How to prevent ggplot hoverOpts messages to go off screen with css

前端 未结 2 910
孤城傲影
孤城傲影 2021-01-28 09:51

When running the demo App below, the problem I run into is that hover messages for the bottom part of the plot end up running off the screen.

Does anybody know if there

相关标签:
2条回答
  • 2021-01-28 10:28

    Here is a solution with the JS library qTip2.

    library(shiny)
    library(ggplot2)
    library(DT)
    
    js_qTip <- "
    $('#hoverinfo').qtip({
      overwrite: true,
      content: {
        text: $('#tooltip').clone()
      },
      position: {
        my: '%s',
        at: '%s',
        target: [%s,%s],
        container: $('#FP1Plot1')
      },
      show: {
        ready: true
      },
      hide: {
        target: $('#FP1Plot1')
      },
      style: {
        classes: 'qtip-light'
      }
    });
    "
    
    ui <- fluidPage(
      tags$head(
        tags$link(rel = "stylesheet", href = "jquery.qtip.min.css"),
        tags$script(src = "jquery.qtip.min.js"),
        tags$script(
          HTML(
            'Shiny.addCustomMessageHandler("jsCode", function(mssg){setTimeout(function(){eval(mssg.value);},10);})'
          )
        )
      ),
      plotOutput('FP1Plot1' ,
                 width = 1000,
                 height = 700,
                 hover = hoverOpts(id = 'FP1Plot1_hover')),
      tags$div(id = "hoverinfo", style = "position: absolute;"),
      tags$div(DTOutput("tooltip"), style = "visibility: hidden;") # put this div at the very end of the UI
    )
    
    server <- function(input, output, session){
      output$FP1Plot1 <- renderPlot({
        ggplot(mtcars, aes(wt, mpg, color = as.factor(cyl))) + geom_point(size = 2)
      })
    
      tooltipTable <- eventReactive(input[["FP1Plot1_hover"]], { 
        hover <- input[["FP1Plot1_hover"]]
        if(is.null(hover)) return(NULL)
        dat <- mtcars
        point <- nearPoints(dat, hover, threshold = 15, maxpoints = 1)
        if(nrow(point) == 0) return(NULL)
        X <- point[["wt"]]
        Y <- point[["mpg"]]
        left_pct <- 
          (X - hover$domain$left) / (hover$domain$right - hover$domain$left)
        top_pct <- 
          (hover$domain$top - Y) / (hover$domain$top - hover$domain$bottom)  
        left_px <- 
          (hover$range$left + left_pct * (hover$range$right - hover$range$left)) / 
          hover$img_css_ratio$x 
        top_px <- 
          (hover$range$top + top_pct * (hover$range$bottom - hover$range$top)) / 
          hover$img_css_ratio$y 
        pos <- ifelse(left_pct<0.5,
                      ifelse(top_pct<0.5, 
                             "top left",
                             "bottom left"),
                      ifelse(top_pct<0.5,
                             "top right",
                             "bottom right"))
        list(data = t(point), pos = pos, left_px = left_px+10, top_px = top_px)
      }) # end of eventReactive
    
      output[["tooltip"]] <- renderDT({
        req(tooltipTable())
        datatable(tooltipTable()$data, colnames = NULL, 
                  options = list(dom = "t", ordering = FALSE))
      }, server = FALSE)
    
      observeEvent(tooltipTable(), {
        tt <- tooltipTable()
        session$sendCustomMessage(
          type = "jsCode", 
          list(value = sprintf(js_qTip, tt$pos, tt$pos, tt$left_px, tt$top_px))
        )
      })
    }
    
    shinyApp(ui, server)
    

    0 讨论(0)
  • 2021-01-28 10:46

    @ stephane, I came up with another solution using 'sending css' code to update the position of the hover message. The only problem I still run into, is that the position doesn't update until the message content changes for the second time to a point in a quadrant.

    there are 2 values for x offset and 2 for y offset, splitting the plot in 4 quadrants in fact. Switching to another quadrant puts the message in the last quadrants configuration, and doesn't correct this until I hover on a second point in the new quadrant.

    Do you have any idea to push the css change more effectively? perhaps with sendcustommessage or so? I tried to do it that way, but couldn't get it to work at all with that approach. Here is my code attempt so far:

    require('shiny')
    require('ggplot2')
    require('DT')
    
    ui <- pageWithSidebar(
    
      headerPanel("Hover off the page"),
      sidebarPanel(width = 2
      ),
      mainPanel(
        tags$head(
          tags$style('
                     #my_tooltip {
                     position: absolute;
                     pointer-events:none;
                     width: 10;
                     z-index: 100;
                     padding: 0;
                     font-size:10px;
                     line-height:0.6em
                     }
                     ')
        ),
    
        plotOutput('FP1Plot1' ,
                   width = 1000,
                   height = 800,
                   hover = hoverOpts(id = 'FP1Plot_1_hover', delay = 0)          
        ),
    
        uiOutput("my_tooltip"),
        uiOutput("my_tooltip_style"),
        style = 'width:1250px'
          )
        )
    
    server <- function(input, output, session) {
    
      # ranges <- reactiveValues()
    
    
      output$FP1Plot1 <- renderPlot({
        ggplot(mtcars, aes(wt, mpg, color = as.factor(cyl))) + geom_point() #+
          # coord_cartesian(xlim = ranges[[paste('FP1Plot1',  'x', sep = '')]], 
          #                 ylim = ranges[[paste('FP1Plot1',  'y', sep = '')]]
          # )          
      })
    
    
    
    
      # turn the hovers into 1 single reactive containing the needed information
      hoverReact <- reactive({
        ## in my real app I observer hover of all sub plots of all stages (7 pages with a multilot object)
        ## followed by code to store the page ID and plot NR as elements in hoverReact()
        hover <-  input[['FP1Plot_1_hover']]
    
        if(is.null(hover)) return(NULL)
         hover
    
      })
    
      ## debounce the reaction to calm down shiny
      hoverReact_D <- hoverReact %>% debounce(100)  ## attempt to stop hoverData <- reactive({}) from firing too often, which is needed when you have 10k point scatter plots.....
    
      hoverData <- reactive({
        hover <- hoverReact_D() 
        if(is.null(hover)) return(NULL)
        ## in my multi plot multi data frame I look up which dataframe to grab based on hover$plot_id as well as which x and y parameter are plotted
        hoverDF <- nearPoints(mtcars, coordinfo = hover, threshold = 15, maxpoints = 1, xvar = 'wt', yvar = 'mpg')
        hoverDF
      })
    
    
    
      hoverPos <- reactive({
        ## here I look up the position information of the hover whenevver hoverReact_D and hoverData change 
        hover <- hoverReact_D()
        hoverDF <- hoverData()
    
        if(is.null(hover)) return(NULL)
        if(nrow(hoverDF) == 0) return(NULL)
    
        ## in my real app the data is already 
        X <- hoverDF$wt[1]
        Y <- hoverDF$mpg[1]
    
        left_pct <- 
          (X - hover$domain$left) / (hover$domain$right - hover$domain$left)
    
        top_pct <- 
          (hover$domain$top - Y) / (hover$domain$top - hover$domain$bottom)  
    
        left_px <- 
          (hover$range$left + left_pct * (hover$range$right - hover$range$left)) / 
          hover$img_css_ratio$x 
    
        top_px <- 
          (hover$range$top + top_pct * (hover$range$bottom - hover$range$top)) / 
          hover$img_css_ratio$y 
    
        list(top = top_px, left = left_px)
      })
    
    
    
      output$my_tooltip_style <- renderUI({
        req(hoverPos())
        hover <- hoverPos()
        if(is.null(hover)) return(NULL)
    
        print(hover$top)
        offX <- if(hover$left  > 350) {-400} else {30}
        offY <- if(hover$top  > 350) {-290} else {10 }
    
        print(paste(offX, offY))
    
        cssMessage <- paste( "
                        $(document).ready(function() {
                             setTimeout(function(){
                             $('[id^=FP1Plot]').mousemove(function(e) {  
                             $('#my_tooltip').show();         
                             $('#my_tooltip').css({             
                             top: (e.offsetY +", offY, " ) + 'px',            
                             left: (e.offsetX +", offX, ") + 'px'        
                             });     
                             });     
                             })});", sep = '')
    
        tags$script(cssMessage)
    
    })
    
    
    
    
      output$GGHoverTable <- DT::renderDataTable({  
        df <- hoverData()
        if(!is.null(df)) {
          if(nrow(df)){
            df <- df[1,]
            DT::datatable(t(df), colnames = rep("", nrow(df)),
                          options = list(dom='t',ordering=F, autowidth = T))
          }
        }
      })
    
    
      output$my_tooltip <- renderUI({
        req(hoverData())
        wellPanel(
          dataTableOutput('GGHoverTable'),
          style = 'background-color: #FFFFFFE6;padding:10px; width:400px;border-color:#339fff')
      })
    
    
    }
    
    shinyApp(ui, server)
    
    0 讨论(0)
提交回复
热议问题