Changeset 29094 for main/trunk/greenstone2

Timestamp:

2014-06-09T14:40:40+12:00 (10 years ago)

Author:

kjdon

Message:

added -metadata_selection_mode firstvalue|firstvalidmetadata|allvalues option, to decide which values to classify on if using a list of metadata, eg dc.Date,Date. Replicates -firstvalueonly and -allvalues of AZCompactList, but can specify for each level in hierarchy.

Location:

main/trunk/greenstone2/perllib

Files:

• classify/List.pm (modified) (6 diffs)
• strings.properties (modified) (1 diff)

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

@@ -44 +44 @@
 }
 
+my $meta_select_type_list =
+    [ 
+      { 'name' => "firstvalue",
+    'desc' => "{List.metadata_selection.firstvalue}"},
+      { 'name' => "firstvalidmetadata",
+    'desc' => "{List.metadata_selection.firstvalidmetadata}"},
+      { 'name' => "allvalues",
+    'desc' => "{List.metadata_selection.allvalues}"} ];
+my $valid_meta_select_types = { 'firstvalue' => 1,
+                'firstvalidmetadata' => 1,
+                'allvalues' => 1 };
 my $partition_type_list = 
     [ { 'name' => "per_letter",
@@ -74 +85 @@
     'type' => "metadata",
     'reqd' => "yes" },
-
+      { 'name' => "metadata_selection_mode",
+    'desc' => "{List.metadata_selection_mode}",
+    'type' => "enumstring", # Must be enumstring because multiple values can be specified (separated by '/')
+    'list' => $meta_select_type_list,
+    'deft' => "firstvalidmetadata" },
       # The interesting options
       { 'name' => "bookshelf_type", 
@@ -162 +177 @@
     }
 
+    # meta selection mode for each level
+    if (!$self->{'metadata_selection_mode'}) {
+    foreach my $metadata_group (@metadata_groups) {
+        $self->{$metadata_group . ".metadata_selection_mode"} = "firstvalidmetadata";
+    }
+    } else {
+    my @metadata_selection_mode_list = split(/\//, $self->{'metadata_selection_mode'});
+    foreach my $metadata_group (@metadata_groups) {
+        my $meta_select_mode = shift(@metadata_selection_mode_list);
+        if (defined($meta_select_mode) && defined $valid_meta_select_types->{$meta_select_mode}) {
+        $self->{$metadata_group . ".metadata_selection_mode"} = $meta_select_mode;
+        } else {
+        $self->{$metadata_group . ".metadata_selection_mode"} = "firstvalidmetadata";   
+        }
+    }
+    }   
     # Whether to group items into a bookshelf, (must be 'always' for all metadata fields except the last)
     foreach my $metadata_group (@metadata_groups) {
@@ -276 +307 @@
     }
     $self->{'sort_leaf_nodes_using_metadata_groups'} = \@sort_leaf_nodes_using_metadata_groups;
-
+    foreach my $sort_group (@sort_leaf_nodes_using_metadata_groups) {
+    # set metadata_select_type, if not already set - might be already set if the same group was used in -metadata
+    if (!defined $self->{$sort_group . ".metadata_selection_mode"}) {
+        $self->{$sort_group . ".metadata_selection_mode"} = "firstvalue";
+    }
+    }
     # Create an instance of the Unicode::Collate object if better Unicode sorting is desired
     if ($self->{'sort_using_unicode_collation'}) {
@@ -413 +449 @@
 
             # Add the metadata value into the list for this combination of metadata group and section
-            push(@{$self->{$metadata_group . ".list"}->{$section_OID}}, $lc_metadata_value);
+            # text that we have some non-whitespace chars
+            if ($lc_metadata_value =~ /\S/) {
+            
+            push(@{$self->{$metadata_group . ".list"}->{$section_OID}}, $lc_metadata_value);
+            
+            
             # add the actual value into the stored values so we can remember the case
             if (!$self->{'standardize_capitalization'}) {
@@ -421 +462 @@
                 $self->{$metadata_group . ".actualvalues"}->{$lc_metadata_value}->{$metadata_value} = 1;
             }
-            }       
-        }
-        last if (@metadata_values > 0);
-        }
+            }
+            last if ($self->{$metadata_group . ".metadata_selection_mode"} eq "firstvalue");
+            }
+        } # foreach metadatavalue
+        last if ((@metadata_values > 0) &&  $self->{$metadata_group . ".metadata_selection_mode"} =~ /^(firstvalue|firstvalidmetadata)$/ );
+        } # foreach metadata element
 
         $metadata_groups_done{$metadata_group} = 1;

main/trunk/greenstone2/perllib/strings.properties

@@ -649 +649 @@
 List.level_partition.approximate_size:Create a partition per letter, then group or split the letters to get approximately the same sized partitions. 
 
-List.metadata:Metadata fields used for classification. Use '/' to separate the levels in the hierarchy and ';' to separate metadata fields within each level.
-
+List.metadata:Metadata fields used for classification. Use '/' to separate the levels in the hierarchy and ';' or ',' to separate metadata fields within each level.
+
+List.metadata_selection_mode:Determines how many metadata values the document is classified by, within each level. Use '/' to separate the levels.
+List.metadata_selection.firstvalue:Only classify by a single metadata value, the first one encountered.
+List.metadata_selection.firstvalidmetadata:Classify by all the metadata values of the first element in the list that has values.
+List.metadata_selection.allvalues:Classify by all metadata values found, from all elements in the list.
 List.partition_name_length:The length of the partition name; defaults to a variable length from 1 up to 3 characters, depending on how many are required to distinguish the partition start from its end. This option only applies when partition_type_within_level is set to 'constant_size'.