a j%@sdZddlZddlmZmZddlmZddlZddl m Z m Z ddl m Z ddlmZmZmZdd lmZdd lmZdd lmZmZmZmZGd d d e edZdddZdS)zGeneric feature selection mixinN)ABCMetaabstractmethod) attrgetter) csc_matrixissparse)TransformerMixin)_safe_indexing check_arraysafe_sqr)_get_output_config)get_tags)_check_feature_names_in _is_pandas_dfcheck_is_fitted validate_datac@sHeZdZdZdddZeddZddZd d Zd d Z dddZ d S) SelectorMixina Transformer mixin that performs feature selection given a support mask This mixin provides a feature selector implementation with `transform` and `inverse_transform` functionality given an implementation of `_get_support_mask`. Examples -------- >>> import numpy as np >>> from sklearn.datasets import load_iris >>> from sklearn.base import BaseEstimator >>> from sklearn.feature_selection import SelectorMixin >>> class FeatureSelector(SelectorMixin, BaseEstimator): ... def fit(self, X, y=None): ... self.n_features_in_ = X.shape[1] ... return self ... def _get_support_mask(self): ... mask = np.zeros(self.n_features_in_, dtype=bool) ... mask[:2] = True # select the first two features ... return mask >>> X, y = load_iris(return_X_y=True) >>> FeatureSelector().fit_transform(X, y).shape (150, 2) FcCs|}|s|St|dS)a Get a mask, or integer index, of the features selected. Parameters ---------- indices : bool, default=False If True, the return value will be an array of integers, rather than a boolean mask. Returns ------- support : array An index that selects the retained features from a feature vector. If `indices` is False, this is a boolean array of shape [# input features], in which an element is True iff its corresponding feature is selected for retention. If `indices` is True, this is an integer array of shape [# output features] whose values are indices into the input feature vector. r)_get_support_masknpwhere)selfindicesmaskrEE:\soccer\ml_env\Lib\site-packages\sklearn/feature_selection/_base.py get_support4szSelectorMixin.get_supportcCsdS)a Get the boolean mask indicating which features are selected Returns ------- support : boolean array of shape [# input features] An element is True iff its corresponding feature is selected for retention. Nr)rrrrrKszSelectorMixin._get_support_maskc CsJtd|dd}|dkot|}t||ddt|jj |dd}||S) aBReduce X to the selected features. Parameters ---------- X : array of shape [n_samples, n_features] The input samples. Returns ------- X_r : array of shape [n_samples, n_selected_features] The input samples with only the selected features. transform) estimatorZdensedefaultNZcsrF)dtypeZ accept_sparseZensure_all_finiteZskip_check_arrayreset)r rrr Z input_tags allow_nan _transform)rXZoutput_config_denseZ preserve_XrrrrWs  zSelectorMixin.transformcCsj|}|s\tdtt|dr<|jddddfStjd|j d |j ddfSt ||ddS)z"Reduce X to the selected features.zYNo features were selected: either the data is too noisy or the selection test too strict.ilocNrraxis) ranywarningswarn UserWarninghasattrr$remptyrreshapeshaper )rr#rrrrr"vs  zSelectorMixin._transformcCst|rx|}|t|jdd}|}tdgt |g}t |j |j |f|j dt|df|jd}|S|}t|dd}||j dkrtd|jdkr|dddf}tj|j d|jf|jd}||dd|f<|S)aReverse the transformation operation. Parameters ---------- X : array of shape [n_samples, n_selected_features] The input samples. Returns ------- X_r : array of shape [n_samples, n_original_features] `X` with columns of zeros inserted where features would have been removed by :meth:`transform`. r&r)r0rNr%z,X has a different shape than during fitting.)rZtocscinverse_transformrdiffindptrr/ZravelZ concatenateZcumsumrdatarr0lenrrr sum ValueErrorndimZzerossize)rr#itZ col_nonzerosr4ZXtZsupportrrrr2s(   zSelectorMixin.inverse_transformNcCst|t||}||S)aMask feature names according to selected features. Parameters ---------- input_features : array-like of str or None, default=None Input features. - If `input_features` is `None`, then `feature_names_in_` is used as feature names in. If `feature_names_in_` is not defined, then the following input feature names are generated: `["x0", "x1", ..., "x(n_features_in_ - 1)"]`. - If `input_features` is an array-like, then `input_features` must match `feature_names_in_` if `feature_names_in_` is defined. Returns ------- feature_names_out : ndarray of str objects Transformed feature names. )rrr)rZinput_featuresrrrget_feature_names_outs z#SelectorMixin.get_feature_names_out)F)N) __name__ __module__ __qualname____doc__rrrrr"r2r<rrrrrs  (r) metaclassr&cCst|trZ|dkrPt|dr&td}qXt|dr:td}qXtd|jjdqjt|}nt|sjtd||}|dur~|S|dkr|jd krt |}qt j j |d |d }n6|d kr|jd krt |}qt |jd d }ntd|S)a Retrieve and aggregate (ndim > 1) the feature importances from an estimator. Also optionally applies transformation. Parameters ---------- estimator : estimator A scikit-learn estimator from which we want to get the feature importances. getter : "auto", str or callable An attribute or a callable to get the feature importance. If `"auto"`, `estimator` is expected to expose `coef_` or `feature_importances`. transform_func : {"norm", "square"}, default=None The transform to apply to the feature importances. By default (`None`) no transformation is applied. norm_order : int, default=1 The norm order to apply when `transform_func="norm"`. Only applied when `importances.ndim > 1`. Returns ------- importances : ndarray of shape (n_features,) The features importances, optionally transformed. autoZcoef_Zfeature_importances_z;when `importance_getter=='auto'`, the underlying estimator z should have `coef_` or `feature_importances_` attribute. Either pass a fitted estimator to feature selector or call fit before calling transform.z4`importance_getter` has to be a string or `callable`Nnormr&r)r(ordZsquarer'zpValid values for `transform_func` are None, 'norm' and 'square'. Those two transformation are only supported now) isinstancestrr-rr8 __class__r=callabler9rabsZlinalgrCr r7)rgetterZtransform_funcZ norm_orderZ importancesrrr_get_feature_importancess:          rK)Nr&)r@r*abcrroperatorrnumpyrZ scipy.sparserrbaserutilsr r r Zutils._set_outputr Z utils._tagsr Zutils.validationrrrrrrKrrrrs    /