升级日志
......................................................................................................................
..########.....##........##..........##.........####........##...##########.......##......##......###...##.....######..
.##......## ...##........##.........####........##.##.......##.......##..........###.......##....##.....##...##.....##.
##........##...##........##........##.##........##..##......##.......##.........####........#...##......##...##......##
##........##...##........##.......##...##.......##...##.....##.......##.........##.##.......##.##.......##....##.......
##........##...##........##......##.....##......##....##....##.......##........##..###.......###........##.....##......
##........##...##........##......##......##.....##.....##...##.......##.......##....##.......###........##......###....
##........##...##........##.....##........##....##......##..##.......##......##......##.....##.##.......##........##...
##........##...##........##....#############....##.......##.##.......##.....###########.....##..##......##.........##..
###.......##...##........##...##...........##...##.......##.##.......##....##.........##...##...##......##...##.....##.
.##......###....##......###..##.............##..##........####.......##...##..........##..###....##.....##....##.....##
..#########......########...##..............###.##..........##.......##..##............##.##......##....##.....##....##
........#####.....................................................................................................####..
......................................................................................................................
..........................Copyright..yutiansut..2017......QUANTITATIVE FINANCIAL FRAMEWORK..............................
......................................................................................................................
........................................................................................................................
最新版本 :0.4.0-beta
Github workers
GitHub stars
GitHub forks
-
升级日志
-
新的功能:
- 1.1 组合回测支持
- 1.2 多种交易状态支持
- 1.3 实盘交易的支持
- 1.4 更加方便的数据更新接口
- 1.5 手续费/滑点
- 1.6 QUANTAXIS-Log 优化
- 1.7 重构了回测流程,简化回测设置步骤
- 1.8 对于QACMD进行改进
- 1.9 新增QADataStruct模块
- 1.10 对于backtest的一个bug修改:
- 1.11 对于SU中心的修改
- 1.12 增加了界面的logo以及log的logo
- 1.13 增加一个标准化的QUANTAXIS事件队列(0.3.9)
- 1.14 增加了两个时间选择的api(0.3.9)
- 1.15 增加了一个事件订阅的方式QA.QA_Event(0.3.9):
- 1.16 修改了CLI中的创建example内容(0.3.9):
- 1.17 增加了一个带参数的延时装饰器
- 1.18 web 部分的改进
- 1.19 回测部分的最后一天停牌情况的处理
- 1.20 新增一个时间接口 QA_util_time_now()
- 1.21 新增一个QAWeb的接口
- 1.22 新增一个QACSV的接口
- 1.23 新增一个交易时间的变量
- 1.24 新增一个k线接口(通达信)
- 1.25 在回测的时候,增加一个回测内全局变量
- 1.26 新增一个创建多维list的函数
- 1.27 修改了一个QABacktest的传参, 现在在策略中,需要指定买卖状态
- 1.28 对于backtest的加载的外部函数的接口进行修改
- 1.29 修改了QA_Query 里返回的数组格式
- 1.30 修改了QA_util_save_csv的模式
- 1.31 通达信5分钟线解析保存
- 1.32 获取指数k线的api更新
- 1.33 QA_util更新 QA_util_time_stamp
- 1.34 QABACKTEST回测引擎更新
- 1.35 QA_fetch_stock_list 函数更新
- 1.36 回测框架新增一个一键平仓函数
- 1.37 回测框架的报价函数增加回调
- 1.38 更新了save/update的方式
- 1.39 QUANTAXIS CLI 增加一个选项 CLEAN
- 1.40 QA_util_diff_list
- 1.41 QUANTAXIS.QABACKTEST.QABACKTEST_STOCKDAY.QA_backtest_get_OHLCV
- 1.42 QUANTAXIS MARKET ENGINE修改
- 1.43 QA_util_get_real_datelist
- 1.44 QA_fetch_stock_full('date',type)(0.4.0-beta-dev28)
- 1.45 同花顺日线爬虫可用(0.4.0-beta-dev29)
- 1.46 数据源更换 QA_fetch_get_stock_day('ts')(0.4.0-beta-dev30,dev31,dev32)
- 1.47 新增数据库api QA_util_mongo_initial,QA_util_mongo_make_index(0.4.0-beta-dev33)
- 1.48 修复一个委托单超过市场上下限 返回none的bug(0.4.0-beta-dev34)
- 1.49 修复了pytdx的get_k_data api(0.4.0-beta-dev35)
- 巨大改动/重构
- 重要性能优化 重新定义回测流程,减少数据库IO压力
- 废弃的接口
- to do list
-
新的功能:
作者: yutiansut
新的功能:
1.1 组合回测支持
2017/6/13
在之前的版本里,quantaxis是通过穿透性测试去做unit测试.然后对于unit结果进行组合.这种构建组合的方式虽然行之有效,但是在一些情境下会比较笨重.
好比如,你只是想对于特定的股票(如小市值股票)的一些方法进行回测分析,由于这些股票组合是固定的,所以本来只需要进行一次回测,但是在穿透性测试下要进行n次回测并重新组合
新的模式在原有基础上推出了基于组合的unit测试,在固定组合的时候,只需要进行一个测试,就可以得到结果
1.2 多种交易状态支持
2017/6/14
之前的版本中,quantaxis只支持单次持仓,及(不能进行买入-继续买入的状态)
0.3.9-gamma进行了一定的修改和优化,目前支持了多次连续买入和卖出的交易状态,并通过order_id和trade_id来锁定买卖的对应关系
1.3 实盘交易的支持
2017/6/14
通过tradex的接口,quantaxis实现了一套实盘的解决方案,在quantaxistrade文件夹下,具体详见quantaxis_trade
1.4 更加方便的数据更新接口
2017/6/15
import QUANTAXIS as QA
QA.QA_SU_update_stock_day(client=QA.QA_Setting.client,engine='ts')
1.5 手续费/滑点
2017/6/16
重写了交易撮合引擎,区分不同市场/状态,同时对于股票交易量的区分有了更加接近实盘的表现
-
对于滑点的设置:
- 如果报价单的购买数量小于1/16当日成交量
按正常报价进行交易
- 如果报价单的购买数量在1/16-1/8 当日成交量, 成交价会进行一个浮动:
买入价=mean(max{o,c},h) 卖出价=mean(min{o,c},l)
- 如果报价单的数量在1/8当日成交量以上,则只能成交1/8的当日成交量
买入价=high 卖出价=low 交易数量=1/8当日成交量
-
对于手续费的设置:
买入的时候无需付费,卖出的时候,按成交额收万分之五的手续费,并在现金中扣除
1.6 QUANTAXIS-Log 优化
2017/6/15
ipython
In [1]: import QUANTAXIS as QA
QUANTAXIS>> start QUANTAXIS
QUANTAXIS>> ip:127.0.0.1 port:27017
QUANTAXIS>> Welcome to QUANTAXIS, the Version is 0.3.9-beta-dev20
1.7 重构了回测流程,简化回测设置步骤
2017/6/21
在quantaxis的backtest的基础上,对于回测的流程和模式进行了重构,通过ini的读取以及函数式编程的方法,把回测简化到一个ini+策略即可进行回测的步骤
具体参见 test/new test 文件夹
1.8 对于QACMD进行改进
2017/6/26
对于QACMD的模式进行了修改,现在直接在命令行中输入quantaxis即可进入quantaxis的交互式界面进行操作
1.9 新增QADataStruct模块
2017/6/27
datastruct将在未来对于不同的场景下的数据进行重构和规范化处理,目前新增了基础数据类,机器学习类,ohlc价格序列类等
1.10 对于backtest的一个bug修改:
2017/6/27
修改了购买的时候的方式,其中mean的报价方式改变为当前可用资金/股票列表数量的资金分配方式
修改了对于cash小于买入报单总量的修正,以免出现在不能买入的时候的误操作
1.11 对于SU中心的修改
修改并优化了对于QUANTAXIS backtest的回测过程的资金曲线的存储过程,增加了存为csv的方法,并优化和修正了存储时的结果
1.12 增加了界面的logo以及log的logo
Markdown
Markdown
1.13 增加一个标准化的QUANTAXIS事件队列(0.3.9)
2017/6/30-2017/7/2,2017/7/4
引入方式:
from QUANTAXIS import QA_Queue
使用方式:
首先需要引入一个标准化的队列:
from six.moves import queue
import queue # python3
import Queue # python2
qa=queue.Queue()# 你可以自定义队列的大小
#启动一个事件队列:
qa_event=QA_Queue(qa)
# 往事件引擎里发事件,需要有函数
"""
标准的事件是:
{'type':'xxx','fn':'func'}
"""
qa.put({'type':'xxx','fn':'func'})
事件引擎会默认一直监听这个队列
2017/7/4 update 重新优化了这个事件引擎 参见test/test_job_queue.py
13:35:45 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:45 QUANTAXIS>>> job--id:0
13:35:46 QUANTAXIS>>> job--id:1
13:35:46 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 2 tasks to do
13:35:46 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:46 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:47 QUANTAXIS>>> job--id:2
13:35:47 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:47 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:47 QUANTAXIS>>> job--id:3
13:35:48 QUANTAXIS>>> job--id:4
13:35:48 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 2 tasks to do
13:35:48 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:48 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:49 QUANTAXIS>>> job--id:5
13:35:49 QUANTAXIS>>> job--id:6
13:35:49 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 2 tasks to do
13:35:49 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:49 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:50 QUANTAXIS>>> job--id:7
13:35:50 QUANTAXIS>>> job--id:8
13:35:50 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 2 tasks to do
13:35:50 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:50 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:51 QUANTAXIS>>> job--id:9
13:35:51 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:51 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:52 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:53 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:54 QUANTAXIS>>> job--id:1
13:35:54 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>: There are still 1 tasks to do
13:35:54 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:55 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:56 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:57 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:58 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:35:59 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:36:00 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
13:36:01 QUANTAXIS>>> From Engine <QA_Queue(EVENT ENGINE, started 12488)>Engine will waiting for new task ...
1.14 增加了两个时间选择的api(0.3.9)
2017/7/3
-
QUANTAXIS.QAUtil.QADate.QA_select_hours
-
QUANTAXIS.QAUtil.QADate.QA_select_min
引入方式
from QUANTAXIS.QAUtil import QA_select_hours,QA_select_min
1.15 增加了一个事件订阅的方式QA.QA_Event(0.3.9):
2017/7/3
from QUANTAXIS.QATask import QA_Event
class MyEvent(QA_Event):
ASK = "askMyEvent"
RESPOND = "respondMyEvent"
class WhoAsk(object):
def __init__(self, event_dispatcher):
self.event_dispatcher = event_dispatcher
self.event_dispatcher.add_event_listener(
MyEvent.RESPOND, self.on_answer_event
)
def ask(self):
print("who are listener to me?")
self.event_dispatcher.dispatch_event(MyEvent(MyEvent.ASK, self))
def on_answer_event(self, event):
print("receive event %s", event.data)
class WhoRespond(object):
def __init__(self, event_dispatcher):
self.event_dispatcher = event_dispatcher
self.event_dispatcher.add_event_listener(
MyEvent.ASK, self.on_ask_event)
def on_ask_event(self, event):
self.event_dispatcher.dispatch_event(
MyEvent(MyEvent.RESPOND, self))
dispatcher = QA_EventDispatcher()
who_ask = WhoAsk(dispatcher)
who_responde1 = WhoRespond(dispatcher)
who_responde2 = WhoRespond(dispatcher)
# WhoAsk ask
who_ask.ask()
result:
who are listener to me?
receive event %s <__main__.WhoRespond object at 0x02361D50>
receive event %s <__main__.WhoRespond object at 0x02361DB0>
1.16 修改了CLI中的创建example内容(0.3.9):
2017/7/3
现在在命令行中输入QUANTAXIS:
λ quantaxis
QUANTAXIS>> start QUANTAXIS
QUANTAXIS>> ip:127.0.0.1 port:27017
QUANTAXIS>> From Engine <_MainThread(MainThread, started 19040)>2017-07-03 14:22:08.218800
QUANTAXIS>> FROM QUANTAXIS SYS== now running 3 threads
QUANTAXIS>> Welcome to QUANTAXIS, the Version is 0.3.9-beta-dev29
QUANTAXIS> examples
QUANTAXIS>> QUANTAXIS example
QUANTAXIS>> successfully generate a example strategy inC:\Users\yutiansut\strategy\A05-CCI
QUANTAXIS> exit
λ ls
目录: C:\Users\yutiansut\strategy\A05-CCI
Mode LastWriteTime Length Name
---- ------------- ------ ----
-a--- 2017/7/3 14:22 358 backtest.ini
-a--- 2017/7/3 14:22 905 backtest.py
-a--- 2017/7/3 14:22 2738 quantaxis-2017-07-03-14-22-08-.log
1.17 增加了一个带参数的延时装饰器
2017/7/4
QUANTAXIS.QAUtil.QADate.QA_util_time_delay()
使用方式
from QUANTAXIS import QA_util_time_dalay
@QA_util_time_dalay(2)
#延时2秒
def pp():
print(1)
1.18 web 部分的改进
2017/7/10
web部分,修改了页面布局和回测的显示方式
去除了之前单个股票和行情的对比
改成了资金曲线和组合分析的方式
1.19 回测部分的最后一天停牌情况的处理
2017/7/10
_remains_day=0
while __message['header']['status']==500:
#停牌状态,这个时候按停牌的最后一天计算价值(假设平仓)
__last_bid['date'] = self.trade_list[self.end_real_id-_remains_day]
_remains_day+=1
__message = self.market.receive_bid(__last_bid, self.setting.client)
#直到市场不是为0状态位置,停止前推日期
在原来的代码里,组合的最后一天需要被平仓,而如果恰巧这一天停牌:(600127,601001 在2017/6/30都是停牌状态)
那么就会一直陷入市场返回500错误,而回测框架一直在组装报价,来回轮询,陷入死循环
现在加入了对于市场的数据的检测,如果是500状态,则会将交易日向前递推一天,重新组装报价,直到停牌的最后一天为止
1.20 新增一个时间接口 QA_util_time_now()
2017/7/10
返回的就是datetime.datetime.now()的结果
1.21 新增一个QAWeb的接口
2017/7/10
可以在命令行直接输入 quantaxis_web 来启动这个服务器,端口在5050(带socket)
1.22 新增一个QACSV的接口
2017/7/13
现在可以便捷的吧数据存成csv格式:
import QUANTAXIS as QA
data=['1',2,'x','t']
QA.QA_util_save_csv(data,'test')
1.23 新增一个交易时间的变量
2017/7/13
from QUANTAXIS import trade_date_sse
print(trade_date_sse)
1.24 新增一个k线接口(通达信)
2017/7/13
import QUANTAXIS as QA
QA.QA_fetch_get_stock_day('tdx','601801','2017-01-01','2017-07-01')
Out[3]:
open close high low vol amount year month day \
0 14.39 14.70 14.76 14.36 164426.0 239606672.0 2016 12 14
1 14.55 14.99 15.10 14.55 188986.0 281594560.0 2016 12 15
2 15.03 16.00 16.11 14.86 180712.0 281681280.0 2016 12 16
3 15.93 16.79 17.05 15.93 207338.0 345849056.0 2016 12 19
4 16.87 16.80 16.87 16.39 81613.0 135869952.0 2016 12 20
5 16.90 16.88 17.12 16.66 98858.0 166289104.0 2016 12 21
6 16.95 16.61 16.96 16.28 92686.0 153868304.0 2016 12 22
7 16.58 16.51 17.05 16.03 89121.0 147409888.0 2016 12 23
8 16.59 17.30 17.65 16.26 226506.0 387817440.0 2016 12 26
9 17.35 17.49 17.62 17.14 106783.0 186348288.0 2016 12 27
10 17.56 17.89 17.96 17.40 181507.0 321431424.0 2016 12 28
11 17.69 17.59 17.84 17.35 109785.0 193161984.0 2016 12 29
12 17.58 17.57 17.90 17.24 142565.0 251420576.0 2016 12 30
13 17.60 17.45 17.60 17.15 106347.0 185482976.0 2017 1 3
14 17.40 17.00 17.40 16.41 204781.0 344751936.0 2017 1 4
15 16.85 16.86 17.00 16.55 102506.0 171934160.0 2017 1 5
16 16.89 16.90 17.13 16.73 77188.0 130475344.0 2017 1 6
17 16.86 16.81 17.10 16.76 77502.0 131223856.0 2017 1 9
18 16.81 17.40 17.53 16.10 193194.0 325042304.0 2017 1 10
19 17.20 17.68 17.84 17.10 368129.0 651724672.0 2017 1 11
20 17.68 17.96 18.24 17.61 408000.0 735998016.0 2017 1 12
21 17.96 18.31 18.84 17.80 337911.0 619458496.0 2017 1 13
22 18.30 18.56 18.66 17.36 284054.0 518380576.0 2017 1 16
23 18.39 18.57 19.00 17.60 263141.0 491844896.0 2017 1 17
24 18.23 17.99 18.80 17.90 95064.0 173332688.0 2017 1 18
25 17.81 18.24 18.39 17.80 49353.0 89376280.0 2017 1 19
26 18.32 18.46 18.69 18.19 143792.0 265606432.0 2017 1 20
27 18.28 18.84 19.00 18.26 228183.0 427935936.0 2017 1 23
28 18.80 19.00 19.10 18.50 113616.0 214593088.0 2017 1 24
29 19.00 18.78 19.14 18.72 265691.0 504179328.0 2017 1 25
.. ... ... ... ... ... ... ... ... ...
89 12.73 12.98 13.16 12.68 176499.0 228946128.0 2017 5 17
90 12.78 13.26 13.27 12.71 145098.0 189804288.0 2017 5 18
91 13.20 13.09 13.33 13.08 78134.0 102854576.0 2017 5 19
92 13.15 12.63 13.18 12.50 96282.0 123843520.0 2017 5 22
93 12.64 12.76 13.16 12.53 111309.0 141901152.0 2017 5 23
94 12.74 13.17 13.19 12.62 104595.0 136030272.0 2017 5 25
95 13.12 13.09 13.25 13.08 69110.0 90972200.0 2017 5 26
96 13.25 13.42 13.44 13.18 110524.0 147250112.0 2017 5 31
97 13.43 13.26 13.46 13.07 88958.0 117311632.0 2017 6 1
98 13.09 12.91 13.09 12.70 91351.0 117710656.0 2017 6 2
99 12.98 12.83 12.98 12.74 110453.0 141482496.0 2017 6 5
100 12.77 12.83 12.85 12.68 62182.0 79344728.0 2017 6 6
101 12.80 13.26 13.30 12.74 144478.0 188284080.0 2017 6 7
102 13.19 13.13 13.33 13.09 101171.0 133169712.0 2017 6 8
103 13.12 13.04 13.15 12.96 87196.0 113572160.0 2017 6 9
104 13.00 12.76 13.00 12.74 86848.0 111470976.0 2017 6 12
105 12.62 12.96 13.02 12.60 67389.0 87066136.0 2017 6 13
106 13.00 12.75 13.00 12.72 62876.0 80648232.0 2017 6 14
107 12.74 12.85 12.95 12.70 71232.0 91385064.0 2017 6 15
108 12.87 12.71 12.87 12.67 59412.0 75701632.0 2017 6 16
109 12.72 12.90 12.92 12.69 77521.0 98993784.0 2017 6 19
110 12.86 13.21 13.24 12.86 180431.0 237042944.0 2017 6 20
111 13.28 13.21 13.32 13.00 117902.0 154974640.0 2017 6 21
112 13.20 13.37 13.58 13.15 191968.0 256520880.0 2017 6 22
113 13.28 13.21 13.29 12.99 114182.0 150111952.0 2017 6 23
114 13.22 13.71 13.73 13.16 186511.0 252084368.0 2017 6 26
115 13.74 13.78 13.81 13.54 135037.0 184420320.0 2017 6 27
116 13.73 13.98 14.05 13.56 242043.0 335440128.0 2017 6 28
117 13.94 13.96 14.01 13.84 178812.0 248913488.0 2017 6 29
118 13.89 13.82 13.94 13.71 83402.0 115219784.0 2017 6 30
hour minute datetime
0 15 0 2016-12-14 15:00
1 15 0 2016-12-15 15:00
2 15 0 2016-12-16 15:00
3 15 0 2016-12-19 15:00
4 15 0 2016-12-20 15:00
5 15 0 2016-12-21 15:00
6 15 0 2016-12-22 15:00
7 15 0 2016-12-23 15:00
8 15 0 2016-12-26 15:00
9 15 0 2016-12-27 15:00
10 15 0 2016-12-28 15:00
11 15 0 2016-12-29 15:00
12 15 0 2016-12-30 15:00
13 15 0 2017-01-03 15:00
14 15 0 2017-01-04 15:00
15 15 0 2017-01-05 15:00
16 15 0 2017-01-06 15:00
17 15 0 2017-01-09 15:00
18 15 0 2017-01-10 15:00
19 15 0 2017-01-11 15:00
20 15 0 2017-01-12 15:00
21 15 0 2017-01-13 15:00
22 15 0 2017-01-16 15:00
23 15 0 2017-01-17 15:00
24 15 0 2017-01-18 15:00
25 15 0 2017-01-19 15:00
26 15 0 2017-01-20 15:00
27 15 0 2017-01-23 15:00
28 15 0 2017-01-24 15:00
29 15 0 2017-01-25 15:00
.. ... ... ...
89 15 0 2017-05-17 15:00
90 15 0 2017-05-18 15:00
91 15 0 2017-05-19 15:00
92 15 0 2017-05-22 15:00
93 15 0 2017-05-23 15:00
94 15 0 2017-05-25 15:00
95 15 0 2017-05-26 15:00
96 15 0 2017-05-31 15:00
97 15 0 2017-06-01 15:00
98 15 0 2017-06-02 15:00
99 15 0 2017-06-05 15:00
100 15 0 2017-06-06 15:00
101 15 0 2017-06-07 15:00
102 15 0 2017-06-08 15:00
103 15 0 2017-06-09 15:00
104 15 0 2017-06-12 15:00
105 15 0 2017-06-13 15:00
106 15 0 2017-06-14 15:00
107 15 0 2017-06-15 15:00
108 15 0 2017-06-16 15:00
109 15 0 2017-06-19 15:00
110 15 0 2017-06-20 15:00
111 15 0 2017-06-21 15:00
112 15 0 2017-06-22 15:00
113 15 0 2017-06-23 15:00
114 15 0 2017-06-26 15:00
115 15 0 2017-06-27 15:00
116 15 0 2017-06-28 15:00
117 15 0 2017-06-29 15:00
118 15 0 2017-06-30 15:00
[119 rows x 12 columns]
1.25 在回测的时候,增加一个回测内全局变量
2017/7/14
现在的回测加载的函数句柄里面会增加一个叫info的句柄,这个句柄在引擎里面是交易日循环外部的变量,可以存贮一些连续周期信息,和一些点位信息
在quantaxis/test/new test的strategy里面可以看到改动
1.26 新增一个创建多维list的函数
2017/7/14
在QUANTAXIS.QAUtil中, 接口的名称是 QA_util_multi_demension_list(row_,col_)
如果需要创建一个[[],[]], 那就用 row_=2,col=0
其他时候,返回的都是[[None]]
import QUANTAXIS as QA
QA.QAUtil.QA_util_multi_demension_list(3,0)
QA.QAUtil.QA_util_multi_demension_list(3,3)
[[], [], []]
[[None, None, None], [None, None, None], [None, None, None]]
1.27 修改了一个QABacktest的传参, 现在在策略中,需要指定买卖状态
2017/7/19
现在 QUANTAXIS在接受策略的传参的时候,需要指定买卖状态
在开始的时候,quantaxis使用的是if_buy和账户状态来判断是否买卖的
- 账户持仓, if_buy=1 视为卖出
- 账户空仓, if_buy=1 视为买入
买卖行为和账户状态相关联对于买卖的行为有很大的限制,现在对于这个行为和账户持仓状态进行了解耦
同时,为了兼容性,如果策略并没有给出if_buy或者if_sell,则会将其初始化为0,及不操作
以下是改动后的源代码
def __QA_backtest_excute_bid(self, __result, __date, __hold, __code, __amount):
"""
这里是处理报价的逻辑部分
2017/7/19 修改
__result传进来的变量重新区分: 现在需要有 if_buy, if_sell
因为需要对于: 持仓状态下继续购买进行进一步的支持*简单的情形就是 浮盈加仓
if_buy, if_sell都需要传入
现在的 买卖状态 和 持仓状态 是解耦的
"""
# 为了兼容性考虑,我们会在开始的时候检查是否有这些变量
if 'if_buy' not in list(__result.keys()):
__result['if_buy'] = 0
if 'if_sell' not in list(__result.keys()):
__result['if_sell'] = 0
self.__QA_backtest_set_bid_model()
if self.bid.bid['bid_model'] == 'strategy':
__bid_price = __result['price']
else:
__bid_price = self.bid.bid['price']
__bid = self.bid.bid
__bid['order_id'] = str(random.random())
__bid['user'] = self.setting.QA_setting_user_name
__bid['strategy'] = self.strategy_name
__bid['code'] = __code
__bid['date'] = __date
__bid['price'] = __bid_price
__bid['amount'], __bid['amount_model'] = self.__QA_bid_amount(
__result['amount'], __amount)
if __result['if_buy'] == 1:
# 这是买入的情况
__bid['towards'] = 1
__message = self.market.receive_bid(
__bid, self.setting.client)
if float(self.account.message['body']['account']['cash'][-1]) > \
float(__message['body']['bid']['price']) * \
float(__message['body']['bid']['amount']):
# 这里是买入资金充足的情况
# 不去考虑
pass
else:
# 如果买入资金不充足,则按照可用资金去买入
__message['body']['bid']['amount'] = int(float(
self.account.message['body']['account']['cash'][-1]) / float(
float(str(__message['body']['bid']['price'])[0:5]) * 100)) * 100
if __message['body']['bid']['amount'] > 0:
# 这个判断是为了 如果买入资金不充足,所以买入报了一个0量单的情况
#如果买入量>0, 才判断为成功交易
self.account.QA_account_receive_deal(__message)
elif __result['if_buy'] == 0:
# 如果买入状态为0,则不进行任何买入操作
pass
# 下面是卖出操作,这里在卖出前需要考虑一个是否有仓位的问题:
# 因为在股票中是不允许卖空操作的,所以这里是股票的交易引擎和期货的交易引擎的不同所在
if __result['if_sell'] == 1 and __hold == 1:
__bid['towards'] = -1
__message = self.market.receive_bid(
__bid, self.setting.client)
self.account.QA_account_receive_deal(
__message)
1.28 对于backtest的加载的外部函数的接口进行修改
2017/7/20
对于backtest的加载的外部策略,现在暂时有以下三个句柄
- on_start
- strategy
- on_end
其中,on_start/on_end这两个句柄可有可无, strategy句柄是必须要有的
1.29 修改了QA_Query 里返回的数组格式
2017/7/21
def QA_fetch_stock_day(code, startDate, endDate, collections=QA_Setting.client.quantaxis.stock_day, type_='numpy'):
# print(datetime.datetime.now())
startDate = str(startDate)[0:10]
endDate = str(endDate)[0:10]
if QA_util_date_valid(endDate) == True:
list_a = []
for item in collections.find({
'code': str(code)[0:6], "date_stamp": {
"$lte": QA_util_date_stamp(endDate),
"$gte": QA_util_date_stamp(startDate)}}):
list_a.append([str(item['code']), float(item['open']), float(item['high']), float(
item['low']), float(item['close']), float(item['volume']), item['date'], float(item['turnover'])])
## 多种数据格式
if type_ == 'numpy':
data = numpy.asarray(list_a)
elif type_ == 'list':
data = list_a
elif type_ == 'pandas':
data = DataFrame(list_a, columns=[
'code', 'open', 'high', 'low', 'close', 'volume', 'date', 'turnover'])
return data
else:
QA_util_log_info('something wrong with date')
现在可以通过QA_fetch_stock_day里面选择是返回list,numpy还是dataframe. 为了兼容性考虑,默认依然是numpy
1.30 修改了QA_util_save_csv的模式
2017/07/21
def QA_util_save_csv(data: list, name: str, column=None, location=None):
# 重写了一下保存的模式
assert isinstance(data, list)
if location == None:
path = './' + str(name) + '.csv'
else:
path = location + str(name) + '.csv'
with open(path, 'w', newline='') as f:
csvwriter = csv.writer(f)
if column == 'None':
pass
else:
csvwriter.writerow(column)
for item in data:
csvwriter.writerow(item)
示例:
import QUANTAXIS as QA
QA.QA_util_save_csv(QA.QA_fetch_stock_day('000001','2017-01-01','2017-07-01','list'),'000001')
000001,9.11,9.18,9.09,9.16,459840.47,2017-01-03,0.31
000001,9.15,9.18,9.14,9.16,449329.53,2017-01-04,0.31
000001,9.17,9.18,9.15,9.17,344372.91,2017-01-05,0.24
000001,9.17,9.17,9.11,9.13,358154.19,2017-01-06,0.24
000001,9.13,9.17,9.11,9.15,361081.56,2017-01-09,0.21
000001,9.15,9.16,9.14,9.15,241053.95,2017-01-10,0.14
000001,9.14,9.17,9.13,9.14,303430.88,2017-01-11,0.18
000001,9.13,9.17,9.13,9.15,428006.75,2017-01-12,0.25
000001,9.14,9.19,9.12,9.16,434301.38,2017-01-13,0.26
000001,9.15,9.16,9.07,9.14,683165.81,2017-01-16,0.4
000001,9.12,9.16,9.1,9.15,545552.38,2017-01-17,0.32
000001,9.14,9.19,9.13,9.17,574269.38,2017-01-18,0.34
000001,9.15,9.24,9.15,9.18,437712.88,2017-01-19,0.26
000001,9.17,9.23,9.17,9.22,393328.56,2017-01-20,0.23
000001,9.22,9.26,9.2,9.22,420299.31,2017-01-23,0.25
000001,9.23,9.28,9.2,9.27,470244.09,2017-01-24,0.28
000001,9.27,9.28,9.25,9.26,304401.97,2017-01-25,0.18
000001,9.27,9.34,9.26,9.33,420712.56,2017-01-26,0.25
000001,9.45,10.26,9.33,10.26,19345.0,2017-02-02,0.01
000001,9.34,9.36,9.23,9.26,315472.25,2017-02-03,0.19
000001,9.26,9.32,9.26,9.31,516786.12,2017-02-06,0.31
000001,9.31,9.32,9.27,9.3,396884.97,2017-02-07,0.23
...........
1.31 通达信5分钟线解析保存
2017/7/23
首先要下载:
- 深圳5分钟线
http://www.tdx.com.cn/products/data/data/vipdoc/sz5fz.zip
- 上证5分钟线
http://www.tdx.com.cn/products/data/data/vipdoc/sh5fz.zip
然后解压到一个目录里
import QUANTAXIS as QA
QA.QA_SU_save_stock_min_5('你解压好的目录')
#例如 QA.QA_SU_save_stock_min_5('C:\\users\\yutiansut\\desktop\\sh5fz')
In [2]: QA.QA_SU_save_stock_min_5('C:\\users\\yutiansut\\desktop\\sz5fz')
QUANTAXIS>> Now_saving 000001's 5 min tick
QUANTAXIS>> Now_saving 000002's 5 min tick
QUANTAXIS>> Now_saving 000004's 5 min tick
QUANTAXIS>> Now_saving 000005's 5 min tick
QUANTAXIS>> Now_saving 000006's 5 min tick
QUANTAXIS>> Now_saving 000007's 5 min tick
QUANTAXIS>> Now_saving 000008's 5 min tick
QUANTAXIS>> Now_saving 000009's 5 min tick
QUANTAXIS>> Now_saving 000010's 5 min tick
QUANTAXIS>> Now_saving 000011's 5 min tick
QUANTAXIS>> Now_saving 000012's 5 min tick
QUANTAXIS>> Now_saving 000014's 5 min tick
QUANTAXIS>> Now_saving 000016's 5 min tick
QUANTAXIS>> Now_saving 000017's 5 min tick
QUANTAXIS>> Now_saving 000018's 5 min tick
QUANTAXIS>> Now_saving 000019's 5 min tick
QUANTAXIS>> Now_saving 000020's 5 min tick
QUANTAXIS>> Now_saving 000021's 5 min tick
QUANTAXIS>> Now_saving 000022's 5 min tick
QUANTAXIS>> Now_saving 000023's 5 min tick
QUANTAXIS>> Now_saving 000024's 5 min tick
QUANTAXIS>> Now_saving 000025's 5 min tick
QUANTAXIS>> Now_saving 000026's 5 min tick
QUANTAXIS>> Now_saving 000027's 5 min tick
QUANTAXIS>> Now_saving 000028's 5 min tick
1.32 获取指数k线的api更新
2017/7/24
QA_fetch_index_day(code, startDate, endDate,type_='numpy' ,collections=QA_Setting.client.quantaxis.stock_day)
一般而言,直接使用QA_fetch_index_day(code, startDate, endDate)就可以了
1.33 QA_util更新 QA_util_time_stamp
2017/7/24
QA_util_time_stamp用于将时间转化成时间戳
import QUANTAXIS as QA
QA.QA_util_time_stamp('2017-01-01 10:25:08')
1.34 QABACKTEST回测引擎更新
2017/8/1-8/4
重大更新:
现在大量使用@装饰器来扩展回测引擎的能力
import QUANTAXIS as QA
from QUANTAXIS import QA_Backtest_stock_day as QB
"""
写在前面:
===============QUANTAXIS BACKTEST STOCK_DAY中的变量
常量:
QB.account.message 当前账户消息
QB.account.cash 当前可用资金
QB.account.hold 当前账户持仓
QB.account.history 当前账户的历史交易记录
QB.account.assets 当前账户总资产
QB.account.detail 当前账户的交易对账单
QB.account.init_assest 账户的最初资金
QB.strategy_stock_list 回测初始化的时候 输入的一个回测标的
QB.strategy_start_date 回测的开始时间
QB.strategy_end_date 回测的结束时间
QB.today 在策略里面代表策略执行时的日期
QB.benchmark_code 策略业绩评价的对照行情
函数:
获取市场(基于gap)行情:
QB.QA_backtest_get_market_data(QB,code,QB.today)
获取市场自定义时间段行情:
QA.QA_fetch_stock_day(code,start,end,model)
报单:
QB.QA_backtest_send_order(QB, code,amount,towards,order: dict)
order有三种方式:
1.限价成交 order['bid_model']=0或者l,L
注意: 限价成交需要给出价格:
order['price']=xxxx
2.市价成交 order['bid_model']=1或者m,M,market,Market
3.严格成交模式 order['bid_model']=2或者s,S
及 买入按bar的最高价成交 卖出按bar的最低价成交
查询当前一只股票的持仓量
QB.QA_backtest_hold_amount(QB,code)
"""
@QB.backtest_init
def init():
#
QB.setting.QA_util_sql_mongo_ip='192.168.4.189'
QB.account.init_assest=2500000
QB.benchmark_code='hs300'
QB.strategy_stock_list=['000001','000002','600010','601801','603111']
QB.strategy_start_date='2017-03-01'
QB.strategy_end_date='2017-07-01'
@QB.before_backtest
def before_backtest():
global risk_position
QA.QA_util_log_info(QB.account.message)
@QB.load_strategy
def strategy():
print(QB.account.message)
print(QB.account.cash)
input()
for item in QB.strategy_stock_list:
if QB.QA_backtest_hold_amount(QB,item)==0:
#获取数据的第一种办法[这个是根据回测时制定的股票列表初始化的数据]
QB.QA_backtest_send_order(QB,item,10000,1,{'bid_model':'Market'})
else:
print(QB.QA_backtest_hold_amount(QB,item))
QB.QA_backtest_send_order(QB,item,10000,-1,{'bid_model':'Market'})
@QB.end_backtest
def after_backtest():
pass
1.35 QA_fetch_stock_list 函数更新
一个获取股票代码的函数
import QUANTAXIS as QA
QA.QA_fetch_stock_list()
# 自定义数据库的模式
import pymongo
QA.QA_fetch_stock_list(pymongo.MongoClient(ip='192.168.4.189',port=27017).quantaxis.stock_list)
1.36 回测框架新增一个一键平仓函数
2017/8/6
一键平仓:
QB.QA_backtest_sell_all(QB)
1.37 回测框架的报价函数增加回调
2017/8/6
现在回测的报价函数增加了回调
1.38 更新了save/update的方式
2017/8/7-2017/8/8
- update的时候之前出现了 如果该股票尚未上市,数据库无数据的时候 出现负索引的问题 已经解决
- 把之前写在easy里面的代码 写进了quantaxis cli中
quantaxis> save
quantaxis> update
1.39 QUANTAXIS CLI 增加一个选项 CLEAN
2017/8/9
删除旧的回测报告和log文件
QUANTAXIS
quantaxis> clean
1.40 QA_util_diff_list
2017/8/9
一个快速返回前后相减的list函数
1.41 QUANTAXIS.QABACKTEST.QABACKTEST_STOCKDAY.QA_backtest_get_OHLCV
2017/8/9
一个快速拿到OHLCV的函数
拿到开高收低量
Open,High,Low,Close,Volume=QB.QA_backtest_get_OHLCV(QB,QB.QA_backtest_get_market_data(QB,item,QB.today))
1.42 QUANTAXIS MARKET ENGINE修改
2017/8/10
增加了一个严格模式的委托方式
增加了收盘价委托的模式
1.43 QA_util_get_real_datelist
2017/8/10
一个直接拿到真实的交易区间的list
1.44 QA_fetch_stock_full('date',type)(0.4.0-beta-dev28)
2017/8/14
一个获取某一天全市场数据的接口 type有三种
- list
- numpy (默认)
- pandas
import QUANTAXIS as QA
In [2]: QA.QA_fetch_stock_full('2017-08-01')
Out[2]:
array([['600808', '4.57', '4.62', ..., '4.56', '1217308.12', '2017-08-01'],
['600231', '3.85', '3.9', ..., '3.81', '899979.31', '2017-08-01'],
['601619', '3.88', '3.88', ..., '3.88', '1856.2', '2017-08-01'],
...,
['603320', '42.0', '42.95', ..., '42.54', '50230.13', '2017-08-01'],
['hs300', '3738.74', '3770.4', ..., '3770.38', '152546544.0',
'2017-08-01'],
['sz50', '2640.87', '2681.93', ..., '2681.81', '338205.06',
'2017-08-01']],
dtype='<U11')
In [3]: QA.QA_fetch_stock_full('2017-08-01','p')
Out[3]:
code open high low close volume
2017-08-01 600808 4.57 4.62 4.46 4.56 1.217308e+06
2017-08-01 600231 3.85 3.90 3.71 3.81 8.999793e+05
2017-08-01 601619 3.88 3.88 3.88 3.88 1.856200e+03
2017-08-01 300067 6.45 6.72 6.44 6.55 1.744641e+05
2017-08-01 600610 7.23 7.39 7.22 7.29 8.639420e+04
2017-08-01 603778 9.98 10.30 9.90 10.17 1.652618e+05
2017-08-01 601958 8.47 8.48 8.25 8.31 2.703109e+05
2017-08-01 300528 14.60 14.66 14.38 14.64 1.585500e+04
2017-08-01 300675 12.43 12.43 12.43 12.43 7.934700e+02
2017-08-01 600088 15.95 15.95 15.65 15.80 1.194841e+04
2017-08-01 603688 11.38 11.39 11.02 11.06 3.880022e+04
2017-08-01 300416 22.41 22.41 21.90 22.01 8.022790e+03
2017-08-01 300377 10.58 10.74 10.40 10.66 1.219292e+05
2017-08-01 603612 29.44 29.44 29.44 29.44 6.987030e+03
2017-08-01 300387 18.34 18.45 18.01 18.14 6.550860e+03
2017-08-01 603918 21.25 21.46 21.12 21.43 7.064660e+03
2017-08-01 300676 61.63 61.63 61.63 61.63 1.284490e+03
2017-08-01 603909 28.50 28.50 28.01 28.29 1.134000e+04
2017-08-01 600602 7.19 7.34 7.16 7.22 4.433393e+04
2017-08-01 600119 13.83 14.98 13.81 14.66 1.432730e+05
2017-08-01 002423 13.55 13.59 12.96 13.10 9.671325e+04
2017-08-01 002084 8.39 8.45 8.30 8.38 2.086325e+04
2017-08-01 603757 49.70 52.50 48.51 48.96 6.130912e+04
2017-08-01 603903 37.30 38.97 37.00 38.53 3.565021e+04
2017-08-01 300508 75.64 77.70 74.36 77.14 1.624389e+04
2017-08-01 002887 55.45 60.49 55.45 60.49 1.570000e+02
2017-08-01 300402 17.09 17.12 16.40 16.66 4.063693e+04
2017-08-01 600846 10.20 10.58 10.06 10.30 7.099526e+05
2017-08-01 603322 40.40 40.78 40.00 40.50 9.743430e+03
2017-08-01 002302 22.15 22.29 21.65 21.81 3.073918e+05
... ... ... ... ... ... ...
2017-08-01 300666 30.78 30.78 28.58 29.08 1.717917e+05
2017-08-01 600068 11.42 11.63 11.35 11.56 7.490524e+05
2017-08-01 603089 41.33 42.68 38.97 41.38 3.956856e+04
2017-08-01 002236 23.95 24.18 23.57 23.95 1.549279e+05
2017-08-01 002453 8.23 8.44 8.15 8.33 5.073431e+04
2017-08-01 000952 17.06 17.80 16.65 17.00 3.366324e+05
2017-08-01 603578 47.57 48.20 47.50 47.95 5.762650e+03
2017-08-01 300097 25.30 26.09 25.20 25.88 5.197803e+04
2017-08-01 002161 13.12 13.20 12.41 12.74 1.055990e+06
2017-08-01 002753 28.56 28.70 27.27 28.10 1.924529e+05
2017-08-01 000036 13.12 13.13 12.50 12.53 8.219014e+05
2017-08-01 000710 56.98 58.34 56.63 58.00 1.359221e+04
2017-08-01 002386 10.18 10.30 9.90 10.01 6.746068e+05
2017-08-01 002430 11.73 12.37 11.20 11.24 3.188267e+05
2017-08-01 002230 46.89 49.90 46.79 48.75 8.563892e+05
2017-08-01 603133 36.68 36.96 35.57 36.39 1.071875e+05
2017-08-01 603305 51.80 51.80 45.73 48.23 1.686128e+05
2017-08-01 603505 22.10 24.43 21.21 24.43 2.496812e+05
2017-08-01 300679 107.27 107.27 107.27 107.27 1.837600e+02
2017-08-01 300136 38.10 38.31 37.75 38.00 4.952795e+04
2017-08-01 603730 43.42 43.42 41.05 43.42 1.936224e+05
2017-08-01 603355 52.80 52.88 51.89 52.27 7.977550e+03
2017-08-01 600854 7.71 8.10 7.65 7.92 7.716632e+05
2017-08-01 002258 14.06 14.44 13.83 14.03 8.656090e+04
2017-08-01 600230 55.99 57.86 54.09 55.07 4.172212e+05
2017-08-01 600291 35.90 38.30 35.60 36.03 9.485712e+05
2017-08-01 300080 9.05 9.37 8.54 8.65 4.989257e+05
2017-08-01 603320 42.00 42.95 40.61 42.54 5.023013e+04
2017-08-01 hs300 3738.74 3770.40 3737.10 3770.38 1.525465e+08
2017-08-01 sz50 2640.87 2681.93 2638.02 2681.81 3.382051e+05
[3053 rows x 6 columns]
1.45 同花顺日线爬虫可用(0.4.0-beta-dev29)
2017/8/14
import QUANTAXIS as QA
QA.QA_fetch_get_stock_day('ths','000001','2015-01-01','2017-08-01','01')
#引擎用ths或者THS
#最后一个选项是复权方式
#00 不复权 01 前复权 02 后复权
Out[3]:
open high low close volume amount factor
date
2015-01-05 10.70 10.90 10.43 10.72 286043640 4565387800.00 0.034
2015-01-06 10.60 10.98 10.39 10.55 216642140 3453446100.00 0.025
2015-01-07 10.40 10.59 10.22 10.34 170012060 2634796400.00 0.020
2015-01-08 10.36 10.41 9.94 9.98 140771420 2128003400.00 0.017
2015-01-09 9.94 10.62 9.81 10.07 250850020 3835378000.00 0.029
2015-01-12 9.92 10.05 9.66 9.85 155329080 2293104600.00 0.018
2015-01-13 9.77 9.94 9.74 9.79 81687477 1204987140.00 0.010
2015-01-14 9.86 10.15 9.80 9.88 126302964 1889296600.00 0.015
2015-01-15 9.91 10.25 9.81 10.25 124217032 1868795900.00 0.015
2015-01-16 10.29 10.44 10.14 10.27 155584630 2403346100.00 0.018
2015-01-19 9.32 9.71 9.20 9.20 213712360 3016203000.00 0.025
2015-01-20 9.20 9.36 9.01 9.20 149101810 2064280500.00 0.018
2015-01-21 9.23 9.73 9.14 9.61 194053030 2758192700.00 0.023
2015-01-22 9.55 9.68 9.43 9.52 125501611 1801436200.00 0.015
2015-01-23 9.57 9.75 9.52 9.59 145918190 2108746700.00 0.017
2015-01-26 9.57 9.62 9.43 9.55 105760580 1508446500.00 0.012
2015-01-27 9.56 9.57 9.20 9.31 133949466 1881058900.00 0.016
2015-01-28 9.23 9.52 9.18 9.36 124087755 1742175700.00 0.015
2015-01-29 9.19 9.32 9.14 9.25 101675329 1408825300.00 0.012
2015-01-30 9.27 9.40 9.15 9.27 93011669 1298735710.00 0.011
2015-02-02 9.04 9.18 9.00 9.06 86093216 1176949550.00 0.010
2015-02-03 9.16 9.31 9.05 9.28 88334914 1217876590.00 0.010
2015-02-04 9.32 9.34 9.11 9.12 80762311 1122666750.00 0.010
2015-02-05 9.52 9.62 9.15 9.17 191372910 2710523800.00 0.022
2015-02-06 9.10 9.28 8.90 8.98 103040856 1411298500.00 0.012
2015-02-09 8.97 9.09 8.77 8.98 94658628 1273141150.00 0.011
2015-02-10 8.96 9.19 8.91 9.16 72487537 991823740.00 0.009
2015-02-11 9.16 9.21 9.09 9.13 55434951 763414580.00 0.007
2015-02-12 9.14 9.25 9.05 9.22 60871568 838611230.00 0.007
2015-02-13 9.27 9.44 9.21 9.28 88774310 1244514790.00 0.010
... ... ... ... ... ... ... ...
2017-06-21 9.01 9.02 8.95 8.99 49693219 454488890.00 0.002
2017-06-22 8.99 9.24 8.98 9.09 142695820 1325210560.00 0.007
2017-06-23 9.07 9.11 9.00 9.09 58400441 538303570.00 0.003
2017-06-26 9.10 9.24 9.10 9.14 71076995 663762850.00 0.003
2017-06-27 9.14 9.23 9.11 9.20 54601613 509162010.00 0.003
2017-06-28 9.19 9.33 9.17 9.27 116879622 1102438100.00 0.006
2017-06-29 9.27 9.29 9.21 9.27 48880457 459810420.00 0.002
2017-06-30 9.24 9.27 9.15 9.23 49963349 468003520.00 0.002
2017-07-03 9.24 9.27 9.18 9.24 38834937 364465850.00 0.002
2017-07-04 9.24 9.25 9.14 9.18 48836253 456577020.00 0.002
2017-07-05 9.13 9.22 9.11 9.21 56772000 529294110.00 0.003
2017-07-06 9.20 9.25 9.15 9.24 73891186 691387180.00 0.004
2017-07-07 9.21 9.32 9.18 9.31 76036958 717084350.00 0.004
2017-07-10 9.29 9.50 9.28 9.43 136081590 1303090010.00 0.006
2017-07-11 9.45 10.30 9.45 10.09 381208660 3842010200.00 0.018
2017-07-12 10.11 10.42 10.04 10.18 299884400 3113681300.00 0.014
2017-07-13 10.14 10.74 10.08 10.74 299453430 3180144600.00 0.014
2017-07-14 10.65 10.78 10.50 10.74 172257030 1864449200.00 0.008
2017-07-17 10.79 11.17 10.56 10.65 327312320 3608692300.00 0.016
2017-07-18 10.59 10.98 10.46 10.89 234943190 2558434000.00 0.011
2017-07-19 10.83 11.03 10.72 10.93 193307590 2131336300.00 0.009
2017-07-20 10.92 11.06 10.75 10.81 153733860 1695061000.00 0.007
2017-07-21 10.83 10.95 10.69 10.89 150102000 1625416400.00 0.007
2017-07-24 10.82 11.06 10.73 10.95 169266440 1846886700.00 0.008
2017-07-25 10.98 11.27 10.95 11.00 195476840 2172114600.00 0.009
2017-07-26 10.92 11.18 10.66 10.74 169741210 1846282200.00 0.008
2017-07-27 10.72 10.77 10.53 10.59 119449040 1273888890.00 0.006
2017-07-28 10.61 10.81 10.58 10.74 81919535 877769280.00 0.004
2017-07-31 10.80 10.82 10.45 10.67 157586440 1671814000.00 0.008
2017-08-01 10.64 11.08 10.60 11.04 203570990 2222887900.00 0.010
[629 rows x 7 columns]
1.46 数据源更换 QA_fetch_get_stock_day('ts')(0.4.0-beta-dev30,dev31,dev32)
2017/8/14
把tushare的数据源更换成腾讯网的
dev31 更新 增加retry次数,增加暂停时间到0.005秒
dev32 更新 retry200次
1.47 新增数据库api QA_util_mongo_initial,QA_util_mongo_make_index(0.4.0-beta-dev33)
2017/8/14
QA_util_mongo_initial()
删除数据库数据文件
QA_util_mongo_make_index()
对于日线和分钟线建立索引
也可以在quantaxis cli中使用
QUANTAXIS> drop_database
QUANATXIS> make_index
1.48 修复一个委托单超过市场上下限 返回none的bug(0.4.0-beta-dev34)
2017/8/14
现在返回的是market 400状态
1.49 修复了pytdx的get_k_data api(0.4.0-beta-dev35)
2017/8/15
之前这个api 因为停牌原因会导致时间索引错位
现在已经修复
In [2]: QA.QA_fetch_get_stock_day('tdx','000001','1997-01-21','1997-03-21')
Out[2]:
open close high low vol amount
date
1997-01-21 16.82 16.98 17.20 16.80 131296.0 2.229589e+08
1997-01-22 17.20 17.49 17.58 17.02 232687.0 4.047471e+08
1997-01-23 17.50 18.13 18.40 17.50 264847.0 4.773427e+08
1997-01-24 18.20 18.05 18.24 17.68 169601.0 3.046991e+08
1997-01-27 18.09 18.04 18.23 17.99 154886.0 2.805207e+08
1997-01-28 18.00 18.25 18.35 18.00 110507.0 2.011880e+08
1997-01-29 18.26 18.81 19.09 18.24 200763.0 3.763796e+08
1997-01-30 19.01 18.98 19.12 18.66 125712.0 2.374635e+08
1997-01-31 19.00 19.29 19.50 18.90 271490.0 5.232251e+08
1997-02-17 18.98 19.30 19.55 18.73 179070.0 3.419574e+08
1997-02-18 19.00 17.61 19.09 17.30 636572.0 1.145977e+09
1997-02-19 18.00 18.91 18.95 18.00 454504.0 8.426491e+08
1997-02-20 17.03 18.88 19.62 17.03 629958.0 1.170372e+09
1997-02-21 18.80 19.00 19.10 18.68 228541.0 4.309114e+08
1997-02-24 19.10 18.89 19.23 18.87 142884.0 2.713711e+08
1997-02-25 19.00 19.39 19.78 18.83 284111.0 5.501372e+08
1997-02-26 19.60 19.54 19.69 19.35 129555.0 2.524559e+08
1997-02-27 19.50 19.30 19.60 19.20 158984.0 3.085820e+08
1997-02-28 19.25 19.20 19.28 18.90 149898.0 2.866645e+08
1997-03-03 19.28 19.36 19.56 19.18 122863.0 2.377923e+08
1997-03-04 19.33 19.15 19.33 19.00 156856.0 3.004253e+08
1997-03-05 19.18 19.01 19.32 19.00 114645.0 2.188174e+08
1997-03-06 18.98 19.05 19.10 18.89 95569.0 1.818445e+08
1997-03-07 19.08 19.12 19.35 19.00 98483.0 1.882669e+08
1997-03-10 19.20 19.44 19.63 19.15 128497.0 2.496492e+08
1997-03-11 19.45 19.62 19.65 19.27 183575.0 3.580383e+08
1997-03-12 19.70 20.05 20.20 19.64 251126.0 5.020225e+08
1997-03-13 20.10 19.89 20.30 19.75 178583.0 3.577445e+08
1997-03-14 19.88 20.33 20.39 19.60 171159.0 3.415037e+08
1997-03-17 20.80 22.30 22.37 20.80 465572.0 1.015967e+09
1997-03-18 22.50 22.55 23.45 21.90 439436.0 9.973509e+08
1997-03-19 22.60 23.40 23.75 22.45 423729.0 9.799536e+08
1997-03-20 23.51 23.18 24.24 23.00 318485.0 7.548200e+08
1997-03-21 23.01 24.40 24.45 23.01 333392.0 7.951931e+08
巨大改动/重构
2.1 QA.QAARP.QAAccount
在0.3.9-gamma中,quantaxis对于account账户方法进行了重构.优化了回测逻辑以及数组的存储方式.
现在的account只有如下几个变量:
- assets: 总资金曲线
- hold: 持仓列表
- history : 历史交易列表
- cash: 历史现金列表
- detail: 买卖明细表(有买入id和卖出id对应关系)
新版的quantaxis并不在回测框架中定义利润的计算以及其他的逻辑,这些在backtest_analysis中会涉及计算,当然也可以自己定义利润的计算方法
2.2 QA.QABacktest.Backtest_analysis
对于quantaxis_backtest_analysis进行了巨大的修改,现在的backtest_analysis的接口调用函数有了一定程度的修改.
对于交易组合而言,quantaxis的backtest_analysis 可以直接对于组合进行分析,计算组合的情况
增加了可以自己选定组合的benchmark标的
2.3 QA.QABacktest.QABacktest
对于QABacktest进行了彻底的重构,重新写入了读取数据方法和函数句柄的接受方法,具体参见test/new test
2.4 QA.QA_Queue
事件队列,参见 ###1.14
2.5 彻底放弃QUANTAXIS-WEBKIT-CLIENT
webkit/client 是一个基于electronic的客户端,但是其功能本质上和网页版并无区别,且缺乏维护
0.4.0-alpha中将其去除,之后也不会再维护
2.6 更换QABACKTEST的回测模式
参见 ###1.34
import QUANTAXIS as QA
from QUANTAXIS import QA_Backtest_stock_day as QB
"""
写在前面:
===============QUANTAXIS BACKTEST STOCK_DAY中的变量
常量:
QB.account.message 当前账户消息
QB.account.cash 当前可用资金
QB.account.hold 当前账户持仓
QB.account.history 当前账户的历史交易记录
QB.account.assets 当前账户总资产
QB.account.detail 当前账户的交易对账单
QB.account.init_assest 账户的最初资金
QB.strategy_stock_list 回测初始化的时候 输入的一个回测标的
QB.strategy_start_date 回测的开始时间
QB.strategy_end_date 回测的结束时间
QB.today 在策略里面代表策略执行时的日期
QB.benchmark_code 策略业绩评价的对照行情
函数:
获取市场(基于gap)行情:
QB.QA_backtest_get_market_data(QB,code,QB.today)
获取市场自定义时间段行情:
QA.QA_fetch_stock_day(code,start,end,model)
报单:
QB.QA_backtest_send_order(QB, code,amount,towards,order: dict)
order有三种方式:
1.限价成交 order['bid_model']=0或者l,L
注意: 限价成交需要给出价格:
order['price']=xxxx
2.市价成交 order['bid_model']=1或者m,M,market,Market
3.严格成交模式 order['bid_model']=2或者s,S
及 买入按bar的最高价成交 卖出按bar的最低价成交
查询当前一只股票的持仓量
QB.QA_backtest_hold_amount(QB,code)
"""
@QB.backtest_init
def init():
#
QB.setting.QA_util_sql_mongo_ip='192.168.4.189'
QB.account.init_assest=2500000
QB.benchmark_code='hs300'
QB.strategy_stock_list=['000001','000002','600010','601801','603111']
QB.strategy_start_date='2017-03-01'
QB.strategy_end_date='2017-07-01'
@QB.before_backtest
def before_backtest():
global risk_position
QA.QA_util_log_info(QB.account.message)
@QB.load_strategy
def strategy():
print(QB.account.message)
print(QB.account.cash)
input()
for item in QB.strategy_stock_list:
if QB.QA_backtest_hold_amount(QB,item)==0:
#获取数据的第一种办法[这个是根据回测时制定的股票列表初始化的数据]
QB.QA_backtest_send_order(QB,item,10000,1,{'bid_model':'Market'})
else:
print(QB.QA_backtest_hold_amount(QB,item))
QB.QA_backtest_send_order(QB,item,10000,-1,{'bid_model':'Market'})
@QB.end_backtest
def after_backtest():
pass
2.7 数据源更换(0.4.0-beta-dev30)
对比了通达信,同花顺,ifeng,腾讯财经的数据源
前复权:
====================腾讯======================
600340
2009-09-02
http://web.ifzq.gtimg.cn/appstock/app/fqkline/get?_var=kline_dayqfq2009¶m=sh600340,day,2009-01-01,2010-12-31,640,qfq&r=0.18617627655340497
[{'date': '2009-09-02', 'open': 1.264, 'close': 1.293, 'high': 1.303, 'low': 1.264, 'volume': 30684.29, 'code': '600340', 'date_stamp': 1251820800.0, 'fqtype': 'qfq'}, {'date': '2009-09-03', 'open': 1.307, 'close': 1.357, 'high': 1.357, 'low': 1.307, 'volume': 40504.84, 'code': '600340', 'date_stamp': 1251907200.0, 'fqtype': 'qfq'}, {'date': '2009-09-04', 'open': 1.423, 'close': 1.425, 'high': 1.425, 'low': 1.386, 'volume': 52082.41, 'code': '600340', 'date_stamp': 1251993600.0, 'fqtype': 'qfq'}, {'date': '2009-09-07', 'open': 1.462, 'close': 1.471, 'high': 1.497, 'low': 1.425, 'volume': 69120.01, 'code': '600340', 'date_stamp': 1252252800.0, 'fqtype': 'qfq'}, {'date': '2009-09-09', 'open': 1.544, 'close': 1.544, 'high': 1.544, 'low': 1.544, 'volume': 2528.08, 'code': '600340', 'date_stamp': 1252425600.0, 'fqtype': 'qfq'}, {'date': '2009-09-10', 'open': 1.621, 'close': 1.621, 'high': 1.621, 'low': 1.592, 'volume': 65376.57, 'code': '600340', 'date_stamp': 1252512000.0, 'fqtype': 'qfq'}, {'date': '2009-09-11', 'open': 1.694, 'close': 1.633, 'high': 1.694, 'low': 1.615, 'volume': 105574.09, 'code': '600340', 'date_stamp': 1252598400.0, 'fqtype': 'qfq'}, {'date': '2009-09-14', 'open': 1.619, 'close': 1.715, 'high': 1.715, 'low': 1.593, 'volume': 90609.98, 'code': '600340', 'date_stamp': 1252857600.0, 'fqtype': 'qfq'}, {'date': '2009-09-15', 'open': 1.743, 'close': 1.73, 'high': 1.782, 'low': 1.707, 'volume': 45260.01, 'code': '600340', 'date_stamp': 1252944000.0, 'fqtype': 'qfq'}, {'date': '2009-09-16', 'open': 1.702, 'close': 1.733, 'high': 1.747, 'low': 1.662, 'volume': 48149.54, 'code': '600340', 'date_stamp': 1253030400.0, 'fqtype': 'qfq'}]
====================ths同花顺======================
600340
2009-09-02
open high low close volume amount factor
date
2009-09-02 -0.48 -0.44 -0.48 -0.45 3068429 30789151.00 15.202
2009-09-03 -0.43 -0.38 -0.43 -0.38 4050484 42518565.00 20.068
2009-09-04 -0.31 -0.30 -0.35 -0.30 5208241 57345823.00 25.804
2009-09-07 -0.26 -0.23 -0.30 -0.25 6912001 79692366.00 34.245
2009-09-09 -0.18 -0.18 -0.18 -0.18 252808 3041280.00 1.253
2009-09-10 -0.09 -0.09 -0.12 -0.09 6537657 82537179.00 32.390
2009-09-11 -0.01 -0.01 -0.10 -0.08 10557409 135804180.00 52.306
2009-09-14 -0.10 0.01 -0.12 0.01 9060998 117785149.00 44.892
2009-09-15 0.04 0.08 0.00 0.03 4526001 61421440.00 22.424
2009-09-16 -0.00 0.04 -0.05 0.03 4814954 63716609.00 23.855
In [3]: =======凤凰网=====
ts.get_h_data('600340','2009-09-02','2009-09-16',autype='qfq')
[Getting data:]Out[3]:
open high close low volume amount
date
2009-09-16 39.29 40.33 40.00 38.37 4814954.0 63716608.0
2009-09-15 40.24 41.13 39.94 39.41 4526001.0 61421440.0
2009-09-14 37.37 39.59 39.59 36.77 9060998.0 117785152.0
2009-09-11 39.11 39.11 37.69 37.28 10557409.0 135804192.0
2009-09-10 37.43 37.43 37.43 36.74 6537657.0 82537176.0
2009-09-09 35.65 35.65 35.65 35.65 252808.0 3041280.0
2009-09-07 33.75 34.55 33.96 32.89 6912001.0 79692368.0
2009-09-04 32.86 32.89 32.89 32.00 5208241.0 57345824.0
2009-09-03 30.17 31.32 31.32 30.17 4050484.0 42518564.0
2009-09-02 29.19 30.08 29.84 29.19 3068429.0 30789152.0
对比发现,ifeng,同花顺,通达信(软件)的复权方法有问题,最终采用wind/腾讯的数据作为数据源
缺点是这样没有turnover数据
重要性能优化 重新定义回测流程,减少数据库IO压力
在新的回测框架中,大幅优化了数据的读取方式,通过大量的内存结构来进行数据缓存,之后的数据调用请求都通过内存中的数据接口来获得,这样大大减少了数据库IO
同时,通过对于Account账户的修改,大幅优化了回测数据的存储方式,以及存储的的模式,大大减少了回测数据存储的数量
网友评论