Module: IsoDoc::Function::Cleanup

Included in:
Common
Defined in:
lib/isodoc/function/cleanup.rb

Instance Method Summary collapse

Instance Method Details

#admonition_cleanup(docxml) ⇒ Object

todo PRESENTATION XML



28
29
30
31
32
33
34
35
36
 
# File 'lib/isodoc/function/cleanup.rb', line 28

def admonition_cleanup(docxml)
  docxml.xpath("//div[@class = 'Admonition'][title]").each do |d|
    title = d.at("./title")
    n = title.next_element
    n&.children&.first
      &.add_previous_sibling("#{title.remove.text}—")
  end
  docxml
end

#cleanup(docxml) ⇒ Object 



16
17
18
19
20
21
22
23
24
25
 
# File 'lib/isodoc/function/cleanup.rb', line 16

def cleanup(docxml)
  @i18n ||= i18n_init(@lang, @script, @locale)
  comment_cleanup(docxml)
  inline_header_cleanup(docxml)
  figure_cleanup(docxml)
  table_cleanup(docxml)
  symbols_cleanup(docxml)
  example_cleanup(docxml)
  admonition_cleanup(docxml)
end

#empty_tags(html) ⇒ Object

Atomic group (?>…) prevents catastrophic backtracking across ‘>` characters. Quoted alternatives listed first so `/` inside attribute values (e.g. URLs in href=“…”) is consumed as part of the quoted string, not rejected by [^>/]. Includes Word HTML tags



98
99
100
101
102
103
104
105
106
107
108
 
# File 'lib/isodoc/function/cleanup.rb', line 98

def empty_tags(html)
  void_elements = %w[area base br col embed hr img input link meta
                     source track wbr wrapblock]
  html.gsub(%r{<([\w:]+)((?>"[^"]*"|'[^']*'|[^>/])*)\s*/>}) do
    if void_elements.include?($1)
      $&
    else
      "<#{$1}#{$2}></#{$1}>"
    end
  end
end

#example_cleanup(docxml) ⇒ Object 



38
39
40
41
42
43
 
# File 'lib/isodoc/function/cleanup.rb', line 38

def example_cleanup(docxml)
  docxml.xpath("//table[@class = 'example']//p[not(@class)]").each do |p|
    p["class"] = "example"
  end
  docxml
end

#figure_cleanup(docxml) ⇒ Object 



45
 
# File 'lib/isodoc/function/cleanup.rb', line 45

def figure_cleanup(docxml); end

#inline_header_cleanup(docxml) ⇒ Object 



47
48
49
50
51
52
53
54
55
56
57
58
 
# File 'lib/isodoc/function/cleanup.rb', line 47

def inline_header_cleanup(docxml)
  docxml.xpath('//span[@class="zzMoveToFollowing"]').each do |x|
    x.delete("class")
    n = x.next_element
    if n.nil?
      x.name = "p"
    else
      n.add_first_child(x.remove)
    end
  end
  docxml
end

#merge_fnref_into_fn_text(elem) ⇒ Object 



60
61
62
63
64
65
 
# File 'lib/isodoc/function/cleanup.rb', line 60

def merge_fnref_into_fn_text(elem)
  fn = elem.at('.//span[@class="TableFootnoteRef"]/..') or return
  n = fn.next_element
  n or return
  n.children.first.add_previous_sibling(fn.remove)
end

#passthrough_cleanup(docxml) ⇒ Object 



8
9
10
11
12
13
14
 
# File 'lib/isodoc/function/cleanup.rb', line 8

def passthrough_cleanup(docxml)
  docxml.split(%r{(<passthrough>|</passthrough>)}).each_slice(4)
    .map do |a|
    a.size > 2 and a[2] = HTMLEntities.new.decode(a[2])
    [a[0], a[2]]
  end.join
end

#symbols_cleanup(docxml) ⇒ Object 



91
 
# File 'lib/isodoc/function/cleanup.rb', line 91

def symbols_cleanup(docxml); end

#table_cleanup(docxml) ⇒ Object 



86
87
88
89
 
# File 'lib/isodoc/function/cleanup.rb', line 86

def table_cleanup(docxml)
  table_footnote_cleanup(docxml)
  docxml
end

#table_footnote_cleanup(docxml) ⇒ Object

preempt html2doc putting MsoNormal under TableFootnote class



68
69
70
71
72
73
74
75
76
77
 
# File 'lib/isodoc/function/cleanup.rb', line 68

def table_footnote_cleanup(docxml)
  docxml.xpath("//tfoot[descendant::aside]").each do |t|
    t.xpath(".//aside").each do |a|
      merge_fnref_into_fn_text(a)
      a.name = "div"
      a["class"] = "TableFootnote"
    end
  end
  table_footnote_cleanup_propagate(docxml)
end

#table_footnote_cleanup_propagate(docxml) ⇒ Object 



79
80
81
82
83
84
 
# File 'lib/isodoc/function/cleanup.rb', line 79

def table_footnote_cleanup_propagate(docxml)
  docxml.xpath("//p[not(self::*[@class])]" \
               "[ancestor::*[@class = 'TableFootnote']]").each do |p|
    p["class"] = "TableFootnote"
  end
end

#textcleanup(docxml) ⇒ Object 



4
5
6
 
# File 'lib/isodoc/function/cleanup.rb', line 4

def textcleanup(docxml)
  passthrough_cleanup(docxml)
end