B T\m@sddlmZmZmZddlmZddlmZddlZ ddl m Z m Z ddl mZdd lmZdd lmZdd lmZee jdd dZdddZdddZdS))absolute_importdivisionprint_function)wraps)NumberN)merge merge_sorted)Array)tokenize)HighLevelGraph)IteratorlinearcCst|}t|sd|fSt|tr*t|}|jjdkrhtj|j||d}ddl }|j ||j |j |fSt|jtjr|d}tj|||d}||j|fSt|jtjsd}tj|||d|fS)Ncategory) interpolationrZi8nearest)len isinstancerlistdtypenamenp percentilecodespandas Categorical from_codes categoriesordered issubdtypeZ datetime64Zastypenumber)aqrnresultpdZa2r'4lib/python3.7/site-packages/dask/array/percentile.py _percentiles    r)c s|jdkstdttr"gtt|t}d|tfddt | D}d|}|dft gt |j dt|fi}|j}t|tjrtjg|dd j}t||}tj|||gd }t||t ff|d S) z` Approximate percentile of 1-D array See :func:`numpy.percentile` for more information r z+Percentiles only implemented for 1-d arrayszpercentile_chunk-c3s&|]\}}|ft|ffVqdS)N)r)).0ikey)rrr#r'r( 0szpercentile..z percentile-r)rg?)Z dependencies)chunksr)ndimNotImplementedErrorrrrarrayr rdict enumerateZ __dask_keys__merge_percentilesrr.sortedrr Zintegerrr Zfrom_collectionsr ) r"r#rtokenZdskZname2Zdsk2rZgraphr')rrr#r(r#s&    rlowercCst|trt|}t|}ttt|}t|}|dkrFt|\}}t|}ttddt|||D}|sxtd|\}}}|djj dkrt ||dd|D||}ddl }|j ||dj|djSt|djtjsd}t|t|ks t|t|krtd g}xVt||D]H\} } tt| } t| | d d<| d| d<| | 9} || q$Wttt||} t| \} }t| } t|}t|}|t|}|d krt||| }ntj||d d }tj||dd d }t|t| d |t||}t||}|dkr6| |}n|dkrJ| |}nx|dkrjd| || |}nX|dkrt|||}t|||}||k}|}||||<| |}ntd|S)a% 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]) NcSs g|]\}}}|r|||fqSr'r')r*r#valNr'r'r( fsz%merge_percentiles..zNo non-trivial arrays foundrrcSsg|] }|jqSr')r)r*vr'r'r(r:nsrz3qs, vals, and Ns parameters must be the same lengthr rleft)Zsiderightr7ZhigherZmidpointg?zOinterpolation can only be 'linear', 'lower', 'higher', 'midpoint', or 'nearest')rrrrr1mapzip ValueErrorrrr4rrrrrr r!remptyZdiffappendrZcumsumsumZinterpZ searchsortedZminimumZmaximumabs)ZfinalqZqsZvalsrZNsLr%r&Zcountsr#r9countZcombined_vals_countsZ combined_valsZcombined_countsZ combined_qZ desired_qrvr<r=r7upperZlower_residualZupper_residualmaskindexr'r'r(r4@sn    $                  r4)r)r)r7N)Z __future__rrr functoolsrZnumbersrZnumpyrZtoolzrrZcorer baser Zhighlevelgraphr Z compatibilityrrr)r4r'r'r'r(s