fix: normalize data URI parameter case

Author: pony-maggie

packages/markitdown/src/markitdown/_uri_utils.py

@@ -29,7 +29,7 @@ def parse_data_uri(uri: str) -> Tuple[str | None, Dict[str, str], bytes]:
 
     is_base64 = False
     # Ends with base64?
-    if parts[-1] == "base64":
+    if parts[-1].lower() == "base64":
         parts.pop()
         is_base64 = True
 
@@ -43,9 +43,9 @@ def parse_data_uri(uri: str) -> Tuple[str | None, Dict[str, str], bytes]:
         # Handle key=value pairs in the middle
         if "=" in part:
             key, value = part.split("=", 1)
-            attributes[key] = value
+            attributes[key.lower()] = value
         elif len(part) > 0:
-            attributes[part] = ""
+            attributes[part.lower()] = ""
 
     content = base64.b64decode(data) if is_base64 else unquote_to_bytes(data)
 

packages/markitdown/tests/test_module_misc.py

@@ -200,6 +200,13 @@ def test_data_uris() -> None:
     assert attributes["charset"] == "utf-8"
     assert data == b"Hello, World!"
 
+    data_uri = "data:text/plain;CHARSET=utf-8;BASE64,SGVsbG8sIFdvcmxkIQ=="
+    mime_type, attributes, data = parse_data_uri(data_uri)
+    assert mime_type == "text/plain"
+    assert len(attributes) == 1
+    assert attributes["charset"] == "utf-8"
+    assert data == b"Hello, World!"
+
     data_uri = "data:,Hello%2C%20World%21"
     mime_type, attributes, data = parse_data_uri(data_uri)
     assert mime_type is None