I have a pptx which leads to an error when trying to ``MarkItDown().convert()`` it. Traceback is: ``` Traceback (most recent call last): [...] File "/var/lang/lib/python3.12/site-packages/markitdown/_markitdown.py", line 273, in convert return self.convert_stream(source, stream_info=stream_info, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/var/lang/lib/python3.12/site-packages/markitdown/_markitdown.py", line 361, in convert_stream return self._convert(file_stream=stream, stream_info_guesses=guesses, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "/var/lang/lib/python3.12/site-packages/markitdown/_markitdown.py", line 600, in _convert raise FileConversionException(attempts=failed_attempts) markitdown._exceptions.FileConversionException: File conversion failed after 1 attempts: - PptxConverter threw TypeError with message: '<' not supported between instances of 'NoneType' and 'Emu' ``` The proplematic pptx is confidential, so I cannot provide it. But I could pin the bug down to [this line of code](https://github.com/microsoft/markitdown/blob/da7bcea527ed04cf6027cc8ece1e1aad9e08a9a1/packages/markitdown/src/markitdown/converters/_pptx_converter.py#L175). There is indeed for one shape ``shape.top == None``, so the sorting fails. The problematic shape seems to be empty anyways. Currently, I use a very ugly monkey patch: ```python def _shape_filter(s): return not ( # if "top" and "left" attributes exist, both of them must not be None hasattr(s, "top") and hasattr(s, "left") and (s.top is None or s.left is None) ) def _mock_sorted(iterable, **kwargs): iterable = (it for it in iterable if _shape_filter(it)) return sorted(iterable, **kwargs) from markitdown.converters import _pptx_converter _pptx_converter.sorted = _mock_sorted # type: ignore ```