Add experimental ggplot2 scales for calendars

[DavisVaughan]

Closes #233

library(ggplot2)
library(clock)
library(vctrs)

set.seed(1234)

from <- year_month_day(2019, 1)

df <- vec_rbind(
  data_frame(
    g = "stock 1",
    date = from + duration_months(cumsum(sample(1:2, size = 100, replace = TRUE))),
    price = cumsum(1 + rnorm(100))
  ),
  data_frame(
    g = "stock 2",
    date = from + duration_months(cumsum(sample(1:2, size = 100, replace = TRUE))),
    price = cumsum(1 + rnorm(100))
  )
)

# Defaults automatically know you have monthly data
ggplot(df, aes(date, price, group = g, color = g)) +
  geom_line()

# Fully customize as needed
ggplot(df, aes(date, price, group = g, color = g)) +
  geom_line() +
  scale_x_year_month_day(
    date_breaks = duration_months(24),
    date_minor_breaks = duration_months(6),
    date_labels = "%Y"
  )

ggplot(df, aes(date, price, group = g, color = g)) +
  geom_line() +
  scale_x_year_month_day(
    date_labels = "%B\n%Y",
    date_locale = clock_locale("fr")
  )

economics$date <- as_year_month_day(economics$date)
economics$date <- calendar_narrow(economics$date, "month")

ggplot(economics, aes(x = date, y = unemploy, fill = date)) +
  geom_col()

set.seed(1234)

from1 <- year_quarter_day(2019, 1)
from2 <- year_quarter_day(2000, 2)

df <- vec_rbind(
  data_frame(
    g = "stock 1",
    date = from1 + duration_quarters(cumsum(sample(1:5, size = 50, replace = TRUE))),
    price = cumsum(1 + rnorm(50))
  ),
  data_frame(
    g = "stock 2",
    date = from2 + duration_quarters(cumsum(sample(1:5, size = 50, replace = TRUE))),
    price = cumsum(1 + rnorm(50))
  )
)

ggplot(df, aes(date, price, group = g, color = g)) +
  geom_line()

# Zooming with `coord_cartesian()`
ggplot(df, aes(date, price, group = g, color = g)) +
  geom_line() +
  coord_cartesian(xlim = year_quarter_day(c(2020, 2040), 1))

^{Created on 2023-05-10 with reprex v2.0.2.9000}