OpenSBLI is a Python-based modelling framework that expands a set of differential equations written in Einstein notation and automatically generates C code to obtain a solution. OpenSBLI depends on a number of external dependencies, which are often difficult to configure correctly and lacked automated testing to confirm correct installation and prevent against regression.
We developed an integration test framework to improve the installation process for the code. We also improved usability by updating documentation on both the testing framework and the installation process.
“With Ed, we have been able to tackle some of the wider issues that will make it easier to install and maintain the software in the future. His software engineering skills were useful in devising a strategy towards a continuous integration testing approach”
Professor Neil Sandham
Aerodynamics and Flight Mechanics Group, School of Engineering