[Zope] ZCatalog: Non english text

[Michel Pelletier]

> -----Original Message-----
> From: technews@egsx.com [mailto:technews@egsx.com]
> Sent: Monday, November 22, 1999 2:00 PM
> To: Zope
> Subject: Re: [Zope] ZCatalog: Non english text
> 
> 
> Hi,
> 
> I used ZCatalog to index an object I created, which contains 
> a site_index,
> and site_description properties.  These properties were 
> indexed and are
> searchable hence when I do the following search in DHTML
> 
> <dtml-in "Catalog.searchResults(site_title='computer')">
>   <a href="/<dtml-var "Catalog.getpath(data_record_id_)">">
>     <dtml-var site_title></a><br>
>     <dtml-var site_description>    <p>
> </dtml-in>
> 
> I do get the correct results.  However when I enter a 
> non-english word,
> which is indexed as well, I do not get anything else.  I did hear that
> someone tries using ZCatalog with chinease and were unscussesful, has
> there been any changes??

With Zope 2.1, you can specify a locale on the command line '-L'.
Chinese is a whole different story because it's a multi-byte character
set, and not even Python has worked out what the story on that is.  Most
european languages work with locale.
 
> Also what are the chances that ZCatalog can allow us to 
> include a function
> to be used when indexing and searching of the objects, so we 
> can apply any
> word tranformation such as suffix removal i.e. computing, computers,
> computer, computed, can be indexed as comput if I choose to.  
> What I mean
> is that we specify a method (possibly external) that will do what is
> needed and the resulting word can be indexed by ZCatalog, and the same
> method is called to transform the input word usinf the same 
> method when
> searching that catalog.

The catalog allready stems words just like you have here.  In future
releases, there will be an interface defined where you can create your
own stemming methods.

-Michel