我刚刚为python-markdown编写了一个扩展，它使用它的解析器来检索标题，并将TOC输出为带有本地链接的markdown格式的无序列表。文件是

Md_toc.py(是Md_toc.py)

．.． 它应该放在markdown安装的markdown/extensions/目录中。然后，你所要做的就是输入anchor <a> tags，带一个id="…"属性作为引用-对于这样的输入文本:

$ cat test.md 
Hello
=====

## <a id="sect one"></a>SECTION ONE ##

something here

### <a id='sect two'>eh</a>SECTION TWO ###

something else

#### SECTION THREE

nothing here

### <a id="four"></a>SECTION FOUR

also...

．.． 扩展可以这样调用:

$ python -m markdown -x md_toc test.md 
* Hello
    * [SECTION ONE](#sect one)
        * [SECTION TWO](#sect two)
            * SECTION THREE
        * [SECTION FOUR](#four)

．.． 然后您可以将这个toc粘贴回您的标记文档中(或者在您的文本编辑器中有一个快捷方式，在当前打开的文档上调用脚本，然后将结果toc插入到同一文档中)。

注意，旧版本的python-markdown没有__main__.py模块，因此，上面的命令行调用将不适用于这些版本。

您可以尝试使用这个ruby脚本从标记文件生成TOC。

 #!/usr/bin/env ruby

require 'uri'

fileName = ARGV[0]
fileName = "README.md" if !fileName

File.open(fileName, 'r') do |f|
  inside_code_snippet = false
  f.each_line do |line|
    forbidden_words = ['Table of contents', 'define', 'pragma']
    inside_code_snippet = !inside_code_snippet if line.start_with?('```')
    next if !line.start_with?("#") || forbidden_words.any? { |w| line =~ /#{w}/ } || inside_code_snippet

    title = line.gsub("#", "").strip
    href = URI::encode title.gsub(" ", "-").downcase
    puts "  " * (line.count("#")-1) + "* [#{title}](\##{href})"
  end
end

基于albertodebortoli answer创建了附加检查和标点符号替换功能。

# @fn       def generate_table_of_contents markdown # {{{
# @brief    Generates table of contents for given markdown text
#
# @param    [String]  markdown Markdown string e.g. File.read('README.md')
#
# @return   [String]  Table of content in markdown format.
#
def generate_table_of_contents markdown
  table_of_contents = ""
  i_section = 0
  # to track markdown code sections, because e.g. ruby comments also start with #
  inside_code_section = false
  markdown.each_line do |line|
    inside_code_section = !inside_code_section if line.start_with?('```')

    forbidden_words = ['Table of contents', 'define', 'pragma']
    next if !line.start_with?('#') || inside_code_section || forbidden_words.any? { |w| line =~ /#{w}/ }

    title = line.gsub("#", "").strip
    href = title.gsub(/(^[!.?:\(\)]+|[!.?:\(\)]+$)/, '').gsub(/[!.,?:; \(\)-]+/, "-").downcase

    bullet = line.count("#") > 1 ? " *" : "#{i_section += 1}."
    table_of_contents << "  " * (line.count("#") - 1) + "#{bullet} [#{title}](\##{href})\n"
  end
  table_of_contents
end

# Table of Contents
1. [Example](#example)
2. [Example2](#example2)
3. [Third Example](#third-example)

## Example [](#){name=example}
## Example2 [](#){name=example2}
## [Third Example](#){name=third-example}

如果您使用markdown extra，不要忘记您可以为链接、标题、代码围栏和图像添加特殊属性。 https://michelf.ca/projects/php-markdown/extra/#spe-attr

正如在其他回答中提到的，有多种方法可以自动生成目录。大多数都是开源软件，可以根据您的需要进行调整。

然而，我所缺少的是，使用Markdown提供的有限选项，在视觉上吸引人的目录格式。我们得出了以下结论:

Code

## Content

**[1. Markdown](#heading--1)**

  * [1.1. Markdown formatting cheatsheet](#heading--1-1)
  * [1.2. Markdown formatting details](#heading--1-2)

**[2. BBCode formatting](#heading--2)**

  * [2.1. Basic text formatting](#heading--2-1)

      * [2.1.1. Not so basic text formatting](#heading--2-1-1)

  * [2.2. Lists, Images, Code](#heading--2-2)
  * [2.3. Special features](#heading--2-3)

----

在你的文档中，你可以像这样放置目标子部分标记:

<div id="heading--1-1"/>
### 1.1. Markdown formatting cheatsheet

根据你在哪里以及如何使用Markdown，以下也应该工作，并提供了更好看Markdown代码:

### 1.1. Markdown formatting cheatsheet <a name="heading--1-1"/>

例子呈现

内容 1. 减价 1.1. 折算格式备忘单 1.2. 折扣价格式细节 2. BBCode格式化 2.1. 基本文本格式 2.1.1. 不是最基本的文本格式 2.2. 列表，图像，代码 2.3. 特殊功能

优势

You can add as many levels of chapters and sub-chapters as you need. In the Table of Contents, these would appear as nested unordered lists on deeper levels. No use of ordered lists. These would create an indentation, would not link the number, and cannot be used to create decimal classification numbering like "1.1.". No use of lists for the first level. Here, using an unordered list is possible, but not necessary: the indentation and bullet just add visual clutter and no function here, so we don't use a list for the first ToC level at all. Visual emphasis on the first-level sections in the table of content by bold print. Short, meaningful subpart markers that look "beautiful" in the browser's URL bar such as #heading--1-1 rather than markers containing transformed pieces of the actual heading. The code uses H2 headings (## …) for sections, H3 headings (### …) for sub-headings etc.. This makes the source code easier to read because ## … provides a stronger visual clue when scrolling through compared to the case where sections would start with H1 headings (# …). It is still logically consistent as you use the H1 heading for the document title itself. Finally, we add a nice horizontal rule to separate the table of contents from the actual content.

有关这项技术以及我们如何在论坛软件Discourse中使用它的更多信息，请参见这里。

这是一个小的nodejs脚本，它生成目录，并考虑重复的标题:

const fs = require('fs')
const { mdToPdf } = require('md-to-pdf');

const stringtoreplace = '<toc/>'

const processTitleRepetitions = (contents, titleMap) => {
  for (const content of contents) {
    titleMap[content.link] = typeof titleMap[content.link] === 'undefined'
      ? 0
      : titleMap[content.link] + 1
    if (titleMap[content.link] > 0) {
      content.link = `${content.link}-${titleMap[content.link]}`
    }
  }
}

const convertContentToPdf = async (targetFile) => {
  const pdf = await mdToPdf({path: targetFile}).catch(console.error)
  if(pdf) {
    const pdfFile = `${targetFile.replace(/\.md/, '')}.pdf`
    fs.writeFileSync(pdfFile, pdf.content)
    return pdfFile
  }
  throw new Error("PDF generation failed.")
}

const generateTOC = (file, targetFile) => {
  // Extract headers
  const fileContent = fs.readFileSync(file, 'utf-8')
  const titleLine = /((?<=^)#+)\s(.+)/
  const contents = fileContent.split(/\r?\n/).
    map(line => line.match(titleLine)).
    filter(match => match).
    filter(match => match[1].length > 1).
    map(regExpMatchArray => {
      return {
        level: regExpMatchArray[1].length, text: regExpMatchArray[2],
        link: '#' + regExpMatchArray[2].replace(/(\s+|[.,\/#!$%^&*;:{}=\-_`~()]+)/g, '-').toLowerCase(),
      }
    })
  const titleMap = {}
  processTitleRepetitions(contents, titleMap)
  // Write content
  let toctext = '## Table of Contents\n'
  // Find the toplevel to adjust the level of the table of contents.
  const topLevel = contents.reduce((maxLevel, content) => Math.min(content['level'], maxLevel), 1000)
  levelCounter = {}
  contents.forEach(item => {
    let currentLevel = parseInt(item.level)
    levelCounter[currentLevel] = levelCounter[currentLevel] ? levelCounter[currentLevel] + 1 : 1
    Object.entries(levelCounter).forEach(e => {
      if(currentLevel < parseInt(e[0])) {
        levelCounter[e[0]] = 0
      }
    })
    const level = Array(currentLevel - topLevel).fill('\t').join('')
    const text = `${levelCounter[currentLevel]}. [${item['text']}](${item['link']}) \n`
    toctext += level + text
  })

  const updatedContent = fileContent.toString().replace(stringtoreplace, toctext)
  fs.writeFileSync(targetFile, updatedContent)
  convertContentToPdf(targetFile).then((pdfFile) => {
    console.info(`${pdfFile} has been generated.`)
  })
}

const args = process.argv.slice(2)

if(args.length < 2) {
  console.error("Please provide the name of the markdown file from which the headers should be extracted and the name of the file with the new table of contents")
  console.info("Example: node MD_TOC.js <source_md> <target_md>")
  process.exit(1)
}

const source_md = args[0]
const target_md = args[1]

generateTOC(source_md, target_md)

要使用它，您需要在markdown文件中注入<toc/>。

下面是你如何使用它:

generateTOC('../README.md', '../README_toc.md')

第一个参数是源markdown文件，第二个参数是带有markdown的文件。