#!/usr/bin/perl

#------------------------------------------------------------------
# getrejectsbypeer v1.1               elmi@newsfeeds.de 2000 02 10
#
# Call as a filter with
# getrejectsbypeer [-q] [-i <pattern>] <limit>
#
# where <limit> limits the number of groups listed per peer
# ("top <limit> rejected groups").
#
# -q sets quiet (non-debug) mode
# -i <pattern> sets a to-be-ignored groupname pattern
#
# Input on STDIN:   news log file
#
# Feb  3 01:00:57.028 - news-in.ivm.net <389dbf76.22284799@news.pacbell.net> 437 Unwanted newsgroup "us.jobs"
# Feb  3 01:00:59.391 - news-in.ivm.net <jw3m4.1279$uc.22301@nnrp4.clara.net> 437 Duplicate
# Feb  3 01:00:57.560 + news-in.ivm.net <87ag7l$1r$1@nslave2.tin.it> 2200 INFLOW news
#
# Output on STDOUT:
#
#        +------------------------------------------------------+
#        |                                                      |
#        |Rejected <count> articles from <peer> (top <limit>)   |
#        |  <group>     <count>                                 |
#        |  <group>     <count>                                 |
#        |  ...                                                 |
#        |  <group>     <count>                                 |
#        |                                                      |
#        |Rejected <count> articles from <peer> (top <limit>)   |
#        |  <group>     <count>                                 |
#        |  ...                                                 |
#        |                                                      |
#        +------------------------------------------------------+
#
#------------------------------------------------------------------


%peergroup	= ();	# list of "unwanted newsgroups" per peer as "$peergroup{peer:group}=count"
%peerduplicate	= ();
%peersum	= ();

$debug = 1;
$|=1;
$ignore = "";
$limit  = 99999999;

while ($#ARGV >= 0)
  {
  $opt = shift @ARGV;
  if ($opt =~ /^-/)
    {
    if ($opt eq "-q")
      { $debug=0; }
    elsif ($opt eq "-i")
      { $ignore=shift @ARGV; }
    }
  else
    { $limit = $opt; }
  }

print "Limit: $limit\n";
print "Debug: $debug\n";
print "Ignore: $ignore\n";


$i=0;
while (<STDIN>)
  {
  print STDERR "$i\r" if ((($i % 100)==0) && ($debug));
  $i++;
  $line = $_; chomp($line);

  if ($line =~ /\s*[a-zA-Z]+\s+[0-9]+\s+[.:0-9]+\s+-\s+(\S+)\s+<\S+>\s+([0-9]+)\s+([^"]+)\s+"([^"]+)"/)
    {
    $peer   = $1;
    $errnum = $2;
    $reason = $3;
    $group  = $4;
    
    $peersum{$peer}=0 if (!defined $peersum{$peer});
    $peersum{$peer} += 1;
    $ignored{$peer} = 0 if (!defined $ignored{$peer});


    if (($ignore ne "") && ($group =~ /$ignore/i))
      { $ignored{$peer} += 1; }
    else
      {
      $key = "$peer:$group";

      if ($reason =~ /unwanted newsgroup/i)
        {
        if (!defined $peergroup{$key}) { $peergroup{$key} = 1; }
        else                           { $peergroup{$key} += 1; }
        }
      elsif ($reason =~ /duplicate/i)
        {
        if (!defined $peergroup{$key}) { $peergroup{$key} = 1; }
        else                           { $peergroup{$key} += 1; }
        }
      }
    }
  }

print STDERR "$i\r" if ($debug);


# resort that stuff now...

@result = ();

foreach $key (keys %peergroup)
  {
  my ($peer,$group) = split(":",$key);
  my $num = $peergroup{$key};
  my $s = sprintf("%s %05d %s %s",$peer,$num,$group,$num);
  push @result,$s;
  }


$currentpeer = "";
$counter=0;
foreach $line (reverse sort @result)
  {
  my ($peer,$snum,$group,$num) = split(" ",$line);
  if ($peer !~ /^$currentpeer$/)
    {
    print "\n";
    print "Rejected $peersum{$peer} (ignored $ignored{$peer}) articles from $peer (top $limit):\n";
    $currentpeer = $peer;
    $counter=0;
    }

  print sprintf("  %5d %s\n",$num,$group) if ($counter < $limit);
  $counter++;
  }