Hello, all. This has gone private by mistake. I guess there are many people interested in the same question, so i'm forwarding it back to list.
---------- Forwarded message ---------- From: Pieter v.d. Brink <[EMAIL PROTECTED]> To: Peter Farafonov <[EMAIL PROTECTED]> CC: Time: Mon, 04 Jun 2007 15:47:41 +0200 Subject: [fw-general] Zend Search Lucene, wildcard and partial matching support Attachments: <none> Hi Peter, Thanks for the information, I'm looking at it now and it's quite useful. Also found some existing PHP implementations for English stemmers via that site. Anyway, English is the most important for now - my database also contains a few German and Dutch articles, but I can implement stemming for these as seperate stemming filters later on. greetings, Pieter Peter Farafonov schreef: > Hello, Pieter. > > >> Also on the issue tracker it was suggested that partial matching is >> already possible with the proper analyzer, is there a standard analyzer >> available that can do this? (so for example the keyword "test" could >> match "tests" or "testing" and vice versa.) What about keywords in >> languages other than English? >> > > What you need is stemmer. Check out this site for more information: > http://snowball.tartarus.org/ > > It provides stemming algorithms for many languages and java code. You > have to rewrite the stemmer for your language in php (make the class > to be a child to TokenFilter). Then you just turn it on this way: > > $analyzer = new Zend_Search_Lucene_Analysis_Analyzer_xxxxx(); > $stemmer = new xxxxStemmer(); > $analyzer->addFilter($stemmer); > > ---------- End of forwarded message ---------- -- Best regards, Peter mailto:[EMAIL PROTECTED]
