maxvolpy: finds good submatrices.
Module maxvolpy is designed for constructing different low-rank skeleton and cross approximations. There are different strategies of doing this, but this module is focused on approximations, based on submatrices of good volume.
- Module maxvolpy can be installed in 2 ways:
- Installation from git repository
Package extensively uses numpy python package and blas/lapack libraries.
Simply run pip with necessary arguments:
pip install maxvolpy
Installation from git repository
To compile and install package from repository, you should have cython and numpy python packages installed.
git clone https://bitbucket.org/muxas/maxvolpy cd maxvolpy python setup.py install
If module was installed from git repository, but was not properly compiled during installation process due to any errors, it will still work if module scipy is presented in system. Module has backup functions, written in Python.
Is available at https://maxvolpy.readthedocs.org