Speaker
Description
SIRIUS, the 4th generation synchrotron light source operated by the Brazilian Synchrotron Light Laboratory (LNLS/CNPEM), currently hosts multiple beamlines with different experimental techniques and continuously evolving software demands. As beamline activities and data processing requirements grow in complexity, challenges related to interoperability, maintainability, software reuse and collaboration between scientific software groups have become increasingly relevant.
Within this context, SIRIUS beamlines have progressively developed ad hoc solutions for data processing and task automation according to their demands. These solutions, often implemented as custom Python scripts and beamline-specific integrations, successfully addressed local requirements but evolved independently, with limited standardization and varying levels of sophistication.
Current efforts at SIRIUS investigate how workflow technologies can be incorporated into a broader control and data architecture designed to be modular, reusable and interoperable across beamlines. Ongoing discussions explore workflow engines not only as automation tools for processing pipelines, but also as potential building blocks for orchestration layers connecting data acquisition, processing, ingestion, analysis and higher-level decision-making components within the Bluesky ecosystem.
The proposed architecture aims to support workflows spanning multiple stages of the experimental lifecycle, including data pre-processing, transformation into standardized application definitions, data ingestion, post-processing, and analysis tasks. Different workflow technologies are currently being evaluated according to their potential roles within the ecosystem, including Prefect, Airflow and Ewoks. These efforts also seek to support future adaptive and autonomous experimental procedures, ranging from conditional execution and automated region-of-interest selection to AI-assisted analysis and decision-making workflows.
