Metadata-Version: 1.0
Name: pysofia
Version: 0.9
Summary: Python bindings for sofia-ml
Home-page: http://pypi.python.org/pypi/pysofia
Author: Fabian Pedregosa & Csy
Author-email: fabian@fseoane.net
License: UNKNOWN
Description-Content-Type: UNKNOWN
Description: PySofia
        =======
        
        .. image:: https://travis-ci.org/rth/pysofia.svg?branch=master
            :target: https://travis-ci.org/rth/pysofia
        .. image:: https://ci.appveyor.com/api/projects/status/jhn2d8n8y836pnj8/branch/master?svg=true
            :target: https://ci.appveyor.com/project/rth/pysofia/branch/master
        
        PySofia is a python wrapper around the methods present in the C++ sofia-ml library. These include Stochastic Gradient Descent implementations of some ranking algorithms, notably RankSVM.
        
        Dependencies
        ------------
        
        - cython >= 0.17 (previous versions will not work)
        - numpy
        - sklearn >= 0.15
        - six
        - enum34 (for Python versions before 3.4)
        - A C++ compiler (gcc will do)
        
        You will not need to install sofia-ml since it is incuded in this distribution
        
        Installation
        ============
        
        ::
        
            $ pip install -U git+https://github.com/rth/pysofia.git
        
        
        Methods
        =======
        
        pysofia.train_svm
        
            Trains a model using stochastic gradient descent. See docstring for
            more details.
        
        pysofia.compat.RankSVM implements an estimator following the conventions
        used in scikit-learn.
        
        Development
        ===========
        
        Check out the latest version at github: https://github.com/rth/pysofia/
        
        License
        =======
        
        Apache License 2.0
        
        Authors
        =======
        
        PySofia is the work of Fabian Pedregosa, currently maintained by Roman Yurchak. The sofia-ml library is written by D. Sculley.
        
Platform: UNKNOWN
