Skip to content

README.md

Latest commit

 

History

History
86 lines (58 loc) · 2.93 KB

README.md

File metadata and controls

86 lines (58 loc) · 2.93 KB
Raw

Solstice Algorithmic Trading Strategy Engine

Solstice Strategy Engine is a high-performance backtesting and strategy execution framework built using C++23 with Python bindings via pybind11. The engine enables rapid development and testing of quantitative trading strategies with seamless integration between C++ execution logic and Python data pipelines.

The architecture leverages the Strategy Pattern with a dispatcher-based routing system, allowing multiple strategy implementations to coexist with minimal overhead. Market data is ingested via Python (yfinance), transformed into C++-native structures, and processed through strategy-specific execution logic.

⚠️ Current Development Status

The Strategy component is currently standalone and operates independently for backtesting purposes.

  • ✓ Backtesting Framework: Fully functional Python-C++ integration for historical data analysis
  • ✓ Market Data Pipeline: Complete yfinance integration with configurable parameters
  • ✗ Matcher Integration: No current interface with the order matching engine
  • ✗ Pricer Integration: No pricing component integration exists

This component will be integrated with the Matcher and Pricer in future releases.

Key Features

  • Hybrid C++/Python Architecture: Core strategy logic in C++23 for performance, data ingestion and configuration in Python for flexibility.
  • Strategy Dispatcher Pattern: Modular, extensible strategy registry with factory-based instantiation.
  • pybind11 Integration: Seamless Python-C++ interoperability with automatic type conversions and exception handling.
  • Market Data Pipeline: Configurable data fetching via yfinance with support for multiple timeframes, intervals, and tickers.
  • Report Generation: Structured performance metrics including PnL, trade counts, win/loss ratios, and candle analysis.
  • Development-Optimized Build System: Fast iteration with optional rebuild flags and separate CMake/setuptools workflows.

Local Execution

Initial Setup

Build and install the Python C++ extension:

pip install -e .

Running Backtests

Fast execution (no rebuild):

./backtest.sh

or directly:

python src/strategy/entry_point.py

With C++ extension rebuild (use after modifying C++ files):

./backtest.sh --rebuild

Configuration

Backtest parameters are configured via src/config/backtest/data_config.json:

  • ticker: Symbol to fetch (e.g., "AAPL", "TSLA")
  • interval: Candle interval ("1m", "5m", "1h", "1d", etc.)
  • time_frame: Either "period" or "start_end"
    • period: Relative period ("1mo", "3mo", "1y", etc.)
    • start_end: Absolute date range (start_date, end_date)

Example configuration:

{
  "ticker": "AAPL",
  "interval": "1d",
  "time": {
    "time_frame": "period",
    "period": "1y"
  }
}

Benchmarks

Benchmark results will be added as strategies are implemented and tested.