From 6a7f015f11f1933798820396f60fae7254078152 Mon Sep 17 00:00:00 2001 From: mhjensen Date: Fri, 6 Dec 2019 09:11:10 +0100 Subject: [PATCH] typo in decision tree --- .../html/._DecisionTrees-bs000.html | 2 +- .../html/._DecisionTrees-bs040.html | 2 +- .../html/._DecisionTrees-bs041.html | 2 +- .../html/._DecisionTrees-bs042.html | 8 ++++---- .../DecisionTrees/html/DecisionTrees-bs.html | 2 +- .../html/DecisionTrees-reveal.html | 14 +++++++------- .../html/DecisionTrees-solarized.html | 14 +++++++------- doc/pub/DecisionTrees/html/DecisionTrees.html | 14 +++++++------- .../DecisionTrees/ipynb/DecisionTrees.ipynb | 14 +++++++------- .../ipynb/ipynb-DecisionTrees-src.tar.gz | Bin 294061 -> 294061 bytes .../pdf/DecisionTrees-minted.pdf | Bin 563497 -> 563936 bytes doc/src/DecisionTrees/DecisionTrees.do.txt | 12 ++++++------ 12 files changed, 42 insertions(+), 42 deletions(-) diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html index 270401bc0..6180ddc0b 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html @@ -297,7 +297,7 @@ MathJax.Hub.Config({
[2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

-

Dec 5, 2019

+

Dec 6, 2019


diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html index f7d02c5d9..5b8c37cd4 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html @@ -282,7 +282,7 @@ MathJax.Hub.Config({

Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with -a decision tree wth different depths and perform a bootstrap aggregate (in this case we as many bootstraps as data points \( n \)). +a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).

diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html index 0d8d8cac3..0cdba8e3c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html @@ -315,7 +315,7 @@ bagged trees will look quite similar to each other. Hence the predictions from the bagged trees will be highly correlated. Unfortunately, averaging many highly correlated quantities does not lead to as large of a reduction in variance as averaging many -uncorrelated quanti- ties. In particular, this means that bagging will +uncorrelated quantities. In particular, this means that bagging will not lead to a substantial reduction in variance over a single tree in this setting. diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html index 33bb1e112..d4c0e951a 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html @@ -285,21 +285,21 @@ The algorithm described here can be applied to both classification and regressio We will grow of forest of say \( M \) trees.

    -
  1. For \( m=1:M \) we
    2. +
  2. For \( b=1:B \)
    3. -
  3. Output then the ensemble of trees \( \{T_m\}_1^{M} \) and make predictions for either a regression type of problem or a classification type of problem.
    4. +
  4. Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.

diff --git a/doc/pub/DecisionTrees/html/DecisionTrees-bs.html b/doc/pub/DecisionTrees/html/DecisionTrees-bs.html index 270401bc0..6180ddc0b 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees-bs.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees-bs.html @@ -297,7 +297,7 @@ MathJax.Hub.Config({

[2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

-

Dec 5, 2019

+

Dec 6, 2019


diff --git a/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html b/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html index 1169dea27..64b17ab31 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html @@ -148,7 +148,7 @@ MathJax.Hub.Config({

[2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

 
-

Dec 5, 2019

+

Dec 6, 2019


@@ -1784,7 +1784,7 @@ plt.show()

Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with -a decision tree wth different depths and perform a bootstrap aggregate (in this case we as many bootstraps as data points \( n \)). +a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).

@@ -1889,7 +1889,7 @@ bagged trees will look quite similar to each other. Hence the predictions from the bagged trees will be highly correlated. Unfortunately, averaging many highly correlated quantities does not lead to as large of a reduction in variance as averaging many -uncorrelated quanti- ties. In particular, this means that bagging will +uncorrelated quantities. In particular, this means that bagging will not lead to a substantial reduction in variance over a single tree in this setting. @@ -1903,17 +1903,17 @@ The algorithm described here can be applied to both classification and regressio We will grow of forest of say \( M \) trees.

    -

  1. For \( m=1:M \) we
    2. +

  2. For \( b=1:B \)
    3. -

  3. Output then the ensemble of trees \( \{T_m\}_1^{M} \) and make predictions for either a regression type of problem or a classification type of problem.
    4. +

  4. Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.
diff --git a/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html b/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html index 627c7f756..cc7d97d36 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html @@ -223,7 +223,7 @@ MathJax.Hub.Config({
[2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

-

Dec 5, 2019

+

Dec 6, 2019












@@ -1806,7 +1806,7 @@ plt.show()

Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with -a decision tree wth different depths and perform a bootstrap aggregate (in this case we as many bootstraps as data points \( n \)). +a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).

@@ -1908,7 +1908,7 @@ bagged trees will look quite similar to each other. Hence the predictions from the bagged trees will be highly correlated. Unfortunately, averaging many highly correlated quantities does not lead to as large of a reduction in variance as averaging many -uncorrelated quanti- ties. In particular, this means that bagging will +uncorrelated quantities. In particular, this means that bagging will not lead to a substantial reduction in variance over a single tree in this setting. @@ -1922,21 +1922,21 @@ The algorithm described here can be applied to both classification and regressio We will grow of forest of say \( M \) trees.

    -
  1. For \( m=1:M \) we
    2. +
  2. For \( b=1:B \)
    3. -
  3. Output then the ensemble of trees \( \{T_m\}_1^{M} \) and make predictions for either a regression type of problem or a classification type of problem.
    4. +
  4. Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.










diff --git a/doc/pub/DecisionTrees/html/DecisionTrees.html b/doc/pub/DecisionTrees/html/DecisionTrees.html index 339c930bf..4c136f88f 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees.html @@ -228,7 +228,7 @@ MathJax.Hub.Config({
[2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

-

Dec 5, 2019

+

Dec 6, 2019












@@ -1811,7 +1811,7 @@ plt.show()

Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with -a decision tree wth different depths and perform a bootstrap aggregate (in this case we as many bootstraps as data points \( n \)). +a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).

@@ -1913,7 +1913,7 @@ bagged trees will look quite similar to each other. Hence the predictions from the bagged trees will be highly correlated. Unfortunately, averaging many highly correlated quantities does not lead to as large of a reduction in variance as averaging many -uncorrelated quanti- ties. In particular, this means that bagging will +uncorrelated quantities. In particular, this means that bagging will not lead to a substantial reduction in variance over a single tree in this setting. @@ -1927,21 +1927,21 @@ The algorithm described here can be applied to both classification and regressio We will grow of forest of say \( M \) trees.

    -
  1. For \( m=1:M \) we
    2. +
  2. For \( b=1:B \)
    3. -
  3. Output then the ensemble of trees \( \{T_m\}_1^{M} \) and make predictions for either a regression type of problem or a classification type of problem.
    4. +
  4. Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.










diff --git a/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb b/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb index 1885b3cc2..2a7a30dba 100644 --- a/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb +++ b/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb @@ -10,7 +10,7 @@ " \n", "**Morten Hjorth-Jensen**, Department of Physics, University of Oslo and Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University\n", "\n", - "Date: **Dec 5, 2019**\n", + "Date: **Dec 6, 2019**\n", "\n", "Copyright 1999-2019, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license\n", "\n", @@ -1827,7 +1827,7 @@ "## Making your own Bootstrap: Changing the Level of the Decision Tree\n", "\n", "Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with\n", - "a decision tree wth different depths and perform a bootstrap aggregate (in this case we as many bootstraps as data points $n$)." + "a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points $n$)." ] }, { @@ -1944,7 +1944,7 @@ "predictions from the bagged trees will be highly correlated.\n", "Unfortunately, averaging many highly correlated quantities does not\n", "lead to as large of a reduction in variance as averaging many\n", - "uncorrelated quanti- ties. In particular, this means that bagging will\n", + "uncorrelated quantities. In particular, this means that bagging will\n", "not lead to a substantial reduction in variance over a single tree in\n", "this setting.\n", "\n", @@ -1953,13 +1953,13 @@ "The algorithm described here can be applied to both classification and regression problems.\n", "\n", "We will grow of forest of say $M$ trees.\n", - "1. For $m=1:M$ we\n", + "1. For $b=1:B$\n", "\n", " * Draw a bootstrap sample of from the training data organized in our $\\boldsymbol{X}$ matrix.\n", "\n", - " * We grow then a random forest tree $T_m$ based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached\n", + " * We grow then a random forest tree $T_b$ based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached\n", "\n", - "1. we select $m \\le p$ varibales at random from the $p$ predictors/features\n", + "1. we select $m \\le p$ variables at random from the $p$ predictors/features\n", "\n", "2. pick the best split point among the $m$ features using either the CART algorithm or the ID3 for classification and create a new node\n", "\n", @@ -1967,7 +1967,7 @@ "\n", "\n", "\n", - "4. Output then the ensemble of trees $\\{T_m\\}_1^{M}$ and make predictions for either a regression type of problem or a classification type of problem. \n", + "4. Output then the ensemble of trees $\\{T_b\\}_1^{B}$ and make predictions for either a regression type of problem or a classification type of problem. \n", "\n", "## Random Forests Compared with other Methods on the Cancer Data" ] diff --git a/doc/pub/DecisionTrees/ipynb/ipynb-DecisionTrees-src.tar.gz b/doc/pub/DecisionTrees/ipynb/ipynb-DecisionTrees-src.tar.gz index a0743d6653ee058735b059f143d21bfa731631de..bcf32e0b9cd246bb2c0b76e245ed27a75ea704f5 100644 GIT binary patch delta 30 mcmZ4cQ*iB1L3a6W4u(fOuNv7~*%@2enOfPIx3aS=s{sJ8^9mFI delta 30 mcmZ4cQ*iB1L3a6W4u;41FB{og*%@2enOfPIx3aS=s{sJF>k4-O diff --git a/doc/pub/DecisionTrees/pdf/DecisionTrees-minted.pdf b/doc/pub/DecisionTrees/pdf/DecisionTrees-minted.pdf index a60a2f889e5d309a13fcaed8108cacaca8eadb6c..52ca6827cafb42f996e658c718ad13d262829bdf 100644 GIT binary patch delta 18044 zcmV(*K;FNpvLfKMBCs|B0Rpo(0*wU(zQ^&CmvI>g6NB3ahua4Mhua4Nx7!B;IqU*5 zG?$V611Oh1^aBuoxcX*Iyye;NIT<-Q*gil_s2dqBEiR+c-={XP$)1*SX@G@7q4p}E z9o5`im)F?_t@CooraHTuy%}pc6X^R!tW52hm9Eyvif6^;xsovgG%dzb)(beYy^u&`z72RY)SYaP&UUUy{@-g>GiOAtp@ zS$OWf$RGvif|#F<$OX{Y_i#cPPWB5PqVuqS;)O?ZIJoX0Sd9TW|BQ7C9J{__D2aQ^ z^Os#xg`TE=5d>4$Jeg0j!yh~KlWF)s)Alp-;Bn04WvDJifKeyeMut56(RGgZf==xh z2D*{Q+U^HyMqBgeBcAli*yE{jlc4no|BOF&5WX1c;27F5bg{v$$2#vb2Y<>HYE8pV zw)IAlFU_kcP4rg2YhDSovMA58;7+Ni%EW)g@0dP+Xp5E-#E%1Tq^K-QMg)&Jk+T+A z`0rhnhW*4ghVSpc7+k#NucI|J&-#JK$+(8?0yN`DmkHDSeOySeMXu2Pu=o$?;|h)6 z8Z?m{-SyK2xweNcQs{cU%b#W5bG*GHL7_`VkhzH;3C;@fAr$7_hcnooq6 zsTiDp23eQ~JhlQ0!`Dm4OZURxG){v#eY?9Wia7jmG##IU_#_MB{$fAlR@?9=h8vE! zE(|CgRcXNt{N3)|$N>keKA>8`)>j+@1O6gd(xQrTKq6=Q#e!0kyb9+Lb4wgP4#XRQ zPCSxuxOM<}z<2q#W?0|UjP&5J0{6Bgw^nC{RsvMStTdIUv75@qQ|SzqqQBZBlO3`wDq^uiZBNsY+&K5HnoB8v zvBOh8wBp6QaMpFATc|cFL zv|MmrS(ZydQW!I_!sE(Ch&3Q*PLxekCZ}~+o5}EuPwgFN`te#WR4JGmixzo)cg4vw zc5+$=k=50dcc&nN69}M9=KB{YyBFnuf|~8n&DNlLTi_ro^BQf-hqXe>Ike-Viz#mh2}4)M6&`=(Z*AG7+Wbf zUzt#sS(K?t3Mi5xTGo7G!%^{N)7gDL47LFu6Jtuf)vIztX(wB{q%ta066ivIM_Dc! z`;Io3q3bqn#Vtema9NpMQ#07`&|yz8w}|s7PRk3vM_EOvE*E9Nf(L!&q~;{f&lS-L7=bYU26f9yvq1<(lG>jY7NcUO~_dWXy} zb)lhpRMPo!Wg^@AX7p)p8U~G4n$Y*fR*+b9x}|%1qbPf3k|d8*ez|FXksDk3EDj%- zr*e@Hn_3AWi0S?uv9)icd2t?PV~@44dG#N58!;0q^W8=iVi;~Z8!p@Hjc5uV>9!GF z%j2zanH`|LDN3P?mn4;fP3(B&>F9kA=Ir&L{z?3h=qjo{jIpsZPcT za-whi;gj+rj1;Y9=-EYoGhr~PO3FFzq&vn5Zg&kC0zgZmgSj|0C#83-k(4=P(gMaR zDvPa*T5W14WcBP^8nxoQ5Cb?5axkS910nq8`-hw78-Ivac~n6t=TT9pVBg;S@mCzw zu!QE*pt~1BY@zzb3Zu9QKHmKE%GFv3d0y$bZ~sq`6{koFaH3Ry?ma(K>5L_tQYwWn z!^4uQV9pg7wV!IXPrAvb@a4k@rcTIm*y9A!r)*we>z)wzj6C9%V2^bXeb*EX zW9(lr9{dFwFUVYfGO0OYp!AVCTdvZB0S;t!--x{Q7?JDn<9zZjB9^!LUIIx$%~@Z| zGoP7q-}}5W7)o^ib~R3P9>teo3&*G2U*JeO!U?yYID@2)mA1ePio^-X-9F9)i<+{_ zW&;;c`!phH6%w?fD`|oq2q?v3l+#JbJzhnY_lwI?lq6|?q>}Vn_6MR5NB_*=Ii~|}a5CVDM>YgbxGlr!i}P4it{Bb9qL`lMYvU@`#+GJ(l1&)jU)LJXe+)MP zCEES~L1&+TXG>k>gpXz!Kbk3KUaO3?>JLJ|qTaJ02xdVL*ids%ci~UG_B=bfmLYfi zK*w+jegIN}s6s1ZrYbg}-;;H`CmIvFn!_^!SDcsUnYbEEX{9yAEib zUq@27|1Z~0`toPv{CqZlKhL<6Z%}S>T0UDo|NR+w)TH@zjO5Q%!lw5?>#WSJ2t16aT-Xm2wqYc*rBk*}< z9F~@U43Kb%MC+Sja{C!%xqT}JRF*&$*WPV>Guc|xjjIP&zAl~Le9Of#3VDzN5Nnfq z=YDllE5dPp94F7Gk!Bk71oE6!^lF}O_=#~Bl!^A(0w@Cd1 zQw9MvmyrDfDSxF|-;dk45q`hF!f%yylKv3YYXbW8Wz_n#th$R`4 ze4D+$zB8PmBzl$IEBa!I6o1SNznS@lPB+I*y7}7=DL+5l|M2EhCO1-w(wJ;>f7nQ+ zgw2ahZmh_Z+1&3pzb3!9cHcGIJ7rDsX#Jd;_J8~HvFq#c)NKEA|Mxeaq}`OF z%(EO9OGR10M(IZSb+z2&st|>Sj}0t)f2v@OR3&a&!#DTUuAQz>)eM4$YEf#vxg&#;@&2_flowH{FwBeTafAzTVT%v z^&Wy@;hp+)GJrI0o7v=vESwb!3(L$3om-I4@AY}*5J0BXW?{tGR}~36p#a3yTBtOK zmqMtlq>=Jw)US6YP2u=Ko339GMWrH7i?GiFkAF~$5muyHg#qv2RBVw*di-T&!3@k@ zmSOydNSZgFlw1tGQ#l-s0R>6{CZQQ$QfY4{MTWt*Nt};YnQ zCa$@htK~Vt+4bH<;OltePcWfw>-I>ky&D}LHzMrl+WL2oRp_Z(UhK-I1r#m(h!K^O ztbhMrJ7pz|nohZKV_$z0FPaR@t4lPQ5mIM=%w$reS@4D;6{b`hV^X1tjJyG`o_VYn zk`QjaMWF>YY3U7ZP^3v8VhVIku*ZP_&^4g*wn{6Z0M602_dKE8}(uu@in zd;9mkx=>Jq;6%tN1E7jz7^}-becg3D*Dc|vNJb7UCD@xIBJ}o^+hOx`vrOZ@)eU~Q zHuR1IiDrk)GfJU^4Fo*rWi0Z+K8IxImfW>dQ@DkxG?T7>FLxl)l>6 zyRqwsHwVDlwXZ;yAwH8O5K2>QG=HpP3(kCajt|ez5j@IUYk82-yi4uT@)f|ts`{4) zavdKA;AMbB&adQD9fU|6PkpZGymP#6ECVSG1d2px)+50rx3XYiYvM_4rGa+64s{Tq z0_FnjWGPG(D1#9S3`tCVetiE6hRS=KRWkqCVr*%sr}OI-!8}Fs^G8K52Y>oh5|4p7 zF_pp9Je7f=JMher=R10|)nn}9K{%^KVOi5C8Z2po71E$LL$1-~?9IF9C?K*B4iQsf z$cgjf6hvBdY){o74es6bc=8?Ipv&Pm(8P={GoQuEloNg?o`sY-rT`9dJW`mar6EQC zaUCyLFQ5qtYEIuKsoJ6PVSn<0a(&=whFH{EKnVasM>8tk?XUt<;ERO}vE`LCW}-9} z?F$qS@U|FqTv>|tQX-dTL8)L+z8C~lBNu5=Fb7Dj+`NoWz`8l&I27eXPuyqeY1fbYY zm#Mf|usmK*u<4@O=?)}CFv37cdu3cZWkxS*r-3?t?3#<~F^OaAvBnw&-#UsYnK3}= zec6KsLMMTI!E;eqe0sS0#aw$w#KXI2sre9h-j|?M`-p^t2V_WcBn@CGXwV0DJW^K@qg{6GE_mk{0R#Bk`jeph@ zuT~(a&VH0O;HDN4Tl*@<$w|_z_?_r zdY(*#9W2!EtA4p&K8IaJ3c};qOG%f2Y+ zG=zNuWIc>kJBjb3Y;aZgTqd4>s%Hvm5dNgYbw_-`?~GYdCfD=G?MI}7HIu@%5;`lf zTw$N?gNwE2CoZp0lNvJtWK`;$YW;Ca4dcnN|35CmjDPR0=RMci)flGx5}M*iKrIfq zLL}jm*q=ZUReu-4LWao-Ac@ulVAZT*cFC)|<+HC3Pc)J8CVaudMt^MY4APdCzD*^9Z?OTw9*>LN ztLAcYTx0k(yP;Z#5R_|q&garS3{?+I8*DNq7#+iq0)N0-Qx6L(?I(!v2{MeTn%T%y z!-I}ZxXK}CoGuFv4%zf-V!%iU3|xmBp_Ajwvzmk61BFK9XdEE_y7dg|{T6!m`x?X;Ud%uu%L%OHPv~w3 zdZIDMd78&NK3M@iWi;%!i)$HVxPBXv)WQ4gVqK{Bnu_SZ>v>NM?L$DfyoERNT5Fj# z9IoxmJ5qu>dtMq?ND4AdmpZ^)yC%HVb1yN1?tlE$c1`UtNI~4V(>hDu55u*2am5i2 z{T9d?Ybz_lZNnVLRC4kL{%z3{aK_rK6+N;5lxcqr24_lWJFHT{!rJI^CHqTr;EmSa`m1Stbi_U&14IySNVFbOk~++WjP222-0VF zQ<2rQ$RO{eubSTcVbi}|tw3~E%)3alsub0I&y`X)O4%E@$k$|k`yG0BlZB;LLgwKv z2*~kEHNUi00CkQh{jcZsU5QGLFw4a`uYcr=+Hn!#kT<#BHPy)UF`u~S{09bd2^`Rb zOFc951}>-Swr=X*U2y)}zgN6e3ZR}A(|;3w`5iDEPF3yxCA$^5zj2QiSwm)YR|nx2 zjx(C`9k9-q4j6y3c+R(%_sP-};FZU7-oA7Pm_jk8id(_{4}34Bkw`xDGp*$-w||3B zAX@ksO=IlPA3dCQ^djx3IDgHhEn`iCu#l92eb>voMlQMo?q0Bzwh-2eeDydYP4Mnm zD4WOsZ$%YCxD|%t+V{csJJ=oh9Me<1pjitY@V~p+M^=Od8gA#mYhFJ{^L=lg0~HUt zxi^vuxG^MzyQ~z&bQPkbK|kI9@C9Gqe%{V1}51nW02d zL`5k{DWw~fkPeY<6e;O05fBiNxPzXv-Fx42f8K`&7Vnqu`f7%o-N-~p8H})pXdvKd zAu(amn}D9aJ_rrfMuS{o4mX61Ax@sIAS56rEG{Za!p*IUgn-a61Y8Y&gofM%#H7(s zzy#ushPc^7kO0O+mV_J7guo$4j2;Zw`vUq9G|1G~9U=zsfqnra1PU!=4?B4JKY^kobQA)(7im)ucyfdL>m7|;>c7X}Ov-WVYa;6uOx zdk7Td>Ifhl0aJ(tU}mC!ZfpW*8XKA!nOqmf1UB(>cSj)q!$sA^)J#(lP*XNARR`|9JgTa5v^py>? zHPlT^h4e792N%*uV5$QbM*EaD+G9@Bg86gu%g%mpuS`x{H{>VIH1;5N);pxL`yie`HP&G$0B< zJOGG~161TP@vphL6pLMoF$MAWb4R!Xjv!YQ#2@Ae!8}O(P#`Y|fJS;k{Qdr}_$ML} zy8(b<4rt8CW7dY`S9EQ-BLa~9EyghUuj)U;z$Y$@*{185l?5Zm;z%N5fIwqL zgYW;(#{DaQmxiaSs{zOj!uLPx`fp>98_d=B|2F@ZnK|T=HJ~HBmkx%~fcZec zMliGk^mmDW%YR2whC8`J03k6+VNnTb@!uM=%T05|j5Ovf!7dLQMlbqr9cEe`T;LEC z3XqWerGjAU@-On3W?WJSL^O2F)l@YE{`VyPvQ>wFJ0QR?xDy~QB@KX(NRTgyC}upw zrKA8qG0e<^AwIw62M`g4BhVNZz}*w=4>%%_B$qvwmIg#rE=9l44KYAO^AC~)M0EZj zDL_Q`54r(}=>LUe!~qfGKM3Pz`WKRw0YvQoLgJXg5V6Nx?!P3G7;6X2cKwlD)=b3l zZxEAz8TL1bq3rxOh+*RLC%zaaoh!)A9{l%sn4E5ZpqMBi0{Ox1jU!GE^V=<>q-T3OLQef`hr|AkG^NQ4W-90taG@cqpM z(`zKm$66G#Z( zAr8??zA8tVRo#W+-VqBPo$knE`E*qqR}-|^_s1OOH6HE8%{ zVg5tHxTH{AA3MQOxk){SFLuK|Y^Bo`BZIF#NtUr9{3+%I)b!k|Pk*@hh(z&!tBopM z(a=_I&{$NHbMl{h;sfgZ@*`DS zD)sa9FqJ0u$$GY2npXR@!6E+0bXm2I<4K)#VIIc(uYGFH4uKHk)h=1mpky|N$v3J= zazirv*_B7CL9N&G`^aCyH%8un3J})HDA@V6Q3S>G5SLYc?`A!eJWg3JbLb^$SP9BIk2h7H@W!+S1$R{6dn9L z{5x^b^wD{Oy?^cm?N$sP=i=8>a^0-jf@ds7Ii>HIbRm3kNtSKZ1_tAQL~N{OwoK&X z>SxEQ1;S)JNfs;JEP0ach&hLkAK&ZT=Q)c*bl!Juvh46>9{VYIKJjxk)x7<_+Z&&2 z9O+QYRr+AEVO3iTWam#CQv{xv1jFtH>9^!em6NesaO2pG* z4LVdhif}#sybMd6?;3S5h_ajH?HMW5DN&d>q+@lc$3I9WCV5kTuH?dGamcIv+FBJF z_P%K{F5^TPITIKpqEa0b}N_%7WE;ZOSK0m3YTLRpjaPeM_m?Lv2-1!VpM8}jeKnh@be{-Ef(Ks3|JEeC z=bdHAJ1qD zaO8 zHT<|<0;}?w8P(xA-^`!CJUP`*G6@!&_EfZ=GdMoZ)m&gVnqaPa8BYvS$=hg_qFs=B zc$=l9nb0hM8cUC0p4PQZ+_DNx=c;)F&z?WKEYm2)RZnjksOGd4!ndX%1(g&^eCu5>tkONjlE z@|3V_Y6a~hDJd&`1|paC^EU?A2Ct+2s;~CnDyaj1KPp2kmTjh2aOA9ss~PsK(#GFF zeSY^Gv2w52MEP3B>H=K5$##WJNWHuj%VwHm9zSEhRG<5LcWp-RZBkNwe}ltk7wCea z2RsL#^pW9Po>}WF5{=i2S9Y9b{cJq>(mxS)istPIl1Ka`LwV+=mpY_c3)u@-Ob-e< zWM?sdcR!~g1ktRM-If746$%S^lvG48dIxYmK%Xbey#$IGkid~wnTH_Je^z5+E9{>5rbY$(mu#bLJTr(ApFXWADq z9x|3J>f**!IgevBQDGNSkud9#g#5-bDx^9#?Q{%_N4~}OhV*NSl>uT`>e$tYxtZsG z6rcJ$<^`He^*tnJ9*%g9eucjjw1LWrOY#pqVF@vjq?pKQ+OT~wG9cqdzazGmEf zC9y(JFiNo@;gItF_l=zey_Wguajj?606@R6gN>E$babF3Hulit z+R|Muj*Y{_jv?>4*`U(;1f_&*^dRqnl>Ko%3A zEL!>zJ_i@*vgM_%jilSQ`3b-G;rz*`o_wr2Z4jz$m>XuokHevYJ^oDEApquo;q_vq zy!B?Sh5NT0uDEqigrf#5wPEB$K7-fg^eJpI`Qu_Lf00dti#~^8OCxUO5-m2xwL@8% z)z)cxDQ;9Zwzn*3QuFHLp+=o0O)pZ7eiHNmC53LJG3Q(%tRvQ58PViOaMQ^#rY#`o zrP5aa*FDhj{_?EEXG_V$(D<}};y2ctVv;_Dv!&I;z5!m9b~URmb`uuyMeJ6&48p^H zvJ(+Mv*`Em^$i!o#_1KaZ!z|L%b6O?wyl(GYJI`Yv&$Vlo+&HPK*&s#pqEbRvGD1z z9{+kpGkLVS!;kxN2PNN;{`kGg+ARY0qTHn*KkxCd3OFNUl%ax;mDu}#L2IsI-oYXE z(a^HcB5QbLmkMG_JwlNGtgq;CEmm+$I^Lk$s-hJ^hev4rAgn^rP%7Nri@MWee|f`_ zGAX=oR0@gB>^Q(+s3cBO8q&BaL(w5CDmmT*M##uE#Rn>SOn+RW*<#=4f~IrE@nlh49^*c$`z7P?h3~}#a|dVx}(9m zU&jyU{YvWgie-B~>8>jBm9J!Prkqt-SY@U}N>~-m$v@;epOcWHO^1ITm|>iYUAzb zrs&4o%k%3upH#)ALDH@_ac``f4iC&dwj;x1&TCLpAv`P`EF6e%$Bw<5ra#SY2eIa9 z|4`)+^6J_%37y!VPZyYjAggpmZU}bXiQ%J|wz8k58|PDG*NN|Bh*qLm4khbH8`iFL zUEpF}F=0}FYF=dY9m+#M@L&mnSAQTS8h+5c!S#@5u5L0q7bx+Y4ZgRLJ&ITzB`n;R zY+Saj4W2icZ6i!wCam{F>_ohvPA$dC`z*N)J}8@^F~b_WpP1hq_5G-dOzyb}R*pq* zZkWkV&`|C86?4LS2}y_MNr5C9-|Tm3+j;gm2bYe2+JH^!0;%W)cK0laJ}sMx8F!;v z6Y~&j_G*tdZPS@@#Zb-l40}9$5sUPAb?JBeg`uDB4C3%hgjL7;kf%z*#}ZcPdImlm z=-Z^)71eNvJSK{`H$VY7>f_zf)O6oAdYY1xoUIQr=Hz1r6xAwHML{p|-h z?MB?J2Of>vRg$FCSE&cDXLk3xnj{~iWI#B7`JA*H@av=H)csk0+;{wm%um6i-7fm9 zZ=teymiN}kDw!5vh%i^rn0a|K#JA;~?Q9{Ba*F+xuk;5|d4^ZAC@Usk>BhfCXYo~G z^~GTk_%?IGhT9zh2bAWO-hJA*TTfOF0(Osh=nhsNUVkBPoXW?OAU+hy@2nGs_Ay3( z^s_pD^Y##DD~=HMrN8rzVS=+4+r*vCV)h8HwL4l#=779DN=T8MY-8^J!|iC@>T4{y z>+)udHyUQl#a13?1y0{E^V#Uh)WUz$sr72<>CNhO|DGSa@1-^h_!v_Rvwq?X9H*>l zW4`wtg)9ZN7Ccqj3p`G8_`Ff=^@Lr20eIY;#P`4=w@m)cgPLjQwE}0s>y3i&dS(dyhzP|~Vv6&KnddC^}OrjNSqt|b$ z5M0^y)}t_hbty;iWZ1~ERpZ_(aNoBPbLaeYr@Cvhp76UXdXF^34!^Q1+2#0u9|QID z&wRSU7VK^fqN-cBPu-H9mFCTzhefQ@`V9!{n-HbGXwU$cf0H}h6y?c3({zZ`O(&k^ zD9@ZyZc5dCm9#K!u>6JQ_H~c(M|KARlVhCPfOlBBop{4eU4^^5NjBpXnvpS@g zZn_C=)kD1~whyq7Y3ng^(> ze)>96#WLn$I~=iev%&ar_XWxy_Zcq^X8QBSx~ zXMdJqH`ue1&hu@cKMBICW^-Gj`DO3j+^<4%Fz8lf0=dE&R@PE~PA~t?V^(gfcM%=O z$me*1oTFQ}CJCvdo3?+>roH`H+EFFp*VrDthx?i&E_wIkt;|-nUMp6^PMC8?+AIz4 z2aYhZLEDa_J*&!G;7+$h3KC^PKM+WspN9h@(gS^JkI9Twukrp|nm2w|1MU-dYyPm9XdRQ_=)&QvAH(K{ z{f(TS^abSf?I`|Z71j6w0@o&(nNP* z>k8HDD`mMU3-WB%zL__x-Fp}-jB332r`OX=y_$)C79kV-5`#4@a3*>*N${6ayA6#m zeWnUAom=$FU=OH+(|nz*HOFmwzT*$XFRlukM^f)pZ>hy1$$6A?pEPiFQK6`Wy5pO`PY5pLPKH43g&TShg0D}-ZasQX=F>@w z1Mx6_A2o?vhL0;TiuTD8!XNgg^vT_?1BfaU$MP#*8pfm*N{!}`O(E@0^@p>nWfD^T z*+s)=78ydT10trW`d`u>XHggp6kIR~_G9C6>v1XD)czdAD!MtDo*yi@d=}pNw6X1d z5i!6()97!_hhIo66eO(Y_nJy{cHl=*a61Km2d-Jhz}087wl{qShPUBTIBVX`*(HaA z8nM0!zB5Ms_P+WIlIba}-xmh!mI{HhIIB9FPTTwA?dhL5A%{QIMsMb&-?j3DeL+qx zqST>;!div8Vrmdsql3^xgD}K!!ftNsMYWWzivRXJ7_19Tp_qQ_dZ!T0*O$wKen&um z*yZ++H(SMl5W7?%(w+Dn`(#C!W>KjWPQ_Va%&i6i<%(EY8?(uIJS1<BO6!VWGEg;hfU)ui%AdbVQk*;f9;q<1%M8fj)l;{&uDHe9s_e zpbSs-@MTBw?6&@KqJyG|R(*CyiVfE6$XrUrA>?IttWRv4YIL`>z;ae9S!iH?wcV9I z7-IC^cx*i1SR$Mbmj$i&ih}q{)yTFXmgQSM-Yf9Kv_n+5k^QI%@mn8_SpiuP zyLfPHdad@VpC?Prp=EYVWyfQ8m60f+8`4yHiuv=_4-GCt*5+-cUDw0YjyF?I{lQmb zBB6ozHQRdwJe00`f9qwCsrABLnD+9}h_YvJl~a73APk`UyB~GL1d4foevHe=l1jFQ zi)7wQll=gY`IHR(I3<>)U)v!}8_Hqp4>4S}zJKk^NE04d7+&MM@xh12j)DO1`p9f= z(#s3CEWxTKUMbHHSbRlWqwZt2_C=>{D`Hb+XWbQEavzm6*DP>0XICiqS?L=}b2T$U zMA-v-b;v~-Rzk~4G~;rA?;L5tB6x;df`*skRLZQDUm6opvt(Li6PL3XP~L1st~}*Y zAk-I%vK@n8yn4SxG&4?9HdV-3nY)*@O$Wkr%hwTLqWO|*W1{LM6TG}49o0;E+b^rP zx*$}H&iQyCQ#O#ZzVHpB%3>a*!XsTpT0tW?6!F4-9Af~kgo@% z3m42(Gq_1I5P%NBfb_QA>w=~XL| zVLl?OMHc%?qD`*toLB~HO$VY6GMe{z`{gMm;;UozQNa|CcF9MEb#fKQ4@fEN>Q1>4 zbXbiaORr|`NsGeS*NZdPH(o`5xW6DgujWdoGF6kf6RNa8!Z^Q}=x_L}Qkx(!ikHR&M3T_CXA zj;v*d=v~I*WDmpElM1}oEQTLu0}k%1&9=8Px{}HBX;L z6_%w6j1VuY$Ci>yjbe3bCE?hNe^J#7~x&aTy2`hxq0Nhxq0OxA^7bmGVRU66 zC`39kFflnXGB7qUIWaIbGMC7M2*Lt5F_)4311Nv3TWeDrITroSuh7R~w?enxQq|U0 zB_t4*5GH}VW-1>XXu!M1W^6N??62Q*CAF>YcH4O3s?ex!tEF>~?n`PSv|%L^+A*aB zetYI25r#9aJRTUyxaA5B&A8_d4amBQ{JPGvrG$*QJrU6TXgoEX~!Y)k{Pb`p75F(86H_?jL=MA zbmKH$SBhDVjzV!}rGz#T5>0reW7Yy4*NVBa#G}$O$7KmEnS=IwTr#JC29r5$pcH4$ z3OskrIZ{V6U!p%|0}DWmdx;xGgq4nQ!@Pe1Bk1Ct@OuISdJjDno|PO|A<&SRS7X#t z0Tw{>(g8&6&WqBZy#AAFN-&^#J&(kVt0ail$>)!7My=i|Ix_a$^ zHC}b_P;xeo(U!8XAQ_fzjBeXHWe}Z-;9fFTZMV%4yBPDUU8ZAwkp=K1A|-!9MqX>I22V&L zpXP&Mr{vB^%KmlwUrBNBS_RhLMHe zJa`YR-6jS|nplWZFN3WuS0%i+=o_CS4+KA@H`KCMcorFo?Q|{6uY3Ra$6j&y^78cT?Cr_l-<-Xl+z-aQeD-28 zx?Zi5LXGr{szOdN0&L4%E#FX!Zu`|LmrJI4wR%y!F8)=#Db9-b#YJ&h42o~XRdF5G z@Mc)t7I($nkB5J|YFrG9e;5BLM#X(GE+)l8@lZ{Nlj~wy%!)ad`%f%*Yq`}n6urGX zK0i8NP$bq-WQiiw)F^)fMK+_zwN_p|jiS=`QuG`o9TqQ(SH)3r47yGdVdurKiL?*J z$Kn&LKazJf8BNBx-ro-bk=6IB(cpdspE;X<6n{Kcv-y7zZ~pN(nS)y(3FOX3gV`M@ zp5INYDiICOD*o3b+iXSl+lvpcKY{GC!Tj#k_jxs*VP@Fn`|9>_G?=dD9fc_@(@+_U zZHKjDNxLd@*2|1@)IL3&o1AjVbIQL>2Uq{8=A-In{%v@BTi*v#g5Y5=t;Xqbys4iR zvqun5L|%UnXAh&nkELK#%QF#l z-_&laY;g3iPcILT8rs?QcEpo<-0~f%XT=&bIu0gu7R;wMnqV?FwQ&sYzg-W4aXds5 z2ut5t^sO?$Z*Sk7A75liaxEoF(wch;Mzrn5-abzP%{=LzmM7N)q`n;RwkHc@0jv?c z$oYSU)2qk(n^E<>n<+w<>W|04D9al3Z&-wXf5Kx&#m(e#+LxKf!*K;~o#6VgjFMT1 ztX=nE^f+6Wb*lSegg2}H#2mBX_f{sF|91jUe;i*|)7jNzS}hY(@xN+1X&S;-+3E85 z@bmd)Z3ycf|CvVSdlpWCkdVj@%um|^wjY1JhCswZ^QLxNrI9ahKYsfBxz^5X&@NQ) zbSh?4uyXy=+BvzEa&D`@{ruf;A5Op2%58{h%1Kg=j=)SysM&EF*wH4jqx_DR-kLQ5?l9~GV%2|8 z7qW{ofg(;L-gKi)z8#3Mq081*`QXdZ>tBC;U$f;mu%&8SqS_Q}xu0gsmuuO2nI(V= z1pf|a3;U{T{N+}unI_#<(ed?%(+_87S(=RC`;5q?EGeD&DI>~i2i?0JqtS(P6r=0(;DT|U#pl@zd0Pj6b26u)5-T3 zTD}dYA?sI*{GT%Gx5v?_LWUnz_n5D)x#LSzZpYUBaEzF9|2UctA4WfladjKEFERa2 z2iL=^!KlHf-!4X9o_+av-sE#*b?ud+ZyVUfc)kNN*O*YyzNy|;LH5h#`$9g5#j2FD6Lq`kU2PIWRfJQfzr$VFAn zj9#^MKq6iy59UO31Hrt`%4+6q<52E4+o3NHkMdKvP=kMc5})OogA^vpjps1TmCfT7 z=3M6h=4ntkuTLpxyBW?TnBZxr$%HJ%k4%`j*`(73rd_>)aaF@{MO=T(6jzy1`Jn|f zhL(a^UK}$k^MjJG2H5f!m!!bwCvZ-Osx|BMDLV50RMvy*GqH&ANY4%bKt-H#1+Bs| zGvo<_37sb$rpr3AFs+(lg&9kF!|3Rv>dMqan;yzrR3p?m=4slfvd7(IsGXT|4d-H7 z1ka-N>vthvyh21O&ai)5DK41jxg|0$TYKuojyfzHR4Ma?!@XOza0vHh@%?Kzz#^DE zew&9|5+aJd7|beqvf`^tBOr&rGofHs91W+lIR!t=vRAwrEN(E`r{VSd4$Cj?>HxoP z{`Oaq_Vltdec70w?iTZ#?dR|NlF`PY+c<>lqd~`NaA9BJFJOPR2G;lepbgA!0G9Ei z{Vss5Io4`nvt4Yq7n|4xw*C!i3!3Y!Cw(JXLgO}r#&-d&fA`mhX4gY&n3~)LwEi7f z8=7xJ^R^F~+Km@%{;<<_hz`awgXmm_Pw#2)I52D0Y7VaKq<=HbDLKgrxN@@NL|$w-9>1&H_2lqMqtRWce3NaMK33jZdz_ zS7!NC?*iSX{c9VYUWZPrj6As$@;2-<+weNknf`ucxwH85x@a}c7W;qnOtCyC7^f+Q}N5I?#x4k>@ESHlwf%T7J0h#cKOLFDKL3L-~0 zP!Kt$At7=aYYm_=vvg7E{Y9VEs3dU{|CpujN!-LTwU!V!HeYB-+}NBON!<8GAJRA0 z8Xa6T>!Mc=Fe1n2ic2ELHxvgsqHKWoB+0=8mulRX)I+R>W%if7R>M-9Ue7gGV2A!3 zHIsixlH}{E6iISrlqN~8jKbtXlq-SJtczYfz%0FMsny zx5`*$F!16Em0^(P&UI>Sdr@y|o%2z9N-NSAaQ+929TA%CaXq)T7j``@EHC)<{MnmDu%xdNWQ?Esyu276@2FGgicit`(Xig8D& zV!nvy9By)U&uQO=4?I2dPZZ&rP#6x(d8oe8yOByJ!s;`soK-G+}@LqzNrIE0b zr`3<^xPV~kaZK`}P&%bC&i*|64z z-v=Y4!qvgxINR2t_pUstoI6=wRKfNstK?RnJKkTezBV}+(WkiI9p-1fisH-N(%*EV z;g*dnXwW651doSTP6BvNm|>+IzB{ehj-9$sVqDO21e35Z^b#V!qC>_ra=M563fwC; zh&Q&8jg({wMm$3wZV+jbTzA(_zHk-ika_-V^ElHgugkvwY|uRZt+0^GAO*T2 zpkgxFK`KCP@rCz3&#WlPsg!a!rKg`3^6ZY4*f#~-lN9+fkoXGgC#pCtoyfO0t^akt(YwYlwPh+D4lH1MY zJ%lgg-O#pXU|?v!TJw1vJ+s#)2J8G?E^6?njt7yfOZwB#?%jH%`}KAPX0kYi+nVBS zu+?*^v}Z3nMup0~+2Vmlz(CJmB~HViUA()?@pNoewnJ4GP-8?EaeGX6zIBE^d1Ccg zd%}`|MH6SlM`2(A|7SwO@}fcWmnNVyrH2}sa1a(Hy#2wohxZjCnXzrz2TOG~%$1$H z9Ie|5>vW>87|^~D$J2Yl;CfQNo$QoT70*^4T9e}~Z<7LSyAEwr-m`7@dz8erDZ2Cc zm6By~7y~((-2*($Ei)fk8JG(KLoC~hVusTEKl%>axJ<+(N=?$sE&iatXG}Nc(G3ah z-ZtR1Mc!ESMu!PUG8Xv2#oP@P+jYG{d$!NzWBssV=MSBb?oNOw;L7+$=QG@u-p!?% z^~swxTVn4j9B*K@r7-{n!_l&hZZed7d~=X|p(9O&o3!F~{WdBPipJgYQkdjn|76m- zB9qi=BY!$+Q=~KYm1OK}cd8lND1nkdndzsQ1NeY1g^r|?tHXHAAAf_wO97HzHj(@nhwHjFHRm_vPTj_ujJs$B!QFNPi{Z27d)O za&1J&Unr?Z3KqT{sdNTlxD7~Z$D!S9_4(z2`?@L8Tdj{oV6{eLhL_k=LfPsIU9(WM zbx_1$Qh0fQ*j;kgzUk&Is&lWpDHvfzD^;j|T!@@_Uy=DO6=IyBvg-jX9cV89Z=du{ wKhfV+X|vwt94(}C6T`k~b$YGy#M?Wo23J|-YXxFcDA3c(6^K9th7y2(0#R)lN&o-= delta 17590 zcmV)2K+M13wIZpqBCs|B0UWb80*wU(uPpJBm!VMz6NB3ahua4Mhua4Nx7!B;IqU&3 zlko>Bmp=3Z5PtwJzF8A*dG>owMotd44^R{8Mutm^%V_lXsX`UlWKT=EG{8ckP^Fv=*OFs$NPtyZ+^(is7k6l%cF;yHm1=ze7+SH)c5S>gS!- zv1w-hbY1QCX0kgGj|4Ftue3~im;i)yiPmYbux088KH{Tm9nf!HcV!me2C67alqOYK z1nz^#AO+~cGd~@X3!t&@;e;}t>=!&l=5hZ-3yJvB7P?I`1=waLN^GO+!w$ z^+uLYlX2O1~i8nO$}OcPmVMDzD?Awd?oLi@wwKcJ5*G=6K) zM0|ADPZ#9c9!3%O-ALc&U|jdN-JMQTJW8&QG9B}M&klU$xckMo&vK8~5PvnFJf0RQ zIDZYYFb{Za1r~;{myVY1nZIeA26Os$cUKf?{NZRiItB5)EQtGy{ft{}L!TIKIN~}p zAY@Xhf*AO_-Mf(j4p@CawSuj$I0gp%MX;!%N^(H*&h(1~rDk~*&m-oRG=3aBZyt1_ z(F=!b2apGRmws!4^-WD!kINeD@B$Oul7AlQ<7wzy+UEsFP~$pl2AU7{FxZg<^J{iL z%>(-|rrvBN3kaz4BFXcNMEMV<2L_SXXgm<09fajT-|IuAE+Hfj z{otl!OUo}X6%_Aj|>FhyVOD)v4=fJ{=v?-$PK*>;*aGx#88;bW1W^pH$*W zpgKH^qC#8>q+rx+4k9Px+Dzsw4Y2DdtB+UmUapcdM}PHB-d*u$P5$iHK{#3#STVD ztpjL{R=v7gyHszrSr}*smRMggl%N#OV8nT)`3tJhtN^||E3X4xLdg+JCq!3vo3z_JgHTIp~Tz|%{+cXfj z1f|1yWp+(XU_(QP-NW1>&Lf8cErf1l6^pW5lm!a`bm6|}yBt(?*aI~q)Prl*&qb1_ ztgn$u0`yrqoN+&+rXcmc9DN%AGcSVAF9!mE%6zQBbj2$!aFe7U$y-DDidS6AI}8oJ zAEBvNA`~`rCh{iZ3evo0BY!xqkkHnR;Gxj0p#ICk69V|K3J#2MbM6N~)o(W_S(PTL zDz=BU@5hx)Ws*XRve~orPt54TFxvjukCscFVR+C9q5$u%_FftsGQZS?ggQ}ik4rVk z+dj0NPjk}{XtdmJzAv_b#3I#}?&*#48<@$mJQ4ZjrbTXS-gO#35PwhQA|bY~96}J& z{W)Um&@$8FJo3gKt6uXOKI}GPCRFCTEhoe<+-Wvkw%1!u1s`hLa<1hG*0{_L(8v@e zQ^xC%P+$`~9%(uT--G%4M#!KT$Z|`seZL3FK&k?KZnJ0OdvL1bae|x}T73BQc@ah* ztz{h8MKfVAiHhqs?tk~lcli738YWnyujAoAMP1x#LLIEa+>K1ucMxAySfod4%T8FtdfIO zE_s+yL*wN=;=gb{wppRN0|Eu_Qh-&sUTU~N!d0Ih8esi0`mQM&#x%TDJcO$n zd~}B+brB+7E&{b@RK&k`;5XCUk?WTd^!SDciGLzbWXcvW>$?tUo!&td-1J*|(R`(^ za5m1*XVdrdj63ND{ zVI?s@I)D#?K3gpR1wf|@#KE+88r-+wHVjF~4HEQA{{uxYI^MTX{R2}50W*{F2PuE0 zS>2D@xDkKPU*WgfK+T9>`e<+{lFOyI0zrYpqAhUg1J{;qA(muFayEN^eP=jBN%U%W zQ}o56C~`R8^BX!@9aqWfZ*LNQez<${_LGt;DMeweTHPI1QfrZ=`6{)UP};8U_N(95 zzqw}L)|;))?E2HDFzdD}`+?rypO1fSR}H6n^PjuFzx^b$RUwL0rARChMGglgE9vPP zxk_~|aswZ0AbWQzfkvX&ZlvK~_hrB9s{74WiT{>n^b3 za<0Obp`9;khZDW-&aUsPFNrogH&iWC0tsnXS04RIum-+?Ou+mkf>(`7MP7eMlHt;| z_vfH*(Cp9h*f1jWLb=FGYWZrny? z>qpYC%4ZsuDibEnfIolM=LI7GPf4c2ilHk@Vs`WbVAmL-lN1KI(5j%7@@CcVTbm?c zevqc^7Ff}VNRvFA!|tH;oD+X3tTbW8PhgfTUrT!YWr3p7BL8v89yZ<;<~+asq~&b= z)>;s=05J;zilc2`5}Dm>iVne%4rhGq+{|i@48{J1a)6ihLkEsXmUH#xT+Xct7Vdl? z!Pnu$pI}4PRLzmz_HJ;`M*uOB&O{>QQ60 z8@lSdcu{3w15;R_Y=t!H$5bXoqC(8%iLixUS(^x*D~cJ22<0Qf6ch)U*C@P!P8u3H zdUPT4z<3M}TnKKw6b^i>Hd4ss=GZ&lP^?l>WID$UR^&>7bReXO9QQeaw9Q4Li%ED9v9=i{p}{@*#oT{8F8&2}V9J>1$e=7pk_P<~X$OBmN~(Y);KyBgq0|Q{j65eC zsq=L|l$W07+Uszx8uFifJ#hLdLa_MRu5>$mJ>8J0-S?`-57(N$G_5~#U&d)Ma1CkE zWUP@MoR)mG>gh?RWN?aP(Wle`I*|FTkJ-e*k9@Ln)K*al1^Iu)g41hF#|E*KO4ogP zE_cHVPNG1~(!>PV`cVwTY3oa(%o0&hf=t7r%ge=3rI59wbH8nDwsuhT+_~(Le|Q_{ z2_XO~^HG4Wlt6HVhI%0ciUg_84|#a?e8wsuoftuBSOy zw>~&tR~AII28JRQp2-9w-#P;gXExr%TI%W4YhMKhO29XOVdl(4i83f5$CShX>F4*q zVXC~tT?L~g&_4ksm}xz|Jg6sM|N2qW;6R@WLOZZ020VXA%>y1-x`W6Zc)z1ZQ$EBY z9+=CZB4;dVlnv%!A`{YLG()X1#T3oEr(QsiA{Y@%VW^1{aY`Z$2DS%uPzLwzdOZ0F zuh9AO8`s?G6e_*1VnxCQUy0{a3b;Xv7KLp_T9Wf0*Wq&Y{Hai%Xhegu=}Vs_@6S9# zrCt%#TFQSX!G97+^CpW=J1lr9fXaL<+OJeG8&$FBT!?N!inB$>1yXdD7OAum@B$Lk zB0w6cNb;P4;wx#zfYanRVv_fPl4|#LEJZ4$8*tK|CdiLb%@jV7R|ab-BSK5zLa{k^ zjNr+yJbvc78$)ge@C>ZjpBRBH2M=O50Q+mbIq18hmX#} zumN{IjG&kMh**OcD8x6C7BCmI=z}{RDXYPYR|q`ska6$t{*@TT(M{{9gLLy#xa5Bz z_S=6RAb0ij!gul zF+TP@5fuR|7THWJ5LPEvK+RxaO=_RWzQ}**xP)T@cs&edGqUd!Y_pFD&Uxbfr)olv z7Tup@xb6rl_>-|9%Jua;@cRRzz>4wUQje_)>{K|WyAWc{`H?#-l%&EU0L7Fh$IF0W ztO~=)ar{qk*U-4Vp7-2{mqXa@du+-d!E15B6*39;%-#ZlRg;Shii^qtA7(ZglJ zhUdN|uSxNMoDN9W&;*ot^us~Xfl+paT`Wn0hVCufyRh(LFj#bWIm-y;VGc7%9YI^q7 z;gL2{&4e%5(}>O1B4TOi+t?lWH&%Z@IOAbeVG7f7kJunA-c`-SXDG5z=NEqlUw^nLtNE?#_{y1PGDlY~wX!Ki0HrQ5WB3v{~c}%ZQUcujHGr`YT znx%e6=HO+<-++Nl$=eQ#Ffh-u=w9UW!i65*<(k!O5hkCdt3jYPHqhS}2w9O9vku$x zrQlBsoB?8xfJ>wj1`EWDH9fx!+iFN@8o~;IXdYdNMbr|J8SD=vsDpp;ShWCnmB?i= zUR=t>i&)SChID2}T}ZQt)l+SkZk!@=h}HACX73lYW`u&elf0^W`@^b#x7d9cl~1Qg zyC@X3d{2eaiw%gGS^5>8-#r8GZmO_^E<>i_8VJDgOYOeQG6B#z-t@oP>sgKp&Miy#1gom4)=L>!VUBaU#hzL!-e3#{r|;trNPyc zeEjpmlivZt;ZoJ?U(#EiiVOseh%9+#bWI287t9&$dB$0%`v%NED&F(`<#SY$fP1C! zo_8-@|0U3j>HLot)|P$E z0JBh(0ex5Dn$huPg_->B1x=X@(psJ_nj?|~uZ@My()bUss6z-B!q8m%IXHd?rz4(Y zcxq-aYbFE!xi`hg0G!9R@4~jM8v@G09Z^yT%2A&Kpl^S5zsg+e*{KADga;^f-_)(bSEI)9Fcf{WFk*5 z0B9kxNIXf8033Y)D3Sm(^>Ign0U_8gV2JY|h&sYNNNOb335`VxlU&qs?ml?5lQZEs zhNP(Iain8!H8DUJ26x4IdAOnh7#0EOib2Hy1DqF0hz5jkSilkK48x!R911W+S^{P! ze;URnfR^!fGeZ+$F;ZX?qPsf|{~s>uCZ=XuR{)5rfvE<7G`|99nVFa#|20KoN&KCz z00yQc{c#?W;jueZ!&KGO%1{FgI(`QL40s~(9_Zt=|KcV{Vg~%7Msh{rac;j707A|L zg1Z6; z`e-;3>w!FW(!~8%x{*praw5qI|0PB$1mQR)=C?WEfkgfzjWf*Sm#@B|p+4XSLt_a@ zEDQ@L84_Rwq6gscOGf%dA};+Vhy>J$c>Hk==wB-Qf7AR$T@6QiGaHPbKg{dje;tEi zi5|XxtIa=O8;-+zpgjm4zat_66dHp(X76#_Gc@*>4616Nqp4wHDymPKJ*+4cN2(51 zjNnc9Wqlk^6{4>INXf|nU`ZL^cuqC22z8vB8;P|C{c-sqXi}XBIJ^(&f6uoo7UzZa z`+q1=XetInh zK>oLw#N;p4e}+LwLX5Oc!lac&;IJ4U0D(l&gA8y4(r5_%|Jk^I<A-2 ztn0sxVQy%Q&;M=yFEb0|F>4_M9NrCv`Iipup^5fJA`H<4xbyE4|Caxbq>6RIAOTUZ zl$f}rjKpt^+3}`fNFz-;OX%anM$(J_TSuB!xGNUv;Q>fW{Zb)Gb@>;0QZtUJ10Z#M zh@OSz<^MfNzic(Ia2x`Se{}*Rq-6jY9uM=O7blH}gtRo^2PVxt0_pv0egKdd7Dpht z0PaMBKY+sF>5qFX3kE=1$D-dz3IOT;LDB$7?+=m%K+yjpISBw{{0EWzO#h1{NX-X1 zlCJZAB~m0koU}oIB*%3Dq5cL*xzK-uG62ZsZ;+JR^-p}T900<=f7~1qe~(AX>GlVL z#Q_lZZ;+JD{Z9Z1NdSa*#{J<(itmAec{u;^kRS!{{u?CK$>$gRXR8d4uh_4(6#vt^ z|D5+<*o1(`xgssl2-1h!-%LpT#G}2f#YsB_Cdo-ZfBk3sj|2jL&cWZj)zom_exhJN zR6?56Y%r;bveM%IfB$fV|30(7mYFnRf8paB2OyE&NI3n-M;u)7woBZbySV`x8NGSb z0`g+Z`E1ITx=~akiM>@^9FWv$KBP)!@Y|qgf|)pdZ3Wwa*kEk7rNC`g%;7{u%&VSF zgweGxumETP7n6o+M~bT^ZkihYJ^2)8%p)}$ckvfI zBFHZ~7j00n5omxj=u}u=Lf(r|+L-v;l->?kJ}Q3F=NNV9^X7d$UNYE|G9XpJgY!Q< zbs3`XUjB2()f*Hk0{N(NQd3Jc!o5rij}x66t%}qne+o_F;>!3P9qgUdw$K76#b0pj zT-E#nQFQVjyYCHa8Qn?Jkxu$B-gicee19=h@ws*rXHT!lqhxtV^WG=jWHG^WFIGAs zF7QC4@qDX1Lr@|w$EQN|c*S11&CJqW^`Hjf7oAKA*rmau%d{18R~`HsPX|S{(-o9% zwDD{!e_mX}IcRgLlzop};lE-o-;gv3cnkHQsx1b zUF`mo5!U4>s*7J%elqE0ROE(p8)m(J!=;B5 ziix*sEH^M1KEcaVV9&)gtZ}fXo-4*E7;ib(f5x37)r6ab_xIQ9`U)Pz;97h!HCD|& z=ZAhs9gh5%PqJw8bu0Ad9PTTYIaI|jef7Jj_JHb~qDpN{iP{>F?@K(@#qh$6)<@<1 z^ay#>%y(_GP_{zg-QvwGHp%L)?KsCN3y!GVSJMn^dj|ZCcVB_-QAN$=SSQ8`2KMYv zf40+T=uRjxm?7sSOH^JN7<|$iiEcql&$bWccR%HG8MYkq5Uh05k7&_l8;_Y?h+#d+ z{LNTZ74=~DY(p6;uwBin2|@ORW( ziFEXZ#VW2`mfM$fQmxgU@7C9Rib>lSf5T4%27%PdCh2L^JH4(iR$ogmzQHH9#Ld71 z_C6ajXY9}9O@AwC?;;9uPE9~|Fw`l4j=S6Vd~0d${5vmK+0(boy6rWy#xV2m*E`uE z5*i*3Tud_O{Pxvj`Q|V?h}&)7tY1FeqvJIDPGugg(vIllqG1xG4XN42^Ts|be=m3u z+u3Yufug(k6W^V0pJV3J|D0o}YMCm-|9nF&!=>%{3;xe58e7wi=iAY~v!5(!p6}&_ zugEi`d`?&KJwYKUh%FN-GRbUzWA*9{8FrI8>~VL{g|ocb3?_SgT+z(V6Sm8HY`5f% zpCmC;UJ|OkC(1GOrmjnLxlAsLf2`K5oz(_f!5}zrZ=9y1I+y#JDC+j@AJK>C=c4D+ zE-;IeuV(7n7-#Ih)LT=)4b+0)icor9o>xjwD-X+BvxrF8{|SvZ2?mc7l^rJy_V%*1 zrY{(doG(j=rGu&EEY(W0PD_UxalfjiHG52^Pcy}eX_T-kL$G1AWT_lQe=;kdO8KGv z&UI@~Tp02vie+wJJ+dFo05_)Ph8Q(-`Mlj(k2tEZ6S=Y8O-#HVMfF8BdxX9QndLHJ zYiO;{Of1cV;g{EC!r1FV9Cua6#N?AoSno(n-+*$QaBVs)G$1!fee73$s_WXTO2nN4 z+?Rr{$t8SQ3lfm)o%5`*f3hAQ-bCQ$?D9=iIh*IFu{t&ObG)J&#SLV(<9t)pX`8R1 z0>W(-X&puk3{ZcA?eHT)ZtpF@tpO-LY?+v`I44=nnLqc9&w6L>_j7CaVG$xm4pVkeGf5%-&d8nTLQIGP3 z<|Z?H;iLSbr{Y&A=XlO=p3i=?rVy`qyQR>vqos8(50R|Y)_0G5(d{J}oB9BFQf7~r zH(mH){1i{AQoztzrtf!6ox8lqux-sB_`6ixnx6Q7uD0yN`$(v!+lBYj+_w3KTndX} zxYOU06xkN|N*YVie`y#H+&+$X+`o#^+w710DC{!tx~xYA$rrvY87P*^(tFbdU$}LO z`~KsDw@oY+Lz#s`?Liai6xgf-A7MXP)kWWVFjBa6{LRb&vFi@>N1EeugM@2HGQ07) z&zJJ-dTH{E)4^wthC0HRc%QOjYpwkq_ha9g6jNkY<@b0uf4*+`DKe(9wJCJSt!<^m z#}(;s-!K+RWwdf5&4%xpE`%H1sFZ&9^L`so@H1o`?sj@q$I{T<@7jzelzFoh1V{Fh z8qb|*v*2%-b=<|AtWFCsFmU>6B4=-n9guz696fAI())Pfa4Y?}`K43ux+G~o`UdQt z0}Fg>_FllFf90#Q+eB?lH}2#wUvX3Cb&-4YcD|0OUoYTundo3E$JfN7Lx#s!KbP2q zF)e?asj$7pp%Yw^W(WP@uJ0a37smrDTr5MI$)sjz3pyP>jA+bhuLE9i=**b+-r0x( z!K)415IG*-R;HEK&t9HF4kW&aIN9U%5v*FYk~(+Pf0ui@2lqskh*F)Iqm#Yu%1p(u zn3P|}#-eI@hMK|}Co(FxQvSFT0%_WpTI?DD3U{X8N3uMcd;+Q-7^+>rfij>qwJFwF zN3mG;aNds=xK~2naBej+r(w+D`m`kXAYiRNL+N)BR;taqaA_b$a1?)e;)Ay4j~`ad zaTegrfBY+TfgaoLxyh0(YfOuuU3{+#xC=E@6FLUDk8IvPWEgt~eXD7hANP{2XI}-i z|D(4U-FoAa`cf3s>moE@zJRRiVY;4e<-N-eZV#7lM7bW;+JyV;I)M%68M%0k-@Mb+ z-Szz`(z418JjwUZghLva`|mC*NVu^?cy^D{SyI) zA2Ke3n^vKLx$N^-xre7Hj@~&5#28%G{iGQ5CIMcceaAoB-0-w*tWNXLxI%Hu3&q_0 z{3T34Bj6|{qb?<(yfByVR9QoIQcbo_5;?6#lRw7!BU#P?OXxxao z{X1}2^Mg2ZYwdX%JXr7oDAIal@h1gV#fXfdC9QY!@rGO}muvxk-s)XM`J8Eimg|xI z2;+ic<)i2m#q%YNMN0#936^ogZUlO{8KPMm&QPuPeGe|7n%mCf_QK4Xr|f4*sE zW+QS?TI71&wIeT+wD+v>qV>Uce2x4}fe!;!J1YASs_Z7AbxKbdY|HR!4b;;8XVOnh z$WsS;9DY$%mnyGRL>b?=_r|wTi6-#gceW<$^}5b5s?wWMgS8u2Y4X*o?cV3ii$c)f zveMBKunlI6SmnR-&@)vy7J@60e`jPZ3Q3&|9-U$8dpR=3Q1EijgM0!o)1KD))E+Ld z`B9~;cg|}t#%9)f-5=}nqqik6AoL9_8m&&LI1GCe_S<6k&w2s?hSef5Im%D$o3N zh0l8Yuzj$ycuH2DT(?W0MAXD@w5w3jGOe~P%X9bPnJCll9LxDIbN}Z(%?8YIamhEn zxx_#Epspv>d59oGu6&1SXa^2my?O%^_^leZX7j1@;~iQJlMJ&CzIa=Ivo4E(wwSrc zQ6b$>H9R@e?;VTHz&+~0f7288nk@DC@0@RWD_$}@?@-t7j8)(xR&knIv2kxWm8{ws zkMF5GX@pO8DSgp$Ww`%?CTi^PsY2mZ1z_BNQ9lR&tW~ZvnxE{8auc0h7UxLt`$#IG zR4r;PtA%eW0xyR6>{Ba;8I{LN!d#6Q!nPF8NlV={DbL+&dsbCgf4bK^zTj~wnH+g85l)80>d#NZuB`f6qiIzE@k7f5b;uJ%sr}!Tb&gEx98s~Ce&Ic0=S9o4b z2BzIjK)t-!7TrsC1$5PY#g=_RuC`Roe1}OY!-^;zI{F24UCuXZr)(B4Ld6homiSH2 zVum$J$TTA@>1;{xe<&`0bGPE4lz@LWaL!))Oi+eUb@#;+Ps-GqO24duXk9Dl%**V1$eB`4rET=`{;Xk-IKVIR7M1|Ck=ms5ZlcQ?`K`2muslL z$WHKD8wrz%X)*LQBo`?o`0)oTFJ;A2roGGStye}vxv1C;H0o2c5Ut#7(RZr#30G9p^Nk7DZx_?bRvD^&E@xc5Bz{xsN@=GO);o4+Q)MKk6{60i z%BG4*eP3ibbB2qz@!l@l5k?-@_ps{q1 zo5H5(U@qn9!lN9BeCtiS61#pEfdJJ+$w~h~F6jrk=5!hd%}!B)GnXdh)45K$udH>% zIl#%%SQ}R_(He$sxK4RFue)q8o2^n92nNvKiTv(4j8P7KvNH{WDASv@7w{c)+*9&; z9Dm?ae^Xyv8!EdYhIW%Wy#l%qaBCOf zVf}{{H{TWZi$A*z&ehvU&>nP?qZZ9I^Y!s!U(?`@b#(7MReX|<$ z4046<%iG)<|N7_!3yn{1kPc(%z#)SB>-w^Km4RN7)8hJQPyXZ@;g`-wA7*Q6Z3Wq^ zfBg;BpC?9fEPgISX&?2i#C+TpYn)57C)1`+ePWRsuET?{7ger&e!mHuJ)_nUzhaeg z{X2C4GfkwqL>PZiRZNA8uQjf3ik2`9!?L^w+enKl294UP~nn!n1E1zOOiZ ze$v5KVs8O?!T!q!&IC7l-f@lF&$WX-0xya?+x+|KV>cn2vT^VG!{++r4Z}?g6nAU{ z$h}45G@Sqge88?PSTlWoIvRh+$y?Hdo zs3q5W@XVvPEo(>}N_Rt?NS;9R+4QF;`viKdGTU@d?w%8h=YBGvmZ5^%>ZjTsG|w%_ zJ$p;UMz71wWYEM+Uz#dx(znsY=8l!f5wEnJ(zeit-rXVQ>=W-9hUU+^e>I3~mEh1> zMAU=aw9|GF#}zw`^w!Cibn!a*$9nGhp{sEs0SQ&NV+ct&yW7{xPwYxNiaAqp46Sj# z8=qf5LHdO_Gw@AKWmMl9cdv0;#g4dtKN#qb9cKu|(9AzuhI(>KFhR3mw`-nyF{(=N zCl(Xv+*ZUr9tDPd*IQ9&f0+Z)fje>(YZQJ{9S z;9@a51jw6g6XqOR+GyAL9s>JWUE*EB+E3mz=gvWD_w?|B79F-zf8X}I`RJl~3PTXd zSNfAK$HQ>m^rKey$qOem4wAx3rbT0AJUyjnjD5S8++i`V1bD%>3}fv4*N}VQnwh}N zem7JfbBf_A-*q`aa)m&2TbP5 zUUSH%ES2W;t+3f%f8D9se$Tu2I<+{H`-Z5P7psD4hJy`qYh^D*4dy|z6ioO^Ml6Dr z6dI!sLpL9Ma6e@Gw3_^oX~z?8WC65BungxNZbdt-zR$|U+)bhz^*1B4mJns$VFA2v zy4aQTXHPl%a~sWXj1mId+c$1p=p-^0l%L)#ts5B;JGA^%eT3~+dbTuW_kjByO~q{Mm%ibi!itG}NsZd(CR;a?Zgg~Qr%8kK&NMP; zlv+pNJr^1YSKnx-e%UXzVJZhY3DKY^e%7bdwaFFALeRrD5$7gNzPuG_E=SeNs7>LB z>3KP~`uaNDfBoAQn$+(fU42{BGZmJ}>utgEqT=Y8%?F<%k!YcZmt0zbO7rq&qJ0%y z)}<&~=@Xtc7Lv1S5BGgf*6~56GrT1C@I59Q`C?+XFT1lX=nFA!hI7x8w~m{m>P zJQ$%ITsko2mvKed^UQ~zw??mRAHnY?tz9?$f0#$rN?Vrv*yk=8FQ}F?8`rO%9U_cn zhvliinfqEjU)4Y(n2v9(>(8_ZGVmHoY%1T6x-PBFyR3mr;jv^(eGeA6dI!Q^m=RVM zVxW2_+?LbrTwtYlXMip%j=iUwcgc@(OfrOF@4T^F^ShPu{p@?Ht#r>EY)y9QU;5k| zf8KcgB>1ddu|4%pxxE|Rk(mVS)-Ij?>?yhw=?4A6^;2v2hOGo|dOK(6y7rVQqf#A( zHw`H7fY}E)ZT4%{Mw>3`*cq`ajm!#fRU(c^7n{hc!l@HyBJ5=Bi;i}IcQm5RX zQJ`g7O+~TzVXKCbUTpIg$(eR7qeUs(=5!=NAeU{tvUeTvz-F5)`e_9UW z@I~Pzbtm>o!a(oO6<%${>e|+CA?yhfB|7b5n)-I{>SQ0t>DAE}Rtrh%#J0a{@OK`l zVZClY(0eiA(E2&OLCEQ#N1`)K{r6^)4X@zON4$!T6v=$8c*>i8Of-*?V+kL}(zT!? zn)q#7arNfyv+n6;_wU#)4zgc3f3Tpw4g!z6;FZNpGfu?{R_1h^HB_9oVOi7LGwe4u zucZn|S0gXc+lsniG$`19qUClx8*hk}p|7>%X*=4L+|QY_KWP%4n>2qc`stVS1hu$* zFfcMUmwSZ>!U8urmjQ4HD1WV6+j84D zwtd%E;IVzCViI?tYHF$yr{nlIjyth)>#jT~Hf`&SEV(7QCz-E5YXhJJlA>gFR29=8 z0AjDbux}t8p$#jU(2gl3@Y^$w9$`4+%45Jt#w}N9XvRHvXjmph>CwQCwivRU3Fjml zl1ZV=Kv>PBGMH>FlUC7?XMai(B=o z!Wn8o1{Ot&Lx$uA&wp~rkW!RrL59Req6Hb!(AI-pE**?Xl4hy|4U$MzLim9swUy#D z3K>dclO?gCtc47$2@|Aeu@_KHk#RwWrfro(kfu$B!O^w~>VpieLkluA3}zs~C0ZVA zU@q5|+rb}7nM1|m6=S$M0xKk1(y2qA+Hve>u4&EeN|Lm6hkx~z?wyi{5_ahh4UAGB zYUWIN2qQKYSgf>Y!@tmuGZh!~yrZ^*FW8ba;ia&LE={3jalP&UM%qvO4Z;a*CzJR!n!13tCrO2P-NrjsnBc664%({vA&MaE$7DZv`bommC*h= z_`_f}NujIP4u4qVRa3x$Un$42+Nv+MoL3~9dyFY%!`gJE8TptDjK(&NJ@qF&9j2BABm2v=K*Lzn5CS^-a-QGfT;ylc)OEVTvMrqqOuo{qhV zkI-vMo9^RYr014%XUC@p0 zRk&B#jMbQ2w&KXoT+@o-ns$F_Z(8!;q~wak4RR9sV4RDm5e4cjvrE)^#YVIT z91Hn(}3ALIP#v((povvm1 zb@0!B4vNbk4;3q3O(%<&)%OO353j4y?Hz8N85AUk#-4E+JXs9K zqpRnW+i`^+bX+W|`;U~U2gT>~m{PRL-EekME!Z!`=i*zn7#3A=Q#=;`DZUp!4kDaa zBY!L+P=%E2{or5!dXj#xemy_=@FwbK*Q;lfc9IIDy=Z6pmAh%`B6iW`<>}ek+mnC1 zIeS07A5M7r?8S6^ym(6&Ui_9w`%ru=KEe7Uc~{f%bb{;s{V)(&eZLwH@7M5|v-xN7=VLWrjPT^2kAKqz zxCN3x?tDC)-+|)A-K?q-(eSL|zfH2uPGrBm`0)A@$UYk`?tcBgs3vnrhF!j|ZXd_P z*?Qhlh{7@rm9cc}uwq5KDs$G$jC0gJJ)E1Ia>;Ybzs-hM|E?C}>Spn6bbDLh2U0@E z!*Etj(r~<~$BOwQh$kYiNArjA@PEguV^MuyG)Z&2kal!>a{m5JA}wmPjWH=wgeq~Q z*&LzdQzKN_%Nm*KaWK!*V4!bdo|nNahvSDkGR_d5uQJ=bfkoa9>F#z`4Pkb$&)U() z)qEc8bMhD+emGlh+ zf!`C)PWAFk1l>2a+bJ6yy?J^1>2pInyV;I-Qjc4{C-rdb?o;nL7|>ZTo@g+^TyAQ^ z7~X%o9tOjBh}$2wy|dvvMgQO5zB@m@$P(m6I+UdS_7sTd$c=M-o&cIT(mgFlZU{zw z@&9d42FQX}BXE)9jb>Mm_kTCz>U%d!geldZkHc}6G3ei@2>Eo;~D^Es~ z3Z6Q}^-nhPsGdmVXEoaT@Wa8*TEVK#UDtwsy(~UtYcb?YH+eTYd{$sw2ggxM>6XGf1(!*w2qjdUv)spHr zha;H6@Zn)L{T@Thx8W>g{c4&2Q)d14I38EX@Z;(p^6Gj!K18K8+`iRV)W(NmyhR7KDSoZUMUtn%2?0$K;{M$3fec-+bPI? zxqScf^th?swtvzZN!DvlpNRy$Puz#K^e3Wv_v1D<7ZzzLfg&xvgT^*-7 znHe4n3uff9s%A#7J31f{uaXCIqIrN|UT0-B^S1Fq?lwD}uU;JGZ{R`={`E%4q<8Whg!HxzW-4DTbD;OV5vge)eHOqh7sq&E#ryLt!Xu7+0? zaWhj~Wq(HHFD#fbv=z+q;+R>Pza$BJfa88~NeX;^0_XHXwPu|@Lr319%6f2pCYCWC z>AB$_sEAXppj}vHhCE>~q4T7}bXi9hrd2bnFk?w?7#)33U7MQd&_j8PYJ@t+JWU%_ z_PCo2wKG$$;ap6M;90bP{VwE-UlEatGn`h63xDQ$Zi&px)|q;Fq7EAeRmyzhaPQV_ z9KwBBeE-@Fun1<4-xlGPIuXTQ4CfV%toh>72*}}YOemNYuST=^f`T8U42n0yhTDr1d`VZ9T8I5PH7J0y)p3p6Wzo z`8P~(+X1+ZPj13jX8Ba_1KqasYa5;3gifoBJh>P0wwyED@H)|%{&r-wyE3!Bav%79 zCf%xr;fZW(RTrDv9b5m0jW)DSP^R|@t6j~F&hG^?y-lBY%^d zx09O;nt2Me--fr_*nFHI?u%4BCgM6`L_k|cnG174DomTiyg*olI4<&SY~aKA$w&c+=53n zL?mm|T9lRI!y;@eb4g{D=#CUDmw%vlV3_>N%jwm#iv`a72Vf`6{N-d(-5>G;CL8zV z0gIJS-GLeL^F-Yt(4q4Vc|d>k9>o@KM)#vdOX6}~2UZ+Vv_J!erXHF180u-@u+-B4 z;;6^uJ)YFnZD>oW%USDzK~tH75G5^_3lN-oY~CZN$K^f5k8Z3(3hVW-#DBqRQV%L3 zM>kLqIl6&@$k7cHM2>0n5IK##2GE#Ux~TNQa?W~GlDLU~$WqTFZep2Q_YgNW-)KqP z*qj?l-1x>E(l^!`9b6=JIjaX4k>hj4C6VJBih~?cHo$x8$-x7cYTTDJL#&5o4py_) z!&00c&ox(IhyD~b>ygxx?|-XO)RQZtH1*`lC#QeGHdf5vNG4eMK0(XD98nU{uJ4OU5`p1EPv;$M@8XD@LX>S zPeSDDg9w$g2O4y)2ZWd$Gt{zbEgq6q&smQL1h`xl)b}|$C|~P7ude8&?Ef4FuW0u{ zlJ@OR%rxtfkx}LII_Oq|pxfy7%HYU$Fn}oAnF+S6X?KFMR2yWS#6>4oG;lN~Z-5LI zYpYsyf=bs0LZ;-}eSe_RwK*p%@$Drkk9pUR*c~({_amnK=iA#CQefF;Iixn;CNHG7 z+OAi<_4<887-4&?2933*U#fb9>O zv-@F(1S;FSgd{3eyO)@i9TkaL)!rM*kzBR+hVmp={gzE=8h^91ynkEF7g{E}ec7yKxoK^mR=((OY*}i!TYej7=Z>@67YgX!mXYEzl=U|V z`FV2NTZyD?kKZILj=p8G;YsoDz~=pZ&l_y-`goqgvgIntqrMGe@=$* zpI*1N#BXwCqHOr|lR7CzEWhH3s)U2{1@OmkG(MGXVd5Z zjI3tcenXaa7o{fVf>_f}OyfunvNSa@H#4+!G`BD}G_f>taW!-{bTfB0HZwADGqW^y zwNtPms3aD)HeA8n!b|}&xS@LGe6Vk#%<+%Uf8Qy7XFD}7a8ZYg;HmX4_8OY5`vX`0 z@seJx)0-HPlG=8*yoc91`CyN)vDvXh5kJM(==98$G!`@CxhFsSdd>WK&*wxx-?Lk> z<%4?LgUJT0>HjCPo&GzI-zRP154T{(Ant`ur#W^?oZ04JyT#FX%OSHHr*^i~uVgGc zY*!b?SjpY$zbl?iHQ?05xjv~I*iJ1x;kNkb$M#)|#I$tJ?YUm0yQW!dMcdbyjm<0D zGOf2b2){{8X3lBR4M;C6^!a}K$i%%qvlq&h9SrL}z1Hdy!|erv+nVjxa?WzBch}w$ zvBz}5>9S_KnCL|2o;Y*enKxi^@Z{j- z*BkTvRpatK-@3-SWuM#Mc6q|&F7v2Me}Z-tm!lBv?@`PCm@*Bai&J_ z#PR7CGCI5K?F=Mcyw{!2n7R4M!C>D`RedXyrPHSQXV;&KuKK=hS<_PfeVM0pYHzPy zdE`=mzUkDcUtgCmGdg|$$I{H+r}v+jPL2Ned(*N-qHg;#R&6+x)xFH?bXK^@+O0W8 z8ByssFa5u(b$VL&tW~