Hi Lewis, Thanks for information about this work. Emre worked at our company. I reviewed the code. The architecture of work based on an abstract RankingJob. It is similar to our old architecture of IndexerJob. Moreover Emre didn't use gora, ToolRunner or it didn't get crawlId etc. I want to create a job that similar our exists jobs. But I build based on knowledge of Emre's work . Wdyt ?
Talat 2014-05-03 19:59 GMT+03:00 Lewis John Mcgibbney <[email protected]>: > Hi Talat, > > > On Sat, May 3, 2014 at 4:35 AM, <[email protected]> wrote: >> >> As far as I know Giraph use only map slots as works. At the present our >> architecture of scoring plugins dont permit. IMHO We should create a >> pluggable RankingJob like as IndexingJob. The Pluggable architecture can >> permit us for implementing custom pagerank algorithms. Wdyt? >> >> Now we use different giraph algorithms similar thia solution. If this >> makes sense for everybody, After 2.3 releaes i can implement it. > > Last year there was work done to address this. > You can see it here > https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=31820383 > https://issues.apache.org/jira/browse/GIRAPH-729 > > > > -- Talat UYARER Websitesi: http://talat.uyarer.com Twitter: http://twitter.com/talatuyarer Linkedin: http://tr.linkedin.com/pub/talat-uyarer/10/142/304
