R/plotlyGraphWidget.R

In ejanalysis/batch.summarizer: Summarize Multiple Buffer Reports from the (nonpublic) Batch Tool for EJSCREEN

#library(shiny)
# see https://plot.ly/r/getting-started/

graphOutput <- function(inputId, width="100%", height="550px") {
  tagList(
    singleton(tags$head(
      tags$script(src="plotlyGraphWidget.js")
    )),
    tags$iframe(id=inputId, src="https://plot.ly/~ejanalyst/31",
                class="graphs", seamless=TRUE, width=width, height=height)
  )
}

renderGraph <- function(expr, env=parent.frame(), quoted=FALSE) {
  ## This gets called when inputs change --
  ## Place data wrangling code in here
  ## and pass the result to the client
  ## to be graphed.

  installExprFunction(expr, "func", env, quoted)

  function(){
    data = func();

    ## this function returns a list of named lists that descripe
    ## valid postMessage commands to be sent to the embedded iframe.
    ## see binding.renderValue for the receiving JS side of this function
    ## and https://github.com/plotly/Embed-API for more about the postMessage
    ## graph messages.
    ##
    ## it can also pass non-valid postMessage commands with the task "custom",
    ## which will have custom, user-provided handling code, see plotlyGraphWidget.js
    ## `parseAndRelayMessages`

    ## In this case, I couldn't figure out how JSON-ify the R data structures into a
    ## correct postMessage format, so I set up the data structure "as is" and then
    ## parsed it on the JS side after the JSON was parsed

    return(list(
      list(
        id="mychart",
        task="custom",
        data=data$message
      ),
      list(
        id="mychart",
        task="relayout",
        update=list(margin=list(b=180, t=50))
      )
    ))
  }
}