Mpl-interactions

145 Fixed a bug where `imshow`, `hyperslicer` and `scatter` would break if they were not passed the current axis.:
python
fig, axs = plt.subplots(1, 2)
iplt.imshow(img, ax=axs[0])

would break because this would ultimately call `gca()._sci(im)` but `gca` would return `axs[1]` which was empty.

New features:
Created the `scatter_selector` widgets + allowed mixing of ipywidgets and matplotlib widgets. 141

This enables fun things like this:

New features:
`hyperslicer` now natively supports taking xarray objects and infers the approriate sliders from them. https://github.com/ianhi/mpl-interactions/pull/138

Thanks to: jrussell25 for contributing to this release

bug fixes:
`image_segmenter` wasn't making use of the `cmap` argument. Now all kwargs are passed through to `imshow`. Thanks to stefanopiovesan for finding this bug. https://github.com/ianhi/mpl-interactions/issues/135

Improvements:
`image_segmenter` now allows styling of the lineprops for the lasso selector.

All the above is documented here: https://mpl-interactions.readthedocs.io/en/latest/examples/image-segmentation.html#Styling

Updated the old mpl-sliders docs page and fixed a few bugs with passing pre-created mpl-sliders along the way. Moving docs to a notebook also introduced tests for that.

Allow hyperslicers to share controller objects. This allows doing stuff like this:

python
%matplotlib ipympl
import matplotlib.pyplot as plt
from mpl_interactions import hyperslicer
from skimage.data import binary_blobs

arr1 = []
arr2 = []
for i in range(10):
arr1.append(binary_blobs(length=128,n_dim=2))
arr2.append(binary_blobs(length=128,n_dim=2))
plt.figure()
controls = hyperslicer(arr1, play_buttons=True, cmap='jet')
_ = hyperslicer(arr2, play_buttons=True,alpha=.5, controls = controls)