Regressziós teszthalmaz optimalizációja függőségek alapján
With the expansion of DevOps, automated testing is a must in most of newly developed applications. For long-term projects this also means an evergrowing regression testset througout the development lifecycle.
Külső partner: Evosoft Hungary Kft.
Several methods are used, to shrink the testset including, but not limited to: risk-based, priority-based, machine learning-based and static code analysis-based solutions. Our project is developing a solution using the latter two approaches.
The goal of the team is to create a solution with a cascaded approach of filtering the regression set based on dependencies, then using machine learning to derive a minimal set of necessary tests.
The student will take part of this project to help achieving this goal, so the cost of regression testing can be reduced significantly.
The tasks of the student will include:
- creating a graph based on dependencies
- creating an algorithm optimized for reducing the testset, without losing valuable information, based on historical test results
- integrate the solution into our development pipeline