uniword 1.2.5 → 1.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d8cdb6f45c08ae845ce242ca0a0140e7784413d9fe1b552b56eefd8040f9b03c
-  data.tar.gz: 6603afb7881e218ed635e77349cff132817086ec1399bd2a5157ce487907165f
+  metadata.gz: c28f3b26ea127104f901589ebdb8752163be6f9e963c11eca94a26c86bdbd0df
+  data.tar.gz: d8fbe7de4f8f005d78d10acd4f6c1f7459d06385b419c487732e0398470d2c23
 SHA512:
-  metadata.gz: c4c71185ce70622298500cb9512e6e0ab5f5b4509783f3273cfe8ac8828ec44159b875d9bf06b3d8f0ac346791f5ce779050e3030858d2320477ee0e89d6c7ec
-  data.tar.gz: 849da67a5a82968042f85239b1b2a536f7f85056e657f9e8b7c0b3487141e956b8903232efa1478dc476174d51e11f7580557b8ee5bd0d8ac93a88eeb8fe723b
+  metadata.gz: 807f859be7a5b6b45512ea946e4a4b1bd12a8e9df975e9dd9e06dd75c0842d916e5f811e9ac10062873426402b0a3526fa5301a0634e15062fb3f65696f3bd03
+  data.tar.gz: 3d6d287a816d2c2cc42a59b5da669dc92026774b0a2e27ff02e33a916367a63336e390338e556d1688b6cb7bbf78676dd5f200b47e8fdcd5c1509e0ebd238f90

data/CHANGELOG.md

@@ -8,6 +8,67 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.3.0] - 2026-07-19
+
+### Added
+
+#### Comment Authoring and Round-Trip Preservation
+- `DocumentBuilder#comment` now anchors the comment around the preceding
+  body paragraph (`w:commentRangeStart`/`w:commentRangeEnd` plus a
+  `w:commentReference` run styled with the `CommentReference` character
+  style, auto-registered in styles.xml) in addition to registering it in
+  the document's comments collection
+- Saving a document with comments emits `word/comments.xml` with its
+  `[Content_Types].xml` override and `document.xml.rels` relationship
+  (part metadata registered in `Ooxml::PartRegistry` under `:comments`);
+  loading parses the part into the document's comments collection, so
+  comments (ids, authors, dates, text) and their body anchors survive
+  load → save round-trips
+- `Builder::CommentAnchorer`: anchor placement for fresh and parsed
+  paragraphs (nested anchors on the same paragraph match Word's output)
+- `Wordprocessingml::Run` maps `w:commentReference`
+- `CommentsPart` is the single comments collection for both
+  `DocumentBuilder` and `Review::ReviewManager`; it gains Array-style
+  access (`size`/`[]`/`each`) and collision-safe sequential ID
+  assignment (OOXML `ST_DecimalNumber`)
+
+#### Validation Engine Consolidation
+- `Validation::Engine`: single validation engine running the rules of
+  `Rules::Registry` against a validation context, with two front-ends
+  sharing one result model (`Report::ValidationIssue`)
+- `Rules::ModelContext`: in-memory validation context wrapping a
+  `Wordprocessingml::DocumentRoot` (pre-save invocation)
+- `Rules::ModelRule`: base class for in-memory model-level rules;
+  `Rules::Base#context_type` (`:package`/`:model`) selects the rule set
+- Model-level rules: DOC-200 (document requires body), DOC-201 (bookmark
+  pairing), DOC-202 (bookmark name uniqueness), DOC-203 (empty
+  paragraphs), DOC-204 (`tbl` requires `tblGrid` per wml.xsd CT_Tbl),
+  DOC-205 (`tbl` requires `tblPr` per wml.xsd CT_Tbl)
+
+### Changed
+
+- `DocumentRoot#valid?`, `#validation_errors`, `#validation_warnings` are
+  now powered by `Validation::Engine` over the model-level rules instead
+  of the deleted `Validation::StructuralValidator`
+- `uniword validate` runs the same engine on the in-memory model, prints
+  rule-coded issues, and exits 1 on error-severity issues (aligned with
+  `uniword verify`); it previously always exited 0
+- `Validators::DocumentSemanticsValidator` (verify layer 3) runs through
+  `Validation::Engine` like every other front-end
+
+### Removed
+
+- Dead validation code (no production callers):
+  `Validation::DocumentValidator` (7-layer pipeline),
+  `Validation::StructuralValidator`, the `validation/validators/` layer
+  validators unused by `VerifyOrchestrator` (`FileStructureValidator`,
+  `ZipIntegrityValidator`, `OoxmlPartValidator`, `RelationshipValidator`,
+  `ContentTypeValidator`), the `Uniword::Validators` stub namespace
+  (`ElementValidator`/`ParagraphValidator`/`TableValidator`), the
+  `Uniword::Warnings` module (`Warning`/`WarningCollector`/
+  `WarningReport`), and the orphaned `config/validation_rules.yml` and
+  `config/warning_rules.yml`
+
 ### Added
 
 #### Open-Source Resource System (April 2025)
@@ -41,8 +102,140 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Injects missing separator (id=-1) and continuation (id=0) entries
 - Runs during `Docx::Package#to_zip_content` before serialization
 
+#### Write-Time Validation and Reconciler Transparency (July 2026)
+- `Docx::PackageIntegrityChecker`: write-time OPC integrity gate invoked on
+  every save after reconciliation, before packaging; refuses invalid output
+  (missing content types, unresolvable relationship targets, dangling
+  `r:id`/`r:embed`/`r:link` references, duplicate rIds, malformed XML)
+- Saves raise `Uniword::ValidationError` with structured `issues` (code,
+  part, message) when the gate rejects the package
+- Escape hatch: `save(path, validate: false)` threaded through
+  `DocumentRoot#save`/`#to_file`, `DocumentWriter`, `Docx::Package.to_file`
+  and `Ooxml::DotxPackage.to_file`; default from
+  `Uniword.configuration.validate_on_save` (default `true`)
+- `Uniword.configuration.log_save_fixes` (default `true`): log each repair
+  the Reconciler applies during save
+- `Docx::Package#applied_fixes`: reconciliation report (Fix value objects
+  with code, message, part) exposed after save
+- Reconciler now repairs dangling `r:embed` image references by removing
+  the drawing, and strips relationships whose target part the package does
+  not carry (e.g. unmodelled parts dropped on load) — recorded as fixes
+- Allocator (builder) and legacy (template) save paths now apply identical
+  referential repairs (previously the builder path only logged warnings)
+- `SchemaRegistry` now maps `word/webSettings.xml` and the docProps
+  core/app/custom parts to their XSDs; `docProps/core.xml` validates
+  offline (Dublin Core XSD imports bundled with relative schemaLocations)
+- Integration spec now XSD-validates real library output end-to-end
+  (Builder corpus + fixture round-trips) via `uniword verify --xsd`
+  semantics
+- `uniword theme fonts INPUT OUTPUT --name X` and `uniword theme colors
+  INPUT OUTPUT --name Y`: apply a bundled font scheme (25) or color
+  scheme (23) to an existing document — Word's Design → Fonts / Colors
+  galleries as CLI commands, replacing only the targeted theme element
+  (`--list` shows available schemes); also available as
+  `DocumentRoot#apply_font_scheme` / `#apply_color_scheme`
+- `uniword fonts replace INPUT OUTPUT --from X --to Y`: Word's
+  Replace Fonts dialog as a one-shot CLI command — rewrites rFonts
+  references across styles and defaults, body content, headers/footers,
+  notes, comments, and numbering (also `DocumentRoot#replace_font`)
+- `w:updateFields` support: `uniword toc insert` / `toc update` now set
+  it by default (`--no-update-fields` to opt out), so Word refreshes
+  the TOC and all fields when the document is opened — no manual F9
+- Comments can now target an arbitrary paragraph:
+  `doc.comment(author:, text:, on: paragraph_or_index)` — anchor around
+  any existing paragraph object or body index, including paragraphs
+  inside table cells (default remains the last paragraph)
+- `uniword page setup INPUT OUTPUT`: Word's Layout dialog as a CLI
+  command — paper size presets (letter/legal/a4/a5/executive),
+  orientation with Word-style dimension swap, and margins in
+  in/cm/mm/twips units, applied to every section (also
+  `DocumentRoot#apply_page_setup`)
+- `uniword styles list FILE [--type T] [--verbose]`: Word's Styles
+  pane as a terminal listing — style id, type, base style, and key
+  formatting details
+- `uniword styles remove FILE OUTPUT [--id X | --unused] [--dry-run]`:
+  Styles-pane management as a one-shot command — delete a style by id,
+  or declutter every style that no content references (directly or via
+  basedOn/link/next chains); default styles are always kept (also
+  `DocumentRoot#remove_style` / `#remove_unused_styles`)
+- `uniword styles rename FILE OUTPUT --id X --name Y`: Word's style
+  rename — updates the display name while the styleId (and every
+  reference to it) stays intact (also `DocumentRoot#rename_style`)
+- `uniword repair INPUT OUTPUT`: load a document, run the save-time
+  Reconciler over it, and write a repaired copy — dangling references
+  stripped, missing parts rebuilt — with every repair reported by code
+  (substantive repairs vs routine normalization); exits non-zero when
+  the write-time gate rejects what the Reconciler cannot fix
+
+### Fixed
+
+- `w:lvl` no longer emits `w:ind`/`w:tabs` as direct children (invalid per
+  ECMA-376 CT_Lvl); numbering level indentation is now written via the
+  level's `w:pPr` — previously every Builder-generated list produced
+  schema-invalid `numbering.xml`
+- Dangling image drawings and their relationships are now removed in a
+  single reconcile pass: a drawing whose `r:embed` rel exists but points
+  at a part the package does not carry is treated as dangling (previously
+  the first pass kept the drawing and stripped the rel, leaving a new
+  dangling reference for the next save)
+- `word/_rels/settings.xml.rels`, `word/_rels/footnotes.xml.rels` and
+  `word/_rels/endnotes.xml.rels` now survive load → save round-trips on
+  both save paths (previously dropped, leaving dangling `r:id`
+  references in the saved package)
+- Saving no longer crashes with FrozenError when the Reconciler injects
+  missing footnote/endnote separator entries into a parsed document
+- Theme application no longer produces invalid `theme1.xml`:
+  `Drawingml::Theme#dup` silently dropped the format scheme (fill/line/
+  effect lists), so every `theme apply` emitted a `fillStyleLst` Word
+  rejects as "unreadable content"; the save-time fill repair also used
+  a wrong attribute kwarg (`scheme_color:` for `scheme_clr:`), creating
+  color-less fills. Both fixed — applied themes now carry the complete
+  format scheme. Same family fixed in `FontScheme#dup` (per-script
+  font entries were dropped) and `ColorScheme#dup` (system-color
+  entries like windowText were lost)
+- Tables whose `gridCol` columns lack widths now get even shares of the
+  section content width (page width minus margins) at save time —
+  matching Word's fallback — instead of emitting width-less `gridCol`
+  elements and warning; explicit widths are preserved and the remainder
+  is shared
+- `uniword verify`-style integration spec hang: `soffice --view` example
+  no longer blocks the LibreOffice integration spec indefinitely
+- Header/footer dual path: adding a header/footer via the Builder to a
+  loaded document no longer drops the new part, duplicates
+  relationships or content-type overrides, or points two sectPr
+  references at the same rId. Round-trips that previously re-added a
+  stale relationship with the original rId (duplicate target rels, e.g.
+  on the APA paper template) now emit exactly one rel per part
+- `uniword headers list` now reports headers/footers of loaded
+  documents, and `headers add-header` / watermark managers no longer
+  crash when saving (both used to bypass the round-trip storage)
+- rId stability through round-trip: `Docx::IdAllocator` is now the single
+  rId authority for all relationships parts. Loaded documents round-trip
+  with their original rIds preserved verbatim (previously the reconciler
+  renumbered every rId on save, e.g. `r:embed="rId8"` becoming
+  `"rId9"`); only genuinely new relationships (builder-added images,
+  hyperlinks, charts, headers/footers) receive fresh allocations.
+  The 3 long-skipped `docx_roundtrip_spec.rb` round-trip examples
+  (APA template, two ISO fixtures) now pass unskipped
+- Loaded image parts are now keyed by their actual document relationship
+  rId, so `r:embed` references resolve correctly in MHTML conversion and
+  the image manager; theme-only media no longer gains a spurious
+  document-level relationship on save
+- Adding the same image file twice now stores one part with one
+  relationship shared by both drawings (previously produced a dangling
+  duplicate rId reference)
+
 ### Changed
 
+- `Uniword::Ooxml::Relationships::PackageRelationships.next_available_rid`
+  removed: all rId assignment flows through `Docx::IdAllocator` (scoped
+  per rels part — `:document`/`:package` — with `seed_from_rels`
+  preserving loaded rIds verbatim). The reconciler's legacy renumbering
+  path, the serializer-side `next_rid`, and `ChartBuilder`'s literal
+  `rIdChartN` ids are gone; chart relationship ids are now allocator
+  numerics. The reconciler's light-touch vs full normalization choice is
+  now explicit (`builder_managed`) instead of inferred from allocator
+  presence
 - `Uniword::Ooxml::DocxPackage` renamed to `Uniword::Docx::Package`
   - All references in lib/ and spec/ updated (~20 files)
   - `lib/uniword/ooxml/docx_package.rb` moved to `lib/uniword/docx/package.rb`
@@ -52,6 +245,36 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - All 12 stylesets reauthored with OFL-licensed fonts (Calibri references fixed)
 - `uniword.gemspec` updated to include `data/**/*.yml` and `data/**/*.xsd` in gem
 - Added `rainbow ~> 3.1` gem dependency
+- Single `Uniword::Ooxml::PartRegistry` now owns the part↔content-type↔
+  relationship-type mapping for every DOCX part kind the library writes
+  - New `Ooxml::PartDefinition` value class (key, path/pattern, content
+    type, relationship type, extension, required, default-vs-override)
+  - `ContentTypes.generate`, `Docx::PackageDefaults.minimal_*`,
+    `Reconciler::PackageStructure`, and the `PackageSerialization`
+    `inject_*` methods derive from the registry instead of holding
+    string literals; `IdAllocator` and builder rel-type constants also
+    derive from it
+  - Open/closed: register a `PartDefinition` to add a part kind —
+    consumers need no changes
+  - No behavior change: saved packages are byte-identical
+- Package-held parts are now model objects instead of raw hashes
+  - New `Uniword::Docx::Part` value object (definition, rId, target,
+    content, verbatim rel-type/content-type overrides) with `ChartPart`
+    and `HeaderFooterPart` subclasses and `CustomXmlItem`;
+    `Docx::PartCollection` backs `chart_parts` (keyed by rId) and
+    `embeddings` (keyed by target)
+  - Single header/footer storage path: `DocumentRoot#header_footer_parts`
+    is a `Docx::HeaderFooterPartCollection` fed identically by the loader
+    (original rIds/targets, sectPr-derived types) and the Builder;
+    `document.headers`/`document.footers` are delegating
+    `Docx::HeaderFooterView`s over that store (Hash-style by sectPr type,
+    Array-style over parts) — public API preserved
+  - Single wiring implementation: the Reconciler wires fresh
+    header/footer parts into `document.xml.rels` and sectPr
+    (`IdAllocator` on the builder path, `find_or_create_rel` on the
+    legacy path); the serializer's duplicate rel/sectPr wiring and its
+    divergent rId strategy are deleted — it now only emits one part
+    file and one content-type override per store entry
 
 #### StylesetPackage Implementation (December 4, 2024)
 - **StylesetPackage**: Proper MODEL-DRIVEN package for .dotx files

data/CONTRIBUTING.md

@@ -226,7 +226,7 @@ After adding any class (autoload or require_relative):
 **Good (autoload)**:
 ```ruby
 # New utility class with no dependencies
-autoload :DocumentValidator, 'uniword/document_validator'
+autoload :DocumentFactory, 'uniword/document_factory'
 ```
 
 **Good (require_relative with documentation)**:

data/config/ooxml/schemas/shared_types.yml

@@ -46,7 +46,7 @@ elements:
     description: 'Hexadecimal color value (RGB)'
     attributes:
       - name: val
-        type: :string
+        type: Uniword::Ooxml::Types::HexColorValue
         xml_name: val
         xml_attribute: true
         required: true
@@ -57,7 +57,7 @@ elements:
     description: 'Measurement in twips (1/20th of a point)'
     attributes:
       - name: val
-        type: :integer
+        type: Uniword::Ooxml::Types::UnsignedDecimalNumber
         xml_name: val
         xml_attribute: true
         required: true
@@ -79,7 +79,7 @@ elements:
     description: 'Measurement in points'
     attributes:
       - name: val
-        type: :integer
+        type: Uniword::Ooxml::Types::UnsignedDecimalNumber
         xml_name: val
         xml_attribute: true
         required: true
@@ -90,7 +90,7 @@ elements:
     description: 'Measurement in pixels'
     attributes:
       - name: val
-        type: :integer
+        type: Uniword::Ooxml::Types::UnsignedDecimalNumber
         xml_name: val
         xml_attribute: true
         required: true
@@ -135,6 +135,7 @@ elements:
     attributes:
       - name: val
         type: :string
+        values: [top, center, baseline, bottom, auto]
         xml_name: val
         xml_attribute: true
         required: true

data/data/schemas/ecma/dc.xsd

@@ -0,0 +1,118 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
+           xmlns="http://purl.org/dc/elements/1.1/"
+           targetNamespace="http://purl.org/dc/elements/1.1/"
+           elementFormDefault="qualified"
+           attributeFormDefault="unqualified">
+
+  <xs:annotation>
+    <xs:documentation xml:lang="en">
+      DCMES 1.1 XML Schema
+      XML Schema for http://purl.org/dc/elements/1.1/ namespace
+
+      Created 2003-04-02
+
+      Created by 
+
+      Tim Cole (t-cole3@uiuc.edu)
+      Tom Habing (thabing@uiuc.edu)
+      Jane Hunter (jane@dstc.edu.au)
+      Pete Johnston (p.johnston@ukoln.ac.uk),
+      Carl Lagoze (lagoze@cs.cornell.edu)
+
+      This schema declares XML elements for the 15 DC elements from the
+      http://purl.org/dc/elements/1.1/ namespace.
+
+      It defines a complexType SimpleLiteral which permits mixed content 
+      and makes the xml:lang attribute available. It disallows child elements by
+      use of minOcccurs/maxOccurs.
+
+      However, this complexType does permit the derivation of other complexTypes
+      which would permit child elements.
+
+      All elements are declared as substitutable for the abstract element any, 
+      which means that the default type for all elements is dc:SimpleLiteral.
+
+    </xs:documentation>
+
+  </xs:annotation>
+
+
+  <xs:import namespace="http://www.w3.org/XML/1998/namespace"
+             schemaLocation="xml.xsd">
+  </xs:import>
+
+  <xs:complexType name="SimpleLiteral">
+        <xs:annotation>
+        <xs:documentation xml:lang="en">
+            This is the default type for all of the DC elements.
+            It permits text content only with optional
+            xml:lang attribute.
+            Text is allowed because mixed="true", but sub-elements
+            are disallowed because minOccurs="0" and maxOccurs="0" 
+            are on the xs:any tag.
+
+    	    This complexType allows for restriction or extension permitting
+            child elements.
+    	</xs:documentation>
+  	</xs:annotation>
+
+   <xs:complexContent mixed="true">
+    <xs:restriction base="xs:anyType">
+     <xs:sequence>
+      <xs:any processContents="lax" minOccurs="0" maxOccurs="0"/>
+     </xs:sequence>
+     <xs:attribute ref="xml:lang" use="optional"/>
+    </xs:restriction>
+   </xs:complexContent>
+  </xs:complexType>
+
+  <xs:element name="any" type="SimpleLiteral" abstract="true"/>
+
+  <xs:element name="title" substitutionGroup="any"/>
+  <xs:element name="creator" substitutionGroup="any"/>
+  <xs:element name="subject" substitutionGroup="any"/>
+  <xs:element name="description" substitutionGroup="any"/>
+  <xs:element name="publisher" substitutionGroup="any"/>
+  <xs:element name="contributor" substitutionGroup="any"/>
+  <xs:element name="date" substitutionGroup="any"/>
+  <xs:element name="type" substitutionGroup="any"/>
+  <xs:element name="format" substitutionGroup="any"/>
+  <xs:element name="identifier" substitutionGroup="any"/>
+  <xs:element name="source" substitutionGroup="any"/>
+  <xs:element name="language" substitutionGroup="any"/>
+  <xs:element name="relation" substitutionGroup="any"/>
+  <xs:element name="coverage" substitutionGroup="any"/>
+  <xs:element name="rights" substitutionGroup="any"/>
+
+  <xs:group name="elementsGroup">
+  	<xs:annotation>
+    	<xs:documentation xml:lang="en">
+    	    This group is included as a convenience for schema authors
+            who need to refer to all the elements in the 
+            http://purl.org/dc/elements/1.1/ namespace.
+    	</xs:documentation>
+  	</xs:annotation>
+
+  <xs:sequence>
+    <xs:choice minOccurs="0" maxOccurs="unbounded">
+      <xs:element ref="any"/>
+    </xs:choice>
+    </xs:sequence>
+  </xs:group>
+
+  <xs:complexType name="elementContainer">
+  	<xs:annotation>
+    	<xs:documentation xml:lang="en">
+    		This complexType is included as a convenience for schema authors who need to define a root
+    		or container element for all of the DC elements.
+    	</xs:documentation>
+  	</xs:annotation>
+
+    <xs:choice>
+      <xs:group ref="elementsGroup"/>
+    </xs:choice>
+  </xs:complexType>
+
+
+</xs:schema>

data/data/schemas/ecma/dcmitype.xsd

@@ -0,0 +1,50 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
+           xmlns="http://purl.org/dc/dcmitype/"
+           targetNamespace="http://purl.org/dc/dcmitype/"
+           elementFormDefault="qualified"
+           attributeFormDefault="unqualified">
+
+  <xs:annotation>
+    <xs:documentation xml:lang="en">
+      DCMI Type Vocabulary XML Schema
+      XML Schema for http://purl.org/dc/dcmitype/ namespace
+
+      Created 2003-04-02
+
+      Created by 
+
+      Tim Cole (t-cole3@uiuc.edu)
+      Tom Habing (thabing@uiuc.edu)
+      Jane Hunter (jane@dstc.edu.au)
+      Pete Johnston (p.johnston@ukoln.ac.uk),
+      Carl Lagoze (lagoze@cs.cornell.edu)
+
+      This schema defines a simpleType which enumerates
+      the allowable values for the DCMI Type Vocabulary.
+    </xs:documentation>
+
+ 
+  </xs:annotation>
+
+
+  <xs:simpleType name="DCMIType">
+     <xs:union>
+        <xs:simpleType>
+           <xs:restriction base="xs:Name">
+		<xs:enumeration value="Collection"/>
+		<xs:enumeration value="Dataset"/>
+		<xs:enumeration value="Event"/>
+		<xs:enumeration value="Image"/>
+		<xs:enumeration value="InteractiveResource"/>
+		<xs:enumeration value="Service"/>
+		<xs:enumeration value="Software"/>
+		<xs:enumeration value="Sound"/>
+		<xs:enumeration value="Text"/>
+		<xs:enumeration value="PhysicalObject"/>
+            </xs:restriction>
+        </xs:simpleType> 
+     </xs:union>
+  </xs:simpleType>
+
+</xs:schema>