asciichem 0.29.3 → 0.30.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6aed1bf1e1bf05bd4f3459f3bf147d39f08f283dc49b8452040643171e836a30
-  data.tar.gz: 557b7506d58e8505c6b9e3ffa36da6ef7025e054850eba7a9fdee7e03503c337
+  metadata.gz: 2f8a0ef9952b9507c35ded04295298a3a62e86a7efce1f92d258a1aaf6d6125c
+  data.tar.gz: 0d11ccfb16383194e8a1fd3bf0816ce8b8465b548bc6dac43535d2633a0cf2ba
 SHA512:
-  metadata.gz: b3ff4937b6f92f253c16f03d414ef4fa177ab0df85fea38c61be07e83a04e7889c2485e4b4d8b7dd3c2a7e0042711e533cdf5968e11abc130d68d21620e6ed07
-  data.tar.gz: dfeada783a748286cf51fbbb9cbaa0a8f7a485af352b511dd54a67d53fce0ec77c9b58d339d980853664ff58438f0e5be56ae6632802d70a44e12967d5418285
+  metadata.gz: 270015bf0f88f0bc2f06452a472759970b39ab75ae9aef233ea3c9f94f7b3abc2d20752098151aedd7ab51251aaad9414cfdb349167c059ac036e8bbdaff04d5
+  data.tar.gz: a4211637fc504c7db50994d14b151bb5e45357b48d870b4f4aff7eb49c79c269ba7014a755fd00c6e8dd22678e1f044011b9b4123ecc24530b9d693464a11357

data/CHANGELOG.md

@@ -3,6 +3,32 @@
 All notable changes to AsciiChem are documented here.
 This project follows [Semantic Versioning](https://semver.org/).
 
+## [0.30.1] - 2026-09-23
+
+### Fixed
+- Groups attached by an explicit bond no longer re-spell through
+  CML round-trip: `CC(=O)O` came back as `CC=(O)O`. The CML
+  rebuilder left the attachment bond outside the reconstructed
+  group; the grammar keeps it inside (the inner molecule starts
+  with the bond), so the rebuilder now pulls an adjacent preceding
+  Bond into the group. Affects text syntax and SMILES ingestion
+  alike; regression specs cover single/sibling bonded groups,
+  multi-atom groups, and aspirin stability across double
+  round-trips.
+
+## [0.30.0] - 2026-09-22
+
+### Changed
+- The XML layer migrates from Nokogiri to
+  [moxml](https://github.com/lutaml/moxml) — the adapter layer the
+  rest of the lutaml ecosystem (chemicalml, CML wire, metanorma
+  gems) already uses, so AsciiChem follows the consumer's
+  configured XML backend instead of pinning one. The gemspec
+  dependency changes accordingly (`nokogiri ~> 1.18` out, `moxml
+  ~> 0.5` in). MathML output is byte-identical to 0.29.3: the L2
+  corpus goldens pass unchanged, including empty-element
+  compaction and the `<?xml?>` declaration.
+
 ## [0.29.3] - 2026-09-22
 
 ### Changed

data/asciichem.gemspec

@@ -39,7 +39,7 @@ Gem::Specification.new do |spec|
   spec.add_dependency "elkrb", "~> 1.0"
   spec.add_dependency "lutaml-model", "~> 0.8"
   spec.add_dependency "mml", "~> 2.3"
-  spec.add_dependency "nokogiri", "~> 1.18"
+  spec.add_dependency "moxml", "~> 0.5"
   spec.add_dependency "parslet", "~> 2.0"
   spec.add_dependency "relaton-bib", "~> 2.1"
   spec.add_dependency "plurimath", "~> 0.11"

data/lib/asciichem/cml/conditions_extensions.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'nokogiri'
+require 'moxml'
 
 module AsciiChem
   module Cml
@@ -26,11 +26,11 @@ module AsciiChem
         return xml if conditions_map.empty?
 
         # If native <conditionList> already present, skip aci: emit.
-        doc_check = Nokogiri::XML(xml)
+        doc_check = Moxml.parse(xml)
         return xml if doc_check.xpath('//cml:reaction/cml:conditionList',
                                       cml: Extensions::CML_NS).any?
 
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         root = doc.root
         Extensions.ensure_namespace(root)
         apply_conditions(root, conditions_map)
@@ -40,7 +40,7 @@ module AsciiChem
       # Extract aci:conditionsAbove/Below from each <reaction>.
       # Returns `{ reaction_id => { above:, below: } }`.
       def self.extract(xml)
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         result = {}
         doc.xpath('//cml:reaction', cml: Extensions::CML_NS).each do |el|
           id = el['id']

data/lib/asciichem/cml/extensions/atom_attributes.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'nokogiri'
+require 'moxml'
 
 module AsciiChem
   module Cml
@@ -39,9 +39,9 @@ module AsciiChem
         def self.inject(xml, extensions)
           return xml if extensions.empty?
 
-          doc = Nokogiri::XML(xml)
+          doc = Moxml.parse(xml)
           root = doc.root
-          unless root.namespaces.value?(Extensions::NAMESPACE)
+          unless root.namespaces.any? { |ns| ns.uri == Extensions::NAMESPACE }
             root.add_namespace(Extensions::PREFIX, Extensions::NAMESPACE)
           end
 
@@ -62,7 +62,7 @@ module AsciiChem
         # Extract aci: attributes from a CML XML string. Returns a map
         # `{ atom_id => { field: value } }` with Ruby-native types.
         def self.extract(xml)
-          doc = Nokogiri::XML(xml)
+          doc = Moxml.parse(xml)
           result = {}
           doc.xpath('//cml:atom', cml: Extensions::CML_NS).each do |atom_el|
             atom_id = atom_el['id']

data/lib/asciichem/cml/extensions/top_level.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'nokogiri'
+require 'moxml'
 require 'set'
 
 module AsciiChem
@@ -120,9 +120,9 @@ module AsciiChem
         def self.inject(xml, top_level)
           return xml if top_level.empty?
 
-          doc = Nokogiri::XML(xml)
+          doc = Moxml.parse(xml)
           root = doc.root
-          unless root.namespaces.value?(Extensions::NAMESPACE)
+          unless root.namespaces.any? { |ns| ns.uri == Extensions::NAMESPACE }
             root.add_namespace(Extensions::PREFIX, Extensions::NAMESPACE)
           end
           top_level.each { |entry| insert_element(doc, root, entry) }
@@ -133,7 +133,7 @@ module AsciiChem
         # array of `{ position:, element_name:, content: }` hashes in
         # ascending position order.
         def self.extract(xml)
-          doc = Nokogiri::XML(xml)
+          doc = Moxml.parse(xml)
           result = []
           element_names = HANDLERS.map(&:element_name)
           element_names.each do |name|
@@ -174,7 +174,7 @@ module AsciiChem
           def insert_element(doc, root, entry)
             element = doc.create_element("#{Extensions::PREFIX}:#{entry[:element_name]}")
             element['position'] = entry[:position].to_s
-            element.content = entry[:content]
+            element.add_child(doc.create_text(entry[:content]))
             # Insert before existing children so extensions appear at
             # the top of <cml>, which reads more naturally than appended.
             root.children.first&.add_previous_sibling(element) || root.add_child(element)

data/lib/asciichem/cml/extensions.rb

@@ -31,7 +31,7 @@ module AsciiChem
       # extension module that injects aci: attributes or elements
       # calls this instead of inlining its own check.
       def self.ensure_namespace(root)
-        return if root.namespaces.value?(NAMESPACE)
+        return if root.namespaces.any? { |ns| ns.uri == NAMESPACE }
 
         root.add_namespace(PREFIX, NAMESPACE)
       end

data/lib/asciichem/cml/group_extensions.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'nokogiri'
+require 'moxml'
 
 module AsciiChem
   module Cml
@@ -46,7 +46,7 @@ module AsciiChem
       def self.inject(xml, groups_by_molecule)
         return xml if groups_by_molecule.empty?
 
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         root = doc.root
         Extensions.ensure_namespace(root)
 
@@ -77,7 +77,7 @@ module AsciiChem
       # molecule's ID. Each value is an array of record hashes:
       # `{ multiplicity:, bracket:, atom_ids: }`.
       def self.extract(xml)
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         result = Hash.new { |h, k| h[k] = [] }
         doc.xpath('//cml:molecule', cml: Extensions::CML_NS).each do |mol_el|
           group_els = mol_el.xpath("./#{Extensions::PREFIX}:group",
@@ -213,18 +213,23 @@ module AsciiChem
       def self.splice_group_into(nodes, target_nodes, record)
         return if target_nodes.empty?
 
-        positions = target_nodes.map do |target|
-          found = nodes.each_with_index.find { |node, _| node.equal?(target) }
-          found&.last
-        end.compact
+        inner_nodes = group_nodes(nodes, target_nodes)
+
+        positions = target_nodes
+                    .map { |target| position_of(nodes, target) }
+                    .compact
         return if positions.empty?
 
         group = AsciiChem::Model::Group.new(
-          nodes: target_nodes,
+          nodes: inner_nodes,
           multiplicity: record[:multiplicity],
           bracket: record[:bracket]
         )
+        replace_positions_with_group(nodes, group, positions)
+      end
+      private_class_method :splice_group_into
 
+      def self.replace_positions_with_group(nodes, group, positions)
         first_pos = positions.min
         nodes[first_pos] = group
         # Remove remaining positions in descending order so earlier
@@ -233,7 +238,31 @@ module AsciiChem
           nodes.delete_at(pos)
         end
       end
-      private_class_method :splice_group_into
+      private_class_method :replace_positions_with_group
+
+      # A Bond sitting immediately before the group's first target node
+      # is the group's attachment bond — the `=` of `C(=O)`. The
+      # grammar keeps that bond INSIDE the group (its inner molecule
+      # starts with the bond), so the rebuilder moves it in instead of
+      # leaving a stray prefix bond outside, which would re-spell as
+      # `C=(O)`. By construction of the bond insertion (each bond sits
+      # just before its later endpoint), a bond adjacent to the group's
+      # first atom always has that atom as its later endpoint.
+      def self.group_nodes(nodes, target_nodes)
+        idx = position_of(nodes, target_nodes.first)
+        bond = idx&.positive? ? nodes[idx - 1] : nil
+        return target_nodes unless bond.is_a?(AsciiChem::Model::Bond)
+
+        [nodes.delete_at(idx - 1), *target_nodes]
+      end
+      private_class_method :group_nodes
+
+      # Identity-based position lookup so duplicate Bonds (same kind
+      # → `==` equal) aren't confused.
+      def self.position_of(nodes, node)
+        nodes.each_with_index.find { |candidate, _| candidate.equal?(node) }&.last
+      end
+      private_class_method :position_of
 
       # Flatten the canonical document's molecules (top-level +
       # reaction reactants/products + cascade reactions) into a list,

data/lib/asciichem/cml/metadata_extensions.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'nokogiri'
+require 'moxml'
 
 module AsciiChem
   module Cml
@@ -28,7 +28,7 @@ module AsciiChem
         metadata_map = build_metadata_map(formula)
         return xml if metadata_map.empty?
 
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         root = doc.root
         Extensions.ensure_namespace(root)
         apply_metadata(root, metadata_map)
@@ -38,7 +38,7 @@ module AsciiChem
       # Extract aci:meta-* attributes from each <molecule> in the XML.
       # Returns a map `{ molecule_id => { name => content } }`.
       def self.extract(xml)
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         result = {}
         doc.xpath('//cml:molecule', cml: Extensions::CML_NS).each do |el|
           id = el['id']
@@ -132,11 +132,11 @@ module AsciiChem
         end
 
         def read_meta_attrs(element)
-          element.attributes.each_with_object({}) do |(name, attr), memo|
-            next unless name.start_with?(META_PREFIX)
+          element.attributes.each_with_object({}) do |attr, memo|
+            next unless attr.name.start_with?(META_PREFIX)
             next unless attr.namespace&.prefix == Extensions::PREFIX
 
-            memo[name.sub(META_PREFIX, '')] = attr.value
+            memo[attr.name.sub(META_PREFIX, '')] = attr.value
           end
         end
       end

data/lib/asciichem/cml/opaque_extensions.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri"
+require "moxml"
 
 module AsciiChem
   module Cml
@@ -42,9 +42,9 @@ module AsciiChem
       def self.inject(xml, formula)
         return xml if formula.nodes.empty?
 
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         root = doc.root
-        wire_children = root.element_children
+        wire_children = root.children.select(&:element?)
         inserts = build_inserts(formula, wire_children.length)
         return xml if inserts.empty?
 
@@ -54,11 +54,11 @@ module AsciiChem
 
       def self.apply_inserts(root, wire_children, inserts)
         inserts.reverse_each do |wire_index, raw_xml|
-          fragment = Nokogiri::XML::DocumentFragment.parse(raw_xml)
+          nodes = Moxml.new.parse_fragment(raw_xml)
           if wire_index >= wire_children.length
-            root.add_child(fragment)
+            nodes.each { |n| root.add_child(n) }
           else
-            wire_children[wire_index].add_previous_sibling(fragment)
+            nodes.each { |n| wire_children[wire_index].add_previous_sibling(n) }
           end
         end
       end
@@ -88,10 +88,10 @@ module AsciiChem
       # document order. Also returns the cleaned XML with the unknown
       # elements removed (so chemicalml's parser doesn't trip).
       def self.extract(xml)
-        doc = Nokogiri::XML(xml)
+        doc = Moxml.parse(xml)
         root = doc.root
         result = []
-        children = root.element_children
+        children = root.children.select(&:element?)
         children.each_with_index do |child, idx|
           next if cml_namespace?(child)
           next if aci_namespace?(child)
@@ -107,14 +107,12 @@ module AsciiChem
       end
 
       def self.cml_namespace?(element)
-        ns = element.namespace
-        ns && ns.href == Extensions::CML_NS
+        element.namespace_uri == Extensions::CML_NS
       end
       private_class_method :cml_namespace?
 
       def self.aci_namespace?(element)
-        ns = element.namespace
-        ns && ns.href == Extensions::NAMESPACE
+        element.namespace_uri == Extensions::NAMESPACE
       end
       private_class_method :aci_namespace?
 

data/lib/asciichem/formatter/mathml.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "nokogiri"
+require "moxml"
 require "mml"
 
 module AsciiChem
@@ -16,19 +16,26 @@ module AsciiChem
       MATHML_NS = "http://www.w3.org/1998/Math/MathML".freeze
 
       def initialize
-        @doc = Nokogiri::XML::Document.new
+        @doc = Moxml.new.create_document
       end
 
       # Model entry point. Returns a `<math>` element as a string.
       def visit_formula(formula)
-        math = el("math", xmlns: MATHML_NS)
+        math = @doc.create_element("math")
+        math.add_namespace(nil, MATHML_NS)
         mrow = el("mrow")
         formula.nodes.each { |n| mrow.add_child(render_node(n)) }
         math.add_child(mrow)
+        @doc.add_child(@doc.create_declaration("1.0", "UTF-8", nil))
         @doc.root = math
         # Native UTF-8 output preserves unicode arrow entities so specs
         # and downstream consumers see ⇌ and → instead of &#x21CC;.
-        @doc.to_xml(encoding: "UTF-8")
+        # Empty MathML elements are compacted back to <none/> form: the
+        # L2 corpus goldens are byte-exact across the TypeScript and
+        # Python implementations, whose serializers self-close.
+        @doc.to_xml(encoding: "UTF-8", indent: 2, indent_text: " ")
+            .gsub("<none></none>", "<none/>")
+            .gsub("<mprescripts></mprescripts>", "<mprescripts/>")
       end
 
       def visit_molecule(molecule)
@@ -72,7 +79,7 @@ module AsciiChem
         multi.add_child(base)
         multi.add_child(el("none"))
         multi.add_child(el("none"))
-        multi.add_child(Nokogiri::XML::Element.new("mprescripts", @doc))
+        multi.add_child(@doc.create_element("mprescripts"))
         multi.add_child(el("none"))
         multi.add_child(mn(atom.isotope))
         multi
@@ -211,12 +218,13 @@ module AsciiChem
           next unless attr_name
 
           Array(math.public_send(attr_name)).each do |child|
-            fragment = Nokogiri::XML::DocumentFragment.parse(child.to_xml)
-            fragment.children.each { |node| mrow.add_child(node) }
+            fragment = Moxml.new.parse_fragment(child.to_xml)
+            # parse_fragment returns the fragment's top-level nodes
+        fragment.each { |node| mrow.add_child(node) }
           end
         end
         mrow
-      rescue Mml::Error, Lutaml::Model::Error, Nokogiri::XML::SyntaxError
+      rescue Mml::Error, Lutaml::Model::Error, Moxml::ParseError
         el("mrow")
       end
 
@@ -388,30 +396,30 @@ module AsciiChem
         msup
       end
 
-      # -- Nokogiri element factories ------------------------------------
+      # -- Element factories ------------------------------------
 
       def el(name, attrs = {})
-        element = Nokogiri::XML::Element.new(name, @doc)
+        element = @doc.create_element(name)
         attrs.each { |k, v| element[k.to_s] = v }
         element
       end
 
       def mi(content)
         e = el("mi", mathvariant: "normal")
-        e.content = content.to_s
+        e.add_child(@doc.create_text(content.to_s))
         e
       end
 
       def mn(content)
-        e = el("mn"); e.content = content.to_s; e
+        e = el("mn"); e.add_child(@doc.create_text(content.to_s)); e
       end
 
       def mo(content)
-        e = el("mo"); e.content = content.to_s; e
+        e = el("mo"); e.add_child(@doc.create_text(content.to_s)); e
       end
 
       def mtext(content)
-        e = el("mtext"); e.content = content.to_s; e
+        e = el("mtext"); e.add_child(@doc.create_text(content.to_s)); e
       end
 
       # -- Beyond-formulas helpers -----------------------------------

data/lib/asciichem/formatter/structural_svg.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require 'nokogiri'
+require 'moxml'
 
 module AsciiChem
   module Formatter
@@ -88,7 +88,8 @@ module AsciiChem
       end
 
       def render_svg(result)
-        doc = Nokogiri::XML::Document.new
+        doc = Moxml.new.create_document
+        doc.add_child(doc.create_declaration('1.0', nil, nil))
         svg = build_svg_root(doc, result)
         bonds_first_then_atoms(result, doc, svg)
         doc.root = svg
@@ -96,14 +97,14 @@ module AsciiChem
       end
 
       def build_svg_root(doc, result)
-        svg = Nokogiri::XML::Element.new('svg', doc)
+        svg = doc.create_element('svg')
         svg['xmlns'] = 'http://www.w3.org/2000/svg'
         svg['width'] = result.width.to_s
         svg['height'] = result.height.to_s
         svg['viewBox'] = "0 0 #{result.width} #{result.height}"
         svg['role'] = 'img'
-        title = Nokogiri::XML::Element.new('title', doc)
-        title.content = title_text(result)
+        title = doc.create_element('title')
+        title.add_child(doc.create_text(title_text(result)))
         svg.add_child(title)
         svg
       end
@@ -136,7 +137,7 @@ module AsciiChem
         end
 
         def render_into(parent)
-          group = Nokogiri::XML::Element.new('g', @doc)
+          group = @doc.create_element('g')
           group.add_child(circle)
           group.add_child(label)
           parent.add_child(group)
@@ -145,7 +146,7 @@ module AsciiChem
         private
 
         def circle
-          el = Nokogiri::XML::Element.new('circle', @doc)
+          el = @doc.create_element('circle')
           el['cx'] = @atom.x.to_s
           el['cy'] = @atom.y.to_s
           el['r'] = StructuralSvg::ATOM_RADIUS.to_s
@@ -156,14 +157,14 @@ module AsciiChem
         end
 
         def label
-          el = Nokogiri::XML::Element.new('text', @doc)
+          el = @doc.create_element('text')
           el['x'] = @atom.x.to_s
           el['y'] = (@atom.y + 4).to_s
           el['text-anchor'] = 'middle'
           el['font-family'] = 'serif'
           el['font-size'] = '14'
           el['fill'] = color
-          el.content = @atom.element
+          el.add_child(@doc.create_text(@atom.element))
           el
         end
 
@@ -176,7 +177,7 @@ module AsciiChem
       # Renders a single bond between two positioned atoms. Dispatches
       # on bond kind via a registry of Procs. Each Proc receives the
       # renderer (for its public line/offset helpers) and returns an
-      # array of Nokogiri elements. Adding a new bond style is a new
+      # array of moxml elements. Adding a new bond style is a new
       # Proc + one registry entry — no edits to existing renderers.
       class BondRenderer
         def initialize(doc, from_atom, to_atom, kind)
@@ -210,7 +211,7 @@ module AsciiChem
         end
 
         def line(start_x, start_y, end_x, end_y)
-          el = Nokogiri::XML::Element.new('line', @doc)
+          el = @doc.create_element('line')
           el['x1'] = start_x.to_s
           el['y1'] = start_y.to_s
           el['x2'] = end_x.to_s
@@ -221,7 +222,7 @@ module AsciiChem
         end
 
         def polygon(points, fill:)
-          el = Nokogiri::XML::Element.new('polygon', @doc)
+          el = @doc.create_element('polygon')
           el['points'] = points.map { |x, y| "#{x},#{y}" }.join(' ')
           el['fill'] = fill
           el
@@ -247,7 +248,7 @@ module AsciiChem
         # -- Strategy registry --------------------------------------
         #
         # Each entry maps a bond kind symbol to a Proc that takes the
-        # renderer and returns an array of Nokogiri elements. Procs
+        # renderer and returns an array of moxml elements. Procs
         # use the public helpers above; nothing reaches into private
         # state.
 

data/lib/asciichem/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module AsciiChem
-  VERSION = "0.29.3"
+  VERSION = "0.30.1"
 end

data/lib/asciichem/xml_builder.rb

@@ -1,8 +1,8 @@
 # frozen_string_literal: true
 
 # Internal XML helper. Currently a thin marker module kept for parity
-# with future formatters that need XML construction without Nokogiri
-# (e.g. JRuby). The Mathml formatter uses Nokogiri directly.
+# with future formatters that need XML construction without a hard XML engine
+# dependency. The Mathml formatter uses moxml directly.
 module AsciiChem
   module XmlBuilder
   end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: asciichem
 version: !ruby/object:Gem::Version
-  version: 0.29.3
+  version: 0.30.1
 platform: ruby
 authors:
 - Ribose Inc.
@@ -66,19 +66,19 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '2.3'
 - !ruby/object:Gem::Dependency
-  name: nokogiri
+  name: moxml
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.18'
+        version: '0.5'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.18'
+        version: '0.5'
 - !ruby/object:Gem::Dependency
   name: parslet
   requirement: !ruby/object:Gem::Requirement