--- trunk/jsFind.pm	2004/07/11 20:18:25	1
+++ trunk/jsFind.pm	2004/08/28 14:31:58	12
@@ -3,8 +3,18 @@
 use 5.008004;
 use strict;
 use warnings;
+use HTML::Entities;
 
-our $VERSION = '0.01';
+our $VERSION = '0.03';
+
+use Exporter 'import';
+use Carp;
+
+our @ISA = qw(Exporter);
+
+BEGIN {
+	import 'jsFind::Node';
+}
 
 =head1 NAME
 
@@ -13,8 +23,19 @@
 =head1 SYNOPSIS
 
   use jsFind;
-
-
+  my $t = new jsFind(B => 4);
+  my $f = 1;
+  foreach my $k (qw{minima ut dolorem sapiente voluptatem}) {
+  	$t->B_search(Key => $k,
+  		Data => {
+  			"path" => {
+  			t => "word $k",
+  			f => $f },
+  		},
+  		Insert => 1,
+  		Append => 1,
+  	);
+  }
 
 =head1 DESCRIPTION
 
@@ -36,24 +57,12 @@
 
 =back
 
-=head1 METHODS
-
-This module contains two packages C<jsFind> and C<jsFind::Node>.
-
-=head2 jsFind methods
-
-=cut
-
-use Exporter 'import';
-use Carp;
+=head1 jsFind methods
 
-our @ISA = qw(Exporter);
+C<jsFind> is mode implementing methods which you, the user, are going to
+use to create indexes.
 
-BEGIN {
-	import 'jsFind::Node';
-}
-
-=head3 new
+=head2 new
 
 Create new tree. Arguments are C<B> which is maximum numbers of keys in
 each node and optional C<Root> node. Each root node may have child nodes.
@@ -82,11 +91,20 @@
   bless { B => $B, Root => $Root } => $package;
 }
 
-=head3 B_search
+=head2 B_search
 
 Search, insert, append or replace data in B-Tree
 
- 
+ $t->B_search(
+ 	Key => 'key value',
+	Data => { "path" => {
+			"t" => "title of document",
+			"f" => 99,
+			},
+		},
+	Insert => 1,
+	Append => 1,
+ );
 
 Semantics:
 
@@ -215,7 +233,7 @@
   }
 }
 
-=head3 B
+=head2 B
 
 Return B (maximum number of keys)
 
@@ -227,7 +245,7 @@
   $_[0]{B};
 }
 
-=head3 root
+=head2 root
 
 Returns root node
 
@@ -241,7 +259,7 @@
   $self->{Root};
 }
 
-=head3 node_overfull
+=head2 node_overfull
 
 Returns if node is overfull
 
@@ -255,7 +273,7 @@
   $node->size > $self->B;
 }
 
-=head3 to_string
+=head2 to_string
 
 Returns your tree as formatted string.
 
@@ -269,7 +287,7 @@
   $_[0]->root->to_string;
 }
 
-=head3 to_dot
+=head2 to_dot
 
 Create Graphviz graph of your tree
 
@@ -287,7 +305,7 @@
 	return $dot;
 }
 
-=head3 to_jsfind
+=head2 to_jsfind
 
 Create xml index files for jsFind. This should be called after
 your B-Tree has been filled with data.
@@ -296,15 +314,36 @@
 
 Returns number of nodes in created tree.
 
+There is also longer version if you want to recode your data charset
+into different one (probably UTF-8):
+
+ $root->to_jsfind('/full/path/to/index/dir/','ISO-8859-2','UTF-8');
+
+Destination encoding is UTF-8 by default, so you don't have to specify it.
+
+ $root->to_jsfind('/full/path/to/index/dir/','WINDOWS-1250');
+
 =cut
 
+my $iconv;
+my $iconv_l1;
+
 sub to_jsfind {
 	my $self = shift;
 
 	my $path = shift || confess "to_jsfind need path to your index!";
 
+	my ($from_cp,$to_cp) = @_;
+
+	$to_cp ||= 'UTF-8';
+
+	if ($from_cp && $to_cp) {
+		$iconv = Text::Iconv->new($from_cp,$to_cp);
+	}
+	$iconv_l1 = Text::Iconv->new('ISO-8859-1',$to_cp);
+
 	$path .= "/" if ($path =~ /\/$/);
-	carp "can't create index in '$path': $!" if (! -w $path);
+	#carp "creating directory for index '$path'" if (! -w $path);
 
 	return $self->root->to_jsfind($path,"0");
 }
@@ -315,9 +354,34 @@
   $_[0] cmp $_[1];
 }
 
+=head2 _recode
+
+This is internal function to recode charset.
+
+It will also try to decode entities in data using L<HTML::Entities>.
+
+=cut
+
+sub _recode {
+	my $self = shift;
+	my $text = shift || return;
+
+	sub _decode_html_entities {
+		my $data = shift || return;
+		$data = $iconv_l1->convert(decode_entities($data)) || croak "entity decode problem: $data";
+	}
+
+	if ($iconv) {
+		$text = $iconv->convert($text) || $text && carp "convert problem: $text";
+		$text =~ s/(\&\w+;)/_decode_html_entities($1)/ges;
+	}
+
+	return $text;
+}
+
 #####################################################################
 
-=head2 jsFind::Node methods
+=head1 jsFind::Node methods
 
 Each node has C<k> key-data pairs, with C<B> <= C<k> <= C<2B>, and 
 each has C<k+1> subnodes, which might be null.
@@ -339,12 +403,15 @@
 
 use Carp;
 use File::Path;
+use Text::Iconv;
+
+use base 'jsFind';
 
 my $KEYS = 0;
 my $DATA = 1;
 my $SUBNODES = 2;
 
-=head3 new
+=head2 new
 
 Create New node
 
@@ -364,7 +431,7 @@
   bless [@_] => $package;
 }
 
-=head3 locate_key
+=head2 locate_key
 
 Locate key in node using linear search. This should probably be replaced
 by binary search for better performance.
@@ -401,7 +468,7 @@
 }
 
 
-=head3 emptynode
+=head2 emptynode
 
 Creates new empty node
 
@@ -414,7 +481,7 @@
   new($_[0]);			# Pass package name, but not anything else.
 }
 
-=head3 is_empty
+=head2 is_empty
 
 Test if node is empty
 
@@ -428,7 +495,7 @@
   !defined($self) || $#$self < 0;
 }
 
-=head3 key
+=head2 key
 
 Return C<$i>th key from node
 
@@ -444,7 +511,7 @@
    $_[0]->[$KEYS][$_[1]];
 }
 
-=head3 data
+=head2 data
 
 Return C<$i>th data from node
 
@@ -457,7 +524,7 @@
   $self->[$DATA][$n];
 }
 
-=head3 kdp_replace
+=head2 kdp_replace
 
 Set key data pair for C<$i>th element in node
 
@@ -478,9 +545,13 @@
    $self->[$DATA][$n]];
 }
 
-=head3 kdp_insert
+=head2 kdp_insert
+
+Insert key/data pair in tree
+
+  $node->kdp_insert("key value" => "data value");
 
- # No return value.
+No return value.
 
 =cut
 
@@ -499,7 +570,7 @@
   splice(@{$self->[$SUBNODES]}, $where, 0, undef);
 }
 
-=head3 kdp_append
+=head2 kdp_append
 
 Adds new data keys and values to C<$i>th element in node
 
@@ -520,7 +591,7 @@
    $self->[$DATA][$n]];
 }
 
-=head3 subnode
+=head2 subnode
 
 Set new or return existing subnode
 
@@ -538,7 +609,7 @@
   $self->[$SUBNODES][$n];
 }
 
-=head3 is_leaf
+=head2 is_leaf
 
 Test if node is leaf
 
@@ -551,7 +622,7 @@
   ! defined $self->[$SUBNODES][0]; # undefined subnode means leaf node.
 }
 
-=head3 size
+=head2 size
 
 Return number of keys in the node
 
@@ -564,11 +635,12 @@
   return scalar(@{$self->[$KEYS]});
 }
 
-=head3 halves
+=head2 halves
 
- # Accept an index $n
- # Divide into two nodes so that keys 0 .. $n-1 are in one node
- # and keys $n+1 ... $size are in the other.
+Split node into two halves so that keys C<0 .. $n-1> are in one node
+and keys C<$n+1 ... $size> are in the other.
+
+  my ($left_node, $right_node, $kdp) = $node->halves($n);
 
 =cut
 
@@ -592,7 +664,7 @@
   ($self->new(@left), $self->new(@right), \@middle);
 }
 
-=head3 to_string
+=head2 to_string
 
 Dumps tree as string
 
@@ -643,7 +715,7 @@
 
 =end comment
 
-=head3 to_dot
+=head2 to_dot
 
 Recursivly walk nodes of tree
 
@@ -682,22 +754,45 @@
 	$dot;
 }
 
-=head3 to_jsfind
+=head2 to_xml
+
+Escape <, >, & and ", and to produce valid XML
+
+=cut
+
+my %escape = ('<'=>'&lt;', '>'=>'&gt;', '&'=>'&amp;', '"'=>'&quot;');
+my $escape_re  = join '|' => keys %escape;
+
+sub to_xml {
+	my $self = shift || confess "you should call to_xml as object!";
+
+	my $d = shift || return;
+	$d = $self->SUPER::_recode($d);
+	confess "escape_re undefined!" unless ($escape_re);
+	$d =~ s/($escape_re)/$escape{$1}/g;
+	return $d;
+}
+
+=head2 to_jsfind
 
 Create jsFind xml files
 
- my $nr=$tree->to_dot('/path/to/index','0');
+ my $nr=$tree->to_jsfind('/path/to/index','0');
 
 Returns number of elements created
 
 =cut
 
+
 sub to_jsfind {
 	my $self = shift;
 	my ($path,$file) = @_;
 
 	return 0 if $self->is_empty;
 
+	confess("path is undefined.") unless ($path);
+	confess("file is undefined. Did you call \$t->root->to_jsfind(..) instead of \$t->to_jsfind(..) ?") unless (defined($file));
+
 	my $nr_keys = 0;
 
 	my ($k, $d, $s) = @$self;
@@ -709,22 +804,22 @@
 		my $key = lc($k->[$i]);
 
 		if ($key) {
-			$key_xml .= qq{<k>$key</k>};
-			$data_xml .= qq{<e>};
+			$key_xml .= '<k>'.$self->to_xml($key).'</k>';
+			$data_xml .= '<e>';
 	#use Data::Dumper;
 	#print Dumper($d->[$i]);
 			foreach my $path (keys %{$d->[$i]}) {
-				$data_xml .= '<l f="'.($d->[$i]->{$path}->{'f'} || 1).'" t="'.($d->[$i]->{$path}->{'t'} || 'no title').'">'.$path.'</l>';
+				$data_xml .= '<l f="'.($d->[$i]->{$path}->{'f'} || 1).'" t="'.$self->to_xml($d->[$i]->{$path}->{'t'} || 'no title').'">'.$self->to_xml($path).'</l>';
 				$nr_keys++;
 			}
-			$data_xml .= qq{</e>};
+			$data_xml .= '</e>';
 		}
 
 		$nr_keys += $s->[$i]->to_jsfind("$path/$file","$i") if ($s->[$i]);
 	}
 
-	$key_xml .= "</n>";
-	$data_xml .= "</d>";
+	$key_xml .= '</n>';
+	$data_xml .= '</d>';
 
 	if (! -e $path) {
 		mkpath($path) || croak "can't create dir '$path': $!";