Allow blobs to be empty when extracting text

It is possible to have an email with an empty body. Other scenarios (empty HTML, docx etc)
are pretty unlikely

Author: martinburchell

cardinal_pythonlib/extract_text.py

@@ -358,7 +358,7 @@ def get_filelikeobject(filename: str = None, blob: bytes = None) -> BinaryIO:
     Returns:
         a :class:`BinaryIO` object
     """
-    if not filename and not blob:
+    if filename is None and blob is None:
         raise ValueError("no filename and no blob")
     if filename and blob:
         raise ValueError("specify either filename or blob")
@@ -373,11 +373,11 @@ def get_file_contents(filename: str = None, blob: bytes = None) -> bytes:
     """
     Returns the binary contents of a file, or of a BLOB.
     """
-    if not filename and not blob:
+    if filename is None and blob is None:
         raise ValueError("no filename and no blob")
     if filename and blob:
         raise ValueError("specify either filename or blob")
-    if blob:
+    if blob is not None:
         return blob
     with open(filename, "rb") as f:
         return f.read()
@@ -1408,7 +1408,7 @@ def document_to_text(
         Raises an exception for malformed arguments, missing files, bad
         filetypes, etc.
     """
-    if not filename and not blob:
+    if filename is None and blob is None:
         raise ValueError("document_to_text: no filename and no blob")
     if filename and blob:
         raise ValueError("document_to_text: specify either filename or blob")

cardinal_pythonlib/tests/extract_text_tests.py

@@ -204,6 +204,12 @@ def test_htm_converted(self) -> None:
         )
         self.assertEqual(text.strip(), content)
 
+    def test_empty_htm_converted(self) -> None:
+        text = document_to_text(
+            blob="".encode("utf-8"), extension="htm", config=self.config
+        )
+        self.assertEqual(text, "")
+
     def test_log_converted(self) -> None:
         content = """
 2025-04-02 06:05:43,772 INFO Starting unattended upgrades script
@@ -477,6 +483,25 @@ def test_eml_with_no_content_type_converted(self) -> None:
 
         self.assertIn(text_content, text)
 
+    def test_eml_with_empty_body_converted(self) -> None:
+        content = """From: bar@example.org
+Subject: No body
+To: foo@example.org
+Mime-Version: 1.0
+Content-Type: multipart/mixed;boundary="==="
+
+--===
+--===--
+"""
+        message = message_from_string(content, policy=policy.default)
+        blob = message.as_bytes()
+
+        text = document_to_text(
+            blob=blob, extension=".eml", config=self.config
+        )
+
+        self.assertEqual("", text)
+
     def test_unsupported_converted(self) -> None:
         with mock.patch.multiple(
             "cardinal_pythonlib.extract_text.subprocess",