#!/usr/bin/perl 

# Filters an RSS feed for interesting content:
# search words are read from a file and in addition,
# all names of applications stored in the /Applications 
# directory (on Mac OS X) can be used as keywords.
#
# This allows you to filter the RSS feeds of software version
# update sites that would not otherwise allow you to filter
# the output.
#
# Run this on your local web server or as cron job
# Use on public web servers is discouraged. Security risks!

# How to install this? Instructions here:
# http://www.davids-world.com/archives/2005/05/filtering_macup.html

# (C) 2005 David Reitter, http://www.reitter-it-media.de/
# Released under the GNU General Public License.


# set to 1 if you'd like to run this as CGI,
#  0 otherwise.
my $runascgi = 1;   

# this is the path to store the temporary file in
# needs to be writeable by this script (i.e. by web server if CGI is used!)
my $rsstempfilepath = "/home/".$ENV{'USER'}."";   # for general Unix systems
# my $rsstempfilepath = "/Users/".$ENV{'USER'}."/Temp";   # for the author's system
# my $rsstempfilepath = "/Users/".$ENV{'USER'}."/Library/Application Support";   # for OS X systems

# set this to 1 if you'd like to include the contents of the /Applications directory (Mac OS X)
# as keywords
my $appsearch = 1;

# the file containing the keywords to look for (line by line)
my $keywordsfile = "";

# NO CHANGES BEYOND THIS LINE NECESSARY
# ---------------------------------------------------------------------------------------------------

if ($ARGV[0]) {
  $runascgi = 0;
}

# this little CGI browses 

use strict;
my $sourcerss = "";
my $rssfile = "";

if ($runascgi) {

  use CGI qw/:standard/;

  # if running as CGI, set this manually in the script
 
  if (param()) {
    $sourcerss = param('source');
  }

  unless ($sourcerss =~ /^http:\/\//) {
    print "Content-Type: text/plain\n\n";
    print "You must give this script an URL as argument 'source' that begins with http://. Usage example:  /cgi-bin/filter-rss.cgi?source=http://mysite.com/index.rss\n\n";
    exit;
  }

  $rssfile = "filter-rss-".&hashCode($sourcerss).".temp.rss";

} else {
  
  $sourcerss = $ARGV[0];	# give source rss
  $keywordsfile = $ARGV[1]; # file with keywords (line-by-line),  optional

  if ($sourcerss eq '') {

    print "Filter-RSS usage:\nfilter-rss sourceURL [keywordsfile]\n";
    print "Ore use as CGI with argument source=sourceURL";
    exit;
  }
  $rssfile =  $rsstempfilepath."/filter-rss-".&hashCode($sourcerss).".temp.rss";


}

# this is used for temporary storage
  

# no changes beyond this line
 
my @keywords = ();
if (open K, "<$keywordsfile") {
  while (<K>) {
    chop $_;
    push @keywords, $_;
  } 
  close K;
}

if ($appsearch) {
  open K, "ls /Applications |" || die "couldnt open keyword file";
  while (<K>) {
    chop $_;
    s/\.app//;	
    s/\s-\s.*$//;		# chop off version number
    s/([0-9]+\.)*[0-9]+//g;	# chop off version number
    s/^\s//g;   s/\s$//g;
    push @keywords, $_;
  } 
  close K;
}

my $prevrss = '';
if ( open P, $rssfile ) {
  warn "reading old file $rssfile";
  while (<P>) { 
    $prevrss .= $_; 
  } 
  close P;
} 
open C, "curl $sourcerss 2>/dev/null |";
my $file = join('', <C>);
close C;  

unless ($prevrss) {

  my $header;
  my $footer;
  if ($file =~ /^(.*?)<item>/is) {
    $header = $1;
  }

  if ($file =~ /(<\/channel>.*?)$/is) {
    $footer = $1;
  }

  $prevrss = $header.$footer;


  unless ($prevrss) {

    $prevrss = '<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE rss PUBLIC "-//Netscape Communications//DTD RSS 0.91//EN" "http://my.netscape.com/publish/formats/rss-0.91.dtd">
<rss version="0.91"><channel></channel></rss>';

  }
}

 
while ($file =~ /(<item>.*?<\/item>)/isg) {
  my $item = $1;
  if ($item =~ /<title>(.*?)<\/title>/is) {

    my $title = $1;

    my $desc = '';
    if ($item =~ /<description>(.*?)<\/description>/is) {

      $desc = $1;
    } 
    

    #print $title;
    foreach my $k (@keywords) {
      my $kq = quotemeta($k);
      if ($title =~ /\b$kq\b/s || $desc =~ /\b$kq\b/s ) {
	warn "found: $title $k";
	my $name = $title;
	$name =~ s/\s-\s.*$//;		  # chop off version number
	$name =~  s/([0-9]+\.)*[0-9]+//g; # chop off version number
 	$name =~ s/^\s//g;   s/\s$//g;
	my $nameq = quotemeta($name);
	$prevrss =~ s/<item>\s*<title>$nameq\b.*?<\/item>//is; 

	if ($prevrss =~ /<item>/) {
	  

	  $prevrss =~ s/<item>/$item\n<item>/is;


	} else {
	  
	  $prevrss =~ s/<channel>/<channel>$item\n/is;

	}
 
	last;
      }
    }

  }

}
 
# store result
if (open RSS, ">".$rssfile)	# || die("cannot write to $rssfile");
  {
    print RSS $prevrss;
    close RSS;
    warn "written to $rssfile";
  }

if ($runascgi) {

  print "Content-Type: text/xml\n\n";
}
print $prevrss;
  


sub hashCode {
  my ($text) = @_; my $length = length($text);
  my @characters = unpack("C*", $text);
  my $char;
  my $code = 0;
  my $i = 1; 
  foreach $char (@characters) {
    $code += $char * 31^($length - $i);
    $i++;
  }
  return $code;
}