Apifiny Algo Trading System
AATS is a light weight trading system with low-latency and high-scalability for Live trading on multi-crypto exchanges.
Key Features
- Light weight: you only need to take care of your strategy component and you can trade alive!
- Low latency: the low-level system is optimizied by C++, such as exchange connectivity, order book management and order placement
- High flexibility: with raw tick data and quote data, you can freely build samplers, pricing models, variables and signals on your own into your strategy
- Distributed broadcast: You can only start one market instance and connect to multiple trading instances
Quick Start
This quick start gives you all the steps that you need to do before start trading.
Install docker
Please download and install Docker CE or Docker Desktop for your computer/server if it isn't installed:
Get and run the Algo SDK docker image using the following command:
docker run -it apifinyalgo/algo-sdk:1.1.0
his step will automatically run control server which will be used to start market/trading instance later. Depending on system, you may need to run this with sudo.
TBA: how to check if it is running, i.e. ps command
Install aats package
# install stable version from pip (support python version >=3.7)
pip install aats (TBA: currently pip install -i https://test.pypi.org/simple/ aats)
# clone the repository with latest version
git clone http://git.ddesk.io/exone-plus/algo-client-api.git (TBA: replace with public repo link)
Maintain your unique cid symbol mapping table
It requires to be consistent in both market engine and trade engine
sym_cid_map = {
1001: "BTCUSDT.BINANCE",
1002: "ETHUSDT.BINANCE",
1003: "DOGEUSDT.BINANCE",
1004: "BTCUSDTSWAP.BINANCE_SWAP"
}
Start market instance to listen subscribed symbols and exchanges
TBA: change the below to one command line to start market instance
from aats.market_engine import MarketEngine
# start market engine
mkt_engine = MarketEngine(sym_cid_map)
mkt_engine.set_control_server(6000) # default control server port
mkt_engine.set_multicast_cfg('239.0.0.4', 4141)
# mkt_engine.add_listen_symbol('BTCUSDT', 'BINANCE', 5)
# mkt_engine.add_listen_symbol('ETHUSDT', 'BINANCE', 5)
mkt_engine.add_listen_symbol('DOGEUSDT', 'BINANCE', 5)
mkt_engine.add_listen_symbol('BTCUSDTSWAP', 'BINANCE_SWAP', 5)
mkt_engine.run()
Once console returns the following message, it means market instance is running successfully
{'result': 'ok', 'type': 'market_instance'}
Write your strategy and ready to trade
In the examples folder, we provide two sample strategies and demo_main.py script which illustrates how to setup config and run strategy
- SimpleTakerStrategy: it is a naive taker strategy, place order on a fixed time interval
- SimpleMakerStrategy: it is a bit complicated maker strategy with order management system
##############################################
# Demo strategy #
##############################################
import time
from aats.trade_engine import TradeEngine
from simple_maker_strategy import SimpleMakerStrategy
from simple_taker_strategy import SimpleTakerStrategy
sym_cid_map = {
1001: "BTCUSDT.BINANCE",
1002: "DOGEUSDT.BINANCE"
}
# init trade engine
trade_engine = TradeEngine(sym_cid_map)
# setup public config
trade_engine.set_md_multicast_cfg(send_to_ip='239.0.0.3', send_to_port=4141)
trade_engine.add_md_symbol(symbol='BTCUSDT', exchange='BINANCE')
# trade_engine.add_md_symbol(symbol='DOGEUSDT', exchange='BINANCE')
# setup private config
trade_engine.set_control_server(server_port=6000)
trade_engine.config_exchange(exchange='BINANCE', trade_type='sandbox')
trade_engine.set_apikey(exchange="BINANCE",
key="02SvhZEYG1p92JWdekP75XQKayqfLxmjHWNEfWU1KrCPjJ5xrLcOU1YHZ5SUBVFA",
secret="iKnZvDKMGQuEINhjDX8gbIVJDLl48fV6GFLL5gcFT8Sfj9yxGrnP7uFm7AAVWeFP",
password="",
subaccount="")
trade_engine.set_fee('BINANCE', 0.0003, 0)
trade_engine.add_trade_symbol(symbol='BTCUSDT', exchange='BINANCE')
# setup your strategy
# my_strategy = SimpleMakerStrategy()
my_strategy = SimpleTakerStrategy(msecs=10000)
trade_engine.add_strategy(my_strategy)
# start run
try:
trade_engine.run()
except KeyboardInterrupt as err:
# trade_engine.stop()
print("close strategy and close open orders")
# trade_engine.manager.stop()
my_strategy.close()
time.sleep(1)
print("trade closed")