[SciPy-user] ANN: Python Equations 2.0 - weaveless version

zunzun@zunzu... zunzun@zunzu...
Fri Sep 28 10:01:26 CDT 2007


On Fri, Sep 28, 2007 at 10:38:25AM -0400, Alan G Isaac wrote:
> 
> This is a pretty sparse description.

I really didn't think too many people
would be interested, and did not want
to blather on and on to the mailing list.


> Could you elaborate a bit, and maybe offer
> an example of the usage?

Sure.  I have a curve and surface fitting web site
http://zunzun.com and have released the actual
fitting parts as a separate independant module.
If someone wants to run the code on their own
equipment, for example because their computers
are faster than the web site's shared server
or to implement some equation I do not yet have
on the site, then they can.

Originally I used weave and C++ for performance,
but thanks to Robert Kern's suggestions I've
been able to remove the requirement for a C++
compiler and use numpy natively - hence the
new version, which requires only python, numpy
and scipy.

The package comes with quite a few examples, below
is a simple linear 3D surface fit with Java
source code output - this is included in the
package under the Examples directory.  For
readability here I have removed the imports
and comments that are in the example.

   James


equation = PythonEquations.Equations3D.Polynomial.Linear3D()

equation.fittingTarget = 'SSQABS'
equation.ConvertTextToData(equation.exampleData)
equation.Initialize()

equation.FitToCacheData() # perform the fit

print equation.CodeJAVA() # output Java source code



More information about the SciPy-user mailing list