musicaiz.features.compute_ssm¶

musicaiz.features.compute_ssm(file: Union[musicaiz.loaders.Musa, str, TextIO, pathlib.Path], ssm_type: str = 'louie', measures: Optional[Union[List[str], str]] = None, is_binary: bool = True, threshold: float = 0.5) → numpy.ndarray[source]¶

Computes the selected SSM.

Parameters

file: Union[Musa, str, TextIO, Path]

A file or a musicaiz Musa() object.

ssm_type: str

Valid values: louie, measure , measures.

louie: Computes the SSM with Louie’s method self_similarity_louie()
measure: Computes the SSM with one of the _DEFAULT_MEASURES measures

self_similarity_single_measure(). Note that if this ssm_type is selected, the measures argument must be a valid measure str.

measures: Computes the SSM with more than one of the _DEFAULT_MEASURES

measures self_similarity_measures(). Note that if this ssm_type is selected, the measures argument must be a list of valid measure strs or all.

measures: Union[List[str], str]

The measure or list of measures to compute the SSM with. It also accepts all as a value if ssm_type is set to measures.

is_binary: bool = True

Binarizes the SSM.

threshold: int = 0.5

The threshold value used to binarize the SSSM.

Returns

s: np.ndarray [shape=(n_bars, n_bars)]: The SSM computed from the feature_vector or measure(s).

Examples

>>> from pathlib import Path
>>> file = Path("tests/fixtures/tokenizers/mmm_tokens.mid")
>>> ssm = musicaiz.features.compute_ssm(
    file=file,
    ssm_type="measures",
    measures="all"
)