Asks the Internet Archive which archived copy of a web address is closest
to a given time and returns a URL that downloads that copy. This is the
fetch failsafe the wider morie package family relies on: callers try the
live source first and fall back to the snapshot when the source is down,
has moved or has been taken offline.
fdl() uses it
automatically.
Value
A length-1 character string, the https snapshot URL of the
form https://web.archive.org/web/ followed by
<timestamp>/<url>, or NULL.
Details
The function sends one request to the availability API,https://archive.org/wayback/available?url=<url>[×tamp=<ts>],
where the timestamp part is sent only when ts is given.
The answer's archived_snapshots$closest is read. When the API reports
an available snapshot its URL is returned, with http:// rewritten to
https:// and passed through the same public-URL check as every other
URL the package fetches. The timestamp is passed to the API as given; the API
accepts any prefix of YYYYMMDDhhmmss and returns the snapshot nearest
to it, before or after.
NULL is returned when the API has no snapshot, when the request
fails (offline, an HTTP error, invalid JSON) or when the answer has no
usable URL. Errors: url that is not a single non-empty string or
is not a public web address; ts that is not NULL,
"" or a string of 1 to 14 digits.
Nothing is written.
Examples
# argument errors come before any request
try(wbsnpurl(NA))
#> Error : wbsnpurl: `url` must be a single non-empty string
try(wbsnpurl("https://www.r-project.org/", ts = "2020&x=1"))
#> Error : wbsnpurl: `ts` must be NULL, "" or a string of 1 to 14 digits (YYYYMMDDhhmmss or a prefix of it)
try(wbsnpurl(c("https://a.org", "https://b.org")))
#> Error : wbsnpurl: `url` must be a single non-empty string
if (FALSE) { # interactive()
# needs the network
wbsnpurl("https://www.r-project.org/")
# the snapshot nearest to January 2020
wbsnpurl("https://www.r-project.org/", ts = "202001")
}