Expose position arg on geom_hline() and geom_vline()

[yutannihilation]

Fix #4285

[thomasp85]

Have you tested how this behaves with other positions than position_nudge()? I'm fine with it not working but I could envision some hard-to-understand errors

[yutannihilation]

Hmm, no. And, sorry, I think I tested this quickly with the similar code to #4285 (comment) and thought it works, but it was probably just I forgot to change as.numeric(x). This needs some more tweaks. I don't see immediately the reason why this doesn't work.

library(ggplot2)

mydata <- data.frame(
  x = factor(c("1", "2", "3")),
  y = c(3, 4, 5)
)
mydata$splits <- factor(mydata$x, labels = c("baseline", "Cycle 1", "Cycle 1"))

devtools::load_all("~/repo/ggplot2/")
#> ℹ Loading ggplot2

pos <- position_nudge(x = -0.1)

# works
ggplot(mydata, aes(x, y)) +
  geom_point(size = 3, position = pos) +
  geom_vline(
    aes(xintercept = as.numeric(x)),
    alpha = 0.2, size = 5,
    colour = "red",
    position = pos
  )

# doesn't work
ggplot(mydata, aes(x, y)) +
  geom_point(size = 3, position = pos) +
  geom_vline(
    aes(xintercept = x),
    alpha = 0.2, size = 5,
    colour = "red",
    position = pos
  )
#> Warning in Ops.factor(x, params$x): '+' not meaningful for factors
#> Warning: Removed 3 rows containing missing values (geom_vline).

^{Created on 2021-03-25 by the reprex package (v1.0.0)}

[thomasp85]

@yutannihilation do you think it is feasible to have this PR done this week? I'm beginning to wrap up the patch release

[yutannihilation]

Sorry, I'm afraid I can't... I thought this was as easy job as just exposing the argument, but it turned this requires a better understanding about how position works, which I don't know yet.

[thomasp85]

No problem - I've moved all other position related issues to next release because it seems to be a huge work getting it all fixed coherently

[yutannihilation]

I see, thanks for handling the release!

[hadley]

@yutannihilation do you still think this is worth finishing off?

[yutannihilation]

Yes, but I don't think I can finish this soon. I'm giving up for now, sorry.

[hadley]

No problems — thanks for giving it a go!

[teunbrand]

The issue in #4286 (comment) is that xintercept in not a member of scale_x_discrete()$aesthetics. That problem should get fixed by #5640, and this PR improves the lack of position adjustments.

@yutannihilation Is there any chance I could convince you to revive this PR? If you don't have a luxurious amount of free time on your hands I wouldn't mind recreating it anew.

[yutannihilation]

Oh, thanks. I just quit my job today so I only have free time in theory :) Let me try this again. Though I couldn't keep up with the recent evolution of ggplot2, I hope I can adjust myself.

[teunbrand]

I just quit my job today

I hope this was by choice and works out in your favour!

I couldn't keep up with the recent evolution of ggplot2

I'd be glad to help out any way I can, e.g. if you have questions..

There isn't any hurry in this, but I think this PR is the right solution for the problem.

[yutannihilation]

Thanks. It's my choice, so don't worry about me.

I hope I'll start addressing on this next week!