FIDIA: A Format Independent Data Interface for Astronomy

We present a Python package for interacting with data for astronomical objects in a consistent and intuitive manner, regardless of the source or format of the data. Considerable past work has focused on standardised formats and protocols for working with data. However, these tools are structured around images and tables, not stars and galaxies. We take such work a step further by defining an interface to data on disk, on the web, in a database, etc, with a structure described by a vocabulary (based in part on existing vocabularies) for the physical properties of astronomical sources. In our Data Interface, the data are structured in samples of astronomical objects, such as stars and galaxies, with data members such as spectra, magnitudes and velocities. This structure provides a uniform model that is intuitive to the astronomer, simplifying analysis and visualisation. The package accepts that data will always be available in a wide variety of formats and places, and therefore it provides a formulaic way to connect to and import from new data sources such as on the user's computer. Cross matching is done between data sources as needed. Throughout, the package has been designed to deliver the most likely result (with a warning) in the face of ambiguity, such as multiple versions of data or multiple potential cross-matches, but allow for arbitrary, repeatable reconfiguration. The package takes care of mundane details such as memory management, local caching, unit conversions, etc. The astronomer can use this package to leverage existing data about their stars or galaxies in their analysis, or to rapidly apply existing analysis tools based on this package to new data.