Mpl-interactions

Matplotlib sliders would break with `numpy < 1.19`. Found and fixed by mwshinn in 155

Hyperslicer support for `vmin_vmax`
python
%matplotlib widget
import numpy as np
import matplotlib.pyplot as plt
import mpl_interactions.ipyplot as iplt
from mpl_interactions import hyperslicer
arr = np.random.randn(5, 100,100)
fig, axs = plt.subplots(1,2)
ctrls = hyperslicer(arr, ax=axs[0], vmin_vmax=("r", arr.min(), arr.max()))
def f(axis0):
return arr[axis0].flatten()
_ = iplt.hist(f, controls = ctrls['axis0'], ax=axs[1])
_ = iplt.axvline(ctrls['vmin'],ax = axs[1], color='k')
_ = iplt.axvline(ctrls['vmax'],ax = axs[1], color='k')



Big PR: https://github.com/ianhi/mpl-interactions/pull/150

- Scalar arguments can passed tuples to directly create sliders
- you can index a controls object directly into a scalar argument
- created a range slider widget
- added vmin_vmax kwargs
- more pass through of matplotlib kwargs

Fixed an issue with indexer where it would break when passed positional args. e.g. in scatter 152

Fixed xarray missing from hyperslicer docstring

Also gets the stable version to render on RTD they were down when last release happened so builds failed.

Implemented the ability to save animations generated from a `controls` object. 147