ó ¦–Õ\c@`síddlmZmZmZddlmZddlmZddlZ ddl m Z m Z ddl mZdd lmZdd lmZdd lmZee jƒd d „ƒZd„Zd„Zd dd„Zddd„ZdS(i(tabsolute_importtdivisiontprint_function(twraps(tNumberN(tmerget merge_sortedi(tArrayi(ttokenize(tHighLevelGraph(tIteratortlinearcC`s-t|ƒ}t|ƒs"d|fSt|tƒr@t|ƒ}n|jjdkr›tj|j |d|ƒ}ddl }|j j ||j |jƒ|fStj|jtjƒrð|jdƒ}tj||d|ƒ}|j|jƒ|fStj|jtjƒsd}ntj||d|ƒ|fS(Ntcategoryt interpolationiti8tnearest(tlentNonet isinstanceR tlisttdtypetnametnpt percentiletcodestpandast Categoricalt from_codest categoriestorderedt issubdtypet datetime64tastypetnumber(tatqR tntresulttpdta2((s4lib/python2.7/site-packages/dask/array/percentile.pyt _percentiles     " cC`s*ddlm}|ƒ}|j|ƒ|S(Ni(tTDigest(tcrickR)tupdate(R"R)tt((s4lib/python2.7/site-packages/dask/array/percentile.pyt_tdigest_chunk#s  cC`s@ddlm}|ƒ}|j|Œtj|j|dƒƒS(Ni(R)gY@(R*R)RRtarraytquantile(tqstdigestsR)R,((s4lib/python2.7/site-packages/dask/array/percentile.pyt_percentiles_from_tdigest-s  tdefaultc `si|jdkstdƒ‚ntˆtƒr9ˆg‰ntjˆƒ‰t|ˆˆƒ}|j}tj|tj ƒr—tjgd|ƒdj}ndddg}||krÁt dƒ‚n|dkrÖd}n|}|dkrœˆd krœtj|tj ƒstj|tj ƒrœd d l m }|d d ƒd|‰t‡fd†t|jƒƒDƒƒ} d|} itˆt| ƒf| d f6} n~d|‰t‡‡‡fd†t|jƒƒDƒƒ} d|} itˆˆgt|jd ƒt| ƒˆf| d f6} t| | ƒ} tj| | d|gƒ} t| | dtˆƒffd|ƒS(s] Approximate percentile of 1-D array Parameters ---------- a : Array q : array_like of float Percentile or sequence of percentiles to compute, which must be between 0 and 100 inclusive. interpolation : {'linear', 'lower', 'higher', 'midpoint', 'nearest'}, optional The interpolation method to use when the desired percentile lies between two data points ``i < j``. Only valid for ``method='dask'``. * 'linear': ``i + (j - i) * fraction``, where ``fraction`` is the fractional part of the index surrounded by ``i`` and ``j``. * 'lower': ``i``. * 'higher': ``j``. * 'nearest': ``i`` or ``j``, whichever is nearest. * 'midpoint': ``(i + j) / 2``. method : {'default', 'dask', 'tdigest'}, optional What method to use. By default will use dask's internal custom algorithm (``'dask'``). If set to ``'tdigest'`` will use tdigest for floats and ints and fallback to the ``'dask'`` otherwise. See Also -------- numpy.percentile : Numpy's equivalent Percentile function is+Percentiles only implemented for 1-d arraysRgà?R3tdaskttdigests1method can only be 'default', 'dask' or 'tdigest'R i(timport_requiredR*s=crick is a required dependency for using the t-digest method.spercentile_tdigest_chunk-c3`s-|]#\}}ˆ|ft|ffVqdS(N(R-(t.0titkey(R(s4lib/python2.7/site-packages/dask/array/percentile.pys tsspercentile_tdigest-spercentile_chunk-c3`s3|])\}}ˆ|ft|ˆˆffVqdS(N(R((R7R8R9(R RR#(s4lib/python2.7/site-packages/dask/array/percentile.pys ss percentile-t dependenciestchunks(tndimtNotImplementedErrorRRRR.RRRtintegert ValueErrortfloatingt dask.utilsR6tdictt enumeratet __dask_keys__R2tsortedtmerge_percentilesRR;RR tfrom_collectionsR( R"R#R tmethodttokenRtallowed_methodstinternal_methodR6tdsktname2tdsk2tgraph((R RR#s4lib/python2.7/site-packages/dask/array/percentile.pyR7sF      *  %   tlowercC`st|tƒrt|ƒ}ntj|ƒ}ttt|ƒƒ}t|ƒ}|dkrot|Œ\}}nt|ƒ}ttgt|||ƒD]$\}}}|r”|||f^q”Œƒ}|sÙtdƒ‚n|\}}}|dj j dkr\t ||g|D]} | j ^q||ƒ} ddl } | jj| |dj|djƒStj|dj tjƒsd}nt|ƒt|ƒks±t|ƒt|ƒkrÀtdƒ‚ng} xjt||ƒD]Y\}}tjt|ƒƒ} tj|ƒ| d)|d| d<| |9} | j| ƒqÖWttt|| ƒŒ}t|Œ\}}tj|ƒ}tj|ƒ}tj|ƒ}|t|ƒ}|dkr¾tj|||ƒ}n?tj||d d ƒ}tj||d d ƒd}tj|t|ƒd|ƒtj||ƒ}tj||ƒ}|d krL||}n±|d kre||}n˜|dkrŠd||||}ns|dkrñtj|||ƒ}tj|||ƒ}||k}|}||||<||}n tdƒ‚|S(s% Combine several percentile calculations of different data. Parameters ---------- finalq : numpy.array Percentiles to compute (must use same scale as ``qs``). qs : sequence of :class:`numpy.array`s Percentiles calculated on different sets of data. vals : sequence of :class:`numpy.array`s Resulting values associated with percentiles ``qs``. Ns : sequence of integers The number of data elements associated with each data set. interpolation : {'linear', 'lower', 'higher', 'midpoint', 'nearest'} Specify the type of interpolation to use to calculate final percentiles. For more information, see :func:`numpy.percentile`. Examples -------- >>> finalq = [10, 20, 30, 40, 50, 60, 70, 80] >>> qs = [[20, 40, 60, 80], [20, 40, 60, 80]] >>> vals = [np.array([1, 2, 3, 4]), np.array([10, 11, 12, 13])] >>> Ns = [100, 100] # Both original arrays had 100 elements >>> merge_percentiles(finalq, qs, vals, Ns=Ns) array([ 1, 2, 3, 4, 10, 11, 12, 13]) sNo non-trivial arrays foundiR NRs3qs, vals, and Ns parameters must be the same lengthiR tsidetlefttrightRPthighertmidpointgà?sOinterpolation can only be 'linear', 'lower', 'higher', 'midpoint', or 'nearest'( RR RRR.tmapRtzipR?RRRFRRRRRRRR!RtemptytdifftappendRtcumsumtsumtinterpt searchsortedtminimumtmaximumtabs(tfinalqR0tvalsR tNsR#tvaltNtLtvR%R&tcountstcounttcombined_vals_countst combined_valstcombined_countst combined_qt desired_qtrvRRRSRPtuppertlower_residualtupper_residualtmasktindex((s4lib/python2.7/site-packages/dask/array/percentile.pyRF‹sn   I. $ 0            (t __future__RRRt functoolsRtnumbersRtnumpyRttoolzRRtcoreRtbaseRthighlevelgraphR t compatibilityR RR(R-R2RRF(((s4lib/python2.7/site-packages/dask/array/percentile.pyts   T