Made changes to the notebook to fit current pypkg setup.

[Sarthakmistry]

1. Replaced debian pull with pip install of openbt_pypkg.
2. Replaced Tawaret pull with openbtmixing in current python package.

[jared321]

• Reviewed all changes
  • No quantitative info printed by notebook to assist in checking if results significantly different.
  • Compared graphics using eyeball metric. While they are generally unchanged, there are slight differences. Not clear to me if these differences are expected or reasonable.
• Loaded notebook locally and ran successfully
• Confirmed all actions passing
• Confirmed that main has not been updated since it was last synched with this branch

Matt confirmed that notebook is nondeterministic at the moment. I also pointed out a sign change in the upper-left corner of the left 2D plot at the bottom. Matt responded that

Yes, that top-left corner is the unstable / high variance part of the fit. It’s not a code reproducibility issue, it is a limitation of the model itself, which was later fixed in the RPBART model.