Unifying trajectories

[xukai92]

This PR unifies all trajectories under the same interface.
For example, NUTS shall be constructed by:

τ = Trajectory{MultinomialTS}(Leapfrog(1e-3), GeneralisedNoUTurn())

Old constructors are supported but deprecated.

Besides, the way to unify FixedNSteps and FixedIntegrationTime is via a new-introduced nsteps(::Trajectory), suggested by @torfjelde.

• Previously, FixedIntegrationTime was converted to FixedNSteps before calling transition, and termination_criterion.L appears here and there, which is not ideal.

[torfjelde]

LGTM! Just address my one comment. And, though it's not directly related to this PR, after we completely remove the NUTS struct it would still be useful to have a NUTS method which simply constructs the corresponding Trajectory so that people who aren't too familiar with the topic find what they expect.

Also, I really like this work:)

[xukai92]

CI passed now and it's ready for a final look @yebai

[xukai92]

The only CI which "fails" due to running out of time.
For some reaon the Geweke tests are pretty slow on Julia 1.0 (>4h) vs nightly (~1.5h).
I check previous CI logs and this was true before (this PR) as well.

[xukai92]

To be safe I just triggered a re-run of all CI. Let's see how it goes this time.

[yebai]

Excellent work - many thanks to Kai! I left some minor comments below.

[yebai]

Maybe consider unifying these type names in a future PR, e.g. FixedNSteps ==> FixedIntegrationSteps.

[xukai92]

Created an issue for this and some other potential naming improvements #246.

[xukai92]

Is this good to go?