Organised by:
in collaboration with:



social media:


Accesso Utenti

LAM - Laboratoire d'Astrophyisique de Marseille, Marseille
Miscellaneous Information:

Abstract Reference: 31147
Identifier: P1.4
Presentation: Poster presentation
Key Theme: 1 Reduction and Analysis Algorithms for Large Databases and Vice-versa

Ariadne: a system for evaluation of AMAZED's spectroscopic redshift estimation efficiency

Borges Renato C., Arnouts Stéphane, Chabaud Pierre-Yves, Fauchier Florian, Gray Morgan, Jamal Sara, Le Brun Vincent, Le Fèvre Olivier, Schmitt Alain, Surace Christian, Vibert Didier, Vidal Clément

AMAZED (Algorithms for Massive Automatic Z Evaluation and Determination) is a project developed at LAM (Laboratoire d'Astrophysique de Marseille) for supplying automated spectroscopic redshift estimation software for next generation spectrographs such as PFS (Prime Focus Spectrograph) and Euclid. Redshift can be estimated by cross-correlating spectral data and templates or by identifying key features in a spectrograph, which are then matched to relevant astrophysical models (spectroscopic templates). Both strategies are present in AMAZED's redshift estimation library, which is written in C++ with a pipeline based on Boost modules. It performs a bayesian reliability assessment of its results, which can be used to flag estimation quality. AMAZED's efficiency is well-defined for the set of spectrographs for which we have reference redshift values. Here we present Ariadne, an automation subsystem for AMAZED, whose primary task is to compute AMAZED's redshift estimation efficiency. Secondarily, it serves as a testing harness to facilitate AMAZED's development. Ariadne is implemented at LAM in Python 3, importing widely used libraries such as ZeroMQ, SQLAlchemy and PyQt. It is developed using Agile practices and designed to be highly modular. Ariadne has modules for interprocess communication (named Pépin), for database, storage and batch system interface (named Minos) and for user interface (named Ariadne). Minos' database stores spectrograph sets, configuration parameters with their allowed variation intervals, and a list of repository tags (which is updated by Jenkins). In a future work, we will add to Minos an automation module (to be called Tauros). Then, Ariadne will give a multi-threaded interface to aid the user exit AMAZED with an estimation of its efficiency, and kill Minos' Tauros if its process hangs.