# -*- coding: utf-8 -*-
from __future__ import absolute_import
from lightning.impl.base import BaseEstimator
from lightning import classification, regression
from sklearn.multiclass import OneVsRestClassifier
from eli5.base import Explanation
from eli5.base_utils import singledispatch
from eli5.sklearn import (
explain_linear_classifier_weights,
explain_linear_regressor_weights,
explain_prediction_linear_classifier,
explain_prediction_linear_regressor
)
from eli5.explain import explain_prediction, explain_weights
[docs]@singledispatch
def explain_weights_lightning(estimator, vec=None, top=20, target_names=None,
targets=None, feature_names=None,
coef_scale=None):
""" Return an explanation of a lightning estimator weights """
return explain_weights_lightning_not_supported(estimator)
@explain_weights.register(BaseEstimator)
def explain_weights_lightning_not_supported(
estimator, vec=None, top=20, target_names=None,
targets=None, feature_names=None,
coef_scale=None):
return Explanation(
estimator=repr(estimator),
error="Error: estimator %r is not supported" % estimator,
)
[docs]@singledispatch
def explain_prediction_lightning(estimator, doc, vec=None, top=None,
target_names=None, targets=None,
feature_names=None, vectorized=False,
coef_scale=None):
""" Return an explanation of a lightning estimator predictions """
return explain_weights_lightning_not_supported(estimator, doc)
@explain_prediction.register(BaseEstimator)
def explain_prediction_lightning_not_supported(
estimator, doc, vec=None, top=None,
target_names=None, targets=None,
feature_names=None, vectorized=False,
coef_scale=None):
return Explanation(
estimator=repr(estimator),
error="Error: estimator %r is not supported" % estimator,
)
@explain_prediction_lightning.register(OneVsRestClassifier)
def explain_prediction_ovr_lightning(clf, doc, **kwargs):
# dispatch OvR to eli5.lightning
# if explain_prediction_lightning is called explicitly
estimator = clf.estimator
func = explain_prediction_lightning.dispatch(estimator.__class__)
return func(clf, doc, **kwargs)
@explain_weights_lightning.register(OneVsRestClassifier)
def explain_weights_ovr_lightning(ovr, **kwargs):
# dispatch OvR to eli5.lightning
# if explain_weights_lightning is called explicitly
estimator = ovr.estimator
func = explain_weights_lightning.dispatch(estimator.__class__)
return func(ovr, **kwargs)
_CLASSIFIERS = [
classification.AdaGradClassifier,
classification.CDClassifier,
classification.FistaClassifier,
classification.LinearSVC,
classification.SAGAClassifier,
classification.SAGClassifier,
classification.SDCAClassifier,
classification.SGDClassifier,
# classification.SVRGClassifier, # tests fail for it
]
_REGRESSORS = [
regression.AdaGradRegressor,
regression.CDRegressor,
regression.FistaRegressor,
regression.LinearSVR,
regression.SAGARegressor,
regression.SAGRegressor,
regression.SDCARegressor,
regression.SGDRegressor,
# regression.SVRGRegressor
]
for clf in _CLASSIFIERS:
explain_weights.register(clf, explain_linear_classifier_weights)
explain_weights_lightning.register(clf, explain_linear_classifier_weights)
explain_prediction.register(clf, explain_prediction_linear_classifier)
explain_prediction_lightning.register(clf, explain_prediction_linear_classifier)
for reg in _REGRESSORS:
explain_weights.register(reg, explain_linear_regressor_weights)
explain_weights_lightning.register(reg, explain_linear_regressor_weights)
explain_prediction.register(reg, explain_prediction_linear_regressor)
explain_prediction_lightning.register(reg, explain_prediction_linear_regressor)