Closed jeromedockes closed 3 months ago
ATM the outputs have the default 1,...,n index. This preserves the index of the input, as is done by scikit-learn when the input is a pandas dataframe and set_output(transform='pandas')
the test failures are not related to the PR but numpy 2, they will be fixed by #946
@glemaitre thanks for the review -- I think I've addressed your comments now
Thanks @jeromedockes
ATM the outputs have the default 1,...,n index. This preserves the index of the input, as is done by scikit-learn when the input is a pandas dataframe and set_output(transform='pandas')