def supports_file(self, fname: str) -> bool:
# Readable by audiotools?
try:
audiotools.open(fname)
return True
except UnsupportedFile:
return False
def check_file(self):
try:
self.verify()
except (audiotools.InvalidFile, audiotools.UnsupportedFile, ValueError):
# We failed to open it in audiotools, so the streamer
# won't be able to play it; try re-encoding for good measure.
try:
filename = self.encode()
except VerifyError:
error("failed to encode the song")
except:
error("unknown error on {0}".format(self.filename))
fileobj = audiotools.open(filename)
length = long(fileobj.seconds_length())
# If there was no error we now have two files, delete the old one
os.remove(args.file)
else:
filename = self.new_path(self.uuid())
self.rename(self.filename, filename)
fileobj = audiotools.open(filename)
length = long(fileobj.seconds_length())
print json.dumps({"length": length, "success": True, "filename": filename})
sys.exit(0)
def transcode(self, in_filepath, out_filepath):
"""Transcode audio file."""
logger.info("Transcoding from {} to {}", in_filepath, out_filepath)
try:
if not self._mode.startswith('replaygain'):
audiotools.open(in_filepath).convert(
out_filepath, self._format, compression=self._compression)
else:
in_file = audiotools.open(in_filepath)
rp_info = self.get_replaygain(in_filepath)
if rp_info:
pcmreader = audiotools.replaygain.ReplayGainReader(
in_file.to_pcm(),
rp_info.gain + self._replaygain_preamp_gain,
rp_info.peak)
self._format.from_pcm(out_filepath,
pcmreader,
compression=self._compression)
else:
logger.warning("No ReplayGain info found {}", in_filepath)
audiotools.open(in_filepath).convert(
out_filepath,
self._format,
compression=self._compression)
except (audiotools.EncodingError, audiotools.UnsupportedFile) as err:
raise IOError("Failed to transcode file {}: {}".format(
in_filepath, err)) from err
def convert_to_flac(self, path, r_hash):
w_path = "{0}/out_{1}.wav".format(path, r_hash)
n_path = "{0}/out_{1}.flac".format(path, r_hash)
if self.debug:
print "**converting {0}/out_{1}.wav to {0}/out_{1}.flac**".format(path, r_hash)
audiotools.open(w_path).convert(n_path, audiotools.FlacAudio)
return n_path
def supports_file(self, fname: str) -> bool:
# Readable by audiotools?
try:
audiotools.open(fname)
return True
except UnsupportedFile:
return False
def any_to_wav(src, dst):
log.info('any to wav: %s > %s' % (src, dst))
if not os.path.isfile(src):
raise IOError('unable to access %s' % src)
src_path, src_ext = os.path.splitext(src)
if src_ext.lower() == '.wav':
# shutil.copyfile(src, dst)
return src
try:
audiotools.open(src).convert(dst, audiotools.WaveAudio)
except Exception as e:
log.info('file %s not supported by audiotools' % src)
name, ext = os.path.splitext(src)
if ext.lower() in ['.mp3', ]:
mp3_to_wav(src, dst)
if ext.lower() in ['.m4a', '.mp4']:
m4a_to_wav(src, dst)
print name
print ext
log.debug('to wav: %s > %s' % (src, dst))
if os.path.exists(dst):
return dst
def any_to_wav(src, dst):
log.info('any to wav: %s > %s' % (src, dst))
if not os.path.isfile(src):
raise IOError('unable to access %s' % src)
src_path, src_ext = os.path.splitext(src)
if src_ext.lower() == '.wav':
# shutil.copyfile(src, dst)
return src
try:
audiotools.open(src).convert(dst, audiotools.WaveAudio)
except Exception as e:
log.info('file %s not supported by audiotools' % src)
name, ext = os.path.splitext(src)
if ext.lower() in [
'.mp3',
]:
mp3_to_wav(src, dst)
if ext.lower() in ['.m4a', '.mp4']:
m4a_to_wav(src, dst)
print name
print ext
log.debug('to wav: %s > %s' % (src, dst))
if os.path.exists(dst):
return dst
def convert_wav_to_flac(audio):
"""
Takes the path of a wav audio file, then converts it to flac 16000hz and names it temp.flac
Note, you must install the audiotools package.
http://sourceforge.net/projects/audiotools/files/
"""
audiotools.open(audio).convert("temp.flac",audiotools.FlacAudio)
def check_file(self):
try:
self.verify()
except (audiotools.InvalidFile, audiotools.UnsupportedFile,
ValueError):
# We failed to open it in audiotools, so the streamer
# won't be able to play it; try re-encoding for good measure.
try:
filename = self.encode()
except VerifyError:
error("failed to encode the song")
except:
error("unknown error on {0}".format(self.filename))
fileobj = audiotools.open(filename)
length = long(fileobj.seconds_length())
# If there was no error we now have two files, delete the old one
os.remove(args.file)
else:
filename = self.new_path(self.uuid())
self.rename(self.filename, filename)
fileobj = audiotools.open(filename)
length = long(fileobj.seconds_length())
print json.dumps({
"length": length,
"success": True,
"filename": filename
})
sys.exit(0)
def wavtoflac(filepath_wav, mode):
mode = 8 #default
filepath_flac = filepath_wav.replace(".wav", ".flac")
audiotools.open(filepath_wav).convert(
filepath_flac,
audiotools.FlacAudio,
compression=audiotools.FlacAudio.COMPRESSION_MODES[mode])
return filepath_flac
def convert_wav_to_flac(audio):
"""
Takes the path of a wav audio file, then converts it to flac 16000hz and names it voice.flac
returns the name of the file generated.
"""
audiotools.open(audio).convert("voice.flac",audiotools.FlacAudio)
os.remove(audio)
return 'voice.flac'
def main():
for file1 in fileIn1:
if not ('.flac' in file1):
continue
adopath_flac = os.path.join(path1, file1)
adopath_wav = adopath_flac.replace('flac', 'wav')
audiotools.open(adopath_flac).convert(
adopath_wav, audiotools.WaveAudio) # sampleRate 16000Hz
with wave.open(adopath_wav, 'rb') as f1:
params1 = f1.getparams()
nchannels1, sampwidth1, framerate1, nframes1, comptype1, compname1 = params1[:
6]
f1_str_data = f1.readframes(nframes1)
f1_wave_data = np.fromstring(f1_str_data, dtype=np.int16)
musicID = random.randint(0, musicNum - 1) # [a, b]
music_wav = os.path.join(path2, fileIn2[musicID]) # sampleRate 44100Hz
with wave.open(music_wav, 'rb') as f2:
params2 = f2.getparams()
nchannels2, sampwidth2, framerate2, nframes2, comptype2, compname2 = params2[:
6]
f2_str_data = f2.readframes(nframes2)
f2_wave_data = np.fromstring(f2_str_data, dtype=np.int16)
f2_wave_data = f2_wave_data[::3] # decrease sampleRate
nframes2 = len(f2_wave_data)
if nframes1 < nframes2:
length = abs(nframes2 - nframes1)
startp = random.randint(0, length)
rf1_wave_data = f1_wave_data
rf2_wave_data = f2_wave_data[startp:startp + nframes1]
elif nframes1 > nframes2:
length = abs(nframes1 - nframes2)
temp_array = np.zeros(length, dtype=np.int16)
rf2_wave_data = np.concatenate((f2_wave_data, temp_array))
rf1_wave_data = f1_wave_data
else:
rf1_wave_data = f1_wave_data
rf2_wave_data = f2_wave_data
new_wave_data = rf1_wave_data + rf2_wave_data
new_wave = new_wave_data.tostring()
out_wav = adopath_wav
with wave.open(out_wav, 'wb') as wf:
wf.setnchannels(1)
wf.setsampwidth(2)
wf.setframerate(16000)
wf.writeframes(new_wave)
out_flac = adopath_flac
audiotools.open(out_wav).convert(out_flac, audiotools.FlacAudio)
def Flac2Wav(flac_file, wav_path):
wav_path = wav_path + "/"
if not os.path.exists(wav_path):
os.makedirs(wav_path)
file_name = os.path.basename(flac_file)
wav_name = file_name.replace(".WAV", ".wav")
wav_file = wav_path + wav_name
audiotools.open(flac_file).convert(wav_file, audiotools.WaveAudio)
return wav_file
def convert_wav_flac(self, filename = "temp.wav"):
name, ext = path.splitext(filename)
flac = "%s.flac" % name
audiotools.open(filename).convert(flac, audiotools.FlacAudio)
f = open(flac)
data = f.read()
f.close()
return data
def _rec_googleapi(self, input):
print "in googleapi func @@@@@@@@"
f = open('python_api.yml')
yml_dict_speech = yaml.safe_load(f)
f.close()
APIKEY = yml_dict_speech['google_speech']['api_key']
audiotools.open(input).convert("output.flac", audiotools.FlacAudio)
headers = {"Content-Type": "audio/x-flac;rate=44100"}
speechfile = open("output.flac", "r")
target = "/speech-api/v2/recognize?output=json&lang=en-us&key=" + str(
APIKEY)
target = target.strip(
) # just in case there are leading or trailing whitespace or newlines
request = self.connection.request("POST", target, speechfile, headers)
hyp = ""
json_result = ""
try:
response = self.connection.getresponse()
if response.status == 200:
lines = str(response.read()).splitlines()
if (len(lines) > 1):
json_result = lines[1]
decoded = json.loads(json_result)
# pretty printing of json-formatted string
print json.dumps(decoded, sort_keys=True, indent=4)
# hyp = decoded['result'][0]['alternative'] # [0]['transcript']
hyp = [
str(x["transcript"])
for x in decoded['result'][0]['alternative']
]
print "speech rec result: ", hyp
else:
print "got this response status from google: ", response.status
hyp = [
"error, got this response status from google: " +
response.status
]
return hyp
except httplib.BadStatusLine:
print "got bad status line error, trying to make google request again"
self.recognize(input) # call again
finally:
speechfile.close()
os.remove("output.flac")
self.connection.close()
def proces_to_google(path):
# converteren naar FLAC
input = audiotools.open(path)
input.convert(
(u"%s.flac" % path),
audiotools.FlacAudio,
compression=audiotools.FlacAudio.COMPRESSION_MODES[8],
progress=progress,
)
path = u"%s.flac" % path
# versturen naar google
f = open(path, "rb")
flac_cont = f.read()
f.close()
lang_code = "en-US"
googl_speech_url = (
"https://www.google.com/speech-api/v1/recognize?xjerr=1&client=chromium&pfilter=2&lang=%s&maxresults=6"
% (lang_code)
)
hrs = {
"User-Agent": "Mozilla/5.0 (X11; Linux i686) AppleWebKit/535.7 (KHTML, like Gecko) Chrome/16.0.912.63 Safari/535.7",
"Content-type": "audio/x-flac; rate=16000",
}
req = urllib2.Request(googl_speech_url, data=flac_cont, headers=hrs)
p = urllib2.urlopen(req)
res = eval(p.read())["hypotheses"]
map(os.remove, (path))
return res
def process(self):
iext = None
try:
iext = os.path.splitext(self.master.path)[1].lower()
iext = iext[1:]
audiofile = audiotools.open(self.master.path)
base_format = iext
base_bitrate = audiofile.bits_per_sample()
base_samplerate = audiofile.sample_rate()
base_filesize = os.path.getsize(self.master.path)
base_duration = audiofile.seconds_length()
try:
base_duration = float(audiofile.total_frames()) / float(audiofile.sample_rate()) * 1000
print 'frames: %s' % audiofile.total_frames()
print 'rate: %s' % audiofile.sample_rate()
print base_duration
print
except:
pass
self.processed = 1
except Exception, e:
print e
base_bitrate = None
base_samplerate = None
base_filesize = None
base_duration = None
self.processed = 2
def __init__(self, filename):
from .id3 import skip_id3v2_comment
self.filename = filename
try:
f = open(filename, "rb")
self.__stream_offset__ = skip_id3v2_comment(f)
from .bitstream import BitstreamReader
from .text import (ERR_TTA_INVALID_SIGNATURE,
ERR_TTA_INVALID_FORMAT)
reader = BitstreamReader(f, True)
(signature,
format_,
self.__channels__,
self.__bits_per_sample__,
self.__sample_rate__,
self.__total_pcm_frames__) = reader.parse(
"4b 16u 16u 16u 32u 32u 32p")
if (signature != "TTA1"):
raise InvalidTTA(ERR_TTA_INVALID_SIGNATURE)
elif (format_ != 1):
raise InvalidTTA(ERR_TTA_INVALID_FORMAT)
self.__total_tta_frames__ = div_ceil(
self.__total_pcm_frames__ * 245,
self.__sample_rate__ * 256)
self.__frame_lengths__ = list(reader.parse(
"%d* 32u" % (self.__total_tta_frames__) + "32p"))
except IOError, msg:
raise InvalidTTA(str(msg))
def _open_file(self, filename):
"""Open a file for reading and wrap it in several helpers."""
# Audiotools seems to hate unicode, so we.. don't give it that
if isinstance(filename, unicode):
filename = filename.encode('utf8')
try:
reader = audiotools.open(filename)
except (audiotools.UnsupportedFile):
raise AudioError("Unsupported file")
self.file = reader
total_frames = reader.total_frames()
# Wrap in a PCMReader because we want PCM
reader = reader.to_pcm()
# Wrap in a converter
reader = audiotools.PCMConverter(
reader,
sample_rate=44100,
channels=2,
channel_mask=audiotools.ChannelMask(0x1 | 0x2),
bits_per_sample=24,
)
# And for file progress!
reader = audiotools.PCMReaderProgress(reader, total_frames,
self.progress)
return reader
def fix_id3_preserve_originals(self, tempFilePath):
f = file(self.filename, 'rb')
#figure out where the start and end points of the FLAC file are
audiotools.ID3v2Comment.skip(f)
flac_start = f.tell()
f.seek(-128, 2)
if (f.read(3) == 'TAG'):
f.seek(-3, 1)
flac_end = f.tell()
else:
f.seek(0, 2)
flac_end = f.tell()
#copy the FLAC data to a temporary location
temp = tempfile.TemporaryFile()
f.seek(flac_start, 0)
reader = audiotools.__capped_stream_reader__(f, flac_end - flac_start)
audiotools.transfer_data(reader.read, temp.write)
#rewrite the original FLAC with our temporary data
temp.seek(0, 0)
f.close()
f = file(tempFilePath, 'wb')
audiotools.transfer_data(temp.read, f.write)
temp.close()
f.close()
returnValue = audiotools.open(tempFilePath)
return returnValue
def testShnConversion(self):
audioFile = audiotools.open(self.showPath + '/show2/CD1/1.shn')
pcmReader = audioFile.to_pcm()
self.antiCrashBin.append(pcmReader)
alacFile = audiotools.ALACAudio.from_pcm(self.testPath + '/output/2.m4a', pcmReader)
self.assertTrue(os.path.exists(self.testPath + '/output/2.m4a'))
pass
def fixBadFlacFiles(self, metadata):
"""
Fix FLAC files that are malformed because they contain ID3 tags.
Replace the audioFile paths the metadata dict with the paths to
temporary fixed versions.
"""
parent = self.parent()
for index, audioFile in enumerate(metadata['audioFiles']):
if self.stopped:
return
self.updateProgress(
'Loading %s\n\nFixing malformed FLACs (%d\%d)'
% (self.basename, index+1, len(metadata['audioFiles'])),
)
audioObj = audiotools.open(audioFile)
if isinstance(audioObj, tracklint.BrokenFlacAudio):
tempFilePath = unicode(
metadata['tempDir'].absolutePath()
) + u'/%d.flac' % index
if platform.system() == 'Darwin':
self.process = Process(
target=self.fixFlacProcess,
args=(tempFilePath, audioObj)
)
self.process.start()
while self.process.is_alive():
pass
else:
self.fixFlacProcess(tempFilePath, audioObj)
metadata['audioFiles'][index] = (tempFilePath)
def __init__(self, file_path):
"""
:param file_path: path to audio file
"""
# check if the file format is supported
if not os.path.isfile(file_path):
raise ValueError("Could not find file {0}".format(file_path))
with open(file_path, "rb") as f:
supported = audiotools.file_type(f) in audiotools.AVAILABLE_TYPES
if not supported:
raise UnsupportedFileType("File format not supported.")
# open the audiofile and extract the pcm stream
self.file = audiotools.open(file_path)
if not self.file.supports_to_pcm():
raise UnsupportedFileType("Can not extract the pcm stream.")
# is the file seekable
self.seekable = self.file.seekable()
# the total number of frames in this file
self._num_frames = self.file.total_frames()
# create a pcm reader
pcm = self.file.to_pcm()
if self.seekable:
self._pcm = pcm
else:
# load the whole file into memory if seeking is not available
self._data = self._load(pcm)
# Todo: just for debugging
self._num_frames = len(self._data)
# save the last frame number
self._next_frame_number = 0
def delete_metadata(self):
"""deletes the track's MetaData
this removes or unsets tags as necessary in order to remove all data
raises IOError if unable to write the file"""
import os
from audiotools import (TemporaryFile,
LimitedFileReader,
transfer_data)
#this works a lot like update_metadata
#but without any new metadata to set
if (not os.access(self.filename, os.W_OK)):
raise IOError(self.filename)
new_mp3 = TemporaryFile(self.filename)
#get the original MP3 data
old_mp3 = open(self.filename, "rb")
MP3Audio.__find_last_mp3_frame__(old_mp3)
data_end = old_mp3.tell()
old_mp3.seek(0, 0)
MP3Audio.__find_mp3_start__(old_mp3)
data_start = old_mp3.tell()
old_mp3 = LimitedFileReader(old_mp3, data_end - data_start)
#write data to file
transfer_data(old_mp3.read, new_mp3.write)
#commit change to disk
old_mp3.close()
new_mp3.close()
def process(self):
iext = None
try:
iext = os.path.splitext(self.master.path)[1].lower()
iext = iext[1:]
audiofile = audiotools.open(self.master.path)
base_format = iext
base_bitrate = audiofile.bits_per_sample()
base_samplerate = audiofile.sample_rate()
base_filesize = os.path.getsize(self.master.path)
base_duration = audiofile.seconds_length()
try:
base_duration = float(audiofile.total_frames()) / float(audiofile.sample_rate()) * 1000
print 'frames: %s' % audiofile.total_frames()
print 'rate: %s' % audiofile.sample_rate()
print base_duration
print
except:
pass
self.processed = 1
except Exception as e:
print e
base_bitrate = None
base_samplerate = None
base_filesize = None
base_duration = None
self.processed = 2
"""
self.base_format = iext
self.base_bitrate = base_bitrate
self.base_samplerate = base_samplerate
self.base_filesize = base_filesize
"""
self.duration = base_duration
meta = None
try:
meta = EasyID3(self.master.path)
log.debug('using EasyID3')
except Exception as e:
meta = MutagenFile(self.master.path)
log.debug('using MutagenFile')
if 'title' in meta:
self.name = meta['title'][0]
if 'artist' in meta:
self.artist, created = Artist.objects.get_or_create(name=meta['artist'][0])
self.save()
try:
self.create_waveform_image()
except:
pass
return
def process(self):
iext = None
try:
iext = os.path.splitext(self.master.path)[1].lower()
iext = iext[1:]
audiofile = audiotools.open(self.master.path)
base_format = iext
base_bitrate = audiofile.bits_per_sample()
base_samplerate = audiofile.sample_rate()
base_filesize = os.path.getsize(self.master.path)
base_duration = audiofile.seconds_length()
try:
base_duration = float(audiofile.total_frames()) / float(
audiofile.sample_rate()) * 1000
print 'frames: %s' % audiofile.total_frames()
print 'rate: %s' % audiofile.sample_rate()
print base_duration
print
except:
pass
self.processed = 1
except Exception, e:
print e
base_bitrate = None
base_samplerate = None
base_filesize = None
base_duration = None
self.processed = 2
def power_chroma(file_name, hop_length=512, power=2):
song = audiotools.open(file_name)
sr = song.sample_rate()
y, sr = librosa.load(file_name, sr=sr)
# Compute the STFT matrix
stft = librosa.core.stft(y, n_fft=4096, hop_length=int(hop_length))
# Decompose into harmonic and percussives
stft_harm, stft_perc = librosa.decompose.hpss(stft)
stft_harm = np.abs(stft_harm)**power
# stft_perc = np.log(np.abs(stft_perc))
# Invert the STFTs. Adjust length to match the input.
# y_harmonic = librosa.util.fix_length(librosa.core.istft(stft_harm, dtype=y.dtype), len(y))
y_percussive = librosa.util.fix_length(
librosa.core.istft(stft_perc,
dtype=y.dtype,
hop_length=int(hop_length)), len(y))
chromagram = librosa.feature.chroma_stft(S=stft_harm,
sr=sr,
hop_length=int(hop_length))
tempo, beat_frames = librosa.beat.beat_track(y=y_percussive,
sr=sr,
hop_length=hop_length)
beat_t = librosa.frames_to_time(beat_frames, sr=sr, hop_length=hop_length)
# Aggregate chroma features between beat events # The median feature is used instead
beat_chroma = librosa.util.sync(chromagram,
beat_frames,
aggregate=np.median)
return chromagram, beat_chroma, beat_frames, beat_t, sr
def chroma(file_name, hop_length=512, type_='cqt', tol=0.0):
# An 11 kHz sample rate is used because it was suggested in the paper.
eprint('Processing file: {}'.format(file_name))
song = audiotools.open(file_name)
sr = song.sample_rate()
y, sr = librosa.load(file_name, sr=sr)
# Separate harmonics and percussives into two waveforms
y_harmonic, y_percussive = librosa.effects.hpss(y)
# Beat track on the percussive signal
tempo, beat_frames = librosa.beat.beat_track(y=y_percussive, sr=sr)
beat_t = librosa.frames_to_time(beat_frames, sr=sr)
if type_ == 'cqt':
# Compute chroma features from the harmonic signal
chromagram = librosa.feature.chroma_cqt(y=y_harmonic,
sr=sr,
hop_length=hop_length,
threshold=tol)
elif type_ == 'stft':
chromagram = librosa.feature.chroma_stft(y=y_harmonic,
sr=sr,
hop_length=hop_length)
else:
raise Exception("Must specify chromagram type!")
# Aggregate chroma features between beat events # The median feature is used instead
beat_chroma = librosa.util.sync(chromagram,
beat_frames,
aggregate=np.median)
# commented out,
# chromagram, beat_chroma, target_file, target_image_file = write_chroma_info(file_name, chromagram, beat_chroma)
# save_chroma_pics(chromagram, beat_chroma, beat_frames, beat_t, sr, "Please Please Me", "Please Please Me")
return chromagram, beat_chroma, beat_frames, beat_t, sr
def get_text(self):
if not self.text is None:
return self.text
print "Converting to FLAC"
(_, recording_flac_filename) = tempfile.mkstemp('.flac')
recording_wav = audiotools.open(self.audio.filename())
recording_wav.convert(recording_flac_filename,
audiotools.FlacAudio,)
#compression=audiotools.FlacAudio.COMPRESSION_MODES[8],
#progress=False)
# turn the audio into useful text
print "Sending to Google"
google_speech_url = "http://www.google.com/speech-api/v1/recognize?lang=en"
headers = {'Content-Type': 'audio/x-flac; rate= %d;' % self.recordingRate}
recording_flac_data = open(recording_flac_filename, 'rb').read()
r = requests.post(google_speech_url, data=recording_flac_data, headers=headers)
# housekeeping
os.remove(recording_flac_filename)
self.audio.housekeeping()
# grab the response
response = r.text
if not 'hypotheses' in response:
raise NotUnderstoodException()
# we are only interested in the most likely utterance
phrase = json.loads(response)['hypotheses'][0]['utterance']
print "Heard: " + phrase
return str(phrase)
def open_audiotools(f_name, fmt=None):
if not isinstance(f_name, str):
return None
try:
f = audiotools.open(f_name)
except audiotools.UnsupportedFile:
return None
sample_width = f.bits_per_sample()
pcm = f.to_pcm()
frag = Fragment(rate=f.sample_rate(), channels=f.channels(),
length=f.total_frames())
frame_size = frag.channels * sample_width / 8
chunk_size = 65536 / frame_size
rem = f.total_frames()
cur = 0
while rem > 0:
frames = pcm.read(chunk_size)
if frames.frames == 0:
# Some lossy formats like MP3 have slightly varying data length
break
raw_bytes = bytearray(frames.to_bytes(False, True))
frag.import_bytes(raw_bytes, frag.rate, frag.channels,
'int{:d}'.format(sample_width), cur)
rem -= frames.frames
cur += frames.frames
return frag
def playOrStopSelected(self, rowClicked):
"""
Play the selected track's file using the default
application for that file type.
"""
rowStopped = self.stopAudio()
if rowStopped is rowClicked:
return
item = QTableWidgetItem(str(rowClicked + 1))
item.setIcon(self.stopButtonIcon)
self.tracklistTableWidget.setVerticalHeaderItem(rowClicked, item)
encoding = sys.getfilesystemencoding()
audiofileObj = audiotools.open(
self.metadata['audioFiles'][rowClicked].encode(encoding))
self.pyaudioObj = pyaudio.PyAudio()
self.stream = self.pyaudioObj.open(format=pyaudio.paInt16,
channels=audiofileObj.channels(),
rate=audiofileObj.sample_rate(),
output=True)
self.playing = rowClicked
self.playAudioThread = PlayAudioThread(self, audiofileObj, self.stream)
self.connect(self.playAudioThread, SIGNAL('done()'), self.stopAudio)
self.playAudioThread.start()
def open_audiotools(f_name, fmt=None):
if not isinstance(f_name, str):
return None
try:
f = audiotools.open(f_name)
except audiotools.UnsupportedFile:
return None
sample_width = f.bits_per_sample()
pcm = f.to_pcm()
frag = Fragment(rate=f.sample_rate(),
channels=f.channels(),
length=f.total_frames())
frame_size = frag.channels * sample_width / 8
chunk_size = 65536 / frame_size
rem = f.total_frames()
cur = 0
while rem > 0:
frames = pcm.read(chunk_size)
if frames.frames == 0:
# Some lossy formats like MP3 have slightly varying data length
break
raw_bytes = bytearray(frames.to_bytes(False, True))
frag.import_bytes(raw_bytes, frag.rate, frag.channels,
'int{:d}'.format(sample_width), cur)
rem -= frames.frames
cur += frames.frames
return frag
def delete_metadata(self):
"""deletes the track's MetaData
this removes or unsets tags as necessary in order to remove all data
raises IOError if unable to write the file"""
import os
from audiotools import (TemporaryFile, LimitedFileReader,
transfer_data)
#this works a lot like update_metadata
#but without any new metadata to set
if (not os.access(self.filename, os.W_OK)):
raise IOError(self.filename)
new_mp3 = TemporaryFile(self.filename)
#get the original MP3 data
old_mp3 = open(self.filename, "rb")
MP3Audio.__find_last_mp3_frame__(old_mp3)
data_end = old_mp3.tell()
old_mp3.seek(0, 0)
MP3Audio.__find_mp3_start__(old_mp3)
data_start = old_mp3.tell()
old_mp3 = LimitedFileReader(old_mp3, data_end - data_start)
#write data to file
transfer_data(old_mp3.read, new_mp3.write)
#commit change to disk
old_mp3.close()
new_mp3.close()
def play_ogg(self, path):
file = audiotools.open(path)
stream = file.to_pcm()
self.context.log("play ogg using device {0}".format(self.output_device_index))
self.output_stream = self.pa.open(format = FORMAT,
channels = 2,
rate = RATE,
input = False,
output = True,
output_device_index = self.output_device_index,
frames_per_buffer = INPUT_FRAMES_PER_BLOCK)
# read data (based on the chunk size)
framelist = stream.read(CHUNK)
data = framelist.to_bytes(False, True)
# play stream (looping from beginning of file to the end)
while data != '':
# writing to the stream is what *actually* plays the sound.
self.output_stream.write(data)
framelist = stream.read(CHUNK)
data = framelist.to_bytes(False, True)
# cleanup stuff.
self.output_stream.close()
self.pa.terminate()
def verify(self, filename=None):
if not filename:
filename = self.filename
audio = audiotools.open(filename)
# Do verification with audiotools; this might be too strict
audio.verify()
def verify(self, filename=None):
if not filename:
filename = self.filename
audio = audiotools.open(filename)
# Do verification with audiotools; this might be too strict
audio.verify()
def toMp3(inputFilePath, outputFilePath):
mp3 = audiotools.MP3Audio.from_pcm(outputFilePath,
audiotools.open(inputFilePath).to_pcm())
os.chmod(outputFilePath, NEW_FILE_PERMISSIONS)
def any_to_wav(src, dst=None):
log.info('any to wav: %s > %s' % (src, dst))
if not os.path.isfile(src):
log.error('unable to access %s' % src)
raise IOError('unable to access %s' % src)
src_path, src_ext = os.path.splitext(src)
if src_ext.lower() == '.wav':
shutil.copyfile(src, dst)
return dst
try:
audiotools.open(src).convert(dst, audiotools.WaveAudio)
except Exception as e:
log.info('file %s not supported by audiotools' % src)
name, ext = os.path.splitext(src)
if ext.lower() in [
'.mp3',
]:
mp3_to_wav(src, dst)
if ext.lower() in [
'.m4a',
'.mp4',
]:
m4a_to_wav(src, dst)
# fallback version via ffmpeg
if ext.lower() in [
'.flac',
]:
ffmpeg_to_wav(src, dst)
log.debug('to wav: %s > %s' % (src, dst))
if not os.path.exists(dst):
log.warning('unable to convert {} to wav.'.format(src))
return
return dst
def metadata_audiotools(self, path, media):
from audiotools import MetaData
import audiotools
meta = MetaData()
# release-level metadata
if media.release and media.release.main_image:
if meta.supports_images() and os.path.exists(media.release.main_image.path):
opt = dict(size=(200, 200), crop=True, bw=False, quality=80)
image = get_thumbnailer(media.release.main_image).get_thumbnail(opt)
meta.add_image(get_raw_image(image.path, 0))
audiotools.open(path).update_metadata(meta)
return
def create_waveform_image(self):
if self.master:
tmp_directory = tempfile.mkdtemp()
src_path = self.master.path;
tmp_path = os.path.join(tmp_directory, 'tmp.wav')
dst_path = os.path.join(self.get_folder_path('cache'), 'waveform.png')
audiotools.open(src_path).convert(tmp_path, audiotools.WaveAudio)
args = (tmp_path, dst_path, None, 1800, 301, 2048)
create_wave_images(*args)
try:
shutil.rmtree(tmp_directory)
except Exception, e:
print e
def insert_track(source_filename):
mimetype = _magic.from_file(source_filename)
hash_ = _ca_copy_contents(open(source_filename, 'rb'))
source_file = audiotools.open(source_filename)
metadata = source_file.get_metadata()
return TrackOriginal(Blob(hash_, mimetype),
metadata,
metadata.artist_name,
metadata.track_name)
def clean_audio_file(self):
audio = self.cleaned_data['audio_file']
if audio.size > 10*1024*1024:
raise ValidationError(self.error_messages['audio_exceed'])
allowed_type = ['audio/mpeg', 'audio/ogg', 'audio/mp3']
if not (audio.content_type in allowed_type):
raise ValidationError(self.error_messages['audio_invalid'])
randomfilename = str(time.time())+''.join(random.choice(string.ascii_uppercase + string.digits) for _ in range(5))
path = default_storage.save('tmp/'+randomfilename, ContentFile(audio.read()))
try:
audiotools.open(os.path.join(settings.MEDIA_ROOT, path))
default_storage.delete('tmp/'+randomfilename)
except audiotools.UnsupportedFile:
default_storage.delete('tmp/'+randomfilename)
raise ValidationError(self.error_messages['audio_corrupted'])
return audio
def create_waveform_image(self):
if self.master:
tmp_directory = tempfile.mkdtemp()
src_path = self.master.path
tmp_path = os.path.join(tmp_directory, 'tmp.wav')
dst_path = os.path.join(self.get_folder_path('cache'),
'waveform.png')
audiotools.open(src_path).convert(tmp_path, audiotools.WaveAudio)
args = (tmp_path, dst_path, None, 1800, 301, 2048)
create_wave_images(*args)
try:
shutil.rmtree(tmp_directory)
except Exception, e:
print e
def any_to_wav(src, dst=None):
log.info('any to wav: %s > %s' % (src, dst))
if not os.path.isfile(src):
log.error('unable to access %s' % src)
raise IOError('unable to access %s' % src)
src_path, src_ext = os.path.splitext(src)
if src_ext.lower() == '.wav':
shutil.copyfile(src, dst)
return dst
try:
audiotools.open(src).convert(dst, audiotools.WaveAudio)
except Exception as e:
log.info('file %s not supported by audiotools' % src)
name, ext = os.path.splitext(src)
if ext.lower() in ['.mp3',]:
mp3_to_wav(src, dst)
if ext.lower() in ['.m4a', '.mp4',]:
m4a_to_wav(src, dst)
# fallback version via ffmpeg
if ext.lower() in ['.flac',]:
ffmpeg_to_wav(src, dst)
log.debug('to wav: %s > %s' % (src, dst))
if not os.path.exists(dst):
log.warning('unable to convert {} to wav.'.format(src))
return
return dst
def _rec_googleapi(self, input):
print "in googleapi func @@@@@@@@"
f = open('python_api.yml')
yml_dict_speech = yaml.safe_load(f)
f.close()
APIKEY = yml_dict_speech['google_speech']['api_key']
audiotools.open(input).convert("output.flac", audiotools.FlacAudio)
headers = {"Content-Type": "audio/x-flac;rate=44100"}
speechfile = open("output.flac","r")
target = "/speech-api/v2/recognize?output=json&lang=en-us&key=" + str(APIKEY)
target = target.strip() # just in case there are leading or trailing whitespace or newlines
request = self.connection.request("POST",target,speechfile,headers)
hyp = ""
json_result = ""
try:
response = self.connection.getresponse()
if response.status == 200:
lines = str(response.read()).splitlines()
if(len(lines) > 1):
json_result = lines[1]
decoded = json.loads(json_result)
# pretty printing of json-formatted string
print json.dumps(decoded, sort_keys=True, indent=4)
# hyp = decoded['result'][0]['alternative'] # [0]['transcript']
hyp = [str(x["transcript"]) for x in decoded['result'][0]['alternative']]
print "speech rec result: ", hyp
else:
print "got this response status from google: ",response.status
hyp = ["error, got this response status from google: " + response.status]
return hyp
except httplib.BadStatusLine:
print "got bad status line error, trying to make google request again";
self.recognize(input) # call again
finally:
speechfile.close()
os.remove("output.flac")
self.connection.close()
def verify(self, progress=None):
"""verifies the current file for correctness
returns True if the file is okay
raises an InvalidFile with an error message if there is
some problem with the file"""
from . import verify
try:
f = open(self.filename, 'rb')
except IOError, err:
raise InvalidMP3(str(err))
def __init__(self, filename):
"""filename is a plain string"""
AudioFile.__init__(self, filename)
from .bitstream import BitstreamReader
import cStringIO
try:
mp3file = open(filename, "rb")
except IOError, msg:
raise InvalidMP3(str(msg))
def __init__(self, filename):
"""filename is a plain string"""
AudioFile.__init__(self, filename)
from .bitstream import BitstreamReader
import cStringIO
try:
mp3file = open(filename, "rb")
except IOError, msg:
raise InvalidMP3(str(msg))
def update_metadata(self, metadata):
"""takes this track's current MetaData object
as returned by get_metadata() and sets this track's metadata
with any fields updated in that object
raises IOError if unable to write the file
"""
import os
from audiotools import (TemporaryFile,
LimitedFileReader,
transfer_data)
from audiotools.id3 import (ID3v2Comment, ID3CommentPair)
from audiotools.id3v1 import ID3v1Comment
from audiotools.bitstream import BitstreamWriter
if (metadata is None):
return
elif (not (isinstance(metadata, ID3v2Comment) or
isinstance(metadata, ID3CommentPair) or
isinstance(metadata, ID3v1Comment))):
from .text import ERR_FOREIGN_METADATA
raise ValueError(ERR_FOREIGN_METADATA)
elif (not os.access(self.filename, os.W_OK)):
raise IOError(self.filename)
new_mp3 = TemporaryFile(self.filename)
#get the original MP3 data
old_mp3 = open(self.filename, "rb")
MP3Audio.__find_last_mp3_frame__(old_mp3)
data_end = old_mp3.tell()
old_mp3.seek(0, 0)
MP3Audio.__find_mp3_start__(old_mp3)
data_start = old_mp3.tell()
old_mp3 = LimitedFileReader(old_mp3, data_end - data_start)
#write id3v2 + data + id3v1 to file
if (isinstance(metadata, ID3CommentPair)):
metadata.id3v2.build(BitstreamWriter(new_mp3, 0))
transfer_data(old_mp3.read, new_mp3.write)
metadata.id3v1.build(new_mp3)
elif (isinstance(metadata, ID3v2Comment)):
metadata.build(BitstreamWriter(new_mp3, 0))
transfer_data(old_mp3.read, new_mp3.write)
elif (isinstance(metadata, ID3v1Comment)):
transfer_data(old_mp3.read, new_mp3.write)
metadata.build(new_mp3)
#commit change to disk
old_mp3.close()
new_mp3.close()
def clean_audio_file(self):
audio = self.cleaned_data['audio_file']
if audio.size > 10 * 1024 * 1024:
raise ValidationError(self.error_messages['audio_exceed'])
allowed_type = ['audio/mpeg', 'audio/ogg', 'audio/mp3']
if not (audio.content_type in allowed_type):
raise ValidationError(self.error_messages['audio_invalid'])
randomfilename = str(time.time()) + ''.join(
random.choice(string.ascii_uppercase + string.digits)
for _ in range(5))
path = default_storage.save('tmp/' + randomfilename,
ContentFile(audio.read()))
try:
audiotools.open(os.path.join(settings.MEDIA_ROOT, path))
default_storage.delete('tmp/' + randomfilename)
except audiotools.UnsupportedFile:
default_storage.delete('tmp/' + randomfilename)
raise ValidationError(self.error_messages['audio_corrupted'])
return audio
def play(self):
# audiotools.player.available_outputs()
audio_file = audiotools.open(
'/home/dhruv/Music/Music/Aisha (2010) ~ 320 VBR/04 - Behke Behke [DoReGaMa].mp3'
)
audio_output = audiotools.player.open_output('ALSA')
replay_gain = audiotools.player.RG_NO_REPLAYGAIN
player = audiotools.player.Player(audio_output, replay_gain)
player.open(audio_file)
player.play()
def process(self, q):
"""This is the worker thread function.
It processes files in the queue one after
another. These daemon threads go into an
infinite loop, and only exit when
the main thread ends.
"""
while True:
global logger
compression_quality = '0' #min compression
file_flac = q.get()
file_wav = file_flac.replace(".flac", ".wav")
try:
audiotools.open(file_flac).convert(file_wav,audiotools.WavAudio, compression_quality)
logger.info(''.join(["Converted ", file_flac, " to: ", file_wav]))
os.remove(file_flac)
q.task_done()
except InvalidWave:
logger.error(''.join(["Failed to open file ", file_flac, " to: ", file_wav," failed."]), exc_info=True)
except Exception as e:
logger.error('ExFailed to open file', exc_info=True)
def play_song(self, song=None):
if song:
audio_file = audiotools.open(song['filepath'])
self.player.open(audio_file)
self.player.play()
time.sleep(0.1)
self.thread = threading.Thread(target=update_timer_seek,
args=(self.ui.timer_label, self.ui.seek,
self.player, song['duration']))
self.thread.start()
print(f'playing {song.get("filepath")}' if song else '')
def PopulateMetaData(self):
audioFile = audiotools.open(self.Filepath)
metaData = audioFile.get_metadata()
if not metaData:
self.ManualMetaData()
else:
self.Title = metaData.track_name
self.Artist = metaData.artist_name
if PREPROCESS:
self.Data = self.PrepareSongData()
print("Preproccessed " + self.Title)
def update_metadata(self, metadata):
"""takes this track's current MetaData object
as returned by get_metadata() and sets this track's metadata
with any fields updated in that object
raises IOError if unable to write the file
"""
import os
from audiotools import (TemporaryFile, LimitedFileReader,
transfer_data)
from audiotools.id3 import (ID3v2Comment, ID3CommentPair)
from audiotools.id3v1 import ID3v1Comment
from audiotools.bitstream import BitstreamWriter
if (metadata is None):
return
elif (not (isinstance(metadata, ID3v2Comment)
or isinstance(metadata, ID3CommentPair)
or isinstance(metadata, ID3v1Comment))):
from .text import ERR_FOREIGN_METADATA
raise ValueError(ERR_FOREIGN_METADATA)
elif (not os.access(self.filename, os.W_OK)):
raise IOError(self.filename)
new_mp3 = TemporaryFile(self.filename)
#get the original MP3 data
old_mp3 = open(self.filename, "rb")
MP3Audio.__find_last_mp3_frame__(old_mp3)
data_end = old_mp3.tell()
old_mp3.seek(0, 0)
MP3Audio.__find_mp3_start__(old_mp3)
data_start = old_mp3.tell()
old_mp3 = LimitedFileReader(old_mp3, data_end - data_start)
#write id3v2 + data + id3v1 to file
if (isinstance(metadata, ID3CommentPair)):
metadata.id3v2.build(BitstreamWriter(new_mp3, 0))
transfer_data(old_mp3.read, new_mp3.write)
metadata.id3v1.build(new_mp3)
elif (isinstance(metadata, ID3v2Comment)):
metadata.build(BitstreamWriter(new_mp3, 0))
transfer_data(old_mp3.read, new_mp3.write)
elif (isinstance(metadata, ID3v1Comment)):
transfer_data(old_mp3.read, new_mp3.write)
metadata.build(new_mp3)
#commit change to disk
old_mp3.close()
new_mp3.close()
def toNormalizedWav(inputFilePath, outputFilePath):
# open file (can raise errors)
orig = audiotools.open(inputFilePath)
## Decode file to raw audio (PCM)
origPCM = orig.to_pcm()
# If this is a surround sound file, or if there is no channel mask
# we shall error for now. This will probably not happen.
channel_mask = origPCM.channel_mask
if channel_mask < 2 or channel_mask > 3:
raise Exception ('Unsupported audio channel configuration')
# Make sure bit depth is not > 16bit.
# If it is, error for now. I tried to convert the file, but I
# get the following error: "data chunk ends prematurely"
bits_per_sample = origPCM.bits_per_sample
if bits_per_sample > 16:
# Uncomment this line to try changing the bit depth:
#bits_per_sample = 16
# Comment this line to try changing the bit depth:
raise Exception(
'Unsupported bit depth, try converting file to 16bit.'
)
# Make sure sample rate is not > 44100
sample_rate = origPCM.sample_rate
# If original sample rate was greater than 44100
if sample_rate > 44100:
# Convert to 44100
sample_rate = 44100
# Create new PCM stream at 44100, with 2 channels
normalizedPCM = audiotools.PCMConverter(origPCM, 44100, 2,
channel_mask, bits_per_sample)
# Output normalized audio to wav (Can raise error)
wav = audiotools.WaveAudio.from_pcm(outputFilePath,
normalizedPCM)
os.chmod(outputFilePath, NEW_FILE_PERMISSIONS)
def create_waveform_image(self):
if self.master:
tmp_directory = tempfile.mkdtemp()
src_path = self.master.path
tmp_path = os.path.join(tmp_directory, "tmp.wav")
dst_path = os.path.join(self.get_folder_path("cache"), "waveform.png")
print "create waveform"
print "src_path: %s" % src_path
print "tmp_path: %s" % tmp_path
print "dst_path: %s" % dst_path
audiotools.open(src_path).convert(tmp_path, audiotools.WaveAudio)
args = (tmp_path, dst_path, None, 1800, 301, 2048)
create_wave_images(*args)
try:
shutil.rmtree(tmp_directory)
except Exception, e:
print e
def insert_track(source_filename_or_fh):
if isinstance(source_filename_or_fh, str):
source_filename_or_fh = open(source_filename_or_fh, 'rb')
buf = source_filename_or_fh.read(4096)
source_filename_or_fh.seek(0)
mimetype = _magic.from_buffer(buf)
hash_ = _ca_copy_contents(source_filename_or_fh)
source_file = audiotools.open(cont_addr.addr_to_path(hash_))
metadata = source_file.get_metadata()
return TrackOriginal(Blob(hash_, mimetype),
metadata,
metadata.artist_name,
metadata.track_name)
