I am trying to replicate the behaviour of a full script via the command line with btrun and have the following command:

btrun --csvformat btcsv \
      --data data/gbpusd-1m.csv \
      --strategy strategies/TemplateStrategy.py \
      --fromdate 2003-1-1 \
      --todate 2003-1-5 \
      --timeframe minutes \
      --commission 0 \
      --cash 10000 \
      --compression 1

How would I pass parameters to GenericCSVData such as open, high, low, close, openinterest as below?

    data = bt.feeds.GenericCSVData(
        timeframe=bt.TimeFrame.Minutes,
        compression=1,
        dataname=datafile,
        fromdate=datetime.datetime(2003, 1, 1),
        todate=datetime.datetime(2003, 1, 5),
        dtformat=('%Y-%m-%d %H:%M:%S'),
        datetime=0,
        high=2,
        low=3,
        open=1,
        close=4,
        volume=5,
        openinterest=-1)

Thanks

Ah well now I feel stupid lol. Thanks for pointing that out, I thought I had bought the entire years worth of data but it would appear not!

I'm having trouble outputting data when using fromdate and todate for anything less than 11 months of data from a csv.

It doesn't return/output at all with my test script below unless I set it to 11 months or more.

My test strategy:

from __future__ import (absolute_import, division, print_function,
                        unicode_literals)

import datetime  # For datetime objects
import os.path  # To manage paths
import sys  # To find out the script name (in argv[0])

# Import the backtrader platform
import backtrader as bt

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

    def next(self):
        print(self.data.close[0])


if __name__ == '__main__':

    # Setup Cerebro
    cerebro = bt.Cerebro(tradehistory=True,oldbuysell=True)
    cerebro.broker.setcash(10000.0)
    cerebro.broker.setcommission(commission=0)
    modpath = os.path.dirname(os.path.abspath(sys.argv[0]))
    datafile = os.path.join(modpath, 'data/gbpusd-1m.csv')

    # Add a strategy
    cerebro.addstrategy(TestStrategy)

    # Create a Data Feed
    data = bt.feeds.GenericCSVData(
        timeframe=bt.TimeFrame.Minutes,
        compression=1,
        dataname=datafile,
        fromdate=datetime.datetime(2002, 1, 1),
        todate=datetime.datetime(2002, 6, 1),
        dtformat=('%Y-%m-%d %H:%M:%S'),
        datetime=0,
        high=2,
        low=3,
        open=1,
        close=4,
        volume=5,
        openinterest=-1)

    # Add the Data Feed to Cerebro
    cerebro.adddata(data)

    # Run strategy
    result = cerebro.run(runonce=False)

Here is a sample of my data:

date,open,high,low,close,volume
2002-10-21 01:00:00,1.55035,1.55035,1.5503,1.5503,0
2002-10-21 01:01:00,1.55035,1.55035,1.5503,1.5503,0
2002-10-21 01:02:00,1.55035,1.55035,1.5503,1.5503,0
2002-10-21 01:03:00,1.55035,1.55035,1.5503,1.5503,0
2002-10-21 01:04:00,1.5502,1.5502,1.55015,1.55015,0
2002-10-21 01:05:00,1.54975,1.54995,1.54965,1.54985,0
2002-10-21 01:06:00,1.54995,1.55005,1.54985,1.55,0
2002-10-21 01:07:00,1.54995,1.54995,1.5497,1.5497,0
2002-10-21 01:08:00,1.54955,1.5496,1.54945,1.54945,0
2002-10-21 01:09:00,1.54935,1.54945,1.54935,1.54945,0
2002-10-21 01:10:00,1.54945,1.54945,1.54935,1.54945,0
2002-10-21 01:11:00,1.5495,1.54955,1.54935,1.54955,0
2002-10-21 01:12:00,1.5495,1.5495,1.5495,1.5495,0
2002-10-21 01:13:00,1.54955,1.54955,1.5495,1.5495,0
2002-10-21 01:14:00,1.54955,1.54955,1.54955,1.54955,0
2002-10-21 01:15:00,1.54955,1.54955,1.54955,1.54955,0
2002-10-21 01:16:00,1.5495,1.54955,1.5495,1.5495,0
2002-10-21 01:17:00,1.5494,1.54945,1.5494,1.5494,0
2002-10-21 01:18:00,1.5494,1.54945,1.5494,1.5494,0
2002-10-21 01:19:00,1.5493,1.5493,1.5492,1.5492,0
2002-10-21 01:20:00,1.5494,1.5495,1.5494,1.5494,0
2002-10-21 01:21:00,1.5493,1.5494,1.5493,1.5493,0
2002-10-21 01:22:00,1.54945,1.54945,1.54885,1.5490000000000002,0
2002-10-21 01:23:00,1.54885,1.5489,1.54875,1.5488,0
2002-10-21 01:24:00,1.54895,1.5490000000000002,1.54885,1.5489,0
2002-10-21 01:25:00,1.5488,1.5488,1.5488,1.5488,0
2002-10-21 01:26:00,1.5489,1.5489,1.5487,1.54875,0
2002-10-21 01:27:00,1.5487,1.54875,1.5487,1.54875,0
2002-10-21 01:28:00,1.5486,1.54865,1.54835,1.5484,0
2002-10-21 01:29:00,1.5485,1.54855,1.5484,1.5484,0
2002-10-21 01:30:00,1.5483,1.54835,1.54805,1.54805,0
2002-10-21 01:31:00,1.54805,1.5484,1.54805,1.5483,0
2002-10-21 01:32:00,1.5483,1.54835,1.5483,1.5483,0
2002-10-21 01:33:00,1.5482,1.54825,1.54815,1.5482,0
2002-10-21 01:34:00,1.54825,1.54825,1.54815,1.54825,0
2002-10-21 01:35:00,1.54815,1.5482,1.5481,1.54815,0
2002-10-21 01:36:00,1.54795,1.548,1.54775,1.54785,0

Thanks!

How do you access the attribute values out of _trades? I've tried this:

        for k, v in self._trades.items():

           print(k.tradeid)

and get this error:

AttributeError: 'Lines_LineSeries_DataSeries_OHLC_OHLCDateTime_Abst' object has no attribute 'tradeid'

Ah perfect, thanks @ab_trader.

I am struggling to work how how to output the PNL of a trade once it is in play, this is so that I can monitor a trade and close it if it's PNL reaches -1% of the account size.

I've been looking at the self.position variable which on output shows:

Position:  --- Position Begin
- Size: -66666
- Price: 1.54975
- Price orig: 0.0
- Closed: 0
- Opened: -66666
- Adjbase: 1.61085
--- Position End

The only aspect that seems to change on each iteration of next() is Adjbase, is that showing the change in price? If so I can work out the PNL from that perhaps.

I've dug around the docs but struggling to find the information I need, perhaps there is an easy way to do this? Any help appreciated.

Thanks

How did you resolve this in the end?

Ok thanks, that's good to have the confirmed :)

Ok after some more reading I found that Oanda is supported so that would indicate you can back test currencies and also trade through the Oanda API with Backtrader too which is cool.

If anyone's done this before then any tips are welcome before I make a start on coding my algo for Backtrader :) If not I'll post back any issues I come across on the way and tips for others.

Thanks

Hi everyone,

Hope you're well.

Can I test forex data in Backtrader? Does anyone have any tips for doing so?

I'm new to Backtrader and started looking into it after finding out Zipline does not really support forex.

I've written an algo and trading framework in python which I'm currently testing in the market via an Oanda demo account but I'd like to do some proper backtesting.

I have a few years of minute GBPUSD data in csv form which I can use.

Any tips welcome!

Thanks,
Eddy