Changeset 24691 for main/trunk/greenstone2

Timestamp:

2011-09-29T16:51:48+13:00 (13 years ago)

Author:

davidb

Message:

Improved multi-lingual support

File:

• main/trunk/greenstone2/perllib/classify/AZCompactList.pm (modified) (7 diffs)

main/trunk/greenstone2/perllib/classify/AZCompactList.pm

@@ -27 +27 @@
 package AZCompactList;
 
+use strict;
+no strict 'refs'; # allow filehandles to be variables and viceversa
+
 use BaseClassifier;
 use sorttools;
 
-use strict;
-no strict 'refs'; # allow filehandles to be variables and viceversa
+use Unicode::Normalize;
 
 sub BEGIN {
@@ -40 +42 @@
     [ { 'name' => "top",
     'desc' => "{AZCompactList.doclevel.top}" },
+      { 'name' => "firstlevel",
+    'desc' => "{AZCompactList.doclevel.firstlevel}" },
       { 'name' => "section",
     'desc' => "{AZCompactList.doclevel.section}" } ];
@@ -200 +204 @@
     push(@sectionlist,$topsection);
     }
-    else
+    elsif ($self->{'doclevel'} =~ /^first(level)?/i)
+    {
+    my $toplevel_children = $doc_obj->get_children($topsection);
+    push(@sectionlist,@$toplevel_children);
+    }
+    else # (all)?section(s)?
     {
     my $thissection = $doc_obj->get_next_section($topsection);
@@ -389 +398 @@
         else
         {
+            # first(level)? or (all)?section(s)?
             eval ("\$listclassobj = new SectionList([],\$ptArgs)");
         }
@@ -489 +499 @@
         {
         my $section=$1;
-        if ($self->{'doclevel'} =~ m/^top/i) { # toplevel
+        if ($self->{'doclevel'} =~ m/^top(level)?/i) { # toplevel
             $self->{'classifiers'}->{$node_name}->{'classifyobj'}
             ->classify($doc_obj,"Section=$section");
-        } else { # section level
-            # Thanks to Don Gourley for this...
-            # classify can't handle multi-level section
+        } else { 
+            # first(level)? or (all)?section(s)? 
+
+            # classify() can't handle multi-level section, so use
+            # classify_section()
+            # ... thanks to Don Gourley for this...
+
             $self->{'classifiers'}->{$node_name}->{'classifyobj'}
             ->classify_section($section, $doc_obj, $sortmeta);
@@ -683 +697 @@
     my $title = $self->{'reclassifylist'}->{$classification};
     $title =~ s/&(.){2,4};//g; # remove any HTML special chars
-    $title =~ s/^\W+//g; # remove leading non-word chars
+    $title =~ s/^(\W|_)+//g; # remove leading non-word chars
 
     # only want first character for classification
@@ -689 +703 @@
     if (defined($1) && $1 ne "") {
         $title=$1;
+
+        # remove any accents on initial character by mapping to Unicode's
+        # normalized decomposed form (accents follow initial letter)
+        # and then pick off the initial letter 
+        my $title_decomposed = NFD($title); 
+        $title = substr($title_decomposed,0,1);
     } else {
         print STDERR "no first character found for \"$title\" - \"" .