March 2017 - graph-tool - archives.skewed.de

number of hierarchy levels running off to infinity
by Peter Straka 27 Mar '17

27 Mar '17

Dear list, I'm trying to fit a nested blockmodel to a (bipartite) network with ~10^6 edges. The algorithm minimize_nested_blockmodel_dl() doesn't terminate; it keeps on adding layers indefinitely until it runs out of memory / hits walltime, see below. What could be going on? Many thanks in advance, Peter level 2 : replaced (94, 1) -> (94, 16) , dS: -15452.0134871 3 l=3 B: 8 <- 16 shrinking 16 -> 11 l=3 B: 8 <- 16 B=11 niter: 1 count: 1 breaks: 1 min_S: 1002.5742 max_S: 1002.5742 S: 1002.5742 ΔS: 0.00000 moves: 0 l=3 B: 8 <- 16 shrinking 11 -> 8 l=3 B: 8 <- 16 B=8 niter: 1 count: 1 breaks: 1 min_S: 838.70438 max_S: 838.70438 S: 838.70438 ΔS: 0.00000 moves: 0 l=3 Current bracket: (2, 8, 16) (691.21840695494018, 838.70438205706921, 1340.6798309766764) l=3 B: 5 <- 8 shrinking 8 -> 5 l=3 B: 5 <- 8 B=5 niter: 1 count: 0 breaks: 0 min_S: 752.39171 max_S: 753.34067 S: 752.39171 ΔS: -0.948956 moves: 2 l=3 B: 5 <- 8 B=5 niter: 2 count: 0 breaks: 0 min_S: 751.86303 max_S: 753.34067 S: 751.86303 ΔS: -0.528683 moves: 1 l=3 B: 5 <- 8 B=5 niter: 3 count: 1 breaks: 1 min_S: 751.86303 max_S: 753.34067 S: 751.86303 ΔS: 0.00000 moves: 0 l=3 Current bracket: (2, 5, 8) (691.21840695494018, 751.86302859686145, 838.70438205706921) l=3 B: 3 <- 5 shrinking 5 -> 3 l=3 B: 3 <- 5 B=3 niter: 1 count: 0 breaks: 0 min_S: 720.12311 max_S: 720.62000 S: 720.12311 ΔS: -0.496886 moves: 1 l=3 B: 3 <- 5 B=3 niter: 2 count: 1 breaks: 1 min_S: 720.12311 max_S: 720.62000 S: 720.12311 ΔS: 0.00000 moves: 0 l=3 Current bracket: (2, 3, 5) (691.21840695494018, 720.123114036723, 751.86302859686145) l=3 Current bracket: (2, 2, 3) (691.21840695494018, 691.21840695494018, 720.123114036723) l=3 Bisect at B = 2 with S = 691.2184069549402 l=3 Best result: B = 2, S = 691.2184069549402 level 3 : replaced (16, 1) -> (16, 2) , dS: -628.252218219 4 l=4 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=4 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=4 Bisect at B = 2 with S = 26.65330734662712 l=4 Best result: B = 2, S = 26.65330734662712 level 4 : replaced (2, 1) -> (2, 2) , dS: -1.86264514923e-09 5 l=5 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=5 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=5 Bisect at B = 2 with S = 26.65330734662712 l=5 Best result: B = 2, S = 26.65330734662712 level 5 : replaced (2, 1) -> (2, 2) , dS: -1.86264514923e-09 6 l=6 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=6 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=6 Bisect at B = 2 with S = 26.65330734662712 l=6 Best result: B = 2, S = 26.65330734662712 level 6 : replaced (2, 1) -> (2, 2) , dS: -1.86264514923e-09 7 l=7 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=7 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=7 Bisect at B = 2 with S = 26.65330734662712 l=7 Best result: B = 2, S = 26.65330734662712 level 7 : replaced (2, 1) -> (2, 2) , dS: -1.86264514923e-09 8 l=8 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=8 Current bracket: (2, 2, 2) (26.653307346627116, 26.653307346627116, 26.653307346627116) l=8 Bisect at B = 2 with S = 26.65330734662712 and l is increasing into the thousands if the left alone.

2 8

Problem with the partition into communities
by yerali 25 Mar '17

25 Mar '17

Dear list, I am new with graph-tools, which looks pretty nice. However, I have the following problem. The result of the partition into communities gives one big community and two very small ones, where the vertices are almost no connected between them. I wonder if I am doing something wrong. My network is directed and unweighted. My algorithm is just: " g2 = load_graph("test.graphml") state = gt.minimize_blockmodel_dl(g2, deg_corr=False) b = state.get_blocks() for v in g2.vertices(): print v,b[v] " My network has only 50 vertex, is that a limitation? Thanks in advance, Yérali. -- 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.

4 14

Representing temporal network data
by Marcus Burkhardt 21 Mar '17

21 Mar '17

Hello, looking for a python library for analyzing and visualizing dynamic graphs I recently came across graph-tool which seems to be quite suitable for my needs. But before getting into my remaining question I believe a big thank you for Tiago Peixoto is in order: Thanks for all the work you put into this library! And secondly I should preface with the disclaimer that I am neither a computer scientist nor mathematician, but a humanities scholar currently diving in to graph analysis and visualization for a project. Quickly approaching the limits of Gephi (the Software most of my colleagues turn to) for analyzing a large dynamic network I thought that I should give graph-tool a try. I have a large data set of the evolution of categorization on Wikipedia, i.e. data on when a category has been assigned to or removed from all pages in Wikipedia. (On a side note: I am currently not aiming to visualize the entire dataset, but the evolution of topics within Wikipedia) I searched and read through most of the documentation of graph-tool and graphml, but there is one thing I just cannot wrap my head around: Is there a smart or default way of representing dynamic networks in graph-tool? I understand that graph tool-allows for adding and removing edges / vertices dynamically, but is there a (recommended) way of storing the data on when an edge or a vertice is created/removed/created again/removed again... directly in the graph? And if not, how would you suggest to do that? This question does not only relate to the visualization of the graph, but to analyzing and comparing the graph at multiple ponints in time as well, i.e. calculating graph_tool.topology.similarity(g1, g2, ...) for the graph at time 1 and time 2. Of course, I could genrate two (or more) seperate graph objects for t1 and t2. However, for me it would be more desirable to create just one graph and to be able to calculate the graphs properties for certain points in time. A possible solution I could imagine is to make use of graph.views. But then again, how should I include the data for edge/vertice creation or removal in a property map? I suspect this is a question which does not have just one straight forward answer, but I appreciate any suggestions, tips or links to where to find more on this issue. And maybe I have even overlooked the solution in the documentation... Sorry in advance for that. Thanks a lot and all the best, marcus -- 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.

3 2

non-integer eweight
by Peter Straka 21 Mar '17

21 Mar '17

Dear list, does minimize_(nested_)blockmodel_dl() handle non-integer eweight (edge weights, as passed on to BlockState via state_args)? If so, how exactly? (A link to literature would be great.) Many thanks in advance, Peter

2 1

Remove vertices without edges
by Hobé Alex 20 Mar '17

20 Mar '17

Hi Guys! What is the best way to remove vertices that do not have any edges? Best, Alex

2 2

Website down?
by Hobé Alex 20 Mar '17

20 Mar '17

Hi Guys! The website seems to be down. Is there somewhere I can get an offline version of the documentation? Best, Alex

2 2

Use of edge's weights in model inference
by Andrea Briega 15 Mar '17

15 Mar '17

Dear Dr. Tiago Peixoto, I have a question regarding weighted graphs. I would like to classify a network with weighted edges into groups taking into account these weights. I have used minimize_blockmodel_dl and minimize_nested_blockmodel_dl before but I don't know exactly how to implement these functions in a weighted network. I'm not sure if I have to use the layered version, but in this case I don't know how to get these layers from my network. Thank you very much and sorry for the basic question, Andrea

3 5

vweight in BlockState
by Peter Straka 14 Mar '17

14 Mar '17

Dear list, what is the role of vweight in the constructor to BlockState? In particular, I'm trying to understand how the result of minimize_(nested_)blockmodel_dl will be different if certain vertices are assigned higher weights in the graph. I believe that if the eweight between a pair of nodes is increased, then the two nodes are more likely to end up in the same block (more or less). Would the opposite be true if their vweight is increased? Many thanks in advance, Peter

2 1

Debian package and boost at compile-time
by Alcolo Alcolo 12 Mar '17

12 Mar '17

Hi, I have installed the package python3-graph-tool and when I use graph_tool.search.dfs_iterator I got the error: Traceback (most recent call last): File "/home/steph/bin/superGraph.py", line 160, in <module> gg.create_revdep(gg.getPackage('umda')) File "/home/steph/bin/superGraph.py", line 88, in create_revdep for e in bfs_iterator(self.graph, node): File "/usr/lib/python3/dist-packages/graph_tool/search/__init__.py", line 345, in bfs_iterator return libgraph_tool_search.bfs_search_generator(g._Graph__graph, int(source)) RuntimeError: This functionality is not available because boost::coroutine was not found at compile-time Ubuntu: 14.04 python3-graph-tool version 2.12-1 (from deb http://downloads.skewed.de/apt/trusty trusty universe) There is a way to install python3-graph-tool binaries compiled with boost ? Of course I can compile myself graph-tool, but it is not my goal.

6 26

need intersphinx_mapping for graph_tool
by reckoner 10 Mar '17

10 Mar '17

I'm writing some documentation using Sphinx and I would like to reference graph_tool. Is this the correct intersphinx_mapping? 'graph_tool': ('https://graph-tool.skewed.de/static/doc/',None), If so, I am still having trouble referencing graph_tool using markdown such as: :py:mod:`graph_tool` or even :mod:`graph_tool` and it is not picking up the link. What am I missing here? Thanks! -- 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.

1 0