Changeset 24419 for main/trunk

Timestamp:

2011-08-17T18:52:58+12:00 (13 years ago)

Author:

ak19

Message:

1. Dr Bainbridge fixed another bug with the EmbeddedMetadataPlugin and its interaction with OAIPlugin: extracted metadata is now merged into any earlier extracted metadata table, so that ex.dc.* meta extracted by the OAIPlugin higher up in the plugin pipeline is preserved after EmbeddedMetaPlug is through with the file. 2. Removed recent PDFPlugin commit where can_process_this_file_for_metadata was overridden. It isn't needed and may have adverse effects.

Location:

main/trunk/greenstone2/perllib/plugins

Files:

• EmbeddedMetadataPlugin.pm (modified) (1 diff)
• PDFPlugin.pm (modified) (1 diff)

main/trunk/greenstone2/perllib/plugins/EmbeddedMetadataPlugin.pm

@@ -281 +281 @@
     # Associate the metadata now
 
-    $extrametadata->{$file} = \%exif_metadata;
-    push(@$extrametakeys, $file);
+    if (defined $extrametadata->{$file}) {
+    print STDERR "\n****  Need to merge new metadata with existing stored metadata: file = $file\n" if $verbosity > 2;
+
+    my $file_metadata_table =     $extrametadata->{$file};
+
+    foreach my $metaname (keys %exif_metadata) {
+        # will create new entry if one does not already exist
+        push(@{$file_metadata_table->{$metaname}}, @{$exif_metadata{$metaname}});       
+    }
+
+    # no need to push $file on to $extrametakeys as it is already in the list
+    }
+    else {
+    $extrametadata->{$file} = \%exif_metadata;
+    push(@$extrametakeys, $file);
+    }
 
 }

main/trunk/greenstone2/perllib/plugins/PDFPlugin.pm

@@ -213 +213 @@
 }
 
-# Even if a plugin can extract metadata in its metadata_read pass,
-# make the default return 'undef' so processing of the file continues
-# down the pipeline, so other plugins can also have the opportunity to
-# locate metadata and set it up in the extrametakeys variables that
-# are passed around.
-sub can_process_this_file_for_metadata {
-    my $self = shift(@_);
-
-    # this plugin will look for metadata in any file through its
-    # metadata_read(). Returning undef here means anything else further
-    # down the pipeline can do the same
-
-    return 0;
-}
-
 sub init {
     my $self = shift (@_);