TST: Image extraction

py-pdf · Jul 9, 2022 · c3515d4 · c3515d4
1 parent f776f54
commit c3515d4
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 3 deletions.
diff --git a/PyPDF2/filters.py b/PyPDF2/filters.py
@@ -563,6 +563,7 @@ def _xobj_to_image(x_object_obj: Dict[str, Any]) -> Tuple[Optional[str], bytes]:
         IA.COLOR_SPACE in x_object_obj
         and x_object_obj[IA.COLOR_SPACE] == ColorSpaces.DEVICE_RGB
     ):
+        # https://pillow.readthedocs.io/en/stable/handbook/concepts.html#modes
         mode: Literal["RGB", "P"] = "RGB"
     else:
         mode = "P"

diff --git a/tests/test_workflows.py b/tests/test_workflows.py
@@ -378,9 +378,36 @@ def test_merge_output():
     merger.close()
 
 
-def test_image_extraction():
-    url = "https://corpora.tika.apache.org/base/docs/govdocs1/994/994636.pdf"
-    name = "tika-994636.pdf"
+@pytest.mark.parametrize(
+    ("url", "name"),
+    [
+        (
+            "https://corpora.tika.apache.org/base/docs/govdocs1/994/994636.pdf",
+            "tika-994636.pdf",
+        ),
+        (
+            "https://corpora.tika.apache.org/base/docs/govdocs1/952/952133.pdf",
+            "tika-952133.pdf",
+        ),
+        (  # JPXDecode
+            "https://corpora.tika.apache.org/base/docs/govdocs1/914/914568.pdf",
+            "tika-914568.pdf",
+        ),
+        (
+            "https://corpora.tika.apache.org/base/docs/govdocs1/952/952016.pdf",
+            "tika-952016.pdf",
+        ),
+        (
+            "https://corpora.tika.apache.org/base/docs/govdocs1/965/965118.pdf",
+            "tika-952016.pdf",
+        ),
+        (
+            "https://corpora.tika.apache.org/base/docs/govdocs1/959/959184.pdf",
+            "tika-959184.pdf",
+        ),
+    ],
+)
+def test_image_extraction(url, name):
     data = BytesIO(get_pdf_from_url(url, name=name))
     reader = PdfReader(data)