Changeset 1279 for branches/New_Config_Format-branch/gsdl/perllib/mgbuildproc.pm

Timestamp:

2000-07-13T10:21:53+12:00 (24 years ago)

Author:

sjboddie

Message:

merged changes to trunk into New_Config_Format branch

File:

• branches/New_Config_Format-branch/gsdl/perllib/mgbuildproc.pm (modified) (7 diffs)

branches/New_Config_Format-branch/gsdl/perllib/mgbuildproc.pm

@@ -58 +58 @@
     $self->{'num_sections'} = 0;
     $self->{'num_bytes'} = 0;
+    $self->{'num_processed_bytes'} = 0;
 
     $self->{'indexing_text'} = 0;
@@ -69 +70 @@
     $self->{'num_docs'} = 0;
     $self->{'num_sections'} = 0;
+    $self->{'num_processed_bytes'} = 0;
     $self->{'num_bytes'} = 0;
 }
@@ -84 +86 @@
 }
 
+# num_bytes is the actual number of bytes in the collection
+# this is normally the same as what's processed during text compression
 sub get_num_bytes {
     my $self = shift (@_);
 
     return $self->{'num_bytes'};
+}
+
+# num_processed_bytes is the number of bytes actually passed
+# to mg for the current index
+sub get_num_processed_bytes {
+    my $self = shift (@_);
+
+    return $self->{'num_processed_bytes'};
 }
 
@@ -126 +138 @@
 }
 
+sub get_index {
+    my $self = shift (@_);
+
+    return $self->{'index'};
+}
+
 sub set_classifiers {
     my $self = shift (@_);
@@ -138 +156 @@
 
     $self->{'indexing_text'} = $indexing_text;
+}
+
+sub get_indexing_text {
+    my $self = shift (@_);
+
+    return $self->{'indexing_text'};
 }
 
@@ -416 +440 @@
             if ($real_field eq "text") {
             $new_text = $doc_obj->get_text ($section);
+            $self->{'num_processed_bytes'} += length ($new_text);
             $new_text =~ s/[\cB\cC]//g;
             $self->find_paragraphs($new_text);
@@ -423 +448 @@
             foreach $meta (@{$doc_obj->get_metadata ($section, $real_field)}) {
                 $meta =~ s/[\cB\cC]//g;
+                $self->{'num_processed_bytes'} += length ($meta);
                 $new_text .= "\cC" unless $first;
                 $new_text .= $meta;