Hi James,
don't give up, your very close to having it work. If you can get CJKAnalyzer
and CJKTokenizer to work in Lucene, you should also be able to get it to
work in Solr. Look at the bright site, at least ant doesn't throw any
errors. And my code isn't going to work, since it really cann't handle
Chinese or Japanese characters. You should have a look at how you did things
in Lucene.
I have went through my archives and I have found that people also have used
something similar to:
<fieldtype name="text" class="solr.TextField">
<analyzer class="org.apache.lucene.analysis.ja.JapaneseAnalyzer"/>
</fieldtype>
Maybe you can try it this way, and forget about the FilterFactories. Let me
know how it goes.
Kind regards,
Nick
On 9/19/06, James liu <[EMAIL PROTECTED]> wrote:
today not Ok。
i check source of cjk: CJKAnalyzer.java and CJKTokenizer.java (these from
lucene 2.0 source code)
and your code,,,i write CJKJLFilterFactory.java and
CJKJLTokenizerFactory.java
ant is ok.i copy new solr.war to tomcat's webapps,,,,and modify schema.xml
use admin page, i use http://localhost:8484/solr/admin/analysis.jsp?highlight=on
to check word analyze
it show me *()*)&*^&*, oh my god. i m failure.
i see org.apache.lucene.analysis.nl's code, i find something difference.
like your Tokenizer sames as StandardTokenizer,,i have to use myself..
thk u very much, no your code, i think i maybe give up. i only use delphi
and php, no java and unix before i meet lucene.
i use lucene well and i think i can use solr well.
thk u again.
my msn: [EMAIL PROTECTED],,,maybe we can be friend.
