QuantDigger是一个基于python的量化回测框架。它借鉴了主流商业软件(比如TB, 金字塔)简洁的策略语法,同时 避免了它们内置编程语言的局限性,使用通用语言python做为策略开发工具。和 zipline , pyalgotrade 相比, QuantDigger的策略语法更接近策略开发人员的习惯。目前的功能包括:股票回测,期货回测。 支持选股,套利,择时, 组合策略。自带了一个基于matplotlib编写的简单的策略和k线显示界面,能满足广大量化爱好者 基本的回测需求。设计上也兼顾了实盘交易,未来如果有时间,也会加入交易接口。开发人员都是量化爱好者,也欢迎感兴趣的新朋友加入开发, 我的QQ交流群:334555399。使用中遇到任何问题,欢迎提个[issue](https://github.com/QuantFans/quantdigger/issues)。
除了开发人员,也特别感谢以下朋友给的建议:
北京的 vodkabuaa
国元证券的王林峰
深大的邓志浩
matplotlib
numpy
logbook
pandas
progressbar2
zmq
BeautifulSoup4 (tushare需要)
lxml (tushare需要)
tushare (一个非常强大的股票信息抓取工具)
python-dateutil(可选)
IPython
TA-Lib
- 可以用pip安装依赖库:
>>> pip install -r requirements/requirements.txt
- 如果出现pypi源超时情况:
>>> pip install -r requirements/requirements.txt -i http://pypi.douban.com/simple --trusted-host pypi.douban.com
- TA-Lib 通过pip直接安装可能会出错,
- 到 http://www.lfd.uci.edu/~gohlke/pythonlibs/#ta-lib 下载相应版本然后通过命令安装,如
>>> pip install TA_Lib-0.4.10-cp36-cp36m-win_amd64.whl
- Anaconda用户可以用
>>> conda install -c quantopian ta-lib
#from quantdigger.engine.series import NumberSeries#from quantdigger.indicators.common import MA#from quantdigger.util import pcontractfromquantdiggerimport*importsixclassDemoStrategy(Strategy): """ 策略A1 """defon_init(self, ctx): """初始化数据"""ctx.ma10=MA(ctx.close, 10, 'ma10', 'y', 2) ctx.ma20=MA(ctx.close, 20, 'ma20', 'b', 2) defon_symbol(self, ctx): """ 选股 """returndefon_bar(self, ctx): ifctx.curbar>20: ifctx.ma10[2] <ctx.ma20[2] andctx.ma10[1] >ctx.ma20[1]: ctx.buy(ctx.close, 1) elifctx.pos() >0andctx.ma10[2] >ctx.ma20[2] and \ ctx.ma10[1] <ctx.ma20[1]: ctx.sell(ctx.close, ctx.pos()) defon_exit(self, ctx): returnclassDemoStrategy2(Strategy): """ 策略A2 """defon_init(self, ctx): """初始化数据"""ctx.ma5=MA(ctx.close, 5, 'ma5', 'y', 2) ctx.ma10=MA(ctx.close, 10, 'ma10', 'black', 2) defon_symbol(self, ctx): """ 选股 """returndefon_bar(self, ctx): ifctx.curbar>10: ifctx.ma5[2] <ctx.ma10[2] andctx.ma5[1] >ctx.ma10[1]: ctx.buy(ctx.close, 1) elifctx.pos() >0andctx.ma5[2] >ctx.ma10[2] and \ ctx.ma5[1] <ctx.ma10[1]: ctx.sell(ctx.close, ctx.pos()) defon_exit(self, ctx): returnif__name__=='__main__': set_symbols(['BB.SHFE-1.Minute'], 0) # 创建组合策略# 初始资金5000, 两个策略的资金配比为0.2:0.8profile=add_strategy([DemoStrategy('A1'), DemoStrategy2('A2')],{'captial': 5000, 'ratio': [0.2, 0.8] }) run() # 绘制k线,交易信号线fromquantdigger.diggerimportfinance, plottingplotting.plot_strategy(profile.data(0), profile.indicators(1), profile.deals(1)) # 绘制策略A1, 策略A2, 组合的资金曲线curve0=finance.create_equity_curve(profile.all_holdings(0)) curve1=finance.create_equity_curve(profile.all_holdings(1)) curve=finance.create_equity_curve(profile.all_holdings()) plotting.plot_curves([curve0.equity, curve1.equity, curve.equity], colors=['r', 'g', 'b'], names=[profile.name(0), profile.name(1), 'A0']) # 绘制净值曲线plotting.plot_curves([curve.networth]) # 打印统计信息six.print_(finance.summary_stats(curve, 252*4*60))- k线和信号线
k线显示使用了系统自带的一个联动窗口控件,由蓝色的滑块控制显示区域,可以通过鼠标拖拽改变显示区域。 上下方向键 来进行缩放。
>>> [('Total Return', '-0.99%'), ('Sharpe Ratio', '-5.10'), ('Max Drawdown', '1.72%'), ('Drawdown Duration', '3568')] 0.5.1 版本 2017-07-13
- 在原来0.5.0版的基础上改为支持Python3.6
0.5.0 版本 2017-01-08
- 完善文档
- 数据源可配置
- 添加shell, 界面,回测引擎三则间的交互框架
0.3.0 版本 2015-12-09
- 重新设计回测引擎, 支持组合回测,选股
- 重构数据模块
0.2.0 版本 2015-08-18
- 修复股票回测的破产bug
- 修复回测权益计算bug
- 交易信号对的计算从回测代码中分离
- 把回测金融指标移到digger/finace
- 添加部分数据结构,添加部分数据结构字段
- 添加几个mongodb相关的函数
0.1.0 版本 2015-06-16
- 夸品种的策略回测功能
- 简单的交互
- 指标,k线绘制


