AstroSkyFlow: an astronomical sky image flow simulator for time domain survey validation and machine learning

Read full text
Authors: Kexin Li, Yicheng Rui, Fabo Feng, Shuyue Zheng, A. V. Pomazan, Yiyang Guo, Jie Zheng, Lin-Qiao Jiang
Cited by: 1

Abstract

Frozen release of the AstroSkyFlow software associated with the AAS journal manuscript. AstroSkyFlow is a modular photometric image simulator designed to generate high-fidelity, multi-epoch astronomical image sequences for time-domain survey validation, signal-injection experiments, and machine-learning applications. This software release contains the frozen version of AstroSkyFlow associated with the AAS journal manuscript. The repository is organized into three main components. The modular_core/ directory contains the main simulator code, configuration files, observing schedules, event-injection files, reference-data directory structure, and output structure. The tests/ directory provides minimal executable examples for lightweight software validation. The reproduce_paper/ directory contains configured workflows and supporting materials for reproducing the manuscript validation examples, including the Muguang-transit and Xinglong-binary cases. AstroSkyFlow uses input configuration files, observing schedules, reference photometric data, and event files to generate simulated time-series images. Supported injected signals include transits, eclipsing binaries, flares, occultations, supernova-like eruptions, satellite trails, and user-defined scattered-light inputs. The simulator outputs FITS images and corresponding injected-variable files for comparison, validation, and downstream analysis. Large reference star and galaxy catalog files are distributed separately because of their file size and are not included in this software release. Users who wish to run the full workflows or reproduce the manuscript examples should download the required reference catalogs from the associated Zenodo reference dataset: 10.5281/zenodo.18830766. After downloading, the relevant files should be placed in the appropriate reference_data/ directory, such as modular_core/reference_data/, reproduce_paper/Binary_Xinglong/reference_data/, or reproduce_paper/Transit_Muguang/reference_data/. For detailed installation instructions, configuration options, file-structure information, usage examples, and workflow descriptions, please refer to the included README.md file.

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
24 April 2026