Cloud computing and minimal recomputation for Casa.
The Casa data processing package, widely used in radio astronomy; Jupyter, a system for hosting science data processing on the World Wide Web and Recipe (Recipe Tarball), developed during the RadioNet Hilado project.
In data-processing astronomical data it is common to rerun scripts with slight changes as understanding of the data set is refined.
The Recipe system takes Casa scripts (in python) and translates them to graph form; a recomputation-elimination engine can then compare this graph with previous iterations of the script.
Our goal in Obelics is to develop this system further and integrate it with a web-hosted edition of the Casa package to transparently allow efficient cloud-based data processing.