#!/usr/bin/perl
=head1 NAME
#!/usr/bin/perl

graph-top-referers - produce a PNG graph of the top 10 referers in access_logs

=head1 SYNOPSIS

[todo]

=head1 DESCRIPTION

[todo]

=head1 AUTHOR

Justin Mason, jm at jmason dot org

=cut

###########################################################################

use File::Find;
use URI::Escape;
use Time::Local;
use POSIX qw(strftime);
use strict;
use warnings;

# my $EXCLUDE_RE = qr{^/(?:[^0-9]|$)}i; my $REQUIRE_RE = qr{\.html$}i;
my $EXCLUDE_RE = qr{\.(?:gif|png|css|jpe?g)$}i; my $REQUIRE_RE = qr{.};

my $MAX_PAGES = 10;

my %mon2mm = (
  Jan => 0, Feb => 1, Mar =>  2, Apr =>  3,   # note 0 based
  May => 4, Jun => 5, Jul =>  6, Aug =>  7,
  Sep => 8, Oct => 9, Nov => 10, Dec => 11
);

foreach my $dir (@ARGV) {
  File::Find::find (\&wanted, $dir);
}

my %count = ();
my %whens = ();
my $t_earliest;
my $t_latest;

sub wanted {
  return unless (-f $_);
  if (/\.gz$/) {
    return unless open (IN, "gunzip -cd '$_' |");
  } else {
    return unless open (IN, "<$_");
  }

  while (<IN>) {

# 71.235.13.232 - - [01/Apr/2006:00:01:10 +0100] "GET
#   /wp-content/themes/jmason/images/bg.gif HTTP/1.1" 200 992
#   "http://taint.org/2004/11/" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT
#   5.1; .NET CLR 1.1.4322)"

    /^\S+ \S+ \S+ \[([^\]]+?)\] "\S+ ([^"]+?) \S+" (\d+) \S+ "([^"]+?)"/
          or next;

    my $when = $1;
    my $tgt =  $2;
    my $code = $3;
    my $refr = $4;

    next if ($code !~ /^2/);        # ignore error codes please
    next unless ($refr =~ /^http/);

    next if ($tgt =~ m{$EXCLUDE_RE});
    next unless ($tgt =~ m{$REQUIRE_RE});

    $when =~ /^(\d\d)\/(...)\/(\d\d\d\d):(\d\d):(\d\d):(\d\d) /;
    $when = timegm($6,$5,$4,$1,$mon2mm{$2},$3) || 0;

    if (!defined $t_earliest || $when < $t_earliest) {
      $t_earliest = $when;
    }
    if (!defined $t_latest || $when > $t_latest) {
      $t_latest = $when;
    }

    $count{$tgt}++;
    push (@{$whens{$tgt}}, $when);
  }
  close IN;
}

my $bucketsize = ($t_latest - $t_earliest) / 100;
my $idx = 0;
my @legend = ();

open OUT, ">refgp.dat";
foreach my $tgt (sort {
          $count{$b} <=> $count{$a}
        } keys %count)
{
  $idx++;
  # warn "$count{$tgt}  $tgt\n";

  last if ($idx > $MAX_PAGES);

  my @hits = ();
  foreach my $hit (sort @{$whens{$tgt}}) {
    my $bkt = int (($hit - $t_earliest) / $bucketsize);
    $hits[$bkt]++;
  }

  my $bcount = 0;
  foreach my $bkt (@hits) {
    $bcount++;
    # next if !$bkt;
    $bkt ||= 0;
    my $t = int ($t_earliest + ($bucketsize * $bcount));
    print OUT "$t $bkt\n";
  }
  print OUT "\n\n";

  $legend[$idx] = "$tgt";
}
close OUT;

open GP, ">refgp.run";
print GP qq{

  set terminal png size 1024,768
  set timefmt "%s"
  set xdata time
  set format x "%04Y%02m%02d"

};

my @plot = ();
foreach $idx (0 .. ($idx-2)) {
  my $leg = $legend[$idx+1];
  push @plot, qq{ "refgp.dat" using 1:2 index $idx }.
        qq{ with linespoints lw 1 title '$leg' };
}

print GP "plot ".join(", ", @plot)."\n";
close GP;

print "[gnuplot -persist < refgp.run > refgp.png]\n";
system "gnuplot -persist < refgp.run > refgp.png";
exit;