lib/porter2stemmer/constants.rb

   1 # coding: utf-8
   2
   3 # Constants for the Porter 2 stemmer
   4 module Porter2
   5
   6   # A non-vowel
   7   C = "[^aeiouy]"
   8
   9   # A vowel: a e i o u y
  10   V = "[aeiouy]"
  11
  12   # A non-vowel other than w, x, or Y
  13   CW = "[^aeiouywxY]"
  14
  15   # Doubles created when adding a suffix: these are undoubled when stemmed
  16   Double = "(bb|dd|ff|gg|mm|nn|pp|rr|tt)"
  17
  18   # A valid letter that can come before 'li' (or 'ly')
  19   Valid_LI = "[cdeghkmnrt]"
  20
  21   # A specification for a short syllable.
  22   #
  23   # A short syllable in a word is either:
  24   # 1. a vowel followed by a non-vowel other than w, x or Y and preceded by a non-vowel, or
  25   # 2. a vowel at the beginning of the word followed by a non-vowel.
  26   #
  27   # (The original document is silent on whether sequences of two or more non-vowels make a
  28   # syllable long. But as this specification is only used to find sequences of non-vowel -
  29   # vowel - non-vowel - end-of-word, this ambiguity does not have an effect.)
  30   SHORT_SYLLABLE = "((#{C}#{V}#{CW})|(^#{V}#{C}))"
  31
  32   # Suffix transformations used in porter2_step2.
  33   # (ogi, li endings dealt with in procedure)
  34   STEP_2_MAPS = {"tional" => "tion",
  35                  "enci" => "ence",
  36                  "anci" => "ance",
  37                  "abli" => "able",
  38                  "entli" => "ent",
  39                  "ization" => "ize",
  40                  "izer" => "ize",
  41                  "ational" => "ate",
  42                  "ation" => "ate",
  43                  "ator" => "ate",
  44                  "alism" => "al",
  45                  "aliti" => "al",
  46                  "alli" => "al",
  47                  "fulness" => "ful",
  48                  "ousli" => "ous",
  49                  "ousness" => "ous",
  50                  "iveness" => "ive",
  51                  "iviti" => "ive",
  52                  "biliti" => "ble",
  53                  "bli" => "ble",
  54                  "fulli" => "ful",
  55                  "lessli" => "less" }
  56
  57   # Suffix transformations used in porter2_step3.
  58   # (ative ending dealt with in procedure)
  59   STEP_3_MAPS = {"tional" => "tion",
  60                  "ational" => "ate",
  61                  "alize" => "al",
  62                  "icate" => "ic",
  63                  "iciti" => "ic",
  64                  "ical" => "ic",
  65                  "ful" => "",
  66                  "ness" => "" }
  67
  68   # Suffix transformations used in porter2_step4.
  69   # (ion ending dealt with in procedure)
  70   STEP_4_MAPS = {"al" => "",
  71                  "ance" => "",
  72                  "ence" => "",
  73                  "er" => "",
  74                  "ic" => "",
  75                  "able" => "",
  76                  "ible" => "",
  77                  "ant" => "",
  78                  "ement" => "",
  79                  "ment" => "",
  80                  "ent" => "",
  81                  "ism" => "",
  82                  "ate" => "",
  83                  "iti" => "",
  84                  "ous" => "",
  85                  "ive" => "",
  86                  "ize" => "" }
  87
  88   # Special-case stemmings
  89   SPECIAL_CASES = {"skis" => "ski",
  90                    "skies" => "sky",
  91
  92                    "dying" => "die",
  93                    "lying" => "lie",
  94                    "tying" => "tie",
  95                    "idly" =>  "idl",
  96                    "gently" => "gentl",
  97                    "ugly" => "ugli",
  98                    "early" => "earli",
  99                    "only" => "onli",
 100                    "singly" =>"singl",
 101
 102                    "sky" => "sky",
 103                    "news" => "news",
 104                    "howe" => "howe",
 105                    "atlas" => "atlas",
 106                    "cosmos" => "cosmos",
 107                    "bias" => "bias",
 108                    "andes" => "andes" }
 109
 110   # Special case words to stop processing after step 1a.
 111   STEP_1A_SPECIAL_CASES = %w[ inning outing canning herring earring proceed exceed succeed ]
 112
 113 end
 114