Check out the highlighter code in the contribs folder. Should do what u need.

Ben

On Saturday, August 7, 2010,  <e...@telexis.com> wrote:
> Sorry if duplicate post, I wasn't subscribed when I did the 1st one, so not 
> sure if it was aloud or blocked by spam filters.  This is my first "post", so 
> I hope I'm doing this correctly.  I've only used Google Groups or on-line 
> forums in the past, never email based lists.
>
> I'm using CLucene to index a custom ISAM legacy database of mine.  The 
> database holds emails and other messages.  I've got the indexing working, and 
> also the searching.  However, our current (very slow) search method provides 
> a "summary" for each search result.  Basically fragments of the text around 
> the words that were found.
>
> I have to read the records from my database for each result found by CLucene 
> anyway, because I don't want to store all that data in Lucene, so I was 
> thinking I would just scan my text manually to provide the same "summary" I'm 
> currently providing.  However, now I realize that isn't as simple as it was 
> before, because I don't always know what words/phrases (I think Lucene calls 
> them "terms") were found.  For example, if the user does a fuzzy or wildcard 
> search.  Without recreating the CLucene logic, how can I do this?
>
> I understand that there are some additional libraries (maybe only for the 
> Java version?) that do "highlighting", which is similar to what I want to do. 
>  I haven't found any source code for them yet, and even if I did, I'm not 
> sure I could figure out how to extract that and recreate for my purposes 
> (they may assume the text is stored in Lucene, or that I want the results in 
> HTML format).  My scenario seems to be unique, because I'm not working with a 
> website, or indexing basic files.
>
> If I could find the section of code in CLucene where it "finds" results for 
> my query, then I would know what words it found for each document.  I'm not 
> too worried about showing the "most relavent" fragments at this point, 
> although that would be a nice feature down the road.
>
> p.s.-I'm loving CLucene so far, and if I can get past this last hurdle, I 
> should be set =]
>
> -Eric Selk
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by
>
> Make an app they can't live without
> Enter the BlackBerry Developer Challenge
> http://p.sf.net/sfu/RIM-dev2dev
> _______________________________________________
> CLucene-developers mailing list
> CLucene-developers@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/clucene-developers
>

------------------------------------------------------------------------------
This SF.net email is sponsored by 

Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev 
_______________________________________________
CLucene-developers mailing list
CLucene-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/clucene-developers

Reply via email to