def load_audio_metadata_speechrecognition(cls, conn, path, audio_path):
'''
Pre-process and loads the metadata
Parameters
----------
conn : CAS
A connection object to the current session.
path : string
Location to the input metadata file.
audio_path : delimiter
Delimiter for the metadata file.
Returns
-------
:class:`CASTable`
'''
output_name = random_name('AudioTable_Metadata', 6)
dc = DataClean(conn=conn, contents_as_path=path)
dc_response = dc.process_contents(audio_path = audio_path)
tbl = dc.create_castable(dc_response['results'], output_name, replace=True, promote=False,
col_names=dc_response['col_names'])
scode = 'length _fName_ $1000; '
scode += '_fName_ = _filename_; '
ctbl = CASTable(tbl, computedvars=['_fName_'],
computedvarsprogram=scode)
conn.table.partition(table=ctbl, casout=dict(name=tbl, replace=True))
return CASTable(tbl)
def load_audio_metadata_speechrecognition(cls, conn, path, audio_path):
'''
Pre-process and loads the metadata
Parameters
----------
conn : CAS
A connection object to the current session.
path : string
Location to the input metadata file.
audio_path : delimiter
Delimiter for the metadata file.
Returns
-------
:class:`CASTable`
Examples
--------
>>> import swat
>>> from dlpy.audio import AudioTable
>>> s=swat.CAS("cloud.example.com", 5570)
>>> aud_tbl = AudioTable.load_audio_metadata_speechrecognition(s, path="/path/to/metadata/file.txt", audio_path="/path/to/audio/file.txt")
>>> aud_tbl.set_connection(s)
'''
if conn is None:
conn = cls.get_connection()
if conn is None:
raise DLPyError('cannot get a connection object to the current session.')
output_name = random_name('AudioTable_Metadata', 6)
dc = DataClean(conn=conn, contents_as_path=path)
dc_response = dc.process_contents(audio_path = audio_path)
tbl = dc.create_castable(dc_response['results'], output_name, replace=True, promote=False,
col_names=dc_response['col_names'])
scode = 'length _fName_ varchar(*); '
scode += '_fName_ = _filename_; '
ctbl = CASTable(tbl, computedvars=['_fName_'],
computedvarsprogram=scode)
conn.table.partition(table=ctbl, casout=dict(name=tbl, replace=True))
return CASTable(tbl)
def __extract_audio_features(conn, table, frame_shift=10, frame_length=25, n_bins=40, n_ceps=40,
feature_scaling_method='STANDARDIZATION', n_output_frames=500, casout=None,
label_level=0,
random_shuffle=True,
**kwargs):
conn.loadactionset('audio', _messagelevel='error')
if isinstance(table, AudioTable) is False and isinstance(table, CASTable) is False:
return None
if casout is None:
casout = dict(name=random_name('AudioTable', 6))
elif isinstance(casout, CASTable) or isinstance(casout, AudioTable):
casout = casout.to_outtable_params()
# always use dither with 0 to turn it off
rt = conn.retrieve('audio.computefeatures', _messagelevel='error', table=table,
frameExtractionOptions=dict(frameshift=frame_shift, framelength=frame_length, dither=0.0),
melBanksOptions=dict(nbins=n_bins), mfccOptions=dict(nceps=n_ceps),
featureScalingMethod=feature_scaling_method, nOutputFrames=n_output_frames,
casout=casout, **kwargs)
if rt.severity > 1:
for msg in rt.messages:
print(msg)
return None
server_type = get_cas_host_type(conn).lower()
if server_type.startswith("lin") or server_type.startswith("osx"):
fs = "/"
else:
fs = "\\"
if label_level:
scode = "i=find(_path_,'{0}',-length(_path_)); ".format(fs)
scode += "length _fName_ varchar(*); length _label_ varchar(*); "
scode += "_fName_=substr(_path_, i+length('{0}'), length(_path_)-i);".format(fs)
scode += "_label_=scan(_path_,{},'{}');".format(label_level, fs)
ctable = CASTable(casout['name'], computedvars=['_fName_', '_label_'],
computedvarsprogram=scode)
else:
scode = "i=find(_path_,'{0}',-length(_path_)); ".format(fs)
scode += "length _fName_ varchar(*); "
scode += "_fName_=substr(_path_, i+length('{0}'), length(_path_)-i);".format(fs)
ctable = CASTable(casout['name'], computedvars=['_fName_'],
computedvarsprogram=scode)
if random_shuffle:
conn.table.shuffle(table=ctable, casout=dict(name=casout['name'], replace=True))
else:
conn.table.partition(table=ctable, casout=dict(name=casout['name'], replace=True))
out = AudioTable(casout['name'])
out.set_connection(connection=conn)
out.feature_size = n_ceps
out.num_of_frames_col = '_num_frames_'
if label_level:
out.label_col = '_label_'
else:
out.label_col = None
return out
def extract_audio_features(cls, conn, table, frame_shift=10, frame_length=25, n_bins=40, n_ceps=40,
feature_scaling_method='STANDARDIZATION', n_output_frames=500, casout=None, **kwargs):
'''
Extracts audio features from the audio files
Parameters
----------
conn : CAS
A connection object to the current session.
table : AudioTable
An audio table containing the audio files.
frame_shift : int, optional
Specifies the time difference (in milliseconds) between the beginnings of consecutive frames.
Default: 10
frame_length : int, optional
Specifies the length of a frame (in milliseconds).
Default: 25
n_bins : int, optional
Specifies the number of triangular mel-frequency bins.
Default: 40
n_ceps : int, optional
Specifies the number of cepstral coefficients in each MFCC feature frame (including C0).
Default: 40
feature_scaling_method : string, optional
Specifies the feature scaling method to apply to the computed feature vectors.
Default: 'standardization'
n_output_frames : int, optional
Specifies the exact number of frames to include in the output table (extra frames are dropped and missing
frames are padded with zeros).
Default: 500
casout : dict or string or CASTable, optional
CAS Output table
kwargs : keyword-arguments, optional
Additional parameter for feature extraction.
Returns
-------
:class:`AudioTable`
If table exists
None
If no table exists
'''
if isinstance(table, AudioTable) or isinstance(table, CASTable):
if casout is None:
casout = dict(name=random_name('AudioTable', 6))
elif isinstance(casout, CASTable) or isinstance(casout, AudioTable):
casout = casout.to_outtable_params()
rt = conn.retrieve('audio.computefeatures', _messagelevel='error', table=table,
frameExtractionOptions=dict(frameshift=frame_shift, framelength=frame_length),
melBanksOptions=dict(nbins=n_bins), mfccOptions=dict(nceps=n_ceps),
featureScalingMethod=feature_scaling_method, nOutputFrames=n_output_frames,
casout=casout, **kwargs)
if rt.severity > 1:
for msg in rt.messages:
print(msg)
return None
server_type = get_cas_host_type(conn).lower()
if server_type.startswith("lin") or server_type.startswith("osx"):
fs = "/"
else:
fs = "\\"
scode = "i=find(_path_,'{0}',-length(_path_)); ".format(fs)
scode += "length _fName_ $1000; "
scode += "_fName_=substr(_path_, i+length('{0}'), length(_path_)-i);".format(fs)
ctable = CASTable(casout['name'], computedvars=['_fName_'],
computedvarsprogram=scode)
conn.table.partition(table=ctable, casout=dict(name=casout['name'], replace=True))
return AudioTable(casout['name'])
return None
