ggplot2对象带有grid.draw输出后损坏的导出

Question

我正在尝试创建ggplot2绘图样式。 theme_custom()函数处理调整后的主题。困难的部分是覆盖title，subtitle和caption的边距的部分。我使用finalise_plot()函数的目标是将文本对齐到图的最左侧。

直到我偶然发现出口问题时，该图看起来几乎就像我想要的那样。当我尝试将图像（通过绘图设备）保存为pdf时，它将删除所有文本。当我将其导出为svg时，它已损坏。

是否可以像下面的示例那样在svg中导出确切的输出？

library(tidyverse)
library(grid)
library(gridExtra)

theme_custom <- function(){
  theme_minimal() +
    theme(plot.title = element_text(face = "bold",
                                    family = "Arial",
                                    size = 24),
          plot.subtitle = element_text(face = "bold",
                                       family = "Arial",
                                       size = 20),
          plot.caption = element_text(hjust = 0,
                                      color = "#6c737b",
                                      size = 18,
                                      family = "Arial"),
          axis.text = element_text(size = 18,
                                   color = "#1d2939",
                                   family = "Arial"),
          axis.title.y = element_text(color = "#6b38e8",
                                      family = "Arial",
                                      size = 18,
                                      hjust = 0),
          axis.text.y = element_text(color = "#6b38e8"),
          plot.background = element_rect(fill = "#f2f2f2",
                                         color = 0),
          panel.grid.minor = element_blank(),
          panel.grid = element_line(color = "#d8cedb",
                                    size = 0.75),
          panel.border = element_blank(),
          plot.margin = unit(c(10,0,0,0),"pt"))
}

finalise_plot <- function() {

  table_grob <- ggplotGrob(p)

  table_grob$layout$l[table_grob$layout$name %in% c("title", "subtitle", "caption")] <- 1

  grid::grid.draw(table_grob)

}

p <- iris %>%
  ggplot(aes(Sepal.Length, Sepal.Width,
             color = Species)) +
  geom_point() +
  theme_custom() +
  labs(title = "Main title",
       subtitle = "Subtitle",
       caption = "UCI Machine Learning Repository",
       x = NULL)

finalise_plot()

^{由reprex package（v0.2.1）于2019-04-16创建}

pdf导出：

svg导出：