Top-level functions and classes

sklearn_xarray

wrap(estimator, reshapes=None, sample_dim=None, compat=False, **kwargs)

Wrap an sklearn estimator for xarray objects.

Parameters

estimatorsklearn estimator class or instance

The estimator this instance wraps around.

reshapesstr or dict, optional

The dimension(s) reshaped by this estimator. Any coordinates in the DataArray along these dimensions will be dropped. If the estimator drops this dimension (e.g. a binary classifier returning a 1D vector), the dimension itself will also be dropped.

You can specify multiple dimensions mapping to multiple new dimensions with a dict whose items are lists of reshaped dimensions, e.g. {'new_feature': ['feature_1', 'feature_2'], ...}

sample_dimstr, optional

The name of the dimension that represents the samples. By default, the wrapper will assume that this is the first dimension in the array.

compatbool, default False

If True, the method will return a CompatEstimatorWrapper instead of an EstimatorWrapper. This might be necessary when the estimator defines parameters with the same name as the wrapper.

Returns

A wrapped estimator.

class EstimatorWrapper(estimator=None, reshapes=None, sample_dim=None, **kwargs)

A wrapper around sklearn estimators compatible with xarray objects.

Parameters

estimatorsklearn estimator

The estimator instance this instance wraps around.

reshapesstr or dict, optional

The dimension(s) reshaped by this estimator. Any coordinates in the DataArray along these dimensions will be dropped. If the estimator drops this dimension (e.g. a binary classifier returning a 1D vector), the dimension itself will also be dropped.

You can specify multiple dimensions mapping to multiple new dimensions with a dict whose items are lists of reshaped dimensions, e.g. {'new_feature': ['feature_1', 'feature_2'], ...}

sample_dimstr, optional

The name of the dimension that represents the samples. By default, the wrapper will assume that this is the first dimension in the array.

Methods

fit(self, X[, y])	A wrapper around the fitting function.
get_params(self[, deep])	Get parameters for this estimator.
set_params(self, **params)	Set the parameters of this estimator.

fit(self, X, y=None, **fit_params)

A wrapper around the fitting function.

Parameters

Xxarray DataArray, Dataset other other array-like

The training input samples.

yxarray DataArray, Dataset other other array-like

The target values.

Returns

Returns self.

get_params(self, deep=True)

Get parameters for this estimator.

Parameters

deepboolean, optional

If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns

paramsmapping of string to any

Parameter names mapped to their values.

set_params(self, **params)

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as pipelines). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Returns

self

class ClassifierWrapper(estimator=None, reshapes=None, sample_dim=None, **kwargs)

A wrapper around sklearn classifiers compatible with xarray objects.

Parameters

estimatorsklearn estimator

The estimator this instance wraps around.

reshapesstr or dict, optional

The dimension reshaped by this estimator.

Methods

fit(self, X[, y])	A wrapper around the fitting function.
get_params(self[, deep])	Get parameters for this estimator.
predict(self, X)	A wrapper around the prediction function.
score(self, X, y[, sample_weight])	Returns the score of the prediction.
set_params(self, **params)	Set the parameters of this estimator.

class RegressorWrapper(estimator=None, reshapes=None, sample_dim=None, **kwargs)

A wrapper around sklearn regressors compatible with xarray objects.

Parameters

estimatorsklearn estimator

The estimator this instance wraps around.

reshapesstr or dict, optional

The dimension reshaped by this estimator.

Methods

fit(self, X[, y])	A wrapper around the fitting function.
get_params(self[, deep])	Get parameters for this estimator.
predict(self, X)	A wrapper around the prediction function.
score(self, X, y[, sample_weight])	Returns the score of the prediction.
set_params(self, **params)	Set the parameters of this estimator.

class TransformerWrapper(estimator=None, reshapes=None, sample_dim=None, **kwargs)

A wrapper around sklearn transformers compatible with xarray objects.

Parameters

estimatorsklearn estimator

The estimator this instance wraps around.

reshapesstr or dict, optional

The dimension reshaped by this estimator.

Methods

fit(self, X[, y])	A wrapper around the fitting function.
fit_transform(self, X[, y])	A wrapper around the fit_transform function.
get_params(self[, deep])	Get parameters for this estimator.
inverse_transform(self, X)	A wrapper around the inverse transformation function.
set_params(self, **params)	Set the parameters of this estimator.
transform(self, X)	A wrapper around the transformation function.

class Target(coord=None, transform_func=None, transformer=None, lazy=False, dim=None, reduce_func=None)

A pointer to xarray coordinates or variables to be used as a target.

Parameters

coordstr, optional

The coordinate or variable that holds the data of the target. If not specified, the target will be the entire DataArray/Dataset.

transform_funccallable, optional

A function that transforms the coordinate values to an sklearn-compatible type and shape. If not specified, the coordinate(s) will be used as-is.

transformersklearn transformer, optional

Deprecated, use transform_func=Transformer().fit_transform instead.

lazybool, optinonal

If true, the target coordinate is only transformed by the transformer when needed. The transformer can implement a get_transformed_shape method that returns the shape after the transformation of the provided coordinate without actually transforming the data.

dimstr or sequence of str, optional

When set, multi-dimensional coordinates will be reduced to this dimension/these dimensions.

reduce_funccallable, optional

A callable that reduces the coordinate(s) to the dimension(s) in dim. If not specified, the values along dimensions not in dim will be reduced to the first element in each of these dimensions.

Attributes

ndim

The number of dimensions of the transformed target.

shape

The shape of the transformed target.

Methods

__call__(self, X)	Call self as a function.
assign_to(self, X)	Assign this target to a DataArray or Dataset.

assign_to(self, X)

Assign this target to a DataArray or Dataset.

Parameters

Xxarray DataArray or Dataset

The data whose coordinate is used as the target.

Returns

self:

The target itself.

property ndim

The number of dimensions of the transformed target.

property shape

The shape of the transformed target.