Class: Kabosu::PosMatcher

Inherits:
Object
  • Object
show all
Defined in:
lib/kabosu/pos_matcher.rb

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(*patterns, &block) ⇒ PosMatcher

Build a matcher from POS patterns or a block.

# From a block
PosMatcher.new { |pos| pos[0] == "名詞" }

# Single pattern (array of strings; "*" or nil = match anything)
PosMatcher.new(["名詞", "固有名詞", "*", "*"])

# Multiple patterns (matches if any pattern matches)
PosMatcher.new(["名詞", "固有名詞"], ["動詞", "*", "*", "*"])



14
15
16
17
18
19
20
21
22
23
24
25
26
 
# File 'lib/kabosu/pos_matcher.rb', line 14

def initialize(*patterns, &block)
  if block
    raise ArgumentError, "cannot supply both patterns and a block" unless patterns.empty?

    @proc = block
  elsif patterns.empty?
    raise ArgumentError, "must supply at least one pattern or a block"
  else
    @patterns = patterns.map(&:freeze).freeze
  end

  freeze
end

Class Method Details

.adjectives ⇒ Object 



81
82
83
 
# File 'lib/kabosu/pos_matcher.rb', line 81

def self.adjectives
  @adjectives ||= new(["形容詞"])
end

.adverbs ⇒ Object 



93
94
95
 
# File 'lib/kabosu/pos_matcher.rb', line 93

def self.adverbs
  @adverbs ||= new(["副詞"])
end

.auxiliary_verbs ⇒ Object 



89
90
91
 
# File 'lib/kabosu/pos_matcher.rb', line 89

def self.auxiliary_verbs
  @auxiliary_verbs ||= new(["助動詞"])
end

.nouns ⇒ Object

── Pre-built matchers ──



73
74
75
 
# File 'lib/kabosu/pos_matcher.rb', line 73

def self.nouns
  @nouns ||= new(["名詞"])
end

.particles ⇒ Object 



85
86
87
 
# File 'lib/kabosu/pos_matcher.rb', line 85

def self.particles
  @particles ||= new(["助詞"])
end

.proper_nouns ⇒ Object 



97
98
99
 
# File 'lib/kabosu/pos_matcher.rb', line 97

def self.proper_nouns
  @proper_nouns ||= new(["名詞", "固有名詞"])
end

.verbs ⇒ Object 



77
78
79
 
# File 'lib/kabosu/pos_matcher.rb', line 77

def self.verbs
  @verbs ||= new(["動詞"])
end

Instance Method Details

#&(other) ⇒ Object

Intersection: matches if both matchers match.



56
57
58
59
 
# File 'lib/kabosu/pos_matcher.rb', line 56

def &(other)
  a, b = self, other
  PosMatcher.new { |pos| a.match?(pos) && b.match?(pos) }
end

#-(other) ⇒ Object

Difference: matches self but not other.



62
63
64
 
# File 'lib/kabosu/pos_matcher.rb', line 62

def -(other)
  difference(other)
end

#difference(other) ⇒ Object 



66
67
68
69
 
# File 'lib/kabosu/pos_matcher.rb', line 66

def difference(other)
  a, b = self, other
  PosMatcher.new { |pos| a.match?(pos) && !b.match?(pos) }
end

#filter(morphemes) ⇒ Object

Return matching morphemes as an Array.



40
41
42
 
# File 'lib/kabosu/pos_matcher.rb', line 40

def filter(morphemes)
  morphemes.select { |m| match?(m) }
end

#match?(morpheme_or_pos) ⇒ Boolean

Returns true if the morpheme (or raw POS array) matches this matcher.

Returns:

  • (Boolean) 


29
30
31
32
33
34
35
36
37
 
# File 'lib/kabosu/pos_matcher.rb', line 29

def match?(morpheme_or_pos)
  pos = extract_pos(morpheme_or_pos)

  if @proc
    @proc.call(pos)
  else
    @patterns.any? { |pattern| pattern_match?(pattern, pos) }
  end
end

#reject(morphemes) ⇒ Object

Return non-matching morphemes as an Array.



45
46
47
 
# File 'lib/kabosu/pos_matcher.rb', line 45

def reject(morphemes)
  morphemes.reject { |m| match?(m) }
end

#|(other) ⇒ Object

Union: matches if either matcher matches.



50
51
52
53
 
# File 'lib/kabosu/pos_matcher.rb', line 50

def |(other)
  a, b = self, other
  PosMatcher.new { |pos| a.match?(pos) || b.match?(pos) }
end