for text search the rum index is quite robust. my guess is that rum (or something like it) will be introduced into the pg core soon. we index many terabytes of pdf files excellent performance.

   https://github.com/postgrespro/rum

also, postgrespro are behind the json/b indexing.