def encode_as_linear16(path, tmp):
# From: https://cloud.google.com/speech/support#troubleshooting:
# "The LINEAR16 encoding must be 16-bits, signed-integer,
# little-endian."
# In avconv, this translates to "s16le". See also:
# http://stackoverflow.com/a/4854627/64911 and
# https://trac.ffmpeg.org/wiki/audio%20types
assert isinstance(path, basestring), "path argument is not a str."
av_path = get_audio_binary()
# fmt: off
av_command = [
av_path,
'-y', # Assume yes (clobber existing files)
'-i',
path, # Input file
'-f',
's16le', # Force output format
'-ac',
'1', # Mono
'-ar',
'16k', # Sample rate of 16000Mhz
tmp.name, # Output file
]
# fmt: on
try:
_ = subprocess.check_output(av_command, stderr=subprocess.STDOUT)
except subprocess.CalledProcessError as e:
print("%s failed command: %s\n"
"error code: %s\n"
"output: %s\n" % (av_path, av_command, e.returncode, e.output))
print(traceback.format_exc())
raise e
def process_audio_file(pk):
"""Given the key to an audio file, extract its content and add the related
meta data to the database.
"""
af = Audio.objects.get(pk=pk)
tmp_path = os.path.join("/tmp", "audio_" + uuid.uuid4().hex + ".mp3")
av_path = get_audio_binary()
av_command = [
av_path,
"-i",
af.local_path_original_file.path,
"-ar",
"22050", # sample rate (audio samples/s) of 22050Hz
"-ab",
"48k", # constant bit rate (sample resolution) of 48kbps
tmp_path,
]
try:
_ = subprocess.check_output(av_command, stderr=subprocess.STDOUT)
except subprocess.CalledProcessError as e:
print(
"%s failed command: %s\nerror code: %s\noutput: %s\n%s"
% (
av_path,
av_command,
e.returncode,
e.output,
traceback.format_exc(),
)
)
raise
set_mp3_meta_data(af, tmp_path)
try:
with open(tmp_path, "r") as mp3:
cf = ContentFile(mp3.read())
file_name = trunc(best_case_name(af).lower(), 72) + "_cl.mp3"
af.file_with_date = af.docket.date_argued
af.local_path_mp3.save(file_name, cf, save=False)
except:
msg = (
"Unable to save mp3 to audio_file in scraper.tasks."
"process_audio_file for item: %s\n"
"Traceback:\n"
"%s" % (af.pk, traceback.format_exc())
)
print(msg)
ErrorLog.objects.create(
log_level="CRITICAL", court=af.docket.court, message=msg
)
af.duration = eyed3.load(tmp_path).info.time_secs
af.processing_complete = True
af.save()
def process_audio_file(pk):
"""Given the key to an audio file, extract its content and add the related
meta data to the database.
"""
af = Audio.objects.get(pk=pk)
tmp_path = os.path.join('/tmp', 'audio_' + uuid.uuid4().hex + '.mp3')
av_path = get_audio_binary()
av_command = [
av_path, '-i', af.local_path_original_file.path,
'-ar', '22050', # sample rate (audio samples/s) of 22050Hz
'-ab', '48k', # constant bit rate (sample resolution) of 48kbps
tmp_path
]
try:
_ = subprocess.check_output(av_command, stderr=subprocess.STDOUT)
except subprocess.CalledProcessError as e:
print('%s failed command: %s\nerror code: %s\noutput: %s\n%s' %
(av_path, av_command, e.returncode, e.output,
traceback.format_exc()))
raise
set_mp3_meta_data(af, tmp_path)
try:
with open(tmp_path, 'r') as mp3:
cf = ContentFile(mp3.read())
file_name = trunc(best_case_name(af).lower(), 72) + '_cl.mp3'
af.file_with_date = af.docket.date_argued
af.local_path_mp3.save(file_name, cf, save=False)
except:
msg = ("Unable to save mp3 to audio_file in scraper.tasks."
"process_audio_file for item: %s\n"
"Traceback:\n"
"%s" % (af.pk, traceback.format_exc()))
print(msg)
ErrorLog.objects.create(log_level='CRITICAL', court=af.docket.court,
message=msg)
af.duration = eyed3.load(tmp_path).info.time_secs
af.processing_complete = True
af.save()