BUG: Lookup index in _xobj_to_image can be ByteStringObject (#1366)

DEV: Adjusted File classes __str__ and __repr__ to easy debugging

PyPDF2/_utils.py

@@ -416,7 +416,24 @@ def rename_kwargs( # type: ignore
  )
 
 
+def _human_readable_bytes(bytes: int) -> str:
+ if bytes < 10**3:
+ return f"{bytes} Byte"
+ elif bytes < 10**6:
+ return f"{bytes / 10**3:.1f} kB"
+ elif bytes < 10**9:
+ return f"{bytes / 10**6:.1f} MB"
+ else:
+ return f"{bytes / 10**9:.1f} GB"
+
+
 @dataclass
 class File:
  name: str
  data: bytes
+
+ def __str__(self) -> str:
+ return f"File(name={self.name}, data: {_human_readable_bytes(len(self.data))})"
+
+ def __repr__(self) -> str:
+ return f"File(name={self.name}, data: {_human_readable_bytes(len(self.data))}, hash: {hash(self.data)})"

PyPDF2/filters.py

@@ -593,7 +593,12 @@ def _xobj_to_image(x_object_obj: Dict[str, Any]) -> Tuple[Optional[str], bytes]:
 
  img = Image.frombytes(mode, size, data)
  if color_space == "/Indexed":
- img.putpalette(lookup.get_data())
+ from .generic import ByteStringObject
+
+ if isinstance(lookup, ByteStringObject):
+ img.putpalette(lookup)
+ else:
+ img.putpalette(lookup.get_data())
  img = img.convert("RGB")
  if G.S_MASK in x_object_obj: # add alpha channel
  alpha = Image.frombytes("L", size, x_object_obj[G.S_MASK].get_data())

tests/test_utils.py

@@ -7,7 +7,9 @@
 import PyPDF2._utils
 from PyPDF2 import PdfReader
 from PyPDF2._utils import (
+ File,
  _get_max_pdf_version_header,
+ _human_readable_bytes,
  deprecate_bookmark,
  mark_location,
  matrix_multiply,
@@ -256,3 +258,24 @@ def test_escapedcode_followed_by_int():
  reader = PdfReader(io.BytesIO(get_pdf_from_url(url, name=name)))
  for page in reader.pages:
  page.extract_text()
+
+
+@pytest.mark.parametrize(
+ ("input_int", "expected_output"),
+ [
+ (123, "123 Byte"),
+ (1234, "1.2 kB"),
+ (123_456, "123.5 kB"),
+ (1_234_567, "1.2 MB"),
+ (1_234_567_890, "1.2 GB"),
+ (1_234_567_890_000, "1234.6 GB"),
+ ],
+)
+def test_human_readable_bytes(input_int, expected_output):
+ assert _human_readable_bytes(input_int) == expected_output
+
+
+def test_file():
+ f = File(name="image.png", data=b"")
+ assert str(f) == "File(name=image.png, data: 0 Byte)"
+ assert repr(f) == "File(name=image.png, data: 0 Byte, hash: 0)"

tests/test_workflows.py

@@ -636,6 +636,7 @@ def test_merge_output(caplog):
  "https://corpora.tika.apache.org/base/docs/govdocs1/969/969502.pdf",
  "tika-969502.pdf",
  ),
+ ("https://arxiv.org/pdf/2201.00214.pdf", "arxiv-2201.00214.pdf"),
  ],
 )
 def test_image_extraction(url, name):