NAME

WWW::Search::Dejanews - class for searching discussions at www.deja.com

SYNOPSIS

use WWW::Search;
my $oSearch = new WWW::Search('Dejanews');
my $sQuery = WWW::Search::escape_query("sushi restaurant Columbus Ohio",);
$oSearch->native_query($sQuery,
                       {'defaultOp' => 'AND'});
while (my $oResult = $oSearch->next_result())
  { print $oResult->url, "\n"; }

DESCRIPTION

This class is a Dejanews specialization of WWW::Search. It handles making and interpreting searches against Deja discussions archive http://www.deja.com/usenet.

This class exports no public interface; all interaction should be done through WWW::Search objects.

Dejanews DOES support wildcards (asterisk at end of word).

The default behavior is the OR of the query terms. If you want AND, insert 'AND' between all the query terms in your query string:

$oSearch->native_query(escape_query('Dorothy AND Toto AND Oz'));

or, call native_query like this:

$oSearch->native_query(escape_query('Dorothy Toto Oz'), {'defaultOp' => 'AND'} );

The results are sorted by www.deja.com's "confidence" score.

The URLs returned point to "text only" articles from Dejanews' server.

If you want to search particular fields, add the escaped value for each field to the second argument to native_query:

my $sFromDate = WWW::Search::escape_query('Jan  1 1999');
my $sToDate   = WWW::Search::escape_query('Jan 31 1999');
$oSearch->native_query($sQuery, 
                       {'groups'   => 'rec.juggling',
                        'subjects' => 'learning+five',
                        'fromdate' => $sFromDate
                        'todate'   => $sToDate, } );

NOTES

In the SearchResults, the description field contains the forum name and author's name (as reported by www.deja.com) in the following format: "Newsgroup: comp.lang.perl.modules; Author: Martin Thurn"

CAVEATS

BUGS

Please tell the author if you find any!

TESTING

This module adheres to the WWW::Search test suite mechanism. See $TEST_CASES below.

AUTHOR

WWW::Search::Dejanews is maintained by Martin Thurn (MartinThurn@iname.com); original version for WWW::Search by Cesare Feroldi de Rosa (C.Feroldi@it.net).

LEGALESE

THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

VERSION HISTORY

2.14, 2000-07-17

make the UserAgent non-robotic

2.13, 2000-06-23

titles, forums, and authors are no longer truncated

2.12, 2000-05-22

short-circuit if explicitly no results returned

2.11, 2000-03-20

What a difference a space makes! www.deja.com mucked with spaces in the output.

2.10, 2000-03-09

Fixed for new output format; updated test cases.

2.08, 2000-02-24

fix for date-range-, subject-, groups-limited searches.

2.07, 2000-01-18

Handle www.deja.com's new output format.

2.06, 1999-12-07

New test cases, pod update, ignore deja links, etc.

2.04, 1999-12-06

Handle www.deja.com's new output format.

2.03, 1999-10-05

Now uses hash_to_cgi_string().

2.02, 1999-09-17

BUGFIX: was returning "power search" link (thanks to Jim Smyser for noticing)

2.01, 1999-07-13

1.12, 1999-07-06

Finally moved from www.dejanews.com to www.deja.com; New test suite mechanism;

1.11, 1998-12-03

Now uses the split_lines() function; sync with WWW::Search distribution's version number

1.4, 1998-08-27

New Dejanews.com output format

1.3, 1998-08-20