Skip to content

PUMA: Plotting UMami Api#

Code style: black Umami docs PyPI version DOI

codecov Testing workflow Linting workflow Pages workflow Docker build workflow

The Python package puma provides a plotting API for commonly used plots in flavour tagging.

ROC curves Histogram plots Variable vs efficiency

Installation#

puma can be installed from PyPI or using the latest code from this repository.

Install latest release from PyPI#

pip install puma-hep

The installation from PyPI only allows to install tagged releases, meaning you can not install the latest code from this repo using the above command. If you just want to use a stable release of puma, this is the way to go.

Install latest version from GitHub#

pip install https://github.com/umami-hep/puma/archive/main.tar.gz

This will install the latest version of puma, i.e. the current version from the main branch (no matter if it is a release/tagged commit). If you plan on contributing to puma and/or want the latest version possible, this is what you want.

Docker images#

The Docker images are built on GitHub and contain the latest version from the main branch.

The container registry with all available tags can be found here.

The puma:latest image is based on python:3.11.10-bullseye and is meant for users who want to use the latest version of puma. For each release, there is a corresponding tagged image. You can start an interactive shell in a container with your current working directory mounted into the container by using one of the commands provided below.

On a machine with Docker installed:

docker run -it --rm -v $PWD:/puma_container -w /puma_container gitlab-registry.cern.ch/atlas-flavor-tagging-tools/training-images/puma-images/puma:latest bash

On a machine/cluster with singularity installed:

singularity shell -B $PWD docker://gitlab-registry.cern.ch/atlas-flavor-tagging-tools/training-images/puma-images/puma:latest

The images are automatically updated via GitHub and pushed to this repository registry.