Module: Varar::Core::Sentences

Defined in:
lib/varar/core/sentences.rb

Constant Summary collapse

ABBREVIATIONS = 
Set.new(['e.g.', 'i.e.', 'etc.', 'cf.', 'vs.']).freeze

Class Method Summary collapse

Class Method Details

.build_cp_to_u16(text) ⇒ Object

cp_to_u16 is the UTF-16 offset of text.



65
66
67
68
69
70
71
72
73
74
 
# File 'lib/varar/core/sentences.rb', line 65

def build_cp_to_u16(text)
  result = Array.new(text.length + 1, 0)
  u16 = 0
  text.each_char.with_index do |ch, i|
    result[i] = u16
    u16 += ch.ord > 0xFFFF ? 2 : 1
  end
  result[text.length] = u16
  result
end

.digit?(ch) ⇒ Boolean

Returns:

  • (Boolean) 


104
105
106
 
# File 'lib/varar/core/sentences.rb', line 104

def digit?(ch)
  !ch.empty? && ch.match?(/[0-9]/)
end

.inside_number_or_abbrev?(text, dot_pos) ⇒ Boolean

Returns:

  • (Boolean) 


76
77
78
79
80
81
82
83
84
85
86
 
# File 'lib/varar/core/sentences.rb', line 76

def inside_number_or_abbrev?(text, dot_pos)
  prev = dot_pos.positive? ? text[dot_pos - 1] : ''
  nxt = dot_pos + 1 < text.length ? text[dot_pos + 1] : ''
  return true if digit?(prev) && digit?(nxt)

  ABBREVIATIONS.each do |abbrev|
    start = [0, dot_pos + 1 - abbrev.length].max
    return true if text[start...(dot_pos + 1)] == abbrev
  end
  lower?(nxt)
end

.lower?(ch) ⇒ Boolean

Unicode-aware "is a lowercase letter": has case and is already lower.

Returns:

  • (Boolean) 


109
110
111
 
# File 'lib/varar/core/sentences.rb', line 109

def lower?(ch)
  !ch.empty? && ch != ch.upcase && ch == ch.downcase
end

.push_segment(out, text, start_cp, end_cp, cp_to_u16) ⇒ Object 



88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
 
# File 'lib/varar/core/sentences.rb', line 88

def push_segment(out, text, start_cp, end_cp, cp_to_u16)
  return if end_cp <= start_cp

  raw = text[start_cp...end_cp]
  stripped = raw.strip
  return if stripped.empty?

  lead = raw.length - raw.lstrip.length
  trail = raw.length - raw.rstrip.length
  out << Sentence.new(
    text: stripped,
    start_offset: cp_to_u16[start_cp + lead],
    end_offset: cp_to_u16[end_cp - trail]
  )
end

.split_sentences(text) ⇒ Object 



16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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
60
61
62
 
# File 'lib/varar/core/sentences.rb', line 16

def split_sentences(text)
  cp_to_u16 = build_cp_to_u16(text)
  n = text.length
  out = []

  # Mark no-split zones (backtick spans, double-quoted strings).
  skip = Array.new(n, false)
  j = 0
  while j < n
    c = text[j]
    if ['`', '"'].include?(c)
      close = text.index(c, j + 1)
      break if close.nil?

      (j..close).each { |k| skip[k] = true }
      j = close + 1
      next
    end
    j += 1
  end

  i = 0
  segment_start = 0
  while i < n
    if skip[i]
      i += 1
      next
    end
    ch = text[i]
    if ["\n", '.', '!', '?'].include?(ch)
      if ch == '.' && inside_number_or_abbrev?(text, i)
        i += 1
        next
      end
      stop = i + 1
      push_segment(out, text, segment_start, stop, cp_to_u16)
      i = stop
      i += 1 while i < n && [' ', "\n"].include?(text[i])
      segment_start = i
      next
    end
    i += 1
  end

  push_segment(out, text, segment_start, n, cp_to_u16)
  out
end