ref: 8b56b1835e9b8f4858547b88c82e2cf2c977ea9f
parent: f6892d491c8a7651ec19c0bc23a4fb03f40e2098
author: Paul Brossier <piem@piem.org>
date: Tue May 10 18:42:55 EDT 2016
python/lib/aubio/slicing.py: clean up
--- a/python/lib/aubio/slicing.py
+++ b/python/lib/aubio/slicing.py
@@ -1,15 +1,16 @@
-from aubio import source, sink
+"""utility routines to slice sound files at given timestamps"""
+
import os
+from aubio import source, sink
-max_timestamp = 1e120
+_max_timestamp = 1e120
-def slice_source_at_stamps(source_file, timestamps, timestamps_end = None,
- output_dir = None,
- samplerate = 0,
- hopsize = 256):
+def slice_source_at_stamps(source_file, timestamps, timestamps_end=None,
+ output_dir=None, samplerate=0, hopsize=256):
+ """ slice a sound file at given timestamps """
- if timestamps == None or len(timestamps) == 0:
- raise ValueError ("no timestamps given")+ if timestamps is None or len(timestamps) == 0:
+ raise ValueError("no timestamps given")if timestamps[0] != 0:
timestamps = [0] + timestamps
@@ -18,14 +19,14 @@
if timestamps_end != None:
if len(timestamps_end) != len(timestamps):
- raise ValueError ("len(timestamps_end) != len(timestamps)")+ raise ValueError("len(timestamps_end) != len(timestamps)")else:
- timestamps_end = [t - 1 for t in timestamps[1:] ] + [ max_timestamp ]
+ timestamps_end = [t - 1 for t in timestamps[1:]] + [_max_timestamp]
regions = list(zip(timestamps, timestamps_end))
#print regions
- source_base_name, source_ext = os.path.splitext(os.path.basename(source_file))
+ source_base_name, _ = os.path.splitext(os.path.basename(source_file))
if output_dir != None:
if not os.path.isdir(output_dir):
os.makedirs(output_dir)
@@ -32,12 +33,13 @@
source_base_name = os.path.join(output_dir, source_base_name)
def new_sink_name(source_base_name, timestamp, samplerate):
+ """ create a sink based on a timestamp in samples, converted in seconds """
timestamp_seconds = timestamp / float(samplerate)
return source_base_name + "_%011.6f" % timestamp_seconds + '.wav'
- # reopen source file
- s = source(source_file, samplerate, hopsize)
- samplerate = s.get_samplerate()
+ # open source file
+ _source = source(source_file, samplerate, hopsize)
+ samplerate = source.get_samplerate()
total_frames = 0
slices = []
@@ -44,7 +46,7 @@
while True:
# get hopsize new samples from source
- vec, read = s.do_multi()
+ vec, read = _source.do_multi()
# if the total number of frames read will exceed the next region start
if len(regions) and total_frames + read >= regions[0][0]:
#print "getting", regions[0], "at", total_frames
@@ -53,9 +55,9 @@
# create a name for the sink
new_sink_path = new_sink_name(source_base_name, start_stamp, samplerate)
# create its sink
- g = sink(new_sink_path, samplerate, s.channels)
+ _sink = sink(new_sink_path, samplerate, _source.channels)
# create a dictionary containing all this
- new_slice = {'start_stamp': start_stamp, 'end_stamp': end_stamp, 'sink': g}+ new_slice = {'start_stamp': start_stamp, 'end_stamp': end_stamp, 'sink': _sink}# append the dictionary to the current list of slices
slices.append(new_slice)
@@ -62,7 +64,7 @@
for current_slice in slices:
start_stamp = current_slice['start_stamp']
end_stamp = current_slice['end_stamp']
- g = current_slice['sink']
+ _sink = current_slice['sink']
# sample index to start writing from new source vector
start = max(start_stamp - total_frames, 0)
# number of samples yet to written be until end of region
@@ -72,12 +74,13 @@
if remaining < read:
if remaining > start:
# write remaining samples from current region
- g.do_multi(vec[:,start:remaining], remaining - start)
+ _sink.do_multi(vec[:, start:remaining], remaining - start)
#print "closing region", "remaining", remaining
# close this file
- g.close()
+ _sink.close()
elif read > start:
# write all the samples
- g.do_multi(vec[:,start:read], read - start)
+ _sink.do_multi(vec[:, start:read], read - start)
total_frames += read
- if read < hopsize: break
+ if read < hopsize:
+ break