Skip to contents

Resolve the real filenames behind a source's download urls

Source: R/download_data.R
pg_resolve_filenames.Rd

Some sources serve files from urls that do not name them, such as https://ndownloader.figshare.com/files/17626052. Left alone, those land on disk as 17626052. This developer-facing helper asks each server what the file is actually called and records the answer as a tab-separated second column in the source's url list, so that pg_rawfiles() keeps deriving local paths offline and the bundled pgchecksum stays valid.

Usage

pg_resolve_filenames(id, write = FALSE, delay = 1)

Arguments

id

Character. Source id (UUID), as in pgsources.

write

Logical. If TRUE, rewrite the url list file under inst/extdata. Defaults to FALSE so that the proposal can be inspected first.

delay

Numeric. Seconds to wait between requests. Default 1.

Value

data.frame of urls, current filenames and resolved filenames (invisibly).

Details

The name is taken from the Content-Disposition header when the server sends one, then from the url the request ended up at after redirects, then from the stated url. None of these is reliable for every provider: sources behind a login redirect to the login page, and GitHub redirects to storage urls with opaque object keys. Read the proposed table before writing, which is why this is run by a maintainer and committed rather than run by users.

Examples

if (FALSE) { # \dontrun{
priogrid:::pg_resolve_filenames("d99fbea7-2a01-4221-b900-29a58d33f591")
priogrid:::pg_resolve_filenames("d99fbea7-2a01-4221-b900-29a58d33f591", write = TRUE)
} # }