David Gleich

In response to Feixiong, when any of the shortest_path algorithms have a target set, the search stops when it first finds the vertex. This does not guarantee that the shortest path is correct, but it's the first path found to the vertex. In light of your comment, I plan to revisit this behavior in a future version. If you require the actual shortest paths, then you should not use the target option in MatlabBGL 4. I'm not planning to provide a patch for this until a new version is released unless I hear from others that it's a source of considerable pain. Please contact me if you need a temporary (patched) shortest_path.m file that would transparently (but potentially inefficiently) address the issue.

Thanks for the comments. I've intended to update this code for more recent versions of Matlab for a while, but somehow never find the time.

I think I can do it with a reweighted graph and kamada_kawai_spring_layout. However, my edge weights vector is failing the following test in the MEX file:

if (!mxIsDouble(arg_ewopt) < nz) {
mexErrMsgIdAndTxt("matlab_bgl:invalidParameter",
"The reweight array must be of type double");

even though in Matlab the array is a double. I'm on 64 bit Matlab, and I'm guessing that the mex file 'thinks' I'm on 32.

Any ideas ?

a gem!