2026-11-04 –, Conference Foyer 1
Time-domain astronomical observations continuously generate large volumes of data that require repeated processing through multiple steps before science-ready data products can be produced and released. As the number of telescopes and observing systems increases, maintaining telescope-specific processing scripts becomes difficult to reuse, monitor, and extend. We present a configurable and reusable batch-processing pipeline designed to support the processing and release of time-domain astronomical data.
The pipeline separates telescope-specific configurations from common processing logic and organizes data reduction into modular steps with clearly defined inputs, outputs, parameters, and quality metrics. Processing states, logs, and data provenance are recorded for each step, allowing users to identify successfully processed data, locate failed tasks, and rerun individual processing stages when necessary. The pipeline is designed for portable deployment in virtualized or standalone computing environments and can be adapted to different telescopes through configuration and processing modules.
Processed science data products are connected to an online data release service that supports data query, visualization, and download. A natural-language query interface based on NL2SQL is also introduced to facilitate data discovery without requiring users to directly construct SQL statements. The current implementation is evaluated using time-domain observational datasets from miniST and the ST prototype, demonstrating the feasibility of a reusable processing and data-service workflow for future multi-telescope applications.
Linying Mi works at the National Astronomical Data Center (NADC), National Astronomical Observatories, Chinese Academy of Sciences (NAOC), and is currently pursuing a PhD in astronomy. Her research interests include astronomical data processing, time-domain astronomy, reusable scientific workflows, and intelligent science data services.