Hi,

I am trying to couple some data passed to the indicator i.e. coupling based on the clock created from the first data passed. The following works in the sense that signal2 is the coupled version to the clock based on datas[0]. However, I guess there must be a better way to achieve this i.e. without creating signal1 that is not needed.

Thanks

class NaiveIndicator(bt.Indicator):
    lines = ('signal1','signal2',)

    def __init__(self):
        self.lines.signal1 = self.datas[1]
        self.lines.signal2 = self.lines.signal1()

...

cerebro.addindicator(NaiveIndicator, cerebro.datasbyname["Data0"], cerebro.datasbyname["Data1"], cerebro.datasbyname["Data2"])

Hi,

is there some way to apply a data filter to multiple data feeds? For example to create a data filter that skips some bars based on information from all data feeds added to the strategy. I guess I could create a single data feed that combines all data or include the logic in next function of the strategy. However, is there some other way I am missing?

Thanks

Many thanks again for your prompt reply and help. However, I am confused since I actually got this without any errors apart from the ones mentioned.

Isn't self.datas[0].lines.close equivalent to self.data.close(0)? I also understand that bt.If is redundant but I fail to see why these two lines would not produce the same results.

self.lines.signal = bt.If(self.datas[0].lines.close > self.params.threshold, 1, 0)
self.lines.signal = self.data.close(0) > self.p.threshold

I mean in a similar fashion I used

self.lines.signal = bt.If(self.datas[0].lines.close > self.datas[0].lines.close(-self.params.period), 1, 0)

instead of

self.lines.signal = self.data.close(0) > self.data.close(-self.p.period)

Thanks

Hi,

When I am adding an indicator (TestIndicator1 below) and I have two strategies, then the indicator values in the second strategy seem to be wrong. However, if I have a single strategy or next instead of init used for indicator, the values are correct.

It seems though that this behaviour does not apply to all indicators. For example values for TestIndicator2 below are correct for both strategies.

import backtrader as bt
import sys
import pandas as pd


class TestStrategy(bt.Strategy):

    params = dict(name="TestStrategy")

    def __init__(self):
        self._indicators = None

    def start(self):
        self._indicators = self.getindicators_lines()

    def next(self):
        print("{}: {}, {}, {}, {}".format(self.params.name, self.datas[0].datetime.date(0), self.datas[0].lines.close[0], self._indicators[0].lines[0][0], self._indicators[1].lines[0][0]))


class TestIndicator1(bt.Indicator):

    lines = ('signal',)
    params=dict(threshold=30)

    def __init__(self):
        self.lines.signal = bt.If(self.datas[0].lines.close > self.params.threshold, 1, 0)

    # def next(self):
    #     if self.datas[0].lines.close[0] > self.params.threshold:
    #         self.lines.signal[0] = 1
    #     else:
    #         self.lines.signal[0] = 0


class TestIndicator2(bt.Indicator):

    lines = ('signal',)
    params=dict(period=1)

    def __init__(self):
        self.addminperiod(self.params.period)
        self.lines.signal = bt.If(self.datas[0].lines.close > self.datas[0].lines.close(-self.params.period), 1, 0)


def main(argv):

    cerebro = bt.Cerebro()
    df1 = pd.DataFrame({'Close': [1,2,3,4,100,5,7,6], 'Open': [0,1,2,3,99,4,6,5], 'High': [0,0,0,1,-1,0,0,0]}, index=pd.date_range(start='1/1/2018', periods=8))
    cerebro.adddata(bt.feeds.PandasData(dataname=df1, nocase=True))
    cerebro.broker.setcash(100000.0)
    cerebro.broker.set_coc(True)

    cerebro.addstrategy(TestStrategy, name="Strat1")
    cerebro.addstrategy(TestStrategy, name="Strat2")

    cerebro.addindicator(TestIndicator1, threshold=4)
    cerebro.addindicator(TestIndicator2, period=1)

    cerebro.run(runonce=True, preload=True)


if __name__ == "__main__":
    main(sys.argv)

Output is

Strat1: 2018-01-02, 2.0, 0.0, 1.0
Strat2: 2018-01-02, 2.0, 1.0, 1.0
Strat1: 2018-01-03, 3.0, 0.0, 1.0
Strat2: 2018-01-03, 3.0, 1.0, 1.0
Strat1: 2018-01-04, 4.0, 0.0, 1.0
Strat2: 2018-01-04, 4.0, 1.0, 1.0
Strat1: 2018-01-05, 100.0, 1.0, 1.0
Strat2: 2018-01-05, 100.0, 1.0, 1.0
Strat1: 2018-01-06, 5.0, 1.0, 0.0
Strat2: 2018-01-06, 5.0, 1.0, 0.0
Strat1: 2018-01-07, 7.0, 1.0, 1.0
Strat2: 2018-01-07, 7.0, 1.0, 1.0
Strat1: 2018-01-08, 6.0, 1.0, 0.0
Strat2: 2018-01-08, 6.0, 1.0, 0.0

while with next I get the expected one:

Strat1: 2018-01-02, 2.0, 0.0, 1.0
Strat2: 2018-01-02, 2.0, 0.0, 1.0
Strat1: 2018-01-03, 3.0, 0.0, 1.0
Strat2: 2018-01-03, 3.0, 0.0, 1.0
Strat1: 2018-01-04, 4.0, 0.0, 1.0
Strat2: 2018-01-04, 4.0, 0.0, 1.0
Strat1: 2018-01-05, 100.0, 1.0, 1.0
Strat2: 2018-01-05, 100.0, 1.0, 1.0
Strat1: 2018-01-06, 5.0, 1.0, 0.0
Strat2: 2018-01-06, 5.0, 1.0, 0.0
Strat1: 2018-01-07, 7.0, 1.0, 1.0
Strat2: 2018-01-07, 7.0, 1.0, 1.0
Strat1: 2018-01-08, 6.0, 1.0, 0.0
Strat2: 2018-01-08, 6.0, 1.0, 0.0

Thanks for your help

Hi,

I am trying to access broker params values from within the strategy. Depending on they way I try to get the coc value, sometimes I get the actual value and other the default value. Can someone please explain the behaviour?

Many thanks

from __future__ import (absolute_import, division, print_function,
                        unicode_literals)
import backtrader as bt
import pandas as pd
import datetime

# Create a Strategy
class TestStrategy(bt.Strategy):

    def __init__(self):
        params_dict = self.broker.params._getpairs()
        print(type(self.broker.params))
        print("COC: {}, {}, {}".format(self.broker.params.coc, params_dict['coc'], self.broker.params._get('coc')))


if __name__ == '__main__':
    # Create a cerebro entity
    cerebro = bt.Cerebro()

    # Add a strategy with parameters
    cerebro.addstrategy(TestStrategy)

    df1 = pd.DataFrame({'Close': list(range(1,11))}, index=pd.date_range(start='1/1/2018', periods=10))
    cerebro.adddata(bt.feeds.PandasData(dataname=df1, nocase=True))

    # Set our desired cash start
    cerebro.broker.setcash(100000.0)
    cerebro.broker.set_coc(True)

    # Print out the starting conditions
    print('Starting Portfolio Value: %.2f' % cerebro.broker.getvalue())

    # Run over everything
    cerebro.run(runonce=False, preload=False)

    # Print out the final result
    print('Final Portfolio Value: %.2f' % cerebro.broker.getvalue())

Starting Portfolio Value: 100000.00
<class 'backtrader.metabase.AutoInfoClass_BrokerBase_BackBroker'>
COC: True, False, True
Final Portfolio Value: 100000.00

Hi,

I am trying to combine SessionFiller and Resampler but it does not seem to work no matter which order these are added. Independently both work correctly.

The data set (just showing close) is

,close
2018-08-08 09:01:01,4
2018-08-08 09:02:01,5
2018-08-08 09:03:01,6
2018-08-08 09:06:01,7
2018-08-08 09:07:01,8
2018-08-08 09:08:01,9
2018-08-08 09:09:01,10
2018-08-08 09:10:01,11
2018-08-08 09:11:01,12
2018-08-08 09:12:01,13

and I would expect to get something like

2018-08-08 09:02:00, 4.0
2018-08-08 09:04:00, 6.0
2018-08-08 09:06:00, 6.0
2018-08-08 09:08:00, 8.0
2018-08-08 09:10:00, 10.0
2018-08-08 09:12:00, 12.0
2018-08-08 09:14:00, 13.0

Thanks

import backtrader as bt
import sys
import os
import datetime

class StrategyBidAsk(bt.Strategy):

    def start(self):
        self.counter = 0

    def next(self):
        self.counter += 1
        output_str = ""
        for i in range(len(self.datas)):
            output_str += "{}: {}, {}|".format(self.datas[i]._name, self.datas[i].datetime.datetime(0), self.datas[i].lines.close[0])
        output_str = output_str[:-1]
        print(output_str)

def main(argv):

    folder = "data"
    file = 'Toy_Ask_new.csv'
    cash = 100000.0

    cerebro = bt.Cerebro()
    cerebro.broker.setcash(cash)
    cerebro.broker.set_shortcash(False)
    cerebro.broker.setcommission(commission=0)
    cerebro.broker.set_coc(False)

    data = bt.feeds.GenericCSVData(
            dataname=os.path.join(folder, file),
            dtformat='%Y-%m-%d %H:%M:%S', # default
            datetime=0,         # default
            time=-1,            # default
            close=1,
            high=2,
            low=3,
            open=4,
            volume=5,           # default
            openinterest=-1,    # default for not present
            sessionstart=datetime.time(9, 0),
            sessionend=datetime.time(16, 0),
            timeframe=bt.TimeFrame.Minutes,
            compression=1)      # default

    data.addfilter(bt.filters.SessionFilter)
    data.addfilter(bt.filters.SessionFiller)
    data.addfilter(bt.ResamplerMinutes, compression=2)
    cerebro.adddata(data, name=os.path.splitext(file)[0])

    idx = cerebro.addstrategy(StrategyBidAsk)
    cerebro.addsizer_byidx(idx, bt.sizers.SizerFix, stake=100)
    cerebro.run(runonce=False, preload=True)


if __name__ == "__main__":
    main(sys.argv)

Thanks for your reply and help. By system you mean the strategy? Because in this case, strategy will have length 5 while indicator length 4.

# Toy_Ask: 2018-08-08 00:00:01, 3.0|Toy_Bid: 2018-08-08 00:00:01, 5.0
# MidPriceIndicator toy_mid 4.0 4.0 1 1
# Toy_Ask: 2018-08-08 00:02:01, 3.0|Toy_Bid: 2018-08-08 00:02:01, 6.0
# MidPriceIndicator toy_mid 4.5 4.0 2 2
# Toy_Ask: 2018-08-08 00:02:01, 3.0|Toy_Bid: 2018-08-08 00:03:01, 7.0
# MidPriceIndicator toy_mid 5.0 4.5 3 3
# Toy_Ask: 2018-08-08 00:04:01, 0.0|Toy_Bid: 2018-08-08 00:03:01, 7.0
# MidPriceIndicator toy_mid 3.5 4.5 3 4
# Toy_Ask: 2018-08-08 00:08:01, 1.0|Toy_Bid: 2018-08-08 00:08:01, 2.0
# MidPriceIndicator toy_mid 1.5 3.5 4 5

import backtrader as bt
import sys
import os


class StrategyBidAsk(bt.Strategy):

    def next(self):
        output_str = ""
        for i in range(len(self.datas)):
            output_str += "{}: {}, {}|".format(self.datas[i]._name, self.datas[i].datetime.datetime(0), self.datas[i].lines.close[0])
        output_str = output_str[:-1]
        print(output_str)
        for ind in self.getindicators_lines():
            line_names = ind.getlinealiases()
            for idx in range(len(line_names)):
                print("{} {} {} {} {} {}".format(ind.plotinfo.plotname or ind.__class__.__name__,
                                        line_names[idx], ind.lines[idx][0], ind.lines[idx][-1], len(ind), len(self)))


class MidPriceIndicator(bt.Indicator):
    lines = ('toy_mid',)

    def __init__(self):
        self.lines.toy_mid = (self.datas[0]+self.datas[1])/2


def main(argv):

    folder = "data"
    files = [['Toy_Ask.csv', 'Toy_Bid.csv']]
    cash = 100000.0

    cerebro = bt.Cerebro()
    cerebro.broker.setcash(cash)
    cerebro.broker.set_shortcash(False)
    cerebro.broker.setcommission(commission=0)
    cerebro.broker.set_coc(False)

    for file in files:
        if not isinstance(file, list):
            file = [file]
        data_list = []
        for i, f in enumerate(file):
            data = bt.feeds.GenericCSVData(
                    dataname=os.path.join(folder, f),
                    dtformat='%Y-%m-%d %H:%M:%S', # default
                    datetime=0,         # default
                    time=-1,            # default
                    close=1,
                    high=-1,
                    low=-1,
                    open=-1,
                    volume=-1,           # default
                    openinterest=-1,    # default for not present
                    timeframe=bt.TimeFrame.Minutes,
                    compression=1)      # default
            if i != 0 and len(file) == 2:
                data.compensate(data_list[0])
            data_list.append(data)
            cerebro.adddata(data, name=os.path.splitext(f)[0])

    idx = cerebro.addstrategy(StrategyBidAsk)
    cerebro.addsizer_byidx(idx, bt.sizers.SizerFix, stake=100)
    cerebro.addindicator(MidPriceIndicator, cerebro.datasbyname["Toy_Bid"]._getline("close"), cerebro.datasbyname["Toy_Ask"]._getline("close"))
    cerebro.run(runonce=False, preload=True)
    cerebro.plot(volume=False, numfigs=1)


if __name__ == "__main__":
    main(sys.argv)

Hi,

I would like to implement an indicator that first calculates the bid-ask midpoint on two separate minute-bar data feeds, one for bid and the other for ask. These data feeds are not synchronised i.e. there are minutes when one has value and the other does not. Although the value of the indicator is correct, its length is not. When the second feed does not have a minute bar of the first feed, it just overwrites the indicator value.

Many thanks for the help.

import backtrader as bt
import sys
import os


class StrategyBidAsk(bt.Strategy):

    def next(self):
        output_str = ""
        for i in range(len(self.datas)):
            output_str += "{}: {}, {}|".format(self.datas[i]._name, self.datas[i].datetime.datetime(0), self.datas[i].lines.close[0])
        output_str = output_str[:-1]
        print(output_str)
        for ind in self.getindicators_lines():
            line_names = ind.getlinealiases()
            for idx in range(len(line_names)):
                print("{} {} {} {} {}".format(ind.plotinfo.plotname or ind.__class__.__name__,
                                        line_names[idx], ind.lines[idx][0], ind.lines[idx][-1], len(ind)))


class MidPriceIndicator(bt.Indicator):
    lines = ('toy_mid',)

    def __init__(self):
        self.lines.toy_mid = (self.datas[0]+self.datas[1])/2


def main(argv):

    folder = "data"
    files = [['Toy_Ask.csv', 'Toy_Bid.csv']]
    cash = 100000.0

    cerebro = bt.Cerebro()
    cerebro.broker.setcash(cash)
    cerebro.broker.set_shortcash(False)
    cerebro.broker.setcommission(commission=0)
    cerebro.broker.set_coc(False)

    for file in files:
        if not isinstance(file, list):
            file = [file]
        data_list = []
        for i, f in enumerate(file):
            data = bt.feeds.GenericCSVData(
                    dataname=os.path.join(folder, f),
                    dtformat='%Y-%m-%d %H:%M:%S', # default
                    datetime=0,         # default
                    time=-1,            # default
                    close=1,
                    high=-1,
                    low=-1,
                    open=-1,
                    volume=-1,           # default
                    openinterest=-1,    # default for not present
                    timeframe=bt.TimeFrame.Minutes,
                    compression=1)      # default
            if i != 0 and len(file) == 2:
                data.compensate(data_list[0])
            data_list.append(data)
            cerebro.adddata(data, name=os.path.splitext(f)[0])

    idx = cerebro.addstrategy(StrategyBidAsk)
    cerebro.addsizer_byidx(idx, bt.sizers.SizerFix, stake=100)
    cerebro.addindicator(MidPriceIndicator, cerebro.datasbyname["Toy_Bid"]._getline("close"), cerebro.datasbyname["Toy_Ask"]._getline("close"))
    cerebro.run(runonce=False, preload=True)
    cerebro.plot(volume=False, numfigs=1)


if __name__ == "__main__":
    main(sys.argv)

# # Output
# Toy_Ask: 2018-08-08 00:00:01, 3.0|Toy_Bid: 2018-08-08 00:00:01, 5.0
# MidPriceIndicator toy_mid 4.0 4.0 1
# Toy_Ask: 2018-08-08 00:02:01, 3.0|Toy_Bid: 2018-08-08 00:02:01, 6.0
# MidPriceIndicator toy_mid 4.5 4.0 2
# Toy_Ask: 2018-08-08 00:02:01, 3.0|Toy_Bid: 2018-08-08 00:03:01, 7.0
# MidPriceIndicator toy_mid 5.0 4.5 3
# Toy_Ask: 2018-08-08 00:04:01, 0.0|Toy_Bid: 2018-08-08 00:03:01, 7.0
# MidPriceIndicator toy_mid 3.5 4.5 3
# Toy_Ask: 2018-08-08 00:08:01, 1.0|Toy_Bid: 2018-08-08 00:08:01, 2.0
# MidPriceIndicator toy_mid 1.5 3.5 4

Hi,

I understand that the concept of multiple strategies has been discussed in many threads.

If I got it correctly, since there is one broker per cerebro and the broker holds positions by data and not by (data, strategy), strategies won't work independently i.e. whatever functionality is based on position will 'work' on the total position and not just the position of the strategy itself. However, orders and trades seems to be independent per strategy.

I would be interested to have some feedback on how to manage multiple strategies. I guess:

• you can have different cerebro, one per strategy, but then communication/synchronisation won't be easy.

• or have a different sizer per strategy and have each strategy keeping track of its own position, value, etc. I am not sure whether observes, analysers, etc would work on strategy level.

I would highly appreciate any feedback and ideas on this.

Many thanks

Thanks for your prompt reply and help.

I am not questioning whether this is the right behaviour from a design/conceptual point of view.

However, it seems to me that if you want to record the highest close value since the trade opened, it is not possible with a dynamic indicator. Order executed on open of 2018-01-05 (based on signal on previous day), trade opens, p.tradeopen set to True, close on that day is 100 and on 2018-01-06 the dynamic indicator has as highest close value the value of 2018-01-06 which is 5 (instead of 100).

Unless of course, I try something like this

    def notify_trade(self, trade):
        self.dyn_highest.p.tradeopen = trade.isopen
        if trade.justopened:
            self.dyn_highest.lines.dyn_highest[0] = self.data[0]
            print("NOTE: Opening trade at price {} on date {}".format(trade.price, bt.utils.num2date(trade.dtopen)))