I have gone through code from several papers on github. It looks they just used some IDE since they only have .py scripts. So this makes me skeptical that jupyter notebooks are being adopted at the research level. Please correct me if I am wrong.