Merge pull request #43 from RudolfCardinal/outlook-msg-fixups

Skip extraction of text from Outlook .msg files where the file extension is null or missing (e.g. SignedAttachment)

Author: martinburchell

cardinal_pythonlib/extract_text.py

@@ -1369,13 +1369,14 @@ def _gen_msg_content(
         )
 
     for attachment in message.attachments:
-        # null termination seen in the real world
-        # https://github.com/TeamMsgExtractor/msg-extractor/issues/464
-        ext = attachment.extension.replace("\x00", "")
-        if ext is not None and ext in ext_map:
-            yield document_to_text(
-                blob=attachment.data, extension=ext, config=config
-            )
+        if (extension := getattr(attachment, "extension", None)) is not None:
+            # null termination seen in the real world
+            # https://github.com/TeamMsgExtractor/msg-extractor/issues/464
+            extension = extension.replace("\x00", "")
+            if extension in ext_map:
+                yield document_to_text(
+                    blob=attachment.data, extension=extension, config=config
+                )
 
 
 # =============================================================================

cardinal_pythonlib/tests/extract_text_tests.py

@@ -33,6 +33,7 @@
 from tempfile import mkdtemp, NamedTemporaryFile
 from unittest import mock, TestCase
 
+from extract_msg import SignedAttachment
 from faker import Faker
 from faker_file.providers.docx_file import DocxFileProvider
 from faker_file.providers.eml_file import EmlFileProvider
@@ -735,3 +736,51 @@ def test_attachment_converted(self) -> None:
             converted = convert_msg_to_text(dummy_filename, config=self.config)
 
         self.assertEqual(converted.strip(), content)
+
+    def test_attachment_with_null_extension_skipped(self) -> None:
+        self.fake.add_provider(DocxFileProvider)
+
+        dummy_filename = "dummy_filename.msg"
+
+        content = self.fake.paragraph(nb_sentences=10)
+        docx = self.fake.docx_file(content=content, raw=True)
+        mock_attachment = mock.Mock(
+            extension=None,
+            data=BytesIO(docx).read(),
+        )
+        mock_msgfile = mock.Mock(
+            body=None, htmlBody=None, attachments=[mock_attachment]
+        )
+        mock_openmsg = mock.Mock(return_value=mock_msgfile)
+        with mock.patch.multiple(
+            "cardinal_pythonlib.extract_text",
+            openMsg=mock_openmsg,
+        ):
+            self.config.width = 0
+            converted = convert_msg_to_text(dummy_filename, config=self.config)
+
+        self.assertEqual(converted.strip(), "")
+
+    def test_signed_attachment_with_no_extension_skipped(self) -> None:
+        self.fake.add_provider(DocxFileProvider)
+
+        dummy_filename = "dummy_filename.msg"
+
+        content = self.fake.paragraph(nb_sentences=10)
+        docx = self.fake.docx_file(content=content, raw=True)
+        mock_attachment = mock.Mock(
+            spec=SignedAttachment,
+            data=BytesIO(docx).read(),
+        )
+        mock_msgfile = mock.Mock(
+            body=None, htmlBody=None, attachments=[mock_attachment]
+        )
+        mock_openmsg = mock.Mock(return_value=mock_msgfile)
+        with mock.patch.multiple(
+            "cardinal_pythonlib.extract_text",
+            openMsg=mock_openmsg,
+        ):
+            self.config.width = 0
+            converted = convert_msg_to_text(dummy_filename, config=self.config)
+
+        self.assertEqual(converted.strip(), "")

docs/source/changelog.rst

@@ -910,6 +910,11 @@ Quick links:
   document converters (``.docx``, ``.pdf``, ``.odt`` etc.) to
   :func:`cardinal_pythonlib.extract_text.document_to_text`.
 
+.. _changelog_2026:
+
+2026
+~~~~
+
 **2.1.2 (2026-01-27)**
 
 - ``cardinalpythonlib_grep_in_openxml``: new facility to search XML node text
@@ -919,9 +924,9 @@ Quick links:
 - Fix extraction of text from HTML files in
   :func:`cardinal_pythonlib.extract_text.document_to_text`.
 
-.. _changelog_2026:
-
-2026
-~~~~
 
 **2.1.3 (IN PROGRESS)**
+
+- Skip extraction of text from Outlook ``.msg`` files where the file extension
+  is null or missing (SignedAttachment).
+  :func:`cardinal_pythonlib.extract_text.document_to_text`.