--- trunk/lib/WebPAC/Normalize.pm	2006/07/23 20:19:56	603
+++ trunk/lib/WebPAC/Normalize.pm	2006/09/11 12:56:05	667
@@ -4,11 +4,13 @@
 	_set_rec _set_lookup
 	_get_ds _clean_ds
 	_debug
+	_pack_subfields_hash
 
 	tag search display
 	marc marc_indicators marc_repeatable_subfield
 	marc_compose marc_leader
 	marc_duplicate marc_remove
+	marc_original_order
 
 	rec1 rec2 rec
 	regex prefix suffix surround
@@ -22,7 +24,6 @@
 
 #use base qw/WebPAC::Common/;
 use Data::Dump qw/dump/;
-use Encode qw/from_to/;
 use Storable qw/dclone/;
 
 # debugging warn(s)
@@ -35,11 +36,11 @@
 
 =head1 VERSION
 
-Version 0.15
+Version 0.19
 
 =cut
 
-our $VERSION = '0.15';
+our $VERSION = '0.19';
 
 =head1 SYNOPSIS
 
@@ -468,7 +469,6 @@
 	foreach (@_) {
 		my $v = $_;		# make var read-write for Encode
 		next unless (defined($v) && $v !~ /^\s*$/);
-		from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
 		my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
 		if (defined $sf) {
 			push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $i1, $i2, $sf => $v ];
@@ -539,12 +539,15 @@
 
 	warn "### marc_compose input subfields = ", dump(@_),$/ if ($debug > 2);
 
+	if ($#_ % 2 != 1) {
+		die "ERROR: marc_compose",dump($f,@_)," not valid (must be even).\nDo you need to add first() or join() around some argument?\n";
+	}
+
 	while (@_) {
-		my $sf = shift or die "marc_compose $f needs subfield";
+		my $sf = shift;
 		my $v = shift;
 
 		next unless (defined($v) && $v !~ /^\s*$/);
-		from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
 		warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ),$/ if ($debug > 1);
 		if ($sf ne '+') {
 			push @$m, ( $sf, $v );
@@ -641,11 +644,134 @@
 	warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
 }
 
+=head2 marc_original_order
+
+Copy all subfields preserving original order to marc field.
+
+  marc_original_order( marc_field_number, original_input_field_number );
+
+Please note that field numbers are consistent with other commands (marc
+field number first), but somewhat counter-intuitive (destination and then
+source).
+
+You might want to use this command if you are just renaming subfields or
+using pre-processing modify_record in C<config.yml> and don't need any
+post-processing or want to preserve order of original subfields.
+
+
+=cut
+
+sub marc_original_order {
+
+	my ($to, $from) = @_;
+	die "marc_original_order needs from and to fields\n" unless ($from && $to);
+
+	return unless defined($rec->{$from});
+
+	my $r = $rec->{$from};
+	die "record field $from isn't array\n" unless (ref($r) eq 'ARRAY');
+
+	my ($i1,$i2) = defined($marc_indicators->{$to}) ? @{ $marc_indicators->{$to} } : (' ',' ');
+	warn "## marc_original_order($to,$from) source = ", dump( $r ),$/ if ($debug > 1);
+
+	foreach my $d (@$r) {
+
+		if (! defined($d->{subfields}) && ref($d->{subfields}) ne 'ARRAY') {
+			warn "# marc_original_order($to,$from): field $from doesn't have subfields specification\n";
+			next;
+		}
+	
+		my @sfs = @{ $d->{subfields} };
+
+		die "field $from doesn't have even number of subfields specifications\n" unless($#sfs % 2 == 1);
+
+		warn "#--> d: ",dump($d), "\n#--> sfs: ",dump(@sfs),$/ if ($debug > 2);
+
+		my $m = [ $to, $i1, $i2 ];
+
+		while (my $sf = shift @sfs) {
+
+			warn "#--> sf: ",dump($sf), $/ if ($debug > 2);
+			my $offset = shift @sfs;
+			die "corrupted sufields specification for field $from\n" unless defined($offset);
+
+			my $v;
+			if (ref($d->{$sf}) eq 'ARRAY') {
+				$v = $d->{$sf}->[$offset] if (defined($d->{$sf}->[$offset]));
+			} elsif ($offset == 0) {
+				$v = $d->{$sf};
+			} else {
+				die "field $from subfield '$sf' need occurence $offset which doesn't exist", dump($d->{$sf});
+			}
+			push @$m, ( $sf, $v ) if (defined($v));
+		}
+
+		if ($#{$m} > 2) {
+			push @{ $marc_record->[ $marc_record_offset ] }, $m;
+		}
+	}
+
+	warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1);
+}
+
+
 =head1 Functions to extract data from input
 
 This function should be used inside functions to create C<data_structure> described
 above.
 
+=head2 _pack_subfields_hash
+
+ @values = _pack_subfields_hash( $h, $include_subfields )
+
+=cut
+
+sub _pack_subfields_hash {
+
+	warn "## _pack_subfields_hash( ",dump(@_), " )\n" if ($debug > 1);
+
+	my ($h,$include_subfields) = @_;
+
+	if ( defined($h->{subfields}) ) {
+		my $sfs = delete $h->{subfields} || die "no subfields?";
+		my @out;
+		while (@$sfs) {
+			my $sf = shift @$sfs;
+			push @out, '^' . $sf if ($include_subfields);
+			my $o = shift @$sfs;
+			if ($o == 0 && ref( $h->{$sf} ) ne 'ARRAY' ) {
+				# single element subfields are not arrays
+warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n";
+
+				push @out, $h->{$sf};
+			} else {
+warn "====> $sf $o / $#$sfs ", dump( $sfs, $h->{$sf} ), "\n";
+				push @out, $h->{$sf}->[$o];
+			}
+		}
+		if ($include_subfields) {
+			return join('', @out);
+		} else {
+			return @out;
+		}
+	} else {
+		if ($include_subfields) {
+			my $out = '';
+			foreach my $sf (keys %$h) {
+				if (ref($h->{$sf}) eq 'ARRAY') {
+					$out .= '^' . $sf . join('^' . $sf, @{ $h->{$sf} });
+				} else {
+					$out .= '^' . $sf . $h->{$sf};
+				}
+			}
+			return $out;
+		} else {
+			# FIXME this should probably be in alphabetical order instead of hash order
+			values %{$h};
+		}
+	}
+}
+
 =head2 rec1
 
 Return all values in some field
@@ -662,13 +788,15 @@
 	return unless (defined($rec) && defined($rec->{$f}));
 	warn "rec1($f) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
 	if (ref($rec->{$f}) eq 'ARRAY') {
-		return map { 
-			if (ref($_) eq 'HASH') {
-				values %{$_};
+		my @out;
+		foreach my $h ( @{ $rec->{$f} } ) {
+			if (ref($h) eq 'HASH') {
+				push @out, ( _pack_subfields_hash( $h ) );
 			} else {
-				$_;
+				push @out, $h;
 			}
-		} @{ $rec->{$f} };
+		}
+		return @out;
 	} elsif( defined($rec->{$f}) ) {
 		return $rec->{$f};
 	}
@@ -924,7 +1052,7 @@
 	my $v = shift @r;
 	warn "### first rec($fld,$sf) = ",dump($v),$/ if ($debug > 2);
 
-	return '' if( ! defined($v) || $v =~ /^\s*$/);
+	return '' if ( ! defined($v) || $v =~ /^\s*$/);
 
 	my @s = split( $regex, $v );
 	warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s),$/ if ($debug > 1);