[Zope-dev] TextIndex Questions

Michel Pelletier michel@digicool.com
Thu, 01 Jun 2000 10:46:16 -0700


"R. David Murray" wrote:
> 
> On Sun, 28 May 2000, Chris Withers wrote:
> > > may be queried with a simple boolean query language that allows
> > > 'and', 'or', phrasing, parenthesized boolean expressions, and
> > > proximity matching. Relevance ranking is supported and returns the
> > > sum of the occurances of all query terms in the "hit". A normalized
> > > score is also provided that is normalized from 0 to 100 over the
> > > whole result set.
> >
> > This sounds really useful...
> >
> > Is it still true?

Except for proximity searching, it is all true.  I was in error when I
stated that proximity searching was  supported.  The code is there, buy
the back end storage and references to the objects to do the proximity
matching had to be ripped out of ZTables because of size constraints. 
You would have to create a meta-data table entry for your text content
that was indexed which is not efficient at all for a minor feature.
 
> I believe Michel said that the code for all this was still in
> ZCatalog but had not been tested.  'and' and 'or' work, but the
> rest I'm not sure about.

Boolean terms, parenthetical expressions, and phrase matching work and
have been in the code since 2.0 and came from yonder back in ZTables. 
They are pretty stable although I have done a lot of re-arranging.

Revelance ranking has been in the CVS for a while now.  I will document
it more in the interfaces Wiki.

There are a number of other features that are new and are the least
tested.  These include wildcard 'globbing' style searches with
GlobbingVocabularies.  There is some documentatin on this in the
interfaces wiki which I will expand on.

-- 

-Michel Pelletier

http://www.zope.org/Members/michel/MyWiki

Visit WikiCentral for the latest Zen:

http://www.zope.org/Members/WikiCentral