Bubble plot still ignores size constraints for singleton color categories

As you can see in the following reprex, the sizing options get ignored for the singleton category (color == "D"); commit https://github.com/plotly/plotly.R/commit/306b4ef057631f9d3200fdfce636696bc0268952 seemed intended to fix this issue but it persists.

library(plotly)
library(dplyr)

dat = structure(list(color = c("A", "A", "A", "B", "B", "C", "C", "D"
), size = c(2L, 1L, 1L, 803L, 254L, 339L, 261L, 577L), x = c(0, 
0, 0, 32.9, 52, 20.4, 33.3, 73.7), y = c(0, 0, 0, 0.7, 0.8, 0.6, 
0, 0.2)), row.names = c(NA, -8L), class = c("tbl_df", "tbl", 
"data.frame"))

print(dat)

desired_maximum_marker_size = 10
sizeref <- 2 * max(dat$size) / (desired_maximum_marker_size^2)

plotly::plot_ly(
    data = dat,
    y = ~y,
    x = ~x,
    color = ~color,
    type = "scatter",
    mode = 'markers',
    sizes = c(1, desired_maximum_marker_size),
    marker = list(
      size = ~size,
      sizemin = 1,
      sizemax = 10,
      opacity = 0.7, 
      sizeref = sizeref,
      sizemode = 'area')
    
  )

Run the supplied R reprex with plotly and confirm that the singleton color category ignores the marker sizing constraints. Trace how marker sizes are handled when a color category has one observation, then add a regression test covering that case. Done means the singleton category respects sizemin, sizemax, sizeref, and sizemode consistently with the other categories.