Lion-Core is a lightweight yet powerful Python library designed to serve as a foundational framework for machine learning, artificial intelligence, workflow automation, scientific computing, and advanced data management. With a focus on simplicity and extensibility, Lion-Core aims to provide a robust toolkit for complex computational tasks while maintaining a minimal dependency footprint.

• Minimal Dependencies: Built with only Pydantic as a dependency, ensuring a lightweight and easily maintainable codebase.
• Extensible Architecture: Designed from the ground up to be modular and adaptable to a wide range of computational needs.
• Performance-Focused: Optimized data structures and algorithms for efficient large-scale data processing.

• Advanced Data Structures:

  • Pile: A flexible container that combines the best of lists and dictionaries, offering efficient access and powerful querying capabilities.
  • Progression: An ordered sequence container designed for high-performance operations on large datasets.

• Robust Type System: Leveraging Pydantic for data validation, Lion-Core ensures type safety and data integrity across your entire project.

• Converter System: A flexible framework for seamless data conversion between various formats, reducing the complexity of data pipeline management.

• Element-based Architecture: A foundational Element class that serves as the building block for creating modular and composable components, enabling complex system design with ease.

• Communication Framework: A groundwork for inter-component messaging, facilitating the development of distributed and microservices-based applications.

• Form and Validator System: Dynamic workflow manipulation tools that allow for flexible and adaptable process designs.

• Worker System: Powerful workflow composition capabilities, enabling the creation of complex, multi-stage computational pipelines.

• Adapted Features: A suite of capabilities originally developed for lionagi, now being refined and integrated into Lion-Core for enhanced functionality and performance.

Lion-Core is on an exciting journey of evolution. Here's a glimpse into our vision for the future:

• Polishing core abstractions and data structures
• Enhancing the converter system for broader compatibility
• Refining workflow automation processes
• Expanding scientific computing utilities

• Venturing into quantum realms: Integrating quantum-inspired algorithms to tackle complex problems in new ways
• Embracing uncertainty: Developing tools for stochastic processes and probabilistic modeling
• Streamlining workflows: Creating intuitive interfaces for managing intricate computational pipelines
• Pushing scientific boundaries: Implementing cutting-edge numerical methods and optimization techniques
• Empowering AI/ML practitioners: Crafting specialized modules to simplify and accelerate machine learning workflows
• Chasing performance: Continuously refining our core to handle ever-larger datasets with grace and speed

Lion-Core is not yet available on PyPI. You can install the development version directly from the repository:

pip install git+https://github.com/lion-agi/lion-core.git

Comprehensive documentation is under development. For now, please refer to the inline documentation and comments in the source code. Detailed examples and usage guidelines can be found in the README files within each module folder.

We welcome contributions! As the project is in its early stages, please open an issue to discuss potential changes before submitting pull requests.

Lion-Core is released under the Apache License 2.0. See the LICENSE file for details.

Lion-Core is under active development, currently working towards its first stable release (0.1.0). Expect frequent updates and potential API changes.

For questions, support, or to discuss potential collaborations, please open an issue on our GitHub repository.

Join our Discord community for discussions, support, and updates: Lion-Core Discord Server

If you use Lion-Core in your research or project, please cite it as follows:

@software{Li_LionCore_2024,
  author = {Haiyang Li},
  month = {07},
  year = {2024},
  title = {LionAGI: Towards Automated General Intelligence},
  url = {https://github.com/lion-agi/lion-core},
}

Note: Lion-Core is an extraction and refinement of core components from the larger lionagi project. After reaching version 0.1.0, development will continue on Lion-Core while also informing a rewrite of the lionagi library.