R/gcite_papers.R

In muschellij2/gcite: Google Citation Parser

#' @title Parse Google Citation Index
#' @description Parses a google citation indices (h-index, etc.) from main page
#'
#' @param doc A xml_document or the url for the main page
#' @param ... Additional arguments passed to \code{\link{GET}} if \code{doc}
#' is a URL
#'
#' @return A matrix of indices
#' @export
#' @importFrom rvest html_table html_nodes
#' @importFrom stats reshape
#' @examples 
#' library(httr)
#' library(rvest) 
#' url = "https://scholar.google.com/citations?user=T9eqZgMAAAAJ"
#' url = gcite_url(url = url, pagesize = 10, cstart = 0) 
#' if (!is_travis() & !is_cran()) {
#' ind = gcite_papers(url)
#' doc = content(httr::GET(url))
#' ind = gcite_papers(doc)
#' ind_nodes = rvest::html_nodes(doc, "#gsc_a_b")
#' ind = gcite_papers(ind_nodes)
#' }
gcite_papers <- function(doc, ...){
  UseMethod("gcite_papers")
}

#' @rdname gcite_papers
#' @export
gcite_papers.xml_nodeset = function(doc, ...) {
  doc = as_list(doc)[[1]]
  gcite_papers(doc)
}

#' @rdname gcite_papers
#' @export
gcite_papers.xml_document = function(doc, ...) {
  doc = rvest::html_nodes(doc, css = "#gsc_a_b")
  gcite_papers(doc)
}

#' @rdname gcite_papers
#' @export
gcite_papers.character = function(doc, ...) {
  res = httr::GET(url = doc, ...)
  stop_for_status(res)
  doc = httr::content(res)
  gcite_papers(doc)
}

#' @rdname gcite_papers
#' @export
gcite_papers.default = function(doc, ...) {
  
  tab = doc
  tab = lapply(tab, function(x) {
    x = unlist(x, recursive = FALSE)
    names(x) = gsub("^td[.]", "", names(x))
    x
  })
  parse_a = function(a) {
    href = attr(a, "data-href")
    if (is.null(href)) {
      href = attr(a, "href")
    }
    value = a
    if (length(value) == 0) {
      value = NA
    }
    a = cbind(value = value, 
              # href = attr(a, "href"),
              href = href,
              class = attr(a, ".class"))
    a = data.frame(a, stringsAsFactors = FALSE)
  }
  fake_a_df = data.frame(
    title = NA, 
    title_link = NA, 
    n_citations = NA,
    n_citations_link = NA)
  # i=1
  get_data = function(x) {
    # print(i)
    # i <<- i +1
    nx = names(x)
    a = nx == "a"
    a = lapply(x[a], parse_a)
    a = do.call("rbind", a)
    if (NROW(a) == 0) {
      return(fake_a_df)
    }
    a$id = 1
    a$value = trimws(a$value)
    a$href = trimws(a$href)
    a$class[ grepl("gsc_a_ac", a$class)] = "gsc_a_ac"
    a = reshape(a, direction = "wide", idvar = "id", 
                timevar = "class", v.names = c("value", "href"))
    cn = colnames(a)
    cn[ cn == "value.gsc_a_at"] = "title"
    cn[ cn == "value.gsc_a_ac"] = "n_citations"
    cn[ cn == "href.gsc_a_at"] = "title_link"
    cn[ cn == "href.gsc_a_ac"] = "n_citations_link"
    # Bug fix https://github.com/muschellij2/gcite/issues/1
    cn[ cn == "href.gsc_a_ac gsc_a_acm"] = "n_citations_link"
    cn[ cn == "value.gsc_a_ac gsc_a_acm"] = "n_citations"
    
    colnames(a) = cn
    # print(cn)
    a$id = NULL
    get_cols = c("title", "title_link", "n_citations", "n_citations_link")
    not_there = setdiff(get_cols, colnames(a))
    if (length(not_there) > 0) {
      warning(paste0("There is an odd error here with parsing an ", 
                     "individual page, please submit bug to ",
                     "https://github.com/muschellij2/gcite/issues"))
      for (i in not_there) {
        a[, i] = NA
      }
    }
    a = a[, c("title", "title_link", "n_citations", "n_citations_link")]
    
    a
  }
  all_a = lapply(tab, get_data)
  all_a = do.call("rbind", all_a)
  rownames(all_a) = NULL
  
  
  all_a$n_citations = as.numeric(all_a$n_citations)
  
  all_a$n_citations_link[ all_a$n_citations_link == "" ] = NA
  all_a$title_link[ all_a$title_link == "" ] = NA
  all_a$title_link[ !is.na(all_a$title_link)] = 
    paste0("https://scholar.google.com",
           all_a$title_link[ !is.na(all_a$title_link)])
  if (all(is.na(all_a))) {
    return(NULL)
  }
  all_a$n_citations[ is.na(all_a$n_citations)] = 0  
  return(all_a)
}