@backtrader Thanks, substituted fromdate instead of backfill_from and it works well.

For the next person searching for how do download ib historical data here is the small snippet that worked.

    stockkwargs = dict(
        timeframe = bt.TimeFrame.Days,
        rtbar = False,  # use RealTime 5 seconds bars
        historical = True,  # only historical download
        what = 'TRADES',  # historical - what to show TRADES MIDPOINT BID ASK BID_ASK HISTORICAL_VOLATILITY OPTION_IMPLIED_VOLATILITY
        useRTH = False,  # historical - download only Regular Trading Hours
        qcheck = 0.5,  # timeout in seconds (float) to check for events
        backfill_start = True,  # do backfilling at the start
        backfill = True,  # do backfilling when reconnecting
        fromdate = datetime.datetime(2017,1,2),  # get data from..
        latethrough = False,  # let late samples through
        tradename = None  # use a different asset as order target
        )
    data0 = ibstore.getdata(dataname="AAPL-STK-SMART-USD", **stockkwargs)
    cerebro.adddata(data0)

@CptanPanic said in Anyone use backtrader to do live trading on Bitcoin exchange?:

GDAX

You might want to look at https://github.com/ccxt-dev/ccxt
and see if you can wrap it to backtrader api as it covers a boatload of exchanges.. It is on my one day list of things to look at.

Jupyter lab is a pretty good enviroment to run experiments and see visual results. For a training enviroment it also allows for you to pose questions but predo any boilerplate like data loading that is not showing concepts your trying to teach.
-Ric

@CptanPanic said in Anyone use backtrader to do live trading on Bitcoin exchange?:

GDAX

You might want to look at https://github.com/ccxt-dev/ccxt
and see if you can wrap it to backtrader api as it covers a boatload of exchanges.. It is on my one day list of things to look at.

@backtrader Thanks, substituted fromdate instead of backfill_from and it works well.

For the next person searching for how do download ib historical data here is the small snippet that worked.

    stockkwargs = dict(
        timeframe = bt.TimeFrame.Days,
        rtbar = False,  # use RealTime 5 seconds bars
        historical = True,  # only historical download
        what = 'TRADES',  # historical - what to show TRADES MIDPOINT BID ASK BID_ASK HISTORICAL_VOLATILITY OPTION_IMPLIED_VOLATILITY
        useRTH = False,  # historical - download only Regular Trading Hours
        qcheck = 0.5,  # timeout in seconds (float) to check for events
        backfill_start = True,  # do backfilling at the start
        backfill = True,  # do backfilling when reconnecting
        fromdate = datetime.datetime(2017,1,2),  # get data from..
        latethrough = False,  # let late samples through
        tradename = None  # use a different asset as order target
        )
    data0 = ibstore.getdata(dataname="AAPL-STK-SMART-USD", **stockkwargs)
    cerebro.adddata(data0)

So I am trying download the historical data from IB.

So I try run getdata on the store like this:

    stockkwargs = dict(
        timeframe = bt.TimeFrame.Days,
        rtbar = False,  # use RealTime 5 seconds bars
        historical = True,  # only historical download
        what = 'TRADES',  # historical - what to show TRADES MIDPOINT BID ASK BID_ASK HISTORICAL_VOLATILITY OPTION_IMPLIED_VOLATILITY
        useRTH = False,  # historical - download only Regular Trading Hours
        qcheck = 0.5,  # timeout in seconds (float) to check for events
        backfill_start = True,  # do backfilling at the start
        backfill = True,  # do backfilling when reconnecting
        backfill_from = datetime.datetime(2017,1,2,tzinfo=timezone('US/Eastern')),  # additional data source to do backfill from
        latethrough = False,  # let late samples through
        tradename = None  # use a different asset as order target
        )
    data0 = ibstore.getdata(dataname="AAPL-STK-SMART-USD", **stockkwargs)
    cerebro.adddata(data0)

But I get an error:

Server Version: 76
TWS Time at connection:20170503 10:13:24 AEST
Traceback (most recent call last):
  File "tutorial3.py", line 129, in <module>
    runstrat()
  File "tutorial3.py", line 122, in runstrat
    cerebro.run()
  File "/home/ric/miniconda3/envs/bt34/lib/python3.5/site-packages/backtrader/cerebro.py", line 1070, in run
    runstrat = self.runstrategies(iterstrat)
  File "/home/ric/miniconda3/envs/bt34/lib/python3.5/site-packages/backtrader/cerebro.py", line 1144, in runstrategies
    data._start()
  File "/home/ric/miniconda3/envs/bt34/lib/python3.5/site-packages/backtrader/feed.py", line 203, in _start
    self.start()
  File "/home/ric/miniconda3/envs/bt34/lib/python3.5/site-packages/backtrader/feeds/ibdata.py", line 361, in start
    self.p.backfill_from._start()
AttributeError: 'datetime.datetime' object has no attribute '_start'

Now as far as I can see backfill_from should be a datetime object.
Now looking at git blame I can see that line 361 of ibdata.py was added a few months ago.

Ensure initialization of backfill_from feeds takes place

Now if I simply remove that section of ibdata.py and set

      self._state = self._ST_START 

The historical download works and it disconnects and then plots beautifully.

So my question? Bug in ibdata.py or I am doing it wrong?
Thanks,
Ric