Re: fast case-insensitive autocomplete
Hi, Try this solution using a Solr core: http://www.lucidimagination.com/blog/2009/09/08/auto-suggest-from-popular-queries-using-edgengrams/ -- Jan Høydahl, search solution architect Cominvent AS - www.cominvent.com On 5. mai 2011, at 15.22, Kusenda, Brandyn J wrote: Hi. I need an autocomplete solution to handle case-insensitive queries but return the original text with the case still intact. I've experimented with both the Suggester and TermComponent methods. TermComponent is working when I use the regex option, however, it is far to slow. I get the speed i want by using term.prefix for by using the suggester but it's case sensitive. Here is an example operating on a user directory: Query: bran Results: Branden Smith, Brandon Thompson, Brandon Verner, Brandy Finny, Brian Smith, ... A solution that I would expect to work would be to store two fields; one containing the original text and the other containing the lowercase. Then convert the query to lower case and run the query against the lower case field and return the original (case preserved) field. Unfortunately, I can't get a TermComponent query to return additional fields. It only returns the field it's searching against. Should this work or can I only return additional fields for standard queries. Thanks in advance, Brandyn
Re: fast case-insensitive autocomplete
Hi, I haven't used Suggester yet, but couldn't you feed it all lowercase content and then lowercase whatever the user is typing before sending it to Suggester to avoid case mismatch? Autocomplete on http://search-lucene.com/ uses http://sematext.com/products/autocomplete/index.html if you want a shortcut. Otis Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch Lucene ecosystem search :: http://search-lucene.com/ - Original Message From: Kusenda, Brandyn J brandyn-kuse...@uiowa.edu To: solr-user@lucene.apache.org solr-user@lucene.apache.org Sent: Thu, May 5, 2011 9:22:03 AM Subject: fast case-insensitive autocomplete Hi. I need an autocomplete solution to handle case-insensitive queries but return the original text with the case still intact. I've experimented with both the Suggester and TermComponent methods. TermComponent is working when I use the regex option, however, it is far to slow. I get the speed i want by using term.prefix for by using the suggester but it's case sensitive. Here is an example operating on a user directory: Query: bran Results: Branden Smith, Brandon Thompson, Brandon Verner, Brandy Finny, Brian Smith, ... A solution that I would expect to work would be to store two fields; one containing the original text and the other containing the lowercase. Then convert the query to lower case and run the query against the lower case field and return the original (case preserved) field. Unfortunately, I can't get a TermComponent query to return additional fields. It only returns the field it's searching against. Should this work or can I only return additional fields for standard queries. Thanks in advance, Brandyn
Re: fast case-insensitive autocomplete
Are you giving that solution away? What is the costs? etc!! On Thu, May 5, 2011 at 2:58 PM, Otis Gospodnetic otis_gospodne...@yahoo.com wrote: Hi, I haven't used Suggester yet, but couldn't you feed it all lowercase content and then lowercase whatever the user is typing before sending it to Suggester to avoid case mismatch? Autocomplete on http://search-lucene.com/ uses http://sematext.com/products/autocomplete/index.html if you want a shortcut. Otis Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch Lucene ecosystem search :: http://search-lucene.com/ - Original Message From: Kusenda, Brandyn J brandyn-kuse...@uiowa.edu To: solr-user@lucene.apache.org solr-user@lucene.apache.org Sent: Thu, May 5, 2011 9:22:03 AM Subject: fast case-insensitive autocomplete Hi. I need an autocomplete solution to handle case-insensitive queries but return the original text with the case still intact. I've experimented with both the Suggester and TermComponent methods. TermComponent is working when I use the regex option, however, it is far to slow. I get the speed i want by using term.prefix for by using the suggester but it's case sensitive. Here is an example operating on a user directory: Query: bran Results: Branden Smith, Brandon Thompson, Brandon Verner, Brandy Finny, Brian Smith, ... A solution that I would expect to work would be to store two fields; one containing the original text and the other containing the lowercase. Then convert the query to lower case and run the query against the lower case field and return the original (case preserved) field. Unfortunately, I can't get a TermComponent query to return additional fields. It only returns the field it's searching against. Should this work or can I only return additional fields for standard queries. Thanks in advance, Brandyn