ruby regex w/lookahead benchmarks

benchmark.rb

require 'benchmark'
require 'nokogiri'

content = "#{'<div class="test" style="margin:10px;">' * 450}<p>Deep content</p>#{'</div>' * 450}"

DUPLICATE_NESTED_DIV_LOOKAHEAD_REGEX = %r{
  (?<outer_div>
    <div(?<attrs>[^>]*)>\s*               # Match the outermost <div> with its attributes
    (?:<div\k<attrs>>\s*)*                # Match zero or more nested identical <div>
    (?<inner_div><div\k<attrs>>.*?</div>) # Match the innermost <div> and its content
    (?:\s*</div>\s*)*                     # Match all closing </div> tags for the nested structure
  )
}xm

DUPLICATE_NESTED_DIV_REGEX = %r{
  <div(?<attrs>[^>]*)>\s*                        # Match an opening <div> and capture its attributes
  (?<inner_div><div\k<attrs>>\s*.*?\s*</div>) # Match a nested <div> with the same attributes and its content
  \s*</div>                                      # Match the outer closing </div>
}xm

def attributes_match?(node1, node2)
  node1.attributes.transform_values(&:value) == node2.attributes.transform_values(&:value)
end

Benchmark.bm(7) do |x|
  @looked = x.report('Looked:') do
    content.gsub(DUPLICATE_NESTED_DIV_LOOKAHEAD_REGEX) do
      duplicate_nested_div_matches = Regexp.last_match
      duplicate_nested_div_matches[:inner_div]
    end
  end

  @looped = x.report('Looped:') do
    new_content = content.dup

    loop do
      content_before = new_content.dup
      new_content = new_content.gsub(DUPLICATE_NESTED_DIV_REGEX) do
        duplicate_nested_div_matches = Regexp.last_match
        duplicate_nested_div_matches[:inner_div]
      end
      break if new_content == content_before
    end
  end

  @nokoed = x.report('Nokoed:') do
    doc = Nokogiri::HTML.fragment(content)

    loop do
      modified = false
      doc.css('div').reverse_each do |div|
        children = div.children.select { |child| child.element? && child.name == 'div' }

        if children.length.positive? && attributes_match?(children.first, div)
          div.replace(children.first)
          modified = true
        end
      end

      break unless modified
    end
  end
end

looked_real = @looked.real
looped_real = @looped.real
nokoed_real = @nokoed.real
relative_ll_speed = (looped_real / looked_real)
relative_ln_speed = (nokoed_real / looked_real)

puts "\nComparison:"
puts "Looked is #{relative_ll_speed.round(2)}x faster than Looped"
puts "Looked is #{relative_ln_speed.round(2)}x faster than Nokoed"