#!/usr/bin/perl -T
# ms2pdf, Andrew Daviel, TRIUMF, 2008
#

$ENV{PATH} = '/bin:/usr/bin' ; # need wget, bash

use Time::Local ;
$sfx{HTML} = 'html' ; $sfx{Text} = 'txt' ; $sfx{PDF} = 'pdf' ; 
$type{html} = 'text/html' ; $type{txt} = 'text/plain' ; $type{pdf} = 'application/pdf' ;
$oopython = "/opt/openoffice.org2.4/program/python" ;

$method = $ENV{'REQUEST_METHOD'} ;
$path = $ENV{'PATH_INFO'} ;
$agent = $ENV{'HTTP_USER_AGENT'} ;
$content_type = $ENV{'CONTENT_TYPE'} ;
$content_length = $ENV{'CONTENT_LENGTH'} ;
$cookie = $ENV{'HTTP_COOKIE'} ;
$query = $ENV{'QUERY_STRING'} ;
$referer = $ENV{'HTTP_REFERER'};
$len = $ENV{'CONTENT_LENGTH'};
$script = $ENV{'SCRIPT_NAME'};
$raddr = $ENV{'REMOTE_ADDR'} ;

$now = localtime() ;
#$nowt = &wtime(time()) ;

$pragma = $ENV{'HTTP_PRAGMA'} ; $pragma =~ tr/A-Z/a-z/ ;

$imsd = $ENV{'HTTP_IF_MODIFIED_SINCE'} ; 
$ims = $imsd ; $ims =~ s/;.*// ; # for Netscape
$ims = &get_gmtime($ims) ;

open (LOG,">>logs/ms2pdf.log") ;

if ($method eq 'POST' and $len>0) {
  read(STDIN, $buffer, $len) ;
} elsif ($method eq 'GET') {
  $buffer = $query ;
} else {
  print<<EOT;
Status: 400 Bad Method
Content-type: text/html

Method $method not supported, or zero-length POST
EOT
  print LOG "$now $raddr Method $method not supported\n" ;
  exit ;
}
@pairs = split(/&/, $buffer);
foreach $pair (@pairs) {
  ($name, $value) = split(/=/, $pair);
  $value =~ tr/+/ /; $name =~ tr/+/ /;
  $value =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C", hex($1))/eg;
  $name =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C", hex($1))/eg;
  $F{$name} = $value ;
}

unless ($F{op}) {
  # print form and exit
  &head ;
  &foot ;
  print LOG "$now $raddr $method\n" ;
  exit ;
}
$otype = $sfx{$F{op}} ;
unless ($F{url} =~ /^(http.*)/i or $F{url} =~ /^(ftp:.*)/i) {
  &head ;
  print "<p>Bad URL \"$F{url}\" - must start with  <b>http[s]://</b> or <b>ftp://</b></p>\n" ;
  &foot ;
  print LOG "$now $raddr $method \"$F{url}\"\n" ;
  exit ;
}
$url = $1 ; $url =~ s/"/\\"/g ;

#foreach $f (keys %ENV) {
#  print LOG "$f $ENV{$f}\n" ;
#}

chdir "soffice" ;
open (IN,"wget -N -T 15 -x  -S \"$url\" 2>\&1|") ;
$status = 0 ;
while (<IN>) {
  $wget .= $_ ;
  chomp ;
  if (/ HTTP\/[\d]\.[\d][\s]+([\d]+) (.*)/) { $status = $1 ; $reason = $2 ; }
  if (/ Content-Type:[\s]+([\w]+\/[\w\.\-]+)/i) { $type = $1 ; $type =~ tr/A-Z/a-z/ ; }
  if (/ => `(.*)'/) { $ifile= $1 ; } 
  if (/\. failed: /) { $err = $_ ; }
  if (/WWW-Authenticate:[\s]+(Basic.*)/i) { $realm = $1 ; }
}
close(IN) ; $stat = $? ;
if ($stat and $status == 0) {
  print<<EOT;
Status: 500 Retrieve error
Content-type: text/html; charset=iso-8859-1

<title>Retrieve error</title>
$err
<p>
$url
EOT
  print LOG "$now $raddr $method $url $err\n" ;
  exit ;
}
if ($status == 401 and 0) {
    print<<EOT;
Status: $status $reason
Content-type: text/html; charset=iso-8859-1
WWW-Authenticate: $realm

Authorization failed
EOT
  print LOG "$now $raddr $method $url $status $reason\n" ;
  exit ;
}
unless ($status == 200) {
    print<<EOT;
Status: $status $reason
Content-type: text/html; charset=iso-8859-1

<title>Error $status $reason</title>
Error $status $reason
<p>
$url
EOT
  print LOG "$now $raddr $method $url $status $reason\n" ;
  exit ;
}

if ($ifile =~ /\.([a-z]{2,4})$/i) {
  $suffix = $1 ; $suffix =~ tr/A-Z/a-z/ ;
  #print "Suffix $suffix\n" ;
}
if (-f $ifile) {
  @stat = stat($ifile) ;
  $fmod1 = $stat[9] ; 
  $ofile = &escape($ifile) . '.' . $otype ;
} else {
  $fmod1 = 0 ;
}
if ($ims and $fmod1 and $fmod1<=$ims) {
  print "Status: 304 Not Modified\n\n" ;
  print LOG "$now $raddr $method $url 304\n" ;
  exit ;
}

if (-f $ofile) {
  @stat = stat($ofile) ;
  $fmod2 = $stat[9] ; $fsize2 =  $stat[7] ;
} else { 
  $fmod2 = 0 ;
}
if ($fmod1 and $fmod2 and $fmod2 >= $fmod1 and $fsize2) {
  $glmod = &wtime($fmod2);
  print LOG "$now $raddr $method $url cached $ofile $glmod\n" ;
  print<<EOT;
Content-Type: $type{$otype}
Content-Length: $fsize2
Last-Modified: $glmod

EOT
  open (IN,$ofile) or die "$0 error reading $ofile\n" ;
  while (<IN>) { print $_ ; }
  exit ;
}
open (IN,"$oopython ../DocumentConverter.py \"$ifile\" $ofile 2>\&1|") ;
while (<IN>) {
  chomp ;
  if (/ERROR/) { $err = $_ ; }
}
close(IN) ; 
if ($?) {
  print<<EOT;
Status: 500 Conversion error
Content-type: text/html; charset=iso-8859-1

<title>Conversion error</title>
$err
<p>
$url
EOT
  print LOG "$now $raddr $method $url convert error $err\n" ;
  exit ;
}

system("touch -r \"$ifile\" $ofile") ;
if (-f $ofile) {
  @stat = stat($ofile) ;
  $fsize2 =  $stat[7] ;
} else { 
  die "$0 create $ofile failed\n" ;
}

$glmod = &wtime($fmod1);
print LOG "$now $raddr $method $url created $ofile $glmod\n" ;
print<<EOT;
Content-Type: $type{$otype}
Content-Length: $fsize2
Last-Modified: $glmod

EOT
open (IN,$ofile) or die "$0 error reading $ofile\n" ;
while (<IN>) { print $_ ; }
exit ;

# Netscape:
# Reload sends no-cache and IMS
# Shift-Reload sends no-cache without an IMS


sub wtime {
  my @DoW = ('Sunday','Monday','Tuesday','Wednesday','Thursday','Friday','Saturday');
  my @MoY = ('Jan','Feb','Mar','Apr','May','Jun',
        'Jul','Aug','Sep','Oct','Nov','Dec');

  my $time = shift ;

  my ($sec, $min, $hour, $mday, $mon, $year, $wday, $yday, $isdst) = gmtime($time) ;
  $year += 1900;
  sprintf("%s, %02d %s %04d %02d:%02d:%02d %s", substr($DoW[$wday],0,3),
     $mday, $MoY[$mon], $year, $hour, $min, $sec, 'GMT');
}

sub get_gmtime {
    local($_) = @_;
    my $Mstr = 'JanFebMarAprMayJunJulAugSepOctNovDec';
    local($day, $mn, $yr, $hr, $min, $sec, $adate, $mon, $midx);
    if (/([\d]+) ([\w]{3}) ([\d]{4}) ([\d]+):([\d]+):([\d]+) GMT/) {
      $day = $1 ; $mn = $2 ; $yr = $3 ; $hr = $4 ; $min = $5 ; $sec = $6 ;
    } else {
      return 0 ;
    }
    $midx = index($Mstr, substr($mn,0,3));
    if ($midx  < 0) { return 0; }
    else            { $mon = $midx / 3; }
    # Translate to seconds since Epoch
    return (timegm($sec, $min, $hr, $day, $mon, $yr) );
}

sub escape {
    my $str = shift ;
    my $pat = '[\x00-\x20"#%/;:<>?\x7F-\xFF]' ;
    my $pat = '[\x00-\x2F\x3A-\x40\x5B-\x60\x7B-\xFF]' ;
    $str =~ s/($pat)/sprintf("%%%02lx",unpack('C',$1))/ge;
    return($str);
}

# wget -r -l inf -N
# -q quiet -O xx output -N time-stamp
# --ignore-length --header=additional-header

# wget --ignore-length --header=if-modified-since:"Sat, 01 Mar 2008 02:37:00 GMT" -N -O foo.html -nv http://andrew.triumf.ca 2>&1 |less
#application/msword              doc
#application/vnd.ms-powerpoint   ppt
#application/vnd.ms-excel        xls
#application/vnd.oasis.opendocument.text odt
#application/vnd.sun.xml.writer sxw
#application/vnd.sun.xml.calc sxc
#Authorization: Basic YWR2YXg6dHYwZzBucw==


sub head {
  print<<EOT;
Content-type: text/html; charset=iso-8859-1

<html><head><title>Document Converter</title></head>
<body><h2>Document Converter</h2>
<p>
Convert "office" documents to PDF or HTML using 
<a href="http://www.openoffice.org">OpenOffice</a> server, including
Microsoft Word, Excel, PowerPoint, OpenOffice and StarOffice documents and
presentations.
</p>
<form>
Get <input type="text" name="url" size="50">
<input type="reset" value="Clear">
as
<input type="submit" value="PDF" name="op">
<input type="submit" value="Text" name="op">
<input type="submit" value="HTML" name="op">
</form>
<p>
N810 hint (Firefox 3): Make a new window for this page. Tap-and-hold the document link on the original page, then
copy the URL to the clipsheet. Switch to this page and paste the URL in the box above (Ctrl-V).
</p>
EOT
}

sub foot {
  print<<EOT;
<p>Notes:
<ul>
<li>The conversion requires OpenOffice running on the server; there is currently no assurance that it
will be restarted if it dies.
<li>Conversion and upload may take several seconds, or longer for large documents
<li>Not all conversion pairs are supported
</ul>
</p>
<p>
Derived from the <a href="http://www.oooninja.com/2008/02/batch-command-line-file-conversion-with.html">OOONinja article</a> by
Andrew Ziem.
</p>
<address>
Andrew Daviel &lt; advax (at) triumf.ca &gt;
</address>
</body></html>
EOT
}