Merge pull request #33 from statnett/reintroduce-invalid-value-in-bisect

feat: reintroduce invalid_value in bisect

linerate/solver.py

@@ -1,5 +1,5 @@
 from functools import partial
-from typing import Callable
+from typing import Callable, Optional
 
 import numpy as np
 
@@ -13,6 +13,7 @@ def bisect(
     xmin: FloatOrFloatArray,
     xmax: FloatOrFloatArray,
     tolerance: float,
+    invalid_value: Optional[float] = None,
 ) -> FloatOrFloatArray:
     r"""Compute the roots of a function using a vectorized bisection method.
 
@@ -31,6 +32,9 @@ def bisect(
         bounded within an interval of size :math:`\Delta x` or less. The bisection method will
         run for :math:`\left\lceil\frac{x_\max - x_\min}{\Delta x}\right\rceil`
         iterations.
+    invalid_value:
+        If provided, then this value is used whenever
+        :math:`\text{sign}(f(\mathbf{x}_\min)) = \text{sign}(f(\mathbf{x}_\max))`.
 
     Returns
     -------
@@ -47,10 +51,12 @@ def bisect(
     f_right = f(xmax)
 
     invalid_mask = np.sign(f_left) == np.sign(f_right)
-    if np.any(invalid_mask):
+    if np.any(invalid_mask) and invalid_value is None:
         raise ValueError(
             "f(xmin) and f(xmax) have the same sign. Consider increasing the search interval."
         )
+    elif isinstance(invalid_mask, bool) and invalid_mask:
+        return invalid_value  # type: ignore
 
     while interval > tolerance:
         xmid = 0.5 * (xmax + xmin)
@@ -63,7 +69,7 @@ def bisect(
         f_left = np.where(mask, f_mid, f_left)
         f_right = np.where(mask, f_right, f_mid)
 
-    out = 0.5 * (xmax + xmin)
+    out = np.where(invalid_mask, invalid_value, 0.5 * (xmax + xmin))  # type: ignore
     return out
 
 
@@ -112,6 +118,7 @@ def compute_conductor_ampacity(
     min_ampacity: Ampere = 0,
     max_ampacity: Ampere = 5_000,
     tolerance: float = 1,  # Ampere
+    invalid_value=None,
 ) -> Ampere:
     r"""Use the bisection method to compute the steady-state thermal rating (ampacity).
 
@@ -134,6 +141,10 @@ def compute_conductor_ampacity(
         bisection iterations will stop once the numerical ampacity uncertainty is below
         :math:`\Delta I`. The bisection method will run for
         :math:`\left\lceil\frac{I_\text{max} - I_\text{min}}{\Delta I}\right\rceil` iterations.
+    invalid_value:
+        if the optimization problem is invalid, this value is returned instead of an error.
+        Suggested value: 0 for 0-ampacity when max_conductor_temperature is exceeded for all
+        ampacities.
 
     Returns
     -------
@@ -142,4 +153,4 @@ def compute_conductor_ampacity(
     """
     f = partial(heat_balance, max_conductor_temperature)
 
-    return bisect(f, min_ampacity, max_ampacity, tolerance)
+    return bisect(f, min_ampacity, max_ampacity, tolerance, invalid_value=invalid_value)