Changeset 2955 for trunk/gsdl/perllib/classify/AZCompactList.pm

Timestamp:

2002-02-07T16:28:48+13:00 (22 years ago)

Author:

jrm21

Message:

Added removeprefix option. Added better usage information of the options.

File:

• trunk/gsdl/perllib/classify/AZCompactList.pm (modified) (13 diffs)

trunk/gsdl/perllib/classify/AZCompactList.pm

@@ -52 +52 @@
 sub print_usage {
     print STDERR "
-  usage: classify AZCompactList [options]
+  usage: classify AZCompactList -metadata X [options]
   options:
-   -metadata     Metadata field used for classification
-   -buttonname   Title to use on web pages
-
-   -mingroup
-   -minnesting
-   -mincompact
-   -maxcompact
-   -doclevel
-   -onlyfirst
+   -metadata X            (required) Metadata field used for classification
+   -buttonname X          Title to use on web pages (defaults to metadata)
+   -removeprefix regex    pattern to remove from metadata before sorting
+   -doclevel top|section  (Defaults to top)
+   -mingroup N            Minimum num of documents required to form a new group
+   -minnesting N          Minimum list size to become a nested list
+   -mincompact N          Used in compact list
+   -maxcompact N          Used in compact list
+   -onlyfirst             Only use the first value if metadata is repeated.
    -recopt
 ";
@@ -71 +71 @@
     my $self = new BasClas($class, @_);
 
-    my ($metaname, $title);
+    my ($metaname, $title, $removeprefix);
     my $mingroup = 2;
     my $minnesting = 20;
@@ -83 +83 @@
              q^metadata/.*/^, \$metaname,
              q^buttonname/.*/^, \$title,
-
+             q^removeprefix/.*/^, \$removeprefix,
              q^mingroup/.*/2^, \$mingroup,
              q^minnesting/.*/20^, \$minnesting,
@@ -99 +99 @@
     }
 
-# These are the old-fashioned optins to AZCompactList.
-# They should be deleted once we're sure the parsearg version is working. 
-#    foreach $option (@options) {
-#   if ($option =~ /^metadata=(.*)$/i) {
-#       $metaname = $1;
-#   } elsif ($option =~ /^title=(.*)$/i) {
-#       $title = $1;
-#   } elsif ($option =~ /^mingroup(size)?=(.*)$/i) {
-#       $mingroup = $2;
-#   } elsif ($option =~ /^minnesting=(.*)$/i) {
-#       $minnesting = $1;
-#   } elsif ($option =~ /^mincompact=(.*)$/i) {
-#       $mincompact = $1;
-#   } elsif ($option =~ /^maxcompact=(.*)$/i) {
-#       $maxcompact = $1;
-#   } elsif ($option =~ /^doclevel=(.*)$/i) {
-#       $doclevel = $1;
-#   } elsif ($option =~ /^onlyfirst$/i) {
-#       $onlyfirst = 1;
-#   } elsif ($option =~ /^recopt$/i) {
-#       $recopt = "on";
-#   }
-#    }
-
     if (!defined $metaname) {
     my $outhandle = $self->{'outhandle'};
@@ -137 +113 @@
     $self->{'metaname'} = $metaname;
     $self->{'title'} = "$title";     # title for the titlebar.
+    if (defined($removeprefix) && $removeprefix) {
+    $self->{'removeprefix'} = $removeprefix;
+    }
     $self->{'mingroup'} = $mingroup;
     $self->{'minnesting'} = $minnesting;
@@ -216 +195 @@
     foreach $metavalue (@$metavalues) 
     {
-###     print STDERR "$metaname :\tmetavalue = $metavalue\n";
         # if this document doesn't contain the metadata element we're
         # sorting by we won't include it in this classification
         if (defined $metavalue && $metavalue =~ /\w/) 
         {
+        if ($self->{'removeprefix'}) {
+            $metavalue =~ s/^$self->{'removeprefix'}//;
+        }
+
         my $formatted_metavalue = $metavalue;
+
         if ($self->{'metaname'} =~ m/^Creator(:.*)?$/)
         {
@@ -231 +214 @@
         }
         
+        #### prefix-str
+        if (! defined($formatted_metavalue)) {
+            print $outhandle "Warning: AZCompactList: metavalue is ";
+            print $outhandle "empty\n";
+            $formatted_metavalue="";
+        }
+
         push(@{$self->{'list'}->{$full_doc_OID}},$formatted_metavalue);
         push(@{$self->{'listmetavalue'}->{$full_doc_OID}} ,$metavalue);
@@ -245 +235 @@
 {
     my ($self,$classlist_ref) = @_;
+    my $outhandle = $self->{'outhandle'};
     
     my %mtfreq = ();
@@ -294 +285 @@
     else 
     { 
-        my $outhandle = $self->{'outhandle'};
         print $outhandle "AZCompactList ERROR - couldn't find classifier \"$listname\"\n"; 
         die "\n";
@@ -342 +332 @@
         }
         if ($@) {
-        my $outhandle = $self->{'outhandle'};
         print $outhandle "$@";
         die "\n";
@@ -352 +341 @@
         {
         my $formatted_node = $metavalue;
+
+        if ($self->{'removeprefix'}) {
+            $formatted_node =~ s/^$self->{'removeprefix'}//;
+        }
+
         if ($self->{'metaname'} =~ m/^Creator(:.*)?$/)
         {
@@ -360 +354 @@
             &sorttools::format_string_english(\$formatted_node);
         }
-        
+
+        # In case our formatted string is empty...
+        if (! defined($formatted_node)) {
+            print $outhandle "Warning: AZCompactList: metavalue is ";
+            print $outhandle "empty\n";
+            $formatted_node="";
+        }
+
         $self->{'classifiers'}->{$metavalue} 
         = { 'classifyobj'   => $listclassobj,
@@ -418 +419 @@
         my $outhandle=$self->{outhandle};
         print $outhandle "Warning: AZCompactList::reclassify ";
-        print $outhandle "could not find sub-node for $metavalue with doc_OID $doc_OID\n";
+        print $outhandle "could not find sub-node for metadata=`$metavalue' with doc_OID $doc_OID\n";
     }
     }