From 33ae4ea75ae7abe780b9663a1e0d1705f266f7f3 Mon Sep 17 00:00:00 2001 From: Ian Dong Date: Wed, 10 Apr 2024 18:27:27 -0700 Subject: [PATCH] Added image for extra slides on worst predictor --- .../roc_curve_worse_predictor_differing_T.png | Bin 0 -> 161601 bytes logistic_regression_2/logistic_reg_2.ipynb | 428 ++++++++++++++++++ logistic_regression_2/logistic_reg_2.qmd | 9 +- 3 files changed, 435 insertions(+), 2 deletions(-) create mode 100644 logistic_regression_2/images/roc_curve_worse_predictor_differing_T.png create mode 100644 logistic_regression_2/logistic_reg_2.ipynb diff --git a/logistic_regression_2/images/roc_curve_worse_predictor_differing_T.png b/logistic_regression_2/images/roc_curve_worse_predictor_differing_T.png new file mode 100644 index 0000000000000000000000000000000000000000..1180046cb6372f50b252a662bad60c50047f3000 GIT binary patch literal 161601 zcmeFYbzdA!*ENc}YjAg$Kp?oMzI1b5fqZoxHpa3}a_uKSVqJ?Ee6 z1DyHw&^_HjPxs{C>1cX$0vIe}S>M&lmPC}fBWu7cz2O^~kloVoK@H*ya8fr+< za1IQ$V>9KRhFnT09%uE>HM7TIS>iQ!{Q&{Bta!&G88}le2KVdU8>{zc!WJ)1d#h<) zXArrr0aRGK`G*iYh{%04JulY|!wk+U=pvAa2T;UOTlQDtV+RNJ5Ti}U_gBFnv?!~l#%Jn3;ycV^i{iSpgx+pz_jR2`sAa_{TmJ|OHsIa%eJsbL-FLTU@IVZB59v0zJye0F`s|Ejy+wf>I%6)Z!I*lwb7E>12 z*h)p_u#rkRltT`?ASaTse3&qTJEvEevP=p!^(VVnLdPpDb_)@BV>(hgauj2+u)v^} z_}@rEf?^nbA@3?NoYQmtNYHCY`y%h}{T+}p0hqU^;f-^0BK<;tgD+>0(#lz*A)QLsqGbwFpDX z-`;iD5?{)kvh)!6i-duuNe(gP4>Rw~PGAcCNz%ZCtUd7#1`-V;P_7#u4awZ{b3(Z! zLX|@#Y4P6EZ zng^!@I#;l>vG@^V5f%{jf>sh?uKI<=2_Rc@ec~wb{|N6vH3g(sHVdDk&gF4d1`*y%s(<5j>r9tifQ4%1iv3)TAIyc50eMmSx-Z_Ty zyMCnf1$TtUgLyHsZqHT&gdxs_gg$(5|N9Z&lh_Y@YX8XVyKnREd^QiIyT!cnvo8I< z{hwfn4XiKr#ebkp4AE>hy)*rZ&Y^tBF@gfFyh|)=)$}GKKuj6CL-((FGPhPDq4boIxQ)OA=^jlAn`5? z9V;w|sz!=4A;gqY^mD-2XFT#EDFT&n8;LLRU@0y~EWTK0%BJ{GDYi?(`yk^Iq|#u* zJkSAsl0WA(J2z>TpLxFO6s0poD@uik^^~3^v?o?an7(A&6#fAO9J`quJXb$4JKVs} z{AwI#Xl#?Z8mw~SvYx{lyuX=M#(J7$x0rNW+?g>SyK#vF=r(n|k%Wj0A~d!>p2M}H zc0~yHJzji;VeE-BR6M7yM%+e73;GdcEXrp@^c&+h_p0n#klSZUv~Tj6 zpLAt(6<5UsJ{?mF(W1pkQG*i%hcj@*x8jhb`U_nO^$NEOnF{S@;AzZguxTuc$cu1j z@@Bk?9E-AK29ql31yvv9Ky*4lA!Soi4}K3m52lu2@9j5ttGi7F+K^cTmg+1QID!izTMCfGIehCV>M02AB|tdTcwfTBze}lP6uxF$@PhOk$IMQeuQiG z&-xVAqZWt?SB&}~VHtf+*h8YsY5W@YF{C6f%c%S}N-s;uNeHyWfJBc3Wr2%CanwVU zOVo2zU9<)f8Q1CiytKdH5ad7^#JnB~3*Wi^v|J;35_N3+n&R$LGj3g~@4c=|YtRZ< z@~#s!zZ+-RL%Bk^n%k?`d&Z;1!^GoaIb=T7Okg2M(MqvqnbhuVs%@01)i#~8oHMDg z7;j{)t*q;`5No(-NHbP7aT-st9IYO&(fjJPgx28HFj`wy+h!WO+_r>#ys+XHLmhSf zj)7Q<>+*YG1oUj)RNlnzyn+~$=56mmiNVwYW{ILGZN9_flCKJt3M~pcS*J1c#3cNl zF9#{pRaqyQEfeY!3j<1nFwtpT->t@&@V)RmN8wU+D!CUB7Q`znDh<|F*ZS;k>^e^I zPOVPoPI(Emqbvth2KdkeLc0cuW2s2V1lI(R1@63Fc6T?Ew;N{5EXr>*UluD&$hnyHtGbl>NM$~ zs{JMN4nuZnSV zi)l3$+m;TATdo5gk$15Mq`tg8w4V~D(o0glrE6yV6=ER&!TiH|wZx%rKu?ZG3Xht% zC{6`SwXp~#3Z1yi+}o|x`isM1G~zt=cli&((}u>UeFcfXXdP%tq^JsMicaJne}-@U zvWJ^l`ZM?ESr^ZZc$Xa4?QfkTq^Dmq|gcVvM#cZAyRZ})bJQw+k%vy7qt`BnUEM&tu zaT{v6u)aGhG&~6g)ab}yq%tI7Ir&Ua!}I~}-}SeDTxV0|d_g#B*DEOAaBdPuarHF`9~FRgc0`AaX=4gT1!)U9NrUD87F zkAAgA?cIH6J(%}4^yI5x_Xj24sY}BfisifnxI!aOF<%pqiH549s)duKlXaxVXkKqj zXcHNSpo)vye>AXm@Gb* z6p!1#d}mIQ6JFV0JG+~w+^0{B#oDYN^EUK+g@65?pSB-r|^uEmb~UPyW9H(}itb z7iV|pU9OF~dv7HjVo!zkgnTZjUx!mqgB5=$j^?BZl$|)7^34$a^18qOyT8WvnN8>p z_w~nx#f@x6D*Wr^V_AoK0$(9WqW?LTfTV|j{^vXt1VpG61k8WkqX6vRK5@YJt;~P+&)|8jyH~DJIYGnG=*o@WP*8c4|5CZP}z@e>~vk|$wt&N=%zq=si zKX>p0$8V?ED9Qi1#o1brQd3@uTm>{ zTG=~WeYGQhd#;i3R~KhNO3Js2{^#?b_cU|2`rn%Doc^m>Km*y{uCQ^iva|iqvw@-l zZ)f?HtlZ6Pv_4ze0(}P5A;j~6N8q3P|36p$x5odYq~`yYH)t$^7 z#lG4Cbvg_E?|A)Jz5nOM|0*cJ_SW?OqbvT?&;OhSI$8)xfbD+|QUmxy|DQc@42kj~kR!te0zw2r`tv6>cgW)`*mjzRo54U_HyvWU2w5Ws z3kSCm(fIy6sk|x5$TP}dO7RG06R{;xJ!zX)9+1Zoke zf3;1qL#gfX9EVhY_*b{TJyo1F=U=_Z+xuF+gJg6hauw4g{+Dl$l}njA34s{*?>6oz z2;3IbHVS66e_KK+e$?V$cD^4Bp@d(=YRDP}Ug*CYEi^f!VRXDl`IkK(C(?_c6u(kU^mFUqmH>vz z8}45t-w5bS+eK6(hksk*4ffFg7R3KQ?4@!GU?ig68G#JIE?j>K6J6_x+dur@RwS}P zVShEwn^BaaJtD_7&J8Y}DQ7^86N!DpZETwvGdm)8M)#!ftIYs`MW#7^LJ`$}I|fS_ ze^8lO%!e%mr|W6S3}m+UDVw&F#%OwcczM^|3$*NeV7+{U zwZs4flS%{jN9wVnX^CZt=1=DPG@a>r_+#(1?TiPgwcTSs%WdqJs-o?~-PzX(fh##K z!*0JloD`ZL!QFapYi5-#dxTyUz39TLa>CEo4Z5lL|FMJvEtuRs5^l@5mdg<)`7WR5 zF_xxPV>~_Q>1_3IESblf1%nLX*ZWDH;)VtNmGZ8a2b!G`W~p2Y?B&GhTe{2Z5gt-X!cm!>qE+Tk&WL6U_>#v#HR zX{cji&q>>^odnsjKy)F?0TMUetER5!irR?09R9c!6OBJz-F7&s%{eCaQ{OMwl*J+F+G|h2@;VLEP4`3iNl;>-18%uXsOw-=Ivx>sD zjT=5*S7Th0eY74x|BqwIGGlthz@KwTlet#YNk~;gJ_tK zUS?#G?Xfp=bt+9D+HFqJueY=DG1@9Cl-!JIy0iDWgUD@)TfX~r% zvOET3Az8eh~Km1{@GTlfsfU?*O?$R0Y0+0`8OlDM||ztEZ61X zOyE^Pz7SgNBM@oOlN0r`3f9@4hYVb^a$M>je0%b5sOo|FqP*ktc+xblzI^_>cU+dO z!}R5GZc{%k;XJw=jOH6Basy*1{+O}xG~oSmzi$q}omWH8 zMSV}a>*0`z>u%L416}xOYCzBXqTg1#ynf*03xK}#lg5<^x%YEwX&xM+XOw_7oV5Z& z4(DZR*Rq=e$GWV@w~JG^pzk&3aWA2V9BD}E`{cm2VpB)J9QMurCnAx$5MBsO1ePBi z`f+$9?k!MFtd{4WqDXTstp@=|6)B^VA=JG7W^gbB-ByNE@rKW%t<$tPej3oAUm~Dh zcQ)!0O1Cwah4;K`&Z<%>0~Ox@=ZXt$eoI2p{Zc@+c=C?)>ZLj#3=9Mc9&H<>{v*$d z4OiXk%cG9j4=>g~ZH$g|+3vK24BA$MMkSZAzE8c(O{ zJQ+4E%r@LQAcfv>%G2w4T!<>YiTSQ$sH@s9W%?j5 zXHS|o);}^1XvcNGVq^51$(H*bJF$9g^JL4C@-Y;u)Y7wh?)OvjQZsBPh*Z&#pAHn< zi88}_*$QT8b2!|uw{8X?Yj5e-zMt1L5x)~%J%_{hX#MlO?hl8#sU@C2D$io~I??qo z3iXr@LiKCOt3v0~YuQkj3ESt|E zG=l5DJ@(QDTVHo%8n}(YN4km3j<1)gU5n2d7v9EI*dfL+n`r1>##B!?RlcZ=Lqn|( zsJyF&ERMbpf9TraCjfY6Cl@Yc8@Mi+jf8bc|ZYuULO|$thn1Y|B>R;|ceX z@yVQq!Fx@!jU5a|y6023F0LG(haWc|xNR1z$4aLud9juA5Rzb#ejjNXYpc!0z5te4 zIjHg|)a#09vzYy6%2D1Y=_!V05-3v5~TyZ=8>(W7?KN|Z> z?J$eJJ2a_mBmxWF0l5127;tFBmL0aNbz_|7zk~64KEyh(>^=RaC0K9z++)2@@;jP| z#V~_a87reF5I*B-QdpSfi`D)npYij}LX&==zWZ<3t0RNg*827|%r}_gP%4#b|9$0p z+=^t=-XdkucB%V`#8jW{`*J@f8y?w^!abwhy76+`#6lov`rQNu==d!dpTECDQ;zj^ zlBeq?g@`(e^Xsof(};ENL@C$7Aj@Dq3Fs2L@Q-= zG9OKqVV~U!E%P)Q!tG&dRpjjSd7ZBv^WQV<2uaWzhB|KMbqeYgg`b_l$y>?o2vd~2 zB&sAPIoDX{Hl)D3aj7(#3YYysPj(fU+8TmQq(4P9=R0)&-HzSB+O8}tVR0y0Rf*JJ za%JGq$E8a#@bOr2`^47s4;huGfXKNi3e9p`=fj$ZMeaVcEZub*qoa=h9EHj1Nc~5n8qb@A||mg$yoh1 z7HHfSTi%Xd;U7aM=cq#-9twT?F!uctJ$ctDd zab4{Bo@JRIuC0mDJH-r8+V30aG2oed?i=RE*oV_7&0&dW=thGp#HKY@9c1K_=?!A% zsZZh3reAOAzt-%@M``pUUhJ7yMOVOPdtVMSvi1l_1B{;ho5-Jt)yLB|wq-H1qGS+~ z%GQ@ZU4V^SzCS2TR9MF1Me&Tt0UCfxzR7qf3V(%4FBGZ70eN?Sa825h+0Jrj?!;N3 zn>|)>hvk0Ku%L~v@1FPUb=Gm``Bfa+Iz;G`#=L>k_r56a3Od#rvc3y5)9&6i-kHcq1g04h263-S6jY@HlH~shakW0es_*K1YFjS?K&-D9i&Rq-j3W{Mhr8RIbIUxs2 zZ_)a$GzPd*`%%1vf|^MuN2)5XI@IHlA#3Q?(eFCc1D;aP+=gbF28;c|XAG85^N-&BxeQD<5~Z ze56)PiC?Rv7{{VKv1Fbx9)sJsXdLNk+HvTn9-(_J=m0{_3RviKV~Y{Tt%aym-PNS7 zB&8TEC2k8&DnKZxrQM25qA06Wzr9f($e#HeOO(w6f^{!tOtU7!}Mk6lB9-)b)JJMVI4UJARZGO3Dea4d*;>aZAox<<@br z)!8(_zq9gmT}^zPj3RtLGIZz@$W8e911D#^Icw)%x>v zd*G~!vpv5!?WI|KACTu~U@DBFFFke%MI}nX87Gx&pH=ucKJ-$&_cnnUhFibx@pf@U zvbL7QL#APxfxnP?s@zJmZ`BSD>3@W~5s5>aesj=!wWF+Dz|ek=4`;}D#a0wBGEGoS zbiHo}{O&kSs;)6#z8vQ#qce-wmuK57)1oQ@0r%?oITgK``{P8JI6Q%iT9efV~!k|j|LB>3&vj_&%PFfljDOYW%jdu4}TQrAK3x!L;^CX92Nae&g*D35yP7;`D|(L>dPh5Ks94jwsrV6Yb*8@tF3;HSvQ!-(X$ z&f2fO?>u#>nl&T?wq68wy9of;ay91Nlf*4g&7Df@Hul~9^{k>BdcRMqH2n(c7B+Yo zp3K{DL%Ckva&USfuk#>KG!VtmM%V>#xywBID`MA}?r`U!onNzD<`Ku22iTVPh+Tdu zOO&4$Sa)3v5ba$Ik~MX`QS=o%NxEHR_WavKj+wbs__$zt&swqiA=F2I7gV3o$*JnqxxB~Lc3gXVKYitnPa(Nx-;25j zV48=M;LG<+M;WG#2qJc>M|qkN)pW_VoCBv+9XmcRt#kkOL>zK_(bJ!gYx;ME+@8do zD5gFROXi=A%0pF!DO8Ph{n+g{yDgvc!e@4wfw{AJ+HtpTeZC$TF6W&FM2C~N!p6HcX2bNK5Dh#p+d>ah zDUj)*mwLxXPKWSKK1AuUuYca4Ph3k896&#<1BD8NT{h>9)nsIRJ-7qH*NGw1;vU^@ zfS8yO$a4-;gYZ5JH~$VHA{;spf}FesI{#~d=0}byiL`iiWhiRglg?ND2EIOk+<;?y2G#N!0>= zfrk>|flSK{vv}X7)YAlmCC9Et8kcePyx8XJ8ChGM;5#MeO5109@EG6FPrS$CZM^ox zw^agg*LgNBYBTMQ5sUzDoC)SstiJE_%^nc5HcU1PZ8SF>OtPoW3c(v;&p%Z~)x@q} z2|N!D#Z#SpBbpNeNO`m-l_$3SARN%p227d`jb(-5`;*{6W*84i=JzM7V}KJh?#6KX zejrRPChBu(1$}u-X4ds;M^Z*y=UVzI;_CxMUul27XG{!bAKJNebc)?dWcb}fufDeD zjXYE}Z$(H%&6kfw-?Fe6mOBlSxM}jiZ*#!CMf3DoB59?@%8s8HR<+%K2g?rQFj{WCV^3z}OFtQz*|CghBG@8A z>yHbZ7LgV1+%4If9=(EbL6$~U`TWqQ1t~@VNjq}2I=(922MdWWZLpx{K(}_mWp~vl z_ey&_zua}jjxb>7NKb2PO0yXev(xqRfeOJ zV0jRtG0(Jz!Xx3th#E=hLpmG$QJno{1D&8=@}ll`h^4JUj?rM%K{KQXwq7znEF5Gs zu$wnH^*!%}0&mR{`#NpGJ3$DDRPIwyP-Y@lPIL)uCxB^*cx|+>wq;NpbFG4y0^aoU zed{_9>BXnje+fy0jYDYOKPaf_iXEjq{kRij+qBk56N$`Oml7<ef%pt$359Q|jlkIgHKJ(D-d!Vr#DCZ;YvH3#TO(@S& zG;=8~jI}IWZ_xGZrs#90l*+AtvY%iIX+|D-&rEvkf_mR5DCpB9%;ho^=s8*XuFYzl zBFDX$BRhI~*2iZ2{esQ#&*rgXF_zL&VlL3!0(5|!ev|8}!ysOEgDdEy%dEyy7BA-4 z=>=7lxZR$73PLSL6AiK@=Mfs-H=X!k5uqix!5=cG3ypmwWT5vNuuJwA{rHonh)f?y zHogc43MPyqXy|*GIRl(B1(uYB<9k?4lkma=k*(O;75Rw^tI2V}oq*!2|2^$)-qZ5v zw`-!S@N^Jf54HG_V|$`wCmzfR)1DcyX!ca{q#b)Z*|MMAMuWWh-fGLj@>O~6u?n{Q z%h>Xt7qNZLzD>z8!fTI3qfqg?ro!$O{0`9&I;<5qGEk6uPIYB)Kio2gQaFhHAIcT)pCF=t2I^B3!LlJYzbFCo}pMQ zsb&qL!Dcr6%wEkTV006%?nueL_<_?tg8BE><)o7wZ5idLmM(r7a!mlryiWCeFQY2^ zM}Wnmg1{7pE-dBoeIzL*j|gcxb@Oi>mQ^)s zncR?pQqkYBcUSra7QI_EqM-rlP<4-hg0OVhyvX;+K$AV%kf6b>goQBCpV~3qSn5e_ z^P-+11m^5-Vfp=R&daasRu{4`JtFm3@c*DBXl2yS{A^!OBN?G zdV(RiR@(Ex?6XqYJ?r#02$pcYYzjj5mlzsfAFbH)oXjE}S-E{5m;YP`<>rc;AS+2W zw&s`I3K73#KG0H6DOTBDH5LI=&$NW0xBwKWw;c@tj<)i|E9fzL>cwYMU>UX^@xiB?6E{Wl@8|Be9P@Y`SDmF( zTZj{ab>(`nrqKyp`R5I%Nx?O-X*U6ApeKTnq1GMhx7`~J9b&dEy-eBMS9~pfp_o22 z%*S#@-UkR_&976EEbb$N#*eDHIva8W)k&H=fw`zLR>Z*HZ*H*qHfYxeL|-=U4-a|W zLueQs`55hoTYs{S(5y{*$xVky8~PE>5VaPhExtgkNJw8B>?jm( z;jr}i1H^=#A1=iX%1{|#aUEVGw{(_gYYQDGmtLnmx~?#rGzH3=L6sYFT_}J!oxUt zhlL_!9vOJNG0R?ZUJ65Rlu5<&(|8ggZ#vK((Q6Sa@W{gVBC_b|FQ>D5G(KGb2 zT)oS?g&r^5EnSy6b=UH8m*6`xcRL0qHHJHD9hDnh+z&?C;uTxxPllDSau*8uIsO`4 zTJFYX!q@R}MZcAEchWj_{$+L!Cpc_eKQju*yYJ8(wH3}tJjx_cyAnDg_fE%L`b+vfQ4f?AfCKVauxfx6!Vq;#8jx^Fh zXV7mH;_|m{p)xfWCe~lnv=U|%)4^gg>Tv#TbTGPM%5qcrwYh!zlBb!*EVI}r&=)N| z+Mu5(ozVyOHkxMS?V|Z9A2)|z&Kn13HOI(PDZ&ZS z8D-`FK2IG+7f>Qk#Tuuhxs;7(*b?a#VLkL&lWANA^FGYDh9*&ur&F%G!E`e0_|!g2%O z5#aamv&vMOxNDdySsmrvmJgW4JfhqE#qJSem64_GqGIwT&}^tKhG!;7fMTxF8})fo zKz7kI+yUpzjAS`0R1|k}A3giwCdsvGzSN3PwBlQxB|Q=+FM(QbhU?f!*FTp3mF%cV zS6QY1E=0%au|C7UeLshLYR9*Fb{^;F9WTs09w)?*8{0i%17_zGRZrk6M^zW?K{cu9~VZ#ewux z`L4|4Tv0j`gQQ-}o(6sL1zwC7A9tARD<0rmM(9e4lvYXgq{tZQaF?hw8q+L?DP?rv zhkn#faWo>G>rzL;6ts@)xZAV~z)W*3sLo49KzDfJxFG&4B!3{L1T4?ph4zQScB z@^hPUr;7nko8+H|pE{Hz8_qOb9Cx>(xy1KJ%yWsV=n{wGW;butUl{$NA33Q%Sl=&A zd2C?$;aG`umNHT2M|VqumpW2|f}xRA!J?(Rp6q&Xhvk*f=9GTXR9lNT8xq+CyH;qxVv`8Trq*2 z&ysUzzR4!@3U_3cHq6Kv*k|I)h+1(tYrbQ1!9lHuvG>mPsqQ&u@e;I=(8NampN-#-rJJv0Hv_VL!i6 zeu}{TjGTC=jqm-P!7*>=m$isAI*mK1J6G*$&{PUYYW!rk`i83utD6>~tE&MrTz(L9 zNNz;P?}%CkZ9aQ(4RkQU)hVU>v%pno$!Lumlo#o^!^NXvZ&}?{$Aq`FpC!%@-eFI+ zvKF3$h0oC%QrV>Ab;GUb&IDf%Xk28Tde0*s>Bi2=< z2`*j96i6hHHjDmvS~gUXV5HlU5EI@XWlOXPLW(FY_w&!y+*;zcR5jhTGsem)XHku3 zVp=b-?#=R2DS63S3^)EB014<4tgY)hRG(pse1EnC$T-!%+|H}D_4tfk}dg)a!8%n z7wWMDWrs(ueK*8;m39sm2_xx8 z5!6SYms8aEG zw|Swm;Psd7Qf<^X@6kzjfE826rHWWZ=pLh%pt9o>n!#$0S#>T3GO{)YV8L-$H{PJO zE?~la29Rpl&~D}B+-QLOSf76^m0>!~Y$r*1B;vrgk}baV^|(LSJ8wsV^aN70?(~Gt z&aEAWeg`r*MZ3ZX>!Ve0QQ4w@@O#?YI;bS!5zI1Ju>kFXo^%Qpk-YMS^DMQ`KFw9> zXgR=pIA4baehP0!lVMUwbvg}7tio!Oh4U8{rK8s?N~oe7X_7k4k28rCoM$4)7I#bd z$Qs`)A9b41+!-C2;6BEpx_qTM*r%6$$53O;K&MM1GV};J>Y+ClU?Kb;Sb|2usjJ&o z8~t;sU{G}lIkiYn1b@LxHK!EPq8Qvo?Q+TObt?uW-Y4n2Lms+GQwbHV?U+YK$1>A* z8kNl^-#$!P2>#fJkqzyHxFhqe?vzF{V{%;cqGLTxV6~3bcqT5=U*%h(zMo)7Sy-sW zZE@E(&#-OmLoB-$H3|@;Ypa?1G23$^v|YETGFlc{PcB)+QYh-(0oN`j!T@2e8JA=V zgR!XQiE0JI2c&Yjzp7Qt59+~yIhzCH{WVTI!Q-ZY`sMH z6b&m%{&?R;ptP__JCrHy9BL9zV(f`}L*odQ2$wr13N+sD1-k?Aw9`?U(@pkt5HUbR zXYT~oFcreGivWZ+D;{Kk18t0OlO-FWujD~i0vbLcGsASvcLMdVC1UX_FnnbC;%cCu z0t!+@n~Z+_@7<4CF8&C|RUh7Um2y~PKcKf4EaL}HRQ-_;422(uevbND()RL23q4j+ zUdwcAG~xnC0C?g>`U~W}%`qs#iY$I_dk8h{M7k(}RlBd_eqrhJ$tJh2U0zwooi-a!B@e zKjQS`s`enP#zheBq7Wo#coPBEYo<_JMW~jyj0KyPJrQ(c6rif%y`|rCNJZU!Q;$R} z9idBK&y9wYLjgHY!4HY75uqMBg*d{9+w!esPCdmP$db1q@l9MJocULN7$Be9U$g3} zY~>kz3_`4LR=VjUsZ84kM=)70b1W4MX|rtKk6-y5Y+)8Es_cN)Kwq9k>Aso?YaUS& zMyj#Lp&_2T=bm}b{G{3`W>w@+Cc~=sgRdMu_vu3=KeTjO6r6^+Qg)@;H!4(Ct7Y5mZ&+ZP&rS=JsB?bJlDBQ7f2oVZm5>w1o7t zC@yn)=a{1U(1^Zv&56cL13CJaHcvo5Ln~f4=(2)*&1Romj6};kAcOqw<0fGodcsbR z5}!03-S71oy4-ec#U!0um|JP{Z4uh(`r6`Zx7PYF!TYtyGK~^uKLK4hdm|F8(iHeI z(zg3~^;`0&><;AuiF?|*i_K+&9G&LCSY)ApT<4ll&2gvx8P$hETbHQ`xRkp%igS2z zU6MQ`@yb@X!FeeP$l8*5VdvBQXMUHVH(2)TCNq6J0~YU3)a9O|Td>ynOOeD`nXMD| zOtz5rCRlo8wXj9AYht}7#U5izaa-&%=B4?_koz0Z&u2tgwu z%+SSE{3f7~P97=$?JBpemRmY}M_V)7=iyI}hGR|!wRoj<7JhAn0dm3oR-RY?mxs8? z{{W1>$y)HJUj~ zhr!>=vX5Pyci(?mCl?L%i{iN;BhqNk7g~@^#?Nyc&1NB3CwPatqJn*w$xJE1LQ|xL z{rxS}OIj=5^l(NUfX^v|s(+t_)(TIE_a_*Jncr^?SM%kMqOXix1gxiKQn-8T&}ZB> zE9`yCz~T%${tUFpxCGIXOtsJPZ9oc*2AArhSpmlmkPL)7h$?X3tRephB;t)_-sD6e zp#rg>g9KfjmWsd&&+=Ez&B3AHhiNm$I=!SLL15ag*^yM235$P7ryb(2Lcvel&FXnW zbFhb&O=1EX{mE6>gWkJ7O)Nh|Q+ixg7{5{vV_A9gEE+;*2-%`BvVb33D6g0+_qp6h zE#IbFhcge$*WZj$qIU%$0n$0J^sQT>syp8FhRPs!1%DKLLbXb{^xpE;*K_cH0qCHR zvg!1p?dtFjNTwcakNka`WR(d0^^1q1>r`Y4cD0pg82D#rQau#lpH&qFUA{Eho1o-b zM8BVxiVjUIJ@;$)n)$PPO_BL5w!ceK#*Ib8Angtv?+j@vJA9iY@kgNB=REzubQn$!SJUFKdTKOK2rAaNID$C z9>G{^Vzv4yyls&vceXBL7w;&=t z;OsC%X0gr22H~ROpFdlwH{rzDc4AO)Gd~*j~}hZn@~Gb*~sC?;AQtL|jCov7eOtbQN5KUWBPjOzdH0kwBO<(o?q zj;wx1)_LZuZZc1MO;$;))Ff`VcGI<+qLbf2&5A_#9^BX_WFic*m zVcDs?Hz12#oeCFnGed<-%tGNmm{d_HNrtE08;BxE)HJ0p{}lo0g*ph^B-VP`dbE{q zN;=!qi^NopLDV!vZE6!j0I&C(_m7sdM&*E(y}52aV9LQWia9I55*-~=+u~KJTrXH5 z-7iZ1+qU5WOi>D|3$s_OE{hP(<`CJLSUPS(8zGyO?~i_bgVnv@ zhX@>6_N{pqdHzBO7c5*d^uB^yoVEO;voNWEqJ33lKeD)}o~=Uma8#VE1> zd#%_uiG%#UMfC@i)ca`7-g~Fy4Y}z3>IpVl4Pb3zb`g|3eLvfYpN3GT(y3ZRu)b0I zQmV8Aw5(G3zV;Nn45r`C)DC{juk|eb+4^V$84#Y)?B|Me2c`4fvOB2y)Fj&NJ!>Wq z3-XZc>~edst0|P~#7EjE?oqG;Fdo}Hj$QABeoU;B-Pl3tx}y7=Y+$du}6 zHaG!d+IlqpgIqjjmbX0T$E5&7{O1hwbSsnWr^BjpXWV(%NiOt`HR!m9FNy(`Ov~?N%@syx!5<-Y#ltH$>_XP5B-NYKU&q`zQ&})a$eeL?thj1jQ z9?-xA&o8V|@y3d?;;?44ZzvTp`Dm5*E5ZEZ%e7zWe}sV`+&)dnwm4n zj05L^pJ7jjv`J+ z$WAgq4#v1LLJ4qfRkrr?pF>6ZGmivr0!d(PQ!jXrG%3LIR`oCh( zm34hsvY$`p0FCrXvbf%4yoM3Ait6sioJ*62Mj_;r5%HXB!eXh64ZzgjgvB1YdZXCC za1B>|el6%aMUN99?=Np~Bx{h1HL{h85LIz{ridGrWEOIkRc)n5tO_Uw`=$%jip1X* z%Xn1B)MTJ*>9I$-p7y|#T}xvT(>YUiMsO79956n3j+3F(#HkE2FZ*s{KSQiKt-Ec2 z2|I2lOL1C&B+^VzfKJo;-^$4C^Ho59Tt?WDFj>0uLTM{KubW(k(F-EJy~GC3OIX>E&|Bo4RlEhTkZuPOFT9i&CWPEfb|}a&$e+RkiMT0OCs`f?NXS9YSR#7rup;?HO6N_r@fa z=#jNE%*~qtdig_NK2piZc*)^lngbN^$Wh-nEVvr8sPLHq$(<^3%9HW0YLcD#tUT{Y zib9g81iuiK*Us{TOJtlQgc%lW7G_aebnxXcb?%>yuC87BM?@E&} z8CvyT65hb!$P#;KEARj;Gx8QDvm4u*Dn?qH1$ya^8q}H66%-QY&LQ-4S zwqwpmlwlZO8@?^HkuAi!b8hcic3bP|kZts{LU)0%gwalyTY_Flp!b(H z_<*HF*vUOW8EPdFO=vcSo+pnPES3aw_FYai1JIZ(552@rJKV|Y*`lrI=(#TS>uUq8 zSS@Vrr+)(P48=?6+AMib3BTSpb={2Ap@it;kAzXD_)9eUSKndYsel;Wq$yr(3L+ua@-Gb)mdQgciMDIH(^MVtT{1>}{*vy!uVdHOc90)QxrD4s zF}a5*zU+OyZ;=CT>|kETY3dkJO`)_>I}ON;cBA4RDlEY~;tq6^T1|EP5lQbBZBu8s zYpe|CJWO~}>n>b9Nb2*~hzglH{|XS;D^lN$7UpB#O-{GmvgJ8q?WiVUKf^8GmXjx5 zNR&_6aMuhqoPa@ucLK}dfPi2r%#F|q;Eb)sj^MN5Gkx7)V1-cpof-o=KrSLp7JX10 z(FQCL6a{aR;`{qgl#yNsG?7hMVI0axJE$B8G~Sy>wSm&WV6T^2#maa=GTV>gZY^Me znki9a0s&}#-|6hm(v4cF*|3Fo86BFb1W5%E{||3(8I|SQbqz~MD2g(j_8D zs-$!&-O?Z;N_UE+0n*(dBHbV%-67owh`j5>?Y{5l8}AtJ^XnVKA8xi6*LfbXVy?O7 zqE3?QkmnmWGD3M1-g@}-^0^@is#mpy+xe`jEPLB3h~J3Fu3+BfTC%=E{doBpA62>< zM`M`C!I`HCO48h*Fdr3JBC zQs&3qsg-eiP4gg}$XX-o+Rh{N81FXJb9xj=anYg4Ek{1QGdmpEHa9O)zwxw z`!Wel^2Pc~)EjjOUIt+d3BB$9(D1AzZM;@x{8=Wtguix?hT)xP| zQj_%h1OL9~-C^yEib87n>cJNsOZZ(>KQu@4{ZP-4vU(umLdr0&GIv`mE>*Fc7N9yS74}5)#HHK6U4yY`{2;I0RvmZhr$tfc;Jou#s1mdKc7e8TUXoikyvB5uIf|l9y8g6i;XK7imEs~ zvQj!%wtBE%(7fUrVsMkvK3`^ATW;$Cu5$P&QK-5RP`(UBP~A81KtpEJ*W5oof~*g)&53M}8L zlAvGWVAzz+S(=QywlexH{iGJsGtEjW-lUsbni0e`{q~7UbCpDOF}zGZU3{s( zjfI>5r&DR?TvGdjc5z?_9SGr{=E;r;D;C$aR3SlPXgw6r`IJeqa|XbDOU})G0vtM3 z2v7<*N0N)#PAdlsS8f(mm;DHT?R>E0TG*f~wQ`g4-aET`rNFaJX#uGe$UT`;+VHa@ z+<$zCy)c9w%y!k`mw1ZI1u66ofWXg;48)N5Wbb=aMzrNMt6Z0gly}8E6bQ z6DR>^S$Qz`f?U;fc_&EafYCpuX@<)er2L}>ZIw(E%!90?Rw zSIOF^B>M5>SHX_Bw+}!yqO9X;Xn+>|)@MbJbq?EDz9k7PwcL6GexV(8p?E^Ai$oO3 zdKpi=3=#l!z`1b)xj z20DF{U1<}vH4K8IQ0N=v5xS)zwRv_LcKUOmMnH6exxa8%mwVjd+x#c9``-+cF55{( zr`|icSw4mOSh2q(@_0b4^|cE$UhU$beqgpU#tqG|C79gz#YWglyrMYxlh6v<@GAgI z)|lB&pY7a5UhZ`pva44eBHGo$1Ah%{tL|U-W5+y@WJ1esyA>vP(MAP_1*0Lh2D>k9 zK_pgs>-~$w)kHjZZ7L~tLhcN%-$olTH{&gE|7rnDnH$}(1FbfeY`vY=N)PY2kZP4* z!$ywBqd5t@jDxmBrWd~KA?X?`EEm*dn2*_hOs6P1>0Nhw*!a;=`@GlQfLeiV#N|q3 zxrC_J9T9qMt{=C@8ku$giHo5v>q3r1;1e82>Cb5TqSqgNYLViP0X}txg>B3e$6|gL z1}+Iw)lV&Z`lJsAt8KO0>)%V!vdf}#8*!5GlhYOR?dsdnh$I||$D5xD*ho%m*?R)vuH?RJ!-_6mTi|T{rlZ>I|w~Y^|zP`E? zn55?1f`whaadLc6_G7eW;pH3wKA-hTL8EtVEVx&b7J9N+1j?7CSIVR%Ar z{XT2*)$*)&`P~+pOtehVpCDeiXNY=}n-!3zQhO}rxyuv+{U5zQP^-TG4GmS{kEO~# zZDMCGJ)fa1-q~x_W$bp>W_fqHawsq>Wxs-WW+~~((o^UZ+p<_@-WH0)wVBVNIE1TD z@R(a|NfhYChO^xB0Js?^258;VG>JE6wU#5*r@yhE_Tcv-Gzv4^#WCxbxJoSSQV%_; zc-r5_le;_bGsY{I_SE?7Y2O`DV6iJulM6rinz-t}g1KzLQsMy~Z)@yv^iOTJt2^b> zlssmSc7v|3g$I_NW8Z-Aw!4AzdSV?as>k|lIRO6ULy*y6Mg+V$E)S(GlYK>VzWIW| z*!}FcxR*vg%0oA_zl!mwu#esCemcL$SO=6&l^bRO?u*-7Ckj$lY7f?B_g_mn_h{TOLE&pQHm@z##v6_VbxhtWt{XqEixdD8BrmA)r9OP7|dc+5NI67sfDCc`{R!@c4SSthql8ZSnl57WiNP-*wkOa>a5hR~i6>-1KT zo~={{CP~l%L%b1w0_BORR?8c|Ji`uk72VJR93!k>zU*&a7Ala0s?Em~pLmuRdSp}$ z?N@J-+Y4)+_E=IQwSj3g(fgeP}z&8@(pxn`z*J# zoRtQ6y&n)&Fuh&5U<4s`w~9`7M=!P<&Vxetc@n%R_m_r~vU&#Ef@wW6q(g7`jYR3y zna!L1E9;`uVo)gFDw0CWbQDJ0T%lwJEy+-`80y}p%5SrHU0}So)RX@BhVdp}?Cv!= z&(F}x@qc^Ni9l15j2M~WjggeJGY*)4d~Zi z=S0)jbC|}G+Ec{E!u@bWCvYSt-|cTVo$M>mskYwAm6K_Hr0@kbvgejw-8r|L-}VXm zeif-QeYMldN$VM^hUSmd#vk~fLl!gI&p-6!ma4sf!(iqHvt3o1Eh9s3`r;*KFX+VZ zWK-!nV!GcOYJQCEg5&mR)~PQoVSn4VO~y=rDoY(KQlfLkXGvoet(tqs4@-p9_r7<= ziV*WI-zFT5dx_3;XY=yRMs`tGfc*0Q!FR!1?2m*^`fJNqYbj#}D$lX>M-G$dy5oJg zG*8*wv#35Cprtn9L(6%+j-pmXa7fq7<6fVYZC1>&?H5{0|86ZJA+JwP7k0yG27}|! z;sLm3Q-lW=rrZ}xDo_k}3-P^r(-j5sXLk9byyo%>Qcmg(iVAIxunNB*Y2D+-bZtC zi!CqQB0USL=D%nTJ{gP|nR+u*&hqTK+uoO>o(qCVam5ZV=qGwzV8Yos;rN4A_o5{}c zvoX^$G>ppUWr`J~sw%W`S`N(3hKf>hG^2-F+vLpHBwhz&Z;x4Xh0}vjM)Pv^t)`Qg zDflE%;`nw|*wrEbc&NM@@ujMBP)dyre?e*GoFaGL4a=(*g`iByj+z4FB5A`<^l~jN z(+|;UPa4NVD}`&S@=-0t`UHEE6o|eZjq!?yXGKmhJ#0Dd)rq|=WtC`2uWHYtmP~ZI zUKC6;FM_f+41|)jl_^&nE<%ziRM~d44U}3GV&zPxy9sXyGcD|GyothxRz{=l>L_2z+MvI6%S{<^#!x_0txi1#RMhB=NgS$|7hB41 z$%k*ax355-=g^1V!fo8_{mm{t-O+0OWRLv<7qPBJ1JSuf^Y9OCDonw<#dT;IW^|K-~BYWm6l$^hGkd!mMIAy?^Offqa(wCoyQ(t@SM!f zQ_F_&3!D;!h_t%mNPfe4_62>KwGA7khq20Jb>vKio%o2vkN3N4&8b1E8dz(_Eu_O9 ze7(!Y!*!>{IAP zOY>a}6O90nwcR7RJ?Y){jjSdz5qIa^U)mmf1FpAs>;c#f?xSB*TmI(t=;XH=zlr68 z(`n5YjF#eqo(~JI$Tgpj`+Dh;;~>?J_lPcqO&vuN<^Zugm^VzaA7Y3Y6`k3Q0pI1Z_3Gl}kJc0QQX3uRaNz0BBW;QPk9&PbrI z&Xp`c(fzd&A^xq~Nl+h2bHn&>XRtQ+`;H9ehLQ+lMMCF0D7al)_7;^hd9tRv4{lNl zHLsn+eDW>uDg`SqN2zG|+#9kiin%Et9*7c=h2GF$7;`J+$pXklgrM85CSE1krSM4z zO`~ym(MRMesg1SBwgOTLh8sjlJjTh6UZZ*Qmo;xy1S~SJ>EWRe7P<;kIdPcnd2@@z zh%mmjk%;=Nla3vrWn;^T?Jtzn0^GXUD-GlR zou|k7IVjG0;eH_Y!H`uDc89D!^ZE@)N5}oE#-EL{-N>4nRS}i0J#Q8P=$c9JYqvnR z;Lp`=MRa*WoKZ1f_U?q!hnl@%7In0Cb6>c$Tu-^iRyzIs^$XSt**xD?l4`fgOcC?& zmmj}A0myrsqYbmnUc)vbDyxn}r2=8r7keq5_c{Z)%nSZL*sCUu3u;T!57 z%@<`+tYM)WKjj+KgiW#w^FIHa38cHc*%SW0=h>m&7=I(-!|x7HF(nD>tVB{L$2p=x zt7juWvy9GEDTulDNdH^X1 z*iy(vd2>#C8a15v>qOY!xJ zGh&O*tUNa(q3)k;)!s8Do`1(D%m8|9|;(IysSE#bgBU$lyay*H_!TTE4 zs1ByD%v{;4Tc*vP`Q}abk|+Ra>R~r@J+e6ruA5mnMA*Q_ zy&}KM@;-&B2IMekLrmu-Tnb1s%r6C>JS>oYUDZvnKK>zA$IK;#)y6vz+uprd7Z@)? zb?CDoo#P$kK6dQQixLxA*~~W)N*8jQRyOi|OHNi!LsfQ`@RKfZ?XPJJl>d6VD)=}g31*qlkZ7R>venvVp$mUgs z{A@npHj?en4L*_E`C?ECnkB9K#;%jNLYrd36vv;|BqIv3b?6 zG;dGnR8wv!AF(mhrq*15+@~zlfH4F(FLk47Q_BhTl)!$!zPripYk~cu#%1s1E2EM1 zg{2& zase=eUScHTRX&+^!7x17-y)u~E}pa@&ytmH>0Q4jg?Aa80szZl70gVyj~Jv*eQYn3 zUhi{Zy7gj)>cB39e<6RYK$?k_6E8Vn>}NXDStcia!$#gIi#F&yN0-CorH*I=Pjq{6 z*U9NlCF4+lYI1vl6*ym4{W`0E6x4VuYxGL;PztG{OygF*9T~HEBK$zFKO}61Ju3BG zbX;;*uly7|kMs%Y;@LREue3&M%~uYfp}KeU+Aj^YlL9d+f z((~9>U8nq5+iYBKEAvtnr6l94R&8_#16QGbVOm1ZH>G~jgi#&2JsBF3cjW&$7gT_xShhSD%l2B|hN4GT{w2gBZIhjxU?41=-gk;@tAB9&n55~AQQ>Bk)lGaOR?QwLsvgY`nR{C5O!3J)G74%Wr*seIt4NqH5}JDL{#Lz#yUg=XgQQYjMP!ll0u7bkTc1X&^{VH- zdU4O?m)&+Nfz<67gpj&>*}u#_JpOr(wICX2Bz&Kr;H@BZxD!5PONMKtY|~Pr)Kl6` zqCbI7$u@;g)g6BPOKjnesETTxaK-}!DleavWLhgkCB9H5dY@EL>{E~DwTm*Ta|5LX zjLdeuuj(@Kca%x((>8KwUxq*d*PczZsejE5eS$#fbj|*ev-B;|^HMdnEFy-53hjrQ zY8}lLJ3<*~%NIz#kt&vSk6J(ql*8)sSlxdbc(3w@{nAx*aUaGvQl4!PU*rJFe z|E|bqbv|<;a#!Py(gmWB$ToQP3sP9P0-q*zba9?eyOLRD@MKioZWxDVy1o3NbaU)i z(Xmpf=mx9GscxYqu@s*4_}y_B>P#!oaRd01 zpGT|ct!w&L4SA!kx6p+aU(lFa1jx=~o;;EEkFu(gp z-eAjG(n<^%%l6O9OnR45h^@OEwUN^BvXNs6b~y-?dHd|Ru&`j!2`&!Kw9oZH|^@F+J}|_X0cwBn~!B?8E%Q4 z;_&^Tu`FBI*wKg{(|ym@%B6PR>9^ADC3im&(IbCP(xfi#CC_)l5M)D zO^1udm<`vZ56aj%yKV$t+yNE9r9|{5iIVsFne;8x-&*QATd2ER=3a%Xc;6*FcUM)H z%ycOjORpsAU00xAyUY$ES7JMDzQT%Hh)16RLl=Ux6*7yMDst%))9}<37C;Y+oA%zw z@%2HMyM*scCT#xh77IfovwVHX+ zvBb}&?Nu74p>q_@Op16F%9cbOO((T4i;uFmX4<%Q_p-Mdd52@C*AIoYV$h+P(Xnjf ztP8G}WIrl~a(0PJM&PPx#G&V$x2p7Ge^1QrOCwh_pbYvVbCk zl*K3=F~eP$-|+Dw|Bd5}@1?|RY$MuHipSFNQCO|ayjH!ujO-6>vn-8f>ii7)YFVak zu&4$*kNE}a92P)WrFw1+Yn|?Q9fgO;I=zU;h|}bX1Fa5Gw=&~$>d>82Hg=m}xh}F{ z-ufx+NNg-Rw>`4qhmYUwEq4Qy`pwVoTMK$+_(@rg)_IrvLQlVccIhcStmQuH_u(7sFAM;%Ec?zJ#nq4L%JagJhO!x5A(8Vi;n_ zezl6MWnAA#{KM3Une61>=iH~XsQ-THK#jpPr`m5(hdB9a5LIJ+pPxh;Q(OP8V=&?q zls(5Ji^;?2ZmB-UN!_7O>$sQWr6}ruNL#>J<`{ck^)4~L^xn;f>k6}vUXLX|J1Oj% zCW~I<4Ps@0?HW-l*bG{@(pIT3emF&ehwE@R+bCFlpIjr(vYCo>HW4^7cY#h3_3du; zM6^z^La>KXYQUFcY7F$dWBcS=`-aE&hiltgW_J3PhY9(0qn@Y|i+Nod7OVb58dPK4 zt8eNk-&!jt~9P0V@6d9Nfq#7u8 zirOc@({WA$I^qFdhD!M8@n*qYkLc&w7{vriq>l{4w4K5R-0y0&eQ%eb86#mpr`a^9 zCn4Aw!rRgedGw}I^0}1cnN51g-6#jDy+CeK-eHN?K@rV0W<;;9D!kvEY##KM7 z9-~tty{{;^sBvBm5*ADc1>g4?2o!hkH8LR|Rbm67s;v3;F}E8i>S$jJuMxZjsl)qg zC4>hZceUKUV?V(n7Va$M!o4-T)2xvDHG)D_s2mgL&zoKkdwq`eYO${EF+q0mI!k~S zYXGn_eX<86`(EV3S~l)iY_clH(33`$#XW5d0KQ6M;Y|0=*e!W}^bmn%CiC-`rE)P^ z`qkgwlgh=PW<*fAFfl=tU7}44^sK$xxK7I|mnhqfy4&{Wex2!9>Y(R`sg#lw%@UGs z)Zr)nse;Ob7OKJ|iq7YE=u@gBRw|N&sh|*6q+wmNXHmg{;tC2u{hFBN_abTt)_9D) z$OWgh%uxg!IEE3iB(_1j3s#h52n!Vt(@?NC}rXuX44!31837iaa&Y90}G@+k5a(m)`OL6Mto9BZ#YOlNoIv z`<38r?QmaVt`(j565zZg2(FiZtxWadQP#x*HLhe0+eRWwmAuMj<|2@{Iq)?r3i-1w zJOoSk1GLJ+U%kd%jZ7d+a-PAiN+#OJpD>D)k4QUDB?=;H9l+tTI(YlySkgk}OHUa# zC>ED(h?*-R=Sc&d`et1unoy*^_1ODOkX;&B1vcjbxVRF^j)8E7~!4PT3o(-yn+bibReQ=fkmK@H6lFKH+HOlehFZ8v)tMv zvjmf$o6{zc3S`#^AhNY3I^OssR@iQCbt10qtVKbNsoYR?i$m0Rg5c$AYDTiD{+tm^ ztX(20_I->bkjv`;nU_wQ*ViKM9YJhUbi9=0&?jR8@SA)Pi7O(TTHrSJ?b zSZ(XfDIGZz306h{wH&PuwO>INK1~=f1mP8agiXL@ZygYZC`HfD-5>``KW=H=sQ$dgE&7 znQ|uys;6rb^#=`A`ATxu`_foi4!u%?ZF8UvSv1t_Spniv zTp&7@9oOyhpGg5UxyUr5^&{VCW@@=PdThhNNe8$d& zcgqM6(^a0UuWFRI$Y#_fxFJ*oSLgbSMTw>f04Ka4F<8dG*Z+`Uc3yZ3V&d;P1U;vc z$~e9pl9M2|ZZ%oGkbR{LYs1i|Nl^~6YImG#w|jy}IqH$*z<!hhXjBed}PENV% zHA*>U=K?w}d^T);9vY;XG@Eb{;-8&ABz!2R51n~R3hSyqZ!3^raSr=7$U-{mwG4rI zE+2v5$71Tz`M|RLax3VA{Go~Udzk>yjU*_Gk3KtgWAI;{!6DdixHz%uE*&ER#lq)w z;9_I!gq<1L9&R{TylWx(LKKY9>vB`wya$ve;t<_cgb#{g6-%~EV*_mUHwZOP?^LMV zqoF}(WTJ5|DnRWYQRj=uce=)|D9z0T+^pG+7}oQ-lf&_930(c*zE6?|SBbh29=?s;>S)+ypQ{+r8;egW+{Fr^E&EPm_~$r9(1L$`!gu5I>4e;^FX+!6`vZHHzKLZPoz4;vohgxQTR5Xpnre39bYS$k;TCjeMIDDnh%B2aU@;M$xlR zhQljb&ROcG{$t0_tobrSj~ZcxQcqegawVlH93Rmxz%3i?0I9S13zH2va43Y}byu(h zde4l1DE@PUXMp>2gQ9Z!I2s@JSJB^FxQLkD6K40wzMcP40p|`!?nE^UE5VsTI@|m( zu{HD~KpOM_27EFI6)*esibV!HL{YU9ut>w)&%a%Zz#=jrufZe>UMmef9>NW1^W;OZ z_<9Q(+v4xDS2S3F;;DbMFmw}jh+|S1)&88hsQI1#!OYpaBkS#*gvjqFJ!@9i`Oo@O zd&^?&(o3yFL!f!>?<4aa0K^yy`%jiYxtgp9&d%BD{eQmJ+xqj#AhU_ns3o8Mh-o^# z>a(IwUM>(&jy;5rHMI)UYWS~_jDnr?h|s3KRJcWGyK<2nWDvXAIy1zLo|U-B_2q^ z%PJ64jwqY@&I9}9$+HntMtcB?zSk^NzMaFxXNrfdqenA-hIq@Ki`aVl!_QIBkV9dp zz&{_w&hu+NR<~p7i%-tVyqd1Z!0yK6=XRIh{~mN$emP|M-4E2C{rB?sAG*!HYrbYoc$*6 zXQ;0!T{Fl895SMH5j9z^(_cHImb9;MsjTL&w%iOeJ!_ZwCsyG@#u)I8%Mbb@XXu|l z)O_BxA6w36fKFuFDx8IdAL+@wKx>FJ#w>dYPV^oBj2~UL`TUqzeAxdu{m;=?X^ZGW zh5D-ZN6cd|O1{pU~oL$yN}9u>w*tcg~{Evp@#^hd%n0*&(zC z@qwv$7q9<{tRA0>_dmy{Uv3H{i_6-ym~0snEDGP`b+scw(?2)I1NJs2lqc_h{>uM+ z>EAE;_a{2=fxY^Md;k96|M3?kDzI5AhMg?``O<&C{r~!x=Oplfnp+(Bf1mOH`mg@$ zhyBuUDvs25Wd9#e^6#rmNx@`>n{q+_>v!HmHe6(|`hWdM&hUYuiv)E4+jkcG|6v=v zb+TGc;FxsYo(+0PC8Oj2_v9i0Rx^Ie>UTqHii>B}q(2K;Lh!Ui6T=J2)X&z27S1E~ z{_M{YvA52qDZ5 z1bgsD%=-FSB!iS>uDPuiV7$7%rVrYp9FRNNAM5|ljT8H3)C|hWpH*W#K&G`5$Qb4e zZ|xPhK+|1C&U^UJ@7X<%8vJ1HG6Y)+V;!QWzx)Zz*ZAgv@WhFj;=|3axNB!!lomko z#&AhN^QE-~Iuh4Ic^(zFzJ^J489vwDrG_0TMGHt4RDZI(QGEDq^4X~60{9iqXTHn7#7;BL#x=Sfx4+_YhnT=60@AW*&h03s7)aO z?#M+%9MlBrb^Db}45L;xd#dD4h<2RdT5+Vt(^TJVDR!e~h|`-`p=Vvkg{BYWea&)f z_BA(trs6$$YkL=wsC>z=!m*|6`h(sc=`sN~*Cp1vj-f7-S2|*bVMk%Kl!?Y_J~fXi zi%7u_yThUCrT5y)(v7cJc+<`bU1bXhqARd^X_#b1+B63DzRQjf(%cCM-H?3OU1i)}oFV6S*ZI&g>HV111{0 zllO4|NpS1?s!cBG@&QaNAJc@wt%tqR)DML{;Vq2w(y<$LTSUS;zj%PzbM8P*%W2sioQdQ`|8SbZ(#g`gi++~ zxxi6KtHvsM!evNyfF2P&vu**|>H?%j_dfqTO>Q(3Dc`Q!)9;)nKvZ*!z-b=eS>v$P ztH}pm_}xL;{D=uS#yO4kut}zD_EfG*G(Kkv(Jk*k<&ckEO|oqwi|!|2-_#pGX`eKX z)#m+q^(&n# z@YSlHM^N;}gZpg8^PMsb@$dU9VTgb)(%GV%l>fWY+W8P^6N|pVG5Fl^V)pBS99p5n zx=Od~f~J7J?Ti2JJQwy{IK{M!@lsD}sRTrR;dsV2({17yHBce1-mZWPL#uce)Qv|D zG16qe;&?KWAQ5RK-9aHK66^0A=!^uNAUNxIUC_V+HU1RLd0{s$YZ>0H#S^$WQJh|u zTyvs-BBWYz>PI_|jejWxYQ^{Xs|vTil#Q0U-|?6M#gc4L(>$l)w#W3#a@mk|U%C7e z5l9Og%n)J)+b|1XfT%%(UqiWzqMtzpJ%QQ&RmjrY6E1!G*E@2@iBV<+z_*XIVWwrQ z-0MX0WVaJ+hPV{EA?r3NxOYeQh@%>va*(LD;Awv8tG(i&uGLAH9g#@N*X<7iymW=d zP}MNiWO}T5#nq_GwVm|0c5SjXZS*uR&S$oX zfjF(rgJK=8YMDYhp`Tysi4V>bR}^M&MT;6~C}&F_Y-PL)kqz#S{OghwQluZx8BpcO z^A0zdjUg;wpOdcT%iqb-1~bxHf>m1vv|lOc;<9uQYuSjK$LG%ku#BkqYPj{+KuTi* zqX0O6z8q+T(KYI3Jo(KWtkrmW`raN9f59+Be#ToVD*&>wF+_Qn_(eG{*Y}O=^>Y`y zK*R*P6g5NX9d$or5Bc!0uxjhNt?_)z(-ozQ)L#JO;AqwhXp3M`?_F{-by3MKcfc*u z+~VK#%wC@x0@Kso)bJB^sPzJs=g-t=uW4h91N=c^07R@+InTt`&tm)~YHy`sJSpGU zGNeh<#GgLq77#X{u6=lo8H4Lg1Kctco!w_JRi$ zV6w$azkT2DR6H-g;lCZz`XKpw(7*G@dcI=4)KsZkeu%CIglia7L|5`b$|}ndx-jk@ zOfS6)-LJ*Xqhr8ELeKiM>bF8eX-1S-rBj^w0+tZA_`!sM2c&cB1fwoBD zg&$Cxp2+pfd@MRS^k?J=*H;{_@50vl8s5iK_{Jxeps8z4=bpMmd6IOaAH59k-sNVI-=7r=`bAuI$ z^!a4fQ|ltc$2dld^Ix&y=ZlUbQZPm35K&#)=i9lIpItXm3#o0e{OjGn7FP_llOA3< zjbRgg8=Tu4KuX3C8Ojp15aNo?2#SJ(Tl>vY&tt{+FqV$>I=H{dVI0r5@!rJPzIBF3 z!6yE^0-^2e2eNN#b6b2_gv#&Ty$#JGNER0~OF9-!<&fLQO5CGn0_Adb05Z0x77J zzJa#c(Uzpv1whX(zA9SI%z6XAxV-~J<)0;lPvOMZ(ZwU1mk5y90SM%N z)$s-ZPTU9;+AX7h!(*J5#H~X=j~|`2deq}!+~;Tg3zQffEMs+*4G=2s{W*gKeQ=pf zm0Y0t@A4&%6# z#gWi!2l7Pn;mcj?wpSB>p5DZ~X7hY&n65-Pabu&_DtwfC7;h&9G?ELVr4ig>a+I`j z6a+ud#t?u)#I+rmYZh0~OerXku8UwN_E2k4uOKWmi8dstVK~E!V1S`QicIbY5YCSF zFL`E5vpZ-1AiP7kNbs7J99BpO?l}>h2$tPbh?B*xgG|oR!1#j z+&l9`;+fIuT@~ydcm8$N<5)&(_WRIsEOk;Ey6SfuKq}+1l_-~#yp8ac^-z}exfi`{ zrIc*(n}r7|rRvxUlSxxhiU~>zm!@=if8wxUJ*?UAvH6KsnE&J;Bm&p8XG=$|0t1`Q z$UJB3#d8JLOZg>?RuEn;^O%z!X2|Y~IM@@mOJEQrt{FD)GfKYM;mznP?t>xKo#GrK z6VNpK=lMMR#6k;OkPxdgm~Vo~uK%b9#-Sfi%?;n2@`u9qP1&YP2`pHdip%I5)4H4SDw*zeKtIkHF zcI|9?HoN{c_L5~Le)w+gKURn<E!a@ot(-5t_~maYF@q|E35n&y{eGmT zAyG!_-yYfg$BMi1dQ*OAv6S|x$7Uqe@VCZ|{Si}kB4GMiFttS(^dEv*dGE2r4Drj| z_V<0RHunA=j1VctTRFe-2DujU0Wr?kj)xS$$?pI1#-w~oof3oRxsx!dUN%}7&{

n&|CsfORgYkkR0w}_3pL{Hse`r)^Rrdjd7NSL=7 z_RF>qCVAa4x&ug%dB3Ydh}ws5|3pdxtRStUZCN@wzuo&UsNLn8AW!!#Wg<+5V?+JX z?~p2t&_9vb?k!aLLXj@JoH3oj&%fqxhG$_b&y^I_D)yq4Ezywi@0avfo;qzyLOj)>@Ob z;Z44OV^<&SpXJT=fj5a$!dI^k60d(6sdcI^*{<_Od*@nZr?li$aPLNkDQFes8++qm z@U^_k7{kiloqX1T%V?3V#f0J8A!qNI@TQ1OY68rWf#ioo5)8|>^!ayg{Zlv+oJY&&_1!{%g3V5lY zK3A_f1yt4sM?&HdnS640Kbed323bYL^4k8bH;I=X=-tdtr@^ z3yASAtm|@z8(L8~0?XP7LJ`@IXY*>0nAIrP{E^I+T;sYK?$<(MD0 znxpt(tV(uSED9RNcgwpU;|v0Upl2(& zEuk>w zF&wE5t$7-MzlyF8X>5^WucgGQ`0xKr{A5L9SfU?+(DFN8$Iu?ao{mZBT&FjB&rELe9YoaB~P z^S@r^bpN?IH>4^`qWLuSCo=UdZ@|;j8-6)_$cX-VOQ&pRqJ!VqFaNRYkMe5v-^A@% zf3RQdYR$dul?1aR?V#GyWD2}o(h8_{2+DOQoIM$Q_$9rb8`tGp zKp&5WE0kY2_qJ;y&F8MBmB-O;I1Cefg!mJEnUEMG=Syd=sDVXoSERH|eTkOuGi{kq z{)gHRiL6AQD?jEc6bJ$2^|MLa%;4F-%lG!r@@cY(0hmF{u>jbwljBnjrl!KR&?7T(z&+r7{-^(Bs{BeJGhJ+ON^{8vLQV*$O;p` zYFWEtIzGu7(2_{e04W0HO({OJex?c_!Qa75So+xhZ0TynNVCyDult{6*7Blt0-!Du z!A>apf`_52MXh<|hnXel$tA8rlSuHVIL*J?na^;xz~;f<-7u2*W*x|aWFi^(%eMUJ zLc+k&akEQH4`*iN@V4zk`))YQLAV1mChvm++59}*DkAM9pG$w9`kePK{QNb+<#1}d zuT5N`9@39#EAQsAx7+rJW46w5NXhDAZ6MIf#4+0o#>fpRC;a4JE2F1N6a zd3(G4W;J4jf837k5U^cUUvLzW^QH)pn?wk0+Ok zo*gvI+e?l|%^epgtkt*DY20&K8Duk$QT>5TLy|oP5MfDm!}rJ-I@N+l--{0a-JX`4 z-ayf)>I?k|RMAXc2T<~f-Zfz!&-}3xt{2Ew>Nna`Vm{zc>GZ#E#b zRb-nqNe^#}E(Nc6TF|N~WB&10NINO8$c=T`Zcf15KH0||PeU~W{E+#ZvSffD*~uG+ zVgEh%O=k;~)d9WJPBkvyKXU}X&50@M63FTs$|%|3ejKji#nUeLXxZV1;AL{&Y%}cv z{JBpX>NzF^C7g=no`OD!D17`>*8^^SQc64EsuYpnSLpOJ$!Ei!oZ4Sr=u7rU=t4P3moWC2n zELQjIkyKGg3lAS#%`SWTeEnZ708_G7M?!>?#ykOO~!$l#X`9bIlHGL6Lxb7>c&Dd)*f8)PL?bShQdo zy-}xrRi0VDG0|VsxZ5zJZ;+qKRkyQ6Hr;N!-^oEB?9{{)SP(`4Y6W-+!V#N?a>b4^ ztMQpV6l=w%wxdA1CNU?#J8b@J=X(4qTw>e!5i+vdHYqBHY&(FBoJ1#>s`?h+`vvM` zelhP7a+6EJkCZv{jkbg4WeZ|`X6@E&OtQis{WTW-5_S6~T15e4n%@UmLiY!VVF}BB zwzF5sIhYwF|}!p(0b|V23o{y%F)Kj8mSltp&o2TnDT8 zpm7+=*-3g6CYE%Uo~y(zDJ?vQ4Y7g7vR4w9N1lzs3$h%-^ef5HRm;9mD$&{pI&2Ra z6yDx0h&Y9!mgz$g!GA9SF3RYStv8Nl^S|z}1^|jz)gpvNc2L$K=`<)`1rNaNQIcR? zJB(NXT;bg?dkqsK8*bM|=VDfbBN6r@Xv!mN7{T#0%WV?1)dhX5TXb?ygYYL%Yf!aeg(3=Ff zlqvQ`(iN%V1spRW+`hUEWg?TVQDq6sZZLFoNLgnLV0Q|7%bf{0jI<_~LY+^+Td@cb z(UAXrsKe`ef|)PyDh~bL?dLfRMgjoCwJynC(10ILj`WGOL0-tNUFi@GGY!>U_m-pi zH&(F$MZTz^=Y4)=9NHIOAWAkxjGvBO5F>M$h zTRWSc?5P=lXbtdD)MrSe2W_}R*_%^Gq4vf+Qua62ctjt@=+x;qCL~4Y_PC>%Hafjk zj`6yE6Y}m6-D5x7-sWl6gKi)-`_S0;z>dnAfrv6!FSy=8J8#GXdjb3zSQh>pFo{v? zqFYF-+@j^9+H!+Hg2kf_a&uw(cO7Ux(e9!@zuZb9l#MdOanHO`d9wBFiB<)=^dq1 zM^DZ_H_jeEOo)P)mQS&_WrCQikoOH(z1a{yvd*jVdz6_bgExXYT`O%Vix@0-A>e(~ z8OienTpy_;ggESOkZ!fSVKZ~FAIctz;itl;`6q0|l0@!X`^J%GjNYbg)MJYRsWIsv zBPCg(GZ!8=fj11>aWJ)5fKyE1D^K)eIVrgU=|FZegfAn@|SS-A9lc5$+a=3<8K zXHQ#7qV|!deO?c1s6zD?n|>R)U+?+XC|?zWb-6}ch9cgo?i=v^{ml$q`&#e{YYV@? z@V0MuUAhV}7GRtIiau*$1Y(YLYs+fWy-d z)gd}CS(#1~0s@vj95T-9t|0QE864H4FajAVB@~NYl<}MAFm&9ES2*H$hG*;?=cHoD zJ=UTd|6Rl}tnU*+l5>??`-xTa{dR$@h49#g%&LmL~5Rh&GNg^O186}BCB}o>6MzUnd zIwmqSxyee7Ejfb$q{$LnvO<#uBxks*b>{4|&)(Nq*{>6KUE1dGVfEB3w^jP$e(AOu&kqzxS;wzu<1vq+NEcCK z+@2to4360iH%JcY4zYH-hxMg<85h&S4Vc=JK(4~TjTJ{47+{HP1+gh^1O)JpNf#TG}vhPB)=nEPp> zFLledrqwH1X|}gy%IBNSP=!WQ&n`Gd_j}s94Ph`|({%-f{0~iQ^joH{eAV2GXMB!T z=SQrNge9Kld?zX6#j?FJ3CWwU-*H1J2~c%?tAP;?yHl*hI6nXn)*bH>cG@ij@0q{z z?x%h->plj3LoLR8iggY;^u`TA))B1T;lt+E#*DwWWm4aXuWR%XWC%eCGx_SRAYn24 zX7?8$d&i#Vni%&w3b~?b#+d@TE!Cv9tti|;%6`U@)^=u(Ky`2of#@7(F;9B_cDN4_ z9qvm=d)>p`blSo(51|kqp)A+AdE09yqn}u}bNWo`)s?vxdHIiLlB|R@b<;1Hp^0>tT!72Y0%NHh19eQQ-mqQ{N4$-DCif)kRL; zmF%~p@j{YSOnJHf6{#6- zSL*k7dSzDq=?+5#5)<`wp(c6Q<$Bj;e~OA;)z%CHp4o4H+J~gYp<#K$V2p?!zRuWs z_>bOkrI_n%^A0fwrF&f)f193h=?p=l{5(N$vnzw711LorY zBBg)5SXKOx&)`yhIP#DGh1^algWP@|==AvCEQkjz$hbe7+;0!_f0o<1IFQ@7gDIQ; z^Pm0payB8$(@d=vbmD&_w;Le04~K6r|K-j6TY>$4>a!&ZNkSxd+C_^0jocoC-2ONo zG9mS^GX2L!`*R!o|79Z+lnKE2k$yg9sL};nH6b&1Buyz+v++#VsZ$1$m#!SlYfjq8 ztU7tJsA|G(eSYZ0ktTQC3jJoL_Dv-#N&& zMaXl@5d(hKU=cK5DG0mF$2Fj+vXsCPj3;^sHib*-dEl{0-ao`u+prOP)rvEtTId8k zWv0WWB_2l{%`hD-vumZ$5BgJ6A9(ia>=YzHtN@0HBsZ*SM&TO-6k*`fSJXtP3jei0 zO7agNdJO|tqns_cI_G#UVjViMuFih9)B9^D9l3mMGR~n1YSd%ZN6XKb#|t~1-EEjq zzT}7K@V@3`CV!Irr_6)i9BsS1dQmk#_ehTjvxIn1BkV;c)azVy{RV|5qLY1Ihg1r% z;{65Aq?Mm0BeA%5sL@G&Bwt?T0BPm-BcYb0=z+&8PS|&;xN*K7vU!4AN9Ti?*KDbh z?;|9eXG3eJ^~^&V5A~e9jMwY(8HD&-LpJCl8&v+npe|S@%ybXJkilWf@Fi&5Yz8lzn2a>pvfo0fBmI`1ecC+of5hvZlz zR0fe^ZYv@2&i$Es$El|t5qrUCcpkoZyezDGAcN`<+Y#byk}Dv8nHKkGhJAQ`JU!(B z^^r^8!rp^#$k{HwodbPqyaf=FZE?mF9HNXCa!AvI1dbgCv|Djcccvz@SSI)l1#uDTt63zF)GUh0#dP&|O9yKz8$ z#F40nNQB@TD~f-6jbAt|dR9WmM{i$|jp9%q{CVCHNP~%hM}s4CTh%A7vY+ce}9e4oj}|wBbRiZcUt&~gIaJOtv?8_pVEW3P0xKVom`TT zCTT={|SkWn*D zG%|GdunoT={@#Yaw(j5C@SnBR-`ntSZTPn~{QuA>-}|D^$KuVbCcdNUwz|OW;5dKp zOL9^lSf6Is1J6E#+aM6>g0z$(Y^8%iZojVXJL1VFRr6$Bf4aObw9_=f^%nngH!-ls zZclGhkiWYG*6OLoyA-e@2jLiE+Enm0M#{9>*qhCvRnX8V z==Ccv6)>lRd`6?;DuhiJ#wh9rai-VSis9q)(JU51V7@;PO+L|lcgaTHnCns@ zH2SwikQsUyh`aCdZ`w=BJ)#&9 z^?A2pg^@8V@R$t;$5o!ostOssufX2F(#-D!!`49b0nCvKajSERu3KbLxH4}E7RFlM zJ2e3n)*HWQiv{kYc>C@V21G!u^QF$)0|AcK;WI zH9Od4_Q+d5E_zPR*)Oz{hS{36gtR%(xr#i}m^PAC_t@#YviY!OV3B7kuxt)nbJuNE zH;ueO8d*z>U3)Rtj3*pjT?f0C;>r)M52xGSG1=>yOQGMtyxh|I{G5yr3!?PI-K^X| zpTYhzr|{G00g8B!+V#FEJYmh#ISk7Z5IjEjM>7|9_3=EDt-u8wE4BvVmaNQpU~r-! zJ7u3Gc5?f{BYh5VM>KEP?>Fa6iMTF5k0Na52RF}-B}UnrwFV4SO>W6GHYt#{dTyI1 zoq#1VJNTCqDUel%W%Xp+zwwxnW}Pr*7a>Mvcxx+1KswEvr{l~OSA!Z1uS|*>tM+~w zrRiL<*Kh{vZt!t%1j)$h^y)TzjYpp$lZV~!rL|teC-n7d1G}IN&TG+fFUl5P$)h2i zJXeZ9g;2Dn%}jpGC_MgRgJtlYT6_41ZTc@3!sFAF5>b`G!Vm&X9h~lW@Y-CYU&q~Q z#(qh==u8NM4Vh?X-j{_7@bcas(;{s01UF`nB}QOIJSl63EQg07+*cXH`wT|bE$z~x zV;45cib&D3x46Z)5rXjw!|Pr@f7~v&ZjT?OvpMWCC?g;-`%$y7#AACEk1DFKU+S)( ztltIM)+lNYx|UWkQq661!*cT|za*=@NWD&PcwO*Ynui#hO%b$fAURMiJ38d$^*+fZ zXVJQG_xnoDp)D11+}xyXoMc@U->jdv-RU43dxqHhHnQeje$#DDKdZC*Q7@+<1GGYXL(#;^1&I)XrH%0qK`e8Y7%yp*WA=jX*o!0ZNCQB>a z6y+x~4(&)2s4m@hwM8HV7bq+S#63snRHQV+F>2{ez20T_G&eeetSK>K!datvpGZRm zI7TbBNIs9i@It&puSq)}yaEU0ISdYt=Q$!4odSt z&-o^arcdedSC+X-$}X;S`%m2b&}QEFHHr^bKJMt{R3W^m+cV6ug_j%gLX**Ik+X$a zIUTlA+D3~#go16A2(p?B{y3%Yk$8Qd!7AC$)%reqF0Trsp5o%;Z}WKJUvDRreY4WG z3gkCUpS2Cd*DG0WZlovj^40ucABlDkmvAMRYd?5j@K28nu31wl-%qV=%z31_Gaq`I z)Mr5?ucBLc=}Y#f@sZ>OSwh<5=Gm*#HU>zwBxff3)8}&b1^8(K2W-PyP10v==6)V< ztO%3H6;MUBZYh&`jJi&jn~aclEu+LiO)gWsUene1F|s;7=Obvu%5Y<7&+- z$Am(KI6XPFOZEfEs1Xq5&gaj?ro?#3dKqayz}vW;pknB)k=JMCIW6qXEkxMFk1-1g zTP!=$SIWl!ues|m5 z`z0>-%azZkVWWg7y3>M} z>^ZLcEgnMa2HY60i0~QId!4Q{<~S3Y-OGqDtXa1a&24<92BBuQ+~7e#*`|JbHBF6z{V)QR^wsTJv}#US;3EO>;hgoJI43O zO)y{2N88mmqtqnAHYc^O@N*H(SDp*P-hrMW{WFt&{LfSrEce64)iy2LO6_@}8CC5k zsW%JScY`&oXD4lIn6$`rbtmF1ZG@f>6gsQdKWZ$8T;c=;aDPwRBJsNQ+m)%_W=^T0 zOeK418@kKoBhPTHo<2(>lF~MN#@z)kDM!2%_6j;wr+Y0A3a!)&@+!FqPJGyIz0O+M zhbTiZA&=}CUA*?Ctp76*1bgnrL0jI(f)tk8r>NfOTI z)1PAM;7T0G{~&Y8p6*OlKo66>|M?!feJ}OBl7Y2?sc$`|SPC+f#ZJol=cb6+JgVlA zPlm3nY7A#zLhwM+Jd)}xop-wWu`3j*zv3%V5buA4_fVYG9?h{eYm!mmMx2$g?;c>h z9c44btKqU4!_iD2W5IcmiL;%~y>^XwBZ!?z8iwUsee@X5(diyTl;&5_Y14tHU`@;( ztX^GAqb~_RiQGqu7_19U*EUXIvMRSTqzJ_WHfSYphlHGx{hOGjkSiBKz&F;!p}@SA z^z|z*d3oTtep5fF>Qk1y)xOG^Rl9~Zxb6)Mv~t-V_mBwQyKAcya&@TEtV!JT|$bb8Nx=sg-l@QB3T1<#<%MZJHu_mvJwcvutn)vy8rW1|)G|Ufcpw2T5Ny zh>|cH&Y*?8ArC1JzFae3M$z?v!LlJLyM2}IXWFGAYXkbjoA_!wZ?)}st4&A@wYqp@ z6_cWM16B8i

kpEHygt9WqParF@fA+v0g3-z3uL0&LZ5gPz0%xtYIc3R|_k>MlRH zdttEFW;*R>M=si>?~J8yH-^sUumUc#)Bhu=hk;=?VD^e$KaXRDum#_v_#k))74sh2tbX*&CZ1zKD2X(=6q5jo#@9YC2x^ zwX)?JMc3QsAR8rk(VdYWY453OWkmxwS-NU!?KAlBL zD=q@p{&ev}0wgwrDT&^@Tci&zmvv|-JP_Ba!f9cC5>ZNJVYCc8e+g?g2cmTRElO!= z8?ot$bSr5auZMRP6v)9oosK<@ri=7uvydZDaRh+C>a%-$5%!kM)Z&Z?L60*Q<_C#4 ztH(Wu)FJC{c~GErRVp^w=+Lo=_q#jKO51RCf1J2Ucr|X=-VYC&CML6BL;DQ6b+)SX zfx)eN8yR`Y-trBV0F-e#_wRKF;kYy}P?XpEY31h15PI)i^$?e~*|^Z%35^+C-G1of zkoZ*ot+Nn_(geq4HRw2z$j?Y+_8HX4c-i_K)Ld;sF$lMdn2LklMp3?+5}Q~6y+6J( zTwFH^y|=im7>M>FHG>}Z^tIuA|K#yhx;(T~khpTM$b4N(;Qq3kB`(Uiq0TaSn%4ZU zBPMJ?k0?b7Z+eoUV@JN0m7IYz<2!3FX&YS@*Wu+Y`qD-pBxxvfua2CT$Mz5=)@{U2 zuCZCrfMb@yp(@Ay$I1WIeozzNyS-Pm^}fu{xWSVpej<)f%?Uf=l^t+nDZPLoJd1G; zb}?;|`cnsPXMoCk{?NGjSG36n(Iyo~j*5X}o~Y}0p@JrZ?MA9egShg=THQBT4?fp| zi$6E`IEcnzT2*#~tx@)}jFqMB7!)ou5>CS5RG>^-tAV*9WRBZ`nN$U@z zB_wEeTj^Z*mheBnMSg5Ji~~=@2~lw3&d8J<99N^+R>CLh#4vhnwak`nh0KJ<)|IeQjw7s6ZR60}=C375m)jjaZHMKF;YC+9BJr{nAdkup?xP0=2c8{Ms7&_;`t!OXSfgedRT8!gSWOJb40WV^AH|pgJ_|jh31Iiyo?!)}K_wiwtZJ z%bB2WW^3l!Omt^1ZJ*!vz4PspY&F{u^^870*~yQPTzNIVpd$j6g9l^atN*iwSvV_q z(JrLQtHHWpF|!uUBq1iM<|JrUS?;m1Q&=#UxQA%At&~a={cG0*$551uUlAxdBW+{s z8dV?wiuI%L?11gKU$G{djnVUpM8%ShwsW?1Hp6=#q&GY9I1=5Mf@#~f?3jcj-{B~H zYDns(0!_wDEdbh)bKk8ODk>_}e|i$LUk;nBNZ3U6crU5-5ds{tHP@E{@lBo$^vvK_ zsjj*IZ2X6G^mL#c^@f@S;^BYtYpNA4It}$KVJU-nJU^UQGW_YA)x-Q;+3{WDhxV#C zSN)-^qhU!l;2Jqe4TL}+BFTpz#b>4SwBlN|FCa>v*x~tE9p`+#J_$s`=hfzZIvl+S zRB2=1#>WS2YtV_JTSc7jS_cUy4 zTwVR;63h0@@epcRg-PJId)Gu)!h>Wtc%oTyW8ti}X}hP0Vsq>!B(+-QE?QV0C9e1P zC0Or#EHpgo?Z5bSE-UM7Yxi`nG8Tj??mv7>w0SA*7E;1FOZkP$QRpZG#9^1O+9i9R zBIwh`|Kqhv;I%l)A8!(Uz^F8ow9VozA>$=2yVn*IMdk%V=kd4#U7Ij3^1CJpNg|rE z`rb|~)3F98+@;{NUM0<^D@*4w1_fR{iXM108kkK6O9Ul7NOXliSq{=|Bu1;2k@TqE zt0IL%>~duLm4i+j&lY(ZUW=~BIL`MM4;59R8R0PeSKIZURvdZ9V|1sGkg3oXMKlM_ z74w?YyzQcyJzfG^PUWuYx#P3kwYbZkYNw6NKEdXp3mUY1>7$3eL3a;UiYdBTAM7+N zPg}Ti*2L7Hc1hI+FHPCl)m+lcG5iYif@-7H9?PqX8UFe<3|IZh^eZ)Jy&8#G5HDl3 zTvMMxix*3q;>0}6Vh1}>qvnEnD4Dn$KJs~Y`!3{G^hcd+kN*j_D*m=-dneS8_?@3w6q=#W@ya*{Otu&<{_LZb&QpmQi6Yk9iYS^^&5_?|b+5`t#9M-4$vSxy1t1o$vI7~hv%=5DmlOf@CLDVma!&QhA-OjRRJ>Mp$%*^T=a+t(v))-5a_MwvtjW)a_s;ZMT2$a} zI^Hfu>5@{x-L30fw)MHJOJfjeT@#74_os7!z}gtHAgIBDB1kH|4AqC1q{oV;GVCX^ zj<+3N58Cz_a^t$>PeQ})e?~{6=7NZ<_>&29%*tNb2Gh-ms_Y-dh6ypsx>>MEG!KR^ zmr|4qk*VHUCe|YF6hNL*B7|rG=rl$NWD+4n7hK$RJoQOrBvFt1$L@^oj`7FNl zK(xo#ISZwnXkcD*O3(A5O;}fZWP)c@){jXU9*9)@DbkBH5%ygf1aHtU`#f32FTvqy zEC$DKKs**v#W!7@zx0KT6?eQ{f}@Mo%fQ?|YJSK#v$CzDq46FaY4$(vv!>&Vvkq4D zT1Ubv(BUJQpoCr08jQ1UPPy&AVlakG-@Wr$d@V$gcHJO#T3G7at!A?D2@o9n{>E8) z=E>?f_|^$g$?I5RK|J-e`S?@v^7H404KCSNzd2(N67;?YeNJrqD%WY!`!hrBy{h}cK<{Pojg*JX^WC&X4DUbv8qG@ zq|m(zr&*TMTt%uDw8Yu)K<1wS)6oo*%an#;Zq^0RalOj?lh}m`13)RZBQ*suq2F7| zd&l<5Gs3gtws;YRL~&1A9T#nmSRUh787(i(wUI4$=h5%$d?`&~jK8zcOl2A{5-7-? z2}mWW9;w-y;lNxs4{d=*6u?w1b=KVV65~T%=_*k$-6f=41kh>CFp|_^cWgNK4)m$g zNO*}VW``mU9-)BS0 zk{)G2;NP4m_-ow#z%zA-CDb`z4r6=|>3`tP{6`}? z3i(&>EZ=~^x}gKBJ+yKf(TF%(iQcg;ACGj9xOIn|zIPXJ?LQ;q04l@P2U@g+E?u+e z&Um6;4|pD_T37?2M-)m=CCg7%0b>3Vpy)JqDsXcF3>?S$2Z6p0W<;qaKCb{KegykE z@RN_71gPWholRT)u8>Kd@$x-!z)(238SwmQ_OnyZEW!QzPAW!UyZLKOdZ@A#pd{)aOduL!)-yk)n}|-G-x7_a!y& z8y&@u;-INtohga0oH%i`sQ3wn?{y=(*K8-12{RvpE(+b-TAgJj!gcLZpp!|S5stA6 zVh56pjvx7h;z8ixgQ?6i`IGNt!JBQTg;vI9_~&;|>O0xLIy5bRx3l&V8$#iq->r^M z`Ty!!jj;*Tlg=AD8OqF$5IoJ?DVdZHixD`kWDQYJF?wGZO z!SV<|5+S5&7bPEsT=VG*%Rx0#S7}FGpYe>XLM>U};V)LU@XL!nzJD23gEQ}aI z*CC1bI`Dml=-QOaBzUfAcYK=?Adbh;TIMhTxLCx_4>O78_-+y=`-SRt10jHFtExAH zfyxL#i737|47lU0Kmn?-vkeFqtV9f{@#-9IfV3b+D|T@nD^ndWca2LrCJBH)G!8Hd z&3;14>wt!E8^8;4m%_#DPM`;tDl%U#uxIBWKD{Qhjc`izTtgEPpL5q*_x!zNNr8WD zuM`kKKjY^CIzi06;N$KdfG#yJc&%4%Y<~lIj?gQ9(!rJh=M({Np`{EcGZ5##E*%@~ zcaZYMIgLe44!Tm82kzzkq{t>mMAak*<4ie_bwWZw10e)6^Z-*~Kym(lV+oO%j`dXV?_ z>@5PIp+ym6PxrBRA5U@Mh$QV!0PnN;szI;eidX8eL=E8Zm^>#D z0H74DX-PaW=R7eu z?|?1Jh~W-u^4j%WA~cp;RBj#!z=#6~2#F=$x^zg>DD;}crYiw0S7K2~As4aKdfd?@ z;fuwpSEON!T3`fk3vPx}+dhJ&PRLyN9eVnPMww~gG~_fK>aUUomP6uVY_?NW^~<}7 zqAHt$&v~X5L<%(}UBLk>zS-qB0H1<(y_BY;Ez@t;r6eemmVAM+a_{=6KtI4^AiIom ze_r#}lLIuHw5U(8Um7xYPt6%Vl6ao0cSO-lC@6PLY^j8QtMDU!F;?P&aT)-iWn&T9 zky0c1>+c^%^>A0DyW#gb7=NoB&QV>0B$|_eCsp0!LbFGf!kO`CP;F zdbF!#fzj*RBOQ!0aKM3)>z+x9iianr?-PRWlwgfZK4A3AOnh&_1>gcly0&R$uWpEZ zIUA93W;KRSkX6A^{i5I5YY(pn<$l&OoHXxR&VxkI;%er)fhJJAJ%SlnpI2vS&SEf$ zqq`GE-*KXsBGOTETkBa%-QIx}wWwnk>O^&h3m?1YCI%MT#*4TF&0Zt3!tc1GxIW^# z`bo6thWp~njp~lc6dI=sm}eq9e7k$R_E!Ny$~qZFdimx>+8|rB=5GOzkv}vej_;&W;D^8F8zZYD zEjFVDDfIyY=g_?-tsY=Ox#`h51fUsC<`)ox$X)=d;o1B0MZ%H62YTbOSej=U^;zZX)XyX1@Ce^N3q!Z?(F(~1*X)* zbkud1NDJLmwgQ~>icsd^WPu^at{W&+k<&`z{!aRNgi- zW0~FAps~<_T+{QATI~iA9rm`!@|?9>vcB8{a~-~;((aA4HMKO-dCm{56T`U*a3oI3 zF>Li?%OZ2~^{*1Y9>u=HNYDRx!`t;nZc94bOEYGA809U3ueCg`nyB8@evw|gV>m4y z2PRWUT$b@;j3rjAvk$7+g@x=oO`?8-s)^@o9^%g+O8d5gU=KtTG`6bExwsVR1H*+~ z|NKy;du;35p0J7)y_ehqL))Z@c~9%Ql5w{$Yuc$NV& zRONxqM zt1Izyi&I5lU%KIxk2hMzCc}YOsJ7MZ4*i?)>N7`Tvzeg4`NJgtkVT`2N1`QGN)$6JO+bBA$4R@MD}N)-;c@g`3!+qe z8^qbj-+LdqmsdOW);}vDM%cX>mYgwN0)Vcw(`-KvA1o~ z-LJvl@3WFQPZXktaMm^av`)6zfs7rE6&8gPJ8(V=Ps(?tSfkL{H+yRe#jCRMHq~hr zxebloQDGEAv)x>A%H07=0Nfe$3*oRu2olU*VxPbl3FwaS)Cpyd zEH_rj11i33nwr{6tXIR#qcbHXM`HAeZic&Z${dkkV03m05OEdUMVv+eQIBqt2!DQW z#I`pc=KW>Gf&DxEP|w19z>gEA+#n*XD%S#?0P2q6g7g;6&=f!D#4v1T#Olo`6WR{I zRnQHT{ghVMU2#>3-dx?% zCu~oNEC<*0=sHxTZID-7R9|0>%Ztxs+?|r6PqEnT1^n*i8^LXTiLY}(GBmRl^;?9< zJ$K-u=QZNa!jZTtTTI@@zLpIL{}P)+1PhIwZedl9yMR4Pq|1`2-Ri+UJh=g$poyPr z!dZTBmF^RfWfwOcfh=g8?fP7a41VCtvNu#oa)et|?&0)ZhR#oRu%_t?EE{NxxFa3pDhC@9JLK?9|*$j-WQh~FLNRn)uNLlc<&l2J0s`Z9MrJYiMv0{ zFv~3o?*i})2$!`hQIqSUetP-}2G~zI>jD66=HEp7#%_$w8tiqX)B(9zKk#0D=&&MSgn#i!p2M zWYMR=ib>nVUW-D>4NK>huxdXC$iPgD&}{#Ut?D%Z|21nCYPP~39WljY4$8}`3={gG-F*0lDn;cWLL;P9%oI`Io55rTe> zxq5;6P^1&lbFb>iH^&77aFv3}}XosP{J0CeYgH>Vp|eipG_V zEV0sn&g>;xoO{YNKFTm%9KWtUn*EdO<11X?TLN2NmFUX7xhjCPu^e4)LA^S*E2Y=2 zv};!QYW;`x_x{*;pWOu?J$IB$BU%tmw*;O@ONgYmkf|$EbzCxaU{89FmUQRfH+1I;k7gj$M0wl4z zjG;6{yAJmlo4W4rObFIy(!O3hdzn{$+t(Watc~Tdp7$f_wzi1~;Zcb>3c0KN4<9SI z&lUFt;{p-YWGOrj^m<00hM28~vDGFz-ukj_P8?&!TDIr``8TnY!h*5rAYOaI?P)wV zkx2CWVPN$+Y;L{;vA2PX(KU`1$;3oKHDd0qN|D~udzpzgyoOG@NCyK8IU+l>pWMZIMhvX@Qng|*j^ z<3y~nMH?=u=7OQ{Y}d{a0o^z+pg5x9JKG^_i*vZwZYfrq=squ`4e;D`if@=vGop7R zfk7}hx})%+F#MO*+raOBjE^fJ2Lg(Ry_X{(0@ZGb!VW(EMDp-O?+G8Y!FSkI@{?Qbax; zhoFlZ=~zGbpt#ABP~+q&e{^E{4v*T?%eAOIS0DoMFW=-q9f^u=Q<(4ooE~yMg{M>N zv98yc9BLv@mapk_X3VV;F;oFgGY@K)>Me{O#piG3>7CsGej`GH`#eQ;?}`KS3Ki@( ze7@y=k7)E2!176ak-%o+IuyCH2hxiO%*(}0GmY(y z6LD~D`h$kAk*zlZtY*DpGb^g1y3A`X@j_!ad+geWk`XU=9((nc!#YH4c+G{$Me4bR zYN*}XF%IUm<3m_FU3M58D7J&NRmgXBZC;+&3<01zUmAIC)LL?mwr(oy6*F0(CP+Ha z?TcBY<=8k??cB3Ev?XQ-JjD&ijnb@am+3pElDoFqQMof4z}yncGmGjCj{?h85ji%+fQ__XT-Lf#IbkN@&e)^A$wwpnRBruXhxcv5|AYi*8hgfG`v zBiBzkFBTkh*_KPCnhQNk%|M*ThH`F-*?IPPc7JbS-y}Iuz;B-+KR3~G!MhIC8pAws zuy12f5E*ZpTim-f63bu34l+7MQ8p_n7^&#n$6TPGF|Ey*Cwh>Rv1q zSh8`uR>&oO!&L;*310Q;|CAKGB47JOw=e*wbR(8*D!O*_tsje0O!yC*<*o)FJtP%J ze(#fk9T9L>2a4F5gF+@#rgQ2Vo`5lzmKW&^>cef~2Jg4q-uw*2{ z654I0M(Sh@@h-2(76#*b*vN>{!wpoi=Xo?{RO%z1im&hmFK$qspja;OT6FeK z%i(FBZRpRClht(c>2B~DG>y_P=`@?3GHCd*Fx_cVY~?D!TI|_aU@6<(WA76o+L$3} zS_+&dG7~q0i5lmNxoH@)NFFq7^o_I)EkJ7E?d2MnY>-x3{NBIKC{XQ^jrVo^Ipr-v zZ@`?-)>RB^K-n4+TYiu+_*+LhZgo14Gq!&n*qu0)g#jZCP;)IKt(+9P6p{Al-)=q( zOq{~A7PPcp;yE3gUa-o8(2_ei9ad3nHPL*Df!R7%uv;io<^D`q@79qZ8sXGVNr@;rhYLzH*%!#=T3Pd@s?WK-<<{x?WNluv z4clAGuM48v+(g~(%TbsOMFP9hlOusa7jjI`iQ^44JL%KA7VKmR&Mg@k3lrATHW}@w z5*IDgD9W31(Vh*ct-X6CD7SZI+!?#}yKREai8yEp8M{kap_15L3(oZ@jWei3Mjo<3 zA8iNu!rmh0^~$f^$%QlN;uT9S+}c zu(KFx#nA2Z8Q#Dw_iz{$2I4e=02TX*Nt&&{@nc{QN`Kpvz8`kuys5s1ScvLmN>J7{ z`k3fGNPp2|VyiOsTDzOk6k2w|N?4V4d;v4kJ}@NE(L^=U8g@nCY)^XIQq9YY=78h`;+e zHE3%Du8YgX`;XEb1Bcyaxw!9#-hC#q-Pg_=k-lD2WVkZMxPL=!E5!F!Fpl4TFl+zo zuvJFIu>xMamwRL@XV2Mb#iMQ#+RNoRZ{kwL6U~?G=UTlK-4-;pXRH(^KE(HpB<*miM4o=#R4M}GU{KCz<|UXOIdXGx#oVC zGn>2orcL3uZEaY{1~99i-vYS{H?NAF9FhEJ{FuM^ktg7U-bifmRpjMcFkzofdQR%M z%}nG-NLyi|e0i>Xh2GNN180d%ri#TJd<*nrtm5mfJ~ct1C@dfU9V2GQo-5 zmemihEfn6Feh9@~u1<_FRfzmL091(um&sySfb($nh0c(Rl`XkG!hY5E_1e6M0nqQF zF|9lh$Huci4=LKu^5og|U~h_yvb1$JfCR3!$6UFx>2Sf|FPAxzg7haXT32EUmA|wh zb!nE8pG)_xAn%I5ZmOz#Y@U8dz!`qEKZ@jE;jEw6%?717BwN0UHV@8i&XXaqEIMoZ zcUQdON1b4^FJ7LtRFEaitVWL~4ih?@mh|wdYdv1E_DgPjHfzS;AQZr@*COq^%EOHy zLi0Wst3ik?Ubrx`ZP}S3zeszhU*ekDh4f*ouGncXt47_)a=~Scc1>*QH{!pP@X9ks zmT;5Jp4Jrk83HQKxOgh1mT{(E!M1gOjb*MMJX2)#F;6Sgsx|U*@wNbu$Y;<@7KJ)r zey5X@gkqf;TG?J+vDTE8&Kp^7RiwU1jNQ<~mg>)JBF7(Rz1L)cgrkyO)#``3D%&?Q zHpQ#Sg-Z(3Hj(Z@^0bAH(knnThss3=PA^@t2iC^b+o?~m2Gt19+jzX@XG<8)n*EWH zAEzU`5y&VZzNd=k4i^<&?Z`pe^Ak|YP#@kwL_A&Oy1HJV`6ZX>3O&s^OnrIM6giKo^&RFTUgg8 zZ6A(x^pBD~ZCOmXSkS-N8LGD;A* z#fth^ihNvnf0!%)YgU~HN`=$jNhseWYYWKFOO&qoU9tCntUM!@W+E?EXk$hw4dg#x&*9_6)V$Sa*7A| z#xu=80(+9zhpEUgdm&MI_pzU~g;Ip9al=8p+rAjw6PmN$-9XO|&;=pU3 zef65H_=Yv9T1%ihk4l*@Mn@a!iBW9W&3n|#jtS8ejXskBWfv?0VC_<=1MgCZhz zp?F?&u}>Xt+Mv|%ZV((#Z z5*6QDl@|3oAINfUC!U}CumyPZL^hs>y>;P46xeDLk`cpKHO7S=-eOx*$%`ZQTkBsk z?k1&QIoDk9X`*EMz8o9sl?=DV2r@6iWk3!p%DzbDn#-(GodDXKOc65(+P{#2~+%EN^K2!y3_=7G8>(YWeH6oOj2>P z{CsXr!2PWrCh<78st+qxjejdHzxrlag^w(^43SIV{wt_K>DUp28uo^nxtCfI^Kaf& z<)UL1(RPr3G8GqIKFT0*-b_4)cQVuLBZ>i?t=CFY_=M}ILx(hQyP6#>21DK$#;a2@ zC*kRr5t(~?GG)^kG4ar(V)YF%N-&7!x5wy|8@H@pL|ZaMPLo0}D{ph0e&P|unN^{q^2Ns<&lNM*`3#mQHWjdfo9Mp+r3|4jE#QL4i){zB zF729{+}+uAAr`S*Vp8W-Pzjk%)q}i2rxYb4tcfYI%LTdKz1a&HmV7hU!)424SEHHJ zetB=YiLs_zTyW2Z02)efvl-sQ0k7`By%v?u0Bc$V?_bK=P77PS02DGr?;Uz&-n1Y0ZNn?FlO!$pM} zee^i2=vl6IWt=LUPnarJZ7uU!M2i24VJS^-D`K-8&xlaFf0pmd-81L%3@;w`M#_BR?NsE~GJG}l_x(U5p5 zNe)Op6lScM0OES|b>Aw|k~=VAbg%52n)N;E83R^pPf7`OJKQdvEO$F2qhe$#XQjvwe^IDObM5jo!4*B8M7AkD`vGFveh0wucb#J z;rbwLvnft(@{FL_V@F8cquSKkUqFB(Xk?i-=ri^(?Xw>MRu?|0Rt^l?>q1k@^xcXN zL(vA|6vwHiX}@R-I()h`VVTCz7?T1rOt`u2&--uo`VH^bXS*&qUEfrMg2HZe>X zp}!r3Qk9$d#7{;pJ;CI2BT-uuNlb|~-(bH6wMLMy-)dn8QNE5tM%65<2_|ACzn$=G z2fay)=7H`e)OE`g!}r1@Xqv^l5fNU#(dnb3l)J&U*LAV*V;rigR?x|b90qE)TMoUA^-Ht zOOfyDjy8=Rqqv)?aYdo~K;R2n?IX<6-QGT@{2nGpFq3DTx^-0Bgf9a}b>hJr^Nu8R zePq@SQ>UbV@5>vt%WduD)ukil>t2%yE~qxl!e(gf!Z5~R7~?wZP-8nU$~So?-Dp_^ zAsCcg$x$55Z>iZwW&`-2I4aA>pU0%deyZ6SCiV+lRL8RG>37esKQv$t8oF4n;bSraIyMc} zO>ddJKr(`X^b)BO&q`mrm-7bu@<`>^b@K9w*2|cuY<_yxvZ?!mYXe%qP1aOUWiWQ< z`%BzV%*#}TsK|0Vq{DgP>U(*J(rJO|gf)(GeQ;8Aje?eS`Arkt*-O7zhETdX-DT0Z z2(!k%Rl~9K?&88*nE8p%T-$)Gt<#Y04_Gd`atbE{GMCwAPDaE^dJe|83^{3OYntW> zY@IXk8O+sIeUov|V0kmp)+5;iKjPqq&j&c zl|DJc%9Ah%=6LU1Dg%>(!6Rl&25z=)~&Y-O5uYzH=jYY zXvK<_1wyN@WKr7)qb5DC`Ad)$N(R8qHol(fr(NAqoEX){nB!1By`DqIfK6lWSUR?z=3;t>OBb*EkaEqnqI>cL{q?-oLs% z3k>(_{||d_8I{%7?u{xS-5t`>-BN;ppoD@5NGpw$bV-SHw@4`^lF}VY-O?c4DM%}& zAn;zd|GoD(XTQ&T&U-$b59f^MgJY<0-)pY9=9=@0U)WAnm>axC`oriNDcP$0GoNZ{ zqGaGqZj+jYw2ab^n{i2#0^xdPi_^KWz511KGOiv83;=A5qoG}l-${MM2r1t`*Z#e+ zp;mn=lbZI%{O2hI1%d#9fg(-3w!YMZe=oaRXr9uLyC=GG-}E0dd{+pNg?cCeMO-sdGY zdoTjeUqo5sJUvUmR8hcsBzGCiC8_crX=H9Zdhq*@*+&U)ReR{9EBouEgo6ELu@hHn zHo|VD^0S}=Otf(0^W908v_~9gFc~dU6r*Wu-QM4?HG-FC6?fSliBDm6gy0905Jbt{ z)uVZr1x<-*Kk!eltB|2}LNNN=d-u4|aN9d{c2Yae{+#qq=}4h{%P+`2WLTS$xO9@w z_22*b-7Rw=U@?^sOns&Ji?K3at?1nXF;|_e7>R8s`!d#;Gp2$^9X0dn=I@UUhN|xdm9{h zcg>&v)*TksSJ=I~40}VBZNh1)7g%0;wPJ92K}66!bh3st?%ljp;#9#0pSp@|@SZX# zK+D`HYIxi;)I-8@vx06tn%PtJuq^*<9?}NGd~Fj zp4LJ>a2>FoPkU3^uvz#s-=ul|Cx8VSY5g-xn4G-drF(6DShFnurtIfLTAys^G}qwT zi3}$<@Zxi%9gBAKb-za)yia#h{sG(8VzQm=`E9&;Z(jntMMI|r_+9koNP5X6||G9UN^ntX|8>12a=+63P2lA{SI zN0j`6R)TTCFZ<2T&YO}W3up5u!KYCJH18G&kP+LDxR@7a8(24_zav3-igWBf4u3q? zm4d_q$U9&=)TadibGeh=R4v}m(;7dCq4M}K)U&%u@^a$bX*jM#q{K9)!jH)kHF z6Imv9;WGssALA-3jo*IxM(!WJagEZ19|bQa!e%{hYdhamKMC-{;9CjKfcAQ%ZTHhT z>ui=V0^+b%ipk4l1E}J30+q6-;*{_U2AJZ>ZEaTHZGLJLqA|XA^L%)i?NGwFuLwUi zxBCQYs`WQF3v#e&1y3zy&UE{D@7M7%xqse>FPL;$D8koMA{)KPn>gKi;aa{2-e|T0 zB%vklP2UY=l71DoNVyPuH_vb;Jyi>)r)F}$8CBA{h4C<{_4*xq*Bj+_5Bb!;|KXg@ zZ#}Lq(|X$d;kGZdA-^>3 zPY=7drz#`Higa|+?6zs+W+*qOMRPAho3{-2IE_vkCh&Zct?1PQ&W2-rt9~IXj^Ap& zd-oqpE9}1R7uZVTh%mTDbEZ422zqqBo5*+5D#1%bIRr-vFsFY$2J?&^AEpkz$e#q; zWTh9z+UqgS-}*R^y9}woeg zgFaPd-%%rSq=_7YN+W#A?Bg0UKD%Qne_PAVDb}l2o;6fi&cs)>XswweVzcd>i{ja( zf8i+48IhJ|GQ;%c&w1m*`TbdispZ}I1hC}fPCYkKvc>AC`2t62SYsP`mFDgZg;}g* zDF8n$-+yEV_~v^xd{#5=FV&!KBH`%%5ubLQ$}*jCcffUckk0+3bXC)?v3={go8tqs z@PseRyidlP#g`}U%AN)I4m5}xcFb)sgHa#eG__AzYRmd=IGI>_0{7#Ovd8e=yuGe* z(+6O0#6B~LxqZ7WKU1%NhmTA-zk#5aj|6#*AWVTTBTfMLS{m99zJo>Wz zG?5r_QXS39{uq)tn`q%?|Ipb(r_XdEDsBBXJoeegKVPQJ5ApvwLb^@!Hf}Q^ix(Yp zM-nfzN}&2`)2XtrfSEfAp=+3MIJFxyMVx;7y9shoou<)4v6O3*F3|#HKEyK}{mtW* z6Cu}~LN{V)mhvnO4RJuSD~`op8SUXFgcpf=!8l&M<-sl(=b(9oCD3v;|VJ zzSmMviJEV-pAUB;y&_hegO^9${0e@$ornTu6eFpx*54mzmPS9S1*O?|V zIfxFJ7xNeX@(O0J=SWMI<#=rtaN6HDc67cxE-o_#-8cNg4%IkWPNa&W$ri#ki?BYG zC2~%L?fJJ<%kZ?UlIwJX6b2Jp>O25Bu5@ta0qv>kCoEPb|BSYSI%H{=!3po_1s~}@ z7i4-%v2-7btUvz@*RD7Cf<@4pFjdQuE4TdW`0b?EABgULfX!LchSu#Y5K(Qadq)mJ zR}cC}C^TaBq+%5I;ydF;-aBJ~@AUS*F(g6kQ)>xR1drtJ6cx7QT@>v=q@h;gYE~x$ zYQPCyylKQk;v`51?A{c^|!j`kSy<{CD&pTj-^^MB}GAodzeEXHcvG%3hMQw)-_*Q|31#z{$%yN8lw8+7w|_l z>@`II)PvRS!3GQYskBRS?#Jr4;)GPX-%eY zcVyo?eYF-Bb(rS{VS$s~rujebC+H~njJ6FOnhE1<)Dk^7{<%XqI$8~eoQ#6{fXo+_&hOJrs)IT%qr2u=jZwZ#-0uUd%sfAsUU6udMRv*Yq{UoGR@0BypMdzY&>-#iHd5y@?(@SZx6}P(kN! zst#0JAl+t@-Jufqcf@HXIOS6yhlWD+S<-cn#efRyudJz{(m)Xa zlQ=;qb6g(dIib&|u&PJGVif+*0w>kKYjO4pASh3_OT_~o_-txsA87n_I`6uGt|mLs zdXfOTz5c&5Xr$G4IvAl=_?~<5p-5XIw*vFtWS;+>&kjnTr1zzXJpfp!PK!hD5)v*B zJRp2Pt;8U1W6)OZZ0cPou8zMo7D@I~-U3?dxNvylLd@vd9P%a(6dLL|eKhwBD za9RPebT6clOXb zr{|81@eGL<%tazZF<+5={KMiR8`#BoD>~G|{(0aJdq2)mD3tO%h=QeRsoc?o3CtM3 zxO(UH9-zk<;3w4X%<~yGJUXO50zjqnj8^q|6ru!=CxC?M?+mbx@<+N~i?hKU4U%7- z7Gg2+Mf*0G#V5fz1YBZB<-iI^vD_fXBK1p@iU5@(+!>bMVQ@;);qPn6j7{_^XLhi7-sk}IBn&sY$syo zj=*WhQH?$0b9;6A=cv_xe6;Q7TJIXLhT$ilW=-^=vylnirD;TUK$Thpmv7$nFVYZxn*oj$+1o(?X^syv zv~U9`k6O%ZvJo=X(kmx^M+tZYCs8d<$rnG|Py?Bi*Lh%y;5E8~9RTlY3dP)j#1{AO z&lD{Tkn1KQZMkp#2z!u7vF@_bbT5Hb#H|jLZ`CW`C1d@)A0Dw+)M7D+>X3-P5AzQu z6rQ+#9@n-S0zQSH9{yo|ygju>Qg`~QG*;wIE#)SN6+<&%F(SOFWlROl#!|Ws( ztAPLMRU2$Gy*Q4@+*~~I3pmiJpofdx z>u&5r)g2MfqSC|0c6G#tlD&S%7}W8fIH;c1q@r(mB@HKjd+_HYAMK#l%wpBbgq z>B>8Q7f>desQv8YOi3iPA3{18PvDvLT1$YLwO*>2gHaAf5WttzW>M3&7&PU!&#rn zv1a#G>{4s>pQs~DHVwuid6%ao-xP$O(ULzBhJt!%l)sV|rmFuHCU=w}!4EGH)ertT z_8AFPK&n?I_LLA4sYt()qzM&hpBh5sv4t6cl+S(qax_Gi5*V12e9GRS>?N)K7%djx0HbWRRwX`(h*n4MCbNuJ&>^ordcO?6i2u%;v-0j4#+WIcSh#S+K;GXRN z9>xFtp32gdnVohjgOH0<0kZ?tNV#{L=;s z%Wh(8a8I*L|MH&1lN48{Epg&(1`|UGLqYW_ish~k<}0Ky^N;MR5FqZN2HhdFN(GkV z8rsL-0Wj-3_B}AR`oC-Lf4V0pPJrtSf@>fw=Kub%?pPCmiy#Je1jCRZ%u+AblLJ%` zaGseS~7dm$Cm$JR^7alhMegdRlfxBk(e+47|{ht240RO!J z|GfbJ|F{EE-i$i~7DYrRKd{&J@oz5r^XnxZ6MH~1wI09y%Di71Ml7n=th3|^ZP)F ztIG*fwk@;`sQE0wxXJ}sy%JQ@pXU_3x);!v%%i^w5iOs)`Ed3fWDx%TqrUE5J5)%F z>vw8qdQ2bJJ%0x=8q3zo({;w7Ybd}FPpGz;=s}h^Flb%WDe%Kg{bfhcpeKFJ6Pm?l zpg-UQRs1)5TW&$#sY+3p9uOWQAh>2?QYZT_Q|w<@eHoHE^OqMVXF3(aDkx(JX5C#W zNxc71cY`n7Y@kdzyzTyCJvd@Enw~QR<3E4Yj>h^RlbbEWCH}W|0$F_mtJ-qu%^a7l z3`Zn|y=@ee{VBr}S=kWhkuxi!gEtchXsB8e#t{?HX^j2>rfGVFxkFz0*(^2Ik z8BNbllA%4e^URmlQYuWcZQ=g-`X6Mxp33@pAjS2F3r?>y)fx@hDbVvec{+{qqQaP4 zd!O)En=0SH>>zBz=orf3Mf;5PXhH)WV0c3!AWI()1YI?J>BFJtCZijS>V}KyVh)+t zfvqXMr3?Ps!Vmn#ozg4^$T8BcIK#1f*@uiXzWwJ}{UrK7gOhLnL@w}?L2~E+3=Q~| z{v&Q^YwHJ}Vmup$^F+nfWBD*aLMsSTYIWyr@sQC=4Vjt>5bc>Wz7{J#{QZj#qu zVqCd$c^UNPobErhzP2+ke%%oH-5nU}|0src&1wW6O+n)GC&oM0Y#1XNSBcPauY;3$ z2g*wnY-N&M<#p2iUDDX>#1(dBjh=ftUS-(>m*>HL#}+4Ik_oC6k_l4|nU+~|8TSNJ zALQTI6Y~@@WV$`ZF7<0M@EPQO&&oT=HJO#7i#5J1^YlQsHg~t_;C?ldhz#6xnR$|@ zU+%rVjplncvrq^skh*q2<$SgjL}5Ra`s2~vx0z^AzhXUt9lQw06>3qG0@;~CW*?M^ zJa+5XM*)vG3aMsk4S=;Ddh%;X1^`C(Hn2`zpVdrRXqt9KS}%O@yCI8`MEDD~?qWL6 zRibK00*jKg^*+Q4?!Udke-{l&J_EEfX*DQ`6abFNJ@$U{kM&aku@oTdXIbmech9wW z{WQ2M%em~qK6FmXB!HrK^b~ds43U*cq2trozAUu6Z~fb#Vo+?~xFcGPV2S|bpdwkv zH+y1+tHvlug8nGzUyP+Y4}4IJq1#EDi#`0RfuQ}I>qx`CDq5=SDTl`(%UF9g385Jp zekEwOps@GDM^dMZ=el?DV{L-M;_kjJLt|e(CKDr2LIgggBuxaKaV@D+yO4SGjYSn01K;g@wyR?BTXPY zzoxTeX)-ySZ@_erI{9D;N@LZ*fVR3K%7x%7`u#tv-B1^}id|2pbAh78%dqg^w(gR6FhLAs;SIazldQBDzV@1wMD^Hh#3iV=EMBd zu>b4%G1u}}%5~u%*$5Hm$1UWzY|BLw%a6^Z$pi-|3 zL*rzHMX-#$xCALs(CjyDaTunn%D;XbI;!Q!`QiCr%nzkxAbbTb`>X;A8OhsNeaCD2 z`>0AVXy>AIcpMFR|N3!orhh{(E=*|K9GJ7mzmZcjz*o8>z$<7h9+#z_B`>LnoKK4BbyHP%y~U18C(K(2^(ww-`Oo_xvt3 z2W2jsF1YI!Up8F^6H+$>$nC1Ls!hqkaioSkj;czv_BfAt#YE$M{q%Y4y&Go?fxV0y=Hg*TtI&oNkF$n@An*Z4W5B) zh4uO1sQ`e|x2~eA@Yd(r+r9vXybiRG-dAQTR^y@jgC^8{q{`JMw3jsa?=y2(4auEe zAmM{!7HM|m+v)lDGYZ=VShy?u1UKDV&oWfvySnmXXC@2BXwN`bu>gtavui#YufC}YV>szl#?iZMQ7LhyI&GUFJ?ni?F{R{m>L$|3r=ieXrJuE71D70^_-KqK$ zD9ZM84g@SlA<(u>Ei@WFJ>UI2mBaT^lw-Qc$~%qUc9MUNPPWkmC{Gqbt&M~mDY0iT zjUNVId=oNMUTP69;xRdOnsW*)Kq_W%eRb>bY5dVe;~qgbb{xUma91V}+E*hD^!$W` zKDB%=mT)%xQSmW3H-X5SfqnjitF%t88I~lE_RVijy*%%sAfR3iWR}{RgKl96^4^zQ zO_xo+nw@cIHLRsYmKC@Ttv4w*8CJ=g^~39=k@*cYLhSDd)%1yfdo4c)WIOB~@`>l5 zLi2r{d%nm6zndEFNp9u`5RJ4Ux2czF>_9s_87jRAfz;(!T`m$~>6Q_Y#e z5h?sD9tcdVaUr*v#MGE5QZEp_oeCf3YK0YrHHm0AIc82;PMXQt+I!dZan0p>&H;_~ zXgy4#O+=G5bTa#Z@LiGW&z`S;j>`PhkPIy!e5QF;U@*O?06f(Vr1%aARZ>oEZ=e)W zpyS7iQqathFits5NV)xSMT{e70mMa4TV8B_c<<$8>wg@Z zk$_b-Gg4iBBY|^IP}YctY_ZjKw~pe6(IgN(?6NcGvR9Z$@*TxP4dGI+ao>LWChze_3n`WR-^c=O!XJxoK4NM7G$ddI}z zo*9EVQ!wQ@eg;sF^j#~Us>EVaP()iPL56v9ZQk@`-rM0p#1S|aRBy8Q7$==)jFAgW zyi&;&+!y^+DLn3fb{jpgYSMR^B!OMJ>V^r~9MC)V|^i>uLR-&rqwKKeg=o8W-T&V6V~o%dmSN z)*S1{ys(>74LVAZlp5XHn}SQJsF}Y)-v=o7=zg1v_a%ULSKn^!{)`|J&RgSC<2P{Y z%+}D)rno%VE?%#5>!cBfjW3Xlcv4V9>=$|?aO8Abg;E}ygpHqHnNca*v92zRBo2!E zM1Pvu1SCS0TsM^pP{OR*WJTWXIY6yy$r$ERD*K(G2B2bwVx)ho{QNg8n2bK*uc~i& zTJF@Tfnvy6jeDT#qB7*>aU<_*fatmDjdzl^)jisodoMwmzV#j!5{0(uU8p?Q_-&y2knO2$Qm{Yac{&AIs{Vw*984UH)Dxo8$Fz}oF4>Worjxl+6i zI9kpcjCyBtx+ZV+a#y-YC)PQs3=wrk7+TETSU%OKC$Kgf*P3u~o%;bi*cxgz{`(gI z*R17vG+@`) zVF2{*mx3bdmN!MlbzE7YE@$lexw1QH?kXpvQ{E#PelF@ijGXWUaO&Zle15*ta5>q1 zD}N?0!%nV=&)mhc$nc6l$~(ybI2Gnp`0WTznF?l_ju)yX<^57y3^Ms6Y>olsqz0vo zDl?$cjdcgI{|0!6iB&L>ZJaLl>2-$1p1xAFofVhZsjRTN$K#GqaCe%FM;z^ZBiL9n zD0W-v|14r)DI^d0yTwsGsRCUJmtDGFmyl0;V~uJdf@0a}Rt^`R2w|cQ>K)+b#-{nH zO*iTPbhcVZBX`mufOFK<8!f`G9H$=MqEFE;FNkj`(RrGo%vf0Z58*6fyH@NrY~g!m zxpmR6UY?UW+1*ULr{uZQYXAFnOS(tZCOOL|AyQGr$pQW=gcF#z2ydC#)svdT))Wiv7Sv%JtD7+wXs0?j5%wKc2`L9dv3zvQT_gBoD|~5 zubxY_M2q}6fuCa97_jOyax*=@r79Sp)f)YB0O+S>qTVHi`?rSe*JBL4$BoHHz;75v z<=sy|omsdT7s=egcht&`!D+u9C<*EQZlB3L4JAgM*;2M;ZKH)sD|w_?NLZbv%KnA0 zPzC#{TkB^pj@36c%#l|;?Fx!VI=NSm&*p-nrO|E}TvFc4qyR{#%{VVEAMO+yx*9@F z)1Q%|ve8lcsi<5lp4Rt@#NSa=8SG~exTMKm-zMk3f}ngqcylGvei*y9otY#$9CoUV>AcI3K{+sA0ytfD)1a4k}0F{cFE^Bj?oV~Af znx!sP96Oh#X|FIGMcV-r*eURu&Fr~TZ1x{!ncb^3==<1C=|E1{NLhh}`zChg-V!h- ztUNW(0a!Ct-#C70{#{gBlAOuj$QXOCQoyup!(E~EnvI9<=QGan+sVy!SD?=Ivd}bh zL$w6(8(xR<>)Wi@^_~1`EAJ`fcxRiI-(8g_!d2vD)MlKB^oTnWFIIe(W<$!ypXHx| z`Pej>F;FJpHo4Pn0{e%m?UJldiBUaN>z{h?Vk>@}0(g0NHu%FyNJ z*koVrp)SOlrV62Tn~V*({>>;qcIF&=D=#);{Rgl@emu+=cpNkf0i>c!WyTtt&Bmt0 zrwz|wTa-ZVRcfY+`=u@SD`rK~Zo2-VllH3O6|?w2agBMhEdx}FQm`zB_@u(8&<>1H zjDxKqu)J?NTxIM^>a=u0LU;KQjmYA9K1N#VI!e&&AFgs8xm`pYGs%kRaj)@(lP{yh zao0c3J#YDmbJdYm@F%5Ef;}bVGdDzBBlBSSNs6%npMa%I@h|b~ecRjh!SbHWirNVd;BJ<~++j_)0hg6hHL-(Fh;f*jYM9K1; zxs=1?0A~}N4~LTW5IZA^UIxs=g1tAy*WDaJAl1}CxF7k+7MpTX0wc&WiJmt$3g<%5 zUg}e4A6=z5UC;j6!ljYW`6`=Cs#b2moFcMfWvs+V!7JG4S(qJb1M6jg6fg|LqqZ10B-+{E1LiQ{dxynvB&1+d_z3;S%{IQJ>e) zN_R{x=Mo5g58jNn3N|r`5jO+Pn?A|*)*PcL*H|P;OeXR zw|*_6mF6&Zn7Uco2FTfG&nYaud0GA)lAFkEYylJs^R8-3nvc<9BX3c6Nnqob&L9}q z+H&cO^j3;{ZytXiq-FI?8Ve5P)YOGZ6|D0XERso6>HwP%Xrex_2qDB@z%}Y zEAGy9rM0t8DqG9cE0psTrAu8j4x`9Sl5P#rUGXTd zMO+VumHvi2PZX+mx@&0PzLbFyv;X{i`EdTvx5u0Q3P@g(E7;JYxl7-0M`)J_+~YxCjidsV=;q9dFDLsFf&ZM189GLd3KvrH+b^4Q}u_4%)n}@96w<5x%}jVgX`={e@%t4c0xzZ6$XemNEf&YX?1#sdDm^k!$Up9 z9bJL1SA?8?NLBjWqG7&*UfDrf=;5B~PChLudUYb*vsU&O2g#dxx7^sXC9QIhFx;?V zNzP-9zPXHPXhgBXIPTbf{t;qA$pvh%H@klf%IzIpQI%h2`&<-_sfAwlq8*N@TdZ8$ z1i{O%bh#h7W9n-1a=gxnmnU>j@f0R_akrw;1gk zmF)pcIhWjsqkU5rbFo#{y!&$_hdbM!%S~lPj zzVhhh&6u0TH^7UKIUc{_bF&Fe&7=vAYrn_@yDpzj7{4iOMe~4)&Q++>K<39up&yhDq}i=II8%qS z%{ubprz(^cTv&wo6ESbuS@-^#f*Q;y{lz{;HApRT>Kdzk!BG`T7`|gSwIm_y+3`B3 zBPuIi>c8Bumh;u-peVt6PtVbDZ&`YUc)y{eng#MxXGH@}U6yX2z%!-sVs%0O zPn>F&Hyou)I$Xov*wsuWV%Bd$8a=iiy=92*!A|u$TX&f*vD#cBLGn^;;#KkO(gjB* zrXo0{6=X~$g&G0qk`!8}+;ojySMn}0*vim&_Tw@?&Hmvsx?V}m^t04;jR?64!aU5- zGaeSE*cPQ+XKl_HUsl0AOH3!M%77ri+9D2Kf2yKu2@PW^=oK8ve@pRs_ zAA9=u-mo(j7LzKffTrC$XIv2Otq04chO|R+&U$SxvwnMfXmTBc`*vz+{5~?wvyuo zze`5P&5u_Hd8JaG-v^|&J#Bc0AR3eZ3I6;#rr_P)dV1;t%JHeTWYjeLZNoahO-eI0 zr65uK&ww3%+p%V`5b^!M{tIVux%sjagz5i6l~Q`wJEroahYI#)q-VGwa$~Bj3wu0+~M_Kc9Jm@5Auk++ho+Z?7zG z(7HqFan|?VFN^#v2{+Qi31|Yiaq<<0v?@jeHT9^{zrB!hzB2ixvp3_n&3TmXH{ey3 z?wovFnL(e#eJa`EjMZ?P4<+*eSx~emX-8inc{Ayjcn|&5hmmA>bYnQjXW#67P^|v@ zD1jo&`z2O+|{lIi+qsTjew=?A`Vc+_e1f612gK}u9FoHYc34v! zD0Di%zYN{aE_0wbLmv+lHE-_=F+#7)>A=-Jd~ipY*QdZ`?n!?Y(+sLq?8GS?By*{JgJ_ zk3<`457OR)EG3jbcnQ8tmb5zGk4{|*O}-Icif+m1|ANdTz`Ui?_qH*}RmqkZ7Gi+{DSuRCgnb1z9;{Cn?@6jY{ zZ#cWI3{SH+!r_ogV1<&-a42hcOdMbQV>|i)!quNl_t5~L6P^YNGgBN61qVn!M!!}7 z7BjMg+XLCupkw^DrO|A^>2o8>ZS*l4Z;YlJZw+vW~`0`kpmE-$V`|CTlG*5DrjIRor_j<%}NZ{{^5xqZZ z#=vV66E5Eg6_vW_sA!+;VjO6f9;a`BWuaIa>gUykY9T%Og1@juU)Jbn+g$|&E%CLS z^1f}=KGiL9xjw{BvU^z#jm&SRLJ3;A$b>gKbTQeuo##md@*M>Vbd#yW0^#~--n0Zq zfvKqGfvH-KZC5cKn+H^e)lBnIEsi%YOBnVO#EX)dpB}(Eyp`t{GSvnY^ zTCR>)Tm0!|i89~u%y`!Xm;h{^v^FfNsW+L$aJyZ2UEjoUC0tFPj@e+dxMtc*I?BFJxiSeh)_p-BhfBKBbC@!Eb|iV zetJ||+vWFP-z;=dvN2Qk=g&H<*rV^BnP94sI6ZxP$I#rzOw@zP2601>m^Y5d6bQ;;wnK&*aCaVZ~w-uCWg3Vp!00pZ~ z)VD1s{BIvxa9}*0{F^K@q$utQE=A6o)a3t2VCy4aAEC>+i@39m?Dxr-eYCI{ziWYA zzoRODktX5#Y%5RP_s_3~x{cH;{4=O{5{_c}n;z%)M}-+DC?{=OQs7e~D5kAj96n9Z zU3Jk7_e~R_)X}rf%PvSgGn5o298sNfC(#fj810TCbvnar91H(@5-Q2 z>Do7ECxEhTj7cKVWJj88j1=?Pa%iT#+oH@A-E9&piu^Cv%Lf1)l@Pk%SB1(?^Bjzc zM(w&GJ?!sQQ6kx)v;GVdA2A(i(aQZhWg8MoEDN`8a;Qgwmca38X1g>ACv1*)8O27& zhI3x9uaXCfL}mHn)KfN_&(_kp4BeK;rp}4gcN^zWtXwhw)}Ve$XoXJ`i4%bwDgKB2 zJ9)^mQ<-0CmrbfSsp1_2{nw75BGYZ@`@{no{gb#9ES^e@tGkVV527$Cp!(8a{SR-|7Y zizd$_aQJ&lz>Q7f)^vIiA6R=?_IuG5#Q0bZV+A8@1jeU}S!Rxvc_Nlt>Z(7dE zJg1-feZ5B_xVHvL@LNbh(5K4>u~M%92(Lk6Glc+&ffFo{)-XS+Wnm1M_cPp*3VUor3!~aM zSRwAa5@Nr*Z1_{TJ4xh&YZ_@_vHY|PggM*Qv68f$igYikv6j&vt7S^*enDul5Ufon z7U|#qSa%LZ%p9cFjtDqD*$;Eh@iKReLQCuexuJ_jHMGA@nz?_PJOYL9SHy95`IPCi zuLxEs-jmn(LJc{otxNwyMar#}@HCC)oArI;ww+|P50oXoco99LdO9(^`%`{amcwyv zDPEyF>3%)&1lPd#j-=Nud>wnupR9(u}$tUs%4%Hs&F zCS`2o&EOKM0Uukv_iNv_XKgnLBji%sA{=s;3QtYd2!w)8vNvIA*=af3HA%<3zH z^iK{F^rg9LZpc8^3+w`tD2Yku>+jsg&3e{Jrt?UUP!skyyJwK~b1Ms)`N7hx<3j?1 zS$01pdCXEeU7XMQOUR`)+MW#_u9?MV#UJUOX}eHOzBR)Nf7spb{HaU*{ydS#0uWp$`MMn{*0zY9{4sADD?Z=#i88 zn0#&rx!vjnj|s*;-EjJNY|xzbn@%d))3kSQN$DCBPir<)R4IZsz1FvKhZ%- zEJqoeZMp>(v3ptC<}gPLLI0HZ zOVe_Axg0%7H!sPXTSeK{rAhJ)tyRrN@X?twUfKufwUc!DJej)d5776k>;}&cK+!~h z0cT{gFM|9tVf7)UwnI&TtLIgnhMRH7kZ@&pCGh_b*;3?bxn zn?#8I>3VzRS#jmSv#a<5%9BDx0^!5GZ z!GiUkVvU_534zT~GL}U>Jyn}&#&xd{@zTBN;qQSh(B#D0ZWg@F8+VWCyWxE6f8=r4 zTmO~EaaycqyHI*I0bP4*surn`FbZ|kVw@-<`W7K4a3y?v1YK=%`Hj$kxR=PPm}Cw4 zWhX*Iw^r>-gQ+szh*y`PBWIqH@zZ=%MfVdMwALn%!@CZVZ-^(zMLTUHz8>n5rN_-^ z&*fb_As1DAo`m(9NbJS0=j*wzJN>B^UVmU!*Fm=Wg@QTe;s?p2+KeB#O$o$U3{Zm&MK@-QaDNXdPQ>&nkvHD&cnRUg~euZ@be{J>k!-Mv^d zK1XBkD#m_83;NXM^TD6nzzKq2f{}3ad`HY)O>zofJzC%n8*{ zkS~x{;4Nkf1xG_X8c|glCG*A1;Omew{F>$*C|}s)^?2KMri89ePQgpo?&19Dw8LBu zULwfvz7-bS>W5eT!(Iq9t!5kems;m* z#c%ApM3EPa`Fp`TefK*$D}3j}F>@x|2?Usd2tc+z2|eg0?YBGZfj`4>Qo!B>Io{l0g)2H-vK*@a-R$D``hJ*-t%#=?EIVIsN`gY{mtnDXa%vDz~KAWxnLM?_hi`&7u+H+UvoDd0G>i@9j+g^f0 ztsuxHE@m`;KlNV0L2UK2gY4>^1d+EcZ%7_UmD@G$WaZO9NHcAObrgOA*b6!139jOg zFSS0YqT&W}?91_y3^67iy*P3-?-wZIHe1b$MuL#^I`<++JMi&>W(V}_2Ot?)+KQLk z4N;BalGGjKh^w|sfChqZYh$F%8d%m!7^W*WO+%e5V$t9SSCB5~TEE-ZUy$-!EL!Pm9U>m%kRY#o1au~$;f&kNm zkE7Iov_s=lecwz;9rUfZIKHXz)OBw}rv6tS{+(7aRtqEq^E9&#_MQ^#7Qh#7{sPsH z7_Ap+cfbACa*MSa7Jy=6Zmj;JEwa7!$k}D=g7!#*#E0a6?sp^=`hN5QNLHjaqH{Wv z(1LNh30f0DWd0bu%ccs3Q~pP?aTpyAX{O7cSa{#MQd4I1y!tzx$@QFt51@0fcZ2`c zl5ZVc6&vT()OKjZl=`i!Hl_?&NGO{2)Dv>jWz!qaQ-}UerJ#=7b!&(&%h<5%zrK#9 zlh%`YU~V1Wftw$=NU6DT(x@_v^~|EOUsT^@3lAOr$=9-%OO>itUNdoW_NhhLbk-jg zsi%=X90pGVSXE!YU6JeS^jRV?<47?B+p~!+(eN{gtYYl1qQjI*jI|%9bD|(ed@-hh z$mlvu)Vx1D=!jcLCuX01fyZZ99^?z3AWLMerN^d5|NGD0wyH>GsJHQI%$$BTlSQ;B2-J`nM& zdHK9+@j0kR=orNBhQNed+amjG(oEC@Vk`EPyBH=o75Cr!zNUNLcD4SF81Lp{^i{>C zimS}+Q4F?1B$UklL$Bjq<9g`?ts4ZQ)N#vZ%A?N&MCv+D} zvOa-!a?VAx8lrRR5Fom4E6~F#(f#!6ioC1_|5_TT-;`MJSgcL_ZG{!<#XzUVHGdDI z?tQP&$1iUz!iq!kSEfT-JS}5BwZ0|`&J*ST`eEW-Sl3PL`sS8rg}qt8H#r6@r@);? zOMtt59d!EY7O8>QIceP5GQMoEsA)sNxk2a_@ltmHuUCC36G96}BAkKblutC_P=zJH zPQFuz#H`vZ{b&jhW1d{JQ&=daa@}rwBD-1-T{SNP?rlWFQEix=5_WybX307M09B42 zrcl-A=RUTW(={4S0mU)oh-(DZY8d0$Jr4lJqlwyfH#m8bq${1rO#=RiHf zGgcyXnn^Vwt*{nbmQ-XETW##k-wDmT(pPci^)S+zpoDhw^3of)c*Ira!xo)aaw~*5 zV+~>d2HZB^J1fy>Y7u1_CH}`(OQ$8<;c|(^(Abl}Z@SwIF-3-k-X#)kWCqRIolV+r zNC?qlCh=eq(u6#zG)3l(RU}j~db#qao1ibc{(!MI-D;d1JotkUFV?yxgZI zr5Ox>^(QCP{$>{t77ih&mzVsn>FgXQuPGefKxW~ky+GhMItlO zuh{;2j6$>{J&WVHHA!AgD=8#Kg81m`S|<1R4e@rZM(atDvll=#VAXqk&(U~->m-0( z|JD}(n6b&M40#3+@mf#HV2qY4^DI>N09vN?(;hHIFI8OP8N?%03d>s?2;dLBt!$qdoI8V$nISc z=fs3_JUTzDu;K-H`baCQ_eII!a6Ya*`@q`+6;Yc*!WN)sB!J`kS=KUX4TzH@D5qES z^YuPdtqrJq6V&-$P$k^7vy-qUWW5RHA^~c0#oy5UrmHR+ibW!w=}tP@mWI7ADu>9= z6g5SjMB3rnYn6UU&-6Uw|(VZCbwqpwlDNE@d$dkvESA_`9xCRgXGJhCSCPu|!PL zk11AdND=-T! zrG33h>2r$CaY-KWQJSgxgFQf5x@G9H@V+#Co2`hJm7f7{Is#^wm#TnWcbXf+8%H-M z0kOp`j0vxNbbM%U#@q40wFC9VGgFh4OQmZMfb5W2w(H2n;HgrctNk-dwaqOCx5)1# zS!IaXrl44=TUZG2;R&67Xm2=Nj{yB*u!o z(=T-E&X3Lh89~VBx@y~$wIf}-krc*Q4E0?uZ0KJ;sD&vU0|tjv&s}%?b*89Kpu*l!#kCPuPqDhF{DbJVl$dosj0(J|s5k3? z!UuA1Qd8?6V7K_#hd<&8knQzrj~?fw$0{sK1!P{3KD?OpNa;m+BvhXj#Pjie)=ol(i3oIo0t7X}kRFv;~e@g$iTepv%lG<%= zonBD|W4y66-{HnWnR$h#Ra!!symZeK$>%}H=ilFXN&V2Jx%_xZC%P~fW!k#h%t9vQ zsFCxIz%_{s2fJMv7G^%WjIH~i+!=)7z!N(VBky{|`MUT0^M-?8;3!iM%_DrCo*UZOKGj;(6ZQ7(B z6w+wbh-+3%_>%IrP7mn~E(J5*Cb%iG+p7FP#CzT^v%_VmwKLrgE;0zI;piXVhci}$ zJm#L=(fUnF3CCxWb)3oa+#lYlx#0_hqKdSA697I(4#dxR!ItCUbli>KsCyC?ZW51j zB!$FKIGQ8wu@N7iBy5crA7lj%fwD&%JFlL+0%~WIuyQno;{oV1DdXV$&vHDW!X+G% zG!$?y@D*~wAF%_5euG%A2tzN`vLciWQ5fJiUnrsjX0QwGlO-mQj74Sl+DF`lg_*=R zO5G>(k7Lx7&4G2|dxDpYi|LN{vI$u4E52X_YHEZ;Bm$nYHFNrVUb^*!4#-;!DtY`y zaUwb}&oS<-dHg^oM^4LWQpp^6GLBh^A8%&AC<3vA-XN?jT=y53*A6G?rCF@HX;z?m z<%mZOWRJvcgyeB;7Nc8BuY#oBz{@v*P_ijLkTkUg2P9-^cz%ewOW;vGB`te2 z4NiU^m@X^SCGY#dRcm+Ny<{zIWk8D6P-5YsBq`%YuB)>qUlheL`T5UlQFtKz* z{_eCV`694T8aj)Js2Zwor2_pMFjBD zUio&Qk(4mtTik1iY$hh`WVO=~Vf1o?S8_6?T~8>jY{&z{-V*c|d3%jUJ{X3esw<^T zSd0ebZUJwZ1vxO{ywTYA;E~K;zLd9_^k8SB{4?TYFxJnF{2FMZ2;hJNs(M7KAZY|R z-_}}!RIg7E3wB53dzo}hoZd--XddG);yJ7(u-FgQl4f$yPVe7I4QZfToXIHfz#~GR z++KzXDghLt$Rm&f`2EHdAb>BjAfUT~W&ZHG3Ap`u2*AY<;)PLuwt@O#GAk=aqh)9} zJr=OEiT)zZ!&)*2;x1TA+yy{8Z-22l!BFG}e8RtTVu(Ldoti%OE7&J93xZu0etC?5 z33DJ433>}F`LCH~4e!JX0?Gdx7{J2wOq+ZI4&9%eZ~xhnDgdwh(-Lpc&g)}?U1#uV z^KJpZz+^)d-5h-euQdB+44rO0*X-wICQMNvm_ZMcmZ|aUJm`XV+Nb z3wfpBAYtn(oQ^e-UFMa zI_ZK+u!w78ujr5twqhv%*^>W9R?J`1w1v;FfCt=_{#;egZ8%8}%v8hnL(g;h|DB2d8=R!?VXR8vc9lm2ko+3@;WOjUA2Y6=Gd%*+ zD*vxd?*FE2#(&>Duz;@?rYxXWm@AO*gCG*9ZawpZX*9r6j{ng7qDw~mJo%{&K&T6u>(?J^RdwU* zXnL(vlYcj0%b{0{eP+Hp@x5UHKYW#!818^Wg6_q;D8RMHnw;;*kv{|87M2OwWTs&RWi^}aVy8WVA-_qh3?_&m-zsb`#{;Sk#b%C zf);-MwI2NuFz{6dX*HVUp!Oe{5)>$Kr(`0=o>LEn`eY<15(Tf*=K*)Ye$C!Q6Om#a zAi#g zJPe*01gU`9dZ&I4>c2+dfBH8B#^#e?i0Y>KpPu|{NKKhR%r;o8kxuwuqyNuo@PGJ5 z@GE#>F?Bmeu7C4+2iO(Gdf{Z?-|SK<@WR|@Oywrd1 z`Ts0|f6@7{UkUza3H%RB=08i|f7pxuKXeLEp!mVGhjxJ$Nd`=vs_4`MNK{XPsK}Bu z14lA5qzyQGW&*|@&pFI5j@F#%@-#i)sJcx4uM_hP)$;YCL*!D9KF2%~j_!Q9Q2Y-? z%0Cm#ln-vYMwd{-XpQm&u^F{aJ9rTpJxOT52rDFOR867R%JKYUU7{_zjiIzDyJo>z zqZ*L!3joUyLSG4E5Z(gcWo1yxOjPm|qk?(rwdLXY953bS;i|M=&#%eQ#j8r0BvvC zH{W+*Ch6C`;{0Ry#(WY!;HWuZu_22@%%)^@+Wk-scj@QjRFU^^r@kT+M z^c&38so<%yst{T|z$X~!5s6WN*MpgNa2W!4roMSc0IE3v*;Yv^hrf#Ezs^?&)Y2*i zV$wKZnqGvrp-=hu;Z5-jE~FOQ5Fj@5Vg4gdKraf<80Nsad&$e!D%eZPrvLWuzs?3U zra*DB5X95smw7O^zI=nI*O1LP!r;IADw}%QT%6IMFaq-tnkK5>jFk%rtXp|D;uI&q z@E4Rcok+R&6|K|n9thE`UnCs0ZQTL!heGgb;=qR}W7RY!qy%YVcDJy!gv|4&kP|5E z7XZ17tgPv=fJX8Tm=*TS?r2&iiH|zU*V7fEjqowQj@Z+Y= zMNP{xumDHO3d;uOOXRJJ48Pwrw&>6{2h=KhAF@adz2Ae+c>^)*T4RS0~j2 z4x9$dI9}xev*hE4Y?q+e{r$R6r6qd^z{c;Cokhh9E277a6iXsw&~h$(?oP=|PlHz; zP^*l$)94h{qu`+GQ}YDDl$W^Zzn=cMj)Z)SOdWXVx%#Am!xs~dj4UaVbC~bc(1`Gs z3(5zqc`6z}_f5vpDNdXY3zQC~f+oL90zMpLf>c6_HF%$blgP7_@fKt%7qvlsm z3)Pi74WoD^M=hA@lxVQiBCHTz2yR#NuSqpm0VMupXlFMelm5_^rY&=iXgLK(Y{d~~ zUs?)N7GT=W9wk?-1E9J|aNEcem4&_l>;May?po$O8}C!%z~mlRWZf!8qR&ZYfgUK< z@K`+C6oYL%BH0r<2KpuvWpqpnSbw5haSD_JdH&^<=X17!0uCbU$~ za^p*weX8<2&NR7L%Um5R5r=U^$LTAU<+%W4_$t!l_R5?e2!aK!*1iI7>%r$syVz&A zu7)bdXOaQMXuUVD4m_4_(gAz&oB9pX4s2c(FQ=gOTl$J#y$2jeHB^6L@J`mJr*x%=D9wAqFYm z(BDfKo_T10Xnq$#|AW1sF=spkt+L}6kY9^r4*vcEilrYSX=p1_h}kXp8R%gN&<^x- zesoMx?_uq!y-J8`-pew_Z2qdSm8u~17d+up0jTt45 z+Bvb9e?JM;_*&DDAVeC9zAR~Y;b?IK$VY(lraGQ(5ORT(hxn776jqQi6Nh;0a-Fn>VkS1Je%pygEF_2%C|>x}aV zo;ohVu$!|J=NiUKH2OcS3!e}Y#Pw1HPjSU5q6Z7zsIc?$&4Tk{E{d|lj%si5g+98J ztxn3GIAYB@U_k7-W3nCQhw-Z30_AhLkswx2BczfRfo=o&k1hVQQG1xT=2cjYfiv?_ zuF5N(qovlX4*3QAVFPuNhQ1&6)H`k?TBr_fpj(?notI$4PtGjrnxuS8d{|yQssVkb;M4w7r6y_7a5A*E6^?s1uE;J$Am)A0<4M1fw4(a`7S~( zOI3M;-~qgm9*Bs#n3tN5p`lh`gzTLiaJ@F?XP08Ud});6M#UlTsHu_l@DgSBgK@>y zW@OsxVwTZa?5D`kh-}Lz;gmngf%YDcz&ags8rrF2CVRO;IV1=DX6?yM}%} zbNpKK!&N*vhv3XTd!N4je3^tg8#}KM-fqO~EOq-#3??Bp^sM{Q2TOk~6e zwv5OJTj(>0SG)7NKQujj#=h$wf{jRh4`?px$6eBIUs{UberTfKb_gz`q~3F=gaUiBh;F z$?sX(Fj^dfl@#^woG+^#OfRNgot3&t14fkJYQUeP!jU75oh`rDj5byL8o9SBM2l|? zi(8WjfC8j8qEl^@2D~>8o!?Ht6tg8y`&5AQ=atR6ns787^ZNE|G4*PV5q`M2b9P zH8c#Yitzy2s79iM6&PZLI>0Krcs4kx*$@Byz7RCjOdC7zvub6I4zNDYJ$< z5>%7c=;>~IXxCqGimOsZSz$o-37ZvCHf;?h3@+vmBHT%LB#QVlbj zs;vHQm)0!0A#1?fDO`=>;X;ZAU^dp;jEI&lAcZ;KAP{bx)(v&Wy>z|i>t3+Qa`~pN zDTo)CUl9HAuJD*9F!`1>d+XuA@6P5(EF=}SF4;^O;^EDR$QTCeC9CYcS4p2jEc(#t zeUrwS2q^J~jpr$(p;d@fnyscT-DKdM;x>>P@n`U5^J!Ck7RK6{qfWbTcvR(R+h8Rs(1{ld;T6&|SJooQRC zN76KOi)tE=FxOG?iery~ktzX~Tk-TIb|QcRJ7+z#o75z!V4jM>Ns&+%>*Wmg-+M#- z;fL^qu|wR_wrwiJB*9~E=yQj}OX}^Xp+0G5jki7fa(rKY#u%Btc#wlM(?+VI-A4Nv z+yyj<*TT@1X~^pY#1?J8YvD%l!bhanq5`42j3JFFr}9v30_4gi1!V%_QL_S(wtNjw zceV6_dVda0F~2tv@>%dmT@>#KmpH%Y|FVZT3KJvDz;=*OnI}=l!QbLp6WNO213|)CX)NjBMCt z4oI3N?rQ~@2tJl?R{<sV!u0 ze?svc(`P;ofclYPB)5M1f&>X&u!rYOTDy3!sCLCj^AItWDXmdUWyI~TXoUdXfa+{Z zUM|(nM`o6Yf~gLIJ)~EGj+15%P=Zwui;|_tj-u!j6ts;fqX=oS02yo^*-wJO9uJw7 zV_FZ2Wa^p)Jv}W0H=3?}(UjI7QoEAE8`Gqm!VJ^rKT}Qm@lHMG@K;`j@PBBVBZfp$ zy=|K50Cz5@TLWyVeYl((Tc(_Ja-ZLJh2wq*%u|_wVf@FjH?Kz zC8Vo1iKr8Rpq*8lb+_3A8qze#>s7O&3bt%|#0{Gtps4aPU%K`*1yHKS-ouJWw+t?>t!(iz@=WoU(HO- zB4TFKU{&vyd3{Sij>F93Dl(OGICG@mfZB$HTnwX*m&u;kNL!A8)a1zk8r&6r{qvVO zGHi{VqE&2=L(p|Sxf1nu%ZE>=QHt_jVF6Y-__Tx}1tDIHy!a z^7bYh^U1Ax=<9P@%HG1y7_=o?wFwG2=yb-LYKPPxgd{s|rnfzqOcGJ@!mYX-tX{29 zCcG{+)-_+)$6(VzR*GAwjhrqXXb~>4!w3s~rgri8)Y`$W_zQF4ae6u6VdZj*pEnk4 zi4jU*vq$2S7To}0Lk~@s=BC-WC6sB~7isvK`yn*sf9(mmP|+ob*2@H}9i#2w7>%bc zYYlkALrrjqX`XCqSO;)2ujn69TDjaFZV5%RIw2hMY|1;4cgLrxNo@lpxQlEGF`)V| zr#}-CLFRUu!u8Uj#D{o{6Qb$$GQWyX-xo7z|I!IC)K*^T1XL8vAFWU>4Is3*`kT22 zlDbT~(4-_aUHG9rK(3G(k>|27%CfXQ_x-dXdUq+ypVd;%^^zdxiyNGN>dqHuhljJt zWE_z(E{p2SG=n(eBzj7iPnJA859p(CI^tJgH|o>AhQz_+>#GGU15MJpwe+{7tNbXW zcOZkX*K~~DfihwZm&Ol8*lDLp;tb|ww6jUsLe>ly!Q}@E{`S*`d(I?@2d+Lzqls=7 zeGThsN}A(rWopsSw8ZLUd#Z8t(+r@~y=a%OJD0Jyp9=PABJ6Xo}3F)-!T_hN2RIhJE2$~V!J+bTv)L$ZpX`3aGK zne!8n>!+Op;ryB5`fDm?qNiGBtBm*b?6&8yP&(xEoel*?%QT{-g1hOS^)2_zInyY?7+}xjO0Aa%}i(S7fWJqi1J~EXLU&U$&*-m2Q z&xE?)mBho9gKI*=S*V~>=xMH*`+vJ#~@H4Zk1$uUE#&e)UWWJU$U zwG$}cF6NFWR3i#$-$9y(G^3x8cX;8nE#+aHg`PxP8P8we{!s{&dN!!b{b6q+eL%?=_cD zpR5-(8_Rq|J^y?B*qvz|(zReKq(OT5Xn&5Y9pbDq?M;Rxl4K+jj-)4lqk z%C*=7GRrE7hRSwy>v+2CK-g)@{`{%^=;%w9dF|p)YPRl+jgcOv1T~sE{HiXGwvH-; zp2|J&UX(R>h~e-3@H8eUT!bKu;$!wP#u_eJ6r{Lw{GCAB^7F`9WZD<2^8|#R!<_^b zfaBfL5!IqrIWskwqZa&N>juy&pW;o{owTECB>LXK53m;g;MQe&zy6Th*9m0@$*uJ8 zp2xQ5x_TMY^3L)GjU zX+70YWyN>#Q743NaG!ACqdh7CPM8Jt-tBHI*6Ql(;HKq*P(53evp#~Y)Y9ZnXgN|- z99=5CoRBQ-$b!S3ryqj1g8JQv@H_H(GvqIUF2#v)rsAhyyvnKi@Ko`ElkUnn{#hqQBIMm0nh^~0O^s!IC?ZXWp7X8?Lr|4o+2P(ncbBV2^#lV^mrpEh z%Y#(Yi9swLZ;WxoBAPC1xl46>*z^7gfj$+L1xH6jU5n%zvqFP#>)mVsits3Zw1X6gfTV#}M14`;Q0BzkX< z74udL990*dtUrclc-%*o^I|mQpV$Wah2H`dCep{_pqH@V*N@h6F>;pi0h45mJcQWD z|5!YFtr4kG;v8J@_EM^t_@_|dMFxI~`m!SU6B9jg1$SF|W6$=}P5iQEAl zJ!b`lxr-&>iNZrVmaVt@WSJfRN;iS!E$;xdvl_W}&~!hFkV6d3BS~nNV;)iy0h>bs zP%wvh(3h<>W7f!C11FSfX=}%4+_Q$C*MZQ$==m`)y_j9K=;qi{LMP{MsxCea#sRv- z;=zxoP>@s|(Nueoj38UF(#$(}gC4w4q0}|LyI*VCS6O~i9bC6eR zRV7sMF**i~4t z#gPokY3)-DD_q(ieYrMksQX$$HE7RbFbHea1ubcGW8C>n?Xn!xE6!T8qGs*&hYPp3 zc=B2$@itiev<~7YocobHzqIl{`z~nhb;fYyPzvESIXpu;ejSZ)z*!@C&6D*=Z^zmP z8VSd^AtqL(bI&$^VaJXheza9EtaPZRUM2WzQeq@F*%;V-^ddlj^B{$7RvinrL6KO) z>Sy08k%p37cWj8~wbE|fbZ+LR&ni1zf@t@*`GJgoE*){fVa`kfu?1Mokr!etK#iWH zz_!(lCsyr{$SXiIVt}36_Gj!l>h{l7D!Lm#>AniMoY%Hc_Ml%dW01UV_xJp8d(zqH zNjTw65kdmt$iF|6gKUNn5yvKFKcChSZi=b?-o(l%;RX z!lyxKk%6)n6H1xu$CmD3JfVT-VA)yJ2Pjlip$wWi$-f&=+rf>sr_2t{jOS{b@ti%W zowVdq%p{pYo*2`pg%UF!l|aupN@c=sZ{QvjJHE+-;$b;|8J;meZb&=VpGm? z!uX2`5JaD{TQuo2tz`Q?N3mtncdU|y$#6P)R&`s!^I1{cvm@{i7U`#+meENokPxhl zU=9;iunauv^qR4H zO$Bal*@5*v$IvrWPwVw@j*`2T+x}`BDAN|0ZRa5tS+q&#%Q&smdmGusfon4}PFe>c zC4ct4@Li;vK?hokjwD(4=~vBo_694M)}{t2LZh~&`;WscU7QZL((L=_uKAgA7QPbd zef;8<5qAZN4K@-Xy`~j4>OmxQNR#>Us+`vS`z=Tn>dkcT6D7V&r!dwu)L)O5_RWIS zm2}R@s0F1~0DFKOR;i@$MPTISAMEC}V7#TE^I&#U!xFYesdH{yK`Z<|k8ilXZ3RMG`v&3jvP81yFnrMnw+? zEcEG0yK?7Lx9j3nHdkr^F6A*)NFDzoM>x}8g z=7t+bg~TE;ABc4w(4hN<{HTo(w4CxzH_J^v;}r@?EWqxf&4J3yV-7ytQ3w+Y;k%2P)N4(>VvRMpyM-b==?#*HxImOj8@m zPUK>W!~RTRdoUM*2MRta8sEw>RLnA%3lxhS(6!6-&JBLNiOT~g_d;D~&BJ+Keu7@# zjm0wK?}P#7G_Fr4mA3Av{_K@MBSQ8t5ynTN8Hh7;epQiLHL*J}XlPy2F7yQ-lnH(0 zD5=&CoI|aDr8kX*=x5zd@CyhMHu?Ts7JnVnGXLQQmMUXM;#-fN&0-7Y;VWBVesUgw zNh#J2Hp8c^>>Wjoh!-su{olyPYP^Tp^p`Z3THw8>UZ+?Yf31m}7ljJs%G6lD8Q69q zt{tRRoJ?~nQfXTLh_-tgPJ%3PE*$*cUaU9jLJwXXm73 z1SM0p^7D25u49r!K}@y#ImCCS)Kot0Pv0nZ-oi#Y}YM3hmS>6&+dFIp}K)$^6c>f)UK zawWz~>JQ%^kTf4<%!R1Q*h2){`SgNYm`k`rUbOhPSw;1{YJQTYZ`ejPsh?V4ZQFtzDLJ8dFfX863RfW*+ANgIlp((Cf3uZY|D|eX=kAW)Ar2;6(2Cl zW74|;B^_G606W5vpmM7uU;=mcI`hs-Sc_8%DsWIx5Y?Lr+X3J#qka}IX3#18ijCc3I z9lT6-lVrF~GH{vHAU~YKf0na`>>(?{(nRr*dm`JE+ULG6T#(rsLRr2#uOmE}7T^*- zV9C9D8#$Yd-BLz&2tzfA*RD1Q-Hv`)BJGrSR+4eeb5oL(6oGj)7YL?ur#HrUi#^t#iscQ@P6w^U3%*nLRehcGa?Ha#JLM0aEX*6 z>b=!wImrg8aJe-I;<7x6Hm>hGIh}#>z4~(y>NU*K@9Ae8BCUaxgylKmeRT8bgT&nY z6Ys1NMyjw7UsNHXcEMk`Pc?-QA3hIzrm`H*C?zlWj8RCmJ^mMSl3>hpx>!TWRdKN& zLK~v;*iYb@AJLa=AJ(2umYzl24jt^2$c-;J+@JHELV3q0CzV=DO8A|IzAF|D6jS4f z27Aa)=S5W7PS>sV3Oh>i57;o$`?2W<+7C>H2aySCOH7(`|(L(cf8+k`-zK@cel(l||!QbBuO8JP>6FRY2=y ztco>>!}aeg@-BcvWxNL}$2ufIn*YM3r`?d~JvjtJ_@pC^wR4T?G_D?BE1uMh`kFey*0tuC?5?Q6rd_z#w?2>#$0E0A$XvXjF%>|wSQzpE^ko54a8 z9~gaW%9YJ9tC``=$xJ`KUGI5Rx!+4~cundM+>%=|tl;{N?XL5+GymAM2iyFv6}$FA zMGm3;4u>c|Ok8GWlkSK=iLzE=Ho{@wi1bu@xX~Lw%Eb-ZCU0GsQQE5J8L0FREr7dg zb-Q@Zb@jLw@K)7R*Q$C1zgY)PTB6ieRK=oIKd8()-M+L(8@5erAV;rPYj_(qTc?L( zSaI^spK|)z^n@k+i)}#J6^cw4kKX3aVFS;wKLW&R57{YNL2^j;ZYl7!l=98YBP7V$ z*_hfE-fJ!k^m-IF+V(O2fS0;UGzt7V0N)3Sg9u+5wVpB+ zw;MA{r4qsvRTkgq?84%dUb2wAR6Z*Dc_7+f{aeznNrFI+S75Vrvh^L*Y#Qj|4cRVL zcBnj`w|Q~`^87EZPJiAr6Wd#jy{k(FZhJPZUcVyn7LHbdh8OBq%eE*>itG4iOoRXw zrQH{B9v6Gc5Vd-O>ale(Z|bVY2ADTyIv^VrD4d^6nP|$i27YW3E#ZO8b)IgzHss<3 z72GkCif%|dY@p|@Q1@>UBY-5m^{Hwh+_SU%Jdc3+iemGze;jefcI4~Pd9`{QHJgt7i(;s9$NKamx zT#-Rfd5vNrYhtwHI$8y``5=c>_hiG#VEh;7VhSpdwUCO&R}z( znQws=i{ogk(Po-u2d_c>djZs!-y$puG;=J$E+=mSluIOCAz=ZjwR`KglXYyLM*t?1 zghn^QqFis)ID!2xk-0C*)imeqIS&fulK$f;a0&S`6VLw&-5wM35nCvB!n*je*2~!o z0@+@bDi=UwLviZFp>r0sALaVGkrA*zL*s(J&H_Nx2pWd%Glm*8Bjg4mPcPJ*yNNO-E)^ddetTu_lco$gm6P~j5uO{T=pyKQm= zd6*#nh_RMFWqq^hQPAK=Lz;N(Nh6m|dFzd>(m|S3IBd$sR!&KV3V-brMz=5LUw4F? z)b-tP{CIhBpNmniwaFia6XglTGIZM%W}^egM7LXdb@%~=8A}HJDRtj>Dqi;Vcv-6< zg_B*&HOiIRR5|u_WeUx%n0_NmTM;);0CxLOck(-rhP${1xw2P>??*okS}~-nejCAz zxMZ!FW5*E)(qbW(v+Fs0x4}m?Qx0-CQnMpTmC2+X!(y~VFI3i#4J>_iHX>qBfpxk} z_txX5rUmtiWBa+n%p`2Q9V6ga*sr>Nkz`DoT$R zXIU9;DAzrc11L<=yNmJTGTxmBJRz+ zP|IGAV7sitcpm3(V2n-D-Hc5+Mp^YgzsBsKzqr+V0*&@Q%TY)yw{i5+&K_G!+4F7>w>9 zLkQ#PpO}Xf@RhqoDqCT=zm>q!ZL6-Y9XE4>#H|oN_ypB`LRZZPA*YmoK#doI5g2 zMCi;`%3otV5!3&+p9Z-|P#8{I+p_KSdfR zs~E@=Z2uRZuY<|lX&WR1HvL5R$?9RrcxU?Z)*CUKeZg3$Ve@jz>0BB5TG^BnSoZthvb%hXqlH z&i1)|{;4i)(4X)?Go#rt?piXF&PY>ngJ^>)NOOqI#u3uRBjRTPE#qY;punGd#1@w> z8Y$t!LU+?z@*`>@LCGW{P;}i+`Ijwb#>~_2`_$5RGjHrebKiPjWHH6ggmvy-^ioIT zzUJgU0`yY+K8Rna$LpP!sI%p=hnvgYp7o{`N{`tJCutjD1Kfm4r}ezrEM{kpPk2z69%KT*^cIK^KLR|c`^ zR9@i(^;CS96|qirh-1{QvDyaU{V(ywH3ndn3aWt-{KiV8{;DTi&zRvub-Cj4MWt6) zm}@hKTE%IBt3N18ywSV=n$7U#Go^(y#nF?j5i=*>P&+}$Dsxy*9{n@KIzT)WRo9ef69P$==>A%k~62CCcD;E7? zYmg+v{%+nOZnOrVc0X~_V|Y4CVE4Z)Z*x#@x# z%v;HL%YYTE+iT<(OIT_HU=qT$_4$9!XRBLBB6Qbe1h@SQDzl9bSH)&vNTF)ggY?7f zmteM!w+)2D9&A0&q92U}PQ2 z$lRn@Dm`2iwp?u~Z-}`OQys3@0SqH2T7P?3(tcat8)(j=4;QX}BJdJ2vJCK*ZQ1sB z%FXSo=E37!w@)u_%Puv@-}NwefJ#(!j+oqFQ%mb-TCzLII=}ozd$d4>8WP^gP7~aG}KcX2Y{TXsJ4OxQ2Y*Z6= zNKevJAiGZrC|ye<%A1m2vR109lNL<9UsK{Zo&+j$3 z+m!P1ur*7SL(4=6hQUV}1RW}=m+`SjcgzI#vWej(MW*xx;*2?kmWl#t5QNvs{1qd@ zXc!|}Ri9G0*2i*d>gJU(jLzVCBcQ}2>-z804@)dgn+#K#!4e!%j))4lJ~S|Qh+>bz z=fb(h_A+=c83cXwwV%iy^)yf@JT)>a=jaD6jq7#y;kW@MXW`-Orw^-@(VCVOsJ2Dy zbj46+j4?d1sOE*RAcQVgi=Z+<*k>#L&f#6z9(J3r|M>V9>WxfBG-oc2NRZEE<5Fb( zzzmCgh?4y#yBg+5Y}I03cn(ta&`|z(7lQ z&98dTndJp~h|gSRqY`F#;6AC&9Il|&v=q?2FyFH<_lVMms{V-l-IuqGN2wficL58( zwDB?Op~@nZ>-}xk z0$?~4Wrf^ne27*RQ<=^6=E3NiU%aSk3-5ht^-Yq%etvc6!iz6hloXBC_+;PSI-=7w z;;=;ev}v>j;u%>}0<>Fmg5A%Xi(g9-J-bljnkO4Ypour^Ef#DjqD+*octMVZvl7^P z*?gpwFM5*)?qfM3{?mY2@JR#sFsJ&Vm&O$Qi2FrgV(=J=2km<(vi0o5H=((UVTMM# zy1-YV7?)O*axjB3Q%?XDaM?aBztIM21>TfCr0Gn?^3ODt_wPP(?PNukGsNfU{fhXr z$1|Usedaare~sG3zXxC+YIR3n@Sha)T$s9cGibBn`KVyjOPU1Ec)zas#6+&UYWcFk-1%o1U90KC{+3c-x_x(491 zUWL{gR9O4Q`dKOv>?B4k*(Zh?f-BfO*E&`BO30AZ5xc!I;hzX?{(8qyN7q6${mrJ$ zXIlbUZQiTYK3#-R8TS=DwM~sG_K0NRMxk;NO(H?##U%K7{a%=_N0nc_E3DNxU^Kx84C z(ZZbENG~$B1nfb1qJr;E9qzP?E{xARNIEW%kjwD}=vZ0^hH5!-hbM0Xhj&TAP@l;v z%3X9ic~9H!GvU)rzoT9jv~s_jT`4cYJ6F07onI|r{0&u?Y0w_uKyG#M)c#d4$0=xh zXoq2H>ykr+=U`l09I-iA#e8z)01jg4L4(yxO;W+`gXn3+o}yzc6y;t|!x!r)?UMN8 z%ZOtL??<=2P<~i6HlmO-&;gB&e|r>x-Gp?ex4`GQB?fZ%x4VwB{F8`r|bU6(8HF{U6rStM&76jDfF z;eWJeeB2~2@}Ms+y_w+QUAbA$#v5X@C<^1Q-FlN3ED2PQ=3px+d7x(XwhDJMIVhywkvSX!yKB#$>?=jq0E>>u;2-ZT4}Wc^L2Oz0$m#ezowoLC zqIi0Yra))x(j=qZr<=?QE4n@@)9P4xs%`sj!e2ECO?R|1@nU{`=!SQVV;E6iaR9oh zG$VJO{C|%u5{MIRjWL;t*4}_uk@(T0`)%-WtLamg)eEM9zuQ%j&XcDSo8mNJ0@LTg zZbkh3YjA!EtoiVK$i*+=yd9P9dX4BG18Ev8oB}df`V~Ff$(ZmF* z`4wy&tZn9bx;muFC$J#v0~--?Opio6H|ghqTjZPnb(7*|v=5%QCwFc<*Z;VQd5bk~ zAnkIgoAMzU-UWUPXt2&|I@<2sDw#JRcN+hC?m7P~0Yn^LFJ8dk9j$JJgvLmvm@9O= zQm`6o9}?rK^?l{}(3RV$urpcfnD_K^ukU=jkbzG;#=>?L`~TQu?PM340X#k zzSDtf6~@S2m}IZy;{(hO>$-YuuRW%dy_U9bqz|~MNMme$R6$@$Sg;i#-DFeVMI=sd z_>nowh7+C}O@h)xN&X{KcDPM8L0~~}v4xs9*yR#4z~|->aKs)QdD#iaO7ntdF$+{! z<`Z>TX`2PaazO$KsB1~1|Msv1-CIo=6qNHqR^!@E>aa#0Io|uhHnzI!)!jK?c*H?g z@1g&RtgV8RU>a!%ZbjLD=}%N<`(vn~B*?T}g7Cv0l|Nn9@`a-Wce4eP!Ds1PWM_%p zM(`9c#G!&t1?6A5<}DJHAF+0FOCsx>&RqiMWO1=)g~|u$6`Nf@TROhHcoHCtc>q)ga8DESiDs6p53m#mz&-lhw)K(%Q_nugfPTU%rpN#ZFhbdLV= zbi7ZPKv&M>()n|Xaw)ap`FwZ1ziEt9FwT+LXB>06&l5zmP6rDOIX4xi@vxd?pHC}D zf9j}yo@MwRnyPQEYZsde>k4Snkdwa`%07D#FA}$^MAivsC3SM|QRq#U4^mW793Wtg z#erAy_yoC7gSlhXzoCfvuMHHbG&kF_Qd?XH_Z9F@EpE5`Zntpt!i5557=8`uef;`_B|D8hjZY5oc$S+U(#>%x=iPmfZ zPW4^+Sc-5(v{2JrHSQ2xsvmzQz}aPhGPa4-dnVVWvHp`5uV+;kW^Z0@wi~%)67+4G zzmF6Vmc^<%Ru`AnwJ`daDBuFXTq*{8=eYeHReBO4ZWsGa_Cklkl33sBz8F_8BuT zkW!JEAn^OBIxOgH4?ySKYc>@2x5vs1^165Z1Y+a^Iu9D!$pw;Z+eL@3$lOhr)%0K&>qYowMqxBH%q+#DY6f>U#uUA zrgA4>?DBn;kD(xb)!iM!F%@t39XajlMec`;=*pFmY8wI>?e+#9qOIhYxy zz^81T;6g~>Lg*DUex#DfWusN|42raY0HFR}&$a zKYJl_$8X*9j9%oRX4)ucaa*K_aXgT46p6Ly-^`7v65RaC=m|d;n15riz+GP%Jte$b z!uwV_&O4MrHP_h6Up3yshaVuL>!Rf)dvYTC7ec?o7$W$Z5MJ|Xd2a!bZOZXgq5|D% z%=-TGZIRjp^a+*r&{0?Mw?%8^XF*QBCwi+k=S(|8m}$Yoh*h%%z{!ALhpZ2=!!yr5 zIr>hztU5FM1dyO+8dL3fpX=3ul0~F^&uvMu$15-&QB0V*3?1`Q%{8sLX4Bk?rW#^D z$nZ0L;>U@=*@CVvT}<{5!E1J#%{b~v9g$uC0`AIBuODDMr43==SZ55|Ed<;9Zg|2pcnU^3E7r#3(&2Co0?_JiD>~$cN z{nl=|j4lpe_x?UHjOwi;18Xc24)Iq3R}puv6<** z#X7Du<9E_|cx@?cWzK}aB5v2}UO58fr9|oxVs#=zpmLLVHIGryfz=Jr)Y1_-b}onl z$YHU4CuOIsDZ1TXcK-3o=$j{A^CB@qZR8HN)KsbRU!ENpg;QOUxQML#TDUL$S-8tB zRGzKa7Imn$^W8|9cO6&mKN0_3`m&rAwR4VZ2*l^$3mZiRsze=5G$vA6YxZ zE=!2m!Ry~tMM&tDh8YQcZ&3Y&`AL~`qn zV`8x>Fe>w6-)q+2_z32JijUp)T2WOVRr(79>WoBEvITIy@ArOj6zHJ~S$&vw(R^zJ z#4gN0-^s zPI_v`%y(Yj^_WUjP4E6eCUZAr98OP*g8Yh8{3yrkiCPdWz}$g;<-(M=iaVBmW@Q-v zkz;~gozQHo*$dHeh2!OMOhxX}dKC6j55=xOMrO+^Z~v`%^ox$F^qM;Ow$SDCfBqUR zN*-!NUks8*?|C;tgz+7~T~TGu)q~*o5cKL&7{rG8qlzYw|_;i zekwB7O^;|+XP!3nF!Z!E`WvBxgO<(j|gWG-zkCFvh(SmE8xaLa zw~-APEJoVAMg=E@Hpe{hRjv7PG!1`BSyPF!(E%8qR6D*C>avsS50mLo%rVZz_#gR7 zw2sJP#81CIx_W}`DPDie2b z<68piA@iEhtNRPuW;E?!6kYV^6f2r4AgqD3ik9SlIS4wnF=zVgojgkN@9te9Xn*+& z%ETTKv5z@-+&5iZ*2qTc9RBI=dZOxPB?R(tD11VqA2-t=jwkFTw*T{<7|51aP=~ol zY(Btz4uM8%t6DQu2HY68i;9g@&^g%oC``VO&gV=yG;QEC6 zxELSjl02aX!*^EUV6BlC`4Y7cABnsXFGWs+0gyBFNLQ#3QE+I&XGGno1i{(dQL}XP zsSXQF+FjltR_x)Se@iQYZwBJ%4)X}nY{h^mNMN_tKuvqOl4es@NDsS02!?TJ#{NZU*ApUu>k%Tpxx?a6t7k|9wkoQ4n}0<$CMYWu(sPWmL{Rbg!}& zK?${4yj#!NOcKr?Sg`|qILK0z+ah0@q8KJnMGhg5indV44dgT+G^CBrYV9uSKilO# zo*cu-i`jVLfA8^I#fSqvm%7+#VFYdX8QuQ^ zg4zhreYoA)wSvS}vWdCPZyn+K2A3^i8hsnvEMXiqk~TLN45teWdv%tJ z00*pGtCe3?n$LTDNuN%kt4lwVsYwe)1y(Z5 zb-?Q@G}#vknzQ*M@eNa%hYCVW-~e<-rvw1R&zDz;+1D1f54C}$W9zKPP0*M%{NHPO zsewwY|Bmh~pVOz;4h$nNOgV%OHzNJ!F4RhgLOE3i^JqW>Nn`rrWmM&@)%eeiycE6P z8}qj$Ndt9`o_4n4+0kNa8CJt$t3=Tv2mN8O@l9XL`7cg)@bY^|qBmE6ySnLKaLGbY zdH}KfH>xQIuwn$X?M0p)i(+8@Ore5PTX6vNz$KePG71i)=LFy5pVR$&ju~L!-u#3w z9$*8fEKO8f52pccn4tDL8@T#vDX>WH^}9v?9tOM^a_$+u(rsZcr_try_vWbny!Nr^ z+3)sIdnVm(daVbdELFu`xQ#>_V3bSLle8WN9{|+3grT}P0VJg&s0VF#x48TlK;Z36 zVPC%*rGP|-4gj3cboCR*_lwV<<;edWJZ|_VOb5VZYz-Kv)lQgA11maQ-_M6{6~LwJ zsyz@M&fMLMbZQTHNp{5g`YWe+%|O4MU_$Y;R( z#cL_Zf7kin6)Z$$6!~U08cRPS1TIy?Ls7px4y&Pb8AaaEI0wNeMKeuEBiOW=g& ztkx<9#6t&ByhDz_5YG>NvJu4Vn;X;{5ZOi^-)BZn(Tp+AC%`5K@NBMm{sAzwJ9yrF zDCq%sw!u~n$$fuv|NYdeh<0p|y%k+2=$Jf919B#nAgSrhA`lcUf&^h73?6FM`y~}k z#E2fromCfz*|j!b(u4cQ zzCVbVTP-UFZ!&o82N-E*otO!3=ur$Az{_Ie@pbVMq=C=<3=L7c@!MHlK?l15xMeT2 ztUTl-O;<$&r{EmZA+TEzifTm0kHrp5&INmjtJ}ZPKu>vPOX9BKD*Nst`%ooql6Jxq zZ<`Bzq_DS(<_qM%NriwXIaEnewt1TH(E?_)R6EE1Ov-*=PV38m;?Td*=)Zi|WT2I@ zf8i@^@SnK(@8SGkzQ1Py|Ka~W;s0m3{%@83%NF|IU;W=I{qIuzmmmIbmHx}({O_;+ zzg1FV@s|LVqafJQC-Id?n@D2R#w)`s8?>uzdB}pCi2%T_0)N4ce$TNFaTwbN| z)4xZ!D=zEhQsSN<)me*8DeD|gljfcHbc-UT)_=-0UtCti61$6v#JlgG?C`c-vM<4( zb*KYVZtJZ?*U0Qv$K5CxL39js)T%~T!t^iq*|#~^&Wm*iz^{#FBGN&x5%e|3kOTPa zL8=_~Zz9`{^mAAOBjOu@+vsv>^_YU!tcb_;69>=0d4OXqf=X(!p|Mc%-(!?sjRoXn zZeyByTIk1mx0fKtp`^6FJ2AeS6x+8}Ee?3K_3v>M_ObxO8`#ZyfCkg&CP3p|c%%3% z*+7Rl#vAxZd5|uu2(AV#T=&tURG?mR`1!0+EIS%sj-@OJZDp(x8`xINFtZ!0B7uxV ziv0HxFAY$!fx+Hcud44@%Yj=tphf^X99I<^j_Bb>1BaHA1<*3wRuS3fs{=K$#pd)Y zyfCJJPTkUqZt_=Zvag!KHFRgqeP~4{JqOvhf3iDxU3=(4>yb{ZrPK06NF)b&;$^$Hz@Y#Y(XI1rHWq%_883&MKC}q;hnyHY- z8q8sYPZM5PkmU?0L)iwPNLc{35)0Pv#!9R^0k7-;crS{abqlnP8an+VYB+uY2dj?;@{H|fo7X!L?kATQK^t5e9 ztK8WOJIhUH6L`EPLyt#g7yE%v+=5rQVbk!@vdneOgO#00y zQQaXV?s$*kpo*Jl*@JMpfP#CxAgzX+1!Y|F%b=6FK56Y_3{t&m4~BJO^85k-bBaw@ zeikqm$wO9;3rL>E*~PG*cu+!gpS_a!#y*+ zy#$t))WlkJ#mk@_Xb&20G4DN=c~6wv3`0DML*6nox`j%3E+Mi=gst(rku7Dyt={F= zcPzf=kw(Y4Cej@S_u(|6#^#cwz!p2(g(cH<>|H#uz{=4+$YvyKSH8y&+iZLH@g4M}b(W9P`Zm`as$m%NRbCBF`s`hi$dh(?f zH|B0aW@i}!y3eUtY;M{2h6lmcHGs+>r317N^nW7B`HS2zN^@h_Ez?`P$|ovv-u!d# z^1h!DD?|V=$lA7na^+%WF>)bnG=SS-En2`GQ091fhj({>)>_DVI^D%@0E6Z++A4A> z3;Y9imvBJMveRZe_B#xusRxO7i#=KSb$cNvyK?cKTsrsRxGdM#8!&p9%FLwfd@QAv zPf(2BpzSERoRT{In_nx)vW0VVXsZZ(&>oT$Tc%3jHF{?IoMM=3F511ezX4{?_%mQ& zP0P>t;WB@vpV|~id?W0GSzE!)e$(~IkVosd4+yXB0~_D8jI0&izqWb9q027KXT#!G zwg5r8v7ZU{Tt2#$;p|!KTKX?a0eOpV!FtlU(O3M^_}EKu8R=N}*)-t}ShPPP`^N?SYO1bvW9jy^ zZi^pB&lX^z=)>6T*a%8{X1dOi&1S8xh9_4BkJhPc33s7ssAqhyk#E(SBL~E36j)WR z#W6iI?Lo$rMF4>9gWY{LGi-i$3c@>Bvi;i)<69o=Enxu{&56Y;5>I%8Ys z^13#?n4J%Ba;AoVM{t;xeGzf|%mJ&Y<+&k;BN%av?R^b~ z3w;pBnl%X~(nk04(G3}Bb{}aYmGmuFBReO0cIvWY%sKMJTX$i#<*0ARUq=n6Ik2-` zL=4OCEMQ~>XarTC=WUT90CFRr7lhAAFBT5312X=*mLZMEqVUt6Lp zp8D|p$0oYxR`#1S!E#Pd{|~JSSK+FKyBuJ*KPcZOJI7e}-iY@7`>6BH#ucu7ATZ-~ zSz&pWCv$oUHK@!x8=|%aKD~%VOb1GG+C1pS%{{Nl%NUPlGH8h2>oxz6x)Ki=S4+DlaxIhW1g{#&7);NyTaoa9?S; z3T^9zv1{;N$&Cgs2?_MzoMIgTqih>p@w&2RgH)rn0OHixL3kvT1yS;{c%)*h7x_wh)mg$=XSs_}6i^;`K=+CRn3j}~$ zg$`8;5iM{pYQLkz>?mUHoaghlh-a?xNuG(SOH0N4958#s$Cb@S$k*+c_iM2z=OvpD zGww7W>ZZOy;oWMdZ8!66Pp*snw`0$mOKYeNIR>C3X#A`%j|kg+U%2&LAwk}vr{bu; z+sj|0Z>>Ld@|LT<-)(f+hEpSUNJsN%NzvDn3LYFL(S(xw! zYKyC8Cue<#ESA9O!YuU3FFBqIJDNuT6TC;2xmv?vXTZ~<;oiG zHQbL4(4So4yp?s6KXz)n`|Ujx*4&nSDP!`c(tIo~^7>QKpM<+BqLl|38)O=ZaP0Nw zl3?lts=FmPvzYz)JN*YsiV$&RCW$t{;;Cc0lJTUM*p5rKPhq=vl*W7^Y!bs3SSR{J zvPvdHur$8;%?Js88kf~oo7$Rqbwfep%(8# z)c<~bcME2b@V%u8jhIdP7T<)^LUq8@-PFoNL(Mti6VCLBbVZ*3S#K0Q>8SVj;)=nJ z{t3!&o}p98x|74HPJl5zsr9w*&r^S%O5@5x%bs5@0{j*9e*r?e=a@|Hyi}j0+5L;{ zd7KOB8P}(gVY2C~thppAW&dH_B$Ce$8}MN~^;fv*J}_#-J21zUrFs^wob5GK-4|0b zVKs?%mg%+ny%WYtNcp^{pI7)vaoy+aF=|Sxuy?%AbE>PhFbd`XmD=|0)b{cqcjN`i znzZ&M$fxmk1{8cgwRxiH`W@$(Ri(`qn}a-8p5+NmrEnry#IoOIf#2gBEZH#* zGJz^usy#mcBIHg^OROeW`kEz(=t9o;rfdn*Bz#j9W&o@H&F zIN0oJ51qsI$0Ym zb2)h#P`k~T#vfMnVEZ_0e>nV!CPfo`0gklnx3s+sw^cpUu1{pVH_zX)CFqXD<=n>DbQbsROZnkQP0+ zHIUz6*_}Zj!YqadJNt-qkIZC07ae<`D2e}-Gh6Awl#uMP5&SJfl6a`n1tmQ2icn*2 zAr6TiLG}6*FH21(jMhb-jOeq-7u=U3TM9;%(6-yaXwcdbE?v-DRc%}I{hd5t5cI{G z;8Yb1Eb%G2oN3gr$4jdBB4B^~7n&FfgC{#Z3@x%_UIjX{m-eUh3HfacOJn zf8Vfv6QUv|QB?dy3j(VQdAp;UYQjs{3}AFsy95hZo0+xEf?Uf}Jc&wH|BB5Z0+q0y z#XAn^v9B#ndzFfo=8m2WG}s0e^Up4B6gq`usoXRzXDMwJLwZYf6&~f1cLZo9D*vhU zg42mu%0wv?>aS+c{T|{MI}_bbpOb z2TA&m*tibOB&!9D(n@5DJ}|)|nZF7mvR6P5m`I2TZk5u%YB0SKo5Br1NviyN)T?;x zbW{naoQ{y+%j*@Y2-#@nnWgYQ+If8Q$)lxtZF3QLYD2dm&wRYo8VFv;UWNO&I`l?y zd!DKWdOTmlJfnB75lO@_xS`=C>u+jHvWlk$KF3+jd&1%ys6}jRMWoKJn!reu(5?G< zD*8paQRXKU^A8T2H(I90%NCQDL4bHB47inI!7g+W89{_C=!# zOLT{Lqm=XZ!X(k))(2<6y@E9QDAha$W2^mZp){%Ej?A4MTgC50n?oa?k#hYo`5?M2 z%8t2Tq^+=IFQh`&S+|8S{m^}D&Re%4F&V%WOwkMDKXmL3D%X(uu@N1vfz8r6`BNrn z-IiTTvZ(CcS?Eix*@_4{5{1)S6G{fxwRZ`7=Q{Lu(G+=S44YS=N+szg$6{3$K7dU? z-gtR>jE5rVWmlx7T4x>-??o&6cJDWbS=#k+K?LVfIHOjS^N~oB!wOnwW5%}Z?i{nn zIE4dH4CGCcUs2^X*lmQ1(XpTC-TwBfC>;b9Ij6;oDiT=Zm*tp;zp5-R18_|&{>KX- zP!}z*xf0Sd$gmzu(3$qrenDZJE#&QU2P2s^Q8<0k%m-ovpN^todIAp(>27=5>`66A zU#ING+JbwgGigUfeZS*HEd%bG8dfH>WXy2Va(ZVZn}_geo3Arz^Et*wptJIunr=X{ zQZ$!NgodbpPUM8WdH!vo(u>)&C)4sEC~MWX@-d(4m*FN^r7|-rtfRgNAr-<$2I?6s zg562emed_F`Uc>yoZd1fsZ9o%6Z-fM^LdhMZ*b2Kmls79zwd$MbNJwuukbbWEZMDx zKPk;p{Jd+%zYp}N+=IK{g9POgYdu)c^j$lt7vZ~C{$FveTI#6la6QtHh-P+!JZo?Oc!=<%{JcE@%4a$ zTc5Gr0K$F7G_*Z5SlLIF8LiG*Ck*m#Uaj)vts#(lp(h{#2}cuUjiYq>#vba1-u^jh zdN?FSw7}*rUGn~`d^!4K;y^p9RhX^IIX$(Y%ud>;l>4dt2>FBl{@d6+qF)LksLH3< zMxa;>lngLWRO&J*zz&ia;vLf0tt|1mNpOQXnz65R2UFKwlU6Bjw*pea{? zOhma1_@)y+l#cKEBvjp^u&syr=LYDHzW3y#k#5wx4;VpDV|o(WfMrJ2nMOQOfAC8c zR+se3yyy)85nMg4nd%M}7f~mAf5_2qyC$hOTTw(uvT*#Yo7>eOPrh=^WMs3R`-8+$ zD+3N~Ki*;fR|AmlYHBfgojpaA`K^hY;-ab3}fF2sJaUR9)y%Q zz2{+Bw7jKbUd#XPA%F(&wwuL@osx=a?s`Do>m`&+d+ZWLa{9>x!msrh^)k)r8N2(8 z4Kuc8ufn;*zbpDE8;&}_4jIf=6ahke=$+?L#@-u8@rp-iB|mtn@O@7WE?bG^z75at zAE(sw&YE!v)f$>IoQC6_H0r1suL6~pzV_WA0R6cH}9cu7mkErTfAq^;-^vAU|l#@XX z-lxvg6Qxj7)4X==kuBj z>>c292y<-iP|uGA98SEgD!30E8calH%kb0F?YmBlmi*~~Yz%!xz=$@_8*!tHDZld} zE6_!MR?SnbaD!}&I}SE$S&ajw&sJXSVHR7$Oj^)YFoBr{9x@=lkP>%Rh_}eWM`$}+ zB{$65H$4pLv91W`p3b7HP3l$TKGMKiJTna!b__Ez(nM2u&u6&AETuRt+qwPD5OV2;JUwn| zxkCDRI#Odw56Imkki=d?@3}iQtx!|%)Xq_0;xVV}N>!pd%6&5WXs73QN@`HqCw(7(V;1#=HXEpBtxcGs^-q&l1q>u(zB205!j!Ln@D9&iE6gNNdar(A4UrG zJ6w{s-kEwgp4VYs?(04$_K70us=#LFhV+9i)T_O27H4B7^E^cM(KvHU5ks2Z)5zdZ z%V4{BSJ3nhuFd;Vls3aa5EHbJs;QuIh{4kL-ejk}(u+`tI3oJVhI&K}?wVjW=v=h@ zeb1wt>X=ezHg(GJ+UXp~@1-XPPwB$`rm|Q(-hW-8<%+vDyRuMp^{z67aD$e^3EmjD z$3(MsU;A5YmH!pMrTl`fB39kr^89o|;t`LeGH|wcm8jJMW^es~#(?utv*Tav{PMUq z`8)~9>ZiMitWhVs-3~bMW^%Hi4E}y7Z+5fT{!t&K4X2N8p(z1fQ>$C+#X7Uj`nQvP zu*0y3PWSj;kU2DxxV>`c7<=6@IeRO)(8kugbhG!jvrp!L3C>iD`BMHR{T^imyMfT_ zEI|FJ>VxDBLw<2j_iMZ#cr?pRKDW%Hzb#Q4Yt8nVr>?hw;j!uEy?EXT#&x}+c|bml z%-}aR;yqmT4kfTuLMpL4-PA*tfJUUDpsUB&Fl!4>jCqr6@ZqRg2;lL_q&pZqUhOI8 z^Fb}@ivbD;I~n`5B`IXoNpX5m-~^4GdI78Rkbt5hhRpLhLo{;LR$RM9Fblqck!_sY z(u3b4v3ZnZ4T0(I2Sn<&ZwdJ(l{=X5C#@Cuqi#JmI>35wKF&0k{J2i3=cJlFL@(ZX z{3uO4YWKwpk3CF}wZsK%Cc9Sqiv&OO_42YI1emv%9ez^BSS5T15AYSK6(fGi>L5;K zkhPIWH!WvzUGp>9r(*2QBvAuDuI7{RZiSJQ=5?lg2{67|0&mG>C!ZweHT>ycfhpcx z?C(b?bq?y6XAOdFDoEZMU4njyaEXqQk;sq&Q(vs!Q$^~dG1*AwAg!TNa{C}3xu>B- zt&FHfg!ZCP(Pam~w(2x;4APgrmnB`9;;j*~Fx#FvLc5?N?)l;y6Tw^V-{oLH$Iq8Y z+HC>kN!#i`JDXk~a@d$4zcCe>5pd!T-OT{v_~Z!lf#M_VFbo;NvVf9i|4IINTi)mW zGt1ncTtB7y# z@+9Ys3P2p=$15gZa;oi*;T2Zt#_6eF(CDDjNnYMt9KH%W8-oJWH%Cfcp9l%rIoc62 z_Lho{9ckY@VUg~JlqIvI4?mi%Kv%tdcoXRR##KdgUHIF8U0IERr`!p2Sf#M%bW2Xx}jk3(`ztH-&*AwxUnjG!&Jf68nJ|T<2yU? z@Wx$f(e7HN;O*Jc2G9ED!u=#NS2f|Q`nVP#^=nq!X9|fgR4|JrVJ{vq#>#*s2CqLW z3S;pOnoJgBEMhz*%;PvEH8H7l9=|o4L5|!ke)~KlJ!8cslN7%2_2crfZ$oWF;EnfD zLU52}u#!>-u9@xMh+aj{oVd0qGb_-yd9pQ@#TG86S@@vXh!yi;Q;q2{Cv}W0wU-41 zVrS&D&$OIG=TYk@khfdU&HA+7d0FpQddfgtpk$Lgr#9jFaa&Q$V?o-e(>3~K3{_+P zRT=%-C&VvkEi?LxXs#C2(J~;813mArfMK)H@=}uC#*VH828dj<2xIsKqQU=KCf7_{?N2q1?F1Hek-_dUuQUn0}e0Yb7Ui}^CVt5O{o^m zB5L3Tg$0bh#O@b5oK#y=AX(}2d=g#Cq|_<)An#^uZ9La%M!)!k7QSE26mvI19& z=i#|gHu#Mta^Dw3FN*sS%|QPpUfnRz%;sfm*|e8x5r5VulUr0dCkiJQUV4tqt=|l5 zFjwx|InNpLfD%)5gMIX7?C89R>!$e zDBxl}??N{JD*mtV>)V|j(b#89dsG?_(+81RB^f+^NQ&I(0l`fRIO|gT=4P7iZm(b# zoJq1F3zTnbhjE<7;efn6u>g6}Q1XaxtrM`pQuN*9Y4l!&c)XicB?=Do!jI+!OMq6q zwf8d{^mXfP)!PiHSaEXhtt|su8TU)I`oj*)&i8Lm&3zg9`v)-Rr~?g!xYNs7i=e1;eAUp=ZzwSa#7DPk#smY$lVF1TVh zF$7NMYR}@o&Jh>#=@Na)CKrYfLyq=le+{c!nf&< zX}SxFFbYIE*fps8OqXVnO-eLD8ia~fw%zB zV5WRT+KB3zClSV-#W*~9cSEP9VQ;#FTeNb!y|2RPs1oh znI{c*q7X7Ed$!2{61j2zi@2=SyE=+oQi2c|SpoaY0ABT9bQ(k>y%v%l~5!u%~0?i^P8 z_obQ zxPCs$meO|PaW?EsTv%4Yt{s2Zl^s5eA1+nuG>skOeJ_e)H}7GRCSZp-GMn$2)J)fF zz9y3;B#9zRioSLR9mN>-R+6^5XCmE_W>M`c5%?HPGE}|H=Q5{VJWP7F>q$H(e5!WZ zG_au9tS6P`ShAEt)4Jx-hqlBR!~A8ol!jFn9iZR*>}_Pr@5&=5j@PegN@Y`$oJvSd z(kT;RW4FKSMO41g*clu_v#uvMyo{V*NE@G^8^%#3-k2`jv3d<+q&IvWJI4RcCi01U zy^@J^s6?=ht9A9I;86HmWb$?56nNG9HVH~*?#gk*2dvr*ElVEYxR~t(B)%jwe4}*3 zwSyZRxcfu*Dd&FuS`3}Tz^Ox0a2_QD(VMn5@7BcFshj!wZqB!ZgZ3#yThu4m84wll zX)yE2iU?NH*nYY`-!;Ay-pBvk<`yq9Z>zT~e+yB8)l=|Jli9gOT~AJWQQL7FpqqZW zN(B(o&|1iokX!ztr|j_S=C!K6xU1T_mS>(#D}Q+|f4Y!&l3cCgms7yC+X(^)hR;e; zvdEkPh&l?W%D4m8G)V({k?$Xi_ib(b`Z#|!K?I=;keX`Y8A3KsLQOIH_x`-wK<*yG zqD{v3o#Af@b|q8>oS3t)cG5%+Z^|=qW;aMX#Ej6UxsFJX)DO=!f>~RV^Jn&s z=xtHYI~`@_P7TY+VWjZ5WAaY~PmG;oShx9Tl4F473F?~2kVCD^<6%~I?}#$4-o&;> z(c!#oy=9Ypc^VBJ3mAvHGzW4-Gm4F5BJpP_i{xVSMMe-_l*paA@fkOMt=}QlGb?H| z@fa7wl)0W+-1jVdJpcvVp@-==^_g>F{-+`zU+a*!#ARLI09nKQyo}X6BeG!#(5GqOc>ir1+(LW^t88c z-b%_oL2>C0Qg1HjzX5c(iTd7pQe|Y8L*iF^6VAKm@ju!oGNH@U6KIeHRs(Ilc?1bX z*86HKA2`S-`UH8`^=ZOynt(PrWS{b=hcVfy0zwcFC9jiQ%5}8!qIbtslY{82DRZ9 zoEP7^QJI6Q{F}%O)qdya18=~@#9QIu3qEY+o7&OW6v4iyt2lo3hfL_J@-z_RQF#{0 zBxgHG(WlB_PJEvrg$5UYOd{1q=Gn8H%YW=tjZ%6Zaz8DMGM7wXl9=0Zy6qOU`t`OA z(@W{mZth-K)I0bV7ztV9zSfXT)pV46HCYXkU&*LYFvP-L^DJ&p5M8B#wRvUgYv=U# z6K;sT26TN1u)T7WtBb=-lRjjDK?|#IJK9r}2QcisX@ymm>ifi72V}oz zMp1iQjMy_=sI`^M#?vdt*(M414fbj4xj4nIBI~qEv)t+x-xJB2%yoRyRtajk)vM<| znZ7gE{(U=2yu_WTB=4g;iq6b^vL}2t%KyW&%R<&xhTEsVuLoQZuoP^Zhhksp&}gKb z%nM%6UT=WP!Ws)+!n0n91|jIhYT(WpiV-`V;d+ z$dGZiPaT*;HY*+NnDwa#Z+duQr78t06LB?bk@MM0Tk&B`MT&36=LwrXMMD+FW!jmaLLd^n z>yq-d$wT6mboYu&`$@_Y(0228z$Lf{Re}mSiGHo&Wp@T3nOq zQurcFP#j?=sk6zMlLgFPV}Baa$lz0VO%Rm_^W#H@oqP!0A#X@n&OP17zUAIT`VB*~w8n*rj3+H7uf zb$x}$W$4Bg*7zX{n4HaA$J`kYi<%E-ixYkIPo)I2DtjeM%WJ0egCp7>Ebc^BydT5J zBb~OQ(airtucmdn3!a^b`4Il`q;RK*f_ug+oD7qUWX2;g8B3WFdPc&1kH3MQ#Qen( zS;&vO%ksMj`y<InKHR_1r~e|dm?jdg^JGXxJ27qpBfzNw~VW+I3(K2 z$?~KU9Iku@w1o1tb}D=CS3qqMi4a};3mBoOLqA?mai#-#4I|7qc2o4%W0(9{sV|@> z#U%CIqx9+jWA814s&3zQ{~JUF1p(>q?rv1NQ@Tq)LAs5hdU zy|3l{d-lwJvG<<;|K*-%#&MKcthK&%#d#jb=U7l(?;W=Mb+h{5gC9#E1P`T8B6$C& zXXVa-bK~W=Zz#-npxh&4M1%&~mkL zc2J%jT=%WB5ZDLC@v;$EVEpAn3D6y8wd@=Z()RKv#@}@>VodJMY25J!+HR4t3jl1_ z;i9QQdh=OU&All;~r9H1eat-wQbf3VXkh zGPDnBc7c#pH;a}9sN>e4fSQW(Xu!qvvR{%z+(%UMa=QYd(kkT{0)A{#l%uBIg?FLQ z0({qbWV5F9mGNi_qYahwu!jhld|w|=zMd~i7*wQ)M$3h6vJ8&h`vNuGicr|8y3sFNOM z0ph4~S;a`G^XvP6ui>6z=N00p7dv^JBKQ!xAdWA0Wov}Zk`uQyJ9-)LN>}r{ep@+L zz;dllSkg2G7{$|Jf1W$j<9p!IUC2nQj0upFoCGO-IHRFIs4n5fjQ+7&)0~YOnl5t( z;~Ocbm+!m$z!vaBXO#R>bHB)by8ZCEHuS~}R|8(>P1&X)$ZXrzNd^?$vu&}5dh7nf z?>P+HwN?+oK=ZpzUkzDTEQ)vYVnWj^z|uQIRj%Ql3Q4@ik0OGQo4L z&4M5FJ_-!&#jlR0#zd6h$pu->Ddz71&5mCFbaC@s2q?( zc|Orq>r?OMiw8CUEH-MsLoD^!@3(7)B68T-nc;fi7so8nffO}JxS`v5GpH4CZ|LsS zW>479n-L`CmlV7Tp&2y-*cR!OG|+#4ynSZ!SlMt8On`u{S{LN@xt2$>GEKPc`2*Ny zMsT(_MtOkkA_RqmMSAr{9{wEjfalnB(Y&uLoQRz^w?wk?>yP{}SO~4>rCU$bhft4u zV5z88Q3q9w1oqMFwqV(Vt`En}^H<7xoC=zrqUTxY0j}L26%#Zxlgz+`tbX=EZy9cC zMPPys9SEfo*PETUfc<~MmPd|v&qnI}WBr`%`nwj}J6GMCB>6UAR6rSXv+4g+<7ny| zOTqW;tyC*OWZ&kXSi_^b7`}^5OHy}wVU$S1*kSR|c(Td?h)bD& zls@mmU~>YH@`E8)_MeW*z4rqSc(w3axqg`tQA67rnjP|Guq8PQ&Eq-BCtb1-%@uG7 z#eWc>-z3MRJA|QC4)m6AK3m%0VQ%UK{j8W*+xKJI12B<~G;SXe2FG>a(jf!xjjFRq zjpOr$?=S2P8XN#~LiP#2hfd;8xQu*2Jp9(~U!v-$7%gjZzJaU4*=yH%!13yY1u+Mk zR+fUNoffSzKXa3`j~3fsqQN-Kv^aq?n8lDk!!e})gZ}d%$C{@xNNmYfG!X1?WyZk6u;zlq}omZIKJ%I2cXpsQ+<|dmy5w z;Q9KK-{h5+OrbjXuf76gW7Xf0=Kw)A#lfB_IR>EhTf)L8pf{#!dYr=zO_g%!hBgfN z#ZO7JG#6H5uPealXzQR{w)l=hDcZbgQoFAQ?6?8nmHPSl#ksg>=-T`m<3(KL;dYiCO_|bXq0WHMAW=Qg0R%vjj196oT7! zvO`}0i;t0n!7A$Rvj03zUe04T>o0bmYzZ2B+Jzh1n@^dUb64P~HC67S)l*mA^(#0X zHks~ab&j? zHKZL*g{}H!Fs!Tkjce{JuZ!voG;heIMsy*T;%^%Dt#9#HX8V#?!bEVc>@h{E`EY(9sUKVV7Thp$ujh|(`!qExPZ3{y!TgfP>X2g+@L*vJ@7C>eg*oNP4+0b znh}0EWAaPZ{G`*6*-RGp0XbwNc#c88p0tgtSa0IlXM2B-KtfV>uJJWRO5SQ)5>3o>M%8FqPrQXDHYXu;<`#>kt`eq`YVV^FA0p>l$)adHQ^a5zI72Q|`ZR z^ScJT2Q^%P*t9ZQ_P+el<;0El9A##>^SD@-RsH2@kMF33Q0V;~=Vutn)v%?}sUniHsqArz91jcVkB4 zMkK~cpGY_mM;GC)NFcKvT$KJ&$F(a~*bGSB(#8Di%`H%Ep8k7c>lIyy32!9pH4HdH zxpGpQeA+)N8QlRVpP^QC(dP6>TIfj_*5Czr8s%Hwo%PW+ro0SH*l7SggHU0ZB1ra8 zenBfHxyMtMX8WL2NS&^ePSm?r%|h9WAR#GU5Fa7&s-)K`!F_8znlr)>K30;Z^i0C<)<1Lr&$EHB>l zO616;>uo)N-G%LbCA3FK$F}A%MWrfDOsLZ2`E%>J_hNwz*%|pt{R7N$Ss68lUZJPB zEmRq_zm;qq>yZUddPg}~_XtCuAwO3V-6*mHIS>JP z*S3G|Ht1fVyeK)(fd-hmoPDFHf07Kk<@)z~w&phoniLs+#p^s#j zgG?94G;wyo(>{!Ksi$Rhn@3XMS6Ep{c%2VBqwX0HsvM^~fN&an%ip2D0|3(E>Tm86 zz3p&&CkJ%)jCOWF;}9Y_EDuL4?Re(@)1lHA`c1{8=;(=OYmXmJl)Bh$$N1Lyiwy}v zJ@OCTP?6WMt9$>UO5J<~f;$ejD}6EZ73{Z!QJd&3*e|$x=WqPXpk#2;OQpr`>o)i< zCh1GU{gOo0_^)%x)SwVr+V!oGd%(Xh*x5}RGAcO0=lXc^vA*&#jcC>DB;}IJ3wnE^ zf<^8PA0RAzj%oBHXa-i=e>jilyz_T56|W7yh%WQ8EPH2oZ2!~4#8)h? z92CL!RQ&u7!u%~YLpGHN@V^F!bv-jGlc%OjO(!`1igbSD)+R(Ro%F$xSW@sAbRPRZ zlMiiOP1XnmJ@wmL__yKfCO{1I+G)1PIG#Tg&fgSO^b`UmDFegkp_UZh1&7J?UaN<` zg+vaBhIg!t0sDq&F4Z;8VQ+&juYvnoKo;s?2oGZwo$#E<8U!nwO|Jm8OEPq?V4*x- zaNV7VsFAkGfn*c}AUUstQjY!4h|vVkKp1Nk)pcQRYeFM8(^oU~lG~X8Fz%%wn<*Jd zY%%0`;{RcQ(IPNc7m`2B8P(0vGZi!l<#0-(pCNP$v1Ws*( zl#MKdnsD^2#`(hgd9%pyrVG$Gz!FR1a3K5r51H0KYmD;>r0bambN8GJ(hM*_UG-Sr z4e-h6M#`9Z4;;l-`EM}I1Z=|=+<1%qO5$>eJC)4bLuKo-HvuXwazJ!g{EGeox#ztc zezP*$r-{%fz#^}6Vo#9HzD4%Om|KkBH1D-db}7_&#OlM`7TevogxiKYO|$Ze9xoRM z>7PDRDJ$5iD!?UMu#-PK^5dHMlDT+HsWU2c3gJnYcEJ-QeIhEIH@u1{$~wei>=@J; zED-DTD|F!fRy9#TF$;c#p~R0*Vv2>>=+Q=7KK2mC7A9gF$QxoCbQTX<3Iua5;;iOM z0-JGbu%5w=)`BMUzC;JAfsf8%^k*$L*m3!Ef2N=35X!YVSeLh4JiDWFeVB6CocC-{ zt}-h@HN!LyIS-zIPW>wEk<;3;uj7fR#S(a8xd2y3tvB_bTxuV6Vcov!f6KpW`nv7| z?StuOaRDs}1{@~%iY8&BB)ja%(~QBWy9^+P#69eRusBC*<27x1tnUL^6Nocf7wUC2JhZF=w$my z+xZ`*RQ;_!IN8*~<9CVTUnLd2AP77Q*(~8az zn5`*BRP%nX6Yx8a=V8@AG!Ra@_xrN$d`ZH$G9X04JI!48bLf>)DK%m^MyBUg#p5I3 zI`@MyWu;yc17xb+j;@cr%RPlN6VGx#xCQn&ADl#;4IXd+BO(pUV)6?QqVI(UQ0W!wIhA zo;Xbar}XL^t31_lrF!y8^_dp5Ee{F zXq$7x-^T~E%2|&rk$P62v3K<#B(n)9IA%`|ls=pRCMxVugsMw>Q)>kK{rA=S$y1x_ zV8<+2|G{OuvyXD&Nf=idyAlgfn*b_EpwRjkbZb0wmSaD%Es^0vXpWXiuI-SKB$zhCVNX76=Y;cbSeZ+=O`3BLz4a<* z6w*v0QhNNFM-1s51wuH{=10L#{O1uqLkQd9SZ|E>!>Nh_u8`eF?L4=q?b&4aRO-$c$uQwz<$QMFJAd7W~mHQ}*BkEJzZhCnbo1IM;890jkzl> z)39IOCd$_a@RcM;qQ8rj8$may+ZIfz7}b#3>HY9dtlae5rn4hg`$ri<7oZBTzZ{d4^ho==9l11)-pa{zK@0;CPsbK5aVN z_2P~@M6xR!Y&#?mz<57GA6V!uO6;bMvX9or6_Idda&uUZ&V=MWZx!^Bjx8&vhCZ~v z)q;v|%RstMl3X^^_NsVCm`%s+Ub?_{Mng-W4JT^}*v6&M_dUnpAeCyO{Jl{nM4iB8 z+OHUz1do5hTc%<{+W{YF%^y<)gW4>q+iqS$Qxj4%?P5M>fqy%QUjiiPVhUouL90%z z`ZQCt#+8SmwZb(m1%bAVFP?xMvg$hH;{7i;g@G!{qTe>VD25~@^E>B0^|;0rlWV#g zu%k4>87qJ=y{#m;n=3d`$o%&g5HS0>H+cPm><1uC7B@o9!5n%em>~rIFwg_tzHQA= zd@FAe((Z-SMDN=(3L6(F19ts4Zl3@gMX2q&Pix(-`~&>3+C1|K$HWjTv2L5jGk4lQ zJdJDyn8m@XahgWS^)(nU(o-EY_O9rhMGQ&iqv9a{iHbe~Bw&}yl@uIS9Pjrf6qlgm z;taQ=Yex@pecPtTG(fPx{tpeN4fAW-&$-76Rb~d%8cOO&O^m)h!?5|egK#^q9Jp#0 z?}9&6)s?SDb@N~F4dDNj;4~`M&WD);zMsjwFW<2sNH{Ij11nsTh&wo}(2*)@GrnM* zGDYc~SCqVS8{p@y{v}=jxpmbNz>GKz{6Zb0jM+>uT7jcD;N+B15orR6 zOnN`oAtMeLG|g>AKv*5DjtaX>#Hv;0{L}nWyLYQ>sAP;p&x6pD8*H2y12cOGza=Zo71f zOQ5ZE0z+zeU=MfMFf0?oOquOOqD6h%)NuKFu(KznMU1zdCp^Fp=_=5Bmi6c5!5nAhs*sbG30*& zk|VOoCR`{kbWP*A;mTk^;P@9l2$0T;cc+oQsIjq%ybHPVKzQ&koGL%;^$Jqkz!sQ! zV#)?t#WK;)v}T3vsz-R8;9S>vImn&ECb{0m5mWi!Yy27$(U!omh^h(OnHmmPf;%t8 zsY6Pds=q=s0wo?(GjmlH06x8B=KkL|0x=RF;mg3#l2O5ZW;(i{rcaW5gV7ZAM3--B zG5iz|b+B%Su|5d$+pR#2<3!*<1F4DC*J1<~A-G3hKg>DQ0FMRC zFB<$WE>j*#jdg$Cq_XfSWC^tMmjPC34;SEX3SCn{=wQd6p=S~_n1!LQ=QhhrJXQeI zpXj0+PKY_Id0AgZI+5lGpq`^GDvA=u%zh;(_w%vNfE#R_jdX_Eh*cWKO8f73Du#b2 zOCEe4LnJ{dzzL1`fuZpiG5n6fuMVxT8oN1}1u+1{DqRbIVLFvu97uoVc+8VgtiHS& zanrI*!A0cUP#Q{mD+Iq*nGZne87byCQkyiNyhd6rcuiVuy&)PNoMANz|G87W1kOvpw}{Exd2rn3 zUji-`E%*|DS)3{{pSL_{Iutvs!B^-|e2bNz@tO9|lyBrr=RH;wMg7C23d>&}mNi8t z%6-kUY^@o9)Ti?ux@jN6QOABh7~oZ8m$J3k@Gome7DeH9vEo?^m4+Y1#E6=PB?|(9 zqvc_#R@dL7aDxAQnME}~f;=94bdmOX&cA~S1O*HRCO?}}k>jWV{#Djq-98z$QQaPE z2C_>BKKp#tthV9-eh%R027acDb&LiSt;6t3jrFtpmzeCCY?e9cH_(#Go z#DCra9OwV>Pk!iU)W#{7S2KkFnN9x#F#F$qkRJm%xc2{j2>-kD^Y`cf%T>7d^|VBE zQ`VS0rC`B!lJ;lrV0nunDLMWYNmT;l*&A{RN{nH&eiBtNp%=?auxmU-rMeK?M^0caYL{X#YR_&VPRn)&IliE!_H1 z=6|sO{!h2t|9T(&_dW6d@;&k2_rd@FG4>IUBao4nw=u~Dg8%1` zjv>+RC4uoRuXRV%Xz$2n{l7+g)R50tU*D{y4Gkycjg|JN4Xtg;5x_5nCN3iAQ;fyJ zw-zX&`=FTE5kK<>XJxSfpk)V$R{!*x1BSldnoPsh|C}^lV2~P)7q`q}yZ_88#?(kk zg3oTkI|DDfO`Yf^qPWrie1zJ6u}8Ae=|6>(t;y{I4(gF1#H?L?%P`4UXl~&vWT4yt?5ew|&M2C?*E`F+buC z>*RmtzxJ2pCyis?tWM;A0DEv#AsWt)Wqnh;48cc&6Z$uR1>dOZux^PK9wXT)^`(ZA zFTb8?(ra62Rn`Z_=gZYN*K(v0Jy@?PhzMe6J|NPJ31Tacad2~~ea$MH_u*~p|QCAY6@c=!p_*DB& zlmWsIC72r9fEdqpb${8GVFNfU!2xbSqdBeF z;tjJE2ONIQ^T+}7nD9pFSJx9xkzJte8V&u0Q!WEX^i$*M+rO#n3n@AW&b_o{_x11v z(O~a|lATuo^T?HslfKOn>7n54ZK=^#>HhxYvt2;@CTMafzl|LclIqxykcYHW9v&tj zUj%#@^u3$&rGaix#0@55y*K)do(<-y`@ zK4Sl^3dO>y`DGH@0GIDpw=2ef2NGR);U4Ykgo?8)1UtAEPYJFUhm0J-oH))X~r)GkYhJw_eYb)7tP#Hk{-P{{WqE zEd85i#>M-?)q588F93uhI{>HmAZ;nu03It{+i$(LcwwM~l6xWg=jR3pW2pFem9$(MlA9>x~Wj2P_3r3m%HCcz8stBB^yeF2!xe-9iS9FSu=E*Uoat z?&yb}(oE2%*u)6KlExI^R8Kl@rts_T$7Xo!?}mLq_Rj$HgRb9NkR~=FJZ<^iKPod; zkY{iha5@2RyAbDSMJF*+>L>uGB*1l3BfL@YwsR~fax{FuZsmK4cyf87q0!xn<Zt?a4$ z?saeT3}2SJ0Bo46x)Jhe>q-V`^?+yfqOP6>u5!3x6=ZDVaDUHu;ZJr;-USi)r_%^? zpYzEX{dOeC;iqe_!(2%8B_e2#t(BZd+wP&P&8Hu4z+pReV_6B%%vQ(mH?Z^ZSf)M7(S|QJR>4Nj{g}4K!t? z$LXe`u)CG*a!>-sFuU-{!08w>-Bo8!{Ja;eHKcydRp}S%2l(=QF|pH3HZi)@Gk7X8 z(To(d3|Q5=NkyHmF}eihj%f}Li*v26G0(7{HFn=A#5>1MCws*U`~KGQbvXlFc#oK8 ze435vj4pK6{fh%&5Ax7$arE|iXuHwfjFIAv58%KgF1fPX!P6>6M2mQQY(=IpK25ov zvKFLd%6`vH%_R!j{ryyA=S}Qe3jGwVSfrwa8}5gB0f_VT2P?UN(^zM9-cq9`0r0|3 z+3zZexpINKzu$WBG=kpHr&6WWhFqeqlGf>0zx8)(%SLSkFlEqQ=bJ$Xnv*^{(}Duj z8hs{oSJ5tcRzL$gLhUJ@ZscpmxWo(f*vj9)$!F=fRf?I7{>Qn3{)D#%868yu=)JCqB(eIieEu8 zO?Z!b*M3KF@hi1W$vQiWw)@Xdg^fR@iE!>)ki4>5arl<=PGC@1V#PuA;$nzHY_a}V zIq(7$I3;d1Bpy#^na2>`DX2+W=X6F>klME+#AY*-Ts_-&ue6!XWGaVNsS?Pgrqc+@ z;*1q<1eZ|o$p5QXc}=!*R6NG~s6QdQ7@ihgU75U##)?7}U-!olOEy?Mvb zT3^|w?%&QBO*S<@wT*nziVlY!_Zx8E@xH;g2evci#D91&PfQC|d?!W}SyJeV@C+}6SA&Qh zWq9AImxnN;Czei&+nvF=kyxEY#7K^AKU}n&1=u}2Oir-UPav}IXM|--hqY9v^C`Yn zpJD12m#ki;tp>C9q2OU5KuCc`#Ft;TYv$d$t)X5e;HI0epRw3^$Xnp3kBtog$;7uj zN!29XU1hcB_q$wxx!?A1gJPEAFuE8OXqCQle675vk+GQ)XqTS5c9Z?8D9C0mn<+!N z)Q+x*&q$LGRZ#RaSXn{s7_!S%f;NrNL*I~zAtUln!DaJ37ZU4U*ZDz6*$fQE^bZ2c zx(^uO>%m+a{;&k#t^+`szosxB&FA5mPovFK}hD@N(p zA#y$Dsl!`2Wpm0E)X+0DkgSnsO&QQ&O(FJS(8MtUCdzx|(h|B|KvAE4 zb8Ub3doClBIfMolQ7Kcp>>Rc^UF_uu{Qe^>$Ol&h57y>oi-MD+aW%QhEH~h!cQD>i zqb{$UMz|?#_pIKutQL=O(IPnchoMeYe2Z=S6WW$|3qLiCk@*x1bk^qi)^{qbW1aJ^ z3h9C(QWhaRoeywTFOS)~E$$7bbApT}5yZ@7H+>7MD`_;u(0-wamU`u!8H;18Gdgtv zL=Q<;&!Z@PqCAA&Tz>7$gYE?}7)eOzxqI#zqN5x8*60vR#wB!HksU*FmYxpNJPN}%9KRa6S3y%neaWNb0pf6TUL_i{#`(t0lNQ$fg79#nz{J{ocTvL~@5AW(3tW5lcKU zgSJkcpiYhO&%S>+L4Q!uAx&|CJbl^U(a`hp=PF1@P$Hl%Y1l9M(Vq~kr7(jZh#qhJ zXDdYf^Hp+Xq+_L0lHco;%Fr3Q&I(R3{`mBSZ+SgYUwYcmJ74>sErbnYWt9{dy1Q0b zGf@e*m=8w>YBI{t{<+3zsKCrrwC?XfA$l8&St$0plRqxIvU^Jrw=s-owl`j)CMgo9 zai)b;l9kk@5>my()^2+Eor<^c<*%$i!5pNw(*>gOr}^hpfhh2x2Lp};Z&-vXyTAl7 z)n~tCs16%Qr5;SCT#m`&e>U}91UFDr?GNSGW3g#f(yrZTvLuY}jIW?X+-znl!rWk} zGqVXp^sB3SGUu_F{ga%fH_1in>lR56DbG+mnEJQwtR+7Zzw?m%m{uht@Dhzm1krjg zxBI>|f_<2($$$2RyXFSD1|&Kx`w69g@Q5pBqlNuVL{g>4R)M}W>hCJ z2~cXHNVs~Av6vJ0SX-wkBn4mRBt5S$QuC}4M@Qm^F1P;An+_V0&hi@X;g(ianLbC|ekRiz;*X*D6e<4_?ij3PSqasaPLTgCu;&SFqHI9uC5Y?Ss9f7Us zUqR*4-76(G)S69$5g2p1RquOKzaNsZ;9a7(dFY2JOJWjC>-R40q?(oO2N@#XHOQZs z9Xb&O{UeRIw(!>&e!L*Fo7#dEcMJbX8MZ0wIu=vA4nqU>(t=d9JSlzXBv%4w*r{&6 zbR_+^4m&8)E_&xA4?&Q-bM7|m-B_=LF2^yx*yw!;1#^(RPFM)HgmLE$ieln824pGZAQxsW-P zvfH9a3FyZ9Vz3p2J>L@wjGH6%iv458-b)eq+2uOB(!H;4{%7u+rAe-;rFWRWmnql2 za7AY4&YrG;y`W_vS+p^3Fs@}h+1#oHG6%}T{}%ph^7GXry)q=>2EL-^Fw zl4)}^$0XJZg}^b&qH637Hg^5#Yq!<-CaMPc$6yy&eQS78l`un-w%`rrZGU|Hpq;e} zXV)Dh;UicnhzE8`66tI;Frp_X3JQ)mQjJUE2eJR+^I7Msax;CnyEOWk=~!yE^3zQ6 z#2gV^*lYvEt2B#(-z6s{}@U; z@scKLARSA~n5~opZB7g0?m4N+<+=5VuGIByQ_*8XdG!FTneRSv~uPpG}(;ukl0vdy zxx7u?lVl-@2Y7D~xLieoyb7AmZPW5}6QCZNW*Ftmf-!UY!(t`S6U`X`)>Q8w@o?_U zL5mQe$~C{6S%SH9v*X#9c`;i%4ZK6m($yk-h9^9a-D$n!6xeL~bGuJ6Gn73d*(rkI zShb{2v$fW2bhV_s#K=}gxVr?q@nqg*ZDp~>-!gE6q?z8aC?#77_&!K)QDa7VyT}U| z_vBC3&M6xh6%1!Sy;CFgD55STaevI<-bb-dWAa>jWjbI}WIc?l zs`2g(`X`l~Mm0|-}s#G@LS_Q(zT2b%^)3={&x)J7FCSgCz2EJt?)%wWqxr>0!Q$T?O*q z@noya?FFfha;}jyuKQZz8%6)P6 zG|`Ed$qds@k^Qx5IrcMqx>g`@6Yaql}`o8r+(~sRB?iK2ed9~ z-e0vN|-9k zqUOL&xWrRzm~KfdJ8=a9aajbFw+sR-A_?$Z2+)H=AH{`tQCw&llF$_r&6uTM_cC4(uUGyO=D`EB+ z4C!Sd`f11g?kE#r56|;k#h=YeK9xw>m$N=0?RncUbRF=DqQci-UB$i%4vfl?QOAM1ipj-T%q!TOk4(y7UF`X{9Jh?J=ud}Yd&aE%n{{lYP)tMCOsL5tkuz7 zoP_L%4;DV()cdv#b}tzs%A|^q4_*q#6TC=>U)Z?T7Unr-NGt(>J;m=r1M37iNEph- zE0vhTOsti?1I+GytsBroCSMd&lGzsiLK zY+J2|%l}YGX6oLBL#gU!3=Rdd9S?vzL1*6b4~OVMB%I5OvN?Uobz*K#d<=&2;!Oh= zzgv$oX0@CPyNufnx@sHt$xop8#^*5^n;s{bZKhxGLYzLSKa@&yZ!bk8>c(Fo#Bx6( zD;eily8v%MLnOuNYkH6a@&qrw(SZs3`yTrEJFAQk1IoHDLRG7RseQjlV1K&r(S4Cz zC4_d>-Ci>KM%$a*j<5FDngxqck6|=At6EuOHxla8OALKQC__Uf=?Rh*|0dNGXnX(+ zC-g(_J{f$$#LvN?;MPjlGV*q*i`orSGq9v8d$@Xsnav_8_A^#l>$fi%sa0w-`8#Ec zEPD-YY6J1lDwYVpLDVZlu@J>;@=iZ?2PSkRO5L9>;Z@0uQz41F^@(Y9A^NyBU{ZRL zgk51P{OPT9YXFu-Ji{ZG3(KYOr<<8bpMdQ&BW;(7)_>m?r3-;~Yfq^4ROOp=ZduNF z-to{^!D!yCRO8L83KR&(^Jw6j5-#9s2LKrFrk8Dc3=pU;FA9GdtM_1U%gR&(9L z*ccw7L4boUl{+0gy7w_rUveYxGbpy?ogYy+Cp^WO`)Ov{xh}`v$z_qX*;8z)kDT>( zpuC7a9p=RC(UAiYPV43?T(MZS=zaf2MUJ%-J=1D10f9|t!-b=!5inHb9k1Gk_EP?R zTmz7LsMsac+9xAjyqtGD48&#f9klbkrtr0;R!%j_KWXr;47xk7#XV@!bM{*AKyFm@0@Yacpiqwx1p@O} zR}KZF5KHCpb9#L*pn6M){~gE&&9b+g$#6&7ybKO|J?7q{G0(M{`(zUKA<8#^`+`e&@jPossEd!wa#1PINzP*FoAuo>gEegiI+=5YC* zTL#q3=kAEnxAeQyLQ7f5vR9%vDq$#OqV3N^S+CmShd4>>Y5do%GXYTp6&KT|I2;z_ zV{=KFgb86E+->i+JOQvb?I%=Kir+t~EKHisQ|a~so=0T-(j?>y&b{J`H&;!v9o^ou zyhk%6c+qUAJ_uLHWAUfr)mtR{=ak=H9ZUNtWux_rFU)OmbQ(j}Ew9P9#Lpr)yzoQa zY^IbvW{!cv$d=UZd^r0j;%ccGH87h@@9>(I3>k(vOm@rCHho5T$r?_nMcQa#`C#BQ z!ZRiN=*g4k+`QY*6#IHp!W(GyXlK4VmF|Q#p)iffkRwH1PQs(3Y^uJsAA%gRF{mDm zmB~M{mARspDxK!L{uNTOB$^&gqi!tRx?(pppL76oV$DxJuA-vbAsX%-*ahSZ03+2{1Rie zCX7M7-_~CRGVQv(xGuK*Dt^Y=Cq2fK#9Q}o;?%52&o5)wu0gY~?Z5G>LoMphvkBN1 zQw?q>468I3t7sQO<{*_F`V*$g`3^nL7kjaVnfQ?fNBTg@>6hBP5^fJ8Hn{*^?cVPP zP@2*~t7}C=uR&hVbIfQInXcxr#}u~ttK=*4TewihI6Pc`1KI64c=0A|P|j={c(5bn z$|a!X)6F)DDou%Y9!WW$lb%vCquQFKsrmNI?gJ#!`-=lmb-=rkv2LPpFj1V9ivPW2 zjwFWcjMFMYO_SqNE{naKm&9vUg}cEY{Ks>mjd>VY;y|Bp)cvy;{? zp|z30caO4BA#aoWWT|@8zWsc7sCW@7yFp-nqo%L&?J#{T~{2eBw5#W5xM@L z_)u8|_(C@yE~SuXHcF=@>@&xYPJAgkbJRol&Mc1$?r<$AbGwJq9J-&fO^yrW4mKa? zhEXOoid*~|`68w($GhUTMeNRgh>rXa{>Bl{G$LDJOJ^%aP|XUKZ#XIOIGY#}QcYd@ zd^n33ofF$Sf;6_^<0S8Z?)*?+lt&+aTw7|s;}xB)Ab%2yvM%Q)Ja86BhZ ziZa9lP5QsOv@u->w$%+&e4!7Edu+sxF5&6WkgXK<0hKoTM+f?w{%~agX&5dHwb-Af;2ZMu_U&XMc7^2Wk7x+>(Gk$3K^zhZLW28*!uhv6AVsdQh&E;U4yNkQl4u# zH0Y89=xy4|h*4t77kka`}g+cWg{-5Nb)Wg~C<=V80Bhx4b8MCu{j= zs$4BFkP%Uw;;~5Q2nJedf6w*v&`wS}%bDPDx}U@Obb-dbGGqKGn~F;`!<|gfgEZ4t z6?!RNLq+U?tWWPt;%qPQ#dEVtn#SkcIhkHpF!@RH$#lQ(@cmMVj2EBj)nxLuMh_jHiXDe z6YJEL54@V{dw5w06Tu#Nr2He-w*L)vjooDA7SOgsIj3HX2)&?C&MDyLi$6Tf>z8B= z3437)`EHhieE3~h)kY)QJPtB-92)qOBq*+}x2Cz660cn2I`<$$m)D3r)^C#baAmQ|9<5@+L|gP3CIs*9hs>l^eB@;r(!=N+A|XNg#5mf{01fNoFC{s34UHDG~>Sj%gRO z1}C_Z34w_yGqJJ+R(`s-CpI0dym0db)_13^?8v}f(Z46AO``@UUg@Nu5ee7k6unkd zea>F9BSugC_UTrX=~tB>Q+#4KW$AP-zkAfe&KBv5W{&AGo5q>#qxsym@|+S}uNSVQ z;3bCeY+(X%;Yj9mk2x`@1u{X{3mJ5yx9`HHB%`IS5))0M^ukDS)Dsp1^oRW6Pu~i?E=N8&= ztrt9`Pj|%wH1pBT;&MWEI@2@d?91*c%!xg)VnjAv{2k~oJ<4%9B*y|!ZEF59weIEUFDQ4-@}BJ{)Q?rFw=$QN3)m? zeoU5%ejRny*AH9k+Q7#S5u+9LE*u{`R~^E5~e? znhFn0a9FicooVR(>X$w1aG(jq^>gb+);DV&+-1*iskFvLz$1RY^8r;RWiY(Mue!ve zHQb8_w@-A61xkt6YnWxnH*f4egYPYWoBV)TA76q8tVIXkG4dZ%CCZb_PyNE zgGTHXMu>g@rA76Rjb%w4C(H}wtnV=Q%>akW!AG6OR%b_qN9CoIT84JhIn2{6D_!5| zG{X&0#!%0mqwvBkKlQM)8|@%B!f*vKhs6~oMfIlnIQja5whPV&=3yx>I zzhJjg4fHuF7=}HT&OXGOyarKuDh$<}EZ;OPgp8i<`I+L|nf5Y-U|+Q+3_@W zLcYc0cp6x8)|+sf=VG5Nhp=QgSGPZA(q!-R9D}cIWd51y)z#m&;(8S%X2*aku&THv z%&_AM${;H7L;;%bNRVe(;y5->TOogfygp5y?8-+MtK@DF2;x#2x!(ICVF-66VrthgRMOTZ)OR3wapS0$m>kgpn;~~pc~9a{?14d0 zF9-MDx*AwUslSt1VYG~=#l##O-1}r8?1Z&0Ey+ab_zjEtsUt5O zMrs@9gC_DC&tLm2`7uikCEjx++ho<*K&DT$FPM_Q!u{0QTrAYH|AkNJloD80rgt!EeOSVVZo!Z(b+H4 zE_S8x%tG!DR7;o4u!lt^xcA8T$BDd|JNn{|yD5yT5YS~eX|W}bz-qHU|H z7CAF34?fe|*7%8_B6(KA2rC@#IDx#j7Gt-pVLBJ>*~+O@t6Nb2!Cx#Gxbr_MJIk;r zzpm{oBHbY+HFQWf11K;cAl=;{Qc}XuAVUtFl2XzQBHaTFNQ+{SLxaR9sq$XqegE(0 zeLpJ{~ehUeRlQe_D(sOIv7pqy-EimwxeNbbA>G-9?gG`10@>=|CU zHOKj`ikCZ-#m_6vB**#$T2BGHA1ce%fJK>P@wOU4Nc{x|+%z}s;#3k7TF8=P-vF}VM_Ru2ef$R}?@sL3*()Q;cBp;R|>ZNOi8(0;}mFcZ+oBU=o3WfG-njd8DyQaZTI+MstaU{ zmI!(t)u=QFK3)($wC^iE4ygU0Io@rrKWt=mRSYDGH(h?DU9>m%^Z?tbVN!wTNK12A zotfvA5sG9)mV_rP5vuN8f7qt5ueKle+8K0yXK zBk0@ptk0xvs73ae%~HRha{uXqS z8Ln-j16dGmR8H3P5j7sRugC{|2ioTaHS~fkk9?M#ND2MtY@%izLaQIV14T?xzOleo z_Y)ITE*&4Yl2SKMxA?h|N;17dvF9M}=~sYH^K0g?(v^ZDi^togG^&eM ze=43Tkn9gv^p@q0GDC{%jCHMAEuqqU_wjov3&8yDUvqJx z=sZ$;Y1u0N>x`H7<=OTPS8La4op<1_eyS|{kyP0)@nrL*8z5AYvm-B59aI77-mY)o zLuRs{`+u{3GF<{o^XhsiUeqq2it1N)i`1lxi56IEBoT_8anw z8%w0qS67m?aofjxR8WYI@3vnkxcqUQdGEU07*fljFMQJ*IutPHfxQuuy8M9`=Z%RG zW%AdQ#@d{sPSfR6^(=yeUTsZ`VUeW zsPn(~Z(5lQF^s>*6NBVW*q<>9*u+p8YB(6xQ~4g@j=~L6h4z$E^HzRj)QKDVMR=CT zyFE@2lQ`mr)E~!RBF3oJv?jX+BUHw2HOE4}dE$F2h4dYCmo9(c%sU4QQk+J*b%m7{ zQO%Y=Tj5zwNbzJHqU*j@Y}WSxI6#gYtM2KjQsGQ;G%Z`Z?~PoS*-P6g zX-;>|;Qu)ToYXxRr<+)HzSv}kIgtlED|m?~QT*8T@+{*waIq(%v!=oNyTQ&^c3q^1 zglcp@q=X2lnscvImwvc1K#`2gzP0QF%zfw;3|F9#`sa1moCkc)k)z6{?AQvIFd~+k zJ?hTv?G~jF!SXepihPQ|V5g5%vlz?w?{!alYS1NPhQN{u;UaPWA*ZSt_4LW#TTe)8 zOuXa}G=8iKuvAd3|(lae%6h_ZkHE2GXLZfd|uh7^p3@XrB+}%-wQ!k z%$`_Z2S|!dtytHlJ{Vie6x!1flJ zk^Un$ek*P3Zh!&0R04q+MvA^zmABe-1^%!lS(0@RWkt zKE43#neo-BL2N8G_@=@%=H+ADj_5Gqkdt!&x}MwPT->sht(Ka1j(P{UMw%wWv@;#+{BjxmCjwDyZ*vN5bd{~A(_KGl0ct#i?R&(A6-@xgd1A0QU zHerJjz6SJ6jc{q?{dD#&7CTF8G0%9503`^V_lvd{C-3hFEHP#K)1}qY2OWzBreYv%#N*8zzR_miqQu$RX4^qDBdk(wA-8p|j={)} zpu`98bX5<-%4Z);S$S-d?5H%s6*{z*N^oJT*6^4qgDG=C3SRFga}J?fubj_}fFlM< zeKq1WFRCfx9M}F2Q1&IU3(pofzwB3Ty&N`6hVi~`w!a7Z91r8da+MpEoTOHi%lRg+ z_?ji0%{mG$3^$Yag>v&MR(|VzL-vw&gQA>nFAsg>FoOK`QIdz>ek`G`o*#DTQS!WU zLi8+D-vU)Pn#02jwVUwm+*AR=8I@@W1%WVv5!o$)-8XJFDd`gV2-u_OJPeb^sL|+8q^!PJ$@^ z>Wb~x!Au&l&#ybG+&ufZfjt`voa-7!yr`}Uj8R6g>Mc{0VEiB2p!@0YkyXhr_2}wz zfUc00SQP_Cv39m1+Z}wyViD{R8P+F$f2Yys@)yw#3EpHp=^DW5S=@!3{*LyNM?R%` zmlopcF%(C#u@CILJU5T`Z!o+UC;R-Xk(ut(qy@{_;7{k1xm(mUCA8e}oJYPigSU(} zgchqBG@zD6md_vptG+dT&9(^7ZLQ#GVWaRdloFwCJ@kJ_7HUYb-L`7be$ zbSIPNFU*N}s0g=rUljjhJ?2Iv45La$l4j`b)A+0KHWq2(O!H%Akk8UHevAyE$e-+j z!6YiAx>hF1Sgwm@uu8WUV4WI1jFg~Zg5kx$r3A=RYX1Lvy ztD;bZ%`JT}j@Mh7O-jysPd$^z5@DK>UMA{#g4nh8I+PZ^%R)GJh2w98KW~VPLTtEg z^QRVEx=evbo<9HeS!c_o-u_-LUh61cUB1HkeVDC2Qo;I1C_|pVTglkw!2_NaE5Il~ zW#ILn!Mtw1R9tl6n2pyGbQGBPbJ==xlxo$FPy6u?`P4MR#`&6(kGDro_d_LhQCnyb zKO3T;^%A3Imtdkxr^N%+epe^RGkPIlDI5lym;EuLl+ex_HYBR=Nn~`*Irz4?pDRpJ z8F4%ivS@1seTMy(UNnT(@s98hF6HZgwWxaJfu&j@{7jP2$==*dY6}N2f$0d zDy@|(U8@Vx0r6};A>${*waxOYJ0wyzktZcvQm8Nt;$g{t_G9@k@XqU(C55k4ms8`d zfhrN5#MmFC!?Z53+pDKmUbDFqvhxOM>yP@bgVP>0siXc)_+MgY!Hv+o2?3oe1S z8jrT^ea>1|I{Xa!+#S0qeUrLV7~yG#lJtn;Oa`XCrdD1%>}uXP$kz=x&O-z;xqe`M zq9By#eOwz>zwUilYl&(38t3!@ay-bivV+QKOwFdDW*vZ;P!g>lUk# zm7O24X{H8C>m9QO4L2{Akqmt0aeYul5MH`CLZ*J1qxdJ^%2LfzVH$JBA$xh!j zOk;H_$jabZ@_ShcUDTY;$!zQ9e9<3Y_b(L+at4~9H-rAcqky*?`+|czJF8K5MFM z3c&pXZG(08bUFM9C0?(u1CtC*ZYcn$KwS>kn1^YiK|S2XbMOI7Kr$HZ^eicP8v15> zyq$2jvh)*o3>qO@eMlfUHLGnu1707rxY=CQjCNo*c_%(|rGxS=Oa;UrZ&MW?MR8_hLJ@Tb9%cE#YH*JXh>GA7 zYmQ5`!T9#?#r9tDi5Oycy739J>Ex1fqb1RieS-Io3KEKm5?m%i*=W};fwrGgezlGL zQzsYi?*eevpBP4Swt|8Xd2y7@^~lsVZX2sW!9?Xq0lA^OxF@0$&cq ztMHKK5Uw5X!WH-SZxSPAlP1L`4H&%JlBnjZFw@i*_|T+qd|L;&POph4_PZYg6+n~A zdfikX`PGDdMO+0T!)>{bBXe#>?_OuzQpyoaC&y!b{ho;dF2B2=ul*jubg$5T=Qe4; z!?WP0_DxUXEgEmLE$a)ps*7Y{HGtYI`=r-jUWtkS;mKMtNS#2h!%k8lI;N1?nuT=j z+H70yKESqAVdCZUKaPkPV2)9ok`m?R7i3B16G0(wzx8citCCwv5j-mM$=~Z%$MoCO zY61kHBBp{Rf97}#!z_Or0CYJ0dG04z0mg>7l8?sVMw~V!5|Jf3>3Hw1%Rct~`Dq#= zH6)OAbJ!BRDbA@~C$p0ml7_&&e%GsLn&(e7_(wp=hI`rTF_hk&d|K!4=}EwyF5cg* zE0k~0N`Hg-b6Ullc#A@6LHHz(Q;3dljfvn;@b4cAwtqNk#?U;S0OAlX9OLg`MUthh zf3DDsOVm8hIx}hp2qZ_0ZOMbWWRGY~5 zYL)p_koeZ;DgTy(ymDg{1yeT3Kw9Oqb;1Y%V|PW%sM^iHK`su+0NRu`8fs>0nUJcT z&=Dlaz7?tm7E%+7s}(|P8N@55l_Nz(K`@Py@C8QK*|wXTlypU;6KL71pqw;yE-=ih zJ(Y=cL_&eaw`DUrcs3_A3fx#jFv>_r)SmJV&9d{FLIxzeo3~Lfk)wbCadEg#EW&4T?1Zg~os7w0U}D^&nZ$vnWXffV z7p)ex%Mw+`-b)E}F4${?Uf*MTbP}NSRbHQIA{16pCb)!Al-79_PWf;;c+$CmMT-O& zj=1Ghd~EGq?g2#!r}ibMJeM{HX)79H)%bIub8&p$+4-k-QM|=oe)8~KR5I^dGW$xU?Laih0`Jzeg-#Ph|#(Kq?;900I|AG6mq@&kf`+b3b--zPYR^AD~m;I0Mjq=7*t^u!hi|HKt4lk2Zpsyk>R9;n(u9c-@nRlW{xkTLoEGsR<7?nI1bXsuXM4ts4YHHqYey zQ7)#yu^Apch}V>;EPlK9c>>TJ^J`};FUVyr111~VG#`QHNx!yH#61KcPY3CwX9`7T z$bcPwC!|#qD{=7smeBz3m6JgiYN}R~f5;S7O|jXuB~4ns+pDg5vvCAuV#dBz{V^`2 z#PdujJAMG{S{f%tyPks%$4o?(~-`nCYyB}Cf@xu zTR?nmN!44Kc;=VPL9eSzzH&B1LFC20W8vs5z}w5jbX{U+$_?4ZBk`bR@jJmN~(!J zQU}8QFy7)D?P*R|kyLF8bUNi3`Q#Sv5KFzDvC8<=#NJ?Df~3ZO)wWz=a#s}GQ%Hwo z%MRo%A>0V>HL~^st5PS9%Np%MR-wW&Fn@sI+Gp}1%JBR6rr+%iFy{+vHGDQ)XzveNjHx5gq${G_;?sK&eUEr zYm*7emSa|=e9l&8N;D8OW^&u{g}Gulq@&xsNjZwSFiy@``zYAjiu9lbW5iTT0{CvY zxwYYHOLA&!ZnUnvjOqC?rxE1Y03BUswd^GNcZm#t(m7Z$DCj$B)fi+-b;(kRRA2ik zX7|6`zxmL8wAC zX;?Ch3tkKe7iD^4@5ime?D1jrCL7B0S*N3&Bi}cH0wdZT=8t{n49gPLa0&kMM?w4D z)B1z;9HZe!Xh^m_b^?^-G6=%$Ej_IrF9PO3Vmo8e@6*6GYfrJO|DHjgQVNPQM@5FD z4!VUKTkh!Bb~yYbVAHv$5$tB0pCO}pk7TWi%s@yepFJ?8wS_b)1Uevw(0ml6S^DO3%G5s9YOz1kq^V)K^KhT>#5we+g-@mXK z@9?-ih-V($G9o9pk-|;$N&vlft7kS(<@3DCxvZ`3xeS7CutbR)`Vj3W^8XkhU@IoW zlrgH5RwT=RbtdClP?N7IcGKAo*LR;Ht5HjGTgnh;+d6#MF^OX{yqoP;ToOGV{{E?y zkN|qyL-@v;_6*2s;#eMFjBXc58g_bjYZ=>Vm!dnr z1U!KuRpoCV;W2-MAy8F3c|Cl{-!cs?67FJ0pK#hT&nIWPH^@%K0@ZwMo_bFIW~J^B zS#?a@!Dw+z31Rx_A!j+xD1vFvP;{O?*IkgOA$fNzm1S*%@DwBME%n=2Dt<}4yhYtG z9;-CJX5T`kN439kPFrM5(~-Q_2p!E){ntA6KDp!QHAgK|y|4X2A5XXf&f|-2B>>+9 zD0C<+PNzQkh5c1DcmtAhC;QmnT9fj)q<(a_7Y|F32zdR(E?gd73{*;jf77C6NV^h* zS!R7#8N2eUG<&x0(wy~sza0wi6-?6&Jm~&p+~JNkC6Bl0x7FuNV7Zq$(*fh2ZC!JS z+zOsZ8C0|RsA-}kN9UK@zCSdm)Q;nn??3a%FRbo-VD_g!Wfk z%2W%E8_B=*tz5LqY6ONaJU3jbN?rSzLyJGHblV3=s68ZX0MeJ}p9pkZV+?oe#FqA! z^o0h}VbbtK=(2nNPqw7;&Xbc)XY{!Xnil*MQWbIgTK0L|nlRcAV3u~<^Xu`+e*5>z zg8Rp4FIjm!h6s#`=t>0HE^qc$dIFNwOIORTt^ib8dlI?L^pt#Gd`jt#o8=XbG`%iK z{Hd!$gj6zjAd8E%Y%4yd@;`qPbDw0>EcW9%y%wi;e5CcxdQM6M+lG~{z%iJ3d|tMJww)`X1}?5a-3;z)Nnq!-sSm{FwMRp|M)?9H&_B zc!aAkz)mMDV)-^Cm&ii+{ti;<5V$g!c4&ehjIJ{7o`c8tneJ#J{LB?!kbEB#?wkX! zg0>s$ryx-N)i33Iwhfjlny67n=nub;5pM6>aGj23l&w&y!60wkK)+K4e(rv-jvE!t z0B8re2F$>t0QccuH`*osxxzPuK24TSTHWr6=a^zQU95xku%APVu_-kyW0!zr2*n#i z9;0|#RG5Zzf)K0)cMq@hHjHI90ILN^SIG?N;@s_wv1=PQ)#oib$#f1KPrd-kA36tv ztIwTskP-O(gl}BOtTG!s!DJ(~CnT(`$~tfmMYQVSW+0;ylAYI;cJI9`fk*yf+>El~ zeK4Z$7$)ctd20?3jF#5{7lzYn&VV(&Ut2v3vJfWj39=WZw8OH7gEHR-o*Lk%#Y1C4 z(}5fKd#35GC|u1^tZEkoRmXXEH40ci+ns9J&#d$5JBP zco^&dWK)Rq6>d(Dkd-3qutnK*V5yCZw0*P)hpwi%ZAxIxRXiHTFy9YDjaOavqJ<# z;_Jn5wClLBZ04y*%NqzaJ(ipH$4uy^6Dk;Za|yYD6F8gpdp^gYYhXW;0zD=fvSR2U z+P*Ei;%~U{^KUtJ(>HwEjApNT45GzlUzO8Qh(`?-G)LfxkfV%_36L80F>+77uq6nU zWrOrLS(IX?RSV<1pjL%2ex9=XhmnF)%1~&p!|Ji}F>*co)5>uEX#9O85q5{HRYc)k zQ-V`ogp51HDaIo__)-qIhm%`!C!E*vywAntd33@+PGk5nFKjR1F9~IEE;zRw>%Edm z!R`g}oH#U$q%js9ae4uZ-S$ln(&YSqmSuuyPDQSG~iu7^^;eKf)e@TkCk2a;= zbfgnL-Kg&FTsGGQwnw|6B-~qhf%qn3rpNC=!xKzw{q?|!(20GFxuNA(q#DELCuhYG z!&tX`-rN;VrbJ>%=sPZsUH$@Y=YD{0D%KT^KTog3?R4^aQj{M0Ye$qHeubBi$i=}x zuKDcJ`+^yVq?|-OxfpMRb8z+7dQG5^kq!}pH)V;N<>7nLcqHKfIXsi9iW{p#k`S~1 z*?`+Pz=}HF+jlLVF+5=>@-MJec3VQWyR}pVf4bGr@>s2E`$r-Rog@8Xen={uO=nJ$ zqxOdsO=B!8G6YPdb396Bdk0D?e7{n!a|=D#MLqD`%8M<(lGL6JVgG7o35;w10rzBz zWtDwpJ@DZ>mJBgr2A&*qTDL_N3FW2*8UbF?+oefLl``iQ?jeRreKgDwrg$3<@e5NL zRcLLa7R>_G7fVz80?9XMScNMl2p!zwM)p0po*FL$4by27M!MHdhP*}atX5PsJ zI6Okpx(uD7yPzLRX}1hh<(~?gR6EWn=6+cG3m+GLn%&#a&8ZxC1T+Z$R$zaUb|OIV zt@mjbxD7NWHUcW@MZTNvwzn>|yF|aDKEg{28J{21I8p>MPQZS(DZ1@KTGM&_po7Te zvz|4eRy_UKc?P`yw7%}50g1)yS6(JNz# zrmprVR%FYr_-J8_`qJ?|y<8sV-f_Z0h9Ui3oZ*;Eo>Z#4#PCkC zW**zgoM%ccC`_x+jH~Gb#Q#(fBk2?$qNRz^sC!YED>9PBF+8SBd?nkec4s1GP*!wa%wbCgpi6 z%Se+g_}t8(+B4RoGFzNd;X4}at|p1AR=uD(2a!{|>swQ9K1^8E zsDJ?ir7b8)tI-s*QwMPEBDmG)-p=L@1r?e}J01mA=V<8N33tnXQ#e$2)lkg}*I#~_ zAw4ei_-=&Z_{(#>2+cHtlJtxYH9beSG=Uv8UMZqC54v)qdDQd}^#?C@{TkN`v?L4g zLtaJiPcOF{&{J<6OL*}Kz!$U?(Ci-tS12$Cp!=XDa6K0V95p?!zi&+$k~6Z5r^ zP|3NoSLq(IfBG{7$gHv$zI6l}2WA;cXp3YIC`lBADMF`D>3xg?r=CdQw&0X6w=Cs; zb7s`@+=4TQE0`*5GV<6jg4K!+UpZRY;NFZ$x)N@{we!Sls+T7avU5fObOEZ1j{xmpEb4HYUirNs_6}F zA})vePOJtLnw_<%->A+1cH3i3E5llb$GBTWkdT%w^Qb-hnpK8FsJ3_cm9^pnGh=#% zRgc7BeU86dOkPngG1de~+n&+xCXW|Mu@!`>IDR1(D9+CiCtz(9Nm)odKI9HJz+=+&xwtW7PMY#)eXeu%hbRxy27A0#qGdjcv|xON&}> ze_81I{tCLuxK_UEa>+vT(i1fgNPIt)#ow@tr##k+>19t$DxH{JNtl#<6AjR$UO&;B z-e&(7sONYaHhSctxkjoi#Kw7Lwy#B17@~m&u78kuoB=Ngn+}|#7 zcC1Dcrv1n%Vo!WVfgX!jPgi-Tsn>8lDMiIL3t{E+rYWzm(8jQO`zct%RlNfX9U`4& z3adhb=-%@g=-E`T=BA5*x4MmdS_3Tucgxl2=M-AL{{cMhXS*aXo)Y^E{MYw$Ve32{ zlpR!hv-UClM{t|gL1K$Ow6+&{# zwzmXK_l-QY0*x%1fo5>$wWJ0>HK42H(qPqf(m+Yu%Tf3|7T7jRDG6a@SSn@O>_=Bx2? z{4JurvZdPF)x!>LtJv>k5qI=p5`A!oW^WF}S*p<3Yj)#(X~xF;zDd-E@G; z=_;;*6xzs2Yy>Dm!AO4ooiE|!IgD6%>sH5+H0&5qWHWjSFxR?%Uw=%lz8>5Hc#*qC zrEPaTR=n#0z(U6FFdc`Z_G%im07X>6WqqTo^M9r~Jn~j__%RUxLUi^9>^?Xc+*@b4 zf94n#pMUky=8Wn(?>UkK`i@HOqmu|{HS5d%{SG2#$u9sCy~6Gth@g7!#fTs$;Njri zJS~3>x2h=Svsv2M!viWV-^Ky6w%me($;*L+fF}vm9X01Q4gX?D_Y;vz?Nyehwf7vl zo4%7840}B(T6Z7Y1Hd3S07F=4M&!{na*wa##G~pz?|p{$_8_k(kVqH@F3<%9*%+dy z#MBTezW{(ap=m#|@cHMh*t0xje@Jv2&_OagsX{T=HhY61$E2uRqnQVdLnNjKMQlk@ zASh5)kk`}fq;`dPNdEfO`mZnYRSp+|M#)G6ARbD^0u1P%(dWlTLgXzg&g5VeTRO(S zjZU;@BS}_3{e$Jd-v=%O!{a1 zHWgQ+89W}RNPhz;mE#y2RH(CnG2ypd5p0Yn7Ygte(Ik;Zr5pr*6<;(d<=?fiZU_0- zca3I>Vy05vMKY zF{h7Cw!K3oJO!XJZFeSENdGflK5_j3tReFN@Glt$Ob&9t0=1{237Jm4-)zqTb8`;B z`X&^ZWU}W)mqb3TApdBaF7AIm`+qIy{|-O?kO<2Ng?Nn;`(OO;KM?Z>IuNj00j7%X z|IdeLv1`W?AC0^GCzRj*pC40Hxj+6ZUF2cO|Iden0VF!Z%e49bJrUgo+bOo69smes zbBgkN7RPr242w+gu9~(w)74CjfI*L?>Hqw!|BS^wh6wQ&-|UU~yp=+7O1wu8{n@{` xcUgJ^ULsL?7rFmgGXF%W|DFnOLx%l_rmI#|@@pfj%`Mtpr_fpr\n", + "\n", + "Last week, we developed the logistic regression model to predict that probability, but we never actually made any *classifications* for whether our prediction $y$ belongs in Class 0 or Class 1. \n", + "\n", + "$$ p = P(Y=1 | x) = \\frac{1}{1 + e^{-x^{\\top}\\theta}}$$\n", + "\n", + "A **decision rule** tells us how to interpret the output of the model to make a decision on how to classify a datapoint. We commonly make decision rules by specifying a **threshold**, $T$. If the predicted probability is greater than or equal to $T$, predict Class 1. Otherwise, predict Class 0. \n", + "\n", + "$$\\hat y = \\text{classify}(x) = \\begin{cases}\n", + " 1, & P(Y=1|x) \\ge T\\\\\n", + " 0, & \\text{otherwise }\n", + " \\end{cases}$$\n", + " \n", + "The threshold is often set to $T = 0.5$, but *not always*. We'll discuss why we might want to use other thresholds $T \\neq 0.5$ later in this lecture.\n", + "\n", + "Using our decision rule, we can define a **decision boundary** as the “line” that splits the data into classes based on its features. For logistic regression, the decision boundary is a **hyperplane** -- a linear combination of the features in $p$-dimensions -- and we can recover it from the final logistic regression model. For example, if we have a model with 2 features (2D), we have $\\theta = [\\theta_0, \\theta_1, \\theta_2]$ including the intercept term, and we can solve for the decision boundary like so: \n", + "\n", + "$$\n", + "\\begin{align}\n", + "T &= \\frac{1}{1 + e^{-(\\theta_0 + \\theta_1 * \\text{feature1} + \\theta_2 * \\text{feature2})}} \\\\\n", + "1 + e^{-(\\theta_0 + \\theta_1 \\cdot \\text{feature1} + \\theta_2 \\cdot \\text{feature2})} &= \\frac{1}{T} \\\\\n", + "e^{-(\\theta_0 + \\theta_1 \\cdot \\text{feature1} + \\theta_2 \\cdot \\text{feature2})} &= \\frac{1}{T} - 1 \\\\\n", + "\\theta_0 + \\theta_1 \\cdot \\text{feature1} + \\theta_2 \\cdot \\text{feature2} &= -\\log(\\frac{1}{T} - 1)\n", + "\\end{align} \n", + "$$\n", + "\n", + "For a model with 2 features, the decision boundary is a line in terms of its features. To make it easier to visualize, we've included an example of a 1-dimensional and a 2-dimensional decision boundary below. Notice how the decision boundary predicted by our logistic regression model perfectly separates the points into two classes. Here the color is the *true* class, rather than the model's predictions.\n", + "\n", + "

varying_threshold
\n", + "\n", + "In real life, however, that is often not the case, and we often see some overlap between points of different classes across the decision boundary. The *true* classes of the 2D data are shown below: \n", + "\n", + "
varying_threshold
\n", + "\n", + "As you can see, the decision boundary predicted by our logistic regression does not perfectly separate the two classes. There's a “muddled” region near the decision boundary where our classifier predicts the wrong class. What would the data have to look like for the classifier to make perfect predictions?\n", + "\n", + "## Linear Separability and Regularization\n", + "\n", + "A classification dataset is said to be **linearly separable** if there exists a hyperplane **among input features $x$** that separates the two classes $y$. \n", + "\n", + "Linear separability in 1D can be found with a rugplot of a single feature where a point perfectly separates the classes. For example, notice how the plot on the bottom left is linearly separable along the vertical line $x=0$. However, no such line perfectly separates the two classes on the bottom right.\n", + "\n", + "
linear_separability_1D
\n", + "\n", + "This same definition holds in higher dimensions. If there are two features, the separating hyperplane must exist in two dimensions (any line of the form $y=mx+b$). We can visualize this using a scatter plot.\n", + "\n", + "
linear_separability_1D
\n", + "\n", + "This sounds great! When the dataset is linearly separable, a logistic regression classifier can perfectly assign datapoints into classes. Can it achieve 0 cross-entropy loss?\n", + "\n", + "$$-(y \\log(p) + (1 - y) \\log(1 - p))$$\n", + "Loss is 0 if:\n", + "- $p = 1$ when $y = 1$\n", + "- $p = 0$ when $y = 0$\n", + "\n", + "\n", + "Consider a simple model with one feature and no intercept. \n", + "\n", + "$$P_{\\theta}(Y = 1|x) = \\sigma(\\theta x) = \\frac{1}{1 + e^{-\\theta x}}$$\n", + "\n", + "What $\\theta$ will achieve 0 loss if we train on the datapoint $x = 1, y = 1$? We would want $p = 1$ which occurs when $\\theta \\rightarrow \\infty$.\n", + "\n", + "However, (unexpected) complications may arise. When data is linearly separable, the optimal model parameters **diverge** to $\\pm \\infty$. The sigmoid can never output exactly 0 or 1, so no finite optimal $\\theta$ exists. This can be a problem when using gradient descent to fit the model. Consider a simple, linearly separable \"toy\" dataset with two datapoints.\n", + "\n", + "
toy_linear_separability
\n", + "\n", + "Let's also visualize the mean cross entropy loss along with the direction of the gradient.\n", + "\n", + "
mean_cross_entropy_loss_plot
\n", + "\n", + "Because gradient descent follows the tilted loss surface downwards, it never converges.\n", + "\n", + "The diverging weights cause the model to be **overconfident**. For example, consider the new point $(x, y) = (-0.5, 1)$. Following the behavior above, our model will incorrectly predict $p=0$, and thus, $\\hat y = 0$.\n", + "\n", + "
toy_linear_separability
\n", + "\n", + "The loss incurred by this misclassified point is infinite.\n", + "\n", + "$$-(y\\text{ log}(p) + (1-y)\\text{ log}(1-p))=1\\text{log}(0)$$\n", + "\n", + "Thus, diverging weights ($|\\theta| \\rightarrow \\infty$) occur with **lineary separable** data. \"Overconfidence\" is a particularly dangerous version of overfitting.\n", + "\n", + "### Regularized Logistic Regression\n", + "\n", + "To avoid large weights and infinite loss (particularly on linearly separable data), we use regularization. The same principles apply as with linear regression - make sure to standardize your features first.\n", + "\n", + "For example, $L2$ (Ridge) Logistic Regression takes on the form:\n", + "\n", + "$$\\min_{\\theta} -\\frac{1}{n} \\sum_{i=1}^{n} (y_i \\text{log}(\\sigma(x_i^T\\theta)) + (1-y_i)\\text{log}(1-\\sigma(x_i^T\\theta))) + \\lambda \\sum_{i=1}^{d} \\theta_j^2$$\n", + "\n", + "Now, let us compare the loss functions of un-regularized and regularized logistic regression.\n", + "\n", + "
unreg_loss
\n", + "\n", + "
reg_loss
\n", + "\n", + "As we can see, $L2$ regularization helps us prevent diverging weights and deters against \"overconfidence.\"\n", + "\n", + "`sklearn`'s logistic regression defaults to $L2$ regularization and `C=1.0`; `C` is the inverse of $\\lambda$: $C = \\frac{1}{\\lambda}$. Setting `C` to a large value, for example, `C=300.0`, results in minimal regularization.\n", + " \n", + " # sklearn defaults\n", + " model = LogisticRegression(penalty = 'l2', C = 1.0, ...)\n", + " model.fit()\n", + "\n", + "Note that in Data 100, we only use `sklearn` to fit logistic regression models. There is no closed-form solution to the optimal theta vector, and the gradient is a little messy (see the bonus section below for details).\n", + "\n", + "From here, the `.predict` function returns the predicted class $\\hat y$ of the point. In the simple binary case, \n", + "\n", + "$$\\hat y = \\begin{cases}\n", + " 1, & P(Y=1|x) \\ge 0.5\\\\\n", + " 0, & \\text{otherwise }\n", + " \\end{cases}$$\n", + "\n", + "## Performance Metrics\n", + "You might be thinking, if we've already introduced cross-entropy loss, why do we need additional ways of assessing how well our models perform? In linear regression, we made numerical predictions and used a loss function to determine how “good” these predictions were. In logistic regression, our ultimate goal is to classify data – we are much more concerned with whether or not each datapoint was assigned the correct class using the decision rule. As such, we are interested in the *quality* of classifications, not the predicted probabilities.\n", + "\n", + "The most basic evaluation metric is **accuracy**, that is, the proportion of correctly classified points.\n", + "\n", + "$$\\text{accuracy} = \\frac{\\# \\text{ of points classified correctly}}{\\# \\text{ of total points}}$$\n", + "\n", + "Translated to code: \n", + "\n", + " def accuracy(X, Y):\n", + " return np.mean(model.predict(X) == Y)\n", + " \n", + " model.score(X, y) # built-in accuracy function\n", + "\n", + "You can find the `sklearn` documentation [here](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html#sklearn.linear_model.LogisticRegression.score).\n", + "\n", + "However, accuracy is not always a great metric for classification. To understand why, let's consider a classification problem with 100 emails where only 5 are truly spam, and the remaining 95 are truly ham. We'll investigate two models where accuracy is a poor metric. \n", + "\n", + "- **Model 1**: Our first model classifies every email as non-spam. The model's accuracy is high ($\\frac{95}{100} = 0.95$), but it doesn't detect any spam emails. Despite the high accuracy, this is a bad model.\n", + "- **Model 2**: The second model classifies every email as spam. The accuracy is low ($\\frac{5}{100} = 0.05$), but the model correctly labels every spam email. Unfortunately, it also misclassifies every non-spam email.\n", + "\n", + "As this example illustrates, accuracy is not always a good metric for classification, particularly when your data could exhibit class imbalance (e.g., very few 1’s compared to 0’s).\n", + "\n", + "### Types of Classification\n", + "There are 4 different different classifications that our model might make:\n", + "\n", + "1. **True positive**: correctly classify a positive point as being positive ($y=1$ and $\\hat{y}=1$)\n", + "2. **True negative**: correctly classify a negative point as being negative ($y=0$ and $\\hat{y}=0$)\n", + "3. **False positive**: incorrectly classify a negative point as being positive ($y=0$ and $\\hat{y}=1$)\n", + "4. **False negative**: incorrectly classify a positive point as being negative ($y=1$ and $\\hat{y}=0$)\n", + "\n", + "These classifications can be concisely summarized in a **confusion matrix**. \n", + "\n", + "
confusion_matrix
\n", + "\n", + "An easy way to remember this terminology is as follows:\n", + "\n", + "1. Look at the second word in the phrase. *Positive* means a prediction of 1. *Negative* means a prediction of 0.\n", + "2. Look at the first word in the phrase. *True* means our prediction was correct. *False* means it was incorrect.\n", + "\n", + "We can now write the accuracy calculation as \n", + "$$\\text{accuracy} = \\frac{TP + TN}{n}$$\n", + "\n", + "In `sklearn`, we use the following syntax\n", + "\n", + " from sklearn.metrics import confusion_matrix\n", + " cm = confusion_matrix(Y_true, Y_pred)\n", + "\n", + "
confusion_matrix
\n", + "\n", + "### Accuracy, Precision, and Recall\n", + "\n", + "The purpose of our discussion of the confusion matrix was to motivate better performance metrics for classification problems with class imbalance - namely, precision and recall.\n", + "\n", + "**Precision** is defined as\n", + "\n", + "$$\\text{precision} = \\frac{\\text{TP}}{\\text{TP + FP}}$$\n", + "\n", + "Precision answers the question: \"Of all observations that were predicted to be $1$, what proportion was actually $1$?\" It measures how accurate the classifier is when its predictions are positive.\n", + "\n", + "**Recall** (or **sensitivity**) is defined as \n", + "\n", + "$$\\text{recall} = \\frac{\\text{TP}}{\\text{TP + FN}}$$\n", + "\n", + "Recall aims to answer: \"Of all observations that were actually $1$, what proportion was predicted to be $1$?\" It measures how many positive predictions were missed.\n", + "\n", + "Here's a helpful graphic that summarizes our discussion above.\n", + "\n", + "
confusion_matrix
\n", + "\n", + "### Example Calculation\n", + "\n", + "In this section, we will calculate the accuracy, precision, and recall performance metrics for our earlier spam classification example. As a reminder, we had 100 emails, 5 of which were spam. We designed two models:\n", + "\n", + "- Model 1: Predict that every email is *non-spam*\n", + "- Model 2: Predict that every email is *spam*\n", + "\n", + "#### Model 1\n", + "\n", + "First, let's begin by creating the confusion matrix.\n", + "\n", + "+-------------------+-------------------+---------------------------+\n", + "| | 0 | 1 |\n", + "+===================+===================+===========================+\n", + "| 0 | True Negative: 95 | False Positive: 0 |\n", + "+-------------------+-------------------+---------------------------+\n", + "| 1 | False Negative: 5 | True Positive: 0 |\n", + "+-------------------+-------------------+---------------------------+\n", + "\n", + "Convince yourself of why our confusion matrix looks like so.\n", + "\n", + "$$\\text{accuracy} = \\frac{95}{100} = 0.95$$\n", + "$$\\text{precision} = \\frac{0}{0 + 0} = \\text{undefined}$$\n", + "$$\\text{recall} = \\frac{0}{0 + 5} = 0$$\n", + "\n", + "Notice how our precision is undefined because we never predicted class $1$. Our recall is 0 for the same reason -- the numerator is 0 (we had no positive predictions).\n", + "\n", + "#### Model 2\n", + "\n", + "Our confusion matrix for Model 2 looks like so.\n", + "\n", + "+-------------------+-------------------+---------------------------+\n", + "| | 0 | 1 |\n", + "+===================+===================+===========================+\n", + "| 0 | True Negative: 0 | False Positive: 95 |\n", + "+-------------------+-------------------+---------------------------+\n", + "| 1 | False Negative: 0 | True Positive: 5 |\n", + "+-------------------+-------------------+---------------------------+\n", + "\n", + "$$\\text{accuracy} = \\frac{5}{100} = 0.05$$\n", + "$$\\text{precision} = \\frac{5}{5 + 95} = 0.05$$\n", + "$$\\text{recall} = \\frac{5}{5 + 0} = 1$$\n", + "\n", + "Our precision is low because we have many false positives, and our recall is perfect - we correctly classified all spam emails (we never predicted class $0$).\n", + "\n", + "### Precision vs. Recall\n", + "\n", + "Precision ($\\frac{\\text{TP}}{\\text{TP} + \\textbf{ FP}}$) penalizes false positives, while recall ($\\frac{\\text{TP}}{\\text{TP} + \\textbf{ FN}}$) penalizes false negatives.\n", + "\n", + "In fact, precision and recall are *inversely related*. This is evident in our second model -- we observed a high recall and low precision. Usually, there is a tradeoff in these two (most models can either minimize the number of FP or FN; and in rare cases, both). \n", + "\n", + "The specific performance metric(s) to prioritize depends on the context. In many medical settings, there might be a much higher cost to missing positive cases. For instance, in our breast cancer example, it is more costly to misclassify malignant tumors (false negatives) than it is to incorrectly classify a benign tumor as malignant (false positives). In the case of the latter, pathologists can conduct further studies to verify malignant tumors. As such, we should minimize the number of false negatives. This is equivalent to maximizing recall.\n", + "\n", + "\n", + "\n", + "### Two More Metrics\n", + "\n", + "The **True Positive Rate (TPR)** is defined as\n", + "\n", + "$$\\text{true positive rate} = \\frac{\\text{TP}}{\\text{TP + FN}}$$\n", + "\n", + "You'll notice this is equivalent to *recall*. In the context of our spam email classifier, it answers the question: \"What proportion of spam did I mark correctly?\". We'd like this to be close to $1$.\n", + "\n", + "The **False Positive Rate (FPR)** is defined as\n", + "\n", + "$$\\text{false positive rate} = \\frac{\\text{FP}}{\\text{FP + TN}}$$\n", + "\n", + "Another word for FPR is *specificity*. This answers the question: \"What proportion of regular email did I mark as spam?\". We'd like this to be close to $0$.\n", + "\n", + "As we increase threshold $T$, both TPR and FPR decrease. We've plotted this relationship below for some model on a `toy` dataset.\n", + "\n", + "
tpr_fpr
\n", + "\n", + "\n", + "## Adjusting the Classification Threshold\n", + "\n", + "One way to minimize the number of FP vs. FN (equivalently, maximizing precision vs. recall) is by adjusting the classification threshold $T$.\n", + "\n", + "$$\\hat y = \\begin{cases}\n", + " 1, & P(Y=1|x) \\ge T\\\\\n", + " 0, & \\text{otherwise }\n", + " \\end{cases}$$\n", + " \n", + "The default threshold in `sklearn` is $T = 0.5$. As we increase the threshold $T$, we “raise the standard” of how confident our classifier needs to be to predict 1 (i.e., “positive”).\n", + "\n", + "
varying_threshold
\n", + "\n", + "As you may notice, the choice of threshold $T$ impacts our classifier's performance.\n", + "\n", + "- High $T$: Most predictions are $0$. \n", + " - Lots of false negatives\n", + " - Fewer false positives\n", + "- Low $T$: Most predictions are $1$. \n", + " - Lots of false positives \n", + " - Fewer false negatives\n", + "\n", + "In fact, we can choose a threshold $T$ based on our desired number, or proportion, of false positives and false negatives. We can do so using a few different tools. We'll touch on two of the most important ones in Data 100.\n", + "\n", + "1. Precision-Recall Curve (PR Curve)\n", + "2. \"Receiver Operating Characteristic\" Curve (ROC Curve)\n", + "\n", + "\n", + "### Precision-Recall Curves\n", + "\n", + "A **Precision-Recall Curve (PR Curve)** is an alternative to the ROC curve that displays the relationship between precision and recall for various threshold values. In this curve, we test out many different possible thresholds and for each one we compute the precision and recall of the classifier.\n", + "\n", + "\n", + "Let's first consider how precision and recall change as a function of the threshold $T$. We know this quite well from earlier -- precision will generally increase, and recall will decrease.\n", + "\n", + "
precision-recall-thresh
\n", + "\n", + "Displayed below is the PR Curve for the same `toy` dataset. Notice how threshold values increase as we move to the left.\n", + "\n", + "
pr_curve_thresholds
\n", + "\n", + "Once again, the perfect classifier will resemble the orange curve, this time, facing the opposite direction.\n", + "\n", + "
pr_curve_perfect
\n", + "\n", + "We want our PR curve to be as close to the “top right” of this graph as possible. Again, we use the AUC to determine \"closeness\", with the perfect classifier exhibiting an AUC = 1 (and the worst with an AUC = 0.5).\n", + "\n", + "### The ROC Curve\n", + "\n", + "The “Receiver Operating Characteristic” Curve (**ROC Curve**) plots the tradeoff between FPR and TPR. Notice how the far-left of the curve corresponds to higher threshold $T$ values. At lower thresholds, the FPR and TPR are both high as there are many positive predictions while at higher thresholds the FPR and TPR are both low as there are fewer positive predictions.\n", + "\n", + "
roc_curve
\n", + "\n", + "The “perfect” classifier is the one that has a TPR of 1, and FPR of 0. This is achieved at the top-left of the plot below. More generally, it's ROC curve resembles the curve in orange.\n", + "\n", + "
roc_curve_perfect
\n", + "\n", + "We want our model to be as close to this orange curve as possible. How do we quantify \"closeness\"?\n", + "\n", + "We can compute the **area under curve (AUC)** of the ROC curve. Notice how the perfect classifier has an AUC = 1. The closer our model's AUC is to 1, the better it is. \n", + "\n", + "\n", + "#### (Extra) What is the “worst” AUC, and why is it 0.5? \n", + "On the other hand, a terrible model will have an AUC closer to 0.5. Random predictors randomly predict $P(Y = 1 | x)$ to be uniformly between 0 and 1. This indicates the classifier is not able to distinguish between positive and negative classes, and thus, randomly predicts one of the two.\n", + "\n", + "
roc_curve_worst_predictor
\n", + "\n", + "We can also illustrate this by comparing different thresholds and seeing their points on the ROC curve.\n", + "\n", + "
\"roc_curve_worse_predictor_differing_T\"
\n", + "\n", + "\n", + "## (Extra) Gradient Descent for Logistic Regression\n", + "Let's define the following: \n", + "$$\n", + "t_i = \\phi(x_i)^T \\theta \\\\\n", + "p_i = \\sigma(t_i) \\\\\n", + "t_i = \\log(\\frac{p_i}{1 - p_i}) \\\\\n", + "1 - \\sigma(t_i) = \\sigma(-t_i) \\\\\n", + "\\frac{d}{dt} \\sigma(t) = \\sigma(t) \\sigma(-t)\n", + "$$\n", + "\n", + "Now, we can simplify the cross-entropy loss\n", + "$$\n", + "\\begin{align}\n", + "y_i \\log(p_i) + (1 - y_i) \\log(1 - p_i) &= y_i \\log(\\frac{p_i}{1 - p_i}) + \\log(1 - p_i) \\\\\n", + "&= y_i \\phi(x_i)^T + \\log(\\sigma(-\\phi(x_i)^T \\theta))\n", + "\\end{align}\n", + "$$\n", + "\n", + "Hence, the optimal $\\hat{\\theta}$ is \n", + "$$\\text{argmin}_{\\theta} - \\frac{1}{n} \\sum_{i=1}^n (y_i \\phi(x_i)^T + \\log(\\sigma(-\\phi(x_i)^T \\theta)))$$ \n", + "\n", + "We want to minimize $$L(\\theta) = - \\frac{1}{n} \\sum_{i=1}^n (y_i \\phi(x_i)^T + \\log(\\sigma(-\\phi(x_i)^T \\theta)))$$\n", + "\n", + "So we take the derivative \n", + "$$ \n", + "\\begin{align}\n", + "\\triangledown_{\\theta} L(\\theta) &= - \\frac{1}{n} \\sum_{i=1}^n \\triangledown_{\\theta} y_i \\phi(x_i)^T + \\triangledown_{\\theta} \\log(\\sigma(-\\phi(x_i)^T \\theta)) \\\\\n", + "&= - \\frac{1}{n} \\sum_{i=1}^n y_i \\phi(x_i) + \\triangledown_{\\theta} \\log(\\sigma(-\\phi(x_i)^T \\theta)) \\\\\n", + "&= - \\frac{1}{n} \\sum_{i=1}^n y_i \\phi(x_i) + \\frac{1}{\\sigma(-\\phi(x_i)^T \\theta)} \\triangledown_{\\theta} \\sigma(-\\phi(x_i)^T \\theta) \\\\\n", + "&= - \\frac{1}{n} \\sum_{i=1}^n y_i \\phi(x_i) + \\frac{\\sigma(-\\phi(x_i)^T \\theta)}{\\sigma(-\\phi(x_i)^T \\theta)} \\sigma(\\phi(x_i)^T \\theta)\\triangledown_{\\theta} \\sigma(-\\phi(x_i)^T \\theta) \\\\\n", + "&= - \\frac{1}{n} \\sum_{i=1}^n (y_i - \\sigma(\\phi(x_i)^T \\theta)\\phi(x_i))\n", + "\\end{align}\n", + "$$\n", + "\n", + "Setting the derivative equal to 0 and solving for $\\hat{\\theta}$, we find that there's no general analytic solution. Therefore, we must solve using numeric methods. \n", + "\n", + "### Gradient Descent Update Rule\n", + "$$\\theta^{(0)} \\leftarrow \\text{initial vector (random, zeros, ...)} $$\n", + "\n", + "For $\\tau$ from 0 to convergence: \n", + "$$ \\theta^{(\\tau + 1)} \\leftarrow \\theta^{(\\tau)} - \\rho(\\tau)\\left( \\frac{1}{n} \\sum_{i=1}^n \\triangledown_{\\theta} L_i(\\theta) \\mid_{\\theta = \\theta^{(\\tau)}}\\right) $$\n", + "\n", + "### Stochastic Gradient Descent Update Rule\n", + "$$\\theta^{(0)} \\leftarrow \\text{initial vector (random, zeros, ...)} $$\n", + "\n", + "For $\\tau$ from 0 to convergence, let $B$ ~ $\\text{Random subset of indices}$. \n", + "$$ \\theta^{(\\tau + 1)} \\leftarrow \\theta^{(\\tau)} - \\rho(\\tau)\\left( \\frac{1}{|B|} \\sum_{i \\in B} \\triangledown_{\\theta} L_i(\\theta) \\mid_{\\theta = \\theta^{(\\tau)}}\\right) $$\n" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + } + }, + "nbformat": 4, + "nbformat_minor": 4 +} diff --git a/logistic_regression_2/logistic_reg_2.qmd b/logistic_regression_2/logistic_reg_2.qmd index 0290ea29..59227a9f 100644 --- a/logistic_regression_2/logistic_reg_2.qmd +++ b/logistic_regression_2/logistic_reg_2.qmd @@ -346,12 +346,17 @@ We want our model to be as close to this orange curve as possible. How do we qua We can compute the **area under curve (AUC)** of the ROC curve. Notice how the perfect classifier has an AUC = 1. The closer our model's AUC is to 1, the better it is. -#### (Bonus) What is the “worst” AUC, and why is it 0.5? +#### (Extra) What is the “worst” AUC, and why is it 0.5? On the other hand, a terrible model will have an AUC closer to 0.5. Random predictors randomly predict $P(Y = 1 | x)$ to be uniformly between 0 and 1. This indicates the classifier is not able to distinguish between positive and negative classes, and thus, randomly predicts one of the two.
roc_curve_worst_predictor
-## (Bonus) Gradient Descent for Logistic Regression +We can also illustrate this by comparing different thresholds and seeing their points on the ROC curve. + +
roc_curve_worse_predictor_differing_T
+ + +## (Extra) Gradient Descent for Logistic Regression Let's define the following: $$ t_i = \phi(x_i)^T \theta \\