ROB: Pad truncated data in bits2byte instead of reading out of bounds (#3820)

---------

Co-authored-by: Stefan <96178532+stefan6419846@users.noreply.github.com>

Author: metsw24-max

pypdf/generic/_image_xobject.py

@@ -131,6 +131,12 @@ def bits2byte(data: bytes, size: tuple[int, int], bits: int) -> bytes:
 
     byte_buffer = bytearray(buffer_size)
     mask = (1 << bits) - 1
+
+    required = size[1] * ((size[0] * bits + 7) // 8)
+    if (length := len(data)) < required:
+        logger_warning("Image data is not rectangular. Adding padding.", source=__name__)
+        data += b"\x00" * (required - length)
+
     data_index = 0
     bit = 8 - bits
     for y in range(size[1]):

tests/generic/test_image_xobject.py

@@ -276,3 +276,38 @@ def test_bits2byte__limit() -> None:
             match=r"^Requested buffer size 76500000 exceeds limit of 75000000\.$"
     ):
         bits2byte(data=b"TEST", size=(9000, 8500), bits=8)
+
+
+def test_bits2byte__truncated_data(caplog: pytest.LogCaptureFixture) -> None:
+    # 4x4 image at 2 bits per sample needs 4 bytes; provide only 1.
+    result = bits2byte(data=b"\x00", size=(4, 4), bits=2)
+    assert result == bytes(16)
+    assert "Image data is not rectangular. Adding padding." in caplog.text
+
+
+def test_handle_flate__truncated_2bit_image(caplog: pytest.LogCaptureFixture) -> None:
+    # A 3x3 indexed image at 2 bits per sample needs 3 bytes; provide only 1.
+    # Padding the missing bytes lets the image still be loaded instead of
+    # raising IndexError out of bits2byte.
+    lookup = DecodedStreamObject()
+    lookup.set_data(bytes([0, 0, 0, 10, 10, 10, 20, 20, 20, 30, 30, 30]))
+    result = _handle_flate(
+        size=(3, 3),
+        data=b"\xe4",
+        mode="2bits",
+        color_space=ArrayObject(
+            [NameObject("/Indexed"), NameObject("/DeviceRGB"), NumberObject(3), lookup]
+        ),
+        colors=1,
+        obj_as_text="dummy",
+    )
+    image = result[0]
+    image.load()
+    assert image.mode == "RGB"
+    assert image.size == (3, 3)
+    assert get_image_data(image) == (
+        (30, 30, 30), (20, 20, 20), (10, 10, 10),
+        (0, 0, 0), (0, 0, 0), (0, 0, 0),
+        (0, 0, 0), (0, 0, 0), (0, 0, 0),
+    )
+    assert "Image data is not rectangular. Adding padding." in caplog.text