def test_mirex_with_real_data(self):
features = read_features('data/drumtrack.mp3')
tempo_classifier = TempoClassifier('fcn')
t1, t2, s1 = tempo_classifier.estimate_mirex(features)
self.assertAlmostEquals(s1, 0.99617153)
self.assertAlmostEquals(t1, 100.)
self.assertAlmostEquals(t2, 101.)
def test_mirex_tempo_sanity(self):
tempo_classifier = TempoClassifier('fcn')
t1, t2, s1 = tempo_classifier.estimate_mirex(self.get_test_data())
tempo = tempo_classifier.estimate_tempo(self.get_test_data())
if s1 > 0.5:
self.assertEqual(tempo, t1)
else:
self.assertEqual(tempo, t2)
def test_predict(self):
tempo_classifier = TempoClassifier('fcn')
predictions = tempo_classifier.estimate(self.get_test_data())
self.assertEqual(predictions.shape, (2, 256))
np.testing.assert_array_almost_equal(np.ones(2), np.sum(predictions, axis=1))
tempi = np.argmax(predictions, axis=1) + 30
self.assertEqual(tempi[0], 163)
self.assertEqual(tempi[1], 43)
def calculateTempo(self):
model_name = 'cnn'
classifier = TempoClassifier(model_name)
features = read_features(self.details["song_file"])
tempo = classifier.estimate_tempo(features, interpolate=False)
self.details["bpm"] = tempo
def test_init(self):
tempo_classifier = TempoClassifier('fcn')
self.assertIsNotNone(tempo_classifier.model)
def test_tempo_with_real_data(self):
features = read_features('data/drumtrack.mp3')
tempo_classifier = TempoClassifier('fcn')
tempo = tempo_classifier.estimate_tempo(features)
self.assertAlmostEqual(100., tempo)
def test_predict_mirex(self):
tempo_classifier = TempoClassifier('fcn')
t1, t2, s1 = tempo_classifier.estimate_mirex(self.get_test_data())
self.assertAlmostEquals(s1, 0.7373429)
self.assertAlmostEquals(t1, 43.)
self.assertAlmostEquals(t2, 86.)
def test_predict_tempo(self):
tempo_classifier = TempoClassifier('fcn')
tempo = tempo_classifier.estimate_tempo(self.get_test_data())
self.assertAlmostEqual(43., tempo)
def test_bad_model_name(self):
try:
TempoClassifier('bad_model_name')
self.fail('Expected FileNotFoundError')
except FileNotFoundError:
pass
