A ruby class to extract h1/h2/h3/... headings from HTML

html_heading_parser.rb

#!/usr/bin/env ruby

class HTMLHeadingParser
  
  class TreeNode
    attr_accessor :content, :children, :parent, :level

    def initialize(content, parent=nil)
      @content = content
      @children = []
      self.parent = parent
    end

    def parent=(p)
      @parent = p
      @level = (parent.level + 1) rescue 0
    end

    def to_s(indent=4)
      " " * (indent * @level) + content + "\n" + @children.map { |c| c.to_s(indent) } * ""
    end

    def ancestor(deepth=-1)
      ancst = self
      (deepth+1).times { ancst = ancst.parent }
      ancst
    end
  end

  class HTMLHeading
    attr_accessor :content, :text, :level

    def initialize(content, level)
      @content = content
      @level = level.to_i
      parse_content(content)
    end

    def content=(content)
      @content = content
      parse_content(content)
    end

    def parse_content(content)
      regex = %r{>?([^><]*)(</)?}
      @text = content.scan(regex).map(&:first).join
    end

    def to_str
      text
    end
  end
  
  def extract_headings(html)
    regex = %r{<h([1-6])[^>]*>(.*?)</h\1>}
    html.scan(regex).map do |(level, content)|
      HTMLHeading.new(content, level)
    end
  end

  def parse(html)
    root = TreeNode.new("")
    previous = root
    headings = extract_headings(html)
    headings.each do |heading|
      deepth = previous.level - heading.level
      next if  deepth < -1
      parent = previous.ancestor(deepth)
      this_node = TreeNode.new(heading, parent)
      previous.ancestor(deepth).children << this_node
      previous = this_node
    end
    root
  end

end

if __FILE__ == $0
  html = File.read(ARGV[0]).gsub("\n", "  ")
  puts HTMLHeadingParser.new.parse(html)
end