def __init__(self, table, *additional_tables, **kwargs):
key_list = kwargs.pop('key_list', None)
rng = kwargs.pop('rng', None)
super(ShuffledData, self).__init__(table, *additional_tables, **kwargs)
try:
key_list = tuple(key_list)
except TypeError:
pass
if key_list is None:
_, rx_fn, rx_type, _ = parse_kaldi_input_path(
self.table_specifiers[0][0])
if rx_type == RxfilenameType.InvalidInput:
raise IOError('Invalid rspecifier {}'.format(rx_fn))
elif rx_type == RxfilenameType.StandardInput:
raise IOError(
'Cannot infer key list from stdin (cannot reopen)')
with io_open(*self.table_specifiers[0][:2]) as reader:
self.key_list = tuple(reader.keys())
else:
self.key_list = tuple(key_list)
if self.ignore_missing:
self._num_samples = None
else:
self._num_samples = len(self.key_list)
if isinstance(rng, np.random.RandomState):
self.rng = rng
else:
self.rng = np.random.RandomState(rng)
self.table_handles = tuple(
io_open(rspecifier, kdtype, mode='r+', **o_kwargs)
for rspecifier, kdtype, o_kwargs in self.table_specifiers
)
def __init__(self, table, *additional_tables, **kwargs):
key_list = kwargs.pop('key_list', None)
rng = kwargs.pop('rng', None)
super(ShuffledData, self).__init__(table, *additional_tables, **kwargs)
try:
key_list = tuple(key_list)
except TypeError:
pass
if key_list is None:
_, rx_fn, rx_type, _ = parse_kaldi_input_path(
self.table_specifiers[0][0])
if rx_type == RxfilenameType.InvalidInput:
raise IOError('Invalid rspecifier {}'.format(rx_fn))
elif rx_type == RxfilenameType.StandardInput:
raise IOError(
'Cannot infer key list from stdin (cannot reopen)')
with io_open(*self.table_specifiers[0][:2]) as reader:
self.key_list = tuple(reader.keys())
else:
self.key_list = tuple(key_list)
if self.ignore_missing:
self._num_samples = None
else:
self._num_samples = len(self.key_list)
if isinstance(rng, np.random.RandomState):
self.rng = rng
else:
self.rng = np.random.RandomState(rng)
self.table_handles = tuple(
io_open(rspecifier, kdtype, mode='r+', **o_kwargs)
for rspecifier, kdtype, o_kwargs in self.table_specifiers)
def kaldi_rxfilename_arg_type(string):
'''argument type to make sure string is a valid extended readable file'''
table_type, _, rxfilename_type, _ = kaldi_io_util.parse_kaldi_input_path(
string)
if table_type != kaldi_io_enums.TableType.NotATable:
raise argparse.ArgumentTypeError(
'Expected an extended file name, got an rspecifier (starts with '
"'ark:' or 'scp:')")
elif rxfilename_type == kaldi_io_enums.RxfilenameType.InvalidInput:
raise argparse.ArgumentTypeError('Not a valid rxfilename')
return string
def kaldi_rxfilename_arg_type(string):
'''argument type to make sure string is a valid extended readable file'''
table_type, _, rxfilename_type, _ = kaldi_io_util.parse_kaldi_input_path(
string)
if table_type != kaldi_io_enums.TableType.NotATable:
raise argparse.ArgumentTypeError(
'Expected an extended file name, got an rspecifier (starts with '
"'ark:' or 'scp:')")
elif rxfilename_type == kaldi_io_enums.RxfilenameType.InvalidInput:
raise argparse.ArgumentTypeError('Not a valid rxfilename')
return string
def __init__(self, path):
self.path = path
self.closed = False
if self.readable():
self._table_type, self._xfilenames, self._xtypes, options = \
parse_kaldi_input_path(path)
else:
self._table_type, self._xfilenames, self._xtypes, options = \
parse_kaldi_output_path(path)
self.binary = True
for key, value in options.items():
setattr(self, key, value)
super(KaldiIOBase, self).__init__()
def __init__(self, path):
from pydrobert.kaldi.io.util import parse_kaldi_input_path
from pydrobert.kaldi.io.util import parse_kaldi_output_path
self.path = path
self.closed = False
if self.readable():
self._table_type, self._xfilenames, self._xtypes, options = \
parse_kaldi_input_path(path)
else:
self._table_type, self._xfilenames, self._xtypes, options = \
parse_kaldi_output_path(path)
self.binary = True
for key, value in options.items():
setattr(self, key, value)
super(KaldiIOBase, self).__init__()
def __init__(self, table, *additional_tables, **kwargs):
super(SequentialData, self).__init__(table, *additional_tables,
**kwargs)
self._num_samples = None
sorteds = tuple(
parse_kaldi_input_path(spec[0])[3]['sorted']
for spec in self.table_specifiers)
if not all(sorteds):
uns_rspec = self.table_specifiers[sorteds.index(False)][0]
uns_rspec_split = uns_rspec.split(':')
uns_rspec_split[0] += ',s'
sor_rspec = ':'.join(uns_rspec_split)
warn('SequentialData assumes data are sorted, and "{}" does '
'not promise to be sorted. To supress this warning, '
'check that this table is sorted, then add the sorted '
'flag to this rspecifier ("{}")'.format(uns_rspec, sor_rspec))
if self.ignore_missing and len(self.table_specifiers) > 1:
self._sample_generator_for_epoch = self._ignore_epoch
else:
self._sample_generator_for_epoch = self._no_ignore_epoch
def open(path,
kaldi_dtype=None,
mode='r',
error_on_str=True,
utt2spk='',
value_style='b',
header=True,
cache=False):
"""Factory function for initializing and opening kaldi streams
This function provides a general interface for opening kaldi
streams. Kaldi streams are either simple input/output of kaldi
objects (the basic stream) or key-value readers and writers
(tables).
When `path` starts with ``ark:`` or ``scp:`` (possibly with
modifiers before the colon), a table is opened. Otherwise, a basic
stream is opened.
See also
--------
pydrobert.kaldi.io.table_streams.open_table_stream
For information on opening tables
pydrobert.kaldi.io.basic.open_duck_stream
For information on opening basic streams
"""
if 'r' in mode:
table_type = parse_kaldi_input_path(path)[0]
else:
table_type = parse_kaldi_output_path(path)[0]
if table_type == TableType.NotATable:
return open_duck_stream(path, mode=mode, header=header)
else:
return open_table_stream(path,
kaldi_dtype,
mode=mode,
error_on_str=error_on_str,
utt2spk=utt2spk,
value_style=value_style,
cache=cache)
def __init__(self, table, *additional_tables, **kwargs):
super(SequentialData, self).__init__(
table, *additional_tables, **kwargs)
self._num_samples = None
sorteds = tuple(
parse_kaldi_input_path(spec[0])[3]['sorted']
for spec in self.table_specifiers
)
if not all(sorteds):
uns_rspec = self.table_specifiers[sorteds.index(False)][0]
uns_rspec_split = uns_rspec.split(':')
uns_rspec_split[0] += ',s'
sor_rspec = ':'.join(uns_rspec_split)
warn(
'SequentialData assumes data are sorted, and "{}" does '
'not promise to be sorted. To supress this warning, '
'check that this table is sorted, then add the sorted '
'flag to this rspecifier ("{}")'.format(
uns_rspec, sor_rspec))
if self.ignore_missing and len(self.table_specifiers) > 1:
self._sample_generator_for_epoch = self._ignore_epoch
else:
self._sample_generator_for_epoch = self._no_ignore_epoch
def open(
path, kaldi_dtype=None, mode='r', error_on_str=True,
utt2spk='', value_style='b', header=True, cache=False):
"""Factory function for initializing and opening kaldi streams
This function provides a general interface for opening kaldi
streams. Kaldi streams are either simple input/output of kaldi
objects (the basic stream) or key-value readers and writers
(tables).
When `path` starts with ``'ark:'`` or ``'scp:'`` (possibly with
modifiers before the colon), a table is opened. Otherwise, a basic
stream is opened.
See also
--------
pydrobert.kaldi.io.table_streams.open_table_stream
For information on opening tables
pydrobert.kaldi.io.basic.open_duck_stream
For information on opening basic streams
"""
from pydrobert.kaldi.io.enums import TableType
from pydrobert.kaldi.io.util import parse_kaldi_input_path
from pydrobert.kaldi.io.util import parse_kaldi_output_path
from pydrobert.kaldi.io.duck_streams import open_duck_stream
from pydrobert.kaldi.io.table_streams import open_table_stream
if 'r' in mode:
table_type = parse_kaldi_input_path(path)[0]
else:
table_type = parse_kaldi_output_path(path)[0]
if table_type == TableType.NotATable:
return open_duck_stream(path, mode=mode, header=header)
else:
return open_table_stream(
path, kaldi_dtype, mode=mode, error_on_str=error_on_str,
utt2spk=utt2spk, value_style=value_style, cache=cache)
def kaldi_rspecifier_arg_type(string):
'''argument type to make sure string is a valid rspecifier'''
table_type, _, _, _ = kaldi_io_util.parse_kaldi_input_path(string)
if table_type == kaldi_io_enums.TableType.NotATable:
raise argparse.ArgumentTypeError('Not a valid rspecifier')
return string
def kaldi_rspecifier_arg_type(string):
'''argument type to make sure string is a valid rspecifier'''
table_type, _, _, _ = kaldi_io_util.parse_kaldi_input_path(string)
if table_type == kaldi_io_enums.TableType.NotATable:
raise argparse.ArgumentTypeError('Not a valid rspecifier')
return string
