--- trunk/lib/WebPAC/Normalize.pm	2006/09/29 20:18:30	729
+++ trunk/lib/WebPAC/Normalize.pm	2007/11/07 09:19:29	1012
@@ -1,24 +1,30 @@
 package WebPAC::Normalize;
 use Exporter 'import';
-@EXPORT = qw/
-	_set_rec _set_lookup
-	_set_load_ds
+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
 
 	rec1 rec2 rec
+	frec
 	regex prefix suffix surround
 	first lookup join_with
 	save_into_lookup
 
 	split_rec_on
+
+	get set
+	count
+
 /;
 
 use warnings;
@@ -32,18 +38,16 @@
 # 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 0.21
-
 =cut
 
-our $VERSION = '0.21';
+our $VERSION = '0.32';
 
 =head1 SYNOPSIS
 
@@ -56,7 +60,7 @@
 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
@@ -74,16 +78,16 @@
 	rules => $normalize_pl_config,
 	marc_encoding => 'utf-8',
 	config => $config,
-	load_ds_coderef => sub {
-		my ($database,$input,$mfn) = shift;
-		$store->load_ds( database => $database, input => $input, id => $mfn );
+	load_row_coderef => sub {
+		my ($database,$input,$mfn) = @_;
+		$store->load_row( database => $database, input => $input, id => $mfn );
 	},
   );
 
 Options C<row>, C<rules> and C<log> are mandatory while all
 other are optional.
 
-C<load_ds_coderef> is closure only used when executing lookups, so they will
+C<load_row_coderef> is closure only used when executing lookups, so they will
 die if it's not defined.
 
 This function will B<die> if normalizastion can't be evaled.
@@ -93,7 +97,7 @@
 
 =cut
 
-my $load_ds_coderef;
+my $load_row_coderef;
 
 sub data_structure {
 	my $arg = {@_};
@@ -101,30 +105,31 @@
 	die "need row argument" unless ($arg->{row});
 	die "need normalisation argument" unless ($arg->{rules});
 
-	no strict 'subs';
-	_set_lookup( $arg->{lookup} );
-	_set_rec( $arg->{row} );
-	_set_config( $arg->{config} );
+	_set_lookup( $arg->{lookup} ) if defined($arg->{lookup});
+	_set_ds( $arg->{row} );
+	_set_config( $arg->{config} ) if defined($arg->{config});
 	_clean_ds( %{ $arg } );
-	$load_ds_coderef = $arg->{load_ds_coderef};
+	$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";
 }
 
@@ -164,10 +169,11 @@
 
 =cut
 
-my ($out, $marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators);
+my ($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;
 }
 
@@ -181,7 +187,7 @@
 
 sub _clean_ds {
 	my $a = {@_};
-	($out,$marc_record, $marc_encoding, $marc_repeatable_subfield, $marc_indicators) = ();
+	($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};
 }
@@ -212,23 +218,23 @@
 	return $lookup;
 }
 
-=head2 _set_load_ds
+=head2 _set_load_row
 
 Setup code reference which will return L<data_structure> from
 L<WebPAC::Store>
 
-  _set_load_ds(sub {
+  _set_load_row(sub {
 		my ($database,$input,$mfn) = @_;
-		$store->load_ds( database => $database, input => $input, id => $mfn );
+		$store->load_row( database => $database, input => $input, id => $mfn );
   });
 
 =cut
 
-sub _set_load_ds {
+sub _set_load_row {
 	my $coderef = shift;
 	confess "argument isn't CODE" unless ref($coderef) eq 'CODE';
 
-	$load_ds_coderef = $coderef;
+	$load_row_coderef = $coderef;
 }
 
 =head2 _get_marc_fields
@@ -286,13 +292,15 @@
 
 =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++;
 	}
@@ -301,9 +309,9 @@
 
 	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);
 
@@ -324,7 +332,7 @@
 
 	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 ), $/;
 	}
@@ -405,6 +413,19 @@
 	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
@@ -424,40 +445,51 @@
 
 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>
@@ -466,13 +498,18 @@
 
 =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
 
@@ -487,9 +524,58 @@
 	my ($offset,$value) = @_;
 
 	if ($offset) {
-		$out->{' leader'}->{ $offset } = $value;
+		$marc_leader->[ $marc_record_offset ]->{ $offset } = $value;
 	} else {
-		return $out->{' leader'};
+		
+		if (defined($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);
 	}
 }
 
@@ -621,9 +707,11 @@
 	 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
@@ -635,6 +723,10 @@
 
 This will erase field C<200> or C<200^a> from current MARC record.
 
+  marc_remove('*');
+
+Will remove all fields in current MARC record.
+
 This is useful after calling C<marc_duplicate> or on it's own (but, you
 should probably just remove that subfield definition if you are not
 using C<marc_duplicate>).
@@ -652,39 +744,47 @@
 
 	warn "### marc_remove before = ", dump( $marc ), $/ if ($debug > 2);
 
-	my $i = 0;
-	foreach ( 0 .. $#{ $marc } ) {
-		last unless (defined $marc->[$i]);
-		warn "#### working on ",dump( @{ $marc->[$i] }), $/ if ($debug > 3);
-		if ($marc->[$i]->[0] eq $f) {
-			if (! defined $sf) {
-				# remove whole field
-				splice @$marc, $i, 1;
-				warn "#### slice \@\$marc, $i, 1 = ",dump( @{ $marc }), $/ if ($debug > 3);
-				$i--;
-			} else {
-				foreach my $j ( 0 .. (( $#{ $marc->[$i] } - 3 ) / 2) ) {
-					my $o = ($j * 2) + 3;
-					if ($marc->[$i]->[$o] eq $sf) {
-						# remove subfield
-						splice @{$marc->[$i]}, $o, 2;
-						warn "#### slice \@{\$marc->[$i]}, $o, 2 = ", dump( @{ $marc }), $/ if ($debug > 3);
-						# is record now empty?
-						if ($#{ $marc->[$i] } == 2) {
-							splice @$marc, $i, 1;
-							warn "#### slice \@\$marc, $i, 1 = ", dump( @{ $marc }), $/ if ($debug > 3);
-							$i--;
-						};
+	if ($f eq '*') {
+
+		delete( $marc_record->[ $marc_record_offset ] );
+		warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
+
+	} else {
+
+		my $i = 0;
+		foreach ( 0 .. $#{ $marc } ) {
+			last unless (defined $marc->[$i]);
+			warn "#### working on ",dump( @{ $marc->[$i] }), $/ if ($debug > 3);
+			if ($marc->[$i]->[0] eq $f) {
+				if (! defined $sf) {
+					# remove whole field
+					splice @$marc, $i, 1;
+					warn "#### slice \@\$marc, $i, 1 = ",dump( @{ $marc }), $/ if ($debug > 3);
+					$i--;
+				} else {
+					foreach my $j ( 0 .. (( $#{ $marc->[$i] } - 3 ) / 2) ) {
+						my $o = ($j * 2) + 3;
+						if ($marc->[$i]->[$o] eq $sf) {
+							# remove subfield
+							splice @{$marc->[$i]}, $o, 2;
+							warn "#### slice \@{\$marc->[$i]}, $o, 2 = ", dump( @{ $marc }), $/ if ($debug > 3);
+							# is record now empty?
+							if ($#{ $marc->[$i] } == 2) {
+								splice @$marc, $i, 1;
+								warn "#### slice \@\$marc, $i, 1 = ", dump( @{ $marc }), $/ if ($debug > 3);
+								$i--;
+							};
+						}
 					}
 				}
 			}
+			$i++;
 		}
-		$i++;
-	}
 
-	warn "### marc_remove($f", $sf ? ",$sf" : "", ") after = ", dump( $marc ), $/ if ($debug > 2);
+		warn "### marc_remove($f", $sf ? ",$sf" : "", ") after = ", dump( $marc ), $/ if ($debug > 2);
 
-	$marc_record->[ $marc_record_offset ] = $marc;
+		$marc_record->[ $marc_record_offset ] = $marc;
+	}
 
 	warn "## full marc_record = ", dump( @{ $marc_record }), $/ if ($debug > 1);
 }
@@ -759,6 +859,18 @@
 	warn "## marc_record = ", dump( $marc_record ),$/ if ($debug > 1);
 }
 
+=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
 
@@ -781,6 +893,9 @@
 
 	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;
@@ -880,6 +995,9 @@
   @v = rec('200')
   @v = rec('200','a')
 
+If rec() returns just single value, it will
+return scalar, not array.
+
 =cut
 
 sub rec {
@@ -889,13 +1007,30 @@
 	} elsif ($#_ == 1) {
 		@out = rec2(@_);
 	}
-	if (@out) {
+	if ($#out == 0 && ! wantarray) {
+		return $out[0];
+	} elsif (@out) {
 		return @out;
 	} else {
 		return '';
 	}
 }
 
+=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 regex
 
 Apply regex to some or all values
@@ -925,7 +1060,8 @@
 =cut
 
 sub prefix {
-	my $p = shift or return;
+	my $p = shift;
+	return @_ unless defined( $p );
 	return map { $p . $_ } grep { defined($_) } @_;
 }
 
@@ -938,7 +1074,8 @@
 =cut
 
 sub suffix {
-	my $s = shift or die "suffix needs string as first argument";
+	my $s = shift;
+	return @_ unless defined( $s );
 	return map { $_ . $s } grep { defined($_) } @_;
 }
 
@@ -951,8 +1088,10 @@
 =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 . $_ . $s } grep { defined($_) } @_;
 }
 
@@ -1017,12 +1156,12 @@
 sub lookup {
 	my ($what, $database, $input, $key, $having) = @_;
 
-	confess "lookup needs 5 arguments: what, database, input, key, having" unless ($#_ == 4);
+	confess "lookup needs 5 arguments: what, database, input, key, having\n" unless ($#_ == 4);
 
 	warn "## lookup ($database, $input, $key)", $/ if ($debug > 1);
 	return unless (defined($lookup->{$database}->{$input}->{$key}));
 
-	confess "lookup really need load_ds_coderef added to data_structure\n" unless ($load_ds_coderef);
+	confess "lookup really need load_row_coderef added to data_structure\n" unless ($load_row_coderef);
 
 	my $mfns;
 	my @having = $having->();
@@ -1031,7 +1170,7 @@
 
 	foreach my $h ( @having ) {
 		if (defined($lookup->{$database}->{$input}->{$key}->{$h})) {
-			warn "lookup for $database/$input/$key/$h return ",dump($lookup->{$database}->{$input}->{$key}->{$h}),"\n";
+			warn "lookup for $database/$input/$key/$h return ",dump($lookup->{$database}->{$input}->{$key}->{$h}),"\n" if ($debug);
 			$mfns->{$_}++ foreach keys %{ $lookup->{$database}->{$input}->{$key}->{$h} };
 		}
 	}
@@ -1040,21 +1179,21 @@
 
 	my @mfns = sort keys %$mfns;
 
-	warn "# lookup loading $database/$input/$key mfn ", join(",",@mfns)," having ",dump(@having),"\n";
+	warn "# lookup loading $database/$input/$key mfn ", join(",",@mfns)," having ",dump(@having),"\n" if ($debug);
 
 	my $old_rec = $rec;
 	my @out;
 
 	foreach my $mfn (@mfns) {
-		$rec = $load_ds_coderef->( $database, $input, $mfn );
+		$rec = $load_row_coderef->( $database, $input, $mfn );
 
-		warn "got $database/$input/$mfn = ", dump($rec), $/;
+		warn "got $database/$input/$mfn = ", dump($rec), $/ if ($debug);
 
 		my @vals = $what->();
 
 		push @out, ( @vals );
 
-		warn "lookup for mfn $mfn returned ", dump(@vals), $/;
+		warn "lookup for mfn $mfn returned ", dump(@vals), $/ if ($debug);
 	}
 
 #	if (ref($lookup->{$k}) eq 'ARRAY') {
@@ -1065,9 +1204,13 @@
 
 	$rec = $old_rec;
 
-	warn "## lookup returns = ", dump(@out), $/;
+	warn "## lookup returns = ", dump(@out), $/ if ($debug);
 
-	return @out;
+	if ($#out == 0) {
+		return $out[0];
+	} else {
+		return @out;
+	}
 }
 
 =head2 save_into_lookup
@@ -1123,7 +1266,6 @@
   $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.
 
@@ -1239,5 +1381,45 @@
 	}
 }
 
+my $hash;
+
+=head2 set
+
+  set( key => 'value' );
+
+=cut
+
+sub set {
+	my ($k,$v) = @_;
+	warn "## set ( $k => ", dump($v), " )", $/ if ( $debug );
+	$hash->{$k} = $v;
+};
+
+=head2 get
+
+  get( 'key' );
+
+=cut
+
+sub get {
+	my $k = shift || return;
+	my $v = $hash->{$k};
+	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;