ENH: Add decode_as_image() to ContentStreams (#2615)

Closes #2613

Author: pubpub-zz

docs/user/extract-images.md

@@ -19,3 +19,25 @@ for image_file_object in page.images:
         fp.write(image_file_object.data)
         count += 1
 ```
+
+# Other images
+
+Some other objects can contain images, such as stamp annotations.
+
+For example, this document contains such stamps:
+[test_stamp.pdf](https://github.com/user-attachments/files/15751424/test_stamp.pdf)
+
+You can extract the image from the annotation with the following code:
+
+```python
+from pypdf import PdfReader
+
+reader = PdfReader("test_stamp.pdf")
+im = (
+    reader.pages[0]["/Annots"][0]
+    .get_object()["/AP"]["/N"]["/Resources"]["/XObject"]["/Im4"]
+    .decode_as_image()
+)
+
+im.show()
+```

pypdf/generic/_data_structures.py

@@ -955,6 +955,31 @@ def flate_encode(self, level: int = -1) -> "EncodedStreamObject":
         retval._data = FlateDecode.encode(b_(self._data), level)
         return retval
 
+    def decode_as_image(self) -> Any:
+        """
+        Try to decode the stream object as an image
+
+        Returns:
+            a PIL image if proper decoding has been found
+        Raises:
+            Exception: (any)during decoding to to invalid object or
+                errors during decoding will be reported
+                It is recommended to catch exceptions to prevent
+                stops in your program.
+        """
+        from ..filters import _xobj_to_image
+
+        if self.get("/Subtype", "") != "/Image":
+            try:
+                msg = f"{self.indirect_reference} does not seem to be an Image"  # pragma: no cover
+            except AttributeError:
+                msg = f"{self.__repr__()} object does not seem to be an Image"  # pragma: no cover
+            logger_warning(msg, __name__)
+        extension, byte_stream, img = _xobj_to_image(self)
+        if extension is None:
+            return None  # pragma: no cover
+        return img
+
 
 class DecodedStreamObject(StreamObject):
     pass

tests/test_images.py

@@ -441,3 +441,24 @@ def test_inline_image_extraction():
     name = "iss2598d.png"
     img = Image.open(BytesIO(get_data_from_url(url, name=name)))
     assert image_similarity(reader.pages[0].images[0].image, img) == 1
+
+
+@pytest.mark.enable_socket()
+def test_extract_image_from_object(caplog):
+    url = "https://github.com/py-pdf/pypdf/files/15176076/B2.pdf"
+    name = "iss2613.pdf"
+    reader = PdfReader(BytesIO(get_data_from_url(url, name=name)))
+    image = reader.pages[0]["/Resources"]["/Pattern"]["/P1"]["/Resources"]["/XObject"][
+        "/X1"
+    ].decode_as_image()
+    assert isinstance(image, Image.Image)
+    with pytest.raises(Exception):
+        co = reader.pages[0].get_contents()
+        co.decode_as_image()
+    assert "does not seem to be an Image" in caplog.text
+    caplog.clear()
+    co.indirect_reference = "for_test"
+    with pytest.raises(Exception):
+        co = reader.pages[0].get_contents()
+        co.decode_as_image()
+    assert "does not seem to be an Image" in caplog.text