vignettes/host_cerebro_on_shinyapps.Rmd
host_cerebro_on_shinyapps.RmdIn this vignette, I explain how one can host Cerebro on shinyapps.io. For a more detailed tutorial on how to host Shiny apps on shinyapps.io, take a look here: https://shiny.rstudio.com/articles/shinyapps.html
shinyapps.io is a great place to host Shiny apps and make them accessible for other users. You will need an account on the website. The free account comes with a limit on (1) the number of applications that can be hosted in parallel and (2) the number of active hours, but it is enough to get some experience and figure out if this service fits your purpose.
To upload Cerebro to shinyapps.io, we will use the rsconnect package, therefore it needs to be installed first. Of course, we also need the cerebroApp package.
if ( 'rsconnect' %in% installed.packages() == FALSE ) install.packages('rsconnect')
if ( 'cerebroApp' %in% installed.packages() == FALSE ) install.packages('romanhaa/erebroApp')
library(rsconnect)Then, you need to provide rsconnect with your shinyapps.io account info. On the shinyapps.io website, in the left navigation bar, go to “Account” and then “Tokens”. A token should have already been generated for you. Take the token and secret, your account name, and run the command below.
rsconnect::setAccountInfo(name="<ACCOUNT>", token="<TOKEN>", secret="<SECRET>")Next, we need to copy the Shiny files and some other data for Cerebro v1.3 to a dedicated directory, here ~/test_cerebro_shinyapps. If you prefer to host a different version of Cerebro, adapt the code accordingly.
app_directory <- '~/test_cerebro_shinyapps'
dir.create(app_directory)
dir.create(glue::glue('{app_directory}/shiny'))
file.copy(
system.file('extdata', package = 'cerebroApp'),
app_directory,
recursive = TRUE
)
unlink(
c(glue::glue('{app_directory}/extdata/v1.0'),
glue::glue('{app_directory}/extdata/v1.1'),
glue::glue('{app_directory}/extdata/v1.2')),
recursive = TRUE
)
file.copy(
system.file('shiny', 'v1.3', package = 'cerebroApp'),
glue::glue('{app_directory}/shiny'),
recursive = TRUE
)app.R fileSince we are not using the usual launchCerebro() function to start Cerebro, we have to prepare a dedicated app.R file that will do the same for us. Use your favorite text editor to open a new file, paste the code below into it, and save it as ~/test_cerebro_shinyapps/app.R.
## load packages -------------------------------------------------------------##
library(dplyr)
library(DT)
library(plotly)
library(shiny)
library(shinydashboard)
library(shinyWidgets)
## set parameters ------------------------------------------------------------##
Cerebro.options <<- list(
"mode" = "open",
"crb_file_to_load" = "extdata/v1.3/example.crb",
"cerebro_root" = "."
)
shiny_options <- list(
maxRequestSize = 800 * 1024^2,
port = 1337
)
## load server and UI functions ----------------------------------------------##
source(glue::glue("{Cerebro.options$cerebro_root}/shiny/v1.3/shiny_UI.R"))
source(glue::glue("{Cerebro.options$cerebro_root}/shiny/v1.3/shiny_server.R"))
## launch app ----------------------------------------------------------------##
shiny::shinyApp(
ui = ui,
server = server,
options = shiny_options
)Now, it is time to upload the app the shinyapps.io. With rsconnect, that can be quickly done using the code below.
Since we have a dependency from Bioconductor, we have to add the current release (3.11) to the repositories using the setRepositories() command. When doing so, you will be asked which repositories should be used in this session. For what we are doing here, selecting CRAN and BioC software should be sufficient.
When you run the deployApp() command, you might receive a warning about a hidden browser() call in the shiny/v1.3/about/server.R file, but you can ignore that and proceed. Alternatively, you can simply comment out the respective line.
setRepositories(addURLs = c(BioC = "https://bioconductor.org/packages/3.11/bioc"))
rsconnect::deployApp('~/test_cerebro_shinyapps/', appName = 'Cerebro')Uploading and preparing the app might take a few minutes, but once it finishes successfully, Cerebro should open in your browser. Using the shinyapps.io dashboard, you can manage the app, check usage metrics and logs, etc. An important parameter is the “Instance size”, which controls the available memory and can be controled in the “Settings” tab of the dashboard. The larger the data set you want to upload to Cerebro, the more memory will be necessary.