[Networkit] EdgeListWriter

Elisabetta Bergamini elisabetta.bergamini at kit.edu
Thu Dec 15 11:46:31 CET 2016


Hi Matteo,

I agree with what you wrote, I also think the class is doing what it’s supposed to do and the documentation should be changed accordingly. My guess is that the class was initially developed for LFR benchmark generators and then at some point it was changed to be compatible with EdgeListReader, but the documentation hasn’t been updated accordingly.
Thanks a lot,

Elisabetta

==========================================================
Karlsruhe Institute of Technology (KIT)
Institute of Theoretical Informatics (ITI)

Elisabetta Bergamini
Theoret. Informatics / Parallel Computing

Phone: +49-721-608-41821
Web: http://parco.iti.kit.edu/bergamini

KIT - The Research University in the Helmholtz Association
==========================================================

On Dec 14, 2016, at 15:09, Matteo Riondato <matteo at cs.brown.edu> wrote:

> Hi all,
> 
> The EdgeListWriter class is not documented in the io module (https://networkit.iti.kit.edu/api/doxyhtml/group__io.html). 
> I committed a one-line fix to the Dev branch. 
> 
> While here, I noticed that the comment at the beginning of EdgeListWriter says:
> 
> * A writer for the edge list format used by the LFR benchmark generators, defined as:
> * 		list of edges (nodes are labelled from 1 to the number of nodes;
> * 		the edges are ordered and repeated twice, i.e. source-target and target-source).
> 
> The edges are definitively not repeated twice when the graph is undirected (I tested it and this can also be concluded by looking at the code), and I believe also when the graph is directed (the code is using Graph.forEdges( )).
> 
> Also, the nodes are not necessarily labeled from 1 (the constructor allows the user to choose the starting node index).
> 
> I’m not sure whether the current implementation is what is desired (in my opinion, yes, this is what I expect from a class called EdgeListWriter) and the comment should be fixed, or whether the implementation should be fixed (my opinion: no, rather one should create a LFRWriter class that achieves what the comment says.)
> 
> Before doing any of these, I’d love to know what the consensus is.
> 
> Thank you,
> 
> Matteo
> 
> 
> _______________________________________________
> NetworKit mailing list
> NetworKit at ira.uni-karlsruhe.de
> https://lists.ira.uni-karlsruhe.de/mailman/listinfo/networkit

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ira.uni-karlsruhe.de/pipermail/networkit/attachments/20161215/90d0f94d/attachment.html>


More information about the NetworKit mailing list