
||||
|
||||
def __post_init__(self):
|
||||
if not self.created_at:
|
||||
self.created_at = datetime.now().isoformat()
|
||||
|
||||
@property
|
||||
def progress(self) -> str:
|
||||
"""进度摘要"""
|
||||
total = len(self.meetings)
|
||||
completed = sum(1 for m in self.meetings if m.completed)
|
||||
return f"{completed}/{total}"
|
||||
|
||||
@property
|
||||
def current_meeting(self) -> Optional[WorkflowMeeting]:
|
||||
"""获取当前应该进行的会议（第一个未完成的）"""
|
||||
for meeting in self.meetings:
|
||||
if not meeting.completed:
|
||||
return meeting
|
||||
return None
|
||||
|
||||
@property
|
||||
def is_completed(self) -> bool:
|
||||
"""工作流是否完成"""
|
||||
return all(m.completed for m in self.meetings)
|
||||
|
||||
|
||||
class WorkflowEngine:
|
||||
"""
|
||||
工作流引擎
|
||||
|
||||
管理工作流的加载、执行和进度跟踪
|
||||
"""
|
||||
|
||||
WORKFLOWS_DIR = "workflow"
|
||||
|
||||
def __init__(self):
|
||||
self._storage = get_storage()
|
||||
self._recorder = get_meeting_recorder()
|
||||
self._loaded_workflows: Dict[str, Workflow] = {}
|
||||
# 注册的工作流文件路径
|
||||
self._workflow_files: Dict[str, str] = {}
|
||||
|
||||
async def load_workflow(self, workflow_path: str) -> Workflow:
|
||||
"""
|
||||
从 YAML 文件加载工作流
|
||||
|
||||
Args:
|
||||
workflow_path: YAML 文件路径（相对于 .doc/workflow/）
|
||||
|
||||
Returns:
|
||||
加载的工作流
|
||||
"""
|
||||
import aiofiles
|
||||
|
||||
# 构建完整路径
|
||||
full_path = f"{self.WORKFLOWS_DIR}/{workflow_path}"
|
||||
yaml_path = Path(self._storage.base_path) / full_path
|
||||
|
||||
if not yaml_path.exists():
|
||||
raise ValueError(f"Workflow file not found: {workflow_path}")
|
||||
|
||||
# 读取 YAML 内容
|
||||
async with aiofiles.open(yaml_path, mode="r", encoding="utf-8") as f:
|
||||
yaml_content = await f.read()
|
||||
content = yaml.safe_load(yaml_content)
|
||||
|
||||
# 解析工作流
|
||||
meetings = []
|
||||
for m in content.get("meetings", []):
|
||||
meetings.append(WorkflowMeeting(
|
||||
meeting_id=m["meeting_id"],
|
||||
title=m["title"],
|
||||
attendees=m["attendees"],
|
||||
depends_on=m.get("depends_on", []),
|
||||
node_type=m.get("node_type", "meeting"),
|
||||
min_required=m.get("min_required"),
|
||||
on_failure=m.get("on_failure")
|
||||
))
|
||||
|
||||
workflow = Workflow(
|
||||
workflow_id=content["workflow_id"],
|
||||
name=content["name"],
|
||||
description=content.get("description", ""),
|
||||
meetings=meetings
|
||||
)
|
||||
|
||||
self._loaded_workflows[workflow.workflow_id] = workflow
|
||||
# 保存源文件路径，以便后续可以重新加载
|
||||
self._workflow_files[workflow.workflow_id] = workflow_path
|
||||
return workflow
|
||||
|
||||
async def get_next_meeting(self, workflow_id: str) -> Optional[WorkflowMeeting]:
|
||||
"""
|
||||
获取工作流中下一个应该进行的会议
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
|
||||
Returns:
|
||||
下一个会议，如果没有或已完成返回 None
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return None
|
||||
|
||||
return workflow.current_meeting
|
||||
|
||||
async def complete_meeting(self, workflow_id: str, meeting_id: str) -> bool:
|
||||
"""
|
||||
标记会议为已完成
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
meeting_id: 会议 ID
|
||||
|
||||
Returns:
|
||||
是否成功标记
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return False
|
||||
|
||||
for meeting in workflow.meetings:
|
||||
if meeting.meeting_id == meeting_id:
|
||||
meeting.completed = True
|
||||
|
||||
# 更新工作流状态
|
||||
if workflow.is_completed:
|
||||
workflow.status = "completed"
|
||||
else:
|
||||
workflow.status = "in_progress"
|
||||
|
||||
return True
|
||||
|
||||
return False
|
||||
|
||||
async def create_workflow_meeting(
|
||||
self,
|
||||
workflow_id: str,
|
||||
meeting_id: str
|
||||
) -> bool:
|
||||
"""
|
||||
创建工作流中的会议记录
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
meeting_id: 会议 ID
|
||||
|
||||
Returns:
|
||||
是否成功创建
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return False
|
||||
|
||||
meeting = None
|
||||
for m in workflow.meetings:
|
||||
if m.meeting_id == meeting_id:
|
||||
meeting = m
|
||||
break
|
||||
|
||||
if not meeting:
|
||||
return False
|
||||
|
||||
# 创建会议记录
|
||||
await self._recorder.create_meeting(
|
||||
meeting_id=meeting.meeting_id,
|
||||
title=f"{workflow.name}: {meeting.title}",
|
||||
attendees=meeting.attendees,
|
||||
steps=["收集初步想法", "讨论与迭代", "生成共识版本"]
|
||||
)
|
||||
|
||||
return True
|
||||
|
||||
async def get_workflow_status(self, workflow_id: str) -> Optional[Dict]:
|
||||
"""
|
||||
获取工作流状态
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
|
||||
Returns:
|
||||
工作流状态信息
|
||||
"""
|
||||
# 如果不在内存中，尝试重新加载
|
||||
if workflow_id not in self._loaded_workflows and workflow_id in self._workflow_files:
|
||||
await self.load_workflow(self._workflow_files[workflow_id])
|
||||
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return None
|
||||
|
||||
return {
|
||||
"workflow_id": workflow.workflow_id,
|
||||
"name": workflow.name,
|
||||
"description": workflow.description,
|
||||
"status": workflow.status,
|
||||
"progress": workflow.progress,
|
||||
"meetings": [
|
||||
{
|
||||
"meeting_id": m.meeting_id,
|
||||
"title": m.title,
|
||||
"completed": m.completed
|
||||
}
|
||||
for m in workflow.meetings
|
||||
]
|
||||
}
|
||||
|
||||
async def list_workflows(self) -> List[Dict]:
|
||||
"""
|
||||
列出所有加载的工作流
|
||||
|
||||
Returns:
|
||||
工作流列表
|
||||
"""
|
||||
return [
|
||||
{
|
||||
"workflow_id": w.workflow_id,
|
||||
"name": w.name,
|
||||
"status": w.status,
|
||||
"progress": w.progress
|
||||
}
|
||||
for w in self._loaded_workflows.values()
|
||||
]
|
||||
|
||||
# ========== 执行节点相关方法 ==========
|
||||
|
||||
async def join_execution_node(
|
||||
self,
|
||||
workflow_id: str,
|
||||
meeting_id: str,
|
||||
agent_id: str
|
||||
) -> Dict:
|
||||
"""
|
||||
Agent 加入执行节点（标记完成）
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
meeting_id: 执行节点 ID
|
||||
agent_id: Agent ID
|
||||
|
||||
Returns:
|
||||
状态信息 {"status": "waiting"|"ready"|"completed", "progress": "2/3"}
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return {"status": "error", "message": "Workflow not found"}
|
||||
|
||||
for meeting in workflow.meetings:
|
||||
if meeting.meeting_id == meeting_id:
|
||||
if meeting.node_type != "execution":
|
||||
return {"status": "error", "message": "Not an execution node"}
|
||||
|
||||
if agent_id not in meeting.completed_attendees:
|
||||
meeting.completed_attendees.append(agent_id)
|
||||
|
||||
if meeting.is_execution_ready:
|
||||
return {
|
||||
"status": "ready",
|
||||
"progress": meeting.progress,
|
||||
"message": "所有 Agent 已完成，可以进入下一节点"
|
||||
}
|
||||
return {
|
||||
"status": "waiting",
|
||||
"progress": meeting.progress,
|
||||
"missing": meeting.missing_attendees,
|
||||
"message": f"等待其他 Agent 完成: {meeting.missing_attendees}"
|
||||
}
|
||||
|
||||
return {"status": "error", "message": "Meeting not found"}
|
||||
|
||||
async def get_execution_status(
|
||||
self,
|
||||
workflow_id: str,
|
||||
meeting_id: str
|
||||
) -> Optional[Dict]:
|
||||
"""
|
||||
获取执行节点的状态
|
||||
|
||||
Returns:
|
||||
执行状态信息
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return None
|
||||
|
||||
for meeting in workflow.meetings:
|
||||
if meeting.meeting_id == meeting_id:
|
||||
return {
|
||||
"meeting_id": meeting.meeting_id,
|
||||
"title": meeting.title,
|
||||
"node_type": meeting.node_type,
|
||||
"attendees": meeting.attendees,
|
||||
"completed_attendees": meeting.completed_attendees,
|
||||
"progress": meeting.progress,
|
||||
"is_ready": meeting.is_execution_ready,
|
||||
"missing": meeting.missing_attendees
|
||||
}
|
||||
return None
|
||||
|
||||
# ========== 条件跳转相关方法 ==========
|
||||
|
||||
async def jump_to_node(
|
||||
self,
|
||||
workflow_id: str,
|
||||
target_meeting_id: str
|
||||
) -> bool:
|
||||
"""
|
||||
强制跳转到指定节点（重置后续所有节点）
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
target_meeting_id: 目标节点 ID
|
||||
|
||||
Returns:
|
||||
是否成功跳转
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return False
|
||||
|
||||
# 找到目标节点并重置从它开始的所有后续节点
|
||||
target_found = False
|
||||
for meeting in workflow.meetings:
|
||||
if meeting.meeting_id == target_meeting_id:
|
||||
target_found = True
|
||||
meeting.completed = False
|
||||
meeting.completed_attendees = []
|
||||
elif target_found:
|
||||
# 目标节点之后的所有节点都重置
|
||||
meeting.completed = False
|
||||
meeting.completed_attendees = []
|
||||
|
||||
workflow.status = "in_progress"
|
||||
return target_found
|
||||
|
||||
async def handle_failure(
|
||||
self,
|
||||
workflow_id: str,
|
||||
meeting_id: str
|
||||
) -> Optional[str]:
|
||||
"""
|
||||
处理节点失败，根据 on_failure 配置跳转
|
||||
|
||||
Args:
|
||||
workflow_id: 工作流 ID
|
||||
meeting_id: 失败的节点 ID
|
||||
|
||||
Returns:
|
||||
跳转目标节点 ID，如果没有配置则返回 None
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return None
|
||||
|
||||
for meeting in workflow.meetings:
|
||||
if meeting.meeting_id == meeting_id and meeting.on_failure:
|
||||
await self.jump_to_node(workflow_id, meeting.on_failure)
|
||||
return meeting.on_failure
|
||||
return None
|
||||
|
||||
async def get_workflow_detail(self, workflow_id: str) -> Optional[Dict]:
|
||||
"""
|
||||
获取工作流详细信息（包含所有节点状态）
|
||||
|
||||
Returns:
|
||||
工作流详细信息
|
||||
"""
|
||||
workflow = self._loaded_workflows.get(workflow_id)
|
||||
if not workflow:
|
||||
return None
|
||||
|
||||
return {
|
||||
"workflow_id": workflow.workflow_id,
|
||||
"name": workflow.name,
|
||||
"description": workflow.description,
|
||||
"status": workflow.status,
|
||||
"progress": workflow.progress,
|
||||
"current_node": workflow.current_meeting.meeting_id if workflow.current_meeting else None,
|
||||
"meetings": [
|
||||
{
|
||||
"meeting_id": m.meeting_id,
|
||||
"title": m.title,
|
||||
"node_type": m.node_type,
|
||||
"attendees": m.attendees,
|
||||
"depends_on": m.depends_on,
|
||||
"completed": m.completed,
|
||||
"on_failure": m.on_failure,
|
||||
"progress": m.progress if m.node_type == "execution" else None
|
||||
}
|
||||
for m in workflow.meetings
|
||||
]
|
||||
}
|
||||
|
||||
|
||||
# 全局单例
|
||||
_engine_instance: Optional[WorkflowEngine] = None
|
||||
|
||||
|
||||
def get_workflow_engine() -> WorkflowEngine:
|
||||
"""获取工作流引擎单例"""
|
||||
global _engine_instance
|
||||
if _engine_instance is None:
|
||||
_engine_instance = WorkflowEngine()
|
||||
return _engine_instance
|
||||
Reference in New Issue
Block a user