R/get_data.R

In como-ph/oxcgrt: An Interface to the Oxford COVID-19 Government Response Tracker API

################################################################################
#
#'
#' Get policy actions and stringency data from JSON
#'
#' @param json A JSON string, URL or file created using [get_json_time()] or
#'   [get_json_actions()].
#'
#' @return A tibble of time series stringency index data if `json` is a time
#'   **series endpoint** or a named list of two tibbles (the first tibble is
#'   named `policyActions` and the second tibble is named `stringencyData`) if
#'   `json` is a **policy actions endpoint**.
#'
#' @examples
#' ## Get time series JSON endpoint
#' x <- get_json_time(from = "2020-10-29", to = "2020-10-31")
#'
#' ## Get time series stringency index data
#' get_data(x)
#'
#' ## Get policy actions JSON endpoint
#' x <- get_json_actions(ccode = "AFG", from = NULL, to = "2020-07-16")
#'
#' ## Get data on policy actions and stringency index
#' get_data(x)
#'
#' @export
#'
#
################################################################################

get_data <- function(json) {
  ## Check if curl is available
  requireNamespace("curl", quietly = TRUE)

  ## Check if json is a vector
  if (length(json) > 1) {
    json <- json[1]
    warning("json has more than one value. Using first value only.",
            call. = TRUE, immediate. = TRUE)
  }

  ## Check if with internet
  if (!curl::has_internet()) {
    message("No internet connection detected. Data cannot be retrieved.")
    return(NULL)
  }

  if (httr::http_error(x = json)) {
    message("JSON data source seems to be broken. Data cannot be retrieved")
    return(NULL)
  }

  ## Extract data from JSON
  x <- jsonlite::fromJSON(txt = json, flatten = TRUE)

  if(stringr::str_detect(string = json, pattern = "[A-Z]{3}")) {
    ## Convert data.frames in list to tibble
    x[["policyActions"]] <- tibble::tibble(x[["policyActions"]]) %>%
      dplyr::mutate(
        policyvalue = as.integer(policyvalue),
        notes = as.character(notes)
      )
    x[["stringencyData"]] <- dplyr::bind_rows(x[["stringencyData"]])
  } else {
    ## Convert list to data.frame
    y <- unlist(x[["data"]], recursive = FALSE)
    z <- lapply(X = y, FUN = unlist)
    df <- dplyr::bind_rows(z)

    ## Convert column classes to appropriate classes
    x <- df %>%
      dplyr::mutate(
        date_value = as.Date(date_value, format = "%Y-%m-%d"),
        confirmed = as.integer(confirmed),
        deaths = as.integer(deaths),
        stringency_actual = as.numeric(stringency_actual),
        stringency = as.numeric(stringency),
        stringency_legacy = as.numeric(stringency_legacy),
        stringency_legacy_disp = as.numeric(stringency_legacy_disp)
      )
  }

  ## Return data
  x
}


################################################################################
#
#'
#' Get time series stringency index data from JSON
#'
#' @param json A JSON string, URL or file created using [get_json_time()]
#'
#' @return A tibble of time series stringency index data
#'
#' @examples
#' x <- get_json_time(from = "2020-07-18", to = "2020-07-20")
#'
#' get_data_time(x)
#'
#' @export
#'
#
################################################################################

get_data_time <- function(json) {
  ## Check if curl is available
  requireNamespace("curl", quietly = TRUE)

  ## Check if with internet
  if (!curl::has_internet()) {
    message("No internet connection detected. Data cannot be retrieved.")
    return(NULL)
  }

  if (httr::http_error(x = json)) {
    message("JSON data source seems to be broken. Data cannot be retrieved")
    return(NULL)
  }

  ## Extract data from JSON
  x <- jsonlite::fromJSON(txt = json, flatten = TRUE)

  ## Convert list to data.frame
  y <- unlist(x[["data"]], recursive = FALSE)
  z <- lapply(X = y, FUN = unlist)
  df <- dplyr::bind_rows(z)

  ## Convert column classes to appropriate classes
  df <- df %>%
    dplyr::mutate(
      date_value = as.Date(date_value, format = "%Y-%m-%d"),
      confirmed = as.integer(confirmed),
      deaths = as.integer(deaths),
      stringency_actual = as.numeric(stringency_actual),
      stringency = as.numeric(stringency),
      stringency_legacy = as.numeric(stringency_legacy),
      stringency_legacy_disp = as.numeric(stringency_legacy_disp),
      country_name = countrycode::countrycode(
        sourcevar = country_code,
        origin = "iso3c",
        destination = "country.name",
        custom_match = c("RKS" = "Kosovo")
      )
    ) %>%
    dplyr::relocate(country_name, .after = country_code)

  ## Return data
  df
}


################################################################################
#
#'
#' Get policy actions data from JSON
#'
#' @param json A JSON string, URL or file created using [get_json_actions()] or
#'   a vector of JSON strings or URLs.
#'
#' @return A tibble of policy actions with their respective policy values for
#'   specified country/countries and specified date/dates.
#'
#' @examples
#' ## Get relevant JSON for Afghanistan on 16 July 2020
#' x <- get_json_actions(ccode = "AFG", from = NULL, to = "2020-07-16")
#'
#' ## Get data on policy actions
#' get_data_action(x)
#'
#' ## Get relevant JSON for Afghanistan and Philippines for whole month of
#' ## October
#' x <- get_json_actions(ccode = c("AFG", "PH"),
#'                       from = "2020-10-29",
#'                       to = "2020-10-30")
#'
#' ## Get data on policy actions
#' get_data_actions(x)
#'
#' @rdname get_data_action
#' @export
#'
#
################################################################################

get_data_action <- function(json) {
  ## Check if curl is available
  requireNamespace("curl", quietly = TRUE)

  ## Check if json is a vector
  if(length(json) > 1) {
    json <- json[1]
    warning("json has more than one value. Using first value only.",
            call. = TRUE, immediate. = TRUE)
  }

  ## Check if with internet
  if (!curl::has_internet()) {
    message("No internet connection detected. Data cannot be retrieved.")
    return(NULL)
  }

  if (httr::http_error(x = json)) {
    message("JSON data source seems to be broken. Data cannot be retrieved")
    return(NULL)
  }

  ## Extract data from JSON
  x <- jsonlite::fromJSON(txt = json, flatten = TRUE)[["policyActions"]]

  ## Tidy up policyActions data.frame and convert to tibble
  x <- x %>%
    dplyr::mutate(
      policyvalue = ifelse(policyvalue == "NA", NA, policyvalue),
      policyvalue = as.integer(policyvalue),
      notes = as.character(notes),
      date_value = stringr::str_extract(
        json, pattern = "[0-9]{4}\\-[0-9]{2}\\-[0-9]{2}"
      ),
      date_value = as.Date(date_value, format = "%Y-%m-%d"),
      country_code = stringr::str_extract(json, pattern = "[A-Z]{3}"),
      country_name = countrycode::countrycode(
        sourcevar = country_code,
        origin = "iso3c",
        destination = "country.name",
        custom_match = c("RKS" = "Kosovo")
      )
    ) %>%
    dplyr::relocate(country_name, .before = policy_type_code) %>%
    dplyr::relocate(country_code, .before = country_name) %>%
    dplyr::relocate(date_value, .before = country_code) %>%
    tibble::tibble()

  ## Return data
  x
}


################################################################################
#
#'
#' @rdname get_data_action
#' @export
#'
#
################################################################################

get_data_actions <- function(json) {
  ## Check if curl is available
  requireNamespace("curl", quietly = TRUE)

  ## Extract data from JSON
  x <- lapply(X = json, FUN = get_data_action) %>%
    dplyr::bind_rows()

  ## Return data
  x
}