Source code for growthcurves.fit

"""Entry point of growthcurves package"""

import warnings

import numpy as np

from . import inference, models, non_parametric, parametric
from .models import get_all_models

__all__ = [
    "fit_model",
]




def fit_model(
    t: np.ndarray,
    N: np.ndarray,
    model_name: str,
    lag_threshold: float = 0.15,
    exp_threshold: float = 0.15,
    phase_boundary_method=None,
    **kwargs,
) -> tuple[dict, dict]:
    if model_name in models.MODEL_REGISTRY["non_parametric"]:
        fit_res = non_parametric.fit_non_parametric(t, N, method=model_name, **kwargs)
    else:
        fit_res = parametric.fit_parametric(t, N, method=model_name, **kwargs)
    # return None if fit fails, along with bad fit stats
    if fit_res is None:
        warnings.warn(
            f"Model fitting failed for model {model_name}. Returning None.",
            stacklevel=2,
        )
        return None, inference.bad_fit_stats()

    stats_res = inference.extract_stats(
        fit_res,
        t=t,
        N=N,
        lag_threshold=lag_threshold,
        exp_threshold=exp_threshold,
        phase_boundary_method=phase_boundary_method,
        **kwargs,
    )

    stats_res["model_name"] = model_name
    return fit_res, stats_res



# ! not so good for dynamic inspection tools...
fit_model.__doc__ = f"""Fit a growth model to the provided t and N.

    Parameters
    ----------
    t : np.ndarray
        Time points corresponding to N (in hours).
    N : np.ndarray
        Growth data points corresponding to t.
    model_name : str
        One of the models in {', '.join(get_all_models())}.
    lag_threshold : float, optional
        Fraction of μ_max to define end of lag phase (threshold method, default: 0.15).
    exp_threshold : float, optional
        Fraction of μ_max to define end of exponential phase (threshold method,
        default: 0.15).
    phase_boundary_method : str, optional
        Method to determine phase boundaries ("tangent", "threshold",
        or None for default for model class).
    **kwargs
        Additiona keyword arguments to be passed to fitting and inference functions.

    Returns
    -------
    tuple[dict, dict]
        Return tuple of two dictionaries: (fit_res, stats_res)
        - fit_res: Dictionary containing fitted model parameters.
        - stats_res: Dictionary containing goodness-of-fit statistics and growth metrics

    """