OPUS data dumps and usage

I recently came across [OPUS](https://opus.nlpl.eu/) which is an open parallel corpus.
This is an open repository of translated text from many languages to many languages building a huge open source language pairs dictionary.
There are several open source projects contributing to this effort such as Ubuntu, GNOME, KDE, Wikipedia, PHP, Tatoeba, etc.
I think that contributing the Weblate repository (where license allows it) to this repository could be a wonderful contribution to open source localization plaforms and TM.
Furthermore, I think that Weblate could benefit by using this TM yet I'm not sure they're offering unlimited hosting (they are offerring al their data as open source).