#!c:/perl/bin/perl

use strict;
use LWP::UserAgent;
require LWP;
use URI;
use LWP::RobotUA;
require LWP::RobotUA;
use Globals;
use Globals qw(%ALLPLATTERS);
use Globals qw($MINTRAWL);
use Globals qw($delim);
use Globals qw(@MUSIC);
use Globals qw(@VIDEO);
use Globals qw(@SHORTCUT);
use Globals qw($debug);
use Globals qw($bgcolour);

use CGI qw(:standard);
$|=1;

my $QUERY     = CGI::new();
my $NAMELIST    = $QUERY->param( "hosts" );
my $NAME;
my $HOST;
my $BASE;
my %ABCS;
my $HEADER = new HTTP::Headers( 'Accept' => 'text/plain', 'User-Agent' => 'PlatterTwaartter/3.0.0', 'Dixon' => 'Am-Hammy-Twaart' );
my $ROBOT_NAME = "PlatterTwaartter";
my $ROBOT_OWNER = "chris.dixon\@hammytwaart.com";
my $USERAGENT;
my $RECURSE=0; # Stop weird parent links setting us up the bomb
my $LINECOUNT = 0;
my $QUICKEXT = $delim.join($delim, (@MUSIC, @VIDEO, @SHORTCUT)).$delim;
# Print the header
#
my $TEMPLATE = openTemplate ("trawl");
$TEMPLATE =~ s/%MINTRAWL%/$MINTRAWL/g;
$TEMPLATE =~ s/%BGCOLOUR%/$bgcolour/g;
print $TEMPLATE;

if ( defined $NAMELIST ) {

  # Lock the inventory against meddling by pesky kids
  if (-e "../inventory/trawl.lock") {
    print '<p>The Distroplat inventory is currently locked by another user. Please <a href="javascript:location.reload()">try again</a> in a bit. Bye.</p>';
    exit(0);
    }

  open (LOCK, ">../inventory/trawl.lock");
  close(LOCK);

  # Load the ABCs

  %ABCS = getABCs();

  #Set up the user agent

  $USERAGENT = new LWP::RobotUA $ROBOT_NAME, $ROBOT_OWNER;
  # go as fast as you can
  $USERAGENT->delay(0);
  $USERAGENT->timeout(15); # seconds - doesn't seem to work though

  # Deal with each server in turn
  my @SPLITNAME = split(',', $NAMELIST);

  print "<h2>Attempting to trawl ".scalar(@SPLITNAME)." servers</h2>\n";

  foreach $NAME (@SPLITNAME) {

    print "<h4>Trawling $NAME  . . </h4>";
    open (INVENTORY, ">../inventory/$NAME.inv");
    open (ARTISTS, ">../inventory/$NAME.artist");
    # Find the host
    my $found=0;
    my @RESULT;
    foreach $HOST (keys %ALLPLATTERS) {
      if ($HOST eq $NAME) {
        $found = 1;

        $BASE = $ALLPLATTERS{$HOST};

        print "Trawling ".$NAME."'s server @ http://$BASE\n";
        print INVENTORY "http://$BASE\n";
        my $TIME = time;
        @RESULT = parse_page( 'http://'.$BASE.'/', $NAME);

        my $INVENTORY = shift (@RESULT);
        if ( length $INVENTORY > 0 ) {
          print INVENTORY $INVENTORY;
          my $count = shift @RESULT;

          # remove duplicate artists
          @RESULT = sort @RESULT;
          my $i=0; my $last;
          while ($i<=$#RESULT) {
              if ($last eq $RESULT[$i]) {
                  splice @RESULT, $i, 1;
              }
              else {
                  $last = $RESULT[$i];
                  $i++;
              }
          }

          print ARTISTS join("\n", @RESULT)."\n";
          print "<p id=\"$NAME\">Indexed $count tracks in ".(time-$TIME)." seconds.<script>document.getElementById(\"$NAME\").scrollIntoView(false);</script>";

        }
      }
    }
    close INVENTORY;
    close ARTISTS;
    unless ($found) {
      print "<br>Host $NAME is not registered platter host.";
    }
  }

# Re-write the abcs file

open ( ABC, ">../inventory/abcs.txt" );
my @abckeys = sort {uc($a) cmp uc($b)} keys %ABCS;
my $key;
foreach $key (@abckeys) {
    print ABC $key.$delim.$ABCS{$key}."\n";
    }
close ABC;

print "<h4 id=\"end\">All done</h4><script>document.getElementById(\"end\").scrollIntoView(false);</script></body></html>";
unlink "../inventory/trawl.lock";
}
else
{
  # Generate form
  #
  print q{ <form name="twatsource">
          <table border=0 cellpadding=1 cellspacing=0>
            <tr>
              <td>&nbsp;</td>
              <td><b>User</b></td>
              <td><b>Last Trawled</b></td>
            </tr>
            <tr><td colspan=3><hr size=1 color=white></td></tr>
          };

  my %TIMES;

  foreach $HOST (keys %ALLPLATTERS)
  {
    my $HOSTFILE = '../inventory/'.$HOST.".inv";
    my $thistime = 0;

    if (-e $HOSTFILE)
    {
      my @stats = stat(_);
      $thistime = $stats[9];
    }

    $TIMES{ $thistime.";".$HOST } .= $ALLPLATTERS{$HOST}."<br>";
  }

  my $TIME;
  my $LASTMOD;

  foreach $TIME (sort {$b <=> $a} keys %TIMES)
  {
    my ($THISTIME, $HOST) = split (/;/, $TIME);
    my $URL = $TIMES{$TIME};
    my $TRAWLABLE = 1;

    if ($TIME > 0)
    {
      $LASTMOD = localtime($TIME)." CET";
      if (time - $TIME < (60*$MINTRAWL))
      {
        $TRAWLABLE = 0;
      }
    }
    else
    {
      $LASTMOD = "Not trawled yet";
    }

    print "<tr><td>";

    if ($TRAWLABLE)
    {
      print "<input type=\"checkbox\" name=\"".$HOST."\">";
    }
    else
    {
      print "&nbsp;";
    }
  print "</td>";
  print "<td nowrap valign=bottom><b>".$HOST."</b></td>";
  print "<td nowrap valign=bottom>&nbsp;<small>".$LASTMOD."</small></td></tr>\n";

}

print qq{  </tr>
           <tr><td colspan=3><hr size=1 color=white></td></tr></table>
         </form>
         <form><input type="button" onClick="twat()" value="Trawl"></form>
      </body>
      </html>
      };

}
exit 0;



# sub to recursively trawl http directories for A tags
# should store all the recognised files it can find.
#
# should be called with 1 parameter
#  - path
#
sub parse_page
{
     my($URL, $NAME) = @_;
     my $INVENTORY = "";
     my @ARTISTS = ();
     my $localcount = 0;
     $RECURSE++;

     if ($RECURSE < 3) {
        my $FOLDER;
        if ($RECURSE == 1) {
             $FOLDER = $NAME."'s platter";
         }
         else {
             $URL =~ /.*\/([^\/]+)\/*$/;
             $FOLDER = makeNice($1);
         }
         print "<br><span id=\"line".$LINECOUNT."\">Indexing <b>".$FOLDER.":</b></span><script>document.getElementById(\"line".$LINECOUNT++."\").scrollIntoView(false);</script>\n";
     }


     #request page
     my $REQUEST = HTTP::Request->new('GET', $URL , $HEADER);
     my $RESPONSE = $USERAGENT->request($REQUEST);
     #if response is successful
     if ($RESPONSE->is_success)
     {
          #parse page that is returned
      require HTML::Parse;
      my $HTML = HTML::Parse::parse_html($RESPONSE->content);
      my $base = $RESPONSE->base;
      for ( @{ $HTML->extract_links } )
      {
           my($LINK, $elem) = @$_;
           my $TAG = $elem->tag;

           my $LINK = URI->new($LINK)->abs($base)->as_string;
           #$LINK = makeUnNice($LINK);
           $LINK =~ s/%25/%/g;
           my $OUTPUT = $LINK;
           $OUTPUT =~ s/(.*?)$BASE//;
           my $EXT;
           if ($OUTPUT =~ /\.(\w*)$/) { $EXT = $1;}

           if ($TAG =~ /[Aa]/ && $URL !~ /\Q$LINK\E\/?.*/ && $LINK !~ /\?.*/)
           {
          		if ($debug) { print "<br>$LINK\n" }
                if (
                     $QUICKEXT =~ /$delim$EXT$delim/i
                   ||
                     $OUTPUT =~ /(artist|cover)[^\/]*\.(jpg|gif|png)$/i
                   ||
                     $OUTPUT =~ /.+\/info\.txt$/i
                   )
                {
                     #if the file is recognised then

                     #add link to inventory
                     my @entry = makeInvLine($OUTPUT);

                     $INVENTORY .= join($delim, @entry)."$delim$OUTPUT\n";
                     push (@ARTISTS, $entry[0]) if $entry[0];
                     push (@ARTISTS, $entry[1]) if $entry[1];

                     if ($QUICKEXT =~ /$delim$EXT$delim/i) {$localcount++; }
                }
                elsif ($OUTPUT =~ /.abc$/i)
                {
                     if ($OUTPUT =~ /([^\/]+)\/([^\/]+\.abc)$/)
                     {
                       my $PATH = makeNice($1); my $ARTIST = makeNice($2);
                       $ABCS{$PATH} = $ARTIST;
                     }
                }
                elsif ($OUTPUT =~ /^\/info\.txt$/i)
                {    # news page
                    my $NEWSREQUEST = HTTP::Request->new('GET', $LINK, $HEADER);
                    my $NEWSRESPONSE = $USERAGENT->request($NEWSREQUEST);
                    if ($NEWSRESPONSE->is_success)
                    {
                        my $lastmod = $NEWSRESPONSE->headers()->last_modified;
                        open(NEWS, ">../inventory/$NAME.news");
                        print NEWS $lastmod."\n";
                        print NEWS $NEWSRESPONSE->content;
                        close(NEWS);
                        print "<br>Fetched news page from $NAME";
                    }
                    else
                    {
                        print '<br><b  style="color: red">Code '.$RESPONSE->code()." error  fetching news page from $NAME: ".$RESPONSE->message().'</b>';
                    }
                }
                else
                {
                  #print "<br>$LINK\n";
                  if ($LINK =~ /\Q$URL\E/i # previous link is a substring of the new one
                  && !($LINK=~ /\.\w{3}$/)) # do not retrieve anything with a dot and three chars at the end
                  {
           if ($RECURSE < 5) # limit navigation to 5 folders deep - and that's taking the piss
             {
               my @CHILD = parse_page($LINK);
               $INVENTORY.= shift(@CHILD);
               $localcount += shift(@CHILD);
               push(@ARTISTS, @CHILD);
             }
           else
             {
               print "<h3>Recursion limit hit, will not trawl child folders.<br>".
               "Please check the format of your platter and contact Platter Support</h3>";
             }

                  }
                }
           }
          }
     }
     else {
        print '<br><b  style="color: red">Code '.$RESPONSE->code().' error indexing '.makeNice($URL).': '.$RESPONSE->message().'</b>';     }

     if ($RECURSE == 2) {
        print " $localcount tracks found.";
        }
     $RECURSE--;
     return ($INVENTORY, $localcount, @ARTISTS);
}

sub makeInvLine( @_ ) {
    my @items = split('/', $_[0]);
    shift @items; # remove leading slash

    # get artist folder
    my $folder;
    if ($#items > 0) {
        $folder = shift @items;
        $folder =~ s/^(The%20|)(.*?)(,%20The|)$/$2/i;
    }

    # get track

    my $track = pop(@items);
    
    # square brackets weren't seen for 6 years, but when they arrived they broke it
    $track =~ s/\[/%5B/g;
    $track =~ s/\]/%5D/g;


    # get track artist
    my $trackartist;
    if ($track =~ /^(\d{2}%20-%20|)(The%20|)([^\/]*?)(,%20The|)%20-%20/i
        && !($3 =~ /^\d{2}$/)
        && length $3 > 0) {
        $trackartist = $3;
        $track =~ s/(The%20|)$trackartist(,%20The|)%20\-%20(.*?)/$3/;
    }


    # get album
    my $album; my $year;
    if ($#items > -1) {
        $album = join('/', @items);

        # get year
        if ($album =~ /%20(\(|%28)(\d{4})(\)|%29)(\/|$)/) {
            $year = $2;
        }
    }

    return (
        makeNice($folder),
        makeNice($trackartist),
        makeNice($album),
        $year,
        makeNice($track));

}