From 87290d89f91072ff8d4f5e975bef1224b455fd41 Mon Sep 17 00:00:00 2001
From: Christian Weiske <cweiske@cweiske.de>
Date: Wed, 3 Jun 2015 23:42:27 +0200
Subject: [PATCH]

---
 build-toc.php | 79 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 79 insertions(+)
 create mode 100644 build-toc.php

diff --git a/build-toc.php b/build-toc.php
new file mode 100644
index 0000000..f585731
--- /dev/null
+++ b/build-toc.php
@@ -0,0 +1,79 @@
+#!/usr/bin/env php
+<?php
+/**
+ * Creates a TOC (table of contents) from a given HTML file.
+ * <h1> is ignored; all other headings are used.
+ *
+ * @author Christian Weiske <cweiske@cweiske.de>
+ *
+ * Alternative if you have an XSLT 2 processor: http://xmlplease.com/tocxhtml
+ */
+if ($argc < 2) {
+    echo "Please pass a file\n";
+    exit(1);
+}
+$file = $argv[1];
+if (!file_exists($file)) {
+    echo "File does not exist: $file\n";
+    exit(2);
+}
+if (substr($file, -4) != '.htm') {
+    echo "build-toc.php works only on htm files\n";
+    exit(3);
+}
+
+$sx = simplexml_load_file($file);
+if ($sx === false) {
+    echo "Failed to load XML\n";
+    exit(4);    
+}
+
+$body = $sx->body;
+$body->registerXPathNamespace('h', 'http://www.w3.org/1999/xhtml');
+$elems = $body->xpath(
+    '//*['
+    //XHTML
+    . 'self::h:h2 or self::h:h3 or self::h:h4 or self::h:h5 or self::h:h6'
+    //HTML
+    . ' or self::h2 or self::h3 or self::h4 or self::h5 or self::h6'
+    . ']'
+);
+
+$data = array();
+$n = -1;
+foreach ($elems as $elem) {
+    $name = $elem->getName();
+    $level = (int) $name{1};
+    $data[] = (object) array(
+        'title' => (string) $elem,
+        'id'    => (string) $elem['id'],
+        'level' => $level,
+        'nextLevel' => null
+    );
+    if (++$n >= 1) {
+        $data[$n - 1]->nextLevel = $level;
+    }
+}
+
+$html = "<ul class=\"toc\">\n";
+foreach ($data as $h) {
+    $i = str_repeat(' ', ($h->level -1) * 2 - 1);
+    $html .= $i
+        . '<li><a href="#' . $h->id . '">'
+        . htmlspecialchars($h->title)
+        . '</a>';
+    if ($h->nextLevel > $h->level) {
+        $html .= "\n"
+            . $i . ' <ul>' . "\n";
+    } else {
+        $html .= '</li>' . "\n";
+    }
+    if ($h->nextLevel < $h->level) {
+        $html .= substr($i, 0, -1) . "</ul>\n"
+            . substr($i, 0, -2) . "</li>\n";
+    }
+}
+
+$html .= "</ul>\n";
+echo $html;
+?>
-- 
2.30.2