Density plot functionality about dittoseq HOT 1 CLOSED

For the underlying stat/geom:
I like the idea of stat_binhex() by default + stat_density_2d() when requested with something like do.contour. Alternatively, or additionally only when color is requested, we could use ggplot.multistats::geom_summaries_hex() to show density through opacity, thus leaving color open for some other target variable. But, I'll just admit now that I have almost zero experience using any of these, so ¯_(ツ)_/¯. That's definitely why dittoSeq doesn't already have a hex-functionality.

For the implementation:
We could modify dittoScatterPlot to be the base, and it's what I'd always thought would happen.
But looking into the geoms more, I don't love the idea because there are quite a few inputs in dittoScatterPlot that are not applicable. Also, if there end up being hex specific inputs, these would need to be exposed within the dittoScatterPlot doc. I think that all then gets confusing for users, so I don't like it.

Ultimately, no matter how we implement, we'll need to have a separate .ditto_scatter_hex() sub-function. Otherwise it'd get super dirty because half the code in .ditto_scatter_plot() parses shape or color. (Shape is useless for hexPlots, and coloring is implemented quite differently from geom_point, from what I can tell, for all hex geoms that handle it.)

So I propose we make a .scatter_data_gather() subfunction that would be shared between dittoScatterPlot() and a dittoScatterHex(). Then we can have .ditto_scatter_hex() build the plot with whatever geoms we choose. And dittoDimHex() can wrap around all that similarly to dittoDimPlot().

The "good news" here is that I don't think dittoDimHex() will have as many extra overlays as dittoDimPlot() -- really just the trajectory overlays -- so we could probably group dittoScatterHex() and dittoDimHex() together into a single piece of documentation.

And this is now implemented in #46

from dittoseq.