animate and ggsave interpret theme spacing differently

[bbcuffer]

The images saved by animate have different spacing to those saved by ggsave: particularly the vertical space surrounding titles/captions.

Here's the code that generates the image above. It

• sets up a basic plot
• saves it two ways: once using ggsave and once using animate.
• collates the two images.

Fwiw, I don't think it's a ragg thing: I've had similar issues when saving with svglite. Nor do I think it's a ggsave thing, since it doesn't matter whether I open a graphics device, print the image and then close it or use ggsave to do that work.

library(ggplot2)
library(gganimate)
library(magick)

# The example data:
# Ten points: when state==1, they count up to 10. When state==2, they count down

df <- data.frame(
  x = rep(1:10, 2),
  y = c(1:10, 10:1),
  state = rep(1:2, c(10, 10))
)

# set up the basics of a plot.
base_plot <- ggplot(df, aes(x = x, y = y, group = x)) +
  labs(title = "gganimate v ggsave", subtitle = "Subtitle here", caption = "Caption at bottom") +
  theme_minimal(base_size = 18) +
  scale_x_continuous(limits = c(0, 10)) +
  scale_y_continuous(limits = c(0, 10))


# Version 1: save the images directly using ggsave
ggsave_fn <- ggsave(
  plot = (base_plot %+% subset(df, state == 1)) + geom_point(),
  file = tempfile(fileext = ".png"),
  device = ragg::agg_png,
  width = 5, height = 5, bg = "white",
)


# Version 2: save the images via gganimate.
# first, set up an animation to transition from counting up to counting down
pt_anim <- base_plot +
  labs(title = "gganimate v") +
  geom_point(colour = "red") +
  theme(
    text = element_text(colour = "red"),
    axis.text = element_text(colour = "red"),
    panel.grid = element_line(colour = "pink")
  ) +
  transition_states(state, wrap = FALSE)

# and save one frame per state
gganimate_fns <- animate(pt_anim,
  renderer = \(frames, fps) frames,
  nframes = 2,
  fps = 1,
  device = "ragg_png",
  width = 5,
  height = 5,
  units = "in", res = 300,
  background = NA
)

# Compare the two versions
foreground <- image_read(gganimate_fns[1])
# overlaying the gganimate version on top of the ggsave version
output_fn <- image_read(ggsave_fn) |>
  image_composite(foreground) |>
  image_write(tempfile(pattern = "comparison", fileext = ".png"))

# Produces the chart shown above
browseURL(output_fn)