feat: added new exception type to epub conversions

.gitignore

@@ -212,3 +212,5 @@ annotated/
 .aider*
 pcaps
 python-output
+
+.vs/

CHANGELOG.md

@@ -1,3 +1,12 @@
+## 0.18.6-dev0
+
+### Enhancements
+
+### Features
+
+### Fixes
+- **Improved epub partition errors** EPUB partition will now produce new type of error on unprocessable files.
+
 ## 0.18.5
 
 ### Enhancements

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.18.5"  # pragma: no cover
+__version__ = "0.18.6-dev0"  # pragma: no cover

unstructured/errors.py

@@ -9,3 +9,9 @@ def __init__(self, document_pages: int, pdf_hi_res_max_pages: int):
             f"pages={document_pages}, maximum={pdf_hi_res_max_pages}."
         )
         super().__init__(self.message)
+
+
+class UnprocessableEpubError(Exception):
+    """Error raised when the epub file is not valid."""
+
+    pass

unstructured/file_utils/file_conversion.py

@@ -1,9 +1,11 @@
 from __future__ import annotations
 
 import os
+import re
 import tempfile
 from typing import IO
 
+from unstructured.errors import UnprocessableEpubError
 from unstructured.partition.common.common import exactly_one
 from unstructured.utils import requires_dependencies
 
@@ -23,6 +25,14 @@ def convert_file_to_text(filename: str, source_format: str, target_format: str)
         )
         raise FileNotFoundError(msg)
     except RuntimeError as err:
+        err_str = str(err)
+        if source_format == "epub" and (
+            "Couldn't extract ePub file" in err_str
+            or "No entry on path" in err_str
+            or re.search(r"exitcode ['\"]?64['\"]?", err_str)
+        ):
+            raise UnprocessableEpubError(f"Invalid EPUB file: {err_str}")
+
         supported_source_formats, _ = pypandoc.get_pandoc_formats()
 
         if source_format == "rtf" and source_format not in supported_source_formats: