FMZは Python ロカルのバックテストエンジンを開始しました

FMZは Python ロカルのバックテストエンジンを開始しました

FMZ バックテスト エンジン Python パッケージ python2とpython3に対応し,Windows,Linux,Mac OSシステムをサポートしています.

インストール

次のコマンドをコマンド行に入力します.

pip install https://github.com/fmzquant/backtest_python/archive/master.zip

• 注記: Mac OS システムにエンジンをインストールする際に,セキュリティ制限がある場合は,pipの前に sudo を追加し,インストールコマンドが実行される前にシステムパスワードが要求されます.

シンプル な 例

'''backtest
start: 2018-02-19 00:00:00
end: 2018-03-22 12:00:00
period: 15m
exchanges: [{"eid":"OKEX","currency":"LTC_BTC","balance":3,"stocks":0}]
'''
from fmz import *
task = VCtx(__doc__) # initialize backtest engine from __doc__
print exchange.GetAccount()
print exchange.GetTicker()
print task.Join() # print backtest result

設定文字列は,バックテストの設定を"戦略編集"ページに保存することで自動的に生成できます.

文書

API ドキュメンテーション: (デモで GetAccount のようなドキュメントを呼び出す機能)

• https://www.fmz.com/api

Python コードの簡単なコメント:

'''backtest
start: 2018-02-19 00:00:00
end: 2018-03-22 12:00:00
period: 15m
exchanges: [{"eid":"OKEX","currency":"LTC_BTC","balance":3,"stocks":0}]
'''
from fmz import *           # Call the FMZ Quant library 
task = VCtx(__doc__)        # Initialize the backtest engine from __doc__ by calling VCtx
print exchange.GetAccount() # Test GetAccount function, and print the account information of the backtested platform 
print exchange.GetTicker()  # Test GetTicker function, and print the market quote information of backtest system 
print task.Join()           # Call the initialized task object, and print the backtest result    

• _doc__ について

  With two underscores, "__doc__" is used to access the first unassigned string in modules like class declaration or function declaration; the string can be enclosed by """ ""","" "" or ' ', which is to pass the backtest configuration information of '''backtest ... ''' in the code into the VCtx class constructor to construct objects.
  

• ログとGetTickerを特定して呼び出す方法を見ます.

  # coding=UTF-8
  
  '''backtest
  start: 2018-02-19 00:00:00
  end: 2018-03-22 12:00:00
  period: 15m
  exchanges: [{"eid":"OKEX","currency":"LTC_BTC","balance":3,"stocks":0}]
  '''
  
  from fmz import *                                               # Call the FMZ Quant library
  task = VCtx(__doc__) # initialize backtest engine from __doc__  # Call VCtx function according to the initialization of __doc__
  print exchange.GetAccount()                                     # Test GetAccount and print the account information of the tested platform by the backtest system
  Log("\n call Log")
  Log("Call exchange.GetTicker() : ", exchange.GetTicker())
  print task.Join()                                                # Call the initialized task object, and print the backtest result 
  

  • プリント交換.GetAccount (取得)

  {'Balance': 3.0, 'Stocks': 0.0, 'FrozenBalance': 0.0, 'FrozenStocks': 0.0}
  

  • 呼び出し日記 呼び出し日記 呼び出し交換GetTicker

  プリントされたデータは,プリントタスクの構造にある.Join():

  {
      "Chart": {
   	   "Cfg": "",
   	   "Datas": []
      },
      "Elapsed": 42000000,
      "Finished": true,
      "Indicators": {},
      "LoadBytes": 441845,
      "LoadElapsed": 24000000,
      "LogsCount": 2,
      "Profit": 0.0,
      "ProfitLogs": [],
      "Progress": 100.0,
      "RuntimeLogs": [                                                  # Here call the printed data  
   	   [1, 1518969600200, 5, "", 0, 0.0, 0.0, "\n call Log", "", ""],
   	   [2, 1518969600400, 5, "", 0, 0.0, 0.0, "Call exchange.GetTicker() :  {'Sell': 0.02113476, 'Volume': 519.6953, 'Buy': 0.02113474, 'Last': 0.02113475, 'High': 0.02113476, 'Time': 1518969600000L, 'Low': 0.02113474}", "", ""]
      ],
      "Snapshort": [{
   	   "Balance": 3.0,
   	   "BaseCurrency": "LTC",
   	   "Commission": 0.0,
   	   "FrozenBalance": 0.0,
   	   "FrozenStocks": 0.0,
   	   "Id": "OKEX",
   	   "QuoteCurrency": "BTC",
   	   "Stocks": 0.0,
   	   "Symbols": {
   		   "LTC_BTC_OKEX": {
   			   "Last": 0.01893785
   		   }
   	   },
   	   "TradeStatus": {}
      }],
      "Status": "",
      "Task": {
   	   "Args": null,
   	   "Exchanges": [{
   		   "Balance": 3,
   		   "BaseCurrency": "LTC",
   		   "BasePeriod": 300000,
   		   "BasePrecision": 4,
   		   "DepthDeep": 5,
   		   "FaultTolerant": 0,
   		   "FeeDenominator": 5,
   		   "FeeMaker": 75,
   		   "FeeMin": 0,
   		   "FeeTaker": 80,
   		   "Id": "OKEX",
   		   "Label": "OKEX",
   		   "PriceTick": 1e-08,
   		   "QuoteCurrency": "BTC",
   		   "QuotePrecision": 8,
   		   "SlipPoint": 0,
   		   "Stocks": 0
   	   }],
   	   "Options": {
   		   "DataServer": "q.botvs.net",
   		   "MaxChartLogs": 800,
   		   "MaxProfitLogs": 800,
   		   "MaxRuntimeLogs": 800,
   		   "NetDelay": 200,
   		   "Period": 900000,
   		   "RetFlags": 189,
   		   "SnapshortPeriod": 300000,
   		   "TimeBegin": 1518969600,
   		   "TimeEnd": 1521691200,
   		   "UpdatePeriod": 5000
   	   }
      },
      "TaskStatus": 1,
      "Time": 1521691200000
  }
  

• ローカルバックテストエンジンでバックテストする戦略をどのように使うか?

# !/usr/local/bin/python
# -*- coding: UTF-8 -*-
'''backtest
start: 2018-02-19 00:00:00
end: 2018-03-22 12:00:00
period: 15m
exchanges: [{"eid":"Bitfinex","currency":"BTC_USD","balance":10000,"stocks":3}]
'''
import sys
sys.path.append("/usr/local/lib/python2.7/site-packages")    # Add a path during the backtest, and you can delete it if not needed 

from fmz import *
import math
import talib

task = VCtx(__doc__) # initialize backtest engine from __doc__

# ------------------------------ Strategy Section Starts --------------------------
print exchange.GetAccount()     # Call some interfaces, and print their return values 
print exchange.GetTicker()

def adjustFloat(v):             # Custom functions in the strategy 
    v = math.floor(v * 1000)
    return v / 1000

def onTick(e):
    Log("onTick")
    # ....

#
# ...
# 
# Here the codes including the implementation of custom functions are omitted 

def main():
    InitAccount = GetAccount()
    
    while True:
        onTick(exchange)
        Sleep(1000)
# ------------------------------ Strategy Section Ends --------------------------

try:
    main()                     # Raise EOFError() when the backtest ends, to stop the backtest loop. Therefore, the error needs to be processed by calling task.Join() when the error is detected, and print the backtest result 
except:
    print task.Join()          # print the backtest result