Description
Beginning this year, the LHCb experiment at CERN’s Large Hadron Collider will process 4 TBytes of data per second in real-time. Around a dozen algorithms will reconstruct the raw detector hits into particle trajectories, followed by O(1000) selections which aggregate these trajectories into composite objects with topologies relevant for different physics analyses and classify them as interesting for further offline processing. We present PyConf, a functional python framework for automatically, reliably, and efficiently managing the complex data and control flow dependencies inherent in LHCb’s real-time processing. Through PyConf’s uniform and comprehensible API, physicists can express selections, debug them, and get their output without manually manipulating the data and control flow.