Fully automated integral field spectrograph pipeline for the SEDMachine: pysedm

Current time domain facilities are discovering hundreds of new galactic and extra-galactic transients every week. Classifying the ever-increasing number of transients is challenging, yet crucial to furthering our understanding of their nature, discovering new classes, and ensuring sample purity, for instance, for Supernova Ia cosmology. The Zwicky Transient Facility is one example of such a survey. In addition, it has a dedicated very-low resolution spectrograph, the SEDMachine, operating on the Palomar 60-inch telescope. This spectrograph's primary aim is object classification. In practice most, if not all, transients of interest brighter than ∼19 mag are typed. This corresponds to approximately 10-15 targets a night. In this paper, we present a fully automated pipeline for the SEDMachine. This pipeline has been designed to be fast, robust, stable and extremely flexible. PYSEDM enables the fully automated spectral extraction of a targeted point source object in less than five minutes after the end of the exposure. The spectral color calibration is accurate at the few percent level. In the 19 weeks since PYSEDM entered production in early August of 2018, we have classified, among other objects, about 400 Type Ia supernovae and 140 Type II supernovae. We conclude that low resolution, fully automated spectrographs such as the "SEDMachine with pysedm" installed on 2-m class telescopes within the southern hemisphere could allow us to automatically and simultaneously type and obtain a redshift for most (if not all) bright transients detected by LSST within z < 0.2, notably potentially all Type Ia Supernovae. In comparison with the current SEDM design, this would require higher spectral resolution (R ≳ 1000) and slightly improved throughput. With this perspective in mind, pysedm is designed to easily be adaptable to any IFU-like spectrograph.

https://github.com/MickaelRigault/pysedm