ggplotly makes the border of a polygon transparent. Work around?

I am making a scatter plot using ggplotly so that one can hover over the observation and see the label and so forth. I am plotting a polygon on top of it to highlight a specific area. I put it at the start of the ggplot to and fill the polygon with white and I do it again with at the end of the ggplot statements so that the boundary sits on top of the points and use a transparent fill. I do all this in ggplot and then convert to ggplotly. ggplotly makes the boundary of the polygon transparent and I lose the white interior that is supposed to be on the bottom. Is there a work around? Why is it doing this?

Here is the code:

rm(list = ls())
vertices <- data.frame(x = c(-1, 1, 1, -1, -1), y = c(-1, -1, 1, 1, -1))
df <- data.frame(x = rnorm(1000), y = rnorm(1000), ticker = paste0(1:1000), warren = (runif(1000) < .3))

plot1 <- ggplot(df, aes(x = x, y = y)) +
  theme_minimal(base_size = 14) +
  theme(plot.title = element_text(hjust = 0.5)) +
  theme(panel.background = element_rect(fill = "grey95"), panel.grid = element_line(color = "grey85")) +
  geom_polygon(data = vertices, aes(x, y), fill = "white", color = "lightblue") +
  suppressWarnings(geom_point(size = 2, aes(col = warren, shape = warren, text = ticker))) +
  scale_color_manual(values = c(rgb(1, 0, 0, .2), rgb(0, 0, 1))) +
  geom_polygon(data = vertices, aes(x, y), fill = rgb(0, 0, 0, 0), color = "lightblue", linewidth = 3)

plot1

g <- ggplotly(plot1)

g

Here is the ggplot:

Here is the ggplotly:

It looks like the conversion between plotting mechanisms dropped one of your layers (a polygon layer). The sizing in ggplot is different than in Plotly, so your markers got a lot bigger, as well.

First, we're going to add that missing layer back to your plot, then adjust the sizes. Lastly, because I'm just going to duplicate the polygon layer, we'll need to add the white fill to the bottom layer.

Let me know if you have any questions.

g$x$data <- append(g$x$data, g$x$data[1]) # duplicate the first trace
g                                         # inspect what you expect

lapply(1:length(g$x$data),                # update traces based on type
       function(j) {
         if(g$x$data[[j]]$mode == "markers") {  
           g$x$data[[j]]$marker$size <<- 5      # reduce size
         } else {
           g$x$data[[j]]$line$width <<- 8       # reduce width
         }
       })
g$x$data[[1]]$fillcolor <- "white"        # change the bottom trace only
g                                         # inspect what you expect