fix: accept any IO[bytes] object in convert_to_bytes()

CHANGELOG.md

@@ -1,3 +1,15 @@
+## 0.21.9
+
+### Fixes
+- **Fix `ValueError` when partitioning a text file loaded from a zip archive**: `convert_to_bytes()`
+  rejected any file-like object not in its hardcoded type whitelist (`BytesIO`, `BufferedReader`,
+  `SpooledTemporaryFile`, `TextIOWrapper`), causing a `ValueError: Invalid file-like object type`
+  crash when callers passed a `zipfile.ZipExtFile` (returned by `zipfile.ZipFile.open()`). Fixed
+  by adding a duck-typing fallback that accepts any object with a `.read()` method, which covers
+  `ZipExtFile` as well as other standard `IO[bytes]` types not previously handled (e.g.
+  `GzipFile`, `tarfile.ExFileObject`). The file cursor is reset via `seek(0)` where supported so
+  callers can re-read the file after `convert_to_bytes()` returns.
+
 ## 0.21.8
 
 ### Enhancements

test_unstructured/partition/common/test_common.py

@@ -1,4 +1,5 @@
 import pathlib
+from io import BytesIO
 from multiprocessing import Pool
 
 import numpy as np
@@ -466,3 +467,68 @@ def test_normalize_layout_element_layout_element_text_source_metadata():
     assert hasattr(element, "metadata")
     assert hasattr(element.metadata, "is_extracted")
     assert element.metadata.is_extracted == "true"
+
+
+def test_convert_to_bytes_handles_a_ZipExtFile():
+    """ZipExtFile from zipfile.ZipFile.open() is readable as bytes."""
+    import zipfile
+
+    zip_buffer = BytesIO()
+    test_content = b"Hello from inside a zip archive!"
+    with zipfile.ZipFile(zip_buffer, "w", zipfile.ZIP_DEFLATED) as zf:
+        zf.writestr("test.txt", test_content)
+    zip_buffer.seek(0)
+
+    with zipfile.ZipFile(zip_buffer, "r") as zf:
+        with zf.open("test.txt") as zipext_file:
+            result = common.convert_to_bytes(zipext_file)
+
+    assert result == test_content
+
+
+def test_convert_to_bytes_resets_ZipExtFile_cursor_after_reading():
+    """The ZipExtFile should be seekable again after convert_to_bytes returns."""
+    import zipfile
+
+    zip_buffer = BytesIO()
+    test_content = b"Cursor should be reset after read."
+    with zipfile.ZipFile(zip_buffer, "w", zipfile.ZIP_STORED) as zf:
+        zf.writestr("test.txt", test_content)
+    zip_buffer.seek(0)
+
+    with zipfile.ZipFile(zip_buffer, "r") as zf:
+        with zf.open("test.txt") as zipext_file:
+            common.convert_to_bytes(zipext_file)
+            # -- cursor must be reset so caller can read the file again --
+            zipext_file.seek(0)
+            assert zipext_file.read() == test_content
+
+
+def test_convert_to_bytes_handles_a_generic_readable_IO_bytes_object():
+    """Any object with a .read() method is accepted, not just whitelisted types."""
+    import io
+
+    class _CustomStream(io.RawIOBase):
+        """Minimal non-standard IO[bytes] type not in the existing whitelist."""
+
+        def __init__(self, data: bytes):
+            self._data = BytesIO(data)
+
+        def read(self, n: int = -1) -> bytes:
+            return self._data.read(n)
+
+        def readinto(self, b: bytearray) -> int:
+            data = self._data.read(len(b))
+            n = len(data)
+            b[:n] = data
+            return n
+
+    test_content = b"Custom stream content."
+    stream = _CustomStream(test_content)
+    assert common.convert_to_bytes(stream) == test_content
+
+
+def test_convert_to_bytes_raises_on_a_non_readable_object():
+    """An object with no .read() method still raises ValueError."""
+    with pytest.raises(ValueError, match="Invalid file-like object type"):
+        common.convert_to_bytes(12345)  # type: ignore[arg-type]

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.21.8"  # pragma: no cover
+__version__ = "0.21.9"  # pragma: no cover

unstructured/partition/common/common.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import contextlib
 import numbers
 import subprocess
 from enum import Enum
@@ -389,6 +390,13 @@ def convert_to_bytes(file: bytes | IO[bytes]) -> bytes:
         with open(file.name, "rb") as f:
             return f.read()
 
+    if hasattr(file, "read"):
+        f_bytes = file.read()
+        if hasattr(file, "seek"):
+            with contextlib.suppress(OSError):
+                file.seek(0)
+        return f_bytes
+
     raise ValueError("Invalid file-like object type")