asteroid.utils package¶
-
asteroid.utils.
prepare_parser_from_dict
(dic, parser=None)[source]¶ Prepare an argparser from a dictionary.
Parameters: - dic (dict) – Two-level config dictionary with unique bottom-level keys.
- parser (argparse.ArgumentParser, optional) – If a parser already exists, add the keys from the dictionary on the top of it.
Returns: argparse.ArgumentParser – Parser instance with groups corresponding to the first level keys and arguments corresponding to the second level keys with default values given by the values.
-
asteroid.utils.
parse_args_as_dict
(parser, return_plain_args=False, args=None)[source]¶ Get a dict of dicts out of process parser.parse_args()
Top-level keys corresponding to groups and bottom-level keys corresponding to arguments. Under ‘main_args’, the arguments which don’t belong to a argparse group (i.e main arguments defined before parsing from a dict) can be found.
Parameters: - parser (argparse.ArgumentParser) – ArgumentParser instance containing groups. Output of prepare_parser_from_dict.
- return_plain_args (bool) – Whether to return the output or parser.parse_args().
- args (list) – List of arguments as read from the command line. Used for unit testing.
Returns: dict – Dictionary of dictionaries containing the arguments. Optionally the direct output parser.parse_args().
-
asteroid.utils.
str_int_float
(value)[source]¶ Type to convert strings to int, float (in this order) if possible.
Parameters: value (str) – Value to convert. Returns: int, float, str – Converted value.
-
asteroid.utils.
str2bool
(value)[source]¶ Type to convert strings to Boolean (returns input if not boolean)
-
asteroid.utils.
isfloat
(value)[source]¶ Computes whether value can be cast to a float.
Parameters: value (str) – Value to check. Returns: bool – Whether value can be cast to a float.
-
asteroid.utils.
isint
(value)[source]¶ Computes whether value can be cast to an int
Parameters: value (str) – Value to check. Returns: bool – Whether value can be cast to an int.
-
asteroid.utils.
tensors_to_device
(tensors, device)[source]¶ Transfer tensor, dict or list of tensors to device.
Parameters: - tensors (
torch.Tensor
) – May be a single, a list or a dictionary of tensors. - ( (device) – class: torch.device): the device where to place the tensors.
Returns: Union [
torch.Tensor
, list, tuple, dict] – Same as input but transferred to device. Goes through lists and dicts and transfers the torch.Tensor to device. Leaves the rest untouched.- tensors (
-
asteroid.utils.
to_cuda
(tensors)[source]¶ Transfer tensor, dict or list of tensors to GPU.
Parameters: tensors ( torch.Tensor
, list or dict) – May be a single, a list or a dictionary of tensors.Returns: torch.Tensor
– Same as input but transferred to cuda. Goes through lists and dicts and transfers the torch.Tensor to cuda. Leaves the rest untouched.
-
asteroid.utils.
has_arg
(fn, name)[source]¶ Checks if a callable accepts a given keyword argument.
Parameters: - fn (callable) – Callable to inspect.
- name (str) – Check if fn can be called with name as a keyword argument.
Returns: bool – whether fn accepts a name keyword argument.
-
asteroid.utils.
flatten_dict
(d, parent_key='', sep='_')[source]¶ Flattens a dictionary into a single-level dictionary while preserving parent keys. Taken from https://stackoverflow.com/questions/6027558/ flatten-nested-dictionaries-compressing-keys?answertab=votes#tab-top
Parameters: Returns: dict – Single-level dictionary, flattened.
-
asteroid.utils.
average_arrays_in_dic
(dic)[source]¶ Take average of numpy arrays in a dictionary.
Parameters: dic (dict) – Input dictionary to take average from Returns: dict – New dictionary with array averaged.
-
asteroid.utils.
get_wav_random_start_stop
(signal_len, desired_len=32000)[source]¶ Get indexes for a chunk of signal of a given length.
Parameters: Returns: tuple – random start integer, stop integer.
-
asteroid.utils.
unet_decoder_args
(encoders, *, skip_connections)[source]¶ Get list of decoder arguments for upsampling (right) side of a symmetric u-net, given the arguments used to construct the encoder.
Parameters: - encoders (list of length N of tuples of (in_chan, out_chan, kernel_size, stride, padding)) – List of arguments used to construct the encoders
- skip_connections (bool) – Whether to include skip connections in the calculation of decoder input channels.
Returns: list of length N of tuples of (in_chan, out_chan, kernel_size, stride, padding) – Arguments to be used to construct decoders