fix: issue when filename is provided but file is not on disk (#446)

Author: ryannikolaidis

CHANGELOG.md

@@ -1,4 +1,4 @@
-## 0.5.10-dev2
+## 0.5.10
 
 ### Enhancements
 
@@ -11,6 +11,8 @@
 
 ### Fixes
 
+* FileNotFound error when filename is provided but file is not on disk
+
 ## 0.5.9
 
 ### Enhancements

test_unstructured/file_utils/test_filetype.py

@@ -55,9 +55,17 @@ def test_detect_filetype_from_filename(file, expected):
     ],
 )
 def test_detect_filetype_from_filename_with_extension(monkeypatch, file, expected):
+    """Test that we detect the filetype from the filename extension when libmagic is not available
+    or the file does not exist."""
+    # Test when libmagic is not available
     monkeypatch.setattr(filetype, "LIBMAGIC_AVAILABLE", False)
     filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, file)
     assert detect_filetype(filename) == expected
+    # Test when the file does not exist
+    monkeypatch.setattr(filetype, "LIBMAGIC_AVAILABLE", True)
+    extension = pathlib.Path(file).suffix
+    filename = os.path.join(EXAMPLE_DOCS_DIRECTORY, "not-on-disk" + extension)
+    assert detect_filetype(filename) == expected
 
 
 @pytest.mark.parametrize(
@@ -243,6 +251,7 @@ def test_detect_filetype_detects_png(monkeypatch):
 
 def test_detect_filetype_detects_unknown_text_types_as_txt(monkeypatch):
     monkeypatch.setattr(magic, "from_file", lambda *args, **kwargs: "text/new-type")
+    monkeypatch.setattr(os.path, "isfile", lambda *args, **kwargs: True)
     assert detect_filetype(filename="made_up.png") == FileType.TXT
 
 

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.5.10-dev2"  # pragma: no cover
+__version__ = "0.5.10"  # pragma: no cover

unstructured/file_utils/filetype.py

@@ -177,12 +177,12 @@ def detect_filetype(
             return filetype
 
     if filename or file_filename:
-        _, extension = os.path.splitext(filename or file_filename or "")
+        _filename = filename or file_filename or ""
+        _, extension = os.path.splitext(_filename)
         extension = extension.lower()
-        if LIBMAGIC_AVAILABLE:
+        if os.path.isfile(_filename) and LIBMAGIC_AVAILABLE:
             mime_type = magic.from_file(filename or file_filename, mime=True)  # type: ignore
         else:
-            # might not need this
             return EXT_TO_FILETYPE.get(extension.lower(), FileType.UNK)
     elif file is not None:
         extension = None

unstructured/partition/auto.py

@@ -32,7 +32,7 @@ def partition(
 
     Parameters
     ----------
-     filename
+    filename
         A string defining the target filename path.
     content_type
         A string defining the file content in MIME type