Astro-TS3: Time-series Subimage Search Engine for archived astronomical data.

Efficiently retrieving a subset of time series data in a specific sky region from large-scale astronomical observation data is important for some time-domain studies in astronomy, especially for research on long-term variable sources. Considering that the field of view of a telescope is generally larger than the effective sky region, the retrieval method at the full image level is relatively inefficient in terms of data transmission and storage. Therefore, this paper proposes a search engine dedicated to the efficient retrieval of time-series subimages in observed image data archives (Astro-TS3, Astro nomical T ime- S eries S ubimage S earch Engine). An in situ and resilient indexing scheme is proposed, and it can build indexes for archived image data without any modification of the original files and support incremental updating with newly archived data. A subimage cutting tool is implemented to extract a part of the image according to the range specified by the user and generate a new Flexible Image Transport System (FITS) file for the extracted subimage. With Astro-TS3, users only need to submit requests including the ranges of the time and sky region to retrieve the required data and then download the time-series subimages after the preview. The evaluation of Astro-TS3 is performed using actual archived data from three Antarctic Survey Telescopes (AST3). The experimental data show that the time for the index establishment and incremental update increases approximately linearly relative to the number of data files. For retrieval, the response time of the data retrieval is approximately linearly related to the number and size of the subimages. [ABSTRACT FROM AUTHOR]