Class: Tabula::Line

Inherits:
ZoneEntity
  • Object
show all
Defined in:
lib/tabula/entities/line.rb

Constant Summary collapse

SPACE_RUN_MAX_LENGTH = 
3

Instance Attribute Summary collapse

Attributes inherited from ZoneEntity

#texts

Instance Method Summary collapse

Methods inherited from ZoneEntity

#<=>, #inspect, #merge!, #points, #tlbr, #tlwh, #to_json

Constructor Details

#initialize(index = nil) ⇒ Line

Returns a new instance of Line.



8
9
10
11
 
# File 'lib/tabula/entities/line.rb', line 8

def initialize(index=nil)
  @text_elements = []
  @index = index
end

Instance Attribute Details

#index ⇒ Object (readonly)

Returns the value of attribute index.



4
5
6
 
# File 'lib/tabula/entities/line.rb', line 4

def index
  @index
end

#text_elements ⇒ Object

Returns the value of attribute text_elements.



3
4
5
 
# File 'lib/tabula/entities/line.rb', line 3

def text_elements
  @text_elements
end

Instance Method Details

#<<(t) ⇒ Object 



13
14
15
16
17
18
19
20
21
22
23
24
 
# File 'lib/tabula/entities/line.rb', line 13

def <<(t)
  if @text_elements.size == 0
    @text_elements << t
    self.top = t.top
    self.left = t.left
    self.width = t.width
    self.height = t.height
  else
    self.text_elements << t
    self.merge!(t)
  end
end

#==(other) ⇒ Object

used for testing, ignores text element stuff besides stripped text.



75
76
77
78
79
80
81
82
83
 
# File 'lib/tabula/entities/line.rb', line 75

def ==(other)
  return false if other.nil?
  self.text_elements = self.text_elements.rpad(TextElement::EMPTY, other.text_elements.size)
  other.text_elements = other.text_elements.rpad(TextElement::EMPTY, self.text_elements.size)
  self.text_elements.zip(other.text_elements).inject(true) do |memo, my_yours|
    my, yours = my_yours
    memo && my == yours
  end
end

#remove_sequential_spaces!(seq_spaces_count = SPACE_RUN_MAX_LENGTH) ⇒ Object

remove runs of the space char longer than SPACE_RUN_MAX_LENGTH should not change dimensions of the container Line



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
63
64
65
66
67
68
69
70
71
72
 
# File 'lib/tabula/entities/line.rb', line 29

def remove_sequential_spaces!(seq_spaces_count=SPACE_RUN_MAX_LENGTH)
  self.text_elements = self.text_elements.reduce([]) do |memo, text_chunk|
    long_space_runs = text_chunk
      .text_elements
      .chunk { |te| te.text == ' '}  # detect runs of spaces...
      .select { |is_space, text_elements| # ...longer than SPACE_RUN_MAX_LENGTH
      is_space && !text_elements.nil? && text_elements.size >= SPACE_RUN_MAX_LENGTH
    }
      .map { |_, text_elements| text_elements }

    # no long runs of spaces
    # keep as it was and end iteration
    if long_space_runs.empty?
      memo << text_chunk
      next memo
    end

    ranges = long_space_runs.map { |lsr|
      idx = text_chunk
        .text_elements
        .index { |te| te.equal?(lsr.first) } # we need pointer comparison here
      (idx)..(idx+lsr.size-1)
    }

    in_run = false
    new_chunk = true
    text_chunk
      .text_elements
      .each_with_index do |te, i|
      if ranges.any? { |r| r.include?(i) } # te belongs to a run of spaces, skip
        in_run = true
      else
        if in_run || new_chunk
          memo << TextChunk.create_from_text_element(te)
        else
          memo.last << te
        end
        in_run = new_chunk = false
      end
    end
    memo
  end # reduce
  self
end