Document Properties

The Concept

Version 1.3 of SWISH-E now allows you to specify certain META tags that can be used as document properties. The contents of any META tag that has been identified as a document property will be returned as part of the search results along with the rank, file name, title, and document size.

Usage Changes

• The new "PropertyNames" tag in the configuration file lists all META tag names that are to be retrievable at search time. Note: the index must be regenerated for this to have any effect.

  This example defines three document properties:

   	PropertyNames description author datamodified
  

  This is a META tag example that sets the value of the "author" property to "Jefferson" and the "description" as shown:

  <meta name="author" content="Jefferson"> <meta name="description" content="Blueprints"> <meta name="datamodified" content="Jul-4-1776">

• A new command-line option, "-p", is used to specify which document properties are to be included in the search results output. Follow the "-p" option with property (meta) names.

  This example will return two document properties:

  swish-e -w Independence -p description author

• The search results output has these features:
  • the 'comment' section that starts the search results (the lines that start with a '#') will include one line for each field that was specified, like this:

    # DocProperty 1: description # DocProperty 2: author

  • each line of the search results includes the rank, file name, title, and file size (the normal output) plus one quoted field for each property that was specified with the -p option. If the document does not have one of the properties then the output for that value will be "" (two double quotes, with nothing between them).

    This example of search results uses the previously shown META tags:

    1000 http://www.usa.gov/constitution.html "Title" 10232 "Blueprints" "Jefferson"

• The index file size will increase as given by this formula:

    IncreaseInSize (bytes) = 
  				(2 * <NumberOfFiles>) + 
  				(5 * <NumberOfPropertyMetaTagsInAllFiles>) +
  				(<LengthOfAllPropertyValuesInAllFiles>)
  

SwishTest.pl CGI Example

Accessing the document properties from the SwishTest.pl cgi example program is relatively easy. Here are the changes that need to be made to support the running example:

• There is a line that reads:

  open(SWISH, "$swish -w $query -m $results -f $index|");

  Change it to read:

  open(SWISH, "$swish -w $query -m $results -f $index -p description author |");

  You are adding each of your document properties after the "-p" option

• There is a line that reads:

  ($stringone, $title, $filesize) = split(/\"/, $_);

  Change it to read:

  ($stringone, $title, $filesize, $description, $junk, $author) = split(/\"/, $_);

  Note that the $junk variable will end up containing a single space. You have to put one of those between each document property.

• There is a line that reads:

  print "<LI><A HREF=\"$url\">$title</A><BR>";

  Change it to read:

  print "<LI><A HREF=\"$url\">$title</A> - $description (<I>$author</I>)<BR>";

  Table of Contents

  

  Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000 Hewlett-Packard Company
  Originally by Kevin Hughes, kev@kevcom.com, March 11, 1994.
  SWISH-E is distributed with no warranty under the terms of the GNU Public License,
  Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
  Public questions may be posted to the SWISH-E Discussion.
  Document maintained at http://sunsite.berkeley.edu/SWISH-E/Manual/UsageOverview.html by the SunSITE Manager.
  Last update December 16, 1998. SunSITE Manager: manager@sunsite.berkeley.edu