File Coverage

blib/lib/MARC/Parser/XML.pm

Criterion	Covered	Total	%
statement	12	14	85.7
branch			n/a
condition			n/a
subroutine	5	5	100.0
pod			n/a
total	17	19	89.4

package MARC::Parser::XML;

use strict;
use warnings;
use 5.008_005;
our $VERSION = '0.02';

use Carp qw(croak);
use XML::LibXML::Reader;

sub new {
    my ( $class, $input ) = @_;

    my $self = bless { input => $input, rec_number => 0, }, $class;

    # check for file or filehandle
    my $ishandle = eval { fileno($input); };
    if ( !$@ && defined $ishandle ) {
        binmode $input;    # drop all PerlIO layers, as required by libxml2
        my $reader = XML::LibXML::Reader->new( IO => $input )
            or croak "cannot read from filehandle $input\n";
        $self->{xml_reader} = $reader;
    }
    elsif ( defined $input && $input !~ /\n/ && -e $input ) {
        my $reader = XML::LibXML::Reader->new( location => $input )
            or croak "cannot read from file $input\n";
        $self->{xml_reader} = $reader;
    }
    elsif ( defined $input && length $input > 0 ) {
        $input = ${$input} if ( ref($input) // '' eq 'SCALAR' );
        my $reader = XML::LibXML::Reader->new( string => $input )
            or croak "cannot read XML string $input\n";
        $self->{xml_reader} = $reader;
    }
    else {
        croak "file, filehande or string $input does not exists";
    }
    return $self;
}

sub next {
    my ($self) = @_;

    return
        unless $self->{xml_reader}
        ->nextElement( 'record', 'http://www.loc.gov/MARC21/slim' );

    if ( my $record = $self->_decode() ) {
        return $record;
    }
    else {
        return $self->next;
    }

    return;
}

sub _decode {
    my ($self) = @_;
    my @record;

    foreach my $field_node (
        $self->{xml_reader}->copyCurrentNode(1)->getChildrenByTagName('*') )
    {

        if ( $field_node->localName =~ m/leader/ ) {
            push @record,
                [ 'ldr', undef, undef, '_', $field_node->textContent ];
        }
        elsif ( $field_node->localName =~ m/controlfield/ ) {
            push @record,
                [
                $field_node->getAttribute('tag'), undef,
                undef,                            '_',
                $field_node->textContent
                ];
        }
        elsif ( $field_node->localName eq 'datafield' ) {
            push @record,
                [
                $field_node->getAttribute('tag'),
                $field_node->getAttribute('ind1') // '',
                $field_node->getAttribute('ind2') // '',
                map { $_->getAttribute('code'), $_->textContent }
                    $field_node->getChildrenByTagName('*')
                ];
        }
    }
    return \@record;
}

1;
__END__