coradoc-docx 0.1.2 → 0.1.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e46f27d58495c1034b6fcb7e18a4b3f7361399bae351f07feecc70bf96998c00
-  data.tar.gz: e635bc232ec683c7ac69d29ca3051cb7ff79858fb6726d97cf57abeef736405a
+  metadata.gz: 18f25e708f14692cd5a7b48a8636e34df3eee0b5ac3fa8a7f0904a5f56acbeaa
+  data.tar.gz: b15177104fe5b8b50635fb6856fa059856c0fb4abfea48efaf15a7bb0cad0a20
 SHA512:
-  metadata.gz: 8f59bc132f05dbc709ed0c7b31e6c633e94efde7a95e4bbf1661c18e3a737ad0ca8e592257a1029c9935e57ce679f771705cef1bbdeea0b4d00c11e4e112e4d9
-  data.tar.gz: 926131183805ae1163e544f464417c21fe38dfb5705cb363427c83e053d7f60c8a71f65cfd0189f9003cc6498f1ec0fa54215ef3d21d87fd721619c844de7b3a
+  metadata.gz: 9889857c0c723281bc2c45903c31898b8881d49ff431a7aef699882a74dd1243009f9d1cbd548bc9970f09131b34316743a90d00ee99e6ce9f7509dec7f3276a
+  data.tar.gz: b1a47163cdb94d6ecc07d0d0fb4ca1a36eb9539867620f5b7ad861aa7d789fedaaa6018d5731855938ba164b47ae35874d6fb29d028f67e6567221619a2cfdbd

data/lib/coradoc/docx/transform/from_core_model.rb

@@ -89,6 +89,8 @@ module Coradoc
           doc = Uniword::Builder::DocumentBuilder.new
           doc.title(element.title) if element.title
 
+          FrontmatterCoreProperties.apply(doc, FrontmatterCoreProperties.extract(element))
+
           transform_children(element.children, doc)
 
           doc.build
@@ -232,6 +234,14 @@ module Coradoc
           builder.table do |t|
             table.rows.each do |row|
               t.row do |r|
+                # TableRowBuilder#header_row is broken upstream (calls a
+                # nonexistent table_header= setter), so set the persisted
+                # w:tblHeader property on the row model directly
+                if row.header
+                  r.model.properties = Uniword::Wordprocessingml::TableRowProperties.new(
+                    tbl_header: Uniword::Wordprocessingml::ValInt.new(value: 1)
+                  )
+                end
                 row.cells.each do |cell|
                   r.cell(text: cell.flat_text)
                 end
@@ -293,7 +303,9 @@ module Coradoc
 
           props = Uniword::Wordprocessingml::RunProperties.new
 
-          case inline.format_type
+          # resolve_format_type covers typed subclasses (BoldElement etc.)
+          # that carry no explicit format_type attribute
+          case inline.resolve_format_type
           when 'bold'
             props.bold = Uniword::Properties::Bold.new
           when 'italic'
@@ -353,6 +365,11 @@ module Coradoc
 
           table.rows.each do |row|
             tr = Uniword::Wordprocessingml::TableRow.new
+            if row.header
+              tr.properties = Uniword::Wordprocessingml::TableRowProperties.new(
+                tbl_header: Uniword::Wordprocessingml::ValInt.new(value: 1)
+              )
+            end
             row.cells.each do |cell|
               tc = Uniword::Wordprocessingml::TableCell.new
 
@@ -393,8 +410,7 @@ module Coradoc
         def build_page_break
           para = Uniword::Wordprocessingml::Paragraph.new
           run = Uniword::Wordprocessingml::Run.new
-          run.break = Uniword::Wordprocessingml::Break.new
-          run.break.type = 'page'
+          run.break = Uniword::Wordprocessingml::Break.new(type: 'page')
           para.runs << run
           para
         end
@@ -402,7 +418,9 @@ module Coradoc
         def build_inline_text(inline)
           text = inline.content.to_s
 
-          case inline.format_type
+          # resolve_format_type covers typed subclasses (BoldElement etc.)
+          # that carry no explicit format_type attribute
+          case inline.resolve_format_type
           when 'bold'    then Uniword::Builder.text(text, bold: true)
           when 'italic'  then Uniword::Builder.text(text, italic: true)
           when 'underline' then Uniword::Builder.text(text, underline: true)
@@ -413,7 +431,7 @@ module Coradoc
             run = Uniword::Wordprocessingml::Run.new(text: text)
             props = Uniword::Wordprocessingml::RunProperties.new
             va = Uniword::Properties::VerticalAlign.new
-            va.value = inline.format_type
+            va.value = inline.resolve_format_type
             props.vertical_align = va
             run.properties = props
             run

data/lib/coradoc/docx/transform/frontmatter_core_properties.rb

@@ -0,0 +1,107 @@
+# frozen_string_literal: true
+
+module Coradoc
+  module Docx
+    module Transform
+      # Single source of truth for FrontmatterBlock -> OOXML core
+      # properties mapping. (MECE: DOCX-specific concerns live in the
+      # DOCX gem; CoreModel has no knowledge of OOXML.)
+      #
+      # Uniword's DocumentBuilder exposes setter methods (title, author,
+      # subject, keywords, description, date_field) that write the
+      # docProps/core.xml Dublin Core elements. This module reads
+      # FrontmatterBlock entries and invokes the right setters, so the
+      # mapping table lives in exactly one place.
+      #
+      # Mapping table (single source of truth):
+      #
+      #   | Frontmatter key       | OOXML element        | Builder method |
+      #   |-----------------------|----------------------|----------------|
+      #   | title                 | dc:title             | title          |
+      #   | author                | dc:creator           | author         |
+      #   | date                  | dc:date              | created        |
+      #   | description, excerpt  | dc:description       | description    |
+      #   | subject               | dc:subject           | subject        |
+      #   | tags, categories      | cp:keywords          | keywords       |
+      module FrontmatterCoreProperties
+        DESCRIPTION_KEYS = %w[description excerpt].freeze
+        KEYWORDS_KEYS = %w[tags categories].freeze
+
+        class << self
+          # Apply FrontmatterBlock data to a Uniword DocumentBuilder.
+          # Skips nil/empty values. Does not overwrite builder values
+          # that the caller has set explicitly (caller invokes this
+          # AFTER setting document.title, for example).
+          #
+          # @param builder [Uniword::Builder::DocumentBuilder]
+          # @param block [Coradoc::CoreModel::FrontmatterBlock, nil]
+          # @return [void]
+          def apply(builder, block)
+            return unless block.is_a?(Coradoc::CoreModel::FrontmatterBlock)
+
+            data = Coradoc::CoreModel::FrontmatterBlock::Codec.to_hash(block)
+
+            if (v = find_first_scalar(data, KEYWORDS_KEYS))
+              builder.keywords(v)
+            end
+            if (v = find_first_scalar(data, %w[author]))
+              builder.author(v)
+            end
+            if (v = find_first_scalar(data, %w[subject]))
+              builder.subject(v)
+            end
+            if (v = find_first_scalar(data, DESCRIPTION_KEYS))
+              builder.description(v)
+            end
+            if (v = find_first_scalar(data, %w[date]))
+              builder.created(v)
+            end
+          end
+
+          # Locate the FrontmatterBlock in a DocumentElement's children,
+          # if any.
+          #
+          # @param document [Coradoc::CoreModel::DocumentElement]
+          # @return [Coradoc::CoreModel::FrontmatterBlock, nil]
+          def extract(document)
+            return nil unless document.is_a?(Coradoc::CoreModel::DocumentElement)
+
+            Array(document.children).find { |c| c.is_a?(Coradoc::CoreModel::FrontmatterBlock) }
+          end
+
+          private
+
+          def find_first_scalar(data, keys)
+            keys.each do |k|
+              value = data[k.to_s]
+              next if value.nil?
+
+              str = scalar_value(value)
+              return str if str && !str.empty?
+            end
+            nil
+          end
+
+          def scalar_value(value)
+            return nil if value.nil?
+
+            case value
+            when String then value
+            when Integer, Float, TrueClass, FalseClass then value.to_s
+            when Date, Time, DateTime then value.iso8601
+            when Symbol then value.to_s
+            when Array then array_to_csv(value)
+            when Hash then nil
+            else value.to_s
+            end
+          end
+
+          def array_to_csv(value)
+            strings = value.map { |i| scalar_value(i) }.compact
+            strings.any? ? strings.join(', ') : nil
+          end
+        end
+      end
+    end
+  end
+end

data/lib/coradoc/docx/transform/numbering_access.rb

@@ -0,0 +1,62 @@
+# frozen_string_literal: true
+
+module Coradoc
+  module Docx
+    module Transform
+      # Shared access to a paragraph's numbering (list) attributes.
+      #
+      # uniword persists w:numPr as NESTED properties:
+      #   paragraph.properties.numbering_properties.num_id.value
+      #   paragraph.properties.numbering_properties.ilvl.value
+      # while the flat integers (properties.num_id / properties.ilvl) are
+      # only populated programmatically (builder/YAML API). Readers must
+      # consult the nested shape first and fall back to the flat one —
+      # documents deserialized from DOCX only ever carry the nested shape.
+      #
+      # lutaml-model also returns an UninitializedClass sentinel for
+      # attributes that were not deserialized from source, so every read
+      # is guarded by an Integer type check instead of truthiness.
+      module NumberingAccess
+        # Numbering definition ID (w:numId) for a paragraph.
+        #
+        # @param paragraph [Uniword::Wordprocessingml::Paragraph]
+        # @return [Integer, nil]
+        def paragraph_num_id(paragraph)
+          nested_numbering_value(paragraph, :num_id)
+        end
+
+        # Numbering level (w:ilvl) for a paragraph.
+        #
+        # @param paragraph [Uniword::Wordprocessingml::Paragraph]
+        # @return [Integer, nil]
+        def paragraph_ilvl(paragraph)
+          nested_numbering_value(paragraph, :ilvl)
+        end
+
+        private
+
+        def nested_numbering_value(paragraph, attribute)
+          numbering = paragraph_properties(paragraph)&.numbering_properties
+          nested = numbering&.public_send(attribute)
+          value = wrapper_value(nested)
+          return value if value.is_a?(Integer)
+
+          # Fall back to the flat builder/YAML attribute
+          flat = paragraph_properties(paragraph)&.public_send(attribute)
+          flat.is_a?(Integer) ? flat : nil
+        end
+
+        def paragraph_properties(paragraph)
+          paragraph.properties if paragraph.respond_to?(:properties)
+        end
+
+        # Unwrap w:numId / w:ilvl element models to their value; sentinel
+        # objects (UninitializedClass) and nils fail the Integer check.
+        def wrapper_value(node)
+          value = node.respond_to?(:value) ? node.value : node
+          value.is_a?(Integer) ? value : nil
+        end
+      end
+    end
+  end
+end

data/lib/coradoc/docx/transform/rules/list_item_rule.rb

@@ -17,13 +17,14 @@ module Coradoc
         # dispatches directly after checking style_resolver.list_item?.
         class ListItemRule < Rule
           include OrderedContent
+          include NumberingAccess
 
           def matches?(_element)
             false # Never auto-matched; orchestrator dispatches directly
           end
 
           def apply(paragraph, context)
-            ilvl = paragraph.properties&.ilvl.to_i
+            ilvl = paragraph_ilvl(paragraph).to_i
 
             children = transform_paragraph_content(paragraph, context)
 

data/lib/coradoc/docx/transform/rules/math_rule.rb

@@ -6,23 +6,20 @@ module Coradoc
       module Rules
         # Transforms OMML math elements to CoreModel.
         #
-        # Display math (m:oMathPara) → CoreModel::Block (stem)
-        # Inline math (m:oMath) → CoreModel::InlineElement (stem)
+        # Display math (MathEquation#block?) → CoreModel::Block (stem)
+        # Inline math (MathEquation#inline?) → CoreModel::InlineElement (stem)
         #
-        # Uses Plurimath for OMML → LaTeX conversion when available.
-        # Falls back to raw XML string when Plurimath is not loaded.
+        # LaTeX conversion is provided natively by Uniword::MathEquation.
         class MathRule < Rule
           def matches?(element)
-            return false unless defined?(Uniword::Math)
-
-            element.is_a?(Uniword::Math::OMathPara) ||
-              element.is_a?(Uniword::Math::OMath)
+            defined?(Uniword::MathEquation) &&
+              element.is_a?(Uniword::MathEquation)
           end
 
           def apply(element, _context)
-            latex = omml_to_latex(element)
+            latex = element.to_latex.to_s
 
-            if display_math?(element)
+            if element.block?
               CoreModel::PassBlock.new(
                 delimiter_type: '++++',
                 language: 'latexmath',
@@ -35,45 +32,6 @@ module Coradoc
               )
             end
           end
-
-          private
-
-          def display_math?(element)
-            defined?(Uniword::Math::OMathPara) &&
-              element.is_a?(Uniword::Math::OMathPara)
-          end
-
-          def omml_to_latex(element)
-            if defined?(Plurimath)
-              plurimath_to_latex(element)
-            else
-              # Fallback: serialize to XML string
-              element_respond_to_xml(element) || ''
-            end
-          end
-
-          def plurimath_to_latex(element)
-            xml = element_to_xml(element)
-            return '' if xml.nil? || xml.empty?
-
-            begin
-              formula = Plurimath::OMML.parse(xml)
-              formula.to_latex
-            rescue StandardError
-              ''
-            end
-          end
-
-          def element_to_xml(element)
-            return '' unless element.is_a?(Uniword::Wordprocessingml::AlternateContent) ||
-                             element.is_a?(Uniword::Wordprocessingml::OMath)
-
-            element.to_xml
-          end
-
-          def element_respond_to_xml(element)
-            element_to_xml(element)
-          end
         end
       end
     end

data/lib/coradoc/docx/transform/rules/run_rule.rb

@@ -25,9 +25,11 @@ module Coradoc
           def apply(run, context)
             # Delegate non-text children (breaks, drawings, footnotes, etc.)
             non_text = extract_non_text_children(run, context)
-            return non_text.first if non_text.any? && run.text.nil?
+            return non_text.first if non_text.any? && Array(run.text).empty?
 
-            text = run.text&.content.to_s
+            # w:r may carry several w:t children (uniword models text as
+            # a collection) — join them in document order
+            text = Array(run.text).map { |node| node&.content.to_s }.join
             return '' if text.empty? && non_text.empty?
 
             props = effective_props(run)
@@ -54,7 +56,7 @@ module Coradoc
           def extract_non_text_children(run, context)
             result = []
 
-            result << context.transform(run.break) if run.break
+            result.concat(Array(run.break).map { |brk| context.transform(brk) })
 
             result << context.transform(run.footnote_reference) if run.footnote_reference
 

data/lib/coradoc/docx/transform/rules/simple_field_rule.rb

@@ -64,7 +64,7 @@ module Coradoc
           def field_text(field)
             if field.runs && !field.runs.empty?
               return field.runs.map do |r|
-                r.text&.content.to_s
+                Array(r.text).map(&:to_s).join
               end.join
             end
 

data/lib/coradoc/docx/transform/rules/table_rule.rb

@@ -32,10 +32,23 @@ module Coradoc
           def transform_row(row, context)
             CoreModel::TableRow.new(
               cells: row.cells.map { |c| transform_cell(c, context) },
-              header: row.header?
+              header: row_header?(row)
             )
           end
 
+          # The persisted header flag lives in w:trPr/w:tblHeader
+          # (row.properties.tbl_header); TableRow#header? is a runtime-only
+          # accessor that deserialized documents never populate.
+          def row_header?(row)
+            tbl_header = row.properties&.tbl_header
+            return row.header? unless tbl_header.is_a?(Uniword::Wordprocessingml::ValInt)
+
+            # <w:tblHeader/> without w:val means "on" (ST_OnOff semantics)
+            return true if tbl_header.value.nil?
+
+            tbl_header.value.to_i.positive?
+          end
+
           def transform_cell(cell, context)
             inline_children = cell_paragraphs(cell).flat_map do |para|
               extract_inline_from_paragraph(para, context)

data/lib/coradoc/docx/transform/style_resolver.rb

@@ -15,6 +15,8 @@ module Coradoc
       # The style map is built from the Uniword StylesConfiguration by walking
       # all style definitions and their basedOn chains.
       class StyleResolver
+        include NumberingAccess
+
         HEADING_PATTERN = /^(heading|heading|h)\s*(\d+)$/i
         QUOTE_PATTERN = /\bquote\b/i
         CODE_PATTERN = /\b(code|source|listing)\b/i
@@ -53,14 +55,14 @@ module Coradoc
 
           ol = paragraph.properties.outline_level
           if ol
-            ol_level = ol.is_a?(Uniword::Wordprocessingml::OutlineLevel) ? ol.value.to_i : ol.to_i
+            ol_level = ol.is_a?(Uniword::Properties::OutlineLevel) ? ol.value.to_i : ol.to_i
             return true if ol_level.positive?
           end
 
           style = find_style_for_paragraph(paragraph)
           if style&.outline_level
             ol_val = style.outline_level
-            ol_val = ol_val.is_a?(Uniword::Wordprocessingml::OutlineLevel) ? ol_val.value.to_i : ol_val.to_i
+            ol_val = ol_val.is_a?(Uniword::Properties::OutlineLevel) ? ol_val.value.to_i : ol_val.to_i
             return true if ol_val.positive?
           end
 
@@ -80,7 +82,7 @@ module Coradoc
           # Check outline_level on paragraph properties
           ol = paragraph.properties&.outline_level
           if ol
-            level = ol.is_a?(Uniword::Wordprocessingml::OutlineLevel) ? ol.value.to_i : ol.to_i
+            level = ol.is_a?(Uniword::Properties::OutlineLevel) ? ol.value.to_i : ol.to_i
             return level if level.positive?
           end
 
@@ -93,8 +95,10 @@ module Coradoc
         def list_item?(paragraph)
           return false unless paragraph.properties
 
-          num_id = paragraph.properties.num_id
-          num_id.to_i.positive?
+          # w:numPr is persisted as nested numbering_properties; the flat
+          # num_id is only set by the builder/YAML API (see NumberingAccess)
+          num_id = paragraph_num_id(paragraph)
+          !num_id.nil? && num_id.positive?
         end
 
         # Check if paragraph has a specific role based on style name
@@ -133,10 +137,7 @@ module Coradoc
         private
 
         def resolve_style_name(paragraph)
-          style_ref = paragraph.properties&.style
-          return nil unless style_ref
-
-          value = style_ref.is_a?(Uniword::Wordprocessingml::PStyle) ? style_ref.val : style_ref.to_s
+          value = paragraph_style_value(paragraph)
           return nil unless value
 
           mapped = @style_map[value]
@@ -149,7 +150,7 @@ module Coradoc
           style_ref = run.properties.style
           return nil unless style_ref
 
-          value = style_ref.is_a?(Uniword::Wordprocessingml::PStyle) ? style_ref.val : style_ref.to_s
+          value = style_ref.value.to_s
           return nil unless value
 
           mapped = @style_map[value]
@@ -168,10 +169,12 @@ module Coradoc
         end
 
         def style_id_from_paragraph(paragraph)
-          style_ref = paragraph.properties&.style
-          return nil unless style_ref
+          paragraph_style_value(paragraph)
+        end
 
-          style_ref.is_a?(Uniword::Wordprocessingml::PStyle) ? style_ref.val : style_ref.to_s
+        # uniword models w:pStyle as a collection of StyleReference
+        def paragraph_style_value(paragraph)
+          Array(paragraph.properties&.style).filter_map(&:value).first.to_s
         end
 
         def build_style_map(config)

data/lib/coradoc/docx/transform/to_core_model.rb

@@ -24,6 +24,8 @@ module Coradoc
       #   core = ToCoreModel.transform(doc)
       #   # => Coradoc::CoreModel::StructuralElement
       class ToCoreModel
+        include NumberingAccess
+
         class << self
           def transform(document)
             new.transform(document)
@@ -139,7 +141,7 @@ module Coradoc
 
         def section_break_element(paragraph, context)
           # First, transform the paragraph content if it has text
-          content = paragraph.runs&.map { |r| r.text&.content.to_s }&.join
+          content = paragraph.runs&.map { |r| Array(r.text).map(&:to_s).join }&.join
           if content && !content.strip.empty?
             # Has content — transform normally (content comes before the break)
             context.transform(paragraph)
@@ -154,7 +156,7 @@ module Coradoc
         # Collect consecutive list items with the same numId into a ListBlock
         def group_list(elements, start_index, context)
           first = elements[start_index]
-          num_id = first.properties&.num_id.to_i
+          num_id = paragraph_num_id(first).to_i
           items = []
           consumed = 0
 
@@ -163,7 +165,7 @@ module Coradoc
             para = elements[idx]
             break unless paragraph?(para)
             break unless context.style_resolver.list_item?(para)
-            break unless para.properties&.num_id.to_i == num_id
+            break unless paragraph_num_id(para).to_i == num_id
 
             items << @list_item_rule.apply(para, context)
             consumed += 1
@@ -242,7 +244,7 @@ module Coradoc
             next unless context.style_resolver.heading_level(para) == 1
 
             runs = para.runs || []
-            return runs.map { |r| r.text&.content.to_s }.join
+            return runs.map { |r| Array(r.text).map(&:to_s).join }.join
           end
 
           nil
@@ -312,7 +314,7 @@ module Coradoc
           runs = para.runs || []
           return nil unless runs
 
-          runs.map { |r| r.text&.content.to_s }.join
+          runs.map { |r| Array(r.text).map(&:to_s).join }.join
         end
 
         # Check if header/footer text is purely layout content

data/lib/coradoc/docx/transform.rb

@@ -10,7 +10,9 @@ module Coradoc
       autoload :FromCoreModel, 'coradoc/docx/transform/from_core_model'
       autoload :StyleResolver, 'coradoc/docx/transform/style_resolver'
       autoload :NumberingResolver, 'coradoc/docx/transform/numbering_resolver'
+      autoload :NumberingAccess, 'coradoc/docx/transform/numbering_access'
       autoload :OrderedContent, 'coradoc/docx/transform/ordered_content'
+      autoload :FrontmatterCoreProperties, 'coradoc/docx/transform/frontmatter_core_properties'
 
       # Element transform rules
       module Rules

data/lib/coradoc/docx/version.rb

@@ -2,6 +2,6 @@
 
 module Coradoc
   module Docx
-    VERSION = '0.1.2'
+    VERSION = '0.1.4'
   end
 end

data/lib/coradoc/docx.rb

@@ -1,7 +1,6 @@
 # frozen_string_literal: true
 
 require 'coradoc'
-require 'coradoc/core_model'
 require 'uniword'
 
 # Coradoc::Docx provides DOCX (OOXML) format support for Coradoc.

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: coradoc-docx
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.4
 platform: ruby
 authors:
 - Ribose Inc.
@@ -41,16 +41,16 @@ dependencies:
   name: uniword
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '1.6'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '1.6'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -108,6 +108,8 @@ files:
 - lib/coradoc/docx/transform.rb
 - lib/coradoc/docx/transform/context.rb
 - lib/coradoc/docx/transform/from_core_model.rb
+- lib/coradoc/docx/transform/frontmatter_core_properties.rb
+- lib/coradoc/docx/transform/numbering_access.rb
 - lib/coradoc/docx/transform/numbering_resolver.rb
 - lib/coradoc/docx/transform/ordered_content.rb
 - lib/coradoc/docx/transform/rule.rb
@@ -136,6 +138,7 @@ licenses:
 metadata:
   homepage_uri: https://github.com/lutaml/coradoc
   source_code_uri: https://github.com/lutaml/coradoc
+  rubygems_mfa_required: 'true'
 rdoc_options: []
 require_paths:
 - lib