numpy

NumPy is a library for the Python programming language that adds support for large, multi-dimensional arrays and matrices, along with a collection of mathematical functions to operate on these arrays. The central data structure is the N-dimensional array, or ndarray, which provides vectorized numerical operations and efficient memory handling.

NumPy array objects enable broadcasting, slicing, indexing, and reshaping; it also includes universal functions (ufuncs) for elementwise operations, as well as a suite of mathematical and statistical tools, linear algebra, random number generation, and Fourier transforms. The arrays are designed to be efficiently implemented in C and Fortran behind the scenes, enabling performance close to native code for many operations.

NumPy emphasizes broadcasting, which allows arithmetic operations between arrays of different shapes under well-defined rules. It also supports views and strides, enabling memory-efficient operations and avoiding unnecessary copies. The library provides tools for input and output and for creating, transforming, and combining arrays in a variety of ways.

The library is a core component of the SciPy ecosystem and a foundational dependency for many other Python scientific libraries, including pandas, scikit-learn, and JAX. It provides interoperability with C, C++, and Fortran code through C extensions and the NumPy C-API, and can be used with various data sources and file formats.

History and licensing: NumPy originated in the mid-2000s as a successor to Numeric and Numarray projects and was primarily developed by Travis Oliphant. It is released under the BSD license and is maintained by a global community of contributors.