data_saving_util

save_data_by_format(file_format, save_path, file_name_list, file_content_list, group_ids=None, sample_rate=None)

Save data in the specified format to disk.

Parameters:

Name	Type	Description	Default
file_format	str	The format of the output files. It can be one of 'npy', 'npz', 'wav' or 'flac'.	required
save_path	str	The directory where the files will be saved.	required
file_name_list	List[str]	A list of strings with the names of the files to be saved.	required
file_content_list	List	A list with the content of the files to be saved.	required
group_ids	List[str] or str	A list of strings with the group ids of the files. If provided, it will be used to create a subdirectory for each group of files inside the save_path. Defaults to None.	None
sample_rate	int	The sample rate of the audio files, required for 'wav' and 'flac' formats. Defaults to None.	None

Returns:

Type	Description
	Dict[str, str]: A dictionary that maps the original file names to their corresponding file paths in disk.

Raises:

Type	Description
NotImplementedError	If the file format is not supported.

Notes

• The content of the files can be of any type as long as it can be converted to numpy arrays (using the to_cpu() function if necessary).
• If sample_rate is not None, it will be saved along with the data for 'npz', 'wav' and 'flac' formats.

Source code in speechain/utilbox/data_saving_util.py

def save_data_by_format(
    file_format: str,
    save_path: str,
    file_name_list: Union[List[str] or str],
    file_content_list: Union[
        List[Union[np.ndarray, torch.Tensor]], np.ndarray, torch.Tensor
    ],
    group_ids: List[str] or str = None,
    sample_rate: int = None,
):
    """Save data in the specified format to disk.

    Args:
        file_format (str):
            The format of the output files. It can be one of 'npy', 'npz', 'wav' or 'flac'.
        save_path (str):
            The directory where the files will be saved.
        file_name_list (List[str]):
            A list of strings with the names of the files to be saved.
        file_content_list (List):
            A list with the content of the files to be saved.
        group_ids (List[str] or str, optional):
            A list of strings with the group ids of the files. If provided, it will be used to create a subdirectory
            for each group of files inside the save_path. Defaults to None.
        sample_rate (int, optional):
            The sample rate of the audio files, required for 'wav' and 'flac' formats. Defaults to None.

    Returns:
        Dict[str, str]:
            A dictionary that maps the original file names to their corresponding file paths in disk.

    Raises:
        NotImplementedError:
            If the file format is not supported.

    Notes:
        - The content of the files can be of any type as long as it can be converted to numpy arrays (using the
          to_cpu() function if necessary).
        - If sample_rate is not None, it will be saved along with the data for 'npz', 'wav' and 'flac' formats.
    """

    # record all the data file paths with their names
    name2file_path = {}

    if not isinstance(file_name_list, List):
        file_name_list = [file_name_list]
    if not isinstance(file_content_list, List):
        file_content_list = [file_content_list]
    assert len(file_name_list) == len(file_content_list)

    if isinstance(group_ids, str):
        group_ids = [group_ids for _ in file_name_list]

    # loop over each file name and content pair
    for i, (name, content) in enumerate(zip(file_name_list, file_content_list)):
        # convert content to numpy array if it is a PyTorch tensor
        if isinstance(content, torch.Tensor):
            content = to_cpu(content, tgt="numpy")

        # if group_ids is not None, create subdirectory for this file's group
        if group_ids is not None:
            file_save_path = os.path.join(save_path, group_ids[i])
        else:
            file_save_path = save_path

        # make sure the saving folder exists
        os.makedirs(file_save_path, exist_ok=True)
        file_path = os.path.join(file_save_path, f"{name}.{file_format}")
        # check the existence of the target file to make sure that data saving won't fail due to the system errors that cannot be captured by Python
        while not os.path.exists(file_path):
            # save the file in the specified format
            if file_format == "npy":
                np.save(file_path, content.astype(np.float32))

            elif file_format == "npz":
                assert (
                    sample_rate is not None
                ), "sample_rate must be provided for npz format"
                np.savez(
                    file_path, feat=content.astype(np.float32), sample_rate=sample_rate
                )

            elif file_format == "wav":
                assert (
                    sample_rate is not None
                ), "sample_rate must be provided for wav format"
                sf.write(
                    file=file_path,
                    data=content,
                    samplerate=sample_rate,
                    format="WAV",
                    subtype=sf.default_subtype("WAV"),
                )

            elif file_format == "flac":
                assert (
                    sample_rate is not None
                ), "sample_rate must be provided for flac format"
                sf.write(
                    file=file_path,
                    data=content,
                    samplerate=sample_rate,
                    format="FLAC",
                    subtype=sf.default_subtype("FLAC"),
                )

            else:
                raise NotImplementedError(
                    f"File format '{file_format}' is not supported. "
                    f"Please use one of the supported formats: 'npy', 'npz', 'wav', 'flac'"
                )

        # map the original file names to their corresponding file paths in disk
        name2file_path[name] = file_path

    return name2file_path