Changeset 26267 for main/trunk/greenstone2

Timestamp:

2012-09-27T14:24:22+12:00 (12 years ago)

Author:

kjdon

Message:

added filter_metadata and filer_regex to filter documents going into the classifier. Can use these to restrict which documents get added to the classifier. eg If have several different DocType metadatas, can build a classifier on just one/some of the doctypes

File:

• main/trunk/greenstone2/perllib/classify/List.pm (modified) (3 diffs)

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

@@ -108 +108 @@
     'desc' => "{List.use_hlist_for}",
     'type' => "string" },
+      {'name' => "filter_metadata",
+       'desc' => "{List.filter_metadata}",
+       'type' => "metadata"},
+      {'name' => "filter_regex",
+       'desc' => "{List.filter_regex}",
+       'type' => "regexp"},
+      
       { 'name' => "removeprefix",
     'desc' => "{BasClas.removeprefix}",
@@ -302 +309 @@
     }
     $self->{'all_doc_OIDs'}->{$doc_obj->get_OID()} = 1;
+    # check against filter here
+    if ($self->{'filter_metadata'}) {
+    #print STDERR "filtering documents on $self->{'filter_metadata'}\n";
+    my $meta = $doc_obj->get_metadata_element($doc_obj->get_top_section(), $self->{'filter_metadata'});
+    return unless defined $meta;
+    if ($self->{'filter_regex'} ne "" && $meta !~ /$self->{'filter_regex'}/) {
+        #print STDERR "doesn't match regex\n";
+        return;
+
+    }
+    }
+    # if we get here, we have passed the test for filtering
     # If "-classify_sections" is set, classify every section of the document
     if ($self->{'classify_sections'}) {
@@ -324 +343 @@
     my @metadata_groups = @{$self->{'metadata_groups'}};
 
+    
     # Only classify the section if it has a value for one of the metadata elements in the first group
     my $classify_section = 0;