Class: TinySegmenter

Inherits:
Object
  • Object
show all
Defined in:
lib/tiny_segmenter.rb,
lib/tiny_segmenter/version.rb

Constant Summary collapse

WhitespaceOnlyRegex = 
Regexp.compile("^[　 ]+$")
PunctuationRegex = 
Regexp.compile("^[-–—―.。・（）()［］｛｝{}【】⟨⟩、､,，،…‥〽「」『』〜~！!：:？?\"'|_＿“”‘’;/⁄／«»]+$")
VERSION = 
"0.0.6"

Instance Method Summary collapse

Constructor Details

#initialize ⇒ TinySegmenter

Returns a new instance of TinySegmenter.



9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
 
# File 'lib/tiny_segmenter.rb', line 9

def initialize
  @chartype = []
  @model = SegmentationModel.new
  @BIAS = -332
  # Compile regex patterns
  {
    "[一二三四五六七八九十百千万億兆]" => "M", # numbers (japanese)
    "[一-龠々〆ヵヶ]" => "H", # kanji & misc characters
    "[ぁ-ん]" => "I", # hiragana
    "[ァ-ヴーｱ-ﾝﾞｰ]" => "K", # katakana
    "[a-zA-Zａ-ｚＡ-Ｚ]" => "A", # ascii / romaji letters
    "[0-9０-９]" => "N", # ascii / romaji numbers
  }.each do |pattern, value|
    @chartype << [Regexp.compile(pattern), value]
  end
end

Instance Method Details

#segment(text, options = {}) ⇒ Object 



26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
 
# File 'lib/tiny_segmenter.rb', line 26

def segment(text, options = {})
  return []  if text.nil? || text.strip.empty?
  text = text.strip
  result = []
  segments = %w[B3 B2 B1]
  ctypes = %w[O O O]
  text.split(//).each do |char|
    char.strip!
    next  if char.empty? || char.match(WhitespaceOnlyRegex)
    next  if options[:ignore_punctuation] && char.match(PunctuationRegex)
    segments << char
    ctypes << ctype(char)
  end
  segments.concat(%w[E1 E2 E3])
  ctypes.concat(%w[O O O])

  word = segments[3]
  p1, p2, p3 = %w[U U U]
  (4..segments.size-4).to_a.each do |i|
    score = @BIAS
    w1 = segments[i - 3]
    w2 = segments[i - 2]
    w3 = segments[i - 1]
    w4 = segments[i]
    w5 = segments[i + 1]
    w6 = segments[i + 2]
    c1 = ctypes[i - 3]
    c2 = ctypes[i - 2]
    c3 = ctypes[i - 1]
    c4 = ctypes[i]
    c5 = ctypes[i + 1]
    c6 = ctypes[i + 2]

    score += sum_scores(p1, p2, p3, w1, w2, w3, w4, w5, w6, c1, c2, c3, c4, c5, c6)
    p_new = "O"
    if score > 0
      result << word
      word = ""
      p_new = "B"
    end
    p1, p2, p3 = p2, p3, p_new
    word = "#{word}#{segments[i]}"
  end
  result << word  unless word.empty?
  result
end