Class: OoxmlParser::DocumentStructure

Inherits:
CommonDocumentStructure show all
Defined in:
lib/ooxml_parser/docx_parser/docx_data/document_structure.rb

Constant Summary

Constants inherited from OOXMLDocumentObject

OOXMLDocumentObject::DEFAULT_DIRECTORY_FOR_MEDIA

Class Attribute Summary collapse

Instance Attribute Summary collapse

Attributes inherited from CommonDocumentStructure

#file_path

Class Method Summary collapse

Instance Method Summary collapse

Methods inherited from OOXMLDocumentObject

add_to_xmls_stack, copy_file_and_rename_to_zip, copy_media_file, current_xml, dir, encrypted_file?, get_link_from_rels, media_folder, option_enabled?, unzip_file

Constructor Details

#initialize(elements = [], page_properties = nil, notes = [], background = nil, document_properties = DocumentProperties.new, comments = []) ⇒ DocumentStructure

Returns a new instance of DocumentStructure.



16
17
18
19
20
21
22
23
24
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 16

def initialize(elements = [], page_properties = nil, notes = [], background = nil, document_properties = DocumentProperties.new, comments = [])
  @elements = elements
  @page_properties = page_properties
  @notes = notes
  @background = background
  @document_properties = document_properties
  @comments = comments
  @document_styles = []
end

Class Attribute Details

.default_paragraph_style ⇒ Object

Returns the value of attribute default_paragraph_style.



218
219
220
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 218

def default_paragraph_style
  @default_paragraph_style
end

.default_table_paragraph_style ⇒ Object

Returns the value of attribute default_table_paragraph_style.



217
218
219
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 217

def default_table_paragraph_style
  @default_table_paragraph_style
end

.default_table_run_style ⇒ Object

Returns the value of attribute default_table_run_style.



216
217
218
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 216

def default_table_run_style
  @default_table_run_style
end

Instance Attribute Details

#background ⇒ Object

Returns the value of attribute background.



11
12
13
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 11

def background
  @background
end

#comments ⇒ Object

Returns the value of attribute comments.



11
12
13
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 11

def comments
  @comments
end

#document_properties ⇒ Object

Returns the value of attribute document_properties.



11
12
13
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 11

def document_properties
  @document_properties
end

#document_styles ⇒ Array, DocumentStyle

Returns array of document styles in current document.

Returns:

  • (Array, DocumentStyle) —

    array of document styles in current document



14
15
16
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 14

def document_styles
  @document_styles
end

#elements ⇒ Object

Returns the value of attribute elements.



11
12
13
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 11

def elements
  @elements
end

#notes ⇒ Object

Returns the value of attribute notes.



11
12
13
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 11

def notes
  @notes
end

#page_properties ⇒ Object

Returns the value of attribute page_properties.



11
12
13
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 11

def page_properties
  @page_properties
end

Class Method Details

.parse ⇒ Object 



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
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 122

def self.parse
  OOXMLDocumentObject.root_subfolder = 'word/'
  OOXMLDocumentObject.xmls_stack = []
  OOXMLDocumentObject.namespace_prefix = 'w'
  @comments = []
  @default_paragraph = DocxParagraph.new
  @default_character = DocxParagraphRun.new
  @default_table_properties = TableProperties.new
  PresentationTheme.parse('word/theme/theme1.xml')
  OOXMLDocumentObject.add_to_xmls_stack('word/styles.xml')
  doc = Nokogiri::XML(File.open(OOXMLDocumentObject.current_xml))
  doc.search('//w:docDefaults').each do |doc_defaults|
    doc_defaults.xpath('w:pPrDefault').each do |p_pr_defaults|
      @default_paragraph = DocxParagraph.parse(p_pr_defaults, 0)
    end
    doc_defaults.xpath('w:rPrDefault').each do |r_pr_defaults|
      r_pr_defaults.xpath('w:rPr').each do |r_pr|
        @default_character = DocxParagraphRun.parse(r_pr, DocxParagraphRun.new)
      end
    end
  end
  parse_default_style
  doc_structure = DocumentStructure.new
  doc_structure.document_styles = DocumentStyle.parse_list
  number = 0
  OOXMLDocumentObject.add_to_xmls_stack('word/document.xml')
  doc = Nokogiri::XML(File.open(OOXMLDocumentObject.current_xml))
  doc.search('//w:document').each do |document|
    document.xpath('w:background').each do |background|
      doc_structure.background = DocumentBackground.parse(background)
    end
    document.xpath('w:body').each do |body|
      body.xpath('*').each do |element|
        if element.name == 'p'
          child = element.child
          unless child.nil? && doc_structure.elements.last.class == Table
            paragraph_style = DocxParagraph.parse(element, number, @default_paragraph, @default_character)
            number += 1
            doc_structure.elements << paragraph_style.copy
          end
        elsif element.name == 'tbl'
          table = Table.parse(element, number, TableProperties.new, DocumentStructure.default_table_paragraph_style, DocumentStructure.default_table_run_style)
          number += 1
          doc_structure.elements << table
        end
      end
      body.xpath('w:sectPr').each do |sect_pr|
        doc_structure.page_properties = PageProperties.parse(sect_pr, @default_paragraph, @default_character)
        doc_structure.notes = doc_structure.page_properties.notes # keep copy of notes to compatibility with previous docx models
      end
    end
  end
  OOXMLDocumentObject.xmls_stack.pop
  doc_structure.document_properties = DocumentProperties.parse
  doc_structure.comments = Comment.parse_list
  doc_structure
end

.parse_default_style ⇒ Object 



180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 180

def self.parse_default_style
  doc = Nokogiri::XML(File.open(OOXMLDocumentObject.path_to_folder + 'word/styles.xml'))
  doc.search('//w:style').each do |style|
    next if style.attribute('default').nil?
    if (style.attribute('default').value == '1' || style.attribute('default').value == 'on' || style.attribute('default').value == 'true') && style.attribute('type').value == 'paragraph'
      style.xpath('w:pPr').each do |paragraph_pr_tag|
        DocxParagraph.parse_paragraph_style(paragraph_pr_tag, @default_paragraph, @default_character)
      end
      style.xpath('w:rPr').each do |character_pr_tag|
        DocxParagraphRun.parse(character_pr_tag, @default_character, @default_character)
      end
    elsif (style.attribute('default').value == '1' || style.attribute('default').value == 'on' || style.attribute('default').value == 'true') && style.attribute('type').value == 'character'
      style.xpath('w:rPr').each do |character_pr_tag|
        DocxParagraphRun.parse(character_pr_tag, @default_character, @default_character)
      end
    end
  end
  DocumentStructure.default_table_paragraph_style = @default_paragraph.copy
  DocumentStructure.default_table_paragraph_style.spacing = Spacing.new(0, 0, 1, :auto)
  DocumentStructure.default_table_run_style = @default_character.copy
  doc.search('//w:style').each do |style|
    next if style.attribute('default').nil?
    next unless (style.attribute('default').value == '1' || style.attribute('default').value == 'on' || style.attribute('default').value == 'true') && style.attribute('type').value == 'table'
    style.xpath('w:tblPr').each do |table_pr_tag|
      @default_table_properties = TableProperties.parse(table_pr_tag)
    end
    style.xpath('w:pPr').each do |table_paragraph_pr_tag|
      DocxParagraph.parse_paragraph_style(table_paragraph_pr_tag, DocumentStructure.default_table_paragraph_style, DocumentStructure.default_table_run_style)
    end
    style.xpath('w:rPr').each do |table_character_pr_tag|
      DocxParagraphRun.parse(table_character_pr_tag, DocumentStructure.default_table_run_style, @default_character)
    end
  end
end

Instance Method Details

#==(other) ⇒ Object 



26
27
28
29
30
31
32
33
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 26

def ==(other)
  @elements == other.elements &&
    @page_properties == other.page_properties &&
    @notes == other.notes &&
    @background == other.background &&
    @document_properties == other.document_properties &&
    @comments == other.comments
end

#difference(other) ⇒ Object 



35
36
37
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 35

def difference(other)
  Hash.object_to_hash(self).diff(Hash.object_to_hash(other))
end

#document_style_by_name(name) ⇒ DocumentStyle?

Return document style by its name

Parameters:

  • name (String) —

    name of style

Returns:



108
109
110
111
112
113
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 108

def document_style_by_name(name)
  @document_styles.each do |style|
    return style if style.name == name
  end
  nil
end

#element_by_description(location: :canvas, type: :docx_paragraph) ⇒ Object 



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
79
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 39

def element_by_description(location: :canvas, type: :docx_paragraph)
  case location
  when :canvas
    case type
    when :table
      elements[1].rows[0].cells[0].elements
    when :docx_paragraph, :simple, :paragraph
      elements
    when :shape
      elements[0].nonempty_runs.first.alternate_content.office2007_content.data.text_box
    else
      raise 'Wrong location(Need One of ":table", ":paragraph", ":shape")'
    end
  when :footer
    case type
    when :table
      note_by_description(:footer1).elements[1].rows[0].cells[0].elements
    when :docx_paragraph, :simple, :paragraph
      note_by_description(:footer1).elements
    when :shape
      note_by_description(:footer1).elements[0].nonempty_runs.first.alternate_content.office2007_content.data.text_box
    else
      raise 'Wrong location(Need One of ":table", ":simple", ":shape")'
    end
  when :header
    case type
    when :table
      note_by_description(:header1).elements[1].rows[0].cells[0].elements
    when :docx_paragraph, :simple, :paragraph
      note_by_description(:header1).elements
    when :shape
      note_by_description(:header1).elements[0].nonempty_runs.first.alternate_content.office2007_content.data.text_box
    else
      raise 'Wrong location(Need One of ":table", ":simple", ":shape")'
    end
  when :comment
    comments[0].paragraphs
  else
    raise 'Wrong global location(Need One of ":canvas", ":footer", ":header", ":comment")'
  end
end

#note_by_description(type) ⇒ Object 



81
82
83
84
85
86
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 81

def note_by_description(type)
  notes.each do |note|
    return note if note.type.to_sym == type
  end
  raise 'There isn\'t this type of the note'
end

#outline(location: :canvas, type: :simple, levels_count: 1) ⇒ Object 



95
96
97
98
99
100
101
102
103
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 95

def outline(location: :canvas, type: :simple, levels_count: 1)
  elements = element_by_description(location: location, type: type)
  set = []
  levels_count.times do |col|
    set[0] = elements[col].numbering.numbering_properties.ilvls[col].num_format
    set[1] = elements[col].numbering.numbering_properties.ilvls[col].level_text
  end
  set
end

#recognize_numbering(location: :canvas, type: :simple, paragraph_number: 0) ⇒ Object 



88
89
90
91
92
93
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 88

def recognize_numbering(location: :canvas, type: :simple, paragraph_number: 0)
  elements = element_by_description(location: location, type: type)
  lvl_text = elements[paragraph_number].numbering.numbering_properties.ilvls[0].level_text
  num_format = elements[paragraph_number].numbering.numbering_properties.ilvls[0].num_format
  [num_format, lvl_text]
end

#style_exist?(name) ⇒ True, False

Check if style exists in current document

Parameters:

  • name (String) —

    name of style

Returns:

  • (True, False) 


118
119
120
 
# File 'lib/ooxml_parser/docx_parser/docx_data/document_structure.rb', line 118

def style_exist?(name)
  !document_style_by_name(name).nil?
end