October 5-9, 2014


P2.15 Sherpa - On the Move to Open, Collaborative Development

Omar Laurino (SAO)

O. Laurino (SAO), A. Siemiginowska (SAO), J. Evans (SAO), T. Aldcroft (SAO), D. J. Burke (SAO), J. McDowell (SAO), W. McLaughlin (SAO), D. Nguyen (SAO)

Sherpa is the Chandra Interactive Analysis of Observations (CIAO) modeling and fitting application. Written in Python, with efficient C, C++, and Fortran extensions, Sherpa enables the user to construct complex models from simple definitions and fit those models to data, using a variety of statistics and optimization methods. In time, Sherpa evolved from a very X-ray specific analysis package to a more general-purpose fitting engine with advanced capabilities, and was also used as a backend for the development of new applications like Iris, the Virtual Astronomical Observatory spectral energy distribution builder and analyzer. However, building and installing Sherpa as a standalone Python package was problematic, and such a build would not maintain all of the Sherpa capabilities. For version 4.7 Sherpa’s build scripts have been completely rewritten, standardized, and made independent of CIAO, so that Sherpa can now be built as a fully functional standalone Python package, and yet allow users the flexibility they need in order to build Sherpa in customized environments. This work is part of a larger framework that includes migrating the whole Sherpa codebase to GitHub, so that the wider astronomical community can be engaged in its development, and providing developers with a clean software framework for extending Sherpa. Sherpa will be available in binary form using the Conda package manager, and as a source distribution that can be easily installed using standard means like pip and setuptools. Advanced configurations enable users to link Sherpa against optimized versions of the Fast Fourier Transform package FFTW, as well as to configure Sherpa to interface to the XSpec models distributed by HEASARC. This new version of Sherpa can thus be easily used along with other python tools and software, like ipython notebooks and astropy.

Mode of presentation: poster

Applicable ADASS XXIV theme category: Data Analysis / Pipelines