pyprox is a free Open Source proximal algorithm library for Python programming language.
pyprox aims to bring simplicity to software using proximal algorithms. To start using it, install the package, open the the Python interactive shell and type:
>>> import pyprox as pp
>>> import numpy as np
>>> x0 = np.zeros((100,1))
>>> y = x0 + 0.01 * np.random.randn(100,1)
>>> ProxF = pp.soft_thresholding
>>> GradG = lambda x : x - y
>>> xrec = pp.forward_backward(ProxF, GradG, y, 1)
pyprox requires Python 2.5-2.7 installed. The only external requirement is a recent version of NumPy numeric array module.
The most recent development version can be found on Github at:
- Github - https://github.com/svaiter/pyprox
Latest release is available for download from the Python Package Index at http://pypi.python.org/pypi/pyprox/.
Post your suggestions and guestions to pyprox discussions group (pyprox@googlegroups.com).
Comments, bug reports and fixes are welcome through Github