BUG: fix for merge page with annotation #3467

pypdf/annotations/__init__.py

@@ -13,6 +13,7 @@
 
 from ._base import NO_FLAGS, AnnotationDictionary
 from ._markup_annotations import (
+    AbstractPolyLine,
     Ellipse,
     FreeText,
     Highlight,
@@ -27,6 +28,7 @@
 
 __all__ = [
     "NO_FLAGS",
+    "AbstractPolyLine",
     "AnnotationDictionary",
     "Ellipse",
     "FreeText",

pypdf/annotations/_markup_annotations.py

@@ -52,6 +52,41 @@ def __init__(self, *, title_bar: Optional[str] = None) -> None:
             self[NameObject("/T")] = TextStringObject(title_bar)
 
 
+class AbstractPolyLine(MarkupAnnotation, ABC):
+    """
+    Base class for Polygon and PolyLine
+
+    Args:
+        vertices: List of coordinates of each vertex;
+
+    """
+    def __init__(
+            self,
+            vertices: Union[list[Vertex], ArrayObject],
+            **kwargs: Any
+        ) -> None:
+        super().__init__(**kwargs)
+        if len(vertices) == 0:
+            raise ValueError(f"A {type(self).__name__.lower()} needs at least 1 vertex with two coordinates")
+
+    @staticmethod
+    def _determine_vertices(
+            vertices: Union[list[Vertex], ArrayObject]
+        ) -> tuple[list[Vertex], list[NumberObject]]:
+        coord_list: ArrayObject = ArrayObject()
+        if isinstance(vertices, ArrayObject):
+            coord_list = vertices
+            args = [iter(vertices)] * 2 # Adapted def grouper()
+            vertices = list(zip(*args))  # from https://docs.python.org/3.9/library/itertools.html#itertools-recipes
+
+        else:
+            for x, y in vertices:
+                coord_list.append(NumberObject(x))
+                coord_list.append(NumberObject(y))
+
+        return vertices, coord_list
+
+
 class Text(MarkupAnnotation):
     """
     A text annotation.
@@ -186,19 +221,15 @@ def __init__(
         )
 
 
-class PolyLine(MarkupAnnotation):
+class PolyLine(AbstractPolyLine):
     def __init__(
         self,
-        vertices: list[Vertex],
+        vertices: Union[list[Vertex], ArrayObject],
         **kwargs: Any,
     ) -> None:
-        super().__init__(**kwargs)
-        if len(vertices) == 0:
-            raise ValueError("A polyline needs at least 1 vertex with two coordinates")
-        coord_list = []
-        for x, y in vertices:
-            coord_list.append(NumberObject(x))
-            coord_list.append(NumberObject(y))
+        super().__init__(vertices=vertices, **kwargs)
+
+        vertices, coord_list = self._determine_vertices(vertices)
         self.update(
             {
                 NameObject("/Subtype"): NameObject("/PolyLine"),
@@ -280,20 +311,14 @@ def __init__(
             )
 
 
-class Polygon(MarkupAnnotation):
+class Polygon(AbstractPolyLine):
     def __init__(
         self,
-        vertices: list[tuple[float, float]],
+        vertices: Union[list[Vertex], ArrayObject],
         **kwargs: Any,
     ) -> None:
-        super().__init__(**kwargs)
-        if len(vertices) == 0:
-            raise ValueError("A polygon needs at least 1 vertex with two coordinates")
-
-        coord_list = []
-        for x, y in vertices:
-            coord_list.append(NumberObject(x))
-            coord_list.append(NumberObject(y))
+        super().__init__(vertices=vertices, **kwargs)
+        vertices, coord_list = self._determine_vertices(vertices)
         self.update(
             {
                 NameObject("/Type"): NameObject("/Annot"),

pypdf/generic/_data_structures.py

@@ -29,6 +29,7 @@
 __author__ = "Mathieu Fenniak"
 __author_email__ = "[email protected]"
 
+import inspect
 import logging
 import re
 import sys
@@ -263,6 +264,40 @@ def read_from_stream(
 
 
 class DictionaryObject(dict[Any, Any], PdfObject):
+
+    _init_alt_arg_names_: Optional[dict[str, str]] = None
+
+
+    """
+    Used to map DictionaryObject keyword to __init__ arg names
+    when cloning. key -> arg_name.
+
+    Args:
+        key: string identifying DictionaryObject keyword
+        arg_name: string identifying argument name (value)
+    """
+    def set_alt_arg_name(self, key: str, arg_name: str) -> None:
+        if self._init_alt_arg_names_ is None:
+            self._init_alt_arg_names_ = {}
+
+        if not isinstance(key, str) or not isinstance(arg_name, str):
+            raise TypeError
+
+        self._init_alt_arg_names_[key] = arg_name
+
+    """
+    Used to return __init__ arg names when cloning
+    Args:
+        key: string identifying DictionaryObject keyword
+    Returns:
+        Returns None if not found, else the string representing the
+        argument name
+    """
+    def get_alt_arg_name(self, key: str) -> Optional[str]:
+        if self._init_alt_arg_names_ is None or not isinstance(key, str):
+            return None
+        return self._init_alt_arg_names_.get(key, None)
+
     def replicate(
         self,
         pdf_dest: PdfWriterProtocol,
@@ -291,9 +326,22 @@ def clone(
             pass
 
         visited: set[tuple[int, int]] = set()  # (idnum, generation)
+
+        kwargs = {}
+        inspector = inspect.getfullargspec(self.__class__.__init__)
+
+        for key, val in self.items():
+            alt_arg_name = self.get_alt_arg_name(key)
+            key_stripped = key.removeprefix("/").lower()
+
+            if alt_arg_name:
+                kwargs[alt_arg_name] = val
+            elif key_stripped in inspector.args or key_stripped in inspector.kwonlyargs:
+                kwargs[key_stripped] = val
+
         d__ = cast(
             "DictionaryObject",
-            self._reference_clone(self.__class__(), pdf_dest, force_duplicate),
+            self._reference_clone(self.__class__(**kwargs), pdf_dest, force_duplicate),
         )
         if ignore_fields is None:
             ignore_fields = []

tests/generic/test_files.py

@@ -8,6 +8,7 @@
 import pytest
 
 from pypdf import PdfReader, PdfWriter
+from pypdf.annotations._markup_annotations import Polygon
 from pypdf.constants import AFRelationship
 from pypdf.errors import PdfReadError, PyPdfError
 from pypdf.generic import (
@@ -575,3 +576,27 @@ def test_embedded_file__order():
         "test.txt", attachment4.pdf_object.indirect_reference,
         "xyz.txt", attachment3.pdf_object.indirect_reference,
     ]
+
+
+def test_merge_page_with_annotation():
+    # added and adapted from issue #3467
+    writer = PdfWriter()
+    writer2 = PdfWriter()
+    writer.add_blank_page(100, 100)
+    writer2.add_blank_page(100, 100)
+
+    annotation = Polygon(
+        vertices=[(50, 550), (200, 650), (70, 750), (50, 700)],
+    )
+
+    writer.add_annotation(0, annotation)
+
+    page1 = writer.pages[0]
+    page2 = writer2.pages[0]
+    page2.merge_page(page1)
+
+    assert page2.annotations[0].get_object()["/Type"] == annotation["/Type"]
+    assert page2.annotations[0].get_object()["/Subtype"] == annotation["/Subtype"]
+    assert page2.annotations[0].get_object()["/Vertices"] == annotation["/Vertices"]
+    assert page2.annotations[0].get_object()["/IT"] == annotation["/IT"]
+    assert page2.annotations[0].get_object()["/Rect"] == annotation["/Rect"]

tests/test_generic.py

@@ -54,6 +54,18 @@
 RESOURCE_ROOT = PROJECT_ROOT / "resources"
 
 
+class DummyDictObject(DictionaryObject):
+    def __init__(self, test_name: str) -> None:
+        super().__init__()
+        self.set_alt_arg_name("/TestName", "test_name")
+
+        self.update(
+            {
+                NameObject("/TestName"): test_name
+            }
+        )
+
+
 class ChildDummy(DictionaryObject):
     @property
     def indirect_reference(self):
@@ -1282,6 +1294,22 @@ def test_coverage_arrayobject():
         assert isinstance(k, int)
         assert isinstance(v, PdfObject)
 
+def test_alt_keyword_when_cloning():
+    obj = DummyDictObject(test_name="testval")
+    obj2: DummyDictObject = None
+
+    clone_obj = obj.clone(obj2)
+
+    assert clone_obj.get("/TestName") == "testval"
+
+def test_type_error_when_using_alt_arg_keys():
+    dict_obj = DictionaryObject()
+    with pytest.raises(TypeError):
+        dict_obj.set_alt_arg_name(1, "test_key")
+
+    with pytest.raises(TypeError):
+        dict_obj.set_alt_arg_name("test_key", 1)
+
 
 def test_coverage_streamobject():
     writer = PdfWriter()