def __init__(self, wrapped_node, in_args_transform=None):
if not hasattr(wrapped_node, "transform"):
raise ValueError("wrapped_node should implement transform")
super(TransformNode, self).__init__(wrapped_node=wrapped_node)
self.in_args_transform = \
_func_get_args_names(self.wrapped_node.transform) \
if in_args_transform is None else in_args_transform
def __init__(self, wrapped_node, in_args_transform=None):
if not hasattr(wrapped_node, "transform"):
raise ValueError("wrapped_node should implement transform")
super(TransformNode, self).__init__(wrapped_node=wrapped_node)
self.in_args_transform = \
_func_get_args_names(self.wrapped_node.transform) \
if in_args_transform is None else in_args_transform
def __init__(self,
wrapped_node,
in_args_fit=None,
in_args_transform=None,
in_args_predict=None,
out_args_predict=None):
is_fit_estimator = False
if hasattr(wrapped_node, "fit") and hasattr(wrapped_node, "transform"):
is_fit_estimator = True
elif hasattr(wrapped_node, "fit") and hasattr(wrapped_node, "predict"):
is_fit_estimator = True
if not is_fit_estimator:
raise ValueError("%s should implement fit and transform or fit "
"and predict" % wrapped_node.__class__.__name__)
super(Estimator, self).__init__(wrapped_node=wrapped_node)
if in_args_fit:
self.in_args_fit = in_args_fit
else:
self.in_args_fit = _func_get_args_names(self.wrapped_node.fit)
# Internal Estimator
if hasattr(wrapped_node, "transform"):
if in_args_transform:
self.in_args_transform = in_args_transform
else:
self.in_args_transform = \
_func_get_args_names(self.wrapped_node.transform)
# Leaf Estimator
if hasattr(wrapped_node, "predict"):
if in_args_predict:
self.in_args_predict = in_args_predict
else:
self.in_args_predict = \
_func_get_args_names(self.wrapped_node.predict)
if out_args_predict is None:
fit_predict_diff = list(set(self.in_args_fit).difference(
self.in_args_predict))
if len(fit_predict_diff) > 0:
self.out_args_predict = fit_predict_diff
else:
self.out_args_predict = self.in_args_predict
else:
self.out_args_predict = out_args_predict
def __init__(self,
wrapped_node,
in_args_fit=None,
in_args_transform=None,
in_args_predict=None,
out_args_predict=None):
is_fit_estimator = False
if hasattr(wrapped_node, "fit") and hasattr(wrapped_node, "transform"):
is_fit_estimator = True
elif hasattr(wrapped_node, "fit") and hasattr(wrapped_node, "predict"):
is_fit_estimator = True
if not is_fit_estimator:
raise ValueError("%s should implement fit and transform or fit "
"and predict" % wrapped_node.__class__.__name__)
super(Estimator, self).__init__(wrapped_node=wrapped_node)
if in_args_fit:
self.in_args_fit = in_args_fit
else:
self.in_args_fit = _func_get_args_names(self.wrapped_node.fit)
# Internal Estimator
if hasattr(wrapped_node, "transform"):
if in_args_transform:
self.in_args_transform = in_args_transform
else:
self.in_args_transform = \
_func_get_args_names(self.wrapped_node.transform)
# Leaf Estimator
if hasattr(wrapped_node, "predict"):
if in_args_predict:
self.in_args_predict = in_args_predict
else:
self.in_args_predict = \
_func_get_args_names(self.wrapped_node.predict)
if out_args_predict is None:
fit_predict_diff = list(set(self.in_args_fit).difference(
self.in_args_predict))
if len(fit_predict_diff) > 0:
self.out_args_predict = fit_predict_diff
else:
self.out_args_predict = self.in_args_predict
else:
self.out_args_predict = out_args_predict
def __init__(self, estimator, in_args_fit=None, in_args_predict=None,
out_args_predict=None):
'''
Parameters
----------
estimator: any class contains fit and predict functions
any class implements fit and predict
in_args_fit: list of strings
names of input arguments of the fit method. If missing discover
discover it automatically.
in_args_predict: list of strings
names of input arguments of the predict method. If missing,
discover it automatically.
out_args_predict: list of strings
names of output arguments of the predict method. If missing,
discover it automatically by self.in_args_fit - in_args_predict.
If not differences (such with PCA with fit(X) and predict(X))
use in_args_predict.
'''
if not hasattr(estimator, "fit") or not \
hasattr(estimator, "predict"):
raise ValueError("estimator should implement fit and predict")
super(LeafEstimator, self).__init__(estimator=estimator)
self.in_args_fit = _func_get_args_names(self.estimator.fit) \
if in_args_fit is None else in_args_fit
self.in_args_predict = _func_get_args_names(self.estimator.predict) \
if in_args_predict is None else in_args_predict
if out_args_predict is None:
fit_predict_diff = list(set(self.in_args_fit).difference(
self.in_args_predict))
if len(fit_predict_diff) > 0:
self.out_args_predict = fit_predict_diff
else:
self.out_args_predict = self.in_args_predict
else:
self.out_args_predict = out_args_predict
def __init__(self, estimator, in_args_fit=None, in_args_transform=None):
"""
Parameters
----------
estimator: any class contains fit and transform functions
any class implements fit and transform
in_args_fit: list of strings
names of input arguments of the fit method. If missing discover
discover it automatically.
in_args_transform: list of strings
names of input arguments of the tranform method. If missing,
discover it automatically.
"""
if not hasattr(estimator, "fit") or not \
hasattr(estimator, "transform"):
raise ValueError("estimator should implement fit and transform")
super(InternalEstimator, self).__init__(estimator=estimator)
self.in_args_fit = _func_get_args_names(self.estimator.fit) \
if in_args_fit is None else in_args_fit
self.in_args_transform = \
_func_get_args_names(self.estimator.transform) \
if in_args_transform is None else in_args_transform