Ugropy : An Extensible Python Package for Thermodynamic Model Functional Group Identification via Mathematical Optimization

Abstract

Group contribution models are widely used for property estimation in chemical engineering. However, the identification of functional groups remains a challenge, particularly for large data sets and automated workflows. This work introduces an algorithm for functional group detection based on the Set Cover Problem, formulated as an integer linear programming optimization problem. The method ensures optimality, nonoverlapping assignments and avoids heuristic dependencies. The algorithm is implemented in ugropy, an open-source Python library supporting several group contribution models (Joback, UNIFAC, Dortmund, PSRK, Alshehri). Ugropy allows integration with cheminformatics and machine learning pipelines and supports user-defined models and solvers.