July 2018 - graph-tool - archives.skewed.de

Using an edge covariate as a hyper parameter for the SBM
by Zahra Sheikhbahaee 20 Jul '18

20 Jul '18

Hi there, I am trying to include the edge weights by taking to account an edge covariate matrix for the nested block model inference. Well, Each time I run the code on my data set I get slightly different results both in terms of number of blocks and the nodes in each block. This is my code: state = minimize_nested_blockmodel_dl(g, state_args=dict(recs=[g.edge_properties["weight"]], rec_types=["discrete-geometric"])) state.draw(edge_color=prop_to_size(g.edge_properties["weight"], power=1, log=True), ecmap=(matplotlib.cm.gist_heat, .6), eorder=g.edge_properties["weight"], edge_pen_width=prop_to_size(g.edge_properties["weight"], 1, 4, power=1, log=True), edge_gradient=[], vertex_text=g.vertex_properties["attribute"], vertex_text_position="centered", vertex_text_rotation=g.vertex_properties['text_rotation'], vertex_font_size=10, vertex_font_family='mono', vertex_anchor=0, output_size=[1024*2,1024*2], output="DiscreteGeometric_%s.pdf"%(eventName)) I appreciate if you explain what your approach would be and how I can run graph-tool using the covariance matrix of edges in order to get statistically reliable results? Is there also any way to get the full posterior of each node belonging to each block? Thanks in advance.

2 12

TypeError while using add_edge_list
by ashutosh 18 Jul '18

18 Jul '18

Hi, I'm trying to convert two columns of my pandas dataframe into network using them as edge list. A similar query <https://stackoverflow.com/questions/35459597/graph-tool-reading-edge-lists-…> was posted in StackOverflow but with different problem and it didn't help me. When I am trying the following code it gives me error: * myGraph.add_edge_list(i for i in df[['node1','node2']].values.tolist())* Where, my Graph is the graph; df is the dataframe with 4 columns out of which I am using two 'node1' and 'node2'. *File "/usr/lib/python3/dist-packages/graph_tool/__init__.py", line 2109, in add_edge_list libcore.add_edge_list_iter(self.__graph, edge_list, eprops)* *TypeError: No registered converter was able to produce a C++ rvalue of type unsigned long from this Python object of type str * I tried upgrading my boost library but it doesn't solve the problem. Any suggestions ? -- Sent from: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/

3 5

compiling error: pycairo headers not found
by JM 16 Jul '18

16 Jul '18

Dear Tiago, while trying to compile graph-tool for anaconda3 and python 3 we came across following error message: configure: error: pycairo headers not found I ran the following command: ./configure --with-boost=/usr/local/anaconda3 --enable-openmp Python3 is installed in /usr/local/anaconda3/bin and typing in python starts the correct version. The pycairo headers (py3cairo.h) are located in: /usr/local/anaconda3/include/pycairo I am surprised it can't pick the file up, because it seems to pick up other files from /usr/local/anaconda3/include. Do you have any idea what we could have missed? The config.log <http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com…> is attached and I am running Ubuntu 16.04. Many thanks, Jana -- Sent from: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/

3 2

joblib parallel for with graph-tool filtering?
by Tasos 16 Jul '18

16 Jul '18

With graph-tool and joblib working together, do we need to send graph.copy() in the "Parallel" call like in the code below when using graph vertex filtering with .set_vertex_filter? graph.copy() makes memory usage extreme in large graphs (2M Vs, 4M Es) but in my head ensures any concurrency problems. (or 'graph' without '.copy()' is ok?) What is the best way to run parallel graph searches and filtering (different vertex per thread) with graph-tool and joblib? (or without joblib) ### # defined and filled earlier g_graph = graph_tool.Graph(directed=False) eprop_ang = g_graph.new_edge_property("float") ### from joblib import Parallel, delayed import multiprocessing import os import tempfile import shutil import datetime path2 = tempfile.mkdtemp() out_path2 = os.path.join(path2,'z6path_out2.mmap') out2 = np.memmap(out_path2, dtype=np.float32, shape=(g_graph.num_vertices(),dims), mode='w+') num_cores = 30 num_pre_workers = 60 def runparallel(graph, row, out2): dist, pred = graph_tool.search.dijkstra_search(graph, graph.vertex(row), weight=eprop_ang) ## etc etc ##### v_filter = graph.new_vertex_property('bool',val=False) for v in SOMETHING_LOCAL: v_filter[v] = True graph.set_vertex_filter(v_filter) # do something with the filtered 'graph' (subgraph) # and save output to out2 out4[row] = RESULT ## graph.clear_filters() Parallel(n_jobs=num_cores, pre_dispatch=num_pre_workers, verbose=1)(delayed(runparallel)(g_graph.copy(), r, out2) for r in range(g_graph.num_vertices())) -- Sent from: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/

3 3

Strange Results when Filtering Vertices Iteratively
by P-M 11 Jul '18

11 Jul '18

Hello, I have been trying to filter vertices iteratively but am seeing very strange behaviour. Summary of what my code is doing: I loop through groups of vertices. I set the vertex property map to False for all vertices. I then set it to True for all vertices in the current group. I then loop through all vertices in the graph setting the vertex property for the current vertex also to True, creating a graph_view object using that property map, resetting the value for that vertex to zero and then moving on to the next vertex repeating the procedure. For some reason the number of vertices in the resulting graph_view object bears no resemblance to the number of vertices in my current group. Can anybody help me with what is going on here? I am using '2.27dev (commit 31421e3c, Sun Jun 17 19:18:54 2018 +0200)'. If the code looks like it should work I can upload a copy of the state and network if that would be useful. Here is the code: g = gt.load_graph('graph_no_multi.gt') with open('state.pkl','rb') as state_pkl: state=pickle.load(state_pkl) groups = collections.defaultdict(list) l0 = state.levels[0] for v in g.vertices(): groups[l0.b[v]].append(int(v)) v_filter = g.new_vertex_property('bool') P_part = collections.defaultdict(int) for target_vertices in groups: target_vertices = groups[target_vertices] v_filter.set_value(False) for v_index in target_vertices: vertex = g.vertex(v_index) v_filter[vertex] = True for v in g.vertices(): v_filter[v] = True f_g = gt.GraphView(g,v_filter) if int(v) in target_vertices: if f_g.num_vertices() != len(target_vertices): print 'f_g.num_vertices()=', f_g.num_vertices(), 'expected: ', len(target_vertices) elif f_g.num_vertices() != len(target_vertices)+1: 'f_g.num_vertices()=', f_g.num_vertices(), 'expected: ', len(target_vertices)+1 v_filter[v] = False Best wishes, Philipp -- Sent from: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/

2 2

large graph
by Edwin A Pell 11 Jul '18

11 Jul '18

Hello, is there a hyberbolic option. That is I want to make a graph in a circle where the display distance decreases as the bounding circle is neared. Thus allowing an arbitrarily large graph to be draw (may be not resolvable at the edge). Thanks. Ed

2 1

Installation of the latest version of graph tool with conda
by Zahra Sheikhbahaee 10 Jul '18

10 Jul '18

Hi Tiago, I am trying to install graph-tool on a server and the only way it is possible without root permission is via this command line: conda install -c ostrokach-forge graph-tool But its version is >>> graph_tool.__version__ '2.25 (commit , )' well, how could I update this version in the conda environment to 2.27? Thanks, Zahra

2 4

Graph database
by Matyas, Csongor 06 Jul '18

06 Jul '18

I know this is a bit off topic, but it might be related to graph-tool also there might be someone here who could help me. I need all the planar graphs that are 4-regular (in the meaning that every vertex has a degree 4) but each graph is allowed to have double edges (not triple or more) also an edge can connect to the same vertex. I couldn't find any database that could give these to me so probably I will have to generate them for myself. Any idea of how could I do this with graph-tool? I know how can I check weather these graphs are planar or not, but the rest seems very complicated. Any help is very much appreciated. Csongor

2 3

Graph thread safety
by mae 02 Jul '18

02 Jul '18

Good morning, I'm considering to use graph-tool as a part of a python web-service providing graph analysis features. Graph-tool performance looks really promising, but I'm not sure if it fits my case. I would like to load graph into the memory and share/read it between multiple threads/processes (to handle many requests in parallel to the same graph). Is Graph thread safe? For me it seems that it's not, but I would like to confirm it. Best Regards, Damian -- Sent from: http://main-discussion-list-for-the-graph-tool-project.982480.n3.nabble.com/

3 3