FileConversionException: Could not convert ''

Hi,

When I try to extract information from some pdf from DocLayNet dataset with md.convert, I get this error :

---------------------------------------------------------------------------
FileConversionException                   Traceback (most recent call last)
Cell In[3], [line 45](vscode-notebook-cell:?execution_count=3&line=45)
     [41](vscode-notebook-cell:?execution_count=3&line=41) page_hash = doc[:doc.find('.pdf')]
     [44](vscode-notebook-cell:?execution_count=3&line=44) start_time = time.time()
---> [45](vscode-notebook-cell:?execution_count=3&line=45) conv_result = md.convert(str(doc_path))
     [46](vscode-notebook-cell:?execution_count=3&line=46) diff_time = time.time() - start_time
     [47](vscode-notebook-cell:?execution_count=3&line=47) print(f"Time computing : {diff_time} s")

File c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\markitdown\_markitdown.py:1094, in MarkItDown.convert(self, source, **kwargs)
   [1092](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1092)         return self.convert_url(source, **kwargs)
   [1093](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1093)     else:
-> [1094](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1094)         return self.convert_local(source, **kwargs)
   [1095](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1095) # Request response
   [1096](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1096) elif isinstance(source, requests.Response):

File c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\markitdown\_markitdown.py:1114, in MarkItDown.convert_local(self, path, **kwargs)
   [1111](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1111)     self._append_ext(extensions, g)
   [1113](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1113) # Convert
-> [1114](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1114) return self._convert(path, extensions, **kwargs)

File c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\markitdown\_markitdown.py:1255, in MarkItDown._convert(self, local_path, extensions, **kwargs)
   [1253](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1253) # If we got this far without success, report any exceptions
   [1254](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1254) if len(error_trace) > 0:
-> [1255](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1255)     raise FileConversionException(
   [1256](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1256)         f"Could not convert '{local_path}' to Markdown. File type was recognized as {extensions}. While converting the file, the following error was encountered:\n\n{error_trace}"
   [1257](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1257)     )
   [1259](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1259) # Nothing can handle it!
   [1260](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1260) raise UnsupportedFormatException(
   [1261](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1261)     f"Could not convert '{local_path}' to Markdown. The formats {extensions} are not supported."
   [1262](file:///C:/Users/AppData/Local/miniconda3/envs/myenv/Lib/site-packages/markitdown/_markitdown.py:1262) )

FileConversionException: Could not convert 'E:\users\.cache\huggingface\hub\datasets--pierreguillou--DocLayNet-large\snapshots\38ff443244c1b496c33ed237d3d4468daf24265c\data\part_dataset_3\part_dataset_3\test\pdfs\ccbe08f3390d47046dbb9d4c839788ba05a0f5e139ab6931a06e8304247c54f0.pdf' to Markdown. File type was recognized as ['.pdf', '.pdf', '.fdf']. While converting the file, the following error was encountered:

Traceback (most recent call last):
  File "c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\markitdown\_markitdown.py", line 1239, in _convert
    res = converter.convert(local_path, **_kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\markitdown\_markitdown.py", line 490, in convert
    text_content=pdfminer.high_level.extract_text(local_path),
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\pdfminer\high_level.py", line 169, in extract_text
    for page in PDFPage.get_pages(
                ^^^^^^^^^^^^^^^^^^
  File "c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\pdfminer\pdfpage.py", line 171, in get_pages
    for (pageno, page) in enumerate(cls.create_pages(doc)):
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\pdfminer\pdfpage.py", line 127, in create_pages
    yield cls(document, objid, tree, next(page_labels))
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "c:\Users\AppData\Local\miniconda3\envs\myenv\Lib\site-packages\pdfminer\pdfpage.py", line 64, in __init__
    resolve1(mediabox_param) for mediabox_param in self.attrs["MediaBox"]
                                                   ~~~~~~~~~~^^^^^^^^^^^^
TypeError: 'PDFObjRef' object is not iterable

Is someone have already encountered this issue? It's really strange because the document is not an .fdf file but a .pdf one.

Start in markitdown/_markitdown.py at MarkItDown.convert_local, _convert, and the PDF converter path shown in the traceback. Reproduce the failure with the reported DocLayNet PDF and inspect how pdfminer handles its MediaBox and recognized extensions. Done should mean the failure is correctly handled or the affected PDF converts successfully, with regression coverage if the repository has a suitable test location.