Class: Metanorma::CollectionRenderer

Inherits:
Object
  • Object
show all
Defined in:
lib/metanorma/collection_renderer.rb,
lib/metanorma/collection_fileparse.rb,
lib/metanorma/collection_fileprocess.rb,
lib/metanorma/collection_render_word.rb,
lib/metanorma/collection_render_utils.rb

Overview

XML collection renderer

Defined Under Namespace

Classes: DocOptionsNode, Dummy, PdfOptionsNode

Constant Summary collapse

FORMATS = 
%i[html xml doc pdf].freeze
SECTION_BREAK = 
'<p class="MsoNormal"><br clear="all" class="section"/></p>'
.freeze
DIV1 = 
'<div class="WordSection1">&#xa0;</div>'.freeze
DIV2 = 
'<div class="WordSection2">&#xa0;</div>'.freeze

Instance Attribute Summary collapse

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(collection, folder, options = {}) ⇒ CollectionRenderer

This is only going to render the HTML collection We presuppose that the bibdata of the document is equivalent to that of the collection, and that the flavour gem can sensibly process it. We may need to enhance metadata in the flavour gems isodoc/metadata.rb with collection metadata

Parameters:

  • xml (Metanorma::Collection) —

    input XML collection

  • folder (String) —

    input folder

  • options (Hash) (defaults to: {})

Options Hash (options):

  • :coverpage (String) —

    cover page HTML (Liquid template)

  • :format (Array<Symbol>) —

    list of formats (xml,html,doc,pdf)

  • :output_folder (String) —

    output directory



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
 
# File 'lib/metanorma/collection_renderer.rb', line 30

def initialize(collection, folder, options = {}) # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
  check_options options
  @xml = Nokogiri::XML collection.to_xml # @xml is the collection manifest
  @lang = @xml.at("//xmlns:bibdata/xmlns:language")&.text || "en"
  @script = @xml.at("//xmlns:bibdata/xmlns:script")&.text || "Latn"
  @locale = @xml.at("//xmlns:bibdata/xmlns:locale")&.text
  @doctype = doctype
  @compile = Compile.new
  @compile.load_flavor(@doctype)

  @isodoc = isodoc_create # output processor for flavour
  @outdir = dir_name_cleanse(options[:output_folder])
  @coverpage = options[:coverpage] || collection.coverpage
  @format = Util.sort_extensions_execution(options[:format])
  @compile_options = options[:compile] || {}
  @compile_options[:no_install_fonts] = true if options[:no_install_fonts]
  @log = options[:log]
  @documents = collection.documents
  @bibdata = collection.documents
  @directives = collection.directives
  @dirname = collection.dirname
  @disambig = Util::DisambigFiles.new
  @c = HTMLEntities.new
  @files_to_delete = []
  @nested = options[:nested] # if false, this is the root instance of Renderer
  # if true, then this is not the last time Renderer will be run
  # (e.g. this is sectionsplit)

  # list of files in the collection
  @files = Metanorma::FileLookup.new(folder, self)
  @files.add_section_split
  isodoc_populate
  create_non_existing_directory(@outdir)
end

Instance Attribute Details

#compile ⇒ Object (readonly)

Returns the value of attribute compile.



16
17
18
 
# File 'lib/metanorma/collection_renderer.rb', line 16

def compile
  @compile
end

#compile_options ⇒ Object (readonly)

Returns the value of attribute compile_options.



16
17
18
 
# File 'lib/metanorma/collection_renderer.rb', line 16

def compile_options
  @compile_options
end

#documents ⇒ Object (readonly)

Returns the value of attribute documents.



16
17
18
 
# File 'lib/metanorma/collection_renderer.rb', line 16

def documents
  @documents
end

#isodoc ⇒ Object

Returns the value of attribute isodoc.



15
16
17
 
# File 'lib/metanorma/collection_renderer.rb', line 15

def isodoc
  @isodoc
end

#nested ⇒ Object

Returns the value of attribute nested.



15
16
17
 
# File 'lib/metanorma/collection_renderer.rb', line 15

def nested
  @nested
end

#xml ⇒ Object (readonly)

Returns the value of attribute xml.



16
17
18
 
# File 'lib/metanorma/collection_renderer.rb', line 16

def xml
  @xml
end

Class Method Details

.render(col, options = {}) ⇒ Object

Parameters:

Options Hash (options):

  • :coverpage (String) —

    cover page HTML (Liquid template)

  • :format (Array<Symbol>) —

    list of formats

  • :ourput_folder (Strong) —

    output directory



77
78
79
80
81
82
83
84
85
 
# File 'lib/metanorma/collection_renderer.rb', line 77

def self.render(col, options = {})
  warn "\n\n\n\n\nRender Init: #{DateTime.now.strftime('%H:%M:%S')}"
  cr = new(col, File.dirname(col.file), options)
  cr.files
  cr.concatenate(col, options)
  options[:format]&.include?(:html) and cr.coverpage
  cr.flush_files
  cr
end

Instance Method Details

#add_hidden_bibliography(xmldoc, refs) ⇒ Object 



72
73
74
75
76
77
78
79
80
81
82
83
 
# File 'lib/metanorma/collection_fileparse.rb', line 72

def add_hidden_bibliography(xmldoc, refs)
  ins = new_hidden_ref(xmldoc)
  refs.each do |k, v|
    _, url = @files.targetfile_id(v, {})
    ins << <<~XML
      <bibitem id="#{k}">
        <docidentifier type="repository">current-metanorma-collection/#{v}</docidentifier>
        <uri type='citation'>#{url}</uri>
      </bibitem>
    XML
  end
end

#collect_erefs(docxml) ⇒ Object 



53
54
55
56
57
58
59
 
# File 'lib/metanorma/collection_render_utils.rb', line 53

def collect_erefs(docxml)
  docxml.xpath(ns("//eref"))
    .each_with_object({ citeas: {}, bibitemid: {} }) do |i, m|
    m[:citeas][i["citeas"]] = true
    m[:bibitemid][i["bibitemid"]] = true
  end
end

#collection_coverpages(conv, docs) ⇒ Object 



70
71
72
73
74
75
76
77
78
 
# File 'lib/metanorma/collection_render_word.rb', line 70

def collection_coverpages(conv, docs)
  conv.wordintropage and [DIV2, SECTION_BREAK].reverse.each do |s|
    docs.unshift(Nokogiri::XML(s).root)
  end
  conv.wordcoverpage and [DIV1, SECTION_BREAK].reverse.each do |s|
    docs.unshift(Nokogiri::XML(s).root)
  end
  docs
end

#compile_options_update(identifier) ⇒ Object 



26
27
28
29
30
31
32
33
34
35
36
 
# File 'lib/metanorma/collection_fileprocess.rb', line 26

def compile_options_update(identifier)
  ret = @compile_options.dup
  Array(@directives).include?("presentation-xml") ||
    @files.get(identifier, :presentationxml) and
    ret.merge!(passthrough_presentation_xml: true)
  @files.get(identifier, :sectionsplit) == "true" and
    ret.merge!(sectionsplit: "true")
  @files.get(identifier, :bare) == true and
    ret.merge!(bare: true)
  ret
end

#concat_extract_files(filename) ⇒ Object 



10
11
12
13
14
15
16
17
18
 
# File 'lib/metanorma/collection_render_word.rb', line 10

def concat_extract_files(filename)
  xml = Nokogiri::XML(File.read(filename, encoding: "UTF-8"), &:huge)
  docs = xml.xpath(ns("//doc-container")).each_with_object([]) do |x, m|
    n = Nokogiri::XML::Document.new
    n.add_child(x.elements.first.remove)
    m << n
  end
  [wrapping_doc(docs.first.dup, xml), docs]
end

#concatenate(col, options) ⇒ Object 



87
88
89
90
91
92
93
 
# File 'lib/metanorma/collection_renderer.rb', line 87

def concatenate(col, options)
  warn "\n\n\n\n\nConcatenate: #{DateTime.now.strftime('%H:%M:%S')}"
  (options[:format] & %i(pdf doc)).empty? or
    options[:format] << :presentation
  concatenate_prep(col, options)
  concatenate_outputs(options)
end

#concatenate1(out, ext) ⇒ Object 



111
112
113
114
115
116
117
118
119
120
 
# File 'lib/metanorma/collection_renderer.rb', line 111

def concatenate1(out, ext)
  out.directives << "documents-inline"
  out.bibdatas.each_key do |ident|
    id = @isodoc.docid_prefix(nil, ident.dup)
    @files.get(id, :attachment) || @files.get(id, :outputs).nil? and next
    out.documents[Util::key id] =
      Metanorma::Document.raw_file(@files.get(id, :outputs)[ext])
  end
  out
end

#concatenate_outputs(options) ⇒ Object 



105
106
107
108
109
 
# File 'lib/metanorma/collection_renderer.rb', line 105

def concatenate_outputs(options)
  pres = File.join(@outdir, "collection.presentation.xml")
  options[:format].include?(:pdf) and pdfconv.convert(pres)
  options[:format].include?(:doc) and docconv_convert(pres)
end

#concatenate_prep(col, options) ⇒ Object 



95
96
97
98
99
100
101
102
103
 
# File 'lib/metanorma/collection_renderer.rb', line 95

def concatenate_prep(col, options)
  %i(xml presentation).each do |e|
    options[:format].include?(e) or next
    ext = e == :presentation ? "presentation.xml" : e.to_s
    File.open(File.join(@outdir, "collection.#{ext}"), "w:UTF-8") do |f|
      f.write(concatenate1(col.clone, e).to_xml)
    end
  end
end

#copy_file_to_dest(identifier) ⇒ Object 



50
51
52
53
54
 
# File 'lib/metanorma/collection_fileprocess.rb', line 50

def copy_file_to_dest(identifier)
  dest = File.join(@outdir, @files.get(identifier, :out_path))
  FileUtils.mkdir_p(File.dirname(dest))
  FileUtils.cp @files.get(identifier, :ref), dest
end

#coverpage ⇒ Object

populate liquid template of ARGV with metadata extracted from collection manifest



136
137
138
139
140
141
142
 
# File 'lib/metanorma/collection_renderer.rb', line 136

def coverpage
  @coverpage or return
  warn "\n\n\n\n\nCoverpage: #{DateTime.now.strftime('%H:%M:%S')}"
  File.open(File.join(@outdir, "index.html"), "w:UTF-8") do |f|
    f.write @isodoc.populate_template(File.read(@coverpage))
  end
end

#dir_name_cleanse(name) ⇒ Object 



3
4
5
6
7
8
9
10
11
 
# File 'lib/metanorma/collection_render_utils.rb', line 3

def dir_name_cleanse(name)
  path = Pathname.new(name)
  clean_regex = /[<>:"|?*]/
  fallback_sym = "_"
  return name.gsub(clean_regex, fallback_sym) unless path.absolute?

  File.join(path.dirname,
            path.basename.to_s.gsub(clean_regex, fallback_sym))
end

#docconv ⇒ Object 



3
4
5
6
7
8
 
# File 'lib/metanorma/collection_render_word.rb', line 3

def docconv
  @tempfile_cache ||= []
  doctype = @doctype.to_sym
  x = Asciidoctor.load nil, backend: doctype
  x.converter.doc_converter(DocOptionsNode.new(@directives, @dirname))
end

#docconv_convert(filename) ⇒ Object 



80
81
82
83
84
85
86
 
# File 'lib/metanorma/collection_render_word.rb', line 80

def docconv_convert(filename)
  pref_file, docs = concat_extract_files(filename)
  body = docconv_convert1(docs)
  collection_conv = overall_docconv_converter(body)
  collection_coverpages(collection_conv, body)
  collection_conv.convert(filename, pref_file.to_xml, false)
end

#docconv_convert1(docs) ⇒ Object 



57
58
59
60
61
62
63
64
65
66
67
68
 
# File 'lib/metanorma/collection_render_word.rb', line 57

def docconv_convert1(docs)
  docs.each_with_index.with_object([]) do |(d, i), m|
    conv = docconv
    conv.convert_init(d.to_xml(encoding: "UTF-8"), "xxxx", false)
    html = conv.postprocess_cleanup(conv.convert1(d, "xxx", "."))
    @tempfile_cache += conv.tempfile_cache # hold on to the temp img files
    b = Nokogiri::XML(html).at("//body")
    i == docs.size - 1 or
      b << '<p class="MsoNormal"><br clear="all" class="section"/></p>'
    m << b.children
  end
end

#docid_to_citeas(bib) ⇒ Object 



47
48
49
50
51
 
# File 'lib/metanorma/collection_render_utils.rb', line 47

def docid_to_citeas(bib)
  docid = bib.at(ns("./docidentifier[@primary = 'true']")) ||
    bib.at(ns("./docidentifier")) or return
  docid_prefix(docid)
end

#docref_ident(docref) ⇒ Object 



176
177
178
179
 
# File 'lib/metanorma/collection_renderer.rb', line 176

def docref_ident(docref)
  ident = docref.at(ns("./identifier")).children.to_xml
  @c.decode(@isodoc.docid_prefix(nil, ident))
end

#docrefs(elm, builder) ⇒ Object

Parameters:

  • elm (Nokogiri::XML::Element)
  • builder (Nokogiri::XML::Builder) 


163
164
165
166
167
168
169
170
171
172
173
174
 
# File 'lib/metanorma/collection_renderer.rb', line 163

def docrefs(elm, builder)
  elm.xpath(ns("./docref[@index = 'true']")).each do |d|
    ident = docref_ident(d)
    builder.li do |li|
      li.a href: index_link(d, ident) do |a|
        a << ident.split(/([<>&])/).map do |x|
          /[<>&]/.match?(x) ? x : @c.encode(x, :hexadecimal)
        end.join
      end
    end
  end
end

#doctype ⇒ Object

infer the flavour from the first document identifier; relaton does that



123
124
125
126
127
128
129
130
131
132
 
# File 'lib/metanorma/collection_renderer.rb', line 123

def doctype
  if (docid = @xml.at("//xmlns:bibdata/xmlns:docidentifier/@type")&.text)
    dt = docid.downcase
  elsif (docid = @xml.at("//xmlns:bibdata/xmlns:docidentifier")&.text)
    dt = docid.sub(/\s.*$/, "").lowercase
  else return "standoc"
  end
  @registry = Metanorma::Registry.instance
  @registry.alias(dt.to_sym)&.to_s || dt
end

#dup_bibitem(docid, bib) ⇒ Object 



13
14
15
16
17
18
19
20
 
# File 'lib/metanorma/collection_render_utils.rb', line 13

def dup_bibitem(docid, bib)
  newbib = @files.get(docid, :bibdata).dup
  newbib.name = "bibitem"
  newbib["hidden"] = "true"
  newbib&.at("./*[local-name() = 'ext']")&.remove
  newbib["id"] = bib["id"]
  newbib
end

#eref2link(docxml) ⇒ Object 



85
86
87
88
89
 
# File 'lib/metanorma/collection_fileparse.rb', line 85

def eref2link(docxml)
  isodoc = IsoDoc::PresentationXMLConvert.new({})
  isodoc.bibitem_lookup(docxml)
  isodoc.eref2link(docxml)
end

#file_compile(file, filename, identifier) ⇒ Object

compile and output individual file in collection warn “metanorma compile -x html #Metanorma::CollectionRenderer.ff.path”



12
13
14
15
16
17
18
19
20
21
22
23
24
 
# File 'lib/metanorma/collection_fileprocess.rb', line 12

def file_compile(file, filename, identifier)
  return if @files.get(identifier, :sectionsplit) == "true"

  opts = {
    format: :asciidoc,
    extension_keys: @format,
    output_dir: @outdir,
  }.merge(compile_options_update(identifier))

  @compile.compile file, opts
  @files.set(identifier, :outputs, {})
  file_compile_formats(filename, identifier)
end

#file_compile_formats(filename, identifier) ⇒ Object 



38
39
40
41
42
43
44
45
46
47
48
 
# File 'lib/metanorma/collection_fileprocess.rb', line 38

def file_compile_formats(filename, identifier)
  f = @files.get(identifier, :outputs)
  @format << :presentation if @format.include?(:pdf)
  @format.each do |e|
    ext = @compile.processor.output_formats[e]
    fn = File.basename(filename).sub(/(?<=\.)[^.]+$/, ext.to_s)
    (/html$/.match?(ext) && @files.get(identifier, :sectionsplit)) or
      f[e] = File.join(@outdir, fn)
  end
  @files.set(identifier, :outputs, f)
end

#files ⇒ Object

process each file in the collection files are held in memory, and altered as postprocessing



58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
 
# File 'lib/metanorma/collection_fileprocess.rb', line 58

def files # rubocop:disable Metrics/AbcSize, Metrics/MethodLength
  warn "\n\n\n\n\nRender Files: #{DateTime.now.strftime('%H:%M:%S')}"
  internal_refs = locate_internal_refs
  @files.keys.each_with_index do |ident, i|
    i.positive? && Array(@directives).include?("bare-after-first") and
      @compile_options.merge!(bare: true)
    if @files.get(ident, :attachment) then copy_file_to_dest(ident)
    else
      file, filename = @files.targetfile_id(ident, read: true)
      warn "\n\n\n\n\nProcess #{filename}: #{DateTime.now.strftime('%H:%M:%S')}"
      collection_xml = update_xrefs(file, ident, internal_refs)
      collection_filename = File.basename(filename, File.extname(filename))
      collection_xml_path = File.join(Dir.tmpdir,
                                      "#{collection_filename}.xml")
      File.write collection_xml_path, collection_xml, encoding: "UTF-8"
      file_compile(collection_xml_path, filename, ident)
      FileUtils.rm(collection_xml_path)
    end
  end
end

#flush_files ⇒ Object 



65
66
67
68
69
70
 
# File 'lib/metanorma/collection_renderer.rb', line 65

def flush_files
  warn "\n\n\n\n\nDone: #{DateTime.now.strftime('%H:%M:%S')}"
  warn @files.files_to_delete
  @files.files_to_delete.each { |f| FileUtils.rm_f(f) }
  @files_to_delete.each { |f| FileUtils.rm_f(f) }
end

#gather_internal_refs ⇒ Object

gather internal bibitem references



80
81
82
83
84
85
86
 
# File 'lib/metanorma/collection_fileprocess.rb', line 80

def gather_internal_refs
  @files.keys.each_with_object({}) do |i, refs|
    @files.get(i, :attachment) and next
    file, = @files.targetfile_id(i, read: true)
    gather_internal_refs1(file, i, refs)
  end
end

#gather_internal_refs1(file, ident, refs) ⇒ Object 



88
89
90
91
92
93
94
 
# File 'lib/metanorma/collection_fileprocess.rb', line 88

def gather_internal_refs1(file, ident, refs)
  f = Nokogiri::XML(file, &:huge)
  !@files.get(ident, :sectionsplit) and
    gather_internal_refs_indirect(f, refs)
  key = @files.get(ident, :indirect_key) and
    gather_internal_refs_sectionsplit(f, ident, key, refs)
end

#gather_internal_refs_indirect(doc, refs) ⇒ Object 



96
97
98
99
100
101
102
103
104
 
# File 'lib/metanorma/collection_fileprocess.rb', line 96

def gather_internal_refs_indirect(doc, refs)
  doc.xpath(ns("//bibitem[@type = 'internal']/" \
               "docidentifier[@type = 'repository']")).each do |d|
                 a = d.text.split(%r{/}, 2)
                 a.size > 1 or next
                 refs[a[0]] ||= {}
                 refs[a[0]][a[1]] = false
               end
end

#gather_internal_refs_sectionsplit(_doc, ident, key, refs) ⇒ Object 



106
107
108
109
110
111
 
# File 'lib/metanorma/collection_fileprocess.rb', line 106

def gather_internal_refs_sectionsplit(_doc, ident, key, refs)
  refs[key] ||= {}
  @files.get(ident, :ids).each_key do |k|
    refs[key][k] = false
  end
end

#get_bibitem_docid(bib, identifier) ⇒ Object 



22
23
24
25
26
27
28
29
30
31
32
33
 
# File 'lib/metanorma/collection_render_utils.rb', line 22

def get_bibitem_docid(bib, identifier)
  docid =
    bib.at(ns("./docidentifier[@type = 'metanorma-collection']")) ||
    bib.at(ns("./docidentifier[not(@type)]")) ||
    bib.at(ns("./docidentifier"))
  docid &&= docid_prefix(docid)
  if @files.get(docid) then docid
  else
    fail_update_bibitem(docid, identifier)
    nil
  end
end

#hide_refs(docxml) ⇒ Object 



35
36
37
38
39
40
 
# File 'lib/metanorma/collection_render_utils.rb', line 35

def hide_refs(docxml)
  docxml.xpath(ns("//references[bibitem][not(./bibitem[not(@hidden) or " \
                  "@hidden = 'false'])]")).each do |f|
    f["hidden"] = "true"
  end
end

#index_link(docref, ident) ⇒ Object 



181
182
183
184
185
186
 
# File 'lib/metanorma/collection_renderer.rb', line 181

def index_link(docref, ident)
  if docref["fileref"]
    @files.get(ident, :out_path).sub(/\.xml$/, ".html")
  else "#{docref['id']}.html"
  end
end

#index_object(elm) ⇒ Object

object to construct navigation out of in Liquid



205
206
207
208
209
210
211
212
213
214
215
216
 
# File 'lib/metanorma/collection_renderer.rb', line 205

def index_object(elm)
  c = elm.xpath(ns("./manifest")).each_with_object([]) do |d, b|
    b << index_object(d)
  end
  c.empty? and c = nil
  r = Nokogiri::HTML::Builder.new do |b|
    indexfile_docref(elm, b)
  end
  r &&= r.doc.root&.to_html&.gsub("\n", " ")
  { title: indexfile_title(elm),
    docrefs: r, children: c }.compact
end

#indexfile(elm) ⇒ String

single level navigation list, with hierarchical nesting

Parameters:

  • elm (Nokogiri::XML::Element)

Returns:

  • (String) —

    XML



192
193
194
195
196
197
198
199
200
201
202
 
# File 'lib/metanorma/collection_renderer.rb', line 192

def indexfile(elm)
  Nokogiri::HTML::Builder.new do |b|
    b.ul do
      b.li indexfile_title(elm)
      indexfile_docref(elm, b)
      elm.xpath(ns("./manifest")).each do |d|
        b << indexfile(d)
      end
    end
  end.doc.root.to_html
end

#indexfile_docref(elm, builder) ⇒ Object

uses the identifier to label documents; other attributes (title) can be looked up in @files[:bibdata]

Parameters:

  • elm (Nokogiri::XML::Element)
  • builder (Nokogiri::XML::Builder) 


155
156
157
158
159
 
# File 'lib/metanorma/collection_renderer.rb', line 155

def indexfile_docref(elm, builder)
  return "" unless elm.at(ns("./docref[@index = 'true']"))

  builder.ul { |b| docrefs(elm, b) }
end

#indexfile_title(elm) ⇒ String

Parameters:

  • elm (Nokogiri::XML::Element)

Returns:

  • (String) 


146
147
148
 
# File 'lib/metanorma/collection_renderer.rb', line 146

def indexfile_title(elm)
  elm.at(ns("./title"))&.text
end

#indirect_ref_key(schema, id, docxml) ⇒ Object 



179
180
181
182
183
184
185
186
 
# File 'lib/metanorma/collection_fileparse.rb', line 179

def indirect_ref_key(schema, id, docxml)
  /^#{schema}_/.match?(id) and return id
  ret = "#{schema}_#{id}"
  suffix = docxml.root['document_suffix']
  (k = docxml.root["type"]) && k != schema && suffix and 
    ret = "#{ret}_#{suffix}"
  ret
end

#liquid_docrefs ⇒ Object 



218
219
220
221
222
223
224
225
226
227
228
 
# File 'lib/metanorma/collection_renderer.rb', line 218

def liquid_docrefs
  @xml.xpath(ns("//docref[@index = 'true']")).each_with_object([]) do |d, m|
    ident = d.at(ns("./identifier")).children.to_xml
    ident = @c.decode(@isodoc.docid_prefix(nil, ident))
    title = d.at(ns("./bibdata/title[@type = 'main']")) ||
      d.at(ns("./bibdata/title")) || d.at(ns("./title"))
    m << { "identifier" => ident, "file" => index_link(d, ident),
           "title" => title&.children&.to_xml,
           "level" => d.at(ns("./level"))&.text }
  end
end

#locate_internal_refs ⇒ Object

resolve file location for the target of each internal reference



123
124
125
126
127
128
129
130
131
132
133
134
 
# File 'lib/metanorma/collection_fileprocess.rb', line 123

def locate_internal_refs
  warn "\n\n\n\n\nInternal Refs: #{DateTime.now.strftime('%H:%M:%S')}"
  refs = populate_internal_refs(gather_internal_refs)
  refs.each do |schema, ids|
    ids.each do |id, key|
      key and next
      refs[schema][id] = "Missing:#{schema}:#{id}"
      @log&.add("Cross-References", nil, refs[schema][id])
    end
  end
  refs
end

#locate_internal_refs1(refs, identifier, ident) ⇒ Object 



136
137
138
139
140
141
142
143
144
 
# File 'lib/metanorma/collection_fileprocess.rb', line 136

def locate_internal_refs1(refs, identifier, ident)
  t = locate_internal_refs1_prep(ident)
  refs.each do |schema, ids|
    ids.keys.select { |id| t[id] }.each do |id|
      t[id].at("./ancestor-or-self::*[@type = '#{schema}']") and
        refs[schema][id] = identifier
    end
  end
end

#locate_internal_refs1_prep(ident) ⇒ Object 



146
147
148
149
150
151
152
153
154
155
 
# File 'lib/metanorma/collection_fileprocess.rb', line 146

def locate_internal_refs1_prep(ident)
  file, = @files.targetfile_id(ident, read: true)
  xml = Nokogiri::XML(file, &:huge)
  r = xml.root["document_suffix"]
  xml.xpath("//*[@id]").each_with_object({}) do |i, x|
    /^semantic_/.match?(i.name) and next
    x[i["id"]] = i
    r and x[i["id"].sub(/_#{r}$/, "")] = i
  end
end

#new_hidden_ref(xmldoc) ⇒ Object 



65
66
67
68
69
70
 
# File 'lib/metanorma/collection_fileparse.rb', line 65

def new_hidden_ref(xmldoc)
  ins = xmldoc.at(ns("bibliography")) or
    xmldoc.root << "<bibliography/>" and ins = xmldoc.at(ns("bibliography"))
  ins.at(ns("./referenced[@hidden = 'true']")) or
    ins.add_child("<references hidden='true' normative='false'/>").first
end

#overall_docconv_converter(body) ⇒ Object 



97
98
99
100
101
102
103
104
105
106
107
108
109
110
 
# File 'lib/metanorma/collection_render_word.rb', line 97

def overall_docconv_converter(body)
  collection_conv = docconv
  collection_conv.options[:collection_doc] = body.map(&:to_xml).join
  overall_docconv_cover(collection_conv)

  def collection_conv.postprocess_cleanup(result)
    ret = to_xhtml(super)
    b = ret.at("//div[@id = '_collection_placeholder']")
    b.replace(@options[:collection_doc])
    from_xhtml(ret)
  end

  collection_conv
end

#overall_docconv_cover(collection_conv) ⇒ Object 



88
89
90
91
92
93
94
95
 
# File 'lib/metanorma/collection_render_word.rb', line 88

def overall_docconv_cover(collection_conv)
  p = Util::hash_key_detect(@directives, "collection-word-coverpage", nil)
  collection_conv.wordcoverpage =
    Util::rel_path_resolve(@dirname, p)
  p = Util::hash_key_detect(@directives, "collection-word-intropage", nil)
  collection_conv.wordintropage =
    Util::rel_path_resolve(@dirname, p)
end

#populate_internal_refs(refs) ⇒ Object 



113
114
115
116
117
118
119
120
 
# File 'lib/metanorma/collection_fileprocess.rb', line 113

def populate_internal_refs(refs)
  @files.keys.reject do |k|
    @files.get(k, :attachment) || @files.get(k, :sectionsplit)
  end.each do |ident|
    locate_internal_refs1(refs, ident, @isodoc.docid_prefix("", ident.dup))
  end
  refs
end

#set_displayorder_wrapping_doc(doc) ⇒ Object 



44
45
46
47
48
49
50
 
# File 'lib/metanorma/collection_render_word.rb', line 44

def set_displayorder_wrapping_doc(doc)
  doc.xpath(ns("//preface/* | //sections/* | //annex"))
    .each_with_index do |x, i|
    x["displayorder"] = i + 1
  end
  doc
end

#strip_eref(eref) ⇒ Object 



42
43
44
45
 
# File 'lib/metanorma/collection_render_utils.rb', line 42

def strip_eref(eref)
  eref.xpath(ns("./locality | ./localityStack")).each(&:remove)
  eref.replace(eref.children)
end

#strip_unresolved_repo_erefs(_document_id, bib_docid, erefs, bibitem) ⇒ Object

strip erefs if they are repository erefs, but do not point to a document within the current collection. This can happen if a collection consists of many documents, but not all are included in the current collection. Do not do this if this is a sectionsplit collection or a nested manifest. Return false if bibitem is not to be further processed



151
152
153
154
155
156
157
 
# File 'lib/metanorma/collection_fileparse.rb', line 151

def strip_unresolved_repo_erefs(_document_id, bib_docid, erefs, bibitem)
  %r{^current-metanorma-collection/(?!Missing:)}.match?(bib_docid.text) and
    return true
  @nested and return false
  erefs[bibitem["id"]]&.each { |x| x.parent and strip_eref(x) }
  false
end

#supply_repo_ids(doc) ⇒ Object 



91
92
93
94
95
96
97
98
99
100
101
102
103
104
 
# File 'lib/metanorma/collection_fileparse.rb', line 91

def supply_repo_ids(doc)
  doc.xpath(
    ns("//bibitem[not(ancestor::bibitem)]" \
       "[not(./docidentifier[@type = 'repository'])]"),
  ).each do |b|
    b.xpath(ns("./docidentifier")).each do |docid|
      id = @isodoc.docid_prefix(docid["type"], docid.children.to_xml)
      @files.get(id) or next
      @files.get(id, :indirect_key) and next # will resolve as indirect key
      docid.next = "<docidentifier type='repository'>" \
                   "current-metanorma-collection/#{id}</docidentifier>"
    end
  end
end

#svgmap_resolve(docxml, ids) ⇒ Object 



106
107
108
109
110
111
112
113
114
 
# File 'lib/metanorma/collection_fileparse.rb', line 106

def svgmap_resolve(docxml, ids)
  isodoc = IsoDoc::PresentationXMLConvert.new({})
  isodoc.bibitem_lookup(docxml)
  docxml.xpath(ns("//svgmap//eref")).each do |e|
    svgmap_resolve1(e, isodoc, docxml, ids)
  end
  Vectory::SvgMapping.new(docxml, "").call
  docxml.xpath(ns("//svgmap")).each { |s| isodoc.svgmap_extract(s) }
end

#svgmap_resolve1(eref, isodoc, _docxml, ids) ⇒ Object 



116
117
118
119
120
121
122
123
 
# File 'lib/metanorma/collection_fileparse.rb', line 116

def svgmap_resolve1(eref, isodoc, _docxml, ids)
  href = isodoc.eref_target(eref) or return
  href == "##{eref['bibitemid']}" ||
    (href =~ /^#/ && !ids[href.sub(/^#/, "")]) and return
  eref["target"] = href.strip
  eref.name = "link"
  eref.elements&.remove
end

#update_anchor_create_loc(_bib, eref, docid) ⇒ Object

if there is a crossref to another document, with no anchor, retrieve the anchor given the locality, and insert it into the crossref



244
245
246
247
248
249
250
251
252
 
# File 'lib/metanorma/collection_fileparse.rb', line 244

def update_anchor_create_loc(_bib, eref, docid)
  ins = eref.at(ns("./localityStack")) or return
  type = ins.at(ns("./locality/@type"))&.text
  type = "clause" if type == "annex"
  ref = ins.at(ns("./locality/referenceFrom"))&.text
  anchor = @files.get(docid, :anchors).dig(type, ref) or return
  ins << "<locality type='anchor'><referenceFrom>#{anchor.sub(/^_/, '')}" \
         "</referenceFrom></locality>"
end

#update_anchor_loc(bib, eref, docid) ⇒ Object 



230
231
232
233
234
235
236
237
238
239
240
 
# File 'lib/metanorma/collection_fileparse.rb', line 230

def update_anchor_loc(bib, eref, docid)
  loc = eref.at(".//xmlns:locality[@type = 'anchor']") or
    return update_anchor_create_loc(bib, eref, docid)
  @ncnames[docid] ||= Metanorma::Utils::to_ncname(docid)
  ref = loc.at("./xmlns:referenceFrom") or return
  anchor = "#{ref.text}_#{@ncnames[docid]}"
  @files.get(docid, :anchors).inject([]) do |m, (_, x)|
    m += x.values
  end.include?(anchor) or return
  ref.content = anchor
end

#update_anchors(bib, docid, erefs) ⇒ Object

update crossrefences to other documents, to include disambiguating document suffix on id



217
218
219
220
221
222
223
224
225
226
227
228
 
# File 'lib/metanorma/collection_fileparse.rb', line 217

def update_anchors(bib, docid, erefs)
  erefs.each do |e|
    if @files.get(docid) then update_anchor_loc(bib, e, docid)
    else
      msg = "<strong>** Unresolved reference to document #{docid} " \
            "from eref</strong>"
      e << msg
      strip_eref(e)
      @log&.add("Cross-References", e, msg)
    end
  end
end

#update_bibitem(bib, identifier) ⇒ Object 



4
5
6
7
8
9
10
11
12
13
14
 
# File 'lib/metanorma/collection_fileparse.rb', line 4

def update_bibitem(bib, identifier)
  docid = get_bibitem_docid(bib, identifier) or return
  newbib = dup_bibitem(docid, bib)
  _file, url = @files
    .targetfile_id(docid, relative: true, read: false,
                          doc: !@files.get(docid, :attachment))
  dest = newbib.at("./docidentifier") || newbib.at(ns("./docidentifier"))
  dest or dest = newbib.elements[-1]
  dest.previous = "<uri type='citation'>#{url}</uri>"
  bib.replace(newbib)
end

#update_direct_refs_to_docs(docxml, identifier) ⇒ Object

repo(current-metanorma-collection/ISO 17301-1:2016) replaced by bibdata of “ISO 17301-1:2016” in situ as bibitem. Any erefs to that bibitem id are replaced with relative URL Preferably with anchor, and is a job to realise dynamic lookup of localities.



130
131
132
133
134
135
136
137
138
139
 
# File 'lib/metanorma/collection_fileparse.rb', line 130

def update_direct_refs_to_docs(docxml, identifier)
  erefs, erefs1 = update_direct_refs_to_docs_prep(docxml)
  docxml.xpath(ns("//bibitem")).each do |b|
    docid = b.at(ns("./docidentifier[@type = 'repository']")) or next
    strip_unresolved_repo_erefs(identifier, docid, erefs1, b) or next
    update_bibitem(b, identifier)
    docid = docid_to_citeas(b) or next
    erefs[docid] and update_anchors(b, docid, erefs[docid])
  end
end

#update_direct_refs_to_docs_prep(docxml) ⇒ Object 



141
142
143
144
 
# File 'lib/metanorma/collection_fileparse.rb', line 141

def update_direct_refs_to_docs_prep(docxml)
  @ncnames = {}
  [Util::gather_citeases(docxml), Util::gather_bibitemids(docxml)]
end

#update_indirect_refs_to_docs(docxml, _docidentifier, internal_refs) ⇒ Object

Resolve erefs to a container of ids in another doc, to an anchor eref (direct link)



161
162
163
164
165
166
167
168
169
170
 
# File 'lib/metanorma/collection_fileparse.rb', line 161

def update_indirect_refs_to_docs(docxml, _docidentifier, internal_refs)
  bibitems, erefs = update_indirect_refs_to_docs_prep(docxml)
  internal_refs.each do |schema, ids|
    ids.each do |id, file|
      k = indirect_ref_key(schema, id, docxml)
      update_indirect_refs_to_docs1(docxml, k,
                                    file, bibitems, erefs)
    end
  end
end

#update_indirect_refs_to_docs1(_docxml, key, file, bibitems, erefs) ⇒ Object 



188
189
190
191
192
193
194
 
# File 'lib/metanorma/collection_fileparse.rb', line 188

def update_indirect_refs_to_docs1(_docxml, key, file, bibitems, erefs)
  erefs[key]&.each do |e|
    e["citeas"] = file
    update_indirect_refs_to_docs_anchor(e, file)
  end
  update_indirect_refs_to_docs_docid(bibitems[key], file)
end

#update_indirect_refs_to_docs_anchor(eref, file) ⇒ Object 



196
197
198
199
200
201
202
203
204
205
 
# File 'lib/metanorma/collection_fileparse.rb', line 196

def update_indirect_refs_to_docs_anchor(eref, file)
  a = eref.at(ns(".//locality[@type = 'anchor']/referenceFrom")) or return
  suffix = file
  @files.get(file) && p = @files.get(file, :parentid) and
    suffix = "#{p}_#{suffix}"
  existing = a.text
  anchor = Metanorma::Utils::to_ncname("#{existing}_#{suffix}")
  @updated_anchors[existing] or a.children = anchor
  @updated_anchors[anchor] = true
end

#update_indirect_refs_to_docs_docid(bibitem, file) ⇒ Object 



207
208
209
210
211
212
213
 
# File 'lib/metanorma/collection_fileparse.rb', line 207

def update_indirect_refs_to_docs_docid(bibitem, file)
  docid = bibitem&.at(ns("./docidentifier[@type = 'repository']")) or
    return
  docid.children = "current-metanorma-collection/#{file}"
  docid.previous =
    "<docidentifier type='metanorma-collection'>#{file}</docidentifier>"
end

#update_indirect_refs_to_docs_prep(docxml) ⇒ Object 



172
173
174
175
176
177
 
# File 'lib/metanorma/collection_fileparse.rb', line 172

def update_indirect_refs_to_docs_prep(docxml)
  bibitems = Util::gather_bibitems(docxml)
  erefs = Util::gather_bibitemids(docxml)
  @updated_anchors = {}
  [bibitems, erefs]
end

#update_sectionsplit_eref_to_doc(eref, internal_refs, doclist, opts) ⇒ Object 



55
56
57
58
59
60
61
62
63
 
# File 'lib/metanorma/collection_fileparse.rb', line 55

def update_sectionsplit_eref_to_doc(eref, internal_refs, doclist, opts)
  a = eref.at(ns("./localityStack/locality[@type = 'anchor']/" \
                 "referenceFrom")) or return
  doc = internal_refs[opts[:key]]["#{a.text}_#{opts[:target_suffix]}"]
  bibitemid = Metanorma::Utils::to_ncname("#{doc}_#{opts[:source_suffix]}")
  eref["bibitemid"] = bibitemid
  doclist[bibitemid] ||= doc
  doclist
end

#update_sectionsplit_refs_to_docs(docxml, internal_refs) ⇒ Object 



42
43
44
45
46
47
48
49
50
51
52
53
 
# File 'lib/metanorma/collection_fileparse.rb', line 42

def update_sectionsplit_refs_to_docs(docxml, internal_refs)
  Util::gather_citeases(docxml).each do |k, v|
    (@files.get(k) && @files.get(k, :sectionsplit)) or next
    opts = { key: @files.get(k, :indirect_key),
             source_suffix: docxml.root["document_suffix"],
             target_suffix: @files.get(k, :document_suffix) }
    refs = v.each_with_object({}) do |eref, m|
      update_sectionsplit_eref_to_doc(eref, internal_refs, m, opts)
    end
    add_hidden_bibliography(docxml, refs)
  end
end

#update_xrefs(file, docid, internal_refs) ⇒ String

Resolves references to other files in the collection. Three routines:

  1. Eref to a document that has been split into multiple documents

(sectionsplit) are resolved to direct eref to the split document

  1. Indirect erefs to a file anchor in an unknwon file in the collection

(bibitem[@type = ‘internal’] ) are resolved to direct eref to the containing document

  1. Direct erefs to other files in collection

(repo(current-metanorma-collection/x) are resolved to hyperlinks

Parameters:

  • file (String) —

    XML content

  • identifier (String) —

    docid

  • internal_refs (Hash{String=>Hash{String=>String}] schema name to anchor to filename) —

    nternal_refs [HashString=>Hash{String=>String] schema name to anchor to filename

Returns:

  • (String) —

    XML content



29
30
31
32
33
34
35
36
37
38
39
40
 
# File 'lib/metanorma/collection_fileparse.rb', line 29

def update_xrefs(file, docid, internal_refs)
  docxml = file.is_a?(String) ? Nokogiri::XML(file, &:huge) : file
  supply_repo_ids(docxml)
  @nested or update_indirect_refs_to_docs(docxml, docid, internal_refs)
  @files.add_document_suffix(docid, docxml)
  @nested or update_sectionsplit_refs_to_docs(docxml, internal_refs)
  update_direct_refs_to_docs(docxml, docid)
  hide_refs(docxml)
  @files.get(docid, :sectionsplit_output) and eref2link(docxml)
  svgmap_resolve(datauri_encode(docxml), @files.get(docid, :ids))
  docxml.to_xml
end

#wrapping_doc(doc, xml) ⇒ Object 



20
21
22
23
24
25
 
# File 'lib/metanorma/collection_render_word.rb', line 20

def wrapping_doc(doc, xml)
  doc.at(ns("//bibdata")).replace(xml.at(ns("//bibdata")).to_xml)
  sections = wrapping_doc_body(doc)
  wrapping_doc_intro_outro(xml, sections)
  set_displayorder_wrapping_doc(doc)
end

#wrapping_doc_body(doc) ⇒ Object 



34
35
36
37
38
39
40
41
42
 
# File 'lib/metanorma/collection_render_word.rb', line 34

def wrapping_doc_body(doc)
  doc.xpath(ns("//annex | //preface | //bibliography")).each(&:remove)
  s = doc.at(ns("//sections"))
  repl = <<~BODY
    <sections><clause id='_collection_placeholder'><p>PLACEHOLDER</p></clause></sections>
  BODY
  s.replace(repl)
  doc.at(ns("//sections"))
end

#wrapping_doc_intro_outro(xml, sections) ⇒ Object 



27
28
29
30
31
32
 
# File 'lib/metanorma/collection_render_word.rb', line 27

def wrapping_doc_intro_outro(xml, sections)
  p = xml.at(ns("//prefatory-content")) and
    sections.previous = "<preface>#{p.children.to_xml}</preface>"
  p = xml.at(ns("//final-content")) and
    sections.next = "<annex>#{p.children.to_xml}</annex>"
end