Kramersmoyal

Publishing the release here as well as in `pypi`. As always


pip install kramersmoyal


* `bins` can be `int`s, a `list` of `int`s, e.g. `[50, 50]`. If a specific binning is desired, `[np.ndarray]` works.
* `powers` can be `int`
* `full=True` outputs `bw` and `powers`

Type hinting was added. Back compatibility should be ok.

With the [JOSS](https://joss.theoj.org/) publication of [kramersmoyal: Kramers-Moyal coefficients for stochastic processes](https://joss.theoj.org/papers/f8158e61d51e0a9822dcc69a4a3c1f4d) eminent, follows here the first official GitHub release of version v0.4 of `kramersmoyal`.

You can find the JOSS publication's open-review process main thread [here](https://github.com/openjournals/joss-reviews/issues/1693).