Import broken with scikit-learn>=0.22
The Imputer class imported here:
https://github.com/combust/mleap/blob/beb4fd248e020282f10778c91d68a3518e981dcb/python/mleap/sklearn/extensions/data.py#L24
Was removed in scikit-learn version 0.22 as mentioned here:
https://github.com/scikit-learn/scikit-learn/blob/ee328faa3601b40944ad43e28bce71860d39f2de/sklearn/preprocessing/imputation.py#L60
At the moment, we have scikit-learn pinned to < 0.20, we would need to upgrade the scikit-learn version to get this working. Happy to work with you on doing this if you'd like!
I'm working in an environment where I need scikit-learn>=0.22 and I get the error importing Imputer from the preprocessing.data module which I believe is the same. I 'hacked' a fix, namely I edited the mleap/sklearn/preprocessing/data.py file to have:
from sklearn.preprocessing import StandardScaler, MinMaxScaler, Binarizer, PolynomialFeatures
from sklearn.impute import SimpleImputer as Imputer
instead of the previous line importing Imputer. Does anyone know if this is likely to lead to trouble down the line?
Also, as an aside, as @ancasarb mentioned scikit is pinned to < 0.20, it would be useful if this was stated on the 'Getting Started with Sklearn' Documentation page.
E: And pip installing sklearn 0.23 when you pip install mleap is unfortunate too.