[Bug]: `CurveFit.fit()` accepts incorrectly structured initial guesses and bounds

[sepandhaghighi]

Contact details

me@sepand.tech

What happened?

CurveFit.fit() currently accepts initial_guess and bounds as sequences. The values are matched to free model parameters based on their declaration order.

This makes the API error-prone because users must know the parameter ordering, and incorrectly sized sequences are not explicitly validated before being passed to scipy.optimize.curve_fit().

Steps to reproduce

1. Create a CurveFit instance with a model containing multiple free parameters:

fit = CurveFit(
    model_name="first_order",
    time=[0, 1, 2, 3],
    release_profile=[0, 10, 18, 25],
)

2. Provide an incorrectly sized initial guess:

fit.fit(initial_guess=[1.0])

3. Alternatively, provide incorrectly sized bounds:

fit.fit(
    bounds=(
        [1e-10],
        [np.inf],
    )
)

4. Observe how the invalid sequences are handled by CurveFit.fit() and ultimately scipy.optimize.curve_fit().

Expected behavior

CurveFit.fit() should use parameter-name mappings for initial_guess and bounds, for example:

fit.fit(
    initial_guess={
        "M0": 100.0,
        "k": 0.01,
    },
    bounds={
        "M0": (1e-10, np.inf),
        "k": (1e-10, np.inf),
    },
)

The method should validate the supplied parameter names and values before calling scipy.optimize.curve_fit().

Invalid, missing, or unknown parameter entries should result in a clear ValueError.

Actual behavior

CurveFit.fit() currently relies on positional sequences:

initial_guess=[1.0, 0.01]
bounds=([1e-10, 1e-10], [np.inf, np.inf])

The API does not identify parameters by name, so callers must know the internal parameter declaration order. Incorrectly sized sequences are not explicitly validated by CurveFit.fit() and may result in errors from the underlying SciPy optimizer.

Operating system

Linux

Python version

Python 3.11

Drux version

Drux 0.4

Relevant log output