--- trunk/lib/WebPAC/Normalize.pm	2006/12/10 12:56:59	788
+++ trunk/lib/WebPAC/Normalize.pm	2007/11/08 17:19:50	1015
@@ -1,19 +1,22 @@
 package WebPAC::Normalize;
 use Exporter 'import';
-@EXPORT = qw/
-	_set_rec _set_lookup
+our @EXPORT = qw/
+	_set_ds _set_lookup
 	_set_load_row
 	_get_ds _clean_ds
 	_debug
 	_pack_subfields_hash
 
-	tag search display
+	search_display search display sorted
+
 	marc marc_indicators marc_repeatable_subfield
-	marc_compose marc_leader
-	marc_duplicate marc_remove
+	marc_compose marc_leader marc_fixed
+	marc_duplicate marc_remove marc_count
 	marc_original_order
+	marc_template
 
 	rec1 rec2 rec
+	frec frec_eq frec_ne
 	regex prefix suffix surround
 	first lookup join_with
 	save_into_lookup
@@ -21,6 +24,8 @@
 	split_rec_on
 
 	get set
+	count
+
 /;
 
 use warnings;
@@ -34,18 +39,16 @@
 # debugging warn(s)
 my $debug = 0;
 
+use WebPAC::Normalize::ISBN;
+push @EXPORT, ( 'isbn_10', 'isbn_13' );
 
 =head1 NAME
 
 WebPAC::Normalize - describe normalisaton rules using sets
 
-=head1 VERSION
-
-Version 0.25
-
 =cut
 
-our $VERSION = '0.25';
+our $VERSION = '0.34';
 
 =head1 SYNOPSIS
 
@@ -58,7 +61,7 @@
 C<perl -c normalize.pl>.
 
 Normalisation can generate multiple output normalized data. For now, supported output
-types (on the left side of definition) are: C<tag>, C<display>, C<search> and
+types (on the left side of definition) are: C<search_display>, C<display>, C<search> and
 C<marc>.
 
 =head1 FUNCTIONS
@@ -77,7 +80,7 @@
 	marc_encoding => 'utf-8',
 	config => $config,
 	load_row_coderef => sub {
-		my ($database,$input,$mfn) = shift;
+		my ($database,$input,$mfn) = @_;
 		$store->load_row( database => $database, input => $input, id => $mfn );
 	},
   );
@@ -103,30 +106,31 @@
 	die "need row argument" unless ($arg->{row});
 	die "need normalisation argument" unless ($arg->{rules});
 
-	no strict 'subs';
 	_set_lookup( $arg->{lookup} ) if defined($arg->{lookup});
-	_set_rec( $arg->{row} );
+	_set_ds( $arg->{row} );
 	_set_config( $arg->{config} ) if defined($arg->{config});
 	_clean_ds( %{ $arg } );
 	$load_row_coderef = $arg->{load_row_coderef};
 
-	eval "$arg->{rules}";
+	no strict 'subs';
+	no warnings 'redefine';
+	eval "$arg->{rules};";
 	die "error evaling $arg->{rules}: $@\n" if ($@);
 
 	return _get_ds();
 }
 
-=head2 _set_rec
+=head2 _set_ds
 
 Set current record hash
 
-  _set_rec( $rec );
+  _set_ds( $rec );
 
 =cut
 
 my $rec;
 
-sub _set_rec {
+sub _set_ds {
 	$rec = shift or die "no record hash";
 }
 
@@ -166,10 +170,11 @@
 
 =cut
 
-my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader);
+my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader);
 my ($marc_record_offset, $marc_fetch_offset) = (0, 0);
 
 sub _get_ds {
+#warn "## out = ",dump($out);
 	return $out;
 }
 
@@ -183,7 +188,7 @@
 
 sub _clean_ds {
 	my $a = {@_};
-	($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $leader) = ();
+	($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators, $marc_leader) = ();
 	($marc_record_offset, $marc_fetch_offset) = (0,0);
 	$marc_encoding = $a->{marc_encoding};
 }
@@ -288,13 +293,15 @@
 
 =cut
 
+my $fetch_pos;
+
 sub _get_marc_fields {
 
 	my $arg = {@_};
 	warn "### _get_marc_fields arg: ", dump($arg), $/ if ($debug > 2);
-	my $offset = $marc_fetch_offset;
+	$fetch_pos = $marc_fetch_offset;
 	if ($arg->{offset}) {
-		$offset = $arg->{offset};
+		$fetch_pos = $arg->{offset};
 	} elsif($arg->{fetch_next}) {
 		$marc_fetch_offset++;
 	}
@@ -303,9 +310,9 @@
 
 	warn "### full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 2);
 
-	my $marc_rec = $marc_record->[ $offset ];
+	my $marc_rec = $marc_record->[ $fetch_pos ];
 
-	warn "## _get_marc_fields (at offset: $offset) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1);
+	warn "## _get_marc_fields (at offset: $fetch_pos) -- marc_record = ", dump( @$marc_rec ), $/ if ($debug > 1);
 
 	return if (! $marc_rec || ref($marc_rec) ne 'ARRAY' || $#{ $marc_rec } < 0);
 
@@ -326,7 +333,7 @@
 
 	if ($debug) {
 		warn "## marc_repeatable_subfield = ", dump( $marc_repeatable_subfield ), $/ if ( $marc_repeatable_subfield );
-		warn "## marc_record[$offset] = ", dump( $marc_rec ), $/;
+		warn "## marc_record[$fetch_pos] = ", dump( $marc_rec ), $/;
 		warn "## sorted_marc_record = ", dump( \@sorted_marc_record ), $/;
 		warn "## subfield count = ", dump( $u ), $/;
 	}
@@ -407,6 +414,19 @@
 	return \@m;
 }
 
+=head2 _get_marc_leader
+
+Return leader from currently fetched record by L</_get_marc_fields>
+
+  print WebPAC::Normalize::_get_marc_leader();
+
+=cut
+
+sub _get_marc_leader {
+	die "no fetch_pos, did you called _get_marc_fields first?" unless ( defined( $fetch_pos ) );
+	return $marc_leader->[ $fetch_pos ];
+}
+
 =head2 _debug
 
 Change level of debug warnings
@@ -426,40 +446,51 @@
 
 Those functions generally have to first in your normalization file.
 
-=head2 tag
+=head2 search_display
 
-Define new tag for I<search> and I<display>.
+Define output for L<search> and L<display> at the same time
 
-  tag('Title', rec('200','a') );
+  search_display('Title', rec('200','a') );
 
 
 =cut
 
-sub tag {
-	my $name = shift or die "tag needs name as first argument";
+sub search_display {
+	my $name = shift or die "search_display needs name as first argument";
 	my @o = grep { defined($_) && $_ ne '' } @_;
 	return unless (@o);
-	$out->{$name}->{tag} = $name;
 	$out->{$name}->{search} = \@o;
 	$out->{$name}->{display} = \@o;
 }
 
+=head2 tag
+
+Old name for L<search_display>, but supported
+
+=cut
+
+sub tag {
+	search_display( @_ );
+}
+
 =head2 display
 
-Define tag just for I<display>
+Define output just for I<display>
 
   @v = display('Title', rec('200','a') );
 
 =cut
 
-sub display {
-	my $name = shift or die "display needs name as first argument";
+sub _field {
+	my $type = shift or confess "need type -- BUG?";
+	my $name = shift or confess "needs name as first argument";
 	my @o = grep { defined($_) && $_ ne '' } @_;
 	return unless (@o);
-	$out->{$name}->{tag} = $name;
-	$out->{$name}->{display} = \@o;
+	$out->{$name}->{$type} = \@o;
 }
 
+sub display { _field( 'display', @_ ) }
+
 =head2 search
 
 Prepare values just for I<search>
@@ -468,13 +499,18 @@
 
 =cut
 
-sub search {
-	my $name = shift or die "search needs name as first argument";
-	my @o = grep { defined($_) && $_ ne '' } @_;
-	return unless (@o);
-	$out->{$name}->{tag} = $name;
-	$out->{$name}->{search} = \@o;
-}
+sub search { _field( 'search', @_ ) }
+
+=head2 sorted
+
+Insert into lists which will be automatically sorted
+
+ sorted('Title', rec('200','a') );
+
+=cut
+
+sub sorted { _field( 'sorted', @_ ) }
+
 
 =head2 marc_leader
 
@@ -489,9 +525,58 @@
 	my ($offset,$value) = @_;
 
 	if ($offset) {
-		$leader->{ $offset } = $value;
+		$marc_leader->[ $marc_record_offset ]->{ $offset } = $value;
 	} else {
-		return $leader;
+		
+		if (defined($marc_leader)) {
+			die "marc_leader not array = ", dump( $marc_leader ) unless (ref($marc_leader) eq 'ARRAY');
+			return $marc_leader->[ $marc_record_offset ];
+		} else {
+			return;
+		}
+	}
+}
+
+=head2 marc_fixed
+
+Create control/indentifier fields with values in fixed positions
+
+  marc_fixed('008', 00, '070402');
+  marc_fixed('008', 39, '|');
+
+Positions not specified will be filled with spaces (C<0x20>).
+
+There will be no effort to extend last specified value to full length of
+field in standard.
+
+=cut
+
+sub marc_fixed {
+	my ($f, $pos, $val) = @_;
+	die "need marc(field, position, value)" unless defined($f) && defined($pos);
+
+	confess "need val" unless defined $val;
+
+	my $update = 0;
+
+	map {
+		if ($_->[0] eq $f) {
+			my $old = $_->[1];
+			if (length($old) <= $pos) {
+				$_->[1] .= ' ' x ( $pos - length($old) ) . $val;
+				warn "## marc_fixed($f,$pos,'$val') append '$old' -> '$_->[1]'\n" if ($debug > 1);
+			} else {
+				$_->[1] = substr($old, 0, $pos) . $val . substr($old, $pos + length($val));
+				warn "## marc_fixed($f,$pos,'$val') update '$old' -> '$_->[1]'\n" if ($debug > 1);
+			}
+			$update++;
+		}
+	} @{ $marc_record->[ $marc_record_offset ] };
+
+	if (! $update) {
+		my $v = ' ' x $pos . $val;
+		push @{ $marc_record->[ $marc_record_offset ] }, [ $f, $v ];
+		warn "## marc_fixed($f,$pos,'val') created '$v'\n" if ($debug > 1);
 	}
 }
 
@@ -623,9 +708,11 @@
 	 my $m = $marc_record->[ -1 ];
 	 die "can't duplicate record which isn't defined" unless ($m);
 	 push @{ $marc_record }, dclone( $m );
-	 warn "## marc_duplicate = ", dump(@$marc_record), $/ if ($debug > 1);
+	 push @{ $marc_leader }, dclone( marc_leader() );
+	 warn "## marc_duplicate = ", dump(@$marc_leader, @$marc_record), $/ if ($debug > 1);
 	 $marc_record_offset = $#{ $marc_record };
 	 warn "## marc_record_offset = $marc_record_offset", $/ if ($debug > 1);
+
 }
 
 =head2 marc_remove
@@ -773,6 +860,154 @@
 	warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1);
 }
 
+=head2 marc_template
+
+=cut
+
+sub marc_template {
+	my $args = {@_};
+	warn "## marc_template(",dump($args),")";
+
+	foreach ( qw/subfields_rename marc_template/ ) {
+#		warn "ref($_) = ",ref($args->{$_});
+		die "$_ not ARRAY" if ref($args->{$_}) ne 'ARRAY';
+	}
+
+	my $r = $rec->{ $args->{from} }; # || return;
+	die "record field ", $args->{from}, " isn't array\n" unless (ref($r) eq 'ARRAY');
+
+	my @subfields_rename = @{ $args->{subfields_rename} };
+#	warn "### subfields_rename [$#subfields_rename] = ",dump( @subfields_rename );
+
+	confess "need mapping in pairs for subfields_rename"
+		if $#subfields_rename % 2 != 1;
+	
+	my ( $subfields_rename, $from_subfields, $to_subfields );
+	while ( my ( $from, $to ) = splice(@subfields_rename, 0, 2) ) {
+		my ( $f, $t ) = (
+			$from_subfields->{ $from }++,
+			$to_subfields->{ $to }++
+		);
+		$subfields_rename->{ $from }->[ $f ] = [ $to => $t ];
+	}
+	warn "### subfields_rename = ",dump( $subfields_rename ),$/;
+	warn "### from_subfields = ", dump( $from_subfields ),$/;
+	warn "### to_subfields = ", dump( $to_subfields ),$/;
+
+	my $fields_re = join('|', keys %$to_subfields );
+
+	my $pos_templates;
+	my $count;
+	my @marc_order;
+	my $marc_template_order;
+	my $fill_in;
+	my @marc_out;
+
+	foreach my $template ( @{ $args->{marc_template} } ) {
+		$count = {};
+		@marc_order = ();
+		sub my_count {
+			my $sf = shift;
+			my $nr = $count->{$sf}++;
+			push @marc_order, [ $sf, $nr ];
+			return $sf . $nr;
+		}
+		my $pos_template = $template;
+		$pos_template =~ s/($fields_re)/my_count($1)/ge;
+		my $count_key = dump( $count );
+		warn "### template: |$template| -> |$pos_template| count = $count_key marc_order = ",dump( @marc_order ),$/;
+		$pos_templates->{ $count_key } = $pos_template;
+		$marc_template_order->{ $pos_template } = [ @marc_order ];
+	}
+	warn "### from ",dump( $args->{marc_template} ), " created ", dump( $pos_templates ), " and ", dump( $marc_template_order );
+
+	my $m;
+
+	foreach my $r ( @{ $rec->{ $args->{from} } } ) {
+
+		my $i1 = $r->{i1} || ' ';
+		my $i2 = $r->{i2} || ' ';
+		$m = [ $args->{to}, $i1, $i2 ];
+
+		warn "### r = ",dump( $r );
+
+		my ( $new_r, $from_count, $to_count );
+		foreach my $sf ( keys %{$r} ) {
+			my $nr = $from_count->{$sf}++;
+			my $rename_to = $subfields_rename->{ $sf };	# ||
+#				die "can't find subfield rename for $sf/$nr in ", dump( $subfields_rename );
+			warn "### rename $sf/$nr to ", dump( $rename_to->[$nr] ), $/;
+			my ( $to_sf, $to_nr ) = @{ $rename_to->[$nr] };
+			$new_r->{ $to_sf }->[ $to_nr ] = [ $sf => $nr ];
+
+			$to_count->{ $to_sf }++;
+		}
+
+		warn "### new_r = ",dump( $new_r );
+
+		my $from_count_key = dump( $to_count );
+
+		warn "### from_count = ",dump( $from_count ), $/;
+		warn "### to_count   = ",dump( $to_count ), $/;
+
+		my $template = $pos_templates->{ $from_count_key } ||
+			die "I don't have template for:\n$from_count_key\n## available templates\n", dump( $pos_templates );
+
+		warn "### selected template: |$template|\n";
+
+		$fill_in = {};
+
+		foreach my $sf ( split(/\|/, $template ) ) {
+			sub fill_in {
+				my ( $r, $sf, $nr ) = @_;
+				my ( $from_sf, $from_nr ) = @{ $new_r->{ $sf }->[ $nr ] };
+				my $v = $r->{ $from_sf }; # || die "no $from_sf/$from_nr";
+				warn "#### fill_in( $sf, $nr ) = $from_sf/$from_nr >>>> ",dump( $v ), $/;
+				if ( ref( $v ) eq 'ARRAY' ) {
+					$fill_in->{$sf}->[$nr] = $v->[$from_nr];
+					return $v->[$from_nr];
+				} elsif ( $from_nr == 0 ) {
+					$fill_in->{$sf}->[$nr] = $v;
+					return $v;
+				} else {
+					die "requested subfield $from_sf/$from_nr but it's ",dump( $v );
+				}
+			}
+			warn "#### $sf <<<< $fields_re\n";
+			$sf =~ s/($fields_re)(\d+)/fill_in($r,$1,$2)/ge;
+			warn "#### >>>> $sf with fill_in = ",dump( $fill_in ),$/;
+		}
+
+		warn "## template: |$template|\n## marc_template_order = ",dump( $marc_template_order );
+
+		foreach my $sf ( @{ $marc_template_order->{$template} } ) {
+			my ( $sf, $nr ) = @$sf;
+			my $v = $fill_in->{$sf}->[$nr] || die "can't find fill_in $sf/$nr";
+			warn "++ $sf/$nr |$v|\n";
+			push @$m, [ $sf, $v ];
+		}
+
+		warn "#### >>>> created marc: ", dump( $m );
+
+		push @marc_out, $m;
+	}
+
+	warn "### marc_template produced: ",dump( @marc_out );
+	return @marc_out;
+}
+
+=head2 marc_count
+
+Return number of MARC records created using L</marc_duplicate>.
+
+  print "created ", marc_count(), " records";
+
+=cut
+
+sub marc_count {
+	return $#{ $marc_record };
+}
+
 
 =head1 Functions to extract data from input
 
@@ -795,6 +1030,9 @@
 
 	my ($h,$include_subfields) = @_;
 
+	# sanity and ease of use
+	return $h if (ref($h) ne 'HASH');
+
 	if ( defined($h->{subfields}) ) {
 		my $sfs = delete $h->{subfields} || die "no subfields?";
 		my @out;
@@ -915,6 +1153,54 @@
 	}
 }
 
+=head2 frec
+
+Returns first value from field
+
+  $v = frec('200');
+  $v = frec('200','a');
+
+=cut
+
+sub frec {
+	my @out = rec(@_);
+	warn "rec(",dump(@_),") has more than one return value, ignoring\n" if $#out > 0;
+	return shift @out;
+}
+
+=head2 frec_eq
+
+=head2 frec_ne
+
+Check if first values from two fields are same or different
+
+  if ( frec_eq( 900 => 'a', 910 => 'c' ) ) {
+  	# values are same
+  } else {
+    # values are different
+  }
+
+Strictly speaking C<frec_eq> and C<frec_ne> wouldn't be needed if you
+could write something like:
+
+  if ( frec( '900','a' ) eq frec( '910','c' ) ) {
+  	# yada tada
+  }
+
+but you can't since our parser L<WebPAC::Parser> will remove all whitespaces
+in order to parse text and create invalid function C<eqfrec>.
+
+=cut
+
+sub frec_eq {
+	my ( $f1,$sf1, $f2, $sf2 ) = @_;
+	return (rec( $f1, $sf1 ))[0] eq (rec( $f2, $sf2 ))[0];
+}
+
+sub frec_ne {
+	return ! frec_eq( @_ );
+}
+
 =head2 regex
 
 Apply regex to some or all values
@@ -944,7 +1230,8 @@
 =cut
 
 sub prefix {
-	my $p = shift or return;
+	my $p = shift;
+	return @_ unless defined( $p );
 	return map { $p . $_ } grep { defined($_) } @_;
 }
 
@@ -957,7 +1244,8 @@
 =cut
 
 sub suffix {
-	my $s = shift or die "suffix needs string as first argument";
+	my $s = shift;
+	return @_ unless defined( $s );
 	return map { $_ . $s } grep { defined($_) } @_;
 }
 
@@ -970,8 +1258,10 @@
 =cut
 
 sub surround {
-	my $p = shift or die "surround need prefix as first argument";
-	my $s = shift or die "surround needs suffix as second argument";
+	my $p = shift;
+	my $s = shift;
+	$p = '' unless defined( $p );
+	$s = '' unless defined( $s );
 	return map { $p . $_ . $s } grep { defined($_) } @_;
 }
 
@@ -1146,7 +1436,6 @@
   $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.
 
@@ -1272,7 +1561,7 @@
 
 sub set {
 	my ($k,$v) = @_;
-	warn "## set ( $k => ", dump($v), " )", $/;
+	warn "## set ( $k => ", dump($v), " )", $/ if ( $debug );
 	$hash->{$k} = $v;
 };
 
@@ -1285,10 +1574,22 @@
 sub get {
 	my $k = shift || return;
 	my $v = $hash->{$k};
-	warn "## get $k = ", dump( $v ), $/;
+	warn "## get $k = ", dump( $v ), $/ if ( $debug );
 	return $v;
 }
 
+=head2 count
+
+  if ( count( @result ) == 1 ) {
+  	# do something if only 1 result is there
+  }
+
+=cut
+
+sub count {
+	warn "## count ",dump(@_),$/ if ( $debug );
+	return @_ . '';
+}
 
 # END
 1;