.. _setup: ****** Set Up ****** Currently you must run these scripts on a cobalt testbed or a submitter node. ------------------------- Accessing the Environment ------------------------- To make it easy to reproduce the results, I have created a virtual environment which is accessible on the cobalts. The virtual environment was installed using the python from the py2-v2 CVMFS environment: ``/cvmfs/icecube.opensciencegrid.org/py2-v2/RHEL_6_x86_64/bin/python``. Simply run .. code-block:: bash source enter_environment.sh in ``pev-photons`` to activate the environment. Alternatively, if on an EL7 machine, use: .. code-block:: bash source enter_environment.sh el_7 Note this redefines your python path, leaving the environment with ``deactivate`` will reset the python path to what it was before entering the environment. This will also add ``pev-photons`` to your python path. The exact package versions used are listed here: ============ ======= Package Version ============ ======= astropy 2.0.3 basemap 1.0.7 healpy 1.8.6 matplotlib 1.4.3 numpy 1.13.3 pandas 0.18.1 scikit-learn 0.17.1 scipy 0.15.1 skylab 2.0.2 ============ ======= You can create your own virtual environment with these packages using ``exact_environment.txt`` in ``pev-photons``. Within a freshly made virtual environment: .. code-block:: bash pip install -r exact_environment.txt If you go this route, make sure ``pev-photons`` is in your python path before proceeding. ------------------------- Setting up the File Paths ------------------------- The analysis scripts produce data files and plots which contain the results. The file ``pev-photons/utils/support.py`` has paths that should be changed to the location you want to store these files: 1. ``prefix``: the base directory where generated data files should be stored. 2. ``fig_dir``: the base directory where generated plots should be stored. 3. ``dag_dir``: the directory where dagman log files should be stored. Change these paths to your personal directories. Then, run: .. code-block:: bash python pev-photons/make_directories.py This makes the directory and all the subdirectories necessary in both the plot and data file locations if they don't already exist.