anytopdf

#!/usr/bin/perl
#
# anytopdf : convert any openoffice.org readable document to a PDF file using the current user's macro library (license: LGPL)
# 
# http://code.google.com/p/anytopdf
#
#  known issues (MAJOR - READ THIS!):
#   - does not support attempts to use file names with commas, quotes, brackets, or newlines
#   - for some unknown reason, using perl system() fails silently, so we use backticks, which is REALLY bad!
#
#  notes:
#   - auto creates or extends the current user's Standard/AnyToPDF.xba macro libary
#
#  requirements:
#   - openoffice.org 3+ (with macro support, and support for the file format you want to use)
#   - standard perl install
#   - unix style system
#  
#  changelog:
#   1.0    2009-04     initial version
#  
#  thanks:
#   - http://www.togaware.com/linux/survivor/Convert_MS_Word.html / DannyB : original macro code
#   - openoffice.org contributors
#   - unix hackers with a conscience who don't work for morally bankrupt corporations, governments and militaries.
#
#  ... dedicated to peace, love, understanding and respect for all beings.

# requirements
use Cwd;
use File::Basename; # present in default perl distribution - should be present on every machine

# basic configuration and initialisation (you probably won't need to change this)
$VERSION = '1.0';
$OPENOFFICE_BINARY_NAME = 'soffice.bin';
$HOME = $ENV{'HOME'};
$MACRO_FILE = $HOME . '/.openoffice.org/3/user/basic/Standard/AnyToPDF.xba';
if(! -e $MACRO_FILE) {
 $MACRO_FILE = $HOME . '/.ooo3/user/basic/Standard/AnyToPDF.xba';
 if(! -e $MACRO_FILE) {
  print "ERROR: Unable to locate macro file.\n";
  exit(1);
 }
}

# macro definitions
$MACRO_HEADER = '<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE script:module PUBLIC "-//OpenOffice.org//DTD OfficeDocument 1.0//EN" "module.dtd">
<script:module xmlns:script="http://openoffice.org/2000/script" script:name="AnyToPDF" script:language="StarBasic">REM  *****  BASIC  *****
' . "\n";
$MACROS = 'Sub ConvertAnyToPDF(inFile,outFile)
   inURL = ConvertToURL(inFile)
   oDoc = StarDesktop.loadComponentFromURL(inURL, &quot;_blank&quot;, 0, Array(AnyToPDFMPV(&quot;Hidden&quot;, True), ))
   outURL = ConvertToURL(outFile)
   oDoc.storeToURL(outURL, Array(AnyToPDFMPV(&quot;FilterName&quot;, &quot;writer_pdf_Export&quot;), ))
   oDoc.close(True)
End Sub
Function AnyToPDFMPV( Optional cName As String, Optional uValue ) As com.sun.star.beans.PropertyValue
   Dim oPropertyValue As New com.sun.star.beans.PropertyValue
   If Not IsMissing( cName ) Then
      oPropertyValue.Name = cName
   EndIf
   If Not IsMissing( uValue ) Then
      oPropertyValue.Value = uValue
   EndIf
   AnyToPDFMPV() = oPropertyValue
End Function' . "\n";
$MACRO_FOOTER = '</script:module>';

# check that the openoffice program is in the path
$OPENOFFICE_BINARY = `which $OPENOFFICE_BINARY_NAME 2>/dev/null`;
if(length($OPENOFFICE_BINARY)<=2) {
 $OPENOFFICE_BINARY = `locate soffice.bin 2>/dev/null |grep '^/usr/' |head -n 1`;
 if(length($OPENOFFICE_BINARY)<=2) {
  $path = $ENV{'PATH'}; error("Could not find openoffice.org binary '$OPENOFFICE_BINARY_NAME' in PATH ($path) or via slocate.",100);
 }
}
chop($OPENOFFICE_BINARY);
# check that we can execute it
if(!-x $OPENOFFICE_BINARY) { error("openoffice.org binary '$OPENOFFICE_BINARY_NAME' found in PATH at '$OPENOFFICE_BINARY', but cannot be executed!",101); }

# check that we have an input file argument and that it's a real file we can actually read
$infile = $ARGV[0];
if($infile eq '' || $infile eq '-h' || $infile eq '--help' || $infile eq '/?') { usage(); }
if($infile eq '') { error("Must supply input file name!",90); }
if($infile =~ /[,\(\)'"\r\n]/) { error("Input file name ($infile) cannot include comma, brackets, new lines or quotes.",94); }
if(!(-e $infile)) { error("Input file '$infile' does not exist!",91); }
if(!(-r $infile)) { error("Input file '$infile' is not readable!",92); }

# check that we have an output file argument and that it's in a real directory we can actually write, or it exists and we can overwrite it
$outfile = $ARGV[1];
if($outfile eq '' || $outfile eq '-h' || $outfile eq '--help' || $outfile eq '/?') { usage(); }
if($outfile eq '') { error("Must supply output file name!",95); }
if($outfile =~ /[,\(\)'"\r\n]/) { error("Outfile file name ($outfile) cannot include comma, brackets, new lines or quotes.",94); }
$outdir = dirname($outfile);
# prepends output directory to file should it be relative vs absolute (x.pdf vs. /some/place/x.pdf) - macro needs absolute
if($outdir eq '.') { $outdir = getcwd(); }
$outfile = $outdir . '/' . basename($outfile);
if(!(-e $outdir)) { error("Output directory '$outdir' does not exist!",96); }
else {
 if(!(-w $outdir)) { error("Output directory '$outdir' exists but is not writable!",97); }
 if((-e $outfile) && !(-w $outfile)) { error("Output file '$outfile' exists but is not writable!",98); }
}

# now check that we're installed in the user's openoffice settings
if(!(-e $MACRO_FILE)) { warning("Macro file '$MACRO_FILE' not present, will attempt to create."); }
else { if(!(-r $MACRO_FILE)) { error("Macro file '$MACRO_FILE' exists but is not readable.",80); } }
$MACRO_DIR = dirname($MACRO_FILE);
if(-e $MACRO_DIR) {
 if(!(-d $MACRO_DIR)) { error("openoffice.org macro directory '$MACRO_DIR' is not a directory!",81); }
}
else {
 `mkdir -p '$MACRO_DIR'`;
 if(!(-e $MACRO_DIR)) { error("Unable to create openoffice.org macro directory '$MACRO_DIR'!",82); }
}
if(-e $MACRO_FILE) {
 # check we're inside
 $data = `grep 'AnyToPDF' '$MACRO_FILE'`;
 if(!($data =~ /AnyToPDF/)) {
  $data = `cat '$MACRO_FILE'`;
  $data =~ s/$MACRO_FOOTER.*?$//msg;
  $data .= $MACROS;
  $data .= $MACRO_FOOTER;
  open(MACROS,">$MACRO_FILE") || error("Unable to open existing openoffice.org macro file '$MACRO_FILE' for writing!",83);
  print MACROS $data;
  close(MACROS);
 }
}
else {
 # attempt to create
 open(MACROS,">$MACRO_FILE") || error("Unable to open new openoffice.org macro file '$MACRO_FILE' for writing!",84);
 print MACROS $MACRO_HEADER . $MACROS . $MACRO_FOOTER;
 close(MACROS);
}

# finally, check that our script file is registered
$data = `cat '$MACRO_DIR/script.xlb'`;
if(!($data =~ /library:name="AnyToPDF"/)) {
 warning("Extending user's openoffice.org scripts registry with AnyToPDF macro module.");
 $data =~ s/<\/library:library>.*?$//msg;
 $data .= " <library:element library:name=\"AnyToPDF\"/>\n";
 $data .= "</library:library>";
 open(SCRIPTS,">$MACRO_DIR/script.xlb") || error("Unable to open openoffice.org macro registry file '$MACRO_FILE' for writing!",85);
 print SCRIPTS $data;
 close(SCRIPTS);
}

# now perform the conversion
#  ... 1st add dirname if not present.
if(!($infile =~ /^\//)) { $infile = getcwd() . '/' . $infile; }
$cmd = "$OPENOFFICE_BINARY -headless -invisible -norestore " . "\"macro:///Standard.AnyToPDF.ConvertAnyToPDF(" . $infile . "," . $outfile . ")\"";
`$cmd`;
# unsure why this doesn't work, fails silently.
#@args = ($OPENOFFICE_BINARY,'-headless','-invisible','-norestore','"macro:///Standard.AnyToPDF.ConvertAnyToPDF('.$infile.','.$outfile.')"');
#print system(@args) . "\n";
exit(0);

sub usage {
 print STDERR "anytopdf v$VERSION (LGPL) - converts arbitrary documents to PDF format using openoffice.org v3 macros.\n";
 print STDERR " http://code.google.com/p/anytopdf\n\n";
 print STDERR "usage: $0 <infile> <outfile>\n";
 print STDERR " <infile>   input file in any format that openoffice.org can read (doc/xls/odt/rtf/html/txt/etc.)\n";
 print STDERR " <outfile>  output file (will be written in PDF format)\n\n";
 print STDERR "return values: (more details will be returned via STDERR)\n";
 print STDERR " 0          success\n";
 print STDERR " 80-89      problem with writing openoffice.org macro file\n";
 print STDERR " 90-94      problem relating to input file\n";
 print STDERR " 95-99      problem relating to output file\n";
 print STDERR " 100-109    problem finding/executing openoffice binary\n";
 exit(1);
}

sub error {
 ($str,$code) = @_;
 print STDERR "ERROR: $str\n";
 exit($code);
}

sub warning {
 print STDERR "WARNING: " . shift(@_) . "\n";
}