metanorma-standoc 3.1.6 → 3.1.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9bc2bfaaaf9871ba417d1d455867a0d67b6df3f219886aa290a6195583eacafc
-  data.tar.gz: 3f2a6511903492d4853334fc0780143a7b01d13e8d8f6d27630631208fc10b5d
+  metadata.gz: 154ca8c34101aef7d0dfb0cad7c1bf0cfafaf3df110a0fef56a7d524b17a7d5b
+  data.tar.gz: 7904a365e5bbaf50e159e164c5fc11e74d356f48a50de19889433f015eb9040d
 SHA512:
-  metadata.gz: 3485d18dfbe8b3474770f3e97d80652225539cca579b06cf1f0ca9a2ebe414c8ba1663a9b099eabecdea30cd3c630c2bebc54dfc1eb7dd954ffe89155d9d1da4
-  data.tar.gz: 8953152a143fbc28f9117e1f593b47e26f06b54eebd6498b71b98c5b5ea91a2240e41fae83af96d5f0a1ae618ebf0cb0d55b6f787db68d4ceb3509937493413c
+  metadata.gz: 5d50a9d2df00a760fe948b97d28f50652c09fe313942a48d2648018a68edf4e783d62288d965e9a34225b4804bdae4bf483f4c47317fcb2188590ecf4b4495e5
+  data.tar.gz: 616012d1b9ceaafbfd255aed0d477e66e201affece833b21413529c7ee2172c708400bc63422f3b29d460ca1d5b787a420f111b7cbfa7ec11deb46f103da893c

data/lib/metanorma/standoc/base.rb

@@ -9,6 +9,7 @@ require "metanorma-utils"
 require_relative "render"
 require_relative "localbib"
 require_relative "init"
+require_relative "isolated_converter"
 require "mn-requirements"
 
 module Asciidoctor
@@ -20,6 +21,8 @@ end
 module Metanorma
   module Standoc
     module Base
+      include IsolatedConverter
+      
       # XML_ROOT_TAG = "standard-document".freeze
       # XML_NAMESPACE = "https://www.metanorma.org/ns/standoc".freeze
       FONTS_MANIFEST = "fonts-manifest".freeze
@@ -107,10 +110,14 @@ module Metanorma
         result = makexml1(node)
         ret1 = cleanup(Nokogiri::XML(insert_xml_cr(result)))
         ret1.root.add_namespace(nil, xml_namespace)
-        validate(ret1) unless @novalid
+        validate(ret1) unless @novalid || in_isolated_conversion?
         ret1
       end
 
+      def in_isolated_conversion?
+        !@isolated_conversion_stack.empty?
+      end
+
       def draft?
         @draft
       end

data/lib/metanorma/standoc/cleanup.rb

@@ -134,7 +134,7 @@ module Metanorma
            verbal-definition non-verbal-representation}.freeze
 
       # it seems Nokogiri::XML is treating the content of <script> as cdata,
-      # because of its use in HTML. Bad nokogiri. Undoing that, since we use
+      # because of its use in HTML. Bad Nokogiri. Undoing that, since we use
       # script as a normal tag
       def script_cleanup(xmldoc)
         xmldoc.xpath("//script").each { |x| x.content = x.to_str }

data/lib/metanorma/standoc/cleanup_attachment.rb

@@ -0,0 +1,71 @@
+module Metanorma
+  module Standoc
+    module Cleanup
+      def attachment_cleanup(xmldoc)
+        xmldoc.xpath("//bibitem[uri/@type = 'attachment']").each do |b|
+          b["hidden"] = "true"
+          b.at("./docidentifier[@type = 'title']")&.remove
+          u = b.at("./uri[@type = 'attachment']")
+          c = b.at("./uri[@type = 'citation']") ||
+            u.after("<uri type='citation'/>")
+          uri = attachment_uri(u.text, b)
+          u.children = uri
+          c.children = uri
+        end
+      end
+
+      def attachment_uri(path, bib)
+        init_attachments
+        path = File.join(@localdir, path)
+        valid_attachment?(path, bib) or return ""
+        @datauriattachment or return attachment_location(path)
+        save_attachment(path, bib)
+      end
+
+      def save_attachment(path, bib)
+        init_attachments
+        f = File.basename(path)
+        File.exist?(File.join(@attachmentsdir, f)) and
+          f += "_#{UUIDTools::UUID.random_create}"
+        out_fld = File.join(@attachmentsdir, f)
+        FileUtils.cp(path, out_fld)
+        datauri_attachment(out_fld, bib.document)
+      end
+
+      def attachment_location(path)
+        f = path
+        @datauriattachment and
+          f = File.join(@attachmentsdir, File.basename(path))
+        Pathname.new(File.expand_path(f))
+          .relative_path_from(Pathname.new(File.expand_path(@output_dir))).to_s
+      end
+
+      def datauri_attachment(path, doc)
+        @datauriattachment or return
+        m = add_misc_container(doc)
+        f = attachment_location(path)
+        e = (m << "<attachment name='#{f}'/>").last_element_child
+        Vectory::Utils::datauri(path, @output_dir).scan(/.{1,60}/)
+          .each { |dd| e << "#{dd}\n" }
+        f
+      end
+
+      def valid_attachment?(path, bib)
+        File.exist?(path) and return true
+        p = Pathname.new(path).cleanpath
+        @log.add("Bibliography", bib, "Attachment #{p} does not exist",
+                 severity: 0)
+        false
+      end
+
+      def init_attachments
+        @datauriattachment or return
+        @attachmentsdir and return
+        @attachmentsfld = "_#{@filename}_attachments"
+        @attachmentsdir = File.join(@output_dir, @attachmentsfld)
+        FileUtils.rm_rf(@attachmentsdir)
+        FileUtils.mkdir_p(@attachmentsdir)
+      end
+    end
+  end
+end

data/lib/metanorma/standoc/cleanup_bibdata.rb

@@ -119,9 +119,9 @@ module Metanorma
       end
 
       def hdr2bibitem(hdr)
-        xml = Asciidoctor
-          .convert(hdr[:text], backend: hdr2bibitem_type(hdr),
-                               header_footer: true)
+        xml = isolated_asciidoctor_convert(hdr[:text], 
+                                         backend: hdr2bibitem_type(hdr),
+                                         header_footer: true)
         b = Nokogiri::XML(xml).at("//xmlns:bibdata")
         b.name = "bibitem"
         b.delete("type")

data/lib/metanorma/standoc/cleanup_bibitem.rb

@@ -1,3 +1,5 @@
+require_relative "cleanup_attachment"
+
 module Metanorma
   module Standoc
     module Cleanup
@@ -145,71 +147,6 @@ module Metanorma
         end
       end
 
-      def attachment_cleanup(xmldoc)
-        xmldoc.xpath("//bibitem[uri/@type = 'attachment']").each do |b|
-          b["hidden"] = "true"
-          u = b.at("./uri[@type = 'attachment']")
-          c = b.at("./uri[@type = 'citation']") ||
-            u.after("<uri type='citation'/>")
-          uri = attachment_uri(u.text, b)
-          u.children = uri
-          c.children = uri
-        end
-      end
-
-      def attachment_uri(path, bib)
-        init_attachments
-        path = File.join(@localdir, path)
-        valid_attachment?(path, bib) or return ""
-        @datauriattachment or return attachment_location(path)
-        save_attachment(path, bib)
-      end
-
-      def save_attachment(path, bib)
-        init_attachments
-        f = File.basename(path)
-        File.exist?(File.join(@attachmentsdir, f)) and
-          f += "_#{UUIDTools::UUID.random_create}"
-        out_fld = File.join(@attachmentsdir, f)
-        FileUtils.cp(path, out_fld)
-        datauri_attachment(out_fld, bib.document)
-      end
-
-      def attachment_location(path)
-        f = path
-        @datauriattachment and
-          f = File.join(@attachmentsdir, File.basename(path))
-        Pathname.new(File.expand_path(f))
-          .relative_path_from(Pathname.new(File.expand_path(@output_dir))).to_s
-      end
-
-      def datauri_attachment(path, doc)
-        @datauriattachment or return
-        m = add_misc_container(doc)
-        f = attachment_location(path)
-        e = (m << "<attachment name='#{f}'/>").last_element_child
-        Vectory::Utils::datauri(path, @output_dir).scan(/.{1,60}/)
-          .each { |dd| e << "#{dd}\n" }
-        f
-      end
-
-      def valid_attachment?(path, bib)
-        File.exist?(path) and return true
-        p = Pathname.new(path).cleanpath
-        @log.add("Bibliography", bib, "Attachment #{p} does not exist",
-                 severity: 0)
-        false
-      end
-
-      def init_attachments
-        @datauriattachment or return
-        @attachmentsdir and return
-        @attachmentsfld = "_#{@filename}_attachments"
-        @attachmentsdir = File.join(@output_dir, @attachmentsfld)
-        FileUtils.rm_rf(@attachmentsdir)
-        FileUtils.mkdir_p(@attachmentsdir)
-      end
-
       # remove dupes if both same ID and same docid, in case dupes introduced
       # through termbases
       def remove_dup_bibtem_id(xmldoc)
@@ -229,12 +166,42 @@ module Metanorma
         end
       end
 
+      def remove_empty_docid(xmldoc)
+        xmldoc.xpath("//bibitem/docidentifier[normalize-space(.)='']")
+          .each(&:remove)
+      end
+
+      def empty_docid_to_title(xmldoc)
+        xmldoc.xpath("//references/bibitem").each do |b|
+          b.at("./docidentifier[not(@type = 'metanorma' or @type = 'DOI' or " \
+           "@type = 'metanorma-ordinal')]") and next
+          empty_docid_to_title?(b) or next
+          ins = b.at("./title[last()]") || b.at("./formattedref")
+          id = bibitem_title_to_id(b) or return
+          ins.next = <<~XML
+            <docidentifier type='title' primary='true'>#{id}</docidentifier>
+          XML
+        end
+      end
+
+      def bibitem_title_to_id(bibitem)
+        t = bibitem.at("./title") || bibitem.at("./formattedref") or return
+        t.text
+      end
+
+      # normative references only, biblio uses ordinal code instead
+      def empty_docid_to_title?(bibitem)
+        bibitem.parent["normative"] == "true"
+      end
+
       def bibitem_cleanup(xmldoc)
         bibitem_nested_id(xmldoc) # feeds remove_dup_bibtem_id
-        remove_dup_bibtem_id(xmldoc)
         ref_dl_cleanup(xmldoc)
         formattedref_spans(xmldoc)
         fetch_local_bibitem(xmldoc)
+        remove_empty_docid(xmldoc)
+        empty_docid_to_title(xmldoc)
+        remove_dup_bibtem_id(xmldoc)
         attachment_cleanup(xmldoc)
       end
     end

data/lib/metanorma/standoc/cleanup_block.rb

@@ -163,7 +163,7 @@ module Metanorma
           .each_slice(4).map.with_object([]) do |a, acc|
           acc << safe_noko(a[0], node.document)
           a.size == 4 or next
-          acc << Asciidoctor.convert(
+          acc << isolated_asciidoctor_convert(
             a[2], doctype: :inline, backend: backend&.to_sym || :standoc
           )
         end.join

data/lib/metanorma/standoc/front_contributor.rb

@@ -7,7 +7,9 @@ module Metanorma
       def metadata_author(node, xml)
         org_author(node, xml)
         personal_author(node, xml)
-        committee_contributors(node, xml, default_publisher, {})
+        corporate_author = node.attr("corporate-author") ||
+          node.attr("publisher") || default_publisher
+        committee_contributors(node, xml, corporate_author, {})
       end
 
       def org_author(node, xml)

data/lib/metanorma/standoc/init.rb

@@ -49,6 +49,7 @@ module Metanorma
         @localdir = Metanorma::Utils::localdir(node)
         @xrefstyle = node.attr("xrefstyle")
         @novalid = node.attr("novalid")
+        @isolated_conversion_stack = []
         @smartquotes = node.attr("smartquotes") != "false"
         @sourcecode_markup_start = node.attr("sourcecode-markup-start") || "{{{"
         @sourcecode_markup_end = node.attr("sourcecode-markup-end") || "}}}"

data/lib/metanorma/standoc/isolated_converter.rb

@@ -0,0 +1,65 @@
+module Metanorma
+  module Standoc
+    module IsolatedConverter
+      # Create an isolated Asciidoctor conversion that doesn't interfere with 
+      # the current converter's instance variables
+      def isolated_asciidoctor_convert(content, options = {})
+        # Track that we're in an isolated conversion (for nested calls)
+        @isolated_conversion_stack << true
+        
+        begin
+          # Ensure we get a completely fresh Document and conversion context
+          # Each call to Asciidoctor.convert creates a new Document with its own converter
+          # This naturally isolates the conversion from the current instance variables
+          
+          # Save critical options that should be preserved from the current context
+          preserved_options = extract_preserved_options(options)
+          
+          # Merge with isolated options to ensure clean state and skip validation
+          isolated_options = preserved_options.merge(options).merge(
+            attributes: (preserved_options[:attributes] || {}).merge(
+              'novalid' => ''  # Force no validation for isolated documents
+            )
+          )
+          
+          # Perform the isolated conversion
+          Asciidoctor.convert(content, isolated_options)
+        ensure
+          # Always pop from stack, even if conversion fails
+          @isolated_conversion_stack.pop
+        end
+      end
+
+      private
+
+      # Extract options that should be preserved from the current conversion context
+      def extract_preserved_options(user_options)
+        options = {}
+        
+        # Preserve safe mode to maintain security context
+        options[:safe] = user_options[:safe] if user_options.key?(:safe)
+        
+        # Preserve local directory context if not explicitly overridden
+        options[:base_dir] = @localdir if @localdir && !user_options.key?(:base_dir)
+        
+        # Preserve attributes that are safe to share
+        if user_options[:attributes].nil? && respond_to?(:safe_shared_attributes)
+          options[:attributes] = safe_shared_attributes
+        end
+        
+        options
+      end
+
+      # Define attributes that are safe to share between converter instances
+      def safe_shared_attributes
+        # Only include read-only or configuration attributes
+        # Avoid any attributes that could cause state pollution
+        {
+          'source-highlighter' => 'html-pipeline', # Use simple highlighter
+          'nofooter' => '',
+          'no-header-footer' => ''
+        }
+      end
+    end
+  end
+end

data/lib/metanorma/standoc/regex.rb

@@ -71,11 +71,11 @@ module Metanorma
 
       # These regexes allow () inside usrlbl but not inside code
       NON_ISO_REF = %r{^<ref\sid="(?<anchor>[^"]+)">
-      \[(?<usrlbl>\(.+\))?(?<code>.+?)\]</ref>,?\s*
+      \[(?<usrlbl>\(.+\))?(?<code>.+)\]</ref>,?\s*
       (?:<fn[^>]*>\s*<p>(?<fn>[^\]]+)</p>\s*</fn>)?(?<text>.*)$}xm
 
       NON_ISO_REF1 = %r{^<ref\sid="(?<anchor>[^"]+)">
-      (?<usrlbl>\(.+\))?(?<code>.+?)</ref>,?\s*
+      (?<usrlbl>\(.+\))?(?<code>.+)</ref>,?\s*
       (?:<fn[^>]*>\s*<p>(?<fn>[^\]]+)</p>\s*</fn>\s*)?(?<text>.*)$}xm
     end
   end

data/lib/metanorma/standoc/spans_to_bibitem.rb

@@ -16,11 +16,19 @@ module Metanorma
           @spans[:docid] = override_docids(ids[:docid], @spans[:docid])
         end
 
+        # override old values with new values if type is the same
+        # comparison is case-insensitive
+        # if types differ in case, use the old value's type, not the new
         def override_docids(old, new)
           ret = new
-          keys = new.map { |a| a[:type] }
+          keys = new.map { |a| a[:type]&.upcase }
           old.each do |e|
-            keys.include?(e[:type]) or ret << e
+            if keys.include?(e[:type]&.upcase)
+              ret.each do |a|
+                a[:type]&.upcase == e[:type]&.upcase and a[:type] = e[:type]
+              end
+            else ret << e
+            end
           end
           ret
         end

data/lib/metanorma/standoc/spans_to_bibitem_preprocessing.rb

@@ -119,7 +119,7 @@ module Metanorma
         end
 
         def spans_preprocess_new_contrib?(span, contrib)
-          contrib.empty? ||
+          contrib.empty? || contrib[-1][:entity] == "organization" ||
             (span[:key] == "surname" && contrib[-1][:surname]) ||
             contrib[-1][:role] != (span[:type] || "author")
         end

data/lib/metanorma/standoc/utils.rb

@@ -155,7 +155,7 @@ module Metanorma
 
           #{text}
         ADOC
-        c = Asciidoctor.convert(doc, backend: flavour, header_footer: true)
+        c = isolated_asciidoctor_convert(doc, backend: flavour, header_footer: true)
         ret = Nokogiri::XML(c).at("//xmlns:sections")
         separate_numbering_footnotes(ret)
       end

data/lib/metanorma/standoc/version.rb

@@ -19,6 +19,6 @@ module Metanorma
   end
 
   module Standoc
-    VERSION = "3.1.6".freeze
+    VERSION = "3.1.7".freeze
   end
 end

data/metanorma-standoc.gemspec

@@ -32,7 +32,7 @@ Gem::Specification.new do |spec|
   spec.add_dependency "asciidoctor", "~> 2.0.0"
   spec.add_dependency "crass", "~> 1.0.0"
   spec.add_dependency "iev", "~> 0.3.5"
-  spec.add_dependency "isodoc", "~> 3.2.0"
+  spec.add_dependency "isodoc", "~> 3.3.0"
   spec.add_dependency "metanorma", ">= 1.6.0"
   spec.add_dependency "metanorma-plugin-glossarist", "~> 0.2.3"
   spec.add_dependency "metanorma-plugin-lutaml", "~> 0.7.31"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: metanorma-standoc
 version: !ruby/object:Gem::Version
-  version: 3.1.6
+  version: 3.1.7
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-09-15 00:00:00.000000000 Z
+date: 2025-09-29 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: addressable
@@ -72,14 +72,14 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 3.2.0
+        version: 3.3.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: 3.2.0
+        version: 3.3.0
 - !ruby/object:Gem::Dependency
   name: metanorma
   requirement: !ruby/object:Gem::Requirement
@@ -501,6 +501,7 @@ files:
 - lib/metanorma/standoc/cleanup.rb
 - lib/metanorma/standoc/cleanup_amend.rb
 - lib/metanorma/standoc/cleanup_asciibib.rb
+- lib/metanorma/standoc/cleanup_attachment.rb
 - lib/metanorma/standoc/cleanup_bibdata.rb
 - lib/metanorma/standoc/cleanup_bibitem.rb
 - lib/metanorma/standoc/cleanup_block.rb
@@ -534,6 +535,7 @@ files:
 - lib/metanorma/standoc/inline.rb
 - lib/metanorma/standoc/isodoc-compile.rng
 - lib/metanorma/standoc/isodoc.rng
+- lib/metanorma/standoc/isolated_converter.rb
 - lib/metanorma/standoc/lists.rb
 - lib/metanorma/standoc/localbib.rb
 - lib/metanorma/standoc/macros.rb