.. include:: ../references.txt

.. _install:

************
Installation
************

* Gammapy works with legacy Python (version 2.7) as well as Python 3 (version 3.5 or above).
* The core dependencies are Numpy and Astropy, as well as for now the regions package
  (until it is merged into Astropy core).
* The main optional dependencies are PyYAML, Scipy, healpy, uncertainties and Sherpa
  (only imported when used).
* Linux and Mac OS are fully supported.
* Windows is fully supported by Gammapy, but not by Sherpa (which is used as a backend for
  modeling and fitting in Gammapy), so on Windows only part of the Gammapy functionality is available.
* You can always check what the latest stable release of Gammapy is here: https://pypi.python.org/pypi/gammapy

Quick install guide
===================

Stable version
--------------

You can install the latest stable version of Gammapy with conda::

    conda install -c conda-forge gammapy

or with pip::

    python -m pip install gammapy

or with Macports (a package manager for Mac OS)::

    sudo port install gammapy

Gammapy is not yet available in the Linux distributions, i.e. at this time you can't
install it with e.g. ``apt-get`` or ``yum``.

Development version
-------------------

You can install the development version of Gammapy like this::

    python -m pip install --user git+https://github.com/gammapy/gammapy.git

This will ``git clone`` the Gammapy repository from Github into a temp folder
and then build and install Gammapy from there.

If there are any errors related to Cython, Numpy or Astropy, you should install
those first and try again::

    conda install -c cython numpy astropy click regions
    python -m pip install --user git+https://github.com/gammapy/gammapy.git

How to get set up for Gammapy development is described here: :ref:`dev_setup`

Verify
------

To verify that Gammapy is installed and available,
and to check it's version and where it's located, type this::

    $ python

    >>> import gammapy
    >>> print(gammapy.__version__)
    >>> print(gammapy.__path__)


Need help?
==========

If you're not sure how to best install Gammapy on your machine (e.g. whether to use
conda or pip or Macports ...), we recommend that you give conda a try first.
It's a binary package manager (so generally installation is fast), and allows you to
install any software in your home folder (without needing ``sudo``) and works the
same on Linux, Mac OS and Windows. Many Gammapy users and developers use conda
and are happy with it.

There is a nice blog post `Installing Python Packages from a Jupyter Notebook`_ that
explains how to install Python packages in general with ``pip`` and ``conda``,
specifically from inisde Jupyter notebooks, but also in general it's a good introduction
how Python package installation with ``pip`` and ``conda`` work.

If you have any questions or issues, don't hesitate to ask on the `Gammapy mailing list`_!

Detailed information
====================

Once you've made your choice how to install Gammapy, you can find detailed information
on the following sub-pages:

.. toctree::
    :maxdepth: 1

    conda
    pip
    macports
    other
    check
    dependencies


If you'd like to make a code contribution to Gammapy, please see the developer documentation
for information how to get set up (e.g. to code, run tests, generate docs locally).