FPdfCompress unit

Unit: FPdfCompress

Description

Syntax

TPdfTreeEditor = class
  class function TryLoad(Source: TStream; Kind: TPdfTreeKind;
    const RootReference: TPdfTreeObjectReference;
    const Options: TPdfTreeEditOptions; out Editor: TPdfTreeEditor;
    out Report: TPdfTreeValidationReport): Boolean; static;
  class function Load(Source: TStream; Kind: TPdfTreeKind;
    const RootReference: TPdfTreeObjectReference;
    const Options: TPdfTreeEditOptions): TPdfTreeEditor; static;
  function UpsertName(const Key, RawValue: AnsiString): Boolean;
  function UpsertNumber(Key: Int64; const RawValue: AnsiString): Boolean;
  function RemoveName(const Key: AnsiString): Boolean;
  function RemoveNumber(Key: Int64): Boolean;
  procedure Clear;
  function Validate(out Report: TPdfTreeValidationReport): Boolean;
  function BuildBalancedObjects(out Objects: TPdfTreeSerializedObjects;
    out Report: TPdfTreeValidationReport): Boolean;
  function SaveIncremental(Source, Dest: TStream;
    out Report: TPdfTreeValidationReport): Boolean;
  property Count: Integer;
  property Entries[Index: Integer]: TPdfTreeEntry; default;
end;
TPdfOutlineEditor = class
  class function TryLoad(Source: TStream;
    const Options: TPdfOutlineEditOptions; out Editor: TPdfOutlineEditor;
    out Report: TPdfOutlineValidationReport): Boolean; static;
  class function Load(Source: TStream;
    const Options: TPdfOutlineEditOptions): TPdfOutlineEditor; static;
  function Add(ParentId, SiblingIndex: Integer; const Title: WideString;
    PageNumber: Integer): Integer;
  function Delete(Id: Integer): Boolean;
  function Move(Id, NewParentId, SiblingIndex: Integer): Boolean;
  procedure SetTitle(Id: Integer; const Title: WideString);
  procedure SetStyle(Id: Integer; const Style: TPdfOutlineStyle);
  procedure SetExpanded(Id: Integer; Expanded: Boolean);
  procedure SetColor(Id: Integer; Red, Green, Blue: Single);
  procedure ClearColor(Id: Integer);
  procedure Retarget(Id, PageNumber: Integer;
    const DestinationSuffix: AnsiString = '/Fit';
    DestinationInAction: Boolean = False);
  function ApplyPageMap(const NewPageNumbers: array of Integer;
    DeleteDangling: Boolean; out Report: TPdfOutlineValidationReport): Boolean;
  function Validate(out Report: TPdfOutlineValidationReport): Boolean;
  function SaveIncremental(Source, Dest: TStream;
    out Report: TPdfOutlineValidationReport): Boolean;
  property Count: Integer;
  property Items[Index: Integer]: TPdfOutlineItem; default;
end;
TPdfCrossDocumentGraftMap = class
  constructor Create(const SourceIdentity: AnsiString);
  procedure Clear;
  function TryResolve(const Source: TPdfGraftObjectReference;
    out Destination: TPdfGraftObjectReference): Boolean;
  procedure Add(const Source, Destination: TPdfGraftObjectReference);
  property Count: Integer;
  property Entries[Index: Integer]: TPdfGraftMapEntry; default;
  property SourceIdentity: AnsiString;
end;
TPdfSparseDictionaryReader = class
  constructor Create(Source: TStream);
  function ReadDictionary(ObjectNumber, Generation: Integer;
    out Dictionary: AnsiString): Boolean;
  function ReadLatestDictionary(ObjectNumber: Integer;
    out Generation: Integer; out Dictionary: AnsiString): Boolean;
  function ReadDecodedStream(ObjectNumber, Generation,
    MaximumDecodedBytes: Integer; out Data: TBytes): Boolean;
  property Valid: Boolean;
end;
procedure CompressPdfStream(Source, Dest: TStream);
procedure LinearizePdfStream(Source, Dest: TStream);
procedure PackPdfObjectStreams(Source, Dest: TStream;
  MaxObjectsPerStream: Integer = 100;
  MaxDecodedBytesPerStream: Integer = 65536);
procedure NormalizePdfDocumentContent(Source, Dest: TStream;
  const Options: TPdfContentNormalizeOptions;
  out Report: TPdfContentNormalizeReport);
function GraftPdfAcroForm(Source, Destination, Output: TStream;
  const PageMappings: TPdfGraftPageMappings;
  const Options: TPdfAcroFormGraftOptions;
  GraftMap: TPdfCrossDocumentGraftMap;
  out Report: TPdfAcroFormGraftReport): Boolean;
procedure PruneUnreachablePdfStream(Source, Dest: TStream;
  out Report: TPdfPruneReport);
procedure PruneUnusedPdfResources(Source, Dest: TStream;
  out Report: TPdfResourcePruneReport;
  AllowSignedDocument: Boolean = False);
procedure DeduplicateResourcesPdfStream(Source, Dest: TStream;
  out Report: TPdfDedupReport);
procedure RecompressFlatePdfStream(Source, Dest: TStream;
  const Options: TPdfFlateRecompressOptions;
  out Report: TPdfFlateRecompressReport);
procedure ExternaliseInlineImagesPdfStream(Source, Dest: TStream;
  const Options: TPdfInlineImageExternaliseOptions;
  out Report: TPdfInlineImageExternaliseReport);
procedure ReproduciblePdfStream(Source, Dest: TStream;
  out Report: TPdfReproducibleReport);
procedure AppendPdfDirtyRevision(BaselineState, CurrentState,
  PreviousOutput, Dest: TStream; out ChangedObjectCount: Integer;
  out AppendedByteCount: Int64);
procedure ApplyDefaultOpenAction(Source, Dest: TStream;
  PdfVersion: Integer = 0);
procedure PageLabelledPdfStream(Source, Dest: TStream;
  const Ranges: TPdfPageLabelRanges; ReplaceExisting: Boolean;
  out Report: TPdfPageLabelReport);
procedure InspectOptionalContentPdfStream(Source: TStream;
  Usage: TPdfOptionalContentUsage; ConfigurationIndex: Integer;
  out Inventory: TPdfOptionalContentInventory);
procedure ConfigureOptionalContentPdfStream(Source, Dest: TStream;
  const Changes: TPdfOptionalContentStateChanges;
  const Options: TPdfOptionalContentConfigureOptions;
  out Report: TPdfOptionalContentConfigureReport);
procedure SanitiseActiveContentPdfStream(Source, Dest: TStream;
  const Options: TPdfSanitiseOptions;
  out Report: TPdfSanitiseReport);
procedure SanitiseActiveContentPdfStream(Source, Dest,
  PruneWorkspace: TStream; const Options: TPdfSanitiseOptions;
  out Report: TPdfSanitiseReport);
procedure ExportPdfObjectGraph(Source, Dest: TStream;
  const Options: TPdfObjectGraphInspectOptions;
  out Report: TPdfObjectGraphReport);
procedure PatchPdfObjectGraph(Source, Dest: TStream;
  const Operations: TPdfObjectPatchOperations;
  out Report: TPdfObjectGraphReport;
  AllowSignedDocument: Boolean = False);
procedure PostProcessSavedPdf(Source, Dest: TStream;
  PdfVersion: Integer = 0);
procedure PostProcessSavedPdf(Source, Dest, NormalizedWorkspace,
  RepairedWorkspace, ValidationWorkspace: TStream;
  PdfVersion: Integer = 0);
procedure RepairSubsetToUnicodeCMaps(Source, Dest: TStream);
function ReadPdfVersionInfo(Source: TStream; out Info: TPdfVersionInfo;
  out ErrorText: string): Boolean;
function ReadPdfSparseRevisionInfo(Source: TStream;
  out Info: TPdfSparseRevisionInfo): Boolean;
function ComposePdfSparseUpdateId(Source: TStream;
  const Info: TPdfSparseRevisionInfo;
  const DocumentId, InstanceId: TBytes): AnsiString;
function ReadPdfSparseCatalogSubDict(Source: TStream;
  const Name: AnsiString; out SubDict: AnsiString): Boolean;
function ReadPdfSparsePageDictionaries(Source: TStream;
  out Pages: TPdfSparsePageDictionaries): Boolean;

Functions and procedures

Types

Safety and limits

TPdfOutlineEditor.SetUri replaces a destination with a UTF-8 URI action and safely encodes literal delimiters; save verification ignores expanded state on leaf items because PDF omits their branch Count

The GraftPdfAcroForm options overload supports pagcpReject, pagcpRename, pagcpNamespace and compatible terminal text/choice joining with selected-widget pruning, isolated resources, appearances and calculation order; see FPdfFormMerge for policies and transactional document integration

AcroForm collision reports

TPdfAcroFormGraftOptions.NamespacePrefix names the parent inserted by the namespace policy; TPdfAcroFormCollisionPolicy.pagcpMergeCompatible joins supported compatible terminal text or choice fields

TPdfAcroFormGraftReport.FilteredFieldCount and TPdfAcroFormGraftReport.FilteredWidgetCount count source entries removed by selected-page filtering; TPdfAcroFormGraftReport.ResourceRenameCount counts isolated default resource names and TPdfAcroFormGraftReport.MergedFieldCount counts compatible fields joined to an existing owner

CapturePdfNavigation and WritePdfNavigation capture and remap named destinations and page-link annotations using explicit owner-page mappings; see FPdfMerge for raw key semantics, destination parameters, limits and transactional integration

Navigation capture records

TPdfNavigationNamedDestination= record
    // Keep the original PDF string bytes, including a UTF-16 BOM when present
    NameKey: AnsiString;
    // Absolute one-based page number; zero denotes an unresolved target
    PageNumber: Integer;
    // Exact destination mode and parameters, including null values
    DestinationSuffix: AnsiString;
  end;
  TPdfNavigationNamedDestinations= array of TPdfNavigationNamedDestination;
  TPdfNavigationPageLink= record
    // Owner page is one-based; annotation and link-only ordinals are zero-based
    PageNumber, AnnotationIndex, LinkIndex: Integer;
    HasDestination, HasGotoAction: Boolean;
    DestinationPageNumber, ActionPageNumber: Integer;
    DestinationSuffix, ActionSuffix: AnsiString;
  end;
  TPdfNavigationPageLinks= array of TPdfNavigationPageLink;

Read active object values

ReadPdfActiveObjects returns the latest active object values with their exact generation numbers and the active catalog revision, resolving classic xref tables, xref streams, incremental revisions and compressed object-stream members

Body excludes the indirect-object header and endobj marker; ordinary streams retain their dictionary and encoded payload, while compressed members are materialized as ordinary values, and xref/object-stream infrastructure is excluded from the returned collection

The source position is restored and a failure returns an empty object collection with diagnostics; encrypted input is rejected, while signatures can be inspected without altering their bytes

Defaults allow 256 MiB of source bytes, 64 MiB per materialized body, 256 MiB of cumulative bodies, 64 MiB per decoded object stream, 256 MiB of cumulative decoded object-stream caches and 1,000,000 object-number slots

Body limits are checked before each body copy and count temporary infrastructure bodies as well as returned values; the independently named decoded-cache limits stop object-stream decoding before exceeding the caller's budget

These are logical buffer limits, not a process-wide memory cap: source copies, the direct-object staging stream, scan buffers, decoder working storage and object metadata consume additional bounded allocations, and scan buffers can duplicate retained bodies

type
  TPdfActiveObject = record
    ObjectNumber, Generation: Integer;
    Body: TBytes;
  end;
  TPdfActiveObjects = array of TPdfActiveObject;
  TPdfActiveObjectReadOptions = record
    MaxSourceBytes, MaxObjectBytes, MaxTotalObjectBytes: Int64;
    MaxDecodedObjectStreamBytes, MaxTotalDecodedObjectStreamBytes: Int64;
    MaxObjects: Integer;
    class function Default: TPdfActiveObjectReadOptions; static;
  end;

function ReadPdfActiveObjects(Source: TStream;
  const Options: TPdfActiveObjectReadOptions; out Objects: TPdfActiveObjects;
  out Revision: TPdfSparseRevisionInfo; out ErrorText: string): Boolean;

See Also