Class: Lutaml::Xml::Schema::XsdSchema

Inherits:
Object
  • Object
show all
Extended by:
Model::Schema::SharedMethods
Includes:
Model::Schema::SharedMethods
Defined in:
lib/lutaml/xml/schema/xsd_schema.rb

Overview

XSD Schema generation for XML models

Generates W3C XML Schema (XSD) from LutaML model classes. Supports namespace declarations, type definitions, and nested models.

Defined Under Namespace

Classes: Context

Constant Summary collapse

UNSUPPORTED_XSD_REGEX_FLAGS = 
Regexp::IGNORECASE | Regexp::MULTILINE | Regexp::EXTENDED

Class Method Summary collapse

Methods included from Model::Schema::SharedMethods

extract_register_from

Class Method Details

.attr_is_xml_attribute?(xml_mapping, attr_name) ⇒ Boolean

Returns:

  • (Boolean) 


798
799
800
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 798

def self.attr_is_xml_attribute?(xml_mapping, attr_name)
  xml_mapping.attributes.any? { |rule| rule.to == attr_name }
end

.build_element_attributes(name, xsd_type, attr, xml_mapping, attr_name) ⇒ Object 



802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 802

def self.build_element_attributes(name, xsd_type, attr, xml_mapping,
attr_name)
  attrs = { name: name.to_s, type: xsd_type }

  # Handle collection cardinality
  if attr.collection?
    range = attr.resolved_collection
    if range
      attrs[:minOccurs] = range.min.to_s
      attrs[:maxOccurs] =
        range.end.infinite? ? "unbounded" : range.max.to_s
    else
      attrs[:minOccurs] = "0"
      attrs[:maxOccurs] = "unbounded"
    end
  end

  # Add form attribute from mapping rule if present
  if xml_mapping
    rule = xml_mapping.find_element(attr_name)
    attrs[:form] = rule.form.to_s if rule&.form
    attrs[:annotation] = rule.documentation if rule&.documentation
  end

  attrs
end

.classify_xsd_type(type_name, klass, register) ⇒ Symbol

Classify an XSD type name into one of three categories

Parameters:

  • type_name (String) —

    The XSD type name to classify

  • klass (Class) —

    The model class being processed

  • register (Register) —

    The register for type resolution

Returns:

  • (Symbol) —

    :builtin, :custom, :unresolvable, or :unknown



114
115
116
117
118
119
120
121
122
123
124
125
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 114

def self.classify_xsd_type(type_name, klass, register)
  return :builtin if BuiltinTypes.builtin?(type_name)

  # Custom type - check if resolvable
  if type_name && !type_name.start_with?("xs:")
    return :custom if type_resolvable?(type_name, klass, register)

    return :unresolvable
  end

  :unknown
end

.collection_placeholder?(attr, mapping) ⇒ Boolean

Whether a nested model is emitted only as a placeholder: a collection of a type_name-less model renders as <element name="item" type="xs:string"/>, so the model's own content is never inlined or referenced. This document therefore neither declares its namespaces nor defines its nested types — the walkers must not descend into it.

Returns:

  • (Boolean) 


490
491
492
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 490

def self.collection_placeholder?(attr, mapping)
  attr.collection? && mapping&.type_name_value.nil?
end

.declare_referenced_namespaces!(schema_attrs, referenced, target_uri, skip_validation) ⇒ Object

Declare xmlns: for every namespace this schema references. Foreign model namespaces are strict: an unusable prefix or a prefix collision always raises (the emitted QNames would resolve to the wrong namespace), while a missing schema_location raises unless skip_validation downgrades it to a warning (the output stays structurally correct, only the import is unresolvable). Type::Value namespaces are declared best-effort (their xsd_type references are emitted verbatim).



310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 310

def self.declare_referenced_namespaces!(schema_attrs, referenced, target_uri, skip_validation)
  referenced[:foreign_models].each do |ns_class|
    # An unusable or colliding prefix makes emitted QNames resolve to
    # the wrong namespace — never recoverable, even under
    # skip_validation.
    error = foreign_prefix_error(ns_class, schema_attrs, target_uri)
    raise Lutaml::Model::Error, error if error

    # A missing schema_location leaves the import unresolvable but the
    # output structurally correct — recoverable under skip_validation.
    unless ns_class.schema_location
      error = missing_schema_location_error(ns_class)
      raise Lutaml::Model::Error, error unless skip_validation

      warn "[Lutaml::Model] WARN: #{error} " \
           "(skip_validation: emitting best-effort output)"
    end

    schema_attrs[:"xmlns:#{ns_class.prefix_default}"] = ns_class.uri
  end

  referenced[:type_values].each do |ns_class|
    prefix = ns_class.prefix_default
    next unless usable_prefix?(prefix)

    # A prefix bound to two different namespaces emits QNames that
    # resolve to the wrong one — the same unrecoverable collision the
    # foreign path raises on, so raise here too rather than silently
    # keeping the first binding.
    error = prefix_collision_error(prefix, ns_class.uri, schema_attrs)
    raise Lutaml::Model::Error, error if error

    schema_attrs[:"xmlns:#{prefix}"] = ns_class.uri
  end
end

.effective_facets(attr, attr_type) ⇒ Object 



639
640
641
642
643
644
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 639

def self.effective_facets(attr, attr_type)
  return {} unless attr_type.is_a?(Class) &&
    attr_type < Lutaml::Model::Type::Value

  attr.effective_restriction_facets(attr_type)
end

.emit_bound(xml, element, value, type) ⇒ Object 



691
692
693
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 691

def self.emit_bound(xml, element, value, type)
  emit_facet(xml, element, lexical_value(type, value)) unless value.nil?
end

.emit_facet(xml, element, value) ⇒ Object 



695
696
697
698
699
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 695

def self.emit_facet(xml, element, value)
  return if value.nil?

  xml.public_send(qn(element), value: value.to_s)
end

.emit_import(xml, ns_class) ⇒ Object

Emit a single xs:import for a namespace class, with schemaLocation when the class declares one.



406
407
408
409
410
411
412
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 406

def self.emit_import(xml, ns_class)
  import_attrs = { namespace: ns_class.uri }
  if ns_class.schema_location
    import_attrs[:schemaLocation] = ns_class.schema_location
  end
  xs(xml, "import", import_attrs)
end

.emit_restriction(xml, base, facets, type) ⇒ Object 



657
658
659
660
661
662
663
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 657

def self.emit_restriction(xml, base, facets, type)
  xs(xml, "simpleType") do
    xs(xml, "restriction", base: base) do
      emit_restriction_facets(xml, facets, type)
    end
  end
end

.emit_restriction_facets(xml, facets, type) ⇒ Object 



665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 665

def self.emit_restriction_facets(xml, facets, type)
  emit_bound(xml, :minExclusive, facets[:min_exclusive], type)
  emit_bound(xml, :minInclusive, facets[:min_inclusive], type)
  emit_bound(xml, :maxExclusive, facets[:max_exclusive], type)
  emit_bound(xml, :maxInclusive, facets[:max_inclusive], type)
  emit_facet(xml, :totalDigits, facets[:total_digits])
  emit_facet(xml, :fractionDigits, facets[:fraction_digits])
  emit_facet(xml, :length, facets[:length])
  emit_facet(xml, :minLength, facets[:min_length])
  emit_facet(xml, :maxLength, facets[:max_length])
  Array(facets[:enumeration]).each do |value|
    emit_facet(xml, :enumeration, lexical_value(type, value))
  end
  emit_facet(xml, :whiteSpace, facets[:white_space])
  patterns = Array(facets[:pattern])
  if patterns.size > 1
    raise Lutaml::Model::Error,
          "Cannot export #{patterns.size} conjunctive patterns on " \
          "#{type} to a single xs:restriction: sibling xs:pattern " \
          "facets are alternatives (OR), and XSD cannot express " \
          "conjunction (AND) without nested restriction derivation, " \
          "which is not yet supported."
  end
  patterns.each { |re| emit_facet(xml, :pattern, pattern_value(re)) }
end

.emit_value_element(xml, attrs, xsd_type, attr_type, attr) ⇒ Object 



646
647
648
649
650
651
652
653
654
655
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 646

def self.emit_value_element(xml, attrs, xsd_type, attr_type, attr)
  facets = effective_facets(attr, attr_type)
  if facets.any?
    xs(xml, "element", attrs.except(:type)) do
      emit_restriction(xml, xsd_type, facets, attr_type)
    end
  else
    xs(xml, "element", attrs.merge(type: xsd_type))
  end
end

.emit_xsd_attribute(xml, attr, rule, xsd_type, attr_type) ⇒ Object

Emit an xs:attribute, inlining an xs:restriction when the attribute is a constrained value type (its type= moves into the restriction base) and/or an annotation when the mapping carries documentation.



618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 618

def self.emit_xsd_attribute(xml, attr, rule, xsd_type, attr_type)
  facets = effective_facets(attr, attr_type)
  restricted = facets.any?

  attr_attrs = { name: rule.name }
  attr_attrs[:type] = xsd_type unless restricted
  attr_attrs[:use] = "required" if attr.options[:required]
  attr_attrs[:form] = rule.form.to_s if rule.form

  return xs(xml, "attribute", attr_attrs) unless restricted || rule.documentation

  xs(xml, "attribute", attr_attrs) do
    if rule.documentation
      xs(xml, "annotation") do
        xs(xml, "documentation", rule.documentation)
      end
    end
    emit_restriction(xml, xsd_type, facets, attr_type) if restricted
  end
end

.foreign_namespace?(mapping, target_uri) ⇒ Boolean

Whether a nested model's mapping belongs to a namespace other than the schema's target namespace. A model with no namespace belongs to the target schema (not foreign).

Returns:

  • (Boolean) 


466
467
468
469
470
471
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 466

def self.foreign_namespace?(mapping, target_uri)
  ns_class = mapping&.namespace_class
  return false unless ns_class

  ns_class.uri != target_uri
end

.foreign_prefix_error(ns_class, schema_attrs, target_uri) ⇒ Object

Why a foreign namespace's prefix cannot be used in this schema's QNames, or nil when it can.



348
349
350
351
352
353
354
355
356
357
358
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 348

def self.foreign_prefix_error(ns_class, schema_attrs, target_uri)
  prefix = ns_class.prefix_default
  unless usable_prefix?(prefix)
    schema_desc = target_uri ? "the '#{target_uri}'" : "this no-namespace"
    return "XSD generation: foreign namespace '#{ns_class.uri}' " \
           "needs a usable prefix_default (not nil/empty/'xs') to " \
           "be referenced from #{schema_desc} schema."
  end

  prefix_collision_error(prefix, ns_class.uri, schema_attrs)
end

.generate(klass, options = {}) ⇒ Object 



88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 88

def self.generate(klass, options = {})
  register = extract_register_from(klass)
  xml_mapping = klass.mappings_for(:xml)

  # Validate XSD types unless explicitly skipped
  validate_xsd_types!(klass, register) unless options[:skip_validation]

  # Use Builder with adapter from options or config
  adapter_type = options[:adapter] || Lutaml::Model::Config.xml_adapter_type || :nokogiri

  schema_builder = Builder.new(
    adapter_type: adapter_type,
    options: { encoding: "UTF-8" },
  ) do |xml|
    generate_schema(xml, klass, xml_mapping, register, options)
  end

  schema_builder.to_xml(options)
end

.generate_annotation(xml, xml_mapping) ⇒ Object 



422
423
424
425
426
427
428
429
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 422

def self.generate_annotation(xml, xml_mapping)
  xs(xml, "annotation") do
    doc_text = xml_mapping.documentation_text
    doc_text ||= xml_mapping.namespace_class&.documentation if xml_mapping.namespace_class

    xs(xml, "documentation", doc_text) if doc_text
  end
end

.generate_attributes(xml, klass, register, xml_mapping) ⇒ Object 



602
603
604
605
606
607
608
609
610
611
612
613
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 602

def self.generate_attributes(xml, klass, register, xml_mapping)
  return unless xml_mapping

  xml_mapping.attributes.each do |rule|
    attr = klass.attributes[rule.to]
    next unless attr

    attr_type = attr.type(register)
    xsd_type = get_attribute_xsd_type(attr, attr_type, register, rule)
    emit_xsd_attribute(xml, attr, rule, xsd_type, attr_type)
  end
end

.generate_complex_type(xml, klass, type_name, register, xml_mapping = nil, ctx:) ⇒ Object 



508
509
510
511
512
513
514
515
516
517
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 508

def self.generate_complex_type(xml, klass, type_name, register, xml_mapping = nil, ctx:)
  xs(xml, "complexType", { name: type_name }) do
    if klass.attributes.any?
      xs(xml, "sequence") do
        generate_elements(xml, klass, register, xml_mapping, ctx: ctx)
      end
    end
    generate_attributes(xml, klass, register, xml_mapping)
  end
end

.generate_complex_type_content(xml, klass, register, xml_mapping, ctx:) ⇒ Object 



494
495
496
497
498
499
500
501
502
503
504
505
506
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 494

def self.generate_complex_type_content(xml, klass, register, xml_mapping, ctx:)
  xs(xml, "complexType") do
    if klass.attributes.any?
      xs(xml, "sequence") do
        generate_elements(xml, klass, register, xml_mapping, ctx: ctx)
      end
    end
    if xml_mapping
      generate_attributes(xml, klass, register,
                          xml_mapping)
    end
  end
end

.generate_elements(xml, klass, register, xml_mapping, ctx:) ⇒ Object 



519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 519

def self.generate_elements(xml, klass, register, xml_mapping, ctx:)
  klass.attributes.each do |name, attr|
    next if xml_mapping && attr_is_xml_attribute?(xml_mapping, name)

    # Find the mapping rule for this attribute
    mapping_rule = xml_mapping&.find_element(name)

    attr_type = attr.type(register)

    if attr_type <= Lutaml::Model::Serialize
      # Nested model - check if it has a type_name for reference
      nested_mapping = attr_type.mappings_for(:xml)
      nested_type_name = nested_mapping&.type_name_value

      if attr.collection?
        # Collection of models
        element_attrs = { name: name.to_s }
        element_attrs[:minOccurs] = "0"
        element_attrs[:maxOccurs] = "unbounded"

        if nested_type_name
          # Reference named type by its owning namespace's prefix
          element_attrs[:type] = qualify_type_ref(
            nested_type_name,
            nested_ref_prefix(nested_mapping, ctx),
          )
          xs(xml, "element", element_attrs)
        else
          # Inline anonymous complexType
          xs(xml, "element", element_attrs) do
            xs(xml, "complexType") do
              xs(xml, "sequence") do
                xs(xml, "element",
                   { name: "item", type: get_xsd_type(attr_type) })
              end
            end
          end
        end
      elsif nested_type_name
        # Single nested model - reference by its owning namespace prefix
        xs(xml, "element",
           { name: name.to_s,
             type: qualify_type_ref(
               nested_type_name,
               nested_ref_prefix(nested_mapping, ctx),
             ) })
      else
        # Inline anonymous complexType
        xs(xml, "element", { name: name.to_s }) do
          generate_complex_type_content(xml, attr_type, register, nil,
                                        ctx: ctx)
        end
      end
    else
      # Value type
      xsd_type = get_attribute_xsd_type(attr, attr_type, register,
                                        mapping_rule)

      if attr.collection?
        # Collection of simple types
        element_attrs = { name: name.to_s }
        element_attrs[:minOccurs] = "0"
        element_attrs[:maxOccurs] = "unbounded"

        xs(xml, "element", element_attrs) do
          xs(xml, "complexType") do
            xs(xml, "sequence") do
              emit_value_element(xml, { name: "item" }, xsd_type,
                                 attr_type, attr)
            end
          end
        end
      else
        # Simple element
        element_attrs = build_element_attributes(name, xsd_type, attr,
                                                 xml_mapping, name)
        emit_value_element(xml, element_attrs, xsd_type, attr_type,
                           attr)
      end
    end
  end
end

.generate_imports(xml, namespace_class) ⇒ Object

Emit xs:import for each namespace the root's XmlNamespace class explicitly declares. Returns the Set of imported URIs so the tree import loop can dedupe against them.



391
392
393
394
395
396
397
398
399
400
401
402
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 391

def self.generate_imports(xml, namespace_class)
  imported = Set.new
  return imported unless namespace_class.imports&.any?

  namespace_class.imports.each do |imported_ns|
    next unless imported.add?(imported_ns.uri)

    emit_import(xml, imported_ns)
  end

  imported
end

.generate_includes(xml, namespace_class) ⇒ Object 



414
415
416
417
418
419
420
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 414

def self.generate_includes(xml, namespace_class)
  return unless namespace_class.includes&.any?

  namespace_class.includes.each do |schema_location|
    xs(xml, "include", { schemaLocation: schema_location })
  end
end

.generate_nested_type_definitions(xml, klass, register, ctx:, seen: nil) ⇒ Object 



431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 431

def self.generate_nested_type_definitions(xml, klass, register, ctx:, seen: nil)
  # Cycle guard, seeded with the root class (already defined by
  # generate_schema) so it is never redefined via a back-reference.
  seen ||= Set[klass]

  klass.attributes.each_value do |attr|
    attr_type = attr.type(register)
    next unless attr_type <= Lutaml::Model::Serialize

    nested_mapping = attr_type.mappings_for(:xml)

    # Skip models this document does not emit — checked before the cycle
    # guard so an unemitted reference never consumes the model's slot and
    # masks an emitted reference elsewhere. A collection of a type_name-
    # less model is a placeholder; an imported foreign type is defined in
    # its own schema document. A foreign model without a type_name is
    # inlined here, so we still descend to define the types it references.
    next if collection_placeholder?(attr, nested_mapping)
    next if imported_foreign_type?(nested_mapping, ctx.target_uri)
    next unless seen.add?(attr_type)

    nested_type_name = nested_mapping&.type_name_value
    if nested_type_name
      generate_complex_type(xml, attr_type, nested_type_name, register,
                            nested_mapping, ctx: ctx)
    end

    generate_nested_type_definitions(xml, attr_type, register,
                                     ctx: ctx, seen: seen)
  end
end

.generate_referenced_imports(xml, referenced, target_uri, imported_uris) ⇒ Object

xs:import for every referenced foreign namespace, deduped by URI against the explicit imports already emitted.



379
380
381
382
383
384
385
386
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 379

def self.generate_referenced_imports(xml, referenced, target_uri, imported_uris)
  (referenced[:foreign_models] + referenced[:type_values]).each do |ns_class|
    next if ns_class.uri == target_uri
    next unless imported_uris.add?(ns_class.uri)

    emit_import(xml, ns_class)
  end
end

.generate_schema(xml, klass, xml_mapping, register, options) ⇒ Object 



188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 188

def self.generate_schema(xml, klass, xml_mapping, register, options)
  skip_validation = options[:skip_validation]

  # Bind the XSD vocabulary namespace to the "xs" prefix (via the
  # namespace object, not a hardcoded string) so the xs:-prefixed type
  # references and the prefixed structure elements are both declared.
  schema_attrs = { xsd_ns.attr_name.to_sym => xsd_ns.uri }
  schema_attrs.merge!(target_namespace_attrs(xml_mapping))
  target_uri = schema_attrs[:targetNamespace]

  referenced = referenced_namespaces(klass, register, target_uri)

  # Named-type references into the target namespace need a usable
  # prefix; synthesise a non-colliding one when the namespace class
  # declares none.
  target_prefix = target_ns_prefix(xml_mapping)
  if target_uri && target_prefix.nil?
    target_prefix = synthesize_target_prefix(referenced)
    schema_attrs[:"xmlns:#{target_prefix}"] = target_uri
  end

  declare_referenced_namespaces!(schema_attrs, referenced, target_uri,
                                 skip_validation)

  ctx = Context.new(target_prefix, target_uri)

  xs(xml, "schema", schema_attrs) do
    # Explicit imports declared on the root's XmlNamespace class.
    imported_uris = Set.new
    if xml_mapping.namespace_class
      imported_uris = generate_imports(xml, xml_mapping.namespace_class)
      generate_includes(xml, xml_mapping.namespace_class)
    end

    generate_referenced_imports(xml, referenced, target_uri,
                                imported_uris)

    if xml_mapping.documentation_text || xml_mapping.namespace_class&.documentation
      generate_annotation(xml, xml_mapping)
    end

    element_name = if has_explicit_xml_mapping?(klass, xml_mapping)
                     xml_mapping.element_name || xml_mapping.root_element
                   end

    type_name = xml_mapping.type_name_value

    # Generate XSD based on three patterns:
    # Pattern 1: element only -> inline anonymous complexType
    # Pattern 2: type_name only -> named complexType (no element)
    # Pattern 3: both element and type_name -> element + named complexType

    if element_name && type_name
      xs(xml, "element",
         { name: element_name,
           type: qualify_type_ref(type_name, target_prefix) })
      generate_complex_type(xml, klass, type_name, register,
                            xml_mapping, ctx: ctx)
    elsif type_name && !element_name
      generate_complex_type(xml, klass, type_name, register,
                            xml_mapping, ctx: ctx)
    else
      # Use class name as fallback element name if not specified
      elem_name = element_name || klass.name
      xs(xml, "element", { name: elem_name }) do
        generate_complex_type_content(xml, klass, register, xml_mapping,
                                      ctx: ctx)
      end
    end

    generate_nested_type_definitions(xml, klass, register, ctx: ctx)
  end
end

.get_attribute_xsd_type(attr, attr_type, register, _mapping_rule = nil) ⇒ Object 



836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 836

def self.get_attribute_xsd_type(attr, attr_type, register,
_mapping_rule = nil)
  if attr.union?
    raise Lutaml::Model::UnionSchemaUnsupportedError.new(attr.name,
                                                         "XSD")
  end

  # 1. Check for deprecated attribute-level xsd_type override
  return attr.options[:xsd_type] if attr.options[:xsd_type]

  # 2. Check if type has xsd_type method (Type-level)
  if attr_type.is_a?(Class) && attr_type < Lutaml::Model::Type::Value
    # Special handling for Reference type
    if attr_type == Lutaml::Model::Type::Reference
      target_xsd_type = get_target_xsd_type(attr, register)
      return attr_type.xsd_type(target_xsd_type)
    end

    return attr_type.xsd_type
  end

  # 3. Fall back to default mapping
  get_xsd_type(attr_type)
end

.get_namespace_info(klass) ⇒ Object

Namespace information for a Type::Value class (the only kind the namespace walk queries — nested models are handled structurally by referenced_namespaces). Returns {} for anything else.



910
911
912
913
914
915
916
917
918
919
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 910

def self.get_namespace_info(klass)
  return {} unless klass.is_a?(::Class)

  if defined?(Lutaml::Model::Type::Value) &&
      klass <= Lutaml::Model::Type::Value
    return get_type_namespace_info(klass)
  end

  {}
end

.get_target_xsd_type(attr, register) ⇒ Object 



942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 942

def self.get_target_xsd_type(attr, register)
  return nil unless attr.options[:ref_model_class]
  return nil unless attr.options[:ref_key_attribute]

  begin
    model_class = Object.const_get(attr.options[:ref_model_class])
    target_attr = model_class.attributes[attr.options[:ref_key_attribute]]
    return nil unless target_attr

    target_type = target_attr.type(register)
    get_attribute_xsd_type(target_attr, target_type, register)
  rescue NameError
    nil
  end
end

.get_xsd_type(type) ⇒ Object 



958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 958

def self.get_xsd_type(type)
  {
    Lutaml::Model::Type::String => "xs:string",
    Lutaml::Model::Type::Integer => "xs:integer",
    Lutaml::Model::Type::Boolean => "xs:boolean",
    Lutaml::Model::Type::Float => "xs:float",
    Lutaml::Model::Type::Decimal => "xs:decimal",
    Lutaml::Model::Type::Date => "xs:date",
    Lutaml::Model::Type::Time => "xs:time",
    Lutaml::Model::Type::DateTime => "xs:dateTime",
    Lutaml::Model::Type::TimeWithoutDate => "xs:time",
    Lutaml::Model::Type::Duration => "xs:duration",
    Lutaml::Model::Type::Uri => "xs:anyURI",
    Lutaml::Model::Type::QName => "xs:QName",
    Lutaml::Model::Type::Base64Binary => "xs:base64Binary",
    Lutaml::Model::Type::HexBinary => "xs:hexBinary",
    Lutaml::Model::Type::Hash => "xs:anyType",
    Lutaml::Model::Type::Symbol => "xs:string",
  }[type] || "xs:string"
end

.has_explicit_xml_mapping?(klass, xml_mapping) ⇒ Boolean

Returns:

  • (Boolean) 


829
830
831
832
833
834
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 829

def self.has_explicit_xml_mapping?(klass, xml_mapping)
  return true unless xml_mapping.root_element

  base_name = Lutaml::Model::Utils.base_class_name(klass)
  xml_mapping.root_element != base_name
end

.imported_foreign_type?(mapping, target_uri) ⇒ Boolean

Whether a nested model is referenced across a schema-document boundary: it lives in a foreign namespace AND exposes a named type, so it is referenced by a prefixed QName and resolved through an xs:import. A foreign model WITHOUT a type_name has no named type to reference, so generate_elements inlines it into this document — it is not a boundary, and both the namespace walk and the type-definition walk must descend into it rather than treat it as imported.

Returns:

  • (Boolean) 


480
481
482
483
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 480

def self.imported_foreign_type?(mapping, target_uri)
  foreign_namespace?(mapping, target_uri) &&
    !mapping.type_name_value.nil?
end

.leading_backslash_run(source) ⇒ Object 



745
746
747
748
749
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 745

def self.leading_backslash_run(source)
  run = 0
  run += 1 while source[run] == "\\"
  run
end

.lexical_value(type, value) ⇒ Object 



794
795
796
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 794

def self.lexical_value(type, value)
  type.serialize(type.cast(value))
end

.missing_schema_location_error(ns_class) ⇒ Object 



372
373
374
375
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 372

def self.missing_schema_location_error(ns_class)
  "XSD generation: foreign namespace '#{ns_class.uri}' needs a " \
    "schema_location so its imported types can be resolved."
end

.nested_ref_prefix(nested_mapping, ctx) ⇒ Object

The prefix a named-type reference should carry. A foreign model's type lives in its own namespace, so the ref uses that namespace's prefix; a model in the target namespace (same URI or none) is defined in this schema, so the ref uses the target prefix.



79
80
81
82
83
84
85
86
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 79

def self.nested_ref_prefix(nested_mapping, ctx)
  ns_class = nested_mapping&.namespace_class
  return ctx.target_prefix unless ns_class
  return ctx.target_prefix if ns_class.uri == ctx.target_uri

  prefix = ns_class.prefix_default
  usable_prefix?(prefix) ? prefix : ctx.target_prefix
end

.pattern_value(regexp) ⇒ Object 



701
702
703
704
705
706
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 701

def self.pattern_value(regexp)
  reject_unsupported_regex_flags!(regexp)
  source = strip_ruby_anchors(regexp.source)
  reject_invalid_xsd_pattern!(regexp, source)
  source
end

.prefix_collision_error(prefix, uri, schema_attrs) ⇒ Object

The collision message when prefix is already bound to a different namespace than uri in schema_attrs, or nil when there is no conflict.



363
364
365
366
367
368
369
370
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 363

def self.prefix_collision_error(prefix, uri, schema_attrs)
  existing = schema_attrs[:"xmlns:#{prefix}"]
  return unless existing && existing != uri

  "XSD generation: namespace prefix '#{prefix}' is bound " \
    "to two different namespaces (#{existing} and #{uri}). " \
    "Give them distinct prefixes."
end

.qn(local) ⇒ Object

Qualify an XSD structure element name with the bound prefix, so the prefix used on emitted elements always matches the declared xmlns.



25
26
27
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 25

def self.qn(local)
  "#{xsd_ns.prefix}:#{local}"
end

.qualify_type_ref(type_name, prefix) ⇒ Object

Qualify a named-type reference with the given prefix so it resolves to the namespace the type is defined in. Built-in refs (which already carry a prefix, e.g. "xs:string") and no-namespace schemas (prefix nil) are returned unchanged.



68
69
70
71
72
73
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 68

def self.qualify_type_ref(type_name, prefix)
  return type_name if prefix.nil? || type_name.nil?
  return type_name if type_name.include?(":")

  "#{prefix}:#{type_name}"
end

.raise_inexpressible_pattern!(regexp, detail) ⇒ Object

Raises:



761
762
763
764
765
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 761

def self.raise_inexpressible_pattern!(regexp, detail)
  raise Lutaml::Model::Error,
        "Cannot export pattern #{regexp.inspect} to xs:pattern: " \
        "it is not expressible in XSD's regexp subset (#{detail})."
end

.referenced_namespaces(klass, register, target_uri, seen = Set.new) ⇒ Hash

Namespaces referenced by THIS schema document, collected by walking the model tree from the root class (structural, no instance needed). The walk stops at foreign-model boundaries: a foreign model's type is imported — defined in its own schema document — so namespaces used only inside it are that document's concern, not this one's.

Returns:

  • (Hash) —

    :foreign_models — namespace classes of directly referenced foreign models (their type refs are prefixed QNames resolved through an import); :type_values — namespace classes declared by Type::Value attribute types (imported so their verbatim xsd_type references can resolve). Both deduped by class.



872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 872

def self.referenced_namespaces(klass, register, target_uri, seen = Set.new)
  result = { foreign_models: [], type_values: [] }
  return result unless klass.is_a?(::Class) && seen.add?(klass)

  klass.attributes.each_value do |attr|
    type_class = attr.type(register)
    next unless type_class

    if type_class <= Lutaml::Model::Serialize
      mapping = type_class.mappings_for(:xml)
      if imported_foreign_type?(mapping, target_uri)
        result[:foreign_models] << mapping.namespace_class
      elsif collection_placeholder?(attr, mapping)
        # Placeholder collection: its content is never emitted, so it
        # contributes no namespaces to this document.
        next
      else
        nested = referenced_namespaces(type_class, register,
                                       target_uri, seen)
        result[:foreign_models].concat(nested[:foreign_models])
        result[:type_values].concat(nested[:type_values])
      end
    else
      ns = get_namespace_info(type_class)[:class]
      if ns.is_a?(::Class) && ns < Lutaml::Xml::Namespace
        result[:type_values] << ns
      end
    end
  end

  result[:foreign_models].uniq!
  result[:type_values].uniq!
  result
end

.reject_invalid_xsd_pattern!(regexp, source) ⇒ Object 



767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 767

def self.reject_invalid_xsd_pattern!(regexp, source)
  return unless defined?(::Nokogiri::XML::Schema)

  probe = "    <xs:schema xmlns:xs=\"http://www.w3.org/2001/XMLSchema\">\n      <xs:simpleType name=\"p\">\n        <xs:restriction base=\"xs:string\">\n          <xs:pattern value=\"\#{xsd_escape(source)}\"/>\n        </xs:restriction>\n      </xs:simpleType>\n    </xs:schema>\n  XSD\n  schema = ::Nokogiri::XML::Schema(probe)\n  return if schema.errors.empty?\n\n  raise_inexpressible_pattern!(regexp, schema.errors.first.message)\nrescue ::Nokogiri::XML::SyntaxError => e\n  raise_inexpressible_pattern!(regexp, e.message)\nend\n"

.reject_unsupported_regex_flags!(regexp) ⇒ Object

Raises:



708
709
710
711
712
713
714
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 708

def self.reject_unsupported_regex_flags!(regexp)
  return if regexp.options.nobits?(UNSUPPORTED_XSD_REGEX_FLAGS)

  raise Lutaml::Model::Error,
        "Cannot export pattern #{regexp.inspect} to xs:pattern: " \
        "the i/m/x flags are not expressible in XSD's regexp subset."
end

.strip_leading_anchor(source) ⇒ Object 



721
722
723
724
725
726
727
728
729
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 721

def self.strip_leading_anchor(source)
  return source[1..] if source.start_with?("^")

  run = leading_backslash_run(source)
  return source unless run.odd? && source[run] == "A"

  # Keep the escaped-pair backslashes; drop the anchor's `\A`.
  source[0, run - 1] + source[(run + 1)..]
end

.strip_ruby_anchors(source) ⇒ Object 



716
717
718
719
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 716

def self.strip_ruby_anchors(source)
  source = strip_leading_anchor(source)
  strip_trailing_anchor(source)
end

.strip_trailing_anchor(source) ⇒ Object 



731
732
733
734
735
736
737
738
739
740
741
742
743
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 731

def self.strip_trailing_anchor(source)
  last = source[-1]
  before = source.length - 1
  if ["z", "Z"].include?(last) &&
      trailing_backslash_run(source, before).odd?
    return source[0...-2]
  end
  if last == "$" && trailing_backslash_run(source, before).even?
    return source[0...-1]
  end

  source
end

.synthesize_target_prefix(referenced) ⇒ Object

A synthetic prefix for a target namespace that declares none, avoiding every prefix claimed by a referenced namespace.



293
294
295
296
297
298
299
300
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 293

def self.synthesize_target_prefix(referenced)
  taken = (referenced[:foreign_models] + referenced[:type_values])
    .filter_map(&:prefix_default)
  candidate = "tns"
  suffix = 0
  candidate = "tns#{suffix += 1}" while taken.include?(candidate)
  candidate
end

.target_namespace_attrs(xml_mapping) ⇒ Object

targetNamespace, form defaults, and the target xmlns declaration, derived from the root mapping's namespace configuration.



264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 264

def self.target_namespace_attrs(xml_mapping)
  attrs = {}

  if xml_mapping.namespace_class
    ns = xml_mapping.namespace_class
    attrs[:targetNamespace] = ns.uri
    attrs[:elementFormDefault] = ns.element_form_default.to_s
    attrs[:attributeFormDefault] = ns.attribute_form_default.to_s
    attrs[:version] = ns.version if ns.version

    prefix = xml_mapping.namespace_prefix || ns.prefix_default
    attrs[:"xmlns:#{prefix}"] = ns.uri if usable_prefix?(prefix)
  elsif xml_mapping.namespace_uri
    # Legacy: namespace URI without XmlNamespace class
    attrs[:targetNamespace] = xml_mapping.namespace_uri
    attrs[:elementFormDefault] = "unqualified"
    attrs[:attributeFormDefault] = "unqualified"

    if usable_prefix?(xml_mapping.namespace_prefix)
      attrs[:"xmlns:#{xml_mapping.namespace_prefix}"] =
        xml_mapping.namespace_uri
    end
  end

  attrs
end

.target_ns_prefix(xml_mapping) ⇒ Object

The prefix of the schema's target namespace, or nil when the schema has no target namespace (named types then live in no-namespace and their references stay unprefixed). Never returns the reserved "xs".



52
53
54
55
56
57
58
59
60
61
62
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 52

def self.target_ns_prefix(xml_mapping)
  prefix =
    if xml_mapping&.namespace_class
      xml_mapping.namespace_prefix ||
        xml_mapping.namespace_class.prefix_default
    elsif xml_mapping&.namespace_uri
      xml_mapping.namespace_prefix
    end

  prefix if usable_prefix?(prefix)
end

.trailing_backslash_run(source, pos) ⇒ Object 



751
752
753
754
755
756
757
758
759
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 751

def self.trailing_backslash_run(source, pos)
  run = 0
  i = pos - 1
  while i >= 0 && source[i] == "\\"
    run += 1
    i -= 1
  end
  run
end

.type_resolvable?(type_name, klass, register) ⇒ Boolean

Check if a custom XSD type can be resolved in the model hierarchy

Parameters:

  • type_name (String) —

    The custom type name to resolve

  • klass (Class) —

    The model class being processed

  • register (Register) —

    The register for type resolution

Returns:

  • (Boolean) —

    true if the type can be resolved



133
134
135
136
137
138
139
140
141
142
143
144
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 133

def self.type_resolvable?(type_name, klass, register)
  # Search in nested model attributes
  klass.attributes.each_value do |attr|
    attr_type = attr.type(register)
    next unless attr_type <= Lutaml::Model::Serialize

    nested_mapping = attr_type.mappings_for(:xml)
    return true if nested_mapping&.type_name_value == type_name
  end

  false
end

.usable_prefix?(prefix) ⇒ Boolean

Whether a prefix can appear in emitted QNames: present and not the reserved XSD vocabulary prefix.

Returns:

  • (Boolean) 


40
41
42
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 40

def self.usable_prefix?(prefix)
  !prefix.nil? && !prefix.empty? && prefix != xsd_ns.prefix
end

.validate_xsd_types!(klass, register, seen = Set.new) ⇒ Object

Validate all XSD types referenced by the model

Parameters:

  • klass (Class) —

    The model class to validate

  • register (Register) —

    The register for type resolution

Raises:

  • (UnresolvableTypeError) —

    if any types cannot be resolved



151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 151

def self.validate_xsd_types!(klass, register, seen = Set.new)
  # Cycle guard: recursive models (A -> B -> A) would otherwise recurse
  # forever. A class already being validated needs no re-validation.
  return unless seen.add?(klass)

  errors = []

  klass.attributes.each do |name, attr|
    attr_type = attr.type(register)

    # Validate Type::Value xsd_type
    if attr_type.is_a?(Class) && attr_type < Lutaml::Model::Type::Value
      type_name = attr_type.xsd_type
      classification = classify_xsd_type(type_name, klass, register)

      if classification == :unresolvable
        errors << "Attribute '#{name}' uses unresolvable xsd_type '#{type_name}'. " \
                  "Custom types must be defined as LutaML Type::Value or Model classes."
      end
    end

    # Recursively validate nested models
    if attr_type <= Lutaml::Model::Serialize
      begin
        validate_xsd_types!(attr_type, register, seen)
      rescue Lutaml::Model::UnresolvableTypeError => e
        errors << "In nested model #{attr_type.name}: #{e.message}"
      end
    end
  end

  if errors.any?
    raise Lutaml::Model::UnresolvableTypeError,
          errors.join("\n")
  end
end

.xs(xml, local, attrs = nil, &block) ⇒ Object

Emit an XSD structure element (<xs:local ...>) through the builder.



30
31
32
33
34
35
36
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 30

def self.xs(xml, local, attrs = nil, &block)
  if attrs
    xml.public_send(qn(local), attrs, &block)
  else
    xml.public_send(qn(local), &block)
  end
end

.xsd_escape(text) ⇒ Object 



790
791
792
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 790

def self.xsd_escape(text)
  text.gsub("&", "&amp;").gsub("<", "&lt;").gsub('"', "&quot;")
end

.xsd_ns ⇒ Object

The XSD vocabulary namespace, routed through the project's namespace object so the declared prefix and every emitted prefix stay in sync. W3c::XsNamespace binds "xs" (matching the built-in xs:-prefixed type references) and opts out of the W3C-reserved-prefix warning.



19
20
21
 
# File 'lib/lutaml/xml/schema/xsd_schema.rb', line 19

def self.xsd_ns
  @xsd_ns ||= Lutaml::Xml::W3c::XsNamespace.new
end