#!/usr/bin/perl -w

use utf8;
binmode(STDOUT, ":utf8");

$/ = "</dublin_core>\n"; # record separator

$what = 1000; # dummy id for when there’s no file

while (<>) {
    $cul = "";
    $tac = "";

    # discard the top and bottom tags
    s/<collection>\n//;
    s/<\/collection>\n//;

    # extract the file path from the identifier
    # use the file name as an id
    # note that identifier element is discarded!
    #if (s!<dcvalue element="relation" qualifier="uri">\s*(.*?)\s*<\/dcvalue>\n!!s) {
    if (/<dcvalue element="relation" qualifier="uri">\s*(.*?)\s*<\/dcvalue>\n/) {
	$cul = $1;
	if ($cul =~ /docs/) {
		#$cul =~ s@.*(folders/|d/)([^?&/]+).*@\2@;
		$cul =~ s@.*(d/)([^?&/]+).*@$2@;
		$tac = "https://drive.google.com/uc?export=download&id=$cul";
    		print "$tac\n";
    		print "\n";
	}
        $path = '';
	$id = $what++;
    } else {
        $path = '';
        $id = $what++;
    }

    # let the operator know where we’re up to
    print "$path/$id\n";

    # create the item directory
    mkdir "import/$id", 0755;

    # create the dublin_core.xml file
    open DC, ">import/$id/dublin_core.xml"
      or die "Cannot open dublin core for $id, $!\n";
    print DC $_;
    close DC;

    # assuming we have a file ...
    if ($tac) {
	print "$tac\n";
	print "$cul\n";

	system "curl -L \"$tac\" -o import/$id/$cul";
	#system "curl -s -L \"$tac\" -o import/$id/$cul";

	my $output = `file import/$id/$cul`;
	print $output;

	if ($output =~ /PDF/) {
		$ext = "pdf";
	}
	if ($output =~ /Word/) {
		system "soffice --headless --convert-to pdf import/$id/$cul";
		$ext = "pdf";
	}

	system "mv import/$id/$cul import/$id/$cul.$ext";

        # ... create the contents file ...
        open OUT, ">import/$id/contents"
          or die "Cannot open contents for $id, $!\n";
        print OUT "$cul.$ext";
        close OUT;
    }
}

__END__