Remove mime type

PyPDF2/_page.py

@@ -377,12 +377,10 @@ def images(self) -> List[File]:
         x_object = self[PG.RESOURCES][RES.XOBJECT].get_object()  # type: ignore
         for obj in x_object:
             if x_object[obj][IA.SUBTYPE] == "/Image":
-                mime_type, byte_stream = _xobj_to_image(x_object[obj])
-                if mime_type is not None:
-                    filename = f"{obj[1:]}.{File._mime2extension(mime_type)}"
-                    images_extracted.append(
-                        File(name=filename, data=byte_stream, mime_type=mime_type)
-                    )
+                extension, byte_stream = _xobj_to_image(x_object[obj])
+                if extension is not None:
+                    filename = f"{obj[1:]}{extension}"
+                    images_extracted.append(File(name=filename, data=byte_stream))
         return images_extracted
 
     @property

PyPDF2/_utils.py

@@ -420,19 +420,3 @@ def rename_kwargs(  # type: ignore
 class File:
     name: str
     data: bytes
-    mime_type: str
-
-    @property
-    def format(self) -> str:
-        return File._mime2extension(self.mime_type)
-
-    @staticmethod
-    def _mime2extension(mime_type: str) -> str:
-        mapping = {
-            "image/png": "png",
-            "image/jpeg": "jpg",
-            "image/x-jp2": "jp2",
-            "image/gif": "gif",
-            "image/tiff": "tiff",
-        }
-        return mapping.get(mime_type, "unknown")

PyPDF2/filters.py

@@ -562,7 +562,7 @@ def _xobj_to_image(x_object_obj: Dict[str, Any]) -> Tuple[Optional[str], bytes]:
     It's unclear if PyPDF2 will keep this function here, hence it's private.
     It might get removed at any point.
 
-    :return: Tuple[mime type, bytes]
+    :return: Tuple[extension, bytes]
     """
     from PIL import Image
 
@@ -576,10 +576,10 @@ def _xobj_to_image(x_object_obj: Dict[str, Any]) -> Tuple[Optional[str], bytes]:
         mode: Literal["RGB", "P"] = "RGB"
     else:
         mode = "P"
-    mime_type = None
+    extension = None
     if SA.FILTER in x_object_obj:
         if x_object_obj[SA.FILTER] == FT.FLATE_DECODE:
-            mime_type = "image/png"
+            extension = ".png"  # mime_type = "image/png"
             color_space = None
             if "/ColorSpace" in x_object_obj:
                 color_space = x_object_obj["/ColorSpace"].get_object()
@@ -608,21 +608,21 @@ def _xobj_to_image(x_object_obj: Dict[str, Any]) -> Tuple[Optional[str], bytes]:
         ):
             # I'm not sure if the mime types have any relationship to the filters
             if x_object_obj[SA.FILTER] in [[FT.LZW_DECODE], [FT.CCITT_FAX_DECODE]]:
-                mime_type = "image/tiff"
+                extension = ".tiff"  # mime_type = "image/tiff"
             else:
-                mime_type = "image/png"
+                extension = ".png"  # mime_type = "image/png"
             data = b_(data)
         elif x_object_obj[SA.FILTER] == FT.DCT_DECODE:
-            mime_type = "image/jpeg"
+            extension = ".jpg"  # mime_type = "image/jpeg"
         elif x_object_obj[SA.FILTER] == "/JPXDecode":
-            mime_type = "image/x-jp2"
+            extension = ".jp2"  # mime_type = "image/x-jp2"
         elif x_object_obj[SA.FILTER] == FT.CCITT_FAX_DECODE:
-            mime_type = "image/tiff"
+            extension = ".tiff"  # mime_type = "image/tiff"
     else:
-        mime_type = "image/png"
+        extension = ".png"  # mime_type = "image/png"
         img = Image.frombytes(mode, size, data)
         img_byte_arr = BytesIO()
         img.save(img_byte_arr, format="PNG")
         data = img_byte_arr.getvalue()
 
-    return mime_type, data
+    return extension, data

tests/test_reader.py

@@ -201,8 +201,10 @@ def test_get_images(src, expected_images):
         assert image.name == expected_image
         with open(f"test-out-{src}-{image.name}", "wb") as fp:
             fp.write(image.data)
-        assert image.format.upper() == Image.open(io.BytesIO(image.data)).format
-        assert image.mime_type == expected_mime
+        assert (
+            image.name.split(".")[-1].upper()
+            == Image.open(io.BytesIO(image.data)).format
+        )
 
 
 @pytest.mark.parametrize(