[ 
https://issues.apache.org/jira/browse/LUCENE-2244?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12916044#action_12916044
 ] 

Robert Muir commented on LUCENE-2244:
-------------------------------------

Andi, now that LUCENE-2167 is resolved, i think we should revisit this one.

There are two issues:
* does the tokenization work the way it should for these quotes (I think 
LUCENE-2167 will do the right thing)?
* does the 's-stripping (analysis.en.EnglishPossessiveFilter) work the way you 
want?

For the latter: it could be extended itself to include more apostrophes, maybe 
from this list: http://unicode.org/cldr/utility/confusables.jsp?a='&r=None

Or someone could put ASCIIFoldingFilter before the EnglishPossessiveFilter.


> Improve StandardTokenizer's understanding of non ASCII punctuation and quotes
> -----------------------------------------------------------------------------
>
>                 Key: LUCENE-2244
>                 URL: https://issues.apache.org/jira/browse/LUCENE-2244
>             Project: Lucene - Java
>          Issue Type: Bug
>          Components: contrib/analyzers
>    Affects Versions: 3.0
>            Reporter: Andi Vajda
>         Attachments: StandardTokenizerImpl.jflex.diff
>
>
> In the vein of LUCENE-1126 and LUCENE-1390, StandardTokenizerImpl.jflex 
> should do a better job at understanding non-ASCII punctuation characters.
> For example, its understanding of the single-quote character "'" is currently 
> limited to that character only. It will set a token's type to APOSTROPHE only 
> if the "'" was used.
> In the patch attached, I added all the characters that ASCIIFoldingFilter 
> would change into "'".
> I'm not sure that this is the right approach so I didn't write a complete 
> patch for all the other hardcoded characters used in jflex rules such as ".", 
> "-" which have some variants in ASCIIFoldingFilter that could be used as well.
> Maybe a better approach would be to make it possible to have an 
> ASCIIFoldingFilter-like reader as a character filter that could be in 
> inserted in front of StandardTokenizer ?

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org

Reply via email to