news.pl

Sam Ruby
It's just data

Home

Mark Baker

Keith Ballinger

Cory Doctorow

Miguel de Icaza

Paul Prescod

Clay Shirky

Dave Winer

Axis

BlogDex

DayPop

Gump

Interop

O'Reilly

Slashdot

SOAP-ENV

SOAP::Lite

SoapBuilders

SoapDir

WSDL

WSTK

news.pl

#!/usr/bin/perl
use HTTP::Date;
use POSIX "strftime";

print "Content-type: text/html\r\n\r\n";

chdir "/tmp";

# refresh from source if the version on disk is stale
my $url = "http://www.google.com/news";
my $file = "newsheadlines.html";
@stat=stat $file;
if ((time() - $stat[9]) > 900) { # 15 minute cache
unlink $file;
system "wget -q -O $file $url";
system "touch $file";
}

# read into memory
open FILE, $file;
$_ = join "", <FILE>;
close FILE;
s/n/ /gm;
$start=0;

# output html header
print "<html><head>\n";
print '<META http-equiv="refresh" content="3600">';
print "<title>Google Headline News</title>\n";
print '<style type="text/css" media="screen">';
print 'a.banner { text-decoration:none; color:#FFFFFF; font-family: futura, helvetica, arial, sans-serif;}';
print 'a { text-decoration:none; color:#0000CC; }';
print 'a:hover { text-decoration:underline; color:#FF0000; }';
print "</style>\n";
print "</head>\n";

# output google search form
print "<table>\n";
print "<tr><td colspan=4 align=right>\n";
print '<FORM method=GET action="http://www.google.com/search">';
print '<A HREF="http://www.google.com/">';
print '<IMG SRC="http://www.google.com/logos/Logo_40wht.gif" border="0" ';
print 'ALT="Google" align="absmiddle"></A>';
print '<INPUT TYPE=text name=q size=25 maxlength=255 value="">';
print '<INPUT type=submit name=btnG VALUE="Google Search">';
print '</FORM>';
print "</td></tr>\n";

# output personal links
print "<tr valign=\"top"><td>n";
print "<br><a href=\"http://www.cnn.com/">cnn</a>n";
print "<br><a href=\"http://www.msnbc.com/">msnbc</a>";
print "<br><a href=\"http://newsobserver.com/">n&o</a>n";
print "<br><a href=\"http://www.nbc17.com/">nbc17</a>";
print "<br><a href=\"http://www.usatoday.com/">usatoday</a>";
print "<br><a href=\"http://www.wral.com/">wral</a>n";
print "<br><a href=\"http://abclocal.go.com/wtvd/">wtvd</a>n";
print "<br><hr>\n";
print "<a href=\"http://movies.triangle.com/Cary-Crossroads.html">movies</a>n";
print "</td><td>\n";
print "<table width=\"100%" cellpadding="2" cellspacing="0">n";

my @sections = split '<font color="#FFFFFF"';
shift @sections;
foreach (@sections){

# output section banner
if (/^.*?>(.*?)</font>/) {
$title=$1;
print "</table></td><td><table>" if ($title =~ /Business/);

    if ($title =~ /(.* - Updated) (w*) (d*) *(d*):(d*)/) {
      $year = strftime "%Y", localtime;
      $time = str2time("$3-$2-$year $4:$5") + 3*60*60;
      $title = "$1 " . strftime("%b %d %I:%M %p", localtime($time));
      $title =~ s/ 0/ /g;
    }

    print "<tr bgcolor=\"#336699"><td>" .
      "<a class=\"banner" href="$url/$file#$anchor">$title</a>" .
      "</td></tr>\n" unless ($title =~ /Top Headlines/);
    $anchor = $1 if /<a name=(w*)>/;
}

# output headlines
my @links = split '<a class=g';
shift @links;
print "<tr><td>\n";
foreach (@links) {
    if (/^(.*?)</a>/) {
      print "<li><A $1</A>\n" unless $1 =~/href=#/;
    }
}
print "</td></tr>\n";
}

print "</table>\n";
print "</td></tr></table>\n";
print '</html>';