at this moment we can only use text in text out, this would lose the format

potentially we wanna have interaction (the current ui)
meanwhile we should be able to upload the file, process and download the file

the current file processing example is under gpt/test.py
u can do

python test.py -s sample2.docx -d ~/Downloads/result.doc

we wanna leverage the code on the ui and server side to be able to upload and download.

at this moment the submit click, then the page will show running at the chrome tab only.

to avoid the confusion, we need to be able to
a) click on the submit
b) show the loading gif (the file is under static/loading.gif or u can generate online for anything)
c) once the text generation is ready, we wanna show the context