[Zope] RE: ZCatalog Indexing

28 Mar 2000

      Tobias,

The Catalgo uses a stripper to make the indexes smaller. It doesn't
index =-+*&)(^%$#@. etc... or numbers.  Some words too like 'Bill'. From
what I've been told there is a version of the Catalog on it's way which
will allow you to put keywords or charaters in a special place that will
include them.  I couldn't wait for that, and I love Zope, so I came up
with a major kludge.  catalog_quote() and catalog_unquote().  They are
from an External Method which I wrote catalog_quote.py.  Here it is:

from string import strip, split, join, uppercase, lowercase, find
bad_characters=(('.'),
                ('+'),
                ('-'),
                (';'),
                (':'),
                ('&'),
                ('*'),
                ('%'),
                ('#'),
                ('@'),
                ('$'),
                ('0'),
                ('1'),
                ('2'),
                ('3'),
                ('4'),
                ('5'),
                ('6'),
                ('7'),
                ('8'),
                ('9'))

def catalog_quote(self, ascii):
    text = str(ascii)
    for re in bad_characters:
        if find(text, re) >= 0: text=join(split(text,re),ordinalize(re))
    return text

def ordinalize(letter):
    number = ord(letter)
    number_string = str(number)
    myletters = lowercase[-5:] + uppercase[-10:]
    converted = ''
    for character in number_string:
        converted = converted + myletters[int(character)]
    return converted  + 'g'

def catalog_unquote(self, ascii):
    text = str(ascii)
    for un in bad_characters:
        if find(text, ordinalize(un)) >= 0:
text=join(split(text,ordinalize(un)),un)
    return text

I just use it to turn anything that might contain the charaters in
bad_charaters into catalogable stuff.

Then when you search, you must catalog_quote() your search strings, and
you'll get 100% accurate results.  Hope this helps.

All my best,

-- 
Jason Spisak
444@hiretechs.com