Introducing FormulaicTransformer and deprecating PatsyTransformer
#593
Conversation
| @deprecated( | ||
| version="0.6.17", | ||
| reason="Please use `sklego.preprocessing.FormulaicTransformer` instead. " | ||
| "This object will be removed from the preprocessing submodule in version 0.9.0.", |
There was a problem hiding this comment.
Totally made up version number in which to remove😂
There was a problem hiding this comment.
No worries :) That said ... I'd be fine with 0.8.0.
| assert pipe.fit(X, y).predict(X).shape[0] == X.shape[0] | ||
|
|
||
|
|
||
| def test_unseen_categories(df): |
There was a problem hiding this comment.
Differently from patsy, formulaic raises a warning but doesn't break with unseen categories
FBruzzesi
changed the title
FormulaicTransformer and deprecating PatsyTransformer
| assert isinstance(df_fit_transformed, expected_type) | ||
|
|
||
|
|
||
|
|
There was a problem hiding this comment.
One extra test for this one, could we confirm that we're able to save/load this component? This was an issue with patsy because it wouldn't pickle.
There was a problem hiding this comment.
Just added pickling test.
Also found in the formulaic changelogs that from 0.3.0:
Pickleability is now guaranteed and tested via unit tests. Failure to pickle any formulaic metadata object (such as formulas, model specs, etc) is considered a bug.
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
Description
Addresses #559 by deprecating
PatsyTransformerand introducing aFormulaicTransformer.In writing of
FormulaicTranformerI followed what is described in the formulaic documentation for integrating with sklearn. The main (tiny) differences with such implementation can be summarized as:return_typeparameter available (similar to whatPatsyTransformerdoes)__init__method.transform()Additionally the PR makes both dependencies optional.
Type of change
Checklist: