Refactor use_course()

[jennybc]

Some code shared by @jeroen in a Slack conversation about use_course()

library(curl)

download_with_cd <- function(url, destdir = "."){
  h <- new_handle(noprogress = FALSE)
  tmp <- tempfile()
  curl::curl_fetch_disk(url, tmp, handle = h)
  headers <- curl::parse_headers_list(handle_data(h)$headers)
  filename <- stringr::str_match(headers[["content-disposition"]], "filename=.*")[1]
  filename <- sub("filename=", "", filename)
  outfile <- normalizePath(file.path(destdir, filename), mustWork = FALSE)
  file.rename(tmp, outfile)
  return(outfile)
}

download_with_cd("https://cran.ocpu.io/ggplot2/data/diamonds/csv", destdir = "~")

Other notes:

• If you use curl_fetch_disk you can check the status code also from handle_data().
• Using curl_download() might be easier -- automatically throws an error if the result is not http 200
• the trick is all functions in curl take a handle, and you can always use handle_data() to access the metadata after the request

[jennybc]

Progress-wise, at the very minimum, we should print “starting download”. Even better would be to access curl's progress reporting. cc @hadley