The recent rise in popularity of data-driven and machine learning approaches in surface hopping dynamics highlights the need for tools that support the entire lifecycle of trajectory data once simulations are completed. To address this, we present SHNITSEL-tools, a Python package designed to facilitate the standardized handling, analysis, and visualization of surface hopping trajectory data. SHNITSEL-tools provides a unified framework for reading outputs from commonly used programs such as SHARC, Newton-X, PyrAI2MD, and ASE, and for converting them into a consistent, NetCDF-based storage format with an in-memory representation built on xarray. In addition to data standardization, SHNITSEL-tools offers a wide range of functionalities for filtering, statistical analysis, and exploratory visualization, enabling efficient inspection and comparison of trajectory datasets. Processed and raw data can be stored in a structured and reproducible way, facilitating long-term data preservation and reuse in accordance with the FAIR data principles. Furthermore, standardized outputs generated with SHNITSEL-tools are intended to be curated within the publicly available SHNITSEL-data database, fostering transparency, interoperability, and the development of new data-driven approaches in surface hopping dynamics.
Dr. Kevin Höllring