def test_magic_setitem_wrong_n_bins():
"""Test the setitem for FrequencyData with wrong number of bins."""
freq_a = FrequencyData([1, 0, -1], [0, .1, .3])
freq_b = FrequencyData([2, 0, -2, 0], [0, .1, .3, .7])
with pytest.raises(ValueError):
freq_a[0] = freq_b
def test_magic_setitem():
"""Test the setitem for FrequencyData."""
freqs = [0, .1, .3]
freq_a = FrequencyData([[1, 0, -1], [1, 0, -1]], freqs)
freq_b = FrequencyData([2, 0, -2], freqs)
freq_a[0] = freq_b
npt.assert_allclose(freq_a.freq, np.asarray([[2, 0, -2], [1, 0, -1]]))
def test_magic_setitem_wrong_fft_norm():
"""Test the setitem for FrequencyData with wrong FFT norm."""
freqs = [0, .1, .3]
freq_a = FrequencyData([1, 0, -1], freqs)
freq_b = FrequencyData([2, 0, -2], freqs, fft_norm='psd')
with pytest.raises(ValueError):
freq_a[0] = freq_b
def test_data_frequency_setter_freq():
"""Test the setter for the frequency data."""
data_a = [1, 0, -1]
data_b = [2, 0, -2]
freqs = [0, .1, .3]
freq = FrequencyData(data_a, freqs)
freq.freq = data_b
npt.assert_allclose(freq.freq, np.atleast_2d(np.asarray(data_b)))
def test___eq___notEqual(frequencydata, sine):
"""Check if frequencydata signal is equal."""
actual = FrequencyData(0.5 * frequencydata.freq, frequencydata.frequencies)
assert not frequencydata == actual
actual = FrequencyData(frequencydata.freq, 0.5 * frequencydata.frequencies)
assert not frequencydata == actual
actual = frequencydata.copy()
actual.comment = f'{actual.comment} A completely different thing'
assert not frequencydata == actual
def test_reshape_exceptions():
data_in = FrequencyData(np.random.rand(6, 256), range(256))
data_out = data_in.reshape((3, 2))
npt.assert_allclose(data_in._data.reshape(3, 2, -1), data_out._data)
# test assertion for non-tuple input
with pytest.raises(ValueError):
data_out = data_in.reshape([3, 2])
# test assertion for wrong dimension
with pytest.raises(ValueError, match='Can not reshape signal of cshape'):
data_out = data_in.reshape((3, 4))
def test_separation_from_signal():
"""Check if attributes from Signal are really not available."""
data = [1, 0, -1]
freqs = [0, .1, .3]
freq = FrequencyData(data, freqs)
with pytest.raises(AttributeError):
freq.sampling_rate
with pytest.raises(AttributeError):
freq.domain = 'freq'
with pytest.raises(AttributeError):
freq.fft_norm = 'amplitude'
def test_data_frequency_find_nearest():
"""Test the find nearest function for a single number and list entry."""
data = [1, 0, -1]
freqs = [0, .1, .3]
freq = FrequencyData(data, freqs)
# test for a single number
idx = freq.find_nearest_frequency(.15)
assert idx == 1
# test for a list
idx = freq.find_nearest_frequency([.15, .4])
npt.assert_allclose(idx, np.asarray([1, 2]))
def test_data_frequency_init_wrong_number_of_freqs():
"""Test if entering a wrong number of frequencies raises an assertion."""
data = [1, 0, -1]
freqs = [0, .1]
with pytest.raises(ValueError):
FrequencyData(data, freqs)
def test_data_frequency_with_non_monotonously_increasing_frequencies():
"""Test if non monotnously increasing frequencies raises an assertion."""
data = [1, 0, -1]
freqs = [0, .2, .1]
with pytest.raises(ValueError):
FrequencyData(data, freqs)
def test_overloaded_operators_frequency_data():
x = FrequencyData([2, 1, 0], [0, 1, 2])
y = FrequencyData([2, 2, 2], [0, 1, 2])
# addition
z = x + y
npt.assert_allclose(z.freq, np.array([4, 3, 2], ndmin=2), atol=1e-15)
# subtraction
z = x - y
npt.assert_allclose(z.freq, np.array([0, -1, -2], ndmin=2), atol=1e-15)
# multiplication
z = x * y
npt.assert_allclose(z.freq, np.array([4, 2, 0], ndmin=2), atol=1e-15)
# division
z = x / y
npt.assert_allclose(z.freq, np.array([1, .5, 0], ndmin=2), atol=1e-15)
# power
z = x**y
npt.assert_allclose(z.freq, np.array([4, 1, 0], ndmin=2), atol=1e-15)
def test_add_frequency_data_and_frequency_data():
# generate and add signals
x = FrequencyData([1, 0, 0], [0, .1, .5])
y = signal.add((x, x), 'freq')
# check if old signal did not change
npt.assert_allclose(x.freq, np.atleast_2d([1, 0, 0]), atol=1e-15)
npt.assert_allclose(x.frequencies, np.atleast_1d([0, .1, .5]), atol=1e-15)
# check result
assert isinstance(y, FrequencyData)
npt.assert_allclose(y.freq, np.atleast_2d([2, 0, 0]), atol=1e-15)
npt.assert_allclose(y.frequencies, np.atleast_1d([0, .1, .5]), atol=1e-15)
def test_data_frequency_init_with_defaults():
"""
Test to init without optional parameters.
Test getter for domain, freq, frequencies, and n_bins.
"""
data = [1, 0, -1]
freqs = [0, .1, .3]
freq = FrequencyData(data, freqs)
assert isinstance(freq, FrequencyData)
npt.assert_allclose(freq.freq, np.atleast_2d(np.asarray(data)))
npt.assert_allclose(freq.frequencies, np.atleast_1d(np.asarray(freqs)))
assert freq.n_bins == 3
assert freq.domain == 'freq'
def test_separation_from_time_data():
"""Check if attributes from FrequencyData are really not available."""
data = [1, 0, -1]
freqs = [0, .1, .3]
freq = FrequencyData(data, freqs)
with pytest.raises(AttributeError):
freq.time
with pytest.raises(AttributeError):
freq.times
with pytest.raises(AttributeError):
freq.n_samples
with pytest.raises(AttributeError):
freq.signal_length
with pytest.raises(AttributeError):
freq.find_nearest_time
def test_reshape():
# test reshape with tuple
data_in = FrequencyData(np.random.rand(6, 256), range(256))
data_out = data_in.reshape((3, 2))
npt.assert_allclose(data_in._data.reshape(3, 2, -1), data_out._data)
assert id(data_in) != id(data_out)
data_out = data_in.reshape((3, -1))
npt.assert_allclose(data_in._data.reshape(3, 2, -1), data_out._data)
assert id(data_in) != id(data_out)
# test reshape with int
data_in = FrequencyData(np.random.rand(3, 2, 256), range(256))
data_out = data_in.reshape(6)
npt.assert_allclose(data_in._data.reshape(6, -1), data_out._data)
assert id(data_in) != id(data_out)
def test_flatten():
# test 2D signal (flatten should not change anything)
x = np.random.rand(2, 256)
data_in = FrequencyData(x, range(256))
data_out = data_in.flatten()
npt.assert_allclose(data_in._data, data_out._data)
assert id(data_in) != id(data_out)
# test 3D signal
x = np.random.rand(3, 2, 256)
data_in = FrequencyData(x, range(256))
data_out = data_in.flatten()
npt.assert_allclose(data_in._data.reshape((6, -1)), data_out._data)
assert id(data_in) != id(data_out)
def test_add_frequency_data_and_number_wrong_fft_norm():
# generate and add signals
x = FrequencyData([1, 0, 0], [0, .1, .5])
y = FrequencyData([1, 0, 0], [0, .1, .5], fft_norm='rms')
with raises(ValueError):
signal.add((x, y), 'freq')
def test_getter_fft_norm():
data = [1, 0, -1]
freqs = [0, .1, .3]
freq = FrequencyData(data, freqs, fft_norm='psd')
assert freq.fft_norm == 'psd'
def frequencydata():
data = [1, 0, -1]
freqs = [0, .1, .3]
return FrequencyData(data, freqs)
def test_data_frequency_with_wrong_fft_norm():
data = [1, 0, -1]
freqs = [0, .1, .2]
with pytest.raises(ValueError):
FrequencyData(data, freqs, fft_norm='bull shit')
def test_magic_getitem_slice():
"""Test slicing operations by the magic function __getitem__."""
data = np.array([[1, 0, -1], [2, 0, -2]])
freqs = [0, .1, .3]
freq = FrequencyData(data, freqs)
npt.assert_allclose(FrequencyData(data[0], freqs)._data, freq[0]._data)
def test_add_frequency_data_and_number_wrong_domain():
# generate and add signals
x = FrequencyData([1, 0, 0], [0, .1, .5])
with raises(ValueError):
signal.add((x, 1), 'time')