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 



94
95
96
97
98
99
100
 
# File 'lib/metanorma/files_lookup.rb', line 94

def add_document_suffix(identifier, doc)
  document_suffix = Metanorma::Utils::to_ncname(identifier)
  Metanorma::Utils::anchor_attributes.each do |(tag_name, attribute_name)|
    add_suffix_to_attributes(doc, document_suffix, tag_name, attribute_name)
  end
  url_in_css_styles(doc, document_suffix)
end

#add_section_split ⇒ Object 



4
5
6
7
8
9
10
11
12
13
14
15
 
# 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"]
      s, manifest = sectionsplit(@files[k][:ref], @files[k][:out_path], k)
      s.each_with_index { |f1, i| add_section_split_instance(f1, m, k, i) }
      m["#{k}:index.html"] = add_section_split_cover(manifest, k)
      @files_to_delete << m["#{k}:index.html"][:ref]
    end
    m[k] = @files[k]
  end
  @files = ret
end

#add_section_split_cover(manifest, ident) ⇒ Object 



17
18
19
20
21
22
23
24
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 17

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 



36
37
38
39
40
41
42
43
44
45
46
47
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 36

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 



49
50
51
52
53
54
55
 
# File 'lib/metanorma/files_lookup_sectionsplit.rb', line 49

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

#add_suffix_to_attributes(doc, suffix, tag_name, attribute_name) ⇒ Object 



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

def add_suffix_to_attributes(doc, suffix, tag_name, attribute_name)
  doc.xpath(ns("//#{tag_name}[@#{attribute_name}]")).each do |elem|
    elem.attributes[attribute_name].value =
      "#{elem.attributes[attribute_name].value}_#{suffix}"
  end
end

#bibdata_process(entry, identifier) ⇒ Object 



42
43
44
45
46
47
48
49
50
51
52
53
54
 
# File 'lib/metanorma/files_lookup.rb', line 42

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

#bibitem_process(entry) ⇒ Object 



56
57
58
59
60
61
 
# File 'lib/metanorma/files_lookup.rb', line 56

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

#each ⇒ Object 



201
202
203
 
# File 'lib/metanorma/files_lookup.rb', line 201

def each
  @files.each
end

#each_with_index ⇒ Object 



205
206
207
 
# File 'lib/metanorma/files_lookup.rb', line 205

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



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

def file_entry(ref, identifier)
  out = ref["attachment"] ? ref["fileref"] : File.basename(ref["fileref"])
  out1 = @disambig.source2dest_filename(out)
  ret = if ref["fileref"]
          { type: "fileref", ref: @documents[identifier].file,
            rel_path: ref["fileref"],
            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
     bare-after-first).each do |s|
    ret[s.gsub("-", "").to_sym] = ref[s] if ref[s]
  end
end

#get(ident, attr = nil) ⇒ Object 



191
192
193
194
195
 
# File 'lib/metanorma/files_lookup.rb', line 191

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

#key(ident) ⇒ Object 



183
184
185
 
# File 'lib/metanorma/files_lookup.rb', line 183

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

#keys ⇒ Object 



187
188
189
 
# File 'lib/metanorma/files_lookup.rb', line 187

def keys
  @files.keys
end

#ns(xpath) ⇒ Object 



209
210
211
 
# File 'lib/metanorma/files_lookup.rb', line 209

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

#one_doc_collection? ⇒ Boolean

Returns:

  • (Boolean) 


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

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



152
153
154
155
156
157
158
159
 
# File 'lib/metanorma/files_lookup.rb', line 152

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 



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

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)
  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)



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

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 



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

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 



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

def sectionsplit(file, outfile, ident)
  @sectionsplit = Sectionsplit
    .new(input: file, base: outfile, dir: File.dirname(file),
         output: outfile, compile_options: @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 



197
198
199
 
# File 'lib/metanorma/files_lookup.rb', line 197

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>) 


119
120
121
122
123
124
125
126
127
 
# File 'lib/metanorma/files_lookup.rb', line 119

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 



129
130
131
 
# File 'lib/metanorma/files_lookup.rb', line 129

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

#url_in_css_styles(doc, document_suffix) ⇒ Object

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



103
104
105
106
107
108
 
# File 'lib/metanorma/files_lookup.rb', line 103

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 



140
141
142
143
144
 
# File 'lib/metanorma/files_lookup.rb', line 140

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