From 58cda54dfe84c77db78fea405aeb7f10c4cf7ae9 Mon Sep 17 00:00:00 2001 From: Pietra Ferreira Date: Thu, 4 Apr 2024 11:37:18 +0100 Subject: [PATCH] vault backup: 2024-04-04 11:37:18 Affected files: content/notes/general/Untitled.md content/notes/images/Pasted image 20240404112123.png content/notes/images/Pasted image 20240404112143.png content/notes/images/Pasted image 20240404112933.png content/notes/university/year3/cs3002/cs3002-unsupervised-learning-association-rules.md content/notes/university/year3/cs3002/cs3002-unsupervised-learning-clustering.md content/notes/university/year3/cs3002/cs3002-unsupervised-learning-hierarchical-clust.md content/notes/university/year3/cs3002/cs3002-unsupervised-learning-kmeans.md content/notes/university/year3/cs3002/cs3002-unsupervised-learning.md --- content/notes/general/Untitled.md | 0 .../images/Pasted image 20240404112123.png | Bin 0 -> 49261 bytes .../images/Pasted image 20240404112143.png | Bin 0 -> 50678 bytes .../images/Pasted image 20240404112933.png | Bin 0 -> 50678 bytes ...unsupervised-learning-association-rules.md | 26 ++ ...cs3002-unsupervised-learning-clustering.md | 190 ++++++++++ ...nsupervised-learning-hierarchical-clust.md | 62 ++++ .../cs3002-unsupervised-learning-kmeans.md | 68 ++++ .../cs3002/cs3002-unsupervised-learning.md | 335 +----------------- 9 files changed, 348 insertions(+), 333 deletions(-) create mode 100644 content/notes/general/Untitled.md create mode 100644 content/notes/images/Pasted image 20240404112123.png create mode 100644 content/notes/images/Pasted image 20240404112143.png create mode 100644 content/notes/images/Pasted image 20240404112933.png create mode 100644 content/notes/university/year3/cs3002/cs3002-unsupervised-learning-association-rules.md create mode 100644 content/notes/university/year3/cs3002/cs3002-unsupervised-learning-clustering.md create mode 100644 content/notes/university/year3/cs3002/cs3002-unsupervised-learning-hierarchical-clust.md create mode 100644 content/notes/university/year3/cs3002/cs3002-unsupervised-learning-kmeans.md diff --git a/content/notes/general/Untitled.md b/content/notes/general/Untitled.md new file mode 100644 index 000000000000..e69de29bb2d1 diff --git a/content/notes/images/Pasted image 20240404112123.png b/content/notes/images/Pasted image 20240404112123.png new file mode 100644 index 0000000000000000000000000000000000000000..fcbd32fc84fe68f89f085b569a7e10bf473e08a9 GIT binary patch literal 49261 zcmeEu_ghn2(>6pR4Ep z5--SU5zMQc|kcQnFHZHg@(Jub#g&l{K|BwYPk!F8c@z>*0qFT80*P zw8+GhYAV?8^#%;(=0*~o@kdwMsP_jVm(NTqAd~l_Y$_!qI@YFAESY%cPO>do#mOW| zzU0qF&7SLjVWzh#S`X#OtW?d!31B(sJCe;9Aje15@K)4CeYS4snmk-gMtd(!J8Vo} zjre?DKX#fCbe|H+u`u|fqHVFIMcDC+S0ABQ9|wC_3|kBnHr3qMM&vr0?%VX}UiJRD zISQ&Oi~V-$%~!}FiUk=Z zUvq5db~?cvIqRe(;QfdcBEL%co{>wtYU@$Ek4_bHHiV3 z)9S1pwcOABDEFLUv#;tM1|(?8^f&7htp7+N2?>>~N?%-ke?wi3}QpC(8$h8ROwcgBp` z8VP&Z>V($GXs6YUIIK9{GpDrTaBN2orPw3RzZx%%mKiSz>Rmg#;(f^nHKAS2kK*gr zCh)4Y({GX85mPo1&*9Lrd0vr~smrf@1fuQ2@~UyI!!y6Ee=@({ArWcau^&mxG81S-u$*i2t0+-C3}R^aHtJ-Q~72JCh=>VqfwChc(;cijnUUnliDMnLTso(7czF zZRY|RIu~Th|tv(pgZo%XN49bBU_rEEcfJJa`^{wUawLC7*E1 zF?C40{P2WXWM2;^L_Ax3sqfgLOTgEBt^tizES~GSr|!z!@?r+`a5$j(i;fBgSY%n@MK6UrnximC`^*R0g(mA8+)z+6IT>wBlG<_;(rlf?$2E50^!Vb2? z!U5i41Ammj9~KrU9rV|gONr^1{(gV)A)XsJ2n!30CHGiT(-nJdits(-u-DQluP@=f z8~vB?jTCrE-eXN~T5RTu@qz+)z{{DcQ7w2)MSu)lO`rg6D32|PhBGd?~ zpms^qOCfRPSlwwg86i|{dx{p)GVl4g&{{_4iv_~_$II-K*@v<#5!e2>2z*Bj8>B=D z`tz!<25>Pi`CibULx9uesK8j*I3$0+ykY~VuS7?+{NvKahj@IJm;QNk;6BQ@czh}L zk+=W66P6znpm0C8YlKYrQRxN@orY5+r3gPRaOTM7<`y$oh6eG?%aGtP%0 zi{Y||Dfh1jlV9J|^hH$V-IRg>sO6Uk;IXwoqb^P@TN#GSh$MR@YW9EEH zgqZjm&w5a}*P_WzzeAA3(TubkB;gsusKcIjf)nUHaN==OfkOFvs62iXnr9Za-1NEjpD1F(butY?602e z-UJ5v(E1MVAz5F9%ZG=Do9lK72%jCNENeDH4o|3|TqF^&D3UNNvBd8~H-)5s>Sg+M zzqVA5X3@Rm2Reoa!Y=MKbkOfjx>p4`x^DdN!XdG!!16OovW_dKCrS4W+Ay6`Y`=x7 zcWHp1L=-3&U-7yAiX+{3r|vkMchum(EJsAFdds1tAg(-23nvk(e{Y4}7yJZtErqZE z*3J&1Ch5aGuYE2zlF6zZufnSkvVT3FGu>L4KU>@7DW^zsEW!!nPKOJ)7dk$_wThZP z`N!VZ!E_ebxZ%bpLm>rU7PZLXS?Xx{KClUmA3c8lzVgmQEWW^@0dc{=7zeOA?#I}m z9{L{@);J_jZuOnSt4!9L9Uhgr?BXV`u4t+(l0#k1w}a~6VS%a$@?9!y$~u;6M+)g* z1Gj7v!G`wlLgm{daQH?|4X=)3h8a-;h8SPZ@9!4)FyXj>l<`fdMSMOVICG_lj*=50 zGyHp!XvMyWZm&pMU-M2uV4hJ;O9Em=GJ(^_VU!R#x7{umud9OL1&G+_>)>bcAl)|Q zEm4=ABRJt3Olk+SWv(BS(5c_L)9lAVy*+~Yl}bRCNRSxCm;F~M#0ZO z1fu+8r?hBsxS(dklf9eC-ses7#Fih|pORlV_;f13+fU`2%ZL2@eiPU)hQ0*7W(bQ| zf+7Sm5Z-hA#}SRcf`(mu5{tdpn>arM1HTP|{9p~iADvV-tL{Ceh=y@PZuzi_LAX_Y zt@*ADY+eR^ziL70Pwu^^oMm0+ebjl~Vbge8K={FXPT>quRCO3{qUtH2in~up9fI1k6+6-#kwOn6FLcKJh||q84>lOA0;Y9ERX?u}TZpsx^b#);ooW zz(pUMuXd{9!CSSR$1OdL2JpJI*F|CEa)$|qG>_`+d&~#wdg^1V#U~u!E3j7#K}2ki zoZ~?D!8${#>}lV#$xk(X1(R{kVhw4=X~+u&eW;h&CXI-73MI{XVaMjJ4$>| znmUGyBN3h}5APg9RctlQz&{=e8Fzf|JlEQxP%2uRdwxg6)MXsNS&!#N4VphcaA z;#n_UHovPP()EV#Sx#nKIgcr<&bdE!$F-$veqzZse&!>L^YSU?&{cv3Q6JVshwaXH zrDIbVgiWw~b#+m8luS4zj9kNS^d(|thHBA&KlPcjkymys79)HnHx-35S26Rj{7#wa z`RsK>l+CVG$C^W10+;vddN(QRU{5-2f;0>O0&6#oQ+D{BA{AWcQXPV|Ob_GNo+Ynl zsg$H3YoC@T7ugfNcVEXKg0>H*PJ4#j_z=zB#Z78{A)RSe%fTh#se_XzFfC--sbaoz zY0YL&#{M_6apZnc^muOnyt(r6?z)2xqr^xbj8?byBy9K4)vl}880Q_qs}E*s4@P90 zG_X%QOG1V!+)YAFj4fG}a;Ab7c4j;qMy0Yy->`+kbW~JQIE4$My;7r2>P^-3}0*XAxcYpXefpjnx{f|@HcGjC6Fb|h>hbE?nGWYV5NVFh6NF#|mryD83 z>^@AZoE{cwLT(3pHBrxp=fL%;hF-Eqp7r5z2}^in*XjE1XSUYjxw55W$ulczsnkcl zGgtYx%SgR_mmy9v>AndRC2YvW>CSn1K@OwK_yh)ls2-Ub>%hrGm*i9UNb0|+a+u#bSU$jGAFdcACS*6+@#ToMo!IOJ1R=SZC-8%cIbgHvv z5&I9(eNG4;QN0p)In22~Dxs3NN{kl8--1Z?^g5axb~_1&L~Bi|U5?9P59QiR`1 zVykY^Y`yM@7n@xLVxiH+ppv}!3Ec5(sjvekAgC*2eni9VWOb&#+4DQWv3%Y6D$y75 z`@*boUL!B;_8h)#f9NqXGLk?mReBbUs!BK%(i)4qDAAsW(r9>zO`CgtbEWmG2j=FM z^OP7Hf#Wz}>LbY!S!IEhzWy$iWLf9Dv2kbqZrvSZt+dLJjZD^;BqR14EaQBgC&MYG zaYUvwuF{90Gp8dPDyE$FvQza958bv$(n`;0UEj?sNHJu&>{XfWtdzPt7fn7G%_}!q zkQS@9R_TH>>1g289Qo0)s?>g0MP}~px2mLkK4LQLlEBM9U8qq@RKawb+d4tW3utE= z8Fq$BagM@^PBpY{3PIUQ>f1a@y8L~w0@)M!yKKKE5*$8hsp*{a&_W3;?!^`(V)bcM zs=@xl1rDhM%(nMO_I2;9GcJdQQZC0FozXtE6tV?5eHZN0uQ&HLQj3Vc6A7Pvm_crV$hG-XEC4LsASEeBP41 z+K@(K`kRKF_jE2IS^nXBL<(Z&qc|-(O8;KzT~+00m~n88;}`)p4kcM%oaS_DdvVj_ zac$L1EO0aS4vvV^zQO18Zih^>{Vb#D7)yi5fO{o*q*{bAnq09P2Fpj;A_zG(A4B9M z@tO6dfnwpFRmq_c7vwHmn}7E(w!5f{z2^_JDO>@uS89Ne!g7WypbI07qJCu@Q9-88 zrfTUiPTjh0jRwzR+8&^I@G-dKA1c_qh8tB5-);mi-V+bKBH`|bBstgk@|@mt9M#`Y%*bta5smFD6R&o`z0f~ zIz#lFiA;Cf^HjRMUdF|MEtSyZM{7enCG=gzQLZU?6HO0{3)!tLsUS(PWfP$u1?C~k z(4ehOMs%i=lX@DAMTK)B{@QjozJ9j%&9brshKTBheCjdj^+ye1s~<+qbDoI6EQhwL z3DbmXuAv;3c3iU2ZO6h$UQ1^ky3@6+I?m`j*mk(vo^q@^M>lH}pJhgs0fJs>+~gJU z)Igl*4>816;lIz#iBcmR?3yE+vhACzq(j(dJ%g@Anory>tt>AjFI(Q;BHr7l zWQ3p#Kd_>uH*ORi^}LM5%z@PtTubV~dQ_T)l*m+bwNtjwo@>0!`%poh(kw1)Grn`n z179Cop$dL zs&Rjsu6)O_BsQqOWHN-uhf6!5K6*-qnNn8@84dGq)llhhv8P0Y2CcB(W_6Itb=7n8 zuP>L{tmI0Gp5A%vS$c|q5Wg6}uXgube^F{WRHlYmUXhWLn;64fIql)?RvT4Z)V1pq zUqcni!z8b(&-nf!@h(^+k)>>nnB_#wvv@a< zTABTYDFfri-P;|cWafmtN*v4ENrly-E`-DQdLyE5t#q`IvD<~usya+__Vb>4^jMFW z2-U(%>qkQh^K{$59w+Lj`?PgmBd=}oGJn6A%vspAbywOE<~Nf^EUIfsqir`L^0W!+ z`A$(K_v1nz68(%s~*n7!R6-v z0yAVX)yE8>Rf6Y{vc`0b_=&ApTfK{S)#VR78PW*Dw11mX@kkvGh=oBWrYkV<+MK$h zH7u_Zyf{yp)Sb^)Gpe(AhLTEY^^iAQkvzyoxYuHI>2vRkIoeEYO~WM5?s7j#pSF7= z>9OF1L&CxlVg4lFolbWE*MMSS#Z;~J#Bx;(msC2KZ1tXwSO%qhkB_9s2)V#wWT2oCCIf>FDdsx%op-D5 z-`A{z3^(eY=oyxz-IdWr6+(BC>(kth5{t9h8cS&(?Pkdg(|8k>R!P|mF3uDuKM%$a z+)mK=7>ucz6B&3Oz`%-g2M^%C!1wTRJW25m-LrG=KVV~~9_X;GxlyiVK!aA4y~t7EGIEUr>EJt2*DVoq9)B>&!Qvy*ZRLDRRCn@@$l zJ2|M&=(b>8g--KP1?HGGuBJc-2WWOv{V96F3xyt3ofn379%5=dC7p)77RFN%1~iN# zKthv-yWGE99^K@kj&JatoH4!%aHO@UcE3*i4mvIZqP3!!ni8BFR7RK7{j4cn$L2l zM_&r&u%#9*KHPmSOeXHoic;3E{4QcY)^=>aZSw|P4&v z`uVkJ#TXr`MxYT(%4N{^h11e~myIlAGUzInr=ATCupgyvu+=%T%%q6(m-wAjsEqI% zDm&|QChCm1FsHxNPMuajrGNNx}79!zoW4J9azY3m7G6%f=i#Cq!>I- zIxoshMC66I&=$;|O3kl1hBG^w5cnDcLa`E-{|bQv!(6Z(1Ls1rtngPjyZE z@INWPAGoE|tU7Ew5ho|*l&1CVrDU!<>kh~2r{`N5;Pv_Kl-d;yxZ3T}`^4)s-!eTH z8*qIf0j+`0^P5)QkyqA-Ud1N(Dcz!GGhRbkES7Ad*n`e|aP(4i93 z?<%B{2G_4F)M%?P%L`+(9~adx7HCi&<RivnQb{{RQgp7&?=A}l`+?UPb4!JQ1WTQeUdVMbpAxvetRD zm_rkD&_6?#qK*oFP!aCC6pG^MR2`D#WZ$3%w+tPxdh3Kjy}By$uTPCp>ey!$!Fo%~};!-I<289Qah(Ni1W!{jyqU zXLrpxInQMvS7)9pt>%j!+($OZyPozwLIkh)*0bd`D^jB^a5~i357JL|Z>1}L?L}#I z_mQaq^DAkU>FxOym2+@|`nWX|NBtrPW%IZluWYbw6qh3sN>jI@;4B{WDQHy&6Ycnd zvPb1Q10*kMT69Hb|Krjp_n;w-C>U7}N*z&*^wB1XjTt=??ip1`a85ByrD9(YO3;QoDp^xY)Ccz6 zO6q($)utRHwJz;)+t#XoYr}t^lh%%5G_-BM1kt)6QsJXHXs)}?59#2o1y3v;+OQZ? zhNQpz2sd6MPCgPHqi@`zy}y&)WzJc0dA*(1r;!23oAQdbIYF&Jg(@`ZlXNKVc5>h2 z&{ER0%BZyES|}}wd;-0JNL_8&&|yAkQ{K+nA`ryk2bux-PNqabZf(FPZNtt*+xT^pY0+RGV2#lF;lJHH!BI}?p4T7n>tP`6sgP? zX6r7&looX)%~DZojz89G*R5vmKgY}8z)%c`MmL|F6fKo*IA_gSe&i_}?Z%i(mG~o) zlMq3_7CzM;MF3(4*&^K4q|+HeN5j(#9RUOC5wYf-A2I`l4?B7B-TED~Q0(UoD80A< z7zsJW@O7KcR#TBrFo8y=Y{>pc!SKNadS(@0>z&S-@A&ypN3mRFHnD=C zUrJ{#;&H$IYK2-`xR>YRWR<{h0g&dwsnJs%Q8z%}0^LTQ^w6tT_;zu*;a(-+!E~7# zh7OE~ui_~tw)+@z_B$G;**uLC_7+uz$IJ+?Y94C$8S?cx41aUvp+#<@EJRj2U*g6K z3(-1rnN8}qEdq(U$&y%-w6|->l3bb=#MrqXlO=aGt>#{geQ8a(BC_t3vroK}6PQ6) zY`{9n+pE#%nk^bXgg_jmueMrEJT4V_VlS8ui#vjH=^hz-YLQb#u=Fr(8}m21!E}U% zL^Kw%RgWzTjSNQ%It_DNI*$rYq{$~z$!6J}KHE|OV+!ZmB&NMQxtewl9ZH?&GvnBh zoJB)#i%wd0n#?&l)z4~Ba|I9)D4fA?sdRMbfQ11Q0<&vLUR+JtWzbdcX{8^-kUjD$_P2ahtOJ) z>h%KklYWDUEfHKQQW>#vx-}qVdFm?6TBR`6E-Up~MXXiHt!SRPgm5cCVUYFA=)`q( z!o@_?r&)BGeAM(R&Jpg&9bKa%;~x|tylXegH26<;0b7hpL1jxzvD{2A5x1m@)^jMG z&_8ah_jJ))5UXWD=y`U#@TED4!^Dxni!vWLVai!M;W9I{BQHfJl^tMA97-eRLx@G` zW+1q)o z!E;^|#;_GA5bU|2QX?7uC`ni0X*Na_#oByi0qE&}d;-n~>_21V&Y{fTr8eN5C{QZj zt$fjb>yK}f6ar)C^UK12M&AD^aH{}7C{}X2{LSyO<*yqA0;PTmtDcW`e^k5yni8a> z!X9MvXOTApsE!BeIm}c2B-a0W0jg5q!NGU*`u|QD3;SOwU!;4#lg(!``4BH?cV=&*#C%M{#A^B732Tn;PV+SjobY7i%yh>qz8vw-j?vw-QCQQOt^U5&*l=RQ&`25(28gCdPztMfnm0b1o)}d?CIqq?@8f+Y8}nRgT`X7N zygTL?BCpr5m;LL93Ku`Lio6Bt0Z4j*xo#s87RF6tk27mKk>hI3yEcJf|EB2%506su z>V4W9(+i9^o%myWCKEa}nqve2x3=fecV^gdxs=Y=*A7T0;KO)_l5}@pO*xHXQpX!V zR=d6hGy3fyBpRzvwbD*pa~#KR`{nyWRA64xGjkX9uC@L2Q+-Abm^9$Wa4xc{#PORR zl30E`ATAmMzNOSEXm_qnKoc0-P-?$4`0NdgU-hkEC&6+&CMoYURtOUv)yee58Sfs! zrdNFX*HAoHcA(7|G92+|w=JC_c2IASuH8K4KcXibl1*(O%a|ynfB9#Y5f7~7`3P0b z#@nV?4*l+0C)72EV5jV_RQGd{v1gy?X7zkw3wqJJ_tMdlUbOa;nPM0HVWSp78anfp z0U+y6Fl)QbaV2JFGNQ^bT0O)+5kA+8`%93yKb8Q-VP`lWTk%}Z-^c|Q79F|=dzA8{npmjv%Ji58xY%D zKx~%^SjzttTj1TMzVZJqAksh74o}(_%p09}H|@_2rUTT@Pm^qXE!USRcVzN8V79kZ zCtmz;7%j|Z#zx*+Pv^wGP`5&bILGU+xSD!=CzB%Yf_icIwhjjdBKr;Rn46g;=O{eF z%ZFI%8W`YQ0=n5NIe_l`J$7u|go=s^vvp+Eoob7lcz^A3k1YYWq>~`3Aw+!w*4nb9 zWZ_~>TUlZ|?4prVwJv`=_jx#4_*gPM5mng|zuMf7W%GhIA=bW&$^_o{$nBX!9xHljwFxxwuKXoYskK0k; zzYFE@A13k$Xf?W)liSYx$Ao~r@iU{lDJ@XPnjTWX?{fPp2nrZR{J~@u+lrhoOj;yg z;io#T4lo+xOkRD(OShk!{nnQkTgek%S!v=(-jrT|k%r zsXSP6ac!+wx6YZPjYALjpQ^8@I@RNDmpKH3JlJw0MNT2FK#KG+Q<|2Lx`DHYw){E0 z-X!@EXTO4064HnCd&zk~wg01UP-&7zYDViL>2MwZqlFn!*bCv`rn1}2Xo`63UvDZck2iA!s3Nh5B0YT z*HXp7B<@lo?|vg4(oeK{qi>(St)$jbjvRX*Gz~X*s=Re{ac5F{g zEdgAw&_z#I;Zh0MaK{M(xy=g#{Tg8U|nIsIx=e8;(iT;X_D| z_DARvk`6o`)*1*Aa3EjFx1CQLQBE>o)?->=?MQGs3TL;{KQbrxJn||IJw;W7+qFHS zuE>P2cW|%p7Gg4c>5B*FUvnbhncfhQTS#Y<3Ax=GuO4sas*<&6>xF}4N5!|W!|azC z{#`wBMw24w5o7$h1*8aN#gsbZH#=Lmr|zsD2W z{SaMRKR!PI(4njA(g<8>}}c0NbPgK^#_%G4B=Hs^*j*GDagGH}P9 zSEL;rER*c@a{goImxG!@uaNhI8$EnB?J_O2I9Z`>8#wwQlDIJ%C$YKolKmg}j#nuU zbhGMfrI20z1Hf*>vHTV{QNDM8c2}^V&y(5%ui4A>eqd!lS599_0x%8fs$yBu{Rvm) zlqb?nJ^%tA{lfm*0OY`cI+`eKvOcGoFe50l=qFkg(>q13m19AQx~hzg$8VbS_V(^= zwDi6%5;s?uY03eBUqNZ$QPU@1TT5GVrhM3q6DRs-6Svko-!NG|# z@_N!fM_w3jh#jyvcP}~WKW4t1jJ0Fo4DDon^#qi->uGD-tEi|L7!BE`^!?u;E6Yt! z7x19hE}{oDmH8WgjD;uTn^Eya^Crcwp76f_>MIF=OK^uXcRy5!3nWpU82H+Wv{_oTi~kpi+V|NMEZ;j6|EJwWxUwl7D0 z_U|MvmdM+uHsdv1s8jrhv$&eLr6YL=^ave|Agk@tC@3gaJU%ev4FMHW-MIU}u7Z8> z3#sWMA7}br;9qKcbLdT+M|S!%5}UBY?>{CMJtW+_Jkf1-O!CEI>kl19?5NE-U1zoy zFS^HR#qtiZUyKMn2)JYDZ)aYXrYY19qSjM>mRY%s=h| zBB^kGOvvdMBs5!qTSCGvZJXRr51{d_7Z_gyAiLSB*ArfH1$%_ef z^u*7m?J)S~8s6dlaTCluVy1D|zJ{zB!3e9mrwFZPP*6$}j zcNsnw5fZX=UK{J**ZauazimI+$(8rLi&B`%kdJv~q(nc93+Ry5!$~~HhsM5$U;(X@ z^>H%*FdPTE3gQHO_bzaNU=&1O#eo~2cD{&OY_4w95^j&48OW&a5mTujQIK8HYPi+! zlyWbp+68S?@w_=m59(Hb+$(bAT!8+BI-BS9D!u8IiI&W1eq9uw><-0bJ#wU=r41Kf zcmS%x@5|Dq;=uo1F!3z~8K-$vklz`JX>1dBurHCa$`bR`>QXg*V(mF2*O0KA-WrqW zrJTZt>X{A$&I8a?%wGR(<&_5MtK8#vMRhoc^MLMB^!;+V4W~K3T=x>&hB?F)+w;b= zL(>X#MCvr5(+~EVdRxSFW!pwT(*Dzs?`I!IL%W1&Z9B&55I_tlK1$^F*O5kZ5CIR@ zj9a$g(6DYM1^)0&8z$UDH{0z@>pk0gZ-tquXIePq4T;+yW-MW=RS6XJ^(h*3D|$-g zPN@$2#UpA!=)Y#ss$ughnyCwKNi zm1^Ub#TmXf?KgSaco}=R&OI)x>*b8Q@KJOd-a~TW6s5AfKi+53zB^ATH@^Og4t^rO zZz%)?wA1q;qbQs(9Ak6HJHyZya6lI)~mcR(SgLO@V(H16ACPttL!Y}&GiQnOFi`GEhEl2UXtWCFG@%~ac% zOH*{MpJ(OMB3VzJ&E)4#b{=9bRVQcYb8lmHAP+Hn{yYQCpGr5Y4!FCY>2MhCF0Lv0 zB4Nd!ZpPF4r&-nj{b^accUI?o>W^KS9Uyw<`tf%KUHZ9y40_L*)LEkKKCT}NoD7rV z-cRIYtgYXSio3FB!DNUi$Vfq;B_ftdxk~r43Y5*v%rfhCe{2C6=1~yIQUkquWCluP z@7vBd{z@ahgb_1;nw}AD4&2|`;CJP0kFb;kek9l|ff=ZZ-eN4-e*m`~*jY;Rful`v zb>zhIoO1x1?3g@%<`a=m151LOh#(Q8Yv$okv$yWN1YXUF%u?9obC>?uOG$e7_Cg4b@L+DJW_Wb z3+C4xh5`Kw1Dj*6s?Pgy8z_BQ6(DE0;UTV5@-UD*66*KCxG^!hzdw4ddp`-Oz`EpP%I-?F$2Z}RFFIWQ=@157YVq#N*#Tthl zB9$5_B0Wcr?*NSv4o`7;Vm3m8%(fIF^F4fMvS;xD)E6+ zMI2MC#=T)Lw>^J@Ant4S&70}6yoFz0{?3Ul-uQ+$B;10987bkq5r|(o6+w%c#{+If zOEPymd-}_B4ofMEn}Jrnh~*zU5#tu(f)FH`7t-UT&*uXWcmUt;q>eiH8}pw=l)f=`SVv)!NFU$fSan-s2SF z1!adm5+o2y*0XW%xwYF=Ht}ec-8LLbO0LYW7TO?h?B%tr<0)mtEnQtOkxaj^Fd&uq zD-#W3ztwhX-C0;ot%yCP2zJRBRHY}E zN($1wwYMGzWD&QF-FZ(2=moWKJT3HV#RvDL{mP({j!rsixDnOUTT2D4%9H{LwUW(t zNcu_~?qS~%~Kc=&Z73$N=9 zQrDQThywN35!mZ|Ru=F>55U1jLn|t%RD4Hg=Zu_Y-46 z{94l%c6sl%(qLyxdJ0)o#I12x#rHlPD2U27g#fWJsmu+@bQ)~Jntt+V?^6@WrW`6V7*q$3L*8}X9GaMglwgSvDnsH$BWI7UvdpbDl^$Y>ZX^OjrtXKSa z4bCbIfWC45?Zg)hE68rwyFU@9<9Ml-HDZw)Fe|_=Nz*nSH>aWC_^S1JLemqJJq+R3 z^Vz`RW?GW4|IlEcA2e8^o*A46ajaf6YKY~17PeO;8Jg5kKg$nyPw=}Jr`EC zON4&h@m`YhsvEi!Pe8YGUrx%5e^PU>v?aLHYOKJvj1;-u!G3&c#WkX!$$f|UJb&-S zX>;m1vO2scNpN3m2_Fbg9^R-My2tSw6mHixxuM;?mfhN;{K{RDd!n2NQTLQm%8&JZ zViG=N(^pVFvRXPAQXN0o?A_Yhnv}FLyX;-{m1Az7a_R&Bk=<0SQ|}%_nyXNz+h8S! zlQDZild%NWM8WM3I9zX7~v0jeci|`vsT$TCbN0|{ENjZA+A0F zd{+l?^EQzl8J!Hdhs*G?b$?qxP`CKW@|>)WhM^Vb%5tPYAWRV+<6NJoQfk@AIiY@% z3q}STo~vRk{hCWe^kji37wz zQf9nOPef{k?NbyAFn2x|T-9a{#BkH75$$r|feLp~Lf;qjTCy1{%Mio_QU54tq=dNx z5q3gBK|zaGYw+{l9zr7dn5K04iY=-i7VjclNs{;iVK(q6_oJa>i=|q{`!*rurN>c- zd$u`VAKt&e2PgMZJJj^80em_s5!?MPsf~Oh(1X_=b_+ z*vUz~m)=n+=q7H0vl@3M)#);4Il;!-_Phon(bX>AeW=$1LE#)TZLEZdAuB5|*Oat& zlw+5wuSY6lBX^F8cipLr=Jl}){i~OGSZ4a2Iz>l_AE?RZWF*myXMDA^l&RJlhY+u9 zOx1;7PF0Jt&?8V+IsOop#f8aLRwzt@$GG}~ae*_6?6aR(e@B)FtgjA(r+0Z@cGV1& zjk8IbrQjZS>cH*!4!w@|D3o-v=$`o8G(axsha<*kMv5IN`FQYqNNn2ImM#-ZqXPqCuNpp ziSkgL3GUQ!6cMqrj`2LlF!)eR#nhaC8J2XMA-I{^vY)d(UqiIpczXJY?tD)Ktrrup ze3H5P{XN%~kVw+!Lsd~?QFN;B3^})BsW{66kQZq0Y9(44h~~fmO^@~Q-VIz&EB!`q z)dq(&=7h=2y#aMU^NWPQHz_0r;~_tMy_zUS7Hz)Acrv8QLL-s9gt zp+(TiTAGyGM}ui+mHV~p;|5g&+gEG#t#=b;CyI1+u<}Hsdt55>R97b5+tN@f91169 zqS8UjM$WAjM2Z0bXnVyqNFA({Dwv|7VXJq;1*)Hd&Es<^+{djj8Cy1G2uMCh{^n10?#`U=H z9TZgv^16S>={fd69;K}}cQ+DT8}yetL@pGncu$HOIGG#cR_~h8EPR*j0U2cJ@^S;E zil1NwJ_W7T^cxDzx$q398vEE|2%oMZ`d7u)>V+q%POZn5dh2XaH|4*`j}Y4Y5DOp> zzM>RbuGcxm?zUZmrE&g4%xA;?)8rg*vdSR<>1g+PkL|c@z_`kTYDl)UWH_7?*1JaJ zn1`(N9fr0U;~3x_oK_@;-iQRK=AYqIr5@dMCAEzvjMHSO^CRrre$14ls= zn32`*XOhN3DLCTX-S*BU24zNNiC*@5d!_W=>ym=+xhqhad_&8KuzZ-XYbY4i$Y1Bv zl8>R2oe4M+@AjB1j6_Q=upZZX<9+&)s>I;!Vsz|2-6$uyxPcZO+`r_zD!22Vtx1!* zC19%l_=OlwCqc0;T-x#T+{4b;#B~z9$8^Gy{ItVhHjW4zBuu)FAp0w*(@X=H7h2G$ z(R3K!Gg*G{#bdPg(Z!P{fb9B~aS3c`wc{FpyQZpn-UC)j6TQASZ*UzJ$I+;(zH1)R zXGrH9NO_n(=eFTJJ6+W3Q8#VFt8}}nN^ALJZAI6+8iowk0XnAy3S6=E$VmIxVX8C(St~>306`}3M1)~NeyQ2dpx0Tr*(p?a4i@P-Pz0v zXC_iPIrFRT(J~gfGt;uNQLWhZ-b;_!23kSUWAF)vGLKu{TxUL*A4KG(y6 zXu!<-vW+|Cs_1XFC<9>Poh^3PZvT#`25|V$KgPzMBH0@1EUjuk+--=L!(WwQGyeSI zFlE7qK-lVM)$|WVKz+CT^`zBHLXBJjg1)TrZ+bb3zjhsmF5O5`Nk1}=e5bD{ z7@y}-zD7B4rVZ~Ypp=z20G3<|oA7#4QW9@sbi&_C8Bq=4xPLz$2s@K;)Z& zEGp&m*xyj%2ylPUD7sZM#$7vR*-?{ZQisdRX z$)=t!jZ$e*aQClz6G_m;19iF#8k7>ve{E!QzPA0v+M{pHnt^jx#z_2*a2+23ywpWO zbajTpV>$X~bpeWV(YaBxb;Mz2Y01hj=l(#$&6QXB+27^8+~QWC<#{#=P&=D8TDxPneu*%Q;Qoca#2F=1 zm?$eNCmtTU`FDrGNZ$d)$ElmomTms-5m%L;jotF#y`+eFagiec<)D!m*PIO6Ne}{;tg6@LU(IM?t@oXjA2xl@*(V z5iq8*ib~SQkJrx$z+pc>qXc-A3N>5jAn=3#Wd|)h2Hm%2UzCzN*x1~!ihz;tup1u| zK-jh333lQA-p0=5PlTZ6>$A{`=GW$>c=I=&`L7x%7Hd9gf&o=ZF2B(XB}e;4AVnCu zQEFH>GI>7R-S76mQoq2UURS*d#ts1rO84$2(f5@BKtmW%A@a*iJjwei8CQim33!G> z4^Pf6$p$FSbV>h0(L~Z&%AUpIaZ=D5TxiWACGW3oC9vzdjF*j{5BWqFac^@yZ>iff z&LDbMd&U}Ik8ED}%1x}k{1L+7_jmDol|>qe!Tf4*Zwu>JbFci0j!1j~lS7{kcqO<* zYtgyQ>|W{)r2wZyN&^!Eo_w=wL8Ou9_3Nv%e!y4l>b*mC_4MR>KDfugY^M@yiR^yW z(dhsX$)^5AAoV@>)BlEECyY_L3T#%#RSCBQx{BuYz|4z*q%RIHz568oD?H*LdHVF} zvXXu|P$E!Q;04(F#igY`FoNVQ&{HJC@B03iE?xl)wB~v9BCe!<+e8+3hl?lq(ET=@ z0DRj}cOkihyL+vsuO(?NL>K?87C1E_Z)rbi&#It@H^dAKfvR9M{k6wIoB$G9m%@3R z;){6-jJkmsE5uSE$mpfXMOJ->QuVq+s-5GsHnp*g3ZMA$sg zJ|)kwjI6TmIVJABD6sH3HPqV;z_#T-8C}O#D)|M(QyjlWPpdnh?eSJCUJ7Olnf=&rE%>V86$}aHP7f zz+B&BAb$GNtE?190P)U$Q!<`nxHpqjT0~k zFF?e7K(yGu1eA*#0F5=b>?%Uci(UjG$pN7k2TQEje`KxSOQ8Iy>K8x*{FxOl@IWaL z31HgTnm2UUpch>M)f*_z8i+oUeM6YX@v4A7-9*Z1(|I#c5Qwe|I>vz*hvK z#E9u9Teq2Awhe*m0nt5dyl@rW>>`kU6Y<-*BzsBH@DfQ8=d z?3aMWz6`0wBIDD-!`UkmW4jyU#-Tv1N_^9gIKGRkaq?IlemsR1Xxy`IG)r|`On}#O za&jh~pN@nq63cKcdo+}&OvXo=z;Lovow+*kX9A<#m_t7dbpXUjRU;UM+mQh{ zq97{81~x#tbPzqLGzA6eD!oR5fT0B}coY$&hZdA3T?ho}C@KO89U_EcA)$m2C4`bt z-U{fso_pW@1K#_5f5UAcd#^Rim}AVjwvxM_pWn6T>^>zB5vr#c*xC$=sj^+bozc#( zS!H=b^m!|oTyDCzNjI@}7wE1kb-Olv>Xj7TJzt2vnbL9bfW!f*#aLqvcOZo9J+c4% zS6R00hbn_c7DmpAPh9MOY1@W@?*aZ=)8K1YypjDNxZ22vjCFGjKcpQ+zWjOMIAjUx zEB7ZTn3L%x)lYIhB|m;D^Zonxzf40PU|otYGj2KIBUMyX22!x^$M>&37{Q)(SF>p; zspn(_@14Cp(#a}$=Wc*(Qm%oIzWYFpK zuEg4q!h;PK967CXpiE&K_c+`lu1cv~iuPPBNrzZ`buK!}N~BQr$9vttnu1D?l`3)UaedTp6ssH5l=FW5LCH91+%-rm0J zR7pUgt--LPxw7^)SMEH|hhu|`TDN*@5l(;^Ym9W3cK)l6p;|J7QtIUFtiNZdc}Wqz zGEre}=h{n2#=a7ww+W#2tsX9{uR9wQY4=d34*Ib7$bZY$EvRkceFdd}w-s$)U10@k ztX@XPvRY$uyYYRq_%JvCXw79b3% z&y^HEy$I(RB=Q;#ybDU6K^m^*q zke_*|D1N42<0!j!vOM|Yaa7d4;kSzQE+Kf;zqo#3MUS|U!IwAdfK3Gk!FH1o4@d2-D%ScoPD5e0q}l8y!|@x0?NOyh}rTpO!dw7yxi2QF39^i z7p@%IwhSOH>YeLf@)4~>9S5}y2lSob=W^Gce2B}N z0xzto11cp&tvX;iz>wiGa+`WhZ?uEOu3((Le~fejz!WaPxHV;od;a`+FJq&G*%XEa z2nk!^bX*7oEdE!GWaHJY{QXNV$EYCLf3i8mIP=$|Q6mF`18KAGw}ye9W@-Nb%Pan6 z8bCJ8;*GziUEr!e@!{0Dw$}AUm)Z>gr=Z!1^X!{DE40@)*fE_V5}Tvc6?u1+y2di6 zFVj0xfEXzF?TT^!y`e&cEIc*_9BKB-ci8#G(^9z&03o^^Wc53wXho=kY=zfTlV;x? z6Mk0Oo^9igJK+9iJg`WaWX*F+{rvq|=Q*YTc?RxZ#?_c zAm8PD$~~rZtQ$gtY1!x%S|!d-Yc}yQa9c@dUbSz(Q}$qT9%s~oiCjN2ae5+oO?6Case z3de5H;##16JTLtn%-yW8!AsT-jC1*VWd^owP!wS5i2DQf*FC>H1SW#vcu5^6YbO%o znCD5Ah1#it&=t{BKBssQX!){OQoQIfbLeC0zrpYZ#5A@W`pB(d!_FsecV$WLa z_DC-ctQ@Eh4xAFP^F;U_3IK8NVR)Ncl2`g%8?;2p&Sq)9332LDTKS9A>aCXg-E(@L z$HW}9Ga|vQdfFG6NU(E2Hl+ICCtW^`xV^XYbAMJ-icb#pyt2g=y35AT&&?^&Z(+-A zuAEv=(F-8C&lp}@M=u#195(u{>tf+M|De7!_Y%NEzt8B)DQH0CTzy>J@mthB9j+Z~ zwoce~s)Xs#eg8T*@!_f4j}4LsoD~2lU0g)PdMyYM1l1BCM9-MZkg1g;SKlNm+}KyP z9v<0eSkI_Z*ynuTKFMyv)u|}f-$4In1PM*TWg_U180!V3>#lAaFF*vS+{4B26i0X# z9NCkyhrg4}z0~*(ozQ-t59-Rs&VQX z2G?3cF#4V+t?nyYvwQ64s&5Z@lm1fW+F=T+oqZd^{QI7$dYyS>|+4bAT1Dx+OZc~{( zbGau${a2};UEI8pO2MaX!9$X$pVvDS!NF^Qro{koffEM-8N_rbD!q6Azrzz4z6!=5NJv`a8?S`TEd{k2gD-&V=_O! zz5QSfR)6%4F{slW9I}?%H@;K+0!QTCF1{FMTlee6S?hDBc($EF|2SY#{B5Lxr1q+4 z5Mzs>QDEi-i~540=b$LgW-ellOlZit%!Z$H`8(@M8<6wTiSw0Iob|edB~5$HUq3D zEp|THdUa)hrdh2vlj;WsH#S2R-qy~fveiZo_aAZ)67c0Yney{J2_6GwAD%`=W;@x` zC4pw18*Jqpq=KHh&}SA0u~G}W*ymI=2`=J$LfR&}xuuwafG^+gm!MjKoKYp_P363P z!Ch#9A-g0bBJZj)lQ_Aq?s@Hb5)9<5!|%8@0EpXaldkQ*@Q|wOVYTN;TMiG~pobs& z4!QmPXjuG9HZU_=v+by}H_qQc?7cvHJ9}E<0%)|(z&a8Wdmgx7z|G%s+Ja)H9$1Gi zJn+j$TA~Ndo_m+h+}<+1aN}5W(O0)pf9Nv*nYU3M(c8hPzq>l$j|a|A%lf^X%7WkJ z0pmArRtHZ3)(LP`bqB{>m$HTA4RZ$;&bnIR^IuoygTbHZwrBa5$}}K&HaDo^KF_9t z#G_3COTB(W(1|N6Y10?gCgiWk< zZg|2ltB!5?FUpd!4j~~Use%2fy&I5KyU`v@fkmBsRCq9B3BuNB*mG6v7Iiiyta-}R z)U?hPR~NXzi+dIznsyb5-i~~ zX}e`8qUm0-W{EtE*nfIFGw|n^msNmwH?#~revWZ#*NuSTRHodxE^(@j-`El~F2`?^ z$9*?DdHOgM!oxoQBs^naz3^|8_!07a9YOc9*@k%vwo>zGlL}u zX9=#q!c@=R-6MIR*~8P$NTB_KK~I9qgcS%>wz%;HG4W5hA9FvcH!GyXq&~%}aj-Pb zJpGaUou&YBh9{`dRg-uj~&I~6%Kc)Dy8iJq!?rK)eV-eDfJ$MZcYe6cf#sQeQi!0A9ibWi3)UIHqgC|KkcEm&M#4s&=Snx=bvi7I>8QMO5AS z{+GOxQbO{B09k*)63ffU#Z-&Ju)2-siEfXe*|(41Xy7amt+XaTE(gwM1KP>6SwEy9 zx_kc123|OYr&V6k_AF?9>8dYXGxoCgi0|3E0J02SQM#5F3^Vhzq6)k?C;dBF9nsqd z&U_CA$x0~{C(XV~ESJk*47>9Xl#!+pStE}?em>2nAATzY;6bUof%&B|F6FU3a~y&Q zkge!hFVhUhnMc@~2r!Q80CNbN1wXf9Th<*8#x1wOn8FujQykfxQ>tJUC!7dP+jeBYa# zVJUD813EvoJx{uS^C3o{1RcQlu+2*+r>&J+y-3~jmk#QKCkD;0>)dKMz(;jSpzt~T>44eMQakqEa{y`# z%oSwBe7fcTYx8F=UG^q`IE@WHP6d=|GEnN0*Oi`v6xy3+9N^kEiWE!@_Zq072jg zy3E*i{Yij_+Jgd4qxe7>n~v%|nz2DV$&fdqolU%U^}17Ym0xGipIBGzX7eC_Z)bCz zn;o{Gszs;P*{1KI>c4u%!mNQ$lW*N(%U3fWk|aL{H(zPvP` z0-1zNA-bV`$*+>Pfa96hsc2w}co<9e7u(48>l<#AK8Dk?3tNqJ{$QRa=JmCp{S#P|#J>x!EK)uHGD zv32a$7HBO$*z5>5aRKmjp@G1ny*|G_sQ78Uaetkp@6pgYzG2gyt{|xynQJ{;p`OSA9CO$c8NH&z>bC?Djif^7gI zfMf9;z@4Wrq|2M!H{en3DL)ED=@y^z+vsM~)&o;8Y9?K~fg(6435>zFU~S#egF@Cg za->}C1TxNT*u&kEjf%-5W<2YmeCu{F_)!IbMny${tk%z;U2nO6vftGJOb`I#`yeJL zX2cwU6hnKhvVVz%yI$8NvB$wq{5{ioccQOp^L+#Mber+fF?d(e&^9wJRBWq)5o;OL#39=G(a6GlHPCe#a$mFsc1oz?Aqn z-zLY<^5#F}jNb^{$}JqQ_IS{s#vXb$Az)zno&)oacHdD?C(OM`7!@%rF+ z_gQEId)r5=4?1)01+$IL9isM<`Qces;68BMP6LhW7g=e}rRN#3B}^6k0r;z!oE6cb za#w+kDObY!WG~o}>cc=cp%N=|+ar$i7+SrnO&WF;iu)zkMnfmMcC5dpf~(y2zv+I@ z`vWg)|1t(HQ!w(%=CeSqko#>rF}ApbzLOs|`jRpVK-FqfglGjS*zV9%9?mNNv*Q z5l!YSh#nM$#&YiUxmdFY4aH!MijQ2BJOQwq?e{l(;sIH`3P8ssY&aUw=q@ipsUK>m zVv?Dlg4t+?Y6}_Cw3Jmc<#?CaQvHLg_Iz4ia*1%@5QBLCWqIbv!dWF z0qmJr@J0}W{|SrUojOwW&-q{MB42p(S9as!mrHE57f^aUKoPBijcVRAFrUb}|I1u$ z!2oLiMjY#u_`REQij7whKy3s>8zJYVso+0h0La49q>MooS1E2bctFSkrQLl7dHZ7uLe>+ER?{t$ z&*chad%~-k?k8(K%?CM}iw2V}?gfL>bHCo%H4(F}_K%cMEOc^m0`G#51iOcm?%V13 zYwR!VD;L1;U*BNN0k3Z@ZY-1YWD&HB?z*wnmf#?^4%|4PPXHNbG4xal%x!9G+}!7# z_c-5T85G|+IZ`b>5eN{9Mz*5Bd}%X#q-HPy%(%y{R~n!0M9hb3pJ8u|k`<)5y5ukLU`+er|dcQ^)#EU!nP71`m$T>0_u{$9^xI_yVBQ(vv*3g z%yD;OnP={PyP(bHipSoNxZ)i6U9n6_8g*Fvl0&0M~8jpU4W$;X{0NrYV5Y z3StmV4|a4PyAcV`B8WJ*2NEdcUQ?q>*0~cAgEX?fjFE^$fdbG2Sw+e!L|9i)*)P2u z5WM$5W;xiq3o%`&Z65OWTJ*wuKBLo^j=Gx(ymdCO>Ivz7sa526{z+|Gb54_AjDb3f zj{snvC6@Zr&lY#qrR^jUd%i>aib{V67^!yT{uv_70S29=nv3M_dED2ggUueORSA5` z?egm7YH6V$aEEQT6|TTh=Q-j!2$*Ay%gmt^_9UF+4B}hGJDnVZ-AQg%**}+H^8*pzXlTZpxI8Te*DSbI{V3w0bD!mfH?4#pp)yv zPrcZk*c{x$7dI0>rQmS(YbIfz+f<~8n%|K^#oEN(H%@35+E-C$1@K}~Id;MsSpVO0 zvS519)PUKRq`Zfuf0BRWbKuX1FT@;|&KSvJPe#(t*}Wy>a0Wl$e|~AK?LQ7wVyjX2 zq#Mn<`_ebrcScz%?)O z{qib)mrOur?RST3qh1n}cU*VuYf~Lq{gk7nAe6qgbG%M9IrDT#8BPCm`L;uiLcaSK zU0`Wb-XS3&uZ~`c_6IM-$`1dO(DJ5ubopDZk%E|pGwH-HNJrL8%<>JM<}hlNEZL~M z0jj76Tr|^CUU96zAVU}*4vC+fbd{{Tm7Mzn3_I>L9J%d1+I(f{qVM%GLj%Ga5YqcJ zJ?CIOI!@)c*9Sn}@5Pi55g^{o?RGjDll0f5pI1Z4Qu=u5K!z6aXcO0D9-& zx?FeaBh9GMCiKq=E!r(eqS>>fIsTdZ7zPQlYww%&0>eyu?*s*@Gs{bx*N}k=5cRfe zqnshv+xYq&FL=nj=gs)AcXI(GGg9OP7j$S_&29_ZJ32)kg8UaA3W%Q6aX>XIM!AbO z&R&PU-Mvi?nz_4p?~Vw(G$~@IxX?i|$z;b1hs3vcJOoqsnm-i?>>8o{O(!fZAe{mN za-4Gl$vxK=+Lr^3(3;~u?YUO-V*+}+Wz-4zw> zuIKBvPgvE(w*E-fp1(gJHt03U3;lTb%Tw~*K{sklao}GV=|)vuYv(ll5DA5?w$%-^ zs~k8{=_B@pVLCErrxar7WJMe%QA(1BemqM|ztBI9VhW)6?*27tRWMh;VXoh`RXhX6 zld*IDHCMJ>wj0(e@N;8+c_IXYuk+|1h|gE-U?;y^3L)N%On{w5*6B|cqPX>Cxy)7;Ulxf+B0z&lS@pE>FrBwv2F z?>IG4eQGkTVfHPvBeHHyYmN5R^H%-$QZzbfK1W8>HyX9oc(heZAgmsIW&L@QP}npl zMg}w7Eynd!uTdEPv%GKcXi?Nw35;Whiv+dMeWk3F#>BPR@O?M1;k^^kS757H#hfaL z%CC?7+N!IbVzEhcsSHLB!rl8C4_=E2<>D8A%@OuB&Fi!FQOBzK#zM6_BSCb0n!F}p zC`%n@4Vo-WA>GiAj0UV)3Arj2&a8!5IrUewQWF`Ra?h^#U`;3 zk5Y9z`Hr}uvAJ}l&xi!Z?j-GCqG!<)i=Dc6zL z0|Xm-fGOH22Kf^mA0<@(@nlq7UT&Lo zP2sA~GkP zd9anNK=W{rw^?Fg&?_DYwP3?dN)!WE&E*lbAi@!4_q<$ap1CGvl2JIxf}%TWms^_2 zk~C<6Zc;L#GDoy6)irp9VuyUW7`6Injq(`L)_w*))mxgSJnY+3+Q3>q>^b|$j^%{5 zm`^ruowu?(;I$M`REXQS5e|qNKflY&reKc(gU)EoWBjC%f3>Mk|GS2U+4zzJtYx4# zYR&u8qlVr~M4?~hstaZdXVp#jm8jjxnZu#v?^kVw^;;0tm2y1Yc;(v_OSv1()1=>g zxo00yEH}4wee3qsy-l6HNEWGyS3zD~W6c1{6zr^I3dY+(%V-T08=6~9wi^H5=OEul zzAEjPax30bn2Sx}AZ^vAt&k{^1ZyAsV{Nbal$J5(yMSX8a#4yGqlrNWF7>C?giA~$ zGbWy0^D?{~4!80h8H@q96;96&jMlf z-F_q7N%`Ws`N7%w2}S3-&2lSm&Ao(!t~6i%`1bG#@eO;|zVR$n+lTk*0DX)1$J_BK z)4k5}v8=k81$(-=Nw6G@i<&96v|ya-D$*T2kiX5eB=vDoZtC%LMUsJckDFO_G2LlP zzDk?xAz)S;f3dR?OzptBG&*FOY-%*sFnlpW)^O6}njEG3rPnrp(QK_W5XE?HQ=sfT z3_}prj2(phO&cI$E6a8ZDPKQ%w=Ep~bC=kob(6GBM5&dq#=B7KOqF!3Wla`?!DA!W zR*_oOH0`LhGF?U3Egfs;vtm=Ck_wN!X#oo!c0wta6H@bVx5olnc`?Fo#+^7Bw@DVR zQ|i^Vc9lE{HUxH5m$2s=^QtG;Bo@W^{_+hDhzK;Cb&kOsAML9c28{**JA1oOy5f-c zSIGY5*E?Q#YRbqL9GHq8NIo3peX;ZU=H5>#hr_(giuS-aWI zJK^`bly3((HgaC|Lo}qtF@qKuC64uSKM`PiEeB-3&<}n0@9)hnLz9)8mZFf$*|S|HCG_-h)WedX{+`5E zd&1*g3Q8iEZljoeIxelu2f54NN1mjW*{)OzG+kQBx~e8!iTC4a)@mgc^&AVIEI<9p0Ou8ZPJc((_U}0_(nt;+m7nfH9;_G;ld+ zB?a#N>@~^9&b&_A-(3qT^apEq8WeXee{R;q`qj5NMW^Woe=Ke@^b#EAdR2(rp$xrb z;XJ(in14L}yU`RyzDJXh1}A_o_3H*Rnyz1y(=AqJ6OGeJ>X%-BOc`Z7$*M-=bUKcXOZSW zU1;#3BhMu0(Ra{G`3TF0Ifgo6Sf5Wijyg^JY;aYqD z(Z4-XoOn?9A3f+L6k6ryfyUAuFIFBU9^gx*$fJH<5FcLPd8YaEE8l_4_QUQOWzHTPD zGWAVW`(#gU;%t&{vJ797om7rgks04qwQ$Nl*h>KW0Ye( zd(5q6f&Nh05FA4}CZzFg42-4C&28nDMS=u=t?bOa>d6fBo<%V$7rd5t=8J>rDnUQ$ zI#T9S6bqD@A9^ZhJI-amID!jcSC-C;v))egc~;JGLcZ_hcX^guwAHSAXpF4G7tmR{ ze^p`o@PT5){PciROlHG4*bT*JjJPyoI@fsZ!Sc4aJN*L<2hQUIn5ojNN#_onG^P5- zyY0>){i&!%xg0z=2*ej3T0hxmFiIHJRrgsOT%#t`=SzzH+VxFgDggYcoOS|lx%eXG z%td49`m4#RV^D4OV#tf1@5$wvDA%saBB;9$TCbfzjz1kO425-} z+RM_RNy4pxj6UqxP^2-8=mGofsjS0qb>;UK zKI-&4Zcm>nOgWQc(Ym(cU*nGmoKvM4j9PcTp%`+ZV-j`1x z1g({uj`MuuEw9=>!{6~*sczZumuY_H1mH%q{96qClTYpxGFVer{%xC&?c!{UGmQDo z_t^N$LyBU==PA8oJDgCv6v(Pis|%NER2;BbG9|);mNKymjEVzo2|tG$7CSB_^bnNm z;XZ^?MM8W=spj_t`41{SDwIN4H<`=@7m3=lY=s=_I?lN2rEs-_ILBI0ANNbK{qSyC zz+trq&1|>FZWlTe%J;vU$D83QvG3B?_ZbKBlw}DgTX5byv3J+oMax2}?w(=Hl$6N$R@(I@r?F7#B6Yi0>YB}ahr7(_9&U$ot>y4Z;uRPx} zzP(`~yFa5Yi%Mry)-U2)1V(>EYGSNaOZ+ui^8y`s_4aE_gW-~Axj%%$`k-?~eM6@z zu4l{@WR(qDdXP*Zx;tM$}?EdDm+Np;! z8uR{Wqynt7$MNqOImX4)gX$?LOopd(IIGY;%^Xq(4FX_X!Zf=p1k6piu2yDQqQ(6uUK%X&{1-EWA4nMQ4DLG^gsxLd3Vq|ZjR z5p8W#{HccGtb0OhK{AMvhT`#(EZjj-A(HMW3W{=Ph9c3FBV=t0_Gq-c`IxO9X<~{Iw|fETZLm zgZ=E^EVMGBZH;ao8x^uzLTH(C8Q>gc)wcAk<-&w+h(5+4F>a&D=uBgK=?TWbY|^;AFmh&-!gk+O5I1HV@P4=gsFIN+x-sK?s)T^os~}>5Hz=x ziQ!f|{R)h)A$EQ)Y^1hp%*`j@i_gwHl89Dowhd1V+Ai%3(dv(RXwh%;>bf z9iAKSWUgJGvSQ*cH)Pu1x!N8J^7};rpMIfZD<@P2yozYYIUz)T;6wXc`*nUNO>lEz zn|Fb8lrGyBocQ&O>;H#62wsPz6jyY=@d>0#kTgAjiIsP5zZ>%TNw{E&)a7EISzObA zCb?KT16Jha+_D6z`inj{*%ND;Yh>tANmg5W&=?JcChNNKI3E1H0sy=KJ( zwe26634Yic17h@rq8F$By!d%Dh?=4bFHfwOYQNY20$Bfp72#$4E?i?d`w-f2uX{HR zb>?a_&8l4W(Mzonq4n8(9%p~$JEa#VjUg#J0XvL;(;BZW0};JEygk#vi7T$2^mbhS zHqH_pM`;<+@y-W&siCzQYvmJ-lQ?~v!&!l|U;K&gpQ}^~EUbCtUaRq+7sI$ym0!;u z11G!0H^!#VWwi`ooF$#xQ}f6H`Zt2Z=A;C|47nz(QZC1j!wELZD#6UTvEDq>)8LUG z17<+4`|^{EfBwBvusUj>JI&~K>SRCm({lhx#Dtp|{OjU*aPfcpB(7=qr~7&I&?)s- zgXc1g!F}*)(oC<~p4T2dvCEdVdm~5`XnBBXv_%clIIoWD^hrR1w`IW)KM?4b zvnL=-v*$ksbAPd7jJy*vkd!@a%VD;rHCh78sS`v$6xy(y?a@KyF`@MpyHXxn0?WB1 zM=TONB_^;JoMf83*_G?VA9onqSkJ|!u#kPU@BXjCaNQj??>qw&CtcnFfjV)Gy&VU~ zlBK1p_&9ycb=rB|284ujHw1wHN?au&z)8-AfZo$kfluuQAOGtR=uMVV?nQ^J$Vkcr zGAKFj!Edjb8`UrHj#iL%T^ISsi^H1u7yJ0q(}sThG&Yc~pWujXR(}`v1a#z9ZO=Ef z+I_}9YazjhM}5V=>%36fb)OaRiYgk}m&cnVSU0D1mPG%>i}Z^jhO z?CpZZf3961ATvh{0vc{|5M0tJaAIoln{Q7t`)vFM9Q9~*f5eCkt+4GfwYR-zHsKVS zYE?4_&{R1UVbz+aF$#a0ocQZVn0IAX){1pmNX^qy9liT8I zeiXfV^MMPur?ZmWy?BEs%qI$$oT?cd;0=lmMeovM91!B}+mU3YilT@=4TV!21Y!<_ zvJ?NZ@Dy+Zo^Dn`|7HKz+lE|tuXE8QAM@9 z*5WJ2S5B<7fndtPYSL5$aZ`;QH$|M?55A7k4tX8C{9O!f?b~%eHY#}9{v_$^&ulLF-dwEm_{GthFi0PXymGIm6cajHhor5{Iqf7>h1KWOVIF>z|?LN z|7zhfn7I=bXt1aCS|Ycy)*UWM9Ihf0vC|2bYA3~9@5vI z68-b$wz!oW2n~ZR6%wbVIdq+y`)n`i|Cx>6ZU@70k8e-+`IfnFne{eFS;pI}%?O1yKZ{Ppt2eQf;C z?HiEdF7?-2&284P%j>6&QKJj#u~D@vN866Xhp%JERKSqXp^t_C#L(ie+1j9pM%$kd zJhYJ;)V$ogZNdH&#Iapml*GPXBB>`-etVa1lOt>QBVQ0g(ip ziR1;`@656B>u_+9GGUSMC&+~wfvVCtI_Xa)hXXSbhh{t#qpmxDBh{Ab8G9re!cLpc&D{qN7=ttd`+Q|GFMfV4kHr z{<-GbUmk>N|6WkA;;@j}9LE?&j8h6_D!Wn6LSH+uFe$YbJ+G6!rTPTuzh}j4UL_^+jgtLeRL4ttEn{BJCX~Z(k7Gffi zkEKU*vw~+P%?+gQ_U?&Yn@29UOuDbG_LQuVT*T+H@k-YcSK~tHt;p4p%+oQ5b9P|dcJ8ekV`e7X8zWDIU?U$1X}LSS@>Mi@W(>S z=+d`xLI34+TKdr3s=LH!;3L{ph1t^0*)M43Knrb+RboJ2`NujOE^s!5S&zPcg-?DiTaqvK_q%a%0TnpxGke&-fpR6e7$#U$;%>`I4MDB?FEj+_w(z0XDkb=8 ze~iR7r!m#^E88Rn;b`J599&=90phAe z@Z{gO4#Ui-#xe-9K#q=Mh`jacH}_mtjC1W=(KxSJeaK-_+@#l*9cjZwYdXa%?64uA zcX8BbzLKx|*5C2%1DVh@70jJA^qT-x9LZekF1F&;U`4!(%&xIx^OVt_VVTuqjqfEk zH;kphpOEN3Y?-=J3{b*FG&&NrjbpICLqqOR;3xnp^rrpeL zzaQq4#w47hk#rW)v6c1=nx3-)Ld%R@3UB>iT8k_GXRdtY9HOO5l2Nid!lpM#BUUS7 zQ{k3#o00pe$#P?b9!ttQUNoKfV1-<|-^{zaZgO0D^o(}hvsA;2>-P}>?&Aqn(T$}U zJhNR{%(44`t8X){hCYXpJ2ZLhYvag-KG8iI&NLlv@P0j5j)a7_&%7~|Kj1guYIm>H z<9ac2$uSlk@`2Hl@BGv+9{3tEgt3)HbSQSD zs2G^k%=bv>c4TX&pl-s!TbrFk_>f;#p0$@{?MP9NP@PlyOqiv2>_17j(^~3NUY*9>awDKl1m|7z+}jy4v~sHw6zgl+T5%a!9;R{z~TGZ_SdW_s`4qaccOxv))V zv%i+^Q6HJmSr}vRfSrk*>P%^5gW3A`>if(FBLVH^m6_CE$|kCLJBwl(lc3(tZ=GH* zEi2c-46a@7PlvV&e{fic6PR_h46Fol_Di*0(L_9(!jAgk|!;7IYnlL z)OEVu(cr^!3*83=1-JFRZ=6M{8a* zMEXqt!w2b>srG{$6=9N2Y`w#n`*_xIRjwprD)F!fT75xn9xwJLAV5%tqrW!IEks*u zW}rkRAStMBraCC%9y$Ou8)B9gYqxsz?&lBfnp>+}O@nWLp|7n{Vw>a6$}H`}%nlU> zJg*J`$%T?z*$LZP3TH~w9t{~qFu!>QyHwEzU)EO7Q4m;?ggA+QfZrv>i9tH$V>MhI zjD*>Y7ik5=8n>25Xw#n?BQ4AtW_tzC6{h~YAY`2bBQfg^ES}=LJe#aJI%G7q$(3o2 zDqAU{d03G7s1@}er5xMBKAsO&V-)6ID@J*+H2X$JpSmINr~v|*K>EyJIWoJY3;n;A zQ3L#_#t|WQ)OT&F%7=l&Iaw#REAF4gB9~s&79DqN>1tz!i_!bjM1sS6Jo1A&G*+J2 zhv93zvufLTGbttw(4MMralL0Lrc7WXR+wqrJx{_WUywN0k8pWiPi zQ*hO7R}=GabXCo;Y7rqu(Sgr}m3vggEn#RMrGk+V^3#iU%1Hy|ZuEVX>bYgbRz52Y zeoeViP9Gdlo3-?|2T|}ZX{IYHc$;rN+#mMlY&`YNN@{>K5$*j$awX4GtZchHW!er< z$*YEHr0cZZ<5Edti(1|b2=u(NmezC%=^}r0<}K2Ig^FcTO?$kyLWy}DIZ=?e$tL^3 zaed)K<_UIJ6HL-${QBK7ApE@fW#l`T5wW{K!%8G%zoQ0>B`b#4Wd)`Ca2717ZTM?Mr#?1Wp9J??3ce= z_U{{o|AO* zhy6~Merl%j$X$=Mt5eTCk#$n|iaQ0XE@(z@EB4Lms?nqN5cfrob$D-S2RB3ru`57iF-*w#zvCs*hJ(7`M5;kzI zvnTtM3gtS@cU-3zLDu_ZNNXU!Z&%TB@TR&2d+JUliok?F#-#;upTiU*jFz7J0F1zV zb*AbWhSx1h)1|fw{1q0b9fIqPVU(jf&##6?ZTl5Dhj|MYh!!{hcoNZFG% zWGP~z-%X7YTuN#KHNY!V%d=a5WCbXbEM*bLXWUs+X}%#79mLm?U+yl+iw67N4^0hb z_oS6prUeknYR6E*ny`}vTw-T%NPBT9-&wb5y&W2KB zC+42mt-Y*E z3XKm4?zR_Gb)WaO?XPOfL)2YvPluztd9HZPnI4eym1ugsHWN&fZMo8U03n67nN3T? zjwzP=O${~(1rwOl1p0xRcG~)K*tChSglr!ZzsW0T^vE~8z-=HWtA9Y6lrf}~;|SXZ zrwFf*u%r0>Zxr;3$G(`@jGkvb_1yauA)<>JIfuN&+EJ8lKO^5dVgvdrhCO5wUD8FY z$R8m~VQ(J0#+Qu+d^{iKB;Rq~)wwTEbHAmg@t1hDwYu+h4TxzE*QKC%0@O-3O>^Pl z_YTl8_u27;7B29J5I{-qKe$p0)~HDMIk>ePuoph zkry_yG9EDF4R*8Dw`)zgZ^Ebupm!^l)XB7!u&Sf_F$TQZ>6j_6k(n*feVv>nNB6{A@2$x#KG7lmp?oS}nV| zqi6i3)1rB>gC^o=hK#c+oqrsz@v+lLcY4%@2)_be7C2hwEK06=GxFsJ+e&w!KbEh4mwE5dCGI z1DaYTgfuzSN14~4yKtcf(UW%}uBOvV*U>bwUIF&SyQx`AmPbMUDK>qVONknlxRB>= zH<$yrvhvK=a;zifkJmGDu}XAA4N!-$`>E%jSZlp;aZ69L^mqCE5?&FEy<{;JNi1BZ z&5e$Zf1Vyv3h6uEE_#dAS4Dx^sZZMMpB}u)^wO9jS#f^M;x0X#&^BxmYJm zL&Ot~+umV)BHGE-abFfW6L)y&s3(E&0^4(9CD3-51&qGG3%+_{!({Hyv1FoY3x)do zyg$k~Z)#R2-LS5c!e#zdHjFzxC|88xkw)CDqdKX4xMf$w$m@P*Tghf_u!~y>l@o=I z^y+6$vdPZ5-idwZP5a}m62_m}%!m#;->3$U@bw&~b+~rE6n~h(3VUR&S!5Sd-iyVH zeUN!G)oM|YDs76+PC>aq<>_B$m?*d7-ZKvd;g`fN+WI3VOrsD9L1l+=!$&B22uunS zgI&<|%)Q23sHIt=SE}(CUU8w|QM0_!m&?hd{GNVCw;bZws0#gcdxOo6okLm8)d&Lh zdL^wk2qTqOLT$r;nM#Qzss3Pw_SXi5t2ZF9a?Wm)#_E!kO?@M9wH*UaR|ck3d*N2) zIBcO%{LTGn&&>1+#IPXo6&k+76v$^dKa2dU!b;}G_?wC(4D=Esu>n_N=i_%m??PK# zHNUCC=lh1xAZzj2k$|WY?wF~jXF%V>2S(P^R}_#zCJz>#Bg7OpZpKbmfy zaN=SKjI~_siLq1O{0T9ivbdVC6`cHHkrUHSc>_(Re)brzRvJi9|DFxgNRHnm-*Kwl z#-%OJ(;tG7H6AePee2^RHlqHi{E|&dJKpd5(GDA#o3 zL;#y%h6BQB=#JW$$BjfJ{EiIhQ2;&rp)!bgnLt~*f7O=d()aSKF6r5=-tZ8MBYH6@ z1d(}-2a>wRK3@Yg@F(;kFVv^&sx}uV7K|#x5NLIu@S?j%Jx410N)U0Qcoz|f*X$0ssQhhF$y&e14^EV$tTmhCGy zj^0Z%aILA9N^yPEYZM$pU``QC-)~tXNf7c=Enr^q=k?f4*Nx>8$xu+kJgT2ar0%cd?QGkS^7 zgS05rD|JH>^$jgrlvAdy-P2+u`I7Iuu47@O#;XG%S>Xc`eZi-Rx!XkP?bln-_;Q!O!+5a=S_W``zbe|qj zcO8;C=@mIb51S-0N>bc7k!Pb0Lk!i7&}dGjyV9ae3?ql#lFjoxwdr_C#H?II7VbCH$h=;H~WjLsZDxp$mg0LQYMTCkx-Xm z4>})*o3j$lWLV=MJq>daBER=5zb>bm`eh_B^aRi3v-+v=!h$f7Z&_-ot6=YAjU&;S zHD*xFR|V1)vbdL$bCiiAL8jOFU^X=M9HwAWQ)|i6B7Zo(#*H!sf@@ITJm*`_q7H+e zgj__X&=#5d>csE9e9XLX9Ey84~b}lhRmI$K0CV@4;*`?!pCjl-y@RtG8bD8Ju8-T zNZ?BNIMMOk()^T{5Ed(Uh$X@3Q5fLIK0>5Cx;wxL7?-f5?2rZ7>Z;{Hzgc&6wk#x0 zjVVDkB;4)6Uob2W)Psz?jG+A;eHFj1XiG6Ow?}>b5CuWcs=YF#lO25?BMgV~;PSCn z^$Fv{tHpO5&%?J{kj85zJj7&KEQkKH{!~=$A!akHf^(4DaYAl_b9EsyWP>EivBN{p zsPoaR1rK-m6@qy!QP4bQ{;5p?5^F&>)Pg1 zit7)bBnYb>&|A1bQb$wafVF197)5eXA0P$fMT6Hr+FBev-p>NK9Gi!igpIo{QU1H|{1 z(|#WjzLHkSumtqQK4QbzX&DuU<*(M4@7e~KO^F(lP-U^#h4YCaAT)*U^hb?h$Sy%G zx)BYrJdEl3a~O77{$ygWKF_xD26-FBZlwNA6no5zR(M^PgC;phVd&2;75>L%p2NoN zVclagBC5-Q8(kBW6I@@oxarzC`ISQK@&mBWeSo>?tcbAD#O|JEq0b{2xp>f>lAR01 z8K(TPb!7`;Mr?xG$ZSCQWvS2nmb`tBPu%-0t%5uX@6eB8<#s|WpnJvJ)b|cAT^XYF zkR9} zm^mj$T0@Sld13%%qxZb;^Raa)(5kk}Zcc^=fp@&;P_wd>bZV}$jl5vpVm_#ztDVSj zAr08^cQxDKnuY##(b<(RjS)Q=|6)Admj@>WCQILv=*VVY%)k(Nc`jLst`tD0s`Jpg z!~1rU13+i|rqYTAu_`%1v-JE?1?+w>aQHHhqDWev9_k6zwtLv^PX8b+F{l-w%BYT6 zZ3*8e(8s(B66y=lF9kp8=%vUvvs*$#_S{e0ii{IMfeRC*<+Z# zuksLgd%t5aZ$;`ZC>Yx%lo~4{hULvtLbyv6YE9ok9T3}nxlf>k`nyZYx;25=iZE_v z>mbkDS2Y<$R4Bq0v3Gz^bH@gz!md->w;~@=m`@e?xUvIaQAXAAQswVlb+fN3*H;qD zLkdr=j~b(Dj8p4h^EX^WD3e?^gObgOH_s`f$KIG54DL>lnb-n&_XcEB?2hU-Ij%&C zz8>KDhEGEklD#8m`Mzc+p zc5zdmyu156RO;RZe*cF1eyWpT4Za3J<{vCpSPwe}FlfS5Or@$UTrG`R&>L_Rh@f>A zV9O(}mCuzW6P7@QXq|Dz^X$I@bF8{ft5i6<52)Y>H8BMDoHYR|AZkr4lq$?P*;ftW zN&*$wbD|NT&Bk=P5`nh8-uR8BHo417VM2XfkV2S0sKDJom0vZ~3GK1_jew~EB2do@ zi8M#30I~j#t;m?IK3pJ<)e+EA*kg^eO2~|LR5Kb}(dVILj$Jp;D$T01X>=Zs;{3AX zk0w*9peWg+p5Hrn%g972pTe4o167{o&wrjOr9wPWWtDI4{V zo43vwfVWPsRk>zMMY8VNpJ=W3tPLA~j;?Ft`|>qgtXNkNyEVz?4)`pJ6R;#Xwji@2lKlPBZq^7 zJkEVSvtNB?>1ktz9*3K@7hc}n+TOK~lPO23oAj19zuT*l4f7>+yykiX2 z>-d9&Uq<1U4IkkKd95>vcToqL0P5J;NSF09Y4I#!dSY}`tQfQr1zR27^lni z7(VvWD(Y_&Vq+W?>GGuWU4but&0>G82G;|B#5OSbpl?C8-lxzw{@w{i&Qaapa1N_s zuxJ=U{xE5?fV(0E=yY@jJHlTH3>v9=b>jbXE-UB!(@R^jkYAdWdNX^f{+Dv*1~t-97rVPLDz~0E4w2K7gL|NmNR)V{Eij0Y8a*b|7vyn@!K9D&J&qX3 z!dpwfpadTGpLyQDBJ;e*#qHCre(I-pK#Pjy`{>EPKTIOUAEcMF<^31ts-&F=z8>Zk zF0*i^_M{8UQjAgwfsgp#68D*KND7D(D94nIqm2LXNF%b3{Ql--eAD)4!2^O+DTgz% zGxDFC@bJ5taq4#C%#d^rC-Tb7%8b=`^0KWikB_Z-WQVv~PREkdu``(&t|h(m=-NV- zqo*rp1zI$gP6zZzt(|d& zPo+OrnbWClWy8{rpo;i!XiD$my$e=IdE31pQYk^4Z^1SloEMWgvHd+m_!<@(OoW5pT&k1y ze;xRLiV>Sq%ntyyUMyVwc2e(eSnD1|CQdofT(6Y851P%~NZXZ=voGco9R6l|T^E2# zfnxW2C3ng}!!01!V&FiZ`VZWh3Zj!cEq7b;O~8tGm+r^Qo2SnW8CkLSUtiu8zLM;HG&l<#M!Tl@%05BK$zbz1Dxsd-)9r?d6|7gt# z002S&sqexn?hvP4FqtM|X%8P8-4FL?8;OvtkOpHqG27^hT5=|kGS%=P+)j39!ID= z+5z4QGX30fqaKhG_EK^8QNG_T@rt(Kkwh7o76RCT^G=N>1S%uCK-W?zVL!$HAWD6~2(rqP3wjPglp|fRaRLkB8GOn5c@8IQ;dE|y%NF1kbpPkyT~*E2^O|sP zf>12T0DR$@J;K&`i89CgYa6Ys1(rmDg+LJy4udVdA^MOpL3Tkt+j;g0E`hLG{tN-aElo{NtqYi=j-UP&yNUi};oS`*wO*>!U=Y4U#-+THM2jZW=ka#F1!!t%odv)U))KgF)cvVAyW4w% z9L=K~9T^UBq5K*(ySaJ(6{+~Ur@7cmDl8GXOAzzJ{r%c_GOf+JO$$LWhI0li8!SR9 zNB@JvMc2ccZO0j=u=LjO?oeb^N5?u356`RC4qn;Cpu@&!K(jCb7K66A-pBLun-^?W zYpwwnd(I77H*1rVw3&92MPP;;Dcc^9Dx3^g9_0E@P0R5xx zifB&lABNnS(c8NRj(+C5U?O7nKk~oIW7~tb#Er)xN$Btp!lU=Zk@AEU{lEx; zh*|~9S9O?v;Ic7BwmHZ&pwaDcxmF^L<_8qF!zIX!ja*qK!Q@cT&TNH4m4_^ z&}U@xzV3Ig(h_x_XK~)*z>YEwCd;^A<^j2UbW9|>l;aZEVcI=NBeQvb28O}7)m3&x zySQlrc;$N`yLic%ClwdZySPp7dAfHli2h>QD@!R1gY3KA*@SrITbK}KaUO4w0}k#e zByyzZ@nmtRzrG7+4J?J-5ol0?$6pXHcowqw*+>@VLvG@J?};NWBu;DQ!nu4MX-pF@ z5jDqJXvn6tm(PH8h~9t}U#fb~a?1=L zlFEiYJ5F<;fTz8_-XS={x*$0d%e^%{JH_B}FU*^geJyos-TcQ^T4QF86+jrR{cyKv z<7AFvuAJlQJJWt{_QB>mNK4&Wj`{4 z{;tA2x1T*oHe`hsgZyVXo<>8MYv1*y(G>qm{Dq`<-^wslh+d))zZPBlzS95DE8tB@ zI?svAFTP~vbNBZfxH}#P#!%Tmd=^%eK(y=RfmBa~G<$hju=6IKZRKaWc;VP64}?kp z#oaPc2fT-S#+?x1|w1jF?>$e)#2of7Xl(^-a=- zeQr4Fe_Q^4-;FLYRR5Y6o~B=o3eD5Av+)d%3(4)n!^=*{`JQ3jPlU0NJKxC4N#1H| zn-aDVSgn=;`WEv60!QTx+AId00-M{v_GpNxHTOP|iK zP^eHyRGHWx?IkadWa8*XcS|Hs-yfTJ2%4)7rW2R?dz@uq3(g7ZXTP%w=t3r;K|yP= zvvYI1`hyNuR4A9|sPozz%QlO_F^)BkC>;Jq+EP3x3~`yGGf`so^4!8C1?+NCOD3qvY_fj3+C- zWGUGBp8I`60{;7_MG3=eyDX|!pF@TL%9@)w)bGyfMhLr+EPI#;hq8NDRG9Ly0~gm$ zzlp`ISt|_%#RLB{Mh2+g3_6xMultl|JE26)9yf<=>5P)fFQ*N$ZxEK@ z+3Xx*)PYS~ky^X-AC0Bw>(^}2+~WV|!4W_Jf|i6tXvOX3UA5cl*84qY$JH?2Sq$%a zZO;2`iR;_l^19&ro?yH2(17HOI|c+b4ZhbP*L!%v0++8x-M|0-Nc0bg1#eol%G$9+ z%JoKz`ObcPzkO1oI=wfrdb8ysO$NKw?+Or3;x~TF?k1~cO?CCJ)n8KD5bF7D$|KpgoLsYe%ca|`?bdT}^PsT)U{i(pf@L0ds zJqD?>=5<@w4Tq5@E_5{C_(JRyOOXGsI0sqFNcNIc#s9ptu=jTnz<+R<3q8?ofLX|S z#UPa19KyKt!9Hs+lNu0Vq~afq{OMTzT2U_U$5aZ71*;)$bS1U6Qo$-`?fC21S_x{9 z+p0dEO)sy9Asao1`%A`5!d`58Dqv0wl%9KkeR9>*EHi`l*$F>qFy9JSPce9?YgFzv8LBiFuT~{ zLG`g(M8jKis#m()iDKRwidj!_hrCL#5G&&$+Hk1{e%gcIy7xq$X2M zIxX9Jm2Sf~TCYkjuI<6s`J8usz2_FjmD^)wn=PM*yNl?HNIM#LVWi&yKj7y(pxelT-c@8@xxi6U$;WQF@2M=^H+`hFqDGt?70zb1ACZ`4p*BGgkkTqsls z($fjQ0p&AMJdB1sOQoh-zMg89El)lG2*wZci&(E4ies7S-K?3!FO*+?a9L#X>&q%6 z^Dm_Tnl_M^%KQk~!}6$MC?d>x_2lkFBtSj^{y0muGR%2~{NpZSU+%34ZY|DKad6ZYpqd2#LkFpTX`Ywo_3SXWvbZUHlHROO%gh`@{OMIazv#Vc z-H_$9$t@7xVrL$G_qBTQ#$00RK9K0O*Y&T9`p5lt3#eOEgzF)Z`5BGp4nRjU(MwCk zK%l)V@wxLPLxwsWz$6@GUdH2A7$GJBiGR-$?HcG-w1{Kc41y4a5r_h)Eo5=s7k;@4 z&?T{boue>N3qKV9YkC(Zs@GAL&e6z;A!e37w-%j>)rt;PLrc>(*{htk&rwM2gLZIu48$CpYbReeZ1y}5y`Qr4o;q)R zL&<2H^7fkgBe5!F$Ftg#EX`p;UqwKpKF&kow$|f&M%!%O$7;BiFzhtWsaz+e+`83x4*&aKyl`R5}>jY+A-kat{-MM z$PWscU0_G%U*J8(tJ}5bM}e}xr=2%5&3-QnFE)$U`Clc=%U9@KcY?9jKO>rQ$Nu$1WmlD!8?l4MYZ81m)U<;d&JX-w{;tD38Uh` z0a8Humli2nb{tvqQ&4F9qAhhCS$iv2E%EAlQ*Kp1_Ar^_+Hb#;Sh zJVw`YJy=7b=*vc7m08tq&lIzKjEN~zSpGP=|GvIu{Rc&7;5VFQkzkopWCs-b~G57p#vh@#`uU;VUz4HQ@ zOZ*PY9A$UaiN5gb!;&76;0S`vK(zaZ(nXsr=(`1ibfJ*++=dMAPEdpOpLt1j<*G_u zH(kg*o2)60+`3@XXhHYhH6%oV&3j#UF8+)6wwGcGie$tx0=A0zgcS84d97yaaJjUE z$J{Tn;G$ADO~z830#GvECAHM@h?l5OsPh3>unXyap@!}_wQewf%}z?oiP`Kgm$C^==iYU?^7KSJqUE!kFSgiLFG z+2{w~bI{vsWzPD`^dkUCwS5pvFYnh?us<{X9`FnXI$~*ciDofJ;st37cMA(^u9kvR z>)U;@TDh%VqLUYO^l)JW`H4<}R7$x!&A?+Y;xHap@b&Y=7ooQeews##ykOf4Omj=?3Y8`j z2-UJmK9eX>*d+(=P;aft9Nd#(v`Z;>h*aUU9a4xwU~derFflMXIvR%l6Gy?M`>Eg6 z)g|9LZ17z->+U^>5Re;`Q$?BMJ(tGp|MwrUoT#CwBPdE_p#S)Qb%u|11CS@jFDWc3 zL|3TCXfWbU)@6^bS&Pe+m}*N?quU_RXc^iaZo@7lu*%ckQg>J^YQ0J$(Bhi0uw?rN zIl=C~X}D_xFR92bW!!m2{Vq=4o5X$ul0=ZD)~U;liGco?Fsv6#Geu(klTx?|R$yap zC*32A-#VgJ#8TfW*9EHW{4QFIw1>)Y_AvR%rkadhRz7-SeVFydG$KJ?8doB|C`th4 zFKHllQ-p}etf||%v6WhNEn7mmey#3on&+bR;uCvWCyw`PIiOrvS zvke8qT&`_`_xpQ9ExyLl@u%|PdrfcdC+&&_f9E)W^ua`-9#2>$W3XRO-MI^pPzWc5 zUhD7`pexN^io^$%&1@M5%p8q_*)fXlnn2q?35V+mmBdg>NvP0`22V_b->As3lh6kxEnJWt`1q^pL&1m^u;NJam zYazVL4wdF#TKj8d)pn3~_zdkl@A=Sq^e!N-h8@K0y{Lkv#JG<} zrh}viPk$2?lgVGLT;HmJ#TD7yBvR=^+7YalP2WYGK}wAO>0JCvR&U9|FctMI@)kf1 zjU#{J40wGQqD%M;K|8GsK|N@SY90hr?j)fy8ZdU5EFr#j-iOn`=kEP=W0V;a%^iAM zI=2h7>7OK4YDMl#=qD}=f$f5T76t6Q?j_AspZ5z1G)LR}1vRI|5RrFx%bLe(2ZPTx zq3dAZvg>M(b~Nj-JQVT;+ya}K`Op!R7gN_rC*~nKDjgxct5+`lqNTLTlump}$n{+( zWvMO@twzyAT-WqfqDC^+qGhx_ebsl4;l4h zF$b5*+%T*`PA2=U9jGkCx2w)kD+Oo|Uq&(b4_SbZ31wJS{khQ@=f{`hVl@quCOlCX z0v);gki@A?H;St{Olocf2kj7Cr>&o>5JhW~pdTT#ceA1za)ubNy*|4A?M8ViSZ}M4 zlGY6T7Dx)*h8Cg4^`(yEro!TAaHSB|g%-uWxT~`Z6gud+bf*ow$vt(U7oFpttg)KQ zD$z1k^crR-+FI-B^;^}xy&*K$#?||g3mi-r_;j<9AjdO$*Jq21rW7K_T!w$BzrBI1 zLITv0gpv}e&_}-svMYD|nt3GUjE{kJF5i~@oVv5GWQn|+#@NUdq5e#T?DPwvZMVE(| z2G(=evaU?pzCez_;}PdhnzpfPW>|&@E~8|U;8a_!$nUV{q`rv{w13rdf5=Y~G(d_p z=;nOs9Xq+flJp~u&I_uZCy39=ZW{eoo5liZq;$4=4P|PUWrd{eTxOEOP~X*HVO8x4 zcWNv|x{MmhB6pwU9Jro!jqCMve8J573d~1Pg=8=xBxhH-Y|8h*{|=?{O`V_vttCS%t0p83*hWP)2HpXg^Q zo0iTps$Aw|X)R96I*PRVm8BX=pbK|S;_kA|S@9R-y^YaIK{=`6DUCHdWM93Jg)Ng9 z-Hh^GRUV=fT8-XaQ$r8`^fKw&PHz|P8KCP#aTkSW;zrarGf2sJOby3(aj)Sa?Q$gF z<_3LvGXLVxBoo~_v5h1JfqJl3GHNM@rWiOrUen1Y`0Ga(e-}cF?jJ%2>20z|M;H*@ z`+iHGX1u+9TAe~Q>AQR6-8;DDesYR)y;G$}mAxfAMm&$-N(QRGf?d($uuoz|SAxjM z^%akt`*@Q<17bY8>&FwP%y~9P72V0Sgmu;2Llt&CLA%==SMpkPw(`Ahq(6Qve!`&W z{2ZyJa6(x`T4M|Ha_bAMQuK?h(Nr@Jl)}v-6f?)A;x%iM0jW`@FcbI@eC0XB+sF%((Pu1($V z@BE3b`XM*`WuhdFOZfoCC0(aWDOTS@?5KqHG+NCDSz3r$&(;*OzL%K2iGh4>~5{#?LXE`mX=5=W~M1J zR@vv(`0F$;2b+AY?38B2hb;PW2AF&jrxhI2qZHU56_RLjf?$;bP#Nus3X423^x~@f z-=Rzi?C|!`{S8`8vJo?>Jt^@`VU$d|+@?|`82(ltU`VHTuA7LaPtwS#f1Oy9MWh5{ z$H(BYJn_N zId>&#NI47dEozdr4hnqgrE@688f71ELL*JF*?P7%J_)YXHO75)@mV;d{AHNK0q zQO;~zf^pe8N^%k;+m`6F`{8Ac^CZqMnXbPd=>WFLSeT6A!`p0a9U9~nqZ|~1uV#KM8)ybTwQYq4ZZCJ|x6ikRXcV@F5N=2vcpHvl_rVdcwRbuE!MB{zJMduyS z?CWdjuu+Z8q0|(zkZ>|!fgz*x>7VEjEj+iUJAU@quh_C63Q?CbCQ@q9VLUk=EXMAu zQon~)Xhy>TDQLf{gw%s_MZFG@9(VMF+TxUvR7m7`TeHYiEzFAqIDxT?lAjEk?A0Zo zsf`u1a6OB@M;1qk%&sE{vL=^HC^rY%E2?=QC5KoG?a97KIZAgO|7`GaV^1vqabj(r zV6O@0MGvnGEi_|-P;Wl>CnZQ+sEtWkBKDLlMD$Av<)SLic%2+#c8knY&RP-Nhelj2PvULdW3Qp4_!Z{KEL5C*EMBOa5 z&6XM45giu_8z2vD6uekL#{D);`1S+@3R{0m4DiziH66c_h3C@yK=O7OTg-=R2j z!HPtuOkE~)I}wf(MU+1wo9*0rPtRt|<7ho0>ItX~hLZGs*^Ax(_pD?<69o4pE65nK+=ULm1`+oKWES32mrgVEq1b^uf zh)Qf-@XUjjgnVy6s3%30g*NP6PNZ7JzRcB*SJhF1%<0ucP18L&t+fPq*VK5rU2Ww- z2fD`jBByuY~D7P9LnA*8{&XB(9U^E*92L+doPfG>{QEEaY}& zrd4L8-+H|MHoM*%XE<)4G|zm*7Yv<&Tct^4*JnL9$#x*%MQE;D**J`5W6zDZCzh-d z)(CE%%GojT?)rnycsi=q(vr>K!_AbKn_Xa2(c9OWV0`vX`iT&EP_T1Q-Nd>@G^NK= zlcLA^Y_2ue%b~DvFe452K8dX+hv&hH|8YQOC?+gBpbz?E)#x$p4XOzjrO;d_g$=hU ziNzVh;pR9+U!!sISBugy9a6*GTI04=yPEo!nI|GV_ift@dVBMD{p4X32O$3mWh9T?1+Ad@)Rac8$E;6a78bpcTPUDG&z*VFRC*fy%R&XiwPuVG-t#@TzE%w#vdOll@j3RsF#zVzW4Az($8E#nL zWOS(3;rw^#PXa^dX%`&n*2u3AzbC)W12s7hXKk3RQ-LUj9I_eiEX=TN-aL8NM|wTynHYsT z)%}~*kIe0>^VdvON@Ak%>v>1Xnctk$b)8$~>gjPpHc0xK_2(kl_V(?#gE_YMcVvu0 zniD~O7x&C?Hk<$@=n_wD`!U9}?8tUTe-z)R#|f*;%2MHmH9ot)$O9q&#T?P3@T>os za&D+}@;i{hkQfC3K55hv5p{z$ioYtph>#W;i08IxLpvidlKr%oQ|YoQoe^6CU1g?v z9?!aR&u}s}nre8CpYgG_oZ2O48`YdanVf@vpsFxKJ!*pT4V59X9!QXq37}3%Mb$UN zjz~am9b7?6iQn$*J@2NX=I$XZvmy^rQM3qcY~RiGhd(;$2IrTA3!W5w2AexY@?o+6 zh&VWSW3qUaqHh(WOeFOB{V+L<7kPCA4-SLZ-RU#?wtYk%3u8QpW%u_hV%ub6;c#+r*{~kBmE9yn%a(l!NYrwu zN~%V(;bZPVB?OZIiQcqx*UJ43QpnsVjGuJ%NlMhh!nu@3hf5WBFl3j^&O>kipP3db z2E6iv>>)|R*~(b~5s}vJacMLy&e zmSsu40gI=Bv9$#6S2_E`l2*KVONrc`imSQpkRz#fcTGH7j(;nxBMEIQ+VL}uqBy1j4p z$g?zXa$Ky1gKVUR$)i)uev{Il=b@zeu|CpfXp89g$0dao0hMGe%}-%95371rE#{H| zQzFl}vv&%QaIV)y)4z$R&iYEKU%sqRriOkGywSq}jWb)X`n)t9?h6AIKxuKMw+bPd zbLkIBs)f9hCcl0I%Oy?`ATQ(6a}{Yq$!A34>FFvIfWy326bZFUwu z(1^6&B`wgz3cj}vXGJ!jRClcP_WCz9S=4bJpKBmyvIRBHIS5cei>y^wGkntEghbU| z{cA`wnr)6ndN~3`VCB| z#?YH_kf`z20d|ncJZ96%Qju_&KU<_w`hLX(CD)r*vQZDBlEc))v#^!I=hhpmB#_-5 zX;~j@n62+-*fkx$@pGs~B63(5T8sJKc;3oqD|M7vw^W_EfyFoZGibRG4dIP3oXp=ax)zab1V4uEqF zK0CW}x-NrhwvRqZN<6Bg6!6D40X0&v2wmyMWj8h!1fl`NTKbkKjlPy=p) z17L_-57VtqCs}8lDG)}reV-0O3L&wbot?S3nC1Pyv$%j&%}Y8J z!A*B2D1#rl-kx`r|76%=ztM!${ig;pdidLEP9Eq{aL9X*q7@m_v2-2y&zV=(&=!rj{kmTAV%Q%pFy0^x97lN1fXcoadC!qhE21> z7_|Sl%0Hc${;)znn=F%k9y+#~G#X6XoZ0=}y46EL{mubOp$uOjtc^JUByE(VC+Yh}dM$Kj)R6a)*u zlSz)_7Se-xoDO9R@b@R0M3MLx5CvcD8F>K+1m)=siB%eYPr>2$kb4qZP%WW}y=TGv z;XIM%47F8|zbL}hs3%9SaPy4Z4A{ty)kYUKzirRRcn%l7UX@Jw`b5*WJvv%>J}OS! zS2bElXC8wa1&Roe$14}UZ(58HT6AQ4=SO`*#H1o&E<>*0NUWsrt2N;csH)={mwsww zt|S-4K$qz`L60Fs@f*hfOaRes>estiuyH`-82OfH(7nKE?UL^Q zLXJ-UnoPB$!dd*q$NR;{T1(HH4ew#9k@h#zzdSw)rBk5wf!knq2C=&3kf8pb7{ZPT zdK76H({KdX!89^r>c(^uB&thm@|tY5JgqzqF3Yr1k1fmV;D(u|&6o3LYh6OJ{h%71 z2@^KB!QV1(?d{xV>-Y{~XHUaulkx%@@3CYCq`qT@HNkDXlQVG-u4GM+>T( z0JfdCz3HIG<0D$KH6)LiBYrBnmN?p$iTbW{O*VglErq!kyw4;>YTx2jTni=Wx8 zn#Sd)OW!XjU%k>4p}`9*efNb=m8$Yzi;}GJo1>YUP4~TMLbrjhobJ5w`+G;gWH7DX z(>VB38e6We=hHMI1+0+%RW|p#aPx-iv&T<9tDc7V`UZx`yOw5AOVx60#~buZH6Uu9 zx3EBV-L}K#?o)|%<0a~)-RbV2x~-Coe(>O=qoj5>BKUq~zd@^e6;Uo!>ba<@?veg- z(bMC(9RQ>DDWhspXIDFLn5`bKsIg&#)buXXoaHDis2_f*SWKQMUp$YYEte}RDW|u2 zTeLp6TV(WWD!*5(!c3s?^xeB2IYA_7RMfNaun}#lD&@bjG%v426~+9se<--6cDB5u zpaG()EK`H|UdSFtOipoTg~v#x@NP?W`_hu zwLq?R&lYyO?Ubg6euzyrF%mf{yZ8mG4E2ooa``il)trp;OaYLmPO5^+=VdT0rwm z+0(Yf2mED`LSQ#2Fh(2JLz({Du-m04cn|JXIBUd!4t@odyzJ^v{EB6f@#K)IdO~rW z_NUy8DA^6>%=ff3zr* zG={4`n$foB%@wV|O%ybj>Ao9Ip?}NFM8=?EDJjAi=%0QLBls8~F?zpC;!k3Cj_%UnZsM&c2UUFA2Nm@&_lgc+f zAe_v!zDZrm#DnFff;fubb0R&@gNT;%z!XP?Lj4=>B`UMN6&rO+f^CZlWV=Ibm(7K; zeC;h>EbF>&Gjw8Z2gR`(D{5|BdIIH3p@wc)_Ify$Q##7b;_%BZIz#HH!&r^H5%vIj zYOCG~wl6iK>mL9P0-Qns1Wf|5w|0aBnTn4;(l8nTfbf=e;awv2tuEH*)W}x04|~B( z;0A|Nb<6ML?QfJeYC^?R>+lL%?!a5=$_flEzw9C$HaW;&vWHN3&{t;9L2?C_up#1@PF8R=wANJ7DRp_Wb8fi*&Cq_ zBb#0qX8;V@ztog25U>zJyFM(GNSU)F1}|d~QUQ0AUe0v>Nig#JPg;x<#TIS%znjv- z)|ZXGyzf-6YQn;T8c`YDHDlu4C z{81*E&dFFKL18KgIn{$Vjq-^zX8k^(n{<49oG&uH;SN;8Q{TRm*O*p*h{xfc*qPHN zBbiQ!1CX&NAkeWR*FLulr<8(E;gCq%paQc=7-Dx3BN0hvNerccHsb*2AQYg!L>#a` z0M>$(FK`b)QAnEYCDQ9?Z`}H@>E(lPY^C>-p5?k`F_41u0-u2Cbq!BE#65>PvyHti z`n;9qF5Sy#!~RoP6c4eN-tCd*bhW`UTq2@y$|#=pH2VF+1b`IXS+{_+<6 zQY}X^s^3bR&na{%*4_{cr2p@Rgo->6jx|^FW~zq(!4Fu0~LT>hz@*kjM9wh zZ2(|r%joW-TuTs0ZKwGosf9}VV`~vB#xv^6@B{GY;*cmF_%f{W=(J%LM0EX$X@r16 z;(Wq^`rxh~`|n(H;lJ@F_k$qEyuSaGhWwTZL{?fI+tdd0>&(8?;@LI!_`G;6HMs{j zF)nMDBvn2yKwnScXjNvsl(?+jzCAQpl)MM>iQ!jxCC9Ht(UGvBsLdr_lUa=robg}c z>jom1&|ny|bp4Ej1ER!YIRlA{zvdxcBBZ=&zTz~XOKL8q((BT#EENMLaaA!7_n9mz zxAiY2l>iV3_@;;oOv6r>eJ~x3cWP6IX0Pzdwy8Wq61e9OKNU_Oex546p-6W8bzSE+ zCuxIlFP;6oJ@rKju@<1cVzf5xj*)X zva!EjJvRiVbbn(YfBIL6enI|4kq|M5TXPMC6mN`iU^h9$O~Ad&bf|*iPkxcWKW)+V zL)fGLWp2~wntEI~$_K&d_a`Vdd|Kh*03VX%FO38lPAt_MXV4EoFRyvJIeNjy5c>6u zpzNUG$V0{K_s=*D!f2xx8|i_VKFeVk2T`8MyBB_&=k0d>D=BE#ox(`7wRH$QVCi> zQg%A;x?i)>@f^g@GP(xRNkK9h@*oL8=N$BV|1re*JYvR>v9 zAD59~$2$c5y*1Nvd15>Z*V8fjiyNB0UhY69)g)^5Y+ZG(L9!Qe4|(4r$}TY@@BVpP zOpR{8O8ws-7Ad1}K(sz3x#kJ^qGWWSdKrc=!5{}r(Y~e<;c%uut@mDC-Wul5y5Rk! zL>4p=@`Q#*-eIaRVJGeAVbBb<-c);Euwn4%%NEa>s zXi3OkY&*xC#`*K%3T3B|jogYewHOn=rM{C{(i~a;ecbcqq2X}J!t)UF zK9~_-Kj*d8yEd?vr%}un|QLSaCRNFaxYi- zvS2i)w0xXj`3y+u3{=bq-V(ZPn@WGzVDnN&2z6 z#p_Y=w!ACr`&{ccxFz145aonecDC>2!Fx2Fv-0tgIHPG%{7+SOH7S|(6zRjI%f`s5 zDp;x9hNG>UGws$=LU4@>8>&`rf7HZ*YdmnBlALEJ!l$RB*H?!Pr`;YnSk0 zj-c}O6LF@bN|f#ph{!P&gj{Od^ZXz{#LHEW1mDc3NXnQS&hs`Y)G;0=z&q@Shj=xLme97+f!o!2*1!)N zWBXfaq2X^fWDyv_vTo1!h|m@_wyg)d+BM;b&>GQVXmgleve;PwnkgL)s*PgtWIBcl9!7n_n1Pz;Cf@9Q$g{{`{GJ}2WcVUV{3isb-*68DGwrAE zH2HmIn6HBDYj$U|gSk)$YWDre&TQHL_L~ga&m#!{!whkKO^+!&kY{d+JBlJ6g8pFNTJ+W z99V9v`(3O($8sX;x%ewUDl-NL`f6~E7O+5_Hc;e|PDaK(nm81w;+;Tvn3Z zT)hanq%)=wy5PE~l5!XEf6;V~(Q!3g8=e@A)g+B=+iq;zc4JQ3xUtPiW7}+O+fHNq zo9A8Y`!&DMoU=Bs``(UnlEvc?zRP2q&zd^_2rENDic&?kO16xpI@yq3$rEFl0#e-q z>LwD;C$aNY6pzR2tsPc*W~aPt_9QXZmOtWQOd8Di?jy7uab2cm@~b9n zt7pHqnCGO8{?J1DR@hU|J3pv8XcrH8mfzEp#{ltn-9>7AA_4r)~ocjq2f*B*FO7JZx&R0i}H z-Br7!+cS37_RW5s!ehXPe)ZO4n?7;51wosJlB#k_A2@pWTN|2E)Ti_&sofvP)ztVOjss?E~9ac!0KS$){5 zqP%p$b@#RsL-xfj=ag*c%a@2$mt`{4uu!xHDfI&QMpqRoFi2t2s3)AvDwg{P&zeb9 zbQUYcRY!*JJi$?hnR$kZRIhq4+p9NL(BUuf+cHthjcqIEX(J(L0Ec9v_pMm*M~tQ( z-TryD(BB7vYp+bdG9+@3jM!lY=Y;tR)09ZnY}*c&g=)Vr=y2Rz3K$g|({L}aimkYAh3Ge^C0JI{o|8BDgOGSz;s8pe7ON|qJ zB%#M~YWyvMs>j#z&1K4vRxu8thjl0M314dW3@!uM93)_;{FN3jpGmPer^0xu!#Yo= zJv;48zY9=gBMN4K;aPYX07!3UK4OwBhU}vtR?tEbZg7$0dVAwxm?wcFCQLN@Awk}q z;i9!zRYzP!>oG?A(bv@Fq!4<0{*9%o2E0<^RAiir=EiQkSRRIQb1=XZUT6~TMjZdnl7vDwwF~@D?-0BXt|nKLLO+& zK&!W zOAeO?pwQdB&OWQ*)v)PuZPU?XV8?Qfn*M7={P~rj5sY#hrag~#R8cC1_~K(=XE|eY2Hp@TBQ>(8af0sszRS{ z2tjTG$(@`WBSt2A0xB_-%L zZth=Gs$0+Gz-xBMFl;lV&@A8^q?QmBdZ@z>la+HL;-axxWM$Sn2SpWQL@OO|gu;jn zqS4l{f;X0{xbC}+3aWWgb4svTSlb%2)+??y5lpT7qSxe=JL;2qdS4RVOy)9r=;gQ> zdS|&d1mV?H{WTTkh+!XHJ@|4Sl6X$&%o;8*BA3YHS!>P9OzJp|OoKVjh1O16$~zD&N=vgQ zA+&5i%toh)@fAM;%>S|Gb5zSf!Ja%5R4k%dpa}gM0g2#=_qRXoZDc^o@?ftm^gcB0 zJoXhA5-f*ABj-4v>Wyes}%(eQeaPT6}PYUh+-)K+#}7bNn!>fB>YsZ$pm@pqyw zh2Gk)hGg4jk}`)Q5+7|$x;T|u5lUSrNL=#I7VmlSm|MqtZ?|4I!1g-Vv3b0&h)Yg- zmTsHI*00RS)HwDJ7DL9mgTd>C)ve*ZE&X{@bES4?pXNrBSPlhfaqehBpLvtmEQSam z0^`ACfJj?@sA@Nh?tx}wgTVbCfSgKY60Hr0cL(r0K4FrxIPdG>j=!gEn08Wt6>6#fref4|d+tPmF~EBD$R3 zy(%`U$4Sz*lB)FHHGwaqz7xODXSpf6v)e{c*nR(@u8Kuw@zU#J>Q5Ij)$R?coCrq! z*r}d<%3p3+TKS37#3Ry)v4k>(U*ZXA`>AMC`*;wgRNGZuFS6~AZ}k-hOwpF9bzDW2 zYCXKhV*@;}ZT$8M-`mfhGPKt25;?YH47V3rig1#spB5E^bS8b~(IU)1okkhyp6G9E*- zn^Xwv%1*N`iRx_1ZObLgH)`JCv-JcXm6o#VVq}W}GTB5y>XEEsAyP23t4w(f_A6{d z=%@IrepmMFwsDApg8I7@$ZXe<$cQlqhG+xJvk*#1`I7{2vB-Kayk@lBwvtc`r{kpG zhp)aE(o)eZUqFa0U<1x4}mGskV7$^wU6?%4D*y>rN5NbcV8Eyc@5S{s7nCia=dQ{yDIyXBWg%MU7i5 zY@G%FG!RTBo$?Q&44_#ETiTPfF{I1*g9sXtOfncQS`FV(vlhbYRJQY%1dKJ5YCd;O zeiY<8mz%%+&~J3Pp*=PCJB{Rp8oCAu>&-pQ zA)+&(Ykp6>egOYz!F9u9B&?!F`FEZJ{Bb?Cms2y2qMOodtk$Bz*Fv`=UR4CWs1#~ z=9E(kb$FLwFYLaNlZIYu(Zli>3hy~bD{I2tz;XhpJVE|UE@Lez5gkTMDaTtwb!K-h zjx#&qvEpSE7~BN>g}h1bBpiY)uuWB^utr0tP(8Xv@AQokzlTAtO;-MK$>VDc+4nQ4Q!LrJi@Q*rqFTiVYnnX zC2oph62qe|1nLkg64{eOJQCBqCnGWOC$JN29w zcT$icQ@>(l;mQ6iVyC&oRgdjRN_q5Y?2j5u^2;)>jk|7Ib2wpUYI?P3<^FzY7g%}? z3zw>=Q`Sta^|+F!?z__X)Zg*h+RH&enH7@1gI1+mL z9;304Z%b|$>mh-Df_iw6W}%o#&4nR9dtWFjC3wuKbTWEEIzp_Ps?`~**X)cPD z+&H4T+$iOFL7~6fw|HKnbb@}bOI+mX;wMGx{!#Dd&LU{NZbWPpJx8swcl7ln*eBe^ zd?SlsNX$!JNuA%>_7LIzp}KKoZq@$8we_p)luKB<>*DrlPTs|`auN4pfZugDMh>Am zbZgFN_%ZmCQX(y2GP3Y^e7EFYxAh!RlNhRN^a$^vrG$#?%1O=`;5R=}tYqtW9kX8| zH?&c3@-JLry2ZS%BJ6kL{bI>TQ~C&xtGk`{HI7kE#AbSYXGU&Re{uuK2X-2qcyk0BRijg*+kz0*!E9i(nU!xQWd zDp=$=i5gf`8Rq!S&ZG_G>hC2b?#-9^j6I7AaYbiKpTme0aR?<0rW$FS&Ok8!CB1|a z3NPAn!8_(svw7Ca%c1E8y)9J#+^y$Lj|EA7VJ+_!iyb^Qm(IIRs{`_`i8u&2KdcW7;dfHaw(pa+EJ207$J;P=}vSo8L zIeci^o1xY|%M>Od_HCzy+K`(PwGg8m>hV_`nr-(IbWH%I{f41B0@v1D_V|0MsQr|% zl4UHjj=W)%WinCe5q*?FU-xJmAh*N2grCi?rOLNYMT(r}0f`|*&<{`a46?f+Mt$6V z%(}f55Y9^9b$uW)Nht1S*7;Y{#kMmEeo+p9-K2i*4{Z8Extr3X+TLwTsQ>x`fiGf> zzdeX6FD+d4=$v!Km=9MlN26*it-1J?he%xVag{K+TkcnmeX?rsN%;a=!!G3jxiU%p zoDL$Oel>?i0S%5psr}&86u)dejEy^xuB-c}*XWa+_6iL32C#Y;#LX>^WN)-llaee@ z=G5o$x#}K=*;UEz;TvzG6MRv_Ix>+ae%=k?ayxnUxTNf^`DYfp+%(IsI6Hgsw}S#N zy^lF*DXh12qiO2fmi0uncjKZ5pDqm_`E@9LPzA#uAT^Tw(F{|o5%noqcT|KubY5h> z<&UyL+0ktoRX^swE7Sw>8n`P^MuB5P2;yUnOZ0UMiRrOhN657P+1U9d*kwx@m5WSF z)R22#_=kd=j#-GP4>0gB{ZC%ECEgx|Gd7#ey5Jj|+YKI7kQJE;*w(5sradMB|Zx-%_ zLnIwrjQf|OR=`9kb&Nfs zPXP_-f`J!e-08m5(Q+AYjRE7Vlj9aT*rb^*Bg-XRbk>-kr`08fNIAYYyEm1i6#{7zOOP-=TeNE zIM+KgedQ@FBF%0IlW1~->OVv#!BB_(4t-s4X%3Wm9l-)MMJ%0DK*tCxSwJI$bOAu>{hM*p{O z7(BtOmO=zG@RM=2QTqDu>oXM4m$YG-+ERojm>%;37=P_+s9JCg4l`z?Gu#hDPOxXQ?Dp?Y04Q zkcCP#?mcWn6vD3Nw`a$pbag{YZa0u5$j6+dvhmvAV#p%BTf=%C)Mac1H3$9oUG(1e zK5M@-ca9!n!`g=qHd!iSu-HyJwSuZY4pHbQ)9u!Xd@C9|&m1iI9qD}o6+kI?7S*Hu zX_{f`z;TNkR??hD0deJi=Wl?J*XnH>7$9T`k=Hf^@!h6fbtgfd@)24I{Rsj!5*Bp=$+BnuU~rhJb^e0y%-+thY2e}1w*IsSKXW!aVXZWhYIn5#P4U)#mbAJ- z2i$~rzM&xl>U1$(v-IETz8#{mBT4bTG%Cq#C8-68J!5r_D5EM<7AXX7t6{Y#$0Lf6B4Z-F+G6SI!;FxLo>r-%>rBdeed4m zaw^c|x&5N?d6a=Z^Y-I_zQWzMUXYi-vO>N>5XyAV?Qx0JC&*SyHnP1t!@A|LDQMTvmkSCpZkIK=l+j+W1|{e&60-r!5jxD z(eyI-&+Ug<3j8-bYv+WQ9Q;p!3HO9X!&ItW#3$GgANe1yM*ib6khCFMxM%Le#-b1I zOe)vIU-1FoxaK_q!9?mg80qDh_qEFPZ#u6VW4e*?O#`Fj%V7GCFty5K%imM$(nSfk z-zR@6=23wFJRHj4x#U}PlT+&vl}zWuUpTj$a^b{>hYq)?JHNWQfuzQv-^bAObrmw< zl_0>`&}|(5_n_x6Liu%dZ|QB(Bc5{WL3g5Ar>rhP|BPW;V$}jNdLw2x#Ek3-bUr&|LLcb2us!n(YQVzBK(u--37Be!WZ`SA8aCi0Z#sB@pfZ6Mi|M zB%5YP2O+Bf@=P(wdOyCfjfmIvtdzfkel>pq>Z-A6#vK^QjHRn{Y?6w!JGgYk%~8`8SgJK~jBCer}PFrK%gpQp_tY_{&?5f;@6P;B~pT*mkCO&(N%C zW0LK2&I?VqUY9q+LU(ethk-PlL+!k_j2cZI}_TkjX19{cg%QS|O5`w^(U-&UYuxT?Q3M#%3TNCzRvK(Ia) z%>S@>dpN<3*8NS9kmo7sFRf-6z87I#->?<^q`jsFfXGzZ@M27e0lfO zJ$N;0eia6L4DE6+El=JM)gu4M(tKRQcPGEZ39ncWs5<|sja^0A`pjRkvC25_h^3&x zlaeMe*%9SZj|N!mW*{)5D!49!G=MP5f;VT>)d4qYOg6|Tr5XTA61pWSWk^}*imRUy zJhK=_@{5$~&i%+9{=B2vM|wxJ!n&8B5^d&@;>w1a*#A zK{fm1e+H%tzW`*gI5TxqBlv+_qCqq=4nIY%T|qseG(ZuyJXW%+5}mlXK{+6Rma#N~ zSDCW4XGC5R2j*X@R`@0c0qm0BKAClww+Ry8AVAYnhPn6g;SF&si9e^E{G z;%}Y?5e*D95*o&}9Jq!RhnCDTj!Y zcy>l9@I`G64E!0~fDMqg9Esny;RHXTUArZS7>?mYAarr%Dzt0+9@c*-F{5tdh=Y0t z|I=)l%||SOiJJ=Ze;<5YlUFgtlm^7?q2fAlqd7fJ68lnG<0NA%r z;9UOb6%FD<|0-44nbP7*1lA$m3xs1QN+Hgx0b`5Br`4IP$Km#%c@vTYNgQ;^X(+qB zRclyLl1fo@iYJ@D&9=;Gbk}|lbZffbxCSZ^9XHah6fmaVE#yj(DcUA~27SM_rK)nokhmPB97dbEZHrvOXzj02nV;#Vopo>G@o#Js7;?YYNUbC*+VSWX?Su9mIR z^R_q<{%m9$RV;dJo$FcDipLWkWbs}s`!lv`4!*N&v&~L#0dlAcYCe&uz2xf~$ z)CLS5S8l+bNkffv7YxrenZhz-wQ~1bNmv>Vc8!wR?DobN0&DldW)%ktu8{5L0ue%O zMq`{5i)ipFcoPm}KLDiFGOdstc-9m(1HlHc#fu?HBBOBS%R#dE-{i&$rp>zyXDN0z z8=*X|nVE5n)z4cBV{5BWqJvACC_+|QE}{9O;t73G1jz z0paE15A;V3i~riHoC`1~o!`YRM(9Ot-Lm=@x;*fe*h(KRFHTt-v#WIZQq!-?FEquT zxb4E{_4TjTvO1p3KVIbq4;d-J8x+Gu!M|1D;W}aqf7Tt2o&N@cu8D5ytf>h`84AhQFSYos=Qc0N_llQPIH!8KTfZ>hEiChCR9h5P!i%k)?&VgYnt5{4k#Ya&jd3Df2sj!G#0BSJdj?syHLGdfWw!nmquiPtGtPWHii~{3QpwVQ>T2Osialvx%(+X(@}(O-8-<~ zc2Ci#0F;xYc}|Fu-+_^&s@u8n3koYituBh6j@#-3A5(nPwN#SojFGoha@Kr<3RIUN zTUo+da=3^>{53$$zZ! zTKUZ6Z&>n{gk^sNHjxys>6cO~eoIWX`sPB^QxrtOC_CNpo@I~{y zyzSKhLMVE0JYw6+K~NN=qVSCa5>OJ>ea>~KDK7eRgPcuqMxhcK5iAv#`zQ~f&3?9+ z{Wob-mP4W)+TAw0^PnG7wY}|*&EAn)QmNO+%5k*MVvLq867qihGR(eRI*Bd^c1pX! z`U=CTgl4)^2%}jy3joUv3)^;-;fzKse{Tp=K)Z967B>aBX{ZZb|OST0T+4z2BJ%Xzej~)oP7A#3!^}yck z-Bt?&ty3>7+AA^InySX`X@XsUu>CDI41t0<6UovhE`0EyJO`Ikg~wTK9gtXPu}b@tbkYpLp#ek+|Mw>)Eq@ ziI9A*VPU2%vle`0v zwLEz#Js92V2gagHUg9?cTH3%%Hz=1i2j$YNNjdPT*M)s+#R4w8o+Mq3`(e2 z5XwNzF4AS{Xv5kap}(Y*5cc@IJL5xz0JimSJA5T%?mD^`G4{%J$%7f8OeyRO<$(?k zV>l~J>yRumZFG%Fb&_?aG?xvX=OllGWKc96ZmvIjj9Mn0bbkadCnSL-*tP zR@}6SZ{cV@z_svPZ+Sa4mnKejD( zPDYbz!`Xr}lUp*0nTqx{1fTW=IsmM6zr?%`MHLVRVxG{L#g)bFgBcoJWSA%U|9-o% zfW2^3Mo@(F?sbWnxEg-H6hbGCP;1>_PWic@8M&Y-jL?8w}$fKT5!LM$@VQnOrcAgw(yg&{0TV<|4-x9<$_WZa^ymgBl;2#2r-A@ zkp%waV~t9I^rHgzN>NIPJ=-?DB-1SYUw7dvSZt+%v@SYH32?3_s|^T*0BYFsBr1t^ ziGC3r5><-qo0*$`n$yeRhs$_BhI;_@oR~yU{A;i-C8Y5bPBj1Mw4^Z`vyct=Y$)Y& z9xu25Bo5fH^m`w(gzM}>CZ|gfNE+}l*Wp%A$%u1(VMzk0$Jpd5!IPjrtp65;&g|_p7+w8C#5A-U}a7H0kmBx({S`a>Dh1@BCMMH@(dLK5hXn( z1@HrK_k>5^Nh?b0LgbwWjX6bvCnC&okS~7!y~1295$C>JZU~VZaS$2gS5auP8Hl=V zk|F&^yH$GV#2*YDy{N>~bO`d1H?`a)=QqFYOXq#Y9NSj@jNJA=@J0_|$u@CqF$?<= zHc>3u&Y=np$9Rp=T7K_RZ92S>kKqO1#woLPM6)+&w(*WhYE(?&ZCZ3;rr+Y|ilq6i zZqZ0>_xT374XoEZ)7}IRsVQ$J^49Z|qb6NbYTL4jn%_cJ3?I*nU$-mX$ zZS)cO_ODYE0J2z<8%?9+sAEhJFt4hD^EkN{hwRiiAUa!GRSbEEx{2pT^!IO3MucD? zA*qE%fJF%D;f%tGQ;yS40lP42dY3#a&;2KKZR$tie|kc{&H8v*>VEN;Rs$cd;2JCf zAV*RBXCF@N+xD++>_72rZQx6BpsP5-e1O1Ln{c#;UJDGDk`J2x7m6c*6W`k+!^WC{539fC-zEf9=4#LB*`0Jhn|a2?JzvTEe~vrO)qt>1#c-XhNC;7*2PZ=(c1#Vlad45I0hlhV- zBB?9QD*+-Ry3K9eb!jf8=p-j^eR?|Poc`+sm6eZFgkMznI<2?3kWUJ_e7GoPh4b9X z^FLrzVz^wiT7E%UN}KCVIKKCM@uSNwOdmNjmB6(7p_%ZrC{?Cq?_;uUUlJ!8dmB5F z&`b{6L=1`o04Z?51Bdhkfh1^cg9scn`ZzFJ1xq&h%~r^oEQSz%_Q)P^Z+aEBN7t)D zU#A)cwEFL2=3Xs~F?T(g8f9lqz|HirJYO?4u!%&lNm$P(kuecs!u-jAL;(HC*gf%2 zl!sU!BNX7HCq{em5)UpA0YW-4e}?^Os_C%I}ap0QE|g-b_#k5+CPs z$3LGmf5IgPI4YC$K_G%i5yuA{P^!+sg#i>J{G)4oj<-Wxq*1ok}YQNScd(`v*c~vx8L9y23HuZNo3LEwR;9G5`y!-9O z6$J+y;B&{lIS3b7r^W=pK3?-2OJV@o4k(!53QN(E8~jXDGH1yHf=RoGE_w1k2A$~} zN!CaBTik6lASldW%p@zKDj4975K7*Ttl|r<9V8ZliJ|>`$(z(mB8@YUCQjrOfG*VO;YtJ4gw`G=IUGU) zu|*?e^QQ`Rjf4sk)tq>k@w2_ygS-0|NQz{W7q7%Ffiz zqXz~#jFoykNI%R>9EhEP{v+2puLWGQ`Gp`*U10Q8e9w?3zC^t21kj8J$|NG5g7$(6 zMBIc?=U?Z1F=298U?`4Yl+s=)o*-abPi6ktnP?2bOmSDU786lX+OSnRB=bM7a#}9= zFAzPIKduUXeh@F-u5BTJ7Drn}f$UZ(jM#v6W@Fkc-=^%lKw;zW{`L;>rUR?rnB4vIPc(BnaF$Oq&`r5gX3 zKXN_Wtu`y0H&1mjM0(#S0g!0BVXY_%a7S<)q#i?(-`aK~=>_B(XamN>`Zls$s0Pda z*MNgNw|&bFXLM>}ARGzwy0gg<;W}IvWL^u(`KVFC2L&Pl2_l8U|G(Sq#2+K0sZM|< z{SPSA)jGlk4DDfqbuL@H!uHErPO!(C-Nj4oy>*mz&|r+Eb;xZ-)3BGmp$9!QLAI6P zm}21lfu_JE3{^GMKL<#%11^w2l2Mh_A3KXQ^xw0y=AA#XT#1qL%`^^-;S6>X_!akM zeGLCDFK$sQ&o0L%frf|z=Y?*-V?*rR_pI!IQM|70Gv1>t@tKX=7;pKC?jZ~*0AdU( zH~~5|RfZCfr_Bw>qStm~UhIRXMwk9kk$3iuwH9MIOMe9BUWhLs0wAQVHKKLDjkJ8t z(19!k^Xa!xl0nd_m>>Z{FI0FLVJklg;fXO92>i#P%|gQ#xT0#f5thJ)l^Cq8Mg&5< zfNKNev!N0$_n%U=;2O7hvV$ggGw97ySVws0GDei!OXzm%mmyn`B%-62v*0K~xc9{E zti(D&)fzs1ooMu4XkI8(D{O>;hGoN0CU$a5N#h2&g9MuzgH=o%$p*SRYB3oYI#k4I4jeAE z-?fIX_h;vy{#&2aYI@E77q?o=l+2wHnP?BKi$A#<+00hTfAoiEe zUDuEQq}da|v^!^@2$U)IkBj{1orUiKUKf84f9XWjbz3X#fFQK5p^{iAZ-lBws5#09 zt>(O;V6t}fttQFNmXK-RunZ&g2+kruCr4)Q9l>G2mRkbP2`fsgO4HA{qok9RjjF=- zB6`us48f6tF8;*=GFnX{EtrtMnvSRbln=8Pubvz+a|a(-R$}?nta|%OWFpc4e@JEL zM}cT61;=7l*|<@{PS-Hhz3B>>LwHp&fCeEjygxS5i8^Vnn(Q_$N~dk6Di9clVa?`1 z^c{MlhY$zihwez0ayg1N1WA0TImX+=vFqByU*@$)y~`R8#@Bj@Ri)Tb)oJsMpwSxS z$+kmnr+!M|=WYw`!D;qakDXzK zjF_qfWNA$BVU*wR`wzrxgF{Pem5*Dy<$b861FcPH<$;heKFiJpVu9T~GJ0-PISZA&Zt+cu8);*SBgP*$ zV~|R)DV%Gp^uC&&wrc-oygMfyW)+)I?p`BNIP`xqYkDm0SSZK5H97Rjfz_~8=KZF#a1=DQI%Yfk4mBz z|61FIk*FT8PJ03l_SVIHa$mw%v~UUl~2a37C~6$h>JD4aF)Jdm?vRVhWM- zNo%jp!1q$zUZ1A|FX-c-MA9C$v2Z4i9GrYJGjFv?f4V;dXa#Nn6dcTuOV<2%ZO^PZ z&@9)Jn$xNeVvBBe_$nkp^Uy+Ehjh^5GIHBME*JzjkY*#;DPx205rH@T6VC=L+3v}3 zl$qX*$O=k|A-&2tkrv!PIh0#)kAS7)m)nOB_}1)*BdD$hki53$H?tHx9acKIvI7#` z(UHQ<=nTK)k>9(Z>&OTo$+6>M8#j)$PZ~VC8U0>L-v}rc^jlbHdVW~@hi0QjEv+|O zhoZ?CwaR8ty^fMM?DI_e9nTQa?oJ6;)=EjZdba zC$_i~TO1?4C-LzlZE6%rjI{VU9h_!8rt51hCO-)ig%4$>{J7znI~CIBpH;cOl5`dN z&CwZExCcIwPy|K~|`KfzGD`%lbxE(n+)s zCh&GdJ*+x9sTci^#H17l933(Kw*%$v@@t2|K9sL!we490U2-M?{hb7rClgKq^|93d z4f-GKMBsWN-ZPy2T5d+^H_T!yIj{D87s?F}EX8{w-rynpu6GArK>JT<>M}!?e|iGo zN(GTKxY%a~M`M&1Ld-LgDXsW9OEGIsPxv-EJ8sy>U8WkLDBWr?%6?Y)`+-vs$7JUI za*R6?B9OIAznb9W29l_@%)=uwGKxP z+_TqxJns5pigS^Z#<2?FftaAjEvNFX4^BPres?adi~S+Hv*-e#h1{j(x9Q-O^{8Kr z&CWsD&NI)J70O-a`5g43;KEx&-sfkwM{69aHCj7ekP&9zZqy8uRaBOG|9Gjm({HDI z>JEE?Z$G~_vMTfoD&ocV`Gp~%C{PbZr=|IVNB?cjL~Hd>0@ixghX2aGSVPDLem(|X zuJWy-|2l^T#`u{o8FW0hz=_Okk~!$Vio+dGbv3kErl77gheb% z6QBzg8UK%q0~s;1~x$qCwz9lC2M}&~`1DhR()5-ubT^gmr<305tr_I=1bSr3J2nV5C(#{LI9jx59CH_ZJ@J1SzI51^ zGph6L_Dt0DOXhDKe2Z+0hWac0NT9c-*5Cj~Z2LVYFo>~DnyNufaqmx##vkiWC<{hh z`qayn)jDByBotKdU2Db!L5>Y=|(T6$kh!z6OepE3!NV zR+-c--_2X1|NN5v{pSng557=#Ez@ex17EY_CAys@FrC#`F|>FY>;-pZtqoy0at9}= zYV84T&oZ{ZCaq7C6{RD+2@|)B23(yBO~|F&Wee&K{E!x0SCMgowIxVs(5iTdxFSkx z3U)3I%JL9!R}t#{NdO$00r5d3t=AyWzcgs&JrJDo^@aLyE~?EyR_Iy}i^>}Be_&4o z=5iZ;W)Qfrd-?$(yr%KJ-1Efj=#GmG821;cv|Rhj?zw#KQ_u6iIn~T0kDXu5m~`rm zHAJR0=KOR>VRZSe8Itzc_jUQC+PBiCsa#0Dd9XUE9XJ>Y#Ih)JDa(?t*%n{{b;lp2p5R4S6D<|jlRs9d)77}{vk zOmyKMF>{JJG;&`GWvpi0xLtLwY3A4>ngMOBM|@wSP3>Oyne^>K9<6uYi2QCZk*8yu zO9~wDE<_QhQ^Emt(w31i(m{Re1vOARLJ5q=*g}(Z;eDfkR0bVo#T{g4UtTbu|MC!7 z)2_FSwi9eXUWGa2SzH$P?Q?sq{(PzS_;^0@ef0)o6$D<8ejcfw>;o!FK&%>QHwbMi znrg#NpYfAIywlotwm!J7*$6Lk<{XNsk~svWhVm-6;n8upfuQ1okm!4ns)}z z-=(DQa2xYqqBZUH^OI#>9R4hhI$<)K8cqL<&2b87h#NsanD8z4j-EU1>|M$!LHHa1 zxvokDd|jjcls@N0)WDewKrY{<5%ttwndI z(@uct?4iBT-RA?R)YKG{zPQz{f~pDqiwZK;;ksi* zGZD@A=})G^m_oD8Q6as<;h_*eMl?izakImmEc!PP>4d-@-b2>c9IJp*$-)?v2BbLXF19Ibte_@zb&>X?)TXp@ovyQEmwX!V>h z8<2qrc?Bw_eT9|NQp>Wx&bgMcQ=96*suki`QyGf&MVX~>2%r>z{41^%xd_~GuRHsa z(<2F`&N4t%6Qj7~0{V0`_7aJhrAa$2U$KhMWgrBp(ptIsPmGp^Yhk636?rw+UiJnK z6zwX&56#lSi&8ADg&Ik<*9#3-iK-)kMdQj9x1ka05Aj4|nKVmznPB~)4EPqPCR@HX z5n%$aO**t)wQu8$+E+h0DU@>RB>$);+tlFoVBvlU(jN+fe^or{D6Kie^fmNi^->q; zj z3zhb?|9TAen765Uv(VGQ5wD8rlYPXKV)g$bS9%vffuu(DvRLemNR$6(zNfVp6tBGl zPUL^%ZRYXs*`tW!~-@#^l%PkuXL)#E^aqiimo$$GiJ;=mx zPr5=VU0fJO)jv+!(nKZVMWVT=Lv*CST7^;*dHM*1j-2_;Cn51hI4??u>SfT(>Jw;Q2Mpgy7k)9OtdfKUz$_@-D*0<% zUHqDendw+@C@bcRQ4Fi|i9cssT3Kup4`v8__zh>DtY5dZ`39vo-Tv3IKD-?E{f3=W zlfx;rGSyNuJupu2oHI4EQ;Nv<3+htdG64c)Q0zLroc!nlB1vJ!fZhz!uBmS`6vkiZ zq+<#Z+peIltW6`3Fqz$}M zXAmA zC+1U^Bgbb$OOt7*A{3R$4qH38N_464`f4({DocSbjFFrE`WFA^9WzZ-52ysmU4nKr z5a#CJmkd&gSLuM3+7DYD6jX0oL{?oWVnjl&wQp&o1~KQ7$->D3t|T0_`NvB$#I)rKLX z&du>KXN!~fw?5{ME6kqP+mgatg3Kn?%kcNyK zt!}FC%x}*nwV|lO;}<92<%&-4KLkv96gJ|(OAH%a1L?}*`Z3Da`sm|yzK((8ci@^* zEgD-9R|id?hqtYw!pa`SF210`jWsAu9J z5yue&pQhdsxh8_%{a9Jx+i)n+F#ozFq9`mYE7K;Zv~8{Ez9(k@$hN_R2T7%Ke4ixE z7fPN zfnQpV$&!3y|JHg%gi2>;C{J5*DOaGK;_xaeC4n2kGkt65OVtwXh5GUMmT7m6JDG!H z{@zI+o>ye;&$}NT>hvwkC3ioI@G>XpvCOOn4ty6>7VK8`+spf%^Q73&qF;k?mA{`( z>iDzXJI`n}7~KoCKomx?>dQFlfu)Gav8p}|xBmpF=CTCmIw z$9j5>NQ}q7eyB9J<6^Q*ggSdG6KGh-ImHi`Po!Ad&0DR$mTtnGt6>d$BTfG$n=L3& zRYE)y&pJ>=TIP%3&!GWW14K}V^kAU2%#^5lB3q#vZBnAb5ds-#BVaTS@=fL?5b=CG zRNU4Y5gzgnL9&TL&!Zwyn;o>Ei`-L5HDH$}DW!+QUh2&3A*3VB9oR-0NB7krFz$OO z2W=3HYNu%bpglGfdk;oH{hEoCG?V1o!q~eCd?k>srS|&7MIfr#fCcs{FU%lFDGm+a zp{9D(*`Wc;SGCOftj1x{TGTQtz&Tp@miJHMr1*B?#5aeepKDf&n?foAi^{%Zm1At} zbOsrQj=%Mhv7V;1$qFj#4+@-duRosoap@7BPX(HmYuKNBo1vsPl%|TC#Ok-C6?1O4 z`{Q_co#`Snjvg_uF4F->fu9SI|K7wVTNpA{#BAmyAy6l#_-r!9enF(T6t z({e^p!vpdR?@;}qKOZ0-jf7qY7w>ZZRgALZe;AD+y(eDT~F#Y;nh6Rxm5tjPJkaL~IBQ+so%UV#)UPrn9z zE9?;yh8RQX2<`y7+~_PJ6ItJ9IXJq|#-M@bU$@w%_<{4N8=4G3cJF2d4nysc1*Hs& z9Gc91UqI~UnY{Q@N=v*sj7|?xumvDyklI?^+!SP?HB+Vp230u-HKpJMn7jZ(}w}D5A zeVEKHu{wq&!WhCNFh3F*ltQ>vOuS1xt5&v^NG8FB9yN&Aq*YaPukJuDUkH}3TCpAk zJ$~<|T(eBcTel|d@TqkB*fD@XIj`@vG;TyLTY3BSj}wx`p)TZ)a@&S*gN~7f`(I}X zwjJu&0-8bOw)1RbZ+53U95Er9!Qeqbzrh0j#h?M9mBpRUdqu>M=oP{9Dz zTTKnF`n}Dlt=W7i|JdKSp!>KLzx)(y;E<~r{9dW+gDuVJ78h}KidECUHrHdXG&xuA zH9ad=U}PG4p~WlFG+km_HUo$MM*V3fq}6Ic+Vhae}sY&l5P-MRYD@JXKf&V{$|74B6G+E6Icv`qjB?IbzXgYVbucD|GHP( z?-<))n%d89h5)Qd=W47orG0z1h?s$9o#UD+vZY0s)&38muK8fB!I1rGnK@ryO4=t3GG;j`{OP8B5v@~ex~=!o%&JYjpE{l#Y+V;2TOSZz1aniC$>rO#1681wi`sxo2>?l`9s(i}tY%*cy+Y(j9n zG6rJ4>nT2Dy>&_dtxNjrSC|#()w}kB$T*uAQ(?4HRhVH>IJT%mfg13?KWT=FDNr`< zm~?hiw4IU&onBZrm8p~nf&q2X22C}jQ^Lgp;Y&lKExc`>i8(K89_b<{Ojbx^sq+KV z(k|l(C82P6?B~Y~nJJz1{f=q9gm7Q_AcJv@GyW1@rT!)InBi2TQp=x%-X2xV zb?htO*ktGRc#VFdB5sIZf;1v@)HZu#9%hz1q(4EjB3L+C-4CJ#LHI}p9ka(_q{q|_ ziya4b;r5X@?(aZ5;nloHM+^R+`-z)wYC}#o8uF~e+h5!|2G$ZrJ;+@#l5V92da}Iz ztMnM%n9%~UOc7vVluz+~gbG^yE3#gV&0{oVzDC>lOSC7bME%yS#e=Mh<@u z#5I?P|0_(fhji(VRjZLn6&ybqLRMOC`T>$+#I8G02<|Z<6q#trI%MmJl#YQ33UN^Yh9krlAC^ zl@j%RKgY7ji2(h$2fphFd6?kJbv96Kk^uOU&Y(IH`>eAXPCa0q1$B?inlz0bKJ_t zL?7*iE~TyYj71Ha#q_al69X$*KSV2ig=d7}`NnzjDe=#vp&hDNfyMHKhiXFe2H}R3 zdJjfqec3hN+W@(}4Uk*M4uu^~rUpoh?kjgbov)hzyADXc@Q{*$-p`a!n(5@@+q2~g zCp+vv?0%~R@!KbKFgOY(j!SL*X&Xm^$EtRs_H{Y#i+LF$S9(?|0_qAxHt3$$k$S3$Lm?XHz;CP`ph~BCOa>D z2o?O_lFNJ>rOb7=aDATYA43niPso|ykJ2}(zFl(DogBmoIh@-qELBkYtuJYdI;vnx z;N4W(kQW-a!hc3} z^EPHT9S4qf@ErFADXbRd(#j$Ly!PqLuF#uG`>_Nd79%t zu*|H6AcmZPA3N;7Mvo!52+B*|a7m_w_e{7j_x+?GX8SY}M1@+?!#$z$(G?pjj#q}lwwZOAQ*&4N_?I>Tqz z#EHW_^CrQQ5yLZBCLX5GxR4QHS0k1wB<%zqhcPl;)IRv@Q^eM#o;s*R;TCqX=qV zdtOuZR)5rS&U2_rhuBF@`|1W88Z<43+%2lCzb^24jFAX=J_ouDn*@`)yMAL^X7;#r zllPx&*;|fGvok48%VYoYJ^UN5N7X-OtoIX6nsEem(I9QIm4S<{Nz`*%!NJmY>OOoH z{f$H88i~|`I3Xb!D;rOY=z>B5S$`qns-yqKm^f&R!8m6aGeqJ7jby_mbjD8eEt?_E zq-M2BBrL@9$0x184CT?W6CUUrs~CGuo4OSCYWJ*Z6QF|Ei;SG10t%Cd`VXDZ@ii1w zEhUI>AgtbwT5Ym~n38VOaqbaMKuSBLx~{z)KK}XeWx7O-R=+ahhKW&Dh_N@9?u1)2 zxWa5!QlkJ;eS(}l^0!(zGcGDA2~Ai?txg0c5M^V3LzKL}Rs%i_#x~pMFF@q*EHN<{ zo{0{X4i!QIy_O>1V2`?BSvF-~&O%wt64x$a?Zd}_XG_B#HHX#j+6mTI;W$d2BL__NLL^m5PzkMJl*%6rugE580EywPo93y_TZe?A$rs7xAv;e%7nj_N=bn&mv>J5DsO9;O3f69395I%iM&U$#vg$ncI z1?kETM92ZWgDmtTS0Hh@4!U**HBU|kZdZ3zk98^T`@BJ^zzSdMsS0lS+x+>>O16*l z%5CScN50>*qq@vhN<>COLq_WJwML@#&r`~kj5`<$jk%-jS_)vs{!9Iht1@`TN1OD}qV9%jO{NfAbBpJ(jb90Fa=R$49&P{Kh z>&!~3B1q0jgZ^cmX*nS-!r0y^6Mm(ZbMfu1p+%vm8;>*(rJ#}fH_UA%sp z{Mv4yL*0pR;(w-2hkDHLwqsWbC0A)eevw~Qg8g8lc z3Nw5c$e8AG>(Tl+csrSr*ElEIDvj%mbaIYCjPMgwP<>fuh?26oK!;K|Mtkxslt7mDu$jo zc7vw}5-Px>5u)e|KYt($JA~eY+xorht^g;27J=AmKqnYlh30YWrxRomI{wmjd-xV5 zq*4ge@>rD>(v=_d?HNqy8`g?5&mb0RyLI6Ud8tW5P!XU+t3^7CSSkHRl1exbsC23Y z`IUOaU-IWb_}l*`=2;24gdVecMktYlc2Rf`*@*R+V^WJ7AuT1ruEqw+dGNK9$@C}K zWwFTI-z=q6=YlRA0DDO94NHcbb$r+Sy;il{6L{%_8-+m#Fe>r8{8~%Dsyt-E!C6qb zSMw7O@1yz$bI3Nojq(VVNJM*%E~mr~boe4rKBgMk4pSnoH;i(!{o@zB_v2B67?*Tz z;HaVhGh?{7u|X<`<5LLrRdd{k@B^kqwOs<})1Ih8VUz3?9MoMOcojl$n7n>)-utV)?5Y$!`!z+7{&?-pac8X(4%h7)!Y<+0Pb5Adq%|4P75eGK5% z&l%)ewhO|=QTW?z#+f36^O)KWu}zL9p4|xUhgs`oC*0$cTr&FNDr;Ub4(M{`q$i2) zbW%2i|L`F8*qKS};6WeFTX*slJM!$JaTxVYDmA2X03#j;ELF;XL9(Fhr9cd}{8@U@ zd7BgitD&{EXl@@SNaC&v=Xbrlkr}=VkWoPNX+o>%@Myh0(L~hZ3Ib94Dg5-n=KvXB zY&hYBF)RjykVN*%-A7n>cE7$vk7498;4A5UO|3Es)S2~VAY;^Bl_=EZrNZp|uIT^1 z7$zK&74~Z}8YWxzbK0TiVI|x*e{Qi9Sfhu`{*~g*PWG)SfP!F8F9$^~w7)aDl(>z~X zm>M-_gka+!0RFmZuIV;5Qh5Kr%|t<$fBxGziOo7&N2Uhr7Zk_1!09Uqo75>cIlK;w zG~Gro=$<)v7))H=@M>hLGz!{Vf>}9>+-_6lg$}(&nB4y!ndE`%JpOIEFfEv-?uu(Q zO7wxMsZ3#{ud{_=>|o!UD1Q?BrpNwEVis2QB zq*#qNeVC3)bx*M$mL^=@bn*^y?>AyZw{M!10jZ@@(0%#LE$-e$KF#C9hp#Z5tyxv) zzzIP<-j9F)PhfWJp7HnOBylE@6NSmO0KRzr7Us2#%jA4;rM&m$)>#1wwG?vc|IecS z8z~elo-jMZfkA0UsX(fkOd8ih2VPdQ_*9%sWGG3KT@{Nt=G_R7QDB{S4!Zx2_hY}- zg6`d=R+HR>hwV&vCZtqeWs28!CtMa0@woYl4brgw%+`RbGD*g0qg?zQeRtbCkwqU0 zB5n*w3Ri2X^cWgtINs3WNBCQ{XS`RXt2S#rF+9GnjO`s}9wxJR^`v!?ocx1zp{zc; zp)-Sbpv93HUdDObPl(KeaN+#j#FdNt9++p9ljV8s%+)A}f=0yP{pWjd#@BO8C~6pfatpBxorbM{Qeo5U z-DDM-kV3`eh)XH%%GSM-w&uiYA!X4 z9X{H|hwyDufi-m|lj-TOc}X>_>x5_mzFev-kEbwjneck}>6{GpyIXDE5-0!~VSDGY zkvg3$xsW^dAM}pYHQ5%(_cI6A`4(Ubr~+yqZEuIoIQ5ef)%!KF9v}mTBnKwQ;PyWA zLXeK3oml@@qk=KALx{eClP+K5eD3pk*6mTHsCDBn$GVW8F-yY;oVg_DO&u_#lPl$y zZQppkuhrbUeb&v~r*Rze>f|zU^Aut~OOff%2vo#nN(_|7P)~X8CrQ)xY@``o84^q+ zmV68GF@G3%{53EmMlxfqG}8X(#(1Fha!t%8n+L@_(>g1#JU?uxy^0Lja&iw6Kr^B) z$mFAvU<{b_+A9qCc`Au?yE0O_AbfbeY4BOj;p9QXNzmowfsbZ*gfN_)zXy5@2ypTU z{N{hWDieW0^kLLX@?@RRT3cEHHXWN(vu!Cny+p9+mWdc!2-Q>Oiv|Eow(%m<<^AN# z{@<2&=S0sJKHT z@`lvA7-3&Jq+GACjA>fzzGovXyI$=2(mj%s)AYlDB<|K-|MD_TcQe){CVjE^UO8%p zK1>URpA6?ZGmGojIlN;=<9#j=GV`lDKaXoa%FR}%M+@+Gyyw2*y^Gme%QpN9kWfGJv&y=aUo zWYkG<6sBR-F7DD?40@!5s(MSuO7157V9+LLF$|99anWmc?|! z(J2MjQ?p-TS=#i`x^ll%aS;+lid1GIeZ!CtWjvqI!PT#*v%2H_PjIH5)wyqRZ?8w@ze~G%>YO&B#E>&P~}wk=V`tyJ>+N$ zL#0eZI^TNx_}<^7Hda?Uth$6dRKE0CjSGxuI~m&zE&!cuu3W*dVVHJKr4t6HiR()e z#V3>8T>lnLb-H;MAT;a2y~kyQq7zoeRk27Dvunz%T68Bx6?N53>MHFXB}mtCrWzZS z8i4*oZ0umY(nC3~F4@6s??n*spVL2GO>uOstHwV-qCM=9aX%(}|2&sUkvN|Ea^^dW z_Q9i?R_XRzcUHTM$NC8MnegbZGsQ2o5BnF-JW`fg8G}gglZEitkJr~ZQNq3p>ZZ8S zj`}$~%#U(cS@k+)lI4wwC*MpU!Yj?Wd#w`unseWlqi%GW;L0goN84^R8Fl_EP%FQ+Uxfxm`Kse53HI`s~zs&f@-eL*2)= zaXoITpa!9FT4~2@nGSJnY0iNm&m)NZq33GU_-ksbzC)8k^8_=wQD#yR1iHvIs1Rwq zwK*)6=sU)T~3wHijv zUQ`)JaO(;JV&*2f?Mu5*HH=P&y-k{6ignE51M(G|I@TYqmkmiZUnQ5zJYgydWaqbX zt1cDYK5mtZ`Xsq+3(@)W=ovj`q2vE9kgRFWFD&ou4EW-PM{H0z@|)nxFj*>qAjAE} zK|%ip0fUftIW_TvZD$eona3Vj$AKIPH89IrKKGl>zdg(Y;z*R7!ed8I%;=ZiJxA=O$9$;n%(UM^km3&uJmmas+^4!kV&~@r`;?=2f_RiCq~#;b34>w{ri`SoiPR184=5@#fp+y z4D}$l#c@HrA{WuveR3&Y))JhL&B+6F_?2#S=+u~l$U;IkuA3bQEUn{5AfY_RL!YE| z^U5C)ReSKfCq2wDI{TL<2sikxC5E!%rR5_^BP)jBWIt%iG989wlzPU_k+lGlqTcD_8(vy4_PVuiS(Q~P}U z9Rdo&6E>|6G=`sg=STK4sVchH?I;ZDhVPp`r#dWUK0m+SnK;Zfxy`e^UTq#C4xa>T z6R+F8zp3Vp7Z?KChv$n@9#28Wu|mc+jZRHc zJc3_4ia*BQ$?o*)BqOS0Il0?RC)IYz@DMNfIC|@V8F9b2nrS(^?DFt7%F2Zop44ia zu@BdoOSOwsvrN-l-;CDs94!p@w~bj%WZ6mTn+3=--8cv~aZqaQ^MMM5%v3`tV-bb5Vm&0BW3vmDWG%78x?T0mi z2h?+im|T7Cb2R8v&6-_ZteregEKOz=WpK(uYFZp{>x5(&=!_iK&B+uw)IJI+%JmL} zi59eF#3-nsIyD!h%JewQiEQ=N6>aILtwECDsJEG%wMu@oeMk_hiY2(p{1BshxFB5D zBQqtt(4sb$rcW)-N^LZm%@)erQ&HOR`Yqo!`ucK7 zN9V^!A~^x>=J}+?VZ-OPd4?9ok}}(6^1D27gv0rU!*qg2H0|#PxQ}r99?}YhZnhH? zy70&}0RyLfW;baEiW9i+o0h5_5{rgmFYz<<=Mk&lVZ-Nsks62t>HUOiOo-v+LqbBD z{GYuIn(Q;6(5|w({56{3)D-+oet@bEER#0s>xIajRSa%&vJ7XN!z7`ZWo>$pdeXwW zZQ>8!8dJNlc>&k)kK(&sYsNPmJlzY6=rViZDBU}Rscjp|v5#`;3&c1B2(xk%E2Ju= z0}0_trRQ>UzHa)l!?CuxI>z|b73auXZHM?d1ol@XskLO>ImP6kH;LxTyubz4eYF31 zK+3s2T0x$pi5$?Hr)ib|2O8=tMb7|$uvhBue z|7zB*(=y7e`^Q4@a{r7-Tqzu**~ts z9`pE8H6^v0UFB^ATTG4KFSdbwq5}#aQSyU;+8e*4+q2lU6jh)*fd9NBsq>eVN8XB! z@HyCok;s@YvhSY+Y2&w^0AW*pbd^Va$JEg|b3ItYqhLzI6KoXeH4r-eSsv6l88U=T$Biq?tl-w#_=eVr?|~P5)Jbvrg+k&BtnK* zKN8BtZ3nrWt|XZm*(O|omP!>mNRGzKzCMQ+rOUx9X00qA@?aUcV%ZzqgIIUi-{Cp9PZ9Q8NeYo2l_LfPRz1ksaWa7kLZ<>S^SH<&J}CKeLqn)OoNO8dI^N$qF#qYgoWxI+7$kk5s4?c66 zgl>4xz)F_J;aWIUds*yLScyfCiRBu;kM`l?LbFz%Fh&zjvN84;v%VAPHN|i;kA`|1 zQ&Bpp)4k?I^T(}Ll!8hmhjaw%K~O|*aRGv-8+<-8E|j+~0KnJBZGS(052KF<=(2@c z9F;D7HA>%bA1mtaadw)>M0NY2!tJn%?~)^q?p#qdCc4f^SyDXYXWNd*?L1$*5*2Nu zEu(cvR?z^OGwNzVmSO;{zE-fkE-%+_fg>Vp!Wi9TN6OuHW$DiZvseiY9rF@|<_iBR zBc=+oDhiE1Bz^OD%RC30ad^vuuo|`WAG8~?e-Wf)8=oUQ^@!~ASGW{XTfzO&FnNib zj(qsZl=RQh@Qf@U;b!-jj?VCJux!KnAwgKwQp7x>?ZHDheI#pF%of*YOKEK+;zqh; z8M;8II}D-35LPek{Y2v#cu{!%U@jb)T;!aa$VX|ekdkuwH)Ysj{ut`9x21lKuT0Bs zoAzW*z0lDN^(VaS8#y23NW|aCW}RZr6F}?$OxX+@_>fa%T$zN)d{6|kqj6XK05PnH z@qE>w$Xwx4nAm*L${M=+n%zXh%!F6#>eH&p;-yxl!)>s;&-_)cjVz^GK@FPFq)wBg zLk)Ft&o8l2VA_7F@wW(b#If7jkjX7XjfDf>qil+-WNvUki74H5Sm5M@r)NdHG60`e z!bhAU#tqlFLy4Slac+4zNIj)p32q>!*TO+fYlK&0drdRVr+$r3o0z#SJi+?L=j5or z9kk?+5lrnw3)K{3k~ zL|BEeKcbg)yQnT-s9)(SE9>fRlRIajO3cKZgR{fKUa#~!|2__;0F!FFK1ilFa`{@i z=lH|8=xniG?`Q^h`#vDW!|;e)m*(IaR=b;4r8#cOF7BPg5dZ?Hw8Df@trj@xLkGcH zexP6GZdiW!?Z@lXO?`XF*EZ{a|J%}iX(sM_B{&QA#rqXK4{*DhwU%nY!~+Fm7s zc^9*yVnk}c?*`$3`x|Rt{=WX_0x=PHXRfrKM~TI)fp)9I*>s@Vj9s{PTlro;Ou@Zs6_22orUHuAE{zY8{7*GuY*FNgm`D<;B_H3)l>|Lxrp?%jP1xrGtzEi*LO^rFhy_x9LV z5%CX53wp(Es=wb4-{^W+k7*hcmzF15vCe&9fw&fXr&Kd_2Vw-@6ml1R2O?g6DVcSC z_v(f-aj#%%fphw5-Y&1lu5KAfC>WLFUv$;uFOUKL$nBM~{YfJ8I5&5AA zW zl|eX^MlTxlKsb|`YZi)=ZhrpdnrHIMUPe-qi^ujUdGEKS%>~H)0g!@P znQ3uPb3T-o(4KY6np=IPBfZ0BTSd3euYVSiI>$?-a&BDX-8sc+b6g}*=&p&dBm0X< z{@1t#lGE}_$H1kk(+I_xPd#}M=3&aSl!r31pp|p%TJ((p4i>;2L!U8u8zIXaP90XS z*0B`!dYD8ghacf3rqS#D9FN+IO+8+bZ7Y@w0T{&}P}}p1{Tx0ncb-}dKRDE5cq?Ai zvFwM6&(h3s5E8|g2Y_lIA@#sw8w^^}>~Wt;M*c8=_*Wcd7{+nSmonE>?zG(D;f>Oj zYc$083AN5~ocR{#$E!u5ujF@LYrUG3M>RR;X8#4kWsm+%mgAE@YX^S-66cuc*lUGq zjw{WJq#*H9YZSeNk)m1mb|%GGv3u(Ouue6L0N);Le;n0(;*dEbD0ZO= zx#6ELY@Rzbo0xUD$y_*+E!<4wIgBAQd0bjBEPKOg^2dvj!1jD|je}k-9>>B0aLBV; zT=%@WAYTN9V6>&iv-l=zuy+HAVFlqu?l!qkyIr@DYOx7(D)_&%-@E2Lp={$CtFNGS zZFqHkYu>QSMG|h_^;+4X#TqP9Si`s6(}R{)rfmPqh`FGaI+S)jNejqwu@DP$T%$T zQI8{G&(VbuRKLWDA(CPBwd*^>cPj$Yk(>X91=5R5bY}%B(RNPNp`z0R5`>5Id_utn zKOe5lO5$km079OkUBS&VswiDFxlaYzZAq6{a=0Y;7u%nx;soIoi>O5$!hIvGc}QRv z)U z=weZW5+w?{JQ)MOAR~QUV zAtoWO3*(WAdTU)mu5GxcB~QsYU5{JYK*Q1m zMvfD4X5By;$&iOXgCDI*^ZZ+LybsIW2Z*izpfF@_NPPPWWF}!h02!8JBiu|otKcDb zV+L?$+^xTytRJhqY^xlLzHEs;VZ(a>y5Bf)%kR?+Kc7L_%7a~iE6Pk+Xu%H{xJ&dO z6IawLFi5b@Fdx^U8SBSTMquR|G14EiYjbf34vXd-!{4y?hztfFRyY8fbL_JiS!!5R zZ2$T+@p^m`#+C4KWYkrS7br1(wFYnu%((JIx}gy{U?kmxny8{rLIL$86(|%rQRuX} z*BENYg4SQo=4$*KSH1RfZqIu$eKJimW@pgki?-Gn#Yz!t#cH_+dhde~(2jq0K7;MF z&dDW6URt62S>I=cz}+F*_-kv9R~Tkw*J&I89K1jf|ETviIsjfb(Hhw|m8yU&oXjO6 z1hY*cUlG15sLN-nZ72yyHWh>1hFJ-#PjT`d5LUlw7@8_gb{5HXCG4e6a9NCzTsJjJ zEJkE0hPEZJHpmU_xW#IbjUo~WR&dC`H{!4>zV-v9HPUQu>$vuPY4&vB3Ksn)$|zya6vyb(|c)>dKmqB^aax-vb2`TS61xy3Xi&XmB^V70f>|Z`@3bf+7VoZw=EWJ9QwlShz8RGEO=(lqx~y#QH|>8R+o3vU-Ga2KVpw z1*X5qVb)T^BRg+ruYG>=6o3xo*f3WyR2ctm1j1&3Y!W~UYnRaIYoeV}j5PLZ*80-` z6af7EJ*}-T=rI0eKmPd0lWo}vq#lAPM!h zEb#l280>B567fREwY_b*>;DG$41H*xp=de_C!gizKr2(^h*vcwD*t}Z54TWh>y#F;VG(*s9!e?bR_#j>elgcct1U zx<19CyLf`^6-foIssqHMo0Rtu&tdSO^f0S{r@??@u^En}{ zldJ_AQb?#DC}a1aWbhw*kxZC_2xPPog}n$=zWb7tE?C|p8o6A%vw#SIaQbYV&?$fC z>su1-_{T?k_Q!yiy?|D3q@8{`Bo05Q*+J6Uk1eace*A#aFV|gFVgASsJ`Nr27sfZe z(e?b%iLM^g0-$PR(k>ZtOwp$=HMCuXUBl56B6o(5T}U4{CSWxuyvALRh1T!m;wMg0 z%a~mB)u65)j~0*%b%o!*cigc#|y^Y63 zn4HBPBHP5jHH2#hP=yNn4&L;9LnBX^1Wx+Pp{be%znq7Q9+SLtPDk-d?Q#u7=%EB7 zex%R#{2S>XCxG5y90*?`1c(x|M%#d=W1`&nBpZ9T>T}kf%Xh{gLe2|$KW(ue*QAyOSsto%s$s zv$i5D*F!~@*GA6hn&$ZZj zX&jChy|ch^a1oQHQKdHena6Ljo2hc0wpl$E623-ES$6B20$yt6K||e22U3B zv=LhFQjp@gj6s8&>?M-eP@d~R>SqbgGwOhWkx7gUA{D)eUJ~)H80BZc5Pq|@he5pJo!AnFH_dM`bD=x;hfpd;fvkjOsYK|7 z4CklHkkhf8i>MNiYm5)#4rt*W1wU>BG^8~J zNg<<;+h_vw_b-%u*yQ2=&?0P5!|G!M6L!amAu;a5|7e#bIsMVXaT;>|`m6I%$m5f2 zR+=Mo{CbZ0y0I3)Z((|A-WJ2f z8)v6Rp%?(69jN+JiFp!vQ3mi}|MI{MW`gc4qSh#5Q)xw1NL&C|&Op!8nFO23>(AI6 zK*qmswnB=qfel0gptvrXJkOa3V8SQ@u*q{dda0W)I^c>Ttc~&LxyZ0Kgm!hX(}fj` z&l#*F+O{};bSI^61CpWWx&{o2b1tf&0Yu8)i6S`NFGffpWPUWx-G1K>iczDObSgk* zGEPR8EiI5sj1Ejma(tIC!+`xoFaik>>5d+ z;-R7l6N?*y=P<*!10@(lLwaDDL~rl`al?-^+4u-AA|N7(RVC6P209iId?>8MH$+9a zE#t~?84}e1P&c+KrYsQf4HSKQqWcb#f{(jtAvZL7#qfd4@g}we16}rgde< z@4to3d7c3XINKjSDZYK{Z+uczDkD~puS&9KJ=#5hE>7jP{xouc>7`5EbZH_{!py-=mHQQ!KZF8?1*ua*HOATTP)IE#^E2dbgs*`ZTmSHw#< z#*-5hYVWwSSm4cgq|Yi~>XZsXL=6qx{o}coV4tI7n7k7C`HC%;v{3cmh7{=CkRK{PY(<#1FPK0)>+4F&8=5JFK0tN!tCNKf&3rD2NuO)XmB`MN z{6%a|Y9BZF+LED$ZKho0iCS~{EK{=*w?l2C&x>sEiaImrHSdz`)O4GJh(mmE6nKg=Z0wJB!HDY9F*oYX4X_ zbZ?%4lMi{?9Y5E_EA+ulT+i-h1{@Y-EsLHVl7Anhf%B`YcB%Y_fBr=a@(egr#Bu^0 zT7g~paO+7|1Ll=KF*}^kLv#!Sf-wiKo<00|FF~nW9fw41*s{||WBao_KFi@y&a^&9!$+!)Ke7_Qiu7+qP}a5R$+ zAz=O~92^|92PaD!Erz$i*tV&)wNu~Tx)jNCq?6EA8qa(rS&6q!c-84%d-t%MBL?wG zf|BtmralM!eMHpl=d4&cO3irWMVN-J|tmq1X&6V{o8_1#YZ`zJ~`$2Oc~; zJ~~|OjzP=Z+zHf4n)WLgTp&A_3&|cvclkr1yKj|7obC~a?m{E6gl7XhJ^zprE^sV2 zIV`uqjHh$ zLsX$__v0Sk6X}?>P`IH};7TDDm}~&K+#45bfRxYijcM(*N=jU>bs$EL%{v)SO(wlIL_PCsWD*8bmsGX^<-IEHv$zI0LKHF%U%jc*AiHu1I&dPnIL2-28} zj`<4q4yDY8x%tMBfMW2|(QOShCwNwE41tj#69{LU4CnU@rFEWC=5KNdWB@RBJtRnC z5cfs{7!N<1DMjUC1HSD75{aYpt;GU~nWQR_h@xA=HpCl3hd2yqC7YPZ3(TZLLNGyw zlc7FS|C-}~)Xn-S^lI#fm>=?C+9d_qq1N+RaFXg@o)ac|8J~#tHhkOUA$1r1A@Owt zEfAS1n35l}Rjjg^aN_4u;)VP+y;~`qWKLoZ{1Tckmsx@>9SM716^vZJuGZGG_D9T3 z1@aKitv@3-j#bN^heEl)Sg)pOFJRAsTu7PFdP5x;GJTg*)v74_f}ESr4W}x-Hy%%O z96C_+k0YH^tXJl*3k7>_Rs5aBC=!jj3T%52GH_(zD8i}HoCEhJ%5}fnYgTk$k_={z zs;|)ma!z@O_Mt$e&sp#sd7ZyQ6o^dBNq*QNl!~zy?=vM60exc&)=q4~k5^kzwTBXJ z%VpR*KM`|C#4?=sK+p1qAgvi{_(HV>4!cPX>f5g<-S%XK-3Ue!atL{1NjanzmD5Zh z@#lQm!Y^Uk;y<;)0kcAfgeNp{*rG=9Q`i++G5`RK&?hNzO^mYF@h1*)Pg*=OR;MB5 zQ6Kmv6o|N$Vnxe@tIGeAzW%4Hj)Vn^x(IvZs!T%q_0c>CZ literal 0 HcmV?d00001 diff --git a/content/notes/images/Pasted image 20240404112933.png b/content/notes/images/Pasted image 20240404112933.png new file mode 100644 index 0000000000000000000000000000000000000000..7f26b7393b2a0967cd5125f605edb80c28450332 GIT binary patch literal 50678 zcmdSAg;ShO)IB&1gNESl1b26b;1ZnR?(S~E-GjTkySqzp3oc=BcUa!`|>qgtXNkNyEVz?4)`pJ6R;#Xwji@2lKlPBZq^7 zJkEVSvtNB?>1ktz9*3K@7hc}n+TOK~lPO23oAj19zuT*l4f7>+yykiX2 z>-d9&Uq<1U4IkkKd95>vcToqL0P5J;NSF09Y4I#!dSY}`tQfQr1zR27^lni z7(VvWD(Y_&Vq+W?>GGuWU4but&0>G82G;|B#5OSbpl?C8-lxzw{@w{i&Qaapa1N_s zuxJ=U{xE5?fV(0E=yY@jJHlTH3>v9=b>jbXE-UB!(@R^jkYAdWdNX^f{+Dv*1~t-97rVPLDz~0E4w2K7gL|NmNR)V{Eij0Y8a*b|7vyn@!K9D&J&qX3 z!dpwfpadTGpLyQDBJ;e*#qHCre(I-pK#Pjy`{>EPKTIOUAEcMF<^31ts-&F=z8>Zk zF0*i^_M{8UQjAgwfsgp#68D*KND7D(D94nIqm2LXNF%b3{Ql--eAD)4!2^O+DTgz% zGxDFC@bJ5taq4#C%#d^rC-Tb7%8b=`^0KWikB_Z-WQVv~PREkdu``(&t|h(m=-NV- zqo*rp1zI$gP6zZzt(|d& zPo+OrnbWClWy8{rpo;i!XiD$my$e=IdE31pQYk^4Z^1SloEMWgvHd+m_!<@(OoW5pT&k1y ze;xRLiV>Sq%ntyyUMyVwc2e(eSnD1|CQdofT(6Y851P%~NZXZ=voGco9R6l|T^E2# zfnxW2C3ng}!!01!V&FiZ`VZWh3Zj!cEq7b;O~8tGm+r^Qo2SnW8CkLSUtiu8zLM;HG&l<#M!Tl@%05BK$zbz1Dxsd-)9r?d6|7gt# z002S&sqexn?hvP4FqtM|X%8P8-4FL?8;OvtkOpHqG27^hT5=|kGS%=P+)j39!ID= z+5z4QGX30fqaKhG_EK^8QNG_T@rt(Kkwh7o76RCT^G=N>1S%uCK-W?zVL!$HAWD6~2(rqP3wjPglp|fRaRLkB8GOn5c@8IQ;dE|y%NF1kbpPkyT~*E2^O|sP zf>12T0DR$@J;K&`i89CgYa6Ys1(rmDg+LJy4udVdA^MOpL3Tkt+j;g0E`hLG{tN-aElo{NtqYi=j-UP&yNUi};oS`*wO*>!U=Y4U#-+THM2jZW=ka#F1!!t%odv)U))KgF)cvVAyW4w% z9L=K~9T^UBq5K*(ySaJ(6{+~Ur@7cmDl8GXOAzzJ{r%c_GOf+JO$$LWhI0li8!SR9 zNB@JvMc2ccZO0j=u=LjO?oeb^N5?u356`RC4qn;Cpu@&!K(jCb7K66A-pBLun-^?W zYpwwnd(I77H*1rVw3&92MPP;;Dcc^9Dx3^g9_0E@P0R5xx zifB&lABNnS(c8NRj(+C5U?O7nKk~oIW7~tb#Er)xN$Btp!lU=Zk@AEU{lEx; zh*|~9S9O?v;Ic7BwmHZ&pwaDcxmF^L<_8qF!zIX!ja*qK!Q@cT&TNH4m4_^ z&}U@xzV3Ig(h_x_XK~)*z>YEwCd;^A<^j2UbW9|>l;aZEVcI=NBeQvb28O}7)m3&x zySQlrc;$N`yLic%ClwdZySPp7dAfHli2h>QD@!R1gY3KA*@SrITbK}KaUO4w0}k#e zByyzZ@nmtRzrG7+4J?J-5ol0?$6pXHcowqw*+>@VLvG@J?};NWBu;DQ!nu4MX-pF@ z5jDqJXvn6tm(PH8h~9t}U#fb~a?1=L zlFEiYJ5F<;fTz8_-XS={x*$0d%e^%{JH_B}FU*^geJyos-TcQ^T4QF86+jrR{cyKv z<7AFvuAJlQJJWt{_QB>mNK4&Wj`{4 z{;tA2x1T*oHe`hsgZyVXo<>8MYv1*y(G>qm{Dq`<-^wslh+d))zZPBlzS95DE8tB@ zI?svAFTP~vbNBZfxH}#P#!%Tmd=^%eK(y=RfmBa~G<$hju=6IKZRKaWc;VP64}?kp z#oaPc2fT-S#+?x1|w1jF?>$e)#2of7Xl(^-a=- zeQr4Fe_Q^4-;FLYRR5Y6o~B=o3eD5Av+)d%3(4)n!^=*{`JQ3jPlU0NJKxC4N#1H| zn-aDVSgn=;`WEv60!QTx+AId00-M{v_GpNxHTOP|iK zP^eHyRGHWx?IkadWa8*XcS|Hs-yfTJ2%4)7rW2R?dz@uq3(g7ZXTP%w=t3r;K|yP= zvvYI1`hyNuR4A9|sPozz%QlO_F^)BkC>;Jq+EP3x3~`yGGf`so^4!8C1?+NCOD3qvY_fj3+C- zWGUGBp8I`60{;7_MG3=eyDX|!pF@TL%9@)w)bGyfMhLr+EPI#;hq8NDRG9Ly0~gm$ zzlp`ISt|_%#RLB{Mh2+g3_6xMultl|JE26)9yf<=>5P)fFQ*N$ZxEK@ z+3Xx*)PYS~ky^X-AC0Bw>(^}2+~WV|!4W_Jf|i6tXvOX3UA5cl*84qY$JH?2Sq$%a zZO;2`iR;_l^19&ro?yH2(17HOI|c+b4ZhbP*L!%v0++8x-M|0-Nc0bg1#eol%G$9+ z%JoKz`ObcPzkO1oI=wfrdb8ysO$NKw?+Or3;x~TF?k1~cO?CCJ)n8KD5bF7D$|KpgoLsYe%ca|`?bdT}^PsT)U{i(pf@L0ds zJqD?>=5<@w4Tq5@E_5{C_(JRyOOXGsI0sqFNcNIc#s9ptu=jTnz<+R<3q8?ofLX|S z#UPa19KyKt!9Hs+lNu0Vq~afq{OMTzT2U_U$5aZ71*;)$bS1U6Qo$-`?fC21S_x{9 z+p0dEO)sy9Asao1`%A`5!d`58Dqv0wl%9KkeR9>*EHi`l*$F>qFy9JSPce9?YgFzv8LBiFuT~{ zLG`g(M8jKis#m()iDKRwidj!_hrCL#5G&&$+Hk1{e%gcIy7xq$X2M zIxX9Jm2Sf~TCYkjuI<6s`J8usz2_FjmD^)wn=PM*yNl?HNIM#LVWi&yKj7y(pxelT-c@8@xxi6U$;WQF@2M=^H+`hFqDGt?70zb1ACZ`4p*BGgkkTqsls z($fjQ0p&AMJdB1sOQoh-zMg89El)lG2*wZci&(E4ies7S-K?3!FO*+?a9L#X>&q%6 z^Dm_Tnl_M^%KQk~!}6$MC?d>x_2lkFBtSj^{y0muGR%2~{NpZSU+%34ZY|DKad6ZYpqd2#LkFpTX`Ywo_3SXWvbZUHlHROO%gh`@{OMIazv#Vc z-H_$9$t@7xVrL$G_qBTQ#$00RK9K0O*Y&T9`p5lt3#eOEgzF)Z`5BGp4nRjU(MwCk zK%l)V@wxLPLxwsWz$6@GUdH2A7$GJBiGR-$?HcG-w1{Kc41y4a5r_h)Eo5=s7k;@4 z&?T{boue>N3qKV9YkC(Zs@GAL&e6z;A!e37w-%j>)rt;PLrc>(*{htk&rwM2gLZIu48$CpYbReeZ1y}5y`Qr4o;q)R zL&<2H^7fkgBe5!F$Ftg#EX`p;UqwKpKF&kow$|f&M%!%O$7;BiFzhtWsaz+e+`83x4*&aKyl`R5}>jY+A-kat{-MM z$PWscU0_G%U*J8(tJ}5bM}e}xr=2%5&3-QnFE)$U`Clc=%U9@KcY?9jKO>rQ$Nu$1WmlD!8?l4MYZ81m)U<;d&JX-w{;tD38Uh` z0a8Humli2nb{tvqQ&4F9qAhhCS$iv2E%EAlQ*Kp1_Ar^_+Hb#;Sh zJVw`YJy=7b=*vc7m08tq&lIzKjEN~zSpGP=|GvIu{Rc&7;5VFQkzkopWCs-b~G57p#vh@#`uU;VUz4HQ@ zOZ*PY9A$UaiN5gb!;&76;0S`vK(zaZ(nXsr=(`1ibfJ*++=dMAPEdpOpLt1j<*G_u zH(kg*o2)60+`3@XXhHYhH6%oV&3j#UF8+)6wwGcGie$tx0=A0zgcS84d97yaaJjUE z$J{Tn;G$ADO~z830#GvECAHM@h?l5OsPh3>unXyap@!}_wQewf%}z?oiP`Kgm$C^==iYU?^7KSJqUE!kFSgiLFG z+2{w~bI{vsWzPD`^dkUCwS5pvFYnh?us<{X9`FnXI$~*ciDofJ;st37cMA(^u9kvR z>)U;@TDh%VqLUYO^l)JW`H4<}R7$x!&A?+Y;xHap@b&Y=7ooQeews##ykOf4Omj=?3Y8`j z2-UJmK9eX>*d+(=P;aft9Nd#(v`Z;>h*aUU9a4xwU~derFflMXIvR%l6Gy?M`>Eg6 z)g|9LZ17z->+U^>5Re;`Q$?BMJ(tGp|MwrUoT#CwBPdE_p#S)Qb%u|11CS@jFDWc3 zL|3TCXfWbU)@6^bS&Pe+m}*N?quU_RXc^iaZo@7lu*%ckQg>J^YQ0J$(Bhi0uw?rN zIl=C~X}D_xFR92bW!!m2{Vq=4o5X$ul0=ZD)~U;liGco?Fsv6#Geu(klTx?|R$yap zC*32A-#VgJ#8TfW*9EHW{4QFIw1>)Y_AvR%rkadhRz7-SeVFydG$KJ?8doB|C`th4 zFKHllQ-p}etf||%v6WhNEn7mmey#3on&+bR;uCvWCyw`PIiOrvS zvke8qT&`_`_xpQ9ExyLl@u%|PdrfcdC+&&_f9E)W^ua`-9#2>$W3XRO-MI^pPzWc5 zUhD7`pexN^io^$%&1@M5%p8q_*)fXlnn2q?35V+mmBdg>NvP0`22V_b->As3lh6kxEnJWt`1q^pL&1m^u;NJam zYazVL4wdF#TKj8d)pn3~_zdkl@A=Sq^e!N-h8@K0y{Lkv#JG<} zrh}viPk$2?lgVGLT;HmJ#TD7yBvR=^+7YalP2WYGK}wAO>0JCvR&U9|FctMI@)kf1 zjU#{J40wGQqD%M;K|8GsK|N@SY90hr?j)fy8ZdU5EFr#j-iOn`=kEP=W0V;a%^iAM zI=2h7>7OK4YDMl#=qD}=f$f5T76t6Q?j_AspZ5z1G)LR}1vRI|5RrFx%bLe(2ZPTx zq3dAZvg>M(b~Nj-JQVT;+ya}K`Op!R7gN_rC*~nKDjgxct5+`lqNTLTlump}$n{+( zWvMO@twzyAT-WqfqDC^+qGhx_ebsl4;l4h zF$b5*+%T*`PA2=U9jGkCx2w)kD+Oo|Uq&(b4_SbZ31wJS{khQ@=f{`hVl@quCOlCX z0v);gki@A?H;St{Olocf2kj7Cr>&o>5JhW~pdTT#ceA1za)ubNy*|4A?M8ViSZ}M4 zlGY6T7Dx)*h8Cg4^`(yEro!TAaHSB|g%-uWxT~`Z6gud+bf*ow$vt(U7oFpttg)KQ zD$z1k^crR-+FI-B^;^}xy&*K$#?||g3mi-r_;j<9AjdO$*Jq21rW7K_T!w$BzrBI1 zLITv0gpv}e&_}-svMYD|nt3GUjE{kJF5i~@oVv5GWQn|+#@NUdq5e#T?DPwvZMVE(| z2G(=evaU?pzCez_;}PdhnzpfPW>|&@E~8|U;8a_!$nUV{q`rv{w13rdf5=Y~G(d_p z=;nOs9Xq+flJp~u&I_uZCy39=ZW{eoo5liZq;$4=4P|PUWrd{eTxOEOP~X*HVO8x4 zcWNv|x{MmhB6pwU9Jro!jqCMve8J573d~1Pg=8=xBxhH-Y|8h*{|=?{O`V_vttCS%t0p83*hWP)2HpXg^Q zo0iTps$Aw|X)R96I*PRVm8BX=pbK|S;_kA|S@9R-y^YaIK{=`6DUCHdWM93Jg)Ng9 z-Hh^GRUV=fT8-XaQ$r8`^fKw&PHz|P8KCP#aTkSW;zrarGf2sJOby3(aj)Sa?Q$gF z<_3LvGXLVxBoo~_v5h1JfqJl3GHNM@rWiOrUen1Y`0Ga(e-}cF?jJ%2>20z|M;H*@ z`+iHGX1u+9TAe~Q>AQR6-8;DDesYR)y;G$}mAxfAMm&$-N(QRGf?d($uuoz|SAxjM z^%akt`*@Q<17bY8>&FwP%y~9P72V0Sgmu;2Llt&CLA%==SMpkPw(`Ahq(6Qve!`&W z{2ZyJa6(x`T4M|Ha_bAMQuK?h(Nr@Jl)}v-6f?)A;x%iM0jW`@FcbI@eC0XB+sF%((Pu1($V z@BE3b`XM*`WuhdFOZfoCC0(aWDOTS@?5KqHG+NCDSz3r$&(;*OzL%K2iGh4>~5{#?LXE`mX=5=W~M1J zR@vv(`0F$;2b+AY?38B2hb;PW2AF&jrxhI2qZHU56_RLjf?$;bP#Nus3X423^x~@f z-=Rzi?C|!`{S8`8vJo?>Jt^@`VU$d|+@?|`82(ltU`VHTuA7LaPtwS#f1Oy9MWh5{ z$H(BYJn_N zId>&#NI47dEozdr4hnqgrE@688f71ELL*JF*?P7%J_)YXHO75)@mV;d{AHNK0q zQO;~zf^pe8N^%k;+m`6F`{8Ac^CZqMnXbPd=>WFLSeT6A!`p0a9U9~nqZ|~1uV#KM8)ybTwQYq4ZZCJ|x6ikRXcV@F5N=2vcpHvl_rVdcwRbuE!MB{zJMduyS z?CWdjuu+Z8q0|(zkZ>|!fgz*x>7VEjEj+iUJAU@quh_C63Q?CbCQ@q9VLUk=EXMAu zQon~)Xhy>TDQLf{gw%s_MZFG@9(VMF+TxUvR7m7`TeHYiEzFAqIDxT?lAjEk?A0Zo zsf`u1a6OB@M;1qk%&sE{vL=^HC^rY%E2?=QC5KoG?a97KIZAgO|7`GaV^1vqabj(r zV6O@0MGvnGEi_|-P;Wl>CnZQ+sEtWkBKDLlMD$Av<)SLic%2+#c8knY&RP-Nhelj2PvULdW3Qp4_!Z{KEL5C*EMBOa5 z&6XM45giu_8z2vD6uekL#{D);`1S+@3R{0m4DiziH66c_h3C@yK=O7OTg-=R2j z!HPtuOkE~)I}wf(MU+1wo9*0rPtRt|<7ho0>ItX~hLZGs*^Ax(_pD?<69o4pE65nK+=ULm1`+oKWES32mrgVEq1b^uf zh)Qf-@XUjjgnVy6s3%30g*NP6PNZ7JzRcB*SJhF1%<0ucP18L&t+fPq*VK5rU2Ww- z2fD`jBByuY~D7P9LnA*8{&XB(9U^E*92L+doPfG>{QEEaY}& zrd4L8-+H|MHoM*%XE<)4G|zm*7Yv<&Tct^4*JnL9$#x*%MQE;D**J`5W6zDZCzh-d z)(CE%%GojT?)rnycsi=q(vr>K!_AbKn_Xa2(c9OWV0`vX`iT&EP_T1Q-Nd>@G^NK= zlcLA^Y_2ue%b~DvFe452K8dX+hv&hH|8YQOC?+gBpbz?E)#x$p4XOzjrO;d_g$=hU ziNzVh;pR9+U!!sISBugy9a6*GTI04=yPEo!nI|GV_ift@dVBMD{p4X32O$3mWh9T?1+Ad@)Rac8$E;6a78bpcTPUDG&z*VFRC*fy%R&XiwPuVG-t#@TzE%w#vdOll@j3RsF#zVzW4Az($8E#nL zWOS(3;rw^#PXa^dX%`&n*2u3AzbC)W12s7hXKk3RQ-LUj9I_eiEX=TN-aL8NM|wTynHYsT z)%}~*kIe0>^VdvON@Ak%>v>1Xnctk$b)8$~>gjPpHc0xK_2(kl_V(?#gE_YMcVvu0 zniD~O7x&C?Hk<$@=n_wD`!U9}?8tUTe-z)R#|f*;%2MHmH9ot)$O9q&#T?P3@T>os za&D+}@;i{hkQfC3K55hv5p{z$ioYtph>#W;i08IxLpvidlKr%oQ|YoQoe^6CU1g?v z9?!aR&u}s}nre8CpYgG_oZ2O48`YdanVf@vpsFxKJ!*pT4V59X9!QXq37}3%Mb$UN zjz~am9b7?6iQn$*J@2NX=I$XZvmy^rQM3qcY~RiGhd(;$2IrTA3!W5w2AexY@?o+6 zh&VWSW3qUaqHh(WOeFOB{V+L<7kPCA4-SLZ-RU#?wtYk%3u8QpW%u_hV%ub6;c#+r*{~kBmE9yn%a(l!NYrwu zN~%V(;bZPVB?OZIiQcqx*UJ43QpnsVjGuJ%NlMhh!nu@3hf5WBFl3j^&O>kipP3db z2E6iv>>)|R*~(b~5s}vJacMLy&e zmSsu40gI=Bv9$#6S2_E`l2*KVONrc`imSQpkRz#fcTGH7j(;nxBMEIQ+VL}uqBy1j4p z$g?zXa$Ky1gKVUR$)i)uev{Il=b@zeu|CpfXp89g$0dao0hMGe%}-%95371rE#{H| zQzFl}vv&%QaIV)y)4z$R&iYEKU%sqRriOkGywSq}jWb)X`n)t9?h6AIKxuKMw+bPd zbLkIBs)f9hCcl0I%Oy?`ATQ(6a}{Yq$!A34>FFvIfWy326bZFUwu z(1^6&B`wgz3cj}vXGJ!jRClcP_WCz9S=4bJpKBmyvIRBHIS5cei>y^wGkntEghbU| z{cA`wnr)6ndN~3`VCB| z#?YH_kf`z20d|ncJZ96%Qju_&KU<_w`hLX(CD)r*vQZDBlEc))v#^!I=hhpmB#_-5 zX;~j@n62+-*fkx$@pGs~B63(5T8sJKc;3oqD|M7vw^W_EfyFoZGibRG4dIP3oXp=ax)zab1V4uEqF zK0CW}x-NrhwvRqZN<6Bg6!6D40X0&v2wmyMWj8h!1fl`NTKbkKjlPy=p) z17L_-57VtqCs}8lDG)}reV-0O3L&wbot?S3nC1Pyv$%j&%}Y8J z!A*B2D1#rl-kx`r|76%=ztM!${ig;pdidLEP9Eq{aL9X*q7@m_v2-2y&zV=(&=!rj{kmTAV%Q%pFy0^x97lN1fXcoadC!qhE21> z7_|Sl%0Hc${;)znn=F%k9y+#~G#X6XoZ0=}y46EL{mubOp$uOjtc^JUByE(VC+Yh}dM$Kj)R6a)*u zlSz)_7Se-xoDO9R@b@R0M3MLx5CvcD8F>K+1m)=siB%eYPr>2$kb4qZP%WW}y=TGv z;XIM%47F8|zbL}hs3%9SaPy4Z4A{ty)kYUKzirRRcn%l7UX@Jw`b5*WJvv%>J}OS! zS2bElXC8wa1&Roe$14}UZ(58HT6AQ4=SO`*#H1o&E<>*0NUWsrt2N;csH)={mwsww zt|S-4K$qz`L60Fs@f*hfOaRes>estiuyH`-82OfH(7nKE?UL^Q zLXJ-UnoPB$!dd*q$NR;{T1(HH4ew#9k@h#zzdSw)rBk5wf!knq2C=&3kf8pb7{ZPT zdK76H({KdX!89^r>c(^uB&thm@|tY5JgqzqF3Yr1k1fmV;D(u|&6o3LYh6OJ{h%71 z2@^KB!QV1(?d{xV>-Y{~XHUaulkx%@@3CYCq`qT@HNkDXlQVG-u4GM+>T( z0JfdCz3HIG<0D$KH6)LiBYrBnmN?p$iTbW{O*VglErq!kyw4;>YTx2jTni=Wx8 zn#Sd)OW!XjU%k>4p}`9*efNb=m8$Yzi;}GJo1>YUP4~TMLbrjhobJ5w`+G;gWH7DX z(>VB38e6We=hHMI1+0+%RW|p#aPx-iv&T<9tDc7V`UZx`yOw5AOVx60#~buZH6Uu9 zx3EBV-L}K#?o)|%<0a~)-RbV2x~-Coe(>O=qoj5>BKUq~zd@^e6;Uo!>ba<@?veg- z(bMC(9RQ>DDWhspXIDFLn5`bKsIg&#)buXXoaHDis2_f*SWKQMUp$YYEte}RDW|u2 zTeLp6TV(WWD!*5(!c3s?^xeB2IYA_7RMfNaun}#lD&@bjG%v426~+9se<--6cDB5u zpaG()EK`H|UdSFtOipoTg~v#x@NP?W`_hu zwLq?R&lYyO?Ubg6euzyrF%mf{yZ8mG4E2ooa``il)trp;OaYLmPO5^+=VdT0rwm z+0(Yf2mED`LSQ#2Fh(2JLz({Du-m04cn|JXIBUd!4t@odyzJ^v{EB6f@#K)IdO~rW z_NUy8DA^6>%=ff3zr* zG={4`n$foB%@wV|O%ybj>Ao9Ip?}NFM8=?EDJjAi=%0QLBls8~F?zpC;!k3Cj_%UnZsM&c2UUFA2Nm@&_lgc+f zAe_v!zDZrm#DnFff;fubb0R&@gNT;%z!XP?Lj4=>B`UMN6&rO+f^CZlWV=Ibm(7K; zeC;h>EbF>&Gjw8Z2gR`(D{5|BdIIH3p@wc)_Ify$Q##7b;_%BZIz#HH!&r^H5%vIj zYOCG~wl6iK>mL9P0-Qns1Wf|5w|0aBnTn4;(l8nTfbf=e;awv2tuEH*)W}x04|~B( z;0A|Nb<6ML?QfJeYC^?R>+lL%?!a5=$_flEzw9C$HaW;&vWHN3&{t;9L2?C_up#1@PF8R=wANJ7DRp_Wb8fi*&Cq_ zBb#0qX8;V@ztog25U>zJyFM(GNSU)F1}|d~QUQ0AUe0v>Nig#JPg;x<#TIS%znjv- z)|ZXGyzf-6YQn;T8c`YDHDlu4C z{81*E&dFFKL18KgIn{$Vjq-^zX8k^(n{<49oG&uH;SN;8Q{TRm*O*p*h{xfc*qPHN zBbiQ!1CX&NAkeWR*FLulr<8(E;gCq%paQc=7-Dx3BN0hvNerccHsb*2AQYg!L>#a` z0M>$(FK`b)QAnEYCDQ9?Z`}H@>E(lPY^C>-p5?k`F_41u0-u2Cbq!BE#65>PvyHti z`n;9qF5Sy#!~RoP6c4eN-tCd*bhW`UTq2@y$|#=pH2VF+1b`IXS+{_+<6 zQY}X^s^3bR&na{%*4_{cr2p@Rgo->6jx|^FW~zq(!4Fu0~LT>hz@*kjM9wh zZ2(|r%joW-TuTs0ZKwGosf9}VV`~vB#xv^6@B{GY;*cmF_%f{W=(J%LM0EX$X@r16 z;(Wq^`rxh~`|n(H;lJ@F_k$qEyuSaGhWwTZL{?fI+tdd0>&(8?;@LI!_`G;6HMs{j zF)nMDBvn2yKwnScXjNvsl(?+jzCAQpl)MM>iQ!jxCC9Ht(UGvBsLdr_lUa=robg}c z>jom1&|ny|bp4Ej1ER!YIRlA{zvdxcBBZ=&zTz~XOKL8q((BT#EENMLaaA!7_n9mz zxAiY2l>iV3_@;;oOv6r>eJ~x3cWP6IX0Pzdwy8Wq61e9OKNU_Oex546p-6W8bzSE+ zCuxIlFP;6oJ@rKju@<1cVzf5xj*)X zva!EjJvRiVbbn(YfBIL6enI|4kq|M5TXPMC6mN`iU^h9$O~Ad&bf|*iPkxcWKW)+V zL)fGLWp2~wntEI~$_K&d_a`Vdd|Kh*03VX%FO38lPAt_MXV4EoFRyvJIeNjy5c>6u zpzNUG$V0{K_s=*D!f2xx8|i_VKFeVk2T`8MyBB_&=k0d>D=BE#ox(`7wRH$QVCi> zQg%A;x?i)>@f^g@GP(xRNkK9h@*oL8=N$BV|1re*JYvR>v9 zAD59~$2$c5y*1Nvd15>Z*V8fjiyNB0UhY69)g)^5Y+ZG(L9!Qe4|(4r$}TY@@BVpP zOpR{8O8ws-7Ad1}K(sz3x#kJ^qGWWSdKrc=!5{}r(Y~e<;c%uut@mDC-Wul5y5Rk! zL>4p=@`Q#*-eIaRVJGeAVbBb<-c);Euwn4%%NEa>s zXi3OkY&*xC#`*K%3T3B|jogYewHOn=rM{C{(i~a;ecbcqq2X}J!t)UF zK9~_-Kj*d8yEd?vr%}un|QLSaCRNFaxYi- zvS2i)w0xXj`3y+u3{=bq-V(ZPn@WGzVDnN&2z6 z#p_Y=w!ACr`&{ccxFz145aonecDC>2!Fx2Fv-0tgIHPG%{7+SOH7S|(6zRjI%f`s5 zDp;x9hNG>UGws$=LU4@>8>&`rf7HZ*YdmnBlALEJ!l$RB*H?!Pr`;YnSk0 zj-c}O6LF@bN|f#ph{!P&gj{Od^ZXz{#LHEW1mDc3NXnQS&hs`Y)G;0=z&q@Shj=xLme97+f!o!2*1!)N zWBXfaq2X^fWDyv_vTo1!h|m@_wyg)d+BM;b&>GQVXmgleve;PwnkgL)s*PgtWIBcl9!7n_n1Pz;Cf@9Q$g{{`{GJ}2WcVUV{3isb-*68DGwrAE zH2HmIn6HBDYj$U|gSk)$YWDre&TQHL_L~ga&m#!{!whkKO^+!&kY{d+JBlJ6g8pFNTJ+W z99V9v`(3O($8sX;x%ewUDl-NL`f6~E7O+5_Hc;e|PDaK(nm81w;+;Tvn3Z zT)hanq%)=wy5PE~l5!XEf6;V~(Q!3g8=e@A)g+B=+iq;zc4JQ3xUtPiW7}+O+fHNq zo9A8Y`!&DMoU=Bs``(UnlEvc?zRP2q&zd^_2rENDic&?kO16xpI@yq3$rEFl0#e-q z>LwD;C$aNY6pzR2tsPc*W~aPt_9QXZmOtWQOd8Di?jy7uab2cm@~b9n zt7pHqnCGO8{?J1DR@hU|J3pv8XcrH8mfzEp#{ltn-9>7AA_4r)~ocjq2f*B*FO7JZx&R0i}H z-Br7!+cS37_RW5s!ehXPe)ZO4n?7;51wosJlB#k_A2@pWTN|2E)Ti_&sofvP)ztVOjss?E~9ac!0KS$){5 zqP%p$b@#RsL-xfj=ag*c%a@2$mt`{4uu!xHDfI&QMpqRoFi2t2s3)AvDwg{P&zeb9 zbQUYcRY!*JJi$?hnR$kZRIhq4+p9NL(BUuf+cHthjcqIEX(J(L0Ec9v_pMm*M~tQ( z-TryD(BB7vYp+bdG9+@3jM!lY=Y;tR)09ZnY}*c&g=)Vr=y2Rz3K$g|({L}aimkYAh3Ge^C0JI{o|8BDgOGSz;s8pe7ON|qJ zB%#M~YWyvMs>j#z&1K4vRxu8thjl0M314dW3@!uM93)_;{FN3jpGmPer^0xu!#Yo= zJv;48zY9=gBMN4K;aPYX07!3UK4OwBhU}vtR?tEbZg7$0dVAwxm?wcFCQLN@Awk}q z;i9!zRYzP!>oG?A(bv@Fq!4<0{*9%o2E0<^RAiir=EiQkSRRIQb1=XZUT6~TMjZdnl7vDwwF~@D?-0BXt|nKLLO+& zK&!W zOAeO?pwQdB&OWQ*)v)PuZPU?XV8?Qfn*M7={P~rj5sY#hrag~#R8cC1_~K(=XE|eY2Hp@TBQ>(8af0sszRS{ z2tjTG$(@`WBSt2A0xB_-%L zZth=Gs$0+Gz-xBMFl;lV&@A8^q?QmBdZ@z>la+HL;-axxWM$Sn2SpWQL@OO|gu;jn zqS4l{f;X0{xbC}+3aWWgb4svTSlb%2)+??y5lpT7qSxe=JL;2qdS4RVOy)9r=;gQ> zdS|&d1mV?H{WTTkh+!XHJ@|4Sl6X$&%o;8*BA3YHS!>P9OzJp|OoKVjh1O16$~zD&N=vgQ zA+&5i%toh)@fAM;%>S|Gb5zSf!Ja%5R4k%dpa}gM0g2#=_qRXoZDc^o@?ftm^gcB0 zJoXhA5-f*ABj-4v>Wyes}%(eQeaPT6}PYUh+-)K+#}7bNn!>fB>YsZ$pm@pqyw zh2Gk)hGg4jk}`)Q5+7|$x;T|u5lUSrNL=#I7VmlSm|MqtZ?|4I!1g-Vv3b0&h)Yg- zmTsHI*00RS)HwDJ7DL9mgTd>C)ve*ZE&X{@bES4?pXNrBSPlhfaqehBpLvtmEQSam z0^`ACfJj?@sA@Nh?tx}wgTVbCfSgKY60Hr0cL(r0K4FrxIPdG>j=!gEn08Wt6>6#fref4|d+tPmF~EBD$R3 zy(%`U$4Sz*lB)FHHGwaqz7xODXSpf6v)e{c*nR(@u8Kuw@zU#J>Q5Ij)$R?coCrq! z*r}d<%3p3+TKS37#3Ry)v4k>(U*ZXA`>AMC`*;wgRNGZuFS6~AZ}k-hOwpF9bzDW2 zYCXKhV*@;}ZT$8M-`mfhGPKt25;?YH47V3rig1#spB5E^bS8b~(IU)1okkhyp6G9E*- zn^Xwv%1*N`iRx_1ZObLgH)`JCv-JcXm6o#VVq}W}GTB5y>XEEsAyP23t4w(f_A6{d z=%@IrepmMFwsDApg8I7@$ZXe<$cQlqhG+xJvk*#1`I7{2vB-Kayk@lBwvtc`r{kpG zhp)aE(o)eZUqFa0U<1x4}mGskV7$^wU6?%4D*y>rN5NbcV8Eyc@5S{s7nCia=dQ{yDIyXBWg%MU7i5 zY@G%FG!RTBo$?Q&44_#ETiTPfF{I1*g9sXtOfncQS`FV(vlhbYRJQY%1dKJ5YCd;O zeiY<8mz%%+&~J3Pp*=PCJB{Rp8oCAu>&-pQ zA)+&(Ykp6>egOYz!F9u9B&?!F`FEZJ{Bb?Cms2y2qMOodtk$Bz*Fv`=UR4CWs1#~ z=9E(kb$FLwFYLaNlZIYu(Zli>3hy~bD{I2tz;XhpJVE|UE@Lez5gkTMDaTtwb!K-h zjx#&qvEpSE7~BN>g}h1bBpiY)uuWB^utr0tP(8Xv@AQokzlTAtO;-MK$>VDc+4nQ4Q!LrJi@Q*rqFTiVYnnX zC2oph62qe|1nLkg64{eOJQCBqCnGWOC$JN29w zcT$icQ@>(l;mQ6iVyC&oRgdjRN_q5Y?2j5u^2;)>jk|7Ib2wpUYI?P3<^FzY7g%}? z3zw>=Q`Sta^|+F!?z__X)Zg*h+RH&enH7@1gI1+mL z9;304Z%b|$>mh-Df_iw6W}%o#&4nR9dtWFjC3wuKbTWEEIzp_Ps?`~**X)cPD z+&H4T+$iOFL7~6fw|HKnbb@}bOI+mX;wMGx{!#Dd&LU{NZbWPpJx8swcl7ln*eBe^ zd?SlsNX$!JNuA%>_7LIzp}KKoZq@$8we_p)luKB<>*DrlPTs|`auN4pfZugDMh>Am zbZgFN_%ZmCQX(y2GP3Y^e7EFYxAh!RlNhRN^a$^vrG$#?%1O=`;5R=}tYqtW9kX8| zH?&c3@-JLry2ZS%BJ6kL{bI>TQ~C&xtGk`{HI7kE#AbSYXGU&Re{uuK2X-2qcyk0BRijg*+kz0*!E9i(nU!xQWd zDp=$=i5gf`8Rq!S&ZG_G>hC2b?#-9^j6I7AaYbiKpTme0aR?<0rW$FS&Ok8!CB1|a z3NPAn!8_(svw7Ca%c1E8y)9J#+^y$Lj|EA7VJ+_!iyb^Qm(IIRs{`_`i8u&2KdcW7;dfHaw(pa+EJ207$J;P=}vSo8L zIeci^o1xY|%M>Od_HCzy+K`(PwGg8m>hV_`nr-(IbWH%I{f41B0@v1D_V|0MsQr|% zl4UHjj=W)%WinCe5q*?FU-xJmAh*N2grCi?rOLNYMT(r}0f`|*&<{`a46?f+Mt$6V z%(}f55Y9^9b$uW)Nht1S*7;Y{#kMmEeo+p9-K2i*4{Z8Extr3X+TLwTsQ>x`fiGf> zzdeX6FD+d4=$v!Km=9MlN26*it-1J?he%xVag{K+TkcnmeX?rsN%;a=!!G3jxiU%p zoDL$Oel>?i0S%5psr}&86u)dejEy^xuB-c}*XWa+_6iL32C#Y;#LX>^WN)-llaee@ z=G5o$x#}K=*;UEz;TvzG6MRv_Ix>+ae%=k?ayxnUxTNf^`DYfp+%(IsI6Hgsw}S#N zy^lF*DXh12qiO2fmi0uncjKZ5pDqm_`E@9LPzA#uAT^Tw(F{|o5%noqcT|KubY5h> z<&UyL+0ktoRX^swE7Sw>8n`P^MuB5P2;yUnOZ0UMiRrOhN657P+1U9d*kwx@m5WSF z)R22#_=kd=j#-GP4>0gB{ZC%ECEgx|Gd7#ey5Jj|+YKI7kQJE;*w(5sradMB|Zx-%_ zLnIwrjQf|OR=`9kb&Nfs zPXP_-f`J!e-08m5(Q+AYjRE7Vlj9aT*rb^*Bg-XRbk>-kr`08fNIAYYyEm1i6#{7zOOP-=TeNE zIM+KgedQ@FBF%0IlW1~->OVv#!BB_(4t-s4X%3Wm9l-)MMJ%0DK*tCxSwJI$bOAu>{hM*p{O z7(BtOmO=zG@RM=2QTqDu>oXM4m$YG-+ERojm>%;37=P_+s9JCg4l`z?Gu#hDPOxXQ?Dp?Y04Q zkcCP#?mcWn6vD3Nw`a$pbag{YZa0u5$j6+dvhmvAV#p%BTf=%C)Mac1H3$9oUG(1e zK5M@-ca9!n!`g=qHd!iSu-HyJwSuZY4pHbQ)9u!Xd@C9|&m1iI9qD}o6+kI?7S*Hu zX_{f`z;TNkR??hD0deJi=Wl?J*XnH>7$9T`k=Hf^@!h6fbtgfd@)24I{Rsj!5*Bp=$+BnuU~rhJb^e0y%-+thY2e}1w*IsSKXW!aVXZWhYIn5#P4U)#mbAJ- z2i$~rzM&xl>U1$(v-IETz8#{mBT4bTG%Cq#C8-68J!5r_D5EM<7AXX7t6{Y#$0Lf6B4Z-F+G6SI!;FxLo>r-%>rBdeed4m zaw^c|x&5N?d6a=Z^Y-I_zQWzMUXYi-vO>N>5XyAV?Qx0JC&*SyHnP1t!@A|LDQMTvmkSCpZkIK=l+j+W1|{e&60-r!5jxD z(eyI-&+Ug<3j8-bYv+WQ9Q;p!3HO9X!&ItW#3$GgANe1yM*ib6khCFMxM%Le#-b1I zOe)vIU-1FoxaK_q!9?mg80qDh_qEFPZ#u6VW4e*?O#`Fj%V7GCFty5K%imM$(nSfk z-zR@6=23wFJRHj4x#U}PlT+&vl}zWuUpTj$a^b{>hYq)?JHNWQfuzQv-^bAObrmw< zl_0>`&}|(5_n_x6Liu%dZ|QB(Bc5{WL3g5Ar>rhP|BPW;V$}jNdLw2x#Ek3-bUr&|LLcb2us!n(YQVzBK(u--37Be!WZ`SA8aCi0Z#sB@pfZ6Mi|M zB%5YP2O+Bf@=P(wdOyCfjfmIvtdzfkel>pq>Z-A6#vK^QjHRn{Y?6w!JGgYk%~8`8SgJK~jBCer}PFrK%gpQp_tY_{&?5f;@6P;B~pT*mkCO&(N%C zW0LK2&I?VqUY9q+LU(ethk-PlL+!k_j2cZI}_TkjX19{cg%QS|O5`w^(U-&UYuxT?Q3M#%3TNCzRvK(Ia) z%>S@>dpN<3*8NS9kmo7sFRf-6z87I#->?<^q`jsFfXGzZ@M27e0lfO zJ$N;0eia6L4DE6+El=JM)gu4M(tKRQcPGEZ39ncWs5<|sja^0A`pjRkvC25_h^3&x zlaeMe*%9SZj|N!mW*{)5D!49!G=MP5f;VT>)d4qYOg6|Tr5XTA61pWSWk^}*imRUy zJhK=_@{5$~&i%+9{=B2vM|wxJ!n&8B5^d&@;>w1a*#A zK{fm1e+H%tzW`*gI5TxqBlv+_qCqq=4nIY%T|qseG(ZuyJXW%+5}mlXK{+6Rma#N~ zSDCW4XGC5R2j*X@R`@0c0qm0BKAClww+Ry8AVAYnhPn6g;SF&si9e^E{G z;%}Y?5e*D95*o&}9Jq!RhnCDTj!Y zcy>l9@I`G64E!0~fDMqg9Esny;RHXTUArZS7>?mYAarr%Dzt0+9@c*-F{5tdh=Y0t z|I=)l%||SOiJJ=Ze;<5YlUFgtlm^7?q2fAlqd7fJ68lnG<0NA%r z;9UOb6%FD<|0-44nbP7*1lA$m3xs1QN+Hgx0b`5Br`4IP$Km#%c@vTYNgQ;^X(+qB zRclyLl1fo@iYJ@D&9=;Gbk}|lbZffbxCSZ^9XHah6fmaVE#yj(DcUA~27SM_rK)nokhmPB97dbEZHrvOXzj02nV;#Vopo>G@o#Js7;?YYNUbC*+VSWX?Su9mIR z^R_q<{%m9$RV;dJo$FcDipLWkWbs}s`!lv`4!*N&v&~L#0dlAcYCe&uz2xf~$ z)CLS5S8l+bNkffv7YxrenZhz-wQ~1bNmv>Vc8!wR?DobN0&DldW)%ktu8{5L0ue%O zMq`{5i)ipFcoPm}KLDiFGOdstc-9m(1HlHc#fu?HBBOBS%R#dE-{i&$rp>zyXDN0z z8=*X|nVE5n)z4cBV{5BWqJvACC_+|QE}{9O;t73G1jz z0paE15A;V3i~riHoC`1~o!`YRM(9Ot-Lm=@x;*fe*h(KRFHTt-v#WIZQq!-?FEquT zxb4E{_4TjTvO1p3KVIbq4;d-J8x+Gu!M|1D;W}aqf7Tt2o&N@cu8D5ytf>h`84AhQFSYos=Qc0N_llQPIH!8KTfZ>hEiChCR9h5P!i%k)?&VgYnt5{4k#Ya&jd3Df2sj!G#0BSJdj?syHLGdfWw!nmquiPtGtPWHii~{3QpwVQ>T2Osialvx%(+X(@}(O-8-<~ zc2Ci#0F;xYc}|Fu-+_^&s@u8n3koYituBh6j@#-3A5(nPwN#SojFGoha@Kr<3RIUN zTUo+da=3^>{53$$zZ! zTKUZ6Z&>n{gk^sNHjxys>6cO~eoIWX`sPB^QxrtOC_CNpo@I~{y zyzSKhLMVE0JYw6+K~NN=qVSCa5>OJ>ea>~KDK7eRgPcuqMxhcK5iAv#`zQ~f&3?9+ z{Wob-mP4W)+TAw0^PnG7wY}|*&EAn)QmNO+%5k*MVvLq867qihGR(eRI*Bd^c1pX! z`U=CTgl4)^2%}jy3joUv3)^;-;fzKse{Tp=K)Z967B>aBX{ZZb|OST0T+4z2BJ%Xzej~)oP7A#3!^}yck z-Bt?&ty3>7+AA^InySX`X@XsUu>CDI41t0<6UovhE`0EyJO`Ikg~wTK9gtXPu}b@tbkYpLp#ek+|Mw>)Eq@ ziI9A*VPU2%vle`0v zwLEz#Js92V2gagHUg9?cTH3%%Hz=1i2j$YNNjdPT*M)s+#R4w8o+Mq3`(e2 z5XwNzF4AS{Xv5kap}(Y*5cc@IJL5xz0JimSJA5T%?mD^`G4{%J$%7f8OeyRO<$(?k zV>l~J>yRumZFG%Fb&_?aG?xvX=OllGWKc96ZmvIjj9Mn0bbkadCnSL-*tP zR@}6SZ{cV@z_svPZ+Sa4mnKejD( zPDYbz!`Xr}lUp*0nTqx{1fTW=IsmM6zr?%`MHLVRVxG{L#g)bFgBcoJWSA%U|9-o% zfW2^3Mo@(F?sbWnxEg-H6hbGCP;1>_PWic@8M&Y-jL?8w}$fKT5!LM$@VQnOrcAgw(yg&{0TV<|4-x9<$_WZa^ymgBl;2#2r-A@ zkp%waV~t9I^rHgzN>NIPJ=-?DB-1SYUw7dvSZt+%v@SYH32?3_s|^T*0BYFsBr1t^ ziGC3r5><-qo0*$`n$yeRhs$_BhI;_@oR~yU{A;i-C8Y5bPBj1Mw4^Z`vyct=Y$)Y& z9xu25Bo5fH^m`w(gzM}>CZ|gfNE+}l*Wp%A$%u1(VMzk0$Jpd5!IPjrtp65;&g|_p7+w8C#5A-U}a7H0kmBx({S`a>Dh1@BCMMH@(dLK5hXn( z1@HrK_k>5^Nh?b0LgbwWjX6bvCnC&okS~7!y~1295$C>JZU~VZaS$2gS5auP8Hl=V zk|F&^yH$GV#2*YDy{N>~bO`d1H?`a)=QqFYOXq#Y9NSj@jNJA=@J0_|$u@CqF$?<= zHc>3u&Y=np$9Rp=T7K_RZ92S>kKqO1#woLPM6)+&w(*WhYE(?&ZCZ3;rr+Y|ilq6i zZqZ0>_xT374XoEZ)7}IRsVQ$J^49Z|qb6NbYTL4jn%_cJ3?I*nU$-mX$ zZS)cO_ODYE0J2z<8%?9+sAEhJFt4hD^EkN{hwRiiAUa!GRSbEEx{2pT^!IO3MucD? zA*qE%fJF%D;f%tGQ;yS40lP42dY3#a&;2KKZR$tie|kc{&H8v*>VEN;Rs$cd;2JCf zAV*RBXCF@N+xD++>_72rZQx6BpsP5-e1O1Ln{c#;UJDGDk`J2x7m6c*6W`k+!^WC{539fC-zEf9=4#LB*`0Jhn|a2?JzvTEe~vrO)qt>1#c-XhNC;7*2PZ=(c1#Vlad45I0hlhV- zBB?9QD*+-Ry3K9eb!jf8=p-j^eR?|Poc`+sm6eZFgkMznI<2?3kWUJ_e7GoPh4b9X z^FLrzVz^wiT7E%UN}KCVIKKCM@uSNwOdmNjmB6(7p_%ZrC{?Cq?_;uUUlJ!8dmB5F z&`b{6L=1`o04Z?51Bdhkfh1^cg9scn`ZzFJ1xq&h%~r^oEQSz%_Q)P^Z+aEBN7t)D zU#A)cwEFL2=3Xs~F?T(g8f9lqz|HirJYO?4u!%&lNm$P(kuecs!u-jAL;(HC*gf%2 zl!sU!BNX7HCq{em5)UpA0YW-4e}?^Os_C%I}ap0QE|g-b_#k5+CPs z$3LGmf5IgPI4YC$K_G%i5yuA{P^!+sg#i>J{G)4oj<-Wxq*1ok}YQNScd(`v*c~vx8L9y23HuZNo3LEwR;9G5`y!-9O z6$J+y;B&{lIS3b7r^W=pK3?-2OJV@o4k(!53QN(E8~jXDGH1yHf=RoGE_w1k2A$~} zN!CaBTik6lASldW%p@zKDj4975K7*Ttl|r<9V8ZliJ|>`$(z(mB8@YUCQjrOfG*VO;YtJ4gw`G=IUGU) zu|*?e^QQ`Rjf4sk)tq>k@w2_ygS-0|NQz{W7q7%Ffiz zqXz~#jFoykNI%R>9EhEP{v+2puLWGQ`Gp`*U10Q8e9w?3zC^t21kj8J$|NG5g7$(6 zMBIc?=U?Z1F=298U?`4Yl+s=)o*-abPi6ktnP?2bOmSDU786lX+OSnRB=bM7a#}9= zFAzPIKduUXeh@F-u5BTJ7Drn}f$UZ(jM#v6W@Fkc-=^%lKw;zW{`L;>rUR?rnB4vIPc(BnaF$Oq&`r5gX3 zKXN_Wtu`y0H&1mjM0(#S0g!0BVXY_%a7S<)q#i?(-`aK~=>_B(XamN>`Zls$s0Pda z*MNgNw|&bFXLM>}ARGzwy0gg<;W}IvWL^u(`KVFC2L&Pl2_l8U|G(Sq#2+K0sZM|< z{SPSA)jGlk4DDfqbuL@H!uHErPO!(C-Nj4oy>*mz&|r+Eb;xZ-)3BGmp$9!QLAI6P zm}21lfu_JE3{^GMKL<#%11^w2l2Mh_A3KXQ^xw0y=AA#XT#1qL%`^^-;S6>X_!akM zeGLCDFK$sQ&o0L%frf|z=Y?*-V?*rR_pI!IQM|70Gv1>t@tKX=7;pKC?jZ~*0AdU( zH~~5|RfZCfr_Bw>qStm~UhIRXMwk9kk$3iuwH9MIOMe9BUWhLs0wAQVHKKLDjkJ8t z(19!k^Xa!xl0nd_m>>Z{FI0FLVJklg;fXO92>i#P%|gQ#xT0#f5thJ)l^Cq8Mg&5< zfNKNev!N0$_n%U=;2O7hvV$ggGw97ySVws0GDei!OXzm%mmyn`B%-62v*0K~xc9{E zti(D&)fzs1ooMu4XkI8(D{O>;hGoN0CU$a5N#h2&g9MuzgH=o%$p*SRYB3oYI#k4I4jeAE z-?fIX_h;vy{#&2aYI@E77q?o=l+2wHnP?BKi$A#<+00hTfAoiEe zUDuEQq}da|v^!^@2$U)IkBj{1orUiKUKf84f9XWjbz3X#fFQK5p^{iAZ-lBws5#09 zt>(O;V6t}fttQFNmXK-RunZ&g2+kruCr4)Q9l>G2mRkbP2`fsgO4HA{qok9RjjF=- zB6`us48f6tF8;*=GFnX{EtrtMnvSRbln=8Pubvz+a|a(-R$}?nta|%OWFpc4e@JEL zM}cT61;=7l*|<@{PS-Hhz3B>>LwHp&fCeEjygxS5i8^Vnn(Q_$N~dk6Di9clVa?`1 z^c{MlhY$zihwez0ayg1N1WA0TImX+=vFqByU*@$)y~`R8#@Bj@Ri)Tb)oJsMpwSxS z$+kmnr+!M|=WYw`!D;qakDXzK zjF_qfWNA$BVU*wR`wzrxgF{Pem5*Dy<$b861FcPH<$;heKFiJpVu9T~GJ0-PISZA&Zt+cu8);*SBgP*$ zV~|R)DV%Gp^uC&&wrc-oygMfyW)+)I?p`BNIP`xqYkDm0SSZK5H97Rjfz_~8=KZF#a1=DQI%Yfk4mBz z|61FIk*FT8PJ03l_SVIHa$mw%v~UUl~2a37C~6$h>JD4aF)Jdm?vRVhWM- zNo%jp!1q$zUZ1A|FX-c-MA9C$v2Z4i9GrYJGjFv?f4V;dXa#Nn6dcTuOV<2%ZO^PZ z&@9)Jn$xNeVvBBe_$nkp^Uy+Ehjh^5GIHBME*JzjkY*#;DPx205rH@T6VC=L+3v}3 zl$qX*$O=k|A-&2tkrv!PIh0#)kAS7)m)nOB_}1)*BdD$hki53$H?tHx9acKIvI7#` z(UHQ<=nTK)k>9(Z>&OTo$+6>M8#j)$PZ~VC8U0>L-v}rc^jlbHdVW~@hi0QjEv+|O zhoZ?CwaR8ty^fMM?DI_e9nTQa?oJ6;)=EjZdba zC$_i~TO1?4C-LzlZE6%rjI{VU9h_!8rt51hCO-)ig%4$>{J7znI~CIBpH;cOl5`dN z&CwZExCcIwPy|K~|`KfzGD`%lbxE(n+)s zCh&GdJ*+x9sTci^#H17l933(Kw*%$v@@t2|K9sL!we490U2-M?{hb7rClgKq^|93d z4f-GKMBsWN-ZPy2T5d+^H_T!yIj{D87s?F}EX8{w-rynpu6GArK>JT<>M}!?e|iGo zN(GTKxY%a~M`M&1Ld-LgDXsW9OEGIsPxv-EJ8sy>U8WkLDBWr?%6?Y)`+-vs$7JUI za*R6?B9OIAznb9W29l_@%)=uwGKxP z+_TqxJns5pigS^Z#<2?FftaAjEvNFX4^BPres?adi~S+Hv*-e#h1{j(x9Q-O^{8Kr z&CWsD&NI)J70O-a`5g43;KEx&-sfkwM{69aHCj7ekP&9zZqy8uRaBOG|9Gjm({HDI z>JEE?Z$G~_vMTfoD&ocV`Gp~%C{PbZr=|IVNB?cjL~Hd>0@ixghX2aGSVPDLem(|X zuJWy-|2l^T#`u{o8FW0hz=_Okk~!$Vio+dGbv3kErl77gheb% z6QBzg8UK%q0~s;1~x$qCwz9lC2M}&~`1DhR()5-ubT^gmr<305tr_I=1bSr3J2nV5C(#{LI9jx59CH_ZJ@J1SzI51^ zGph6L_Dt0DOXhDKe2Z+0hWac0NT9c-*5Cj~Z2LVYFo>~DnyNufaqmx##vkiWC<{hh z`qayn)jDByBotKdU2Db!L5>Y=|(T6$kh!z6OepE3!NV zR+-c--_2X1|NN5v{pSng557=#Ez@ex17EY_CAys@FrC#`F|>FY>;-pZtqoy0at9}= zYV84T&oZ{ZCaq7C6{RD+2@|)B23(yBO~|F&Wee&K{E!x0SCMgowIxVs(5iTdxFSkx z3U)3I%JL9!R}t#{NdO$00r5d3t=AyWzcgs&JrJDo^@aLyE~?EyR_Iy}i^>}Be_&4o z=5iZ;W)Qfrd-?$(yr%KJ-1Efj=#GmG821;cv|Rhj?zw#KQ_u6iIn~T0kDXu5m~`rm zHAJR0=KOR>VRZSe8Itzc_jUQC+PBiCsa#0Dd9XUE9XJ>Y#Ih)JDa(?t*%n{{b;lp2p5R4S6D<|jlRs9d)77}{vk zOmyKMF>{JJG;&`GWvpi0xLtLwY3A4>ngMOBM|@wSP3>Oyne^>K9<6uYi2QCZk*8yu zO9~wDE<_QhQ^Emt(w31i(m{Re1vOARLJ5q=*g}(Z;eDfkR0bVo#T{g4UtTbu|MC!7 z)2_FSwi9eXUWGa2SzH$P?Q?sq{(PzS_;^0@ef0)o6$D<8ejcfw>;o!FK&%>QHwbMi znrg#NpYfAIywlotwm!J7*$6Lk<{XNsk~svWhVm-6;n8upfuQ1okm!4ns)}z z-=(DQa2xYqqBZUH^OI#>9R4hhI$<)K8cqL<&2b87h#NsanD8z4j-EU1>|M$!LHHa1 zxvokDd|jjcls@N0)WDewKrY{<5%ttwndI z(@uct?4iBT-RA?R)YKG{zPQz{f~pDqiwZK;;ksi* zGZD@A=})G^m_oD8Q6as<;h_*eMl?izakImmEc!PP>4d-@-b2>c9IJp*$-)?v2BbLXF19Ibte_@zb&>X?)TXp@ovyQEmwX!V>h z8<2qrc?Bw_eT9|NQp>Wx&bgMcQ=96*suki`QyGf&MVX~>2%r>z{41^%xd_~GuRHsa z(<2F`&N4t%6Qj7~0{V0`_7aJhrAa$2U$KhMWgrBp(ptIsPmGp^Yhk636?rw+UiJnK z6zwX&56#lSi&8ADg&Ik<*9#3-iK-)kMdQj9x1ka05Aj4|nKVmznPB~)4EPqPCR@HX z5n%$aO**t)wQu8$+E+h0DU@>RB>$);+tlFoVBvlU(jN+fe^or{D6Kie^fmNi^->q; zj z3zhb?|9TAen765Uv(VGQ5wD8rlYPXKV)g$bS9%vffuu(DvRLemNR$6(zNfVp6tBGl zPUL^%ZRYXs*`tW!~-@#^l%PkuXL)#E^aqiimo$$GiJ;=mx zPr5=VU0fJO)jv+!(nKZVMWVT=Lv*CST7^;*dHM*1j-2_;Cn51hI4??u>SfT(>Jw;Q2Mpgy7k)9OtdfKUz$_@-D*0<% zUHqDendw+@C@bcRQ4Fi|i9cssT3Kup4`v8__zh>DtY5dZ`39vo-Tv3IKD-?E{f3=W zlfx;rGSyNuJupu2oHI4EQ;Nv<3+htdG64c)Q0zLroc!nlB1vJ!fZhz!uBmS`6vkiZ zq+<#Z+peIltW6`3Fqz$}M zXAmA zC+1U^Bgbb$OOt7*A{3R$4qH38N_464`f4({DocSbjFFrE`WFA^9WzZ-52ysmU4nKr z5a#CJmkd&gSLuM3+7DYD6jX0oL{?oWVnjl&wQp&o1~KQ7$->D3t|T0_`NvB$#I)rKLX z&du>KXN!~fw?5{ME6kqP+mgatg3Kn?%kcNyK zt!}FC%x}*nwV|lO;}<92<%&-4KLkv96gJ|(OAH%a1L?}*`Z3Da`sm|yzK((8ci@^* zEgD-9R|id?hqtYw!pa`SF210`jWsAu9J z5yue&pQhdsxh8_%{a9Jx+i)n+F#ozFq9`mYE7K;Zv~8{Ez9(k@$hN_R2T7%Ke4ixE z7fPN zfnQpV$&!3y|JHg%gi2>;C{J5*DOaGK;_xaeC4n2kGkt65OVtwXh5GUMmT7m6JDG!H z{@zI+o>ye;&$}NT>hvwkC3ioI@G>XpvCOOn4ty6>7VK8`+spf%^Q73&qF;k?mA{`( z>iDzXJI`n}7~KoCKomx?>dQFlfu)Gav8p}|xBmpF=CTCmIw z$9j5>NQ}q7eyB9J<6^Q*ggSdG6KGh-ImHi`Po!Ad&0DR$mTtnGt6>d$BTfG$n=L3& zRYE)y&pJ>=TIP%3&!GWW14K}V^kAU2%#^5lB3q#vZBnAb5ds-#BVaTS@=fL?5b=CG zRNU4Y5gzgnL9&TL&!Zwyn;o>Ei`-L5HDH$}DW!+QUh2&3A*3VB9oR-0NB7krFz$OO z2W=3HYNu%bpglGfdk;oH{hEoCG?V1o!q~eCd?k>srS|&7MIfr#fCcs{FU%lFDGm+a zp{9D(*`Wc;SGCOftj1x{TGTQtz&Tp@miJHMr1*B?#5aeepKDf&n?foAi^{%Zm1At} zbOsrQj=%Mhv7V;1$qFj#4+@-duRosoap@7BPX(HmYuKNBo1vsPl%|TC#Ok-C6?1O4 z`{Q_co#`Snjvg_uF4F->fu9SI|K7wVTNpA{#BAmyAy6l#_-r!9enF(T6t z({e^p!vpdR?@;}qKOZ0-jf7qY7w>ZZRgALZe;AD+y(eDT~F#Y;nh6Rxm5tjPJkaL~IBQ+so%UV#)UPrn9z zE9?;yh8RQX2<`y7+~_PJ6ItJ9IXJq|#-M@bU$@w%_<{4N8=4G3cJF2d4nysc1*Hs& z9Gc91UqI~UnY{Q@N=v*sj7|?xumvDyklI?^+!SP?HB+Vp230u-HKpJMn7jZ(}w}D5A zeVEKHu{wq&!WhCNFh3F*ltQ>vOuS1xt5&v^NG8FB9yN&Aq*YaPukJuDUkH}3TCpAk zJ$~<|T(eBcTel|d@TqkB*fD@XIj`@vG;TyLTY3BSj}wx`p)TZ)a@&S*gN~7f`(I}X zwjJu&0-8bOw)1RbZ+53U95Er9!Qeqbzrh0j#h?M9mBpRUdqu>M=oP{9Dz zTTKnF`n}Dlt=W7i|JdKSp!>KLzx)(y;E<~r{9dW+gDuVJ78h}KidECUHrHdXG&xuA zH9ad=U}PG4p~WlFG+km_HUo$MM*V3fq}6Ic+Vhae}sY&l5P-MRYD@JXKf&V{$|74B6G+E6Icv`qjB?IbzXgYVbucD|GHP( z?-<))n%d89h5)Qd=W47orG0z1h?s$9o#UD+vZY0s)&38muK8fB!I1rGnK@ryO4=t3GG;j`{OP8B5v@~ex~=!o%&JYjpE{l#Y+V;2TOSZz1aniC$>rO#1681wi`sxo2>?l`9s(i}tY%*cy+Y(j9n zG6rJ4>nT2Dy>&_dtxNjrSC|#()w}kB$T*uAQ(?4HRhVH>IJT%mfg13?KWT=FDNr`< zm~?hiw4IU&onBZrm8p~nf&q2X22C}jQ^Lgp;Y&lKExc`>i8(K89_b<{Ojbx^sq+KV z(k|l(C82P6?B~Y~nJJz1{f=q9gm7Q_AcJv@GyW1@rT!)InBi2TQp=x%-X2xV zb?htO*ktGRc#VFdB5sIZf;1v@)HZu#9%hz1q(4EjB3L+C-4CJ#LHI}p9ka(_q{q|_ ziya4b;r5X@?(aZ5;nloHM+^R+`-z)wYC}#o8uF~e+h5!|2G$ZrJ;+@#l5V92da}Iz ztMnM%n9%~UOc7vVluz+~gbG^yE3#gV&0{oVzDC>lOSC7bME%yS#e=Mh<@u z#5I?P|0_(fhji(VRjZLn6&ybqLRMOC`T>$+#I8G02<|Z<6q#trI%MmJl#YQ33UN^Yh9krlAC^ zl@j%RKgY7ji2(h$2fphFd6?kJbv96Kk^uOU&Y(IH`>eAXPCa0q1$B?inlz0bKJ_t zL?7*iE~TyYj71Ha#q_al69X$*KSV2ig=d7}`NnzjDe=#vp&hDNfyMHKhiXFe2H}R3 zdJjfqec3hN+W@(}4Uk*M4uu^~rUpoh?kjgbov)hzyADXc@Q{*$-p`a!n(5@@+q2~g zCp+vv?0%~R@!KbKFgOY(j!SL*X&Xm^$EtRs_H{Y#i+LF$S9(?|0_qAxHt3$$k$S3$Lm?XHz;CP`ph~BCOa>D z2o?O_lFNJ>rOb7=aDATYA43niPso|ykJ2}(zFl(DogBmoIh@-qELBkYtuJYdI;vnx z;N4W(kQW-a!hc3} z^EPHT9S4qf@ErFADXbRd(#j$Ly!PqLuF#uG`>_Nd79%t zu*|H6AcmZPA3N;7Mvo!52+B*|a7m_w_e{7j_x+?GX8SY}M1@+?!#$z$(G?pjj#q}lwwZOAQ*&4N_?I>Tqz z#EHW_^CrQQ5yLZBCLX5GxR4QHS0k1wB<%zqhcPl;)IRv@Q^eM#o;s*R;TCqX=qV zdtOuZR)5rS&U2_rhuBF@`|1W88Z<43+%2lCzb^24jFAX=J_ouDn*@`)yMAL^X7;#r zllPx&*;|fGvok48%VYoYJ^UN5N7X-OtoIX6nsEem(I9QIm4S<{Nz`*%!NJmY>OOoH z{f$H88i~|`I3Xb!D;rOY=z>B5S$`qns-yqKm^f&R!8m6aGeqJ7jby_mbjD8eEt?_E zq-M2BBrL@9$0x184CT?W6CUUrs~CGuo4OSCYWJ*Z6QF|Ei;SG10t%Cd`VXDZ@ii1w zEhUI>AgtbwT5Ym~n38VOaqbaMKuSBLx~{z)KK}XeWx7O-R=+ahhKW&Dh_N@9?u1)2 zxWa5!QlkJ;eS(}l^0!(zGcGDA2~Ai?txg0c5M^V3LzKL}Rs%i_#x~pMFF@q*EHN<{ zo{0{X4i!QIy_O>1V2`?BSvF-~&O%wt64x$a?Zd}_XG_B#HHX#j+6mTI;W$d2BL__NLL^m5PzkMJl*%6rugE580EywPo93y_TZe?A$rs7xAv;e%7nj_N=bn&mv>J5DsO9;O3f69395I%iM&U$#vg$ncI z1?kETM92ZWgDmtTS0Hh@4!U**HBU|kZdZ3zk98^T`@BJ^zzSdMsS0lS+x+>>O16*l z%5CScN50>*qq@vhN<>COLq_WJwML@#&r`~kj5`<$jk%-jS_)vs{!9Iht1@`TN1OD}qV9%jO{NfAbBpJ(jb90Fa=R$49&P{Kh z>&!~3B1q0jgZ^cmX*nS-!r0y^6Mm(ZbMfu1p+%vm8;>*(rJ#}fH_UA%sp z{Mv4yL*0pR;(w-2hkDHLwqsWbC0A)eevw~Qg8g8lc z3Nw5c$e8AG>(Tl+csrSr*ElEIDvj%mbaIYCjPMgwP<>fuh?26oK!;K|Mtkxslt7mDu$jo zc7vw}5-Px>5u)e|KYt($JA~eY+xorht^g;27J=AmKqnYlh30YWrxRomI{wmjd-xV5 zq*4ge@>rD>(v=_d?HNqy8`g?5&mb0RyLI6Ud8tW5P!XU+t3^7CSSkHRl1exbsC23Y z`IUOaU-IWb_}l*`=2;24gdVecMktYlc2Rf`*@*R+V^WJ7AuT1ruEqw+dGNK9$@C}K zWwFTI-z=q6=YlRA0DDO94NHcbb$r+Sy;il{6L{%_8-+m#Fe>r8{8~%Dsyt-E!C6qb zSMw7O@1yz$bI3Nojq(VVNJM*%E~mr~boe4rKBgMk4pSnoH;i(!{o@zB_v2B67?*Tz z;HaVhGh?{7u|X<`<5LLrRdd{k@B^kqwOs<})1Ih8VUz3?9MoMOcojl$n7n>)-utV)?5Y$!`!z+7{&?-pac8X(4%h7)!Y<+0Pb5Adq%|4P75eGK5% z&l%)ewhO|=QTW?z#+f36^O)KWu}zL9p4|xUhgs`oC*0$cTr&FNDr;Ub4(M{`q$i2) zbW%2i|L`F8*qKS};6WeFTX*slJM!$JaTxVYDmA2X03#j;ELF;XL9(Fhr9cd}{8@U@ zd7BgitD&{EXl@@SNaC&v=Xbrlkr}=VkWoPNX+o>%@Myh0(L~hZ3Ib94Dg5-n=KvXB zY&hYBF)RjykVN*%-A7n>cE7$vk7498;4A5UO|3Es)S2~VAY;^Bl_=EZrNZp|uIT^1 z7$zK&74~Z}8YWxzbK0TiVI|x*e{Qi9Sfhu`{*~g*PWG)SfP!F8F9$^~w7)aDl(>z~X zm>M-_gka+!0RFmZuIV;5Qh5Kr%|t<$fBxGziOo7&N2Uhr7Zk_1!09Uqo75>cIlK;w zG~Gro=$<)v7))H=@M>hLGz!{Vf>}9>+-_6lg$}(&nB4y!ndE`%JpOIEFfEv-?uu(Q zO7wxMsZ3#{ud{_=>|o!UD1Q?BrpNwEVis2QB zq*#qNeVC3)bx*M$mL^=@bn*^y?>AyZw{M!10jZ@@(0%#LE$-e$KF#C9hp#Z5tyxv) zzzIP<-j9F)PhfWJp7HnOBylE@6NSmO0KRzr7Us2#%jA4;rM&m$)>#1wwG?vc|IecS z8z~elo-jMZfkA0UsX(fkOd8ih2VPdQ_*9%sWGG3KT@{Nt=G_R7QDB{S4!Zx2_hY}- zg6`d=R+HR>hwV&vCZtqeWs28!CtMa0@woYl4brgw%+`RbGD*g0qg?zQeRtbCkwqU0 zB5n*w3Ri2X^cWgtINs3WNBCQ{XS`RXt2S#rF+9GnjO`s}9wxJR^`v!?ocx1zp{zc; zp)-Sbpv93HUdDObPl(KeaN+#j#FdNt9++p9ljV8s%+)A}f=0yP{pWjd#@BO8C~6pfatpBxorbM{Qeo5U z-DDM-kV3`eh)XH%%GSM-w&uiYA!X4 z9X{H|hwyDufi-m|lj-TOc}X>_>x5_mzFev-kEbwjneck}>6{GpyIXDE5-0!~VSDGY zkvg3$xsW^dAM}pYHQ5%(_cI6A`4(Ubr~+yqZEuIoIQ5ef)%!KF9v}mTBnKwQ;PyWA zLXeK3oml@@qk=KALx{eClP+K5eD3pk*6mTHsCDBn$GVW8F-yY;oVg_DO&u_#lPl$y zZQppkuhrbUeb&v~r*Rze>f|zU^Aut~OOff%2vo#nN(_|7P)~X8CrQ)xY@``o84^q+ zmV68GF@G3%{53EmMlxfqG}8X(#(1Fha!t%8n+L@_(>g1#JU?uxy^0Lja&iw6Kr^B) z$mFAvU<{b_+A9qCc`Au?yE0O_AbfbeY4BOj;p9QXNzmowfsbZ*gfN_)zXy5@2ypTU z{N{hWDieW0^kLLX@?@RRT3cEHHXWN(vu!Cny+p9+mWdc!2-Q>Oiv|Eow(%m<<^AN# z{@<2&=S0sJKHT z@`lvA7-3&Jq+GACjA>fzzGovXyI$=2(mj%s)AYlDB<|K-|MD_TcQe){CVjE^UO8%p zK1>URpA6?ZGmGojIlN;=<9#j=GV`lDKaXoa%FR}%M+@+Gyyw2*y^Gme%QpN9kWfGJv&y=aUo zWYkG<6sBR-F7DD?40@!5s(MSuO7157V9+LLF$|99anWmc?|! z(J2MjQ?p-TS=#i`x^ll%aS;+lid1GIeZ!CtWjvqI!PT#*v%2H_PjIH5)wyqRZ?8w@ze~G%>YO&B#E>&P~}wk=V`tyJ>+N$ zL#0eZI^TNx_}<^7Hda?Uth$6dRKE0CjSGxuI~m&zE&!cuu3W*dVVHJKr4t6HiR()e z#V3>8T>lnLb-H;MAT;a2y~kyQq7zoeRk27Dvunz%T68Bx6?N53>MHFXB}mtCrWzZS z8i4*oZ0umY(nC3~F4@6s??n*spVL2GO>uOstHwV-qCM=9aX%(}|2&sUkvN|Ea^^dW z_Q9i?R_XRzcUHTM$NC8MnegbZGsQ2o5BnF-JW`fg8G}gglZEitkJr~ZQNq3p>ZZ8S zj`}$~%#U(cS@k+)lI4wwC*MpU!Yj?Wd#w`unseWlqi%GW;L0goN84^R8Fl_EP%FQ+Uxfxm`Kse53HI`s~zs&f@-eL*2)= zaXoITpa!9FT4~2@nGSJnY0iNm&m)NZq33GU_-ksbzC)8k^8_=wQD#yR1iHvIs1Rwq zwK*)6=sU)T~3wHijv zUQ`)JaO(;JV&*2f?Mu5*HH=P&y-k{6ignE51M(G|I@TYqmkmiZUnQ5zJYgydWaqbX zt1cDYK5mtZ`Xsq+3(@)W=ovj`q2vE9kgRFWFD&ou4EW-PM{H0z@|)nxFj*>qAjAE} zK|%ip0fUftIW_TvZD$eona3Vj$AKIPH89IrKKGl>zdg(Y;z*R7!ed8I%;=ZiJxA=O$9$;n%(UM^km3&uJmmas+^4!kV&~@r`;?=2f_RiCq~#;b34>w{ri`SoiPR184=5@#fp+y z4D}$l#c@HrA{WuveR3&Y))JhL&B+6F_?2#S=+u~l$U;IkuA3bQEUn{5AfY_RL!YE| z^U5C)ReSKfCq2wDI{TL<2sikxC5E!%rR5_^BP)jBWIt%iG989wlzPU_k+lGlqTcD_8(vy4_PVuiS(Q~P}U z9Rdo&6E>|6G=`sg=STK4sVchH?I;ZDhVPp`r#dWUK0m+SnK;Zfxy`e^UTq#C4xa>T z6R+F8zp3Vp7Z?KChv$n@9#28Wu|mc+jZRHc zJc3_4ia*BQ$?o*)BqOS0Il0?RC)IYz@DMNfIC|@V8F9b2nrS(^?DFt7%F2Zop44ia zu@BdoOSOwsvrN-l-;CDs94!p@w~bj%WZ6mTn+3=--8cv~aZqaQ^MMM5%v3`tV-bb5Vm&0BW3vmDWG%78x?T0mi z2h?+im|T7Cb2R8v&6-_ZteregEKOz=WpK(uYFZp{>x5(&=!_iK&B+uw)IJI+%JmL} zi59eF#3-nsIyD!h%JewQiEQ=N6>aILtwECDsJEG%wMu@oeMk_hiY2(p{1BshxFB5D zBQqtt(4sb$rcW)-N^LZm%@)erQ&HOR`Yqo!`ucK7 zN9V^!A~^x>=J}+?VZ-OPd4?9ok}}(6^1D27gv0rU!*qg2H0|#PxQ}r99?}YhZnhH? zy70&}0RyLfW;baEiW9i+o0h5_5{rgmFYz<<=Mk&lVZ-Nsks62t>HUOiOo-v+LqbBD z{GYuIn(Q;6(5|w({56{3)D-+oet@bEER#0s>xIajRSa%&vJ7XN!z7`ZWo>$pdeXwW zZQ>8!8dJNlc>&k)kK(&sYsNPmJlzY6=rViZDBU}Rscjp|v5#`;3&c1B2(xk%E2Ju= z0}0_trRQ>UzHa)l!?CuxI>z|b73auXZHM?d1ol@XskLO>ImP6kH;LxTyubz4eYF31 zK+3s2T0x$pi5$?Hr)ib|2O8=tMb7|$uvhBue z|7zB*(=y7e`^Q4@a{r7-Tqzu**~ts z9`pE8H6^v0UFB^ATTG4KFSdbwq5}#aQSyU;+8e*4+q2lU6jh)*fd9NBsq>eVN8XB! z@HyCok;s@YvhSY+Y2&w^0AW*pbd^Va$JEg|b3ItYqhLzI6KoXeH4r-eSsv6l88U=T$Biq?tl-w#_=eVr?|~P5)Jbvrg+k&BtnK* zKN8BtZ3nrWt|XZm*(O|omP!>mNRGzKzCMQ+rOUx9X00qA@?aUcV%ZzqgIIUi-{Cp9PZ9Q8NeYo2l_LfPRz1ksaWa7kLZ<>S^SH<&J}CKeLqn)OoNO8dI^N$qF#qYgoWxI+7$kk5s4?c66 zgl>4xz)F_J;aWIUds*yLScyfCiRBu;kM`l?LbFz%Fh&zjvN84;v%VAPHN|i;kA`|1 zQ&Bpp)4k?I^T(}Ll!8hmhjaw%K~O|*aRGv-8+<-8E|j+~0KnJBZGS(052KF<=(2@c z9F;D7HA>%bA1mtaadw)>M0NY2!tJn%?~)^q?p#qdCc4f^SyDXYXWNd*?L1$*5*2Nu zEu(cvR?z^OGwNzVmSO;{zE-fkE-%+_fg>Vp!Wi9TN6OuHW$DiZvseiY9rF@|<_iBR zBc=+oDhiE1Bz^OD%RC30ad^vuuo|`WAG8~?e-Wf)8=oUQ^@!~ASGW{XTfzO&FnNib zj(qsZl=RQh@Qf@U;b!-jj?VCJux!KnAwgKwQp7x>?ZHDheI#pF%of*YOKEK+;zqh; z8M;8II}D-35LPek{Y2v#cu{!%U@jb)T;!aa$VX|ekdkuwH)Ysj{ut`9x21lKuT0Bs zoAzW*z0lDN^(VaS8#y23NW|aCW}RZr6F}?$OxX+@_>fa%T$zN)d{6|kqj6XK05PnH z@qE>w$Xwx4nAm*L${M=+n%zXh%!F6#>eH&p;-yxl!)>s;&-_)cjVz^GK@FPFq)wBg zLk)Ft&o8l2VA_7F@wW(b#If7jkjX7XjfDf>qil+-WNvUki74H5Sm5M@r)NdHG60`e z!bhAU#tqlFLy4Slac+4zNIj)p32q>!*TO+fYlK&0drdRVr+$r3o0z#SJi+?L=j5or z9kk?+5lrnw3)K{3k~ zL|BEeKcbg)yQnT-s9)(SE9>fRlRIajO3cKZgR{fKUa#~!|2__;0F!FFK1ilFa`{@i z=lH|8=xniG?`Q^h`#vDW!|;e)m*(IaR=b;4r8#cOF7BPg5dZ?Hw8Df@trj@xLkGcH zexP6GZdiW!?Z@lXO?`XF*EZ{a|J%}iX(sM_B{&QA#rqXK4{*DhwU%nY!~+Fm7s zc^9*yVnk}c?*`$3`x|Rt{=WX_0x=PHXRfrKM~TI)fp)9I*>s@Vj9s{PTlro;Ou@Zs6_22orUHuAE{zY8{7*GuY*FNgm`D<;B_H3)l>|Lxrp?%jP1xrGtzEi*LO^rFhy_x9LV z5%CX53wp(Es=wb4-{^W+k7*hcmzF15vCe&9fw&fXr&Kd_2Vw-@6ml1R2O?g6DVcSC z_v(f-aj#%%fphw5-Y&1lu5KAfC>WLFUv$;uFOUKL$nBM~{YfJ8I5&5AA zW zl|eX^MlTxlKsb|`YZi)=ZhrpdnrHIMUPe-qi^ujUdGEKS%>~H)0g!@P znQ3uPb3T-o(4KY6np=IPBfZ0BTSd3euYVSiI>$?-a&BDX-8sc+b6g}*=&p&dBm0X< z{@1t#lGE}_$H1kk(+I_xPd#}M=3&aSl!r31pp|p%TJ((p4i>;2L!U8u8zIXaP90XS z*0B`!dYD8ghacf3rqS#D9FN+IO+8+bZ7Y@w0T{&}P}}p1{Tx0ncb-}dKRDE5cq?Ai zvFwM6&(h3s5E8|g2Y_lIA@#sw8w^^}>~Wt;M*c8=_*Wcd7{+nSmonE>?zG(D;f>Oj zYc$083AN5~ocR{#$E!u5ujF@LYrUG3M>RR;X8#4kWsm+%mgAE@YX^S-66cuc*lUGq zjw{WJq#*H9YZSeNk)m1mb|%GGv3u(Ouue6L0N);Le;n0(;*dEbD0ZO= zx#6ELY@Rzbo0xUD$y_*+E!<4wIgBAQd0bjBEPKOg^2dvj!1jD|je}k-9>>B0aLBV; zT=%@WAYTN9V6>&iv-l=zuy+HAVFlqu?l!qkyIr@DYOx7(D)_&%-@E2Lp={$CtFNGS zZFqHkYu>QSMG|h_^;+4X#TqP9Si`s6(}R{)rfmPqh`FGaI+S)jNejqwu@DP$T%$T zQI8{G&(VbuRKLWDA(CPBwd*^>cPj$Yk(>X91=5R5bY}%B(RNPNp`z0R5`>5Id_utn zKOe5lO5$km079OkUBS&VswiDFxlaYzZAq6{a=0Y;7u%nx;soIoi>O5$!hIvGc}QRv z)U z=weZW5+w?{JQ)MOAR~QUV zAtoWO3*(WAdTU)mu5GxcB~QsYU5{JYK*Q1m zMvfD4X5By;$&iOXgCDI*^ZZ+LybsIW2Z*izpfF@_NPPPWWF}!h02!8JBiu|otKcDb zV+L?$+^xTytRJhqY^xlLzHEs;VZ(a>y5Bf)%kR?+Kc7L_%7a~iE6Pk+Xu%H{xJ&dO z6IawLFi5b@Fdx^U8SBSTMquR|G14EiYjbf34vXd-!{4y?hztfFRyY8fbL_JiS!!5R zZ2$T+@p^m`#+C4KWYkrS7br1(wFYnu%((JIx}gy{U?kmxny8{rLIL$86(|%rQRuX} z*BENYg4SQo=4$*KSH1RfZqIu$eKJimW@pgki?-Gn#Yz!t#cH_+dhde~(2jq0K7;MF z&dDW6URt62S>I=cz}+F*_-kv9R~Tkw*J&I89K1jf|ETviIsjfb(Hhw|m8yU&oXjO6 z1hY*cUlG15sLN-nZ72yyHWh>1hFJ-#PjT`d5LUlw7@8_gb{5HXCG4e6a9NCzTsJjJ zEJkE0hPEZJHpmU_xW#IbjUo~WR&dC`H{!4>zV-v9HPUQu>$vuPY4&vB3Ksn)$|zya6vyb(|c)>dKmqB^aax-vb2`TS61xy3Xi&XmB^V70f>|Z`@3bf+7VoZw=EWJ9QwlShz8RGEO=(lqx~y#QH|>8R+o3vU-Ga2KVpw z1*X5qVb)T^BRg+ruYG>=6o3xo*f3WyR2ctm1j1&3Y!W~UYnRaIYoeV}j5PLZ*80-` z6af7EJ*}-T=rI0eKmPd0lWo}vq#lAPM!h zEb#l280>B567fREwY_b*>;DG$41H*xp=de_C!gizKr2(^h*vcwD*t}Z54TWh>y#F;VG(*s9!e?bR_#j>elgcct1U zx<19CyLf`^6-foIssqHMo0Rtu&tdSO^f0S{r@??@u^En}{ zldJ_AQb?#DC}a1aWbhw*kxZC_2xPPog}n$=zWb7tE?C|p8o6A%vw#SIaQbYV&?$fC z>su1-_{T?k_Q!yiy?|D3q@8{`Bo05Q*+J6Uk1eace*A#aFV|gFVgASsJ`Nr27sfZe z(e?b%iLM^g0-$PR(k>ZtOwp$=HMCuXUBl56B6o(5T}U4{CSWxuyvALRh1T!m;wMg0 z%a~mB)u65)j~0*%b%o!*cigc#|y^Y63 zn4HBPBHP5jHH2#hP=yNn4&L;9LnBX^1Wx+Pp{be%znq7Q9+SLtPDk-d?Q#u7=%EB7 zex%R#{2S>XCxG5y90*?`1c(x|M%#d=W1`&nBpZ9T>T}kf%Xh{gLe2|$KW(ue*QAyOSsto%s$s zv$i5D*F!~@*GA6hn&$ZZj zX&jChy|ch^a1oQHQKdHena6Ljo2hc0wpl$E623-ES$6B20$yt6K||e22U3B zv=LhFQjp@gj6s8&>?M-eP@d~R>SqbgGwOhWkx7gUA{D)eUJ~)H80BZc5Pq|@he5pJo!AnFH_dM`bD=x;hfpd;fvkjOsYK|7 z4CklHkkhf8i>MNiYm5)#4rt*W1wU>BG^8~J zNg<<;+h_vw_b-%u*yQ2=&?0P5!|G!M6L!amAu;a5|7e#bIsMVXaT;>|`m6I%$m5f2 zR+=Mo{CbZ0y0I3)Z((|A-WJ2f z8)v6Rp%?(69jN+JiFp!vQ3mi}|MI{MW`gc4qSh#5Q)xw1NL&C|&Op!8nFO23>(AI6 zK*qmswnB=qfel0gptvrXJkOa3V8SQ@u*q{dda0W)I^c>Ttc~&LxyZ0Kgm!hX(}fj` z&l#*F+O{};bSI^61CpWWx&{o2b1tf&0Yu8)i6S`NFGffpWPUWx-G1K>iczDObSgk* zGEPR8EiI5sj1Ejma(tIC!+`xoFaik>>5d+ z;-R7l6N?*y=P<*!10@(lLwaDDL~rl`al?-^+4u-AA|N7(RVC6P209iId?>8MH$+9a zE#t~?84}e1P&c+KrYsQf4HSKQqWcb#f{(jtAvZL7#qfd4@g}we16}rgde< z@4to3d7c3XINKjSDZYK{Z+uczDkD~puS&9KJ=#5hE>7jP{xouc>7`5EbZH_{!py-=mHQQ!KZF8?1*ua*HOATTP)IE#^E2dbgs*`ZTmSHw#< z#*-5hYVWwSSm4cgq|Yi~>XZsXL=6qx{o}coV4tI7n7k7C`HC%;v{3cmh7{=CkRK{PY(<#1FPK0)>+4F&8=5JFK0tN!tCNKf&3rD2NuO)XmB`MN z{6%a|Y9BZF+LED$ZKho0iCS~{EK{=*w?l2C&x>sEiaImrHSdz`)O4GJh(mmE6nKg=Z0wJB!HDY9F*oYX4X_ zbZ?%4lMi{?9Y5E_EA+ulT+i-h1{@Y-EsLHVl7Anhf%B`YcB%Y_fBr=a@(egr#Bu^0 zT7g~paO+7|1Ll=KF*}^kLv#!Sf-wiKo<00|FF~nW9fw41*s{||WBao_KFi@y&a^&9!$+!)Ke7_Qiu7+qP}a5R$+ zAz=O~92^|92PaD!Erz$i*tV&)wNu~Tx)jNCq?6EA8qa(rS&6q!c-84%d-t%MBL?wG zf|BtmralM!eMHpl=d4&cO3irWMVN-J|tmq1X&6V{o8_1#YZ`zJ~`$2Oc~; zJ~~|OjzP=Z+zHf4n)WLgTp&A_3&|cvclkr1yKj|7obC~a?m{E6gl7XhJ^zprE^sV2 zIV`uqjHh$ zLsX$__v0Sk6X}?>P`IH};7TDDm}~&K+#45bfRxYijcM(*N=jU>bs$EL%{v)SO(wlIL_PCsWD*8bmsGX^<-IEHv$zI0LKHF%U%jc*AiHu1I&dPnIL2-28} zj`<4q4yDY8x%tMBfMW2|(QOShCwNwE41tj#69{LU4CnU@rFEWC=5KNdWB@RBJtRnC z5cfs{7!N<1DMjUC1HSD75{aYpt;GU~nWQR_h@xA=HpCl3hd2yqC7YPZ3(TZLLNGyw zlc7FS|C-}~)Xn-S^lI#fm>=?C+9d_qq1N+RaFXg@o)ac|8J~#tHhkOUA$1r1A@Owt zEfAS1n35l}Rjjg^aN_4u;)VP+y;~`qWKLoZ{1Tckmsx@>9SM716^vZJuGZGG_D9T3 z1@aKitv@3-j#bN^heEl)Sg)pOFJRAsTu7PFdP5x;GJTg*)v74_f}ESr4W}x-Hy%%O z96C_+k0YH^tXJl*3k7>_Rs5aBC=!jj3T%52GH_(zD8i}HoCEhJ%5}fnYgTk$k_={z zs;|)ma!z@O_Mt$e&sp#sd7ZyQ6o^dBNq*QNl!~zy?=vM60exc&)=q4~k5^kzwTBXJ z%VpR*KM`|C#4?=sK+p1qAgvi{_(HV>4!cPX>f5g<-S%XK-3Ue!atL{1NjanzmD5Zh z@#lQm!Y^Uk;y<;)0kcAfgeNp{*rG=9Q`i++G5`RK&?hNzO^mYF@h1*)Pg*=OR;MB5 zQ6Kmv6o|N$Vnxe@tIGeAzW%4Hj)Vn^x(IvZs!T%q_0c>CZ literal 0 HcmV?d00001 diff --git a/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-association-rules.md b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-association-rules.md new file mode 100644 index 000000000000..8daf1f1b80fa --- /dev/null +++ b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-association-rules.md @@ -0,0 +1,26 @@ +--- +title: "Unsupervised Learning - Association Rules" +tags: + - university +module: cs3002 +lecturer: Allan Tucker +created: 2024-04-04 +year: '3' +type: lecture +--- +# Overview + +- Works with **sets** of things. +- Learn structure, find different rules to know what group go well together. +- Another form of unsupervised learning. +- Works with "**basket data**". +- Recommender System. + +![[notes/images/Screenshot 2023-10-03 at 14.14.30.png|450]] + +# Support, Confidence and Lift + +![[notes/images/Screenshot 2023-10-03 at 14.14.52.png|400]] + +Given a large amount of basket data, generate rules. + diff --git a/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-clustering.md b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-clustering.md new file mode 100644 index 000000000000..ca822f84c999 --- /dev/null +++ b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-clustering.md @@ -0,0 +1,190 @@ +--- +title: "Unsupervised Learning - Clustering" +tags: + - university +module: cs3002 +lecturer: Allan Tucker +created: 2024-04-04 +year: '3' +type: lecture +--- +# Clustering + +Partition dataset into subsets (clusters), so that the data in each subset shares some common trait, often similarity or proximity. + +Clusters are collections of similar objects without the need for 'teacher' signals. + +A collection of objects which are "similar" between them and are "dissimilar" to the objects belonging to other clusters. + +![[notes/images/img1.png|300]] ![[notes/images/Screenshot 2023-10-03 at 12.24.33.png|400]] + +--- +# Applications of Clustering +## Social Networks + +For purposes like marketing, terror networks, resource allocation in companies/universities. + +--- +## Customer Segmentation + +![[notes/images/Screenshot 2023-10-03 at 12.25.55.png|500]] + +--- +## Gene Networks + +Helps understand gene interactions and identify genes linked to diseases. + +--- +# Methodologies for Clustering +## How to do Clustering? + +![[notes/images/Screenshot 2023-10-03 at 12.27.12.png|400]] + +--- +## Pattern Similarity and Distance Metrics + +- Clusters are formed by similar patterns. +- Commonly adopted similarity metric is **distance**. +- **Euclidean** and **Manhattan** distances are commonly used metrics. + +More distance metrics: +- Correlation. +- Minkowski. +- Mahalanobis. + +They are often application dependant. The important things are the **shape**, **distance** and **scale**. + +--- +### Euclidean + +The square root of the sum of the squared differences between coordinates. + +- Formula: $(d(x,y) = \sqrt{(x_1 - y_1)^2 + (x_2 - y_2)^2 + \ldots + (x_N - y_N)^2})$ + +![[notes/images/Pasted image 20240404112123.png]] + +--- +#### Example + +| x | 5.5 | 2.9 | 4.8 | 6.7 | 0.6 | | +| --- | --- | --- | --- | --- | --- | --- | +| y | 0.2 | 1.0 | 4.8 | 3.8 | 9.2 | | + +Therefore, $d(x,y)$: + +$= \sqrt{((5.5 - 0.2)^2 + (2.9 - 1.0)^2 + (4.8 - 4.8)^2 + (6.7 - 3.8)^2 (0.6 - 9.2)^2)}$ + +$= \sqrt{((5.3)^2 + (1.9)^2 + (0.0)^2 + (2.9)^2 + (-8.6)^2)}$ + +$= \sqrt{(28.09 + 3.61 + 0.0 + 8.41 + 73.96)}$ + +$= \sqrt(114.07) = 10.68$ + +--- +### Manhattan + +The sum of the absolute differences between the coordinates of two points. + +- Formula: $(d(x,y) = |x_1 - y_1| + |x_2 - y_2| + \ldots + |x_N - y_N|)$ + +Therefore, $d(x,y)$: + +$(|5.5 - 0.2| + |2.9 - 1.0| + |4.8 - 4.8| + |6.7 - 3.8| + |0.6 - 9.2|)$ + +$= 5.3 + 1.9 + 0.0 + 2.9 + 8.6$ + +$= 18.7$ + +![[notes/images/Pasted image 20240404112143.png]] + +--- +## Embeddings + +It means to map data onto a new space to capture different characteristics. + +The distance between points has some meaning. + +![[notes/images/Screenshot 2023-10-03 at 12.44.01.png|300]] ![[notes/images/Screenshot 2023-10-03 at 12.44.21.png|300]] + +--- +# [[notes/university/year3/cs3002/cs3002-unsupervised-learning-kmeans|K-Means Clustering]] + +--- +# [[notes/university/year3/cs3002/cs3002-unsupervised-learning-hierarchical-clust|Hierarchical Clustering]] + +--- +# Limitations of K-Means and Hierarchical Clustering +## Challenges with Hard Assignment in Clustering + +At each iteration, a pattern can be assigned to one cluster only (the assignment is **hard**). + +For example, **x** here in the middle of the two cluster centroids will either: +- drag m1 down, or +- drag m2 up. + +![[notes/images/Screenshot 2023-10-03 at 14.05.37.png|300]] + +--- +# Other Clustering Methods +## Fuzzy Clustering + +For example: Fuzzy c-Means. + +- No sharp boundary. +- Fuzzy clustering is often better suited. +- Fuzzy c-Means is a fuzzification of k-Means and the most well-known. + +The cluster membership is now a weight **between 0 or 1** and the distance to a centroid is multiplied by the membership weight. + +--- +## DBSCAN + +- Density based clustering algorithm, density being the number of points within a specified radius (Eps). +- A point is a core point if it has more than a specified number of points (MinPts) within Eps. +- Core point is in the interior of a cluster. + +![[notes/images/Screenshot 2023-10-03 at 14.09.24.png|300]] ![[notes/images/Screenshot 2023-10-03 at 14.09.38.png|300]] ![[notes/images/Screenshot 2023-10-03 at 14.09.49.png|300]] + +--- +# Evaluating Cluster Quality + +How do we know if the discovered clusters are any good? + +The choice of metric is vital. + +## Cohesion and Separation + +- Reduce separation and increase cohesion. + +![[notes/images/Screenshot 2023-10-03 at 14.11.13.png|300]] + + +--- +## Supervised + +We can use the "true clusters" to test the effectiveness of different clustering algorithms. + +### Comparing Clusters + +We can use metrics to measure how similar two arrangements are. + +--- +## Weighted-Kappa + +- 0 is random. +- -1 something weird is going on. +- Between 0.8 and 1 is good. + + +![[notes/images/Screenshot 2023-10-03 at 14.13.16.png|400]] + +--- +# [[notes/university/year3/cs3002/cs3002-unsupervised-learning-association-rules|Association Rules]] + +--- +# Reading + +- [[notes/university/year3/cs3002/cs3002-principles-of-data-mining-chap9-sec93|Principles of Data Mining Chap 9 Section 9.3]] +- Pang-Ning Tan “Introduction to Data Mining” (Chapter 8): http://www-users.cs.umn.edu/~kumar/dmbook/index.php +- Anil Jain: “Data Clustering: 50 Years Beyond K-Means”, Pattern Recognition Letters +- Tang et al., Kumar, Introduction to Data Mining (Chapter 6): https://www-users.cs.umn.edu/~kumar001/dmbook/index.php \ No newline at end of file diff --git a/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-hierarchical-clust.md b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-hierarchical-clust.md new file mode 100644 index 000000000000..55c94268ed00 --- /dev/null +++ b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-hierarchical-clust.md @@ -0,0 +1,62 @@ +--- +title: "Unsupervised Learning - Hierarchical Clustering" +tags: + - university +module: cs3002 +lecturer: Allan Tucker +created: 2024-04-04 +year: '3' +type: lecture +--- +# Overview + +It results in a series of clustering results. + +Starts with each object in its cluster and it ends with all objects in the same cluster, with the intermediate clusters being created by a series of merges. + +--- +# Dendrogram Explanation + +The resultant tree is called a **dendrogram**: + +![[notes/images/Screenshot 2023-10-03 at 13.03.03.png|300]] + +--- +# Algorithm Overview + +1. Each item is assigned to its own cluster (n clusters of size one). +2. Let the distances between the clusters equal the distances between the objects they contain. +3. Find the closest pair of clusters and merge them into a single cluster (one less cluster). +4. Re-compute the distances between the new cluster and each of the old clusters. +5. Repeat steps 3 and 4 until there is only one cluster left. + +--- +# Re-computing Distances (Single, Complete, Average Linkage) + + +![[notes/images/Screenshot 2023-10-03 at 13.04.18.png|400]] + +| Link | Description | +| -------- | ------------------------------------------------------------------------------------------------------ | +| Single | Smallest distance between any two pairs from the two clusters (one from each) being compared/measured. | +| Average | Average distance between pairs. | +| Complete | Largest distance between any two pairs from the two clusters (one from each) being compared/measured. | + +Other methods: +- Ward. +- McQuitty. +- Median. +- Centroid. + +Can find a demo [here](https://macdobry.shinyapps.io/free-clust/). + +--- +# Pros and Cons + +Pros: +- Can produce an ordering of the objects, might be informative for data display. +- Smaller clusters are generated, might be helpful for discovery. + +Cons: +- No provision for a relocation of objects that may have been 'incorrectly' grouped at an early stage. +- Different distance metrics for measuring distances between clusters may generate different results. \ No newline at end of file diff --git a/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-kmeans.md b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-kmeans.md new file mode 100644 index 000000000000..979270cb0a56 --- /dev/null +++ b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning-kmeans.md @@ -0,0 +1,68 @@ +--- +title: "Unsupervised Learning - K-Means Clustering" +tags: + - university +module: cs3002 +lecturer: Allan Tucker +created: 2024-04-04 +year: '3' +type: lecture +--- +# Overview + +1. Place K points into the feature space. These points represent initial cluster centroids. +2. Assign each pattern to the closest cluster centroid. +3. When all objects have been assigned, recalculate the positions of the K centroids. +4. Repeat Steps 2 and 3 until the assignments do not change. + +Interactive [demo](https://user.ceng.metu.edu.tr/~akifakkus/courses/ceng574/k-means/). + +--- +# Pros and Cons + +Pros: +- May be computationally faster than hierarchical clustering, if K is small. +- May produce tighter clusters than hierarchical clustering, specially if the clusters are globular. + +Cons: +- Fixed number of clusters can make it difficult to predict what K should be. +- Different initial partitions can result in different final clusters. +- Potentially empty clusters (not always bad). +- Does not work well with non-globular clusters. + +--- +# Determining the Number of Clusters (not mandatory) + +--- +1. How to determine k, the number of clusters? + +We can use the **elbow method**, **silhouette analysis**, or **domain knowledge** to determine the optimal number of clusters. + +- **Elbow Method:** + - The elbow method involves running K-Means clustering with a range of values for k (the number of clusters), typically from 1 to some maximum value. + - For each k, the sum of squared distances from each data point to its assigned cluster centroid (inertia) is calculated. + - The elbow method looks for an "elbow point" in the plot of inertia against k. This point is where the rate of decrease in inertia sharply changes, indicating an optimal number of clusters. + - The chosen k typically corresponds to the point where adding more clusters does not significantly reduce inertia. + +- **Silhouette Analysis:** + - Silhouette analysis assesses the quality of clusters by measuring how similar each data point is to its own cluster (cohesion) compared to other nearby clusters (separation). + - For each data point, a silhouette score is computed, which ranges from -1 to 1. A higher silhouette score suggests better-defined and well-separated clusters. + - Silhouette analysis is performed for different values of k, and the k with the highest average silhouette score is considered the optimal number of clusters. + +- **Domain Knowledge:** + - Domain knowledge involves leveraging your understanding of the specific problem or dataset to determine the number of clusters. + - Sometimes, the nature of the data or the goals of the analysis may provide clear guidance on the expected number of clusters. + - Domain experts or prior research can offer valuable insights into the appropriate clustering structure. + - While the elbow method and silhouette analysis are data-driven approaches, domain knowledge relies on human expertise and context. + +--- +2. Any alternative ways of choosing the initial cluster centroids? + +Alternative methods include: random initialisation, K-Means++ and custom initialisation based on domain knowledge. + +K-Means++ is an improved initialisation technique for K-Means, starting with better cluster centroids. + +--- +3. Does the algorithm converge to the same results with different selections of initial cluster centroids? If not, what should we do in practice? + +No, it can converge to different results with different initial centroids. To address this, it's normal to run the algorithm multiple times with different initialisations and choose the best result based on a suitable criterion, for example minimising the sum of squared distance within clusters. \ No newline at end of file diff --git a/content/notes/university/year3/cs3002/cs3002-unsupervised-learning.md b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning.md index 98a001056cbe..52ffb801df1e 100644 --- a/content/notes/university/year3/cs3002/cs3002-unsupervised-learning.md +++ b/content/notes/university/year3/cs3002/cs3002-unsupervised-learning.md @@ -8,38 +8,6 @@ created: 2023-10-03 year: '3' type: lecture --- ---- -# Table of Contents - -1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Introduction to Unsupervised Learning|Introduction]] -2. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Clustering: Definition|Clustering: Definition]] -3. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Applications of Clustering|Applications of Clustering]] - 1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Social Networks|Social Networks]] - 2. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Customer Segmentation|Customer Segmentation]] - 3. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Gene Networks|Gene Networks]] -4. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Methodologies for Clustering|Methodologies for Clustering]] - 1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#How to do Clustering|How to do Clustering]] - 2. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Pattern Similarity and Distance Metrics|Pattern Similarity and Distance Metrics]] - 1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Euclidean|Euclidean]] - 1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Example|Example]] - 2. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Manhattan|Manhattan]] - 3. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Embeddings|Embeddings]] -5. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#K-Means Clustering|K-Means Clustering]] - 1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Overview|Overview]] - 2. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Pros and Cons|Pros and Cons]] - 3. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Determining the Number of Clusters|Determining the Number of Clusters]] -6. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Hierarchical (Agglomerative) Clustering|Hierarchical Clustering]] - 1. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Overview|Overview]] - 2. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Dendrogram Explanation|Dendrogram Explanation]] - 3. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Algorithm Overview|Algorithm Overview]] - 4. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Re-computing Distances|Re-computing Distances]] - 5. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Pros and Cons|Pros and Cons]] -7. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Limitations of K-Means and Hierarchical Clustering|Limitations]] -8. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Other Clustering Methods|Other Methods]] -9. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Association Rules|Association Rules]] -10. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Glossary|Glossary]] -11. [[notes/university/year3/cs3002/cs3002-unsupervised-learning#Reading|Reading]] - --- # Introduction to Unsupervised Learning @@ -48,311 +16,12 @@ type: lecture Some methods are: - Dimensionality reduction (e.g. PCA). - Association Rules/Recommender Systems. -- **Clustering**: one of the widely-used unsupervised learning methods. - ---- -# Clustering: Definition - -Partition dataset into subsets (clusters), so that the data in each subset shares some common trait, often similarity or proximity. - -Clusters are collections of similar objects without the need for 'teacher' signals. - -A collection of objects which are "similar" between them and are "dissimilar" to the objects belonging to other clusters. - -![[notes/images/img1.png|300]] ![[notes/images/Screenshot 2023-10-03 at 12.24.33.png|400]] - ---- -# Applications of Clustering -## Social Networks - -For purposes like marketing, terror networks, resource allocation in companies/universities. - ---- -## Customer Segmentation - -![[notes/images/Screenshot 2023-10-03 at 12.25.55.png|500]] - ---- -## Gene Networks - -Helps understand gene interactions and identify genes linked to diseases. +- **[[notes/university/year3/cs3002/cs3002-unsupervised-learning-clustering|Clustering]]**: one of the widely-used unsupervised learning methods. --- -# Methodologies for Clustering -## How to do Clustering? - -![[notes/images/Screenshot 2023-10-03 at 12.27.12.png|400]] - ---- -## Pattern Similarity and Distance Metrics - -- Clusters are formed by similar patterns. -- Commonly adopted similarity metric is **distance**. -- **Euclidean** and **Manhattan** distances are commonly used metrics. - More distance metrics: - Correlation. - Minkowski. - Mahalanobis. -They are often application dependant. The important things are the **shape**, **distance** and **scale**. - ---- -### Euclidean - -The square root of the sum of the squared differences between coordinates. - -- Formula: $(d(x,y) = \sqrt{(x_1 - y_1)^2 + (x_2 - y_2)^2 + \ldots + (x_N - y_N)^2})$ - ---- -#### Example - -| x | 5.5 | 2.9 | 4.8 | 6.7 | 0.6 | | -| --- | --- | --- | --- | --- | --- | --- | -| y | 0.2 | 1.0 | 4.8 | 3.8 | 9.2 | | - -Therefore, $d(x,y)$: - -$= \sqrt{((5.5 - 0.2)^2 + (2.9 - 1.0)^2 + (4.8 - 4.8)^2 + (6.7 - 3.8)^2 (0.6 - 9.2)^2)}$ - -$= \sqrt{((5.3)^2 + (1.9)^2 + (0.0)^2 + (2.9)^2 + (-8.6)^2)}$ - -$= \sqrt{(28.09 + 3.61 + 0.0 + 8.41 + 73.96)}$ - -$= \sqrt(114.07) = 10.68$ - ---- -### Manhattan - -The sum of the absolute differences between the coordinates of two points. - -- Formula: $(d(x,y) = |x_1 - y_1| + |x_2 - y_2| + \ldots + |x_N - y_N|)$ - -Therefore, $d(x,y)$: - -$(|5.5 - 0.2| + |2.9 - 1.0| + |4.8 - 4.8| + |6.7 - 3.8| + |0.6 - 9.2|)$ - -$= 5.3 + 1.9 + 0.0 + 2.9 + 8.6$ - -$= 18.7$ - ---- -## Embeddings - -It means to map data onto a new space to capture different characteristics. - -![[notes/images/Screenshot 2023-10-03 at 12.44.01.png|300]] ![[notes/images/Screenshot 2023-10-03 at 12.44.21.png|300]] - ---- -# K-Means Clustering -## Overview - -1. Place K points into the feature space. These points represent initial cluster centroids. -2. Assign each pattern to the closest cluster centroid. -3. When all objects have been assigned, recalculate the positions of the K centroids. -4. Repeat Steps 2 and 3 until the assignments do not change. - -Interactive [demo](https://user.ceng.metu.edu.tr/~akifakkus/courses/ceng574/k-means/). - ---- -## Pros and Cons - -Pros: -- May be computationally faster than hierarchical clustering, if K is small. -- May produce tighter clusters than hierarchical clustering, specially if the clusters are globular. - -Cons: -- Fixed number of clusters can make it difficult to predict what K should be. -- Different initial partitions can result in different final clusters. -- Potentially empty clusters (not always bad). -- Does not work well with non-globular clusters. - ---- -## Determining the Number of Clusters (not mandatory) - ---- -1. How to determine k, the number of clusters? - -We can use the **elbow method**, **silhouette analysis**, or **domain knowledge** to determine the optimal number of clusters. - -- **Elbow Method:** - - The elbow method involves running K-Means clustering with a range of values for k (the number of clusters), typically from 1 to some maximum value. - - For each k, the sum of squared distances from each data point to its assigned cluster centroid (inertia) is calculated. - - The elbow method looks for an "elbow point" in the plot of inertia against k. This point is where the rate of decrease in inertia sharply changes, indicating an optimal number of clusters. - - The chosen k typically corresponds to the point where adding more clusters does not significantly reduce inertia. - -- **Silhouette Analysis:** - - Silhouette analysis assesses the quality of clusters by measuring how similar each data point is to its own cluster (cohesion) compared to other nearby clusters (separation). - - For each data point, a silhouette score is computed, which ranges from -1 to 1. A higher silhouette score suggests better-defined and well-separated clusters. - - Silhouette analysis is performed for different values of k, and the k with the highest average silhouette score is considered the optimal number of clusters. - -- **Domain Knowledge:** - - Domain knowledge involves leveraging your understanding of the specific problem or dataset to determine the number of clusters. - - Sometimes, the nature of the data or the goals of the analysis may provide clear guidance on the expected number of clusters. - - Domain experts or prior research can offer valuable insights into the appropriate clustering structure. - - While the elbow method and silhouette analysis are data-driven approaches, domain knowledge relies on human expertise and context. - ---- -2. Any alternative ways of choosing the initial cluster centroids? - -Alternative methods include: random initialisation, K-Means++ and custom initialisation based on domain knowledge. - -K-Means++ is an improved initialisation technique for K-Means, starting with better cluster centroids. - ---- -3. Does the algorithm converge to the same results with different selections of initial cluster centroids? If not, what should we do in practice? - -No, it can converge to different results with different initial centroids. To address this, it's normal to run the algorithm multiple times with different initialisations and choose the best result based on a suitable criterion, for example minimising the sum of squared distance within clusters. - ---- -# Hierarchical (Agglomerative) Clustering -## Overview - -It results in a series of clustering results. - -Starts with each object in its cluster and it ends with all objects in the same cluster, with the intermediate clusters being created by a series of merges. - ---- -## Dendrogram Explanation - -The resultant tree is called a **dendrogram**: - -![[notes/images/Screenshot 2023-10-03 at 13.03.03.png|300]] - ---- -## Algorithm Overview - -1. Each item is assigned to its own cluster (n clusters of size one). -2. Let the distances between the clusters equal the distances between the objects they contain. -3. Find the closest pair of clusters and merge them into a single cluster (one less cluster). -4. Re-compute the distances between the new cluster and each of the old clusters. -5. Repeat steps 3 and 4 until there is only one cluster left. - ---- -## Re-computing Distances (Single, Complete, Average Linkage) - - -![[notes/images/Screenshot 2023-10-03 at 13.04.18.png|400]] - -| Link | Description | -| -------- | ------------------------------------------------------------------------------------------------------ | -| Single | Smallest distance between any two pairs from the two clusters (one from each) being compared/measured. | -| Average | Average distance between pairs. | -| Complete | Largest distance between any two pairs from the two clusters (one from each) being compared/measured. | - -Other methods: -- Ward. -- McQuitty. -- Median. -- Centroid. - -Can find a demo [here](https://macdobry.shinyapps.io/free-clust/). - ---- -## Pros and Cons - -Pros: -- Can produce an ordering of the objects, might be informative for data display. -- Smaller clusters are generated, might be helpful for discovery. - -Cons: -- No provision for a relocation of objects that may have been 'incorrectly' grouped at an early stage. -- Different distance metrics for measuring distances between clusters may generate different results. - ---- -# Limitations of K-Means and Hierarchical Clustering -## Challenges with Hard Assignment in Clustering - -At each iteration, a pattern can be assigned to one cluster only (the assignment is **hard**). - -For example, **x** here in the middle of the two cluster centroids will either: -- drag m1 down, or -- drag m2 up. - -![[notes/images/Screenshot 2023-10-03 at 14.05.37.png|300]] - ---- -# Other Clustering Methods -## Fuzzy Clustering - -For example: Fuzzy c-Means. - -- No sharp boundary. -- Fuzzy clustering is often better suited. -- Fuzzy c-Means is a fuzzification of k-Means and the most well-known. - -The cluster membership is now a weight **between 0 or 1** and the distance to a centroid is multiplied by the membership weight. - ---- -## DBSCAN - -- Density based clustering algorithm, density being the number of points within a specified radius (Eps). -- A point is a core point if it has more than a specified number of points (MinPts) within Eps. -- Core point is in the interior of a cluster. - -![[notes/images/Screenshot 2023-10-03 at 14.09.24.png|300]] ![[notes/images/Screenshot 2023-10-03 at 14.09.38.png|300]] ![[notes/images/Screenshot 2023-10-03 at 14.09.49.png|300]] - ---- -# Evaluating Cluster Quality - -How do we know if the discovered clusters are any good? - -The choice of metric is vital. - -## Cohesion and Separation - -- Reduce separation and increase cohesion. - -![[notes/images/Screenshot 2023-10-03 at 14.11.13.png|300]] - - ---- -## Supervised - -We can use the "true clusters" to test the effectiveness of different clustering algorithms. - -### Comparing Clusters - -We can use metrics to measure how similar two arrangements are. - ---- -## Weighted-Kappa - -- 0 is random. -- -1 something weird is going on. -- Between 0.8 and 1 is good. - - -![[notes/images/Screenshot 2023-10-03 at 14.13.16.png|400]] - ---- -# Association Rules - -- Works with **sets** of things. -- Learn structure, find different rules to know what group go well together. -## Overview - -- Another form of unsupervised learning. -- Works with "**basket data**". -- Recommender System. - -![[notes/images/Screenshot 2023-10-03 at 14.14.30.png|450]] - -## Support, Confidence and Lift - -![[notes/images/Screenshot 2023-10-03 at 14.14.52.png|400]] - -Given a large amount of basket data, generate rules. - ---- -# Glossary - -- **Cluster centroid**: central point within clusters, the average position of all data points assigned to a particular cluster. The heart or core of each cluster. - ---- -# Reading - -- [[notes/university/year3/cs3002/cs3002-principles-of-data-mining-chap9-sec93|Principles of Data Mining Chap 9 Section 9.3]] -- Pang-Ning Tan “Introduction to Data Mining” (Chapter 8): http://www-users.cs.umn.edu/~kumar/dmbook/index.php -- Anil Jain: “Data Clustering: 50 Years Beyond K-Means”, Pattern Recognition Letters -- Tang et al., Kumar, Introduction to Data Mining (Chapter 6): https://www-users.cs.umn.edu/~kumar001/dmbook/index.php \ No newline at end of file +They are often application dependant. The important things are the **shape**, **distance** and **scale**. \ No newline at end of file