Getting Started with PyModelFit
===============================


Requirements
------------
Before you do anything with PyModelFit, you'll need:

    * `Python <http://www.python.org/>`_ 2.5 or higher (2.6 highly recommended), although 3.x is not yet supported.
    * `numpy <http://numpy.scipy.org>`_ 
    * `scipy <http://www.scipy.org/>`_
    
Follow the instructions on those sites, or, far easier, install them as packages from your operating system (e.g. apt-get or the synaptic GUI on Ubuntu, `Macports <http://www.macports.org/>`_ on OS X, etc.).  


Install
-------

Once you have the requirements satisfied, you have a few options for installing pymodelfit.  

.. note::
    On most unix-like systems, you will need to either execute these commands as the root user, or preface them with ``sudo``.

If you have `pip <http://pypi.python.org/pypi/pip>`_ (the new, better easy installer) or `easy_install/setuptools <http://pypi.python.org/pypi/setuptools>`_ (you should probably install pip...), just run either::

    pip install pymodelfit

or::

    easy_install pymodelfit

If you are installing from source code, instead, just do::

    python setup.py install


Additional Recommended Packages
-------------------------------

A few other outside packages add useful functionality to PyModelFit.  These
are not required, but are highly recommended for the full PyModelFit 
experience.  

For general publication-quality plotting of data and models, the
following package is suggested:

    * `Matplotlib <http://matplotlib.sourceforge.net/index.html>`_
        Can be installed with ``pip install matplotlib``
    
To use the useful :class:`~pymodelfit.fitgui.FitGui` interactive model-fitting
tool, the following three packages are necessary (with the suggested install
command, if you don't have the package already available on your system):

    * `Traits <http://code.enthought.com/projects/traits/>`_
        ``pip install traits``
    * `TraitsGUI <http://code.enthought.com/projects/traits_gui/>`_
        ``pip install traitsGUI``
    * `Chaco <http://code.enthought.com/projects/chaco/>`_
        ``pip install chaco``
    
For the :class:`~pymodelfit.fitgui.MultiFitGui` model-fitting tool, another
package is needed for 3D rendering:

    * `Mayavi2 <http://code.enthought.com/projects/mayavi/>`_
        ``pip install mayavi``

Finally, Markov Chain Monte Carlo (MCMC) fitting is available for models via the
:func:`pymodelfit.FunctionModel.getMCMC` method.  To work, this requires an 
additional package:

    * `pymc <http://code.google.com/p/pymc/>`_
        ``pip install pymc``
        

Developing PyModelFit
---------------------

PyModelFit is an open source project, and contributions are welcome. The `github 
project page <http://github.com/eteq/pymodelfit>`_ has the development
version of the source code, as well as the bug tracker.   Code guidelines are 
the same as for the :mod:`astropysics` project, and are detailed 
`at that project's documentation page.
<http://packages.python.org/Astropysics/develop.html>`_

You can get the latest version by installing `git
<http://git-scm.com/>`_ and running the command:

    git clone git://github.com/eteq/pymodelfit.git 
    
then install it via:

    python setup.py install
    
or:

    python setup.py develop
    
Use the second option if you want to easily edit the source code and immediately see your changes in
the installed version.