> I am not sure, because the results are different from the same algorithm in
> SNAP, which also implement Girvan-Newman's algorithm while only unweighted
> version.
The edge betweenness algorithm is not deterministic in the sense that it is
free to make a (possibly random) choice when there ar
> I know that the algorithm is not deterministic. In my case, the graph only
> have tens of nodes, while the results are very different.
> In igraph, the unweighted version results in only one community; in SNAP,
> there are quite a few communities.
How does SNAP select the number of communitie
Hi,
Well, I'm not too familiar with either of these packages so I don't know
anything about its internals, but here are a few things to consider:
1. You cannot modify the same variable in multiple parallel execution branches.
For instance, even the following simple function does not work in pa
> It was a simplification of my program. I have a data set of network with
> quite 43,000,000 edges stored in a special unusual format. I myself should
> make an empty graph and append edges. the process is very high and It takes
> too long if I want to do sequentially.
>
The problem lies not
Hi Dov,
I have recently fixed this in the development branch. If you wanna give it a go
and don't mind compiling it yourself, check out the nightly build tomorrow:
http://code.google.com/p/igraph/downloads/list
Revision 3006 of the 0.7-main tree should include all the changes -- this will
be
Hi Sam,
The primary purpose of supporting curved edges is to ensure that multiple edges
between the same node pairs are all visible. igraph does not try to set the
curvature of edges to avoid other nodes (and I guess it would be quite
complicated anyway if you take into account that nodes may
>
> I want to calculate page rank centrality for several networks but in some of
> them I got this error message :
> what does it mean and how I can solve it?
Usually this means that there is something "pathological" in your graph. Send
me a few example graphs that do not work for you and I'll t
> I am not familiar with shapefiles, though, so I don't know how to
> convert them to igraph.
I'm not sure either, but one possible way to go would be to
1) load the shapefiles package in R
2) read your shapefile and convert it to a plain data frame using
convert.to.simple
3) depending on how the
> More probably I am mistaking, but I loaded an edge list graph
> (graph<-read.graph("graph.txt", format="edgelist",0, TRUE), with 7
> nodes but vcount(..) shows me 8 (nodes).
> The nodes in my graph are identified by a numeric number 1-7 (graph
> is shown in below).
igraph uses zero-based vertex
> I am facing problems with `igraph_sparsemat() function.
Could it be the problem that your sparse_g variable is essentially a dangling
pointer that points nowhere? You should _create_ a sparse matrix first and then
pass the address of that sparse matrix to igraph_sparsemat.
--
T.
___
Hi,
First I would try replacing set.vertex.attribute(g, "bucket", vid, bv[i]) with:
V(g)$bucket[vid] <- bv[i]
I think this avoids copying the graph. Also, I would even try this (not sure if
it works, but it's worth a try):
V(g)$bucket[sortedVertexIDs] <- bv
This would apply all the bucket vec
Malcolm,
Add the following line to the foreign-*-parser.y files somewhere below the
includes:
char* stpcpy(char* s1, const char* s2);
This did the trick for me.
Best,
Tamas
On 31 Oct 2012, at 16:47, Malcolm Tobias wrote:
> Gabor,
> I tried what you recommended, but I'm still getting an erro
> Is there an upper limit on graph size for plot operations in igraph?
No, there isn't, apart from the working memory of your system.
> Python hangs with a “BEX” error message (buffer overflow?) when graph objects
> grow in size.
I've never heard about such an error message -- what operating syst
> Tamas will know this better, but how big is your graph? Here is an
> example with ~70,000 nodes and ~4 million edges, so if your graph is
> smaller than this, then graph size is not a problem:
> http://sixdegrees.hu/last.fm/
Well, this figure was created with a custom plotting routine, not the on
For the record: we have managed to resolve the issue with the Intel compilers,
so the 0.6 tree should compile with icc from revision 3000 onwards. Similarly,
the first revision in the development (0.7) tree that supports icc is revision
3043.
Those who do not want to check out the full source t
> And after I pick two nodes, say 25, 26, which have multiple edges between
> them, I'm trying to merge them into a single vertex:
The right contraction vector in this case is:
c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,25)
The reason is as follows. In the contraction ve
Dear Mandy,
As far as I know, Burt's constraint scores can be calculated on an undirected
graph without problems; even the R manual page for the constraint() function
uses an undirected graph in the example:
http://igraph.sourceforge.net/doc/R/constraint.html
If you think you are getting incor
> Of the vertices 1:26, I can also choose two nodes that are not consecutive,
> as in, I can choose to merge (1,5) or (7,18) into a single vertex. Can this
> be done?
Quoting myself:
> In general, if you want to merge vertex v into vertex u (assuming that v >
> u), you will need a contraction
Hi,
> I just started out with iGraph in the python interface, I have trouble
> printing out the membership of the community_walktrap.
The walktrap method does not give you a membership vector, it gives you a
dendrogram instead (since it is a hierarchical clustering method). You must
"cut" the d
Hi,
You did not mention whether you are working with igraph in C, R or Python so I
can give you only some generic ideas:
> I want to filter a graph that I made. If a vertex has a degree 1 and is
> connected to another vertex who has also got degree 1 -> delete both vertices
> (and also the cor
Hi,
> I noticed that new tarballs were added yesterday. Just to let you know that
> the problems with compiing on OSX (R version 2.15.2) remain. Would you know
> the most recent useable tarball that has the revision to alpha centrality
> included? Many thanks.
I have just uploaded a tarball of
Hi,
> sub_graph <- induced.subgraph(graph=g_n1, v=unlist(fg_communities[k]))
Can you let us know what does fg_communities contain when you issue this
command? Even better, can you send us a small, self-contained example that
reproduces the problem?
Best,
Tamas
n more clear in my explanation.
>
> Thank you again.
>
> Best,
>
> Matteo
>
> 2012/11/17 Tamás Nepusz :
>> Hi,
>>
>>> sub_graph <- induced.subgraph(graph=g_n1, v=unlist(fg_communities[k]))
>> Can you let us know what does fg_communities
Hi,
Have you tried upgrading R and XQuartz to their most recent versions? For the
record, I'm using R 2.15.1 and XQuartz 2.7.4 (xorg-server 1.13.0) and it works
fine for me.
Cheers,
Tamas
On 17 Nov 2012, at 18:46, Silvia SM wrote:
> Hello,
>
> Maybe someone can help me. I just switch to Ma
nce would be appreciated.
>
> Kind regards,
> Bob
>
> On 17 November 2012 00:51, Tamás Nepusz wrote:
> Hi,
>
> > I noticed that new tarballs were added yesterday. Just to let you know that
> > the problems with compiing on OSX (R version 2.15.2) remain. Would you k
alled package can be loaded
>
> * DONE (igraph)
>
> Once again many thanks for your help.
>
> Kind regards,
> Bob
>
> On 19 November 2012 11:10, Tamás Nepusz wrote:
> Hi Bob,
>
> This seems to be an issue that is specific to your machine as I cannot
> reprodu
> I also tried to ignore the weights, by using the option
> "weights=E(net)$NA", but it didn't work.
You should use weights=NA if you want to ignore the weights (as far as I know).
Also, do you happen to have any pathological weights in your graph (like
negative weights or zeros)?
--
T.
_
Hi,
> how is the average path length calculated for graphs that contain more than
> one component?
Depending on the value of the "unconn" parameter of the function, there are two
possible behaviours:
- If "unconn" is False, igraph simply assumes for every disconnected vertex
pair that their d
> Hi, I solved this by myself..
Great! :) Now, if you want to generalize it to weighted graphs, there are at
least two possible ways: either you calculate the average weight of the
shortest paths or you still calculate the average path length but take the
weights into account when you calculate
> It's nice to know that the function "get_shortest_paths" support weights. So,
> the formula should be something like the following:
Basically yes, but it would probably be more scalable to calculate the path
lengths for the vertices one by one (in g.shortest_paths) because this way you
don't h
> I am trying to calculate the shortest paths for a large sparse graph
> (48000 nodes and 2000 edges). [...]
> shortest_paths<-shortest.paths(net,weights=E(net)$weight);
> Error in .Call("R_igraph_shortest_paths", graph, v - 1, to - 1,
> as.numeric(mode), :
> negative length vectors are not allow
> I have anohter question relating to this. I like to use start layout but I
> like the edge.arrow to reverse directions. For exampe rather thatn doing this
> <-- I like to this ->, any ideas?
Use the edge.arrow.type parameter. From the documentation:
"This parameter can be used to speci
Dear Bob,
Thanks for the feature request. As a matter of fact, I also have my own C
implementation of the Clauset-Shalizi-Newman method (see
http://github.com/ntamas/plfit for a standalone version) and its igraph
integration has been on my TODO list for a long time -- unfortunately I never
got
Hi,
> Is it possible, that I have a newer version of igraph or RStudio, which is
> not compatible with the fastgreedy alg.? What can I do?
I strongly suspect that the problem lies in RStudio itself because all the
community detection algorithms work perfectly fine for me if I use them from
the
Hi,
> g_adjacency_undirected_weighted_connected <-
> graph.adjacency(adjacency_matrix_undirected_unweighted_connected)
The graph you create here will be directed because graph.adjacency creates
directed graphs by default. Add mode="undirected" to make the graph undirected.
Other than that, your
> If I remember right, some months ago somebody posted here about
> measures to compare networks. I looked for this post in the historial
> of the list, but I didn't find it, sorry.
This is a thread that seems related, although it is quite old:
http://lists.gnu.org/archive/html/igraph-help/2008-04
> First, I was wondering if Igraph considers IGRAPH_INFINITY as zero or
> as other value.
IGRAPH_INFINITY is infinity, period. It is there only to provide us with a
sort-of-platform-independent way to refer to infinity.
> I am working with shortest-paths matrices and I would
> like the distance b
> "The neighborhood of a given order o of a vertex v includes all vertices
> which are closer to v than the order. Ie. order 0 is always v itself, order
> 1 is v plus its immediate neighbors, order 2 is order 1 plus the immediate
> neighbors of the vertices in order 1, etc.", is order 1 yielding "i
Dear Charles,
> number of vertices of a network from in simple pajek file. The program
> compiles correctly, it reads the edges of the network correctly, but
> it reads the number of vertices like a strange number, like -10585152.
I'm 99% sure that this is a printing/formatting issue, i.e. the num
> I used eigen() but It did not work on large networks. I got memory
> allocation error.
This is because eigen() would calculate all the eigenvectors -- guess that
takes a lot of memory ;)
> According to below links, "nev" parameter determines the number of required
> eigenvectors and "which"
> I'm trying to use R igraph to get x,y coordinates to plot vertices in another
> application. The data should show a simple organisation chart, hence the
> desire for a top-down tree.
> [...]
> Looking at l I get something similar to:
>
> [,1] [,2]
> [1,] 0.00e+00
> bmat2<-matrix(rbinom(100,1,0.1),100,100)##Source of the
> problem
You probably need matrix(rbinom(100*100, 1, 0.1), 100, 100), otherwise all the
columns in the matrix will be equal (because the elements generated by rbinom
are enough only for the first column). Also, note that
Hi,
> I'm trying to identify a good distance (similarity / dissimilarity)
> measure for discovered VertexDendrogram and VertexClustering
> communities
Would you like to compare two communities, or to compare two community
structures as a whole? For the latter, we have a function called
compare_c
> vc = g.community_infomap()
You could try the following:
g2 = g.copy()
g2.contract_vertices(vc.membership)
g2.es["weight"] = 1
g2.simplify(combine_edges="sum")
This would give you a graph where the nodes represent the communities of the
original graph, the edges are weighted, and the weight of
> G = igraph.Graph(directed = True)
>
> 1) Adding edges:
> G.add_edge([(1,2)])
> G.add_edge([(1,2)])
> if I add twice the same edges, my digraph keeps two occurence of the same
> edges in G.get_edgelist: [(1, 2), (1, 2)]
Why is that a problem? If you add the edge twice, of course you are going to
Hi Murat,
So I've spent a bit of time with debugging and here's the explanation. It's a
long story.
First of all, H in your example is not a clone of G in any sense; it is a
completely independent object, so that's not the issue here.
Second, igraph's internal data structures are optimized for
> I need to compare time complexity of computing different centrality measures
> in igraph but I could not find how they are implemented exactly.
The documentation of the C core of igraph lists the time complexity of most of
the functions:
- Degree: http://igraph.sourceforge.net/doc/html/ch04s0
Hi,
> I therefore used transitivity function in igraph, and for my graph I've got
> the value: 0.5493 (global), 0.42 (average).
> Reading the previous mail called "Small worlds in igraph", I think the
> "global" is best for what I want to show.
> I also compared my results using 2 others softwar
Yes, if G.is_directed() returns True, then G.shortest_paths() will take edge
directions into account. Use the mode=... keyword argument to override that.
See the docstring of G.shortest_paths() for more details. This applies also to
the other shortest path functions, not only G.shortest_paths()
Seems like this question is answered already on Stack Overflow:
http://stackoverflow.com/questions/14164887/change-the-font-and-colour-of-the-igraph
Basically, get rid of the "main=..." argument and use the title() command
instead:
title("This is my first igraph", cex.main=3, col.main="green")
> first part works, second part (layout) no. It kindly ask me how I want to
> close R... but that'll do great for now.
Oh darn, I forgot that layout.graphopt is having some problems in igraph 0.6
under R. Use any other layout function (e.g., layout.fructerman.reingold,
layout.kamada.kawai etc) a
Hi,
Could be a bug but I'm not sure and unfortunately we don't have any Windows 7
machines to test on. Can you please try it with other formats (e.g.,
write_ncol, write_gml and so on)? Does it work with them or do you get empty
files?
Also, can you please try this:
f = open("gf.graphml")
fg.w
> I switched to Python 2.7 and I am now able to save graphml, ncol, and gml
> files.
Okay, that's good, so the issue is related to Python 3.x only. Quite a few
things changed in the Python 3 API related to file handling so I'll start
digging there. (Hopefully I can reproduce the issue on Windows
> I’m experimenting with Graph.count_subisomorphisms_vf2 to see if it will do
> what I need it to do. I’m starting with a very simple call (assume g1 and g2
> are instances of igraph.Graph):
>
> countMatches = g1. count_subisomorphisms_vf2(g2)
>
> When I execute this I get the
> I have a graph with 23379 edges and 23295 vertices. It is a simple,
> unconnected graph with 917 distinct clusters and a density of 4.26x10^-5.
> I am able to load the graph without a problem. However, I need to calculate
> shortest paths using R so that I can run a series of simulations and
> I need to calculate the shortest paths for all
Okay, so you actually need the shortest paths and not only their lengths.
> in order to run the simulation code that follows, but I am only interested in
> shortest paths which are 6 degrees or less.
> Is there a way to specify sp<=6
No, there isn'
Hi,
> For (deg in 1:6) (
> Loc <-which (sp==deg)
> [...]
> I get the following error message:
>
> Error in which (sp == deg): cannot allocate memory block of size 2.0 GB
>
> Does anyone know what a "memory block" is? Or, how to increase the size
> allocated to a memory block?
The message simpl
> I observed that in the graph which i'm currently analysing, if I were to run
> the evcent function with the arguement for directed set as true, the vectors
> generated are all zeros.
> Is this normal?
Does not seem to be so; send me the graph (in private, not to the mailing list)
so I can tak
Hi,
Thanks for the bug report. It seems that the C core of igraph does not have
this error so the problem is in the glue code that bridges the gap between the
C core and R itself. Gabor will probably fix it soon (if it is not fixed
already in the development tree). In the meanwhile, you can wor
AFAIK IronPython is written in C#, therefore it cannot load any Python
extensions that are implemented in C. This applies to igraph as well since
igraph is mostly a C library and the Python module is "just" a wrapper around
the C core. In other words, the Python module of igraph is only compatib
> is there any combinations() in r? something that explore all possible
> combinations in a vector - as far as I have understood it...
expand.grid could be the function you are looking for.
--
T.
___
igraph-help mailing list
igraph-help@nongnu.org
htt
> I am trying:
> expand.grid(neighbors(g,1))
Try this:
neis <- neighbors(g, 1)
expand.grid(neis, neis)
--
T.
___
igraph-help mailing list
igraph-help@nongnu.org
https://lists.nongnu.org/mailman/listinfo/igraph-help
> Reading package lists... Done
> Building dependency tree
> Reading state information... Done
> E: Unable to locate package igraph-python
Judging from the error message, you wrote "sudo apt-get install igraph-python"
and not "sudo apt-get install python-igraph".
> I've tried to install ca
> Using Ubuntu 11.04 64 bit at the moment but have tried the same in Ubuntu
> 12.10 64 / 32
Ah, that must be the reason. Ubuntu 11.04 is unsupported; the repo contains
packages for Ubuntu Quantal (12.10) and Precise (12.04) only. Are you sure that
it did not work in Ubuntu 12.10?
Best,
Tamas
__
> pretty sure but let me try right now. thank goodness for virtual machines!
> would there be an issue with either 64 bit or 32 bit versions?
There shouldn't be any; I'm using python-igraph regularly on Ubuntu 12.10.
--
T.
___
igraph-help mailing list
> I am wondering how the betweenness centrality calculation work on the
> negative weights values.
Chances are that it won't. Betweenness centrality calculates all the shortest
paths in the network (where "shortest" means "having smaller total weight") and
then counts how many times a given edge
Dear Matthias,
> After applying ``rewire()`` the edge values are replaced by 'None'. Does the
> rewire function allow for randomization of weighted networks at all?
Well, yes, with a little trick. The ``rewire()`` function does not preserve
edge attributes yet, but you can "back them up" in a va
> I tried using rewire.edges from R igraph 0.6 package, but found out that the
> generated random graphs had almost the same edge weights as that of my
> original graph.
How do you define the edge weights in the randomized graph?
> So, if someone could point out how to estimate if there is a
Hi Nina,
> I was just reading some requests to change the rewire-function: I am also in
> trouble because I need a fast rewiring method that maintains the (directed)
> degree distribution and AVOIDS self-loops. I noticed that in November 2012
> somebody actually wanted the rewiring to include self
> I got your point. My graphs are "biological", generated from a set of
> biological experimental data and database information, so each node
> corresponds to a gene, so I would say that node ID in my case not an
> arbitrary property and has meaning behind it.
Ah, understood. Anyway, I would sti
> I am doing some research on networks and I find that fastgreedy.community
> returns different results when run on different PC.
fastgreedy.community is not fully deterministic because the original algorithm
does not specify what should happen when more than one possible merge of
communities wo
> In the code/output below Node 153 has one neighbor (Node 156) but the degree
> function returns 2. Does anyone know why this would happen?
Your graph is directed and node 153 has an outgoing and an incoming edge.
degree() returns the "undirected" degree by default, i.e. it counts both the
inco
> Thank you for explaining this for me. But I am still confused about why
> multiply
> m at last using Ki/2m * Kj/2m * 2.
Because the graph is undirected, so an edge generated from vertex i to vertex j
(which has probability Ki/2m * Kj/2m) is equivalent to an edge generated from
vertex j to ve
> Hi, sorry for letting you misunderstand. What I am not sure is that why at
> last multiply
>
> m since I thought Ki/2m * Kj/2m * 2v was the last result. Thanks!
You are generating m edges, and each generated edge falls between vertices i
and j with probability Ki/2m * Kj/2m * 2. Therefore, the
> than the multilevel algorithm but theoretically the LPA is superior since it
> is linear.
There are at least three catches here:
1) The multilevel algorithm is also claimed to be "near linear" on sparse
graphs. We haven't benchmarked it, this is what is written in the publication
of the multi
> #1
> The installer option fails at the Destination Select window with the
> statement:
> “Python-igraph 0.6 can’t be installed on this disk. python-igraph requires
> Apple Python 2.7 to install”
>
> Python 2.7 is the default version (not sure how to direct the ins
Hi,
I would probably use the DrL layout instead of Kamada-Kawai -- chances are that
you would probably get the same hairball with KK as the one that you would get
by dropping your points randomly in the unit sphere :) As for KK, I would
expect at least a few days of computation time for a graph
Hi,
> Thanks for your help. I did need upgrade to using the python.org installer
> (not from source). Is there a way to trick the installer into using this
> (e.g., what is the installer looking for to see if this is an apple provided
> python).
No, there isn't, as far as I know.
> I haven't i
Hi,
Thanks for the report, this is indeed a bug. You can subscribe to the following
bug report if you are interested in when we will fix it:
https://bugs.launchpad.net/igraph/+bug/1126603
--
Tamas
On 8 Feb 2013, at 16:29, Manisha wrote:
> Hello friends,
>
> This is the first time I am usin
Oh, I forgot to mention: the workaround for the time being is to calculate the
betweenness score for all the vertices and then extract the value corresponding
to "a" from the result vector:
scores = g.betweenness(weights="weight")
print scores[g.vs.find("a").index]
This won't take longer than c
> My mistake, actually there is (somehow I don't remember trying to install
> igraph before) an igraph folder that was in:
It seems like Python is not able to find the instance of libigraph.dylib that
was used when the Python interface was compiled. Did you also delete
/opt/local/lib/libigraph*
> I have to attach node attributes which I have stored in a matrix like:
>
> id posts threads zindex
> "u32" "123" "12""0.45"
>
> is there anything faster than this:
Yes, there is. First, construct an index vector where the i-th element denotes
the index of the vertex whose
> How can I get a list (or vector, or df) containing all of the
> reciprocal node pairs?
See ?is.mutual; e.g.:
> el <- get.edgelist(g)
> el[is.mutual(g), ]
This gives you the subset of the edge list corresponding to mutual edges. If
you are interested in the indices of mutual edges, use which(is
> I have a data.frame with two columns, FROM and TO. Each are names of
> people where the person in FROM sent a message to the person in TO.
> There are many cases where person X sends to Y, but graph.data.frame
> doesn't appear to capture edge weight.
Are the weights defined explicitly in your dat
Thanks for the heads up; this is probably a typo and Gabor will fix that. In
the meanwhile, try vertex.label.color and vertex.label.dist.
--
T.
On 20 Feb 2013, at 20:54, Kun Deng wrote:
> The following snippet in the igraph tutorial won't run:
> >
> g <- barabasi.game(100, directed=FALSE)
>
Hi,
So I've checked our implementation of community_label_propagation() and it is
indeed quadratic, not linear. The reason is that we are using a dense vector to
count the number of occurrences of labels in the neighborhood of a node, and
clearing this vector takes O(n) time, while it would be
Hi,
Seems like this is the same question as the one asked recently on Stack
Overflow; see my answer there:
http://stackoverflow.com/a/15008720/156771
The bottom line is to use the VF2 subisomorphism algorithm in igraph, which
handles node and edge colors. You just have to convert your node lab
Hi Jeff,
I don't know whether controlling the order in which nodes are drawn is actually
possible in the R interface (maybe Gabor can answer that if he's around). It is
true that nodes are drawn in order of their vertex IDs by default and lower IDs
are drawn first. One thing that you can surely
Hi,
> a) Can I constrain the lengths of all the edges to a given value?
> b) Can I constrain the orientation of edges such that they look like the real
> physical foam ?
> (Think of it as constraining the edges such that they form adjacent
> structural frames each of which look like a sphere)
N
> I also want to know is there any way to obtain the exponent element of graphs
> generated by Barabasi model or not?
Yes, there is -- look it up in the paper of Barabasi and Albert ;) In theory,
the model generates networks where the degree distribution is a power-law with
gamma=3. There is a f
> Please find the toy data as an attachment to this mail.
I haven't found any attachment in your email; did you forget to attach it?
Best,
Tamas
___
igraph-help mailing list
igraph-help@nongnu.org
https://lists.nongnu.org/mailman/listinfo/igraph-help
Hi Stefano,
Sorry for the late reply. I'm not sure what happens here, but here's what I
think. The modularity function has a so-called "resolution limit" (see
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1765466/ ) which basically prevents
one from detecting communities that are smaller than a g
Hi,
I assume that "C1 C2" is the header of your file and they are not actual nodes.
Delete that line and use read.graph(..., format="ncol") to read the file into
igraph. This will not create you a bipartite graph yet, only a "simple" graph,
but the only thing you need to do in order to make it
Hi,
Please stay on the list with your follow-up emails.
Your code works fine for me (apart from the fact that you did not post your
full code so I had to guess the missing parts myself). Some suggestions:
1. igraph_real_t is simply a double, so you can use fscanf directly on your
input file wi
> I have a large numeric matrix in R as follows:
>
> > dim(libfactor.mat)
> [1] 4052310
> > head(libfactor.mat)
> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
> [1,] 20884 19184 19185 NA NA NA NA NA NANA
> [2,] 18330 13979 13978 NA NA NA NA NA NANA
Hi Bob,
> Sorry I only come back to you only now. shortest.paths does not seem to
> indicate the vertices that have been traversed in the shortest paths - just a
> matrix of those shortest paths.
Yup, but you can use get.shortest.paths() to get _one_ shortest path per vertex
pair, or get.all.s
Hi Adam,
Things are in a transition period right now; we have released igraph 0.6.5 and
the corresponding Python interface a few days ago, but Homebrew still installs
igraph 0.6 because I haven't had time to update the recipe yet. Try the
following:
easy_install python-igraph==0.6
This will f
Hi Adam,
FYI, I have filed a pull request to the Homebrew project to upgrade the igraph
formula to 0.6.5:
https://github.com/mxcl/homebrew/pull/18313
If you don't want to wait until the changes are merged, type "brew edit igraph"
and replace the red lines with the green ones as seen on the fol
Hi!
> Is this a bug or is something wrong? I'm using Python 2.7.3 on Ubuntu Linux
> 12.10 64-bit. I've installed igraph from the standard package repository.
What is the version number of the igraph module in Python? If
igraph.__version__ is less than 0.6, this means that you don't have this
fe
> I am graduating from PUC Minas Brazil and my work theme of completion is on
> the R language, specifically on the package IGRAPH. I wonder if you can help
> me understand how R calculates the centrality of edges in the function below:
The calculation is not implemented in R; it is implemented i
1 - 100 of 624 matches
Mail list logo