uniword 1.6.1 → 1.7.1

data/lib/uniword/validation/schema_registry.rb

@@ -1,14 +1,14 @@
 # frozen_string_literal: true
 
 require "moxml"
-require "nokogiri"
+require "leptris/xml/xsd"
 
 module Uniword
   module Validation
     # Maps XML namespace URIs to XSD schema files and loads schemas for validation.
     #
     # Uses Moxml for namespace detection (parsing XML to find declared namespaces)
-    # and Nokogiri::XML::Schema for actual XSD validation.
+    # and Moxml::XSD (Leptris engine) for actual XSD validation.
     #
     # The registry knows which OOXML namespace URIs correspond to which XSD files
     # bundled in data/schemas/.
@@ -17,7 +17,7 @@ module Uniword
     #   registry = SchemaRegistry.new
     #   ns_uris = registry.detect_namespaces(xml_string)
     #   schema = registry.schema_for_namespaces(ns_uris)
-    #   errors = schema.validate(Nokogiri::XML(xml_string))
+    #   errors = schema.validate(xml_string)
     class SchemaRegistry
       # Map of namespace URI => XSD file (relative to data/schemas/)
       NAMESPACE_XSD_MAP = {
@@ -145,7 +145,7 @@ module Uniword
       # Load and cache an XSD schema for validation.
       #
       # @param xsd_relative_path [String] Path relative to schemas_dir
-      # @return [Nokogiri::XML::Schema] Compiled schema
+      # @return [Moxml::XSD::Schema] Compiled schema
       def load_schema(xsd_relative_path)
         return @schema_cache[xsd_relative_path] if @schema_cache.key?(xsd_relative_path)
 
@@ -155,18 +155,11 @@ module Uniword
                 "XSD schema not found: #{xsd_path}"
         end
 
-        # Nokogiri::XML::Schema resolves relative imports from CWD.
-        # We chdir to the schema file's own directory so imports like
-        # "word12.xsd" (from microsoft/wml-2010.xsd) resolve correctly.
-        schema = nil
-        original_dir = Dir.pwd
-        schema_dir = File.dirname(xsd_path)
-        begin
-          Dir.chdir(schema_dir)
-          schema = Nokogiri::XML::Schema(File.read(File.basename(xsd_path)))
-        ensure
-          Dir.chdir(original_dir)
-        end
+        # Path-based compilation resolves xsd:import chains relative
+        # to the schema file's own directory; string compilation does
+        # not resolve them.
+        handle = Leptris::XML::XSD.compile_file(xsd_path)
+        schema = Moxml::XSD::Schema.new(handle, @moxml)
 
         @schema_cache[xsd_relative_path] = schema
         schema

data/lib/uniword/validation/validators/xml_schema_validator.rb

@@ -1,7 +1,7 @@
 # frozen_string_literal: true
 
 require "zip"
-require "nokogiri"
+require "moxml"
 # LayerValidator autoloaded via lib/uniword/validation.rb
 
 module Uniword
@@ -14,7 +14,8 @@ module Uniword
       # - XSD validation: parts validated against OOXML XSD schemas
       #
       # XSD validation uses SchemaRegistry for namespace-aware schema selection.
-      # Uses Moxml for namespace detection and Nokogiri::XML::Schema for validation.
+      # Uses Moxml throughout: parse, namespace detection, and the
+      # Moxml::XSD (Leptris engine) validator.
       #
       # @example Validate with XSD
       #   validator = XmlSchemaValidator.new("xml_schema" => { "xsd_validation" => true })
@@ -66,8 +67,8 @@ module Uniword
 
         def validate_entry(entry, issues)
           content = entry.get_input_stream.read
-          Nokogiri::XML(content, &:strict)
-        rescue Nokogiri::XML::SyntaxError => e
+          moxml.parse(content, strict: true)
+        rescue Moxml::ParseError => e
           issues << Report::ValidationIssue.new(
             severity: "error",
             code: "XSD-001",
@@ -87,16 +88,15 @@ module Uniword
             begin
               content = entry.get_input_stream.read
               schema = registry.load_schema(schema_path)
-              doc = Nokogiri::XML(content)
+              doc = moxml.parse(content)
               preprocess_mce(doc)
 
               schema.validate(doc).each do |error|
                 issues << Report::ValidationIssue.new(
                   severity: "error",
                   code: "XSD-002",
-                  message: error.message.gsub("\n", " ").strip,
+                  message: error.to_s.gsub("\n", " ").strip,
                   part: entry.name,
-                  line: error.line,
                 )
               end
 
@@ -120,42 +120,87 @@ module Uniword
           end
         end
 
-        # Strip MCE-ignorable ATTRIBUTES per the part's own mc:Ignorable
-        # declaration (w14:paraId/textId and the mc:Ignorable attribute
-        # itself), so XSD validation is not drowned in false errors for
-        # content every real document carries.
+        # Apply MCE preprocessing per the part's own mc:Ignorable
+        # declaration so strict XSD validation is not drowned in errors
+        # for extension content:
+        #
+        # - mc:AlternateContent is unwrapped to the first understood
+        #   mc:Choice (else the mc:Fallback) per MCE branch selection;
+        # - elements in ignorable namespaces are removed (w15:docId,
+        #   w16se:* settings extensions);
+        # - attributes in ignorable namespaces are removed (w14:paraId,
+        #   w14:textId), plus the mc:Ignorable attribute itself, which
+        #   is undeclared on most roots in the transitional XSDs.
         #
-        # Extension ELEMENTS are deliberately kept: removing them changes
-        # libxml2's content-model evaluation and can expose pre-existing
-        # Word-vs-schema quirks (e.g. rsids placed late, tolerated via the
-        # sequence's trailing xsd:any) as false baseline errors.
         # Operates in place on the parsed document; never touches the
         # packaged bytes.
         #
-        # @param doc [Nokogiri::XML::Document] parsed part to clean
+        # @param doc [Moxml::Document] parsed part to clean
         # @return [void]
         def preprocess_mce(doc)
           root = doc.root
           return unless root
 
-          ignorable = root["Ignorable"] || root["mc:Ignorable"]
-          return if ignorable.nil? || ignorable.empty?
+          uris = ignorable_namespace_uris(root)
+          unwrap_alternate_content(doc, uris)
+          remove_ignorable_elements(doc, uris) unless uris.empty?
+          strip_mce_attributes(root, uris + [MCE_NAMESPACE_URI])
+        end
 
-          uris = ignorable.split(/\s+/).filter_map do |prefix|
-            root.namespaces["xmlns:#{prefix}"]
+        # Namespace URIs the root declares ignorable via mc:Ignorable.
+        def ignorable_namespace_uris(root)
+          ignorable = root["mc:Ignorable"]
+          return [] if ignorable.nil? || ignorable.empty?
+
+          prefix_uris = root.namespaces.to_h { |ns| [ns.prefix, ns.uri] }
+          ignorable.split(/\s+/).filter_map { |prefix| prefix_uris[prefix] }
+        end
+
+        # Replace each mc:AlternateContent with the content of its first
+        # understood mc:Choice, or its mc:Fallback when no Choice applies.
+        def unwrap_alternate_content(doc, ignorable_uris)
+          doc.xpath("//mc:AlternateContent", "mc" => MCE_NAMESPACE_URI)
+            .to_a.reverse.each do |alternate|
+            branch = chosen_branch(alternate, ignorable_uris)
+            if branch
+              anchor = alternate
+              branch.children.each do |child|
+                anchor.add_next_sibling(child)
+                anchor = child
+              end
+            end
+            alternate.remove
+          end
+        end
+
+        # The first mc:Choice whose required namespaces are all
+        # non-ignorable, else the mc:Fallback, else nil.
+        def chosen_branch(alternate, ignorable_uris)
+          choices = alternate.xpath("./mc:Choice", "mc" => MCE_NAMESPACE_URI)
+          choices.find do |choice|
+            requires = choice.attributes
+              .find { |attr| attr.name == "Requires" }
+              &.value.to_s.split(/\s+/)
+            requires.all? { |prefix| !ignorable_uris.include?(prefix) }
+          end || alternate.at_xpath("./mc:Fallback", "mc" => MCE_NAMESPACE_URI)
+        end
+
+        # Remove elements whose namespace is declared ignorable.
+        def remove_ignorable_elements(doc, uris)
+          doc.traverse do |node|
+            next unless node.is_a?(Moxml::Element)
+
+            node.remove if uris.include?(node.namespace&.uri)
           end
-          # The mc:Ignorable attribute itself is MCE machinery and is
-          # undeclared on most root elements in the transitional XSDs.
-          strip_mce_attributes(root, uris + [MCE_NAMESPACE_URI])
         end
 
         # Remove attributes in ignorable namespaces from every element.
         def strip_mce_attributes(doc, uris)
           doc.traverse do |node|
-            next unless node.element?
+            next unless node.is_a?(Moxml::Element)
 
-            node.attribute_nodes.each do |attr|
-              attr.remove if uris.include?(attr.namespace&.href)
+            node.attributes.each do |attr|
+              attr.remove if uris.include?(attr.namespace&.uri)
             end
           end
         end
@@ -192,6 +237,10 @@ module Uniword
           layer_config["xsd_validation"] == true
         end
 
+        def moxml
+          @moxml ||= Moxml.new
+        end
+
         def layer_config
           @config["xml_schema"] || {}
         end

data/lib/uniword/validation/verify_orchestrator.rb

@@ -8,7 +8,7 @@ module Uniword
     #
     # Coordinates:
     # 1. OPC Package validation (ZIP, content types, relationships)
-    # 2. XSD Schema validation (namespace-aware, using Moxml + Nokogiri)
+    # 2. XSD Schema validation (namespace-aware, via Moxml::XSD)
     # 3. Word Document semantic validation (extensible rule system)
     #
     # Returns a VerificationReport that can be serialized to JSON/YAML

data/lib/uniword/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Uniword
-  VERSION = "1.6.1"
+  VERSION = "1.7.1"
 end

data/lib/uniword/wp_drawing/position_h.rb

@@ -18,7 +18,7 @@ module Uniword
         namespace Uniword::Ooxml::Namespaces::WordProcessingDrawing
         mixed_content
 
-        map_attribute "relative-from", to: :relative_from
+        map_attribute "relativeFrom", to: :relative_from
         map_element "align", to: :align, render_nil: false
         map_element "posOffset", to: :pos_offset, render_nil: false
       end

data/lib/uniword/wp_drawing/position_v.rb

@@ -18,7 +18,7 @@ module Uniword
         namespace Uniword::Ooxml::Namespaces::WordProcessingDrawing
         mixed_content
 
-        map_attribute "relative-from", to: :relative_from
+        map_attribute "relativeFrom", to: :relative_from
         map_element "align", to: :align, render_nil: false
         map_element "posOffset", to: :pos_offset, render_nil: false
       end

data/lib/uniword/wp_drawing/size_rel_h.rb

@@ -17,7 +17,7 @@ module Uniword
         namespace Uniword::Ooxml::Namespaces::WordProcessingDrawing
         mixed_content
 
-        map_attribute "relative-from", to: :relative_from
+        map_attribute "relativeFrom", to: :relative_from
         map_element "pctWidth", to: :pct_width, render_nil: false
       end
     end

data/lib/uniword/wp_drawing/size_rel_v.rb

@@ -17,7 +17,7 @@ module Uniword
         namespace Uniword::Ooxml::Namespaces::WordProcessingDrawing
         mixed_content
 
-        map_attribute "relative-from", to: :relative_from
+        map_attribute "relativeFrom", to: :relative_from
         map_element "pctHeight", to: :pct_height, render_nil: false
       end
     end

data/lib/uniword.rb

@@ -3,10 +3,10 @@
 require "lutaml/model"
 require "zip"
 
-# XML hydration routes through lutaml-model's Moxml layer with the
-# runtime default adapter (Leptris). Nokogiri remains available as a
-# Moxml adapter and is used directly for XSD schema validation and
-# HTML handling.
+# XML hydration, XSD validation, and HTML parsing all route through
+# lutaml-model's Moxml layer with the runtime default adapter
+# (Leptris). Nokogiri remains available as an alternative Moxml
+# adapter for development and testing.
 
 # Uniword is a comprehensive Ruby library for reading and writing Microsoft Word
 # documents in DOCX format using a schema-driven architecture.
@@ -378,10 +378,6 @@ module Uniword
       doc
     end
 
-    # Convert HTML to a document with paragraphs.
-    # Alias for {#from_html} for backward compatibility.
-    alias html_to_doc from_html
-
     # Convert HTML directly to a DOCX file
     #
     # @param html [String] HTML content

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: uniword
 version: !ruby/object:Gem::Version
-  version: 1.6.1
+  version: 1.7.1
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-09-26 00:00:00.000000000 Z
+date: 2026-10-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: logger
@@ -25,33 +25,33 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: lutaml-model
+  name: leptris
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: 1.9.368
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.8'
+        version: 1.9.368
 - !ruby/object:Gem::Dependency
-  name: nokogiri
+  name: lutaml-model
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.19'
+        version: '0.8'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '1.19'
+        version: '0.8'
 - !ruby/object:Gem::Dependency
   name: omml
   requirement: !ruby/object:Gem::Requirement
@@ -1072,7 +1072,6 @@ files:
 - lib/uniword/infrastructure.rb
 - lib/uniword/infrastructure/mime_packager.rb
 - lib/uniword/infrastructure/mime_parser.rb
-- lib/uniword/infrastructure/xml_namespace_normalizer.rb
 - lib/uniword/infrastructure/zip_extractor.rb
 - lib/uniword/infrastructure/zip_packager.rb
 - lib/uniword/lazy_loader.rb

data/lib/uniword/infrastructure/xml_namespace_normalizer.rb

@@ -1,107 +0,0 @@
-# frozen_string_literal: true
-
-require "nokogiri"
-
-module Uniword
-  module Infrastructure
-    # XML normalization utilities for round-trip fidelity
-    #
-    # Ensures that serialized XML matches the expected namespace declaration format
-    # by adding missing prefixed namespace declarations to root elements.
-    module XmlNamespaceNormalizer
-      # Prefixed namespaces that should be declared at root level in OOXML documents
-      PREFIXED_NAMESPACES = {
-        "http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing" => "wp",
-        "http://schemas.microsoft.com/office/word/2010/wordprocessingDrawing" => "wp14",
-      }.freeze
-
-      # Normalize XML to ensure prefixed namespace declarations are at root level
-      #
-      # @param xml [String] XML content
-      # @return [String] Normalized XML content
-      def self.normalize(xml)
-        doc = Nokogiri::XML(xml)
-        return xml unless doc.root
-
-        # Check if any prefixed namespaces are used but not declared at root
-        root = doc.root
-        root_ns = root.namespace_definitions.to_h { |ns| [ns.prefix, ns.href] }
-
-        PREFIXED_NAMESPACES.each do |uri, expected_prefix|
-          # Check if this namespace URI is used in the document
-          elements_with_ns = doc.xpath("//*[namespace-uri()='#{uri}']")
-          next if elements_with_ns.empty?
-
-          # Check if the expected prefix is already declared at root
-          if root_ns.key?(expected_prefix) && (root_ns[expected_prefix] == uri)
-            # Check if the URI matches
-            next
-          end
-
-          # Check if default namespace with this URI exists
-          default_ns = root_ns.find { |k, v| k.nil? && v == uri }
-
-          if default_ns
-            # Replace default namespace with prefixed namespace
-            # Find the default xmlns attribute and rename it
-            root.attributes.each do |name, attr|
-              next unless name == "xmlns" && attr.value == uri
-
-              # Remove the default namespace
-              root.remove_attribute("xmlns")
-              # Add the prefixed namespace
-              root["xmlns:#{expected_prefix}"] = uri
-
-              # Update all elements that use this namespace to use the prefix
-              update_elements_namespace(elements_with_ns, expected_prefix)
-              break
-            end
-          else
-            # Namespace is used but not declared at root - add it
-            root["xmlns:#{expected_prefix}"] = uri
-          end
-        end
-
-        doc.to_xml(indent: 0, save_with: Nokogiri::XML::Node::SaveOptions::AS_XML)
-      end
-
-      # Update elements to use the correct namespace prefix
-      #
-      # @param elements [Nokogiri::XML::NodeSet] Elements to update
-      # @param prefix [String] The prefix to use
-      def self.update_elements_namespace(elements, prefix)
-        elements.each do |el|
-          # Update the element's namespace
-          ns = el.document.root.namespaces.find do |_k, v|
-            v == el.namespace.href
-          end
-          next unless ns
-
-          # Change namespace key from default to prefixed
-          el.document.root.namespaces.delete(ns[0])
-          el.document.root["xmlns:#{prefix}"] = ns[1]
-        end
-      end
-
-      # Normalize namespace declarations in a ZIP content hash
-      #
-      # @param zip_content [Hash] File path => content mapping
-      # @param target_files [Array<String>] Files to normalize
-      # @return [Hash] Updated zip content
-      def self.normalize_zip_content(zip_content, target_files = nil)
-        result = zip_content.dup
-        target_files ||= result.keys.select do |k|
-          k.end_with?(".xml", ".rels")
-        end
-
-        target_files.each do |path|
-          next unless result[path]
-
-          result[path] = normalize(result[path])
-        end
-
-        result
-      end
-    end
-  end
-end