use lib instead of blib for local testing

[MARC-Fast] / Fast.pm

diff --git a/Fast.pm b/Fast.pm
index b587be3..aef3d5b 100644 (file)
--- a/Fast.pm
+++ b/Fast.pm
@@ -2,12 +2,12 @@ package MARC::Fast;
 
 use strict;
 use Carp;
-use Data::Dumper;
+use Data::Dump qw/dump/;
 
 BEGIN {
        use Exporter ();
        use vars qw ($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS);
-       $VERSION     = 0.08;
+       $VERSION     = 0.10;
        @ISA         = qw (Exporter);
        #Give a hoot don't pollute, do not export more than needed by default
        @EXPORT      = qw ();
@@ -87,7 +87,8 @@ sub new {
                my $len = read($self->{fh}, $leader, 24);
 
                if ($len < 24) {
-                       carp "short read of leader, aborting\n";
+                       warn "short read of leader, aborting\n";
+                       $self->{count}--;
                        last;
                }
 
@@ -124,6 +125,7 @@ sub new {
 
                # skip to next record
                my $o = substr($leader,0,5);
+               warn "# in record ", $self->{count}," record length isn't number but: ",dump($o),"\n" unless $o =~ m/^\d+$/;
                if ($o > 24) {
                        seek($self->{fh},$o-24,1) if ($o);
                } else {
@@ -275,7 +277,7 @@ sub last_leader {
 
 Read record with specified MFN and convert it to hash
 
-  my $hash = $marc->to_hash($mfn);
+  my $hash = $marc->to_hash( $mfn, include_subfields => 1, );
 
 It has ability to convert characters (using C<hash_filter>) from MARC
 database before creating structures enabling character re-mapping or quick
@@ -302,48 +304,54 @@ sub to_hash {
 
        my $mfn = shift || confess "need mfn!";
 
+       my $args = {@_};
+
        # init record to include MFN as field 000
        my $rec = { '000' => [ $mfn ] };
 
        my $row = $self->fetch($mfn) || return;
 
-       foreach my $rec_nr (keys %{$row}) {
-               foreach my $l (@{$row->{$rec_nr}}) {
+       foreach my $tag (keys %{$row}) {
+               foreach my $l (@{$row->{$tag}}) {
 
                        # remove end marker
                        $l =~ s/\x1E$//;
 
                        # filter output
-                       $l = $self->{'hash_filter'}->($l, $rec_nr) if ($self->{'hash_filter'});
+                       $l = $self->{'hash_filter'}->($l, $tag) if ($self->{'hash_filter'});
 
                        my $val;
 
                        # has identifiers?
                        ($val->{'i1'},$val->{'i2'}) = ($1,$2) if ($l =~ s/^([01 #])([01 #])\x1F/\x1F/);
 
+                       my $sf_usage;
+                       my @subfields;
+
                        # has subfields?
                        if ($l =~ m/\x1F/) {
                                foreach my $t (split(/\x1F/,$l)) {
                                        next if (! $t);
                                        my $f = substr($t,0,1);
-                                       # repeatable subfileds. When we hit first one,
-                                       # store CURRENT (up to that) in first repetition
-                                       # of this record. Then, new record with same
-                                       # identifiers will be created.
+
+                                       push @subfields, ( $f, $sf_usage->{$f}++ || 0 );
+
+                                       # repeatable subfiled -- convert it to array
                                        if ($val->{$f}) {
-                                               push @{$rec->{$rec_nr}}, $val;
-                                               $val = {
-                                                       i1 => $val->{i1},
-                                                       i2 => $val->{i2},
-                                               };
+                                               if ( ref($val->{$f}) ne 'ARRAY' ) {
+                                                       $val->{$f} = [ $val->{$f}, $val ];
+                                               } else {
+                                                       push @{$val->{$f}}, $val;
+                                               }
                                        }
                                        $val->{substr($t,0,1)} = substr($t,1);
                                }
+                               $val->{subfields} = [ @subfields ] if $args->{include_subfields};
                        } else {
                                $val = $l;
                        }
 
-                       push @{$rec->{$rec_nr}}, $val;
+                       push @{$rec->{$tag}}, $val;
                }
        }
 
@@ -377,6 +385,27 @@ sub to_ascii {
 1;
 __END__
 
+=head1 UTF-8 ENCODING
+
+This module does nothing with encoding. But, since MARC format is byte
+oriented even when using UTF-8 which has variable number of bytes for each
+character, file is opened in binary mode.
+
+As a result, all scalars recturned to perl don't have utf-8 flag. Solution is
+to use C<hash_filter> and L<Encode> to decode utf-8 encoding like this:
+
+  use Encode;
+
+  my $marc = new MARC::Fast(
+       marcdb => 'utf8.marc',
+       hash_filter => sub {
+               Encode::decode( 'utf-8', $_[0] );
+       },
+  );
+
+This will affect C<to_hash>, but C<fetch> will still return binary representation
+since it doesn't support C<hash_filter>.
+
 =head1 AUTHOR
 
        Dobrica Pavlinusic