> Hi Markus, > > >> Hi Markus, > >> > >> your right, thats my 3rd attemt. > >> > >> The background is, after crawling with nutch, i export to results to > >> solr, and i make a relevance declaration by calculation the scoring > >> results. > > > > Solr can do its own relevancy calculation without using the boost > > field (which > > is also not indexed by default). Can't you just omit the boost field in > > your set up? > > Sorry, i don't know how - i will work on that... > > >> But i think the problem starts with crawling, not by exporting the > >> results to solr? > > > > It must be somewhere in Nutch indeed. > > Thank you! > > >> If you want, you can use the search on > >> http://www.tachtler.net --> (right top corner), for example search: > >> awstats (only two resaults 1st german, 2nd english) > > > > I see, where does the score come from? Not from Solr it looks like. > > That was your request: > INFO: [] webapp=/solr path=/select > params={fl=title,anchor,content,url,score,tstamp,lang,id&sort=score+desc&st > art=0&q=awstats&hl.fl=content&qt=standard&wt=javabin&hl=true&rows=10&versio > n=1} hits=2 status=0 QTime=362 > > In solr, i ask solr for the scoring result, the max. value is the > highest score i found, after that i calculate for every entry in the > result list it's own score, based on the max. score i found.
Ahaa, you calculate relevancy in percentages based on maxScore in the result set! There may be an error there. Can include the raw Solr output with &wt=xml (instead of javabin)? > > Which config files you need, that you can give me a little more help > on that problem, i will send you all files... > > Thank you so much! > > > Klaus. > > > -- > > ------------------------------------------------ > e-Mail : [email protected] > Homepage: http://www.tachtler.net > DokuWiki: http://www.dokuwiki.tachtler.net > ------------------------------------------------
