Structure traversal and cache

Uses iterative pre-order traversal and returns a caller-owned flattened snapshot with stable Level and ParentIndex values

Each TPdfStructureElement also reports hierarchy facts through MarkedContentIDs, ChildCount, ChildMarkedContentIDs, and AttributeCount, table structure through HasHeaders, HeadersIds, and ColSpan, and the print and reference sentinels HasPrintField and HasReference

Role mapping is exposed through RoleMappedNamespaceUri and RoleMappedTypeName, associated files through HasAssociatedFile, AssociatedFileRelationship, AssociatedFileDescription, AssociatedFileDeclaredInCatalog, AssociatedFileDeclaredInStructure, AssociatedFileInEmbeddedFilesNameTree, AssociatedFileHasReadingOrder, and AssociatedFileReadingOrderIndex, and annotation facts through AnnotationSubtype, AnnotationContents, AnnotationAlternateText, IsArtifact, the widget sentinels WidgetCount and HasAppearance, and FieldAlternateName

MaxDepth, MaxElements, MaxAttributeValues, and DeadlineMilliseconds bound native work and managed allocation

IncludeAttributes=False keeps structure text, hierarchy, and MCIDs while skipping attribute value materialisation

Status.Completed is false when any budget is reached, the deadline expires, or a visitor stops traversal

Status.ElementsVisited and Status.AttributeValuesVisited count the work performed, while Status.DepthLimitReached, Status.ElementLimitReached, Status.AttributeLimitReached, and Status.DeadlineReached identify which budget stopped the walk

Calls Visitor once for each element without allocating the complete page array

Set the callback's KeepGoing parameter to false for a normal early stop reported by Status.VisitorStopped

Each callback receives managed element data that can be retained after the callback returns and no native structure handle is exposed

Reads any one-based page without changing PageNumber or firing OnPageChange

Complete default-budget results enter the shared LRU cache used by StructureElements, TPdfView.StructureElements, readable content, reflow, structured text, and accessibility validation

Nested dynamic arrays and attribute blobs are deep copied so changing a returned snapshot cannot modify the cache or the PDF

Limits the number of complete page snapshots retained by least-recently-used order

Set the capacity to zero to release cached snapshots and disable future caching

Negative capacity values raise EPdfError

Expose cumulative lookup telemetry for the component instance

Edits invalidate the affected page or the complete cache while preserving cumulative counters