July 2020 - graph-tool - archives.skewed.de

What formula is used for std in vertex_average and edge_average?
by VaSa 16 Jul '20

16 Jul '20

I am curious what is being used to calculate the standard deviation of the average in gt.vertex_average and gt.edge_average >>> t2=gt.Graph() >>> t2.add_vertex(2) >>> t2.add_edge(t2.vertex(0), t2.vertex(1)) >>> gt.vertex_average(t2, "in") (0.5, 0.35355339059327373) Now, shouldn't std be σ(n)=sqrt(((0-0.5)^2+(1-0.5)^2)/2)=0.5 ? also q(n-1)=sqrt((0.5^2+0.5^2)/(2-1))~=0.70710 0.3535 is sqrt(2)/4 which happens to be σ(n-1)/2, so it seems there is some relation to that. A little bigger graph. >>> t3=gt.Graph() >>> t3.add_vertex(5) >>> t3.add_edge(t3.vertex(0), t3.vertex(1)) >>> gt.vertex_average(t3, "in") (0.2, 0.17888543819998318) Now, we should have 0,1,0,0,0 series for vertex incoming degree. So Windows calc gives σ(n)=0.4 and σ(n-1)~=0.44721, so where does 0.1788854 come from ? Reason, I am asking because, I have a large graph, where the average looks quite alright but the std makes no sense, as going by the histogram, degree values are quite a bit more distributed than the std would indicate. -- 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 3

calling graph-tool from R using reticulate
by sam 14 Jul '20

14 Jul '20

has anyone had any luck running graph-tool from R(Studio) using the reticulate package? when i try the following, RStudio crashes immediately: library(reticulate) use_condaenv("gt") py_install("graph-tool") source_python(file.path(PYTHON_CODE, "fit_sbm.py")) gt is a conda environment where i have graph-tool installed. for some reason reticulate would give an error "module graph-tool not available" when i would run it, hence the py_install command -- Sent from: https://nabble.skewed.de/

2 1

Specifying number of blocks in a bipartite SBM
by jsfog 13 Jul '20

13 Jul '20

Hi Tiago, Is it possible to specify the number of blocks for both types of nodes when estimating a bipartite SBM? I know that if I specify model.minimize_nested_blockmodel_dl(B_min=G, B_max=G, state_args = {'clabel': model.vp['node_type'], 'pclabel': model.vp['node_type']}) then I will end up with a partition with G blocks. But what I want is a partition in which the type 1 nodes are divided into G1 blocks and the type 2 nodes into G2 blocks, where I specify G1 and G2. Thank you, Jamie -- Sent from: https://nabble.skewed.de/

3 2

Building (and working with) bipartite graphs
by Davide Cittaro 09 Jul '20

09 Jul '20

Hi, I'm trying to build a bipartite graph from a MxN array. So far I've been able to do this: # adjacency is the original np.array S = [] E = [] W = [] for x in range(adjacency.shape[0]): for y in np.where(adjacency[x] > 0)[0]: S.append(x) E.append(y + adjacency.shape[0]) # so that node index starts with an offset) W.append(adjacency[x, y]) g = gt.Graph(directed=True) g.add_edge_list(np.hstack([np.array(S)[:, None], np.array(E)[:, None]])) ew = g.new_edge_property("double") ew.a = W g.ep['weight'] = ew recs = [g.ep.weight] rec_types = ['real-exponential'] state = gt.NestedBlockState(g, state_args=dict(recs=recs, rec_types=rec_types)) […] I'm pretty sure this is not the most efficient way to go, but gt.is_bipartite(g) returns True. Can anybody suggest a more straightforward approach? Also, is it necessary to specify a VertexPropertyMap with node type (i.e. M or N)? Anyhow, after I had the graph and the final model, I tried to draw: gt.draw_hierarchy(state, layout='bipartite') but I get this error IndexError Traceback (most recent call last) <ipython-input-67-d7655416181e> in <module> ----> 1 gt.draw_hierarchy(state, layout='bipartite') ~/anaconda3/envs/schist/lib/python3.8/site-packages/graph_tool/draw/cairo_draw.py in draw_hierarchy(state, pos, layout, beta, node_weight, vprops, eprops, hvprops, heprops, subsample_edges, rel_order, deg_size, vsize_scale, hsize_scale, hshortcuts, hide, bip_aspect, empty_branches, **kwargs) 1773 rel_order_leaf=True) 1774 elif layout == "bipartite": -> 1775 tpos = get_bip_hierachy_pos(state, aspect=bip_aspect, 1776 node_weight=node_weight) 1777 tpos = t.own_property(tpos) ~/anaconda3/envs/schist/lib/python3.8/site-packages/graph_tool/draw/cairo_draw.py in get_bip_hierachy_pos(state, aspect, node_weight) 2145 t.add_edge(root, p2) 2146 for p in ps: -> 2147 if bc.a[tb[p]] == 0: 2148 t.add_edge(p2, p) 2149 else: IndexError : index 1 is out of bounds for axis 0 with size 1 Any help is appreciated d

2 2

ranking vertices by value of vertex property
by Ioana K-Hulpus 07 Jul '20

07 Jul '20

Hi, I ran PageRank on my graph, and now I need to get the top-k highest scored. Is there any faster / more straightforward way of achieving this than by creating and then sorting a new dictionary, as in the following method? def top_k_page_rank(g, pagerank_prop, k): pr_sorted_scores = sort_pagerank_scores(g, pagerank_prop) top_k = list(pr_sorted_scores.items())[0:k] return top_k def sort_pagerank_scores(g, pagerank_prop): pr_scores = dict() for v in g.vertices(): pr_scores[v] = g.vertex_properties[pagerank_prop][v] pr_sorted_scores = {k: v for k, v in sorted(pr_scores.items(), key=lambda item: item[1], reverse=True)} return pr_sorted_scores Thanks, Ioana -- Sent from: https://nabble.skewed.de/

3 2

Missing Example in Flow Algorithms?
by spolo96 06 Jul '20

06 Jul '20

Greetings everyone, I just began to check this awesome module for Graphs in Python and I'm interested in Flow algorithms, specifically the Minimum S-T cut one, however I don't know if the example for the final two algorithms: "graph_tool.flow.min_st_cut" and "graph_tool.flow.min_cut" is missing or something, since they both load this specific graph example: g = gt.load_graph("mincut-example.xml.gz") But after checking the website I cannot find it, or see how it was created like the previous example for the other flow graphs: "flow-example.xml.gz". This is one of my first times in the pages, so I ask for forgiveness if I missed it, but the reason I ask here is because I didn't find it. Thanks in advance! -- Sent from: https://nabble.skewed.de/

2 3

Memory requirement estimate for a large graph
by Carlo Nicolini 06 Jul '20

06 Jul '20

Dear Tiago, How is it possible to get an estimate for the memory requirement of a graph in graph-tool? I know that graph-tool is built upon C++ and Boost, and the adjacency list is stored via a hash-map. Apart from the cost of storing the values of vertices indices and edges indices as `unsigned long`, what is the memory overhead of the structures used in storing the graph? For example, for a network of 1M vertices and 100M links without attributes, how much real memory should I plan to use, excluding temporaries? Sorry if the question is repeated, but I could not find it in the previous mailing list posts. Regards, Carlo

2 3

Questions about PPBlockState
by Davide Cittaro 03 Jul '20

03 Jul '20

Hello, I'm testing the new Planted Partition model in graph-tool on my data, indeed I'm finding interesting results. I have some questions/observations, though. - PPBlockState returns a relatively large number of partitions on large networks, which is fine and expected. When I use NSBM, instead, I make use of the hierarchy not only because I can "abstract" partitions up to a certain level, but also because the hierarchy has a meaning in my case. Is there (or will it be there) a hierarchical formulation of the PPBlockState? - I tried multiple initialisations of PPBlockState over my graph, I also tried to increase the iterations of the initial MCMC sweep and I'd say I get very consistent results. Is this expected? I mean, is it known if the PPBlockState converges to a stable solution faster and in a consistent way? - Does the time required to converge scales with the number of edges as it does for SBM? - As far as I understand, if the assortativity is the dominant pattern the difference between PP and NSBM is negligible. I don't know how to quantify "negligible" as the differences in entropies are at least in the order of 1e2 in the cases I tested (seems pretty large to me); I would be happy to switch to PP, also given the shorter runtime so far, but I'm a bit concerned about these differences. Best d

2 3