Skip to content
Switch branches/tags

Latest commit

* update of hepmc3 to latest master

* fix

* fix

* Fixed error in declaration of overload_cast

* Fixed error in declaration of overload_cast

* fix

* Removed wrong const

* build only master and develop

* Initial setup for building wheels on Azure

* backup .ci

* Squashed '.ci/azure-wheel-helpers/' content from commit 02274e7

git-subtree-dir: .ci/azure-wheel-helpers
git-subtree-split: 02274e749eaeb9903b185c015b0d2f6d633a2913

* Removed .ci backup

* Update requirements

* Submodules updated to latest version

* Check the submodule stuff

* Test azure builds again

* Test azure builds again

* removed HepMC3 from Hans fork

* Added HepMC3 from CERN

* Finished binary wheel builds on Azure DevOps

* Triggers on tags and pr on master

Co-authored-by: Hans Dembinski <>

Git stats


Failed to load latest commit information.
Latest commit message
Commit time


A Python wrapper for the HepMC3 C++ library.

PyPI version Build Status Build Status

Another wrapper is pyhepmc. Why should you use this one?

pyhepmc-ng is easy to install

The command pip install pyhepmc-ng should work on all Python versions > 2.7 and all common architectures.

Under the hood, the bindings are build with the excellent pybind11 library. External installations of pybind11 or HepMC3 are not required, pyhepmc-ng includes the lightweight source code of both libraries with the submodule feature of git.

pyhepmc-ng is actively developed

pyhepmc-ng is part of the Scikit-HEP project, which aims to provide all tools needed by particle physicists to do data analysis in Python. It is also gets official support from the HepMC3 project.

pyhepmc-ng is unit tested

Everything in pyhepmc-ng is unit tested.

pyhepmc-ng is Pythonic

pyhepmc-ng is a hand-crafted mapping of C++ code to Python. It supports Python idioms where appropriate.

  • C++ methods which act like properties are represented as properties, e.g. GenParticle::set_status and GenParticle::status are mapped to a single GenParticle.status field in Python
  • Tuples and lists are implicitly convertible to FourVectors
  • Vectors of objects on the C++ side are mapped to Python lists
  • ReaderAscii and WriterAscii support the context manager protocol
  • A convenient open function is provided for reading and writing HepMC files


pyhepmc-ng currently has no separate documentation, but it mirrors the C++ interface of the HepMC3 library, which is documented here:

For developers

Repository management

If you want to contribute to the source code, please follow these instructions. Start by forking the scikit-hep repository, then clone your fork to your local compute with these commands (replace YourName with your Github username):

git clone --recursive

Now cd to the project folder (the rest assumes you are in the project folder). The command clones the pyhepmc-ng repository and its nested sub-repositories. If you already cloned the fork without the --recursive option, you need to manually initialize the nested sub-repositories:

git submodule update --init

Add a remote endpoint called upstream to keep in sync with the master of the scikit-hep repository:

git remote add upstream

This concludes the initial set up.

To develop a feature or a fix, create a branch from your master (make sure your master is in sync with the scikit-hep master):

git checkout -b my_cool_feature master

Commit to your branch and initiate a pull request from the Github web page when you feel the feature is ready to be reviewed. Note: Never commit to the master, only to feature branches.

The scikit-hep master may have moved forward in the meantime. Keep your local master branch in sync with these commands:

git checkout master
git pull upstream master
git submodule update # update the nested sub-repositories if necessary

If you have followed the rule to never commit to the master, then these commands always work. Rebase your feature branch onto the updated master:

git checkout my_cool_feature
git rebase master

If conflicts between your changes and those in the master appear, you need to resolve them. Follow the instructions printed by git.

Build the package

pyhepmc-ng depends on other Python packages. We recommend to use a virtual environment for development which is isolated from your system-wide Python installation. Install a virtual environment in the project folder:

pip install --user virtualenv # only needed if you don't have virtualenv already
virtualenv py37 -p python3.7 # or use another Python version

Activate the virtualenv and install the required packages for development:

. py37/bin/activate
pip install -r requirements.txt

Now build the package in develop mode.

python develop

This should work since pyhepmc-ng is continously tested on recent versions of gcc, clang and msvc. If it does not, please submit an issue with the build log and your compiler version. You can now change the source code. Run the previous command again to build the project after you made changes. Finally, run the unit tests:

pytest tests

These should all pass. If you add new features, don't forget to add unit tests for them.

To leave the virtualenv, call deactivate or close the shell.

Install your local version

If you want to use your local version for productive work, pip-install it from within the local project folder:

pip install --user --upgrade .

The --user option is not necessary when you are inside a virtualenv or if you have write-permission to the system-wise Python directories. The --upgrade option makes sure that an already existing pyhepmc-ng version is replaced.


pyhepmc-ng is covered by the BSD 3-clause license, see the LICENSE file for details. This license only applies to the pyhepmc-ng code. The connected external libraries HepMC3 and pybind11 are covered by other licenses, as described in their respective LICENSE files.