Class: Unicoder::Builder::DisplayWidth

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

Constant Summary collapse

ZERO_WIDTH_CATEGORIES = 
%w[Mn Me Zl Zp Cf].freeze
ZERO_WIDTH_HANGUL = 
[
  *0x1160..0x11FF, # HANGUL JUNGSEONG
  *0xD7B0..0xD7FF, # HANGUL JUNGSEONG
].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
  # 0x85   =>  0, #    NEXT LINE
  0xAD   =>  nil, #  SOFT HYPHEN, nil = 1 (default)
  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, ambiguous) ⇒ Object 



103
104
105
106
107
108
109
110
111
112
113
114
 
# File 'lib/unicoder/builders/display_width.rb', line 103

def determine_width(codepoint, category, east_asian_width, ambiguous)
  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"
    ambiguous == 1 ? nil : ambiguous
  else
    nil
  end
end

#initialize_index ⇒ Object 



39
40
41
42
43
44
45
 
# File 'lib/unicoder/builders/display_width.rb', line 39

def initialize_index
  @index = {
    WIDTH_ONE: [],
    WIDTH_TWO: [],
  }
  @ignorable = []
end

#parse! ⇒ Object 



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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
 
# File 'lib/unicoder/builders/display_width.rb', line 47

def parse!
  # Find Ignorables
  parse_file :core_properties, :line, begin: /^# Derived Property: Default_Ignorable_Code_Point$/, end: /^# ================================================$/, regex: /^(?<codepoints>\S+)\s+; Default_Ignorable_Code_Point.*$/ do |line|
    if line["codepoints"]['..']
      single_or_multiple_codepoints = Range.new(*line["codepoints"].split('..').map{ |codepoint|
        codepoint.to_i(16)
      })
    else
      single_or_multiple_codepoints = line["codepoints"].to_i(16)
    end

    @ignorable += [*single_or_multiple_codepoints]
  end

  # Assign based on East Asian Width
  parse_file :east_asian_width, :line, regex: /^(?<codepoints>\S+?)\s*;\s*(?<width>\S+)\s+#\s(?<category>\S+).*$/ do |line|
    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 :WIDTH_ONE, codepoint, determine_width(codepoint, line["category"], line["width"], 1)
      assign :WIDTH_TWO, codepoint, determine_width(codepoint, line["category"], line["width"], 2)
    }
  end

  # Assign Ranges
  ## Zero-width
  (ZERO_WIDTH_HANGUL | @ignorable).each{ |codepoint|
    assign :WIDTH_ONE, codepoint, 0
    assign :WIDTH_TWO, codepoint, 0
  }

  ## Full-width
  WIDE_RANGES.each{ |codepoint|
    assign :WIDTH_ONE, codepoint, 2
    assign :WIDTH_TWO, codepoint, 2
  }

  ## Table
  SPECIAL_WIDTHS.each{ |codepoint, value|
    assign :WIDTH_ONE, codepoint, value
    assign :WIDTH_TWO, codepoint, value
  }

  # Compres Index
  4.times{ compress! @index[:WIDTH_ONE] }
  4.times{ compress! @index[:WIDTH_TWO] }
  remove_trailing_nils! @index[:WIDTH_ONE]
  remove_trailing_nils! @index[:WIDTH_TWO]
end