[Networkit] NetworKit 3.4 released

Christian Staudt christian.staudt at kit.edu
Thu Dec 4 18:06:31 CET 2014


Today we have released version 3.4 of NetworKit, the open-source toolkit for high-performance network analysis. This release brings numerous critical bugfixes as well as useful incremental features and performance optimizations. We are also moving towards consistent interfaces for algorithms. We have also further simplified the installation dependencies (try: pip3 install networkit).

Thank you to the numerous people who have contributed code to this release. 

More information can be found on https://networkit.iti.kit.edu/. We welcome user feedback and opportunities for collaboration.

Release Notes

Features

 * graph
   * Graph can be copied on Python level
   * spanning tree/forest (graph.SpanningForest)
 *  algorithms in general
   * Edmonds-Karp max flow algorithm (flow.EdmondsKarp)
   * core decomposition works for directed graphs (properties.CoreDecomposition)
   * algebraic distance, a structural distance measure in graphs (distance.AlgebraicDistance)
 * IO
   * there is no longer a default graph file format
   * read and write the GML graph file format (graphio.GMLGraphReader/Writer)
   * conversion of directed to undirected graph (Graph.toUndirected)
   * reader and writer for the GraphTool binary graph format (graphio.GraphToolBinaryReader)
   * METIS graph reader supports arbitrary edge weights (graphio.METISGraphReader)
 * algebraic
   * algebraic backend supports rectangular matrices (Matrix.h)
 * community detection
   * turbo mode for PLM community detection algorithm gives a factor 2 speedup at the cost of more memory (community.PLM)
   * Cut Clustering community detection algorithm (community.CutClustering)
 * generators
   * Erdös-Renyi generator can generate directed graphs (generators.ErdosRenyiGenerator)
   * configuration model graph generator for generating a random simple graph with exactly the given degree sequence (generators.ConfigurationModelGenerator)
   * generator for power law degree sequences (generators.PowerlawDegreeSequence)

Bugfixes
 * GraphMLReader improved (graphio.GraphMLReader)
 * ConnectedComponents usability improved
 * KONECT reader (graphio.KONECTGraphReader)
 * fixed build problem on case-insensitive file systems
 * closed memory leaks by adding missing destructors on the Cython 
 * improved memory management by adding missing move constructors
 * DynamicForestFireGenerator fixed

Refactoring
 * standardization of analysis algorithm interface: parameters given by constructor, computation triggered in run method, results retrieved via getter methods
 * run methods return self to allow chaining
 * introducing unit tests on Python layer

Build and Installation
 * pip installation does no longer require Cython
 * pip installation does no longer require SCons, minimal build system as fallback if SCons is missing



Best regards 
Christian Staudt

christian.staudt at kit.edu
http://parco.iti.kit.edu/staudt/index-en.shtml
Institute of Theoretical Informatics - Parallel Computing Group 
Building 50.34 Room 034
Karlsruhe Institute of Technology (KIT)



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ira.uni-karlsruhe.de/mailman/private/networkit/attachments/20141204/ad9fa8a6/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4923 bytes
Desc: not available
URL: <https://lists.ira.uni-karlsruhe.de/mailman/private/networkit/attachments/20141204/ad9fa8a6/attachment.p7s>


More information about the NetworKit mailing list