Feature request: remove need for intermediate pl calls when chaining renderers

Description of feature

Thanks for the super nice work, @timtreis and @sagar87 ! I think it would be nice if we didn't have to add intermediate calls to pl when chaining renderers. That is, I think it would be nice if instead of this:

sdata.pl.render_images().pl.render_shapes().pl.show()

We could do this:

sdata.pl.render_images().render_shapes().show()

This perhaps could be possible if the render_*() methods returned the plot accessor instead of the SpatialData object. I suppose this means that one couldn't do a pp operation after a pl operation, but maybe that suggests the pl and pp operations should be on the same level of the hierarchy or one should do their preprocessing first, then their plotting. I'm not sure what's best here, but it might be good to experiment with this once all of the p0 functionality is finished.

Start by inspecting the pl accessor and the render_* methods, especially how they currently return SpatialData and support chained calls. Compare the existing sdata.pl.render_images().pl.render_shapes().pl.show() example with the proposed chaining behavior, and resolve how pl and pp operations should interact. Done means a decided API design and working renderer chaining without intermediate pl calls.