实盘运行

在实盘环境,用户可以基于编写好的CTA策略来创建新的实例,一键初始化、启动、停止策略。

创建策略实例

用户可以基于编写好的CTA策略来创建新的实例,策略实例的好处在于同一个策略可以同时去运行多个品种合约,并且每个实例的参数可以是不同的。在创建实例的时候需要填写如图的实例名称、合约品种、参数设置。注意:实例名称不能重名;合约名称是vt_symbol的格式,如IF1905.CFFEX。

https://vnpy-community.oss-cn-shanghai.aliyuncs.com/forum_experience/yazhang/cta_strategy/add_strategy.png

创建策略流程如下:

  • 检查策略实例重名
  • 添加策略配置信息(strategy_name, vt_symbol, setting)到strategies字典上
  • 添加该策略要订阅行情的合约信息到symbol_strategy_map字典中;
  • 把策略配置信息保存到json文件内;
  • 在图形化界面更新状态信息。
  1. def add_strategy(
  2. self, class_name: str, strategy_name: str, vt_symbol: str, setting: dict
  3. ):
  4. """
  5. Add a new strategy.
  6. """
  7. if strategy_name in self.strategies:
  8. self.write_log(f"创建策略失败,存在重名{strategy_name}")
  9. return
  10.  
  11. strategy_class = self.classes[class_name]
  12.  
  13. strategy = strategy_class(self, strategy_name, vt_symbol, setting)
  14. self.strategies[strategy_name] = strategy
  15.  
  16. # Add vt_symbol to strategy map.
  17. strategies = self.symbol_strategy_map[vt_symbol]
  18. strategies.append(strategy)
  19.  
  20. # Update to setting file.
  21. self.update_strategy_setting(strategy_name, setting)
  22.  
  23. self.put_strategy_event(strategy)

初始化策略

  • 调用策略类的on_init()回调函数,并且载入历史数据;
  • 恢复上次退出之前的策略状态;
  • 调用接口的subcribe()函数订阅指定行情信息;
  • 策略初始化状态变成True,并且更新到日志上。
  1. def _init_strategy(self):
  2. """
  3. Init strategies in queue.
  4. """
  5. while not self.init_queue.empty():
  6. strategy_name = self.init_queue.get()
  7. strategy = self.strategies[strategy_name]
  8.  
  9. if strategy.inited:
  10. self.write_log(f"{strategy_name}已经完成初始化,禁止重复操作")
  11. continue
  12.  
  13. self.write_log(f"{strategy_name}开始执行初始化")
  14.  
  15. # Call on_init function of strategy
  16. self.call_strategy_func(strategy, strategy.on_init)
  17.  
  18. # Restore strategy data(variables)
  19. data = self.strategy_data.get(strategy_name, None)
  20. if data:
  21. for name in strategy.variables:
  22. value = data.get(name, None)
  23. if value:
  24. setattr(strategy, name, value)
  25.  
  26. # Subscribe market data
  27. contract = self.main_engine.get_contract(strategy.vt_symbol)
  28. if contract:
  29. req = SubscribeRequest(
  30. symbol=contract.symbol, exchange=contract.exchange)
  31. self.main_engine.subscribe(req, contract.gateway_name)
  32. else:
  33. self.write_log(f"行情订阅失败,找不到合约{strategy.vt_symbol}", strategy)
  34.  
  35. # Put event to update init completed status.
  36. strategy.inited = True
  37. self.put_strategy_event(strategy)
  38. self.write_log(f"{strategy_name}初始化完成")
  39.  
  40. self.init_thread = None

启动策略

  • 检查策略初始化状态;
  • 检查策略启动状态,避免重复启动;
  • 调用策略类的on_start()函数启动策略;
  • 策略启动状态变成True,并且更新到图形化界面上。
  1. def start_strategy(self, strategy_name: str):
  2. """
  3. Start a strategy.
  4. """
  5. strategy = self.strategies[strategy_name]
  6. if not strategy.inited:
  7. self.write_log(f"策略{strategy.strategy_name}启动失败,请先初始化")
  8. return
  9.  
  10. if strategy.trading:
  11. self.write_log(f"{strategy_name}已经启动,请勿重复操作")
  12. return
  13.  
  14. self.call_strategy_func(strategy, strategy.on_start)
  15. strategy.trading = True
  16.  
  17. self.put_strategy_event(strategy)

停止策略

  • 检查策略启动状态;
  • 调用策略类的on_stop()函数停止策略;
  • 更新策略启动状态为False;
  • 对所有为成交的委托(市价单/限价单/本地停止单)进行撤单操作;
  • 在图形化界面更新策略状态。
  1. def stop_strategy(self, strategy_name: str):
  2. """
  3. Stop a strategy.
  4. """
  5. strategy = self.strategies[strategy_name]
  6. if not strategy.trading:
  7. return
  8.  
  9. # Call on_stop function of the strategy
  10. self.call_strategy_func(strategy, strategy.on_stop)
  11.  
  12. # Change trading status of strategy to False
  13. strategy.trading = False
  14.  
  15. # Cancel all orders of the strategy
  16. self.cancel_all(strategy)
  17.  
  18. # Update GUI
  19. self.put_strategy_event(strategy)

编辑策略

  • 重新配置策略参数字典setting;
  • 更新参数字典到策略中;
  • 在图像化界面更新策略状态。
  1. def edit_strategy(self, strategy_name: str, setting: dict):
  2. """
  3. Edit parameters of a strategy.
  4. """
  5. strategy = self.strategies[strategy_name]
  6. strategy.update_setting(setting)
  7.  
  8. self.update_strategy_setting(strategy_name, setting)
  9. self.put_strategy_event(strategy)

移除策略

  • 检查策略状态,只有停止策略后从可以移除策略;
  • 从json文件移除策略配置信息(strategy_name, vt_symbol, setting);
  • 从symbol_strategy_map字典中移除该策略订阅的合约信息;
  • 从strategy_orderid_map字典移除活动委托记录;
  • 从strategies字典移除该策略的相关配置信息。
  1. def remove_strategy(self, strategy_name: str):
  2. """
  3. Remove a strategy.
  4. """
  5. strategy = self.strategies[strategy_name]
  6. if strategy.trading:
  7. self.write_log(f"策略{strategy.strategy_name}移除失败,请先停止")
  8. return
  9.  
  10. # Remove setting
  11. self.remove_strategy_setting(strategy_name)
  12.  
  13. # Remove from symbol strategy map
  14. strategies = self.symbol_strategy_map[strategy.vt_symbol]
  15. strategies.remove(strategy)
  16.  
  17. # Remove from active orderid map
  18. if strategy_name in self.strategy_orderid_map:
  19. vt_orderids = self.strategy_orderid_map.pop(strategy_name)
  20.  
  21. # Remove vt_orderid strategy map
  22. for vt_orderid in vt_orderids:
  23. if vt_orderid in self.orderid_strategy_map:
  24. self.orderid_strategy_map.pop(vt_orderid)
  25.  
  26. # Remove from strategies
  27. self.strategies.pop(strategy_name)
  28.  
  29. return True