tests/testthat/test-kaldi.R

In torchaudio: R Interface to 'pytorch''s 'torchaudio'

test_that("kaldi__get_lr_indices_and_weights", {
  testthat::skip_if_not_installed("numbers")
  expect_no_error(x <- kaldi__get_lr_indices_and_weights(
    orig_freq = 4,
    new_freq = 3,
    output_samples_in_unit = 6,
    window_width = 2,
    lowpass_cutoff = 1,
    lowpass_filter_width = 2,
    device = torch::torch_device("cpu"),
    dtype = torch::torch_float()
  ))
  expect_tensor(x[[1]])
  expect_tensor(x[[2]])
  expect_error(x[[3]])
})

test_that("kaldi__get_num_lr_output_samples", {
  testthat::skip_if_not_installed("numbers")
  expect_no_error(x <- kaldi__get_num_lr_output_samples(100, 10, 5))
  expect_equal(x, 50)
})

test_that("kaldi_resample_waveform", {
  testthat::skip_if_not_installed("numbers")
  expect_no_error(x <- kaldi_resample_waveform(
    waveform = torch::torch_rand(2, 100),
    orig_freq = 10,
    new_freq = 9,
    lowpass_filter_width = 6
  ))
  expect_tensor(x)
})