--- trunk/lib/WebPAC/Normalize.pm	2005/11/24 11:47:15	125
+++ trunk/lib/WebPAC/Normalize.pm	2006/01/08 20:27:11	364
@@ -11,11 +11,11 @@
 
 =head1 VERSION
 
-Version 0.02
+Version 0.08
 
 =cut
 
-our $VERSION = '0.02';
+our $VERSION = '0.08';
 
 =head1 SYNOPSIS
 
@@ -47,6 +47,10 @@
 code defined as code ref on format after field substitution to producing
 output
 
+There is one built-in filter called C<regex> which can be use like this:
+
+  filter{regex(s/foo/bar/)}
+
 =item *
 
 optional C<lookup{...}> will be then performed. See C<WebPAC::Lookups>.
@@ -82,11 +86,15 @@
 	db => $db_obj,
 	lookup_regex => $lookup->regex,
 	lookup => $lookup_obj,
+	prefix => 'foobar',
   );
 
 Parametar C<filter> defines user supplied snippets of perl code which can
 be use with C<filter{...}> notation.
 
+C<prefix> is used to form filename for database record (to support multiple
+source files which are joined in one database).
+
 Recommended parametar C<lookup_regex> is used to enable parsing of lookups
 in structures. If you pass this parametar, you must also pass C<lookup>
 which is C<WebPAC::Lookup> object.
@@ -111,6 +119,19 @@
 
 	$log->logdie("lookup must be WebPAC::Lookup object") if ($self->{'lookup'} && ! $self->{'lookup'}->isa('WebPAC::Lookup'));
 
+	$log->warn("no prefix defined. please check that!") unless ($self->{'prefix'});
+
+	$log->debug("using lookup regex: ", $self->{lookup_regex}) if ($r && $l);
+
+	if (! $self->{filter} || ! $self->{filter}->{regex}) {
+		$log->debug("adding built-in filter regex");
+		$self->{filter}->{regex} = sub {
+			my ($val, $regex) = @_;
+			eval "\$val =~ $regex";
+			return $val;
+		};
+	}
+
 	$self ? return $self : return undef;
 }
 
@@ -136,22 +157,19 @@
 
 	$log->debug("data_structure rec = ", sub { Dumper($rec) });
 
-	$log->logdie("need unique ID (mfn) in field 000 of record ", sub { Dumper($rec) } ) unless (defined($rec->{'000'}));
+	$log->logdie("need unique ID (mfn) in field 000 of record " . Dumper($rec) ) unless (defined($rec->{'000'}));
 
-	my $mfn = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!");
+	my $id = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!");
 
 	my $cache_file;
 
 	if ($self->{'db'}) {
-		my $ds = $self->{'db'}->load_ds( $mfn );
+		my $ds = $self->{'db'}->load_ds( id => $id, prefix => $self->{prefix} );
 		$log->debug("load_ds( rec = ", sub { Dumper($rec) }, ") = ", sub { Dumper($ds) });
 		return $ds if ($ds);
 		$log->debug("cache miss, creating");
 	}
 
-	undef $self->{'currnet_filename'};
-	undef $self->{'headline'};
-
 	my @sorted_tags;
 	if ($self->{tags_by_order}) {
 		@sorted_tags = @{$self->{tags_by_order}};
@@ -176,15 +194,18 @@
 			$log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH');
 			$format = $tag->{'value'} || $tag->{'content'};
 
-			$log->debug("format: $format");
-
 			my @v;
 			if ($self->{'lookup_regex'} && $format =~ $self->{'lookup_regex'}) {
 				@v = $self->fill_in_to_arr($rec,$format);
 			} else {
 				@v = $self->parse_to_arr($rec,$format);
 			}
-			next if (! @v);
+			if (! @v) {
+				$log->debug("$field <",$self->{tag},"> format: $format no values");
+#				next;
+			} else {
+				$log->debug("$field <",$self->{tag},"> format: $format values: ", join(",", @v));
+			}
 
 			if ($tag->{'sort'}) {
 				@v = $self->sort_arr(@v);
@@ -207,7 +228,7 @@
 
 			foreach my $type (@types) {
 				# append to previous line?
-				$log->debug("type: $type ",sub { join(" ",@v) }, $row->{'append'} || 'no append');
+				$log->debug("tag $field / $type [",sub { join(",",@v) }, "] ", $row->{'append'} || 'no append');
 				if ($tag->{'append'}) {
 
 					# I will delimit appended part with
@@ -249,8 +270,9 @@
 	}
 
 	$self->{'db'}->save_ds(
-		id => $mfn,
+		id => $id,
 		ds => $ds,
+		prefix => $self->{prefix},
 	) if ($self->{'db'});
 
 	$log->debug("ds: ", sub { Dumper($ds) });
@@ -269,8 +291,18 @@
 
  my $text = $webpac->parse($rec,'eval{"v901^a" eq "Deskriptor"}descriptor: v250^a', $i);
 
+Filters are implemented here. While simple form of filters looks like this:
+
+  filter{name_of_filter}
+
+but, filters can also have variable number of parametars like this:
+
+  filter{name_of_filter(param,param,param)}
+
 =cut
 
+my $warn_once;
+
 sub parse {
 	my $self = shift;
 
@@ -288,7 +320,7 @@
 
 	my @out;
 
-	$log->debug("format: $format");
+	$log->debug("format: $format [$i]");
 
 	my $eval_code;
 	# remove eval{...} from beginning
@@ -298,29 +330,45 @@
 	# remove filter{...} from beginning
 	$filter_name = $1 if ($format =~ s/^filter{([^}]+)}//s);
 
+	# did we found any (att all) field from format in row?
+	my $found_any;
+	# prefix before first field which we preserve it $found_any
 	my $prefix;
-	my $all_found=0;
+
+	my $f_step = 1;
 
 	while ($format =~ s/^(.*?)(v|s)(\d+)(?:\^(\w))?//s) {
 
 		my $del = $1 || '';
-		$prefix ||= $del if ($all_found == 0);
+		$prefix = $del if ($f_step == 1);
+
+		my $fld_type = lc($2);
 
 		# repeatable index
 		my $r = $i;
-		$r = 0 if (lc("$2") eq 's');
+		if ($fld_type eq 's') {
+			if ($found_any->{'v'}) {
+				$r = 0;
+			} else {
+				return;
+			}
+		}
 
 		my $found = 0;
 		my $tmp = $self->get_data(\$rec,$3,$4,$r,\$found);
 
 		if ($found) {
-			push @out, $del;
+			$found_any->{$fld_type} += $found;
+
+			# we will skip delimiter before first occurence of field!
+			push @out, $del unless($found_any->{$fld_type} == 1);
 			push @out, $tmp;
-			$all_found += $found;
 		}
+		$f_step++;
 	}
 
-	return if (! $all_found);
+	# test if any fields found?
+	return if (! $found_any->{'v'} && ! $found_any->{'s'});
 
 	my $out = join('',@out);
 
@@ -340,11 +388,21 @@
 		return if (! $self->_eval($eval));
 	}
 	
-	if ($filter_name && $self->{'filter'}->{$filter_name}) {
-		$log->debug("about to filter{$filter_name} format: $out");
-		$out = $self->{'filter'}->{$filter_name}->($out);
-		return unless(defined($out));
-		$log->debug("filter result: $out");
+	if ($filter_name) {
+		my @filter_args;
+		if ($filter_name =~ s/(\w+)\((.*)\)/$1/) {
+			@filter_args = split(/,/, $2);
+		}
+		if ($self->{'filter'}->{$filter_name}) {
+			$log->debug("about to filter{$filter_name} format: $out with arguments: ", join(",", @filter_args));
+			unshift @filter_args, $out;
+			$out = $self->{'filter'}->{$filter_name}->(@filter_args);
+			return unless(defined($out));
+			$log->debug("filter result: $out");
+		} elsif (! $warn_once->{$filter_name}) {
+			$log->warn("trying to use undefined filter $filter_name");
+			$warn_once->{$filter_name}++;
+		}
 	}
 
 	return $out;
@@ -424,6 +482,7 @@
 	}
 
 	my $found = 0;
+	my $just_single = 1;
 
 	my $eval_code;
 	# remove eval{...} from beginning
@@ -435,9 +494,14 @@
 
 	# do actual replacement of placeholders
 	# repeatable fields
-	$format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found)/ges;
+	if ($format =~ s/v(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,$i,\$found)/ges) {
+		$just_single = 0;
+	}
+
 	# non-repeatable fields
-	$format =~ s/s(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,0,\$found)/ges;
+	if ($format =~ s/s(\d+)(?:\^(\w))?/$self->get_data(\$rec,$1,$2,0,\$found)/ges) {
+		return if ($i > 0 && $just_single);
+	}
 
 	if ($found) {
 		$log->debug("format: $format");
@@ -454,7 +518,9 @@
 		# do we have lookups?
 		if ($self->{'lookup'}) {
 			if ($self->{'lookup'}->can('lookup')) {
-				return $self->{'lookup'}->lookup($format);
+				my @lookup = $self->{lookup}->lookup($format);
+				$log->debug("lookup $format", join(", ", @lookup));
+				return @lookup;
 			} else {
 				$log->warn("Have lookup object but can't invoke lookup method");
 			}
@@ -490,8 +556,8 @@
 	my $i = 0;
 	my @arr;
 
-	while (my @v = $self->fill_in($rec,$format_utf8,$i++)) {
-		push @arr, @v;
+	while (my $v = $self->fill_in($rec,$format_utf8,$i++)) {
+		push @arr, $v;
 	}
 
 	$log->debug("format '$format_utf8' returned ",--$i," elements: ", sub { join(" | ",@arr) }) if (@arr);
@@ -537,7 +603,8 @@
 			if ($$rec->{$f}->[$i] =~ /HASH/o) {
 				my $out;
 				foreach my $k (keys %{$$rec->{$f}->[$i]}) {
-					$out .= $$rec->{$f}->[$i]->{$k}." ";
+					my $v = $$rec->{$f}->[$i]->{$k};
+					$out .= "$v " if ($v);
 				}
 				return $out;
 			} else {
@@ -668,4 +735,4 @@
 
 =cut
 
-1; # End of WebPAC::DB
+1; # End of WebPAC::Normalize