Implement geom_function().

[clauswilke]

A first stab at implementing a geom_function(). Closes #3611.

Open questions:

• How much should we rewrite the documentation to stat_function()? Can it remain mostly as is, or should it now recommend geom_function() over stat_function()?
• Is the current warning for multiple groups OK, or should the code filter out the duplicated groups?

library(tidyverse)

data <- tibble(
  x = 1:50,
  y = x*x - 2*x + 5 + 200*rnorm(50),
  a = sample(letters[1:3], 50, replace = TRUE)
)

ggplot(data, aes(x, y)) +
  geom_point(aes(color = "data")) +
  geom_function(aes(color = "function"), fun = ~.x*.x - 2*.x + 5)

ggplot(data, aes(x, y, color = a)) +
  geom_point(aes()) +
  geom_function(aes(color = "function"), fun = ~.x*.x - 2*.x + 5)

ggplot(data, aes(x, y, color = a)) +
  geom_point(aes()) +
  geom_function(aes(), fun = ~.x*.x - 2*.x + 5)
#> Warning: Multiple drawing groups in `geom_function()`. Did you use the correct
#> `group`, `colour`, or `fill` aesthetics?

ggplot(data, aes(x, y)) +
  geom_point(aes(color = "data")) +
  stat_function(aes(color = "function"), fun = ~.x*.x - 2*.x + 5)

ggplot(data, aes(x, y, color = a)) +
  geom_point(aes()) +
  stat_function(aes(color = "function"), fun = ~.x*.x - 2*.x + 5)

ggplot(data, aes(x, y, color = a)) +
  geom_point(aes()) +
  stat_function(aes(), fun = ~.x*.x - 2*.x + 5)
#> Warning: Multiple drawing groups in `geom_function()`. Did you use the correct
#> `group`, `colour`, or `fill` aesthetics?

^{Created on 2020-05-01 by the reprex package (v0.3.0)}

[paleolimbot]

These look good! It might be worth trying the PR on faceted plots just to make sure, but I can't see any red flags.

[clauswilke]

With facets. Works as expected.

library(tidyverse)

data <- tibble(
  x = 1:50,
  y = x*x - 2*x + 5 + 200*rnorm(50),
  a = sample(letters[1:3], 50, replace = TRUE)
)

ggplot(data, aes(x, y, color = a)) +
  geom_point() +
  geom_function(color = "black", fun = ~.x*.x - 2*.x + 5) +
  facet_wrap(~a)

ggplot(data, aes(x, y, color = a)) +
  geom_point() +
  geom_function(fun = ~.x*.x - 2*.x + 5) +
  facet_wrap(~a)

^{Created on 2020-05-01 by the reprex package (v0.3.0)}

[clauswilke]

@hadley I'd prefer to rewrite the documentation so it refers primarily to geom_function() rather than stat_function(), since this would match geom_density(), geom_contour(), geom_smooth(), etc. Any reason not to do this? The name of the .Rd file would change, of course.

[hadley]

That make sense to me. I don't think changing the name of the .Rd file should cause any problems.