Generic PDF objects

Implementation of generic PDF objects (dictionary, number, string, …).

class pypdf.generic.ArrayObject(iterable=(), /)

Bases: list[Any], PdfObject

replicate(pdf_dest: PdfWriterProtocol) → ArrayObject

clone(pdf_dest: PdfWriterProtocol, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → ArrayObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

items() → Iterable[Any]

Emulate DictionaryObject.items for a list (index, object).

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static read_from_stream(stream: IO[Any], pdf: PdfReaderProtocol | None, forced_encoding: None | str | list[str] | dict[int, str] = None) → ArrayObject

class pypdf.generic.BooleanObject(value: Any)

Bases: PdfObject

clone(pdf_dest: PdfWriterProtocol, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → BooleanObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static read_from_stream(stream: IO[Any]) → BooleanObject

class pypdf.generic.ByteStringObject(*args, **kwargs)

Bases: bytes, PdfObject

Represents a string object where the text encoding could not be determined.

This occurs quite often, as the PDF spec doesn’t provide an alternate way to represent strings – for example, the encryption data stored in files (like /O) is clearly not text, but is still stored in a “String” object.

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → ByteStringObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

property original_bytes: bytes

For compatibility with TextStringObject.original_bytes.

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf.generic.ContentStream(stream: Any, pdf: Any, forced_encoding: None | str | list[str] | dict[int, str] = None)

Bases: DecodedStreamObject

In order to be fast, this data structure can contain either:

• raw data in ._data

• parsed stream operations in ._operations.

At any time, ContentStream object can either have both of those fields defined, or one field defined and the other set to None.

These fields are “rebuilt” lazily, when accessed:

• when .get_data() is called, if ._data is None, it is rebuilt from ._operations.

• when .operations is called, if ._operations is None, it is rebuilt from ._data.

Conversely, these fields can be invalidated:

• when .set_data() is called, ._operations is set to None.

• when .operations is set, ._data is set to None.

replicate(pdf_dest: PdfWriterProtocol) → ContentStream

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → ContentStream

Clone object into pdf_dest.

Parameters:

• pdf_dest

• force_duplicate

• ignore_fields

Returns:

The cloned ContentStream

get_data() → bytes

set_data(data: bytes) → None

property operations: list[tuple[Any, bytes]]

isolate_graphics_state() → None

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf.generic.DecodedStreamObject

Bases: StreamObject

class pypdf.generic.DictionaryObject

Bases: dict[Any, Any], PdfObject

replicate(pdf_dest: PdfWriterProtocol) → DictionaryObject

clone(pdf_dest: PdfWriterProtocol, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → DictionaryObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

raw_get(key: Any) → Any

get_inherited(key: str, default: Any = None) → Any

Returns the value of a key or from the parent if not found. If not found returns default.

Parameters:

• key – string identifying the field to return

• default – default value to return

Returns:

Current key or inherited one, otherwise default value.

setdefault(key: Any, value: Any | None = None) → Any

property xmp_metadata: XmpInformationProtocol | None

Retrieve XMP (Extensible Metadata Platform) data relevant to this object, if available.

See Table 347 — Additional entries in a metadata stream dictionary.

Returns:

Returns a XmpInformation instance that can be used to access XMP metadata from the document. Can also return None if no metadata was found on the document root.

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static read_from_stream(stream: IO[Any], pdf: PdfReaderProtocol | None, forced_encoding: None | str | list[str] | dict[int, str] = None) → DictionaryObject

class pypdf.generic.DirectReferenceLink(reference: ArrayObject)

Bases: object

Direct reference link being preserved until we can resolve it correctly.

find_referenced_page() → IndirectObject

patch_reference(target_pdf: PdfWriter, new_page: IndirectObject) → None

target_pdf: PdfWriter which the new link went into

class pypdf.generic.EmbeddedFile(name: str, pdf_object: DictionaryObject, parent: ArrayObject | None = None)

Bases: object

Container holding the information on an embedded file.

Attributes are evaluated lazily if possible.

Further information on embedded files can be found in section 7.11 of the PDF 2.0 specification.

property name: str

The (primary) name of the embedded file as provided in the name tree.

property alternative_name: str | None

Retrieve the alternative name (file specification).

property description: str | None

Retrieve the description.

property associated_file_relationship: str

Retrieve the relationship of the referring document to this embedded file.

property subtype: str | None

Retrieve the subtype. This is a MIME media type, prefixed by a slash.

property content: bytes

Retrieve the actual file content.

property size: int | None

Retrieve the size of the uncompressed file in bytes.

property creation_date: datetime | None

Retrieve the file creation datetime.

property modification_date: datetime | None

Retrieve the datetime of the last file modification.

property checksum: bytes | None

Retrieve the MD5 checksum of the (uncompressed) file.

delete() → None

Delete the file from the document.

class pypdf.generic.EncodedStreamObject

Bases: StreamObject

get_data() → bytes

set_data(data: bytes) → None

class pypdf.generic.FloatObject(value: Any = '0.0', context: Any | None = None)

Bases: float, PdfObject

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → FloatObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

myrepr() → str

as_numeric() → float

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf.generic.IndirectObject(idnum: int, generation: int, pdf: Any)

Bases: PdfObject

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

replicate(pdf_dest: PdfWriterProtocol) → PdfObject

clone(pdf_dest: PdfWriterProtocol, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → IndirectObject

Clone object into pdf_dest.

property indirect_reference: IndirectObject

get_object() → PdfObject | None

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static read_from_stream(stream: IO[Any], pdf: Any) → IndirectObject

class pypdf.generic.NameObject

Bases: str, PdfObject

delimiter_pattern = re.compile(b'\\s+|[\\(\\)<>\\[\\]{}/%]')

prefix = b'/'

renumber_table: ClassVar[dict[str, bytes]] = {'\x00': b'#00', '\x01': b'#01', '\x02': b'#02', '\x03': b'#03', '\x04': b'#04', '\x05': b'#05', '\x06': b'#06', '\x07': b'#07', '\x08': b'#08', '\t': b'#09', '\n': b'#0A', '\x0b': b'#0B', '\x0c': b'#0C', '\r': b'#0D', '\x0e': b'#0E', '\x0f': b'#0F', '\x10': b'#10', '\x11': b'#11', '\x12': b'#12', '\x13': b'#13', '\x14': b'#14', '\x15': b'#15', '\x16': b'#16', '\x17': b'#17', '\x18': b'#18', '\x19': b'#19', '\x1a': b'#1A', '\x1b': b'#1B', '\x1c': b'#1C', '\x1d': b'#1D', '\x1e': b'#1E', '\x1f': b'#1F', ' ': b'#20', '#': b'#23', '%': b'#25', '(': b'#28', ')': b'#29', '/': b'#2F', '<': b'#3C', '>': b'#3E', '[': b'#5B', ']': b'#5D', '{': b'#7B', '}': b'#7D'}

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → NameObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

renumber() → bytes

surfix

Decorator that converts a method with a single cls argument into a property that can be accessed directly from the class.

static unnumber(sin: bytes) → bytes

CHARSETS = ('utf-8', 'gbk', 'latin1')

static read_from_stream(stream: IO[Any], pdf: Any) → NameObject

class pypdf.generic.NamedReferenceLink(reference: TextStringObject, source_pdf: PdfReader)

Bases: object

Named reference link being preserved until we can resolve it correctly.

find_referenced_page() → IndirectObject | None

patch_reference(target_pdf: PdfWriter, new_page: IndirectObject) → None

target_pdf: PdfWriter which the new link went into

class pypdf.generic.NullObject(*args, **kwargs)

Bases: PdfObject

clone(pdf_dest: PdfWriterProtocol, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → NullObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static read_from_stream(stream: IO[Any]) → NullObject

class pypdf.generic.NumberObject(value: Any)

Bases: int, PdfObject

NumberPattern = re.compile(b'[^+-.0-9]')

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → NumberObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

as_numeric() → int

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static read_from_stream(stream: IO[Any]) → NumberObject | FloatObject

class pypdf.generic.OutlineFontFlag(value, names=<not given>, *values, module=None, qualname=None, type=None, start=1, boundary=None)

Bases: IntFlag

A class used as an enumerable flag for formatting an outline font.

italic = 1

bold = 2

class pypdf.generic.OutlineItem(title: str | bytes, page: NumberObject | IndirectObject | NullObject | DictionaryObject, fit: Fit)

Bases: Destination

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf.generic.PdfObject(*args, **kwargs)

Bases: PdfObjectProtocol

hash_func(*, usedforsecurity=True)

Returns a sha1 hash object; optionally initialized with a string

indirect_reference: IndirectObject | None

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

hash_value_data() → bytes

hash_value() → bytes

replicate(pdf_dest: PdfWriterProtocol) → PdfObject

Clone object into pdf_dest (PdfWriterProtocol which is an interface for PdfWriter) without ensuring links. This is used in clone_document_from_root with incremental = True.

Parameters:

pdf_dest – Target to clone to.

Returns:

The cloned PdfObject

clone(pdf_dest: PdfWriterProtocol, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → PdfObject

Clone object into pdf_dest (PdfWriterProtocol which is an interface for PdfWriter).

By default, this method will call _reference_clone (see _reference).

Parameters:

• pdf_dest – Target to clone to.

• force_duplicate – By default, if the object has already been cloned and referenced, the copy will be returned; when True, a new copy will be created. (Default value = False)

• ignore_fields – List/tuple of field names (for dictionaries) that will be ignored during cloning (applies to children duplication as well). If fields are to be considered for a limited number of levels, you have to add it as integer, for example [1,"/B","/TOTO"] means that "/B" will be ignored at the first level only but "/TOTO" on all levels.

Returns:

The cloned PdfObject

get_object() → PdfObject | None

Resolve indirect references.

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf.generic.StreamObject

Bases: DictionaryObject

replicate(pdf_dest: PdfWriterProtocol) → StreamObject

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

get_data() → bytes

set_data(data: bytes) → None

hash_value_data() → bytes

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

static initialize_from_dictionary(data: dict[str, Any]) → EncodedStreamObject | DecodedStreamObject

flate_encode(level: int = -1) → EncodedStreamObject

decode_as_image() → Any

Try to decode the stream object as an image

Returns:

a PIL image if proper decoding has been found

Raises:

Exception – Errors during decoding will be reported. It is recommended to catch exceptions to prevent stops in your program.

class pypdf.generic.TextStringObject(value: Any)

Bases: str, PdfObject

A string object that has been decoded into a real unicode string.

If read from a PDF document, this string appeared to match the PDFDocEncoding, or contained a UTF-16BE BOM mark to cause UTF-16 decoding to occur.

autodetect_pdfdocencoding: bool

autodetect_utf16: bool

utf16_bom: bytes

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: Sequence[str | int] | None = ()) → TextStringObject

Clone object into pdf_dest.

hash_bin() → int

Used to detect modified object.

Returns:

Hash considering type and value.

property original_bytes: bytes

It is occasionally possible that a text string object gets created where a byte string object was expected due to the autodetection mechanism – if that occurs, this “original_bytes” property can be used to back-calculate what the original encoded bytes were.

get_original_bytes() → bytes

get_encoded_bytes() → bytes

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf.generic.TreeObject(dct: DictionaryObject | None = None)

Bases: DictionaryObject

has_children() → bool

children() → Iterable[Any]

add_child(child: Any, pdf: PdfWriterProtocol) → None

inc_parent_counter_default(parent: None | IndirectObject | TreeObject, n: int) → None

inc_parent_counter_outline(parent: None | IndirectObject | TreeObject, n: int) → None

insert_child(child: Any, before: Any, pdf: PdfWriterProtocol, inc_parent_counter: Callable[[...], Any] | None = None) → IndirectObject

remove_child(child: Any) → None

remove_from_tree() → None

Remove the object from the tree it is in.

empty_tree() → None

class pypdf.generic.ViewerPreferences(value: Any = None)

Bases: DictionaryObject

property PRINT_SCALING: NameObject

pypdf.generic.create_string_object(string: str | bytes, forced_encoding: None | str | list[str] | dict[int, str] = None) → TextStringObject | ByteStringObject

Create a ByteStringObject or a TextStringObject from a string to represent the string.

Parameters:

• string – The data being used

• forced_encoding – Typically None, or an encoding string

Returns:

A ByteStringObject

Raises:

TypeError – If string is not of type str or bytes.

pypdf.generic.decode_pdfdocencoding(byte_array: bytes) → str

pypdf.generic.encode_pdfdocencoding(unicode_string: str) → bytes

pypdf.generic.extract_links(new_page: PageObject, old_page: PageObject) → list[tuple[NamedReferenceLink | DirectReferenceLink, NamedReferenceLink | DirectReferenceLink]]

Extracts links from two pages on the assumption that the two pages are the same. Produces one list of (new link, old link) tuples.

pypdf.generic.hex_to_rgb(value: str) → tuple[float, float, float]

pypdf.generic.is_null_or_none(x: Any) → TypeGuard[None | NullObject | IndirectObject]

Returns:

True if x is None or NullObject.

pypdf.generic.read_hex_string_from_stream(stream: IO[Any], forced_encoding: None | str | list[str] | dict[int, str] = None) → TextStringObject | ByteStringObject

pypdf.generic.read_object(stream: IO[Any], pdf: PdfReaderProtocol | None, forced_encoding: None | str | list[str] | dict[int, str] = None) → PdfObject | int | str | ContentStream

pypdf.generic.read_string_from_stream(stream: IO[Any], forced_encoding: None | str | list[str] | dict[int, str] = None) → TextStringObject | ByteStringObject

class pypdf._protocols.PdfObjectProtocol(*args, **kwargs)

Bases: Protocol

indirect_reference: Any

clone(pdf_dest: Any, force_duplicate: bool = False, ignore_fields: tuple[str, ...] | list[str] | None = ()) → Any

get_object() → PdfObjectProtocol | None

hash_value() → bytes

write_to_stream(stream: IO[Any], encryption_key: None | str | bytes = None) → None

class pypdf._protocols.XmpInformationProtocol(*args, **kwargs)

Bases: PdfObjectProtocol

class pypdf._protocols.PdfCommonDocProtocol(*args, **kwargs)

Bases: Protocol

property pdf_header: str

property pages: list[Any]

property root_object: PdfObjectProtocol

get_object(indirect_reference: Any) → PdfObjectProtocol | None

property strict: bool

class pypdf._protocols.PdfReaderProtocol(*args, **kwargs)

Bases: PdfCommonDocProtocol, Protocol

abstract property xref: dict[int, dict[int, Any]]

abstract property trailer: dict[str, Any]

class pypdf._protocols.PdfWriterProtocol(*args, **kwargs)

Bases: PdfCommonDocProtocol, Protocol

incremental: bool

abstract write(stream: Path | str | IO[Any]) → tuple[bool, IO[Any]]