I originally interned at LLNL over the summer of 2017 (click hear to learn more about my group), and joined part-time during university to support the software side of the project longer-term. To facilitate me working at the national lab part-time during university I am contracted under J&A Film Preservation, a company that has been working with the Film Scanning and Re-Analysis Project for several years. I am continuing to develop vision algorithms for scientific measurements, but I am also working on streamlining our team’s workflow by using the latest workflow automation tools like Dagster.
Over the past 2 years working for the Film Scanning and Re-analysis project I have added several new measurement tools to our measurement tool written in Python. These tools include global-segmented film exposure measurement methods, and a a new fireball detection pipeline that I architected utilizing spectral analysis and temporal edge tracking. I also added measurement statistics for all of the data our group generates using scikit-learn to better understand our measurements. In addition, I have refactored our codebase to explicitly support both CLI and GUI (Tk) methods to run which will allow the team to programmatically re-analyze and update old results more easily.
From the refactoring sprint, I was able to abstract our application logic to work well on both desktop environments for developing new codes and on HPC environments provided by Livermore Computing’s supercomputers. Dagster, a workflow orchestration tool, ties our workflow together with its pipeline and asset management systems, and I developed a multithreaded image loader as a CPython module to further speed up our program’s analysis.