chore: added progrss bar to DE optimiser

FoldOptLib/optimisers/axial_surface_optimiser.py

@@ -122,7 +122,7 @@ def generate_bounds_and_initial_guess(self):
         The initial guess is generated using the Von Mises Fisher distribution.
 
         """
-        self.bounds = [(0, 360), (0, 90)]
+        self.bounds = [(0, numpy.pi*2), (0, numpy.pi/2)]
 
         if self.geological_knowledge is not None:
 
@@ -134,6 +134,7 @@ def generate_bounds_and_initial_guess(self):
                 # Sample from the distribution
                 initial_guess = vmf.draw_samples(size=20, random_state=180564327)
                 initial_guess = normal_vector_to_strike_and_dip(initial_guess)
+                initial_guess = numpy.deg2rad(initial_guess)
                 return initial_guess
 
         if self.geological_knowledge is None:
@@ -218,7 +219,6 @@ def build_optimisation_function(
         knowledge_function: GeologicalKnowledgeFunctions = None,
     ):
         def optimisation_function(strike_dip):
-            print(strike_dip)
             # Convert the strike-dip to a unit vector
             unit_vector = strike_dip_to_vector(strike_dip[0], strike_dip[1])
             # Normalize the unit vector
@@ -229,7 +229,6 @@ def optimisation_function(strike_dip):
             angle_difference = ObjectiveFunction[ObjectiveType.ANGLE](
                 predicted_foliation, self.gradient_data
             )
-            print("angle difference: ",angle_difference)
 
             # If the optimisation type is angle, return the angle difference
             if self.optimisation_type == OptimisationType.ANGLE:

FoldOptLib/solvers/solvers.py

@@ -1,6 +1,6 @@
 from typing import Callable, Dict, Any, Tuple, Union, List
 import numpy 
-from ..utils.utils import *
+from ..utils.utils import create_progress_bar
 from scipy.optimize import minimize, differential_evolution
 from ..datatypes import SolverType
 import beartype
@@ -32,9 +32,10 @@ def differential_evolution(
         init: Union[str, numpy.ndarray] = "halton",
         maxiter: int = 5000,
         seed: int = 80,
-        polish: bool = True,
+        polish: bool = False,
         strategy: str = "best2exp",
         mutation: Tuple[float, float] = (0.3, 0.99),
+        callback=create_progress_bar(5000),
         **kwargs,
     ) -> Dict:
         """
@@ -77,6 +78,7 @@ def differential_evolution(
             polish=polish,
             strategy=strategy,
             mutation=mutation,
+            callback=callback,
             **kwargs,
         )
 
@@ -85,7 +87,11 @@ def differential_evolution(
     @beartype.beartype
     @staticmethod
     def constrained_trust_region(
-        objective_function: Callable, x0: numpy.ndarray, constraints=None, **kwargs
+        objective_function: Callable, 
+        x0: numpy.ndarray, 
+        constraints=None, 
+        callback=create_progress_bar(5000),
+        **kwargs
     ) -> Dict:
         """
         Solves the optimisation problem using the trust region method.
@@ -111,6 +117,7 @@ def constrained_trust_region(
             method="trust-constr",
             jac="2-point",
             constraints=constraints,
+            callback=callback,
             **kwargs,
         )
 

FoldOptLib/utils/utils.py

@@ -3,6 +3,7 @@
 from ..from_loopstructural._svariogram import SVariogram
 import mplstereonet
 import dill
+from tqdm import tqdm
 
 
 def calculate_semivariogram(fold_frame, fold_rotation, lag=None, nlag=None):
@@ -117,8 +118,8 @@ def strike_dip_to_vector(strike, dip):
         raise TypeError(f"Expected dip to be a number, got {type(dip).__name__}")
 
     # Convert degrees to radians
-    s_r = numpy.deg2rad(strike)
-    d_r = numpy.deg2rad(dip)
+    s_r = strike
+    d_r = dip
 
     # Calculate the components of the strike-dip vector
     nx = numpy.sin(d_r) * numpy.cos(s_r)
@@ -459,3 +460,18 @@ def objective_function(x):
         return func1(x) + func2(x)
 
     return objective_function
+
+def create_progress_bar(max_iter):
+    pbar = tqdm(dynamic_ncols=True, desc="Optimisation Progress")
+
+    # def callback(xk, convergence):
+    def callback(xk, convergence):
+        # pbar.set_description(f"Strike/Dip: {numpy.rad2deg(xk):.4f}")
+        pbar.update(1)
+        if convergence < 1e-6:
+            pbar.close()
+            return True
+        return False
+
+    return callback
+