MAINT: Reduce usage of b_

pypdf/_reader.py

@@ -54,7 +54,6 @@
 from ._utils import (
  StrByteType,
  StreamType,
- b_,
  deprecate_no_replacement,
  deprecation_no_replacement,
  deprecation_with_replacement,
@@ -1257,7 +1256,7 @@ def _get_object_from_stream(
  assert cast(str, obj_stm["/Type"]) == "/ObjStm"
  # /N is the number of indirect objects in the stream
  assert idx < obj_stm["/N"]
- stream_data = BytesIO(b_(obj_stm.get_data())) # type: ignore
+ stream_data = BytesIO(obj_stm.get_data()) # type: ignore
  for i in range(obj_stm["/N"]): # type: ignore
  read_non_whitespace(stream_data)
  stream_data.seek(-1, 1)
@@ -1868,7 +1867,7 @@ def _read_pdf15_xref_stream(
  xrefstream = cast(ContentStream, read_object(stream, self))
  assert cast(str, xrefstream["/Type"]) == "/XRef"
  self.cache_indirect_object(generation, idnum, xrefstream)
- stream_data = BytesIO(b_(xrefstream.get_data()))
+ stream_data = BytesIO(xrefstream.get_data())
  # Index pairs specify the subsections in the dictionary. If
  # none create one subsection that spans everything.
  idx_pairs = xrefstream.get("/Index", [0, xrefstream.get("/Size")])

pypdf/_writer.py

@@ -705,7 +705,7 @@ def add_attachment(self, filename: str, data: Union[str, bytes]) -> None:
  # endobj
 
  file_entry = DecodedStreamObject()
- file_entry.set_data(data)
+ file_entry.set_data(b_(data))
  file_entry.update({NameObject(PA.TYPE): NameObject("/EmbeddedFile")})
 
  # The Filespec entry

pypdf/filters.py

@@ -41,7 +41,6 @@
 from typing import Any, Dict, List, Optional, Tuple, Union, cast
 
 from ._utils import (
- b_,
  deprecate_with_replacement,
  logger_warning,
  ord_,
@@ -1035,7 +1034,6 @@ def _handle_jpx(
  else:
  extension = ".png" # mime_type = "image/png"
  image_format = "PNG"
- data = b_(data)
  img = Image.open(BytesIO(data), formats=("TIFF", "PNG"))
  elif lfilters == FT.DCT_DECODE:
  img, image_format, extension = Image.open(BytesIO(data)), "JPEG", ".jpg"

pypdf/generic/_base.py

@@ -648,7 +648,7 @@ def write_to_stream(
  deprecate_no_replacement(
  "the encryption_key parameter of write_to_stream", "5.0.0"
  )
- stream.write(self.renumber()) # b_(renumber(self)))
+ stream.write(self.renumber())
 
  def writeToStream(
  self, stream: StreamType, encryption_key: Union[None, str, bytes]

pypdf/generic/_data_structures.py

@@ -820,7 +820,7 @@ def _clone(
 
  def hash_value_data(self) -> bytes:
  data = super().hash_value_data()
- data += b_(self._data)
+ data += self._data
  return data
 
  @property
@@ -906,10 +906,10 @@ def flate_encode(self, level: int = -1) -> "EncodedStreamObject":
 
 
 class DecodedStreamObject(StreamObject):
- def get_data(self) -> Any:
+ def get_data(self) -> bytes:
  return self._data
 
- def set_data(self, data: Any) -> Any:
+ def set_data(self, data: bytes) -> None:
  self._data = data
 
  def getData(self) -> Any: # deprecated
@@ -935,7 +935,7 @@ def decodedSelf(self, value: DecodedStreamObject) -> None: # deprecated
  deprecation_with_replacement("decodedSelf", "decoded_self", "3.0.0")
  self.decoded_self = value
 
- def get_data(self) -> Union[None, str, bytes]:
+ def get_data(self) -> bytes:
  from ..filters import decode_stream_data
 
  if self.decoded_self is not None:
@@ -956,7 +956,7 @@ def getData(self) -> Union[None, str, bytes]: # deprecated
  deprecation_with_replacement("getData", "get_data", "3.0.0")
  return self.get_data()
 
- def set_data(self, data: Any) -> None: # deprecated
+ def set_data(self, data: bytes) -> None: # deprecated
  from ..filters import FlateDecode
 
  if self.get(SA.FILTER, "") == FT.FLATE_DECODE:
@@ -996,14 +996,13 @@ def __init__(
  if isinstance(stream, ArrayObject):
  data = b""
  for s in stream:
- data += b_(s.get_object().get_data())
+ data += s.get_object().get_data()
  if len(data) == 0 or data[-1] != b"\n":
  data += b"\n"
  stream_bytes = BytesIO(data)
  else:
- stream_data = stream.get_data()
- assert stream_data is not None
- stream_data_bytes = b_(stream_data)
+ stream_data_bytes = stream.get_data()
+ assert stream_data_bytes is not None
  stream_bytes = BytesIO(stream_data_bytes)
  self.forced_encoding = forced_encoding
  self.__parse_content_stream(stream_bytes)
@@ -1194,8 +1193,8 @@ def _data(self) -> bytes: # type: ignore
  return new_data.getvalue()
 
  @_data.setter
- def _data(self, value: Union[str, bytes]) -> None:
- self.__parse_content_stream(BytesIO(b_(value)))
+ def _data(self, value: bytes) -> None:
+ self.__parse_content_stream(BytesIO(value))
 
 
 def read_object(
@@ -1276,10 +1275,12 @@ def __init__(self, data: DictionaryObject) -> None:
  if isinstance(self.get("/V"), EncodedStreamObject):
  d = cast(EncodedStreamObject, self[NameObject("/V")]).get_data()
  if isinstance(d, bytes):
- d = d.decode()
+ d_str = d.decode()
  elif d is None:
- d = ""
- self[NameObject("/V")] = TextStringObject(d)
+ d_str = ""
+ else:
+ raise Exception("Should never happen")
+ self[NameObject("/V")] = TextStringObject(d_str)
 
  # TABLE 8.69 Entries common to all field dictionaries
  @property