--- trunk/lib/WebPAC/Normalize.pm	2006/07/02 14:41:01	561
+++ trunk/lib/WebPAC/Normalize.pm	2006/07/02 16:14:41	562
@@ -7,10 +7,13 @@
 
 	tag search display
 	marc marc_indicators marc_repeatable_subfield
+	marc_compose
 
 	rec1 rec2 rec
 	regex prefix suffix surround
 	first lookup join_with
+
+	split_rec_on
 /;
 
 use warnings;
@@ -30,11 +33,11 @@
 
 =head1 VERSION
 
-Version 0.07
+Version 0.08
 
 =cut
 
-our $VERSION = '0.07';
+our $VERSION = '0.08';
 
 =head1 SYNOPSIS
 
@@ -193,6 +196,8 @@
 		$a->[0] . $a->[3] cmp $b->[0] . $b->[3]	
 	} @{ $marc_record };
 
+	@sorted_marc_record = @{ $marc_record };	### FIXME disable sorting
+	
 	# output marc fields
 	my @m;
 
@@ -294,6 +299,7 @@
 sub _debug {
 	my $l = shift;
 	return $debug unless defined($l);
+	warn "debug level $l" if ($l > 0);
 	$debug = $l;
 }
 
@@ -410,6 +416,41 @@
 	@{ $marc_indicators->{$f} } = ($i1,$i2);
 }
 
+=head2 marc_compose
+
+Save values for each MARC subfield explicitly
+
+  marc_compose('900',
+  	'a', rec('200','a')
+  	'b', rec('201','a')
+  	'a', rec('200','b')
+  	'c', rec('200','c')
+  );
+
+=cut
+
+sub marc_compose {
+	my $f = shift or die "marc_compose needs field";
+	die "marc_compose field must be numer" unless ($f =~ /^\d+$/);
+
+	my ($i1,$i2) = defined($marc_indicators->{$f}) ? @{ $marc_indicators->{$f} } : (' ',' ');
+	my $m = [ $f, $i1, $i2 ];
+
+	while (@_) {
+		my $sf = shift or die "marc_compose $f needs subfield";
+		my $v = shift or die "marc_compose $f needs value for subfield $sf";
+
+		next unless (defined($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);
+	}
+
+	warn "## marc_compose(d) ", dump( $m ) if ($debug > 1);
+
+	push @{ $marc_record }, $m;
+}
+
 
 =head1 Functions to extract data from input
 
@@ -579,5 +620,39 @@
 	return join($d, grep { defined($_) && $_ ne '' } @_);
 }
 
+=head2 split_rec_on
+
+Split record subfield on some regex and take one of parts out
+
+  $a_before_semi_column =
+  	split_rec_on('200','a', /\s*;\s*/, $part);
+
+C<$part> is optional number of element. First element is
+B<1>, not 0!
+
+If there is no C<$part> parameter or C<$part> is 0, this function will
+return all values produced by splitting.
+
+=cut
+
+sub split_rec_on {
+	die "split_rec_on need (fld,sf,regex[,part]" if ($#_ < 2);
+
+	my ($fld, $sf, $regex, $part) = @_;
+	warn "### regex ", ref($regex), $regex if ($debug > 2);
+
+	my @r = rec( $fld, $sf );
+	my $v = shift @r;
+	warn "### first rec($fld,$sf) = ",dump($v) if ($debug > 2);
+
+	my @s = split( $regex, $v );
+	warn "## split_rec_on($fld,$sf,$regex,$part) = ",dump(@s) if ($debug > 1);
+	if ($part > 0) {
+		return $s[ $part - 1 ];
+	} else {
+		return @s;
+	}
+}
+
 # END
 1;