Class: TocExtract

Inherits:
Object
  • Object
show all
Defined in:
lib/toc_extract.rb,
lib/toc_extract/preview.rb,
lib/toc_extract/extractor.rb

Class Method Summary collapse

Class Method Details

.extract(pdf_file, template, toc_start_page, toc_end_page) ⇒ Object 



14
15
16
17
18
19
20
 
# File 'lib/toc_extract/extractor.rb', line 14

def self.extract(pdf_file, template, toc_start_page, toc_end_page)
  lines = TocExtract.toc_lines(pdf_file, template, toc_start_page, toc_end_page)
  sections = TocExtract.sections_from_toc_lines(lines, template)
  TocExtract.fill_bounding_boxes(pdf_file, sections, toc_end_page)

  sections
end

.fill_bounding_boxes(pdf_file, sections, toc_end_page) ⇒ Object 



119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
 
# File 'lib/toc_extract/extractor.rb', line 119

def self.fill_bounding_boxes(pdf_file, sections, toc_end_page)
  i = 0
  bboxes = {}
  content_parts = []
  PDF::Reader.open(pdf_file) do |reader|
    reader.pages.each_with_index do |page, page_num|
      next if page_num <= toc_end_page
      page.extend(PDF::Reader::FindText)
      runs = page.runs(merge: false)
      runs.each do |run|
        content_parts << run.text
        bboxes[i] = {
          "x" => run.x,
          "y" => run.y,
          "width" => run.width,
          "endx" => run.endx,
          "endy" => run.endy,
          "page" => page_num
        }
        i += run.text.length
      end
    end
  end
  content = content_parts.join

  # For each section, search in the content and find its position,
  # look it up in the bounding box and store it
  for section in sections
    pos = content.index(section.title)
    # puts "not found #{pos} for #{section.title}" if bboxes.key?(pos)
    section.bounding_box = bboxes[pos]
  end
end

.hi(language = "english") ⇒ Object 



2
3
4
5
 
# File 'lib/toc_extract.rb', line 2

def self.hi(language = "english")
  translator = Translator.new(language)
  translator.hi
end

.preview(pdf_file, section, crop_width, crop_height) ⇒ Object 



2
3
4
5
6
7
8
9
10
11
12
13
14
15
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
 
# File 'lib/toc_extract/preview.rb', line 2

def self.preview(pdf_file, section, crop_width, crop_height)
  require "rmagick"
  require "pdf/reader"
  require "pdf/reader/find_text"

  result = "" 

  PDF::Reader.open(pdf_file) do |reader|
    page_num = section.page_number

    target_page = reader.pages[page_num]
    page_width = target_page.width
    page_height = target_page.height

    images = Magick::Image.read("#{pdf_file}[#{page_num}]") do |info|
      info.density = 150
    end

    img = images.first
    img_width = img.columns
    img_height = img.rows

    scale_x = img_width.to_f / page_width
    scale_y = img_height.to_f / page_height

    # In pdf, x,y are the bottom left coordinates, converting them to top left
    # Convert pdf-reader coordinates to RMagick coordinates
    # pdf-reader: origin at bottom-left, Y increases upward
    # RMagick: origin at top-left, Y increases downward
    # Add some padding as well
    pdf_x = section.bounding_box["x"] - 30
    pdf_y = section.bounding_box["endy"] + 10

    img_x = (pdf_x * scale_x).round
    img_y = ((page_height - pdf_y) * scale_y).round  # Flip Y coordinate
    img_width_pixels = (crop_width * scale_x).round
    img_height_pixels = (crop_height * scale_y).round

    # Ensure coordinates are within image bounds
    img_x = [ img_x, 0 ].max
    img_y = [ img_y, 0 ].max
    img_width_pixels = [ img_width_pixels, img_width ].min
    img_height_pixels = [ img_height_pixels, img_height ].min

    cropped_img = img.crop(img_x, img_y, img_width_pixels, img_height_pixels)

    cropped_img.resize_to_fit!(crop_width, crop_height)
    result = cropped_img.to_blob { |info| info.format = "PNG" }
  end
  
  result
end

.sections_from_toc_lines(lines, template) ⇒ Object 



80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
 
# File 'lib/toc_extract/extractor.rb', line 80

def self.sections_from_toc_lines(lines, template)
  # Same as before, the template is important, but is hard-coded as <NUMBER><TEXT>[...]<NUMBER>
  sections = []
  for line in lines
    section_id = ""
    page = ""

    title_start = 0
    title_end = line.length

    # Section id
    line.each_char do |c|
      if c.match?(/[\d\.]/)
        section_id += c
        title_start += 1
      else
        break
      end
    end

    # Page number
    line.reverse.each_char do |c|
      if c.match?(/[\d]/)
        page += c
        title_end -= 1
      else
        break
      end
    end
    page = page.reverse.to_i - 1 # Since TOC pages are 1-based

    # title
    title = line[title_start..title_end-1].sub(/\.+$/, '')
    sections << Section.new(section_id, title, page)
  end

sections
end

.toc_lines(pdf_file, template, toc_start_page, toc_end_page) ⇒ Object 



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
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
 
# File 'lib/toc_extract/extractor.rb', line 23

def self.toc_lines(pdf_file, template, toc_start_page, toc_end_page)
  # To detect the lines, we assume that all the elements on a line have the same y value.
  # If this assumption changes in the future, we can instead list all the deltas between
  # y values and compute breaking points where a noticeable jumpt in delta occures.

  # template is a placeholder for future extensions. At the moment, always assume that
  # each line starts with a section number, followed by section title and optionally
  # some dots and ends with a page number. Morever, we assume that pages use arabic
  # numbering as opposed to roman numerals.


  lines = []
  max_delta = 0
  line = ""
  last_y = 0
  PDF::Reader.open(pdf_file) do |reader|
    reader.pages.each_with_index do |page, page_num|
      next if page_num < toc_start_page
      break if page_num > toc_end_page
      page.extend(PDF::Reader::FindText)
      runs = page.runs(merge: false)
      runs.each do |run|
        y = run.y
        last_y = y if last_y == 0
        delta = (last_y - y).abs
        if delta > max_delta
          lines << line
          line = run.text
          last_y = y
        else
          line += run.text
        end
      end
    end
  end

  # Merge lines based on the template. Currently, hard-coded as <NUMBER><TEXT>[...]<NUMBER>
  # Also romeve the lines that contain page numbers. This should also be part of the template, assuming arabic numbering
  real_lines = []
  for line in lines
    if line[0].match?(/[\d]/)
      next if line.scan(/\D/).empty?
      real_lines << line
    else
      # we could be unlucky and get a line that happens to have number in the its text portion
      # and the line breaks at that exact moment. Ignoring for now.
      if real_lines.length == 0
        next # this is the toc header
      else
        real_lines[-1] += line
      end
    end
  end

  real_lines
end