ScientificPython
- Version 2.5.11
The current stable release is 2.6 and the current development version is 2.7.3, which supports Numpy. The 2.7.2 release notes state: "NumPy compatibility. Scientific_netcdf was revised by hand. The Python code was run through numpy.oldnumeric.alter_code1 to identify the critical sections, which were then all handled in some way. It is possible that there are still incompatibilities of the kind that numpy.oldnumeric.alter_code1 cannot detect" (See 2.7.2 Release notes)
Ran python -c "import numpy.oldnumeric.alter_code1 as noa; noa.converttree('.')" in top-level directory of source tree.
- Fixed Scientific/N.py to only import from numpy.oldnumeric
- Fixed Scientific/LA.py to only import from numpy.oldnumeric.linear_algebra
- Fixed setup.py to only use numpy and get the include directory using numpy.get_include()
Patch is here: scientific_python_2.5.11_for_numpy_patch
MMTK
- Version 2.5.13
Ran python -c "import numpy.oldnumeric.alter_code1 as noa; noa.converttree('.')" in top-level directory of source tree. Use svn version (not 1.0rc1 version).
- Fixed Include/MMTK/numpy.pxi to use "numpy/oldnumeric.h" instead of "Numeric/arrayobject.h"
- Fixed Src/MMTK_energy_term.pyx to use "import numpy.oldnumeric; Numeric=numpy.oldnumeric"
From the Src directory, ran pyrexc -I ../Include MMTK_energy_term.pyx to regenerate the extension module.
- Changed setup.py so that result of numpy.get_include() is placed in list of include directories for all extension modules.
Patch is here: mmtk_2.5.13_for_numpy_patch2
PyGame
- Revision 983 out of SVN
- Changed Numeric/arrayobject.h to numpy/oldnumeric.h in sndarray.c and surfarray.c
- Eliminated all code that set the flags of the ndarray. This code is un-needed.
- fixed imports to use the numpy.oldnumeric package
- fixed the soften demo in examples/arraydemo.py because it assumed that scalar
multiplication would promote the array to a larger integer type. In NumPy scalar multiplication does not promote arrays.
Changed the DependencyPython class to use the get_include() function of numpy to
- find the include file (why isn't a standard setup.py file being used...)
Patch is here: pygame_r983_for_numpy_patch
RPy
- Revision 341 out of SVN
- Changed Numeric/arrayobject.h to numpy/oldnumeric.h in RPy.h
Changed all NUMERIC, numeric, Numeric names to NUMPY, numpy, and NumPy respectively
- Fixed Python code to use numpy.oldnumeric instead of Numeric
Added include_dir.append(numpy.get_include()) to the setup.py file if NumPy is found
Patch is here: rpy_341_patch_for_numpy
Should be applied the 1st week of February 2007 (see email from Gregory Warnes)
Nurbs
Nurbs-0.1 from http://runten.tripod.com/NURBS/
- Changed Numeric/arrayobject.h to numpy/oldnumeric.h in _Bas.c
- Changed all Numeric imports to numpy.oldnumeric
- Rewrote the 2D plotting in matplotlib
- Changed setup.py to use numpy.distutils
Replaced all occurrences of N.less(x,0) to N.any(x < 0), same for greater
- 3D surfing -- only OpenGL examples run, the rest depend on dislin
Patch is here: nurbs_0.1_patch_for_numpy
PyGtk
Revision 2808 from http://svn.gnome.org/svn/pygtk/trunk
- Changed Numeric/arrayobject.h to numpy/arrayobject.h in gtk/gdk.override
- Fixed configure.in to use numpy.get_include() as the place to look for numpy/arrayobject.h
Note: Probably need to add the numpy.get_include() to the list of include directories as well but I don't know how to do this.
- Changed import Numeric to import numpy in setup.py
Patch is here: pygtk_r2808_patch_for_numpy