repository·master·Indexed 25 days ago
https://github.com/erdewit/ib_insyncA Python library designed to simplify working with the Interactive Brokers Trader Workstation (TWS) API. It provides a linear programming style and an asynchronous framework based on asyncio to keep data in sync with TWS or IB Gateway. The library supports interactive development in Jupyter notebooks and includes functionality for downloading historical data, requesting scanner data, calculating option implied volatility, accessing market depth, and managing orders.
To fetch a continuous stream of historical data (e.g., 1-minute bars) from a specific point in time back to the beginning, use a while loop. Start with the current time and use the endDateTime parameter in reqHistoricalData to request data in chunks, updating the endDateTime with the date of the earliest bar received in the previous request until no more data is returned.
import datetime
from ib_insync import *
ib = IB()
ib.connect('127.0.0.1', 7497, clientId=1)
contract = Stock('TSLA', 'SMART', 'USD')
dt = ''
barsList = []
while True:
bars = ib.reqHistoricalData(
contract,
endDateTime=dt,
durationStr='10 D',
barSizeSetting='1 min',
whatToShow='MIDPOINT',
useRTH=True,
formatDate=1)
if not bars:
break
barsList.append(bars)
dt = bars[0].date
print(dt)
# save to CSV file
allBars = [b for bars in reversed(barsList) for b in bars]
df = util.df(allBars)
df.to_csv(contract.symbol + '.csv', index=False)The IB socket protocol is designed for long-lived connections. If you are using the library for short-lived tasks (e.g., firing a few orders and exiting), add a small delay (e.g., ib.sleep(1)) before calling ib.disconnect(). This ensures that any pending data in the buffer is flushed to the server.
ib = IB()
ib.connect()
... # create and submit some orders
ib.sleep(1) # added delay
ib.disconnect()ib_insync code within a Jupyter notebook, you must call util.startLoop() to ensure the asynchronous event loop is running correctly within the notebook environment.Install the library using pip:
pip install ib_insyncNote: The official ibapi package from Interactive Brokers is not required.
To improve performance, use 'current state' methods instead of 'request' methods. Request methods (prefixed with req) involve network round-trips and are slower, whereas current state methods (e.g., ib.positions()) access data already synchronized in memory and are available immediately.
Examples:
ib.positions() instead of ib.reqPositions()ib.openOrders() instead of ib.reqOpenOrders()To use ib_insync in a Jupyter Notebook, import all members and start the event loop using util.startLoop(). This ensures the notebook remains live-updated with data from the IB API.
Note: util.startLoop() is designed specifically for notebooks and will not work in regular Python scripts.
from ib_insync import *
util.startLoop()To receive live updates for a historical data subscription, call ib.reqHistoricalData with endDateTime='' and keepUpToDate=True.
You can subscribe to updates by attaching a callback function to the updateEvent of the returned bars object. The callback receives (bars, hasNewBar).
Warning: Using keepUpToDate=True may cause the API to become inoperable if a network interruption occurs.
contract = Forex('EURUSD')
# Request bars with keepUpToDate enabled
bars = ib.reqHistoricalData(
contract,
endDateTime='',
durationStr='900 S',
barSizeSetting='10 secs',
whatToShow='MIDPOINT',
useRTH=True,
formatDate=1,
keepUpToDate=True)
def onBarUpdate(bars, hasNewBar):
print(f"New bar: {bars[-1]}")
# Subscribe to updates
bars.updateEvent += onBarUpdate
# To stop the subscription
ib.cancelHistoricalData(bars)Use the IB class to establish a connection to a running Trader Workstation (TWS) or IB Gateway application.
Troubleshooting Connection Failures:
4002.clientId provided is not already in use by another session.ib = IB()
ib.connect('127.0.0.1', 7497, clientId=10)Once you have an option chain, you can programmatically build a list of Option contracts based on specific criteria like strike price increments, expiration dates, and rights ('P' for Put, 'C' for Call). After creating the list of Option objects, use ib.qualifyContracts(*contracts) to fill in missing details like conId and ensure they are valid for trading.
# Example: Building contracts based on filtered strikes and expirations
strikes = [strike for strike in chain.strikes if strike % 5 == 0]
expirations = sorted(chain.expirations)[:3]
rights = ['P', 'C']
contracts = [Option('SPX', expiration, strike, right, 'SMART', tradingClass='SPX')
for right in rights
for expiration in expirations
for strike in strikes]
# Qualify the contracts
contracts = ib.qualifyContracts(*contracts)To use ib_insync, ensure you meet the following requirements:
Note: The official ibapi package from Interactive Brokers is not required.
To start using ib_insync, initialize the IB object, start the event loop using util.startLoop(), and connect to your IB gateway or TWS instance.
Warning: Using live accounts will place real orders. It is recommended to use a paper trading account during testing.
from ib_insync import *
util.startLoop()
ib = IB()
ib.connect('127.0.0.1', 7497, clientId=13)