Installation

We distribute EOS as packaged software for end-users and also as source code for developers.

Linux

Installation with `pip3`

This is the recommended installation method using the ‘package installer for Python’ pip3. We recommend that EOS be installed within a virtual Python enviroment – see this FAQ answer if you do not know how to do this. We distribute a release version and a development version. The changelog of the curated release versions is available here. The development version is automatically up-to-date with the master branch in the Github EOS repository. Here we summarize the most common commands using pip3:

	Command
Install latest release version	`pip3 install --user eoshep`
Install latest development version	`pip3 install --user --pre eoshep`
Update to latest release version	`pip3 install --user --upgrade eoshep`
Update to latest development version	`pip3 install --user --upgrade --pre eoshep`

You can use the EOS Python module via import eos after installation. In addition, some command-line scripts have become available. You can test and debug the installation of the python module as described here.

Installation with `apt` (Ubuntu, Debian, and Siblings)

EOS is available through the Packagecloud web service. The installation requires several steps:

Create the file /etc/apt/sources.list.d/eos.list by running the following command in a shell

DIST=$(lsb_release -sc)
echo deb https://packagecloud.io/eos/eos/ubuntu/ $DIST main \
    | sudo tee /etc/apt/sources.list.d/eos.list

Add the official EOS repository’s GPG key to your package manager’s keychain by running:

curl -L "https://packagecloud.io/eos/eos/gpgkey" 2> /dev/null | apt-key add - &>/dev/null

You might need to install the curl program first, if not already installed. This can be done by running the following command:

apt-get install curl

Install the pre-built binary package and its dependencies by running the following commands:

apt-get update
apt-get install eos

On older versions of Ubuntu, you might need to update your installation of the Matplotlib Python package. We recommend installing the most up-to-date Python package from the Python Package Index, using

pip3 install --user --upgrade dynesty matplotlib networkx numpy pypmc pyyaml scipy

Windows

EOS can be installed and built under Windows using the ‘Windows Subsystem for Linux v2 (WSL2)’. Follow the instructions to setup the latter. In the process, you might be prompted what Linux distribution should be used. We recommend using ‘Ubuntu’. After WSL2 has been setup, you can follow the instructions for installing EOS under Linux without modifcations.

macOS

Installation with `brew`

EOS (and all of its dependencies) can be installed automatically using Homebrew:

brew tap eos/eos
brew install --HEAD eos

We recommend installing the most up-to-date Python package from the Python Package Index, using

pip3 install --user --upgrade dynesty matplotlib numpy pypmc pyyaml scipy

Build and Install from Source

The following instructions explain how to install EOS from source on a Linux-based operating system, such as Debian or Ubuntu, and macOS. Other flavors of Linux or UNIX-like operating systems will likely work as well. However, note that we will exclusively use Debian/Ubuntu-specific package names when listing software dependencies.

Installing the dependencies on Linux

The dependencies can be roughly categorized as either system software, Python-related software, or scientific software.

EOS requires the following system software:

g++: the GNU C++ compiler, in version 10.1 or higher;
autoconf: the GNU tool for creating configure scripts, in version 2.69 or higher;
automake: the GNU tool for creating makefiles, in version 1.16.2 or higher;
libtool: the GNU tool for generic library support, in version 2.4.6 or higher;
pkg-config: the freedesktop.org library helper, in version 0.27.1 or higher;
BOOST: the BOOST C++ libraries (sub-libraries boost-filesystem and boost-system);
yaml-cpp: a C++ YAML parser and interpreter library, in version 0.5.1 or higher.

The Python interface to EOS requires the additional software:

python3: the Python interpreter in version 3.7 or higher, and required header files;
pip: the Python package installer;
BOOST: the BOOST C++ library boost-python for interfacing Python and C++;
dynesty: The Python Dynamic Nested Sampling package for estimating Bayesian posteriors and evidences (needed for some functionality);
matplotlib: the Python plotting library in version 2.0 or higher;
networkx: the Python graph library (needed for some functionality);
scipy: the Python scientific library;
pypmc: the Python library for adaptive importance sampling with Markov Chain and Population Monte Carlo methods (needed for some functionality);
PyYAML: the Python YAML parser and emitter library;
wilson: the Python library for matching, translating, and running Wilson coefficients in the Weak Effective Theory and the Standard Model Effective Theory (needed for some functionality).

We recommend you install the above packages via your system’s software management system.

EOS requires the following scientific software:

GSL: the GNU Scientific Library GSL, in version 1.15 or higher;

If you have administrator access to the computers on which you use EOS, we recommend you install the above packages via your system’s software management system.

On a Debian/Ubuntu-based operating system you can install the prerequisite software with the apt package management system, by running the following commands:

# for the 'System Software'
sudo apt-get install g++ autoconf automake libtool pkg-config libboost-filesystem-dev libboost-system-dev libyaml-cpp-dev
# for the 'Python Software'
sudo apt-get install python3-dev libboost-python-dev python3-matplotlib python3-networkx python3-scipy python3-yaml python3-pip
# for the 'Scientific Software'
sudo apt-get install libgsl0-dev

We strongly recommend the use of a python virtual environment (see this FAQ answer for more information about virtual environments) to install the dynesty, matplotlib, networkx, scipy, pypmc, pyyaml, and wilson packages:

# Make sure to activate your virtual environment first
pip install dynesty matplotlib networkx scipy pypmc pyyaml wilson

Installing the dependencies on macOS with Homebrew and PyPi

You can install most of the prerequisite software via Homebrew. You will need to make Homebrew aware of the EOS third-party repository by running the following command in a shell

brew tap eos/eos

To install the packages, run the following commands in a shell:

# for the 'System Software'
brew install autoconf automake libtool pkg-config boost yaml-cpp
# for the 'Python Software'
brew install python3 boost-python3
# for the 'Scientific Software'
brew install gsl

You can now use the pip3 command to install the remaining packages from the PyPi package index. We strongly recommend the use of a python virtual environment (see this FAQ answer for more information about virtual environments) for this.

Note

Due to problems with the Python 3 installation provided by macOS, we strongly recommend using instead the pip3 program provided by Homebrew, which should be available as /usr/local/bin/pip3.

To install the remaining packages, run the following command in a shell

pip3 install dynesty matplotlib networkx scipy pypmc pyyaml wilson

Installing EOS

You can obtain the EOS source code from the public Github repository. To download it for the first time, clone the repository by running the following command:

git clone -o eos -b master https://github.com/eos/eos.git

To install from the source code repository, you must first create all the necessary build scripts by running the following commands:

cd eos
./autogen.bash

You must now decide where EOS will be installed. To proceed we require you to set the environment variable PREFIX. We recommend installing to the standard directory of your python virtual environment, which is typically given by the $VIRTUAL_ENV environment variable. To do this, run the following command:

export PREFIX=$VIRTUAL_ENV

Next, you must configure the EOS build using the configure script. To use the EOS Python interface you must pass --enable-python to the call configure. The default is --disable-python.

The recommended configuration is achieved by running the following command:

./configure \
    --prefix=$PREFIX \
    --enable-python

If the configure script finds any problems with your system, it will complain loudly.

The flag with-boost-python-suffix might be necessary, depending on the installation of Python and BOOST. For example, when boost-python3 is installed on macOS via brew, you can find the suffix by inspecting the installed libraries: ls /usr/local/lib/libboost_python* might yield /usr/local/lib/libboost_python39.a. Here, the flag --with-boost-python-suffix=39 is required.

After successful configuration, build EOS by running the following command:

make -j all

The -j option instructs the make program to use all available processors to parallelize the build process. (Depending on the amount of RAM in your machine, you might have to use a smaller number of cores when running make, e.g. doing -j4 will only use four cores). We strongly recommend testing the build by running the command

make -j check VERBOSE=1

within the build directory. Please contact the authors if any test fails by opening an issue in the official EOS Github repository. If all tests pass, install EOS by running the command

make install # Use 'sudo make install' if you install e.g. to 'PREFIX=/usr/local'
             # or a similarly privileged directory

If you installed EOS to a non-standard location (i.e. not $VIRTUAL_ENV), to use it from the command line you must set up some environment variable. For BASH, which is the default Debian/Ubuntu shell, add the following lines to \$HOME/.bash_profile:

export PATH+=":$PREFIX/bin"
export PYTHONPATH+=":$PREFIX/lib/python3.8/site-packages"

Note that in the above the python3.8 piece must be replaced by the appropriate Python version with which EOS was built. You can determine the correct value by running the following command:

python3 -c "import sys; print('python{0}.{1}'.format(sys.version_info[0], sys.version_info[1]))"

You can test and debug the installation of the python module as described here.