Setup crowdin integration for pyopensci translations

[lwasser]

Currently, our translations are done via .po files which is great. But it might be nice to have a Crowdin option that people could use to handle and manage languages.

if we had this, we would also want to setup a sync action that synced translations from crowdin to gh and gh to crowdin.