Installing manually

If you're scikit-learn user:

type in bash

pip install rep --no-dependencies

is the simplest way to start, though you will need to do some actions from below later, when you try to use other ML libraries.

Prerequisites

Assuming you have clean Ubuntu 14.04:

sudo apt-get install python-dev libblas-dev libatlas-dev liblapack-dev gfortran g++ python-setuptools

Matplotlib:

sudo apt-get install libpng-dev libjpeg8-dev libfreetype6-dev libxft-dev

Update pip:

which pip || sudo easy_install pip
sudo pip install --upgrade pip

Install Python dependencies:

sudo pip install -r https://github.com/yandex/rep/raw/master/requirements.txt

REP Installation

export LC_ALL=C
git clone git@github.com:yandex/rep.git
pip install rep

Additional stuff

uBoost and other uniforming classifiers, reweighters, neural networks and different stuff for HEP:

pip install hep_ml

TMVA library:

1. install ROOT 5.XX.XX https://root.cern.ch/drupal/content/installing-root-source

2. add to ~/.bashrc

source PATH_TO_ROOT/libexec/thisroot.sh

3. Install additional libraries to work with ROOT

pip install root_numpy

XGBoost library:

pip install xgboost==0.4a30

For more detailed instructions (they will work on Linux) with conda environment, see https://github.com/yandex/rep/blob/master/ci/install_rep_environment.sh