Profanity filter

profanity_benchmark.rb

require 'benchmark/ips'

text = 'Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aliquam ullamcorper eleifend velit sed placerat. In libero leo, fringilla ut mauris nec, bibendum varius nisi. Nam rhoncus facilisis lacinia. Integer massa quam, vestibulum eget est sed, elementum vulputate nibh. Mauris fermentum tellus eu commodo blandit. Cras ornare, risus id semper lacinia, est urna maximus dolor, aliquam maximus metus sem a velit. Suspendisse vel libero eu leo volutpat eleifend accumsan nec massa. Integer eu vulputate quam, ac tristique arcu. Fusce et justo vitae nulla maximus lacinia. Integer dolor massa, finibus vel mattis a, porta eu erat. Suspendisse potenti. Sed ex magna, imperdiet sit amet augue ac, vestibulum vehicula mi. Praesent sapien neque, bibendum eget lorem fermentum, molestie semper leo.'

words_10 = %w(foo bar string amp baguette marine grenadine quezac bambou panda)
words_100 = words_10 * 10
regexp_10 = Regexp.new(words_10.join('|'), Regexp::IGNORECASE)
regexp_100 = Regexp.new(words_100.join('|'), Regexp::IGNORECASE)
r_array_10 = words_10.map { |w| /\b#{w}\b/i }
r_array_100 = words_100.map { |w| /\b#{w}\b/i }

puts 'Worst case: no element found'
puts "Text size: #{text.size}"
Benchmark.ips do |x|
  x.report('regexp_10')   { (regexp_10 =~ text) != nil }
  x.report('regexp_100')  { (regexp_100 =~ text) != nil }

  x.report('regexp_loop_10') {
    r_array_10.each do |foul|
      break(true) if foul =~ text
    end
    false
  }
  x.report('regexp_loop_100') {
   r_array_100.each do |foul|
      break(true) if foul =~ text
    end
    false
  }

  x.report('array#&_10')   { (words_10 & text.split(/\W+/).map! { |w| w.downcase!; w }).empty? }
  x.report('array#&_100')  { (words_100 & text.split(/\W+/).map! { |w| w.downcase!; w }).empty? }

  x.report('array_loop_10') {
    text.split(/\W+/).each do |word|
      word.downcase!
      break(true) if words_10.include?(word)
    end
    false
  }
  x.report('array_loop_100') {
    text.split(/\W+/).each do |word|
      word.downcase!
      break(true) if words_100.include?(word)
    end
    false
  }
end

puts

words_10 = %w(integer bar string amp baguette marine grenadine quezac bambou panda)
words_100 = words_10 * 10
regexp_10 = Regexp.new(words_10.join('|'), Regexp::IGNORECASE)
regexp_100 = Regexp.new(words_100.join('|'), Regexp::IGNORECASE)
r_array_10 = words_10.map { |w| /\b#{w}\b/i }
r_array_100 = words_100.map { |w| /\b#{w}\b/i }

puts 'Best case: first element match'
puts "Text size: #{text.size}"
Benchmark.ips do |x|
  x.report('regexp_10')   { (regexp_10 =~ text) != nil }
  x.report('regexp_100')  { (regexp_100 =~ text) != nil }

  x.report('regexp_loop_10') {
    r_array_10.each do |foul|
      break(true) if foul =~ text
    end
    false
  }
  x.report('regexp_loop_100') {
   r_array_100.each do |foul|
      break(true) if foul =~ text
    end
    false
  }

  x.report('array#&_10')   { (words_10 & text.split(/\W+/).map! { |w| w.downcase!; w }).empty? }
  x.report('array#&_100')  { (words_100 & text.split(/\W+/).map! { |w| w.downcase!; w }).empty? }

  x.report('array_loop_10') {
    text.split(/\W+/).each do |word|
      word.downcase!
      break(true) if words_10.include?(word)
    end
    false
  }
  x.report('array_loop_100') {
    text.split(/\W+/).each do |word|
      word.downcase!
      break(true) if words_100.include?(word)
    end
    false
  }
end

# Benchmark.compare *result

Changing

r_array_10 = words_10.map { |w| /\b#{w}\b/i }
r_array_100 = words_100.map { |w| /\b#{w}\b/i }

for

r_array_10 = words_10.map { |w| /#{w}/i }
r_array_100 = words_100.map { |w| /#{w}/i }

To be equivalent.

Worst case: no element found
Text size: 789
Calculating -------------------------------------
           regexp_10     1.608k i/100ms
          regexp_100   216.000  i/100ms
      regexp_loop_10     3.939k i/100ms
     regexp_loop_100   400.000  i/100ms
          array#&_10   869.000  i/100ms
         array#&_100   785.000  i/100ms
       array_loop_10   752.000  i/100ms
      array_loop_100   163.000  i/100ms
-------------------------------------------------
           regexp_10     16.428k (± 4.2%) i/s -     83.616k
          regexp_100      2.149k (± 5.3%) i/s -     10.800k
      regexp_loop_10     40.490k (± 4.3%) i/s -    204.828k
     regexp_loop_100      4.011k (± 4.8%) i/s -     20.400k
          array#&_10      8.826k (± 4.7%) i/s -     44.319k
         array#&_100      8.102k (± 4.2%) i/s -     40.820k
       array_loop_10      7.459k (± 4.2%) i/s -     37.600k
      array_loop_100      1.633k (± 4.2%) i/s -      8.150k

Best case: first element match
Text size: 789
Calculating -------------------------------------
           regexp_10     4.982k i/100ms
          regexp_100   668.000  i/100ms
      regexp_loop_10    63.389k i/100ms
     regexp_loop_100    64.403k i/100ms
          array#&_10   879.000  i/100ms
         array#&_100   794.000  i/100ms
       array_loop_10     1.282k i/100ms
      array_loop_100   507.000  i/100ms
-------------------------------------------------
           regexp_10     50.999k (± 3.9%) i/s -    259.064k
          regexp_100      6.781k (± 5.1%) i/s -     34.068k
      regexp_loop_10      1.139M (± 5.2%) i/s -      5.705M
     regexp_loop_100      1.133M (± 6.3%) i/s -      5.667M
          array#&_10      8.770k (± 4.0%) i/s -     43.950k
         array#&_100      8.210k (± 4.4%) i/s -     41.288k
       array_loop_10     13.040k (± 4.3%) i/s -     65.382k
      array_loop_100      5.152k (± 5.5%) i/s -     25.857k