From f7a4377812791e7dfc3890d338bd824d09562222 Mon Sep 17 00:00:00 2001 From: Michael Peter Christen Date: Fri, 7 Jun 2013 13:22:22 +0200 Subject: [PATCH] usage of the new normalized link polularity CRn as default ranking function. This replaces the previous formula, which was bad. Before you update to this version, please check if you changed the ranking function yourself before, since it will be overwritten. --- defaults/yacy.init | 9 ++++----- source/net/yacy/search/SwitchboardConstants.java | 2 +- 2 files changed, 5 insertions(+), 6 deletions(-) diff --git a/defaults/yacy.init b/defaults/yacy.init index 726a7d9dc..e9ceda2aa 100644 --- a/defaults/yacy.init +++ b/defaults/yacy.init @@ -964,20 +964,19 @@ search.ranking.rwi.profile = search.ranking.solr.collection.boostname.tmpa.0=Default Profile search.ranking.solr.collection.boostfields.tmpa.0=url_paths_sxt^1000.0,synonyms_sxt^1.0,title^10000.0,text_t^2.0,h1_txt^1000.0,h2_txt^100.0,host_organization_s^100000.0 search.ranking.solr.collection.boostquery.tmpa.0=fuzzy_signature_unique_b:true^100000.0 -search.ranking.solr.collection.boostfunction.tmpa.0=scale(references_external_i,1,20) -#search.ranking.solr.collection.boostfunction.tmpa.0=product(recip(rord(last_modified),1,1000,1000),div(product(log(product(references_external_i,references_exthosts_i)),div(references_internal_i,host_extent_i)),add(clickdepth_i,1))) +search.ranking.solr.collection.boostfunction.tmpb.0=scale(cr_host_norm_i,1,20) search.ranking.solr.collection.boostname.tmpa.1=Date Profile: sort by date in descending order for a '/data' usage search.ranking.solr.collection.boostfields.tmpa.1=text_t^1.0 search.ranking.solr.collection.boostquery.tmpa.1=fuzzy_signature_unique_b:true^100000.0 -search.ranking.solr.collection.boostfunction.tmpa.1=recip(rord(last_modified),1,1000,1000) +search.ranking.solr.collection.boostfunction.tmpb.1=recip(rord(last_modified),1,1000,1000) search.ranking.solr.collection.boostname.tmpa.2=Intranet Profile: when a search is done on a singe domain only, i.e. if a site:-operator is used search.ranking.solr.collection.boostfields.tmpa.2=url_paths_sxt^1000.0,synonyms_sxt^1.0,title^10000.0,text_t^2.0,h1_txt^1000.0,h2_txt^100.0 search.ranking.solr.collection.boostquery.tmpa.2=fuzzy_signature_unique_b:true^100000.0 -search.ranking.solr.collection.boostfunction.tmpa.2=product(recip(rord(last_modified),1,10000,10000),product(references_exthosts_i,pow(div(references_internal_i,host_extent_i),2))) +search.ranking.solr.collection.boostfunction.tmpb.2=scale(cr_host_norm_i,1,20) search.ranking.solr.collection.boostname.tmpa.3=_unused3 search.ranking.solr.collection.boostfields.tmpa.3=text_t^1.0 search.ranking.solr.collection.boostquery.tmpa.3=fuzzy_signature_unique_b:true^100000.0 -search.ranking.solr.collection.boostfunction.tmpa.3=div(add(1,references_i),add(url_chars_i,pow(clickdepth_i,2))) +search.ranking.solr.collection.boostfunction.tmpb.3= # the following values are used to identify duplicate content search.ranking.solr.doubledetection.minlength=3 diff --git a/source/net/yacy/search/SwitchboardConstants.java b/source/net/yacy/search/SwitchboardConstants.java index 61eabc3bd..36868777c 100644 --- a/source/net/yacy/search/SwitchboardConstants.java +++ b/source/net/yacy/search/SwitchboardConstants.java @@ -493,7 +493,7 @@ public final class SwitchboardConstants { public static final String SEARCH_RANKING_SOLR_COLLECTION_BOOSTNAME_ = "search.ranking.solr.collection.boostname.tmpa."; // temporary until we know best default values; add the index number (0..3) to that string public static final String SEARCH_RANKING_SOLR_COLLECTION_BOOSTFIELDS_ = "search.ranking.solr.collection.boostfields.tmpa."; public static final String SEARCH_RANKING_SOLR_COLLECTION_BOOSTQUERY_ = "search.ranking.solr.collection.boostquery.tmpa."; - public static final String SEARCH_RANKING_SOLR_COLLECTION_BOOSTFUNCTION_ = "search.ranking.solr.collection.boostfunction.tmpa."; + public static final String SEARCH_RANKING_SOLR_COLLECTION_BOOSTFUNCTION_ = "search.ranking.solr.collection.boostfunction.tmpb."; /** * system tray