Package: wnpp Severity: wishlist Owner: Edward Betts <[email protected]>
* Package name : numpy-minmax Version : 0.5.0 Upstream Author : Iver Jordal <[email protected]> * URL : https://github.com/nomonosound/numpy-minmax * License : MIT Programming Lang: Python Description : fast minimum and maximum calculation for NumPy arrays Provides a function that returns the minimum and maximum values from a NumPy array in a single call. This avoids making separate passes for low and high values and is intended for cases where both results are needed together. . The implementation focuses on float32 arrays, with optimized handling for C-contiguous arrays, Fortran-contiguous arrays, and one-dimensional strided arrays. It also includes optimized handling for contiguous int16 arrays. For array layouts outside those cases, such as higher-dimensional strided arrays, it falls back to NumPy's normal minimum and maximum operations. . This library is aimed at fast reduction over array data while keeping the result simple: a pair containing the smallest and largest values found in the input. It works on NumPy ndarrays and is particularly concerned with the memory layout and element type of the array, since those determine which code path is used. This library is a dependancy of Home Assistant, the Python smart home platform. I plan to maintain it as part of the Home Assistant team.
