Solution to combine facetted legends breaks for continuous variable

I am trying to use plotly to create facetted plots - I do this by using group_map() to create my plots robustly.

The solution I currently have works in reducing duplicated legend labels when using discrete variables for color/size/etc but fail to work for continuous variables. I assume this is do to each facetted plot potentially having different total ranges for the continuous variable.

library(tidyverse)
library(plotly)

city_list = unique(txhousing$city) %>%  
  sample(8)

##solution works for discrete labels 
txhousing %>%  
  filter(month < 6, 
         city %in% city_list) %>% 
  group_by(month) %>% 
  group_map(~{
    plot_ly(.x, x=~year, y = ~sales, color = ~city,legendgroup = ~city,type = "bar"
            ,showlegend = (.y == "1")
            ) %>%
      layout(yaxis = list(
        title = paste0(c(rep("&nbsp;", 20),
                         paste("<b>", as.character(.y), "</b>"),
                         rep("&nbsp;", 20),
                         rep("\n&nbsp;", 3)),
                       collapse = ""))) %>%
      layout(barmode = 'stack')
  }
  ) %>%
  subplot(nrows = NROW(.), margin = .05, shareX = T, shareY = T, titleY = T) 

##solution fails to merge legends for continuous variable
txhousing %>%  
  filter(month < 6, 
         city %in% city_list) %>% 
  group_by(month) %>% 
  group_map(~{
    plot_ly(.x, x=~year, y = ~sales, color = ~volume,legendgroup = ~volume,type = "bar"
            ,showlegend = (.y == "1")
            ) %>%
      layout(yaxis = list(
        title = paste0(c(rep("&nbsp;", 20),
                         paste("<b>", as.character(.y), "</b>"),
                         rep("&nbsp;", 20),
                         rep("\n&nbsp;", 3)),
                       collapse = ""))) %>%
      layout(barmode = 'stack')
  }
  ) %>%
  subplot(nrows = NROW(.), margin = .05, shareX = T, shareY = T, titleY = T) 

首先使用 plot_ly()、group_map() 和 subplot() 运行提供的两个 R 重现代码，然后比较离散颜色和连续颜色的情况。跟踪 subplot 如何组合轨迹和图例；完成的标准是：连续变量的图例能够合并，同时不破坏分面图的行为。