Fix requires_fit tag for stateless FeatureHasher and HashingVectorizer #31851
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Set requires_fit=False for both FeatureHasher and HashingVectorizer - Both estimators are documented as stateless and work without fit() - Added tests to verify the tag behavior - Addresses inconsistency noted in issue scikit-learn#30689
- Remove whitespace from blank lines - Use double quotes for strings - Add proper spacing between functions
adrinjalali
reviewed
Jul 30, 2025
|
@adrinjalali Changelog entry added as requested. Thanks for the review! All checks are now passing. |
adrinjalali
reviewed
Jul 31, 2025
| @@ -0,0 +1 @@ | |||
| :class:`feature_extraction.FeatureHasher` and :class:`feature_extraction.HashingVectorizer` now correctly set ``requires_fit=False`` tag to reflect their stateless nature. | |||
There was a problem hiding this comment.
Please follow the max 88 chars here as well.
There was a problem hiding this comment.
@adrinjalali I changed it, currently less than 88 char.
jeremiedbb
reviewed
Aug 1, 2025
| @@ -0,0 +1 @@ | |||
| Set ``requires_fit=False`` for ``FeatureHasher`` and ``HashingVectorizer``. | |||
There was a problem hiding this comment.
I'd say it's a fix so the changelog fragment should be named 31851.fix.rst instead.
Suggested change
| Set ``requires_fit=False`` for ``FeatureHasher`` and ``HashingVectorizer``. | |
| - Set the tag `requires_fit=False` for the classes | |
| :class:`feature_extraction.FeatureHasher` and | |
| :class:`feature_extraction.HashingVectorizer`. | |
| By :user:`hakan çanakçı <hqkqn32>`. |
jeremiedbb
reviewed
Aug 1, 2025
doc/whats_new/upcoming_changes/sklearn.feature_extraction/31851.fix.rst
Outdated
Show resolved
Hide resolved
|
@jeremiedbb Thanks for the suggestions. All steps completed. |
lucyleeow
pushed a commit
to lucyleeow/scikit-learn
that referenced
this pull request
Aug 22, 2025
scikit-learn#31851) Co-authored-by: Jérémie du Boisberranger <jeremie@probabl.ai>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Reference Issues/PRs
Fixes #30689
What does this implement/fix? Explain your changes.
This PR fixes an inconsistency in the tag system for two stateless estimators.
Problem:
FeatureHasherandHashingVectorizerare documented as stateless estimators (nofit()required), but theirrequires_fittag was incorrectly set toTrue.Solution:
__sklearn_tags__()method in both classes to setrequires_fit=Falsefit()Files changed:
sklearn/feature_extraction/_hash.py- Addedrequires_fit=Falsetagsklearn/feature_extraction/text.py- Addedrequires_fit=Falsetagsklearn/feature_extraction/tests/test_feature_hasher.py- Added tag validation testssklearn/feature_extraction/tests/test_text.py- Added tag validation testsThis ensures consistency with other stateless estimators in scikit-learn.
Any other comments?
This change is backward compatible and doesn't affect the public API. The estimators continue to work exactly as before, but now their internal tags correctly reflect their stateless nature.
cc/ @glemaitre @adrinjalali (as mentioned in the original issue)