R/monitor_ar_retrieve_param.R

In quishqa/qualR: An R package to download São Paulo and Rio de Janeiro air pollution data

#' Download air quality and meteorology information from MonitorAr-Rio
#'
#' This function download air quality and meteorology measurements from
#' MonitorAr-Rio program from Rio de Janeiro city.
#'
#' @param start_date Date to start downloading in dd/mm/yyyy
#' @param end_date Date to end downloading in dd/mm/yyyy
#' @param aqs_code AQS code
#' @param parameters Parameters to download.
#' It can be a vector with many parameters.
#' @param to_local Date information in local time. TRUE by default.
#' @param verbose Print query summary.
#' @param to_csv Creates a csv file. FALSE by default
#' @param csv_path Path to save the csv file.
#'
#' @return data.frame with the selected parameter information
#' @export
#'
#' @examples
#' \dontrun{
#' # Downloading Ozone information from Centro AQS
#' # from February of 2019
#' date_start <- "01/02/2019"
#' date_end <- "01/03/2019"
#' aqs_code <- "CA"
#' param <- "O3"
#' ca_o3 <- monitor_ar_retrieve_param(date_start, date_end, aqs_code, param)
#'
#' }
monitor_ar_retrieve_param <- function(start_date, end_date, aqs_code,
                                      parameters, to_local=TRUE, verbose = TRUE,
                                      to_csv = FALSE, csv_path = ""){

  # Check if params are measured
  if (!prod(parameters %in% param_monitor_ar$code)){
    stop("One or all wrong param codes, please check monitor_ar_param", # nocov
         call. = FALSE)                                                 # nocov
  }

  if (!(aqs_code %in% aqs_monitor_ar$code )){
    stop("Wrong aqs_code, please check monitor_ar_aqs",                 # nocov
         call. = FALSE)                                                 # nocov
  }

  aqs_name <- aqs_monitor_ar$name[aqs_monitor_ar$code == aqs_code]

  # Adding query summary
  if (verbose){
    message("Your query is:")
    message("Parameter: ", paste(parameters, collapse = ", "))
    message("Air quality station: ", aqs_name)
    message("Period: From ", start_date, " to ", end_date)
  }

  start_date_format <- as.POSIXct(strptime(start_date, format="%d/%m/%Y"),
                                  tz = "UTC")
  end_date_format <-  as.POSIXct(strptime(end_date, format="%d/%m/%Y"),
                                 tz = "UTC")

  # Function to create the WHERE query
  where_query <- function(ds, de, aqs){
    ds_format <- format(ds, format="%Y-%m-%d %H:%M:%S")
    de_format <- format(de, format="%Y-%m-%d %H:%M:%S")
    date_range <- paste0("Data >= TIMESTAMP'", ds_format, "' AND ",
                         "Data <= TIMESTAMP'", de_format)
    aqs <- paste0("' AND Esta", iconv("\u00e7\u00e3", "", "utf-8", "byte"),
                  "o = '", aqs, "'")
    where <- paste0(date_range, aqs)
    return(where)
  }

  # Creating outFiled
  if (length(parameters) == 1){
    outfields <- paste("Data", parameters, sep = ",")
  } else {
    outfields <- paste("Data", paste(parameters, collapse = ","),
                       sep = ",")  # nocov
  }

  url <- paste0("https://services1.arcgis.com/OlP4dGNtIcnD3RYf/arcgis/rest/",
                "services/Qualidade_do_ar_dados_horarios_2011_2018/",
                "FeatureServer/2/query?")

  res <- httr::GET(url,
                   query = list(
                     where = where_query(start_date_format, end_date_format,
                                         aqs_code),
                     outFields = outfields,
                     f = 'json'
                   ))
  # Checking request
  if (res$status_code == 200){
    message("Succesful request")
    message(paste("Downloading ", paste(parameters, collapse = " ")))
  } else {
    stop("Unsuccesful request. Something goes wrong", call. = FALSE)     # nocov
  }

  # Reading json
  raw_data <- jsonlite::fromJSON(rawToChar(res$content))

  # Create an empty data frame is there is no input
  if (length(raw_data$feature) == 0){
    message("Data unavailable")                             # nocov start
    data_aqs <- data.frame(Data = NA)
    for (p in parameters){
      data_aqs[[p]] <- NA                                          # nocov end
    }
  } else {
    data_aqs <- raw_data$features[[1]]                             # nocov
  }


  # Changing epoch to human readable date
  data_aqs$data <- as.POSIXct(data_aqs$data/1000,
                              origin = "1970-01-01", tz = "UTC")

  # Check completion
  start_date2 <- paste(as.character(as.Date(start_date_format)), "00:30")
  end_date2 <- paste(as.character(as.Date(end_date_format) - 1), "23:30")

  all_dates <- data.frame(
    data = seq(as.POSIXct(strptime(start_date2, format="%Y-%m-%d %H:%M"),
                          tz = "UTC"),
               as.POSIXct(strptime(end_date2, format="%Y-%m-%d %H:%M"),
                          tz = "UTC"),
               by = "hour")
  )

  if (nrow(all_dates) != nrow(data_aqs)){
    message("Padding out missing dates with NA")                     # nocov
    data_aqs <- merge(all_dates, data_aqs, all = TRUE)                # nocov
  }

  # Adding aqs code to dataframe

  data_aqs$aqs <- aqs_code

  # Changing to local time
  if (to_local){
    attributes(data_aqs$data)$tzone <- "America/Sao_Paulo"
  }


  # Changing Data column name to date
  colnames(data_aqs)[1] <- "date"

  # Ensure columns as numeric
  cols_unchange <-  !(colnames(data_aqs) %in% c("date", "aqs"))
  data_aqs[, cols_unchange] <- sapply(data_aqs[, cols_unchange], as.numeric)

  # Changing wind speed and direction columns to ws and wd
  if ("dir_vento" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "dir_vento"] <- "wd"      # nocov
  }
  if ("vel_vento" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "vel_vento"] <- "ws"
  }
  if ("ur" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "ur"] <- "rh"      # nocov
  }
  if ("chuva" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "chuva"] <- "rain"      # nocov
  }
  if ("Pres" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "Pres"] <- "p"      # nocov
  }
  if ("temp" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "temp"] <- "tc"    # nocov
  }
  if ("PM2_5" %in% names(data_aqs)){
    names(data_aqs)[names(data_aqs) == "PM2_5"] <- "pm25"  # nocov
  }

  names(data_aqs)[-c(1, ncol(data_aqs))] <- tolower(
    names(data_aqs)[-c(1, ncol(data_aqs))]
  )


  if (to_csv){                                                  # nocov start
    write_csv                                                                                                       (data_aqs, aqs_name, start_date, end_date, parameters, csv_path)
  }                                                             # nocov end

  return(data_aqs)
}