Tangos: the agile numerical galaxy organization system

Abstract: We present Tangos, a Python framework and web interface for database-driven analysis of numerical structure formation simulations. To understand the role that such a tool can play, consider constructing a history for the absolute magnitude of each galaxy within a simulation. The magnitudes must first be calculated for all halos at all timesteps and then linked using a merger tree; folding the required information into a final analysis can entail significant effort. Tangos is a generic solution to this information organization problem, aiming to free users from the details of data management. At the querying stage, our example of gathering properties over history is reduced to a few clicks or a simple, single-line Python command. The framework is highly extensible; in particular, users are expected to define their own properties which tangos will write into the database. A variety of parallelization options are available and the raw simulation data can be read using existing libraries such as pynbody or yt. Finally, tangos-based databases and analysis pipelines can easily be shared with collaborators or the broader community to ensure reproducibility. User documentation is provided separately.