Error when importing eli5

[leetabix]

When creating a virtual environment (using venv) and pip installing eli5 and jupyter lab, I'm unable to import eli5 as I receive the following error in my notebook:

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
Cell In[1], line 1
----> 1 import eli5

File \Lib\site-packages\eli5\__init__.py:13
      6 from .formatters import (
      7     format_as_html,
      8     format_html_styles,
      9     format_as_text,
     10     format_as_dict,
     11 )
     12 from .explain import explain_weights, explain_prediction
---> 13 from .sklearn import explain_weights_sklearn, explain_prediction_sklearn
     14 from .transform import transform_feature_names
     17 try:

File \Lib\site-packages\eli5\sklearn\__init__.py:3
      1 # -*- coding: utf-8 -*-
      2 from __future__ import absolute_import
----> 3 from .explain_weights import (
      4     explain_weights_sklearn,
      5     explain_linear_classifier_weights,
      6     explain_linear_regressor_weights,
      7     explain_rf_feature_importance,
      8     explain_decision_tree,
      9 )
     10 from .explain_prediction import (
     11     explain_prediction_sklearn,
     12     explain_prediction_linear_classifier,
     13     explain_prediction_linear_regressor,
     14 )
     15 from .unhashing import (
     16     InvertableHashingVectorizer,
     17     FeatureUnhasher,
     18     invert_hashing_and_fit,
     19 )

File \Lib\site-packages\eli5\sklearn\explain_weights.py:78
     73 from eli5.transform import transform_feature_names
     74 from eli5._feature_importances import (
     75     get_feature_importances_filtered,
     76     get_feature_importance_explanation,
     77 )
---> 78 from .permutation_importance import PermutationImportance
     81 LINEAR_CAVEATS = """
     82 Caveats:
     83 1. Be careful with features which are not
   (...)
     90    classification result for most examples.
     91 """.lstrip()
     93 HASHING_CAVEATS = """
     94 Feature names are restored from their hashes; this is not 100% precise
     95 because collisions are possible. For known collisions possible feature names
   (...)
     99 the result is positive.
    100 """.lstrip()

File \Lib\site-packages\eli5\sklearn\permutation_importance.py:7
      5 import numpy as np
      6 from sklearn.model_selection import check_cv
----> 7 from sklearn.utils.metaestimators import if_delegate_has_method
      8 from sklearn.utils import check_array, check_random_state
      9 from sklearn.base import (
     10     BaseEstimator,
     11     MetaEstimatorMixin,
     12     clone,
     13     is_classifier
     14 )

ImportError: cannot import name 'if_delegate_has_method' from 'sklearn.utils.metaestimators' (\Lib\site-packages\sklearn\utils\metaestimators.py)

The issue appears to be due to scikit-learn depreciating the if_delegate_has_method decorator (in \Lib\site-packages\sklearn\utils\metaestimators.py) and replacing it with another decorator called available_if.

I've made the following updates to \Lib\site-packages\eli5\sklearn\permutation_importance.py and I'm now able to import the eli5 library:

• changed import statement to mention available_if instead of if_delegate_has_method
• replaced all mentions of:

    @if_delegate_has_method(delegate='wrapped_estimator_')

with

    @available_if(_estimator_has('wrapped_estimator_'))

• defined a new _estimator_has function for use with the above

Please let me know if you'd like me to propose the above update for merge.

Thanks,
James

[GegznaV]

@leetabix I think you should propose a pull request (PR) with the changes you've already made: it at least will be possible to install the working version of the package from your PR.

[GegznaV]

Related:

• Deprecate if_delegate_has_method and replace with available_if scikit-learn/scikit-learn#20506

[frosty24]

This commit addresses this as well:
840695d

[GegznaV]

@frosty24, is 840695d merged to the main branch of this repo? I'm a bit confused as GitHub informs:

[frosty24]

I picked this commit from the following open PR in this repo:
#31

[isabelizimm]

Hi all! I ran into this error as well. I ended up installing scikit-learn==1.2.2 and then was able to import eli5 successfully.

[mprytoluk]

Is it possible for me to do something about this? I've stopped using eli5 in my project because I had to update scikit-learn.

[GegznaV]

If nobody answers, you may use SHAP values as an alternative: https://github.com/shap/shap That project is currently alive.

[RobertCarruthers-Ki]

Slight correction to the above.

Instead of replacing
@if_delegate_has_method(delegate='wrapped_estimator_')
with:
@available_if(_estimator_has('wrapped_estimator_'))
it should be replaced with the name of the function being checked for:
@available_if(_estimator_has('predict_proba')) (or whatever the relevant function is)
based on:
scikit-learn/scikit-learn#20506

This is implemented in my fork here

[DrJStrudwick]

Any update on this?

[mprytoluk]

I believe it's safe to say this project is dead. Some fork on this project would be great.

[lopuhin]

This should be fixed by #48, we'll make a PyPI release soon.

[lopuhin]

this was release on PyPI as 0.14.0, closing