Simple async client

Description

A client to work with many URLs, but all with the same HTTP method

Arguments

Details

Methods

get(path, query, disk, stream, ...)

make async GET requests for all URLs

post(path, query, body, encode, disk, stream, ...)

make async POST requests for all URLs

put(path, query, body, encode, disk, stream, ...)

make async PUT requests for all URLs

patch(path, query, body, encode, disk, stream, ...)

make async PATCH requests for all URLs

delete(path, query, body, encode, disk, stream, ...)

make async DELETE requests for all URLs

head(path, ...)

make async HEAD requests for all URLs

See HttpClient() for information on parameters.

Value

a list, with objects of class HttpResponse(). Responses are returned in the order they are passed in. We print the first 10.

Failure behavior

HTTP requests mostly fail in ways that you are probably familiar with, including when there's a 400 response (the URL not found), and when the server made a mistake (a 500 series HTTP status code).

But requests can fail sometimes where there is no HTTP status code, and no agreed upon way to handle it other than to just fail immediately.

When a request fails when using synchronous requests (see [HttpClient]) you get an error message that stops your code progression immediately saying for example:

- "Could not resolve host: https://foo.com" - "Failed to connect to foo.com" - "Resolving timed out after 10 milliseconds"

However, for async requests we don't want to fail immediately because that would stop the subsequent requests from occurring. Thus, when we find that a request fails for one of the reasons above we give back a [HttpResponse] object just like any other response, and:

- capture the error message and put it in the 'content' slot of the response object (thus calls to 'content' and 'parse()' work correctly) - give back a '0' HTTP status code. we handle this specially when testing whether the request was successful or not with e.g., the 'success()' method

See Also

Other async: AsyncVaried

Examples

## Not run: 
cc <- Async$new(
  urls = c(
    'https://httpbin.org/',
    'https://httpbin.org/get?a=5',
    'https://httpbin.org/get?foo=bar'
  )
)
cc
(res <- cc$get())
res[[1]]
res[[1]]$url
res[[1]]$success()
res[[1]]$status_http()
res[[1]]$response_headers
res[[1]]$method
res[[1]]$content
res[[1]]$parse("UTF-8")

lapply(res, function(z) z$parse("UTF-8"))

# using auth with async
dd <- Async$new(urls = rep('https://httpbin.org/basic-auth/user/passwd', 3))
res <- dd$get(auth = auth(user = "user", pwd = "passwd"))
res
vapply(res, function(z) z$status_code, double(1))
vapply(res, function(z) z$success(), logical(1))
lapply(res, function(z) z$parse("UTF-8"))

# failure behavior
## e.g. when a URL doesn't exist, a timeout, etc.
urls <- c("http://stuffthings.gvb", "https://foo.com", 
  "https://httpbin.org/get")
conn <- Async$new(urls = urls)
res <- conn$get()
res[[1]]$parse("UTF-8") # a failure
res[[2]]$parse("UTF-8") # a failure
res[[3]]$parse("UTF-8") # a success

## End(Not run)