ROB: Cope with some issues in pillow

pypdf/_xobj_image_helpers.py

@@ -73,9 +73,7 @@
  color_components = cast(int, icc_profile["/N"])
  color_space = icc_profile.get("/Alternate", "")
  elif color_space[0] == "/Indexed":
- color_space = color_space[1]
- if isinstance(color_space, IndirectObject):
- color_space = color_space.get_object()
+ color_space = color_space[1].get_object()
  mode2, invert_color = _get_imagemode(
  color_space, color_components, prev_mode, depth + 1
  )
@@ -292,10 +290,17 @@
  mode = "RGBA"
  # we need to convert to the good mode
  try:
- if img1.mode != mode:
+ if (img1.mode == mode) or (img1.mode in ("L", "P") and mode in ("L", "P")):
+ img = img1
+ elif (
+ img1.mode == "RGBA"
+ and mode == "CMYK"
+ or img1.mode == "CMYK"
+ and mode == "RGBA"
+ ):
  img = Image.frombytes(mode, img1.size, img1.tobytes())
  else:
- img = img1
+ img = img1.convert(mode)
  except OSError:
  img = Image.frombytes(mode, img1.size, img1.tobytes())
  # for CMYK conversion :

pypdf/filters.py

@@ -894,10 +894,13 @@
  img_byte_arr = BytesIO()
  try:
  img.save(img_byte_arr, format=image_format)
- except OSError: # pragma: no cover
- # odd error
+ except OSError:
+ # in case of we convert to RGBA and then to PNG
+ img1 = img.convert("RGBA")
+ image_format = "PNG"
+ extension = ".png"
  img_byte_arr = BytesIO()
- img.save(img_byte_arr, format=image_format)
+ img1.save(img_byte_arr, format=image_format)
  data = img_byte_arr.getvalue()
 
  try: # temporary try/except until other fixes of images

tests/test_images.py

@@ -283,3 +283,14 @@ def test_data_with_lf():
  name = "iss2343b0.png"
  img = Image.open(BytesIO(get_data_from_url(url, name=name)))
  assert image_similarity(reader.pages[8].images[9].image, img) == 1.0
+
+
+@pytest.mark.enable_socket()
+def test_oserror():
+ """Cf #2265"""
+ url = "https:/py-pdf/pypdf/files/13127130/Binance.discovery.responses.2.gov.uscourts.dcd.256060.140.1.pdf"
+ name = "iss2265.pdf"
+ reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
+ reader.pages[2].images[1]
+ # due to errors in translation in pillow we may not be have to get
+ # the correct image therefore we cannot use image_similarity