More tests with images

py-pdf · Apr 15, 2022 · ce82f40 · ce82f40
1 parent d0e4318
commit ce82f40
Show file tree

Hide file tree

Showing 5 changed files with 19 additions and 2 deletions.
diff --git a/PyPDF2/filters.py b/PyPDF2/filters.py
@@ -449,6 +449,16 @@ def _xobj_to_image(x_object_obj):
             img_byte_arr = io.BytesIO()
             img.save(img_byte_arr, format="PNG")
             data = img_byte_arr.getvalue()
+        elif x_object_obj["/Filter"] == ["/LZWDecode"]:
+            from PyPDF2.utils import b_
+            extension = ".png"
+            img = Image.frombytes(mode, size, b_(data))
+            if "/SMask" in x_object_obj:  # add alpha channel
+                alpha = Image.frombytes("L", size, x_object_obj["/SMask"].getData())
+                img.putalpha(alpha)
+            img_byte_arr = io.BytesIO()
+            img.save(img_byte_arr, format="PNG")
+            data = img_byte_arr.getvalue()
         elif x_object_obj["/Filter"] == "/DCTDecode":
             extension = ".jpg"
         elif x_object_obj["/Filter"] == "/JPXDecode":

diff --git a/Resources/imagemagick-images.pdf b/Resources/imagemagick-images.pdf
diff --git a/Resources/imagemagick-lzw.pdf b/Resources/imagemagick-lzw.pdf
diff --git a/Tests/test_page.py b/Tests/test_page.py
@@ -19,12 +19,16 @@
             os.path.join(RESOURCE_ROOT, "libreoffice-writer-password.pdf"),
             "openpassword",
         ),
+        (os.path.join(RESOURCE_ROOT, "imagemagick-images.pdf"), None),
+        (os.path.join(RESOURCE_ROOT, "imagemagick-lzw.pdf"), None),
     ],
     ids=[
         "crazyones",
         "attachment",
         "side-by-side-subfig",
         "libreoffice-writer-password",
+        "imagemagick-images",
+        "imagemagick-lzw",
     ],
 )
 def test_page_operations(pdf_path, password):

diff --git a/Tests/test_reader.py b/Tests/test_reader.py
@@ -111,6 +111,7 @@ def test_get_outlines(src, outline_elements):
         (os.path.join(RESOURCE_ROOT, "pdflatex-outline.pdf"), 0),
         (os.path.join(RESOURCE_ROOT, "crazyones.pdf"), 0),
         (os.path.join(RESOURCE_ROOT, "git.pdf"), 1),
+        (os.path.join(RESOURCE_ROOT, "imagemagick-lzw.pdf"), 1),
     ],
 )
 def test_get_images(src, nb_images):
@@ -135,11 +136,13 @@ def test_get_images(src, nb_images):
                     with open(filename, "wb") as img:
                         img.write(byte_stream)
                     images_extracted.append(filename)
-    else:
-        pass  # No image found
 
     assert len(images_extracted) == nb_images
 
+    # Cleanup
+    for filepath in images_extracted:
+        os.remove(filepath)
+
 
 @pytest.mark.parametrize(
     "strict,with_prev_0,should_fail",