Peter K. G. Williams, Casey J. Law, Geoffrey C. Bower
New and upgraded radio interferometers produce data at massive rates and will require significant improvements in analysis techniques to reach their promised levels of performance in a routine manner. Until these techniques are fully developed, productivity and accessibility in scientific programming environments will be key bottlenecks in the pipeline leading from data-taking to research results. We present an open-source software package, miriad-python, that allows access to the MIRIAD interferometric reduction system in the Python programming language. The modular design of MIRIAD and the high productivity and accessibility of Python provide an excellent foundation for rapid development of interferometric software. Several other projects with similar goals exist and we describe them and compare miriad-python to them in detail. Along with an overview of the package design, we present sample code and applications, including the detection of millisecond astrophysical transients, determination and application of nonstandard calibration parameters, interactive data visualization, and a reduction pipeline using a directed acyclic graph dependency model analogous to that of the traditional Unix tool "make". The key aspects of the miriad-python software project are documented. We find that miriad-python provides an extremely effective environment for prototyping new interferometric software, though certain existing packages provide far more infrastructure for some applications. While equivalent software written in compiled languages can be much faster than Python, there are many situations in which execution time is profitably exchanged for speed of development, code readability, accessibility to nonexpert programmers, quick interlinking with foreign software packages, and other virtues of the Python language.
View original:
http://arxiv.org/abs/1203.0330
No comments:
Post a Comment