IMPORTANT: This repository is not used anymore for the MachineLearning code

Please have a look at MachineLearningHEP, also for a proper installation of the prerequisites.

This repository is still used for:

• preprocessing the Tree output (downloading-skimming-merging);
• creating the cut objects for HF analyses;
• development on the TTreeCreator and/or TreeHandler classes (but please note that the version can be old with respect to what is in AliPhysics/PWGHF/treeHF/).

Old README from now on

MachineLearningHF

The current recommended instructions require the installation of a independent ROOT6 framework with Pyhon 3.6. Using the package inside AliPhysics is possible but not described in this README.

Follow:

• step 1a) and 2) if you are using MacOs Sierra
• step 1b) and 2) if you are using Ubuntu 18.04

1a) Prerequisites for MacOs Sierra 10.13.16 (Validated 1 december 2018)

Install git

pip3 install git

Install python 3.6.6

Get from the official website https://www.python.org/downloads/mac-osx/ the version Python 3.6.6. Follow the instructions and install it. The predefined folder should be /Library/Frameworks/Python.framework/Versions/3.6/bin/python3.

Install ROOT with Python 3

This assumes that you are using the python version you just installed placed in /Library/Frameworks/Python.framework/Versions/3.6/bin/python3. If you want to use another version change the path accordingly.

git clone http://github.com/root-project/root.git
cd root
git checkout -b v6-10-08 v6-10-08
cd ..
mkdir build
sudo chown -R $(whoami):$(id -g -n $(whoami)) build
cd build
cmake -DPYTHON_EXECUTABLE=/Library/Frameworks/Python.framework/Versions/3.6/bin/python3 ../root/
make -j<num_core_to_use>
source bin/thisroot.sh

Add in your bash_profile (if you use Terminal)

This assumes you have the build in your home folder. If another choice was made change it accordingly

source build/bin/thisroot.sh

Install all the Machine Learning softwares

pip3 install numpy pandas scipy matplotlib seaborn
pip3 install pkgconfig uproot
pip3 install scikit-learn sklearn-evaluation xgboost
pip3 install keras
pip3 install -Iv tensorflow==1.5
brew install graphviz

1b) Prerequisites for Ubuntu 18.04 (Validated 7 december 2018)

The instruction below requires you to start the installation from your home directory!

Install cmake and other utilities

sudo apt-get update
sudo apt-get install git dpkg-dev cmake g++ gcc binutils libx11-dev libxpm-dev libxft-dev libxext-dev

You might need also:

sudo add-apt-repository universe
sudo apt-get install build-essential

Install python 3.6.6

sudo apt-get install python3.6 python3-tk python3-pip

Install ROOT with Python 3

This assumes that you are using the python version you just installed placed in /usr/bin/python3. If you want to use another version change the path accordingly.

git clone http://github.com/root-project/root.git
cd root
git checkout -b v6-10-08 v6-10-08
cd ..
mkdir build
sudo chown -R $(whoami):$(id -g -n $(whoami)) build
cd build
cmake -DPYTHON_EXECUTABLE=/usr/bin/python3 ../root/
make -j<num_core_to_use>

Add in your .bashrc

source build/bin/thisroot.sh

This assumes you have the build in your home folder. If another choice was made change it accordingly

Install all the Machine Learning softwares

sudo apt-get install graphviz
pip3 install numpy pandas scipy matplotlib seaborn
pip3 install pkgconfig uproot
pip3 install scikit-learn sklearn-evaluation xgboost
pip3 install tensorflow keras

You might need also this in case you get errors:

pip3 install --upgrade setuptools

To install tensorflow with GPU support please refer to https://www.tensorflow.org/install/gpu

2) Install the ALICE Machine learning tool

Configure your github environment

Please use your own name and mail address :D

git config --global user.name "Your name"
git config --global user.email mymail@mail.com

Download the ML package and some test files

git clone https://github.com/ginnocen/MachineLearningHF.git
cd MachineLearningHF/ALICEanalysis/MLproductions
scp <my_cern_user>@lxplus.cern.ch:/afs/cern.ch/work/g/ginnocen/public/exampleInputML/*.root .

In case of problems:

For problems ginnocen@cern.ch, fabio.catalano@cern.ch