You need to enable JavaScript to run this app. Shiny应用中如何正确导出、合并grid.draw生成的图表
优惠活动
大模型
产品
解决方案
定价
更多

Shiny应用中如何正确导出、合并grid.draw生成的图表

Shiny应用中ggnewscale+grid.draw绘图适配问题

问题概述

开发Shiny应用时,使用ggnewscale包结合grid.draw()生成图表,其导出、合并逻辑与常规ggplot2图表存在明显差异:

  • 最初尝试通过pdf()函数在应用内保存绘图结果,点击下载按钮后导出的文件既非有效图片格式,也无法正常打开。已参考同类公开技术方案,均无法适配当前场景。
  • 存在两个核心待解决问题:
    1. 导出逻辑适配:项目中p1()、p2()函数均通过grid.draw()渲染图表,在downloadHandler()中尝试写入grid::grid.newpage(); grid::grid.draw(gt)的导出逻辑完全失效;目前只能靠在downloadHandler中重复编写p1()、p2()的全部绘图代码勉强实现下载,代码冗余度极高。曾尝试用ggsave()直接传入p2()的结果导出,同样无法正常运行。
    2. 合并排版适配:由于p1()、p2()均为grid.draw渲染的输出对象,尝试使用grid.arrange(p1(),p2())合并排版完全不生效——该函数仅适配原生ggplot2类型的图表,无法接收当前这类已经经过渲染的grid对象。

可复现代码与配套数据

options(encoding = "UTF-8")
library(stats)
library(openxlsx)
library(shiny)
library(dplyr)
library(tidyr)      
library(ggplot2)
library(gridExtra)  
library(ggpubr)
library(shinythemes)
library(ggpattern)
library(grid)
library(ggh4x)
library(ggnewscale)
library(psych)
library(DT)
library(shinyBS)
library(shinyjs)    

###
# 模拟测试数据
data <- data.frame(
  Name = c(paste0("Group_", LETTERS[1:20])),
  matx <- matrix(sample(1:1000, 1000, replace = T), nrow = 20)
)
names(data)[-1] <- c(paste0("Gene_", 1:50))


Nameff<-structure(list(Name = c("8_TBI(1 month)", "9_control", "9_VEGF", 
                                "10_control", "10_VEGF", "11_Brain Healty", "12_control_1", "12_control_2", 
                                "12_AOD(Jnk1/2/3 ko)", "13_control", "13_Cpt1_ko(Cdh5 driven)", 
                                "14_control", "14_Tsc2ko(Tbx4 driven)", "15_control", "15_Zmpste24 ko", 
                                "16_control", "16_Adrenomedullin\r\nko(Cdh5 driven)", "17_Lung Healthy", 
                                "18_control(14w)", "18_carboplatin(14w)"), Disease = c("TBI\r\n1month", 
                                                                                       "VEGF\r\nsti_DG\r\nregion", "VEGF\r\nsti_DG\r\nregion", "VEGF\r\nsti_CA1\r\nregion", 
                                                                                       "VEGF\r\nsti_CA1\r\nregion", "Healthy\r\n(Brain)", "AOD", "AOD", 
                                                                                       "AOD", "CPT1\r\nko", "CPT1\r\nko", "Tsc2\r\nko", "Tsc2\r\nko", 
                                                                                       "Zmpste\r\n_24 ko", "Zmpste\r\n_24 ko", "AM\r\nko", "AM\r\nko", 
                                                                                       "Healthy\r\n(Lung)", "Chemo/Radio\r\n(Tibiae)", "Chemo/Radio\r\n(Tibiae)"
                                ), Organ = c("Brain", "Brain", "Brain", "Brain", "Brain", "Brain", 
                                             "Lung", "Lung", "Lung", "Lung", "Lung", "Lung", "Lung", "Lung", 
                                             "Lung", "Lung", "Lung", "Lung", "Bone", "Bone"), fill = c("#f15a24", 
                                                                                                       "#FFFFFF", "#f15a24", "#FFFFFF", "#f15a24", "#FFFFFF", "#FFFFFF", 
                                                                                                       "#FFFFFF", "#00FF00", "#FFFFFF", "#00FF00", "#FFFFFF", "#00FF00", 
                                                                                                       "#FFFFFF", "#00FF00", "#FFFFFF", "#00FF00", "#FFFFFF", "#FFFFFF", 
                                                                                                       "#7570B3"), Condition = c("#CCCCFF", "#d9e021", "#d9e021", "#d9e021", 
                                                                                                                                 "#d9e021", "#CCCCFF", "#CCCCFF", "#CCCCFF", "#CCCCFF", "#fbb03b", 
                                                                                                                                 "#fbb03b", "#fbb03b", "#fbb03b", "#fbb03b", "#fbb03b", "#fbb03b", 
                                                                                                                                 "#fbb03b", "#fbb03b", "#d9e021", "#d9e021"), Organ_fill = c("#f15a24", 
                                                                                                                                                                                             "#f15a24", "#f15a24", "#f15a24", "#f15a24", "#f15a24", "#00FF00", 
                                                                                                                                                                                             "#00FF00", "#00FF00", "#00FF00", "#00FF00", "#00FF00", "#00FF00", 
                                                                                                                                                                                             "#00FF00", "#00FF00", "#00FF00", "#00FF00", "#00FF00", "#7570B3", 
                                                                                                                                                                                             "#7570B3"), Alpha = c(1, 0, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 
                                                                                                                                                                                                                   0, 1, 0, 1, 0, 0, 1)), row.names = 22:41, class = "data.frame")
####################################################################
ui <- fluidPage(
  theme = shinytheme("superhero"),
  useShinyjs(),
  pageWithSidebar(
    tags$h4(Sys.time()),
    sidebarPanel(
      tags$h3("1111"),
      selectInput(
        "selectGeneSymbol", 
        "444d:", 
        choices = colnames(data[,-1]),
        multiple =F,
        width = 400,
        selected = NULL
      ),
      actionButton(inputId = "plot1", label = "down1",width=80),
      actionButton(inputId = "plot2", label = "down2",width=80),
      actionButton(inputId = "all",label = "down1&down2",width=120),
      hr(),
      tags$h5(tags$strong("down:")),
      downloadButton("p1", "down1",width=120),
      downloadButton("p2", "down2",width=120)
    ),
    mainPanel(
      tabsetPanel(
        tabPanel(icon("home"), uiOutput("all"))
      )
    )
  )
)
server <- function(input, output, session) {
  plot_data1 <- reactive({
    subset(data, colnames(data[,-1]) %in% input$selectGeneSymbol)
  })
  plot_data2 <- reactive({
    subset(data, colnames(data[,-1]) %in% input$selectGeneSymbol)
  })
  global <- reactiveValues(out = NULL, p1 = NULL, p2 = NULL)
  
  observeEvent(input$plot1, {
    global$out <- plotOutput("plot1", height=600)
  })
  observeEvent(input$plot2, {
    global$out <- plotOutput("plot2", height=600)
  })
  observeEvent(input$all, {
    global$out <- plotOutput("plot3", height=600)
  })
  
  output$all <- renderUI({global$out})
  
  p1 <- eventReactive(list(input$plot1, input$all), {
    datamean_sd<-data.frame(
      Nameff,
      mean=tapply(data[,input$selectGeneSymbol],data$Name,mean),
      sd=tapply(data[,input$selectGeneSymbol],data$Name,sd)
    ) 
    p<-ggplot(data = datamean_sd, aes(Name,mean, label = Name, fill=Organ)) +
      geom_bar(position="dodge2", stat="identity",width = 0.85,color="black",alpha=datamean_sd$Alpha) +
      facet_nested(.~Organ+Disease, scales = "free_x", space = "free_x",switch = "x")+
      theme_classic2() +
      theme(legend.position = "bottom",
            legend.box = "horizontal",
            plot.title = element_text(hjust = 0.5),
            plot.margin = unit(c(5, 10, 20, 15), "mm"),
            strip.background = element_rect(colour="black", fill="white"),
            strip.text.x = element_text(size = 6, angle=0),
            axis.text.x=element_text(size=8),
            strip.placement = "outside"
      ) +
      rotate_x_text(angle = 90)+
      scale_fill_manual(name = "Organ",values = unique(datamean_sd$Organ_fill))
    gt <- ggplotGrob(p)
    grid::grid.newpage(); grid::grid.draw(gt)
  })
  
  p2 <- eventReactive(list(input$plot2, input$all), {
    datamean_sd<-data.frame(
      Nameff,
      mean=tapply(data[,input$selectGeneSymbol],data$Name,mean),
      sd=tapply(data[,input$selectGeneSymbol],data$Name,sd)
    ) 
    p<-ggplot(data = datamean_sd, aes(Name,mean, label = Name, fill=Organ)) +
      geom_bar(position="dodge2", stat="identity",width = 0.85,color="black",alpha=datamean_sd$Alpha) +
      facet_nested(.~Organ+Disease, scales = "free_x", space = "free_x",switch = "x")+
      theme_classic2() +
      theme(legend.position = "bottom",
            legend.box = "horizontal",
            plot.title = element_text(hjust = 0.5),
            plot.margin = unit(c(5, 10, 20, 15), "mm"),
            strip.background = element_rect(colour="black", fill="white"),
            strip.text.x = element_text(size = 6, angle=0),
            axis.text.x=element_text(size=8),
            strip.placement = "outside"
      ) +
      rotate_x_text(angle = 90)+
      scale_fill_manual(name = "Organ",values = unique(datamean_sd$Organ_fill))
    gt <- ggplotGrob(p)
    grid::grid.newpage(); grid::grid.draw(gt)
  })
  
  output$plot1 <- renderPlot({ p1() })
  output$plot2 <- renderPlot({ p2() })
  output$plot3 <- renderPlot({ 
    lollipop <- gTree(children = gList(p1(), p2()))
    grid.draw(lollipop)
  })
  
  #download p1
  output$p1 <- downloadHandler(
    filename = function() {
      paste0(input$selectGeneSymbol,"_123",".pdf")
    },
    content = function(file) {
      pdf(file,width=20,height=10)
      datamean_sd<-data.frame(
        Nameff,
        mean=tapply(data[,input$selectGeneSymbol],data$Name,mean),
        sd=tapply(data[,input$selectGeneSymbol],data$Name,sd)
      ) 
      p<-ggplot(data = datamean_sd, aes(Name,mean, label = Name, fill=Organ)) +
        geom_bar(position="dodge2", stat="identity",width = 0.85,color="black",alpha=datamean_sd$Alpha) +
        facet_nested(.~Organ+Disease, scales = "free_x", space = "free_x",switch = "x")+
        theme_classic2() +
        theme(legend.position = "bottom",
              legend.box = "horizontal",
              plot.title = element_text(hjust = 0.5),
              plot.margin = unit(c(5, 10, 20, 15), "mm"),
              strip.background = element_rect(colour="black", fill="white"),
              strip.text.x = element_text(size = 6, angle=0),
              axis.text.x=element_text(size=8),
              strip.placement = "outside"
        ) +
        rotate_x_text(angle = 90)+
        scale_fill_manual(name = "Organ",values = unique(datamean_sd$Organ_fill))
      gt <- ggplotGrob(p)
      grid::grid.newpage(); grid::grid.draw(gt)
      dev.off()
    }
  )
  
  #download p2
  output$p2 <- downloadHandler(
    filename = function() {
      paste0(input$selectGeneSymbol,"_123",".pdf")
    },
    content = function(file) {
      ggsave(file,p2(),width=20,height=10)
      grid::grid.newpage(); grid::grid.draw(gt)
      dev.off()
    }
  )
}
shinyApp(ui = ui, server = server)

根因与调整方案

  • 核心根因:grid.draw()函数本身无返回值,调用后仅将grob对象绘制到当前活动网格画布上,不会将grob对象本身传递给调用方。原p1()、p2()函数实际返回空值,既无法被ggsave()读取,也无法在合并排版时被其他grid工具调用。
  • 代码调整方向:
    • 修改p1()、p2()函数的返回逻辑:移除函数内部的grid::grid.newpage(); grid::grid.draw(gt)调用,直接将生成的grob对象作为函数返回值,实现绘图逻辑复用。
    • 单图导出时,直接调用对应函数返回的grob对象,通过ggsave()或pdf()画布配合grid.draw()完成导出,无需重复编写绘图代码。
    • 合并排版时,将两个函数返回的grob对象传入gridExtra::arrangeGrob()完成布局,再统一渲染到画布。

内容的提问来源于stack exchange,提问作者Charité Learner

相关产品推荐
方舟 Agent Plan

超全模态模型 × Harness 升级,最新支持 Deepseek-V4.1-Flash、GLM-5.3 系列、Doubao-Seedream-5.0-pro、Kimi-K3 (部分), 限时 9.9 元起

最近更新时间:2026.10.26 01:45:05