Recover Images from Lightroom Previews, inspired by http://photo.bragit.com/Lightroom/articles/RecoverFromPreview.shtml

lrpreview2jpg.pl

#!/usr/bin/perl
# 从lrprev里截取图片
use strict;
use warnings;

# single file
my $file = '/path/to/Lightroom x Catalog Previews.lrdata/x/x/id.lrprev';
my $dest = '/path/to/jpeg/prefix';
convert($file, $dest);

# all files
my $path = '/path/to/Lightroom x Catalog Previews.lrdata';
my $dst_path = '/path/to/jpeg/';

recovery_path($path);

sub recovery_path {
	my ($path) = @_;
	opendir DH, $path;
	my @files = readdir(DH);
	closedir DH;
	
	foreach my $file(@files) {
		next if $file =~ m{^\.{1,2}$}xmsi;
		my $full = "$path/$file";
		if (-d $full) {
			print "处理文件夹$full\n";
			recovery_path($full);
		} else {
			convert($full, "$dst_path/$file");
		}
	}
}

sub convert {
	my ($src, $dst) = @_;
	open FIN, '<', $src or die $!;
	my $data;
	while (<FIN>) {
		$data .= $_;
	}
	close FIN;
	$/ = '';
	if ($data =~ m{AgHg(.*?)AgHg}xmsi) {
		my $json = $1;
		#print "json $json\n";
		my @sizes = ();
		while ($json =~ m/height\s=\s(\d+).*?width\s=\s(\d+)/xmsg) {
			push @sizes, { h => $1, w => $2};
		}
		
		# 跳过levels = {
		my $pos = index($data, 'levels');
		my $idx;
		do {
			$idx = index($data, 'level_', $pos);
			if ($idx >= 0) {
				my $from = $idx + 8;
				$pos = $idx + 1;
				my $to = index($data, 'AgHg', $pos);
				
				my $size = shift @sizes;
				# skip small images(you can simple skip this by comment if pair)
				if ($size->{w} >= 1500 || $size->{h} >= 1500) {
					# 跳过小图
					#next if (@sizes > 0); # 写最后一个文件
					my $file .= "$dst-$size->{w}x$size->{h}.jpeg";
					print "write $pos, $to to $file\n";
					open FOUT, '>', $file or die $!;
					binmode FOUT;
					print FOUT ($to == -1 ? substr($data, $from) : substr($data, $from, $to-$from));
					close FOUT;
				}
			}
		} while ($idx != -1);
	}
}