Sync information between BibTeX file and PDF XMP

bibpdf.pl

#!/usr/bin/perl -w

# Sync the information between BibTeX and PDF XMP.
# [email protected]

use strict;
use warnings;

use POSIX;
use Getopt::Std;
use File::Basename;
use File::Find;
use File::Spec;

use Text::BibTeX;
use PDF::API2;

$|++;

my $pwd = getcwd();

my %options; getopts('f:', \%options);
my $attachment = $options{'f'} || 'file';

my $bibpath = shift || '$pwd/documents.bib';
$bibpath = File::Spec->rel2abs($bibpath) if !File::Spec->file_name_is_absolute($bibpath);
my $bibdir = dirname($bibpath);

chdir($bibdir);

if (-e $bibpath) {
	my $bibfile = Text::BibTeX::File->new($bibpath);
	
	while (my $entry = new Text::BibTeX::Entry $bibfile) {
	 	next unless $entry->parse_ok && $entry->metatype == BTE_REGULAR;
		if ($entry->exists($attachment)) {
			my $path = $entry->get($attachment);
			next if (!defined $path || $path !~ m/:application\/pdf$/);
			$path =~ s/:application\/pdf$//g;
			$path =~ s/^.*://;
			$path = File::Spec->rel2abs($bibdir."/".$path) 
				if !File::Spec->file_name_is_absolute($path);
			next if (!-e $path);
			print "$path is not processed correctly.\n"
			if !eval { 
				my $file = PDF::API2->open($path);
				my %xmp = $file->info();
				my $title = $entry->get('title');
				$xmp{'Title'} = $title if defined $title;
				my $author = $entry->get('author');
				$xmp{'Author'} = $author if defined $author;
				my $keywords = $entry->get('keywords');
				$xmp{'Keywords'} = $keywords if defined $keywords;
				my $subject = $entry->print_s;
				$xmp{'Subject'} = "BibTeX:".$subject if defined $subject;
				
				$file->info(%xmp);
				$file->update();
				print "bibtex xmp is added to $path\n";
			};
		}
	}
	
	$bibfile->close;
}
else {
	my @pdfpaths;

	find(sub {
		my $path = $File::Find::name;
		return if -d $path || $path !~ m/\.pdf$/;
		push @pdfpaths, $path;
	}, $bibdir);

	my $pdfpath_count = @pdfpaths;
	
	return if $pdfpath_count == 0;
	
	open BFH, ">", $bibpath or die $!;
	
	foreach my $path (@pdfpaths) {
		print "$path is not processed correctly.\n"
		if !eval {
			my $file = PDF::API2->open($path);
			my %xmp = $file->info();
			my $subject = $xmp{"Subject"};
			if ($subject =~ m/^BibTeX:/) {
				$subject =~ s/^BibTeX://;
				printf BFH $subject;
			}
		};
	}
	
	close BFH;
	
}

• Use bibpdf.pl /path/to/docs.bib to sync information from the .bib file to PDFs.
• Use bibpdf.pl to sync information from PDFs in current directory to a .bib file.