This project is a Python-based trading simulator that allows users to simulate trading strategies, manage an order book, and interact with a mock trading environment using various algorithmic traders. The simulator includes a FIX (Financial Information eXchange) protocol handler, a market-making algorithm, and synthetic liquidity generation.
- Order Book Management: A robust order book implementation that supports adding, modifying, and canceling orders.
- Matching Engine: A matching engine that handles the execution of orders based on the best available prices in the order book.
- FIX Protocol Handler: A simple FIX protocol implementation that can handle new orders and order cancellations.
- Market Data Feed: A live market data feed using
yfinancethat can be subscribed to by various components. - Market Maker: A basic market-making algorithm that posts bid and ask prices around the current market price.
- Synthetic Liquidity Provider: A module to inject synthetic liquidity into the order book to simulate a more active trading environment.
- Algorithmic Traders: Various algorithmic trading strategies, including momentum-based, EMA-based, swing trading, and sentiment analysis-driven trading.
To install the required libraries, use the following requirements.txt:
tensorflow==2.6.0
newsapi-python==0.2.6
simplefix==1.0.12
yfinance==0.1.63
numpy==1.21.2
pandas==1.3.3
Install using pip
pip install -r requirements.txtIf you prefer to run the program without the algorithmic traders, use the following requirements.txt:
simplefix==1.0.12
yfinance==0.1.63
numpy==1.21.2
pandas==1.3.3
Install using pip
pip install -r requirements.txtTo start the simulator, initialize the various components like the order book, matching engine, market data feed, and FIX server
order_book = OrderBook()
matching_engine = MatchingEngine(order_book)
fix_app = FixApplication(matching_engine)
# Prompt user for the stock ticker
ticker_symbol = input("Enter the stock ticker symbol you want to track (e.g., AAPL): ").upper()
# Initialize the MarketDataFeed with the user-provided ticker symbol
market_data_feed = MarketDataFeed(symbol=ticker_symbol)
market_maker = MarketMaker(symbol=ticker_symbol, matching_engine=matching_engine)
# Start FIX server
fix_thread = threading.Thread(target=fix_app.start)
fix_thread.start()
# Start market data feed
feed_thread = threading.Thread(target=market_data_feed.start)
feed_thread.start()
# Start market maker
market_maker.start(market_data_feed)Create, modify, and cancel orders in the order book:
# Create an ask order
ask_order = Order(id='2', price=415, quantity=100, side='sell', type='limit', symbol='MSFT')
order_book.add_order(ask_order)
# Modify an order's quantity and price
order_book.modify_order(order_id='2', new_quantity=30, new_price=100.5)
# Cancel an order
order_book.cancel_order(order_id='3')
# Display the order book
order_book.display_order_book()Inject synthetic liquidity into the order book to simulate a more dynamic trading environment:
liquidity_provider = SyntheticLiquidityProvider(symbol=symbol, matching_engine=matching_engine, num_orders=10)
liquidity_provider.generate_liquidity()
# Optionally, inject liquidity at regular intervals
def auto_inject_liquidity(provider, interval=5):
while True:
provider.generate_liquidity()
time.sleep(interval)
liquidity_thread = threading.Thread(target=auto_inject_liquidity, args=(liquidity_provider, 5)) # Inject every 5 seconds
liquidity_thread.start()Start algorithmic traders such as momentum-based, EMA-based, swing trading, and sentiment analysis-driven trading:
# Load the pre-trained sentiment analysis model (assuming it’s stored as 'sentiment_classifier_model.h5')
sentiment_model = tf.keras.models.load_model('sentiment_classifier_model.h5')
# Initialize the traders
momentum_trader = MomentumTrader(symbol=ticker_symbol, matching_engine=matching_engine, interval=10)
ema_trader = EMABasedTrader(symbol=ticker_symbol, matching_engine=matching_engine, interval=30)
swing_trader = SwingTrader(symbol=ticker_symbol, matching_engine=matching_engine, interval=15)
sentiment_trader = SentimentAnalysisTrader(
symbol=ticker_symbol,
matching_engine=matching_engine,
model_file='sentiment_classifier_model.h5',
news_api_key="your_newsapi_key_here",
interval=60
)
# Collect all traders into a list
traders = [momentum_trader, ema_trader, swing_trader, sentiment_trader]
# Start each trader in its own thread
trader_threads = []
for trader in traders:
thread = threading.Thread(target=trader.start, args=(market_data_feed,))
thread.start()
trader_threads.append(thread)Stop the market maker, market data feed, FIX server, and all traders:
market_maker.stop()
market_data_feed.stop()
fix_app.stop()
# Stop all traders
for trader in traders:
trader.stop()
# Join all threads to ensure they have finished
for thread in trader_threads:
thread.join()Run a backtest using historical data
def load_historical_data(symbol, start_date, end_date):
"""Load historical data from yfinance."""
data = yf.download(symbol, start=start_date, end=end_date)
# Ensure the data is formatted as expected
data.reset_index(inplace=True)
return data
def run_backtest(historical_data, market_maker, matching_engine):
"""Simulate historical data through the market maker and matching engine."""
for index, row in historical_data.iterrows():
# Simulate a market data update for the market maker
market_data = {
'symbol': market_maker.symbol,
'price': row['Close'],
'timestamp': row['Date']
}
# MarketMaker reacts to the incoming market data
market_maker.on_market_data(market_data)
print("Backtest completed.")
if __name__ == "__main__":
# Initialize historical data
symbol = "AAPL"
start_date = "2023-01-01"
end_date = "2023-12-31"
historical_data = load_historical_data(symbol, start_date, end_date)
# Initialize the market maker, matching engine, and other components
order_book = OrderBook()
matching_engine = MatchingEngine(order_book)
market_maker = MarketMaker(symbol=symbol, matching_engine=matching_engine)
# Run the backtest
run_backtest(historical_data, market_maker, matching_engine)This project is licensed under the MIT License