Merge branch 'dev' into prcix9320

This commit is contained in:
zhangshixiang
2026-01-12 14:31:56 +08:00
parent 31e8d065c4
commit de7fbe7ac8
18 changed files with 436 additions and 1239 deletions

View File

@@ -19,6 +19,11 @@ if unilabos_dir not in sys.path:
from unilabos.utils.banner_print import print_status, print_unilab_banner
from unilabos.config.config import load_config, BasicConfig, HTTPConfig
from unilabos.app.utils import cleanup_for_restart
# Global restart flags (used by ws_client and web/server)
_restart_requested: bool = False
_restart_reason: str = ""
def load_config_from_file(config_path):
@@ -156,6 +161,11 @@ def parse_args():
default=False,
help="Complete registry information",
)
parser.add_argument(
"--no_update_feedback",
action="store_true",
help="Disable sending update feedback to server",
)
# workflow upload subcommand
workflow_parser = subparsers.add_parser(
"workflow_upload",
@@ -297,6 +307,7 @@ def main():
BasicConfig.is_host_mode = not args_dict.get("is_slave", False)
BasicConfig.slave_no_host = args_dict.get("slave_no_host", False)
BasicConfig.upload_registry = args_dict.get("upload_registry", False)
BasicConfig.no_update_feedback = args_dict.get("no_update_feedback", False)
BasicConfig.communication_protocol = "websocket"
machine_name = os.popen("hostname").read().strip()
machine_name = "".join([c if c.isalnum() or c == "_" else "_" for c in machine_name])
@@ -497,13 +508,19 @@ def main():
time.sleep(1)
else:
start_backend(**args_dict)
start_server(
restart_requested = start_server(
open_browser=not args_dict["disable_browser"],
port=BasicConfig.port,
)
if restart_requested:
print_status("[Main] Restart requested, cleaning up...", "info")
cleanup_for_restart()
return
else:
start_backend(**args_dict)
start_server(
# 启动服务器默认支持WebSocket触发重启
restart_requested = start_server(
open_browser=not args_dict["disable_browser"],
port=BasicConfig.port,
)

144
unilabos/app/utils.py Normal file
View File

@@ -0,0 +1,144 @@
"""
UniLabOS 应用工具函数
提供清理、重启等工具函数
"""
import gc
import os
import threading
import time
from unilabos.utils.banner_print import print_status
def cleanup_for_restart() -> bool:
"""
Clean up all resources for restart without exiting the process.
This function prepares the system for re-initialization by:
1. Stopping all communication clients
2. Destroying ROS nodes
3. Resetting singletons
4. Waiting for threads to finish
Returns:
bool: True if cleanup was successful, False otherwise
"""
print_status("[Restart] Starting cleanup for restart...", "info")
# Step 1: Stop WebSocket communication client
print_status("[Restart] Step 1: Stopping WebSocket client...", "info")
try:
from unilabos.app.communication import get_communication_client
comm_client = get_communication_client()
if comm_client is not None:
comm_client.stop()
print_status("[Restart] WebSocket client stopped", "info")
except Exception as e:
print_status(f"[Restart] Error stopping WebSocket: {e}", "warning")
# Step 2: Get HostNode and cleanup ROS
print_status("[Restart] Step 2: Cleaning up ROS nodes...", "info")
try:
from unilabos.ros.nodes.presets.host_node import HostNode
import rclpy
from rclpy.timer import Timer
host_instance = HostNode.get_instance(timeout=5)
if host_instance is not None:
print_status(f"[Restart] Found HostNode: {host_instance.device_id}", "info")
# Gracefully shutdown background threads
print_status("[Restart] Shutting down background threads...", "info")
HostNode.shutdown_background_threads(timeout=5.0)
print_status("[Restart] Background threads shutdown complete", "info")
# Stop discovery timer
if hasattr(host_instance, "_discovery_timer") and isinstance(host_instance._discovery_timer, Timer):
host_instance._discovery_timer.cancel()
print_status("[Restart] Discovery timer cancelled", "info")
# Destroy device nodes
device_count = len(host_instance.devices_instances)
print_status(f"[Restart] Destroying {device_count} device instances...", "info")
for device_id, device_node in list(host_instance.devices_instances.items()):
try:
if hasattr(device_node, "ros_node_instance") and device_node.ros_node_instance is not None:
device_node.ros_node_instance.destroy_node()
print_status(f"[Restart] Device {device_id} destroyed", "info")
except Exception as e:
print_status(f"[Restart] Error destroying device {device_id}: {e}", "warning")
# Clear devices instances
host_instance.devices_instances.clear()
host_instance.devices_names.clear()
# Destroy host node
try:
host_instance.destroy_node()
print_status("[Restart] HostNode destroyed", "info")
except Exception as e:
print_status(f"[Restart] Error destroying HostNode: {e}", "warning")
# Reset HostNode state
HostNode.reset_state()
print_status("[Restart] HostNode state reset", "info")
# Shutdown executor first (to stop executor.spin() gracefully)
if hasattr(rclpy, "__executor") and rclpy.__executor is not None:
try:
rclpy.__executor.shutdown()
rclpy.__executor = None # Clear for restart
print_status("[Restart] ROS executor shutdown complete", "info")
except Exception as e:
print_status(f"[Restart] Error shutting down executor: {e}", "warning")
# Shutdown rclpy
if rclpy.ok():
rclpy.shutdown()
print_status("[Restart] rclpy shutdown complete", "info")
except ImportError as e:
print_status(f"[Restart] ROS modules not available: {e}", "warning")
except Exception as e:
print_status(f"[Restart] Error in ROS cleanup: {e}", "warning")
return False
# Step 3: Reset communication client singleton
print_status("[Restart] Step 3: Resetting singletons...", "info")
try:
from unilabos.app import communication
if hasattr(communication, "_communication_client"):
communication._communication_client = None
print_status("[Restart] Communication client singleton reset", "info")
except Exception as e:
print_status(f"[Restart] Error resetting communication singleton: {e}", "warning")
# Step 4: Wait for threads to finish
print_status("[Restart] Step 4: Waiting for threads to finish...", "info")
time.sleep(3) # Give threads time to finish
# Check remaining threads
remaining_threads = []
for t in threading.enumerate():
if t.name != "MainThread" and t.is_alive():
remaining_threads.append(t.name)
if remaining_threads:
print_status(
f"[Restart] Warning: {len(remaining_threads)} threads still running: {remaining_threads}", "warning"
)
else:
print_status("[Restart] All threads stopped", "info")
# Step 5: Force garbage collection
print_status("[Restart] Step 5: Running garbage collection...", "info")
gc.collect()
gc.collect() # Run twice for weak references
print_status("[Restart] Garbage collection complete", "info")
print_status("[Restart] Cleanup complete. Ready for re-initialization.", "info")
return True

View File

@@ -6,7 +6,6 @@ Web服务器模块
import webbrowser
import uvicorn
from fastapi import FastAPI, Request
from fastapi.middleware.cors import CORSMiddleware
from starlette.responses import Response
@@ -96,7 +95,7 @@ def setup_server() -> FastAPI:
return app
def start_server(host: str = "0.0.0.0", port: int = 8002, open_browser: bool = True) -> None:
def start_server(host: str = "0.0.0.0", port: int = 8002, open_browser: bool = True) -> bool:
"""
启动服务器
@@ -104,7 +103,14 @@ def start_server(host: str = "0.0.0.0", port: int = 8002, open_browser: bool = T
host: 服务器主机
port: 服务器端口
open_browser: 是否自动打开浏览器
Returns:
bool: True if restart was requested, False otherwise
"""
import threading
import time
from uvicorn import Config, Server
# 设置服务器
setup_server()
@@ -123,7 +129,37 @@ def start_server(host: str = "0.0.0.0", port: int = 8002, open_browser: bool = T
# 启动服务器
info(f"[Web] 启动FastAPI服务器: {host}:{port}")
uvicorn.run(app, host=host, port=port, log_config=log_config)
# 使用支持重启的模式
config = Config(app=app, host=host, port=port, log_config=log_config)
server = Server(config)
# 启动服务器线程
server_thread = threading.Thread(target=server.run, daemon=True, name="uvicorn_server")
server_thread.start()
info("[Web] Server started, monitoring for restart requests...")
# 监控重启标志
import unilabos.app.main as main_module
while server_thread.is_alive():
if hasattr(main_module, "_restart_requested") and main_module._restart_requested:
info(
f"[Web] Restart requested via WebSocket, reason: {getattr(main_module, '_restart_reason', 'unknown')}"
)
main_module._restart_requested = False
# 停止服务器
server.should_exit = True
server_thread.join(timeout=5)
info("[Web] Server stopped, ready for restart")
return True
time.sleep(1)
return False
# 当脚本直接运行时启动服务器

View File

@@ -359,7 +359,7 @@ class MessageProcessor:
self.device_manager = device_manager
self.queue_processor = None # 延迟设置
self.websocket_client = None # 延迟设置
self.session_id = ""
self.session_id = str(uuid.uuid4())[:6] # 产生一个随机的session_id
# WebSocket连接
self.websocket = None
@@ -488,7 +488,16 @@ class MessageProcessor:
async for message in self.websocket:
try:
data = json.loads(message)
await self._process_message(data)
message_type = data.get("action", "")
message_data = data.get("data")
if self.session_id and self.session_id == data.get("edge_session"):
await self._process_message(message_type, message_data)
else:
if message_type.endswith("_material"):
logger.trace(f"[MessageProcessor] 收到一条归属 {data.get('edge_session')} 的旧消息:{data}")
logger.debug(f"[MessageProcessor] 跳过了一条归属 {data.get('edge_session')} 的旧消息: {data.get('action')}")
else:
await self._process_message(message_type, message_data)
except json.JSONDecodeError:
logger.error(f"[MessageProcessor] Invalid JSON received: {message}")
except Exception as e:
@@ -554,11 +563,8 @@ class MessageProcessor:
finally:
logger.debug("[MessageProcessor] Send handler stopped")
async def _process_message(self, data: Dict[str, Any]):
async def _process_message(self, message_type: str, message_data: Dict[str, Any]):
"""处理收到的消息"""
message_type = data.get("action", "")
message_data = data.get("data")
logger.debug(f"[MessageProcessor] Processing message: {message_type}")
try:
@@ -571,16 +577,19 @@ class MessageProcessor:
elif message_type == "cancel_action" or message_type == "cancel_task":
await self._handle_cancel_action(message_data)
elif message_type == "add_material":
# noinspection PyTypeChecker
await self._handle_resource_tree_update(message_data, "add")
elif message_type == "update_material":
# noinspection PyTypeChecker
await self._handle_resource_tree_update(message_data, "update")
elif message_type == "remove_material":
# noinspection PyTypeChecker
await self._handle_resource_tree_update(message_data, "remove")
elif message_type == "session_id":
self.session_id = message_data.get("session_id")
logger.info(f"[MessageProcessor] Session ID: {self.session_id}")
elif message_type == "request_reload":
await self._handle_request_reload(message_data)
# elif message_type == "session_id":
# self.session_id = message_data.get("session_id")
# logger.info(f"[MessageProcessor] Session ID: {self.session_id}")
elif message_type == "request_restart":
await self._handle_request_restart(message_data)
else:
logger.debug(f"[MessageProcessor] Unknown message type: {message_type}")
@@ -890,19 +899,48 @@ class MessageProcessor:
)
thread.start()
async def _handle_request_reload(self, data: Dict[str, Any]):
async def _handle_request_restart(self, data: Dict[str, Any]):
"""
处理重请求
处理重请求
当LabGo发送request_reload时重新发送设备注册信息
当LabGo发送request_restart时执行清理并触发重启
"""
reason = data.get("reason", "unknown")
logger.info(f"[MessageProcessor] Received reload request, reason: {reason}")
delay = data.get("delay", 2) # 默认延迟2秒
logger.info(f"[MessageProcessor] Received restart request, reason: {reason}, delay: {delay}s")
# 重新发送host_node_ready信
# 发送确认消
if self.websocket_client:
self.websocket_client.publish_host_ready()
logger.info("[MessageProcessor] Re-sent host_node_ready after reload request")
await self.websocket_client.send_message({
"action": "restart_acknowledged",
"data": {"reason": reason, "delay": delay}
})
# 设置全局重启标志
import unilabos.app.main as main_module
main_module._restart_requested = True
main_module._restart_reason = reason
# 延迟后执行清理
await asyncio.sleep(delay)
# 在新线程中执行清理,避免阻塞当前事件循环
def do_cleanup():
import time
time.sleep(0.5) # 给当前消息处理完成的时间
logger.info(f"[MessageProcessor] Starting cleanup for restart, reason: {reason}")
try:
from unilabos.app.utils import cleanup_for_restart
if cleanup_for_restart():
logger.info("[MessageProcessor] Cleanup successful, main() will restart")
else:
logger.error("[MessageProcessor] Cleanup failed")
except Exception as e:
logger.error(f"[MessageProcessor] Error during cleanup: {e}")
cleanup_thread = threading.Thread(target=do_cleanup, name="RestartCleanupThread", daemon=True)
cleanup_thread.start()
logger.info(f"[MessageProcessor] Restart cleanup scheduled")
async def _send_action_state_response(
self, device_id: str, action_name: str, task_id: str, job_id: str, typ: str, free: bool, need_more: int