commit
This commit is contained in:
@@ -18,6 +18,14 @@ database:
|
||||
max_connections: 10
|
||||
min_connections: 2
|
||||
|
||||
# ========== Redis配置 ==========
|
||||
redis:
|
||||
host: 127.0.0.1
|
||||
port: 6379
|
||||
password: ""
|
||||
db: 0
|
||||
pool_size: 10
|
||||
|
||||
# ========== 浏览器池配置 ==========
|
||||
browser_pool:
|
||||
idle_timeout: 1800 # 空闲超时(秒),已禁用自动清理,保持常驻
|
||||
|
||||
@@ -18,6 +18,14 @@ database:
|
||||
max_connections: 20
|
||||
min_connections: 5
|
||||
|
||||
# ========== Redis配置 ==========
|
||||
redis:
|
||||
host: 8.140.194.184
|
||||
port: 6379
|
||||
password: "Redis@123456"
|
||||
db: 0
|
||||
pool_size: 10
|
||||
|
||||
# ========== 浏览器池配置 ==========
|
||||
browser_pool:
|
||||
idle_timeout: 1800 # 空闲超时(秒),已禁用自动清理,保持常驻
|
||||
|
||||
@@ -95,6 +95,16 @@ def load_config(env: str = None) -> Config:
|
||||
if os.getenv('DB_NAME'):
|
||||
config_dict.setdefault('database', {})['dbname'] = os.getenv('DB_NAME')
|
||||
|
||||
# Redis配置
|
||||
if os.getenv('REDIS_HOST'):
|
||||
config_dict.setdefault('redis', {})['host'] = os.getenv('REDIS_HOST')
|
||||
if os.getenv('REDIS_PORT'):
|
||||
config_dict.setdefault('redis', {})['port'] = int(os.getenv('REDIS_PORT'))
|
||||
if os.getenv('REDIS_PASSWORD'):
|
||||
config_dict.setdefault('redis', {})['password'] = os.getenv('REDIS_PASSWORD')
|
||||
if os.getenv('REDIS_DB'):
|
||||
config_dict.setdefault('redis', {})['db'] = int(os.getenv('REDIS_DB'))
|
||||
|
||||
# 调度器配置
|
||||
if os.getenv('SCHEDULER_ENABLED'):
|
||||
config_dict.setdefault('scheduler', {})['enabled'] = os.getenv('SCHEDULER_ENABLED').lower() == 'true'
|
||||
@@ -122,6 +132,7 @@ def load_config(env: str = None) -> Config:
|
||||
print(f"[配置] 已加载配置文件: {config_file}")
|
||||
print(f"[配置] 环境: {env}")
|
||||
print(f"[配置] 数据库: {config_dict.get('database', {}).get('host')}:{config_dict.get('database', {}).get('port')}")
|
||||
print(f"[配置] Redis: {config_dict.get('redis', {}).get('host')}:{config_dict.get('redis', {}).get('port')}")
|
||||
print(f"[配置] 调度器: {'启用' if config_dict.get('scheduler', {}).get('enabled') else '禁用'}")
|
||||
|
||||
return Config(config_dict)
|
||||
|
||||
@@ -50,13 +50,13 @@ PROXY_POOL = [
|
||||
"server": "http://111.132.40.72:50002",
|
||||
"username": "ih3z07",
|
||||
"password": "078bt7o5",
|
||||
"enabled": False
|
||||
"enabled": True
|
||||
}, {
|
||||
"name": "天启03",
|
||||
"server": "http://210.51.27.194:50001",
|
||||
"username": "hb6su3",
|
||||
"password": "acv2ciow",
|
||||
"enabled": True
|
||||
"enabled": False
|
||||
}
|
||||
]
|
||||
|
||||
|
||||
216
backend/main.py
216
backend/main.py
@@ -62,8 +62,11 @@ class ConnectionManager:
|
||||
async def connect(self, session_id: str, websocket: WebSocket):
|
||||
await websocket.accept()
|
||||
self.active_connections[session_id] = websocket
|
||||
print(f"[WebSocket] 新连接: {session_id}", file=sys.stderr)
|
||||
print(f"[WebSocket] ========== 新连接建立 ==========", file=sys.stderr)
|
||||
print(f"[WebSocket] Session ID: {session_id}", file=sys.stderr)
|
||||
print(f"[WebSocket] 当前活跃连接数: {len(self.active_connections)}", file=sys.stderr)
|
||||
print(f"[WebSocket] 连接时间: {__import__('datetime').datetime.now()}", file=sys.stderr)
|
||||
print(f"[WebSocket] ===================================", file=sys.stderr)
|
||||
|
||||
# 立即检查缓存消息(不等待)
|
||||
if session_id in self.pending_messages:
|
||||
@@ -91,22 +94,42 @@ class ConnectionManager:
|
||||
else:
|
||||
print(f"[WebSocket] 没有缓存消息: {session_id}", file=sys.stderr)
|
||||
|
||||
def disconnect(self, session_id: str):
|
||||
def disconnect(self, session_id: str, reason: str = "未知原因"):
|
||||
"""断开WebSocket连接并记录原因"""
|
||||
if session_id in self.active_connections:
|
||||
del self.active_connections[session_id]
|
||||
print(f"[WebSocket] 断开连接: {session_id}", file=sys.stderr)
|
||||
print(f"[WebSocket] ========== 连接断开 ==========", file=sys.stderr)
|
||||
print(f"[WebSocket] Session ID: {session_id}", file=sys.stderr)
|
||||
print(f"[WebSocket] 断开原因: {reason}", file=sys.stderr)
|
||||
print(f"[WebSocket] 剩余活跃连接数: {len(self.active_connections)}", file=sys.stderr)
|
||||
print(f"[WebSocket] ===================================", file=sys.stderr)
|
||||
# 清理缓存消息
|
||||
if session_id in self.pending_messages:
|
||||
pending_count = len(self.pending_messages[session_id])
|
||||
if pending_count > 0:
|
||||
print(f"[WebSocket] 清理 {pending_count} 条未发送的缓存消息", file=sys.stderr)
|
||||
del self.pending_messages[session_id]
|
||||
|
||||
async def send_message(self, session_id: str, message: dict):
|
||||
print(f"[WebSocket] ========== 尝试发送消息 ==========", file=sys.stderr)
|
||||
print(f"[WebSocket] Session ID: {session_id}", file=sys.stderr)
|
||||
print(f"[WebSocket] 消息类型: {message.get('type')}", file=sys.stderr)
|
||||
print(f"[WebSocket] 当前活跃连接数: {len(self.active_connections)}", file=sys.stderr)
|
||||
print(f"[WebSocket] 活跃连接session_ids: {list(self.active_connections.keys())}", file=sys.stderr)
|
||||
print(f"[WebSocket] session_id在连接中: {session_id in self.active_connections}", file=sys.stderr)
|
||||
print(f"[WebSocket] ===================================", file=sys.stderr)
|
||||
|
||||
if session_id in self.active_connections:
|
||||
try:
|
||||
await self.active_connections[session_id].send_json(message)
|
||||
print(f"[WebSocket] 发送消息到 {session_id}: {message.get('type')}", file=sys.stderr)
|
||||
except Exception as e:
|
||||
print(f"[WebSocket] 发送消息失败 {session_id}: {str(e)}", file=sys.stderr)
|
||||
self.disconnect(session_id)
|
||||
print(f"[WebSocket] ========== 发送消息失败 ==========", file=sys.stderr)
|
||||
print(f"[WebSocket] Session ID: {session_id}", file=sys.stderr)
|
||||
print(f"[WebSocket] 失败原因: {str(e)}", file=sys.stderr)
|
||||
print(f"[WebSocket] 消息类型: {message.get('type')}", file=sys.stderr)
|
||||
print(f"[WebSocket] ===================================", file=sys.stderr)
|
||||
self.disconnect(session_id, reason=f"发送消息失败: {str(e)}")
|
||||
else:
|
||||
# WebSocket还未连接,缓存消息
|
||||
print(f"[WebSocket] 连接尚未建立,缓存消息: {session_id}", file=sys.stderr)
|
||||
@@ -811,6 +834,87 @@ async def save_bind_info(request: dict):
|
||||
data=None
|
||||
)
|
||||
|
||||
@app.post("/api/xhs/save-login")
|
||||
async def save_login(request: dict):
|
||||
"""
|
||||
保存验证码登录的信息到Go后端
|
||||
与扫码登录不同,验证码登录返回的是storage_state数据
|
||||
"""
|
||||
try:
|
||||
employee_id = request.get('employee_id')
|
||||
storage_state = request.get('storage_state', {})
|
||||
storage_state_path = request.get('storage_state_path', '')
|
||||
user_info = request.get('user_info', {}) # 新增: 获取用户信息
|
||||
|
||||
if not employee_id:
|
||||
return BaseResponse(
|
||||
code=1,
|
||||
message="employee_id不能为空",
|
||||
data=None
|
||||
)
|
||||
|
||||
if not storage_state:
|
||||
return BaseResponse(
|
||||
code=1,
|
||||
message="storage_state不能为空",
|
||||
data=None
|
||||
)
|
||||
|
||||
# 调用Go后端API保存
|
||||
config = get_config()
|
||||
go_backend_url = config.get_str('go_backend.url', 'http://localhost:8080')
|
||||
|
||||
# 从 storage_state 中提取 cookies
|
||||
cookies_full = storage_state.get('cookies', [])
|
||||
|
||||
# 构造保存数据
|
||||
save_data = {
|
||||
"employee_id": employee_id,
|
||||
"cookies_full": cookies_full,
|
||||
"storage_state": storage_state,
|
||||
"storage_state_path": storage_state_path,
|
||||
"user_info": user_info # 新增: 传递用户信息
|
||||
}
|
||||
|
||||
print(f"[保存验证码登录] employee_id={employee_id}, cookies数量={len(cookies_full)}, 用户={user_info.get('nickname', '未知')}", file=sys.stderr)
|
||||
|
||||
import aiohttp
|
||||
async with aiohttp.ClientSession() as session:
|
||||
# 获取小程序传来的token
|
||||
auth_header = request.get('Authorization', '')
|
||||
|
||||
async with session.post(
|
||||
f"{go_backend_url}/api/xhs/save-login",
|
||||
json=save_data,
|
||||
headers={'Authorization': auth_header} if auth_header else {}
|
||||
) as resp:
|
||||
result = await resp.json()
|
||||
|
||||
if resp.status == 200 and result.get('code') == 200:
|
||||
print(f"[保存验证码登录] 保存成功", file=sys.stderr)
|
||||
return BaseResponse(
|
||||
code=0,
|
||||
message="保存成功",
|
||||
data=result.get('data')
|
||||
)
|
||||
else:
|
||||
print(f"[保存验证码登录] 保存失败: {result.get('message')}", file=sys.stderr)
|
||||
return BaseResponse(
|
||||
code=1,
|
||||
message=result.get('message', '保存失败'),
|
||||
data=None
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
print(f"[保存验证码登录] 异常: {str(e)}", file=sys.stderr)
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
return BaseResponse(
|
||||
code=1,
|
||||
message=f"保存失败: {str(e)}",
|
||||
data=None
|
||||
)
|
||||
|
||||
@app.post("/api/xhs/qrcode/cancel")
|
||||
async def cancel_qrcode_login(request: dict):
|
||||
"""
|
||||
@@ -1324,6 +1428,7 @@ async def upload_images(files: List[UploadFile] = File(...)):
|
||||
async def handle_send_code_ws(session_id: str, phone: str, country_code: str, login_page: str, websocket: WebSocket):
|
||||
"""
|
||||
异步处理WebSocket发送验证码请求
|
||||
返回: (result, service_instance) - result是结果字典,service_instance是XHSLoginService实例
|
||||
"""
|
||||
try:
|
||||
print(f"[WebSocket-SendCode] 开始处理: session={session_id}, phone={phone}", file=sys.stderr)
|
||||
@@ -1343,6 +1448,13 @@ async def handle_send_code_ws(session_id: str, phone: str, country_code: str, lo
|
||||
session_id=session_id
|
||||
)
|
||||
|
||||
# 将service实例存储到浏览器池,供后续验证码验证使用
|
||||
if session_id in browser_pool.temp_browsers:
|
||||
browser_pool.temp_browsers[session_id]['service'] = request_login_service
|
||||
print(f"[WebSocket-SendCode] 已存储service实例: {session_id}", file=sys.stderr)
|
||||
else:
|
||||
print(f"[WebSocket-SendCode] 警告: session_id {session_id} 不在temp_browsers中", file=sys.stderr)
|
||||
|
||||
# 检查是否需要验证(发送验证码时触发风控)
|
||||
if result.get("need_captcha"):
|
||||
print(f"[WebSocket-SendCode] 检测到风控,需要扫码", file=sys.stderr)
|
||||
@@ -1353,7 +1465,8 @@ async def handle_send_code_ws(session_id: str, phone: str, country_code: str, lo
|
||||
"message": result.get("message", "需要扫码验证")
|
||||
})
|
||||
print(f"[WebSocket-SendCode] 已推送风控信息", file=sys.stderr)
|
||||
return
|
||||
# 返回service实例,供外部启动监听任务
|
||||
return result, request_login_service
|
||||
|
||||
if result["success"]:
|
||||
print(f"[WebSocket-SendCode] 验证码发送成功", file=sys.stderr)
|
||||
@@ -1369,6 +1482,9 @@ async def handle_send_code_ws(session_id: str, phone: str, country_code: str, lo
|
||||
"success": False,
|
||||
"message": result.get("error", "发送验证码失败")
|
||||
})
|
||||
|
||||
return result, request_login_service
|
||||
|
||||
except Exception as e:
|
||||
print(f"[WebSocket-SendCode] 异常: {str(e)}", file=sys.stderr)
|
||||
import traceback
|
||||
@@ -1381,6 +1497,7 @@ async def handle_send_code_ws(session_id: str, phone: str, country_code: str, lo
|
||||
})
|
||||
except:
|
||||
pass
|
||||
return {"success": False, "error": str(e)}, None
|
||||
|
||||
async def handle_verify_code_ws(session_id: str, phone: str, code: str, country_code: str, login_page: str, websocket: WebSocket):
|
||||
"""
|
||||
@@ -1404,7 +1521,7 @@ async def handle_verify_code_ws(session_id: str, phone: str, code: str, country_
|
||||
request_login_service = browser_data['service']
|
||||
|
||||
# 调用登录服务验证登录
|
||||
result = await request_login_service.login_with_code(
|
||||
result = await request_login_service.login(
|
||||
phone=phone,
|
||||
code=code,
|
||||
country_code=country_code,
|
||||
@@ -1483,6 +1600,43 @@ async def websocket_login(websocket: WebSocket, session_id: str):
|
||||
"""
|
||||
await ws_manager.connect(session_id, websocket)
|
||||
|
||||
# 启动Redis订阅任务
|
||||
import asyncio
|
||||
import redis.asyncio as aioredis
|
||||
import json
|
||||
from config import get_config
|
||||
|
||||
config = get_config()
|
||||
redis_host = config.get_str('redis.host', 'localhost')
|
||||
redis_port = config.get_int('redis.port', 6379)
|
||||
redis_password = config.get_str('redis.password', '')
|
||||
|
||||
# 创建Redis订阅客户端
|
||||
redis_url = f"redis://:{redis_password}@{redis_host}:{redis_port}" if redis_password else f"redis://{redis_host}:{redis_port}"
|
||||
redis_client = await aioredis.from_url(redis_url, decode_responses=True)
|
||||
pubsub = redis_client.pubsub()
|
||||
channel = f"ws_message:{session_id}"
|
||||
await pubsub.subscribe(channel)
|
||||
print(f"[WebSocket] 已订阅Redis频道: {channel}", file=sys.stderr)
|
||||
|
||||
# 启动后台任务监听Redis消息
|
||||
async def redis_subscriber():
|
||||
try:
|
||||
async for message in pubsub.listen():
|
||||
if message['type'] == 'message':
|
||||
try:
|
||||
data = json.loads(message['data'])
|
||||
print(f"[WebSocket] 从Redis收到消息: {data}", file=sys.stderr)
|
||||
await websocket.send_json(data)
|
||||
print(f"[WebSocket] 已转发消息到前端: {session_id}", file=sys.stderr)
|
||||
except Exception as e:
|
||||
print(f"[WebSocket] 处理Redis消息失败: {str(e)}", file=sys.stderr)
|
||||
except Exception as e:
|
||||
print(f"[WebSocket] Redis订阅异常: {str(e)}", file=sys.stderr)
|
||||
|
||||
# 在后台启动Redis监听
|
||||
redis_task = asyncio.create_task(redis_subscriber())
|
||||
|
||||
try:
|
||||
# 保持连接,等待消息或断开
|
||||
while True:
|
||||
@@ -1496,7 +1650,6 @@ async def websocket_login(websocket: WebSocket, session_id: str):
|
||||
else:
|
||||
# 尝试解析JSON消息
|
||||
try:
|
||||
import json
|
||||
msg = json.loads(data)
|
||||
msg_type = msg.get('type', 'unknown')
|
||||
print(f"[WebSocket] 解析消息类型: {msg_type}", file=sys.stderr)
|
||||
@@ -1519,8 +1672,15 @@ async def websocket_login(websocket: WebSocket, session_id: str):
|
||||
login_page = msg.get('login_page', 'creator')
|
||||
print(f"[WebSocket] 收到发送验证码请求: phone={phone}", file=sys.stderr)
|
||||
|
||||
# 启动异步任务处理发送验证码
|
||||
asyncio.create_task(handle_send_code_ws(session_id, phone, country_code, login_page, websocket))
|
||||
# 直接处理发送验证码(不使用create_task)
|
||||
result, service_instance = await handle_send_code_ws(session_id, phone, country_code, login_page, websocket)
|
||||
|
||||
# 如果需要扫码,在当前协程中启动监听
|
||||
if result.get("need_captcha") and service_instance:
|
||||
print(f"[WebSocket] 在主协程中启动扫码监听: {session_id}", file=sys.stderr)
|
||||
# 使用create_task在后台监听,但不阻塞当前消息循环
|
||||
asyncio.create_task(service_instance._monitor_qrcode_scan(session_id))
|
||||
print(f"[WebSocket] 已启动扫码监听任务", file=sys.stderr)
|
||||
|
||||
# 处理验证码验证消息
|
||||
elif msg_type == 'verify_code':
|
||||
@@ -1536,12 +1696,38 @@ async def websocket_login(websocket: WebSocket, session_id: str):
|
||||
except json.JSONDecodeError:
|
||||
print(f"[WebSocket] 无法解析为JSON: {data}", file=sys.stderr)
|
||||
|
||||
except WebSocketDisconnect:
|
||||
ws_manager.disconnect(session_id)
|
||||
print(f"[WebSocket] 客户端断开: {session_id}", file=sys.stderr)
|
||||
except WebSocketDisconnect as e:
|
||||
reason = f"客户端主动断开连接 (code: {e.code if hasattr(e, 'code') else 'unknown'})"
|
||||
ws_manager.disconnect(session_id, reason=reason)
|
||||
except Exception as e:
|
||||
ws_manager.disconnect(session_id)
|
||||
print(f"[WebSocket] 连接异常 {session_id}: {str(e)}", file=sys.stderr)
|
||||
reason = f"连接异常: {type(e).__name__} - {str(e)}"
|
||||
ws_manager.disconnect(session_id, reason=reason)
|
||||
finally:
|
||||
# 清理Redis订阅
|
||||
try:
|
||||
redis_task.cancel()
|
||||
await pubsub.unsubscribe(channel)
|
||||
await pubsub.close()
|
||||
await redis_client.close()
|
||||
print(f"[WebSocket] 已取消Redis订阅: {channel}", file=sys.stderr)
|
||||
except:
|
||||
pass
|
||||
|
||||
# 释放浏览器实例
|
||||
try:
|
||||
# 检查是否有临时浏览器需要释放
|
||||
if session_id in browser_pool.temp_browsers:
|
||||
print(f"[WebSocket] 检测到未释放的临时浏览器,开始清理: {session_id}", file=sys.stderr)
|
||||
await browser_pool.release_temp_browser(session_id)
|
||||
print(f"[WebSocket] 已释放临时浏览器: {session_id}", file=sys.stderr)
|
||||
|
||||
# 检查是否有扫码页面需要释放
|
||||
if session_id in browser_pool.qrcode_pages:
|
||||
print(f"[WebSocket] 检测到未释放的扫码页面,开始清理: {session_id}", file=sys.stderr)
|
||||
await browser_pool.release_qrcode_page(session_id)
|
||||
print(f"[WebSocket] 已释放扫码页面: {session_id}", file=sys.stderr)
|
||||
except Exception as e:
|
||||
print(f"[WebSocket] 释放浏览器异常: {str(e)}", file=sys.stderr)
|
||||
|
||||
if __name__ == "__main__":
|
||||
import uvicorn
|
||||
|
||||
@@ -576,13 +576,12 @@ class XHSLoginService:
|
||||
|
||||
# 尝试查找二维码图片元素
|
||||
qrcode_selectors = [
|
||||
'.qrcode-img', # 根据您提供的HTML
|
||||
'.qrcode-img', # 小红书风控二维码的特定class
|
||||
'img.qrcode-img',
|
||||
'.qrcode-container img',
|
||||
'img[src*="data:image"]', # base64图片
|
||||
'img[src*="qrcode"]',
|
||||
'img[alt*="二维码"]',
|
||||
'img[alt*="qrcode"]',
|
||||
'.qrcode-container img', # 二维码容器内的图片
|
||||
'.verify-captcha img', # 验证弹窗内的图片
|
||||
'img[alt*="二维码"]', # alt属性包含"二维码"
|
||||
'img[alt*="qrcode"]', # alt属性包含"qrcode"
|
||||
]
|
||||
|
||||
for selector in qrcode_selectors:
|
||||
@@ -643,27 +642,37 @@ class XHSLoginService:
|
||||
后台监听扫码后的页面跳转和二维码失效
|
||||
通过监听小红书API https://edith.xiaohongshu.com/api/redcaptcha/v2/qr/status/query
|
||||
来精准判断二维码状态:
|
||||
- status=1: 未过期,等待扫码
|
||||
- status=5: 已扫码,等待确认
|
||||
- 其他: 失效或已完成
|
||||
- status=1: 正常,等待扫码
|
||||
- status=2: 扫码完成,待APP确认
|
||||
- status=5: 二维码已过期/失效
|
||||
|
||||
Args:
|
||||
session_id: 会话 ID
|
||||
"""
|
||||
try:
|
||||
logger.info(f"[WebSocket] 开始监听扫码状态: {session_id}")
|
||||
|
||||
# 等待1秒,确保WebSocket连接完全建立
|
||||
logger.info(f"[WebSocket] 等待WebSocket连接建立...")
|
||||
await asyncio.sleep(1.0)
|
||||
logger.info(f"[WebSocket] 等待完成,开始监听")
|
||||
|
||||
if not self.page:
|
||||
logger.error(f"[WebSocket] 页面对象不存在: {session_id}")
|
||||
return
|
||||
|
||||
# 用于存储最新的二维码状态
|
||||
latest_qr_status = {"status": 1, "scanned": False}
|
||||
latest_qr_status = {"status": 1}
|
||||
# 标记是否已推送失效消息
|
||||
expired_notified = False
|
||||
# 标记是否已推送扫码成功消息
|
||||
scan_success_notified = False
|
||||
# 记录上次推送的状态,避免重复推送
|
||||
last_notified_status = None
|
||||
|
||||
# 设置响应监听,拦截二维码状态查询API
|
||||
async def handle_qr_status_response(response):
|
||||
nonlocal last_notified_status
|
||||
try:
|
||||
if '/api/redcaptcha/v2/qr/status/query' in response.url:
|
||||
json_data = await response.json()
|
||||
@@ -671,11 +680,56 @@ class XHSLoginService:
|
||||
status = json_data['data'].get('status')
|
||||
latest_qr_status['status'] = status
|
||||
|
||||
if status == 5:
|
||||
latest_qr_status['scanned'] = True
|
||||
logger.info(f"[WebSocket] 检测到二维码已扫描,等待确认: status={status}")
|
||||
elif status == 1:
|
||||
logger.debug(f"[WebSocket] 二维码未过期,等待扫码: status={status}")
|
||||
# 推送状态变化给前端
|
||||
if status != last_notified_status:
|
||||
status_message = {
|
||||
1: "等待扫码",
|
||||
2: "扫码完成,请在APP中确认",
|
||||
5: "二维码已过期"
|
||||
}.get(status, f"二维码状态: {status}")
|
||||
|
||||
try:
|
||||
# 使用Redis发布消息,避免事件循环隔离问题
|
||||
import redis
|
||||
import json as json_lib
|
||||
from config import get_config
|
||||
|
||||
config = get_config()
|
||||
redis_host = config.get_str('redis.host', 'localhost')
|
||||
redis_port = config.get_int('redis.port', 6379)
|
||||
redis_password = config.get_str('redis.password', '')
|
||||
|
||||
redis_client = redis.Redis(
|
||||
host=redis_host,
|
||||
port=redis_port,
|
||||
password=redis_password if redis_password else None,
|
||||
decode_responses=True
|
||||
)
|
||||
|
||||
message = {
|
||||
"type": "qrcode_status",
|
||||
"status": status,
|
||||
"message": status_message
|
||||
}
|
||||
|
||||
# 发布到Redis频道
|
||||
channel = f"ws_message:{session_id}"
|
||||
redis_client.publish(channel, json_lib.dumps(message))
|
||||
logger.info(f"[WebSocket] 已通过Redis推送二维码状态: status={status}, channel={channel}")
|
||||
last_notified_status = status
|
||||
|
||||
redis_client.close()
|
||||
except Exception as ws_error:
|
||||
logger.error(f"[WebSocket] 推送状态失败: {str(ws_error)}")
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
|
||||
if status == 1:
|
||||
logger.debug(f"[WebSocket] 二维码正常,等待扫码: status={status}")
|
||||
elif status == 2:
|
||||
logger.info(f"[WebSocket] 检测到扫码完成,等待APP确认: status={status}")
|
||||
elif status == 5:
|
||||
logger.warning(f"[WebSocket] 检测到二维码已过期: status={status}")
|
||||
else:
|
||||
logger.info(f"[WebSocket] 二维码状态: status={status}")
|
||||
except Exception as e:
|
||||
@@ -694,88 +748,104 @@ class XHSLoginService:
|
||||
|
||||
# 1. 检测是否跳转回首页(不再是captcha/verify页)
|
||||
if 'captcha' not in current_url.lower() and 'verify' not in current_url.lower():
|
||||
# 如果已经推送过扫码成功消息,跳过
|
||||
if scan_success_notified:
|
||||
continue
|
||||
|
||||
# 检查是否跳转到小红书首页
|
||||
if 'xiaohongshu.com' in current_url:
|
||||
logger.success(f"[WebSocket] 检测到扫码完成,页面跳转回: {current_url}")
|
||||
|
||||
# 等待500ms确保WebSocket连接完全建立
|
||||
await asyncio.sleep(0.5)
|
||||
|
||||
# 通过WebSocket推送扫码成功消息
|
||||
try:
|
||||
from main import ws_manager
|
||||
await ws_manager.send_message(session_id, {
|
||||
# 使用Redis发布消息
|
||||
import redis
|
||||
import json as json_lib
|
||||
from config import get_config
|
||||
|
||||
config = get_config()
|
||||
redis_host = config.get_str('redis.host', 'localhost')
|
||||
redis_port = config.get_int('redis.port', 6379)
|
||||
redis_password = config.get_str('redis.password', '')
|
||||
|
||||
redis_client = redis.Redis(
|
||||
host=redis_host,
|
||||
port=redis_port,
|
||||
password=redis_password if redis_password else None,
|
||||
decode_responses=True
|
||||
)
|
||||
|
||||
message = {
|
||||
"type": "qrcode_scan_success",
|
||||
"message": "扫码验证完成,请重新发送验证码"
|
||||
})
|
||||
logger.success(f"[WebSocket] 已推送扫码成功消息: {session_id}")
|
||||
}
|
||||
|
||||
channel = f"ws_message:{session_id}"
|
||||
redis_client.publish(channel, json_lib.dumps(message))
|
||||
logger.success(f"[WebSocket] 已通过Redis推送扫码成功消息: channel={channel}")
|
||||
scan_success_notified = True
|
||||
|
||||
redis_client.close()
|
||||
except Exception as ws_error:
|
||||
logger.error(f"[WebSocket] 推送消息失败: {str(ws_error)}")
|
||||
|
||||
break
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
|
||||
# 不退出监听,继续等待用户后续操作
|
||||
logger.info(f"[WebSocket] 扫码成功,保持监听状态")
|
||||
|
||||
# 2. 检测二维码是否失效(通过API状态判断)
|
||||
if 'captcha' in current_url.lower() or 'verify' in current_url.lower():
|
||||
# 如果已经推送过失效消息,跳过后续检测
|
||||
# 如果已经推送过失效消息,跳过后续检测
|
||||
if expired_notified:
|
||||
continue
|
||||
|
||||
# 如果状态不是1和5,说明二维码可能已失效
|
||||
if latest_qr_status['status'] not in [1, 5]:
|
||||
logger.warning(f"[WebSocket] 检测到二维码失效: status={latest_qr_status['status']}")
|
||||
# 如果状态是5,说明二维码已过期
|
||||
if latest_qr_status['status'] == 5:
|
||||
logger.warning(f"[WebSocket] API检测到二维码过期: status=5")
|
||||
|
||||
# 等待500ms确保WebSocket连接完全建立
|
||||
await asyncio.sleep(0.5)
|
||||
|
||||
# 通过WebSocket推送失效消息
|
||||
try:
|
||||
from main import ws_manager
|
||||
await ws_manager.send_message(session_id, {
|
||||
# 使用Redis发布消息
|
||||
import redis
|
||||
import json as json_lib
|
||||
from config import get_config
|
||||
|
||||
config = get_config()
|
||||
redis_host = config.get_str('redis.host', 'localhost')
|
||||
redis_port = config.get_int('redis.port', 6379)
|
||||
redis_password = config.get_str('redis.password', '')
|
||||
|
||||
redis_client = redis.Redis(
|
||||
host=redis_host,
|
||||
port=redis_port,
|
||||
password=redis_password if redis_password else None,
|
||||
decode_responses=True
|
||||
)
|
||||
|
||||
message = {
|
||||
"type": "qrcode_expired",
|
||||
"message": "二维码已失效,请重新发送验证码"
|
||||
})
|
||||
logger.success(f"[WebSocket] 已推送二维码失效消息: {session_id}")
|
||||
expired_notified = True # 标记已推送
|
||||
}
|
||||
|
||||
channel = f"ws_message:{session_id}"
|
||||
redis_client.publish(channel, json_lib.dumps(message))
|
||||
logger.success(f"[WebSocket] 已通过Redis推送二维码失效消息: channel={channel}")
|
||||
expired_notified = True
|
||||
|
||||
redis_client.close()
|
||||
except Exception as ws_error:
|
||||
logger.error(f"[WebSocket] 推送消息失败: {str(ws_error)}")
|
||||
|
||||
break # 退出监听循环
|
||||
import traceback
|
||||
traceback.print_exc()
|
||||
|
||||
# 备用方案:检查页面文本(以防API未返回)
|
||||
try:
|
||||
expired_selectors = [
|
||||
'text="已过期"',
|
||||
'text="二维码已失效"',
|
||||
'text="二维码过期"',
|
||||
]
|
||||
|
||||
for selector in expired_selectors:
|
||||
expired_elem = await self.page.query_selector(selector)
|
||||
if expired_elem:
|
||||
is_visible = await expired_elem.is_visible()
|
||||
if is_visible:
|
||||
# 进一步检查元素文本内容
|
||||
text_content = await expired_elem.text_content()
|
||||
# 只在明确显示"已过期"或"已失效"时才认为失效,忽略"二维码X分钟失效"这种提示
|
||||
if text_content and ('已过期' in text_content or '已失效' in text_content):
|
||||
logger.warning(f"[WebSocket] DOM检测到二维码失效: {selector}, 文本: {text_content}")
|
||||
|
||||
# 通过WebSocket推送失效消息
|
||||
try:
|
||||
from main import ws_manager
|
||||
await ws_manager.send_message(session_id, {
|
||||
"type": "qrcode_expired",
|
||||
"message": "二维码已失效,请重新发送验证码"
|
||||
})
|
||||
logger.success(f"[WebSocket] 已推送二维码失效消息: {session_id}")
|
||||
expired_notified = True # 标记已推送
|
||||
except Exception as ws_error:
|
||||
logger.error(f"[WebSocket] 推送消息失败: {str(ws_error)}")
|
||||
|
||||
# 退出所有循环
|
||||
break
|
||||
|
||||
# 如果检测到失效,退出外层循环
|
||||
if expired_notified:
|
||||
break
|
||||
|
||||
except Exception as e:
|
||||
# 页面可能已关闭,忽略错误
|
||||
pass
|
||||
# 不退出监听,继续等待用户重新操作
|
||||
|
||||
# 每30秒打印一次状态
|
||||
if i > 0 and i % 60 == 0:
|
||||
@@ -783,21 +853,11 @@ class XHSLoginService:
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"[WebSocket] 监听异常: {str(e)}")
|
||||
break
|
||||
# 不退出,继续监听
|
||||
|
||||
# 超时5分钟未扫码,通知前端关闭弹窗
|
||||
logger.warning(f"[WebSocket] 扫码监听超时(5分钟): {session_id}")
|
||||
try:
|
||||
from main import ws_manager
|
||||
await ws_manager.send_message(session_id, {
|
||||
"type": "qrcode_expired",
|
||||
"message": "二维码已超时,请重新发送验证码"
|
||||
})
|
||||
logger.success(f"[WebSocket] 已推送超时消息: {session_id}")
|
||||
except Exception as ws_error:
|
||||
logger.error(f"[WebSocket] 推送消息失败: {str(ws_error)}")
|
||||
|
||||
logger.info(f"[WebSocket] 监听任务结束: {session_id}")
|
||||
# 超时5分钟,通知前端(但不退出监听)
|
||||
logger.warning(f"[WebSocket] 监听已运行5分钟: {session_id}")
|
||||
logger.info(f"[WebSocket] 监听仍将继续,直到用户关闭页面")
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"[WebSocket] 监听任务异常: {str(e)}")
|
||||
@@ -850,6 +910,26 @@ class XHSLoginService:
|
||||
logger.warning(f"[页面导航] 导航超时,但尝试继续: {str(e)}")
|
||||
logger.info(f"[页面导航] 当前URL: {current_url}")
|
||||
|
||||
# 检测小红书反爬JSON页面
|
||||
await asyncio.sleep(0.5) # 等待页面内容加载
|
||||
try:
|
||||
page_content = await self.page.content()
|
||||
# 检查页面是否只返回JSON(小红书的检测机制)
|
||||
if page_content and len(page_content) < 500: # JSON页面通常很短
|
||||
# 尝试解析JSON
|
||||
if '{"code"' in page_content and '"success":true' in page_content:
|
||||
logger.warning("="*50)
|
||||
logger.warning("⚠️ 检测到小红书反爬JSON页面")
|
||||
logger.warning(f"页面内容: {page_content[:200]}")
|
||||
logger.warning("="*50)
|
||||
# 抛出异常,让外层处理
|
||||
raise Exception("ANTI_CRAWL_JSON")
|
||||
except Exception as e:
|
||||
if "ANTI_CRAWL_JSON" in str(e):
|
||||
raise # 重新抛出,让外层捕获
|
||||
# 其他异常忽略,继续执行
|
||||
pass
|
||||
|
||||
# 等待二维码API请求(最多等待timeout秒)
|
||||
wait_count = 0
|
||||
max_wait = timeout * 10 # 每次等待0.1秒
|
||||
@@ -919,12 +999,30 @@ class XHSLoginService:
|
||||
else:
|
||||
# 页面变了,重新访问登录页
|
||||
logger.success(f"[预热] 页面已变更 ({current_url}),重新访问{page_name}登录页...")
|
||||
await self._navigate_with_qrcode_listener(login_url)
|
||||
try:
|
||||
await self._navigate_with_qrcode_listener(login_url)
|
||||
except Exception as e:
|
||||
if "ANTI_CRAWL_JSON" in str(e):
|
||||
logger.error("⚠️ 检测到小红书反爬检测,请稍后再试")
|
||||
return {
|
||||
"success": False,
|
||||
"error": "当前IP被小红书检测,请等待5分钟后再试"
|
||||
}
|
||||
raise
|
||||
else:
|
||||
# 未预热或不是池模式,使用监听机制访问页面
|
||||
logger.debug(f"正在访问{page_name}登录页...")
|
||||
|
||||
await self._navigate_with_qrcode_listener(login_url)
|
||||
try:
|
||||
await self._navigate_with_qrcode_listener(login_url)
|
||||
except Exception as e:
|
||||
if "ANTI_CRAWL_JSON" in str(e):
|
||||
logger.error("⚠️ 检测到小红书反爬检测,请稍后再试")
|
||||
return {
|
||||
"success": False,
|
||||
"error": "当前IP被小红书检测,请等待5分钟后再试"
|
||||
}
|
||||
raise
|
||||
|
||||
logger.success(f"✅ 已进入{page_name}登录页面")
|
||||
|
||||
@@ -951,9 +1049,8 @@ class XHSLoginService:
|
||||
logger.info(f"二维码数据长度: {len(qrcode_data)} 字符")
|
||||
logger.info("返回二维码给前端,等待用户扫码后重新调用接口")
|
||||
|
||||
# 启动后台任务监听页面跳转,扫码完成后通知前端
|
||||
asyncio.create_task(self._monitor_qrcode_scan(session_id))
|
||||
logger.info(f"[WebSocket] 已启动扫码监听任务: {session_id}")
|
||||
# 不再在这里启动监听任务,由main.py中的WebSocket端点启动
|
||||
# asyncio.create_task(self._monitor_qrcode_scan(session_id))
|
||||
|
||||
return {
|
||||
"success": False,
|
||||
@@ -1517,14 +1614,14 @@ class XHSLoginService:
|
||||
logger.success(f"✅ 检测到登录成功,用户: {user_me_data.get('nickname')}")
|
||||
|
||||
# 通过WebSocket推送登录成功消息
|
||||
if session_id:
|
||||
if self.session_id:
|
||||
try:
|
||||
from main import ws_manager
|
||||
await ws_manager.send_message(session_id, {
|
||||
await ws_manager.send_message(self.session_id, {
|
||||
"type": "login_success",
|
||||
"user_info": user_me_data
|
||||
})
|
||||
logger.info(f"[WebSocket] 已推送登录成功消息: {session_id}")
|
||||
logger.info(f"[WebSocket] 已推送登录成功消息: {self.session_id}")
|
||||
except Exception as ws_error:
|
||||
logger.error(f"[WebSocket] 推送消息失败: {str(ws_error)}")
|
||||
except Exception as e:
|
||||
@@ -1599,7 +1696,7 @@ class XHSLoginService:
|
||||
"need_captcha": True,
|
||||
"captcha_type": "qrcode",
|
||||
"qrcode_image": qrcode_data,
|
||||
"session_id": session_id, # 返回session_id,供后续轮询使用
|
||||
"session_id": self.session_id, # 返回session_id,供后续轮询使用
|
||||
"message": "需要扫码验证,请使用小红书APP扫描二维码"
|
||||
}
|
||||
else:
|
||||
@@ -1626,40 +1723,114 @@ class XHSLoginService:
|
||||
logger.info(f"最终URL: {self.page.url}")
|
||||
logger.info("="*50)
|
||||
|
||||
# 2. 即使URL没变,也要检测页面上是否出现二维码弹窗
|
||||
logger.info("检测页面上是否出现扫码验证...")
|
||||
qrcode_selectors = [
|
||||
'.qrcode-img',
|
||||
'img.qrcode-img',
|
||||
'.qrcode-container img',
|
||||
'img[src*="data:image"]',
|
||||
'img[src*="qrcode"]',
|
||||
'img[alt*="二维码"]',
|
||||
'img[alt*="qrcode"]',
|
||||
]
|
||||
|
||||
for selector in qrcode_selectors:
|
||||
# 2. 只有在未检测到风控且未登录成功时,才检测页面上是否出现二维码弹窗
|
||||
current_url = self.page.url
|
||||
# 如果已经跳转到成功页面,不再检测二维码
|
||||
if 'explore' in current_url or 'creator' in current_url or 'xiaohongshu.com' in current_url:
|
||||
if 'login' not in current_url:
|
||||
logger.info("已跳转到登录成功页面,跳过二维码检测")
|
||||
else:
|
||||
logger.info("仍在登录页,检测页面上是否出现扫码验证...")
|
||||
# 先检测提示文本
|
||||
try:
|
||||
tip_elem = await self.page.query_selector('.tip')
|
||||
if tip_elem:
|
||||
tip_text = await tip_elem.inner_text()
|
||||
logger.info(f"检测到提示文本: {tip_text}")
|
||||
if '扫码' in tip_text or '二维码' in tip_text:
|
||||
logger.warning("⚠️ 确认检测到扫码验证提示")
|
||||
except Exception as e:
|
||||
logger.debug(f"检测提示文本失败: {str(e)}")
|
||||
|
||||
qrcode_selectors = [
|
||||
'.qrcode-img', # 小红书风控二维码的特定class
|
||||
'img.qrcode-img',
|
||||
'.qrcode-container img', # 二维码容器内的图片
|
||||
'.qrcode .qrcode-img', # 二维码容器下的二维码图片
|
||||
'.verify-captcha img', # 验证弹窗内的图片
|
||||
'.login-container .qrcode-img', # 登录容器内的二维码
|
||||
'img[alt*="二维码"]', # alt属性包含"二维码"
|
||||
'img[alt*="qrcode"]', # alt属性包含"qrcode"
|
||||
]
|
||||
|
||||
for selector in qrcode_selectors:
|
||||
try:
|
||||
qrcode_elem = await self.page.query_selector(selector)
|
||||
if qrcode_elem:
|
||||
logger.info(f"检测到符合选择器的元素: {selector},尝试提取二维码...")
|
||||
qrcode_data = await self.extract_verification_qrcode()
|
||||
if qrcode_data:
|
||||
logger.warning(f"⚠️ 确认检测到风控二维码: {selector}")
|
||||
logger.success("✅ 成功提取扫码验证二维码,返回给前端")
|
||||
# 注意:不移除API监听,保持session_id对应的浏览器继续运行
|
||||
return {
|
||||
"success": False,
|
||||
"need_captcha": True,
|
||||
"captcha_type": "qrcode",
|
||||
"qrcode_image": qrcode_data,
|
||||
"session_id": self.session_id, # 返回session_id,供后续轮询使用
|
||||
"message": "需要扫码验证,请使用小红书APP扫描二维码"
|
||||
}
|
||||
else:
|
||||
logger.debug(f"选择器 {selector} 匹配到元素但无法提取二维码,可能不是风控二维码")
|
||||
break
|
||||
except Exception as e:
|
||||
logger.debug(f"选择器 {selector} 检测失败: {str(e)}")
|
||||
continue
|
||||
|
||||
logger.info("未检测到扫码验证")
|
||||
else:
|
||||
logger.info("仍在登录页,检测页面上是否出现扫码验证...")
|
||||
# 先检测提示文本
|
||||
try:
|
||||
qrcode_elem = await self.page.query_selector(selector)
|
||||
if qrcode_elem:
|
||||
logger.warning(f"⚠️ 检测到页面上出现二维码: {selector}")
|
||||
qrcode_data = await self.extract_verification_qrcode()
|
||||
if qrcode_data:
|
||||
logger.success("✅ 成功提取扫码验证二维码,返回给前端")
|
||||
# 注意:不移除API监听,保持session_id对应的浏览器继续运行
|
||||
return {
|
||||
"success": False,
|
||||
"need_captcha": True,
|
||||
"captcha_type": "qrcode",
|
||||
"qrcode_image": qrcode_data,
|
||||
"session_id": session_id, # 返回session_id,供后续轮询使用
|
||||
"message": "需要扫码验证,请使用小红书APP扫描二维码"
|
||||
}
|
||||
break
|
||||
except Exception:
|
||||
continue
|
||||
tip_elem = await self.page.query_selector('.tip')
|
||||
if tip_elem:
|
||||
tip_text = await tip_elem.inner_text()
|
||||
logger.info(f"检测到提示文本: {tip_text}")
|
||||
if '扫码' in tip_text or '二维码' in tip_text:
|
||||
logger.warning("⚠️ 确认检测到扫码验证提示")
|
||||
except Exception as e:
|
||||
logger.debug(f"检测提示文本失败: {str(e)}")
|
||||
|
||||
qrcode_selectors = [
|
||||
'.qrcode-img', # 小红书风控二维码的特定class
|
||||
'img.qrcode-img',
|
||||
'.qrcode-container img', # 二维码容器内的图片
|
||||
'.qrcode .qrcode-img', # 二维码容器下的二维码图片
|
||||
'.verify-captcha img', # 验证弹窗内的图片
|
||||
'.login-container .qrcode-img', # 登录容器内的二维码
|
||||
'img[alt*="二维码"]', # alt属性包含"二维码"
|
||||
'img[alt*="qrcode"]', # alt属性包含"qrcode"
|
||||
]
|
||||
|
||||
for selector in qrcode_selectors:
|
||||
try:
|
||||
qrcode_elem = await self.page.query_selector(selector)
|
||||
if qrcode_elem:
|
||||
logger.info(f"检测到符合选择器的元素: {selector},尝试提取二维码...")
|
||||
qrcode_data = await self.extract_verification_qrcode()
|
||||
if qrcode_data:
|
||||
logger.warning(f"⚠️ 确认检测到风控二维码: {selector}")
|
||||
logger.success("✅ 成功提取扫码验证二维码,返回给前端")
|
||||
# 注意:不移除API监听,保持session_id对应的浏览器继续运行
|
||||
return {
|
||||
"success": False,
|
||||
"need_captcha": True,
|
||||
"captcha_type": "qrcode",
|
||||
"qrcode_image": qrcode_data,
|
||||
"session_id": self.session_id, # 返回session_id,供后续轮询使用
|
||||
"message": "需要扫码验证,请使用小红书APP扫描二维码"
|
||||
}
|
||||
else:
|
||||
logger.debug(f"选择器 {selector} 匹配到元素但无法提取二维码,可能不是风控二维码")
|
||||
break
|
||||
except Exception as e:
|
||||
logger.debug(f"选择器 {selector} 检测失败: {str(e)}")
|
||||
continue
|
||||
|
||||
logger.info("未检测到扫码验证")
|
||||
|
||||
logger.info("未检测到扫码验证,继续等待登录...")
|
||||
logger.info("继续等待登录...")
|
||||
|
||||
# 等待URL跳转或API响应(最多30秒)
|
||||
logger.info("[登录检测] 等待扫码完成或登录跳转...")
|
||||
@@ -1908,6 +2079,7 @@ class XHSLoginService:
|
||||
"localStorage": localStorage_data, # API 返回:localStorage数据
|
||||
"sessionStorage": sessionStorage_data, # API 返回:sessionStorage数据
|
||||
"url": current_url,
|
||||
"storage_state": storage_state_data, # 新增:Playwright storage_state对象
|
||||
"storage_state_path": storage_state_path # 新增:storage_state文件路径
|
||||
}
|
||||
|
||||
@@ -3162,6 +3334,26 @@ class XHSLoginService:
|
||||
current_url = self.page.url
|
||||
logger.success(f"[扫码登录] 页面加载完成, 当前URL: {current_url}")
|
||||
|
||||
# 检测小红书反爬JSON页面
|
||||
await asyncio.sleep(0.5) # 等待页面内容加载
|
||||
try:
|
||||
page_content = await self.page.content()
|
||||
# 检查页面是否只返回JSON(小红书的检测机制)
|
||||
if page_content and len(page_content) < 500: # JSON页面通常很短
|
||||
# 尝试解析JSON
|
||||
if '{"code"' in page_content and '"success":true' in page_content:
|
||||
logger.warning("="*50)
|
||||
logger.warning("⚠️ 检测到小红书反爬JSON页面")
|
||||
logger.warning(f"页面内容: {page_content[:200]}")
|
||||
logger.warning("="*50)
|
||||
return {
|
||||
"success": False,
|
||||
"error": "当前IP被小红书检测,请等待5分钟后再试"
|
||||
}
|
||||
except Exception as e:
|
||||
# 其他异常忽略,继续执行
|
||||
pass
|
||||
|
||||
# 检查是否跳转到验证码页面
|
||||
if '/website-login/captcha' in current_url or 'verifyUuid=' in current_url:
|
||||
logger.warning(f"[扫码登录] 检测到风控验证页面,尝试等待或跳过...")
|
||||
|
||||
Reference in New Issue
Block a user