Onyx Trading Engine

A modular and extensible trading strategy creation platform built with scalable design patterns, supporting live trading (Interactive Brokers) and backtesting (custom/Yahoo adapters) for stocks, forex, and more.

Features

• Live Trading: Connects to Interactive Brokers (IB or IBkr) via IB API and Traders Works Station (TWS) for real-time execution.
• Backtesting: Simulate strategies using historical data from Yahoo Finance or custom CSVs.
• Strategy Framework: Easily implement and swap trading strategies (e.g., Moving Average Crossover).
• Thread-Safe Data Flow: Robust handling of live and historical bar data using queues and event-driven callbacks.
• Comprehensive Logging: Unified logging for all trades, statistics, and system events.
• Visualization: Candlestick charts with buy/sell markers for backtest results.

Quick Start

To setup the environment, read and follow src/README.md

Architecture Overview

• Domain Models: Asset, Bar (stocks, forex, crypto, etc.)
• Ports: MarketDataPort, BrokerTradePort (abstract interfaces)
• Adapters: IB API, Yahoo Finance, Custom Backtest
• Strategies: Modular, plug-and-play (e.g., Moving Average, Kalman Filter)
• Engine: Orchestrates data, strategy, and broker actions

Example: Moving Average Crossover Strategy

Code Implementation

Strategies in Onyx Trading Engine are built on our proprietary framework that enforces a clean separation between domain models, ports, and adapters.
This allows the same strategy code to be backtested on historical data and then deployed directly to Interactive Brokers for live trading.

from src.core.logic.trading_engine import TradingEngine
from src.core.adapters.yf_market_adapter import YFMarketDataAdapter
from src.core.logic.moving_average import MovingAverageCrossoverStrategy
from src.core.adapters.custom_broker_adapter import CustomBrokerAdapter
from src.core.models.asset import Asset, AssetType
from datetime import datetime, timedelta

if __name__ == "__main__":
    # Define the trading instrument
    asset = Asset(AssetType.FOREX, "EUR", "USD")

    # Market data adapter (historical data for backtesting)
    market_data_adapter = YFMarketDataAdapter(asset)
    market_data_adapter.request_historical_data(
        asset, datetime.today()-timedelta(days=365*4), datetime.today()
    )

    # Broker adapter (simulated broker for backtesting)
    broker_adapter = CustomBrokerAdapter(10000)

    # Strategy (uses abstract ports, works in backtest or live trading)
    strategy = MovingAverageCrossoverStrategy(broker_adapter, asset)

    # Trading engine orchestrates data, strategy, and broker actions
    trading_engine = TradingEngine(broker_adapter, market_data_adapter, [strategy])
    trading_engine.run(asset, threaded=False)

Backtesting Results

During backtesting, results are visualized as candlestick charts with buy/sell markers.
For example, the Moving Average Crossover strategy produces the following backtest chart:

Seamless Transition to Live Trading

Because strategies depend only on abstract ports (MarketDataPort and BrokerTradePort), the same code can be run in a live trading environment by simply swapping adapters:

• Replace YFMarketDataAdapter with IbApiDataAdapter
• Replace CustomBrokerAdapter with IbBrokerAdapter

No changes to the strategy logic or engine are required—ensuring smooth deployment from backtesting to production trading.

Authors

• Josue Dazogbo
• Sultan Oloyede
• Jedidiah Ange-Emmanuel Kouakou
• Racine Kane

Resources

• Installation Guide
• Quantitative Analysis Reports