hat_matrix_ not available for BasisSmoother

[mynanshan]

Describe the bug
Setting return_basis=True in BasisSmoother leads to

AttributeError: 'BasisSmoother' object has no attribute 'hat_matrix_'

With a further look into the codes, I found that in _LinearSmoother the hat_matrix_ attribute is created inside the fit method with the line

self.hat_matrix_ = self.hat_matrix()

Meanwhile, that line of code does not appear in the fit method of the BasisSmoother.

To Reproduce
Code to reproduce the behavior:

Forgive me for copying the long code to generate data for the example:

import numpy as np
from scipy.integrate import solve_ivp

def simple_linear_ode(t, x, A):
    dx = A @ x
    return dx

t0 = 0; t1 = 1
m = 100
t = np.linspace(t0, t1, m+1)

n = 10
A = np.zeros((n,n))
for k in range(int(n / 2)):
    A[2*k, 2*k+1] = 2 * np.pi * (k+1)
    A[2*k+1, 2*k] = -2 * np.pi * (k+1)


x0 = np.random.normal(0, 0.5, n)
ode_solution = solve_ivp(simple_linear_ode, t_span=(t0, t1), y0=x0, t_eval=t, args=(A,))
x = ode_solution.y
snr = 5
noise_sd = np.std(x, axis=1) / snr
y = x + np.random.normal(0, noise_sd[:, np.newaxis], x.shape)

And the skfda part starts from here:

from skfda import FDataGrid
from skfda.representation.basis import BSplineBasis
from skfda.preprocessing.smoothing import BasisSmoother
from skfda.misc.regularization import L2Regularization
from skfda.misc.operators import LinearDifferentialOperator
from skfda.preprocessing.smoothing.validation import SmoothingParameterSearch
from skfda.preprocessing.smoothing.validation import LinearSmootherGeneralizedCVScorer

basis_degree = 3
basis_knots = np.linspace(t0, t1, int(m/2)+1)
basis = BSplineBasis(domain_range=(t0,t1), order=basis_degree+1, knots=basis_knots)

yfd = FDataGrid(data_matrix=y, grid_points=t)

smoother = SmoothingParameterSearch(
    BasisSmoother(basis, regularization=L2Regularization(LinearDifferentialOperator(2)), return_basis=True),
    param_values=np.exp(np.arange(-10,0,0.5)),
    param_name="smoothing_parameter",
    scoring=LinearSmootherGeneralizedCVScorer()
)

smoother.fit(yfd)
print(f"The best smoothing parameter is {smoother.best_params_}")
yfd_smooth = smoother.transform(yfd)

Error messages

[.../lib/python3.11/site-packages/sklearn/model_selection/_validation.py:821](.../lib/python3.11/site-packages/sklearn/model_selection/_validation.py:821): UserWarning: Scoring failed. The score on this train-test partition for these parameters will be set to nan. Details: 
Traceback (most recent call last):
  File ".../lib/python3.11/site-packages/sklearn/model_selection/_validation.py", line 810, in _score
    scores = scorer(estimator, X_test, y_test)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".../lib/python3.11/site-packages/skfda/preprocessing/smoothing/validation.py", line 129, in __call__
    y_est, hat_matrix = _get_input_estimation_and_matrix(estimator, X)
                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".../lib/python3.11/site-packages/skfda/preprocessing/smoothing/validation.py", line 26, in _get_input_estimation_and_matrix
    hat_matrix = estimator.hat_matrix_
                 ^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'BasisSmoother' object has no attribute 'hat_matrix_'

  warnings.warn(
The best smoothing parameter is {'smoothing_parameter': 4.5399929762484854e-05}
[/.../lib/python3.11/site-packages/sklearn/model_selection/_search.py:979](.../lib/python3.11/site-packages/sklearn/model_selection/_search.py:979): UserWarning: One or more of the test scores are non-finite: [nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan nan
 nan nan]
  warnings.warn(

Version information

• OS: Ubuntu 22.04 LTS
• Python version: 3.11.6 | packaged by conda-forge | (main, Oct 3 2023, 10:40:35) [GCC 12.3.0]
• scikit-fda version: 0.9
• numpy: 1.26.0

[vnmabus]

Sorry for taking too long to answer. I confirmed that the bug is reproducible, and I will try to fix it as soon as I can.

@all-contributors please add @mynanshan for bug reports.

[allcontributors]

@vnmabus

I couldn't determine any contributions to add, did you specify any contributions?
Please make sure to use valid contribution names.

I've put up a pull request to add @mynanshan! 🎉