PySpecKit Projects ================== A few projects appropriate for a Google Summer of Code or similar are outlined below. Refactor and Expand the pyspeckit modeling tools ------------------------------------------------ Since the development of pyspeckit, there has been substantial progress on a more general class of `modeling tools from astropy `_. Pyspeckit already has a wide variety of data fitting and modeling tools that can readily be modified to use the astropy modeling formalism. Details of this project need to be worked out, but will include: * refactoring pyspeckit.models to use astropy.models * building a graphical interface to astropy.models Refactor and Expand the Unit Test suite --------------------------------------- Pyspeckit is a complicated code suite, which has led to many bugs, particularly in the UI. An improved unit test suite would help prevent or remove these bugs. Such a project would start by breaking down the existing tests, which are really end-to-end tests, into their component units. The refactor would be to use astropy's wrappers around pytest to make the testing smoother and easier to extend. The current model has tests in a different repository because the data got too big, but we can and should have the *data* in a different repository, but the tests all in one place. Incorporate astropy.units into pyspeckit.units ---------------------------------------------- *This project was done by an ESO-hosted summer student, Dinos Kousidis.* This project is at the core of both `pyspeckit` and `specutils `_. The most important base functionality of a spectroscopic toolkit is to be able to convert between different spectroscopic systems, e.g. wavelength<->frequency<->velocity. This can be achieved using astropy_'s `unit equivalencies `_. The X-axis unit changes will be straightforward project that should require about 2 weeks to complete. The more complicated and interesting project is creating Y-axis units (i.e., flux units) that appropriately adjust with changes to the X-axis. These would make use of other astropy_ unit equivalencies, e.g. the `spectral density `_ equivalency. The end goal will be to have a `Spectrum` object that will live in `specutils`_ and be inherited by `pyspeckit`, which will provide the interface to modeling and graphical tools.