如何在ggplot中结合ggdag显示DAG的自定义标签？

问题场景 #

用户使用ggdag创建了带有自定义标签的DAG：

test <- dagify("a" ~ "b",
    "b" ~ "c",
    "c" ~ "d",
    "b" ~ "d",
    exposure = "b",
    outcome = "d",
    labels = c(a = "A",
        b = "B",
        c = "D"))

通过ggplot直接调用时，无法显示自定义标签：

test %>%
    ggplot(aes(x = x, y = y, xend = xend, yend = yend)) +
        geom_dag_point(color = "orange") +
        geom_dag_edges_arc(edge_color = "blue", curvature = 0) +
        geom_dag_text(color = "black") +
        theme_dag()

但使用ggdag(test, text = FALSE, use_labels = "label", edge_type = "link_arc")可以正常显示自定义标签，需要实现ggplot流程下的自定义标签显示。

解决方法 #

geom_dag_text()默认读取节点的name字段而非自定义的label字段，只需显式指定label映射即可：

方法1：修改geom_dag_text的映射 #

test %>%
    ggplot(aes(x = x, y = y, xend = xend, yend = yend)) +
        geom_dag_point(color = "orange") +
        geom_dag_edges_arc(edge_color = "blue", curvature = 0) +
        geom_dag_text(aes(label = label), color = "black") +  # 显式绑定label列
        theme_dag()

方法2：替换为geom_text #

也可以直接用基础的geom_text()，同样指定label映射：

test %>%
    ggplot(aes(x = x, y = y, xend = xend, yend = yend)) +
        geom_dag_point(color = "orange") +
        geom_dag_edges_arc(edge_color = "blue", curvature = 0) +
        geom_text(aes(label = label), color = "black") +
        theme_dag()

原理：ggdag生成的DAG对象转化为数据框后，自定义标签存储在label列中，显式指定aes(label = label)就能让文本层调用你定义的标签内容。

内容的提问来源于stack exchange，提问作者pgcudahy