def _load_lyrics(lyrics_path):
if not os.path.exists(lyrics_path):
return None
# input: start time (ms), end time (ms), lyric, [pronunciation]
with open(lyrics_path, 'r') as fhandle:
reader = csv.reader(fhandle, delimiter=' ')
start_times = []
end_times = []
lyrics = []
pronunciations = []
for line in reader:
start_times.append(float(line[0]) / 1000.0)
end_times.append(float(line[1]) / 1000.0)
lyrics.append(line[2])
if len(line) > 2:
pronunciation = ' '.join(line[3:])
pronunciations.append(
pronunciation if pronunciation != '' else None)
else:
pronunciations.append(None)
lyrics_data = utils.LyricData(
np.array(start_times),
np.array(end_times),
np.array(lyrics),
np.array(pronunciations),
)
return lyrics_data
def load_annotations_granularity(annotations_path, granularity):
"""Load annotations at the specified level of granularity
Args:
annotations_path (str): path to a DALI annotation file
granularity (str): one of 'notes', 'words', 'lines', 'paragraphs'
Returns:
NoteData for granularity='notes' or LyricData otherwise
"""
if not os.path.exists(annotations_path):
return None
try:
with gzip.open(annotations_path, 'rb') as f:
output = pickle.load(f)
except Exception as e:
with gzip.open(annotations_path, 'r') as f:
output = pickle.load(f)
text = []
notes = []
begs = []
ends = []
for annot in output.annotations['annot'][granularity]:
notes.append(round(annot['freq'][0], 3))
begs.append(round(annot['time'][0], 3))
ends.append(round(annot['time'][1], 3))
text.append(annot['text'])
if granularity == 'notes':
annotation = utils.NoteData(
np.array([begs, ends]).T, np.array(notes), None)
else:
annotation = utils.LyricData(np.array(begs), np.array(ends),
np.array(text), None)
return annotation
def _load_annotations_granularity(annotations_path, granularity):
if not os.path.exists(annotations_path):
return None
try:
with gzip.open(annotations_path, 'rb') as f:
output = pickle.load(f)
except Exception as e:
with gzip.open(annotations_path, 'r') as f:
output = pickle.load(f)
text = []
notes = []
begs = []
ends = []
for annot in output.annotations['annot'][granularity]:
notes.append(round(annot['freq'][0], 3))
begs.append(round(annot['time'][0], 3))
ends.append(round(annot['time'][1], 3))
text.append(annot['text'])
if granularity == 'notes':
annotation = utils.NoteData(np.array(begs), np.array(ends),
np.array(notes), None)
else:
annotation = utils.LyricData(np.array(begs), np.array(ends),
np.array(text), None)
return annotation
def load_lyrics(lyrics_path):
if not os.path.exists(lyrics_path):
raise IOError("lyrics_path {} does not exist".format(lyrics_path))
# input: start time (ms), end time (ms), lyric, [pronunciation]
with open(lyrics_path, "r") as fhandle:
reader = csv.reader(fhandle, delimiter=" ")
start_times = []
end_times = []
lyrics = []
pronunciations = []
for line in reader:
start_times.append(float(line[0]) / 1000.0)
end_times.append(float(line[1]) / 1000.0)
lyrics.append(line[2])
if len(line) > 2:
pronunciation = " ".join(line[3:])
pronunciations.append(pronunciation if pronunciation != "" else None)
else:
pronunciations.append(None)
lyrics_data = utils.LyricData(
np.array(start_times),
np.array(end_times),
np.array(lyrics),
np.array(pronunciations),
)
return lyrics_data
def test_lyrics():
lyrics_data_1 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(["The", "Test"]),
np.array([None, None]),
),
None,
)
]
lyrics_data_2 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(["The", "Test"]),
np.array([None, None]),
),
"lyrics_1",
)
]
lyrics_data_3 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(["The", "Test"]),
np.array([None, None]),
),
"lyrics_1",
),
(
utils.LyricData(
np.array([0.0, 0.232]),
np.array([0.227, 0.742]),
np.array(["is", "cool"]),
np.array([None, None]),
),
"lyrics_2",
),
]
lyrics_data_4 = (
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(["The", "Test"]),
np.array([None, None]),
),
"lyrics_1",
)
lyrics_data_5 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(["The", "Test"]),
np.array([None, None]),
),
"lyrics_1",
),
[
utils.LyricData(
np.array([0.0, 0.232]),
np.array([0.227, 0.742]),
np.array(["is", "cool"]),
np.array([None, None]),
),
"lyrics_2",
],
]
lyrics_data_6 = [(None, None)]
lyrics_data_7 = [
(
utils.EventData(
np.array([0.2, 0.3]),
np.array([0.3, 0.4]),
np.array(["event A", "event B"]),
),
None,
)
]
jam_1 = jams_utils.jams_converter(lyrics_data=lyrics_data_1)
jam_2 = jams_utils.jams_converter(lyrics_data=lyrics_data_2)
jam_3 = jams_utils.jams_converter(lyrics_data=lyrics_data_3)
jam_6 = jams_utils.jams_converter(lyrics_data=lyrics_data_6)
time, duration, value, confidence = get_jam_data(jam_1, "lyrics", 0)
assert time == [0.027, 0.232]
assert duration == [0.2, 0.51]
assert value == ["The", "Test"]
assert confidence == [None, None]
assert jam_2.annotations[0]["sandbox"]["name"] == "lyrics_1"
time, duration, value, confidence = get_jam_data(jam_3, "lyrics", 0)
assert time == [0.027, 0.232]
assert duration == [0.2, 0.51]
assert value == ["The", "Test"]
assert confidence == [None, None]
time, duration, value, confidence = get_jam_data(jam_3, "lyrics", 1)
assert time == [0.0, 0.232]
assert duration == [0.227, 0.51]
assert value == ["is", "cool"]
assert confidence == [None, None]
time, duration, value, confidence = get_jam_data(jam_6, "lyrics", 0)
assert time == []
assert duration == []
assert value == []
assert confidence == []
assert type(jam_1) == jams.JAMS
with pytest.raises(TypeError):
jams_utils.jams_converter(lyrics_data=lyrics_data_4)
with pytest.raises(TypeError):
jams_utils.jams_converter(lyrics_data=lyrics_data_5)
with pytest.raises(TypeError):
jams_utils.jams_converter(lyrics_data=lyrics_data_7)
def test_lyrics():
lyrics_data_1 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(['The', 'Test']),
np.array([None, None]),
),
None,
)
]
lyrics_data_2 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(['The', 'Test']),
np.array([None, None]),
),
'lyrics_1',
)
]
lyrics_data_3 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(['The', 'Test']),
np.array([None, None]),
),
'lyrics_1',
),
(
utils.LyricData(
np.array([0.0, 0.232]),
np.array([0.227, 0.742]),
np.array(['is', 'cool']),
np.array([None, None]),
),
'lyrics_2',
),
]
lyrics_data_4 = (
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(['The', 'Test']),
np.array([None, None]),
),
'lyrics_1',
)
lyrics_data_5 = [
(
utils.LyricData(
np.array([0.027, 0.232]),
np.array([0.227, 0.742]),
np.array(['The', 'Test']),
np.array([None, None]),
),
'lyrics_1',
),
[
utils.LyricData(
np.array([0.0, 0.232]),
np.array([0.227, 0.742]),
np.array(['is', 'cool']),
np.array([None, None]),
),
'lyrics_2',
],
]
lyrics_data_6 = [(None, None)]
lyrics_data_7 = [
(
utils.EventData(
np.array([0.2, 0.3]),
np.array([0.3, 0.4]),
np.array(['event A', 'event B']),
),
None,
)
]
jam_1 = jams_utils.jams_converter(lyrics_data=lyrics_data_1)
jam_2 = jams_utils.jams_converter(lyrics_data=lyrics_data_2)
jam_3 = jams_utils.jams_converter(lyrics_data=lyrics_data_3)
jam_6 = jams_utils.jams_converter(lyrics_data=lyrics_data_6)
time, duration, value, confidence = get_jam_data(jam_1, 'lyrics', 0)
assert time == [0.027, 0.232]
assert duration == [0.2, 0.51]
assert value == ['The', 'Test']
assert confidence == [None, None]
assert jam_2.annotations[0]['sandbox']['name'] == 'lyrics_1'
time, duration, value, confidence = get_jam_data(jam_3, 'lyrics', 0)
assert time == [0.027, 0.232]
assert duration == [0.2, 0.51]
assert value == ['The', 'Test']
assert confidence == [None, None]
time, duration, value, confidence = get_jam_data(jam_3, 'lyrics', 1)
assert time == [0.0, 0.232]
assert duration == [0.227, 0.51]
assert value == ['is', 'cool']
assert confidence == [None, None]
time, duration, value, confidence = get_jam_data(jam_6, 'lyrics', 0)
assert time == []
assert duration == []
assert value == []
assert confidence == []
assert type(jam_1) == jams.JAMS
with pytest.raises(TypeError):
jams_utils.jams_converter(lyrics_data=lyrics_data_4)
with pytest.raises(TypeError):
jams_utils.jams_converter(lyrics_data=lyrics_data_5)
with pytest.raises(TypeError):
jams_utils.jams_converter(lyrics_data=lyrics_data_7)