bt-ig-store
IG markets store for Backtrader
Warning UNDER Development
Currently integration is very limited. Only streaming is working.
Requires
1 - trading_ig
A lightweight Python library that can be used to get live data from IG Markets REST and STREAM API
2 - Backtrader
Python Backtesting library for trading strategies
3 - Pandas.
Current Functionality
- Basic instrument streaming
- Performs the open position check when initialized to track existing positions
- Opening and closing of simple Market orders using the self.buy() and self.close() is now supported.
- Set IG currency code as a store initialization parameter (Default GBP).
- Stop order creation and cancellation supported.
- Limit order creation and cancellation supported.
- expiry, guaranteed_stop, time_in_force, and good_till_date parameters
- Improved streamer setup. Can use the same streamer for multiple get_instruments rather than creating multiple streamers
- Manual pull of cash and value
- Account cash and value live streaming
- FIX: Level Set during order creation caused MARKET Orders to be rejected can now all be passed as key word arguments during order creation and handled appropriately. Defaults are used where no kwarg is passed.
- Historical data download for backtesting
- Printing of remaining allowance (will later be updated to be a store notification, once notifications are working)
- Sample script for historical data download testing.
- Store and Data notifications tested and working
- Samples updated to print Store and Data notifications
- Updated granualarity check to only make the check if Backfilling, Historical or Backfill_Start are required. This allows you to work with any timeframe using tick data for live trading only.
- New: Backfilling from start now supported
- New: Parameter
backfill_barsallows you to set how many bars to backfill from the start. Since the IG api places restrictions on the number of historical data points a user can download per week, this parameter allows a user to limit the amount of historical data downloaded to only the required amount.
NOTE: Backfilling is not yet supported. The check mentioned above is in preparation for those features being supported.
Important!
To use historical data, you will need to use the forked trading_ig API from my profile.
If you do not, you will receive a ValueError when requesting historical data. A
pull request has been submitted to the official project and this warning will be
removed if/when it is accepted.
Known Issues
See the issues tab: https://github.com/Dave-Vallance/bt-ig-store/issues
Value for header {Version: 2} must be of type str or bytes, not <class 'int'>
This issue is documented in the issues tab. If you see it, you should make sure you have the latest trading_ig module code.
Streaming Example
import backtrader as bt
from datetime import datetime
import logging
from bt_ig import IGStore
from bt_ig import IGData
api_key = 'INSERT YOUR API KEY'
usr = 'INSERT YOUR USERNAME'
pwd = 'INSERT YOU PASSWORD'
acc = "INSERT YOUR ACC NUM"
class IGTest(bt.Strategy):
'''
Simple strat to test IGStore.
'''
def __init__(self):
self._live = False #Track whether we have live data to avoid entering on backfill
self._last_hist_bar = None #To track the last delivered historical bar
def next(self):
dt = self.datetime.datetime()
bar = len(self)
lhst = self._last_hist_bar
print('{}: O: {} H: {} L: {} C:{}'.format(dt, self.data.open[0],
self.data.high[0],self.data.low[0],self.data.close[0]))
if self._live:
if bar == lhst + 1:
print('Testing Get Cash!')
cash = self.broker.getcash()
print("Current Cash: {}".format(cash))
print('Testing Get Value!')
value = self.broker.getvalue()
print("Current Value: {}".format(value))
if bar == lhst + 2:
print('Testing Simple Order!')
pos = self.broker.getposition(self.data)
self.buy(size=5)
if bar == lhst + 3:
print('Closing Order')
pos = self.broker.getposition(self.data)
print('Open Position Size = {}'.format(pos.size))
cOrd = self.close()
print('Closing Order Size = {}'.format(cOrd.size))
# CHECK CASH AND EQUITY ARE AUTMATICALLY BEING UPDATED
cash = self.broker.getcash()
value = self.broker.getvalue()
print("Current Cash: {}".format(cash))
print("Current Value: {}".format(value))
if bar == lhst +4:
print('Testing Limit Order')
limit_price = self.data.close[0] * 0.9 #Buy better price 10% lower
self.limit_ord = self.buy(exectype=bt.Order.Limit, price=limit_price, size=5)
if bar == lhst + 5:
print('Cancelling Limit Order')
self.cancel(self.limit_ord)
if bar == lhst + 6:
print('Testing Stop Order')
stop_price = self.data.close[0] * 0.9 #buy at a worse price 10% lower
self.stop_ord = self.buy(exectype=bt.Order.Limit, price=stop_price, size=5)
if bar == lhst + 7:
print('Cancelling Stop Order')
self.cancel(self.stop_ord)
if bar == lhst + 8:
print("Test Finished")
self.env.runstop()
## NOTIFICATIONS
def notify_order(self,order):
if order.status == order.Rejected:
print('ORDER NOTIF: Order Rejected')
def notify_data(self, data, status, *args, **kwargs):
print('DATA NOTIF: {}: {}'.format(data._getstatusname(status), ','.join(args)))
if status == data.LIVE:
self._live = True
self._last_hist_bar = len(self)
def notify_store(self, msg, *args, **kwargs):
print('STORE NOTIF: {}'.format(msg))
#Logging - Uncomment to see ig_trading library logs
#logging.basicConfig(level=logging.DEBUG)
tframes = dict(
seconds = bt.TimeFrame.Seconds,
minutes=bt.TimeFrame.Minutes,
daily=bt.TimeFrame.Days,
weekly=bt.TimeFrame.Weeks,
monthly=bt.TimeFrame.Months)
#Create an instance of cerebro
cerebro = bt.Cerebro()
#Setup IG
igs = IGStore(usr=usr, pwd=pwd, token=api_key, account=sbet)
broker = igs.getbroker()
cerebro.setbroker(broker)
data = igs.getdata(dataname='CS.D.GBPUSD.TODAY.IP', backfill_start=True, backfill_bars=50)
#Replay the data in forward test envirnoment so we can act quicker
#cerebro.resampledata(data, timeframe=tframes['seconds'], compression=15, name='GBP_USD')
cerebro.resampledata(data, timeframe=tframes['minutes'], compression=1, name='GBP_USD')
#Add our strategy
cerebro.addstrategy(IGTest)
# Run over everything
cerebro.run()Historical Data Example
import backtrader as bt
from datetime import datetime
import logging
from bt_ig import IGStore
from bt_ig import IGData
api_key = 'INSERT YOUR API KEY'
usr = 'INSERT YOUR USERNAME'
pwd = 'INSERT YOU PASSWORD'
acc = "INSERT YOUR ACC NUM"
class IGTest(bt.Strategy):
'''
Simple strat to test IGStore.
'''
def __init__(self):
pass
def next(self):
dt = self.datetime.datetime()
bar = len(self)
print('{}: O: {} H: {} L: {} C:{}'.format(dt, self.data.open[0],
self.data.high[0],self.data.low[0],self.data.close[0]))
## NOTIFICATIONS
def notify_order(self,order):
if order.status == order.Rejected:
print('Order Rejected')
def notify_data(self, data, status, *args, **kwargs):
print('DATA NOTIF: {}: {}'.format(data._getstatusname(status), ','.join(args)))
def notify_store(self, msg, *args, **kwargs):
print('STORE NOTIF: {}'.format(msg))
#Logging - Uncomment to see ig_trading library logs
#logging.basicConfig(level=logging.DEBUG)
tframes = dict(
seconds = bt.TimeFrame.Seconds,
minutes=bt.TimeFrame.Minutes,
daily=bt.TimeFrame.Days,
weekly=bt.TimeFrame.Weeks,
monthly=bt.TimeFrame.Months)
#Create an instance of cerebro
cerebro = bt.Cerebro()
#Setup IG
igs = IGStore(usr=usr, pwd=pwd, token=api_key, account=sbet)
broker = igs.getbroker()
cerebro.setbroker(broker)
data = igs.getdata(dataname='CS.D.GBPUSD.TODAY.IP', historical=True,
timeframe=tframes['minutes'], compression=60,
fromdate=datetime(2017,11,1,8,0), todate=datetime(2017,11,1,12,0)) #Get 10 bars for testing.
#Replay the data in forward test envirnoment so we can act quicker
cerebro.adddata(data, name='GBP_USD')
#Add our strategy
cerebro.addstrategy(IGTest)
# Run over everything
cerebro.run()