Class: IsoDoc::Convert

Inherits:
Common
  • Object
show all
Defined in:
lib/isodoc/convert.rb

Constant Summary

Constants included from Function::XrefGen

Function::XrefGen::CHILD_EXAMPLES_XPATH, Function::XrefGen::CHILD_NOTES_XPATH, Function::XrefGen::SECTIONS_XPATH

Constants included from Function::Utils

Function::Utils::CLAUSE_ANCESTOR, Function::Utils::NOKOHEAD, Function::Utils::NOTE_CONTAINER_ANCESTOR

Constants included from Function::Table

Function::Table::SW

Constants included from Function::Section

Function::Section::MIDDLE_CLAUSE, Function::Section::TERM_CLAUSE

Constants included from Function::References

Function::References::BIBLIOGRAPHY_XPATH, Function::References::ISO_PUBLISHER_XPATH

Constants included from Function::Lists

Function::Lists::OL_STYLE

Constants included from Function::Cleanup

Function::Cleanup::FIGURE_WITH_FOOTNOTES

Constants included from Function::Blocks

Function::Blocks::EXAMPLE_TBL_ATTR

Instance Attribute Summary collapse

Instance Method Summary collapse

Methods included from ClassUtils

#date_range, #liquid, #ns

Methods included from Function::XrefSectGen

#annex_name_lbl, #annex_names, #annex_names1, #back_anchor_names, #clause_names, #initial_anchor_names, #middle_section_asset_names, #preface_clause_name, #preface_names, #preface_names1, #section_names, #section_names1

Methods included from Function::XrefGen

#anchor_names, #anchor_struct, #anchor_struct_label, #anchor_struct_xref, #example_anchor_names, #get_anchors, #hierarchical_asset_names, #hierarchical_figure_names, #list_anchor_names, #list_item_anchor_names, #listlabel, #note_anchor_names, #sequential_asset_names, #sequential_figure_names, #termexample_anchor_names, #termnote_anchor_names, #termnote_label

Methods included from Function::Utils

#attr_code, #date_range, #empty2nil, #extract_delims, #from_xhtml, #get_clause_id, #get_note_container_id, #header_strip, #insert_tab, #liquid, #noko, #ns, #populate_template, #sentence_join, #to_xhtml, #to_xhtml_fragment

Methods included from Function::ToWordHtml

#bookmark_parse, #define_head, #in_sourcecode, #info, #init_file, #make_body, #make_body1, #make_body2, #make_body3, #middle, #middle_title, #note?, #parse, #set_termdomain, #smallcap_parse, #text_parse, #tmpimagedir

Methods included from Function::Terms

#admitted_term_parse, #definition_parse, #deprecated_term_parse, #modification_parse, #para_then_remainder, #term_parse, #termdef_parse, #termdocsource_parse, #termnote_parse, #termref_parse

Methods included from Function::Table

#make_table_attr, #make_tr_attr, #table_parse, #table_title_parse, #tbody_parse, #tfoot_parse, #thead_parse, #tr_parse

Methods included from Function::Section

#annex, #annex_name, #clause, #clause_name, #clause_parse, #clause_parse_title, #external_terms_boilerplate, #foreword, #inline_header_title, #internal_external_terms_boilerplate, #introduction, #scope, #symbols_abbrevs, #symbols_parse, #term_defs_boilerplate, #term_defs_boilerplate_cont, #terms_defs, #terms_defs_title, #terms_parse

Methods included from Function::References

#biblio_list, #bibliography, #bibliography_parse, #date_note_process, #docid_l10n, #docid_prefix, #format_ref, #implicit_reference, #iso_bibitem_entry, #iso_bibitem_entry_attrs, #iso_bibitem_ref_code, #iso_title, #noniso_bibitem, #norm_ref, #norm_ref_preface, #omit_docid_prefix, #prefix_bracketed_ref, #ref_entry_code, #reference_format, #reference_names, #split_bibitems

Methods included from Function::Lists

#dl_parse, #dt_dd?, #dt_parse, #li_parse, #ol_depth, #ol_parse, #ol_style, #ul_parse

Methods included from Function::Inline

#anchor_linkend, #callout_parse, #eref_localities, #eref_parse, #error_parse, #get_linkend, #link_parse, #page_break, #prefix_container, #section_break, #stem_parse, #xref_parse

Methods included from Function::I18n

#eref_localities1, #eref_localities1_zh, #i18n_init, l10n

Methods included from Function::Cleanup

#admonition_cleanup, #cleanup, #example_cleanup, #extract_symbols_list, #figure_aside_process, #figure_cleanup, #figure_get_or_make_dl, #footnote_cleanup, #inline_header_cleanup, #merge_fnref_into_fn_text, #new_fullcolspan_row, #remove_bottom_border, #symbol_key, #symbols_cleanup, #table_cleanup, #table_footnote_cleanup, #table_get_or_make_tfoot, #table_note_cleanup

Methods included from Function::Blocks

#admonition_parse, #annotation_parse, #example_div_parse, #example_label, #example_parse, #figure_key, #figure_name_parse, #figure_parse, #formula_parse, #formula_where, #image_parse, #image_title_parse, #note_label, #note_p_parse, #note_parse, #note_parse1, #para_attrs, #para_parse, #quote_attribution, #quote_parse, #sourcecode_name_parse, #sourcecode_parse

Constructor Details

#initialize(options) ⇒ Convert

htmlstylesheet: Generic stylesheet for HTML wordstylesheet: Generic stylesheet for Word standardsheet: Stylesheet specific to Standard header: Header file for Word htmlcoverpage: Cover page for HTML wordcoverpage: Cover page for Word htmlintropage: Introductory page for HTML wordintropage: Introductory page for Word i18nyaml: YAML file for internationalisation of text ulstyle: list style in Word CSS for unordered lists olstyle: list style in Word CSS for ordered lists bodyfont: font to use for body text headerfont: font to use for header text monospace: font to use for monospace text suppressheadingnumbers: suppress heading numbers for clauses scripts: Scripts file for HTML scripts_pdf: Scripts file for PDF datauriimage: Encode images in HTML output as data URIs



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
 
# File 'lib/isodoc/convert.rb', line 27

def initialize(options)
  @libdir = File.dirname(__FILE__) unless @libdir
  options.merge!(default_fonts(options)) { |_, old, new| old || new }.
    merge!(default_file_locations(options)) { |_, old, new| old || new }
  @options = options
  @files_to_delete = []
  @htmlstylesheet = generate_css(options[:htmlstylesheet], true, extract_fonts(options))
  @wordstylesheet = generate_css(options[:wordstylesheet], false, extract_fonts(options))
  @standardstylesheet = generate_css(options[:standardstylesheet], false, extract_fonts(options))
  @header = options[:header]
  @htmlcoverpage = options[:htmlcoverpage]
  @wordcoverpage = options[:wordcoverpage]
  @htmlintropage = options[:htmlintropage]
  @wordintropage = options[:wordintropage]
  @scripts = options[:scripts]
  @scripts_pdf = options[:scripts_pdf]
  @i18nyaml = options[:i18nyaml]
  @ulstyle = options[:ulstyle]
  @olstyle = options[:olstyle]
  @datauriimage = options[:datauriimage]
  @suppressheadingnumbers = options[:suppressheadingnumbers]
  @termdomain = ""
  @termexample = false
  @note = false
  @sourcecode = false
  @anchors = {}
  @footnotes = []
  @comments = []
  @in_footnote = false
  @in_comment = false
  @in_table = false
  @in_figure = false
  @seen_footnote = Set.new
  @c = HTMLEntities.new
  @openmathdelim = "`"
  @closemathdelim = "`"
  @lang = "en"
  @script = "Latn"
  @tmpimagedir = "_images"
  @maxwidth = 1200
  @maxheight = 800
end

Instance Attribute Details

#options ⇒ Object (readonly)

Returns the value of attribute options.



7
8
9
 
# File 'lib/isodoc/convert.rb', line 7

def options
  @options
end

Instance Method Details

#convert(filename, file = nil, debug = false) ⇒ Object 



145
146
147
148
149
150
151
152
153
 
# File 'lib/isodoc/convert.rb', line 145

def convert(filename, file = nil, debug = false)
  file = File.read(filename, encoding: "utf-8") if file.nil?
  @openmathdelim, @closemathdelim = extract_delims(file)
  docxml, filename, dir = convert_init(file, filename, debug)
  result = convert1(docxml, filename, dir)
  return result if debug
  postprocess(result, filename, dir)
  FileUtils.rm_rf dir
end

#convert1(docxml, filename, dir) ⇒ Object 



119
120
121
122
123
124
125
126
127
128
 
# File 'lib/isodoc/convert.rb', line 119

def convert1(docxml, filename, dir)
  anchor_names docxml
  noko do |xml|
    xml.html do |html|
      html.parent.add_namespace("epub", "http://www.idpf.org/2007/ops")
      define_head html, filename, dir
      make_body(html, docxml)
    end
  end.join("\n")
end

#convert_init(file, filename, debug) ⇒ Object 



134
135
136
137
138
139
140
141
142
143
 
# File 'lib/isodoc/convert.rb', line 134

def convert_init(file, filename, debug)
  docxml = Nokogiri::XML(file)
  filename, dir = init_file(filename, debug)
  docxml.root.default_namespace = ""
  lang = docxml&.at(ns("//bibdata/language"))&.text || @lang
  script = docxml&.at(ns("//bibdata/script"))&.text || @script
  i18n_init(lang, script)
  metadata_init(lang, script, @labels)
  [docxml, filename, dir]
end

#default_file_locations(_options) ⇒ Object

none for this parent gem, but will be populated in child gems which have access to stylesheets &c; e.g. 

htmlstylesheet: html_doc_path("htmlstyle.scss"),
htmlcoverpage: html_doc_path("html_rsd_titlepage.html"),
htmlintropage: html_doc_path("html_rsd_intro.html"),
scripts: html_doc_path("scripts.html"),
wordstylesheet: html_doc_path("wordstyle.scss"),
standardstylesheet: html_doc_path("rsd.scss"),
header: html_doc_path("header.html"),
wordcoverpage: html_doc_path("word_rsd_titlepage.html"),
wordintropage: html_doc_path("word_rsd_intro.html"),
ulstyle: l3
olstyle: l2




92
93
94
 
# File 'lib/isodoc/convert.rb', line 92

def default_file_locations(_options)
  {}
end

#default_fonts(_options) ⇒ Object 



70
71
72
73
74
75
76
 
# File 'lib/isodoc/convert.rb', line 70

def default_fonts(_options)
  {
    bodyfont: "Arial",
    headerfont: "Arial",
    monospacefont: "Courier",
  }
end

#extract_fonts(options) ⇒ Object

extract fonts for use in generate_css



97
98
99
100
101
102
 
# File 'lib/isodoc/convert.rb', line 97

def extract_fonts(options)
  b = options[:bodyfont] || "Arial"
  h = options[:headerfont] || "Arial"
  m = options[:monospacefont] || "Courier"
  "$bodyfont: #{b};\n$headerfont: #{h};\n$monospacefont: #{m};\n"
end

#generate_css(filename, stripwordcss, fontheader) ⇒ Object 



108
109
110
111
112
113
114
115
116
117
 
# File 'lib/isodoc/convert.rb', line 108

def generate_css(filename, stripwordcss, fontheader)
  return nil unless filename
  stylesheet = File.read(filename, encoding: "UTF-8")
  stylesheet.gsub!(/(\s|\{)mso-[^:]+:[^;]+;/m, "\\1") if stripwordcss
  engine = SassC::Engine.new(fontheader + stylesheet, syntax: :scss)
  outname = File.basename(filename, ".*") + ".css"
  File.open(outname, "w:UTF-8") { |f| f.write(engine.render) }
  @files_to_delete << outname
  outname
end

#html_doc_path(file) ⇒ Object 



104
105
106
 
# File 'lib/isodoc/convert.rb', line 104

def html_doc_path(file)
  File.join(@libdir, File.join("html", file))
end

#metadata_init(lang, script, labels) ⇒ Object 



130
131
132
 
# File 'lib/isodoc/convert.rb', line 130

def metadata_init(lang, script, labels)
  @meta = Metadata.new(lang, script, labels)
end