Module: PinYin

Defined in:
lib/ruby-pinyin.rb,
lib/ruby-pinyin/value.rb,
lib/ruby-pinyin/version.rb

Defined Under Namespace

Classes: Value

Constant Summary collapse

VERSION = 
'0.2.1'

Class Method Summary collapse

Class Method Details

.abbr(str, except_lead = false, except_english = true) ⇒ Object 



64
65
66
67
68
69
70
71
 
# File 'lib/ruby-pinyin.rb', line 64

def abbr(str, except_lead=false, except_english=true)
  result = ""
  of_string(str).each_with_index do |word, i|
    w = (except_lead && i == 0) || (except_english && word.english?) ? word : word[0]
    result << w
  end
  result
end

.codes ⇒ Object 



15
16
17
18
19
20
21
22
23
24
 
# File 'lib/ruby-pinyin.rb', line 15

def codes
  return @codes if @codes

  @codes = {}
  src = File.expand_path('../ruby-pinyin/Mandarin.dat', __FILE__)
  override_files.unshift(src).each do |file|
    load_codes_from(file)
  end
  @codes
end

.of_string(str, tone = nil, include_punctuations = false) ⇒ Object 



36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
 
# File 'lib/ruby-pinyin.rb', line 36

def of_string(str, tone=nil, include_punctuations=false)
  res = []
  return res unless str && !str.empty?

  str.unpack('U*').each_with_index do |t,idx|
    code = sprintf('%x',t).upcase
    readings = codes[code]

    if readings
      res << Value.new(format(readings, tone), false)
    else
      val = [t].pack('U*')
      if val =~ /^[_0-9a-zA-Z\s]*$/ # 复原，去除特殊字符,如全角符号等。
        (res.last && res.last.english? ? res.last : res) << Value.new(val, true) # 如果上一个字符也是非中文则与之合并
      elsif include_punctuations
        val = [punctuations[code]].pack('H*') if punctuations.has_key?(code)
        (res.last ? res.last : res) << Value.new(val, false)
      end
    end
  end

  res.map {|phrase| phrase.split(/\s+/)}.flatten
end

.override_files ⇒ Object 



6
7
8
 
# File 'lib/ruby-pinyin.rb', line 6

def override_files
  @override_files || []
end

.override_files=(files) ⇒ Object 



10
11
12
13
 
# File 'lib/ruby-pinyin.rb', line 10

def override_files=(files)
  @override_files = files
  @codes = nil
end

.permlink(str, sep = '-') ⇒ Object 



60
61
62
 
# File 'lib/ruby-pinyin.rb', line 60

def permlink(str, sep='-')
  of_string(str).join(sep)
end

.punctuations ⇒ Object 



26
27
28
29
30
31
32
33
34
 
# File 'lib/ruby-pinyin.rb', line 26

def punctuations
  return @punctuations if @punctuations

  @punctuations = {}
  src = File.expand_path('../ruby-pinyin/Punctuations.dat', __FILE__)
  load_punctuations_from(src)

  @punctuations
end

.sentence(str, tone = nil) ⇒ Object 



73
74
75
 
# File 'lib/ruby-pinyin.rb', line 73

def sentence(str, tone=nil)
  of_string(str, tone, true).join(' ')
end