R/utils.r

Defines functions msg get_counts mk_path val_input pb transform_query fix_list chunks make_path rebi_GET batch_size anno_path rest_path base_uri

# utils

# base uri
base_uri <- function()
  "https://www.ebi.ac.uk"

# rest path
rest_path <- function()
  "europepmc/webservices/rest"
# annotations API
anno_path <- function()
  "europepmc/annotations_api"

# set user agent
ua <- httr::user_agent("https://github.com/ropensci/europepmc")

# check data sources
supported_data_src <-
  c("agr", "cba", "ctx", "eth", "hir", "med", "pat",
    "pmc", "ppr")

# default batch size
batch_size <- function()
  100


# Common methods:

# Implementing GET method and json parser for EPMC
rebi_GET <- function(path = NULL, query = NULL, ...) {
  if (is.null(path) && is.null(query))
    stop("Nothing to search")
  # call api, decode workaround because Europe PMC only accepts decoded cursor
  req <- httr::RETRY("GET", urltools::url_decode(
    httr::modify_url(
    base_uri(), path = path, query = query
    )
  ), #httr::verbose(),
  ua)
  # check for http status
  httr::stop_for_status(req)
  # load json into r
  out <- httr::content(req, "text")
  # valid json
  if (!jsonlite::validate(out))
    stop("Upps, nothing to parse, please check your query")
  # return core format as list
  if (length(query$resulttype) == 1 && query$resulttype == "core") {
    doc <- out
  } else {
    doc <- jsonlite::fromJSON(out)
  }
  if (!exists("doc"))
    stop("No json to parse", call. = FALSE)
  return(doc)
  }


# make paths according to limit and request methods
make_path <- function(hit_count = NULL,
                      limit = NULL,
                      format = "json",
                 ...) {
  limit <- as.integer(limit)
  limit <- ifelse(hit_count <= limit, hit_count, limit)
  if (limit > batch_size()) {
    tt <- chunks(limit)
    arg <- lapply(1:(tt$page_max - 1), function(x)
      list(page = x,
           pageSize = batch_size(),
           format = format,
           ...)
    )
    arg[[tt$page_max]] <- list(page = tt$page_max,
                              pageSize = tt$last_chunk,
                              format = format,
                              ...)
  } else {
    arg <- NULL
    arg[[1]] <- list(pageSize = limit, format = format, ...)
  }
    arg
}

# calculate number of page chunks needed in accordance with limit param
chunks <- function(limit, ...) {
  if (all.equal(limit / batch_size(),
                as.integer(limit / batch_size())) == TRUE) {
    page_max <- limit / batch_size()
    last_chunk <- batch_size()
  } else {
    page_max <- as.integer(limit / batch_size()) + 1
    last_chunk <- limit - ((page_max - 1) * batch_size())
  }
  list(page_max = page_max, last_chunk = last_chunk)
}


# fix to remove columns that cannot be easily flatten from the data.frame
fix_list <- function(x) {
  if (!is.null(x))
    tmp <- plyr::ldply(x, is.list)
  tmp[tmp$V1 == TRUE, ".id"]
}

# URL encode parts of the query. Needed because EPMC inconsistently
# deal with URL encoding
transform_query <- function(query = NULL) {
  # check
  if (is.null(query))
    stop("No query provided")
  query <- URLencode(query)
  return(query)
}

#' Progress bar definition
#'
#' @param limit maximum number of hits to be returned
#'
#' @noRd
pb <- function(limit) {
  if(!is.integer(limit))
    stop("Limit is not integer")
  progress::progress_bar$new(total = limit / batch_size(),
                                 format = "(:spin) [:bar] :percent",
                                 clear = FALSE, width = 60)
}

# common methods for id based request -------------------------------------

#' Validate inputs for ID-based requests
#'
#' @noRd
val_input <- function(ext_id,
                      data_src,
                      limit,
                      verbose) {
  if (is.null(ext_id))
    stop("Please provide a publication id")
  if (!tolower(data_src) %in% supported_data_src)
    stop(
      paste0(
        "Data source '",
        data_src,
        "' not supported. Try one of the
        following sources: ",
        paste0(supported_data_src, collapse = ", ")
      )
    )
    stopifnot(is.numeric(limit), is.logical(verbose))
}
#' Make path for  ID-based requests
#'
#'@noRd
mk_path <- function(req_method, data_src, ext_id, ...) {
  if (!is.null(req_method))
    c(rest_path(), data_src, ext_id, req_method)
  else
    stop("No request method provided")
}
#' Get hit counts for ID-based requests
#'
#' @noRd
get_counts <- function(path, ...) {
  if (!is.null(path))
    doc <- rebi_GET(path = path,
                    query = list(format = "json", pageSize = batch_size(), ...))
  doc$hitCount
}
#' Provide feedback about how many records where found and
#' will be returned
#'
#' @noRd
msg <- function(hit_count, limit, verbose) {
  if (verbose == TRUE)
    message(paste0(
      hit_count,
      " records found. Returning ",
      ifelse(hit_count <= limit, hit_count, limit)
    ))
}

Try the europepmc package in your browser

Any scripts or data that you put into this service are public.

europepmc documentation built on Sept. 21, 2023, 1:07 a.m.