views:

434

answers:

6

I need to add full web search to my site. I need something like Google Custom Search but with no ads and it has to be free. Any recommendation of a web service or open source project that can index my site and allow me to search it will be helpful.

My site is made in ruby on rails, if that helps.

I'll make this question community-wiki so you can edit my bad English. I think many people can benefit from this question.

+6  A: 

Check out Lucene. It's an open source search engine that will certainly be a fun learning experience to implement on your own site. It was originally designed by the Excite folks, I do believe.

Robert Elwell
While lucene is excellent, I shy away from it for ruby project so that another technology is not introduced. Why support a java-based web container if there are ruby-based options. That said, if you are considering Lucene, go straight to Solr!
David Medinets
+6  A: 

Ferret is the Ruby port of Lucene. Check out the acts_as_ferret plugin.

tvanfosson
It took me all of an hour to get my ferret search working. I don't know it's production scalability but it was definitely easy to use.
David Medinets
+5  A: 

People are going to acts_as_solr and thinking sphinx in the blogs i read:

http://acts-as-solr.rubyforge.org/

http://ts.freelancing-gods.com/

I've aslo been looking at tsearch in postgres, it looks very capable:

http://www.sai.msu.su/~megera/postgres/gist/tsearch/V2/

Gene T
+5  A: 

Depends what you mean by full web search really. If you want to search the whole web then the answers above wont help you much as they are really for indexing and searching the content of your site. I would suggest using the Google ajax search (just a 'powered by google' needed, no ads) or Boss from yahoo (might require ads not sure).

http://code.google.com/apis/ajaxsearch/

http://developer.yahoo.com/search/boss/

+2  A: 

I recommend acts_as_xapian. It's very easy to implement, it's fast enough, and it's the got the features you'll normally need.

Dave Nolan
+3  A: 

What do you mean by "full web search"?

The are good answers available for full-text search where a search engine indexes and queries the model objects stored in your database.

If you mean something that indexes and queries your rendered HTML, Nutch is a popular option with a web-crawler, parser, indexer, and query interface.

sock