123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253 |
- # markdown is released under the BSD license
- # Copyright 2007, 2008 The Python Markdown Project (v. 1.7 and later)
- # Copyright 2004, 2005, 2006 Yuri Takhteyev (v. 0.2-1.6b)
- # Copyright 2004 Manfred Stienstra (the original version)
- #
- # All rights reserved.
- #
- # Redistribution and use in source and binary forms, with or without
- # modification, are permitted provided that the following conditions are met:
- #
- # * Redistributions of source code must retain the above copyright
- # notice, this list of conditions and the following disclaimer.
- # * Redistributions in binary form must reproduce the above copyright
- # notice, this list of conditions and the following disclaimer in the
- # documentation and/or other materials provided with the distribution.
- # * Neither the name of the <organization> nor the
- # names of its contributors may be used to endorse or promote products
- # derived from this software without specific prior written permission.
- #
- # THIS SOFTWARE IS PROVIDED BY THE PYTHON MARKDOWN PROJECT ''AS IS'' AND ANY
- # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- # WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- # DISCLAIMED. IN NO EVENT SHALL ANY CONTRIBUTORS TO THE PYTHON MARKDOWN PROJECT
- # BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- # CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- # SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- # INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- # CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- # ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- # POSSIBILITY OF SUCH DAMAGE.
- """
- Table of Contents Extension for Python-Markdown
- * * *
- (c) 2008 [Jack Miller](http://codezen.org)
- Dependencies:
- * [Markdown 2.1+](http://packages.python.org/Markdown/)
- """
- from __future__ import absolute_import
- from __future__ import unicode_literals
- from . import Extension
- from ..treeprocessors import Treeprocessor
- from ..util import etree
- from .headerid import slugify, unique, itertext
- import re
- def order_toc_list(toc_list):
- """Given an unsorted list with errors and skips, return a nested one.
- [{'level': 1}, {'level': 2}]
- =>
- [{'level': 1, 'children': [{'level': 2, 'children': []}]}]
-
- A wrong list is also converted:
- [{'level': 2}, {'level': 1}]
- =>
- [{'level': 2, 'children': []}, {'level': 1, 'children': []}]
- """
-
- def build_correct(remaining_list, prev_elements=[{'level': 1000}]):
-
- if not remaining_list:
- return [], []
-
- current = remaining_list.pop(0)
- if not 'children' in current.keys():
- current['children'] = []
-
- if not prev_elements:
- # This happens for instance with [8, 1, 1], ie. when some
- # header level is outside a scope. We treat it as a
- # top-level
- next_elements, children = build_correct(remaining_list, [current])
- current['children'].append(children)
- return [current] + next_elements, []
-
- prev_element = prev_elements.pop()
- children = []
- next_elements = []
- # Is current part of the child list or next list?
- if current['level'] > prev_element['level']:
- #print "%d is a child of %d" % (current['level'], prev_element['level'])
- prev_elements.append(prev_element)
- prev_elements.append(current)
- prev_element['children'].append(current)
- next_elements2, children2 = build_correct(remaining_list, prev_elements)
- children += children2
- next_elements += next_elements2
- else:
- #print "%d is ancestor of %d" % (current['level'], prev_element['level'])
- if not prev_elements:
- #print "No previous elements, so appending to the next set"
- next_elements.append(current)
- prev_elements = [current]
- next_elements2, children2 = build_correct(remaining_list, prev_elements)
- current['children'].extend(children2)
- else:
- #print "Previous elements, comparing to those first"
- remaining_list.insert(0, current)
- next_elements2, children2 = build_correct(remaining_list, prev_elements)
- children.extend(children2)
- next_elements += next_elements2
-
- return next_elements, children
-
- ordered_list, __ = build_correct(toc_list)
- return ordered_list
- class TocTreeprocessor(Treeprocessor):
-
- # Iterator wrapper to get parent and child all at once
- def iterparent(self, root):
- for parent in root.getiterator():
- for child in parent:
- yield parent, child
-
- def add_anchor(self, c, elem_id): #@ReservedAssignment
- if self.use_anchors:
- anchor = etree.Element("a")
- anchor.text = c.text
- anchor.attrib["href"] = "#" + elem_id
- anchor.attrib["class"] = "toclink"
- c.text = ""
- for elem in c.getchildren():
- anchor.append(elem)
- c.remove(elem)
- c.append(anchor)
-
- def build_toc_etree(self, div, toc_list):
- # Add title to the div
- if self.config["title"]:
- header = etree.SubElement(div, "span")
- header.attrib["class"] = "toctitle"
- header.text = self.config["title"]
- def build_etree_ul(toc_list, parent):
- ul = etree.SubElement(parent, "ul")
- for item in toc_list:
- # List item link, to be inserted into the toc div
- li = etree.SubElement(ul, "li")
- link = etree.SubElement(li, "a")
- link.text = item.get('name', '')
- link.attrib["href"] = '#' + item.get('id', '')
- if item['children']:
- build_etree_ul(item['children'], li)
- return ul
-
- 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 = self.config["anchorlink"] in [1, '1', True, 'True', 'true']
-
- # 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"])
- toc_list = []
- marker_found = False
- 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
- marker_found = True
-
- if header_rgx.match(c.tag):
-
- # Do not override pre-existing ids
- if not "id" in c.attrib:
- elem_id = unique(self.config["slugify"](text, '-'), used_ids)
- c.attrib["id"] = elem_id
- else:
- elem_id = c.attrib["id"]
- tag_level = int(c.tag[-1])
-
- toc_list.append({'level': tag_level,
- 'id': elem_id,
- 'name': text})
-
- self.add_anchor(c, elem_id)
-
- toc_list_nested = order_toc_list(toc_list)
- self.build_toc_etree(div, toc_list_nested)
- prettify = self.markdown.treeprocessors.get('prettify')
- if prettify: prettify.run(div)
- if not marker_found:
- # serialize and attach to markdown instance.
- toc = self.markdown.serializer(div)
- for pp in self.markdown.postprocessors.values():
- toc = pp.run(toc)
- self.markdown.toc = toc
- class TocExtension(Extension):
-
- TreeProcessorClass = TocTreeprocessor
-
- def __init__(self, configs=[]):
- self.config = { "marker" : ["[TOC]",
- "Text to find and replace with Table of Contents -"
- "Defaults to \"[TOC]\""],
- "slugify" : [slugify,
- "Function to generate anchors based on header text-"
- "Defaults to the headerid ext's slugify function."],
- "title" : [None,
- "Title to insert into TOC <div> - "
- "Defaults to None"],
- "anchorlink" : [0,
- "1 if header should be a self link"
- "Defaults to 0"]}
- for key, value in configs:
- self.setConfig(key, value)
- def extendMarkdown(self, md, md_globals):
- tocext = self.TreeProcessorClass(md)
- tocext.config = 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
- # attr_list extension. This must come last because we don't want
- # to redefine ids after toc is created. But we do want toc prettified.
- md.treeprocessors.add("toc", tocext, "_end")
- def makeExtension(configs={}):
- return TocExtension(configs=configs)
|