ek&ddlZddlZddlmZmZmZddlmZddl m Z m Z ddl m Z mZmZmZmZmZmZddl mZd ZGd d eZGd d eZdS)N)Dataset Dimension NdOverlay) Operation)cartesian_productisfinite)Area BivariateContoursCurve DistributionImagePolygons)contoursc|d||zz |d||zz}}t|drt||d}t|drt||d}tj|||S)z0Establish support for a kernel density estimate.rr)r maxminnplinspace) bin_rangebwgridsizecutclipkminkmaxs 9lib/python3.11/site-packages/holoviews/operation/stats.py _kde_supportr s1S()A,c*A$DQ"4a!!Q"4a!! ;tT8 , ,,cLeZdZdZejdddgdZejddZejd d Z ej dd d Z ej ddZ ejddZejddZejdeefdZdZddZdS)univariate_kdea Computes a 1D kernel density estimate (KDE) along the supplied dimension. Kernel density estimation is a non-parametric way to estimate the probability density function of a random variable. The KDE works by placing a Gaussian kernel at each sample with the supplied bandwidth. These kernels are then summed to produce the density estimate. By default a good bandwidth is determined using the bw_method but it may be overridden by an explicit value. scott silverman: Method of automatically determining KDE bandwidthdefaultobjectsdocNzd Allows supplying explicit bandwidth value rather than relying on scott or silverman method.r(r*D Draw the estimate to cut * bw from the extreme data points.rz= Specifies the range within which to compute the KDE.r(lengthr*zA Along which dimension of the Element to compute the KDE.Tz; Controls whether to return filled or unfilled KDE.d3 Number of samples to compute the KDE over.zW Defines a dimension to group the Histogram returning an NdOverlay of Histograms.r(class_r*c|jjr|t|tst d||jjtt }d|j_||jtS ddlm }ddl m }n3#t$r&tt|jddwxYwi}t|tr{|jd}|jt|jkr |j|d<|j|d<|jd}|jd } |jd kr|| d n|g} n|jjr ||jj}n=|j|jz} | s$t d t|jz| d}|jd } t1| d g} ||} |jjp||} | d kst9d| Drd} n(| d| dkr| ddz | ddzf} |jjrt<nt>}tA| r| tC| ng} tA| dkr |"| }n#|$r|g|| fi|cYSwxYw|jj#r|$|jj#|%| &dz}|jjr-tOj(| d| d|jj)}n,tU| ||jj)|jj+|j}|,|}nAtOj(| d| d|jj)}tOj-|d}|||ff|g| d|S)Nz3Cannot use histogram groupby on non-Dataset Element) group_typecontainer_typerstats) LinAlgError* operation requires SciPy to be installed.grouplabel_densityDensity)r<zU%s element does not declare any dimensions to compute the kernel density estimate on.)rrc36K|]}t| VdSNr ).0rs r z*univariate_kde._process..as*%I%I!(1++o%I%I%I%I%I%Ir!)rrr?ddofkdimsvdims).pgroupby isinstancer ValueErrorrmap_processscipyr8 scipy.linalgr9 ImportErrortype__name__rrIr;r<rJnameclone dimension get_dimensionrdimension_valuesrrangeanyfilledr r lenr gaussian_kde bandwidth set_bandwidth scotts_factorstdrr n_samplesr revaluate full_like)selfelementkeygroupedr8r9params selected_dimvdim vdim_namerJ dimensionsdatar element_typekderxsyss rrPzunivariate_kde._process;s7 6> 7gw// X !VWWWoodfnYboccG!DFN;;t}g66 6 l # # # # # # 0 0 0 0 0 0 0 l l ld!4```aagk k l g| , , <"=+L}W 666")-w%mF7O=#D',666I?CyI?U?UTZZ Z;;;[_`EEv -&44TV5EFF $]7=8 !=$&R%)']]%;&<=== *!} ',666Iy :::;E'' 55F$C l(C(C   #%I%Iy%I%I%I"I"I II q\Yq\ ) )"1c)9Q<+;T[=>>>I T0ghhhI %,q'G H H HC#"4A@AAAI T0DEEEIU]4.>???F c06777I"e!$Y7GNZ[[[GL>S>S>S>S>S>Sr!r#creZdZdZejddZejdddgdZej d d Z ej d d Z ejd dZ ej deefdZejddZejd ddZejd ddZdZddZd S) bivariate_kdea Computes a 2D kernel density estimate (KDE) of the first two dimensions in the input data. Kernel density estimation is a non-parametric way to estimate the probability density function of a random variable. The KDE works by placing 2D Gaussian kernel at each sample with the supplied bandwidth. These kernels are then summed to produce the density estimate. By default a good bandwidth is determined using the bw_method but it may be overridden by an explicit value. Tzv Whether to compute contours from the KDE, determines whether to return an Image or Contours/Polygons.r+r$r%r&r'Nzl Allows supplying explicit bandwidth value rather than relying on scott or silverman method.r,r-Fz@ Controls whether to return filled or unfilled contours. zD A list of scalar values used to specify the contour levels.r2r0r1rzY The x_range as a tuple of min and max x-value. Auto-ranges if set to None.r.zY The x_range as a tuple of min and max y-value. Auto-ranges if set to None.c ddlm}n3#t$r&tt|jddwxYwt |dkrtd|dd\}}i}t|tr?|j t|jkr |j |d<|j |d<|j d}nd}| dd gj}|jjp|d\} } |jjp|d \} } t'd | | fDrd \} } n| | kr | d z | d z} } t'd | | fDrd \} } n| | kr | d z | d z} } |jd d kr-|ddt+|dfnt/jd}|jd d kr||} |jjr| |jj| |d z}|jjr!t/j| | |jj}n.tA| | f||jj|jj!|j}|jjr!t/j| | |jj}n.tA| | f||jj|jj!|j}tE||gd\}}t/j#|$|$g}t/j%| |j|j}n|jj&r+|jj'rtPntR}|g||g|gSt/j| | |jj}t/j| | |jj}t/j*|jj|jjf}tW|||jff|dd|gd|}|jj&rAtM||jj'|jj,}|j-|j.d dfi|S|S)Nrr7r:rzQbivariate_kde can only be computed on elements declaring at least two dimensions.r;r<r>rc36K|]}t| VdSr@rArBvs rrDz)bivariate_kde._process..*5518A;;555555r!)grErEc36K|]}t| VdSr@rArs rrDz)bivariate_kde._process..rr!)axis)rrrFFrH)r]levels)/rQr8rSrTrUr^rorNrMr r;r<rJarrayTrKx_ranger[y_ranger\shaper rremptyr_r`rarbrcrrdr rrvstackravelreshaperr]rr zerosrrrWrp)rgrhrir8xdimydimrkrmrpxminxmaxyminymaxrrrrsrtxxyy positionsfeltypeimgcntrs rrPzbivariate_kde._processs l # # # # # # # l l ld!4```aagk k l w!!## $ $q ( (BCC C''))"1"- d gy ) ) }W 666")-w%mF7O=#DDD}}aV$$&V^7w}}Q'7'7 dV^7w}}Q'7'7 d 55t 555 5 5 ,"JD$$ T\\c48$D 55t 555 5 5 ,"JD$$ T\\c48$D6:jma6G6GtAAAx~~))q)11122RXV\M]M] :a=1  $$T**Cv 4!!$&"2333""$$txxQx'7'77Bv~ ^[tTV-=>>!4,DF4DdfjRVR\]]v~ ^[tTV-=>>!4,DF4DdfjRVR\]]&Bx77FB 288::rxxzz":;;I 33y>>+RX66AA V_ ?!%>ARQSMZ););)=)=bqb)A$ZZSYZZ 6? 7C dfmLLLD4:dim66v66 6 s 09r@)rUrurvrwrxr~rryrzr{r`rr]rlistintrrrdr|rrrrPrr!rrr}s  u}T01222H%$Ww >T[=>>>I T0)***I %,q'G H H HCU]5/CDDDF!U T3KFGHHHF c06777I"u!$q?G"u!$q?GL<<<<<<r!r)numpyrrxcorerrrcore.operationr core.utilrr rhr r r r rrrrr r#rrr!rrs2 0000000000&&&&&&33333333UUUUUUUUUUUUUUUUUU---dSdSdSdSdSYdSdSdSPjjjjjIjjjjjr!