halstead volume measure

halstead.rb

#-*- encoding: utf-8

operator_list = [
  /\>\>\=/, /\<\<\=/, /\.\.\./, /\|\|/, /\!\=/, /\%\=/, /\&\&/, /\-\-/, /\-\=/,
  /\&\=/, /\*\=/, /\+\+/, /\+\=/, /\-\>/, /\^\=/, /\|\=/,
  /\/\=/, /\:\:/, /\<\</, /\<\=/, /\=\=/, /\>\=/, /\>\>/,
  /\*/, /\+/, /\(/, /\)/, /\!/, /\%/, /\&/,
  /\,/, /\-/, /\./, /\//, /\:/, /\=/, /\>/, /\</,
  /\?/, /\[/, /\]/, /\^/, /\{/, /\}/, /\|/, /\~/, /\;/
]

reserved_list = [
  /asm/, /break/, /case/, /class/,
  /continue/, /default/, /delete/,
  /do/, /else/, /enum/, /for/, /goto/,
  /if/, /new/, /operator/, /private/,
  /protected/, /public/, /return/,
  /sizeof/, /struct/, /switch/,
  /this/, /union/, /while/,
  /namespace/, /using/, /try/,
  /catch/, /throw/, /const_cast/,
  /static_cast/, /dynamic_cast/,
  /reinterpret_cast/, /typeid/,
  /template/, /explicit/, /true/,
  /false/, /typename/, /#include/,
  /#define/, /auto/, /extern/,
  /inline/, /register/, /static/,
  /typedef/, /virtual/, /mtuable/,
  /const/, /friend/, /volatile/
]

def MakeFileList()
  dir_list = ["rtaudio", "port-audio", "tinyasio", "asio-sdk"]
  file_list = []
  for dir in dir_list
    file_list << ["./tool/#{dir}/test1.cpp", dir]
    file_list << ["./tool/#{dir}/test2.cpp", dir]
  end
  file_list
end

def InitHash(hash, list)
  for name in list
    hash[name] = 0
  end
end

def RemoveComment(str)
  str.gsub(/\/\*[\s\S]*?\*\/|\/\/.*/, " ")
end

def WriteResult(fname, result)
  name = File::split(fname[0])[1]
  name = File::basename(name, ".*")
  open("./tool/#{fname[1]}/#{name}.csv", "w").write(result)
end

def Counting(list, hash, str)
  count = 0
  for name in list
    name = name.class == Array ? name[0] : name
    cnt = str.scan(name).size
    count += cnt
    hash[name] += cnt > 0 ? 1 : 0

    str.gsub!(name, "")
  end
  count
end

def CountOperator(a, b, aa, bb, str)
  Counting(a, aa, str) + Counting(b, bb, str)
end

def UniqueOperand(hash, reg)
  for r in reg
    hash[r] ||= 0
    hash[r] += 1
  end
end

def CountOperand(hash, str)
  literal = str.scan(/\d+/)
  str.gsub!(/\d+/, " ")

  identifier = str.scan(/\w+/)
  str.gsub!(/\w+/, " ")

  UniqueOperand(hash, literal)
  UniqueOperand(hash, identifier)

  literal.size + identifier.size
end

def CountUniqueOperator(a, b)
  count = 0
  count += a.values.inject{|sum, v| sum + (v > 0 ? 1 : 0)}
  count += b.values.inject{|sum, v| sum + (v > 0 ? 1 : 0)}
  count
end

def OpenFile(fname)
  str = open(fname, "r").read
  str.encode!("UTF-8", "Shift_JIS", {
    invalid: :replace, undef: :replace, replace: ""})
  RemoveComment(str)
end

file_list = MakeFileList()

for fname in file_list
  operator_hash = Hash.new
  reserved_hash = Hash.new
  operand_hash = Hash.new
  InitHash(operator_hash, operator_list)
  InitHash(reserved_hash, reserved_list)
  str = OpenFile(fname[0])

  puts fname[0]
  operator = CountOperator(reserved_list, operator_list, reserved_hash, operator_hash, str)
  operand = CountOperand(operand_hash, str)

  unique_operator = CountUniqueOperator(operator_hash, reserved_hash)
  unique_operand = operand_hash.size

  program_length = operator + operand
  vocabulary_size = unique_operator + unique_operand

  program_volume = program_length * Math.log2(vocabulary_size)

  difficulty_level = (unique_operator / 2.0) * (operand / unique_operand)

  effort_to_implement = program_volume * difficulty_level

  program_level = 1.0 / difficulty_level

  time_to_implement = effort_to_implement / 18.0

  number_of_delivered_bugs = (effort_to_implement ** (2.0/3.0)) / 3000.0

  calculated_program_length = unique_operator * Math.log2(unique_operator) + unique_operand * Math.log2(unique_operand)

  result = ""
  result += "Total Operator, #{operator}\n"
  result += "Total Operand, #{operand}\n"
  result += "Unique Operator, #{unique_operator}\n"
  result += "Unique Operand, #{unique_operand}\n"
  result += "Program Length, #{program_length}\n"
  result += "Vocabulary Size, #{vocabulary_size}\n"
  result += "Program Volume, #{program_volume}\n"
  result += "Difficulty, #{difficulty_level}\n"
  result += "Effort, #{effort_to_implement}\n"
  result += "Program Level, #{program_level}\n"
  result += "Time to Implement, #{time_to_implement}\n"
  result += "Number of Delivered Bugs, #{number_of_delivered_bugs}\n"
  result += "Calculated Program Length, #{calculated_program_length}\n"

  WriteResult(fname, result)
end