Here are the significant updates in NumPy and SciPy libraries:

NumPy Updates

• NumPy 2.0: Released in June 2024, this major update includes significant changes such as a new API, improved scalar promotion rules, a new DType API, and a new string DType. It also introduces breaking changes, including an ABI break, which may require code adjustments.

  • API and ABI Changes: The update includes a public and private API split, namespace cleanup, and deprecation of niche functionality. The ABI changes will impact binaries of packages using the NumPy C API.

  • New DType API and String DType: Supports native variable-length string data types, enhancing user-defined custom data types.

  • Scalar Promotion Rules: Updated to prevent user surprises by basing promotion on DTypes rather than data values.

  • Performance Improvements: Faster sorting functions and improved linear algebra operations on macOS using Accelerate.

• NumPy 2.1 and 2.2: These releases provide support for Python 3.13, drop support for Python 3.9, and include bug fixes and improvements. NumPy 2.2 also enhances support for free-threaded Python and improves the StringDType.

SciPy Updates

• SciPy Releases: SciPy releases are not directly tied to NumPy releases. SciPy aims to maintain compatibility with several major releases of NumPy and Python. Recent versions of SciPy support Python 3.10 and later.

• Compatibility Approach: SciPy takes a conservative approach to maintain compatibility with multiple versions of NumPy and Python, ensuring that it works with at least the four previous releases of NumPy.

Overall, NumPy's recent updates focus on modernizing its API and improving performance, while SciPy continues to ensure broad compatibility with different versions of NumPy and Python.