Skip to main content
Seminar | X-Ray Science Division

Tomviz: Data Analysis, Visualization and Reproducibility

APS Scientific Computation Seminar

Abstract: The Tomviz project is an open-source desktop application licensed under the permission 3-clause BSD license developed primarily in C++ with Qt and using Python-based processing pipelines to offer a cross-platform desktop application that offers hardware-accelerated visualization on Windows, macOS, and Linux. It leverages the Visualization Toolkit (VTK) for state-of-the-art visualization and uses the scientific Python stack to offer familiar Python data analysis tools available in a user-friendly graphical user interface. The analysis can be extended at runtime with a built-in editor where new code can be developed and tested interactively within the interface.

The current capabilities of the project will be discussed, along with the ambitious extension of Tomviz to better serve the beamline community. This includes plans for improved packaging within the Conda ecosystem, which will enable deeper integration with community tools such as TomoPy, Bluesky, and DataBroker. New capabilities are being added to the C++ application and the Python interfaces to facilitate deeper integration of X-ray imaging techniques commonly used at synchrotrons, and ongoing work will offer graphical pipeline builders that can save workflows which then can be run in batch.

The use and extension of a powerful JSON-based state file that can now be embedded within a HDF5 file to offer reproducibility of all data processing, analysis, and visualization settings will be shown, and as will how it relates to offering a tool that can be used to do better science designed for transparency and reproducibility only possible with open-source tooling. Standardization of data representation and standards for sharing and dissemination are essential in order to develop tools like Tomviz and promote the principles of open science within the user facility communities.