<scp>prose</scp>: a <scp>python</scp> framework for modular astronomical images processing

Abstract

ABSTRACT To reduce and analyse astronomical images, astronomers can rely on a wide range of libraries providing low-level implementations of legacy algorithms. However, combining these routines into robust and functional pipelines requires a major effort that often ends up in instrument-specific and poorly maintainable tools, yielding products that suffer from a low level of reproducibility and portability. In this context, we present prose, a python framework to build modular and maintainable image processing pipelines. Built for astronomy, it is instrument-agnostic and allows the construction of pipelines using a wide range of building blocks, pre-implemented or user-defined. With this architecture, our package provides basic tools to deal with common tasks, such as automatic reduction and photometric extraction. To demonstrate its potential, we use its default photometric pipeline to process 26 TESS candidates follow-up observations and compare their products to the ones obtained with AstroImageJ, the reference software for such endeavours. We show that prose produces light curves with lower white and red noise while requiring less user interactions and offering richer functionalities for reporting.