graph-tool

I was wondering if there is a built in function in graph_tool that can be used to find all shortest paths from node s to node t. If not, is there any way that I can use shortest_distance() (in module graph_tool.topology), or shortest_path() (in module graph_tool.topology) somehow (or any other built-in function)to compute all the shortest path instead of only one of them, efficiently (I am working with a graph that has around half a million nodes). There is no such function in graph-tool. Note that in general finding all shortest paths on a large graph will probably be unfeasible, since the

I have been using the graph-tool library for a while now, thus far I hadn't really been using most of its drawing features. Today when trying to use graph_tool.draw.graph_draw I realised that I had configured graph-tool initially without cairo , ./configure --disable-cairo and this is exactly the graphics library that graph_draw uses. If I install cairo now, is there a way to activate the features of graph_tool that rely on cairo without having to remove and reinstall the whole graph_tool module? This would spare a few hours of re-installation. (the make process takes about 3 hours on my

For my current project I want to use the graph-tool library since they claim being the fastest: https://graph-tool.skewed.de/performance . I have some algorithms (shortest path, etc.) to run on really large networks, so the faster the better! First question: Is this claim 'being the fastest' true? ;) While trying to build a graph-tool graph fitting my needs, I figured out that its not possible to access vertex properties in a efficient way. Maybe I missed something? My question is now, can the function "getVertexFromGraph(graph, position)" be written in a more efficient way? Or more in general

I am using python graph-tool. To draw graphs, it uses graph_draw function . I want to send vertex positions explicitly to dot engine. It turns out that I can pass a property map named pos . I tried defining it as v_pos = g.new_vertex_property("vector<double>") where g is my graph. I am not sure if it is the right way to do it. There is one code snippet which you might find helpful. pos = gt.random_layout(g, shape=shape, dim=3) >>> pos[g.vertex(0)].a array([ 86.59969709, 1.31435598, 0.64651486]) graph_draw(g, pos=pos, output="graph-draw-random.pdf") What should I do if I were to define my

I want to run Monte Carlo edge swaps, i.e., picking two existing edges in a graph uniformly at random and then (if some conditions are met) swap their end points. I am currently using boost::random_edge for selecting edges uniformly at random. #include <boost/graph/adjacency_list.hpp> #include <boost/graph/erdos_renyi_generator.hpp> #include <boost/random/mersenne_twister.hpp> #include <boost/random/variate_generator.hpp> #include <boost/graph/random.hpp> #include <boost/random/linear_congruential.hpp> int main(int argc, char* argv[]) { typedef boost::adjacency_list<boost::setS,boost::vecS

I'm trying to install graph-tool for Anaconda Python 3.5 on Ubuntu 14.04 (x64), but it turns out that's a real trick. I tried this approach , but run into the problem: The following specifications were found to be in conflict: - graph-tool Use "conda info <package>" to see the dependencies for each package. Digging through the dependencies led to a dead-end at gobject-introspection So I tried another approach: Installed boost with conda, then tried to ./configure, make, and make install graph-tool... which got about as far as ./configure: =========================== Using python version: 3.5.2