Summer of Code 2010 Ideas

Partially recycled from previous years' project ideas.

• improve datasource and integrate it into all the numpy/scipy io http://projects.scipy.org/scipy/numpy/browser/trunk/numpy/lib/_datasource.py
• clean up, refactor scipy package structure: scipy.lib, scipy.misc, scipy.stsci (already gone) - is this still relevant? no discussion on it
• modernize, clean-up scipy.weave, integrate fast-vectorize
• Statistics Cleanup
• scipy.ndimage: Rewrite in Python where possible, port to Cython elsewhere. Decide on a consistent coordinate framework. As a bonus, fix boundary issues.
  • Leverage patches from CellProfiler developers for this, see e.g. patches with ticket #945
• Automatic Differentiation - there are existing implementations and work in progress (algopy, openopt)

Ideas with significant existing momentum

• Port SciPy to Python 3.0 (or continue finishing up the porting of NumPy)
  • Improve numpy.f2py test coverage and fix any issues that crop up on Python 3.
  • Port Scipy to Python 3
    • Set up a build framework that uses 2to3 (see e.g. how it is done in Numpy)
    • After that, modify Scipy so that it works both on Python 2 and Python 3 -- this can be done one submodule at a time, so the work can be partitioned in manageable pieces. Some submodules will be more work than others.
    • Knowledge of C is required in porting some of the submodules. There are, however, also pure-Python submodules.
    • Some information on: changes made in porting Numpy
• datetime types - see the NEP and recent SVN versions of NumPy
• Continue work on Fwrap started by Kurt Smith in GSoC2009
• integrate Jonathan Taylor's statistical models into scipy.stats - see scikits.statsmodels (Skipper Seabold, GSoC2009), current task for this is improving scikits.statsmodels