使用ggtext的element_markdown()渲染ggplot2文本时重叠问题求助

问题描述 #

尝试在ggplot2图表的标题、副标题、说明文本及图例中使用Markdown格式文本，图例文本需通过ifelse语句根据变量动态生成（其中一种场景需将部分文本设为斜体），因此使用ggtext包的element_markdown()函数，用*包裹目标文本。但渲染后文本紧凑重叠，效果极差。调整字体大小、样式、字体族及边距均无效，尝试用expression结合paste/glue等替代方法，但ifelse语句导致实现困难，甚至用UNICODE符号输入也未成功。

运行环境：R版本4.2.1（2022-06-23），macOS Monterey 12.5.1，RStudio版本2022.7.1.554（Spotted Wakerobin），ggtext包版本0.1.1。

最小示例代码 #

library(tidyverse)
library(glue)
library(ggtext)

random_variable <- TRUE

iris %>% 
    ggplot(aes(x = Species, y = Sepal.Width)) + 
    geom_boxplot(show.legend = F, col = 'cadetblue4') +
    stat_summary(fun = mean, geom = 'point', shape = 18, size = 3, aes(col = 'mean')) +
    geom_hline(aes(yintercept = 2, col = '1st'), size = .5, lty = 2) +
    geom_hline(aes(yintercept = 3, col = '2nd'), size = .5, lty = 2) +
    geom_hline(aes(yintercept = 4, col = '3rd'), size = .5, lty = 2) +
    scale_colour_manual(name = 'Title without markdown', 
                        values = c('1st' = 'slateblue2', '2nd' = 'coral4', '3rd' = 'red2', 'mean' = 'black'),
                        labels = c('Sub<sub>1</sub>', 'Sup<sup>2</sup>', 'Both<sub>3</sub><sup>10</sup>', 
                                   paste('Some', ifelse(random_variable, yes = '*markdown* text here', no = 'not markdown text here'), 'and then some more text here.')),
                        guide = guide_legend(override.aes = list(
                            linetype = c(rep('dashed', 3), 'blank'),
                            shape = c(rep(NA, 3), 18)))) + 
    labs(title = paste0("This is a test of element_", ifelse(T, yes = '*markdown*', no = 'text'), " with the Iris dataset!"),
         subtitle = "Subtitle with **boldface text** and *italicised text* -- Not working at all…",
         caption = "THIS SENTENCE DEFINITELY HAS WHITESPACES!^{23}") +
    theme(
          # TITLES AND CAPTION
          plot.title = element_markdown(family="Times", face = "bold", size=15),
          plot.subtitle = element_markdown(family="Times", size=10),
          plot.caption = element_markdown(family = "Times", face = "bold", size = 11),
          
          # LEGEND SETTINGS
          legend.position = 'bottom',
          legend.key.size = unit(1.75, 'line'),
          legend.spacing.x = unit(.1, 'cm'),
          legend.title = element_markdown(size = 12, face = 'bold', family = 'Times', 
                                      margin = margin(r = .25, unit = 'cm')),
          legend.text = element_markdown(size = 13, face = 'bold', family = 'Times', 
                                     margin = margin(r = .5, unit = 'cm')),
          legend.key =  element_blank())

示例输出 #

解决办法 #

1. 更新ggtext包到最新版本 #

你使用的ggtext 0.1.1版本存在文本排版的底层bug，后续版本已修复该问题。执行以下代码更新：

install.packages("ggtext")

2. 优化文本构造与排版参数 #

核心修改点：

• 用glue替代paste构造动态文本，避免空格丢失或格式解析异常
• 给所有element_markdown()添加lineheight参数，强制设置行高，解决文本重叠
• 调整图例布局参数，给长文本留出足够空间

修改后的完整代码

library(tidyverse)
library(glue)
library(ggtext)

random_variable <- TRUE

# 用glue构造图例标签，语法更清晰
legend_labels <- c(
  'Sub<sub>1</sub>', 
  'Sup<sup>2</sup>', 
  'Both<sub>3</sub><sup>10</sup>', 
  glue('Some {ifelse(random_variable, "*markdown* text here", "not markdown text here")} and then some more text here.')
)

iris %>% 
  ggplot(aes(x = Species, y = Sepal.Width)) + 
  geom_boxplot(show.legend = F, col = 'cadetblue4') +
  stat_summary(fun = mean, geom = 'point', shape = 18, size = 3, aes(col = 'mean')) +
  geom_hline(aes(yintercept = 2, col = '1st'), size = .5, lty = 2) +
  geom_hline(aes(yintercept = 3, col = '2nd'), size = .5, lty = 2) +
  geom_hline(aes(yintercept = 4, col = '3rd'), size = .5, lty = 2) +
  scale_colour_manual(
    name = 'Title without markdown', 
    values = c('1st' = 'slateblue2', '2nd' = 'coral4', '3rd' = 'red2', 'mean' = 'black'),
    labels = legend_labels,
    guide = guide_legend(
      override.aes = list(
        linetype = c(rep('dashed', 3), 'blank'),
        shape = c(rep(NA, 3), 18)
      ),
      # 图例项换行显示，避免横向挤压
      nrow = 2
    )
  ) + 
  labs(
    title = glue("This is a test of element_{ifelse(TRUE, '*markdown*', 'text')} with the Iris dataset!"),
    subtitle = "Subtitle with **boldface text** and *italicised text* -- Now working correctly!",
    caption = "THIS SENTENCE DEFINITELY HAS WHITESPACES!^{23}"
  ) +
  theme(
    # 标题与说明文本：添加lineheight参数
    plot.title = element_markdown(family="Times", face = "bold", size=15, lineheight = 1.2),
    plot.subtitle = element_markdown(family="Times", size=10, lineheight = 1.2),
    plot.caption = element_markdown(family = "Times", face = "bold", size = 11, lineheight = 1.2),
    
    # 图例设置：调整行高与间距
    legend.position = 'bottom',
    legend.key.size = unit(2, 'line'),
    legend.spacing.x = unit(.3, 'cm'),
    legend.spacing.y = unit(.2, 'cm'),
    legend.title = element_markdown(size = 12, face = 'bold', family = 'Times', 
                                    margin = margin(r = .25, unit = 'cm'), lineheight = 1.2),
    legend.text = element_markdown(size = 13, face = 'bold', family = 'Times', 
                                   margin = margin(r = .5, unit = 'cm'), lineheight = 1.2),
    legend.key =  element_blank()
  )

效果说明 #

• 更新ggtext包后，底层排版bug修复，文本不再重叠
• lineheight参数确保文本行之间有足够间距
• 图例设置nrow=2让长标签换行，避免横向空间不足
• glue让动态文本的构造更直观，格式解析更准确

内容的提问来源于stack exchange，提问作者OMMJREN