Publish a Shiny App

To publish a R shiny application on the SESYNC server, your files will need to be copied from your working directory to the shiny-apps-data shared folder (/nfs/shiny-apps-data on RStudio Server). Please contact SESYNC IT staff if you would like to host an app on SESYNC’s Shiny Server for the duration of your project’s lifecycle.

Synchronizing changes between development and published version

To sync changes to the application between your working directory (which we will call the development version of the app) and shiny-apps-data (the published version), it is recommended that you use GitLab. Here are the basic steps:

1. Create a new Git project for your application (see How do I create a new Git Project).
2. In your working directory, either:
   • create a new R project from your Git repository (see How do I create an R Studio Project from Git); or
   • associate an existing R project to the repository; when this page appears after creating your GitLab project, follow the command line instructions under Existing Git repo. (You can send command line instructions from a SESYNC SSH session, or from a RStudio Server session using the system() function in R).
3. As you are working on the development version of your app, periodically commit your changes, and push the version you want to share to Git. Both operations can be done from the “Git” tab of the RStudio Server enviornment (see screenshot below).
4. Create another R project from your Git repository (see How do I create an R Studio Project from Git) , this time in the shiny-apps-data folder, which will contain the published version of the app.
5. Anytime you push a new version from your development directory to the Git repository, you can pull it in the shiny-apps-data version of the project to publish it. See the screenshot below showing the location of the pull button in the RStudio Server interface.

Managing package dependencies

To make sure all packages your application needs are installed on the SESYNC shiny server, please follow these steps.

1. Create a dependencies.R file in your project. This R script should verify that each required package is installed and exceeds the minimum package version requirement. Specify the version of the package used in your development as the minimum version. Missing or out of date packages must be installed by the script and then loaded. Here is such a script that requires the “dplyr”, “raster”, “leaflet”, and “ggplot2” packages. For development versions of packages or those not on CRAN, provide the repository address in the format username/repo

# List the package and version as below
dependencies <- read.csv(textConnection("
  Package,     Min.Version, repo
  dplyr,       0.5.0,
  raster,      2.5.8,
  leaflet,     2.0.0.9000, rstudio/leaflet
  ggplot2,     2.0.1,
  "), stringsAsFactors = FALSE, strip.white = TRUE)

## No changes necessary below. ##

# Import installed package versions
pkgs <- installed.packages()
rownames(pkgs) <- c()
pkgs <- data.frame(pkgs, stringsAsFactors = FALSE)

# Compare requirements to installed packages
pkgs <- merge(dependencies, pkgs, by="Package", all.x=TRUE)

# Filter out packages meeting minimum version requirement
pkgs <- pkgs[mapply(compareVersion, pkgs$Min.Version, pkgs$Version) > 0, ]

# Install missing and newer packages
cran <- pkgs[is.na(pkgs$repo), ]
lapply(cran$Package, install.packages)
github <- pkgs[!is.na(pkgs$repo), ]
lapply(github$repo, devtools::install_github)

# Require dependencies [optional]
lapply(dependencies$Package, require, character.only=TRUE)

2. Add “source(dependencies.R)” to the top of your app.R single-file Shiny script.