[Zope] ZCatalog: searching for parts of words?

Michel Pelletier michel@digicool.com
Tue, 28 Sep 1999 09:31:25 -0400


> -----Original Message-----
> From: Thomas Weiner [mailto:weiner@tu-harburg.de]
> Sent: Tuesday, September 28, 1999 8:55 AM
> To: zope@zope.org
> Subject: [Zope] ZCatalog: searching for parts of words?
> 
> 
> Hello Zopestas,
> 
> is there a way to use the full textindex of ZCatalog for finding parts
> of words?

In general, no.

> For Example:
> 
> seeking for 'word' would also bring up 'words'?

This will work because of certain rules of english (words is the plural
of word) and the Catalog does what's called 'stemming' to remove common
english suffixes from words, like the 's' from the end of a word to
singularize it.
 
> I'm not sure, whether this is just a special issue in German - due to
> the fact that German accepts any thinkable and unthinkable combination
> of joined nouns.

The CVS version of Zope has been much improved in this area.  First,
'z2.py' offers the '-L' switch to set your locale.  This is useful so
that the Catalog doesn't split up your words on non-english characters
like umlauts and such.  Second, ZCatalog supports a new feature called a
Lexicon, which allows you to have more control over stopwords and
stemming rules.  More improvement is needed, but I'm not exactly sure
where to go without some specific feedback.

In the future, there may be 'Pluggable Lexicons' for various languages.

-Michel