- 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.org/project/gammapy
Quick install guide¶
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.
You can install the development version of Gammapy like this:
python -m pip install --user git+https://github.com/gammapy/gammapy.git
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: Get set up
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__)
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
specifically from inisde Jupyter notebooks, but also in general it’s a good introduction
how Python package installation with
If you have any questions or issues, don’t hesitate to ask on the Gammapy mailing list!
Once you’ve made your choice how to install Gammapy, you can find detailed information on the following sub-pages:
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).