--- trunk/lib/WebPAC/Normalize.pm	2006/07/04 11:08:43	579
+++ trunk/lib/WebPAC/Normalize.pm	2006/08/25 12:31:01	618
@@ -9,6 +9,7 @@
 	marc marc_indicators marc_repeatable_subfield
 	marc_compose marc_leader
 	marc_duplicate marc_remove
+	marc_original_order
 
 	rec1 rec2 rec
 	regex prefix suffix surround
@@ -35,11 +36,11 @@
 
 =head1 VERSION
 
-Version 0.11
+Version 0.16
 
 =cut
 
-our $VERSION = '0.11';
+our $VERSION = '0.16';
 
 =head1 SYNOPSIS
 
@@ -69,6 +70,7 @@
 	row => $row,
 	rules => $normalize_pl_config,
 	marc_encoding => 'utf-8',
+	config => $config,
   );
 
 Options C<lookup>, C<row>, C<rules> and C<log> are mandatory while all
@@ -90,6 +92,7 @@
 	no strict 'subs';
 	_set_lookup( $arg->{lookup} );
 	_set_rec( $arg->{row} );
+	_set_config( $arg->{config} );
 	_clean_ds( %{ $arg } );
 	eval "$arg->{rules}";
 	die "error evaling $arg->{rules}: $@\n" if ($@);
@@ -111,6 +114,34 @@
 	$rec = shift or die "no record hash";
 }
 
+=head2 _set_config
+
+Set current config hash
+
+  _set_config( $config );
+
+Magic keys are:
+
+=over 4
+
+=item _
+
+Code of current database
+
+=item _mfn
+
+Current MFN
+
+=back
+
+=cut
+
+my $config;
+
+sub _set_config {
+	$config = shift;
+}
+
 =head2 _get_ds
 
 Return hash formatted as data structure
@@ -495,6 +526,9 @@
   	'c', rec('200','c')
   );
 
+If you specify C<+> for subfield, value will be appended
+to previous defined subfield.
+
 =cut
 
 sub marc_compose {
@@ -504,17 +538,23 @@
 	my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
 	my $m = [ $f, $i1, $i2 ];
 
+	warn "### marc_compose input subfields = ", dump(@_),$/ if ($debug > 2);
+
 	while (@_) {
 		my $sf = shift or die "marc_compose $f needs subfield";
 		my $v = shift;
 
 		next unless (defined($v) && $v !~ /^\s*$/);
 		from_to($v, 'iso-8859-2', $marc_encoding) if ($marc_encoding);
-		push @$m, ( $sf, $v );
 		warn "## ++ marc_compose($f,$sf,$v) ", dump( $m ),$/ if ($debug > 1);
+		if ($sf ne '+') {
+			push @$m, ( $sf, $v );
+		} else {
+			$m->[ $#$m ] .= $v;
+		}
 	}
 
-	warn "## marc_compose(d) ", dump( $m ),$/ if ($debug > 1);
+	warn "## marc_compose current marc = ", dump( $m ),$/ if ($debug > 1);
 
 	push @{ $marc_record->[ $marc_record_offset ] }, $m if ($#{$m} > 2);
 }
@@ -602,6 +642,77 @@
 	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
@@ -647,7 +758,14 @@
 	my $f = shift;
 	return unless (defined($rec && $rec->{$f}));
 	my $sf = shift;
-	return map { $_->{$sf} } grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} };
+	warn "rec2($f,$sf) = ", dump( $rec->{$f} ), $/ if ($debug > 1);
+	return map {
+		if (ref($_->{$sf}) eq 'ARRAY') {
+			@{ $_->{$sf} };
+		} else {
+			$_->{$sf};
+		}
+	} grep { ref($_) eq 'HASH' && $_->{$sf} } @{ $rec->{$f} };
 }
 
 =head2 rec
@@ -660,10 +778,16 @@
 =cut
 
 sub rec {
+	my @out;
 	if ($#_ == 0) {
-		return rec1(@_);
+		@out = rec1(@_);
 	} elsif ($#_ == 1) {
-		return rec2(@_);
+		@out = rec2(@_);
+	}
+	if (@out) {
+		return @out;
+	} else {
+		return '';
 	}
 }
 
@@ -696,7 +820,7 @@
 =cut
 
 sub prefix {
-	my $p = shift or die "prefix needs string as first argument";
+	my $p = shift or return;
 	return map { $p . $_ } grep { defined($_) } @_;
 }
 
@@ -759,6 +883,78 @@
 	}
 }
 
+=head2 config
+
+Consult config values stored in C<config.yml>
+
+  # return database code (key under databases in yaml)
+  $database_code = config();	# use _ from hash
+  $database_name = config('name');
+  $database_input_name = config('input name');
+  $tag = config('input normalize tag');
+
+Up to three levels are supported.
+
+=cut
+
+sub config {
+	return unless ($config);
+
+	my $p = shift;
+
+	$p ||= '';
+
+	my $v;
+
+	warn "### getting config($p)\n" if ($debug > 1);
+
+	my @p = split(/\s+/,$p);
+	if ($#p < 0) {
+		$v = $config->{ '_' };	# special, database code
+	} else {
+
+		my $c = dclone( $config );
+
+		foreach my $k (@p) {
+			warn "### k: $k c = ",dump($c),$/ if ($debug > 1);
+			if (ref($c) eq 'ARRAY') {
+				$c = shift @$c;
+				warn "config($p) taking first occurence of '$k', probably not what you wanted!\n";
+				last;
+			}
+
+			if (! defined($c->{$k}) ) {
+				$c = undef;
+				last;
+			} else {
+				$c = $c->{$k};
+			}
+		}
+		$v = $c if ($c);
+
+	}
+
+	warn "## config( '$p' ) = ",dump( $v ),$/ if ($v && $debug);
+	warn "config( '$p' ) is empty\n" if (! $v);
+
+	return $v;
+}
+
+=head2 id
+
+Returns unique id of this record
+
+  $id = id();
+
+Returns C<42/2> for 2nd occurence of MFN 42.
+
+=cut
+
+sub id {
+	my $mfn = $config->{_mfn} || die "no _mfn in config data";
+	return $mfn . $#{$marc_record} ? $#{$marc_record} + 1 : '';
+}
+
 =head2 join_with
 
 Joins walues with some delimiter
@@ -769,7 +965,10 @@
 
 sub join_with {
 	my $d = shift;
-	return join($d, grep { defined($_) && $_ ne '' } @_);
+	warn "### join_with('$d',",dump(@_),")\n" if ($debug > 2);
+	my $v = join($d, grep { defined($_) && $_ ne '' } @_);
+	return '' unless defined($v);
+	return $v;
 }
 
 =head2 split_rec_on
@@ -797,7 +996,7 @@
 	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);