Add tests for file URIs on Windows

packages/markitdown/tests/test_module_misc.py

@@ -2,6 +2,7 @@
 import io
 import os
 import re
+import sys
 import shutil
 import pytest
 from unittest.mock import MagicMock
@@ -219,8 +220,8 @@ def test_data_uris() -> None:
     assert attributes["charset"] == "utf-8"
     assert data == b"Hello, World!"
 
-
-def test_file_uris() -> None:
+@pytest.mark.skipif(sys.platform == "win32", reason="POSIX-specific test")
+def test_file_uris_posix() -> None:
     # Test file URI with an empty host
     file_uri = "file:///path/to/file.txt"
     netloc, path = file_uri_to_path(file_uri)
@@ -251,6 +252,37 @@ def test_file_uris() -> None:
     assert netloc is None
     assert path == "/path/to/file.txt"
 
+@pytest.mark.skipif(sys.platform != "win32", reason="Windows-specific test")
+def test_file_uris_windows() -> None:
+    # Test file URI with an empty host
+    file_uri = "file:///C:/path/to/file.txt"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "C:\\path\\to\\file.txt"
+
+    # Test file URI with no host
+    file_uri = "file:/C:/path/to/file.txt"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "C:\\path\\to\\file.txt"
+
+    # Test file URI with localhost
+    file_uri = "file://localhost/C:/path/to/file.txt"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc == "localhost"
+    assert path == "C:\\path\\to\\file.txt"
+
+    # Test file URI with query parameters
+    file_uri = "file:///C:/path/to/file.txt?param=value"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "C:\\path\\to\\file.txt"
+
+    # Test file URI with fragment
+    file_uri = "file:///C:/path/to/file.txt#fragment"
+    netloc, path = file_uri_to_path(file_uri)
+    assert netloc is None
+    assert path == "C:\\path\\to\\file.txt"
 
 def test_docx_comments() -> None:
     # Test DOCX processing, with comments and setting style_map on init