Remove some warnings in tests.

GAA-UAM · Oct 13, 2023 · 4f0bd4c · 4f0bd4c
1 parent 14b1416
commit 4f0bd4c
Show file tree

Hide file tree

Showing 7 changed files with 16 additions and 55 deletions.
diff --git a/skfda/inference/anova/_anova_oneway.py b/skfda/inference/anova/_anova_oneway.py
@@ -182,7 +182,13 @@ def v_asymptotic_stat(
         .. footbibliography::
 
     """
-    return float(_v_asymptotic_stat_with_reps(*fd, weights=weights, p=p))
+    return float(
+        _v_asymptotic_stat_with_reps(
+            *fd,
+            weights=weights,
+            p=p,
+        ).reshape(())
+    )
 
 
 def _anova_bootstrap(

diff --git a/skfda/misc/metrics/_utils.py b/skfda/misc/metrics/_utils.py
@@ -120,8 +120,8 @@ class NormInducedMetric(Metric[VectorType]):
 
         >>> l2_distance = NormInducedMetric(l2_norm)
         >>> d = l2_distance(fd, fd2)
-        >>> float('%.3f'% d)
-        0.289
+        >>> float(d[0])
+        0.288...
 
     """
 

diff --git a/skfda/ml/classification/_logistic_regression.py b/skfda/ml/classification/_logistic_regression.py
@@ -132,11 +132,9 @@ def fit(  # noqa: D102, WPS210
             (self.max_features, n_features),
         )
 
-        penalty = 'none' if self.penalty is None else self.penalty
-
         # multivariate logistic regression
         mvlr = mvLogisticRegression(
-            penalty=penalty,
+            penalty=self.penalty,
             C=self.C,
             solver=self.solver,
             max_iter=self.max_iter,

diff --git a/skfda/ml/clustering/_hierarchical.py b/skfda/ml/clustering/_hierarchical.py
@@ -173,7 +173,7 @@ def _init_estimator(self) -> None:
 
         self._estimator = sklearn.cluster.AgglomerativeClustering(
             n_clusters=self.n_clusters,
-            affinity='precomputed',
+            metric='precomputed',
             memory=self.memory,
             connectivity=self.connectivity,
             compute_full_tree=self.compute_full_tree,

diff --git a/skfda/preprocessing/dim_reduction/variable_selection/_rkvs.py b/skfda/preprocessing/dim_reduction/variable_selection/_rkvs.py
@@ -83,7 +83,7 @@ def _rkhs_vs(
                 [indexes[j]],
             ])
 
-            new_means = np.atleast_2d(means[new_selection])
+            new_means = means[new_selection]
 
             lstsq_solution = linalg.lstsq(
                 variances[new_selection[:, np.newaxis], new_selection],

diff --git a/skfda/preprocessing/dim_reduction/variable_selection/mrmr.py b/skfda/preprocessing/dim_reduction/variable_selection/mrmr.py
@@ -160,7 +160,7 @@ def _mrmr(
                 redundancies[last_selected, j] = redundancy_dependence_measure(
                     X[:, last_selected, np.newaxis],
                     X[:, j, np.newaxis],
-                )
+                ).item()
                 redundancies[j, last_selected] = redundancies[last_selected, j]
 
         W = np.mean(

diff --git a/skfda/preprocessing/dim_reduction/variable_selection/recursive_maxima_hunting.py b/skfda/preprocessing/dim_reduction/variable_selection/recursive_maxima_hunting.py
@@ -17,6 +17,7 @@
     overload,
 )
 
+import dcor
 import numpy as np
 import numpy.linalg as linalg
 import numpy.ma as ma
@@ -25,7 +26,6 @@
 from sklearn.base import clone
 from typing_extensions import Literal
 
-import dcor
 from skfda.exploratory.stats.covariance import (
     CovarianceEstimator,
     EmpiricalCovariance,
@@ -41,7 +41,6 @@
 
 if TYPE_CHECKING:
     from ....misc.covariances import CovarianceLike
-    import GPy
 
 
 def _transform_to_2d(t: ArrayLike) -> NDArrayFloat:
@@ -56,48 +55,6 @@ def _transform_to_2d(t: ArrayLike) -> NDArrayFloat:
     return t
 
 
-class _PicklableKernel():
-    """Class used to pickle GPy kernels."""
-
-    def __init__(self, kernel: GPy.kern.Kern) -> None:
-        super().__setattr__('_PicklableKernel__kernel', kernel)
-
-    def __getattr__(self, name: str) -> Any:
-        if name != '__deepcopy__':
-            return getattr(self.__kernel, name)
-
-    def __setattr__(self, name: str, value: Any) -> None:
-        setattr(self.__kernel, name, value)
-
-    def __getstate__(self) -> Mapping[str, Any]:
-        return {
-            'class': self.__kernel.__class__,
-            'input_dim': self.__kernel.input_dim,
-            'values': self.__kernel.param_array,
-        }
-
-    def __setstate__(self, state: Mapping[str, Any]) -> None:
-        super().__setattr__('_PicklableKernel__kernel', state['class'](
-            input_dim=state['input_dim']),
-        )
-        self.__kernel.param_array[...] = state['values']
-
-    def __call__(self, *args: Any, **kwargs: Any) -> NDArrayFloat:
-        return self.__kernel.K(*args, **kwargs)  # type: ignore[no-any-return]
-
-
-def make_kernel(k: CovarianceLike) -> CovarianceLike:
-    try:
-        import GPy
-    except ImportError:
-        return k
-
-    if isinstance(k, GPy.kern.Kern):
-        return _PicklableKernel(k)
-
-    return k
-
-
 def _absolute_argmax(
     function: FDataGrid,
     *,
@@ -255,7 +212,7 @@ def __init__(
         super().__init__()
 
         self.mean = mean
-        self.cov = make_kernel(cov)
+        self.cov = cov
 
     def _evaluate_mean(self, t: NDArrayFloat) -> NDArrayFloat:
 
@@ -625,7 +582,7 @@ def __call__(
         **kwargs: Any,
     ) -> bool:
 
-        score = float(dependences.data_matrix[0, selected_index, 0])
+        score = float(dependences.data_matrix[(0,) + selected_index + (0,)])
 
         return score < self.threshold