Hi Martin, welcome back :-)
I'll try to merge the documentation PR tonight. Gelly is in the flink-staging package and most of the Gelly methods are in the Graph class and have javadocs that describe their functionality. Regarding your specific tasks, you can easily get the degree distribution using the inDegrees/outDegrees or getDegrees methods. These will give you a DataSet of <vertexId, degree> and then you can use the regular Flink API to get the distribution. For connected components, there is a isWeaklyConnected() method that returns whether a graph consists of a single component and the existing Spargel implementation. This should be trivial to modify to get the component distribution. Triangle count is a bit tricky if you have high degree nodes. Gelly doesn't have a library method for triangle count yet. Daniel has implemented a clustering coefficient algorithm that might be helpful to you: https://github.com/apache/flink/pull/420 If you'd like to implement and contribute a triangle counting library method for Gelly, I'd be more than happy to help! Let us know if you have any questions. Cheers, Vasia. On 26 February 2015 at 17:42, Andra Lungu <lungu.an...@gmail.com> wrote: > Hello, > > Also, for guidelines on how to implement a graph algorithm in Gelly, you > can > use the provided examples: > > https://github.com/apache/flink/tree/master/flink-staging/flink-gelly/src/main/java/org/apache/flink/graph/example > > Have fun! > Andra > > On Thu, Feb 26, 2015 at 5:31 PM, Fabian Hueske <fhue...@gmail.com> wrote: > > > Hi Martin, > > > > as a start, there is a PR with Gelly documentation: > > https://github.com/vasia/flink/blob/gelly-guide/docs/gelly_guide.md > > > > Cheers, Fabian > > > > 2015-02-26 17:12 GMT+01:00 Martin Neumann <mneum...@spotify.com>: > > > > > Hej, > > > > > > I was busy with other stuff for a while but I hope I will have more > time > > to > > > work on Flink and Graphs again now. > > > > > > I need to do some basic analytic's on a large graph set (stuff like > > degree > > > distribution, triangle count, component size distribution etc.) > > > Is there anything implemented in Gelli already that I can use? Is there > > > some basic documentation about Gelli somewhere? > > > > > > If I need to program those algorithm myself I could do it in a way that > > we > > > can integrate them in Gelli. But I might need some help to make sure I > > > design them in a way that fits, so I might need some guidance/help. > > > > > > cheers Martin > > > > > >