this is actually done in trunk. In trunk fuzzy's enum is a "proxy". for low distances (ed=1,2) it uses automaton.
for higher distances it uses the crappy "brute force" method. but, higher distances still get accelerated if you use a reasonable 'maxExpansions' to FuzzyQuery... the default is quite bad (1024). On Fri, Jul 23, 2010 at 10:59 AM, <[email protected]> wrote: > Thanks! > > > > FuzzyQuery will do for my purposes, for the interim. But I suspect that > FuzzyQuery could be made a lot more efficient if it were rebuilt on top of > Automaton, no? I understand that this would be a trunk project. > > > > Karl > > > > > > *From:* ext Uwe Schindler [mailto:[email protected]] > *Sent:* Friday, July 23, 2010 10:45 AM > > *To:* [email protected] > *Subject:* RE: LevenshteinFilter proposal > > > > Automaton is only in Lucene/Solr Trunk. To get a filter out of FuzzyQuery, > use MultiTermQueryWrapperFilter(new FuzzyQuery(…)) > > > > ----- > > Uwe Schindler > > H.-H.-Meier-Allee 63, D-28213 Bremen > > http://www.thetaphi.de > > eMail: [email protected] > > > > *From:* [email protected] [mailto:[email protected]] > *Sent:* Friday, July 23, 2010 4:25 PM > *To:* [email protected] > *Subject:* LevenshteinFilter proposal > > > > Hi Folks, > > > I’m very interested in using (or developing!) a Levenshtein Filter within > the family of Solr Filter objects. I don’t see such a class today anywhere. > I see how the AutomatonQuery object would permit such a thing to be built, > but to date I don’t know of anyone who has built one. Do you? If not, I’m > willing to give it a whirl. Also, AutomatonQuery doesn’t seem to come up > when I look for it in the javadocs for Lucene – can you point me in the > correct direction? > > Thanks! > Karl > > > > > -- Robert Muir [email protected]
