AttributeError: 'tuple' object has no attribute 'shape'

[skrub-data]

AttributeError: 'tuple' object has no attribute 'shape'

[GaelVaroquaux]

I've made a new release: https://pypi.org/project/dirty-cat/

python -m pip install --upgrade dirty-cat

should fix the problem.

Tell us if it works!

[amy12xx]

are you using it on your own dataset? If so, may be it is not in the expected datatype (numpy or pandas array)?

[AC-Meira]

Yes, I'm using another dataset (https://www.kaggle.com/osmi/mental-health-in-tech-survey).
The variable I'm passing is a numpy.ndarray.
Follow my code:

dataset_C = pd.read_csv('mental-health-in-tech-survey.csv')
dataset_C.info()
values = dataset_C[['Gender', 'Country', 'state']]
sorted_values = values['Gender'].sort_values().unique()
sorted_values
type(sorted_values)
similarity_encoder = SimilarityEncoder(similarity='ngram')
transformed_values = similarity_encoder.fit_transform(sorted_values.reshape(-1, 1))

And log:
log_dirty_cat_test.txt

[anhtholee]

I just had the same issue. Have you been able to fix it? I think the problem is from this line:

X = self._check_X(X)

That _check_X method returns a tuple according to this line from scikit-learn's repo, hence the error.

[AC-Meira]

Unfortunately I didn't have time to fix it and went for another solution to my problem.
But I'm still very interested in this encoder. If I have any news I'll post it here. And please, if you have something too, let me know.

[amy12xx]

This looks to be fixed in the source (not yet released) version.

[GaelVaroquaux]

I should do a release. Let's see if I can do that today...

[AC-Meira]

Yeah! It's working!
Thanks!
@anhtholee Is it working for you too?
I think we can close this issue, right?

[anhtholee]

@AC-Meira Yes it worked for me too.

[GaelVaroquaux]

Great! Thanks a lot!!