fix: prioritize extension-based detection over content sniffing in _get_filetype

Move extension check before content-based detection to ensure
deterministic behavior across Python versions. Content-based detection
in unstructured 0.18.18 may return TXT for plain text content, which
could mask the actual file type indicated by the extension.

Co-Authored-By: unknown <>

Author: devin-ai-integration[bot]

airbyte_cdk/sources/file_based/file_types/unstructured_parser.py

@@ -424,18 +424,18 @@ def _get_filetype(self, file: IOBase, remote_file: RemoteFile) -> Optional[FileT
         if file_type and file_type != FileType.UNK:
             return file_type
 
-        type_based_on_content = detect_filetype(file=cast(IO[bytes], file))
-        file.seek(0)  # detect_filetype is reading to read the file content, so we need to reset
-
-        if type_based_on_content and type_based_on_content != FileType.UNK:
-            return type_based_on_content
-
         extension = "." + remote_file.uri.split(".")[-1].lower()
         try:
             return FileType.from_extension(extension)
         except ValueError:
             pass
 
+        type_based_on_content = detect_filetype(file=cast(IO[bytes], file))
+        file.seek(0)  # detect_filetype is reading to read the file content, so we need to reset
+
+        if type_based_on_content and type_based_on_content != FileType.UNK:
+            return type_based_on_content
+
         return None
 
     def _supported_file_types(self) -> List[Any]: