Change page style: 

Astroconda now recommended for Gemini users, in place of Ureka

Astroconda is a set of astronomy-related packages for Anaconda, a software distribution that comes with many common dependencies —such as the complete SciPy stack and GUI toolkits —and has become very popular in the wider science community since we developed Ureka. While not quite as trivial to install, it is a more modular system, with its own "conda" package manager, allowing individual packages (such as Gemini IRAF) to be updated independently. The core Anaconda distribution is free, backed by a company with strong ties to the SciPy community, and serves as de-facto community platform to a greater extent than previous Python-related software collections have. Re-basing our work on this platform lets us focus limited resources on supporting astronomy-specific packages to a higher standard, brings better alignment with the Python developer community and provides users with more tools and flexibility in general.

Following some IRAF integration and testing work, we are now recommending that all new installations of Gemini's data processing software be performed using Astroconda, in place of Ureka. Detailed instructions can be found at http://astroconda.readthedocs.io (see "Configure Conda to use the Astroconda Channel" and "Legacy Software Stack (with IRAF)"), but in summary, the following steps will give you a set of IRAF and Python packages similar to Ureka's, including (as of writing) the latest Gemini IRAF 1.13.1 and Disco Stu 1.3.4:

  1. Install the latest 64-bit, Python 2.7 version1 of Anaconda from https://www.continuum.io/downloads (unless you already have a recent version installed). If you prefer, you may use Miniconda (omitting ~1GB of packages, out of ~7GB total).
  2. Make sure Anaconda is activated in your bash (or zsh)2 session:

    source ~/anaconda/bin/activate

    (or whichever path it was installed under).

  3. Configure conda to look for packages in the AstroConda "channel" at STScI:

    conda config --add channels http://ssb.stsci.edu/astroconda
    

    This is a one-time step, affecting current and future Anaconda installations belonging to the same user on the same machine.

  4. Install our packages into a new conda environment (with a name of your choosing after "-n", such as "astroconda"):

    conda create -n astroconda python=2.7 iraf-all pyraf-all stsci gemini

    If you add anaconda to the end of this command, you will also be able to access all the default packages distributed with anaconda when working in the "astroconda" environment (this is not 100% guaranteed to work in future, if there are package version conflicts, but we have not encountered any to date and you can always try and see whether the installer is happy). If you want only the minimum needed to run Gemini IRAF, you may replace the last four arguments with iraf.gemini iraf-x11 pyraf ds9, saving ~2GB.

  5. Whenever you would like to make the packages available in any given terminal session, type:

    source activate astroconda

    (or whatever name you gave the environment).

    If you answered "no" when the Anaconda installer asked to modify your login PATH at step 1, you might need to provide the full path to the "activate" script (as for step 2). You may add the above line to the end of your .bashrc if you want, but note that this can add a few seconds to your login time, will override any other python or IRAF installations in your user environment and will display a traceback if the new conda environment is ever removed.

Users of the new GMOS-N Hamamatsu CCDs can access the temporary Gemini IRAF 1.14 commissioning release in Astroconda PyRAF by following the instructions in its "gemini_readme.txt" file, to override the included version (as for Ureka).

For further information, see https://astroconda.readthedocs.io. Please submit any questions on using Gemini's software with Astroconda via the Gemini helpdesk (in the Gemini IRAF category). Support resources for more general questions (including Google groups for discussing Anaconda itself) are listed at http://astroconda.readthedocs.io/en/latest/resources.html.

Notes:

1. It should also be possible to work in a Python 3 environment, but a few packages, notably STSDAS & Kepler, contain PyRAF scripts that are only compatible with Python 2 and there is currently no "pyraf" build for Python 3.6 on MacOS.

2. Sorry, Anaconda does not support csh/tcsh —but if you really want to work in an unsupported shell, you can first "source activate" in bash and then start your shell under that environment.