Overview
--------
returns the cogent3 log-likelihood as an alignment quality score

Options for making the app
--------------------------
cogent3_score_app = get_app('cogent3_score', *args, **kwargs)

    Parameters
    ----------
    aln
An alignment instance.

    Returns
    -------
    returns the log-likelihood, or 0.0 if the alignment does not have the
    score

    Notes
    -----
    The cogent3 tree_align algorithm is a progressive-HMM. It records
    the log-likelihood of the alignment (in addition to all other
    the parameter values, including the guide tree) in
    ``alignment.info['align_params']``. This can be used as a measure of
    alignment quality.

    The instance must have been aligned using cogent3 tree_align. In addition,
    if the alignment has been saved, it has must have been serialised
    using a format that preserves the score.

    Examples
    --------

    Prepare an alignment of BRCA1 genes from 4 species. Create an unaligned
    sequence collection, guide tree, and an app for alignment using cogent3
    ``progressive_align()``.

    >>> from cogent3 import make_unaligned_seqs, get_app
    >>> aln = make_unaligned_seqs(
    ...     {
    ...         "Human": "GCCAGCTCATTACAGCATGAGAACAGCAGTTTATTACTCACT",
    ...         "Bandicoot": "NACTCATTAATGCTTGAAACCAGCAGTTTATTGTCCAAC",
    ...         "Rhesus": "GCCAGCTCATTACAGCATGAGAACAGTTTGTTACTCACT",
    ...         "FlyingFox": "GCCAGCTCTTTACAGCATGAGAACAGTTTATTATACACT",
    ...     },
    ...     moltype="dna",
    ... )
    >>> newick = "(Bandicoot:0.4,FlyingFox:0.05,(Rhesus:0.06,Human:0.0):0.04);"
    >>> aligner = get_app("progressive_align", model="HKY85")

    Create a composable app that aligns the sequences and returns the
    cogent3 log-likelihood alignment score.

    >>> scorer = get_app("cogent3_score")
    >>> app = aligner + score
    >>> result = app(aln)
    >>> print(result)
    -130.47375615734916

Input type
----------
Alignment, ArrayAlignment

Output type
-----------
float