IDEA: Process FOSS projects .po files for translation corpora

[Bugsbane]

Sorry if this is blindingly obvious / done / impractical, but... the first article I read about this project was saying how the projects biggest challenge was finding sources of phrase by phrase translations between languages, and it hit me that there are thousands of FOSS projects often with hundreds / thousands of machine readable translations usually between multiple languages under licenses that can be verified as being copyleft/permissive (depending on your preference).

Why not take advantage of this?