Skip to main content.
home | support | download

Back to List Archive

Re: Future: Collation of Meta Tags?

From: Peter Karman <peter(at)not-real.peknet.com>
Date: Sun Nov 06 2005 - 04:22:01 GMT
A similar function already exists in the API (libswish-e) but not in the swish-e 
command line tool. Look at SetSearchLimit() in the API docs and the proplimit.c 
src file.

Example (using my swishdebug script that uses the perl API):

karpet@cartermac 43% cat date*html
<html>
<head>
<meta name="date" content="17760704" />
</head>
<body>
foo bar
</body>
</html>

<html>
<head>
<meta name="date" content="20050704" />
</head>
<body>
foo bar
</body>
</html>

karpet@cartermac 44% cat date_c
MetaNames date
PropertyNames date

karpet@cartermac 45% swish-e -c date_c -i date*html -v 3
..

Checking file "date1.html"...
   date1.html - Using DEFAULT (HTML2) parser -  (3 words)
Indexing "date2.html"

Checking file "date2.html"...
   date2.html - Using DEFAULT (HTML2) parser -  (3 words)

...
karpet@cartermac 46% swishdebug foo
Found 2 hits
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
date                »20050704«
swishdbfile         »index.swish-e«
swishdocpath        »date2.html«
swishdocsize        »93«
swishfilenum        »2«
swishlastmodified   »1131250346«
swishrank           »1000«
swishreccount       »1«
swishtitle          »«
fuzzy: foo
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
date                »17760704«
swishdbfile         »index.swish-e«
swishdocpath        »date1.html«
swishdocsize        »93«
swishfilenum        »1«
swishlastmodified   »1131250336«
swishrank           »1000«
swishreccount       »2«
swishtitle          »«

karpet@cartermac 47% swishdebug -p date -h '17770101' -l '17760101' foo
Found 1 hits
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
date                »17760704«
swishdbfile         »index.swish-e«
swishdocpath        »date1.html«
swishdocsize        »93«
swishfilenum        »1«
swishlastmodified   »1131250336«
swishrank           »1000«
swishreccount       »1«
swishtitle          »«


William M Conlon scribbled on 11/5/05 5:53 PM:

> Here's a possible wish list item that I don't yet see on the list of  
> future enhancements:  Collation (ordering) of meta tags so one could  
> limit a search by
> 
> metaname < quantity
> 
> in addition to string comparison (metaname = word)
> 
> For example, dates stored as metatags (yyyymmdd):
> 
> swish-e -w "date < 17760704"
> 
> Another is a collection I have where individual documents have access  
> attributes (say ordered 0,1,2,...), where higher means more access.   
> I  spider into a common index, with each @url entry starting at a  
> page that lists all documents (to which  swishe has access),  
> recording the access level in the access metatag.  Although access to  
> the document itself is separately restricted by an application, I  
> want to avoid information leakage in the description property, so:
> 
> swish-e -w "access = (0 or 1 or 2)"  for someone with access level 2
> 
> But with a collation order, I could also use swish-e -w "access <  
> $user_access_level"
> 
> Admittedly this is much less compelling than date sequences.
> 
> 
> 
> Bill

-- 
Peter Karman  .  http://peknet.com/  .  peter(at)not-real.peknet.com
Received on Sat Nov 5 20:22:21 2005