daft.functions.resample#

def resample(
    file_expr: Expression,
    sample_rate: int,
) -> Expression:
    """Resample a audio file.

    Args:
        file_expr (AudioFile Expression): The audio file to resample.
        sample_rate (int): The sample rate to resample to.

    Returns:
        Expression (Tensor[Python] Expression): The resampled audio file.

    Example:
        >>> import daft
        >>> from daft.functions import audio_file, resample

        >>> # Discover the audio files (returns a dataframe with ['path', 'size'] columns)
        >>> df = daft.from_glob_path("hf://datasets/Eventual-Inc/sample-files/audio/*.mp3")

        >>> df = (
        ...     df.with_column("file", audio_file(daft.col("path")))
        ...     .with_column("resampled", resample(daft.col("file"), sample_rate=16000))
        ...     .select("path", "size", "resampled")
        ... )
        >>> df.show(3)
        ╭────────────────────────────────┬─────────┬──────────────────────────╮
        │ path                           ┆ size    ┆ resampled                │
        │ ---                            ┆ ---     ┆ ---                      │
        │ String                         ┆ Int64   ┆ Tensor[Float64]          │
        ╞════════════════════════════════╪═════════╪══════════════════════════╡
        │ hf://datasets/Eventual-Inc/sa… ┆ 822924  ┆ <Tensor shape=(2490278)> │
        ├╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┤
        │ hf://datasets/Eventual-Inc/sa… ┆ 618408  ┆ <Tensor shape=(2207923)> │
        ├╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┤
        │ hf://datasets/Eventual-Inc/sa… ┆ 1190736 ┆ <Tensor shape=(3341800)> │
        ╰────────────────────────────────┴─────────┴──────────────────────────╯
        <BLANKLINE>
        (Showing first 3 rows)
    """
    # Check for required dependencies at expression definition time
    # This gives users a helpful error immediately, rather than during execution
    from daft.dependencies import librosa, sf

    if not sf.module_available():
        raise ImportError(
            "The 'soundfile' module is required to resample audio files. "
            "Please install it with: pip install 'daft[audio]'"
        )
    if not librosa.module_available():
        raise ImportError(
            "The 'librosa' module is required to resample audio files. "
            "Please install it with: pip install 'daft[audio]'"
        )

    return resample_fn(file_expr, sample_rate)