Export fails with complex table heading

[m-rossi]

Describe the bug
If I try to export a table with a rather complex table heading, for example created by pint_pandas I get an error.

To reproduce
I have a notebook with the following content

import pandas as pd
import numpy as np

df = pd.DataFrame(np.random.randn(6, 4), columns=[list("1122"), list("ABCD")])
df.index.name = "custom-index"
df.columns.names = (None, "unit")

An export fails with the following error

[Application] ERROR | nbconvert failed: Must pass non-zero number of levels/codes
Traceback (most recent call last):
  File "C:\minforge\lib\site-packages\jupyter_server\nbconvert\handlers.py", line 131, in get
    output, resources = await run_sync(
  File "C:\minforge\lib\site-packages\anyio\to_thread.py", line 31, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "C:\minforge\lib\site-packages\anyio\_backends\_asyncio.py", line 937, in run_sync_in_worker_thread
    return await future
  File "C:\minforge\lib\asyncio\futures.py", line 284, in __await__
    yield self  # This tells Task to wait for completion.
  File "C:\minforge\lib\asyncio\tasks.py", line 304, in __wakeup
    future.result()
  File "C:\minforge\lib\asyncio\futures.py", line 201, in result
    raise self._exception
  File "C:\minforge\lib\site-packages\anyio\_backends\_asyncio.py", line 867, in run
    result = context.run(func, *args)
  File "C:\minforge\lib\site-packages\jupyter_server\nbconvert\handlers.py", line 132, in <lambda>
    lambda: exporter.from_notebook_node(nb, resources=resource_dict)
  File "C:\minforge\lib\site-packages\jupyter_docx_bundler\__init__.py", line 79, in from_notebook_node
    converters.notebookcontent_to_docxbytes(
  File "C:\minforge\lib\site-packages\jupyter_docx_bundler\converters.py", line 262, in notebookcontent_to_docxbytes
    content = preprocess(content, path, handler=handler)
  File "C:\minforge\lib\site-packages\jupyter_docx_bundler\converters.py", line 193, in preprocess
    raise e
  File "C:\minforge\lib\site-packages\jupyter_docx_bundler\converters.py", line 185, in preprocess
    html_to_pandas_table(output['data']['text/html']).to_markdown(),
  File "C:\minforge\lib\site-packages\jupyter_docx_bundler\converters.py", line 94, in html_to_pandas_table
    df.set_index(index_column_names, inplace=True)
  File "C:\minforge\lib\site-packages\pandas\util\_decorators.py", line 311, in wrapper
    return func(*args, **kwargs)
  File "C:\minforge\lib\site-packages\pandas\core\frame.py", line 5555, in set_index
    index = ensure_index_from_sequences(arrays, names)
  File "C:\minforge\lib\site-packages\pandas\core\indexes\base.py", line 6984, in ensure_index_from_sequences
    return MultiIndex.from_arrays(sequences, names=names)
  File "C:\minforge\lib\site-packages\pandas\core\indexes\multi.py", line 493, in from_arrays
    return cls(
  File "C:\minforge\lib\site-packages\pandas\core\indexes\multi.py", line 321, in __new__
    raise ValueError("Must pass non-zero number of levels/codes")
ValueError: Must pass non-zero number of levels/codes

Desktop (please complete the following information):

• OS: Windows 10
• Browser: Firefox
• Pandoc version: 2.13