Class: Unicoder::Builder::DisplayWidth

Inherits:
Object
  • Object
show all
Includes:
Unicoder::Builder, MultiDimensionalArrayBuilder
Defined in:
lib/unicoder/builders/display_width.rb

Constant Summary collapse

IGNORE_CATEGORIES = 
%w[Cs Co Cn].freeze
ZERO_WIDTH_CATEGORIES = 
%w[Mn Me Cf].freeze
ZERO_WIDTH_RANGES = 
[
  *0x1160..0x11FF, # HANGUL JUNGSEONG
  *0xD7B0..0xD7FF, # HANGUL JUNGSEONG
  *0x2060..0x206F, # Ignorables
  *0xFFF0..0xFFF8, # Ignorables
  *0xE0000..0xE0FFF, # Ignorables
].freeze
WIDE_RANGES = 
[
  *0x3400..0x4DBF,
  *0x4E00..0x9FFF,
  *0xF900..0xFAFF,
  *0x20000..0x2FFFD,
  *0x30000..0x3FFFD,
].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 =>  nil, #    SOFT HYPHEN
  0x2E3A =>  2, #    TWO-EM DASH
  0x2E3B =>  3, #    THREE-EM DASH
}.freeze

Instance Attribute Summary

Attributes included from Unicoder::Builder

#formats, #index, #option

Instance Method Summary collapse

Methods included from MultiDimensionalArrayBuilder

#assign_codepoint, #compress!, #remove_trailing_nils!

Methods included from Unicoder::Builder

#assign, #assign_codepoint, build, #export, #initialize, #meta, #parse_file

Instance Method Details

#determine_width(codepoint, category, east_asian_width) ⇒ Object 



78
79
80
81
82
83
84
85
86
87
88
89
 
# File 'lib/unicoder/builders/display_width.rb', line 78

def determine_width(codepoint, category, east_asian_width)
  if  ( ZERO_WIDTH_CATEGORIES.include?(category) &&
        [codepoint].pack('U') !~ /\p{Cf}(?<=\p{Arabic})/ )
    0
  elsif east_asian_width == "F" || east_asian_width == "W"
    2
  elsif east_asian_width == "A"
    :A
  else
    nil
  end
end

#initialize_index ⇒ Object 



42
43
44
 
# File 'lib/unicoder/builders/display_width.rb', line 42

def initialize_index
  @index = []
end

#parse! ⇒ Object 



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
72
73
74
75
76
 
# File 'lib/unicoder/builders/display_width.rb', line 46

def parse!
  parse_file :east_asian_width, :line, regex: /^(?<codepoints>\S+?)\s*;\s*(?<width>\S+)\s+#\s(?<category>\S+).*$/ do |line|
    next if IGNORE_CATEGORIES.include?(line["category"])

    if line["codepoints"]['..']
      codepoints = Range.new(*line["codepoints"].split('..').map{ |codepoint|
        codepoint.to_i(16)
      })
    else
      codepoints = [line["codepoints"].to_i(16)]
    end

    codepoints.each{ |codepoint|
      assign_codepoint codepoint, determine_width(codepoint, line["category"], line["width"])
    }
  end

  ZERO_WIDTH_RANGES.each{ |codepoint|
    assign_codepoint codepoint, 0
  }

  WIDE_RANGES.each{ |codepoint|
    assign_codepoint codepoint, 2
  }

  SPECIAL_WIDTHS.each{ |codepoint, value|
    assign_codepoint codepoint, value
  }

  4.times{ compress! }
end