plot.CA() 在本地应用闪亮的应用程序中呈现，但在部署应用程序时不呈现

Question

我创建了一个 Shiny 应用程序，它接受用户输入并创建 CA 图。当我在本地运行应用程序时它工作得很好，但由于某种原因，当我部署仪表板时，绘图的图像不会出现。我可以在日志中看到，数据上传并重新格式化为适当的数据框工作正常，但绘图本身无法呈现。

有人知道为什么会这样吗？我在下面发布了我的代码（您会在我的代码中看到一些用于调试的 print() 行）。任何帮助将不胜感激！

    #PERCEPTUAL MAPPING DASHBOARD

library(FactoMineR)
library(factoextra)
library(SensoMineR)
library(shinythemes)
library(ca)



ui <- fluidPage(theme = shinytheme("darkly"),
  # Application title
  titlePanel("Perceptual Map Dashboard"),
  sidebarLayout(
    # Sidebar with a slider and selection inputs
    sidebarPanel(

      #Excel doc row and column names
      numericInput(inputId="startcol",label="Input start column of CSV file:",value="", min=1,max=10000),
      numericInput(inputId="endcol",label="Input end column of CSV file:",value="", min=1,max=10000),
      #Inputing brands and emotions
      br(),
      numericInput(inputId = "rownums",label = "How many emotions/characteristics are you evaluating?",value = "", min = 1,max = 10000),
      br(),
      h6("Note: Please enter brands and emotions/characteristics in the order that they appear in the excel document exported from Survey Gizmo."),
      textInput ( 'brands', 'List the brands included in your perceptual map (separated by commas):', value=""),
      textInput ( 'emotions', 'List the emotions/characteristics included in your perceptual map (separated by commas):', value=""),
      #Removing brands and emotions

      #Select graph type
      textInput(inputId="plottitle",label="Title your graph:"),
      #Upload Excel Grid
      fileInput(inputId = 'data', 'Upload CSV File',
                accept=c('.csv')),
      actionButton("go","Create Map")
    ),

    # Visual Output
    mainPanel(
      wellPanel(h4('Visual'),
                h5("Once your visual appears, just right click it to save it as a .png file.")),
      plotOutput(outputId = "plot",  width = "100%", height=500)
      # downloadButton("downloadPlot", "Download Visual")
    )
  )
)



server <- function(input,output){



  K <- eventReactive(input$go,{


      x <- read.csv(input$data$datapath, header = F)

      x[!is.na(x)] <- 1
      x[is.na(x)] <- 0
      x<-x[,as.numeric(input$startcol):as.numeric(input$endcol)]
      column.sums<-colSums(x)
      print(column.sums)
      pmd.matrix<-matrix(column.sums, byrow = T, nrow=as.numeric(input$rownums))
      pmd.df2<-as.data.frame(pmd.matrix)
      colnames(pmd.df2) = unlist(strsplit(as.character(input$brands),","))
      print(pmd.df2)
      row.names(pmd.df2)= unlist(strsplit(as.character(input$emotions),","))
      print(pmd.df2)
      pmd.df2[-nrow(pmd.df2),] 
      print(pmd.df2)
      fit <- CA(pmd.df2, graph=F)
      return(fit)


    })

  p <- eventReactive(input$go,{

      input$plottitle

  })

  output$plot<- renderPlot({

      plot.CA(K(), col.row = "blue", col.col="black",  cex=1, new.plot=T,
                    title=p())


  })

}

shinyApp(ui = ui, server = server)

Answer 1

我给你的建议是检查这个问题是否特定于你的情节，或者plot.CA 是否一般不适用于闪亮。尝试“部署”（显然，您不使用网络服务器？）以下应用程序

library(FactoMineR)
library(shiny)

data(children)
res.ca <- CA(children, col.sup = 6:8, row.sup = 15:18)

shinyApp(
  fluidPage(plotOutput("plot")),
  function(input, output, sesison) {
    output$plot <- renderPlot({
      plot.CA(res.ca)
    })
  }
)

如果这确实有效，则您的模型可能有问题，或者ca 包和FactorMineR 包之间可能存在名称串通。

如果这不起作用，请尝试以下方法

## use same data/libraries as above

myfile <- tempfile(fileext = ".png")

shinyApp(
  fluidPage(imageOutput("plot")),
  function(input, output, sesison) {
    output$plot <- renderImage({
      png(file = myfile)
      plot.CA(res.ca)
      dev.off()
      list(src = myfile)
    }, deleteFile = FALSE)
  }
)

看看

1. 应用现在是否可以运行
2. myfile 是否被创建并包含合理的内容。