Date: | 5 April 2012 |
---|---|
Authors: | Chris Fonnesbeck, Anand Patil, David Huard, John Salvatier |
Contact: | chris.fonnesbeck@vanderbilt.edu |
Web site: | http://github.com/pymc-devs/pymc |
Copyright: | This document has been placed in the public domain. |
License: | PyMC is released under the Academic Free license. |
Version: | 2.2 |
PyMC is known to run on Mac OS X, Linux and Windows, but in theory should be able to work on just about any platform for which Python, a Fortran compiler and the NumPy module are available. However, installing some extra depencies can greatly improve PyMC’s performance and versatility. The following describes the required and optional dependencies and takes you through the installation process.
PyMC requires some prerequisite packages to be present on the system. Fortunately, there are currently only a few hard dependencies, and all are freely available online.
There are prebuilt distributions that include all required dependencies. For Mac OS X and Windows users, we recommend the Enthought Python Distribution. The Enthought Python Distribution comes bundled with these prerequisites. Note that depending on the currency of these distributions, some packages may need to be updated manually.
If, instead of installing the prebuilt binaries, you prefer (or have) to build pymc yourself, make sure you have a Fortran and a C compiler. There are free compilers (gfortran, gcc) available on all platforms. Other compilers have not been tested with PyMC but may work nonetheless.
The easiest way to install PyMC is to type in a terminal:
easy_install pymc
Provided EasyInstall (part of the setuptools module) is installed and in your path, this should fetch and install the package from the Python Package Index. Make sure you have the appropriate administrative privileges to install software on your computer.
Pre-built binaries are available for Windows XP and Mac OS X. These can be installed as follows:
- Download the installer for your platform from PyPI or the GitHub download page.
2. Double-click the executable installation package, then follow the on-screen instructions.
For other platforms, you will need to build the package yourself from source. Fortunately, this should be relatively straightforward.
First download the source code from GitHub and unpack it. Then move into the unpacked directory and follow the platform specific instructions.
One way to compile PyMC on Windows is to install MinGW and MSYS. MinGW is the GNU Compiler Collection (GCC) augmented with Windows specific headers and libraries. MSYS is a POSIX-like console (bash) with UNIX command line tools. Download the Automated MinGW Installer and double-click on it to launch the installation process. You will be asked to select which components are to be installed: make sure the g77 compiler is selected and proceed with the instructions. Then download and install MSYS-1.0.exe, launch it and again follow the on-screen instructions.
Once this is done, launch the MSYS console, change into the PyMC directory and type:
python setup.py install
This will build the C and Fortran extension and copy the libraries and python modules in the site-packages directory of your Python distribution.
Some Windows users have reported problems building PyMC with MinGW, particularly under Enthought Python. An alternative approach in this case is to use the gcc and gfortran compilers that are bundled with EPD (located in the Scripts directory). In order to do this, you should add the EPD “Scripts” directory to your PATH environment variable (ensuring that it appears ahead of the MinGW binary directory, if it exists on your PATH). Then build PyMC using the install command above.
Alternatively, one may build the currently-available release of PyMC using pip.
In a terminal, type:
python setup.py config_fc --fcompiler gfortran build
python setup.py install
The above syntax also assumes that you have gFortran installed and available. The sudo command may be required to install PyMC into the Python site-packages directory if it has restricted privileges.
In addition, the python-dev package may be required to install PyMC on Linux systems.
You can check out the development version of the code from the GitHub repository:
git clone git://github.com/pymc-devs/pymc.git
Previous versions are available in the /tags directory.
pymc comes with a set of tests that verify that the critical components of the code work as expected. To run these tests, users must have nose installed. The tests are launched from a python shell:
import pymc
pymc.test()
In case of failures, messages detailing the nature of these failures will appear. In case this happens (it shouldn’t), please report the problems on the issue tracker (the issues tab on the Google Code page), specifying the version you are using and the environment.
Report problems with the installation, bugs in the code or feature request at the issue tracker. Comments and questions are welcome and should be addressed to PyMC’s mailing list.