Chapter 9 Shiny Apps

caladaptr can be used in Shiny Apps to retrieve data from the API on-the-fly.

To try out some of these apps, see https://github.com/ucanr-igis/caladaptr.apps

To show a spinner while data is being fetched, you can do this exampe:

show_spinner() 
my_vals <- ca_getvals_tbl(pt_cap(), quiet = TRUE) 
hide_spinner()

To show a progress bar that updates as data is being retrieved, you can pass a Shiny progress object to ca_getvals_tbl() (example):

progress <- shiny::Progress$new()
on.exit(progress$close())

progress$set(value = 0, message = "Fetching data from Cal-Adapt:")
    
## Get values
pt_daytmp_tbl <- ca_getvals_tbl(pt_prj_cap(), quiet = TRUE, shiny_progress = progress)

Additional Tips:

Deploying your Shiny app without being blocked

If you’ll be hosting your ShinyApp on the cloud, there’s a slight possibility that the IP address might be blocked by Cal-Adapt (due to precautions they’ve had to take to defend against DOS attacks)

  • ShinyApps.io works fine (those IP addresses have been white listed)

  • GoogleCloud IP addresses are all blocked

If this is a problem, one solution is to 1) get a VM with a stable IP address, 2) ask the administrator of the Cal-Adapt API to white list that IP address.

Memory

Also, be cognizant of the amount of RAM memory you have to work with. Climate data can quickly consume hundreds of megabytes, and its not uncommon for R to make copies of data frames when doing manipulations. If you’re working on a VM that only has 1 GB of RAM, you can quickly run out of memory. This can be difficult to diagnose, but if your app is working fine on your laptop, but disconnected unexpectedly when you run it on ShinyApps.io, then check the logs for an out-of-memory error.

Your options for dealing with out-of-memory errors include getting a VM with more RAM (which could cost you $), using ca_getvals_db() (which dumps values into a SQLite database) instead of ca_getvals_tbl(), redesigning your app for more efficient use of memory, or making your app available as a zip file that users can run locally from their RStudio.

More coming soon…