I think Postgres Full Text Search is "good enough" for some internal use cases, but for user facing search I had a project where there are a lot of weird edge c

I think Postgres Full Text Search is "good enough" for some internal use cases, but for user facing search I had a project where there are a lot of weird edge cases that eventually makes it worth it to move something dedicated. I believed we used `websearch_to_tsquery` and users complained often enough where we swapped it out for quickwit. A big issue I remember is when users wanted to search for something and order by another field (such as find all documents matching query "x" order by creation date), and PG performance just took a massive nosedive. Even though we knew where was one or two fields we wanted to co-index with FTS, you can't have compound indexes when using FTS.