Now that D.O. is running solr!!!! woohoo. Nice work DamZ, and the rest of the infrastructure team.

http://drupal.org/search/apachesolr_search/organic%20groups

Is pretty bad. I toyed with the boosts, and the title field is heavily weighted as is the project type. Still, the main module page is nowhere near the top. I think this is because we omit norms. If we don't do this, then a closer match will be more valuable (is that correct peter?). So a title of "Organic Groups" is closer than "Organic Groups foo bar foo bar Organic". Currently, I believe the second one is higher ranked.

Comments

pwolanin’s picture

pwolanin commented

Well, as far as I understand, norms are somewhat expensive.

Also, for general search I think we decided that it did not make sense that matching 2 of 4 words should be less valuable than matching 2 of 2 words.

We have in the schema:

  <filter class="solr.RemoveDuplicatesTokenFilterFactory"/>

so I think the title with "Organic" twice should have an equal boost - but we should check that. Likely the other ones are getting boosted because they are newer?

pwolanin’s picture

pwolanin commented
Title: should not be ommitting norms » How do we make a "main" module come up first
janusman’s picture

janusman commented

If "main module" you mean "one that's used at many sites", then a boost on the number of sites implementing that module could help.

jpmckinney’s picture

jpmckinney commented
Status: Active » Closed (won't fix)

I don't really see a proposal here. Every site will want to configure Solr differently to tweak boosts, etc. Many of these tweaks need to be done at the level of the Solr config files. I don't see an opportunity for the module to help in that respect.