Integrační testy pro spouštění komplexních výpočetních úloh

Abstract

This thesis focuses on designing automated integration and acceptance tests for the Apache Airflow orchestrator within the LEXIS distributed computing platform. Following a review of testing concepts and Python frameworks, pytest is selected for the implementation. A subsequent analysis covers the core LEXIS architecture, including authentication, distributed data management, and high-performance computing. The practical part introduces an innovative approach that maps individual DAGs directly to test scenarios. By combining native Airflow methods with the pytest framework, local test execution is achieved without requiring standard infrastructural components like the Scheduler or Webserver. The result is a resource-efficient, stable, and extensible testing suite covering both positive use cases and exhaustive negative testing to ensure system robustness.