[opencms-dev] pls Help

bekalu mamo bekalum at yahoo.com
Sun Mar 8 13:44:06 CET 2009


Hi,

As you have said the standard analyzer works for amharic contents but with a compromise in its efficiency. for the last couple of weeks I have tried to develop an amharic analyzer by getting the source code from lucene project. The analyzer works fine for me when I test it by adding as a library to other classes. So inorder to use that analyzer in opencms, I take the jar file (lucene-analyzers-2.3.1.jar) which is the one having the amharic analyzer integrated in it and then substitute this one with the lucene-analyzers-2.3.1.jar file that exists in the opencms library folder, then I added the analyzer in opencms-search.xml file. When I check to index amharic contents, all that I got was ...skipped, hence I just swapped only the locale (not the class) of am and en, then the analyzer class for Amharic indexes my file with en locale but the standard analyzer will not index any thing with am locale. I'm confused what to do b/c the new analyzer works when the
 locale parameter in the opencms-search is changed to like en or de but existing analyzer classes will not work with am locale. If there is any idea please help me,

with best regards,

bekalu




________________________________
From: Christian Steinert <christian_steinert at web.de>
To: The OpenCms mailing list <opencms-dev at opencms.org>
Sent: Tuesday, February 17, 2009 1:32:28 AM
Subject: Re: [opencms-dev] pls Help



Hi Sebarsian,

I've been working alot to find my language (amharic) analyzer for lucene as well as for opencms, but till now i couldn't find any. May be what I'm thinking is to write some code for the lucene, and then try to use that one in opencms. Well actually i'm not sure enough the lucene used inside the opencms is identical to the source code available for lucene. If there is any idea, please let me know. 


Yes - opencms uses lucene from http://lucene.apache.org/. All Analyzers 
that exist for modern versions of Lucene should also be usable with OpenCms.
Of course, it could be that no analyzer exists for your Language. 
Writing analyzers requires knowing the language well and also requires 
knowing a lot about Lucene and some things about language processing. 
Because of this, for many languages, nobody has managed yet to write an 
Analyzer yet.

P.S. you do not absolutely NEED an analyzer for your language. 
An Analyzer will give much better search result because it will find
words even, if an ending is added to the word. 

But you can always use the English analyzer to have at least some
search functionality. The result will not be perfect because the search
engine will only fine those words that are exactly identical to what
you search for, but at least the search will work in some way

Best Regards
Christian



      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opencms.org/pipermail/opencms-dev/attachments/20090308/e0ea1a65/attachment.html>


More information about the opencms-dev mailing list