fix: properly handle password protected xlsx (#4057)

### Issue
Attempt at partitioning a password protected errors results in an
obscure exception
> Can't find workbook in OLE2 compound document

### Solution
Utilize [msoffcrypto-tool](https://pypi.org/project/msoffcrypto-tool/)
package (MIT License) to load XLSX file and check whether it's
encrypted, if yes throw an `UnprocessableEntityError` exception
detailing the reason for rejecting the file.

---------

Co-authored-by: Filip Knefel <[email protected]>

Author: ds-filipknefel

CHANGELOG.md

@@ -1,3 +1,12 @@
+## 0.18.8
+
+### Enhancements
+
+### Features
+
+### Fixes
+- **Properly handle password protected xlsx** - detect password protection on XLSX files and raise appropriate
+
 ## 0.18.7
 
 ### Enhancements

example-docs/password_protected.xlsx

@@ -5,3 +5,4 @@ openpyxl
 pandas
 xlrd
 networkx
+msoffcrypto-tool

requirements/extra-xlsx.in

@@ -4,18 +4,36 @@
 #
 #    pip-compile ./extra-xlsx.in
 #
+cffi==1.17.1
+    # via
+    #   -c requirements/base.txt
+    #   cryptography
+cryptography==45.0.5
+    # via
+    #   -c requirements/base.txt
+    #   msoffcrypto-tool
 et-xmlfile==2.0.0
     # via openpyxl
+msoffcrypto-tool==5.4.2
+    # via -r ./extra-xlsx.in
 networkx==3.4.2
     # via -r ./extra-xlsx.in
 numpy==2.2.6
     # via
     #   -c requirements/base.txt
     #   pandas
+olefile==0.47
+    # via
+    #   -c requirements/base.txt
+    #   msoffcrypto-tool
 openpyxl==3.1.5
     # via -r ./extra-xlsx.in
 pandas==2.3.1
     # via -r ./extra-xlsx.in
+pycparser==2.22
+    # via
+    #   -c requirements/base.txt
+    #   cffi
 python-dateutil==2.9.0.post0
     # via
     #   -c requirements/base.txt

requirements/extra-xlsx.txt

@@ -28,6 +28,7 @@
 )
 from unstructured.cleaners.core import clean_extra_whitespace
 from unstructured.documents.elements import ListItem, Table, Text, Title
+from unstructured.errors import UnprocessableEntityError
 from unstructured.partition.xlsx import (
     _ConnectedComponent,
     _SubtableParser,
@@ -168,6 +169,11 @@ def test_partition_xlsx_from_file_with_header():
     assert e.metadata.text_as_html is not None
 
 
+def test_partition_xlsx_password_protected_raises_exception():
+    with pytest.raises(UnprocessableEntityError):
+        partition_xlsx(filename="example-docs/password_protected.xlsx")
+
+
 # -- .metadata.last_modified ---------------------------------------------------------------------
 
 

test_unstructured/partition/test_xlsx.py

@@ -1 +1 @@
-__version__ = "0.18.7"  # pragma: no cover
+__version__ = "0.18.8"  # pragma: no cover

unstructured/__version__.py

@@ -3,12 +3,13 @@
 from __future__ import annotations
 
 import io
-from tempfile import SpooledTemporaryFile
 from typing import IO, Any, Iterator, Optional
 
 import networkx as nx
 import numpy as np
 import pandas as pd
+from msoffcrypto import OfficeFile
+from msoffcrypto.exceptions import FileFormatError
 from typing_extensions import Self, TypeAlias
 
 from unstructured.chunking import add_chunking_strategy
@@ -23,6 +24,7 @@
     Text,
     Title,
 )
+from unstructured.errors import UnprocessableEntityError
 from unstructured.file_utils.model import FileType
 from unstructured.partition.common.metadata import apply_metadata, get_last_modified_date
 from unstructured.partition.text_type import (
@@ -187,16 +189,28 @@ def metadata_file_path(self) -> str | None:
     @lazyproperty
     def sheets(self) -> dict[str, pd.DataFrame]:
         """The spreadsheet worksheets, each as a data-frame mapped by sheet-name."""
-        if file_path := self._file_path:
-            return pd.read_excel(file_path, sheet_name=None, header=self.header_row_idx)
+        try:
+            office_file = OfficeFile(io.BytesIO(self._file_bytes))
+        except FileFormatError as e:
+            raise UnprocessableEntityError("Not a valid XLSX file.") from e
 
-        if f := self._file:
-            if isinstance(f, SpooledTemporaryFile):
-                f.seek(0)
-                f = io.BytesIO(f.read())
-            return pd.read_excel(f, sheet_name=None, header=self.header_row_idx)
+        if office_file.is_encrypted():
+            raise UnprocessableEntityError("XLSX file is password protected.")
 
-        raise ValueError("Either 'filename' or 'file' argument must be specified.")
+        return pd.read_excel(
+            io.BytesIO(self._file_bytes), sheet_name=None, header=self.header_row_idx
+        )
+
+    @lazyproperty
+    def _file_bytes(self) -> bytes:
+        if file := self._file:
+            file.seek(0)
+            return file.read()
+        elif self._file_path:
+            with open(self._file_path, "rb") as file:
+                return file.read()
+        else:
+            raise ValueError("Either 'filename' or 'file' argument must be specified.")
 
 
 class _ConnectedComponent: