B \/@sdZddlmZddlmZddlZddlZddlmZm Z ddl Z ddl m Z mZddlmZdd lmZd d lmZd d lmZd d lmZd dlmZmZmZd dlmZd dlm Z d dl!m"Z"d dl#m$Z$d dl%m&Z&m'Z'm(Z(d dl%m)Z*e+ej,ej,e-e.dddddddddddd d!d"d#d$d%d&d'd(gd)Z/e+gge.d)Z0d*d+Z1d,d-Z2Gd.d/d/e"3eeZ4d0d1Z5Gd2d3d3e6Z7d4d5Z8Gd6d7d7e6Z9Gd8d9d9e6Z:Gd:d;d;e6Z;Gdz,Base and mixin classes for nearest neighbors)partial) LooseVersionN)ABCMetaabstractmethod) csr_matrixissparse)BallTree)KDTree) BaseEstimator)pairwise_distances_chunked)PAIRWISE_DISTANCE_FUNCTIONS) check_X_y check_arraygen_even_slices)check_classification_targets)check_is_fitted)six)DataConversionWarning)Paralleldelayedeffective_n_jobs) __version__Z braycurtisZcanberra chebyshevZ correlationZcosineZdiceZhammingZjaccardZ kulsinskiZ mahalanobisZmatching minkowskiZrogerstanimotoZ russellraoZ seuclideanZ sokalmichenerZ sokalsneathZ sqeuclideanZyule wminkowski) ball_treekd_treebrutecCs$|dkr |St|r|StddS)z$Check to make sure weights are valid)NuniformdistancezOweights not recognized: should be 'uniform', 'distance', or a callable functionN)callable ValueError)weightsr%5lib/python3.7/site-packages/sklearn/neighbors/base.py_check_weights3s r'c Cs|dkr dS|dkr|jttkrhxt|D]4\}}t|drVd|krV|dk||<q.d|||<q.WnDtjddd|}WdQRXt|}tj|d d }||||<|St|r||St d dS) axGet the weights from an array of distances and a parameter ``weights`` Parameters =========== dist : ndarray The input distances weights : {'uniform', 'distance' or a callable} The kind of weighting used Returns ======== weights_arr : array of the same shape as ``dist`` if ``weights == 'uniform'``, then returns None )Nr Nr! __contains__gg?ignore)Zdivider)axiszOweights not recognized: should be 'uniform', 'distance', or a callable function) dtypenpobject enumeratehasattrZerrstateZisinfanyr"r#)distr$Z point_dist_iZ point_distZinf_maskZinf_rowr%r%r& _get_weights>s"  r2c @s:eZdZdZedddZd d Zd d Zed dZ dS) NeighborsBasez,Base class for nearest neighbors estimators.Nautorr c Cs<||_||_||_||_||_||_||_||_|dS)N) n_neighborsradius algorithm leaf_sizemetric metric_paramspn_jobs_check_algorithm_metric) selfr6r7r8r9r:r<r;r=r%r%r&__init__mszNeighborsBase.__init__cCs|jdkrtd|j|jdkrV|jdkr2d}q\t|jsJ|jtdkrPd}q\d}n|j}t|jr|jdkrtd|jn |jt|krtd |j|f|jdk rd |jkrtjd td d |jd }n|j }|jdkr|dkrtddS)N)r4rrrzunrecognized algorithm: '%s'r4 precomputedrrrz7kd_tree algorithm does not support callable metric '%s'zMetric '%s' not valid. Use sorted(sklearn.neighbors.VALID_METRICS['%s']) to get valid options. Metric can also be a callable function.r<z\Parameter p is found in metric_params. The corresponding parameter from __init__ is ignored.) stacklevel)rrrz/p must be greater than one for minkowski metric) r8r#r:r" VALID_METRICSr;warningswarn SyntaxWarningr<)r?Z alg_check effective_pr%r%r&r>|s2         z%NeighborsBase._check_algorithm_metriccCs||jdkri|_n |j|_|jd|j}|jdkrJ||jd<|j|_|jdkr|jdd}|dkr|t dn<|dkrd|_n,|dkrd|_n|t j krd |_n ||jd<t |t r|j|_|j|_|j|_|St |tr|j|_||_d |_|St |tr&|j|_||_d |_|St|d d }|jd}|dkrNt dt|r|jdkrntd|jtdkrt|jst d|j||_d|_d|_|S|j|_||_|jdkrF|jdks|j|jjddkr@|jdkr@|jtd krd |_n*t|js0|jtd kr8d |_nd|_nd|_|jd krrt||jfd|ji|j|_nN|jd krt||jfd|ji|j|_n"|jdkrd|_nt d|j|jdk r|jdkrt d|jn(t t |jt j!st"dt |j|S)Nr<)rrrr rz/p must be greater than one for minkowski metricZ manhattan euclideanrrrcsr) accept_sparserz n_samples must be greater than 0)r4rz4cannot use tree with sparse input: using brute forcerzMetric '%s' not valid for sparse input. Use sorted(sklearn.neighbors.VALID_METRICS_SPARSE['brute']) to get valid options. Metric can also be a callable function.r4rAr:zalgorithm = '%s' not recognizedz Expected n_neighbors > 0. Got %dz7n_neighbors does not take %s value, enter integer value)#r>r;effective_metric_params_copygetr<r:effective_metric_popr#r,inf isinstancer3_fit_X_tree _fit_methodr datar rshaperr8rErFVALID_METRICS_SPARSEr"r6rDr9 issubdtypetypeinteger TypeError)r?XrHr< n_samplesr%r%r&_fits                                 zNeighborsBase._fitcCs |jdkS)NrA)r:)r?r%r%r& _pairwiseszNeighborsBase._pairwise)NNr4r5rr NN) __name__ __module__ __qualname____doc__rr@r>r_propertyr`r%r%r%r&r3js (tr3cCs||||S)zHelper for the Parallel calls in KNeighborsMixin.kneighbors The Cython method tree.query is not directly picklable by cloudpickle under PyPy. )Zquery)treerVr6return_distancer%r%r&_tree_query_parallel_helpersrhc@s,eZdZdZddZd ddZd d d ZdS) KNeighborsMixinzMixin for k-neighbors searchescCst|jddddf}tj||ddd}|ddd|f}||t|||ff}|r|jdkrt|||f|f}q|||f|f}n|}|S)aWReduce a chunk of distances to the nearest neighbors Callback to :func:`sklearn.metrics.pairwise.pairwise_distances_chunked` Parameters ---------- dist : array of shape (n_samples_chunk, n_samples) start : int The index in X which the first row of dist corresponds to. n_neighbors : int return_distance : bool Returns ------- dist : array of shape (n_samples_chunk, n_neighbors), optional Returned only if return_distance neigh : array of shape (n_samples_chunk, n_neighbors) rNr)r*rI)r,arangerWZ argpartitionZargsortrOsqrt)r?r1startr6rg sample_range neigh_indresultr%r%r&_kneighbors_reduce_func*s z'KNeighborsMixin._kneighbors_reduce_funcNTcstddkrjn8dkr0tdn"tttjsRtdtdk rld}tddnd }j d 7j j d}|krtd |fj \}}t |dddf}t j } jd kr$tjd } jdkrdd inj} ttj f| j| d| } njdkrtrHtdjtttdk} tjdksj| r| rtdnd}tt|dddi}nttddi}t| f|fddtj d| D} ntdrt| \}}t|t|f} n t| } |s| Sr&| \}}n| }||k}tj |d d}d|dddf|<t!|||d f}rt!|||d f}||fS|SdS)aFinds the K-neighbors of a point. Returns indices of and distances to the neighbors of each point. Parameters ---------- X : array-like, shape (n_query, n_features), or (n_query, n_indexed) if metric == 'precomputed' The query point or points. If not provided, neighbors of each indexed point are returned. In this case, the query point is not considered its own neighbor. n_neighbors : int Number of neighbors to get (default is the value passed to the constructor). return_distance : boolean, optional. Defaults to True. If False, distances will not be returned Returns ------- dist : array Array representing the lengths to points, only present if return_distance=True ind : array Indices of the nearest points in the population matrix. Examples -------- In the following example, we construct a NeighborsClassifier class from an array representing our data set and ask who's the closest point to [1,1,1] >>> samples = [[0., 0., 0.], [0., .5, 0.], [1., 1., .5]] >>> from sklearn.neighbors import NearestNeighbors >>> neigh = NearestNeighbors(n_neighbors=1) >>> neigh.fit(samples) # doctest: +ELLIPSIS NearestNeighbors(algorithm='auto', leaf_size=30, ...) >>> print(neigh.kneighbors([[1., 1., 1.]])) # doctest: +ELLIPSIS (array([[0.5]]), array([[2]])) As you can see, it returns [[0.5]], and [[2]], which means that the element is at distance 0.5 and is the third element of samples (indexes start at 0). You can also query for multiple points: >>> X = [[0., 1., 0.], [1., 0., 1.]] >>> neigh.kneighbors(X, return_distance=False) # doctest: +ELLIPSIS array([[1], [2]]...) rUNrz Expected n_neighbors > 0. Got %dz7n_neighbors does not take %s value, enter integer valueFrJ)rKTrzHExpected n_neighbors <= n_samples, but n_samples = %d, n_neighbors = %dr)r6rgrIsquared) reduce_funcr:r=)rrzQ%s does not work with sparse matrices. Densify the data, or set algorithm='brute'z0.12)rB) check_picklebackend threadingpreferthreadsc3s"|]}j|VqdS)N)rT).0s)r] delayed_queryr6rgr?r%r& sz-KNeighborsMixin.kneighbors..z$internal: _fit_method not recognized)r*)"rr6r#r,rYrZr[r\rrSrWrjrr=rUrrprOrLlistr rrjoblib_versionsys version_inforrhrrzipZvstackallreshape)r?r]r6rgquery_is_trainZ train_sizer^_rmr=rrkwdsroZ old_joblibrsparallel_kwargsr1rnZ sample_maskZ dup_gr_nbrsr%)r]rzr6rgr?r& kneighborsMs4                   zKNeighborsMixin.kneighbors connectivityc Cs|dkr|j}|dk r.t|dd}|jd}n |jjd}|jjd}||}td|d|}|dkrt||}|j||dd} n4|d kr|j||d d\}} t|}n t d |t || |f||fd } | S) a,Computes the (weighted) graph of k-Neighbors for points in X Parameters ---------- X : array-like, shape (n_query, n_features), or (n_query, n_indexed) if metric == 'precomputed' The query point or points. If not provided, neighbors of each indexed point are returned. In this case, the query point is not considered its own neighbor. n_neighbors : int Number of neighbors for each sample. (default is value passed to the constructor). mode : {'connectivity', 'distance'}, optional Type of returned matrix: 'connectivity' will return the connectivity matrix with ones and zeros, in 'distance' the edges are Euclidean distance between points. Returns ------- A : sparse matrix in CSR format, shape = [n_samples, n_samples_fit] n_samples_fit is the number of samples in the fitted data A[i, j] is assigned the weight of edge that connects i to j. Examples -------- >>> X = [[0], [3], [1]] >>> from sklearn.neighbors import NearestNeighbors >>> neigh = NearestNeighbors(n_neighbors=2) >>> neigh.fit(X) # doctest: +ELLIPSIS NearestNeighbors(algorithm='auto', leaf_size=30, ...) >>> A = neigh.kneighbors_graph(X) >>> A.toarray() array([[1., 0., 1.], [0., 1., 1.], [1., 0., 1.]]) See also -------- NearestNeighbors.radius_neighbors_graph NrJ)rKrrrF)rgr!TzRUnsupported mode, must be one of "connectivity" or "distance" but got "%s" instead)rW) r6rrWrSr,rjonesrravelr#r) r?r]r6mode n_samples1 n_samples2Z n_nonzeroA_indptrA_dataA_indkneighbors_graphr%r%r&rs,,      z KNeighborsMixin.kneighbors_graph)NNT)NNr)rarbrcrdrprrr%r%r%r&ri's # #ricCs||||S)zHelper for the Parallel calls in RadiusNeighborsMixin.radius_neighbors The Cython method tree.query_radius is not directly picklable by cloudpickle under PyPy. )Z query_radius)rfrVr7rgr%r%r&"_tree_query_radius_parallel_helper>src@s,eZdZdZddZd ddZd d d ZdS) RadiusNeighborsMixinz)Mixin for radius-based neighbors searchescs`fdd|D|rX|jdkr8fddt|D}nfddt|D}|f}n}|S)a<Reduce a chunk of distances to the nearest neighbors Callback to :func:`sklearn.metrics.pairwise.pairwise_distances_chunked` Parameters ---------- dist : array of shape (n_samples_chunk, n_samples) start : int The index in X which the first row of dist corresponds to. radius : float return_distance : bool Returns ------- dist : list of n_samples_chunk 1d arrays, optional Returned only if return_distance neigh : list of n_samples_chunk 1d arrays csg|]}t|kdqS)r)r,where)rxd)r7r%r& ^szFRadiusNeighborsMixin._radius_neighbors_reduce_func..rIcs"g|]\}}t||qSr%)r,rk)rxir)rnr%r&rbscsg|]\}}||qSr%r%)rxrr)rnr%r&res)rOr.)r?r1rlr7rgresultsr%)rnr7r&_radius_neighbors_reduce_funcJs     z2RadiusNeighborsMixin._radius_neighbors_reduce_funcNTcs~tddk r$d}tddn d}jdkr<jjdkr0jdkrd9d di}nj}tjd }t jf|jj d |}rt |\}} t |g} t | g} t jt| d d } | | dd<t jt| d d } | | dd<| | f}n(t |g} t jt| d d }| |dd<n؈jdkrtrTtdjtj }tttdkrttddddi}nttddi}t|f|fddtjd|D}rtt |\} } t | t | f}n t |}ntd|s|Sr"|\} } n|} x@t| D]4\}}||k}||| |<r0| ||| |<q0Wrv| | fS| SdS)a Finds the neighbors within a given radius of a point or points. Return the indices and distances of each point from the dataset lying in a ball with size ``radius`` around the points of the query array. Points lying on the boundary are included in the results. The result points are *not* necessarily sorted by distance to their query point. Parameters ---------- X : array-like, (n_samples, n_features), optional The query point or points. If not provided, neighbors of each indexed point are returned. In this case, the query point is not considered its own neighbor. radius : float Limiting distance of neighbors to return. (default is the value passed to the constructor). return_distance : boolean, optional. Defaults to True. If False, distances will not be returned Returns ------- dist : array, shape (n_samples,) of arrays Array representing the distances to each point, only present if return_distance=True. The distance values are computed according to the ``metric`` constructor parameter. ind : array, shape (n_samples,) of arrays An array of arrays of indices of the approximate nearest points from the population matrix that lie within a ball of size ``radius`` around the query points. Examples -------- In the following example, we construct a NeighborsClassifier class from an array representing our data set and ask who's the closest point to [1, 1, 1]: >>> import numpy as np >>> samples = [[0., 0., 0.], [0., .5, 0.], [1., 1., .5]] >>> from sklearn.neighbors import NearestNeighbors >>> neigh = NearestNeighbors(radius=1.6) >>> neigh.fit(samples) # doctest: +ELLIPSIS NearestNeighbors(algorithm='auto', leaf_size=30, ...) >>> rng = neigh.radius_neighbors([[1., 1., 1.]]) >>> print(np.asarray(rng[0][0])) # doctest: +ELLIPSIS [1.5 0.5] >>> print(np.asarray(rng[1][0])) # doctest: +ELLIPSIS [1 2] The first array returned contains the distances to all points which are closer than 1.6, while the second array returned contains their indices. In general, multiple points can be queried at the same time. Notes ----- Because the number of neighbors of each point is not necessarily equal, the results for multiple query points cannot be fit in a standard data array. For efficiency, `radius_neighbors` returns arrays of objects, where each object is a 1D array of indices or distances. rUNFrJ)rKTrrIrq)r7rg)rrr:r=r-)r+)rrzQ%s does not work with sparse matrices. Densify the data, or set algorithm='brute'z0.12)rsrtrurvrwc3s"|]}j|VqdS)N)rT)rxry)r]rzr7rgr?r%r&r{sz8RadiusNeighborsMixin.radius_neighbors..rz$internal: _fit_method not recognized)rrrSr7rUrOrLrrr r=rsumr,emptylenrr#rrr}rrrrrWtupleZhstackr.)r?r]r7rgrrrrrZ dist_chunksZneigh_ind_chunksZ dist_listZneigh_ind_listr1rnr=rZindZ ind_neighbormaskr%)r]rzr7rgr?r&radius_neighborslsB                      z%RadiusNeighborsMixin.radius_neighborsrc Cs|dk rt|dddgd}|jjd}|dkr4|j}|dkrR|j||dd }d}n8|d kr~|j||d d \}}tt|}n td ||jd}t d d|D} tt|}|dkrt t |}ttj dt dt| f} t||| f||fdS)aComputes the (weighted) graph of Neighbors for points in X Neighborhoods are restricted the points at a distance lower than radius. Parameters ---------- X : array-like, shape = [n_samples, n_features], optional The query point or points. If not provided, neighbors of each indexed point are returned. In this case, the query point is not considered its own neighbor. radius : float Radius of neighborhoods. (default is the value passed to the constructor). mode : {'connectivity', 'distance'}, optional Type of returned matrix: 'connectivity' will return the connectivity matrix with ones and zeros, in 'distance' the edges are Euclidean distance between points. Returns ------- A : sparse matrix in CSR format, shape = [n_samples, n_samples] A[i, j] is assigned the weight of edge that connects i to j. Examples -------- >>> X = [[0], [3], [1]] >>> from sklearn.neighbors import NearestNeighbors >>> neigh = NearestNeighbors(radius=1.5) >>> neigh.fit(X) # doctest: +ELLIPSIS NearestNeighbors(algorithm='auto', leaf_size=30, ...) >>> A = neigh.radius_neighbors_graph(X) >>> A.toarray() array([[1., 0., 1.], [0., 1., 0.], [1., 0., 1.]]) See also -------- kneighbors_graph NrJZcscZcoo)rKrrF)rgr!TzQUnsupported mode, must be one of "connectivity", or "distance" but got %s insteadcSsg|] }t|qSr%)r)rxar%r%r&rMsz?RadiusNeighborsMixin.radius_neighbors_graph..r)r+)rW)rrSrWr7rr,Z concatenater|r#ZarrayrrZzerosintZcumsumr) r?r]r7rrrrr1rr6rr%r%r&radius_neighbors_graph s2,    z+RadiusNeighborsMixin.radius_neighbors_graph)NNT)NNr)rarbrcrdrrrr%r%r%r&rGs "  rc@seZdZddZdS)SupervisedFloatMixincCs2t|ttfs"t||ddd\}}||_||S)aFit the model using X as training data and y as target values Parameters ---------- X : {array-like, sparse matrix, BallTree, KDTree} Training data. If array or matrix, shape [n_samples, n_features], or [n_samples, n_samples] if metric='precomputed'. y : {array-like, sparse matrix} Target values, array of float values, shape = [n_samples] or [n_samples, n_outputs] rJT) multi_output)rRr r r_yr_)r?r]yr%r%r&fitYs zSupervisedFloatMixin.fitN)rarbrcrr%r%r%r&rXsrc@seZdZddZdS)SupervisedIntegerMixincCst|ttfs"t||ddd\}}|jdksD|jdkrp|jddkrp|jdkr^tjdtddd|_ | d }nd|_ t |g|_ t j|jt jd |_xPt|jjdD]<}t j|d d |fdd \}|jd d |f<|j |qW|j s |j d |_ |j|_||S)aFit the model using X as training data and y as target values Parameters ---------- X : {array-like, sparse matrix, BallTree, KDTree} Training data. If array or matrix, shape [n_samples, n_features], or [n_samples, n_samples] if metric='precomputed'. y : {array-like, sparse matrix} Target values of shape = [n_samples] or [n_samples, n_outputs] rJT)rrr zA column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().)rCF)r)r+N)Zreturn_inverser)rRr r rndimrWrErFrZ outputs_2d_rrZclasses_r,rrrrangeuniqueappendrr_)r?r]rkclassesr%r%r&rms& "   ,  zSupervisedIntegerMixin.fitN)rarbrcrr%r%r%r&rlsrc@seZdZdddZdS)UnsupervisedMixinNcCs ||S)aFit the model using X as training data Parameters ---------- X : {array-like, sparse matrix, BallTree, KDTree} Training data. If array or matrix, shape [n_samples, n_features], or [n_samples, n_samples] if metric='precomputed'. )r_)r?r]rr%r%r&rs zUnsupervisedMixin.fit)N)rarbrcrr%r%r%r&rsr)=rd functoolsrZdistutils.versionrr~rEabcrrZnumpyr,Z scipy.sparserrrr rr baser Zmetricsr Zmetrics.pairwiserZutilsrrrZutils.multiclassrZutils.validationrZ externalsr exceptionsrZ utils._joblibrrrrr}dictZ valid_metricsr|keysrDrXr'r2Zwith_metaclassr3rhr-rirrrrrr%r%r%r&s^               ,5  +