Need help to add web socket live streaming Data to cerebro()

Saish Gadekar

Hi backtrader, thank your for awesome backtrader platform!

I use Upstox API, Indian stock broker. i want to add live streaming web socket data to cerebro to perform live trading. Following is my code to retrieve streaming data:

from upstox_api.api import *
import datetime
import pandas as pd

api_key = open('api_key.txt','r').read()
access_token = open('access_token.txt','r').read().strip()
u = Upstox(api_key,access_token)
master_contract = u.get_master_contract('MCX_FO')
u.subscribe(u.get_instrument_by_symbol('MCX_FO', 'CRUDEOILM18DECFUT'),
    LiveFeedType.LTP)

def event_handler_quote_update(message):
     df1 = pd.DataFrame.from_dict(message)
     
     df1.drop('exchange',axis=1,inplace=True)
     df1.drop('close',axis=1,inplace=True)
     df1.drop('symbol',axis=1,inplace=True)
     df1.drop('timestamp',axis=1,inplace=True)
     df1.drop('instrument',axis=1,inplace=True)
     
     df2 = df1.iloc[0,0]

     print(df2)

 u.set_on_quote_update(event_handler_quote_update)
 u.start_websocket(True)

Output:
3558.0
3559.0
3559.0
3558.0
3558.0
3558.0
3560.0
.....
.....
My Questions:

Now i need to construct open, high, low, close data using above streaming output
Then resample above constructed data to "5Min" and then ADD this data as a live feed data to cerebro.

and please mention where to locate codes that you will be providing.

Please help me out with code to achieve my need. Thanks again!

backtrader

Your best bet is to look at the OandaStore (and associated OandaData and OandaBroker) for an implementation. Things you will for sure need:

A background thread in which things run
Because it is a websockets implementation: a web server (although there are wrappers which isolate you from this and the Upstox API seems to already do this)
A lot of patience during the implementation, mostly related to testing and ironing out the specific details of the provider.

References:

backtrader

@saish-gadekar said in Need help to add web socket live streaming Data to cerebro():

Then resample above constructed data to "5Min" and then ADD this data as a live feed data to cerebro.

Resampling is already built-in and is not bound to any specific data feed: Docs - Resampling