This is a complete rewrite of plumber. The purpose of the rewrite is to take everything we’ve learned from plumber, shed the bad decision that you inevitably make over the course of development, and start from scratch.
You’ll find that plumber2 is very similar to plumber in a lot of ways, but diverts in key areas, resulting in API incompatibility between the two packages. Because of this you may need to update your plumber APIs if switching to plumber2.
plumber2 can be installed from CRAN using
install.packages("plumber2")
. Alternatively you can install
the development version from github using pak
# install.packages("pak")
::pak("posit-dev/plumber2") pak
Below is a simple “hello world” API written for plumber2 that illustrates some of the differences from plumber:
#* Echo the parameter that was sent in
#*
#* @get /echo/<msg>
#*
#* @param msg:string The message to echo back.
#*
#* @response 200:{msg:string} A string containing the input message
#*
function(msg) {
list(
msg = paste0("The message is: '", msg, "'")
)
}
#* Plot out data from the palmer penguins dataset
#*
#* @get /plot
#*
#* @query spec:enum|Adelie, Chinstrap, Gentoo| If provided, filter the
#* data to only this species
#*
#* @serializer png{width = 700, height = 500}
#* @serializer jpeg{width = 700, height = 500}
#*
#* @async
function(query) {
<- penguins
myData <- "All Species"
title
# Filter if the species was specified
if (!is.null(query$spec)){
<- paste0("Only the '", query$spec, "' Species")
title <- subset(myData, species == query$spec)
myData if (nrow(myData) == 0) {
abort_internal_error("Missing data for {query$spec}")
}
}
plot(
$flipper_len,
myData$bill_len,
myDatamain=title,
xlab="Flipper Length (mm)",
ylab="Bill Length (mm)"
) }
Above you can both see some breaking changes and some new features in
action. The biggest breaking change is that parameters coming from the
path, the query string, and the body are now clearly separated. Only
parameters from the path are provided as direct arguments to the handler
function. Query and body parameters are accessible through the
query
and body
argument respectively. As can
be seen above they also use different tags in the documentation.
Speaking of documentation, the parsing of plumber blocks have been greatly improved. It is now built upon roxygen2, so it follows that convention, allowing multiline tags and defaulting to the first line as title and proceeding untagged lines as description. The ability to define input and output types has also been greatly expanded, adding the ability to define nested objects, adding default values and (as seen above) define enum (factors) to name a few. All input will get type checked and default value imputed if missing.
For the /plot
handler you can also see that it specifies
multiple serializers. Doing so will allow the client to request its
preferred response format using the Accept
header. plumber2
will then perform content negotiation to figure out the best response
format based on what it supports and what the client prefers.
Lastly, you can see a new tag (one of many) in the /plot
handler. @async
allows you to convert your handler into an
async handler automatically. It is still possible to create an async
handler manually by returning a promise, but the new tag significantly
simplifies this for the most classic cases. There is still overhead
involved in handling requests asynchronously so this is mainly a good
idea for longer running handlers, but it is shown here as an
example.