M# blosxom plugin: adsense
# raffi krikorian <r@bitwaste.com>
# version: 0.0.2
# see the bottom of the source for documentation

package adsense;
use CGI qw/:standard :netscape/;
use File::stat;
use FileHandle;
use LWP::UserAgent;

# --- configurable variables -----

# the identifier that is included as google_ad_client in the
# javascript that is given to you by the adsense website
my $ad_client = "pub-1258118923120631";

# --------------------------------

$adsenseurl;
$adsenseurlescaped;
$adsensetitle;
$adsensedesc;

sub start {
    return 0 if( $blosxom::flavour ne "rss" );
    1;

}

# this is the helper function that actually gets the advertisment from
# google, parses the information, and assigns the variables to be used
# by the blosxom templates
sub fetchad {
    my( $path ) = @_;
    my $useragent = LWP::UserAgent->new();

    # construct the URL to get the advertisement from the google site
    # with.  this has to get the referrer correct, the time of day for
    # the request, and the asks for only a single advertisement.  we
    # use the first story as the referrer so we can get (hopefully)
    # better targeted advertisements.
    my $refererurl = $path;
    $refererurl =~ s|^$blosxom::datadir(/.*?)\.[^\.]*?$|$blosxom::url$1\.html|;
    $refererurl =~ s|\:|\%3A|g;
    $refererurl =~ s|/|\%2F|g;

    my $adurl = "http://pagead2.googlesyndication.com/pagead/ads?client=ca-";
    $adurl .= $ad_client;
    $adurl .= "&random=" . ( time() * 1000 );
    $adurl .= "&format=125x125_as";
    $adurl .= "&output=html";
    $adurl .= "&url=" . $refererurl;

    # make the request to fetch the html containing the advertisement
    my $request = HTTP::Request->new( GET => $adurl );
    $request->referer( $blosxom::url );
    $request->user_agent( "Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en) AppleWebKit/124 (KHTML, like Gecko) Safari/125" );
    $response = $useragent->request( $request )->content();

    # and if that worked, pull out the relavent information
    if( $response ) {
	$response =~ m|<font[^>]*?>\s*<a.*?href=\"(.*?)\"[^>]*>\s*<b>(.*?)</b></a><br></font><font[^>]*?>(.*?)<br>|;
	
	$googleurl = "http://pagead2.googlesyndication.com$1";
	$googletitle = $2;
	$googledesc = $3;

	$googleurlescaped = $googleurl;
	$googleurlescaped =~ s|\&|&amp;|g;
	return( $googleurl, $googleurlescaped, $googletitle, $googledesc );

    } else {
	return( );

    }

}

sub filter {
    my( $pkg, $files_ref ) = @_;
    
    # get a list of all the files that are being rendered, and pick
    # out the timestamp of the most recent story
    my $fh = new FileHandle;
    my @files = 
	sort { $files_ref->{$b} <=> $files_ref->{$a} } keys %$files_ref;
    my $recenttime = $files_ref->{$files[0]};

    # now check to see if we have a state file -- we check the
    # timestamp on that file and if the timestamp is -earlier- than
    # the recenttime variable, then we are going to replace that file
    # with a new advertisement.  if it is less than or equal to, then
    # we are going to open the file and pull the google ad out of it.
    #
    # we don't want to pull a new advertisement every time as google
    # changes the advertisement URLs frequently, even if they are
    # pointing to the same destination web site.  this does lead to
    # the problem that the advertisement URLs may get stale.
    my $statefile = $blosxom::plugin_state_dir . "/adsense.state";
    if( ( ! -e $statefile ) ||
	( stat( $statefile )->mtime < $recenttime ) ) {
	my @ad = fetchad( $files[0] );

	# store the advertisement into this state file for the next
	# time around
	unlink $statefile;
	open ADSTATE, ">$statefile";
	print ADSTATE $ad[0] . "\n" . $ad[1] . "\n" . $ad[2] . "\n" . $ad[3] . "\n";
	close ADSTATE;
	
	$adsenseurl = $ad[0];
	$adsenseurlescaped = $ad[1];
	$adsensetitle = $ad[2];
	$adsensedesc = $ad[3];

    } else {
	# we don't need a new advertisement, so we are just going to
	# read the one from the state file
	open ADSTATE, "<$statefile";
	$adsenseurl = <ADSTATE>;
	$adsenseurlescaped = <ADSTATE>;
	$adsensetitle = <ADSTATE>;
	$adsensedesc = <ADSTATE>;
	close ADSTATE;

    }

    1;

}

1;

__END__

=head1 NAME

blosxom plug-in: adsense

=head1 SYNOPSIS

fills the variables $adsense::adsenseurl, $adsense::adsenseurlescaped,
$adsense::adsensetitle, $adsense::adsensedesc with the google
advertisement URL (in non-escaped and escaped form), the title of the
advertisement, and a one line description of the advertisement for
inclusion in the RSS of the weblog.

=head1 USAGE

activate this plugin by placing it in your bloxsom plugin directory,
then in your foot.rss (or something similar), add the following

<item>    
<title>Google Ad: $adsense::adsensetitle</title>
<link>$adsense::adsenseurlescaped</link>
<description>$adsense::adsensedesc</description>
</item>

</channel>
</rss>

=head1 VERSION

0.0.1

=head1 AUTHOR

raffi krikorian <r@bitwaste.com>, http://www.bitwaste.com/

=head1 LICENSE

adsense plugin
copyright 2004, raffi krikorian

this software is distributed under the GPL
(http://www.gnu.org/copyleft/gpl.html) with the following addition --
if you do use this plugin, then you are required to publish full
stories in your RSS and -not- use excerpts.  the only exception to
that rule is if the story is "long" (as defined by more than a 250
words) and requires a link to get to the remainder of the content.