R/pmtk_search_pubmed.R

In jaytimm/PubmedMTK: PubMed Mining Toolkit

#' Perform basic PubMed search.
#'
#' @name pmtk_search_pubmed
#' @param search_term Query term as character string
#' @param fields PubMed fields to query
#' @param sleep In seconds.
#' @param verbose Boolean
#' @param retmax Max records
#' @return A data frame of PMIDs   
#' 
#' @export
#' @rdname pmtk_search_pubmed
#' 
#' 
pmtk_search_pubmed <- function (search_term, 
                                fields = c('TIAB','MH'),
                                verbose = TRUE,
                                sleep = 1,
                                retmax = 5000000) { # max_n
  
  fout <- list()
  for(i in 1:length(search_term)){
    
    pre_url1 <- "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?"
    pre_url2 <- paste0("db=pubmed&retmax=", 
                       format(retmax, scientific = FALSE), 
                       "&term=")
      
    url_term_query <- gsub(" ", "+", search_term[i], fixed = TRUE)
    
    if(is.null(fields)) { 
      fields3 <- url_term_query 
      out_search <- search_term[i]} else{
        
        fields0 <- paste0('%5B', fields, '%5D')
        fields1 <- paste0(url_term_query, fields0)
        fields2 <- paste0(fields1, sep = '+OR+', collapse = '')
        fields3 <- gsub('\\+OR\\+$', '', fields2) 
      
        out_search <- paste0(search_term[i], 
                             paste0('[', fields, ']'), 
                             collapse = ' OR ')
      }
    
    ## Make url
    full_url <- paste0 (pre_url1, pre_url2, fields3)
    
    x <- httr::GET(full_url)
    x1 <- xml2::read_xml(x) 
    x2 <- xml2::xml_find_all(x1, './/Id')
    x3 <- xml2::xml_text(x2)
    
    if (length(x3) == 0) {x3 <- NA}
    
    fout[[i]] <- data.table::data.table(search_term = search_term[i], pmid = x3) 
    
    Sys.sleep(sleep)
    if(verbose){
      finally =  print(paste0(out_search, ': ', nrow(fout[[i]]), ' records'))
    }
  }
  return(data.table::rbindlist(fout))
}