Early stopping tutorial

Summary:
Tutorial demonstrating early stopping via the Client using modified hartmann6.

WIP:
* Something is wrong with pending_points, we are seeing early stopped trials getting suggested again. Will investigate.
* I am still a little skeptical about the erogonomics around requiring raw_data and progression when we call mark_trial_early_stopped -- this is substantially different than what we do for mark_trial_failed or mark_trial_abandoned and seems like it would basically always involve us attaching data, asking whether or not to early stop the trial, then attaching the same data again. I know we're concerned about a specific footgun for certain internal users, but Im curious if thinking has changed on this since we last spoke saitcakmak

Differential Revision: D69484561