Module music_df.find

Provides a function for searching salami-slices.

Functions

def all_vals_in_series(vals: Iterable[Any], series: pandas.Series) ‑> bool

Expand source code

def all_vals_in_series(vals: Iterable[Any], series: pd.Series) -> bool:
    # I did a little bit of testing with the functions above and this option seemed
    #   to be the fastest. However my test data was not very realistic. (In particular,
    #   the series were pretty long, whereas in my use case they'll probably rarely
    #   be more than 10 items.)
    unique_vals = series.unique()
    return all(v in unique_vals for v in vals)

def find_simultaneous_feature(music_df: pandas.DataFrame, feature_name: str, values: Iterable[Any]) ‑> list[tuple[float, float]]

Expand source code

def find_simultaneous_feature(
    music_df: pd.DataFrame, feature_name: str, values: Iterable[Any]
) -> list[tuple[float, float]]:
    """
    Find salami-slices that contain all of the specified values.

    For example, if feature_name is "scale_degree", and values contains a list of
    degrees like ["1", "#1"], then this function will return the onset and release times
    of any salami-slices that contain both "1" and "#1" (they may contain other
    degrees as well).

    The dataframe is assumed to be salami-sliced.

    Args:
        music_df: dataframe.
        feature_name: name of the feature to search for.
        values: list of values to search for.

    >>> df = pd.DataFrame(
    ...     {
    ...         "pitch": [60, 67, 61, 72],
    ...         "onset": [0.0, 0, 4, 4],
    ...         "release": [4.0, 4, 8, 8],
    ...         "scale_degree": ["1", "5", "#1", "1"],
    ...     }
    ... )
    >>> find_simultaneous_feature(df, "scale_degree", ["1", "#1"])
    [(4.0, 8.0)]
    >>> find_simultaneous_feature(df, "scale_degree", ["1", "7"])
    []
    """
    assert "release" in music_df.columns
    assert "onset" in music_df.columns
    out = []
    for time_bounds, simultaneity in music_df.groupby(["onset", "release"]):
        if all_vals_in_series(values, simultaneity[feature_name]):
            out.append(time_bounds)
    return out

Find salami-slices that contain all of the specified values.

For example, if feature_name is "scale_degree", and values contains a list of degrees like ["1", "#1"], then this function will return the onset and release times of any salami-slices that contain both "1" and "#1" (they may contain other degrees as well).

The dataframe is assumed to be salami-sliced.

Args

music_df

dataframe.

feature_name

name of the feature to search for.

values

list of values to search for.

>>> df = pd.DataFrame(
...     {
...         "pitch": [60, 67, 61, 72],
...         "onset": [0.0, 0, 4, 4],
...         "release": [4.0, 4, 8, 8],
...         "scale_degree": ["1", "5", "#1", "1"],
...     }
... )
>>> find_simultaneous_feature(df, "scale_degree", ["1", "#1"])
[(4.0, 8.0)]
>>> find_simultaneous_feature(df, "scale_degree", ["1", "7"])
[]