def test_can_invert_frequency_weighting(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.random.random_sample((90, 100)), [td, fd])
weighted = tf * AWeighting()
inverted = weighted / AWeighting()
np.testing.assert_allclose(tf, inverted)
def test_from_example(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((30, 100)), [td, fd])
from_example = ArrayWithUnits.from_example(np.ones((30, 100)), tf)
self.assertEqual(tf.shape, from_example.shape)
self.assertSequenceEqual(tf.dimensions, from_example.dimensions)
def test_can_pool_non_stored_time_slice_feature(self):
@simple_in_memory_settings
class WithOnsets(self.STFT):
transience = ArrayWithUnitsFeature(MeasureOfTransience,
needs=self.STFT.fft,
store=True)
sliding_detection = ArrayWithUnitsFeature(
SlidingWindow,
needs=transience,
wscheme=self.wscheme * Stride(frequency=1, duration=10),
padwith=5,
store=False)
slices = TimeSliceFeature(MovingAveragePeakPicker,
needs=sliding_detection,
aggregate=np.median,
store=True)
signal = self.ticks(self.samplerate, Seconds(4), Seconds(1))
_id = WithOnsets.process(meta=signal)
class WithPooled(WithOnsets):
pooled = VariableRateTimeSeriesFeature(Pooled,
needs=(self.STFT.fft,
WithOnsets.slices),
op=np.max,
axis=0,
store=True)
doc = WithPooled(_id)
self.assertEqual((4, 1025), doc.pooled.slicedata.shape)
def test_inversion_returns_time_series(self):
data = np.random.random_sample((33, 30))
ts = ArrayWithUnits(data,
[TimeDimension(Seconds(1)),
IdentityDimension()])
inverted = self.invert_and_assert_class(ts)
self.assertEqual(Seconds(1), inverted.dimensions[0].frequency)
def test_can_apply_sliding_window_to_time_frequency_representation(self):
band = FrequencyBand(0, 22000)
scale = LinearScale(band, 100)
arr = ArrayWithUnits(np.zeros(
(200,
100)), [TimeDimension(Seconds(1)),
FrequencyDimension(scale)])
sw = SampleRate(Seconds(2), Seconds(2))
@simple_in_memory_settings
class Document(BaseModel):
windowed = ArrayWithUnitsFeature(SlidingWindow,
wscheme=sw,
store=True)
_id = Document.process(windowed=arr)
result = Document(_id).windowed
self.assertIsInstance(result, ArrayWithUnits)
self.assertEqual((100, 2, 100), result.shape)
self.assertEqual(3, len(result.dimensions))
self.assertIsInstance(result.dimensions[0], TimeDimension)
self.assertEqual(Seconds(2), result.dimensions[0].frequency)
self.assertIsInstance(result.dimensions[1], TimeDimension)
self.assertEqual(Seconds(1), result.dimensions[1].frequency)
self.assertIsInstance(result.dimensions[2], FrequencyDimension)
def test_can_create_instance_with_no_slice_data(self):
ts = VariableRateTimeSeries((
(TimeSlice(start=Seconds(1), duration=Seconds(1)), np.zeros(0)),
(TimeSlice(start=Seconds(2), duration=Seconds(1)), np.zeros(0)),
))
self.assertEqual(2, len(ts))
self.assertEqual((2, 0), ts.slicedata.shape)
def test_can_round_trip_3d_constant_rate_time_series_with_frequency_dim(
self):
dim1 = TimeDimension(Seconds(2), Milliseconds(1000))
dim2 = TimeDimension(Seconds(1), Milliseconds(500))
scale = LinearScale(FrequencyBand(20, 20000), 100)
dim3 = FrequencyDimension(scale)
raw = np.random.random_sample((5, 2, 100))
ts = ArrayWithUnits(raw, (dim1, dim2, dim3))
decoded = self._roundtrip(ts)
self.assertIsInstance(decoded, ArrayWithUnits)
self.assertEqual(3, len(decoded.dimensions))
td1 = decoded.dimensions[0]
self.assertIsInstance(td1, TimeDimension)
self.assertEqual(Seconds(2), td1.frequency)
self.assertEqual(Milliseconds(1000), td1.duration)
td2 = decoded.dimensions[1]
self.assertIsInstance(td2, TimeDimension)
self.assertEqual(Seconds(1), td2.frequency)
self.assertEqual(Milliseconds(500), td2.duration)
fd = decoded.dimensions[2]
self.assertIsInstance(fd, FrequencyDimension)
self.assertEqual(scale, fd.scale)
np.testing.assert_allclose(decoded, raw)
def test_generates_correct_number_of_samples(self):
samplerate = SR22050()
duration = Seconds(1)
ss = SineSynthesizer(samplerate)
audio = ss.synthesize(duration, freqs_in_hz=[440.])
expected_samples = int(samplerate) * int(duration / Seconds(1))
self.assertEqual(expected_samples, len(audio))
def test_can_apply_a_weighting_to_time_frequency_representation(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((90, 100)), [td, fd])
weighting = AWeighting()
result = tf * weighting
self.assertGreater(result[0, -1], result[0, 0])
def test_can_get_weights_from_tf_representation(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((90, 100)), [td, fd])
weighting = AWeighting()
weights = weighting.weights(tf)
self.assertEqual((100,), weights.shape)
def test_can_round_trip(self):
td = TimeDimension(Seconds(1), Milliseconds(500), 100)
encoded = self.encoder.encode(td)
decoded = self.decoder.decode(encoded)
self.assertIsInstance(decoded, TimeDimension)
self.assertEqual(Seconds(1), decoded.frequency)
self.assertEqual(Milliseconds(500), decoded.duration)
self.assertEqual(100, decoded.size)
def test_can_construct_instance(self):
frequency = Seconds(1)
duration = Seconds(1)
scale = LinearScale(FrequencyBand(20, 22050), 100)
td = TimeDimension(frequency, duration)
fd = FrequencyDimension(scale)
tf = ArrayWithUnits(np.zeros((30, 100)), [td, fd])
self.assertIsInstance(tf, ArrayWithUnits)
def test_sum_along_frequency_axis(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((30, 100)), [td, fd])
result = tf.sum(axis=1)
self.assertIsInstance(result, ArrayWithUnits)
self.assertEqual(1, len(result.dimensions))
self.assertEqual((30, ), result.shape)
self.assertIsInstance(result.dimensions[0], TimeDimension)
def test_can_use_negative_axis_indices_max(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((30, 100)), [td, fd])
result = tf.max(axis=-1)
self.assertIsInstance(result, ArrayWithUnits)
self.assertEqual(1, len(result.dimensions))
self.assertEqual((30, ), result.shape)
self.assertIsInstance(result.dimensions[0], TimeDimension)
def test_can_apply_empty_time_slice_to_wrapper(self):
synth = SineSynthesizer(SR11025())
samples = synth.synthesize(Seconds(10))
encoded = samples.encode(fmt='OGG', subtype='VORBIS')
wrapper = OggVorbisWrapper(encoded)
samples = wrapper[TimeSlice()]
expected = Seconds(10) / Seconds(1)
actual = samples.end / Seconds(1)
self.assertAlmostEqual(expected, actual, places=6)
def test_can_slice_frequency_dimension_with_integer_indices(self):
frequency = Seconds(1)
duration = Seconds(1)
scale = LinearScale(FrequencyBand(20, 22050), 100)
td = TimeDimension(frequency, duration)
fd = FrequencyDimension(scale)
tf = ArrayWithUnits(np.zeros((30, 100)), [td, fd])
sliced = tf[:, 10:20]
self.assertEqual((30, 10), sliced.shape)
self.assertIsInstance(sliced, ArrayWithUnits)
def test_raises_if_scale_length_does_not_match_frequency_dimension(self):
frequency = Seconds(1)
duration = Seconds(1)
scale = LinearScale(FrequencyBand(20, 22050), 1000)
td = TimeDimension(frequency, duration)
fd = FrequencyDimension(scale)
self.assertRaises(ValueError,
lambda: ArrayWithUnits(np.ones((30, 100)), [td, fd]))
def test_can_access_int_index_and_frequency_band(self):
tf = ArrayWithUnits(np.ones((10, 10)),
dimensions=[
TimeDimension(Seconds(1), Seconds(1)),
FrequencyDimension(
LinearScale(FrequencyBand(0, 1000), 10))
])
sliced = tf[0, FrequencyBand(201, 400)]
self.assertEqual((2, ), sliced.shape)
self.assertIsInstance(sliced.dimensions[0], FrequencyDimension)
def test_can_access_time_slice_and_int_index(self):
tf = ArrayWithUnits(np.ones((10, 10)),
dimensions=[
TimeDimension(Seconds(1), Seconds(1)),
FrequencyDimension(
LinearScale(FrequencyBand(0, 1000), 10))
])
sliced = tf[TimeSlice(start=Seconds(1), duration=Seconds(2)), 0]
self.assertEqual((2, ), sliced.shape)
self.assertIsInstance(sliced.dimensions[0], TimeDimension)
def test_can_use_keepdims_with_sum(self):
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((30, 100)), [td, fd])
result = tf.sum(axis=-1, keepdims=True)
self.assertIsInstance(result, ArrayWithUnits)
self.assertEqual(2, len(result.dimensions))
self.assertEqual((30, 1), result.shape)
self.assertIsInstance(result.dimensions[0], TimeDimension)
self.assertIsInstance(result.dimensions[1], IdentityDimension)
def test_can_multiply_by_array(self):
frequency = Seconds(1)
duration = Seconds(1)
scale = LinearScale(FrequencyBand(20, 22050), 100)
td = TimeDimension(frequency, duration)
fd = FrequencyDimension(scale)
tf = ArrayWithUnits(np.ones((30, 100)), [td, fd])
result = tf * np.ones(100)
self.assertIsInstance(result, ArrayWithUnits)
np.testing.assert_allclose(tf, result)
def test_can_add_axis_at_end(self):
_id = IdentityDimension()
td = TimeDimension(Seconds(1), Seconds(1))
fd = FrequencyDimension(LinearScale(FrequencyBand(20, 22050), 100))
tf = ArrayWithUnits(np.ones((3, 30, 100)), [_id, td, fd])
tf2 = tf[..., None]
self.assertEqual(4, tf2.ndim)
self.assertIsInstance(tf2.dimensions[0], IdentityDimension)
self.assertIsInstance(tf2.dimensions[1], TimeDimension)
self.assertIsInstance(tf2.dimensions[2], FrequencyDimension)
self.assertIsInstance(tf2.dimensions[3], IdentityDimension)
def test_inversion_returns_time_frequency_representation(self):
data = np.random.random_sample((33, 30))
scale = LinearScale(FrequencyBand(20, 20000), 30)
tf = ArrayWithUnits(
data,
[TimeDimension(Seconds(1), Seconds(2)),
FrequencyDimension(scale)])
inverted = self.invert_and_assert_class(tf)
self.assertEqual(Seconds(1), inverted.dimensions[0].frequency)
self.assertEqual(Seconds(2), inverted.dimensions[0].duration)
self.assertEqual(scale, inverted.dimensions[1].scale)
def test_can_slice_freq_dimension_with_freq_band_spanning_bins(self):
frequency = Seconds(1)
duration = Seconds(1)
scale = LinearScale(FrequencyBand(20, 22050), 100)
td = TimeDimension(frequency, duration)
fd = FrequencyDimension(scale)
tf = ArrayWithUnits(np.zeros((30, 100)), [td, fd])
bands = list(scale)
wide_band = FrequencyBand(bands[0].start_hz, bands[9].stop_hz)
sliced = tf[:, wide_band]
self.assertEqual((30, 10), sliced.shape)
self.assertIsInstance(sliced, ArrayWithUnits)
def test_can_round_trip_1d_constant_rate_time_series(self):
dim = TimeDimension(Seconds(1), Milliseconds(500))
raw = np.arange(10)
ts = ArrayWithUnits(raw, (dim, ))
decoded = self._roundtrip(ts)
self.assertIsInstance(decoded, ArrayWithUnits)
self.assertEqual(1, len(decoded.dimensions))
self.assertIsInstance(decoded.dimensions[0], TimeDimension)
td = decoded.dimensions[0]
self.assertEqual(Seconds(1), td.frequency)
self.assertEqual(Milliseconds(500), td.duration)
np.testing.assert_allclose(decoded, raw)
def test_can_iterate_over_timeseries(self):
tf = ArrayWithUnits(np.ones((10, 10)),
dimensions=[
TimeDimension(Seconds(1), Seconds(1)),
IdentityDimension()
])
rows = [row for row in tf]
self.assertEqual(10, len(rows))
for row in rows:
self.assertIsInstance(row, ArrayWithUnits)
self.assertIsInstance(row.dimensions[0], IdentityDimension)
self.assertEqual((10, ), row.shape)
def test_can_use_tuple_indices_for_first_dimension(self):
tf = ArrayWithUnits(np.ones((10, 10)),
dimensions=[
TimeDimension(Seconds(1), Seconds(1)),
FrequencyDimension(
LinearScale(FrequencyBand(0, 1000), 10))
])
subset = tf[tuple([2, 4, 6]), ...]
self.assertEqual((3, 10), subset.shape)
self.assertIsInstance(subset, ArrayWithUnits)
self.assertIsInstance(subset.dimensions[0], TimeDimension)
self.assertIsInstance(subset.dimensions[1], FrequencyDimension)
def test_can_iterate_over_time_frequency_representation(self):
tf = ArrayWithUnits(np.ones((10, 10)),
dimensions=[
TimeDimension(Seconds(1), Seconds(1)),
FrequencyDimension(
LinearScale(FrequencyBand(0, 1000), 10))
])
rows = [row for row in tf]
self.assertEqual(10, len(rows))
for row in rows:
self.assertIsInstance(row, ArrayWithUnits)
self.assertIsInstance(row.dimensions[0], FrequencyDimension)
self.assertEqual((10, ), row.shape)
def test_span(self):
ts = VariableRateTimeSeries(
((TimeSlice(start=Seconds(0), duration=Seconds(1)), np.zeros(10)),
(TimeSlice(start=Seconds(1), duration=Seconds(2)), np.zeros(10)),
(TimeSlice(start=Seconds(3), duration=Seconds(1)), np.zeros(10))))
self.assertEqual(TimeSlice(start=Seconds(0), duration=Seconds(4)),
ts.span)
def test_scale_is_modified_after_slice(self):
frequency = Seconds(1)
duration = Seconds(1)
scale = LinearScale(FrequencyBand(20, 22050), 100)
td = TimeDimension(frequency, duration)
fd = FrequencyDimension(scale)
tf = ArrayWithUnits(np.zeros((30, 100)), [td, fd])
bands = list(scale)
wide_band = FrequencyBand(bands[0].start_hz, bands[9].stop_hz)
sliced = tf[:, wide_band]
self.assertEqual((30, 10), sliced.shape)
self.assertIsInstance(sliced, ArrayWithUnits)
self.assertLess(sliced.dimensions[1].scale.stop_hz, scale.stop_hz)
self.assertEqual(10, sliced.dimensions[1].scale.n_bands)
