March 2014 - graph-tool - archives.skewed.de

what could be causing a dlopen() issue when importing graph_tool?
by Lev Givon 28 Mar '14

28 Mar '14

I've installed the 2.2.31-1 graph-tool package for Ubuntu 13.10 x86_64 on my system. Although I can import it from within a regular Python 2.7.5 session without any problems, I can't seem to import it successfully from within a virtualenv in which I've installed a range of other packages (the virtualenv is configured to also access globally installed Python packages): >>> import graph_tool Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/lib/python2.7/dist-packages/graph_tool/__init__.py", line 100, in <module> dl_import("from . import libgraph_tool_core as libcore") File "/usr/lib/python2.7/dist-packages/graph_tool/dl_import.py", line 57, in dl_import exec(import_expr, local_dict, global_dict) File "<string>", line 1, in <module> ImportError: /usr/lib/python2.7/dist-packages/graph_tool/libgraph_tool_core.so: invalid mode for dlopen(): Invalid argument Before I manually go through all of the packages installed in the virtualenv, does anyone have any ideas as to what sort of things might be causing the above import problem? -- Lev Givon Bionet Group http://www.columbia.edu/~lev/ http://lebedov.github.io/

2 2

Iterating over the edges fast
by Thomas Capelle 26 Mar '14

26 Mar '14

Hello everyone, I am a NetworkX user thinking on moving to graph-tools. I developed some basic Vehicle routing algorithms in Networkx some time ago, but iterating throw the edges was so slow that medium size problems where really unattainable. (more than 50 nodes) I want to know if graph-tools is faster iterating throw nodes. This kind of label-setting algorithms need to go node by node and edge by edge constructing routes, and the "for" in python is really 'lagging'. Should I try the same in graph-tools? thanks in advance. Thomas -- View this message in context: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com… Sent from the Main discussion list for the graph-tool project mailing list archive at Nabble.com.

2 1

Latest Performance Comparison
by pl h 24 Mar '14

24 Mar '14

Hi, I just come across the old comparison for graph-tool, iGraph, and NetworkX http://graph-tool.skewed.de/performance It turns out many of the links are dead, e.g. detailed report "here", PGP graph data set, benchmark scripts. Where can I find latest comparison (report/ code/ data set) between graph-tool and other widely used packages? Thanks! Pili

2 1

Lattice Layout
by Caspar Yim 23 Mar '14

23 Mar '14

Hi, I'm doing some research on concept lattice. And I want to implement the program with graph_tool. Can I use graph_tool API to draw a lattice-like graph similar to the picture below? <https://lh5.googleusercontent.com/-rBOFzP4TlMc/Uyb_MfVxqKI/AAAAAAAABfU/36SH…> Thanks. Best regards, Caspar

2 3

implicit periodic domain
by DEQUIDT Alain 23 Mar '14

23 Mar '14

Hi, I wish to use graph-tool's periodic Delaunay triangulation generation. I guess that the periodic domain is the cube (0,0,0)-(1,1,1) and that all vertex coordinates should be inside the cube. Is that true? Best regards, A. Dequidt

2 1

Computation cost of creating a graph
by Hang Mang 22 Mar '14

22 Mar '14

Sorry for asking too much questions today about efficiency but it seems today is not my lucky day. I'm trying to build a graph by reading from a text file. The graph is 1034 vertices with 53498 edges. Creating this graph takes 6 seconds, so I was wondering if it truly takes that time or something is inefficient in my way of creating the graph. I have a text file that contains the edges in the format: vertex_label_x vertex_label_y. What I'm doing basically is maintaining a dictionary that maps from vertices labels in the text file to graph-tool vertices indices. When I read the file, I check if a label is in the dictionary, if so I retrieve its index from the dictionary. Otherwise I create a new vertex and store its index by the corresponding label in the dictionary. Then I create the edge using the indices of the source and target. I was wondering if my way is the wrong way to do that. This is my code: http://pastie.org/private/n9j7geizaosafmlimujca Many thanks in advance! Best.

2 2

inline and output
by Pietro Battiston 21 Mar '14

21 Mar '14

I saw - and happily tested - the new possibility of having inline figures in the ipython notebooks. Unfortunately (for my typical usage, in which I both save figures and show them in the notebook, which is easy with matplotlib and networkx), it seems like setting the "inline" argument makes the "ouput" argument irrelevant. Am I wrong? Is this an intended behaviour? Thanks, Pietro

2 4

Adding an edge by vertices indices?
by Hang Mang 21 Mar '14

21 Mar '14

Is it possible to add an edge between a source and a target using their indices?

2 3

Accessing a compressed matrix is super slow!
by Hang Mang 21 Mar '14

21 Mar '14

Hi, I found that accessing a compressed matrix is really slow. I'm computing a similarity index called LHN1, it took 34 seconds to compute when I access the 'paths' variable/matrix. But when I converted paths to paths.asarray() it only took 11 seconds. So now I'm really ending up calling toarray() all over my code. I'm not sure of why using compressed matrices in the first place or how I could overcome this! Forget about the get_degrees_dic() for now. Here's my code: def lhn1(graph): A = gts.adjacency(graph) paths = A**2 paths = paths.toarray() S = np.zeros(A.shape) degrees = get_degrees_dic(graph) for i in xrange(S.shape[0]): for j in xrange(S.shape[0]): i_degree = degrees[i] #graph.vertex(i).out_degree() j_degree = degrees[j] #graph.vertex(j).out_degree() factor = i_degree * j_degree S[i,j] = (1.0/factor) * paths[i,j] return S

2 1

Performance question
by Hang Mang 21 Mar '14

21 Mar '14

I have a graph with 1034 vertices and 53498 edges. I'm manually computing the preferential attachement index for the vertices, and other indices. I'm aware that graph-tool has that implemented but I'm doing it for personal stuff. However I noticed that my computations are very slow. It took 2.7 minutes to compute that for the mentioned graph. I'm not sure if it's my algorithm that is slow or the something is wrong with graph-tool. I would be very thankful if someone could have a little look into my code. def pa(graph): """ Calculates Preferential Attachment index. Returns S the similarity matrix. """ A = gts.adjacency(graph) S = np.zeros(A.shape) for i in xrange(S.shape[0]): for j in xrange(S.shape[0]): i_degree = graph.vertex(i).out_degree() j_degree = graph.vertex(j).out_degree() factor = i_degree * j_degree S[i,j] = factor return S

3 8