Add feature/instructions for additional R packages

[brandonerose]

I see the python version is able to add required packages such as plotly. Ideally, I would like to be able to use CRAN and GitHub packages in this format. I would also like to see how to handle data objects in readme or vignette.

[brandonerose]

So I found a way to add additional packages! It should be in the documentation as it will be a common question from people...

You can install many, but not all, R packages. The ones available are at https://repo.r-wasm.org/ and you can add them to your app.R script using the line webr::install("plotly"). Below is a full version that worked for me!

webr::install('plotly')
webr::install('DT')
library('shiny')
library('plotly')
library('DT')

ui <- fluidPage(
  titlePanel("Hello Shiny!"),
  sidebarLayout(
    sidebarPanel(
      sliderInput(inputId = "bins",
                  label = "Number of bins:",
                  min = 1,
                  max = 50,
                  value = 30)
    ),
    mainPanel(
      plotOutput(outputId = "distPlot"),
      DTOutput("table1"),
      plotlyOutput('fig')
    )
  )
)

server <- function(input, output) {
  output$distPlot <- renderPlot({
    x    <- faithful$waiting
    bins <- seq(min(x), max(x), length.out = input$bins + 1)
    hist(x, breaks = bins, col = "#75AADB", border = "white",
         xlab = "Waiting time to next eruption (in mins)",
         main = "Histogram of waiting times")
    })
  output$table1 <- renderDT({datatable(mtcars)})
  output$fig <- renderPlotly({
    Animals <- c("giraffes", "orangutans", "monkeys")
    SF_Zoo <- c(20, 14, 23)
    LA_Zoo <- c(12, 18, 29)
    data <- data.frame(Animals, SF_Zoo, LA_Zoo)
    fig <- plot_ly(data, x = ~Animals, y = ~SF_Zoo, type = 'bar', name = 'SF Zoo')
    fig <- fig %>% add_trace(y = ~LA_Zoo, name = 'LA Zoo')
    fig <- fig %>% layout(yaxis = list(title = 'Count'), barmode = 'group')
    fig
  })
}

shinyApp(ui = ui, server = server)

[schloerke]

Thank you for the reprex and approach!

[schloerke]

I've made #13 and #14 to explore how data is being loaded.

I'm going to alter the issue to only be about instructions for loading additional R packages.

[stla]

Hello,

Thank you for this discussion, it saved me ^^

IMHO a point is not clear in the documentation. After reading it, I tried webr::install in the R console (that returns an error), I didn't understand one had to call webr::install in the Shiny app.

[schloerke]

@stla I'm in the middle of a different PR that I do not want to exit from. Would you mind making a PR to the readme? (Thank you in advance!)

[stla]

Ok - will do.

[stla]

I think the webr::install way is a bit annoying: I have to comment these lines when trying the app locally. And I have to not forget to uncomment them before exporting.

There is surely a way to automatically handle that, something like (I don't know) if("install" %in% NAMESPACE("webr")), but it would be desirable to have a cleaner way (maybe a separate file to enter the required packages, I don't know).

[schloerke]

I am working on a PR for shinylive assets that will handle the webr::install() automatically for both exported apps and within quarto documents.

When that is merged, I'll remove the readme documentation suggesting to use webr::install() at the top of the app.

[stla]

There's no wish to use possibly different package versions with WebR in the future? Like renv, that would be great. In this case a separate file would be more appropriate.

[schloerke]

@stla

From #16 (comment)

In principle I think it should be possible to pre-build a Wasm filesystem image containing an R library (defined by renv, or a package list, or similar) ready to pre-load into the environment for a given shinylive/webR app.

Yes, it'll behave similar to an renv snapshot.

Currently all packages are the latest. Using the above approach will pull the in the packages statically. 1. reducing start up time. 2. freezing the packages being loaded.

[schloerke]

Original bug fixed in posit-dev/shinylive#68

Reduced comment in #21 as webr::install(PKG) is no longer required. Instead, include a library(PKG) or require("pkg") statement somewhere in your Shiny app.

Closing.