chore: raise file not found for images (#19)

When trying to open an image in the from_image_file method, trying to open an empty filename would produce some non-helpful error, when ideally it should raise FileNotFoundError. This PR ensures that a FileNotFoundError is raised.

Author: qued

CHANGELOG.md

@@ -1,6 +1,6 @@
-## 0.2.4-dev0
+## 0.2.4-dev1
 
-* Requirements bumps (delete this line with first real CHANGELOG cahnge)
+* Clarify error when trying to open file that doesn't exist as an image
 
 ## 0.2.3
 

test_unstructured_inference/inference/test_layout.py

@@ -1,4 +1,5 @@
 import pytest
+import tempfile
 from unittest.mock import patch, mock_open
 
 import layoutparser as lp
@@ -267,3 +268,13 @@ def mock_get_elements(self, *args, **kwargs):
         .elements
     )
     assert elements[0] == mock_page_layout
+
+
+def test_from_image_file_raises_with_empty_fn():
+    with pytest.raises(FileNotFoundError):
+        layout.DocumentLayout.from_image_file("")
+
+
+def test_from_image_file_raises_isadirectoryerror_with_dir():
+    with tempfile.TemporaryDirectory() as tempdir, pytest.raises(IsADirectoryError):
+        layout.DocumentLayout.from_image_file(tempdir)

unstructured_inference/__version__.py

@@ -1 +1 @@
-__version__ = "0.2.4-dev0"  # pragma: no cover
+__version__ = "0.2.4-dev1"  # pragma: no cover

unstructured_inference/inference/layout.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 from dataclasses import dataclass
+import os
 import re
 import tempfile
 from typing import List, Optional, Tuple, Union, BinaryIO
@@ -76,7 +77,13 @@ def from_file(cls, filename: str, model: Optional[Detectron2LayoutModel] = None)
     def from_image_file(cls, filename: str, model: Optional[Detectron2LayoutModel] = None):
         """Creates a DocumentLayout from an image file."""
         logger.info(f"Reading image file: {filename} ...")
-        image = Image.open(filename)
+        try:
+            image = Image.open(filename)
+        except Exception as e:
+            if os.path.isdir(filename) or os.path.isfile(filename):
+                raise e
+            else:
+                raise FileNotFoundError(f'File "{filename}" not found!') from e
         page = PageLayout(number=0, image=image, layout=None, model=model)
         page.get_elements()
         return cls.from_pages([page])