[Networkit] networKIT sparsification

Christian Staudt christian.staudt at kit.edu
Wed May 4 20:13:43 CEST 2016


Hi Stephen,

> It does start now, though Gephi doesn’t show anything except a black rectangle - maybe I need to configure a renderer?

Could it be that the black rectangle is actually a large bunch of nodes? Have you tried applying one of Gephi’s layout algorithms to disentangle it?

As for details on the Gephi streaming interface, I refer you to Gerd Lindner (in CC), who is the main developer of this component. It is best to have these discussions on the networkit mailing list (also in CC, registration link on the networkit website), so everyone interested can learn from them or join in.

Best,
Chris

> On 04 May 2016, at 20:05, Stephen North <scnorth at gmail.com> wrote:
> 
> Hi Christian.  Thanks again for the help.
> 
> I got pretty far (at least far enough to see how sparsifiers are invoked).
> 
> In that particular notebook, my experiment hangs in the cell that invokes Gephi.
> Could not connect to the gephi streaming plugin. Did you start the streaming master server in gephi?
> 
> I did eventually figure out that I need to install Gephi, also to separate download and install the streaming plugin,
> also to start the server manually, and I disabled SSL (since the code connects to localhost:8080/workspace0 not port 8443).
> I had to change localhost:8080/workspace0 to localhost:8080/workspace1   (I have no idea why, and I had to also track
> down the right copy of NetworKit, since I grabbed the source from your repository, but installed the executable using pip3).
> It does start now, though Gephi doesn’t show anything except a black rectangle - maybe I need to configure a renderer?
> Even some minimal advice here would be helpful.  The error message might explain the above instructions briefly, as in
> “did you make sure to install gephi and the streaming plugin, start them both on port 8080, and create workspace0?”
> Gephi is particularly confusing because it wants to call the initial workspace “Workspace 1” (note the space).
> 
> Hope this makes sense.
> 
> Stephen
> 
>> On May 4, 2016, at 8:06 AM, Christian Staudt <christian.staudt at kit.edu <mailto:christian.staudt at kit.edu>> wrote:
>> 
>> Hi Stephen,
>> 
>> the main repo contains a Jupyter notebook that should help you with this:
>> 	Docs/Notebooks/Sparsification.ipynb
>> 
>> Best,
>> Chris
>> 
>>> On 03 May 2016, at 22:17, Stephen North <scnorth at gmail.com <mailto:scnorth at gmail.com>> wrote:
>>> 
>>> Hi Christian. Please can you give me an example or two that show how sparsifiers can be
>>> called from python?  I don’t have any intrinsic node or edge scores, at least not yet.
>>> I’m sort of lost here; it appears a sparsifier is a sort of factory or an algorithm generator(?)
>>> 
>>> I stared at test_sparsification.py for a while but I was still not sure.
>>> 
>>> I found the paper Structure-Preserving Sparsification of Social Networks; very interesting.
>>> 
>>> Stephen North
>>> 
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ira.uni-karlsruhe.de/pipermail/networkit/attachments/20160504/f6cee1ae/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 495 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.ira.uni-karlsruhe.de/pipermail/networkit/attachments/20160504/f6cee1ae/attachment.sig>


More information about the NetworKit mailing list