Movatterモバイル変換

[0]ホーム
URL:

Skip to content

TA-Lib

Even ifbacktrader offers an already high number of built-in indicators anddeveloping an indicator is mostly a matter of defining the inputs, outputs andwriting the formula in a natural manner, some people want to useTA-LIB. Someof the reasons:

  • IndicatorX is in the library and not inbacktrader (the author would gladly accept a request)

  • TA-LIB behavior is well known and people trust good old things

In order to satisfy each and every taste,TA-LIB integration is offered.

Requirements

The installation details are on theGitHub repository

Usingta-lib

As easy as using any of the indicators already built-in inbacktrader. Example of aSimple Moving Average. First thebacktraderone:

importbacktraderasbtclassMyStrategy(bt.Strategy):params=(('period',20),)def__init__(self):self.sma=bt.indicators.SMA(self.data,period=self.p.period)......

Now theta-lib example:

importbacktraderasbtclassMyStrategy(bt.Strategy):params=(('period',20),)def__init__(self):self.sma=bt.talib.SMA(self.data,timeperiod=self.p.period)......

Et voilá! Of course theparams for theta-lib indicators are defined by thelibrary itself and not bybacktrader. In this case theSMA inta-libtakes a parameter namedtimeperiod to defined the size of the operatingwindow.

For indicators that require more than one input, for example theStochastic:

importbacktraderasbtclassMyStrategy(bt.Strategy):params=(('period',20),)def__init__(self):self.stoc=bt.talib.STOCH(self.data.high,self.data.low,self.data.close,fastk_period=14,slowk_period=3,slowd_period=3)......

Notice howhigh,low andclose have been individually passed. Onecould always passopen instead oflow (or any other data series) andexperiment.

Theta-lib indicator documentation is automatically parsed and added to thebacktrader docs. You may also check theta-lib source code/docs. Oradittionally do:

print(bt.talib.SMA.__doc__)

Which in this case outputs:

SMA([input_arrays],[timeperiod=30])SimpleMovingAverage(OverlapStudies)Inputs:price:(anyndarray)Parameters:timeperiod:30Outputs:real

Which offers some information:

  • WhichInput is to be expected (DISREGARD the ``ndarray`` comment because backtrader manages the conversions in the background)

  • Whichparameters and which default values

  • Which outputlines the indicator actually offers

Moving Averages and MA_Type

To select a specificmoving average for indicators likebt.talib.STOCH,the standardta-libMA_Type is accesible withbacktrader.talib.MA_Type. For example:

importbacktraderasbtprint('SMA:',bt.talib.MA_Type.SMA)print('T3:',bt.talib.MA_Type.T3)

Plotting ta-lib indicators

Just as with regular usage, there is nothing special to do to plot theta-libindicators.

Note

Indicators which output aCANDLE (all those looking for acandlestick pattern) deliver a binary output: either 0 or 100. Inorder to avoid adding asubplot to the chart, there is anautomated plotting translation to plot them over thedata at thepoint in time in which the pattern was recognized.

Examples and comparisons

The following are plots comparing the outputs of someta-lib indicatorsagainst the equivalent built-in indicators inbacktrader. To consider:

  • Theta-lib indicators get aTA_ prefix on the plot. This is specifically done by the sample to help the user spot which is which

  • Moving Averages (if both deliver the same result) will be plottedON top of the other existingMoving Average. The two indicators cannot be seen separately and the test is a pass if that’s the case.

  • All samples include aCDLDOJI indicator as a reference

KAMA (Kaufman Moving Average)

This is the 1st example because it is the only (from all indicators which thesample directly compare) that has a difference:

  • The initial values of the the samples are not the same

  • At some point in time, the values converge and bothKAMA implementations have the same behavior.

After having analyzed theta-lib source code:

  • The implementation inta-lib makes a non-industry standard choice for the 1st values of theKAMA.

    The choice can be seen in the source code quoting from the source code):The yesterday price is used here as the previous KAMA.

backtrader does the usual choice which is the same as for example the onefromStockcharts:

  • KAMA at StockCharts

    Since we need an initial value to start the calculation, the first KAMA isjust a simple moving average

Hence the difference. Furthermore:

  • Theta-libKAMA implementation doesn’t allow specifying thefast andslow periods for the adjustment of thescalable constant defined byKaufman.

Sample execution:

$ ./talibtest.py --plot --ind kama

Output

image

SMA

$ ./talibtest.py --plot --ind sma

Output

image

EMA

$ ./talibtest.py --plot --ind ema

Output

image

Stochastic

$ ./talibtest.py --plot --ind stoc

Output

image

RSI

$ ./talibtest.py --plot --ind rsi

Output

image

MACD

$ ./talibtest.py --plot --ind macd

Output

image

Bollinger Bands

$ ./talibtest.py --plot --ind bollinger

Output

image

AROON

Note thatta-lib chooses to put thedown line first and the colours areinverted when compared with thebacktrader built-in indicator.

$ ./talibtest.py --plot --ind aroon

Output

image

Ultimate Oscillator

$ ./talibtest.py --plot --ind ultimate

Output

image

Trix

$ ./talibtest.py --plot --ind trix

Output

image

ADXR

Herebacktrader offers both theADX andADXR lines.

$ ./talibtest.py --plot --ind adxr

Output

image

DEMA

$ ./talibtest.py --plot --ind dema

Output

image

TEMA

$ ./talibtest.py --plot --ind tema

Output

image

PPO

Herebacktrader offers not only theppo line, but a more traditionalmacd approach.

$ ./talibtest.py --plot --ind ppo

Output

image

WilliamsR

$ ./talibtest.py --plot --ind williamsr

Output

image

ROC

All indicators show have exactly the same shape, but how to trackmomentum orrate of change has several definitions

$ ./talibtest.py --plot --ind roc

Output

image

Sample Usage

$ ./talibtest.py --helpusage: talibtest.py [-h] [--data0 DATA0] [--fromdate FROMDATE]                    [--todate TODATE]                    [--ind {sma,ema,stoc,rsi,macd,bollinger,aroon,ultimate,trix,kama,adxr,dema,tema,ppo,williamsr,roc}]                    [--no-doji] [--use-next] [--plot [kwargs]]Sample for ta-liboptional arguments:  -h, --help            show this help message and exit  --data0 DATA0         Data to be read in (default:                        ../../datas/yhoo-1996-2015.txt)  --fromdate FROMDATE   Starting date in YYYY-MM-DD format (default:                        2005-01-01)  --todate TODATE       Ending date in YYYY-MM-DD format (default: 2006-12-31)  --ind {sma,ema,stoc,rsi,macd,bollinger,aroon,ultimate,trix,kama,adxr,dema,tema,ppo,williamsr,roc}                        Which indicator pair to show together (default: sma)  --no-doji             Remove Doji CandleStick pattern checker (default:                        False)  --use-next            Use next (step by step) instead of once (batch)                        (default: False)  --plot [kwargs], -p [kwargs]                        Plot the read data applying any kwargs passed For                        example (escape the quotes if needed): --plot                        style="candle" (to plot candles) (default: None)

Sample Code

from__future__import(absolute_import,division,print_function,unicode_literals)importargparseimportdatetimeimportbacktraderasbtclassTALibStrategy(bt.Strategy):params=(('ind','sma'),('doji',True),)INDS=['sma','ema','stoc','rsi','macd','bollinger','aroon','ultimate','trix','kama','adxr','dema','ppo','tema','roc','williamsr']def__init__(self):ifself.p.doji:bt.talib.CDLDOJI(self.data.open,self.data.high,self.data.low,self.data.close)ifself.p.ind=='sma':bt.talib.SMA(self.data.close,timeperiod=25,plotname='TA_SMA')bt.indicators.SMA(self.data,period=25)elifself.p.ind=='ema':bt.talib.EMA(timeperiod=25,plotname='TA_SMA')bt.indicators.EMA(period=25)elifself.p.ind=='stoc':bt.talib.STOCH(self.data.high,self.data.low,self.data.close,fastk_period=14,slowk_period=3,slowd_period=3,plotname='TA_STOCH')bt.indicators.Stochastic(self.data)elifself.p.ind=='macd':bt.talib.MACD(self.data,plotname='TA_MACD')bt.indicators.MACD(self.data)bt.indicators.MACDHisto(self.data)elifself.p.ind=='bollinger':bt.talib.BBANDS(self.data,timeperiod=25,plotname='TA_BBANDS')bt.indicators.BollingerBands(self.data,period=25)elifself.p.ind=='rsi':bt.talib.RSI(self.data,plotname='TA_RSI')bt.indicators.RSI(self.data)elifself.p.ind=='aroon':bt.talib.AROON(self.data.high,self.data.low,plotname='TA_AROON')bt.indicators.AroonIndicator(self.data)elifself.p.ind=='ultimate':bt.talib.ULTOSC(self.data.high,self.data.low,self.data.close,plotname='TA_ULTOSC')bt.indicators.UltimateOscillator(self.data)elifself.p.ind=='trix':bt.talib.TRIX(self.data,timeperiod=25,plotname='TA_TRIX')bt.indicators.Trix(self.data,period=25)elifself.p.ind=='adxr':bt.talib.ADXR(self.data.high,self.data.low,self.data.close,plotname='TA_ADXR')bt.indicators.ADXR(self.data)elifself.p.ind=='kama':bt.talib.KAMA(self.data,timeperiod=25,plotname='TA_KAMA')bt.indicators.KAMA(self.data,period=25)elifself.p.ind=='dema':bt.talib.DEMA(self.data,timeperiod=25,plotname='TA_DEMA')bt.indicators.DEMA(self.data,period=25)elifself.p.ind=='ppo':bt.talib.PPO(self.data,plotname='TA_PPO')bt.indicators.PPO(self.data,_movav=bt.indicators.SMA)elifself.p.ind=='tema':bt.talib.TEMA(self.data,timeperiod=25,plotname='TA_TEMA')bt.indicators.TEMA(self.data,period=25)elifself.p.ind=='roc':bt.talib.ROC(self.data,timeperiod=12,plotname='TA_ROC')bt.talib.ROCP(self.data,timeperiod=12,plotname='TA_ROCP')bt.talib.ROCR(self.data,timeperiod=12,plotname='TA_ROCR')bt.talib.ROCR100(self.data,timeperiod=12,plotname='TA_ROCR100')bt.indicators.ROC(self.data,period=12)bt.indicators.Momentum(self.data,period=12)bt.indicators.MomentumOscillator(self.data,period=12)elifself.p.ind=='williamsr':bt.talib.WILLR(self.data.high,self.data.low,self.data.close,plotname='TA_WILLR')bt.indicators.WilliamsR(self.data)defrunstrat(args=None):args=parse_args(args)cerebro=bt.Cerebro()dkwargs=dict()ifargs.fromdate:fromdate=datetime.datetime.strptime(args.fromdate,'%Y-%m-%d')dkwargs['fromdate']=fromdateifargs.todate:todate=datetime.datetime.strptime(args.todate,'%Y-%m-%d')dkwargs['todate']=todatedata0=bt.feeds.YahooFinanceCSVData(dataname=args.data0,**dkwargs)cerebro.adddata(data0)cerebro.addstrategy(TALibStrategy,ind=args.ind,doji=notargs.no_doji)cerebro.run(runcone=notargs.use_next,stdstats=False)ifargs.plot:pkwargs=dict(style='candle')ifargs.plotisnotTrue:# evals to True but is not Truenpkwargs=eval('dict('+args.plot+')')# args were passedpkwargs.update(npkwargs)cerebro.plot(**pkwargs)defparse_args(pargs=None):parser=argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter,description='Sample for sizer')parser.add_argument('--data0',required=False,default='../../datas/yhoo-1996-2015.txt',help='Data to be read in')parser.add_argument('--fromdate',required=False,default='2005-01-01',help='Starting date in YYYY-MM-DD format')parser.add_argument('--todate',required=False,default='2006-12-31',help='Ending date in YYYY-MM-DD format')parser.add_argument('--ind',required=False,action='store',default=TALibStrategy.INDS[0],choices=TALibStrategy.INDS,help=('Which indicator pair to show together'))parser.add_argument('--no-doji',required=False,action='store_true',help=('Remove Doji CandleStick pattern checker'))parser.add_argument('--use-next',required=False,action='store_true',help=('Use next (step by step) ''instead of once (batch)'))# Plot optionsparser.add_argument('--plot','-p',nargs='?',required=False,metavar='kwargs',const=True,help=('Plot the read data applying any kwargs passed\n''\n''For example (escape the quotes if needed):\n''\n''  --plot style="candle" (to plot candles)\n'))ifpargsisnotNone:returnparser.parse_args(pargs)returnparser.parse_args()if__name__=='__main__':runstrat()
[8]ページ先頭
©2009-2025 Movatter.jp