Can not import feature_engine package's module in Kaggle Editor.
(First noting the below lines of codes are working in local machine's Jupyter Notebook, but not in Kaggle )
To Reproduce, I did below inside Kaggle Editor
To install feature_engine just the regular code
!pip install feature_engine
following this link which is same as the official doc.
And then I just have the following code in Kaggle Notebook per the official documentation
from feature_engine import categorical_encoders as ce
After running this cell getting below error.
ImportError: cannot import name 'categorical_encoders' from 'feature_engine' (/opt/conda/lib/python3.7/site-packages/feature_engine/__init__.py)
And this error is coming for any modules of feature_engine . e.g. for the below line as well
from feature_engine import variable_transformers as vt
I am getting error
ImportError: cannot import name 'variable_transformers' from 'feature_engine' (/opt/conda/lib/python3.7/site-packages/feature_engine/__init__.py)
And there's no simple way to upgrade Python to 3.8, as Kaggle does not allow us to specify the python version in it's notebook.
In Kaggle I have the latest version of feature_engine as running !pip show feature_engine gives below
Name: feature-engine
Version: 1.0.2
Summary: Feature engineering package with Scikit-learn's fit transform functionality
Home-page: http://github.com/solegalli/feature_engine
Author: Soledad Galli
Author-email: solegalli@protonmail.com
License: BSD 3 clause
Location: /opt/conda/lib/python3.7/site-packages
Requires: scipy, pandas, statsmodels, scikit-learn, numpy
Required-by:
