Nah, they released a bunch of the piping that goes around an ML model, but that's very boilerplate stuff to anyone with python data science experience. It's like if you made calculators and "open sourced" one by releasing the code that reads keypresses and does CPU scheduling. Like yeah, those are things a calculator does but not only is it code everyone else in the business could expect to write themselves in a couple sprints but it's also exactly NOT the magic sauce.