Unidirectional Build Architecture: Refactoring a HEP Data Analysis Codebase
Session Number
Project ID: PHYS 18
Advisor(s)
Dr. Peter Dong; Illinois Mathematics and Science Academy
Discipline
Physical Science
Start Date
19-4-2023 10:35 AM
End Date
19-4-2023 10:50 AM
Abstract
Proposed beyond the Standard Model theories, such as various SUSY frameworks, posit the existence of non-Standard Model particles detectable in accelerator experiments through distinct event signatures. The object of our analysis concerns two such non-SM theoretical particles: the doubly charged Higgs boson (H±±), a scalar boson with Drell-Yan and vector boson fusion production mechanisms, and the dark photon (γd), a dark sector gauge boson with Higgs and SUSY production mechanisms. Using collected tracking and calorimeter data from the CMS experiment, alongside simulated PYTHIA models, pp collision event data is sifted through series of statistical methods, identification algorithms, and filters to be coded and stored in the Data Collection directory. However, as the codebase expands, a single directory for such diverse tasks proves increasingly inefficient for code management and compilation times. We explore an alternative architecture for the H±± and γd searches through codebase refactoring, reorganizing the grouping and placement of methods, classes, and files. To this end, the Data Collection directory was refactored into seven task-specific directories—Utility, Histograms, Modules, Filters, Plans, EventFiles, and Data Collection—arranged with hierarchical include directives to prevent circular dependencies.
Unidirectional Build Architecture: Refactoring a HEP Data Analysis Codebase
Proposed beyond the Standard Model theories, such as various SUSY frameworks, posit the existence of non-Standard Model particles detectable in accelerator experiments through distinct event signatures. The object of our analysis concerns two such non-SM theoretical particles: the doubly charged Higgs boson (H±±), a scalar boson with Drell-Yan and vector boson fusion production mechanisms, and the dark photon (γd), a dark sector gauge boson with Higgs and SUSY production mechanisms. Using collected tracking and calorimeter data from the CMS experiment, alongside simulated PYTHIA models, pp collision event data is sifted through series of statistical methods, identification algorithms, and filters to be coded and stored in the Data Collection directory. However, as the codebase expands, a single directory for such diverse tasks proves increasingly inefficient for code management and compilation times. We explore an alternative architecture for the H±± and γd searches through codebase refactoring, reorganizing the grouping and placement of methods, classes, and files. To this end, the Data Collection directory was refactored into seven task-specific directories—Utility, Histograms, Modules, Filters, Plans, EventFiles, and Data Collection—arranged with hierarchical include directives to prevent circular dependencies.