require 'strscan'


module ODDB
  module Util
    module IsoLatin1
      DOWNCASE_PAIRS = {
        "Å" => "å", "Æ" => "æ", "Ä" => "ä", "Á" => "á", "Â" => "â",
        "À" => "à", "Ã" => "ã", "Ą" => "ą", "Ǎ" => "ǎ", "Ă" => "ă",
        "Ā" => "ā", "Ȧ" => "ȧ", "Ḃ" => "ḃ", "Ç" => "ç", "Ĉ" => "ĉ",
        "Č" => "č", "Ć" => "ć", "Ċ" => "ċ", "Ḑ" => "ḑ", "Đ" => "đ",
        "Ð" => "ð", "Ď" => "ď", "Ḋ" => "ḋ", "Ë" => "ë", "É" => "é",
        "Ê" => "ê", "È" => "è", "Ȩ" => "ȩ", "Ę" => "ę", "Ě" => "ě",
        "Ĕ" => "ĕ", "Ẽ" => "ẽ", "Ē" => "ē", "Ė" => "ė", "Þ" => "þ",
        "Ḟ" => "ḟ", "Ģ" => "ģ", "Ǧ" => "ǧ", "Ğ" => "ğ", "Ǵ" => "ǵ",
        "Ĝ" => "ĝ", "Ḡ" => "ḡ", "Ġ" => "ġ", "Ȟ" => "ȟ", "Ĥ" => "ĥ",
        "Ḧ" => "ḧ", "Ḩ" => "ḩ", "Ḣ" => "ḣ", "Ï" => "ï", "Í" => "í",
        "Î" => "î", "Ì" => "ì", "Į" => "į", "Ǐ" => "ǐ", "Ĭ" => "ĭ",
        "Ĩ" => "ĩ", "İ" => "ı", "Ĵ" => "ĵ", "Ǩ" => "ǩ", "Ḱ" => "ḱ",
        "Ķ" => "ķ", "Ł" => "ł", "Ĺ" => "ĺ", "Ľ" => "ľ", "Ļ" => "ļ",
        "Ḿ" => "ḿ", "Ṁ" => "ṁ", "Ň" => "ň", "Ń" => "ń", "Ñ" => "ñ",
        "Ǹ" => "ǹ", "Ņ" => "ņ", "Ṅ" => "ṅ", "Œ" => "œ", "Ö" => "ö",
        "Ó" => "ó", "Ô" => "ô", "Ò" => "ò", "Õ" => "õ", "Ō" => "ō",
        "Ŏ" => "ŏ", "Ø" => "ø", "Ǫ" => "ǫ", "Ǒ" => "ǒ", "Ȯ" => "ȯ",
        "Ṕ" => "ṕ", "Ṗ" => "ṗ", "Ř" => "ř", "Ŕ" => "ŕ", "Ŗ" => "ŗ",
        "Ṙ" => "ṙ", "Ś" => "ś", "Ŝ" => "ŝ", "Š" => "š", "Ş" => "ş",
        "Ṡ" => "ṡ", "Ť" => "ť", "Ţ" => "ţ", "Ṫ" => "ṫ", "Ü" => "ü",
        "Ú" => "ú", "Û" => "û", "Ù" => "ù", "Ų" => "ų", "Ǘ" => "ǘ",
        "Ǔ" => "ǔ", "Ǚ" => "ǚ", "Ǜ" => "ǜ", "Ũ" => "ũ", "Ŭ" => "ŭ",
        "Ů" => "ů", "Ǖ" => "ǖ", "Ṽ" => "ṽ", "Ẃ" => "ẃ", "Ŵ" => "ŵ",
        "Ẁ" => "ẁ", "Ẅ" => "ẅ", "Ẇ" => "ẇ", "Ẍ" => "ẍ", "Ẋ" => "ẋ",
        "Ÿ" => "ÿ", "Ẏ" => "ẏ", "Ỹ" => "ỹ", "Ỳ" => "ỳ", "Ŷ" => "ŷ",
        "Ý" => "ý", "Ȳ" => "ȳ", "Ž" => "ž", "Ź" => "ź", "Ẑ" => "ẑ",
        "Ż" => "ż"
      }
      DOWNCASE_PTRN = /[#{DOWNCASE_PAIRS.keys.join}]/u
    end
  end
end

def _parse_size(size)
  unit_pattern = /(([kmµucMG]?([glLJm]|mol|Bq)\b)(\/([mµu]?[glL])\b)?)|((Mio\s)?U\.?I\.?)|(%( [mV]\/[mV])?)|(I\.E\.)|(Fl\.)/
  numeric_pattern = /\d+(\'\d+)*([.,]\d+)?/
  isolatin1 = ODDB::Util::IsoLatin1::DOWNCASE_PAIRS.values.join
  iso_pattern = /[[:alpha:]()\-#{isolatin1}]+/

  description = /(?!#{unit_pattern}\s)#{iso_pattern}(\s+#{iso_pattern})*/u
  numeric     = /#{numeric_pattern}/u
  unit        = /#{unit_pattern}/u

  count     = /(?<je>je)?\s*(?<numeric>#{numeric})/
  multiple  = /(?<numeric>#{numeric})\s*(?<unit>#{unit})?\s*(?<set>[xXà]|Set)/
  measure   = /(?<numeric>#{numeric})\s*(?<unit1>#{unit})\s*(?<unit2>#{unit})?/
  addition  = /#{numeric}\s*#{unit}?\s*\+/
  scale     = /\/s*#{numeric}?\s*#{unit}/
  dose      = /\(\s*#{numeric}\s*#{unit}\s*\)/

  s = StringScanner.new(size)
  until s.eos?
    s.skip(/\s+/)
    case
    when s.scan(/#{addition}/)
      s_addition = s[0]
    when s.scan(/#{measure}/)
      m = s[0].match(/#{measure}/)
      s_measure = [m[:numeric], m[:unit1], m[:unit2]]
    when s.scan(/#{count}/)
      m = s[0].match(/#{count}/)
      s_count = [m[:je], m[:numeric]]
    when s.scan(/#{multiple}/)
      m = s[0].match(/#{multiple}/)
      s_multi = [m[:numeric], m[:unit], m[:set]]
    when s.scan(/#{scale}/)
      s_scale = s[0]
    when s.scan(/#{dose}/)
      s_dose = s[0]
    when s.scan(/#{description}/)
      s_comform = s[0]
    when s.scan(/.*/)
    end
  end
  s_multi = s_multi ? s_multi : []
  [s_multi, s_addition, s_count, s_measure, s_scale, s_dose, s_comform]
end

str_list = [
  '9 Suppositorien',
  '10 ',
  '200 ml',
  '10x200 ml',
#  '5 Tüchlein',
#  '10 Set',
  #'10 x 5 Mio I.E.',
]

#File.readlines('size.dat').each do |str|
str_list.each do |str|
  print "%-25s: " % str.chomp.inspect
  p _parse_size(str)
end