Changeset 24159 for main/trunk/greenstone2/perllib/plugins/PDFPlugin.pm

Timestamp:

2011-06-14T20:39:16+12:00 (13 years ago)

Author:

ak19

Message:

Implemented use_sections for PDFBox, needed to grab the (at this moment) latest version of the PDFBox pre-built binary jar file: pdfbox-app-1.5.0.jar

File:

• main/trunk/greenstone2/perllib/plugins/PDFPlugin.pm (modified) (4 diffs)

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

@@ -261 +261 @@
     $self->read_file ($conv_filename, "utf8", "", \$text);
 
+    # To support the use_sections option with PDFBox: Greenstone splits PDFs into pages for 
+    # sections. The PDFPlugin code wants each new page to be prefixed with <a name=pagenum></a>,
+    # which it then splits on to generate page-based sections. However, that's not what PDFBox 
+    # generates in its HTML output. Fortunately, PDFBox does have its own page-separator: it
+    # embeds each page in an extra div. The div opener is: 
+    # <div style=\"page-break-before:always; page-break-after:always\">
+    # The PDFPlugin now looks for this and prefix <a name=0></a> to each such div. (The 
+    # pagenumber is fixed at 0 since I'm unable to work out how to increment the pagenum during 
+    # a regex substitution even with regex extensions on.) Later, when we process each section 
+    # to get the pagenum, PDFBox's output for this is pre-processed by having a loopcounter 
+    # that increments the pagenum for each subsequent section.
+
+    #$pdfbox_pageheader="\<div style=\"page-break-before:always; page-break-after:always\">";
+    my $loopcounter = 0; # used later on!
+    $text =~ s@\<div style=\"page-break-before:always; page-break-after:always\">@<a name=$loopcounter></a><div style=\"page-break-before:always; page-break-after:always\">@g;
+
+
     # Calculate number of pages based on <a ...> tags (we have a <a name=1> etc
     # for each page).  Metadata based on this calculation not set until process()
@@ -267 +284 @@
     # be useful to give an indication of document length in browser through setting
     # num_pages as metadata.
-    my @pages = ($text =~ /\<[Aa] name=\"?\w+\"?>/ig);
+    my @pages = ($text =~ m/(\<[Aa] name=\"?\w+\"?>|\<div style=\"page-break-before:always; page-break-after:always\">)/ig); #<div style=\"?page-break-before:always; page-break-after:always\"?>
     my $num_pages = scalar(@pages);
     $self->{'num_pages'} = $num_pages;
@@ -317 +334 @@
 
         $title = $1; # Greenstone does magic if sections are titled digits
+
+        # A title of pagenum=0 means use_sections is being applied on output from PDFBox,
+        # which didn't originally have a <a name=incremented pagenumber></a> to split each page. 
+        # Our Perl code then prefixed <a name=0></a> to it. Now need to increment the pagenum here:
+        if($loopcounter > 0 || ($title eq 0 && $loopcounter == 0)) { # implies use_sections with PDFBox
+        $title = ++$loopcounter;
+        }
+
         if (! defined($title) ) {
         print STDERR "no title: $section\n";
@@ -323 +348 @@
         my $newsection = "<!-- from PDFPlugin -->\n<!-- <Section>\n";
         $newsection .= "<Metadata name=\"Title\">" . $title
-        . "</Metadata>\n--><p>\n";
+        . "</Metadata>\n--><br />\n";
         $newsection .= $section;
         $newsection .= "<!--</Section>-->\n";