Preserve order of nested headers in TOC

Fixes Python-Markdown#380. The TOC replacement is now seperate from building the TOC, which allows us to iter through the doc in order rather than with the non-order-preserving interparent pethod. This is almost a complete refactor of the run method. Also cleaned up the config stuff. Thanks to @colewerner for the report and test case.

Author: waylan

markdown/extensions/toc.py

@@ -86,12 +86,44 @@ def order_toc_list(toc_list):
 
 
 class TocTreeprocessor(Treeprocessor):
+ def __init__(self, md, config):
+ super(TocTreeprocessor, self).__init__(md)
+ 
+ self.marker = config["marker"]
+ self.title = config["title"]
+ self.slugify = config["slugify"]
+ self.use_anchors = parseBoolValue(config["anchorlink"])
+ self.use_permalinks = parseBoolValue(config["permalink"], False)
+ if self.use_permalinks is None:
+ self.use_permalinks = config["permalink"]
+
+ self.header_rgx = re.compile("[Hh][123456]")
 
- # Iterator wrapper to get parent and child all at once
  def iterparent(self, root):
- for parent in root.getiterator():
+ ''' Iterator wrapper to get parent and child all at once. '''
+ for parent in root.iter():
  for child in parent:
  yield parent, child
+ 
+ def replace_marker(self, root, elem):
+ ''' Replace marker with elem. '''
+ for (p, c) in self.iterparent(root):
+ text = ''.join(itertext(c)).strip()
+ if not text:
+ continue
+
+ # To keep the output from screwing up the
+ # validation by putting a <div> inside of a <p>
+ # we actually replace the <p> in its entirety.
+ # We do not allow the marker inside a header as that
+ # would causes an enless loop of placing a new TOC
+ # inside previously generated TOC.
+ if c.text and c.text.strip() == self.marker and \
+ not self.header_rgx.match(c.tag) and c.tag not in ['pre', 'code']:
+ for i in range(len(p)):
+ if p[i] == c:
+ p[i] = elem
+ break
 
  def add_anchor(self, c, elem_id): # @ReservedAssignment
  anchor = etree.Element("a")
@@ -116,10 +148,10 @@ def add_permalink(self, c, elem_id):
 
  def build_toc_etree(self, div, toc_list):
  # Add title to the div
- if self.config["title"]:
+ if self.title:
  header = etree.SubElement(div, "span")
  header.attrib["class"] = "toctitle"
- header.text = self.config["title"]
+ header.text = self.title
 
  def build_etree_ul(toc_list, parent):
  ul = etree.SubElement(parent, "ul")
@@ -136,62 +168,37 @@ def build_etree_ul(toc_list, parent):
  return build_etree_ul(toc_list, div)
 
  def run(self, doc):
-
- div = etree.Element("div")
- div.attrib["class"] = "toc"
- header_rgx = re.compile("[Hh][123456]")
-
- self.use_anchors = parseBoolValue(self.config["anchorlink"])
- self.use_permalinks = parseBoolValue(self.config["permalink"], False)
- if self.use_permalinks is None:
- self.use_permalinks = self.config["permalink"]
-
  # Get a list of id attributes
  used_ids = set()
- for c in doc.getiterator():
- if "id" in c.attrib:
- used_ids.add(c.attrib["id"])
+ for el in doc.iter():
+ if "id" in el.attrib:
+ used_ids.add(el.attrib["id"])
 
+ div = etree.Element("div")
+ div.attrib["class"] = "toc"
+ self.replace_marker(doc, div)
+ 
  toc_list = []
- for (p, c) in self.iterparent(doc):
- text = ''.join(itertext(c)).strip()
- if not text:
- continue
-
- # To keep the output from screwing up the
- # validation by putting a <div> inside of a <p>
- # we actually replace the <p> in its entirety.
- # We do not allow the marker inside a header as that
- # would causes an enless loop of placing a new TOC
- # inside previously generated TOC.
- if c.text and c.text.strip() == self.config["marker"] and \
- not header_rgx.match(c.tag) and c.tag not in ['pre', 'code']:
- for i in range(len(p)):
- if p[i] == c:
- p[i] = div
- break
-
- if header_rgx.match(c.tag):
-
+ for el in doc.iter():
+ if self.header_rgx.match(el.tag):
+ text = ''.join(itertext(el)).strip()
+ 
  # Do not override pre-existing ids
- if "id" not in c.attrib:
+ if "id" not in el.attrib:
  elem_id = stashedHTML2text(text, self.markdown)
- elem_id = unique(self.config["slugify"](elem_id, '-'),
- used_ids)
- c.attrib["id"] = elem_id
+ elem_id = unique(self.slugify(elem_id, '-'), used_ids)
+ el.attrib["id"] = elem_id
  else:
- elem_id = c.attrib["id"]
-
- tag_level = int(c.tag[-1])
+ elem_id = el.attrib["id"]
 
- toc_list.append({'level': tag_level,
+ toc_list.append({'level': int(el.tag[-1]),
  'id': elem_id,
  'name': text})
 
  if self.use_anchors:
- self.add_anchor(c, elem_id)
+ self.add_anchor(el, elem_id)
  if self.use_permalinks:
- self.add_permalink(c, elem_id)
+ self.add_permalink(el, elem_id)
 
  toc_list_nested = order_toc_list(toc_list)
  self.build_toc_etree(div, toc_list_nested)
@@ -235,8 +242,7 @@ def extendMarkdown(self, md, md_globals):
  md.registerExtension(self)
  self.md = md
  self.reset()
- tocext = self.TreeProcessorClass(md)
- tocext.config = self.getConfigs()
+ tocext = self.TreeProcessorClass(md, self.getConfigs())
  # Headerid ext is set to '>prettify'. With this set to '_end',
  # it should always come after headerid ext (and honor ids assinged
  # by the header id extension) if both are used. Same goes for

tests/extensions/test.cfg

@@ -36,6 +36,10 @@ toc_nested2:
  markdown.extensions.toc:
  permalink: "[link]"
 
+toc_nested_list:
+ extensions:
+ - markdown.extensions.toc
+
 wikilinks:
  extensions:
  - markdown.extensions.wikilinks

tests/extensions/toc_nested_list.html

@@ -0,0 +1,30 @@
+<h1 id="title">Title</h1>
+<div class="toc">
+<ul>
+<li><a href="#title">Title</a><ul>
+<li><a href="#section-1">Section 1</a><ul>
+<li><a href="#subsection-1">Subsection 1</a></li>
+<li><a href="#subsection-2">Subsection 2</a></li>
+</ul>
+</li>
+<li><a href="#section-2">Section 2</a></li>
+<li><a href="#section-3">Section 3</a></li>
+</ul>
+</li>
+</ul>
+</div>
+<h2 id="section-1">Section 1</h2>
+<ol>
+<li>
+<p>List Item 1</p>
+<h3 id="subsection-1">Subsection 1</h3>
+<p>Explanation 1</p>
+</li>
+<li>
+<p>List Item 2</p>
+<h3 id="subsection-2">Subsection 2</h3>
+<p>Explanation 2</p>
+</li>
+</ol>
+<h2 id="section-2">Section 2</h2>
+<h2 id="section-3">Section 3</h2>

tests/extensions/toc_nested_list.txt

@@ -0,0 +1,19 @@
+# Title
+
+[TOC]
+
+## Section 1
+
+1. List Item 1
+
+ ### Subsection 1
+ Explanation 1
+
+2. List Item 2
+
+ ### Subsection 2
+ Explanation 2
+
+## Section 2
+
+## Section 3