Jupyter-scatter

- **Breaking changes:**
- For `scatter.color()`, rename `color` to `default`, `color_selected` to `selected`, and `color_hover` and `hover`
- For `scatter.opacity()`, rename `opacity` to `default`
- For `scatter.size()`, rename `size` to `default`
- For `scatter.connection_color()`, rename `color` to `default`, `color_selected` to `selected`, and `color_hover` and `hover`
- For `scatter.connection_opacity()`, rename `opacity` to `default`
- For `scatter.connection_size()`, rename `size` to `default`
- Add `scatter.opacity(unselected=0.5)`. This property defines the opacity scaling factor of unselected points and must be in `[0, 1]`. This scaling is only applied if one or more points are selected.

- Fix an issue when dynamically resizing the scatter plot height
- Fix an issue when switching from categorical to continuous color/opacity/size encoding

- **Breaking change:** Renamed `color_active` and `connection_color_active` to `color_selected` and `connection_color_selected` respectively for clarity.
- Add support for axes via `Scatter(axes=True, axes_grid=True)` or `scatter.axes(True, grid=True)`.
- Add support for log and power x/y scales via `Scatter(x_scale='log')` or `scatter.y(scale='pow')`.
- Add docstrings and type hints

- Unset data-driven color, opacity, and size encoding when only a constant value was passed to an encoder function. E.g., after initializing `scatter = Scatter(..., color_by='property')`, calling `scatter.color('red')` will automatically unset `color_by`.

- Fix `scatter.pixels()`

- Fix more type hints...