Class: Uniword::Docx::PackageIntegrityChecker

Inherits:
Object
  • Object
show all
Defined in:
lib/uniword/docx/package_integrity_checker.rb

Overview

Write-time OPC package integrity gate.

Pure, non-mutating checker invoked on the in-memory ZIP content hash after reconciliation and relationship injection, before ZipPackager. MECE with the mutating Reconciler: the Reconciler repairs what it can, this checker refuses output that would produce an invalid package.

Issue codes mirror the post-hoc OpcValidator semantics (lib/uniword/validation/opc_validator.rb) where they match:

  • OPC-005: every ZIP entry has a content type (Default or Override)
  • OPC-006: every relationship target resolves to a package entry
  • OPC-008: every emitted XML part is well-formed Codes introduced by the write-time gate:
  • OPC-009: every r:id/r:embed/r:link reference in any XML part resolves to a Relationship in that part's .rels
  • OPC-010: relationship IDs are unique within every .rels part

Examples:

Check an in-memory content hash

issues = PackageIntegrityChecker.new.check(zip_content)
issues.each { |i| puts "#{i.code} [#{i.part}] #{i.message}" }

Constant Summary collapse

RELS_NS =

Package relationships part namespace (OPC).

"http://schemas.openxmlformats.org/package/2006/relationships"
OFFICE_RELS_NS =

Office document relationship attribute namespace (r:id, r:embed...).

"http://schemas.openxmlformats.org/officeDocument/2006/relationships"
CT_NS =

[Content_Types].xml namespace.

"http://schemas.openxmlformats.org/package/2006/content-types"
R_REF_ATTRIBUTES =

Relationship attribute local names that reference rIds.

%w[id embed link].freeze

Instance Method Summary collapse

Instance Method Details

#check(content) ⇒ Array<Validation::Report::ValidationIssue>

Check an in-memory package content hash against every invariant.

One issue-collector pass; never mutates the content hash.

Parameters:

  • content (Hash{String => String}) —

    ZIP entry path => content

Returns:



49
50
51
52
53
54
55
56
57
58
59
60
61
 
# File 'lib/uniword/docx/package_integrity_checker.rb', line 49

def check(content)
  @content = content
  @issues = []
  @parsed_parts = {}

  parse_all_parts # OPC-008, memoizes documents for later checks
  check_content_type_coverage # OPC-005
  check_relationship_targets # OPC-006
  check_relationship_references # OPC-009
  check_rid_uniqueness # OPC-010

  @issues
end