tests/testthat/test-read_results.R

In xavierkamp/tsForecastR: Forecasting with Traditional Time Series and Machine Learning Models

library(testthat)
library(tsForecastR)

context("Test read results functions")

test_that("read_tsForecastR_works", {
  df <- read_tsForecastR(NULL)
  data <- seq(1:144)
  ts_data <- stats::ts(data, frequency = 12, start = c(1, 1))
  fc <- generate_fc_arima(ts_data)
  df <- read_tsForecastR(fc)
  expect_equal(nrow(df), 156)
  model_names <- c("arima", "ets", "stl")
  fc <- generate_fc(ts_data, model_names = model_names)
  df <- read_tsForecastR(fc)
  expect_equal(nrow(df), 156 * 3)
})
test_that("read_fc_from_file_works", {
  data <- seq(1:144)
  ts_data <- stats::ts(data, frequency = 12, start = c(1, 1))
  # create a tmp directory where results will be saved
  data_dir <- base::paste(getwd(), "tmp_testing_dir", sep = "/")
  while (dir.exists(data_dir)) {
    data_dir <- base::paste(data_dir, "1", sep = "")
  }
  model_names <- check_model_names(NULL)
  dir.create(data_dir)
  # generate forecasts and save in files
  fc <- generate_fc(ts_data,
                    model_names = model_names,
                    data_dir = data_dir)
  # check if files are properly named
  for (ts_name in base::names(fc)) {
    for (i in model_names) {
      file_name <- base::paste(data_dir,
                               base::paste(ts_name,
                                           i,
                                           sep = "_"),
                               sep = "/")
      expect_equal(base::file.exists(file_name), TRUE)
    }
  }
  expect_error(generate_fc(ts_data, model_names = model_names,
                           data_dir = "1234"))
  # read files
  df <-
    read_fc_from_file(base::colnames(ts_data),
                      data_dir = data_dir,
                      model_names = model_names)
  expect_equal(base::nrow(df), base::length(model_names)*156)
  # delete tmp dir with files
  base::unlink(data_dir, recursive = TRUE)
})
test_that("save_as_df_works", {
  data <- seq(1:144)
  ts_data <- stats::ts(data, frequency = 12, start = c(1, 1))
  fc <- generate_fc_arima(ts_data)
  df <- save_as_df(fc)
  expect_equal(is.data.frame(df), TRUE)
  # create tmp directory where results will be saved
  data_dir <- base::paste(getwd(), "tmp_testing_dir", sep = "/")
  while (base::dir.exists(data_dir)) {
    data_dir <- base::paste(data_dir, "1", sep = "")
  }
  base::dir.create(data_dir)
  # generate forecasts and save in files
  fc <- generate_fc_arima(ts_data,
                          data_dir = data_dir)
  # read files
  df <-
    save_as_df(data_colnames = "time_series_1",
               data_dir = data_dir,
               model_names = "arima")
  expect_equal(base::nrow(df), 156)
  expect_equal(base::ncol(df), 15)
  expect_equal(base::is.data.frame(df), TRUE)
  # delete tmp dir with files
  base::unlink(data_dir, recursive = TRUE)
})