Class: Metanorma::Collection::Sectionsplit

Inherits:
Object
  • Object
show all
Defined in:
lib/metanorma/collection/sectionsplit/sectionsplit.rb,
lib/metanorma/collection/sectionsplit/collection.rb

Overview

Sectionsplit is a collection render nested inside a collection render: a document is split into one output file per section, and those files are then rendered as their own (sub-)collection.

Two output-location invariants matter (both load-bearing -- see the collection architecture review plan and metanorma/iso-10303#208):

  • XML section files are always written FLAT to the split directory (basename only). HTML output may instead carry a directory taken from sectionsplit_filename, reattached only at HTML compile time (preserve_directory_structure?).
  • A sectionsplit document's content is emitted at the sectionsplit output location (the collection root, unless sectionsplit_filename sets a directory), NOT at the document's own :out_path. Anything relativising a URL for such a document (e.g. attachment/citation links) must base it on the split location, or the ../../ overshoots.

Constant Summary collapse

SPLITSECTIONS = 
[["//preface/*", "preface"], ["//sections/*", "sections"],
["//annex", nil],
["//bibliography/*[not(@hidden = 'true')]", "bibliography"],
["//indexsect", nil], ["//colophon", nil]].freeze
FN_CAPTIONS = 
".//fmt-fn-label/span[@class = 'fmt-caption-label']".freeze

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(opts) ⇒ Sectionsplit

Returns a new instance of Sectionsplit.



29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 29

def initialize(opts)
  @input_filename = opts[:input]
  @base = opts[:base]
  @output_filename = opts[:output]
  @xml = opts[:xml]
  @dir = opts[:dir]
  @compile_opts = opts[:compile_opts] || {}
  @fileslookup = opts[:fileslookup]
  @ident = opts[:ident]
  @isodoc = opts[:isodoc]
  @isodoc_presxml = opts[:isodoc_presxml]
  @document_suffix = opts[:document_suffix]
  @sectionsplit_filename = opts[:sectionsplit_filename] ||
    "{basename_legacy}.{sectionsplit-num}"
  @parent_idx = opts[:parent_idx] || 0
end

Instance Attribute Details

#filecache ⇒ Object

Returns the value of attribute filecache.



27
28
29
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 27

def filecache
  @filecache
end

#key ⇒ Object

Returns the value of attribute key.



27
28
29
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 27

def key
  @key
end

#whole_presentation_file ⇒ Object

Returns the value of attribute whole_presentation_file.



27
28
29
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 27

def whole_presentation_file
  @whole_presentation_file
end

Instance Method Details

#att_dir(file) ⇒ Object 



75
76
77
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 75

def att_dir(file)
  "_#{File.basename(file, '.*')}_attachments"
end

#block?(node) ⇒ Boolean

TODO move to metanorma-utils

Returns:

  • (Boolean) 


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

def block?(node)
  %w(p table formula admonition ol ul dl figure quote sourcecode example
     pre note pagebreak hr bookmark requirement recommendation permission
     svgmap inputform toc passthrough annotation imagemap).include?(node.name)
end

#build_collection ⇒ Object 



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

def build_collection
  collection_setup(@base, @dir)
  files = sectionsplit
  input_xml = Nokogiri::XML(File.read(@input_filename,
                                      encoding: "UTF-8"), &:huge)
  collection_manifest(@base, files, input_xml, @xml, @dir).render(
    { format: %i(html), output_folder: "#{@output_filename}_collection",
      coverpage: File.join(@dir, "cover.html") }.merge(@compile_opts),
  )
  section_split_attachments(out: "#{@output_filename}_collection")
end

#coll_cover ⇒ Object 



24
25
26
27
28
29
30
31
32
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 24

def coll_cover
  <<~COVER
    <html><head><meta charset="UTF-8"/></head><body>
          <h1>{{ doctitle }}</h1>
          <h2>{{ docnumber }}</h2>
          <nav>{{ navigation }}</nav>
        </body></html>
  COVER
end

#collection_manifest(filename, files, origxml, _presxml, dir) ⇒ Object 



34
35
36
37
38
39
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 34

def collection_manifest(filename, files, origxml, _presxml, dir)
  File.open(File.join(dir, "#{filename}.html.yaml"), "w:UTF-8") do |f|
    f.write(collectionyaml(files, origxml))
  end
  Metanorma::Collection.parse File.join(dir, "#{filename}.html.yaml")
end

#collection_setup(filename, dir) ⇒ Object 



16
17
18
19
20
21
22
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 16

def collection_setup(filename, dir)
  FileUtils.mkdir_p "#{filename}_collection" if filename
  FileUtils.mkdir_p dir
  File.open(File.join(dir, "cover.html"), "w:UTF-8") do |f|
    f.write(coll_cover)
  end
end

#collectionyaml(files, xml) ⇒ Object 



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
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 41

def collectionyaml(files, xml)
  ret = {
    directives: ["presentation-xml", "bare-after-first"],
    bibdata: {
      title: {
        type: "title-main", language: @lang,
        content: xml.at(ns("//bibdata/title"))&.text || "[TITLE]"
      },
      type: "collection",
      docid: {
        type: xml.at(ns("//bibdata/docidentifier/@type"))&.text || "[FLAVOR]",
        id: xml.at(ns("//bibdata/docidentifier"))&.text || "[DOCID]",
      },
    },
    manifest: {
      level: "collection", title: "Collection",
      docref: files.sort_by { |f| f[:order] }.each.map do |f|
        # XML files are always stored flat, so fileref is always the basename
        # f[:url] may include directory for HTML output, but XML is basename only
        fileref = File.basename(f[:url])
        entry = { fileref: fileref, identifier: f[:title] }
        # Include sectionsplit_filename and sectionsplit-output when there's a directory structure
        # This tells the renderer to preserve directory structure in HTML output
        if @sectionsplit_filename && File.dirname(@sectionsplit_filename) != "."
          entry[:"sectionsplit-filename"] = @sectionsplit_filename
          entry[:"sectionsplit-output"] = true
        end
        entry
      end
    },
  }
  ::Metanorma::Util::recursive_string_keys(ret).to_yaml
end

#conflate_floatingtitles(nodes) ⇒ Object 



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

def conflate_floatingtitles(nodes)
  holdover = false
  nodes.each_with_object([]) do |x, m|
    if holdover then m.last << x
    else m << [x]
    end
    holdover = block?(x)
  end
end

#create_sectionfile(xml, out, file, chunks, parentnode) ⇒ Object 



192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 192

def create_sectionfile(xml, out, file, chunks, parentnode)
  ins = out.at(ns("//metanorma-extension")) || out.at(ns("//bibdata"))
  sectionfile_insert(ins, chunks, parentnode)
  sectionfile_fn_filter(sectionfile_annotation_filter(out))
  Metanorma::Collection::XrefProcess::xref_process(out, xml, @key,
                                                   @ident, @isodoc, true)
  # XML files are always written flat to the splitdir (_files directory)
  # Directory structure from sectionsplit_filename is only used for HTML output
  xml_filename = "#{File.basename(file)}.xml"
  full_path = File.join(@splitdir, xml_filename)
  File.open(full_path, "w:UTF-8") { |f| f.write(out) }
  # Return full file path (with directory) for use in manifest
  # This preserves directory info for HTML rendering without affecting XML storage
  "#{file}.xml"
end

#empty_attachments(xml) ⇒ Object 



182
183
184
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 182

def empty_attachments(xml)
  xml.dup
end

#empty_doc(xml) ⇒ Object 



171
172
173
174
175
176
177
178
179
180
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 171

def empty_doc(xml)
  out = xml.dup
  out.xpath(
    ns("//preface | //sections | //annex | " \
    "//references/bibitem[not(@hidden = 'true')] | " \
    "//indexsect | //colophon"),
  ).each(&:remove)
  ::Metanorma::Collection::Util::hide_refs(out)
  out
end

#ns(xpath) ⇒ Object 



46
47
48
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 46

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

#section_split_attachments(out: nil) ⇒ Object 



79
80
81
82
83
84
85
86
87
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 79

def section_split_attachments(out: nil)
  attachments = att_dir(@tmp_filename)
  File.directory?(attachments) or return
  dir = out || File.dirname(@input_filename)
  ret = File.join(dir, att_dir(@output_filename))
  FileUtils.rm_rf ret
  FileUtils.mv attachments, ret
  File.basename(ret)
end

#section_split_cover(col, ident, one_doc_coll) ⇒ Object 



89
90
91
92
93
94
95
96
97
98
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 89

def section_split_cover(col, ident, one_doc_coll)
  dir = File.dirname(col.file)
  collection_setup(nil, dir)
  r = ::Metanorma::Collection::Renderer
    .new(col, dir, output_folder: "#{ident}_collection",
                   format: %i(html),
                   coverpage: File.join(dir, "cover.html"))
  r.coverpage
  section_split_cover1(ident, r, dir, one_doc_coll)
end

#section_split_cover1(ident, renderer, dir, _one_doc_coll) ⇒ Object 



100
101
102
103
104
105
106
107
 
# File 'lib/metanorma/collection/sectionsplit/collection.rb', line 100

def section_split_cover1(ident, renderer, dir, _one_doc_coll)
  filename = File.basename("#{ident}_index.html")
  # ident can be a directory with YAML indirection
  dest = File.join(dir, filename)
  FileUtils.mv File.join(renderer.outdir, "index.html"), dest
  FileUtils.rm_rf renderer.outdir
  filename
end

#sectionfile(fulldoc, xml, file, chunks, parentnode) ⇒ Object 



186
187
188
189
190
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 186

def sectionfile(fulldoc, xml, file, chunks, parentnode)
  fname = create_sectionfile(fulldoc, xml.dup, file, chunks, parentnode)
  { order: chunks.last["displayorder"].to_i, url: fname,
    title: titlerender(chunks.last) }
end

#sectionfile_annotation_filter(xml) ⇒ Object 



268
269
270
271
272
273
274
275
276
277
278
279
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 268

def sectionfile_annotation_filter(xml)
  ids = sectionfile_annotation_filter_prep(xml)
  xml.root.xpath(ns("./annotation-container/fmt-annotation-body"))
    .each do |f|
    ids.has_key?(f["id"]) or f.remove
  end
  xml.root.xpath(ns("./annotation-container/fmt-annotation-body"))
    .each_with_index do |fnbody, i|
      sectionfile_annotation_filter_renumber(fnbody, i, ids)
    end
  xml
end

#sectionfile_annotation_filter_prep(xml) ⇒ Object

map fmt-annotation-body/@id = fmt-annotation-start/end/@target to fmt-annotation-start/end



260
261
262
263
264
265
266
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 260

def sectionfile_annotation_filter_prep(xml)
  xml.xpath(ns("//fmt-annotation-start | //fmt-annotation-end"))
    .each_with_object({}) do |f, m|
      m[f["target"]] ||= []
      m[f["target"]] << f
    end
end

#sectionfile_annotation_filter_renumber(fnbody, _idx, ids) ⇒ Object 



281
282
283
284
285
286
287
288
289
290
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 281

def sectionfile_annotation_filter_renumber(fnbody, _idx, ids)
  ids[fnbody["id"]].each do |f|
    case f.name
    when "fmt-annotation-start"
      f.children = @isodoc_presxml.comment_bookmark_start_label(f)
    when "fmt-annotation-end"
      f.children = @isodoc_presxml.comment_bookmark_end_label(f)
    end
  end
end

#sectionfile_fn_filter(xml) ⇒ Object 



216
217
218
219
220
221
222
223
224
225
226
227
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 216

def sectionfile_fn_filter(xml)
  ids = sectionfile_fn_filter_prep(xml)
  xml.root.xpath(ns("./fmt-footnote-container/fmt-fn-body")).each do |f|
    ids.has_key?(f["id"]) or f.remove
  end
  seen = {}
  xml.root.xpath(ns("/fmt-footnote-container/fmt-fn-body"))
    .each_with_index do |fnbody, i|
      sectionfile_fn_filter_renumber(fnbody, i, ids, seen)
    end
  xml
end

#sectionfile_fn_filter_fn_renumber(fnbody, idx, ids, seen) ⇒ Object 



244
245
246
247
248
249
250
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 244

def sectionfile_fn_filter_fn_renumber(fnbody, idx, ids, seen)
  ids[fnbody["id"]].each do |f|
    @isodoc_presxml.renumber_document_footnote(f, idx, seen)
    fnlabel = f.at(ns(FN_CAPTIONS)) and
      fnlabel.children = @isodoc_presxml.fn_ref_label(f)
  end
end

#sectionfile_fn_filter_fnbody_renumber(fnbody, _idx, ids) ⇒ Object 



252
253
254
255
256
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 252

def sectionfile_fn_filter_fnbody_renumber(fnbody, _idx, ids)
  fnlabel = fnbody.at(ns(FN_CAPTIONS)) or return
  fnbody["reference"] = ids[fnbody["id"]].first["reference"]
  fnlabel.children = @isodoc_presxml.fn_body_label(fnbody)
end

#sectionfile_fn_filter_prep(xml) ⇒ Object

map fmt-fn-body/@id = fn/@target to fn



230
231
232
233
234
235
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 230

def sectionfile_fn_filter_prep(xml)
  xml.xpath(ns("//fn")).each_with_object({}) do |f, m|
    m[f["target"]] ||= []
    m[f["target"]] << f
  end
end

#sectionfile_fn_filter_renumber(fnbody, idx, ids, seen) ⇒ Object 



239
240
241
242
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 239

def sectionfile_fn_filter_renumber(fnbody, idx, ids, seen)
  sectionfile_fn_filter_fn_renumber(fnbody, idx, ids, seen)
  sectionfile_fn_filter_fnbody_renumber(fnbody, idx, ids)
end

#sectionfile_insert(ins, chunks, parentnode) ⇒ Object 



208
209
210
211
212
213
214
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 208

def sectionfile_insert(ins, chunks, parentnode)
  if parentnode
    ins.next = "<#{parentnode}/>"
    chunks.each { |c| ins.next.add_child(c.dup) }
  else chunks.each { |c| ins.next = c.dup }
  end
end

#sectionsplit ⇒ Object

Input XML is Semantic XML



57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 57

def sectionsplit
  xml = sectionsplit_prep(File.read(@input_filename), @base, @dir)
  # Retain the whole (un-split) Presentation XML. sectionsplit is an
  # HTML-only mechanism, but the collection PDF/presentation path must
  # inline each document intact; without this the sectionsplit parent
  # reaches concatenation with no presentation output and degrades to a
  # bibdata-only, namespace-less <metanorma> stub that mn2pdf drops.
  # https://github.com/metanorma/iso-10303/issues/208
  @whole_presentation_file = write_whole_presentation(xml)
  @key = Metanorma::Collection::XrefProcess::xref_preprocess(xml, @isodoc)
  empty = empty_doc(xml)
  empty1 = empty_attachments(empty)
  @mutex = Mutex.new
  # @pool = Concurrent::FixedThreadPool.new(4)
  @pool = Concurrent::FixedThreadPool.new(1)
  sectionsplit1(xml, empty, empty1, 0)
end

#sectionsplit1(xml, empty, empty1, idx) ⇒ Object

xml is Presentation XML



76
77
78
79
80
81
82
83
84
85
86
87
88
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 76

def sectionsplit1(xml, empty, empty1, idx)
  ret = SPLITSECTIONS.each_with_object([]) do |n, m|
    conflate_floatingtitles(xml.xpath(ns(n[0]))).each do |s|
      # require "debug"; binding.b
      sectionsplit2(xml, idx.zero? ? empty : empty1, s, n[1],
                    { acc: m, idx: idx })
      idx += 1
    end
  end
  @pool.shutdown
  @pool.wait_for_termination
  ret
end

#sectionsplit2(xml, empty, chunks, parentnode, opt) ⇒ Object 



90
91
92
93
94
95
96
97
98
99
100
101
102
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 90

def sectionsplit2(xml, empty, chunks, parentnode, opt)
  @pool.post do
    output_filename = Util.substitute_filename_pattern(
      @sectionsplit_filename,
      document_num: @parent_idx, basename: File.basename(@base, ".*"),
      basename_legacy: @base, sectionsplit_num: opt[:idx]
    )
    warn "Sectionsplit: #{output_filename}"
    a = sectionfile(xml, empty, output_filename, chunks,
                    parentnode)
    @mutex.synchronize { opt[:acc] << a }
  end
end

#sectionsplit_prep(file, filename, dir) ⇒ Object 



121
122
123
124
125
126
127
128
129
130
131
132
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 121

def sectionsplit_prep(file, filename, dir)
  @splitdir = dir
  xml, type = sectionsplit_preprocess_semxml(file, filename)
  flags = { format: :asciidoc, extension_keys: [:presentation],
            type: type }.merge(@compile_opts)
  Compile.new.compile(xml, flags)
  f = File.open(xml.sub(/\.xml$/, ".presentation.xml"), encoding: "utf-8")
  r = Nokogiri::XML(f, &:huge)
  f.close
  r.xpath("//xmlns:svgmap1").each { |x| x.name = "svgmap" }
  r
end

#sectionsplit_preprocess_semxml(file, filename) ⇒ Object 



144
145
146
147
148
149
150
151
152
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 144

def sectionsplit_preprocess_semxml(file, filename)
  xml = Nokogiri::XML(file, &:huge)
  type = xml.root["flavor"]
  type ||= xml.root.name.sub("-standard", "").to_sym
  sectionsplit_update_xrefs(xml)
  xml1 = sectionsplit_write_semxml(filename, xml)
  @tmp_filename = xml1
  [xml1, type]
end

#sectionsplit_update_xrefs(xml) ⇒ Object 



154
155
156
157
158
159
160
161
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 154

def sectionsplit_update_xrefs(xml)
  if c = @fileslookup&.parent
    # nested mode so unresolved erefs are not deleted
    c.with_nested { c.update_xrefs(xml, @ident, {}) }
    xml.xpath("//xmlns:svgmap").each { |x| x.name = "svgmap1" }
    # do not process svgmap until after files are split
  end
end

#sectionsplit_write_semxml(filename, xml) ⇒ Object 



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

def sectionsplit_write_semxml(filename, xml)
  outname = Pathname.new("tmp_#{filename}").sub_ext(".xml").to_s
  File.open(outname, "w:UTF-8") do |f|
    f.write(@isodoc.to_xml(xml))
  end
  outname
end

#titlerender(section) ⇒ Object 



292
293
294
295
296
297
298
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 292

def titlerender(section)
  title = section.at(ns("./fmt-title")) or return "[Untitled]"
  t = title.dup
  t.xpath(ns(".//tab | .//br")).each { |x| x.replace(" ") }
  t.xpath(ns(".//bookmark")).each(&:remove)
  t.xpath(".//text()").map(&:text).join
end

#write_whole_presentation(xml) ⇒ Object

Persist the whole, svgmap-restored Presentation XML (the in-memory doc, not the on-disk tmp which still carries the internal svgmap1 marker) so the collection PDF path can inline the document whole. Written flat to the split output directory, where the per-section files also live.



138
139
140
141
142
 
# File 'lib/metanorma/collection/sectionsplit/sectionsplit.rb', line 138

def write_whole_presentation(xml)
  fname = File.join(@splitdir, "#{@base}.whole.presentation.xml")
  File.write(fname, xml.to_xml, encoding: "UTF-8")
  fname
end