Re: [boost] boost.graph python bindings build against RC-1-34

From: John Reid (j.reid_at_[hidden])
Date: 2007-01-09 12:38:28

Yes they are thank you very much.

However on a slightly separate note I was expecting the spectrum code to 
calculate the eigenvalues of the laplacian. However I think it 
calculates the eigenvalues of the adjacency matrix which is what some 
authors define as the graph spectrum.

I have some python (numpy) code that calculates the eigenvalues of the 
laplacian of a boost.graph in case anyone is interested in it. I needed 
it to calculate diffusion kernels I think it has some applications in 
partitioning as well.

My reference for this:
http://math.ucsd.edu/~fan/cbms.pdf

Cheers,
John.

Doug Gregor wrote:
> On Dec 29, 2006, at 6:47 AM, John Reid wrote:
> 
> 
>>Right FWIW here are patches to fix these compilation problems.
>>
>>Add
>>   using namespace boost::detail;
>>as line 30 in sparse_spectrum.cpp
>>
>>Replace lines 94 and 95 in sparse_spectrum.hpp with
>>     typedef boost::numeric::ublas::compressed_matrix<double> Matrix;
> 
> 
> Thanks, John! I've committed these to the BGL-Python Subversion  
> respository.
> 
> Are the BGL-Python bindings working for you now?
> 
> 	Cheers,
> 	Doug
> 
> 
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
>