h$,fdZddlmZmZddlZddlmZddlm Z ddl m Z m Z m Z ddlmZmZmZdd lmZdd lmZGd d e e e e Zy)z)Principal Component Analysis Base Classes)ABCMetaabstractmethodN)linalg)issparse) BaseEstimatorClassNamePrefixFeaturesOutMixinTransformerMixin)_add_to_diagonaldevice get_namespace)_implicit_column_offset)check_is_fittedcJeZdZdZdZdZed dZdZdZ e dZ y) _BasePCAzwBase class for PCA methods. Warning: This class should not be used directly. Use derived classes instead. c t|j\}}|j}|j}|jr)||j |ddt j fz}||jz }|j||jkD||jdt|}|j|z|z}t||j||S)asCompute data covariance with the generative model. ``cov = components_.T * S**2 * components_ + sigma2 * eye(n_features)`` where S**2 contains the explained variances, and sigma2 contains the noise variances. Returns ------- cov : array of shape=(n_features, n_features) Estimated covariance of data. Nr ) r components_explained_variance_whitensqrtnpnewaxisnoise_variance_whereasarrayr Tr )selfxp_rexp_var exp_var_diffcovs ;lib/python3.12/site-packages/sklearn/decomposition/_base.pyget_covariancez_BasePCA.get_covariance sd../A&& ** ;;%2:: 0F(GGK!5!55 xx d** *  JJs6'?J 3 }}|+{:d22B7 c lt|j\}}|jjd}|jdk(r|j ||j z S|r|j j}nt j}|j dk(r||jS|j}|j}|jr)||j|ddtjfz}||j z }|j||j kD||jdt!|}||j"z|j z }t%|d|z ||j"||z|z}||j dz z}t%|d|j z ||S)a8Compute data precision matrix with the generative model. Equals the inverse of the covariance but computed with the matrix inversion lemma for efficiency. Returns ------- precision : array, shape=(n_features, n_features) Estimated precision of data. rrNrg?r)r rshape n_components_eyerrinvr&rrrrrrrr rr ) rr is_array_api_compliant n_features linalg_invrr"r# precisions r% get_precisionz_BasePCA.get_precision<s&343C3C%D" "%%++A.     "66*%(<(<< < !JJ   3 &d1134 4&& ** ;;%2:: 0F(GGK!5!55 xx d** *  JJs6'?J 3  +--/$2F2FF C,$6;MMJy$99KG t++Q.// C$*>*>$>Cr'Ncy)aPlaceholder for fit. Subclasses should implement this method! Fit the model with X. Parameters ---------- X : array-like of shape (n_samples, n_features) Training data, where `n_samples` is the number of samples and `n_features` is the number of features. Returns ------- self : object Returns the instance itself. N)rXys r%fitz _BasePCA.fitisr'ct|\}}t||j|d|j|jgd}|j 1t |rt||j }n||j z }||jjz}|jr||j|jz}|S)aApply dimensionality reduction to X. X is projected on the first principal components previously extracted from a training set. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) New data, where `n_samples` is the number of samples and `n_features` is the number of features. Returns ------- X_new : array-like of shape (n_samples, n_components) Projection of X in the first principal components, where `n_samples` is the number of samples and `n_components` is the number of the components. )csrcscF) accept_sparsedtypereset) r r_validate_datafloat64float32mean_rrrrrrr)rr5r r! X_transformeds r% transformz_BasePCA.transform{s$a A    ^BJJ 3KSX   :: !{+Atzz: ND,,... ;; RWWT%=%=> >Mr'c t|\}}|jrO|j|jddtj f|j z}||z|jzS||j z|jzS)aTransform data back to its original space. In other words, return an input `X_original` whose transform would be X. Parameters ---------- X : array-like of shape (n_samples, n_components) New data, where `n_samples` is the number of samples and `n_components` is the number of components. Returns ------- X_original array-like of shape (n_samples, n_features) Original data, where `n_samples` is the number of samples and `n_features` is the number of features. Notes ----- If whitening is enabled, inverse_transform will compute the exact inverse operation, which includes reversing whitening. N)r rrrrrrrA)rr5r r!scaled_componentss r%inverse_transformz_BasePCA.inverse_transformsy,a A ;;00BJJ?@4CSCSS ((4::5 5t'''$**4 4r'c4|jjdS)z&Number of transformed output features.r)rr*)rs r%_n_features_outz_BasePCA._n_features_outs%%a((r')N) __name__ __module__ __qualname____doc__r&r2rr7rCrFpropertyrHr4r'r%rrsF 8+Z  "!F5@))r'r) metaclass)rLabcrrnumpyrscipyr scipy.sparserbaserr r utils._array_apir r r utils.sparsefuncsrutils.validationrrr4r'r%rWs?/(!SSFF7.j)#%5}PWj)r'