Augment reference list with DOIs automatically?

[LukasWallrich]

The following code uses crossref to retrieve DOIs for a reference list - worth including in the Annotator rather than only linking to the tool ...

# Requires: httr2, rvest, xml2
library(httr2)
library(rvest)
library(xml2)
library(stringr)

stq_post <- function(refs, email, include_pm = FALSE, multihit = FALSE, timeout_sec = 60) {
  stopifnot(is.character(refs), length(refs) >= 1, nzchar(email))
  
  url <- "https://apps.crossref.org/SimpleTextQuery"
  ua  <- user_agent(sprintf("R httr; mailto:%s", email))
  h   <- handle("https://apps.crossref.org")  # keeps cookies across requests
  
  # 1) GET to start a session (needed; direct POST fails)
  invisible(GET(url, handle = h, ua, timeout(30)))
  
  # 2) POST the form
  body <- list(
    email    = email,
    command  = "Submit",
    freetext = if (length(refs) > 1) paste(refs, collapse = "\n") else refs
  )
  if (include_pm) body$includePM <- "on"
  if (multihit)   body$multihit  <- "on"
  
  resp <- POST(url, handle = h, ua, body = body, encode = "form", timeout(timeout_sec))
  stop_for_status(resp)
  html <- content(resp, as = "text", encoding = "UTF-8")
  
  # Extract DOIs from returned page
  doc   <- read_html(html)
  hrefs <- html_attr(html_elements(doc, "a"), "href")
  dois  <- hrefs[grepl("doi\\.org/", hrefs, ignore.case = TRUE)]
  dois  <- unique(str_replace(tolower(dois), "^https?://(dx\\.)?doi\\.org/", ""))
  
  list(dois = dois, html = html)
}


# Example:
refs <- c(
   "Wallrich, L., Opara, V., Wesołowska, M., Barnoth, D., & Yousefi, S. (2024). The relationship between team diversity and team performance: reconciling promise and reality through a comprehensive meta-analysis registered report. Journal of Business and Psychology, 39(6), 1303-1354.",
   "Knowles MR, Boucher RC (2002) Mucus clearance as a primary innate defense mechanism for mammalian airways. J Clin Investig 109: 571–577."
 )
 res <- stq_post(refs, include_pm = TRUE, multihit = FALSE, email = "you@uni.ac.uk")
 str(res$dois)