ROB: Replace error by warning for EOD in RunLengthDecode/ASCIIHexDecode

pypdf/_xobj_image_helpers.py

@@ -4,7 +4,7 @@
 from io import BytesIO
 from typing import Any, List, Tuple, Union, cast
 
-from ._utils import logger_warning
+from ._utils import WHITESPACES, logger_warning
 from .constants import ColorSpaces
 from .errors import PdfReadError
 from .generic import (
@@ -195,7 +195,13 @@
  else:
  if img.mode == "1":
  # Two values ("high" and "low").
- assert len(lookup) == 2 * nb, len(lookup)
+ expected_count = 2 * nb
+ if len(lookup) != expected_count:
+ if len(lookup) < expected_count:
+ raise PdfReadError(f"Not enough lookup values: Expected {expected_count}, got {len(lookup)}.")
+ lookup = lookup[:expected_count]
+ if not all(_value in WHITESPACES for _value in lookup[expected_count:]):
+ raise PdfReadError(f"Too many lookup values: Expected {expected_count}, got {len(lookup)}.")
  colors_arr = [lookup[:nb], lookup[nb:]]
  arr = b"".join(
  [

pypdf/filters.py

@@ -285,7 +285,8 @@ def decode(
  index = 0
  while True:
  if index >= len(data):
- raise PdfStreamError("Unexpected EOD in ASCIIHexDecode")
+ break # reach End Of String even if no EOD
+ # raise PdfStreamError("Unexpected EOD in ASCIIHexDecode")
  char = data[index : index + 1]
  if char == b">":
  break
@@ -340,7 +341,8 @@ def decode(
  index = 0
  while True:
  if index >= len(data):
- raise PdfStreamError("Unexpected EOD in RunLengthDecode")
+ break # reach End Of String even if no EOD
+ # raise PdfStreamError("Unexpected EOD in ASCIIHexDecode")
  length = data[index]
  index += 1
  if length == 128:

tests/test_filters.py

@@ -11,7 +11,7 @@
 from PIL import Image
 
 from pypdf import PdfReader
-from pypdf.errors import DeprecationError, PdfReadError, PdfStreamError
+from pypdf.errors import DeprecationError, PdfReadError
 from pypdf.filters import (
  ASCII85Decode,
  ASCIIHexDecode,
@@ -131,9 +131,9 @@ def test_ascii_hex_decode_method(data, expected):
 
 def test_ascii_hex_decode_missing_eod():
  """ASCIIHexDecode.decode() raises error when no EOD character is present."""
- with pytest.raises(PdfStreamError) as exc:
-  ASCIIHexDecode.decode("")
- assert exc.value.args[0] == "Unexpected EOD in ASCIIHexDecode"
+ # with pytest.raises(PdfStreamError) as exc:
+ ASCIIHexDecode.decode("")
+ # assert exc.value.args[0] == "Unexpected EOD in ASCIIHexDecode"
 
 
 @pytest.mark.enable_socket()
@@ -561,14 +561,10 @@ def test_runlengthdecode():
  url = "https://github.com/py-pdf/pypdf/files/12162905/out.pdf"
  name = "FailedRLE1.pdf"
  reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
- with pytest.raises(PdfStreamError) as exc:
- reader.pages[0].images[0]
- assert exc.value.args[0] == "Unexpected EOD in RunLengthDecode"
+ reader.pages[0].images[0]
  url = "https://github.com/py-pdf/pypdf/files/12162926/out.pdf"
  name = "FailedRLE2.pdf"
- with pytest.raises(PdfStreamError) as exc:
- reader.pages[0].images[0]
- assert exc.value.args[0] == "Unexpected EOD in RunLengthDecode"
+ reader.pages[0].images[0]
 
 
 @pytest.mark.enable_socket()
@@ -649,3 +645,12 @@ def test_flate_decode_with_image_mode_1():
  reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
  for image in reader.pages[7].images:
  _ = image
+
+
+@pytest.mark.enable_socket()
+def test_flate_decode_with_image_mode_1__whitespace_at_end_of_lookup():
+ """From #2331"""
+ url = "https://github.com/py-pdf/pypdf/files/13611048/out1.pdf"
+ name = "issue2331.pdf"
+ reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
+ reader.pages[0].images[0]