def test_load_melody():
dataset = cante100.Dataset(TEST_DATA_HOME)
track = dataset.track("008")
f0_path = track.f0_path
f0_data = cante100.load_melody(f0_path)
# check types
assert type(f0_data) == annotations.F0Data
assert type(f0_data.times) is np.ndarray
assert type(f0_data.frequencies) is np.ndarray
assert type(f0_data.confidence) is np.ndarray
# check values
assert np.array_equal(
f0_data.times,
np.array([
0.023219954,
0.026122448,
0.029024942,
0.031927436,
0.034829931,
0.037732425,
]),
)
assert np.array_equal(
f0_data.frequencies,
np.array([0.0, 137.0, 220.34, 400.0, -110.0, -110.0]))
assert np.array_equal(f0_data.confidence,
np.array([0.0, 1.0, 1.0, 1.0, 0.0, 0.0]))
def test_load_audio():
dataset = cante100.Dataset(TEST_DATA_HOME)
track = dataset.track("008")
audio_path = track.audio_path
audio, sr = cante100.load_audio(audio_path)
assert sr == 22050
assert audio.shape[0] == 2 # Check audio is stereo
# assert audio.shape[1] == 3957696 # Check audio length
assert type(audio) is np.ndarray
def test_load_spectrum():
dataset = cante100.Dataset(TEST_DATA_HOME)
track = dataset.track("008")
spectrogram_path = track.spectrogram_path
spectrogram = cante100.load_spectrogram(spectrogram_path)
assert spectrogram.shape[0] == 5
assert spectrogram.shape[1] == 514
assert type(spectrogram) is np.ndarray
assert isinstance(spectrogram[0][0], float) is True
def test_metadata():
data_home = "tests/resources/mir_datasets/cante100"
dataset = cante100.Dataset(data_home)
metadata = dataset._metadata
assert metadata["008"] == {
"musicBrainzID": "4eebe839-82bb-426e-914d-7c4525dd9dad",
"artist": "Toronjo",
"title": "Huelva Como Capital",
"release": "Atlas del cante flamenco",
"duration": 179,
}
def test_load_notes():
dataset = cante100.Dataset(TEST_DATA_HOME)
track = dataset.track("008")
notes_path = track.notes_path
notes_data = cante100.load_notes(notes_path)
# check types
assert type(notes_data) == annotations.NoteData
assert type(notes_data.intervals) is np.ndarray
assert type(notes_data.notes) is np.ndarray
assert type(notes_data.confidence) is np.ndarray
# check values
assert np.array_equal(
notes_data.intervals[:, 0],
np.array([25.7625, 26.1457, 37.3319, 37.5612, 37.7876, 44.8755]),
)
assert np.array_equal(
notes_data.intervals[:, 1],
np.array([
26.107896999999998,
26.540439000000003,
37.555392,
37.764375,
40.18796,
45.162847,
]),
)
assert np.array_equal(
notes_data.notes,
np.array([
207.65234878997256,
207.65234878997256,
311.1269837220809,
369.9944227116344,
415.3046975799452,
391.9954359817492,
]),
)
assert np.array_equal(notes_data.confidence,
np.array([1.0, 1.0, 1.0, 1.0, 1.0, 1.0]))
def test_track():
default_trackid = "008"
dataset = cante100.Dataset(TEST_DATA_HOME)
track = dataset.track(default_trackid)
expected_attributes = {
"artist":
"Toronjo",
"duration":
179.0,
"audio_path":
"tests/resources/mir_datasets/cante100/cante100audio/008_PacoToronjo_"
+ "Fandangos.mp3",
"f0_path":
"tests/resources/mir_datasets/cante100/cante100midi_f0/008_PacoToronjo_"
+ "Fandangos.f0.csv",
"identifier":
"4eebe839-82bb-426e-914d-7c4525dd9dad",
"notes_path":
"tests/resources/mir_datasets/cante100/cante100_automaticTranscription/008_PacoToronjo_"
+ "Fandangos.notes.csv",
"release":
"Atlas del cante flamenco",
"spectrogram_path":
"tests/resources/mir_datasets/cante100/cante100_spectrum/008_PacoToronjo_"
+ "Fandangos.spectrum.csv",
"title":
"Huelva Como Capital",
"track_id":
"008",
}
expected_property_types = {
"melody": annotations.F0Data,
"notes": annotations.NoteData,
"audio": tuple,
"spectrogram": np.ndarray,
}
run_track_tests(track, expected_attributes, expected_property_types)
def test_to_jams():
default_trackid = "008"
dataset = cante100.Dataset(TEST_DATA_HOME)
track = dataset.track(default_trackid)
jam = track.to_jams()
# Validate cante100 jam schema
assert jam.validate()
# Validate melody
melody = jam.search(namespace="pitch_contour")[0]["data"]
assert [note.time for note in melody] == [
0.023219954,
0.026122448,
0.029024942,
0.031927436,
0.034829931,
0.037732425,
]
assert [note.duration for note in melody] == [0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
assert [note.value for note in melody] == [
{
"index": 0,
"frequency": 0.0,
"voiced": False
},
{
"index": 0,
"frequency": 137.0,
"voiced": True
},
{
"index": 0,
"frequency": 220.34,
"voiced": True
},
{
"index": 0,
"frequency": 400.0,
"voiced": True
},
{
"index": 0,
"frequency": -110.0,
"voiced": False
},
{
"index": 0,
"frequency": -110.0,
"voiced": False
},
]
assert [note.confidence
for note in melody] == [0.0, 1.0, 1.0, 1.0, 0.0, 0.0]
# Validate note transciption
notes = jam.search(namespace="note_hz")[0]["data"]
assert [note.time for note in notes] == [
25.7625,
26.1457,
37.3319,
37.5612,
37.7876,
44.8755,
]
assert [note.duration for note in notes] == [
0.3453969999999984,
0.3947390000000013,
0.22349200000000025,
0.20317500000000166,
2.400359999999999,
0.2873469999999969,
]
assert [note.value for note in notes] == [
207.65234878997256,
207.65234878997256,
311.1269837220809,
369.9944227116344,
415.3046975799452,
391.9954359817492,
]
assert [note.confidence
for note in notes] == [None, None, None, None, None, None]