Feature: auto-clean some invalid pdfs (#7651)

src/documents/tests/test_api_documents.py

@@ -1402,6 +1402,27 @@ class TestDocumentApi(DirectoriesMixin, DocumentConsumeDelayMixin, APITestCase):
         self.assertEqual(overrides.filename, "simple.pdf")
         self.assertEqual(overrides.custom_field_ids, [custom_field.id])
 
+    def test_upload_invalid_pdf(self):
+        """
+        GIVEN: Invalid PDF named "*.pdf" that mime_type is in settings.CONSUMER_PDF_RECOVERABLE_MIME_TYPES
+        WHEN: Upload the file
+        THEN: The file is not rejected
+        """
+        self.consume_file_mock.return_value = celery.result.AsyncResult(
+            id=str(uuid.uuid4()),
+        )
+
+        with open(
+            os.path.join(os.path.dirname(__file__), "samples", "invalid_pdf.pdf"),
+            "rb",
+        ) as f:
+            response = self.client.post(
+                "/api/documents/post_document/",
+                {"document": f},
+            )
+
+        self.assertEqual(response.status_code, status.HTTP_200_OK)
+
     def test_get_metadata(self):
         doc = Document.objects.create(
             title="test",

src/documents/tests/test_consumer.py

@@ -235,6 +235,8 @@ class FaultyGenericExceptionParser(_BaseTestParser):
 
 def fake_magic_from_file(file, mime=False):
     if mime:
+        if file.name.startswith("invalid_pdf"):
+            return "application/octet-stream"
         if os.path.splitext(file)[1] == ".pdf":
             return "application/pdf"
         elif os.path.splitext(file)[1] == ".png":
@@ -952,6 +954,27 @@ class TestConsumer(
 
         sanity_check()
 
+    @mock.patch("documents.consumer.run_subprocess")
+    def test_try_to_clean_invalid_pdf(self, m):
+        shutil.copy(
+            Path(__file__).parent / "samples" / "invalid_pdf.pdf",
+            settings.CONSUMPTION_DIR / "invalid_pdf.pdf",
+        )
+        with self.get_consumer(
+            settings.CONSUMPTION_DIR / "invalid_pdf.pdf",
+        ) as consumer:
+            # fails because no qpdf
+            self.assertRaises(ConsumerError, consumer.run)
+
+            m.assert_called_once()
+
+            args, _ = m.call_args
+
+            command = args[0]
+
+            self.assertEqual(command[0], "qpdf")
+            self.assertEqual(command[1], "--replace-input")
+
 
 @mock.patch("documents.consumer.magic.from_file", fake_magic_from_file)
 class TestConsumerCreatedDate(DirectoriesMixin, GetConsumerMixin, TestCase):