"RuntimeWarning: invalid value encountered in true_divide" in tests

[cmarmo]

The following tests throw a RuntimeWarning

RuntimeWarning: invalid value encountered in true_divide

in the current development version

• test_sanity_check_pls_regression_constant_column_Y in sklearn/cross_decomposition/tests/test_pls.py
• test_iforest_with_uniform_data in sklearn/ensemble/tests/test_iforest.py
• test_kernel_gradient in sklearn/gaussian_process/tests/test_kernels.py
• test_radius_neighbors_classifier_zero_distance in sklearn/neighbors/tests/test_neighbors.py
• test_calibration_multiclass[0-True-isotonic] in sklearn/tests/test_calibration.py
• test_estimators[LinearDiscriminantAnalysis()-check_classifiers_one_label] in sklearn/tests/test_common.py
• test_random_choice_csc in sklearn/utils/tests/test_random.py

Those warnings are caused by matricial divisions having some element equal to zero in the denominator, either in the test itself or in the function/class called by the test.

The failure can be reproduced locally using

pytest -Werror::RuntimeWarning <test_file.py> -k <test_name>

cc @ogrisel, @reshamas

[amueller]

If you're starting on this issue, maybe try do one file per PR, or potentially several if the errors are related, not one PR that addresses all of it.

[reshamas]

If you're starting on this issue, maybe try do one file per PR, or potentially several if the errors are related, not one PR that addresses all of it.

I think this is confusing. I would say (doesn't matter if the errors are related, because folks can practice having multiple branches):

Submit only one file in each PR.

[nuka137]

@cmarmo

I will join next dev sprint (9th Feb).
I would like to tackle this issue in this sprint.

At first, I will tackle below test.

test_radius_neighbors_classifier_zero_distance in sklearn/neighbors/tests/test_neighbors.py

[marenwestermann]

I'm working on test_sanity_check_pls_regression_constant_column_Y.

[mabu-dev]

Working on sklearn/utils/tests/test_random.py

[t-kusanagi]

I'll work on test_calibration_multiclass[0-True-isotonic] in sklearn/tests/test_calibration.py

[LSturtew]

I'm working on test_iforest_with_uniform_data in sklearn/ensemble/tests/test_iforest.py