vnpy/vn.archive/vn.strategy/strategydemo/demoStrategy.py

310 lines
11 KiB
Python
Raw Normal View History

2016-07-02 03:12:44 +00:00
# encoding: UTF-8
# 首先写系统内置模块
import sys
from datetime import datetime, timedelta, time
from time import sleep
# 然后是第三方库模块如PyQt4等
import sip
from PyQt4 import QtCore
# 然后是自己编写的模块
from demoEngine import MainEngine
from strategyEngine import *
########################################################################
class SimpleEmaStrategy(StrategyTemplate):
"""简单双指数移动均线EMA策略"""
#----------------------------------------------------------------------
def __init__(self, name, symbol, engine):
"""Constructor"""
super(SimpleEmaStrategy, self).__init__(name, symbol, engine)
# 策略在外部设置的参数
self.fastAlpha = 0.2 # 快速EMA的参数
self.slowAlpha = 0.05 # 慢速EMA的参数
# 最新TICK数据市场报价
self.currentTick = None
# K线缓存对象
self.barOpen = 0
self.barHigh = 0
self.barLow = 0
self.barClose = 0
self.barVolume = 0
self.barTime = None
# 保存K线数据的列表对象
self.listOpen = []
self.listHigh = []
self.listLow = []
self.listClose = []
self.listVolume = []
self.listTime = []
# 持仓
self.pos = 0
# 报单代码列表
self.listOrderRef = [] # 报单号列表
self.listStopOrder = [] # 停止单对象列表
# EMA均线
self.fastEMA = 0 # 快速EMA的数值
self.slowEMA = 0 # 慢速EMA的数值
# 是否完成了初始化
self.initCompleted = False
# 初始化时读取的历史数据的起始日期(可以选择外部设置)
self.startDate = None
#----------------------------------------------------------------------
def loadSetting(self, setting):
"""读取参数设定"""
try:
self.fastAlpha = setting['fastAlpha']
self.slowAlpha = setting['slowAlpha']
self.engine.writeLog(self.name + u'读取参数成功')
except KeyError:
self.engine.writeLog(self.name + u'读取参数设定出错,请检查参数字典')
try:
self.initStrategy(setting['startDate'])
except KeyError:
self.initStrategy()
#----------------------------------------------------------------------
def initStrategy(self, startDate=None):
"""初始化"""
td = timedelta(days=3) # 读取3天的历史TICK数据
if startDate:
cx = self.engine.loadTick(self.symbol, startDate-td)
else:
today = datetime.today().replace(hour=0, minute=0, second=0, microsecond=0)
cx = self.engine.loadTick(self.symbol, today-td)
if cx:
for data in cx:
tick = Tick(data['InstrumentID'])
tick.openPrice = data['OpenPrice']
tick.highPrice = data['HighestPrice']
tick.lowPrice = data['LowestPrice']
tick.lastPrice = data['LastPrice']
tick.volume = data['Volume']
tick.openInterest = data['OpenInterest']
tick.upperLimit = data['UpperLimitPrice']
tick.lowerLimit = data['LowerLimitPrice']
tick.time = data['UpdateTime']
tick.ms = data['UpdateMillisec']
tick.bidPrice1 = data['BidPrice1']
tick.bidPrice2 = data['BidPrice2']
tick.bidPrice3 = data['BidPrice3']
tick.bidPrice4 = data['BidPrice4']
tick.bidPrice5 = data['BidPrice5']
tick.askPrice1 = data['AskPrice1']
tick.askPrice2 = data['AskPrice2']
tick.askPrice3 = data['AskPrice3']
tick.askPrice4 = data['AskPrice4']
tick.askPrice5 = data['AskPrice5']
tick.bidVolume1 = data['BidVolume1']
tick.bidVolume2 = data['BidVolume2']
tick.bidVolume3 = data['BidVolume3']
tick.bidVolume4 = data['BidVolume4']
tick.bidVolume5 = data['BidVolume5']
tick.askVolume1 = data['AskVolume1']
tick.askVolume2 = data['AskVolume2']
tick.askVolume3 = data['AskVolume3']
tick.askVolume4 = data['AskVolume4']
tick.askVolume5 = data['AskVolume5']
self.onTick(tick)
self.initCompleted = True
self.engine.writeLog(self.name + u'初始化完成')
#----------------------------------------------------------------------
def onTick(self, tick):
"""行情更新"""
# 保存最新的TICK
self.currentTick = tick
# 首先生成datetime.time格式的时间便于比较
ticktime = self.strToTime(tick.time, tick.ms)
# 假设是收到的第一个TICK
if self.barOpen == 0:
# 初始化新的K线数据
self.barOpen = tick.lastPrice
self.barHigh = tick.lastPrice
self.barLow = tick.lastPrice
self.barClose = tick.lastPrice
self.barVolume = tick.volume
self.barTime = ticktime
else:
# 如果是当前一分钟内的数据
if ticktime.minute == self.barTime.minute:
# 汇总TICK生成K线
self.barHigh = max(self.barHigh, tick.lastPrice)
self.barLow = min(self.barLow, tick.lastPrice)
self.barClose = tick.lastPrice
self.barVolume = self.barVolume + tick.volume
self.barTime = ticktime
# 如果是新一分钟的数据
else:
# 首先推送K线数据
self.onBar(self.barOpen, self.barHigh, self.barLow, self.barClose,
self.barVolume, self.barTime)
# 初始化新的K线数据
self.barOpen = tick.lastPrice
self.barHigh = tick.lastPrice
self.barLow = tick.lastPrice
self.barClose = tick.lastPrice
self.barVolume = tick.volume
self.barTime = ticktime
#----------------------------------------------------------------------
def onTrade(self, trade):
"""交易更新"""
if trade.direction == DIRECTION_BUY:
self.pos = self.pos + trade.volume
else:
self.pos = self.pos - trade.volume
log = self.name + u'当前持仓:' + str(self.pos)
self.engine.writeLog(log)
#----------------------------------------------------------------------
def onOrder(self, order):
"""报单更新"""
pass
#----------------------------------------------------------------------
def onStopOrder(self, orderRef):
"""停止单更新"""
pass
#----------------------------------------------------------------------
def onBar(self, o, h, l, c, volume, time):
"""K线数据更新"""
# 保存K线序列数据
self.listOpen.append(o)
self.listHigh.append(h)
self.listLow.append(l)
self.listClose.append(c)
self.listVolume.append(volume)
self.listTime.append(time)
# 计算EMA
if self.fastEMA:
self.fastEMA = c*self.fastAlpha + self.fastEMA*(1-self.fastAlpha)
self.slowEMA = c*self.slowAlpha + self.slowEMA*(1-self.slowAlpha)
else:
self.fastEMA = c
self.slowEMA = c
# 交易逻辑
if self.initCompleted: # 首先检查是否是实盘运行还是数据预处理阶段
# 快速EMA在慢速EMA上方做多
if self.fastEMA > self.slowEMA:
# 如果当前手头无仓位,则直接做多
if self.pos == 0:
# 涨停价买入开仓
self.buy(self.currentTick.upperLimit, 1)
# 手头有空仓,则先平空,再开多
elif self.pos < 0:
self.cover(self.currentTick.upperLimit, 1)
self.buy(self.currentTick.upperLimit, 1)
# 反之,做空
elif self.fastEMA < self.slowEMA:
if self.pos == 0:
self.short(self.currentTick.lowerLimit, 1)
elif self.pos > 0:
self.sell(self.currentTick.lowerLimit, 1)
self.short(self.currentTick.lowerLimit, 1)
# 记录日志
log = self.name + u'K线时间' + str(time) + '\n' + \
u'快速EMA' + str(self.fastEMA) + u'慢速EMA' + \
str(self.slowEMA)
self.engine.writeLog(log)
#----------------------------------------------------------------------
def strToTime(self, t, ms):
"""从字符串时间转化为time格式的时间"""
hh, mm, ss = t.split(':')
tt = time(int(hh), int(mm), int(ss), microsecond=ms)
return tt
#----------------------------------------------------------------------
def print_log(event):
"""打印日志"""
log = event.dict_['log']
print str(datetime.now()), ':', log
#----------------------------------------------------------------------
def main():
"""运行在CMD中的演示程度"""
# 创建PyQt4应用对象
app = QtCore.QCoreApplication(sys.argv)
# 创建主引擎对象
me = MainEngine()
# 注册事件监听
me.ee.register(EVENT_LOG, print_log)
# 登录
userid = ''
password = ''
brokerid = ''
mdAddress = ''
tdAddress = ''
me.login(userid, password, brokerid, mdAddress, tdAddress)
# 等待10秒钟初始化合约数据等
sleep(5)
# 创建策略引擎对象
se = StrategyEngine(me.ee, me)
# 创建策略对象
setting = {}
setting['fastAlpha'] = 0.2
setting['slowAlpha'] = 0.05
se.createStrategy(u'EMA演示策略', 'IF1506', SimpleEmaStrategy, setting)
# 启动所有策略
se.startAll()
# 让程序连续运行
sys.exit(app.exec_())
if __name__ == '__main__':
main()