Install

Describes how to install Bolt on your machine.

Requirements

Bolt depends on the following python packages:

  • numpy (tested with version 1.1 - 1.5)

To check your installed version use:

python -c "import numpy; print numpy.__version__"

To build Bolt you need a c compiler (e.g., gcc) and the header files of numpy.

On Ubuntu systems the following should suffice:

sudo apt-get install python-numpy

The installation has been tested on Windows 7 using the Enthought Python Distribution. If you get undefined symbol errors, make sure that the lib_dir path in setup.py is properly set (for EPD 6.2 use C:\Python26\PCbuild).

If you want to modify the Cython code in the extension modules in bolt.trainer make sure that the newest release of Cython (>= 0.7) is installed.

To generate the documentation you need Sphinx and the following packages

On Ubuntu systems the following should suffice:

sudo easy_install sphinx
sudo apt-get install dvipng
pip install -e git+git://github.com/michaeljones/sphinx-to-github.git#egg=sphinx-to-github

If you don’t have pip install it via:

sudo easy_install pip

Stable release

Download the latest release from github:

wget http://github.com/pprett/bolt/tarball/v1.4 -O bolt_v1.4.tar.gz

Untar the archive:

tar xzvf bolt_v1.4.tar.gz
mv pprett* bolt

Build bolt:

cd bolt
python setup.py build

Install bolt sytem wide:

sudo python setup.py install

If the build process fails due to undefined references make sure that the include paths in setupy.py are properly defined.

If you need assistance please contact the author at peter.prettenhofer@gmail.com.

Latest development version

To fetch the latest development version from the repository type:

git checkout git://github.com/pprett/bolt.git

To build the project use:

python setup.py build

To install bolt system wide use:

sudo python setup.py install

Table Of Contents

Previous topic

Bolt Online Learning Toolbox

Next topic

API Overview

This Page