Closed
Description
Description
While working on PR #11682 I stumbled upon this small issue in tests/test_metaestimator.py. In the delegation test, score only has the parameter X, while estimators expect X and y to be passed. Because of this, RFE and RFECV cannot be tested for score delegation.
Steps/Code to Reproduce
Remove score from the skip_methods for RFE and RFECV in test_metaestimator.py and run the test.
Expected Results
The test passes.
Actual Results
It fails with:
E
======================================================================
ERROR: sklearn.tests.test_metaestimators.test_metaestimator_delegation
----------------------------------------------------------------------
Traceback (most recent call last):
File "/usr/lib/python3.6/site-packages/nose/case.py", line 198, in runTest
self.test(*self.arg)
File "/home/oliver/repos/scikit-learn/sklearn/tests/test_metaestimators.py", line 124, in test_metaestimator_delegation
delegator_data.fit_args[0])
File "/home/oliver/repos/scikit-learn/sklearn/utils/_unittest_backport.py", line 204, in assertRaises
return context.handle('assertRaises', args, kwargs)
File "/home/oliver/repos/scikit-learn/sklearn/utils/_unittest_backport.py", line 113, in handle
callable_obj(*args, **kwargs)
File "/home/oliver/repos/scikit-learn/sklearn/utils/metaestimators.py", line 118, in <lambda>
out = lambda *args, **kwargs: self.fn(obj, *args, **kwargs)
TypeError: score() missing 1 required positional argument: 'y'
----------------------------------------------------------------------
Ran 1 test in 0.006s
FAILED (errors=1)
Versions
Linux-4.18.7_2-x86_64-with-glibc2.3.4
Python 3.6.6 (default, Jul 16 2018, 09:23:17)
[GCC 7.3.0]
NumPy 1.15.1
SciPy 1.1.0
Scikit-Learn 0.21.dev0
Metadata
Metadata
Assignees
Labels
No labels