Hikyuu is an open-source quantitative trading research framework built on C++ and Python that emphasizes ultra-fast backtesting and strategy analysis.
The framework addresses the need for rapid prototyping and validation of trading strategies by decomposing quantitative analysis into modular components including market environment assessment, signal indicators, risk management models, position sizing, slippage simulation, and portfolio analysis. Users combine these independent modules to construct custom strategy models, validate them through backtesting, and analyze performance. The core engine is written in C++ for computational speed while exposing Python interfaces for strategy development and research workflows.
The tool is designed for quantitative researchers and traders working with Chinese A-share markets, though it supports multiple data storage backends including HDF5, MySQL, ClickHouse, and SQLite. It suits projects requiring systematic strategy research with emphasis on backtesting stability and effectiveness verification. The framework provides extension interfaces for integrating third-party trading terminals and custom implementations, allowing developers to adapt it to specific requirements. The README explicitly disclaims that this is a research tool for personal learning and academic use only, not for providing investment advice or direct trading services, and users bear full responsibility for any trading interface integrations they develop.
Development activity shows consistent cross-platform support with automated builds for both Windows and Ubuntu environments. The project maintains comprehensive documentation including Jupyter notebook tutorials and a dedicated strategy component library repository. Multiple storage backend implementations indicate ongoing infrastructure development to support different deployment scenarios. The framework includes performance benchmarking as part of its development focus, with documented metrics for computational efficiency on specific hardware configurations.
