Add support for ggridges

[AdroMine]

Closes #1912

Add support for ggrides package

I took inspiration from the #1087 and have a version that supports most of the ggridges functionality. The test file has all the examples from ggridges introduction vignette

What's working:

• geom_ridgeline
• geom_density_ridges (including custom densities (stat = identity), bin shape, custom heights)
• geom_density_ridges2
• geom_ridgeline_gradient
• stat_density_ridges
• stat_denstity_ridges with gradients
• ridges with jittered points / raincloud effect, points variations

What's not working:

• geom_ridgeline_gradient with continuous colours
• geom_vridgeline

ggridge functions that already working without new code:

• themes (axis title placement is off)
• scale_fill_cyclical (custom label names don't work)

Examples:

(p <- ggplot(iris, aes(x = Sepal.Length, y = Species, fill = factor(stat(quantile)))) +
  stat_density_ridges(
    geom = "density_ridges_gradient",
    calc_ecdf = TRUE,
    quantiles = c(0.025, 0.975)
  ) +
  scale_fill_manual(
    name = "Probability", values = c("#FF0000A0", "#A0A0A0A0", "#0000FFA0"),
    labels = c("(0, 0.025]", "(0.025, 0.975]", "(0.975, 1]")
  ))

ggplotly(p)

(p2 <- ggplot(iris, aes(x = Sepal.Length, y = Species)) + 
  ggridges::geom_density_ridges(fill = 'midnightblue', alpha = 0.6))

ggplotly(p2)

[cpsievert]

Wow, incredible work, thank you! I'm going to add some nit-picky comments, but overall this is looking great

[cpsievert]

I wouldn't consider this a blocking issue, and I'm not sure if this is sensible to address here, but just throwing it out there that it'd be awesome if this worked better (more specifically, it seems highlight()'s opacityDim isn't having an effect on the non-highlighted ridgelines?)

p <- ggplot(iris |> highlight_key(~Species), aes(x = Sepal.Length, y = Species)) + 
  geom_density_ridges(rel_min_height = 0.01)

ggplotly(p) |> highlight("plotly_hover", opacityDim = 0.2)

[AdroMine]

I wouldn't consider this a blocking issue, and I'm not sure if this is sensible to address here, but just throwing it out there that it'd be awesome if this worked better (more specifically, it seems highlight()'s opacityDim isn't having an effect on the non-highlighted ridgelines?)

p <- ggplot(iris |> highlight_key(~Species), aes(x = Sepal.Length, y = Species)) + 
  geom_density_ridges(rel_min_height = 0.01)

ggplotly(p) |> highlight("plotly_hover", opacityDim = 0.2)

Didn't know about highlight. Any idea what could be affecting it? Or where I could look to debug?

[cpsievert]

I think these need be doubled, like you did for x/y (this might be leading to #2314 (comment))

I'd consider doing something like (using with() here can be kinda dangerous since it'll error if any of the columns are missing

positions <- rbind(data, rev(data))
positions$y <- c(data$ymax, rev(data$ymin))

[AdroMine]

Found the issue! Within the new to_basics, I was using to_basic.GeomDensity and to_basic.GeomLine which removed the crosstalk set. Attaching them back again similar to how it's done in layers2traces solves the highlight not working.

Before this, the to_basic.GeomLine would preserve the set attribute, but to_basic.GeomDensity would not, so we were getting highlight working only the line part and not density/area part.

I have tested the highlight functionality with different types of ridge plots, with different fills, points, lines within them, and they are all working as expected.

[AdroMine]

@cpsievert are there any pending action items on this, or is this good to go now?

[cpsievert]

Sorry for letting this slip @AdroMine, looking good, would you mind just adding a NEWS.md item?

[AdroMine]

@cpsievert , added a line about adding support for ggridges package.

[cpsievert]

Awesome work, thank you!