[python-package][R-package] adapt to scikit-learn 1.6 testing changes, pin more packages in R 3.6 CI jobs #6718
Conversation
jameslamb
changed the title
jameslamb
changed the title
jameslamb
requested review from
StrikerRUS,
borchero,
guolinke,
jmoralez and
shiyu1994
as code owners
| ) | ||
|
|
||
| .install_packages(c( | ||
| "brio/brio_1.1.4.tar.gz" # nolint: non_portable_path |
There was a problem hiding this comment.
All of these hard-coded versions are the latest for which there's a package at https://cran.r-project.org/src/contrib/Archive. That provides a set of packages that were all working together as of a few days ago.
We shouldn't need to actively manage this list... this should be able to remain untouched (I hope) until we drop R 3 support and delete this script entirely.
| _sklearn_ClassifierTags = None | ||
| _sklearn_RegressorTags = None |
There was a problem hiding this comment.
Why are these defined again here?
There was a problem hiding this comment.
Because if an earlier ImportError happens, then this part above would never get evaluated:
except ImportError:
_sklearn_ClassifierTags = None
_sklearn_RegressorTags = NoneAnd then the from .compat imprt _sklearn_ClassifierTags in sklearn.py would fail. This is not new, just following the pattern that's existed in LightGBM for a long time (see all the other {something} = None above it).
Happy to consider something else if you have a recommendation for improving this!
There was a problem hiding this comment.
Oh yeah, sorry.
The first one is defined for scikit-learn<1.6 and the second when scikit-learn isn't installed.
There was a problem hiding this comment.
yeah exactly. No need to be sorry, it's confusing!
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
StrikerRUS
changed the title
3 participants
Fixes #6717, adapting
lightgbmto the following changes inscikit-learn(currently visible inscikit-learn1.6 nightlies).from scikit-learn/scikit-learn#30122:
tags.estimator_typein the tags returns by__sklearn_tags__()tags.regressor_tagswith a non-Nonevaluestags.classifier_tagswith a non-Nonevaluesfrom scikit-learn/scikit-learn#30149:
@parametrize_with_checks()ignores_xfail_checksin_more_tags()/__sklearn_tags__()on estimators... projects must now pass a function that generates a dictionary of xfailed checks to@parametrize_with_checks()in test codeFixes #6719, pinning all the versions of dependencies installed in R 3.6 CI jobs to make those jobs more stable (see previous issues: #6442 (comment), #6434).