TargetEncoder can't convert to string (AttributeError)

[momijiame]

Abstract

AttributeError occurs when TargetEncoder object passed to str() function.
This behavior has the potential to become an issue on debugging etc.

How to reproduce

The steps to reproduce are the following.

>>> from nyaggle.feature.category_encoder.target_encoder import TargetEncoder
>>> encoder = TargetEncoder()
>>> str(encoder)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/amedama/.virtualenvs/py310/lib/python3.10/site-packages/sklearn/base.py", line 279, in __repr__
    repr_ = pp.pformat(self)
  File "/usr/local/Cellar/python@3.10/3.10.9/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pprint.py", line 157, in pformat
    self._format(object, sio, 0, 0, {}, 0)
  File "/usr/local/Cellar/python@3.10/3.10.9/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pprint.py", line 174, in _format
    rep = self._repr(object, context, level)
  File "/usr/local/Cellar/python@3.10/3.10.9/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pprint.py", line 454, in _repr
    repr, readable, recursive = self.format(object, context.copy(),
  File "/Users/amedama/.virtualenvs/py310/lib/python3.10/site-packages/sklearn/utils/_pprint.py", line 189, in format
    return _safe_repr(
  File "/Users/amedama/.virtualenvs/py310/lib/python3.10/site-packages/sklearn/utils/_pprint.py", line 440, in _safe_repr
    params = _changed_params(object)
  File "/Users/amedama/.virtualenvs/py310/lib/python3.10/site-packages/sklearn/utils/_pprint.py", line 93, in _changed_params
    params = estimator.get_params(deep=False)
  File "/Users/amedama/.virtualenvs/py310/lib/python3.10/site-packages/sklearn/base.py", line 211, in get_params
    value = getattr(self, key)
AttributeError: 'TargetEncoder' object has no attribute 'cols'

TargetEncoder takes 'cols' parameter but doesn't save to the attribute.
(That transfers to category_encoders.TargetEncoder internally)
But scikit-learn BaseEstimator#get_params() expects that all __init__() parameters will be saved.
So accessing 'cols' raises AttributeError.

Environment

$ python -V                                   
Python 3.10.9
$ pip list | egrep -i "(nyaggle|scikit-learn)"
nyaggle            0.1.5
scikit-learn       1.2.0

[nyanp]

@momijiame
Thanks for your report! TargetEncoder in nyaggle does not comply with the interface required by scikit-learn, so I think this should be fixed.

In addition, every keyword argument accepted by __init__ should correspond to an attribute on the instance. Scikit-learn relies on this to find the relevant attributes to set on an estimator when doing model selection.
https://scikit-learn.org/stable/developers/develop.html