--- trunk/lib/WebPAC/Normalize.pm	2005/07/24 15:03:11	31
+++ trunk/lib/WebPAC/Normalize.pm	2006/01/07 03:28:10	346
@@ -11,11 +11,11 @@
 
 =head1 VERSION
 
-Version 0.01
+Version 0.08
 
 =cut
 
-our $VERSION = '0.01';
+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;
 }
 
@@ -122,13 +143,7 @@
 
 This structures are used to produce output.
 
- my @ds = $webpac->data_structure($rec);
-
-B<Note: historical oddity follows>
-
-This method will also set C<< $webpac->{'currnet_filename'} >> if there is
-C<< <filename> >> tag and C<< $webpac->{'headline'} >> if there is
-C<< <headline> >> tag.
+ my $ds = $webpac->data_structure($rec);
 
 =cut
 
@@ -140,18 +155,21 @@
 	my $rec = shift;
 	$log->logconfess("need HASH as first argument!") if ($rec !~ /HASH/o);
 
+	$log->debug("data_structure rec = ", sub { Dumper($rec) });
+
+	$log->logdie("need unique ID (mfn) in field 000 of record " . Dumper($rec) ) unless (defined($rec->{'000'}));
+
+	my $id = $rec->{'000'}->[0] || $log->logdie("field 000 isn't array!");
+
 	my $cache_file;
 
 	if ($self->{'db'}) {
-		my @ds = $self->{'db'}->load_ds($rec);
-		$log->debug("load_ds( rec = ", sub { Dumper($rec) }, ") = ", sub { Dumper(@ds) });
-		return @ds if ($#ds > 0);
+		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}};
@@ -160,7 +178,7 @@
 		$self->{tags_by_order} = \@sorted_tags;
 	}
 
-	my @ds;
+	my $ds;
 
 	$log->debug("tags: ",sub { join(", ",@sorted_tags) });
 
@@ -171,7 +189,10 @@
 #print "field $field [",$self->{'tag'},"] = ",Dumper($self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}});
 
 		foreach my $tag (@{$self->{'import_xml'}->{'indexer'}->{$field}->{$self->{'tag'}}}) {
-			my $format = $tag->{'value'} || $tag->{'content'};
+			my $format;
+
+			$log->logdie("expected tag HASH and got $tag") unless (ref($tag) eq 'HASH');
+			$format = $tag->{'value'} || $tag->{'content'};
 
 			$log->debug("format: $format");
 
@@ -192,28 +213,19 @@
 				@v = map { $self->apply_format($tag->{'format_name'},$tag->{'format_delimiter'},$_) } @v;
 			}
 
-			if ($field eq 'filename') {
-				$self->{'current_filename'} = join('',@v);
-				$log->debug("filename: ",$self->{'current_filename'});
-			} elsif ($field eq 'headline') {
-				$self->{'headline'} .= join('',@v);
-				$log->debug("headline: ",$self->{'headline'});
-				next; # don't return headline in data_structure!
-			}
-
 			# delimiter will join repeatable fields
 			if ($tag->{'delimiter'}) {
 				@v = ( join($tag->{'delimiter'}, @v) );
 			}
 
 			# default types 
-			my @types = qw(display swish);
+			my @types = qw(display search);
 			# override by type attribute
 			@types = ( $tag->{'type'} ) if ($tag->{'type'});
 
 			foreach my $type (@types) {
 				# append to previous line?
-				$log->debug("type: $type ",sub { join(" ",@v) }, $row->{'append'} || 'no append');
+				$log->debug("type: $type ",sub { join(" ",@v) }, " ", $row->{'append'} || 'no append');
 				if ($tag->{'append'}) {
 
 					# I will delimit appended part with
@@ -240,14 +252,14 @@
 
 			# TODO: name_sigular, name_plural
 			my $name = $self->{'import_xml'}->{'indexer'}->{$field}->{'name'};
-			$row->{'name'} = $name ? $self->_x($name) : $field;
+			my $row_name = $name ? $self->_x($name) : $field;
 
 			# post-sort all values in field
 			if ($self->{'import_xml'}->{'indexer'}->{$field}->{'sort'}) {
 				$log->warn("sort at field tag not implemented");
 			}
 
-			push @ds, $row;
+			$ds->{$row_name} = $row;
 
 			$log->debug("row $field: ",sub { Dumper($row) });
 		}
@@ -255,14 +267,16 @@
 	}
 
 	$self->{'db'}->save_ds(
-		ds => \@ds,
-		current_filename => $self->{'current_filename'},
-		headline => $self->{'headline'},
+		id => $id,
+		ds => $ds,
+		prefix => $self->{prefix},
 	) if ($self->{'db'});
 
-	$log->debug("ds: ", sub { Dumper(@ds) });
+	$log->debug("ds: ", sub { Dumper($ds) });
+
+	$log->logconfess("data structure returned is not array any more!") if wantarray;
 
-	return @ds;
+	return $ds;
 
 }
 
@@ -274,8 +288,18 @@
 
  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;
 
@@ -293,7 +317,7 @@
 
 	my @out;
 
-	$log->debug("format: $format");
+	$log->debug("format: $format [$i]");
 
 	my $eval_code;
 	# remove eval{...} from beginning
@@ -303,29 +327,45 @@
 	# 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);
 
@@ -345,11 +385,21 @@
 		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;
@@ -429,6 +479,7 @@
 	}
 
 	my $found = 0;
+	my $just_single = 1;
 
 	my $eval_code;
 	# remove eval{...} from beginning
@@ -440,9 +491,14 @@
 
 	# 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: $format");
@@ -459,7 +515,9 @@
 		# do we have lookups?
 		if ($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 {
 				$log->warn("Have lookup object but can't invoke lookup method");
 			}
@@ -535,19 +593,22 @@
 		if ($sf && $$rec->{$f}->[$i]->{$sf}) {
 			$$found++ if (defined($$found));
 			return $$rec->{$f}->[$i]->{$sf};
-		} elsif ($$rec->{$f}->[$i]) {
+		} elsif (! $sf && $$rec->{$f}->[$i]) {
 			$$found++ if (defined($$found));
 			# it still might have subfield, just
 			# not specified, so we'll dump all
 			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 {
 				return $$rec->{$f}->[$i];
 			}
+		} else {
+			return '';
 		}
 	} else {
 		return '';
@@ -671,4 +732,4 @@
 
 =cut
 
-1; # End of WebPAC::DB
+1; # End of WebPAC::Normalize