An array-oriented Python interface for FastJet

Abstract Analysis on HEP data is an iterative process in which the results of one step often inform next. In exploratory analysis, it common to perform computation a collection events, then view (often with histograms) decide what try Awkward Array Scikit-HEP Python package that enables analysis array-at-a-time operations implement cuts as slices, combinatorics composable functions, etc. However, most C++ libraries, such FastJet, have imperative, one-particle-at-a-time interface, would be inefficient and goes against grain logic scientific Python. Therefore, we developed fastjet, pip-installable provides FastJet binaries, classic (particle-at-a-time) new array-oriented interface for use Array. The streamlines interoperability software beyond HEP, machine learning. case, adopting this library along other tools accelerated code by factor 20. It was designed easily integrated libraries ecosystem, including Uproot (file I/O), hist (histogramming), Vector (Lorentz vectors), Coffea (high-level glue). We discuss design fastjet library, integrating array oriented Lorentz vector operations. open source.