Class: Metanorma::FileLookup

Inherits:
Object
  • Object
show all
Defined in:
lib/metanorma/files_lookup.rb,
lib/metanorma/files_lookup_sectionsplit.rb

Overview

XML collection renderer

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(path, parent) ⇒ FileLookup

hash for each document in collection of document identifier to: document reference (fileref or id), type of document reference, and bibdata entry for that file

Parameters:

  • path (String) —

    path to collection



15
16
17
18
19
20
21
22
23
24
25
26
 
# File 'lib/metanorma/files_lookup.rb', line 15

def initialize(path, parent)
  @c = HTMLEntities.new
  @files = {}
  @parent = parent
  @xml = parent.xml
  @isodoc = parent.isodoc
  @path = path
  @compile = parent.compile
  @documents = parent.documents
  @files_to_delete = []
  read_files
end

Instance Attribute Details

#files_to_delete ⇒ Object

Returns the value of attribute files_to_delete.



9
10
11
 
# File 'lib/metanorma/files_lookup.rb', line 9

def files_to_delete
  @files_to_delete
end

#parent ⇒ Object

Returns the value of attribute parent.



9
10
11
 
# File 'lib/metanorma/files_lookup.rb', line 9

def parent
  @parent
end

Instance Method Details

#add_document_suffix(identifier, doc) ⇒ Object 



87
88
89
90
91
92
93
94
95
96
 
# File 'lib/metanorma/files_lookup.rb', line 87

def add_document_suffix(identifier, doc)
  document_suffix = Metanorma::Utils::to_ncname(identifier)
  Metanorma::Utils::anchor_attributes.each do |(tag_name, attribute_name)|
    Util::add_suffix_to_attributes(doc, document_suffix, tag_name,
                                   attribute_name, @isodoc)
  end
  url_in_css_styles(doc, document_suffix)
  doc.root["document_suffix"] ||= ""
  doc.root["document_suffix"] += document_suffix
end

#add_section_split ⇒ Object 



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

def add_section_split
  ret = @files.keys.each_with_object({}) do |k, m|
    if @files[k][:sectionsplit] == "true" && !@files[k]["attachment"]
      process_section_split_instance(k, m)
      cleanup_section_split_instance(k, m)
    end
    m[k] = @files[k]
  end
  @files = ret
end

#add_section_split_cover(manifest, ident) ⇒ Object 



31
32
33
34
35
36
37
38
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 31

def add_section_split_cover(manifest, ident)
  cover = @sectionsplit
    .section_split_cover(manifest, @parent.dir_name_cleanse(ident),
                         one_doc_collection?)
  @files[ident][:out_path] = cover
  { attachment: true, index: false, out_path: cover,
    ref: File.join(File.dirname(manifest.file), cover) }
end

#add_section_split_instance(file, manifest, key, idx) ⇒ Object 



51
52
53
54
55
56
57
58
59
60
61
62
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 51

def add_section_split_instance(file, manifest, key, idx)
  presfile, newkey, xml =
    add_section_split_instance_prep(file, key)
  manifest[newkey] =
    { parentid: key, presentationxml: true, type: "fileref",
      rel_path: file[:url], out_path: File.basename(file[:url]),
      anchors: read_anchors(xml), ids: read_ids(xml),
      sectionsplit_output: true,
      bibdata: @files[key][:bibdata], ref: presfile }
  @files_to_delete << file[:url]
  manifest[newkey][:bare] = true unless idx.zero?
end

#add_section_split_instance_prep(file, key) ⇒ Object 



64
65
66
67
68
69
70
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 64

def add_section_split_instance_prep(file, key)
  presfile = File.join(File.dirname(@files[key][:ref]),
                       File.basename(file[:url]))
  newkey = key("#{key.strip} #{file[:title]}")
  xml = Nokogiri::XML(File.read(presfile), &:huge)
  [presfile, newkey, xml]
end

#bibdata_process(entry, ident) ⇒ Object 



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

def bibdata_process(entry, ident)
  if entry[:attachment]
    entry[:bibdata] = Metanorma::Document.attachment_bibitem(ident).root
  else
    file, _filename = targetfile(entry, read: true)
    xml = Nokogiri::XML(file, &:huge)
    add_document_suffix(ident, xml)
    entry.merge!(anchors: read_anchors(xml), ids: read_ids(xml),
                 bibdata: xml.at(ns("//bibdata")),
                 document_suffix: xml.root["document_suffix"])
  end
end

#bibitem_process(entry) ⇒ Object 



55
56
57
58
59
60
 
# File 'lib/metanorma/files_lookup.rb', line 55

def bibitem_process(entry)
  entry[:bibitem] = entry[:bibdata].dup
  entry[:bibitem].name = "bibitem"
  entry[:bibitem]["hidden"] = "true"
  entry[:bibitem].at("./*[local-name() = 'ext']")&.remove
end

#cleanup_section_split_instance(key, manifest) ⇒ Object 



25
26
27
28
29
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 25

def cleanup_section_split_instance(key, manifest)
  @files_to_delete << manifest["#{key}:index.html"][:ref]
  #@files[key].delete(:ids).delete(:anchors)
  @files[key][:indirect_key] = @sectionsplit.key
end

#each ⇒ Object 



212
213
214
 
# File 'lib/metanorma/files_lookup.rb', line 212

def each
  @files.each
end

#each_with_index ⇒ Object 



216
217
218
 
# File 'lib/metanorma/files_lookup.rb', line 216

def each_with_index
  @files.each_with_index
end

#file_entry(ref, identifier) ⇒ Object

rel_path is the source file address, determined relative to the YAML. out_path is the destination file address, with any references outside the working directory (../../…) truncated identifier is the id with only spaces, no nbsp



66
67
68
69
70
71
72
73
74
75
76
77
78
 
# File 'lib/metanorma/files_lookup.rb', line 66

def file_entry(ref, identifier)
  ref["fileref"] or return
  out = ref["attachment"] ? ref["fileref"] : File.basename(ref["fileref"])
  out1 = @disambig.source2dest_filename(out)
  ret = if ref["fileref"]
          { type: "fileref", ref: @documents[Util::key identifier].file,
            rel_path: ref["fileref"], url: ref["url"],
            out_path: out1 } # @disambig.source2dest_filename(out) }
        else { type: "id", ref: ref["id"] }
        end
  file_entry_copy(ref, ret)
  ret.compact
end

#file_entry_copy(ref, ret) ⇒ Object 



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

def file_entry_copy(ref, ret)
  %w(attachment sectionsplit index presentation-xml url
     bare-after-first).each do |s|
    ret[s.gsub("-", "").to_sym] = ref[s] if ref[s]
  end
end

#get(ident, attr = nil) ⇒ Object 



202
203
204
205
206
 
# File 'lib/metanorma/files_lookup.rb', line 202

def get(ident, attr = nil)
  if attr then @files[key(ident)][attr]
  else @files[key(ident)]
  end
end

#key(ident) ⇒ Object 



194
195
196
 
# File 'lib/metanorma/files_lookup.rb', line 194

def key(ident)
  @c.decode(ident).gsub(/(\p{Zs})+/, " ").sub(/^metanorma-collection /, "")
end

#keys ⇒ Object 



198
199
200
 
# File 'lib/metanorma/files_lookup.rb', line 198

def keys
  @files.keys
end

#ns(xpath) ⇒ Object 



220
221
222
 
# File 'lib/metanorma/files_lookup.rb', line 220

def ns(xpath)
  @isodoc.ns(xpath)
end

#one_doc_collection? ⇒ Boolean

Returns:

  • (Boolean) 


40
41
42
43
44
45
46
47
48
49
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 40

def one_doc_collection?
  return false
  docs = 0
  @files.each_value do |v|
    v[:attachment] and next
    v[:presentationxml] and next
    docs += 1
  end
  docs > 1
end

#process_section_split_instance(key, manifest) ⇒ Object 



15
16
17
18
19
20
21
22
23
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 15

def process_section_split_instance(key, manifest)
  s, sectionsplit_manifest = sectionsplit(@files[key][:ref],
                                          @files[key][:out_path], key)
  s.each_with_index do |f1, i|
    add_section_split_instance(f1, manifest, key, i)
  end
  manifest["#{key}:index.html"] =
    add_section_split_cover(sectionsplit_manifest, key)
end

#read_anchors(xml) ⇒ Object

map locality type and label (e.g. “clause” “1”) to id = anchor for a document Note: will only key clauses, which have unambiguous reference label in locality. Notes, examples etc with containers are just plunked against UUIDs, so that their IDs can at least be registered to be tracked as existing.



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

def read_anchors(xml)
  xrefs = @isodoc.xref_init(@lang, @script, @isodoc, @isodoc.i18n,
                            { locale: @locale })
  xrefs.parse xml
  xrefs.get.each_with_object({}) do |(k, v), ret|
    read_anchors1(k, v, ret)
  end
end

#read_anchors1(key, val, ret) ⇒ Object 



172
173
174
175
176
177
178
179
180
181
 
# File 'lib/metanorma/files_lookup.rb', line 172

def read_anchors1(key, val, ret)
  val[:type] ||= "clause"
  ret[val[:type]] ||= {}
  index = if val[:container] || val[:label].nil? || val[:label].empty?
            UUIDTools::UUID.random_create.to_s
          else val[:label]
          end
  ret[val[:type]][index] = key
  ret[val[:type]][val[:value]] = key if val[:value]
end

#read_file(docref) ⇒ Object 



33
34
35
36
37
38
39
40
 
# File 'lib/metanorma/files_lookup.rb', line 33

def read_file(docref)
  ident = docref.at(ns("./identifier"))
  i = key(@isodoc.docid_prefix(ident["type"], ident.children.to_xml))
  entry = file_entry(docref, ident.children.to_xml) or return
  bibdata_process(entry, i)
  bibitem_process(entry)
  @files[i] = entry
end

#read_files ⇒ Object 



28
29
30
31
 
# File 'lib/metanorma/files_lookup.rb', line 28

def read_files
  @disambig = Util::DisambigFiles.new
  @xml.xpath(ns("//docref")).each { |d| read_file(d) }
end

#read_ids(xml) ⇒ Object

Also parse all ids in doc (including ones which won’t be xref targets)



184
185
186
187
188
189
190
191
192
 
# File 'lib/metanorma/files_lookup.rb', line 184

def read_ids(xml)
  ret = {}
  xml.traverse do |x|
    x.text? and next
    /^semantic__/.match?(x.name) and next
    x["id"] and ret[x["id"]] = true
  end
  ret
end

#ref_file(ref, out, read, doc) ⇒ Object 



144
145
146
147
148
149
 
# File 'lib/metanorma/files_lookup.rb', line 144

def ref_file(ref, out, read, doc)
  file = File.read(ref, encoding: "utf-8") if read
  filename = out.dup
  filename.sub!(/\.xml$/, ".html") if doc
  [file, filename]
end

#sectionsplit(file, outfile, ident) ⇒ Object 



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

def sectionsplit(file, outfile, ident)
  @sectionsplit = Sectionsplit
    .new(input: file, base: outfile, dir: File.dirname(file),
         output: outfile, compile_opts: @parent.compile_options,
         fileslookup: self, ident: ident, isodoc: @isodoc)
  coll = @sectionsplit.sectionsplit.sort_by { |f| f[:order] }
  xml = Nokogiri::XML(File.read(file, encoding: "UTF-8"), &:huge)
  [coll, @sectionsplit
    .collection_manifest(File.basename(file), coll, xml, nil,
                         File.dirname(file))]
end

#set(ident, attr, value) ⇒ Object 



208
209
210
 
# File 'lib/metanorma/files_lookup.rb', line 208

def set(ident, attr, value)
  @files[key(ident)][attr] = value
end

#targetfile(data, options) ⇒ Array<String, nil>

return file contents + output filename for each file in the collection, given a docref entry so my URL should end with html or pdf or whatever formed relative to YAML file, not input path, relative to calling function

Parameters:

  • data (Hash) —

    docref entry

  • read (Boolean) —

    read the file in and return it

  • doc (Boolean) —

    I am a Metanorma document,

  • relative (Boolean) —

    Return output path,

Returns:

  • (Array<String, nil>) 


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

def targetfile(data, options)
  options = { read: false, doc: true, relative: false }.merge(options)
  path = options[:relative] ? data[:rel_path] : data[:ref]
  if data[:type] == "fileref"
    ref_file path, data[:out_path], options[:read], options[:doc]
  else
    xml_file data[:id], options[:read]
  end
end

#targetfile_id(ident, options) ⇒ Object 



140
141
142
 
# File 'lib/metanorma/files_lookup.rb', line 140

def targetfile_id(ident, options)
  targetfile(get(ident), options)
end

#url(ident, options) ⇒ Object

return citation url for file so my URL should end with html or pdf or whatever

Parameters:

  • doc (Boolean) —

    I am a Metanorma document,



109
110
111
112
 
# File 'lib/metanorma/files_lookup.rb', line 109

def url(ident, options)
  data = get(ident)
  data[:url] || targetfile(data, options)[1]
end

#url?(ident) ⇒ Boolean

are references to the file to be linked to a file in the collection, or externally? Determines whether file suffix anchors are to be used

Returns:

  • (Boolean) 


116
117
118
119
 
# File 'lib/metanorma/files_lookup.rb', line 116

def url?(ident)
  data = get(ident) or return false
  data[:url]
end

#url_in_css_styles(doc, document_suffix) ⇒ Object

update relative URLs, url(#…), in CSS in @style attrs (including SVG)



99
100
101
102
103
104
 
# File 'lib/metanorma/files_lookup.rb', line 99

def url_in_css_styles(doc, document_suffix)
  doc.xpath("//*[@style]").each do |s|
    s["style"] = s["style"]
      .gsub(%r{url\(#([^)]+)\)}, "url(#\\1_#{document_suffix})")
  end
end

#xml_file(id, read) ⇒ Object 



151
152
153
154
155
 
# File 'lib/metanorma/files_lookup.rb', line 151

def xml_file(id, read)
  file = @xml.at(ns("//doc-container[@id = '#{id}']")).to_xml if read
  filename = "#{id}.html"
  [file, filename]
end