This is my pull request for the issue 647

[rosignol08]

References to issues or other PRs

Describe the proposed changes

Additional information

Checklist before requesting a review

• I have performed a self-review of my code
• The code conforms to the style used in this package
• The code is fully documented and typed (type-checked with Mypy)
• I have added thorough tests for the new/changed functionality

[codecov]

Codecov Report

Attention: Patch coverage is 48.64865% with 19 lines in your changes missing coverage. Please review.

Project coverage is 86.75%. Comparing base (2f48baa) to head (211cf07).
Report is 30 commits behind head on develop.

Files with missing lines	Patch %	Lines
skfda/misc/scoring.py	48.64%	19 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #649      +/-   ##
===========================================
- Coverage    86.83%   86.75%   -0.09%     
===========================================
  Files          157      157              
  Lines        13522    13603      +81     
===========================================
+ Hits         11742    11801      +59     
- Misses        1780     1802      +22     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[vnmabus]

I think some tests are needed for this functionality, to check with some simple examples that we are computing the right values.

[rosignol08]

hello I'm coming back after a little while of abscence I fixed the main problems and for the tests I did this: ```
import numpy as np
from skfda.misc.scoring import root_mean_squared_error, root_mean_squared_log_error
from skfda.representation.grid import FDataGrid

def test_root_mean_squared_error_ndarray():
# Cas avec des np.ndarray
y_true = np.array([3.0, -0.5, 2.0, 7.0])
y_pred = np.array([2.5, 0.0, 2.1, 7.8])

# Test sans poids
result = root_mean_squared_error(y_true, y_pred)
expected_result = np.sqrt(np.mean((y_true - y_pred) ** 2))  # Calcul attendu

# Debugging: Print intermediate results
print(f"y_true: {y_true}, y_pred: {y_pred}")
print(f"Manual RMSE calculation: {expected_result}")
print(f"skfda RMSE result: {result}")

#assert np.isclose(result, expected_result), f"Expected {expected_result}, but got {result}"

print("\ntest_root_mean_squared_error_ndarray\n")
test_root_mean_squared_error_ndarray()
def test_root_mean_squared_error_fdata():
# Cas avec des objets FData (exemple générique)
# Suppose que FDataGrid, FDataBasis ou FDataIrregular ont des méthodes adaptées
y_true_fdata = FDataGrid(data_matrix=np.array([3.0, -0.5, 2.0, 7.0])) # Exemple générique
y_pred_fdata = FDataGrid(data_matrix=np.array([2.5, 0.0, 2.1, 7.8])) # Exemple générique

# Test pour FData
result = root_mean_squared_error(y_true_fdata, y_pred_fdata)
expected_result = np.sqrt(np.mean((y_true_fdata.data_matrix - y_pred_fdata.data_matrix) ** 2))
print(f"Manual RMSE calculation: {expected_result}")
print(f"skfda RMSE result: {result}")
#assert np.isclose(result, expected_result), f"Expected {expected_result}, but got {result}"

print("\ntest_root_mean_squared_error_fdata\n")
test_root_mean_squared_error_fdata()
def test_root_mean_squared_error_multioutput():
# Cas avec multioutput
y_true = np.array([[3.0, -0.5], [2.0, 7.0]])
y_pred = np.array([[2.5, 0.0], [2.1, 7.8]])

result = root_mean_squared_error(y_true, y_pred, multioutput='uniform_average')
expected_result = np.sqrt(np.mean((y_true - y_pred) ** 2, axis=None))  # Match skfda's behavior
print(f"Manual RMSE calculation: {expected_result}")
print(f"skfda RMSE result: {result}")
#assert np.isclose(result, expected_result), f"Expected {expected_result}, but got {result}"

print("\ntest_root_mean_squared_error_multioutput\n")
test_root_mean_squared_error_multioutput()
def test_root_mean_squared_log_error():
# Cas avec np.ndarray
y_true = np.array([3.0, 0.5, 2.0, 7.0])
y_pred = np.array([2.5, 0.0, 2.1, 7.8])

# Test sans poids
result = root_mean_squared_log_error(y_true, y_pred)
expected_result = np.sqrt(np.mean((np.log1p(y_true) - np.log1p(y_pred)) ** 2))
print(f"Manual RMSE calculation: {expected_result}")
print(f"skfda RMSE result: {result}")
#assert np.isclose(result, expected_result), f"Expected {expected_result}, but got {result}"

print("\ntest_root_mean_squared_log_error\n")
test_root_mean_squared_log_error()

 it returns this :

test_root_mean_squared_error_ndarray

y_true: [ 3.  -0.5  2.   7. ], y_pred: [2.5 0.  2.1 7.8]
Manual RMSE calculation: 0.5361902647381803
skfda RMSE result: 0.5361902647381803

test_root_mean_squared_error_fdata

Manual RMSE calculation: 0.5361902647381803
skfda RMSE result: 0.47346242371139297

test_root_mean_squared_error_multioutput

Manual RMSE calculation: 0.5361902647381803
skfda RMSE result: 0.5361902647381803

test_root_mean_squared_log_error

Manual RMSE calculation: 0.21931244239925476
skfda RMSE result: 0.21931244239925476 which seems coherent but I don't really understand why the function "test_root_mean_squared_error_fdata" gives different results I need to figure why before redo my merge request ?

[vnmabus]

The changes look good.

Please, add your test functions to skfda\tests\test_scoring.py, so that they are tested with the rest of our tests. I think the ones you showed in the comments are already in the right format (they have a test_ prefix and use asserts) but you will need to remove the print statements. I will do a final round of review when the tests are in place.

[rosignol08]

I have added the test fonction now I hope it's good.

[eliegoudout]

I think overloads should replace default values with ... to avoid giving the impression hat they are taken into account (applies to every @overload).

[rosignol08]

I added only for those line.