The library for scientific, mathematical and engineering calculations SciPy 1.8.0 has been released. SciPy provides a large collection of modules for tasks such as calculating integrals, solving differential equations, image processing, statistical analysis, interpolation, applying Fourier transforms, finding the extremum of a function, vector operations, converting analog signals, working with sparse matrices, etc. The project code is distributed under the BSD license and uses the high-performance implementation of multidimensional arrays from the NumPy project.
The new version of SciPy offers an initial implementation of the API for working with sparse arrays, most of whose elements are equal to zero. To perform calculations with large sparse data sets, the PROPACK SVD library is included, the functions of which, when the "solver='PROPACK'" parameter is set, are available through the "scipy.sparse.svds" submodule. A new submodule "scipy.stats.sampling" has been added, providing an interface to the UNU.RAN C library, designed to sample arbitrary one-dimensional non-uniform continuous and discrete distributions. All private namespaces that do not use underscores in their names have been deprecated.
Source: opennet.ru