docx to markdown, how to ignore certain things (like tables and images)

[michellepace]

Hello there.

Using Python in a Colab Notebook, I need to convert a Word into markdown so that I can send it to a large language model in chunks. The demo of Pandas looks brilliant! I've tried doing this myself by using panda-docx to read the docx and then writing the below function. I am so happy I found Pandoc and hope I can entirely replace panda-docx and my function with it.

My Problem:

• I only want to write Paragraphs, headings (a style of paragraph I believe), and Lists (another style of paragraph I think) to my markdown file (maybe Tables but I'm not yet sure).
• I would like to ignore images, table of contents, footers, headers, titles etc. Is it possible to do this and if so could you please share a reference url with me?
  

Thanks in advance and can't wait for a reply!
Michelle

[tarleb]

This can be done with a pandoc filter. Below is a Lua filter, which should work out-of-the-box, without having to install additional software. However, if you prefer a more pythonic solution, then checkout "panflute".

local function omit ()
  return {}
end

return {{
  Table = omit,
  Figure = omit,
  Image = omit,
  -- and so on.
}}

Save the above to a file and pass it to pandoc via the --lua-filter command line parameter.

[michellepace]

Thank you very much @tarleb , pandoc is lovey. Unfortunately though I need this to run in a Colab notebook and I didn’t realise I would have to install pandoc on my machine. But thank you very much for your reply. I think I’m stuck with the python-docx library.