Allow for discrete colors

[andrewheiss]

Viridis includes a discrete = TRUE argument in scale_*_viridis() that allows the colors to be discretized. And with the most recent version of ggplot2/viridis, there are now two types of scales: scale_*_viridis_c() and scale_*_viridis_d()

Currently scale_*_scico() only works with continuous colors and the only way to get discrete colors is to generate a palette with scico(x) and use the hex values manually.

It would be nice to add discretization to the scale_*_scico() functions and bring it in line with viridis.

[andrewheiss]

Here's a reprex:

library(tidyverse)
library(scico)

dsamp <- diamonds[sample(nrow(diamonds), 1000), ]

# Using scale_color_viridis_d()
ggplot(dsamp, aes(carat, price)) +
  geom_point(aes(colour = clarity)) +
  scale_color_viridis_d(begin = 0.2)

# Can't use scale_color_scico()
ggplot(dsamp, aes(carat, price)) +
  geom_point(aes(colour = clarity)) +
  scale_color_scico(begin = 0.2)
# Error: Discrete value supplied to continuous scale

# Can use scio() within scale_color_manual(), but must manually specify the n
ggplot(dsamp, aes(carat, price)) +
  geom_point(aes(colour = clarity)) +
  scale_color_manual(values = scico(8, begin = 0.2, palette = "roma"))

[thomasp85]

Thanks. I think @clauswilke already convinced me to add it but good with an issue to remind me