Changeset 10353

Timestamp:

2005-07-29T14:27:19+12:00 (19 years ago)

Author:

chi

Message:

Modification for allowing PDF document being converted to various types of image through convert utility. Then, allow secondary_plugin PagedImgPlug to deal with the document.

File:

• trunk/gsdl/perllib/plugins/PDFPlug.pm (modified) (8 diffs)

trunk/gsdl/perllib/plugins/PDFPlug.pm

@@ -23 +23 @@
 #
 ###########################################################################
+package PDFPlug;
+
+use ConvertToPlug;
+use unicode;
 use strict;
 no strict 'refs'; # so we can use a var for filehandles (eg STDERR)
-
-package PDFPlug;
-
-use ConvertToPlug;
-use unicode;
 
 sub BEGIN {
@@ -74 +73 @@
     push(@$pluginlist, $class);
 
+    push(@$inputargs,"-title_sub");
+    push(@$inputargs,'^(Page\s+\d+)?(\s*1\s+)?');
+
+    #foreach my $temp (@$inputargs)
+    #{
+    #print STDERR "($temp)\n";
+    #}
+
     if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
     if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
-
-    push(@$inputargs,"-title_sub");
-    push(@$inputargs,'^(Page\s+\d+)?(\s*1\s+)?');
-
+   
     my $self = (defined $hashArgOptLists)? new ConvertToPlug($pluginlist,$inputargs,$hashArgOptLists): new ConvertToPlug($pluginlist,$inputargs);
-
-
-    #if ($self->{"use_sections"}) {
-    #$self->{"description_tags"} = 1;
-    #}
-
+    
     # these are passed through to gsConvert.pl by ConvertToPlug.pm
     my $zoom = $self->{"zoom"};
@@ -102 +101 @@
     $secondary_plugin_options->{'TEXTPlug'} = [];
     }
-
+    if (defined $self->{'convert_to'} && $self->{'convert_to'} =~ /pagedimg.*/i) {
+    if (!defined $secondary_plugin_options->{'PagedImgPlug'}){
+        $secondary_plugin_options->{'PagedImgPlug'} = [];
+        my $pagedimg_options = $secondary_plugin_options->{'PagedImgPlug'}; 
+        push(@$pagedimg_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?')
+        
+    }
+    }
     my $html_options = $secondary_plugin_options->{'HTMLPlug'};
     my $text_options = $secondary_plugin_options->{'TEXTPlug'};
+    #my $pagedimg_options = $secondary_plugin_options->{'PagedImgPlug'};
 
     if ($self->{'input_encoding'} eq "auto") {
@@ -110 +117 @@
     # => restrict primary PDFPlug and secondary HTML plugin to use 
     # utf8 and extract language.
-
     $self->{'input_encoding'} = "utf8";
     $self->{'extract_language'} = 1;
@@ -126 +132 @@
     push(@$html_options,"-description_tags");
     }
-
     # following title_sub removes "Page 1" added by pdftohtml, and a leading
     # "1", which is often the page number at the top of the page. Bad Luck
@@ -132 +137 @@
     push(@$html_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
     push(@$text_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
+    #push(@$pagedimg_options , "-title_sub", '^(Page\s+\d+)?(\s*1\s+)?');
 
     $self = bless $self, $class;
-
     $self->load_secondary_plugins($class,$secondary_plugin_options);
-
     return $self;
 }
@@ -170 +174 @@
     # be useful to give an indication of document length in browser through setting
     # num_pages as metadata.
-
     my @pages = ($text =~ /\<[Aa] name=\"?\w+\"?>/ig);
     my $num_pages = scalar(@pages);
@@ -275 +278 @@
 
     $doc_obj->add_utf8_metadata($cursection, "NumPages", $self->{'num_pages'});
-
     
     if ($self->{'use_sections'} && $self->{'converted_to'} eq "HTML") {