scikit-learn 0.22 compatibility #353
Conversation
Codecov Report
@@ Coverage Diff @@
## master #353 +/- ##
==========================================
+ Coverage 97.31% 97.32% +<.01%
==========================================
Files 49 49
Lines 3134 3138 +4
Branches 584 584
==========================================
+ Hits 3050 3054 +4
Misses 44 44
Partials 40 40
|
|
Failed tests are not related to new sklearn version. |
eli5/sklearn/text.py
Outdated
| try: | ||
| from sklearn.feature_extraction.text import _VectorizerMixin as VectorizerMixin | ||
| except ImportError: # Changed in scikit-learn 0.22 | ||
| from sklearn.feature_extraction.text import VectorizerMixin # type: ignore |
There was a problem hiding this comment.
In the same file there is an import of from sklearn.feature_extraction.text import VectorizerMixin on line 4, so at this point either VectorizerMixin has already been imported, or the whole module would fail to import. So unless I misread the code, this would be better moved to the top of the module, replacing current VectorizerMixin import.
There was a problem hiding this comment.
Yes, that's right. I have just moved import to the top of file.
| from sklearn.pipeline import FeatureUnion # type: ignore | ||
| try: | ||
| from sklearn.feature_extraction.text import _VectorizerMixin as VectorizerMixin | ||
| except ImportError: # Changed in scikit-learn 0.22 |
There was a problem hiding this comment.
FWIW it still works in sklearn 0.22, but it's already deprecated (to be removed in 0.24), so good that we are ready 👍
Adjust ELI5 to scikit-learn 0.22.