asciichem 0.3.2 → 0.3.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: dc89810e56d2b42835eddf9bdc0609e4520a4deccbc9451d6fab66ffc2f441ae
-  data.tar.gz: 39e94b4f58d2e61c44cbc377bea8fb8eb60bd8f5f0123053126fc634b94a8b5f
+  metadata.gz: 82e379bdd585e67154ee2ac11685d34d343fc7603dad41c8e629d27610f66a0e
+  data.tar.gz: 443e0c865b127e706b61ec9d8781b3516aecb8dfb86a83630e1363af5ae52779
 SHA512:
-  metadata.gz: 7f765d66cad35e53f0e069bd1244cbc4bf1f4ba9ca029318a88c6276b28c91ca06f68e50d721a4f61fbbf494c496fbccf36f52f41d13b814c50501b86b8d23cb
-  data.tar.gz: 7762771fb9322427440533bbac79991ffa203f220456c30c8d9cbe5be5b9129928ce7905684090a07172d0d9273127c30d23183d20d5e0e31a376d04e36fbbac
+  metadata.gz: 1dca3abf1149dbbeb1166053cdc6afd58e90c4ad1635c5d13a8884aa11e00b1f43dbdadda6a860d5110dc3924f1ce8f01cba41992faaabbe5bd8a5b05382cdf6
+  data.tar.gz: 2fb960db615687093e7cb30f70752303d1b143fc9c3af8cd582d0fef7afcff00e45feabe88769bbb8f6af6431171b3b431d2689339a1d52903f21f397acc23a3

data/lib/asciichem/cml/translator.rb

@@ -24,6 +24,7 @@ module AsciiChem
           xml = translation.document.to_xml
           xml = inject_atom_extensions(xml, translation.atom_mapping)
           xml = inject_reaction_conditions(xml, formula)
+          xml = inject_metadata(xml, formula)
           inject_molecule_extensions(xml, formula, translation)
         end
 
@@ -32,12 +33,14 @@ module AsciiChem
           top_level = Extensions.extract_top_level(xml)
           atom_extensions = Extensions.extract(xml)
           group_extensions = GroupExtensions.extract(xml)
+          metadata_map = extract_metadata(xml)
           reaction_conditions = extract_reaction_conditions(xml)
           wire_doc = Chemicalml::Cml::Document.from_xml(xml)
           formula = AsciiChem::ModelAdapter.from_canonical(wire_doc)
           Extensions.restore(formula, wire_doc, atom_extensions)
           GroupExtensions.restore(formula, wire_doc, group_extensions)
           restore_reaction_conditions(formula, reaction_conditions)
+          restore_metadata(formula, metadata_map)
           Extensions.restore_top_level(formula, top_level)
           formula
         end
@@ -115,6 +118,65 @@ module AsciiChem
           formula
         end
 
+        # Inject molecule metadata via aci: attributes on <molecule>.
+        # Each {name: "k", content: "v"} produces aci:meta-k="v".
+        def inject_metadata(xml, formula)
+          require 'nokogiri'
+          doc = Nokogiri::XML(xml)
+          root = doc.root
+          molecules = formula.nodes.select { |n| n.is_a?(AsciiChem::Model::Molecule) }
+          has_meta = molecules.any? { |m| !m.metadata.empty? }
+          return xml unless has_meta
+
+          root.add_namespace(Extensions::PREFIX, Extensions::NAMESPACE) unless root.namespaces.value?(Extensions::NAMESPACE)
+          molecules.each_with_index do |mol, idx|
+            next if mol.metadata.empty?
+
+            mol_el = root.at_xpath("//cml:molecule[@id='m#{idx + 1}']", cml: Extensions::CML_NS)
+            next unless mol_el
+
+            mol.metadata.each do |m|
+              mol_el["#{Extensions::PREFIX}:meta-#{m[:name]}"] = m[:content]
+            end
+          end
+          doc.to_xml
+        end
+
+        def extract_metadata(xml)
+          require 'nokogiri'
+          doc = Nokogiri::XML(xml)
+          result = {}
+          doc.xpath("//cml:molecule", cml: Extensions::CML_NS).each do |el|
+            id = el['id']
+            next unless id
+
+            meta = {}
+            el.attributes.each do |name, attr|
+              next unless name.start_with?('meta-')
+              next unless attr.namespace && attr.namespace.prefix == Extensions::PREFIX
+
+              key = name.sub('meta-', '')
+              meta[key] = attr.value
+            end
+            result[id] = meta unless meta.empty?
+          end
+          result
+        end
+
+        def restore_metadata(formula, metadata_map)
+          return formula if metadata_map.empty?
+
+          formula.nodes.each_with_index do |node, idx|
+            next unless node.is_a?(AsciiChem::Model::Molecule)
+
+            meta = metadata_map["m#{idx + 1}"]
+            next unless meta
+
+            meta.each { |name, content| node.metadata << { name: name, content: content } }
+          end
+          formula
+        end
+
         def inject_molecule_extensions(xml, formula, translation)
           xml = inject_groups(xml, formula, translation)
           inject_top_level(xml, formula)

data/lib/asciichem/formatter/text.rb

@@ -57,15 +57,19 @@ module AsciiChem
       end
 
       def atom_annotation(atom)
-        annotation = +""
+        parts = []
         if atom.x2 && atom.y2
-          annotation << "@(#{format_coord(atom.x2)},#{format_coord(atom.y2)}"
-          annotation << ",#{format_coord(atom.z2)}" if atom.z2
-          annotation << ")"
-        elsif atom.atom_parity
-          annotation << "@#{atom.atom_parity}"
+          coord = "@(#{format_coord(atom.x2)},#{format_coord(atom.y2)}"
+          coord << ",#{format_coord(atom.z2)}" if atom.z2
+          parts << "#{coord})"
+        end
+        parts << "@#{atom.atom_parity}" if atom.atom_parity
+        parts << "@m(#{atom.spin_multiplicity})" if atom.spin_multiplicity
+        parts << %(@t("#{atom.atom_title}")) if atom.atom_title
+        if atom.x_fract && atom.y_fract && atom.z_fract
+          parts << "@f(#{format_coord(atom.x_fract)},#{format_coord(atom.y_fract)},#{format_coord(atom.z_fract)})"
         end
-        annotation
+        parts.join
       end
 
       def format_coord(value)

data/lib/asciichem/grammar.rb

@@ -182,7 +182,7 @@ module AsciiChem
         atom_suffix >>
         lewis_radicals.maybe >>
         ring_closures.maybe.as(:ring_closures) >>
-        atom_annotation.maybe).as(:atom)
+        atom_annotation).as(:atom)
     end
 
     rule(:plain_atom) do
@@ -191,15 +191,18 @@ module AsciiChem
         atom_suffix >>
         lewis_radicals.maybe >>
         ring_closures.maybe.as(:ring_closures) >>
-        atom_annotation.maybe).as(:atom)
+        atom_annotation).as(:atom)
     end
 
-    # Atom annotations: stereo parity (@R / @S) or 2D/3D coordinates
-    # (@(x,y) / @(x,y,z)). Both use the `@` prefix. An atom can carry
-    # at most one annotation in the grammar; multiple annotations
-    # would require compound syntax (deferred).
+    # Atom annotations: each is independently optional via .maybe.
+    # Order matters: @(x,y) → @R/@S → @m(N) → @t("...") → @f(x,y,z).
+    # Example: C@(10,20)@R@m(2)@t("C1")@f(0.5,0.5,0.5)
     rule(:atom_annotation) do
-      coordinate_annotation | parity_annotation
+      coordinate_annotation.maybe >>
+      parity_annotation.maybe >>
+      multiplicity_annotation.maybe >>
+      atom_title_annotation.maybe >>
+      fractional_annotation.maybe
     end
 
     rule(:parity_annotation) do
@@ -214,6 +217,27 @@ module AsciiChem
         str(')')
     end
 
+    # Spin multiplicity: @m(2) for doublet, @m(1) for singlet
+    rule(:multiplicity_annotation) do
+      str('@m(') >> match('[0-9]').repeat(1).as(:spin_multiplicity) >> str(')')
+    end
+
+    # Atom title/label: @t("C1")
+    rule(:atom_title_annotation) do
+      str('@t(') >> str('"') >>
+      (str('"').absent? >> any).repeat.as(:atom_title) >>
+      str('"') >> str(')')
+    end
+
+    # Fractional coordinates (crystallographic): @f(0.5,0.25,0.75)
+    rule(:fractional_annotation) do
+      str('@f(') >>
+        float_number.as(:x_fract) >> str(',') >>
+        float_number.as(:y_fract) >> str(',') >>
+        float_number.as(:z_fract) >>
+      str(')')
+    end
+
     rule(:float_number) do
       str('-').maybe >> match('[0-9]').repeat(1) >> (str('.') >> match('[0-9]').repeat(0)).maybe
     end

data/lib/asciichem/model/atom.rb

@@ -20,13 +20,17 @@ module AsciiChem
                     :charge, :oxidation_state,
                     :lone_pairs, :radical_electrons,
                     :ring_closures,
-                    :x2, :y2, :z2, :atom_parity
+                    :x2, :y2, :z2, :atom_parity,
+                    :spin_multiplicity, :atom_title,
+                    :x_fract, :y_fract, :z_fract
 
       def initialize(element:, isotope: nil, subscript: nil,
                      superscript: nil, charge: nil, oxidation_state: nil,
                      lone_pairs: nil, radical_electrons: nil,
                      ring_closures: nil,
-                     x2: nil, y2: nil, z2: nil, atom_parity: nil)
+                     x2: nil, y2: nil, z2: nil, atom_parity: nil,
+                     spin_multiplicity: nil, atom_title: nil,
+                     x_fract: nil, y_fract: nil, z_fract: nil)
         @element = element
         @isotope = isotope
         @subscript = subscript
@@ -40,6 +44,11 @@ module AsciiChem
         @y2 = y2
         @z2 = z2
         @atom_parity = atom_parity
+        @spin_multiplicity = spin_multiplicity
+        @atom_title = atom_title
+        @x_fract = x_fract
+        @y_fract = y_fract
+        @z_fract = z_fract
       end
 
       def value_attributes
@@ -48,7 +57,9 @@ module AsciiChem
           oxidation_state: oxidation_state,
           lone_pairs: lone_pairs, radical_electrons: radical_electrons,
           ring_closures: ring_closures,
-          x2: x2, y2: y2, z2: z2, atom_parity: atom_parity }
+          x2: x2, y2: y2, z2: z2, atom_parity: atom_parity,
+          spin_multiplicity: spin_multiplicity, atom_title: atom_title,
+          x_fract: x_fract, y_fract: y_fract, z_fract: z_fract }
       end
 
       def diagnostic_label

data/lib/asciichem/model_adapter/from_canonical.rb

@@ -104,7 +104,7 @@ module AsciiChem
         canonical_properties.map do |p|
           {
             title: p.title,
-            value: extract_scalar_value(p.value),
+            value: extract_scalar_value(p.scalar),
             dict_ref: p.dict_ref,
             convention: p.convention
           }
@@ -113,7 +113,7 @@ module AsciiChem
 
       def extract_scalar_value(value)
         return nil if value.nil?
-        return value.value if value.is_a?(Chemicalml::Cml::Scalar)
+        return value.content if value.is_a?(Chemicalml::Cml::Scalar)
 
         value.to_s
       end
@@ -215,7 +215,10 @@ module AsciiChem
             isotope: atom.isotope,
             subscript: subscript_from_count(atom.count),
             charge: atom.formal_charge,
-            **extract_coordinates(atom)
+            spin_multiplicity: atom.spin_multiplicity,
+            atom_title: atom.title,
+            **extract_coordinates(atom),
+            **extract_fractional(atom)
           )
         end
 
@@ -232,6 +235,14 @@ module AsciiChem
           end
         end
 
+        def extract_fractional(atom)
+          return {} unless atom.xFract && atom.yFract && atom.zFract
+
+          { x_fract: atom.xFract.to_f,
+            y_fract: atom.yFract.to_f,
+            z_fract: atom.zFract.to_f }
+        end
+
         def subscript_from_count(count)
           return nil if count.nil? || count.to_s == "1"
 

data/lib/asciichem/model_adapter/to_canonical.rb

@@ -130,7 +130,7 @@ module AsciiChem
         properties.map do |p|
           Chemicalml::Cml::Property.new(
             title: p[:title],
-            value: build_scalar(p[:value]),
+            scalar: build_scalar(p[:value]),
             dict_ref: p[:dict_ref],
             convention: p[:convention]
           )
@@ -141,7 +141,7 @@ module AsciiChem
         return nil if value.nil?
 
         Chemicalml::Cml::Scalar.new(
-          value: value.to_s,
+          content: value.to_s,
           dict_ref: nil
         )
       end
@@ -330,9 +330,12 @@ module AsciiChem
             formal_charge: atom.charge,
             count: effective_count(atom, multiplier),
             lone_pairs: atom.lone_pairs,
-            radical_electrons: atom.radical_electrons
+            radical_electrons: atom.radical_electrons,
+            spin_multiplicity: atom.spin_multiplicity,
+            title: atom.atom_title
           }
           merge_coordinates(attrs, atom)
+          merge_fractional_coords(attrs, atom)
           @atoms << Chemicalml::Cml::Atom.new(**attrs)
           emit_pending_bond(id) if @pending_bond_kind && @last_atom_id
           @last_atom_id = id
@@ -386,6 +389,14 @@ module AsciiChem
           end
         end
 
+        def merge_fractional_coords(attrs, atom)
+          return unless atom.x_fract && atom.y_fract && atom.z_fract
+
+          attrs[:xFract] = atom.x_fract
+          attrs[:yFract] = atom.y_fract
+          attrs[:zFract] = atom.z_fract
+        end
+
         def combine(left, right)
           lv = integer_or_nil(left)
           rv = integer_or_nil(right)

data/lib/asciichem/transform.rb

@@ -272,7 +272,12 @@ module AsciiChem
           x2: float_or_nil(hash[:x2]),
           y2: float_or_nil(hash[:y2]),
           z2: float_or_nil(hash[:z2]),
-          atom_parity: hash[:atom_parity]&.to_s
+          atom_parity: hash[:atom_parity]&.to_s,
+          spin_multiplicity: hash[:spin_multiplicity]&.to_s,
+          atom_title: hash[:atom_title]&.to_s,
+          x_fract: float_or_nil(hash[:x_fract]),
+          y_fract: float_or_nil(hash[:y_fract]),
+          z_fract: float_or_nil(hash[:z_fract])
         )
       end
 
@@ -305,7 +310,9 @@ module AsciiChem
       def initialize(element, isotope: nil, subscript: nil, superscript: nil,
                      lone_pairs: nil, radical_electrons: nil,
                      ring_closures: nil,
-                     x2: nil, y2: nil, z2: nil, atom_parity: nil)
+                     x2: nil, y2: nil, z2: nil, atom_parity: nil,
+                     spin_multiplicity: nil, atom_title: nil,
+                     x_fract: nil, y_fract: nil, z_fract: nil)
         @element = element
         @isotope = isotope
         @subscript = subscript
@@ -317,6 +324,11 @@ module AsciiChem
         @y2 = y2
         @z2 = z2
         @atom_parity = atom_parity
+        @spin_multiplicity = spin_multiplicity
+        @atom_title = atom_title
+        @x_fract = x_fract
+        @y_fract = y_fract
+        @z_fract = z_fract
       end
 
       def build
@@ -335,7 +347,12 @@ module AsciiChem
           x2: @x2,
           y2: @y2,
           z2: @z2,
-          atom_parity: @atom_parity
+          atom_parity: @atom_parity,
+          spin_multiplicity: @spin_multiplicity,
+          atom_title: @atom_title,
+          x_fract: @x_fract,
+          y_fract: @y_fract,
+          z_fract: @z_fract
         )
       end
 

data/lib/asciichem/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module AsciiChem
-  VERSION = "0.3.2"
+  VERSION = "0.3.4"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: asciichem
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.3.4
 platform: ruby
 authors:
 - Ribose Inc.