fix bug on raising warnings

fix


fix

black_it/calibrator.py

@@ -21,6 +21,7 @@
 import os
 import textwrap
 import time
+import warnings
 from typing import Callable, Dict, List, Optional, Tuple, Union
 
 import numpy as np
@@ -95,10 +96,11 @@ def __init__(  # pylint: disable=too-many-arguments
             self.N = self.real_data.shape[0]
         else:
             if sim_length != self.real_data.shape[0]:
-                raise RuntimeWarning(
+                warnings.warn(
                     "The length of real time series is different from the simulation length, "
                     f"got {self.real_data.shape[0]} and {sim_length}. This may or may not be a problem depending "
-                    "on the loss function used."
+                    "on the loss function used.",
+                    RuntimeWarning,
                 )
             self.N = sim_length
         self.D = self.real_data.shape[1]

black_it/samplers/gaussian_process.py

@@ -16,6 +16,7 @@
 
 """This module contains the implementation of the Gaussian process-based sampling."""
 import random
+import warnings
 from enum import Enum
 from typing import Optional, Tuple, cast
 
@@ -164,8 +165,9 @@ def sample_batch(
         X, Y = existing_points, np.atleast_2d(existing_losses).T
 
         if X.shape[0] > 500:
-            raise RuntimeWarning(
-                "Standard GP evaluations can be expensive for large datasets, consider implementing a sparse GP"
+            warnings.warn(
+                "Standard GP evaluations can be expensive for large datasets, consider implementing a sparse GP",
+                RuntimeWarning,
             )
 
         # initialize GP class from GPy with a Matern kernel by default

tests/test_samplers/test_gaussian_process.py

@@ -32,10 +32,12 @@ def setup(self) -> None:
         self.xys, self.losses = self._construct_fake_grid()
 
     @classmethod
-    def _construct_fake_grid(cls) -> Tuple[NDArray[np.float64], NDArray[np.float64]]:
+    def _construct_fake_grid(
+        cls, n: int = 6
+    ) -> Tuple[NDArray[np.float64], NDArray[np.float64]]:
         """Construct a fake grid of evaluated losses."""
-        xs = np.linspace(0, 1, 6)
-        ys = np.linspace(0, 1, 6)
+        xs = np.linspace(0, 1, n)
+        ys = np.linspace(0, 1, n)
         xys_list = []
         losses_list = []
 
@@ -106,17 +108,20 @@ def test_gaussian_process_sample_warning_too_large_dataset() -> None:
         parameters_precision=np.array([0.01, 0.01]),
         verbose=False,
     )
-    xs = np.linspace(0, 1, 1000)
-    ys = np.linspace(0, 1, 3)
-    existing_points = xs.reshape((-1, 1)).dot(ys.reshape((1, -1)))
-    losses = np.zeros(existing_points.shape)
-    with pytest.raises(
+    # very high number of samples
+    (
+        xys,
+        losses,
+    ) = TestGaussianProcess2D._construct_fake_grid(  # pylint: disable=protected-access
+        n=23
+    )
+    with pytest.warns(
         RuntimeWarning,
         match="Standard GP evaluations can be expensive "
         "for large datasets, consider implementing "
         "a sparse GP",
     ):
-        sampler.sample(param_grid, existing_points, losses)
+        sampler.sample(param_grid, xys, losses)
 
 
 def test_gaussian_process_sample_wrong_acquisition() -> None: