From: Norman Walsh <ndw@nwalsh.com>
Date: Wed, 21 Aug 2002 13:58:39 +0000 (+0000)
Subject: Extract indexes from PDF and merge page number ranges
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=330e55ff0ab994b1617b43ab486c6dd5c38222c4;p=docbook-dsssl

Extract indexes from PDF and merge page number ranges
---

diff --git a/xsl/fo/pdf2index b/xsl/fo/pdf2index
new file mode 100755
index 000000000..2a8439852
--- /dev/null
+++ b/xsl/fo/pdf2index
@@ -0,0 +1,114 @@
+#!/usr/bin/perl -- # -*- Perl -*-
+
+# this needs some cleanup...
+
+my $PSTOTEXT = "pstotext";
+
+my $pdf = shift @ARGV;
+
+my $index = "";
+my $inindex = 0;
+open (F, "$PSTOTEXT $pdf |");
+while (<F>) {
+    if (/^<\/index/) {
+	$index .= $_;
+	$inindex = 0;
+    }
+    $inindex = 1 if /^<index/;
+
+    if ($inindex) {
+	$index .= $_ if /^\s*</;
+    }
+}
+
+my $cindex = "";
+while ($index =~ /^(.*?)((<phrase role=\"pageno\">.*?<\/phrase>\s*)+)/s) {
+    $cindex .= $1;
+    $_ = $2;
+    $index = $'; # '
+
+    my @pages = m/<phrase role=\"pageno\">.*?<\/phrase>\s*/sg;
+
+    # Remove duplicates...
+    if ($#pages > 0) {
+	my @mpages = ();
+	my $current = "";
+	foreach my $page (@pages) {
+	    my $pageno = &pageno($page);
+	    if ($pageno ne $current) {
+		push (@mpages, $page);
+		$current = $pageno;
+	    }
+	}
+	@pages = @mpages;
+    }
+
+    # Collapse ranges...
+    if ($#pages > 1) {
+	my @cpages = ();
+	while (@pages) {
+	    my $count = 0;
+	    my $len = &rangelen($count, @pages);
+	    if ($len <= 2) {
+		my $page = shift @pages;
+		push (@cpages, $page);
+	    } else {
+		my $fpage = shift @pages;
+		my $lpage = "";
+		while ($len > 1) {
+		    $lpage = shift @pages;
+		    $len--;
+		}
+		my $fpno = &pageno($fpage);
+		my $lpno = &pageno($lpage);
+		$fpage =~ s/>$fpno</>${fpno}-$lpno</s;
+		push (@cpages, $fpage);
+	    }
+	}
+	@pages = @cpages;
+    }
+
+    my $page = shift @pages;
+    $page =~ s/\s*$//s;
+    $cindex .= $page;
+    while (@pages) {
+	$page = shift @pages;
+	$page =~ s/\s*$//s;
+	$cindex .= ", $page";
+    }
+}
+$cindex .= $index;
+
+print "$cindex\n";
+
+sub pageno {
+    my $page = shift;
+
+    $page =~ s/^<phrase.*?>//;
+    $page =~ s/^<link.*?>//;
+
+    return $1 if $page =~ /^([^<>]+)/;
+    return "?";
+}
+
+sub rangelen {
+    my $count = shift;
+    my @pages = @_;
+    my $len = 1;
+    my $inrange = 1;
+
+    my $current = &pageno($pages[$count]);
+    while ($count < $#pages && $inrange) {
+	$count++;
+	my $next = &pageno($pages[$count]);
+	if ($current + 1 eq $next) {
+	    $current = $next;
+	    $inrange = 1;
+	    $len++;
+	} else {
+	    $inrange = 0;
+	}
+    }
+
+    return $len;
+}