Martin Baláž
Dr Martin Baláž develops algorithms and software tools with the A* Vienna software group.
Currently he is mainly focusing on the data reduction pipeline for METIS,
a first-generation instrument at the Extremely Large Telescope.
Martin has a background in planetary science and also develops and maintains
the IT backend of AMOS, a network of all-sky meteor cameras.
Session
In 2023, ESO approved the use of Python in instrument software packages, enabling developers to work at a higher level of abstraction than is possible with legacy C code. This can substantially improve code reuse and allow greater modularity through declarative and object-oriented programming.
Object-oriented design offers several notable advantages for pipeline development. Data items, their relationships and processing steps can be declared, reused, and extended with minimal effort. Human-readable descriptions of data items and algorithms remain closely integrated with the code, reducing the need to maintain separate documentation. Through use of introspection, automated testing for compliance with the specification can also be integrated into the development process, helping prevent inconsistencies. Most importantly, this approach enables us to decouple the general principles of data-reduction pipelines from the scientific objectives of individual instruments. Pipeline developers can therefore focus on instrument-specific requirements without the necessity to pay attention to the complexities of managing the data flow.
We present a generic Python framework for building data-processing pipelines, developed collaboratively by A* software group at the University of Vienna and Academia Sinica Institute of Astronomy and Astrophysics in Taipei, Taiwan. The framework is fully compatible with the existing ESO software stack (EDPS, pyesorex) while keeping dependencies to minimum in order to allow independent use. At the same time, it is already being employed to implement the data-reduction pipeline for METIS, one of the first-generation instruments of the Extremely Large Telescope, potentially paving the way for broader adoption across future instruments.