Installation

Prerequisites

The SkyLLH framework has several dependencies. They are listed in requirements.txt file:

astropy
numpy
scipy
iminuit
matplotlib

They can be installed from skyllh directory with:

pip install -r requirements.txt

On cobalt and NPX servers we can use CVMFS Python 3 virtual environment with all necessary packages already installed. In order to activate it run:

eval `/cvmfs/icecube.opensciencegrid.org/py3-v4.1.1/setup.sh`

Setup

Using pip

The latest skyllh release can be installed from PyPI repository:

The current development version can be installed using pip:

pip install git+https://github.com/icecube/skyllh.git#egg=skyllh

Optionally, the editable package version with a specified reference can be installed by:

pip install -e git+https://github.com/icecube/skyllh.git@[ref]#egg=skyllh

where

  • -e is an editable flag

  • [ref] is an optional argument containing a specific commit hash, branch name or tag

Cloning from GitHub

The framework is split into two packages:

  1. github.com/icecube/skyllh

  • contains open source code with classes defining detector independent likelihood framework

  1. github.com/icecube/i3skyllh

  • contains collections of pre-defined SkyLLH IceCube analyses and pre-defined IceCube datasets

In order to set it up, we have to clone git repositories and add them to the PYTHONPATH:

git clone git@github.com:icecube/skyllh.git
git clone git@github.com:icecube/i3skyllh.git
export PYTHONPATH=$PYTHONPATH:/path/to/skyllh
export PYTHONPATH=$PYTHONPATH:/path/to/i3skyllh

Alternatively, we can add then inside python script:

import sys

# Add skyllh and i3skyllh projects to the PYTHONPATH
sys.path.insert(0, '/path/to/skyllh')
sys.path.insert(0, '/path/to/i3skyllh')