U ÙfMã@sŠdZddlmZddlZddlZddlmZzddlm Z Wne k rZe  d¡YnXddl mZddl mZdd d „Zdd„ZdS)z6 Summary plots of SHAP values across a whole dataset. é)ÚdivisionN)Ú gaussian_kdezmatplotlib could not be loaded!é)Úlabels)Úcolorsú#333333TécVsÔd}tˆtƒr0d}|dkrd}|dksRtdƒ‚n"|dkrt  ˆdd…dtˆƒ…f 5¡d¡},t  ˆdd…dtˆƒ…f 5¡d4¡}-t&|ƒD]\}}ˆdd…|f}t  |¡t  |¡}.}/|/|.}0t  6t  |¡|0d>t  |¡|0d>d¡}1t  7|¡|-|,dkrÂt8|t j( ,t|ƒ¡|-|,dƒ|1ƒ}2n t8|ƒ|1ƒ}2|2t  |2¡d}2|dd…|f}td?t|ƒd#ƒ}3t  -t|1ƒd¡}4t  |¡}d}5d}6d}7d}8t t|1ƒdƒD]¸}9|6t|ƒk r¨|1|9|||6k r¨|7|||67}7|6d7}6|6|5d#k rB|7|||58}7|5d7}5 qB|6|5dk rì|7|6|5|4|9<t |8ƒD]}:|4|9|4|9|:d< qÎn|8d7}8 q>t  |d2¡}'t  |d3¡}(|'|(k rPt  |d¡}'t  |d4¡}(|'|(k rPt  |¡}'t  |¡}(tj0|t  9ˆjd¡|d@tj4|'|(||dddA |4|'8}4|(|'dk r¢|4|(|'}4t t|1ƒdƒD]„}|2|dBk sÖ|2|ddBk r²tj:|1||1|dg||2|||2|dg||2|||2|dgt 4|4|¡ddC q²qn\tj;ˆdd…|ft t|ƒƒdDddEddddF};|;dGD]$}<|< <|¡|< =dH¡|< >|¡ qtn|d kr2dD}=t  6d|jd| d¡ +d¡ 2dI¡}>t  ˆ¡t  ˆ¡}.}/t  6|.|/|=¡}?t&|ƒD]"\}}|dd…|f}@t j?|@ddJ\}A}B|Ajd| k r`t  |A¡}Ct  @|B|C¡}Dt  A|Ddd¡}Dn|>}D|Djdd}"t  |@¡}Ct  9|=¡|}Et  -|"|=f¡}&t |"ƒD]ì}ˆ|C|D||D|d…|f}|jddk rtB CdK|ˆ|f¡|dk r |&|ddd…f|&|dd…f< q t8|t j(jDddL|jddMƒ|?ƒ|&|dd…f<|D|d|D|}F|jd|"}G|F|G}H|&|dd…f|H9< q t j@|&dd$}&d}I|& ¡d|I}Jt |"dd(d(ƒD]\}|&|dd…f|J}K|tjEjFk rüt G|¡||"dƒn|}Ltj:|?||K||K|LdN qÀ qüt |.|/¡nˆ|s²|dkr²|d|…}Mt  "t|Mƒ¡}Nt  ˆ¡  d¡}OtjH|N|O|MdEdO|dPtjI|NdQdRt J¡ K‡fdSd„|MDƒ¡n|rº|dkrº|dkrâdTd„t tˆƒƒDƒ}|d|…}Mt  "t|Mƒ¡}Nt  -t|Mƒ¡}Pt  ‡fdUd„t tˆƒƒDƒ¡}Qt&|QƒD]N\}}t  ˆ|¡  d¡}OtjH|N|O|MdE|PdO||ƒ||dV|P|O|M7}Pq2tjI|NdQdRt J¡ K‡fdWd„|MDƒ¡tjLddXdY| r¼|dk r¼|dkr¼|d ksì|tjEjFkr¼ddlMmE}R|RjN|d kr tj4nt G|¡dZ}S|S Oddg¡tjP|Sddgd[d\}T|T Qtd]td^g¡|TjRtd_dXdd`|TjSjTddda|T >d¡|TjU Vd¡|TjS W¡ Xt #¡jY Z¡¡}U|TjS [|Uj\d1d#¡t J¡j] ^db¡t J¡j_ ^dH¡t J¡j`dc Vd¡t J¡j`dd Vd¡t J¡j`de Vd¡t J¡jT||dftjIt t|ƒƒ‡fdgd„|DƒdQdR|dkrpt J¡jTdhd#d+didjt J¡jTdkddlt ad(t|ƒ¡|dkr°tjtdmdQdRntjtdndQdR| rÐt ¡dS)oa2Create a SHAP summary plot, colored by feature values when they are provided. Parameters ---------- shap_values : numpy.array For single output explanations this is a matrix of SHAP values (# samples x # features). For multi-output explanations this is a list of such matrices of SHAP values. features : numpy.array or pandas.DataFrame or list Matrix of feature values (# samples x # features) or a feature_names list as shorthand feature_names : list Names of the features (length # features) max_display : int How many top features to include in the plot (default is 20, or 7 for interaction plots) plot_type : "dot" (default for single output), "bar" (default for multi-output), or "violin" What type of summary plot to produce. FTNÚbarzBOnly plot_type = 'bar' is supported for multi-output explanations!Údotrz9Summary plots need a matrix of shap_values, not a vector.Zlayered_violinZcoolwarmcst |tˆƒ¡S)N)rZred_blue_circleÚlen)Úi©Ú shap_values©ú9/tmp/pip-target-lpfmz8o1/lib/python/shap/plots/summary.pyÚ:ózsummary_plot..z%rcSsg|]}tdt|ƒ‘qS)ZFEATURE)rÚstr©Ú.0r rrrÚ Nsz summary_plot..éégð?éédgø?gš™™™™™é?)Zfigsize)Ú feature_namesÚsortÚshowÚ color_barÚauto_size_plotÚ max_displayÚé cSsg|]}d‘qS)r!rrrrrrys)rrrrrr ZINTERACTION_VALUEg)ÚpadZw_padZh_padgš™™™™™¹?)ZhspaceZwspacer)Zaxisgš™™™™™Ù?éz#999999éÿÿÿÿ)ÚxÚcolorÚzorderz#ccccccgà?)ré)Úyr'ZlwÚdashesr()Zdtypeg:Œ0âŽyE>gíµ ÷Æ°>gÍÌÌÌÌÌì?r)é_éczs6     þ