Create jupyter kernel for PZ

[heather999]

Chatting with Alex Malz now. we should make a list of what to include:

• RAIL and friends

[heather999]

all the dependencies of rail_hub, which is what we're using to separate out the heavyweight estimators wrapped for RAIL that have such nontrivial unit tests that they're slowing down development. Note that that repo is very unstable at the moment, so this project should probably wait for the dust to settle! Anyway, the only catch is that the dependencies for the sequestered estimators (BPZ, Delight, and FlexZBoost so far, likely more to come soon) have increasingly messy dependencies between them and we'll have to regularly check that core RAIL updates (e.g. requiring Python 3.9 instead of 3.7) don't break the wrapped estimators that are in separate repos. But I want to minimize the activation energy for DESC members to wrap their codes (i.e. not having to install all the other estimators just to get theirs set up) as well as to build and run pipelines made with RAIL that can access as many estimators as possible.