lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bill Taylor <watay...@as-st.com>
Subject Re: Help with Custom Analyzer
Date Mon, 16 Oct 2006 22:18:29 GMT
It is not THAT hard to write a custom analyzer, that is what I did.  I 
found that there is a bug in the setup, however, in that there are two 
incompatible definitions of Token.  The generated file 
<xx>Tokenizer.java refers to the wrong definition of Token so I ahve to 
patch it before it will compile.

Other than that, doing a custom analyzer is pretty simple.


On Oct 16, 2006, at 5:32 PM, Otis Gospodnetic wrote:

> Hi Ryan,
>
> StandardAnalyzer should already be smart about keeping email addresses 
> as a single token:
>
>   // email addresses
> | <EMAIL: <ALPHANUM> (("."|"-"|"_") <ALPHANUM>)* "@" <ALPHANUM>

> (("."|"-") <ALPHANUM>)+ >
>
> (this is from StandardAnalyzer.jj)
>
> As for changing the text you feed to Lucene, that's all up to you.  
> Changing the String seems like the simplest approach.  If you want to 
> wrap that in StringReader, you can, but you can also just work with 
> Strings.
>
> Otis
>
> ----- Original Message ----
> From: Ryan O'Hara <ohara@genome.chop.edu>
> To: java-user@lucene.apache.org
> Sent: Monday, October 16, 2006 4:28:35 PM
> Subject: Help with Custom Analyzer
>
> I have a few questions regarding writing a custom analyzer.
>
> My situation is that I would like to use the StandardAnalyzer but
> with some data-specific rules.  I was wondering if there was a way of
> telling the StandardAnalyzer to treat a string of text, that would
> normally be tokenized into more than one token, as only one token
> (maybe by inserting quotes around the text).  For example, say the
> StandardAnalyzer normally splits the string of text
> ohara@genome.chop.edu into 4 tokens, but I want it to split the
> string into only 1 token.  Could I accomplish this by surrounding the
> string with quotes or by using some other type of flag?
>
> Another question I have is how do I modify the text being analyzed?
>  From how I interpreted what I have read (which could easily be off),
> it looks like in order to accomplish what I have previously
> described, I am going to have to add some code to my custom
> analyzer's tokenStream method.  I see that tokenStream() has a Field
> and a Reader as parameters.  Would the way I go about adding rules be
> to edit the reader text?  If so, would manipulation of the text be
> easier if I were to convert the reader into a string?
>
> Any help is greatly appreciated.  Thanks.
>
> -Ryan
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Mime
View raw message