derivative (pmrf.derivative)

pmrf.derivative(eval_fn: Callable[[...], Any], *args: Unpack[Ts]) → Tuple[Unpack[Ts]]

Computes the exact derivative of a function.

Dynamically routes to the most relevant autodiff method: - grad (reverse-mode) for scalar outputs. - jacfwd (forward-mode) for wide Jacobians (output size > input size). - jacrev (reverse-mode) for tall Jacobians (input size >= output size).

Safely handles models as inputs by filtering out non-differentiable static fields (like strings, booleans, integers, AND NumPy arrays).

Parameters:

  • eval_fn (Callable) – The function to differentiate.

  • *args (*Ts) – The arguments to evaluate the derivative at. Can be raw arrays or full Models.

Returns:

A tuple containing the derivatives. The tuple length and contents will exactly mirror the types and structure of the input args.

Return type:

Tuple[*Ts]

Examples

Compute the sensitivity of a component’s response with respect to its parameters:

>>> import pmrf as prf
>>> from pmrf.models import ShuntCapacitor
>>>
>>> freq = prf.Frequency(2.4, 2.4, 1, 'GHz')
>>> cap = ShuntCapacitor(C=prf.Unconstrained(1.0e-12), name='c1')
>>>
>>> def eval_s21(model):
...     return model.s_mag(freq)[0, 1, 0]
...
>>> (d_cap,) = derivative(eval_s21, cap)
>>>
>>> # The structural layout of the model is preserved in the derivative
>>> print(f"{d_cap.at('c1.C').get():.3e}")
-1.060e-01