³ò ŒB_Kc@skddkZddklZddkTddkTdefd„ƒYZd„Zedjo eƒndS(iÿÿÿÿN(t deprecated(t*t GAAClusterercBs†eZdZdeed„Zeed„Zed„Zd„Z d„Z e dƒd„ƒZ d „Z d „Zd „Zd „ZRS( sN The Group Average Agglomerative starts with each of the N vectors as singleton clusters. It then iteratively merges pairs of clusters which have the closest centroids. This continues until there is only one cluster. The order of merges gives rise to a dendrogram: a tree with the earlier merges lower than later merges. The membership of a given number of clusters c, 1 <= c <= N, can be found by cutting the dendrogram at depth c. This clusterer uses the cosine similarity metric only, which allows for efficient speed-up in the clustering process. icCs2ti|||ƒ||_d|_d|_dS(N(tVectorSpaceClusterert__init__t _num_clusterstNonet _dendrogramt_groups_values(tselft num_clusterst normalisetsvd_dimensions((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyRs  cCsOtg}|D]}|ti|tiƒq~ƒ|_ti||||ƒS(N(t Dendrogramtnumpytarraytfloat64RRtcluster(R tvectorstassign_clustersttracet_[1]tvector((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyR"s6c Cs’g}|D]}||gq ~}ti|ƒ}xHt|ƒt|idƒjo(d}xžtt|ƒƒD]Š}xt|dt|ƒƒD]f} |i||t||ƒ|| t|| ƒƒ} | p| |djo| || f}qŽqŽWqnW|d\}} |||| } |od|| fGHn| ||<|| =|||| ||<|| =|ii|| ƒq6W|i |iƒdS(Niismerging %d and %d( tcopytlentmaxRRtranget_average_similarityRtmergetupdate_clusters( R RRRRtclusterst vector_sumtbesttitjtsimtsum((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pytcluster_vectorspace(s,$" cCsö|ii|ƒ}g|_xÂ|D]º}t|ƒdjpt‚|io|i|dƒ}nti|dƒ}x=|dD]1}|io||i|ƒ7}q||7}qW|t t|ƒƒ:}|ii |ƒq"Wt|iƒ|_ dS(Nii( Rtgroupst _centroidsRtAssertionErrort_should_normaliset _normaliseRRtfloattappendR(R R RRtcentroidR((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyRIs     cCsvd}xet|iƒD]T}|i|}|i|d|dƒ}| p||djo||f}qqW|dS(Nii(RRRR'R(R RR R!R-R#((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pytclassify_vectorspace[s s$Use GAAClusterer.dendrogram instead.cCs t|ƒS(N(t dendrogram(R ((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyt dendogramdscCs|iS(si @return: The dendrogram representing the current clustering @rtype: Dendrogram (R(R ((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyR/hscCs|iS(N(R(R ((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyR oscCs4||}||}ti||ƒ|||dS(Ni(Rtdot(R tv1tl1tv2tl2R$tlength((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyRrs  cCs d|iS(Ns*(R(R ((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyt__repr__ws(t__name__t __module__t__doc__tTrueRRtFalseRR%RR.RR0R/R RR7(((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyRs  !     c Csíddkl}g}ddgddgddgddgddgddggD]}|ti|ƒqQ~}|idƒ}|i|tƒ}dG|GHd G|GHd G|GHH|iƒiƒtiddgƒ}d |G|i|ƒGHHd S( sO Non-interactive demonstration of the clusterers with simple 2-D data. iÿÿÿÿ(Riiiiis Clusterer:s Clustered:sAs:s classify(%s):N( tnltkRRRRR;R/tshowtclassify(RRtfRt clustererRR((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pytdemozs`   t__main__( Rtnltk.internalsRtapitutilRRRBR8(((s'/p/zhu/06/nlp/nltk/nltk/cluster/gaac.pyss   k