Module: Unicode::DisplayWidth::IndexBuilder

Defined in:
lib/unicode/display_width/index_builder.rb

Constant Summary collapse

EAST_ASIAN_WIDTH_DATA_URL = 
"http://www.unicode.org/Public/UNIDATA/EastAsianWidth.txt".freeze
EAST_ASIAN_WIDTH_DATA_FILENAME = 
(DATA_DIRECTORY + 'EastAsianWidth.txt').freeze
IGNORE_CATEGORIES = 
%w[Cs Co Cn].freeze
ZERO_WIDTH_CATEGORIES = 
%w[Mn Me Cf].freeze
ZERO_WIDTH_CODEPOINTS = 
[*0x1160..0x11FF].freeze
SPECIAL_WIDTHS = 
{
  0x0    =>  0, # \0 NULL
  0x5    =>  0, #    ENQUIRY
  0x7    =>  0, # \a BELL
  0x8    => -1, # \b BACKSPACE
  0xA    =>  0, # \n LINE FEED
  0xB    =>  0, # \v LINE TABULATION
  0xC    =>  0, # \f FORM FEED
  0xD    =>  0, # \r CARRIAGE RETURN
  0xE    =>  0, #    SHIFT OUT
  0xF    =>  0, #    SHIFT IN
  0x00AD =>  1, #    SOFT HYPHEN
  0x2E3A =>  2, #    TWO-EM DASH
  0x2E3B =>  3, #    THREE-EM DASH
}.freeze

Class Method Summary collapse

Class Method Details

.build! ⇒ Object 



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
 
# File 'lib/unicode/display_width/index_builder.rb', line 34

def self.build!
  data = File.open(EAST_ASIAN_WIDTH_DATA_FILENAME)
  data.rewind
  Dir.mkdir(DATA_DIRECTORY) unless Dir.exists?(DATA_DIRECTORY)
  index = {}

  data.each_line{ |line|
    line =~ /^(\S+?);(\S+)\s+#\s(\S+).*$/
    if $1 && $2
      cps, width, category = $1, $2, $3
      next if IGNORE_CATEGORIES.include?(category)
      if cps['..']
        codepoints = Range.new(*cps.split('..').map{ |cp| cp.to_i(16) })
      else
        codepoints = [cps.to_i(16)]
      end

      codepoints.each{ |cp|
        index[cp] = is_zero_width?(category, cp) ? 0 : width.to_sym
      }
    end
  }

  index.merge! SPECIAL_WIDTHS
  File.open(INDEX_FILENAME, 'wb') { |f| Marshal.dump(index, f) }
end

.fetch! ⇒ Object 



27
28
29
30
31
32
 
# File 'lib/unicode/display_width/index_builder.rb', line 27

def self.fetch!
  require 'open-uri'
  open(EAST_ASIAN_WIDTH_DATA_URL) { |f|
    File.write(EAST_ASIAN_WIDTH_DATA_FILENAME, f.read)
  }
end

.is_zero_width?(category, cp) ⇒ Boolean

Returns:

  • (Boolean) 


61
62
63
64
65
 
# File 'lib/unicode/display_width/index_builder.rb', line 61

def self.is_zero_width?(category, cp)
  ( ZERO_WIDTH_CATEGORIES.include?(category) &&
      [cp].pack('U') !~ /\p{Cf}(?<=\p{Arabic})/ ) ||
    ZERO_WIDTH_CODEPOINTS.include?(cp)
end