def scrape(self):
id_counter = 0
track_id, filename, sound_length = get_sound()
# Check if the sound is long enough, and if yes we extract some loops from it.
# TODO: make this less restrictive to waste a bit less
if sound_length > 2 * self.sample_length:
offset = 0
upper_limit = sound_length - 2 * self.sample_length
while (offset + 2 * self.sample_length < upper_limit):
# Calculate a random offset where the loop will start
offset = random.randint(offset, int(min(offset + sound_length * 0.2, upper_limit)))
# Extracting a loop and saving it to 'loop<n>.wav'
sample = Sound.from_file(filename, start=offset, end=offset+self.sample_length)
loops = sample.extract_loops()
for loop in loops:
if id_counter > 10:
offset = upper_limit
break
id_counter += 1
loop_id = '%s_%s' % (track_id, id_counter)
loop_path = self._get_free_path()
logger.info('loop extracted to %s' % loop_path)
loop.to_file(loop_path)
#key, key_confidence, length = loop.echonest.key, loop.echonest.key_confidence, loop.length
with self.pool_lock:
self.pool[loop_id] = dict(loop.loop_infos, **{
'path': loop_path,
'length': loop.length,
'loop_id': loop_id,
'track_id': track_id,
'timbre_start': loop.loop_infos['timbre_start'],
'timbre_end': loop.loop_infos['timbre_end']
#'key': (key, key_confidence)
})
# Increment values for next loop
offset += self.sample_length
# Delete the sounds to save memory
# We also have to collect manually because of a "bug" in pandas:
# https://github.com/pydata/pandas/issues/2659
if 'loop' in locals(): del loop
del sample
gc.collect()
def scrape(self):
track_id, pad_path, pad_length = get_sound()
new_pad_path = self._get_free_path()
pad_id = '%s' % (track_id)
sound = Sound.from_file(pad_path, end=min(30, pad_length))
sound = sound.remove_beats()
sound.to_file(new_pad_path)
logger.info('pad extracted to %s' % new_pad_path)
'''pad.to_file(pad_path)
# Delete the sound to save memory
# We also have to collect manually because of a "bug" in pandas:
# https://github.com/pydata/pandas/issues/2659
del pad
gc.collect()'''
with self.pool_lock:
self.pool[pad_id] = {
'path': new_pad_path,
'pad_id': pad_id
}