0.24.17

Author: bartzbeielstein

pyproject.toml

@@ -7,7 +7,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "spotpython"
-version = "0.24.16"
+version = "0.24.17"
 authors = [
   { name="T. Bartz-Beielstein", email="tbb@bartzundbartz.de" }
 ]

src/spotpython/spot/spot.py

@@ -233,7 +233,7 @@ def __init__(
 
         # The writer (Tensorboard) must be initialized before the surrogate model,
         # because the writer is passed to the surrogate model:
-        self.init_spot_writer()
+        self._init_spot_writer()
 
         self._surrogate_setup(surrogate)
 
@@ -862,6 +862,8 @@ def run(self, X_start: np.ndarray = None) -> Spot:
             self.initialize_design(X_start)
             self.update_stats()
             self.fit_surrogate()
+            if self.verbosity > 0:
+                print("\n------ Starting optimization on the Surrogate for the given Budget ------\n")
             timeout_start = time.time()
             while self.should_continue(timeout_start):
                 self.update_design()
@@ -1083,7 +1085,7 @@ def evaluate_initial_design(self) -> None:
         logger.debug("In Spot() evaluate_initial_design(), before calling self.fun: fun_control: %s", self.fun_control)
 
         self.y = self.fun(X=X_all, fun_control=self.fun_control)
-        self.y = apply_penalty_NA(self.y, self.fun_control["penalty_NA"])
+        self.y = apply_penalty_NA(self.y, self.fun_control["penalty_NA"], verbosity=self.verbosity)
         logger.debug("In Spot() evaluate_initial_design(), after calling self.fun: self.y: %s", self.y)
 
         # TODO: Error if only nan values are returned
@@ -1410,7 +1412,7 @@ def update_design(self) -> None:
         )
         # (S-18): Evaluating New Solutions:
         y0 = self.fun(X=X_all, fun_control=self.fun_control)
-        y0 = apply_penalty_NA(y0, self.fun_control["penalty_NA"])
+        y0 = apply_penalty_NA(y0, self.fun_control["penalty_NA"], verbosity=self.verbosity)
         X0, y0 = remove_nan(X0, y0, stop_on_zero_return=False)
         # Append New Solutions (only if they are not nan):
         if y0.shape[0] > 0:
@@ -1580,7 +1582,7 @@ def _get_pickle_safe_spot_tuner(self, unpickleables="file_io", verbosity=0) -> S
 
         return picklable_instance
 
-    def init_spot_writer(self) -> None:
+    def _init_spot_writer(self) -> None:
         """
         Initialize the spot_writer for the current experiment if in fun_control
         the tensorboard_log is set to True
@@ -1590,7 +1592,7 @@ def init_spot_writer(self) -> None:
         if self.fun_control["tensorboard_log"] and self.fun_control["spot_tensorboard_path"] is not None:
             self.spot_writer = SummaryWriter(log_dir=self.fun_control["spot_tensorboard_path"])
             if self.verbosity > 0:
-                print(f"Created spot_tensorboard_path: {self.fun_control['spot_tensorboard_path']} for SummaryWriter()")
+                print(f"_init_spot_writer(): Created spot_tensorboard_path: {self.fun_control['spot_tensorboard_path']} for SummaryWriter()")
         else:
             self.spot_writer = None
             if self.verbosity > 0:
@@ -1643,7 +1645,7 @@ def generate_random_point(self):
         logger.debug("In Spot() generate_random_point(), before calling self.fun: X_all: %s", X_all)
         logger.debug("In Spot() generate_random_point(), before calling self.fun: fun_control: %s", self.fun_control)
         y0 = self.fun(X=X_all, fun_control=self.fun_control)
-        y0 = apply_penalty_NA(y0, self.fun_control["penalty_NA"])
+        y0 = apply_penalty_NA(y0, self.fun_control["penalty_NA"], verbosity=self.verbosity)
         X0, y0 = remove_nan(X0, y0, stop_on_zero_return=False)
         return X0, y0
 
@@ -1978,7 +1980,7 @@ def plot_model(self, y_min=None, y_max=None) -> None:
         if self.k == 1:
             X_test = np.linspace(self.lower[0], self.upper[0], 100)
             y_test = self.fun(X=X_test.reshape(-1, 1), fun_control=self.fun_control)
-            y_test = apply_penalty_NA(y_test, self.fun_control["penalty_NA"])
+            y_test = apply_penalty_NA(y_test, self.fun_control["penalty_NA"], verbosity=self.verbosity)
             if isinstance(self.surrogate, Kriging):
                 y_hat = self.surrogate.predict(X_test[:, np.newaxis], return_val="y")
             else:

src/spotpython/utils/init.py

@@ -74,7 +74,7 @@ def fun_control_init(
     prep_model=None,
     prep_model_name=None,
     progress_file=None,
-    save_experiment=True,
+    save_experiment=False,
     save_result=True,
     scaler=None,
     scaler_name=None,

src/spotpython/utils/repair.py

@@ -78,7 +78,7 @@ def remove_nan(X: np.ndarray, y: np.ndarray, stop_on_zero_return: bool = False)
     return X_cleaned, y_cleaned
 
 
-def apply_penalty_NA(y: np.ndarray, penalty_NA: float, sd=0.1, stop_on_zero_return: bool = False) -> np.ndarray:
+def apply_penalty_NA(y: np.ndarray, penalty_NA: float, sd=0.1, stop_on_zero_return: bool = False, verbosity=0) -> np.ndarray:
     """
     Replaces NaN values in y with a penalty value of penalty_NA and issues a warning if necessary.
 
@@ -87,6 +87,7 @@ def apply_penalty_NA(y: np.ndarray, penalty_NA: float, sd=0.1, stop_on_zero_retu
         penalty_NA (float): penalty value to replace NaN values in y
         sd (float): standard deviation for the random noise added to penalty_NA. Default is 0.1.
         stop_on_zero_return (bool): whether to stop if the returned dimension is less than 1. Default is False.
+        verbosity (int): verbosity level. Default is 0.
 
     Returns:
         numpy.ndarray: y array with NaN values replaced by penalty value
@@ -124,8 +125,9 @@ def apply_penalty_NA(y: np.ndarray, penalty_NA: float, sd=0.1, stop_on_zero_retu
     if nan_dim > 1:
         warnings.warn(f"\n!!! The dimension of the returned y array is {y_cleaned.shape[0]}, " f"which is smaller than the original dimension {original_dim}.")
         warnings.warn("\n!!! Check whether continuing with the reduced dimension is useful.")
+        if verbosity > 0:
+            print(f"y before penalty: {y}. y after penalty: {y_cleaned}")
 
     if (original_dim - nan_dim) < 1 and stop_on_zero_return:
         raise ValueError("!!!! The dimension of the returned y array is less than 1. Check the input data.")
-
     return y_cleaned