Hi Tiago,

Thanks for the blazing fast answer!

It will sure do the job. I can't find it in the documentation though (it's accessible online via ipython autocompletion and magic '?', but doesn't appear on your web site).

Cheers,

Guillaume



Le 17/12/2012 11:14, Tiago de Paula Peixoto a écrit :
On 12/17/2012 11:11 AM, Guillaume Gay wrote:
Hi list.

I am using graph-tool to model an /apical junctions/ network, based on the model by Farhadifar et al. <http://www.sciencedirect.com/science/article/pii/S0960982207023342>. In short, it corresponds to the outer surface of the cells in a particular region of the embryo, where each cell is represented by a polygon. Topology of the network changes due to cell division and cell death.

The physical model involves the local minimization of an energy depending on the local geometry.

I got a working model, but the energy optimization is time consuming. I think that this is due to the way I compute the components of the vector formed by each edge of the graph, by iterating over the edges., eg:

|for edge in graph.edges():
        v0, v1 = edge.source(), edge.target()
        deltax[edge] = x[v1] - x[v0]
        deltay[edge] = y[v1] - y[v0]|

This is further complicated (to a little extent) by periodic conditions to the coordinate systems, and some other details, but I think this is the core issue.

I have the feeling I can do better, perhaps by using the adjacency matrix? This would be particularly interesting as the graph topology doesn't change for a given optimization.

Hi Guillaume,

Take a look at the edge_difference() function, since I think it does
what you want, and should be significantly faster.

Cheers,
Tiago




_______________________________________________
graph-tool mailing list
graph-tool@skewed.de
http://lists.skewed.de/mailman/listinfo/graph-tool