G'day jinnicky,

The module you'll want for that is OpenOffice::OODoc::Meta.

I haven't previously used this. I did install OpenOffice::OODoc some time ago and the OpenOffice::OODoc::Meta, along with other OpenOffice::OODoc::* modules, appear to be bundled with this (see OpenOffice-OODoc distribution details). So, if you have OpenOffice::OODoc, you probably also have the related modules.

The documentation looks good and usage seems straightforward.

I created a very basic text document for testing (pm_1156099_test.odt) and added a title ("PM 1156099 Test Document") via the Properties menu item. I then created this test script:

#!/usr/bin/env perl -l

use strict;
use warnings;

use OpenOffice::OODoc::Meta;

my $meta = OpenOffice::OODoc::Meta::->new(file => 'pm_1156099_test.odt
+');

print 'Author:  ', $meta->creator();
print 'Title:   ', $meta->title();
print 'Created: ', $meta->date();
[download]

This produced this output:

Author:  Ken Cotterill
Title:   PM 1156099 Test Document
Created: 2016-02-25T15:23:56
[download]

There's lots of other metadata you can access if you want.

[I do recall hearing something about OpenOffice::OODoc being superceded by ODF::lpOD. Both sets of modules are by the same author, Jean-Marie Gouarné, and the ODF::lpOD distribution is more recent. I looked around for some definitive information on this but was unsuccessful, so that remains unconfirmed: perhaps another monk can provide something more substantial on this matter.]

The meta data doesn't show the title which is a paragraph with the style of 'Title' or Psomething.

I installed ODF::lpOD once I figured out that lpOD starts with lower case L. It's documentation is voluminous and not much better than OpenOffice::OODoc's.

However I did get it to work!

#!/usr/bin/perl -w

use strict;
use ODF::lpOD;

my $file = $ARGV[0];
die "You must supply an odf file name\n" unless $file;
my $doc = odf_document->get($file) or die "Can't load $file\n$!\n";
my $context = $doc->body;
my $meta = $doc->meta;

# Doesn't do what I wanted
my $title = $meta->get_title;
print "Title: $title\n" if $title;  # shows Title: c if at all
my $p = $context->get_paragraph(style=>'Title', content=>'ODF',positio
+n=>0);
if ($p) {print $p->get_text()."\n";}
else {print "Not found\n";}  # prints Not found

#this works
my ($i,$ps,$style,$pStyle);
for ($i = 0; $i < 6; $i++) {
  $p = $context->get_paragraph(position=>$i);
  if ($p) {
    print "Paragraph $i";
    $ps = $p->get_style();
    if ($ps) {
      if ($ps =~ m/^P\d+$/) {   # Check for internal styles
        $style = $doc->get_style('paragraph',$ps) || '';
        $pStyle = $style->get_parent_style if $style;
        $ps = $pStyle if $pStyle; # This gets the real name
      }
      print " Style: $ps\n";
    }
    else {print "No Style\n";}
    my $data = $p->get_text(recursive=>1);
    if ($data) {print "$data\n";}
    else {print "--No data\n";}
  } else {
    print "Paragraph $i not found\n";
  }
}
[download]

Thanks Ken

—Bob

The ODF::LpOD modules worked on .odf files but choked on .sxw (version 1.1) files.

Ken's suggestion about the similarity between those modules and the OpenOffice::OODoc modules helped me to go back to them and come up with this code:

#!/usr/bin/perl
use warnings;
use strict;
use OpenOffice::OODoc;

# setup file and styles
my $file = $ARGV[0];
die "You must supply an odf file name\n" unless $file;
print "File: $file\n";
my $container = odfContainer($file) or die "Can't get document $file\n
+";
my $doc       = odfDocument(container => $container, part => 'content'
+) or die "Can't get content in $file\n";

my ($i,$element,$text,$style);

for ($i = 0; $i < 10; $i++) {
  $element = $doc->getElement('//text:p',$i);
  if ($element) {
    $text = $doc->getText($element);
    if ($text) {
      $style = $doc->getAttribute($element,'style name')||'';
      if ($style && ($style =~ m/^P\d+$/)) {
        $style = $doc->getAncestorStyle($style);
      }
      print "Paragraph $i: ";
      print "($style) " if $style;
      print "$text" if $text;
      print "\n";
    }
  }
}
[download]

—Bob