From 226c881403532f777348fab9ddcb7bc3ad2e0667 Mon Sep 17 00:00:00 2001
From: minoue-xx <790354+minoue-xx@users.noreply.github.com>
Date: Thu, 16 Apr 2020 14:37:52 +0900
Subject: [PATCH] Added English version of part3
---
.../script_AnimateVelocityField_part2.m | 6 +-
.../vanilaCavityFlowHelmholtzSolver_EN.md | 537 ++++++++++++++++++
.../vanilaCavityFlowHelmholtzSolver_EN.mlx | Bin 0 -> 454200 bytes
.../figure_0.png | Bin 0 -> 38833 bytes
.../figure_1.png | Bin 0 -> 26432 bytes
.../figure_2.png | Bin 0 -> 38833 bytes
.../figure_3.png | Bin 0 -> 43493 bytes
.../figure_4.png | Bin 0 -> 51274 bytes
.../figure_5.png | Bin 0 -> 138936 bytes
.../image_0.png | Bin 0 -> 140158 bytes
.../image_1.png | Bin 0 -> 38227 bytes
functions/getIntermediateU_xyCNAB.m | 2 +-
functions/getIntermediateU_xyRK3.m | 6 +-
functions/getIntermediateV_xyCNAB.m | 2 +-
functions/getIntermediateV_xyRK3.m | 6 +-
functions/updateVelocityField_CNAB_bctop.m | 95 ++++
functions/updateVelocityField_Euler_bctop.m | 97 ++++
functions/updateVelocityField_RK3_bctop.m | 117 ++++
18 files changed, 858 insertions(+), 10 deletions(-)
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN.md
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN.mlx
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/figure_0.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/figure_1.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/figure_2.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/figure_3.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/figure_4.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/figure_5.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/image_0.png
create mode 100644 docs_part3/vanilaCavityFlowHelmholtzSolver_EN_images/image_1.png
create mode 100644 functions/updateVelocityField_CNAB_bctop.m
create mode 100644 functions/updateVelocityField_Euler_bctop.m
create mode 100644 functions/updateVelocityField_RK3_bctop.m
diff --git a/docs_part2/script_AnimateVelocityField_part2.m b/docs_part2/script_AnimateVelocityField_part2.m
index b0fb401..35e3b3d 100644
--- a/docs_part2/script_AnimateVelocityField_part2.m
+++ b/docs_part2/script_AnimateVelocityField_part2.m
@@ -4,6 +4,8 @@
% represents the velocity using |quiver| function. The code below generates GIF.
% The numerical integration process discussed above is now in the function |updateVelocityField.m.|
%
+clear, close all
+addpath('../functions/')
%% Animation Setting
visRate = 4; % downsample rate of the data for quiver
recordGIF = false; % set to true if you wanna make GIF
@@ -80,8 +82,8 @@
end
% Update the velocity field (uses dct)
- [u,v] = updateVelocityField_CNAB_bctop(u,v,Nx,Ny,dx,dy,Re,dt,bctop,'dct');
-% [u,v] = updateVelocityField_RK3_bctop(u,v,Nx,Ny,dx,dy,Re,dt,bctop,'dct');
+% [u,v] = updateVelocityField_CNAB_bctop(u,v,Nx,Ny,dx,dy,Re,dt,bctop,'dct');
+ [u,v] = updateVelocityField_RK3_bctop(u,v,Nx,Ny,dx,dy,Re,dt,bctop,'dct');
% Update the plot at every recordRate steps
if mod(ii,recordRate) == 0
diff --git a/docs_part3/vanilaCavityFlowHelmholtzSolver_EN.md b/docs_part3/vanilaCavityFlowHelmholtzSolver_EN.md
new file mode 100644
index 0000000..8b89f98
--- /dev/null
+++ b/docs_part3/vanilaCavityFlowHelmholtzSolver_EN.md
@@ -0,0 +1,537 @@
+# CFD 101: Performance Comparison of the Implicit Methods
+
+
+Copyright (c) 2020, The MathWorks, Inc.
+
+
+# Introduction
+
+
+This document is part 3 of the CFD 101 series.
+
+
+
+
+In part 2, it was observed that using the implicit method requires solving the discretized Helmholtz equation every time step for each component of the velocity, namely,
+
+
+
+
+
+
+
+When we use the direct method with matrix factorization (observed to be the fastest in this document), the process of calculating the velocity field accounts for about 40% of the total calculation, so this is worth paying attention to how we solve the equations.
+
+
+
+
+
+
+
+## Environment
+
+
+MATLAB R2020a
+
+
+# Goal of this document
+
+
+We compare the performance of the following five methods to solve the equation mentioned above.
+
+
+
+ - Method 1: Direct method
+ - Method 2: Matrix factorization + direct method
+ - Method 3: Iterative method
+ - Method 4: Preconditioned iterative method
+ - Method 5: Discrete sine transformation
+
+
+
+In part 3, we have already introduced the direct method, the direct method with matrix decomposition, and the iterative method to solve the above equation. Here we briefly describe the preconditioning for the iterative method and how we use the discrete sine transform in solving the equation.
+
+
+## Bottom Line on Top
+
+
+This is how it turned out.
+
+
+
+ - The discrete sine transformation is an order of magnitude faster than the direct method.
+ - The direct method isn't so bad if you can pre-decompose the matrix.
+ - Even using the preconditions, the iterative method... didn't perform well.
+
+
+
+As long as the time step size and Reynolds number are kept constant, the decomposed matrix can be reused, and it performs as fast as the discrete sine transformation. Note that it expects to use more memory to store the decomposed matrices, which is observed to use roughly seven times as much memory as the original matrix. Of course, it depends on the finite difference scheme to be used.
+
+
+
+
+Please be noted that the iterative method is calculated with a fixed required accuracy. It will converge faster with the loosened the condition, and it also depends on the Reynolds number and the time step size.
+
+
+
+
+
+
+
+
+
+Note: The number of grids here is that for 1-direction. The actual size of the matrix being solved is (number of grid^2) x (number of grid^2) for a 2D case.
+
+
+# Solver details
+
+
+We describe the preconditioning of the iterative method and the use of discrete sine transform. The other three methods, namely,
+
+
+
+ - Method 1: Direct method
+ - Method 2: Matrix factorization + direct method
+ - Method 3: Iterative method
+
+
+
+are omitted because they were introduced in the previous article, "Implicit Solving of Diffusion Terms."
+
+
+## Equations
+
+
+Let us revisit the discretized Helmholtz equation. Here, only the velocity u in the x-axis is considered.
+
+
+
+
+
+
+
+First, let's look at the matrix 
to be solved.
+
+
+
+```matlab:Code
+addpath('../functions/')
+clear
+rng(0)
+N = 10; % test with small matrix
+
+% Computational domain setting
+Lx = 1; Ly = 1;
+nx = N; ny = N;
+dx = Lx/nx; dy = Ly/ny;
+
+% RHS is random numbers
+rhs = rand(nx-1,ny);
+
+% Matrix for divergence term
+maskU = false(nx+1,ny+2);
+maskU(2:end-1,2:end-1) = true;
+L4u = getL4u(nx,ny,dx,dy,maskU);
+
+% Setting for time-step size and the Reynolds number
+dt = 1; Re = 10;
+A4u = speye(size(L4u))-dt/(2*Re)*L4u; % A matrix for u
+spy(A4u)
+```
+
+
+
+
+
+```matlab:Code
+issymmetric(A4u)
+```
+
+
+```text:Output
+ans =
+ 1
+
+```
+
+
+
+It's a symmetric matrix.
+
+
+# Method 4: Preconditioned iterative method
+
+
+There are symmetric matrices, but here we try CGS (conjugate gradient squared), which requires neither symmetric nor positive definite matrices. The required accuracy is set to 
and the maximum number of iterations to 10,000.
+
+
+
+```matlab:Code
+tol = 1e-6;
+maxit = 1e4;
+[u,flag0,relres0,iter0,resvec0] = cgs(A4u,rhs(:),tol,maxit);
+```
+
+
+
+Looking at the errors and the number of iterations
+
+
+
+```matlab:Code
+relres0 % error
+```
+
+
+```text:Output
+relres0 = 4.9261e-07
+```
+
+
+```matlab:Code
+iter0 % number of iterations
+```
+
+
+```text:Output
+iter0 = 15
+```
+
+
+
+It appears to have converged after 15 iterations.
+
+
+## With Preconditioning
+
+
+Precondition matrices increase convergence speed.
+
+
+
+
+The points are:
+
+
+
+ - The convergence of the iterative method depends on the eigenvalue distribution of the coefficient matrix
+ - The smaller the eigenvalue distribution and the closer to 1, the faster the convergence.
+ - The eigenvalue distribution will be improved by applying a preprocessing matrix.
+
+
+
+and here we solve for 
with precondition matrix 
. Examples can be found in MATLAB's official documentation of, say the cgs function. Commonly used preprocessing methods include the following:
+
+
+
+ - Incomplete LU Factorization
+ - Incomplete Cholesky Factorization
+
+
+
+Using an incomplete inverse matrix allows decrease the number of non-zero elements. With the full LU decomposition, displaying the non-zero elements together with the L/U,
+
+
+
+```matlab:Code
+[L,U] = lu(A4u);
+spy(L+U)
+```
+
+
+
+
+
+
+shows that there are 1564 non-zero elements, and the incomplete LU factorization gives 412.
+
+
+
+```matlab:Code
+[L,U] = ilu(A4u); % 'nofill'
+spy(L+U)
+```
+
+
+
+
+
+
+Compared to L and U by full LU decomposition, the number of non-zero elements is small. Let us solve the equation using the L and U as a preconditioning matrices,
+
+
+
+```matlab:Code
+[u,flag1,relres1,iter1,resvec1] = cgs(A4u,rhs(:),tol,maxit,L,U);
+```
+
+
+
+Looking at the errors and the number of iterations
+
+
+
+```matlab:Code
+relres1
+```
+
+
+```text:Output
+relres1 = 5.1845e-07
+```
+
+
+```matlab:Code
+iter1
+```
+
+
+```text:Output
+iter1 = 5
+```
+
+
+
+It seems to have converged after five iterations. Plotting the number of iterations and the remaining error at that time, it is easy to see the effect of the preconditioning.
+
+
+
+```matlab:Code
+figure
+semilogy(0:length(resvec0)-1,resvec0/norm(rhs),'-o')
+hold on
+semilogy(0:length(resvec1)-1,resvec1/norm(rhs),'-o')
+hold off
+yline(tol,'r--');
+legend('No preconditioner','ILU preconditioner','Tolerance','Location','East')
+xlabel('Iteration number')
+ylabel('Relative residual')
+```
+
+
+
+
+
+
+In fact, the A4u matrix depends on dt (time step size) and Re (Reynolds number). For example, the larger the Reynolds number, the smaller the value of the non-diagonal component of (as it approaches the diagonal matrix) and the convergence tends to be faster. The point is that the convergence speed varies depending on the conditions, so it's not easy to compare speeds.
+
+
+## Eigenvalue distribution
+
+
+If we look at the distribution of eigenvalues for the Reynolds numbers 10 and 100, we can see that the larger the Reynolds number, the smaller the eigenvalues are.
+
+
+
+```matlab:Code
+Re10 = 1e1; % Reynolds number = 10
+A4u1e1 = speye(size(L4u))-dt/(2*Re10)*L4u; % A matrix for u
+Re100 = 1e2; % Reynolds number 100
+A4u1e2 = speye(size(L4u))-dt/(2*Re100)*L4u; % A matrix for u
+figure
+plot(eig(A4u1e1),'or')
+hold on
+plot(eig(A4u1e2),'ob')
+hold off
+legend(["Re=10","Re=100"])
+ylabel('Engenvalues')
+title('Distribution of eigenvalues at Re = 10 vs Re = 100')
+```
+
+
+
+
+# Method 5: Discrete sine transformation
+
+
+The velocity in the x-axis direction `u` has the following boundary conditions:
+
+
+
+ - Dirichet boundary condition (x-axis direction) in which the variable is defined on the cell edge.
+ - Dirichet boundary condition (y-axis direction) in which the variable is defined in the center of the cell.
+
+
+
+As described in the earlier article "Solving the Poisson equation fast in various situations" (in Japanese only), we can apply DST-I on the x-axis and DST-II on the y-axis to solve the problem in wavenumber space. Note that the inverse of DST-I is DST-I itself, and that of DST-II is DST-III.
+
+
+
+```matlab:Code
+kx = [1:nx-1]';
+ax = pi*kx/(nx);
+mwx = 2*(cos(ax)-1)/dx^2;% DST-I
+
+ky = [1:ny]';
+ay = pi*ky/(ny);
+mwy = 2*(cos(ay)-1)/dy^2; % DST-II
+
+mw = mwx+mwy'; % Modified Wavenumber
+
+% mydst* applies DST to each column
+% dst1 in column and transpose and dst2 then transpose back
+rhshat = mydst2(mydst1(rhs).').';
+
+% A4u = speye(size(L4u))-dt/(2*Re)*L4u; % A matrix for u
+uhat = rhshat./(1-dt/(2*Re)*mw);
+u = mydst1(mydst3(uhat.').');
+
+norm(rhs(:)-A4u*u(:))
+```
+
+
+```text:Output
+ans = 9.2596e-15
+```
+
+
+
+The equation is indeed solved.
+
+
+# Comparison of Processing Speed
+
+
+The processing speed of methods 1 to 5 is measured. To get closer to the actual situation, we compare with a time step size of 0.01 and a Reynolds number of 100.
+
+
+
+```matlab:Code
+ntests = 10; % Number of test cases
+dt = 0.01; Re = 100;
+
+pw = 1:ntests;
+times = zeros(ntests,5);
+for ii=2:ntests
+ N = 2^ii; % N = 4,8,16,...,512,1024
+
+ % Computational domain setting
+ Lx = 1; Ly = 1;
+ nx = N; ny = N;
+ dx = Lx/nx; dy = Ly/ny;
+
+ % Viscous term operator L
+ maskU = false(nx+1,ny+2);
+ maskU(2:end-1,2:end-1) = true;
+ L4u = getL4u(nx,ny,dx,dy,maskU);
+
+ % (1-dt/2Re*L)
+ A4u = speye(size(L4u))-dt/(2*Re)*L4u; % A matrix for u
+ rhs = rand(nx-1,ny); % Random numbers for RHS
+
+ % 1. Direct Method
+ f_direct = @() A4u\rhs(:);
+
+ % 2. Direct Method + Decompose
+ dA4u = decomposition(A4u);
+ f_decompose = @() dA4u\rhs(:);
+
+ % 3. Iterative (cgs)
+ f_cgs = @() getIntermediateU_cgs(A4u,rhs);
+
+ [L,U] = ilu(A4u);
+ % 4. Preconditioned Iterative (cgs)
+ f_precondcgs = @() getIntermediateU_precondcgs(A4u,rhs,L,U);
+
+ % 5. Discrete Sine Transform
+ kx = [1:nx-1]';
+ ax = pi*kx/(nx);
+ mwx = 2*(cos(ax)-1)/dx^2;% DST-I
+
+ ky = [1:ny]';
+ ay = pi*ky/(ny);
+ mwy = 2*(cos(ay)-1)/dy^2; % DST-II
+
+ % A4u in frequency domain
+ A4uhat = (1-dt/(2*Re)*(mwx+mwy'));
+ f_dst = @() getIntermediateU_DST(A4uhat,rhs);
+
+ % timeit to measure the process time
+ times(ii,1) = timeit(f_direct,1);
+ times(ii,2) = timeit(f_decompose,1);
+ times(ii,3) = timeit(f_cgs,1);
+ times(ii,4) = timeit(f_precondcgs,1);
+ times(ii,5) = timeit(f_dst,1);
+end
+
+%% Visualize the results
+figure(1)
+loglog(2.^(1:ntests),times,'Marker','o','LineWidth',4);
+
+% Some cosmetics
+handle_axes = gca;
+handle_axes.FontSize = 14;
+box(handle_axes,'on');
+grid(handle_axes,'on');
+
+legend({'Direct Method','Decompose + Direct Method', ...
+ 'Iterative Method','Preconditioned Iterative Method', ...
+ 'Discrete Sine Transform'},'Location','best');
+grid on
+xlabel('Number of grids')
+ylabel('Processing time (s) ')
+title('Processing time vs Grid number')
+hold off
+```
+
+
+
+
+
+# Summary
+
+
+Solving the Helmholtz equation for the velocity is inevitable when using the implicit solution method. We compared the processing time required for each of the following five methods.
+
+
+
+ - Method 1: Direct method
+ - Method 2: Matrix factorization + direct method
+ - Method 3: Iterative method
+ - Method 4: Preconditioned iterative method
+ - Method 5: Discrete sine transformation
+
+
+
+And here are how it turned out.
+
+
+
+ - The discrete sine transformation is an order of magnitude faster than the direct method.
+ - The direct method isn't so bad if you can pre-decompose the matrix.
+ - Even using the preconditions, the iterative method... didn't perform well.
+
+
+
+If you can use discrete sine transformation (even grid spacing), this seems to be the best in performance and easy to implement. The next option is the matrix decomposition + direct method, which can also be used for the complex geometry and uneven grid.
+
+
+
+
+Please note that the discussion above is only valid for the serial processing cases. When the parallel computation is involved, we need to reevaluate them.
+
+
+# Appendix: Helper functions
+
+```matlab:Code
+function u = getIntermediateU_cgs(A4u,rhs)
+
+[u,~] = cgs(A4u,rhs(:),1e-6,1e4);
+
+end
+
+function u = getIntermediateU_precondcgs(A4u,rhs,L,U)
+
+[u,~] = cgs(A4u,rhs(:),1e-6,1e4,L,U);
+
+end
+
+function u = getIntermediateU_DST(A4uhat,rhs)
+
+rhshat = mydst2(mydst1(rhs).').';
+uhat = rhshat./A4uhat;
+u = mydst1(mydst3(uhat.').');
+
+end
+```
+
diff --git a/docs_part3/vanilaCavityFlowHelmholtzSolver_EN.mlx b/docs_part3/vanilaCavityFlowHelmholtzSolver_EN.mlx
new file mode 100644
index 0000000000000000000000000000000000000000..60c538541f1881060134898659ff68b84576f094
GIT binary patch
literal 454200
zcmZ6xQ*bU^&~6(WE4FRhwr#E0wr$(CZ6_)e*6`b1}7ZF;Mk%Fm=|W_pr5TRhf_<
zX8iMwHl>l$;UEabb`vOe2!cjUz{t5YQpK^6Oz$6`o3h!d!66)TqXvM>Y^Clr=VpAl0td-4OF$a?mOpJ>2Bh#1i=^
z#9BfvwOA%|Yrl1_#0yB*oX5j8;dor}R$ZzkgE-tNE5t7KnO#=YJgm#`Y%y
zE(gz{|M6#is`17qpjzv<{O*7*43bmvhkb&@DkK8Y?I6sYi@3KO>02PD2Uzj{gIO%+O#!K(PN|HgGbvab}?ZpR+c3Pkx9IHQb=|g71nwp`y^CbmBH<5R$Vk+}6lC
zyA8#|TZXj*h9uczBPZ|f^?6RK=EYWof4)nbM!uP}6fAYVNu8|R{_1Rkk8CzLY2G3<
z2Y+6vNt<535tzVBR}Du3M3hUO=5B_>L|%10NTx*yjT*;R2Vi6=V!!|Q+fI+vWxD5kpg6v5gMF6g
zk)p@`Zw}7oOzzeG-JqrL$&vw&PZBV&nLCeeVAEl6W$jWd=VvD{<*zj|LX^D$m0jK|
z;rb=tZZT75l{+uu=T7n^Ob|3IZ2LqPhNuDW4ImnS2OnR!doNEkM)Ef%5
z9Yj5g0Tr{{sAKpg7g(o^%MWnyE29go(CS7r$L06`m!5N`C<*+(MwQY3-~IakrvJZl-{^Wf
zZ*eC58svkNSEW>hq=}AB$Zs5NJf1Qg(NLr}PV{K7hA=Kb9VT~2&ZR&Fjl*xe|Pk~cW#ve>_D$tvSY83XJj
zwM~EkG4~0bUKL9!!A%Wt7WdyReR9k~Ycx#v#^B|*FlUYy>yBsm!B0apVVVVK^;e1d
zHyP;n_)P=}HH_~@rgRkw59*))9aFJvG4`5nsx*|CpiL{BUdKt4bN!p3D1NLg^AyG(
z>FM@Kf1#B)Il*e8ziv7_=&;|+az2ccCns?JwEcE18QRO^W<<%x^3wJKFhTBvGcC)`Rpd&%>qKdYLz8Tu9$$~vp1ym_E~Np@
zBdt^soyT=PY+=?|U$Oh9>VOraF0r&L?qrWYcI+C!<9oc<&ua!5(^tON@1wG}QU0y)
z2_TePKT$0y@)As)Y?9d9K=@>p_k5$9?DNOT6Xf!8{PN9IP@}ljbK{o{HBJZlE7~!k
zAf`DnUOHU{$Er(xW-~o3TmDd;3Sv)WoE2-c!vq<~*np1uRYMn7r%<2kBzghmiF?YV
z=tQg`4TB?tBoPx{IICtR+4d%RhP8#qOFc@xgcXlL+VUXeV;c=1SY~|k9D^-^XMw*l
z@h?nQ@r@)AQ)>UFGyN-4?Odum8KN)}E6jyfAx`6ni|}KdZH?NMqQqW1$@@8V8^v7N
zK+CMc#S>LEbfJV=lXTZwlosM0?ntgY%5pDUZDWL!3}7E)I=JxU-x5Lt{>GZXB+rKX
zn{CmbVFi79HBwm|{4(M2=Zn>9?=NZ?$Bb}4vHVTXsUz>%!OjUzTVubKW?&1h^^{L?@F(HF9>8+lw${vI+5k|li)jVNdHE+J@Kzb`<~JEA$15+?ingDmsVGQ)pOk*IxBn&@#H&M6
zTR)gu+Q^F`rzYZ6s?5pfP$J|iR4*O?G3E#?8Yr>ICEvVBn+yTkZf=UQ-sWtAj<6+d
zR*a(Vep#JYlN|sF`+J1`W6uqmABlGa7x2Tg)Eo<{@D?8gq)0!dq6)uHYM+Y5qr;HB
zDm?(zVay#X8pV$LWlYx0>%<6x0i>a!af%+-SjSPKlnZz8FJd^JiN*y%QtSX`kLo!>
zuU);L^y62LV2s&6vkz!4NHi@u8X!
z-jf)n)WqS(5I~YZ`h-0mb_!vcbH@D(pxkg#|nkT1g5qT66BS&b;}o(G&d
ztgp@D=a7)^-&t|O*VWGVv)ih0~J;7y~cHRL8mVl=*pkccPYN$s3
z0mdGcfCK(xi1clIv7DX10a}?Q$mEd04wST{Qs;W$JnHsK%A?pohEAFUB_g;*F49J%
zibP4~5+G-6k^We28$K%{QF@IBG6{;#=D-~^gaV_trGzG$HnUj!r1ebpwr(rai2
z1&2Y*)+yz=aPZ&@z>;F%$NrpQUjZw>H;o^;$2hs*x&PqcCnZC+4{AVgah8c-ajDd(
zele8K3_X`*!kjp)_(^Kv;VQA1O3QGO)v@Mv4Sh#
z`FD-ZV}}0qxz7|(zq)?&PlhP#4$ukidR`W+9yHA7Gj4I0?CX#+9OWr%(~hUCC7O${-7gly
zPNoKW{@dZd&YedR>g%REZ!U9i*ZwJs_OuZNc)jnbKmDW~G9MQ+K>L(@J#JUgb>8;@
zk~?F~Qs-tO*;}I#<3}^b0b;kOKJw6)T8EMmIltc2|fjhq&w&)S3rZh~Il3
z<{f|)kn(vgtfb5KS$njD+ovE3VNig-{7~FUfX2Vle3Q2sxLH9s5S2^5sTU}$tz_ou
zL`>_h6TTdm87b_cxfJ};wUbMEI342@&_gc^BV@V3yM~*J%&m3f>2u0X%J%s?O$+gpaCI}15r?u)K!~7
z;C4rcVP;X@2E+;`T%XUDrytLXW?c;!oO2Qdo9IOumJ3!K4dz?$;Qmjld6|iBm+=|b
zWkP2OIGx+84uP{j?&HJ&^6BABdnMQ4ogrM>bt=8AdLjHSt)gt8e+4XV%}Jcj83
zleg0KIG(ZvYV+o}4gmIzbD!ZH6F==@Dt6fsvL{Bst{zu-g4oOX`^+j@pnMHu{$8#
zc66^x+1Zhp^T6rBf_+-Y0aymzUBhSV)W9K1v^zUOSbK16Jbqn%AD^@7i8adVn(~9}f|qU*xU7_-3dh;&h*ci!NE!{)Qi2
zV{hlTvUgp~Q`frnMX9p1^8z9IQof>&b5XMHe=DOW>Y$(Kf1<$@EKme7YPRxtZu{Tu
zbq{~tulvUr5Y0lb-Ju){_&xY7LG8&V=AX>g4L1!opLJ_!DamwNNjf`|n1CPz8g&l>
zei}6=NVJ%tfy5*(m@=wn^G6xHo5pXy2~b3Ki!F6Y^f)3Y4xC(=y{5&QHIDz$BMVKw
zbTp!OH127L;oaunpkF{BTtI}lfQGvSUnSHbqP*M%yMYRK4xS#=AQ5XJno-?!L6~oJ
zYnrb>$EV&WCuvm{1=d8xu6`osAKH*))q3rY2{B9YsHfkEL4gd-)yaM&vkA=JqtF-LaUW$
zhASx_x3<@5r@CRMwT6$eW`cGko936sf~ZOqOSfs)iW(psP~yeADocEwR;GKHU7BPL
z%>ov+N|W}(>ZgeE#0&>jJ)dN*HX95Dhj8H_TvUOK
zid$*o&h=b`dxa7vKRNg)K)RwAeu41jPC`Vs_l9ryJbZT$3>Y=r$HVVvzYnNC0EE&k3I}F9=zuIv@@Y^oxZ9@r&FSE?-%)d2=xHxC$A#K!c<|PZ
zsio%s3L!5yq`n+}
zmmZXsw>pY;5*Kf$UyNrwB`1CmziiFt_PU+(CvRf>?6I%>G=t4KQXwq9R}m+E4w%68
zbKz^s;WzO{wc1sm6B!(d5;qV83RSeA0*R6~@DNsulcxqfC6IW25ruMl5X0VubR7A3
zUu0Bqx?d{l-`{x4p?ST~OHP#A$!Jam)RaJEVEIsA!}mz5thmp821>?}3B}g4^S?%I
zNX$P2>?}{`B!pn=R|Ediz#oH8x!cNZa!()niq(m_AlVO1I|FT>%fr;m^r!f#n#)^P
zw}Ke>X8_ZsfV&#};GrBzG$Qp5+oa8W!}RR*>mHqJ*c@W~4OV-r&p+s52OC`NL`VgO
z{33N^0jW;1#HQxhOH)>#AzI%p!76I%G+de|aSOC~H9=+&y+VRR)Ab&a0QPlJ?3kev
zL6uUdzFO_#ze2oPMm9}2FSV2II04tQ4z9s|gdm@E-rSb~M!3r!FO|P=A-Eb*DUc!>
zO{4n+r=$q|l2UR*uX?`}0v@-oUVOh=j8DOX4(XE%`8bOxjsKjYi1~HoRYuIOl+mnh
zzY#lvEKC`15`gjLJRv|i%i6D+|5UeMu^?6EbC{B>avP^k07QBDZ#T^aDoPqIt5Re}
zh+{`&sr?_eEPFA8>UlkldQH7}e|B;$ISqd=TpJDYYmlws3_Da)`HUE*j@;eYuLH0P
zUE8G&wJQ-^Hp7b8==g3{wNemD=o!FP?$M}f#C$gt~R?L>RQ3g$u
z?V)!qUaW@)-a}Kl$9lvHhlc2Olrw7YZ7}8PT`;;}i5NGy)rF(;N}G9O<0ZSLCu3Zy
zx*Wxj`$~gK+IXpayOk@GG$uKCx9!ug*=156JX`md#Dy}htHs=K#L$f$TaTN~s>j^m
zo?tuZOBx4O3!obd{gTHE%xqN$V7qEH754i%u6^&NU3Z=%+x9$%mz<+?tz$HUOMC}N
zBJs|%XJS^Nv*adv9&=z)GXnb31{5lWFhXWnl#CeoM2I=?QF1ptxG?zu>KW9HF1Nkv`UmE;Z5^L4k^-7gO%wJ;pK
zxUFBm4$BV*+9fy9hxln!F;KVo^syUpO#HgokLCK<&Mo`Y_~(7j*$(wcOXgJw=BQ8g=6AVa8@iTExY$_BWcc=3}2KJtRW?EFr7CPQ_7nutztA1
zI(Rgx+Xu-sXdfP?r3
z{3D;smstjd8h&W+tD%-uW+&zBM}#nJeB!r>7z`dhXRts>J=B1ng0o!1zW4)ui9TPm3*wABOMkk>8s=;iu33!6j5C?6tk$F#R9z
zB))L=ADq6Q`^U;Vv_5nC`2O*cT&ctHpS`0Dqu<<;;+g(Ek;|XEWjh&`?7!)X-XHdF
zckRDE^LEYs0{gZE&%i(KY&R**`2d~Iz~63N
z1xNzFJQA@yy0;V0eVE%*%wr<`IQ&d;i70uCd|$|5kAqif9{iYdSAGrmO5%zeKPa+HJ@JfeD>Z3~amn%`eoM2ME*(}0SXT$t$StR$cBBM10
z-Re8p-uXzwE%Ozc#%Ro&rJRc1*P9$3{_M5!YJTexa#sk=;Kd!U33bwj3aR}Uj5mg=
zR|NNCxVQK$_PeQ%BG!{6+ZqOXI&(h*Rq9uOwtPT6!<
zTBIX!vvaa~()(=}y*|xMB%V9Yh=(MD8-2X`$~|XhwCC816|zZ@M#j^yvuoE+s>5sk3;{D8Lq(Rcxgoxs|(Lbqy98CYV+)hB(ER;8K#L{cg9Vd
z$)d&C*%`y@6(B6ql_?P{i11(cdK>*0K?gE6E6Q$_`8qlY4OjU^Y7>OFaPWB
zud}Gw((5|4jgD98#sRUxZh7QinA?Lm_cVbl^@!qi93KRIyT9qqI8lo)C1E$Mz)Rr=ya=8=r!Jo*7hVd<>%ZI1
z7r#&Uzh`HZmrnw|U%wZfE?K{4^)voX@|$_T8Z|Th{A&!4&;DZD-&f9OFVFp?w?8YL
zWJmsIh2bdn;S;G@E%fGC!}mQ9W=O!UoOq-_@tS-)cdT?w>niD(S47%3I|hCTvRJcs
zdUIr=SyfMUs>~najHH&>Vz`mbYzL?Gw&lHX$=vEL?`=abZD#!RH0jZ^&BNeoyRY#X
zF%UYd_pATP%e`zcf-1ZqK&5lfIpquyqk!0Ft_5<1OYAyuxE}SJQ(fre*y|WuoqlO`
zMC0@)E;7Yi9q-&*kn(Cj7UJNRMmWhtIoZ*UDx8BjbS5Lyu#|n!J_lO#?T2CXQ>nTm
zm%EG0Cx-WTBnF>*{L;NQyK#Hl4)qZ0vi6=|>s^M;JkbP;tJ5Z6Evpx1^Ssw>-|zdBQF
zbqwE>QFVAN+orm0+r2euxtSTj-PoyLC6RW8ITTK!632oiFCTL%a8_aWrfsy_2YcmX
z-U;#fmE_XbqQ4G0$v6y-%59OSD5bKkoZ)%@*nrfOto+Se@Ty|CC4C1Fy+`}X8r9M3
zicLrQB-M3SwO`x(JYD%Lm1|6v;A?}%+UD&J4-o*1h3Uvl5X`t?XYgHgHdg1XPmz1q
z;--VgMZe26ZiprdNoBdwzzIQ$y~taLE}mHfi}bf{xu2RP`s`Dzan+bmR$n?*&nfGN
z46j~n&OOd?=20bk_jyxMek$3~WAnL8%FdMdny;$WfW>)<25gldyNB6#2ql{_Dj
z$}8tGkn@KQz?T@UhnH*+sEK#ey{OABIfsce&+?eK5oTGFcLPNP^j$^^bR|U0cu1t7gBWv?W2(4tY+_I?~5#f{*<{*f!mr(w0FmExq>%5qPFv
zrc{G_WwuPoNW*1wFG+6D6u(NSi64CMzFcm!&jc&f#Kxu*Bav>$hC-ZYFDss1)ONi&
z$NT$>I)6IC395!vyO*j#u5S&I*K$vz!6ePenO{s+B&RU$HpXSC=NhUWXg3F?+k-xI
zXbJ_si!>j50*%Ek=%AC$8?ClMy=C)BuoJrR^$MKYR;#hS`A6=P3C${ZQh}
zG^dIJB-W^f?^sc~*`r!d=B9&52ZE$5>hottHsYvvHruqx`QIccf`#6-98j@Mq9Mf{
zKNbf0pp?fSI763@;S~N_l1*?I%JK?4;#*uWYlB2GJ_25B4%;^ZT1tt%=Cg^
z7_%9#v@pKCLjfZc(gxmE;6mb)w5SPjsI;XKE8QDJ_pWgD4AH+uTFrc-srOI>me`ble=
zsf^^Iq(0AZubM9F-D5>z7mpX=MSTRHfsr1#yvEu?y|VLk1>ZU^p2o?Rxuce;i$f84
zR^8!(kXLin8+SY
z3a$K?{X!pcMBSX7QNjDD4Z&2FL92n!_l&ClV^{tshJl5k>yi2DfvR99cR^0Q*LKQo
z%Nry|J(sl-_Qz3)r_}0p^d3K&sq=796%(>*cX;w%w|~RD?mc0hcPSMG{Y4S9wjEWF
z*;a;-d`4QfO2Me&lccI%(fchJbZOdP=w^nhxq|=anGb5%bG0_QCn3a_1IwB>K@g3f
zi|^VV-e2!(i5FDBBPrYcqQ>Lh?6rZ
z_9xI{fj7>$`dqwoPI)>}q;7;PfnFQDI`eZZqwB4$JwwCtnREpp{%;4cS9$L7X{#s{
zUKrsrglc8e#`{2dz$a$x%5C00B%XiNhvGrUZUfw74RTB=v|LM;-0A53{B!-ZbaSrI
z&VM^6$iJZ?Hcl!KF>dWgG+TmN0o)vV9Ec;(DSi-r
z^pK{V!o8N@W)WllL*QJNf!@yea(9|zy!at%<>rMtlJW0_4blB}hWYh&>>P
zRreKL&u0X{v>=<%a&GJcHy-i5ByH!w;T#~4MHOFZlgppk!Vbh#kpon>t_Em0&>0Cn
zfoKH!AHJ2FgvieBi@P+Qrk4cu)F(Xe`yYt->ImaVV(}nAk>xW#`X?w15>j&oksWR$_R=!WL$(3M|C4=$2
ziw`jSW3c#gt7S1E%>YN-GOz4bq*)urU?J!FO!#`s=O8iaaqJ+ctPuTb*plAcOT?e(
z6h;y_FCaKf?iWSJ?FD1@sx5SZGvc^+{6+PS#{36}+nuD5Uz{J;6
zkCoX+ajSgz-Vw%}_|)ez9I38sY~sI^xHQ&^Y(ZF)$*FJ_q{JAX*hGHG%}1(EyFb6@
zZfVcB1-FH3cmpQ;23UX@Exaz(nWC>v`_SFfIR=VeJo})El`d8~NfgxRW)_0W?9NHK
zzy?797)2FLBCUsI2;BG4*2;DhO7vwOS^ZWGU|wWL`TZxT?!vI2Wv3ZRTpceh%K8)n
zS_$hIjql185=TM`Sj*rDJYM(OLQo4S+wltI;Vf4s^&R$5L@hdt7CB#Zm3B%0OM60(
z*;-ase_p;$8rj$@;-8VhUP-d-;fQo*w6{}qCvm)u@x=FUEW0F6{L9EMFeM|{vebUk
zec=2Lx)~^j1>Rp!$nOOn21$U*fR^4==AX>tfpdSx#UvM_TXIvsy?2DupL9Q+PM`Is
zr|+V3oVRaC?~9nez%tQNW;f|OZr7>`tv4nZ0@lj>_@CrVPGsw
zRR0yfhA@pzbvPkjM#folbP|}}-mIi$#$;vt>?RdT`iz^0C)Gflb?tpM34~Kq=h|Ud
zL!Kr5%V>)H%dfYmz!{fAe!>rKk&OMWl38Ufw4TT%)wat4dm8^nnfc~>brNMq-Dkml
zeRkmUnmRQiA&BhXOK4N|=G5O_MwG*zcLf(>1(*h)%_)R7g>qa%3?*E6(XQr->~@=z
zi3R$Bs=bA8+LUdPU_qTxavmrtwraae@yA|I!
zkKe93;j;k4M5J+HA0glGv7*^XK^EEMh{M9n{LBB8-@&kd@wzJ#enA!=Z;Z--4vfPL
zZ|Zw67Fu~st{DEkC>CpI4`z^qARCQa42mo3rmeb!s&K50(kOH~NME+~H<&H)6OULc?Zt*1B3zj2X$zVNZ+04XA2rsUgKS=Hq@vSeM>8&Q;tveDJY}hHm
z%!YmbVSJoi%!80l5_8WB`gc8(c~vfnMK_FPKgc`e@4p#PBTMTmVjL#_2qU^bw<(G$
ze|{}iZsQUk{ReXOj6~G2D+<0Pt~Vg`x_LJ)3;ca6`8Ws+7}bIQ>WkgcN$bAPBBI0N
z@f5z{b3rB^VHgP{ycizG(m?gEo_n^
zuuDcm(?bC>FD_^7Ei8Z+ho{?Us%jl8;{qtlm2xpfEBK%c94;Rx(r?(|K_}>GP=6zS
zv1kEeZ#YVGmo_A}08z5%XvEDmf`WwK11ho|Ar~yvD4Lhc?n39sKV`%?kpf^bS3XR~
zXxRR%<_uVQk#)fK)SSO$IxY0Ie@hLz_fMh8i~xyr*5-60N2Nr
z*Go(&l|N%sOZpfIb-u!?$jdPMoRyM((5t*ClqrM#*U5s3OiJ@>S+#Zn9e>6A-J5A}
z3p9HBJ$#qnSM__M!y0_|-O*Grs!w{c9oJti{fq75Zkt?4#R35K6dnxYM$v)l65GIz
zb|~KdeDjSqg4q>3n_Hka;vEkJ4&hr2x@D>m`v_hccV0R$lPh)*!t`m(BUi3C$Nvoe
zz&_b*z*@NOe+X7LThhB&g)wbN(x_IWIY}~dl@1%cy)Bx?WOv(28twh}ME$D|^
ztoh|lSJ)}5i|8L1H2!kT&O68Um@Ls!COxo
z!0*=yStDG=HH3ac95zJyM)Z~vCK$Z@OkE*{D3oXl`r&HH9Z92_EDb2mjnv0!QX9M~
zh)INEcOZ0sq`I1e=u3PJLf+iKztPKD%`*~MA13}loLW-?(?S5$;Jl3%dw?;)VpZ+;
zOy7p3Ko_cjUu9*x>N$;h?f<#k;Nrud2KwyGzm59#vBg5#9+pZPasgKvKyCeE8Pqun
zB0=zpGL6th&3#D}6Xtd1PW&7HXXLEMM*0vzRn7=@;lZcGPb80gzkaGH(^T(n*^6EA
zd=I`Nsd{Idu|_fwNkGzl^o7W4j$*flZN;y4V*zqZxT`P`4afEdvYfe;m=B&`=>G&=
z1&srL034KUu%_8NXYeZAi#b3SA|ol{>$})~A6qD+_DMc`$r^a%B1}oEZ|jJAg`#E)
zgSI4p#*(+jHj(i8jadGzCZasUbuDbQlGx8RFxRP`4dxQ_q3~6+3$xUe;O)4+ero23
zrVD@Wq4@lT*-<_Bva+)>%G#}3Xd<`P_y?j5%b`aKphwUq9>B5z`rC6vdSMkkdLG{I
z9JPE2?Zulf=Hq{*g-*i(7YDiOWXbKGc%0t5p@SJaT=esls-H~?Bb)oK+ow-MF?tYj
z2bI%BK_VX@zVyt*11HMtb|W~alaq=2(!4f1SDrD_m~W#shXYm0ZGsu5qpG<7p?sk^
z9U>@@R{T}Y;M5V^uxf1Xy_6sk#{0wr%>-wyYc-2!j0XY6J*CU1{Hd;Jk3NiIHRPBy
zu<{XS@15hK#oO>7NA1j#q=q#2i9fGt6a$gQNgtc2Dk73Q3Q*oVgmB~8J_*Bv;Q(JU{5f6Fxt=sY0H)vX`g)j#)3LWS!Ii=L+elVP66T{ln
z&p!9^g>;weBsoWN05Xw!I{pgu2sq1iG!;%&lA=No-%qPM@X^UnjJ?ZCs;eT~l~j_I
z2O0o8Eq`$Lvk+~Lt#v)m)7Vo?@v|>yCdOUoV)MJzrC!k
z2$$*6X)sSqIKfYMog<-}JAQ(%3-`=*J`rrL|NakX9N4HBye{W
zu>+!!J-k;-D(KGWI&m7@r;Qdw+_FEC{ZClz=J5e)^K8qJF-T5H(bZIHe
z@{p2^OhHic$X@CW(>*stWlLs!_Ox$#kIzRc;kp#Suz>flKgh*plYQXfhz5
zvBl1!*z7?;u{gJG9<^N>c(myX^=7?p
zC}&&W7S~#D)jPKrmfY}16_Z{;3W{vtqb757*~go~yZQH1W&HX1u{Wj=ku`);-`Ya$
z_Vp&TS3~nL;P1ztWDB5*{=mO0B#=Ap1Z}fx3`YKrpPSlzGz{A;&Cr-v`cz{*^oY!T
z0;-?k8q}3=Tqn@Bi^O+=JuT&DRKU|TVgb9y0xPCyB1y9oCzj6UG1c`-HLcO(&2Ru5
zOVEq1KpN7;_PM<7t&jJc1FcesFDLn7|A!!p71hi}3V5%T_Y8=1k)UG?0Y+r=^ws20
zs}X}|Y^{lKsOKjCHK_raE4Fom8MVLw(Z19XECYCS3DQFju!r!sWT~^RAn$gt6L2np
zB^h8a|E-i!*jr0K@#U#hvs-tA#cThc-O$A^u+Gqz!8j+g)%Q}uA-Y9#_rAVbZ_E#c
z`_q~&NF730bN)Vk_I=de8O)AWKcNg@+&48j
zlpha>eRu$`<1OEB!k>W4tRam5)>6JH6mGTFuFw+{pbxnTBVqQxHbbwGAxO94-RJG-k<(F`z3K|;L8s~(JzMrbMx9J=$&b?^1jW^7a24qzyq8<~
zt$&~L_PXHbgy+dC!sAw->-85Ib`1FqAj~)#UMLn}k3v6N=Li!5#}#a^E%7b92;M#O
z^L5v4WT>8Ap3HTg1@Xj3i@Bj
zqeR|cR<_OsJMl98S6r}gEzeFRPSWGj6c8k!K>@}RqQ2D_53<14347k*xccBv7mid2PIEKZT!VMl%p)AWxQ&Z{M=hlV*4d
z92_VxrQ)zRFP9u74zTMBGZA;~pG%w_GuHCl*CQOEpKI3RRA!a8Vt)Xx
zXA>Dhq0w`^k7VJbd}DJ4i8^luB*$(s?NoP)6%5=cfb}$RG+U23*vg@S;NUj
z@eWOBZz5+kqf$&Kqs{Yk5zaD+GZA~`7c|I=GSMAbQ4CXWG(}V;AgjvF^yv@B1Q^2RP9(Fg?8D
zRIR`u;hnU`73AGc4RFG}^O84)NQ_JaKRO@t)p4>?vBNIA+Gq!;P~jrMQkw%nlf7Nv#!lj1TuE%o3b++G3p23n|$^>^v2
zGv;4{5Fc{1XeUh0pj_KU$*p^Le;)_%PqGOT1+bg}G4XLg8NuqDU()9FJSI!^0#1?5Be{lCfQnY#KTy%0q1RK`
z2d6@D4=>Hzy3(&d9~R=DxI}4i$6}Ud9KYbjoRcEfG{vU{5f5qj8&yf3sv24;-ML3)
zFFg-JF1xiTks{uDM+T2kFX2AE!;5ag_IHikOo|k=I9neUyGUm@i7!?R?^GuIW{%xO
z`a~qzx~n#?)oZ=CSg~$J;m2_We-D+;pYi<9F4m{MfJR4Xkl!oOw)Nk0i~Uw=bra{*$5t1o0zH8|xFM
z;+S<&pZgeRY!p7)Og5Ta3;G$)>qwFCfgP)B!(yMNv&7F-xdJ&>edq`e(&ur+;ywQR
zM+oEhAx2QRI)KH+`}ETIk+HIuV^9-g&s%Mf3KQ-FOy@`}-p)OgUR!MPbj4l+O2R@M
zm?w_GT3sP!`#o;wro`X3Gb|k=F9^;lK7bU69Oiy>_>y$fQ6%KK^V1hImgO0RnDR0@
zHAS}K^iSaeY+U^JRP1d*EM-UTVDgLm-lDS(USeGa
z9uUH~QRtF;&dq(^aH9rlF89r;4Hb1Oy51u#Zaq6D`d*D=j8@hk>y}?*uozc!5ifR@
z-~!BZi{O{#w|{HjE?_%Q3^=gZ{x8zQ&p!(yNoxPjisL>!gp@`{CUDN5BzgbEteO#`
zSAg3+KTOB+1>T*9MBci4ajhaepc23>^;RjBzkXV7EYU}2F9klPm?dX6>aY%;1!^%2
ze;*SkU@*2<4(*z@7GTg2DEc1XGRBjmb)
zHqN5vNz}XTco}5^5|<5~cX*Foqj8b3)h9<*-KT_%p;Rr-V5h
zGvCJq@0o+4Qbq2CJMQ)7M=#|4x1&)Giaqk;?_U7PA035_MFynuxe4+qj(ek#@Tm$s
z``?S7+~+;Y8gF}xm7vY`bEkT%v!n*jAaYB~o6naY9aR)aMwFWa-FK$TDZ=lTRpOTV
zi=P`44dU3B_1}+n%;i5B3Fhufqzrz((tg_^dfx399L<2k(}&&^A_=KdyI#9SL^N0K41F;m+_>y
zM%R7bfD-7AJgtTw&~bXi&)(;T8@xRp16>m!NQW7Exl`#FsYlLs)<%>l4*;wF^8Ka<
z^H88E756VACw(r_fP)vrKVaQi0{B2LQX6}cd>_@ob`|p0+qIvoyyZJT+j>U77Y2XS
zW271%2I>TQ#88-Lsjz9VLpFQ|xIMd<_JG|F^6=a`Ii!Pd
zu?|b;(aMX_{@nH*AsOG3Z-k>o?Bh%OQd4(;0<`C@O@
zFO{1rjuTinV&efh>5+r
z+bdJp4~fL6MW{x>0``jpU2ngheSP0Gr>@0J2D1q
zM|_=!VQFPG8YbQaJWXu<&S)I;3Z`UJ34YCg>g9TGXmbTf$z#nYqV&+!C4etRrt=D9bOLRhG3V
zxdJJwRNaj7Q>AY1fW4m|E7>q4wsUk#dAczp(y-_4xlV-SyJ5ku+f
zEwfPzgy`}PR-lH%c1YRTubfVrSemF`TRDxJc$PlI~0lmX=1UjQVN7~J~NVk@#=6Y!t(Ge+eXm2U-
zRREq=LqPk@&H3@-xm8P?xwg3mAUaMyyh=URWK*hAUg{`35T4+IhlaYMas3ynizX&_
z0gJBnN2KU7;zp$|4(MSG+Q}bK4E?b@vj*Ob&)s325?5b5`WywH^hZyfsd_r`Q=f2G
z9-RX$u`MBE_9E)F9sdhCK*qmuy>+Vl?9>vD*FWNiVBK7Fz%?nWQ($fq2b;B-QVNQG
zlOwk{%K7+KfNlZy?x}*@S}5ibCJ<+Yi|6b%t+~fsS?m!VKP!F!9Fq5x#3AL~>x-7k
zo|ECZ@^sQlCQ>;xK6b-s@RNK-@y1U9Q*6HtI+$qp1#ol>Rdaw`jW*JoxuJEyF~F4i
zX}yoOBFag$i$ybcf8tP0&7N)#;lssT;p93+6y=rPOnKZpAWbx!w3vph=
z*#3NDamC+!Vq&xwspY-_*WkV!wd`InZ^fn~`cXWmr)m5CbWw1(GL*vzN7#hCH1Ia!
zS{=h1i)sMi$YvlOS;>=&w0WPV`D|_RI^8E5lj_LP5(gLq%2%5#(+QqeQr4PS7tVy{
z+f8k;o*c0NO0r@I?PSmJdmhEfL+%tq-I%T4oZpyqlc%ueZ(T(Bk#(1*xg{DrJu+pW
zLfUP9Vb|8Ld1unpR*(@NIH7ontsB(lr90MB{FXoa(=_Q$c7U~B=iPbDZY%7`{?seQ
zq2=@JGYm2u$_;Y)Np6Q|s{(%P03B5CoxL<-=#CSNA%n$CfnOwwAKh;Krvlv97)m86
zYo5&LeJ0OBuU)9?9Wa*D
zN6is9lMm#i6LQ-v5q_hZ-G;jh^X_O&5Kb#j&pVOh6u67u{=$-9>#+eG9hk*5@^jg1
zEz3F8Oq?g+EGEp$>Dk#?OI9a4cU0hTJE(&Sb6-k${)^=+X69t`hx0#FUT>|SvvrQy
zXDEhVh#R)Tfx=MjH=1w#6FHm;&QNc$o6k0r$9GZP2WnE!!A{utMW~M5=Lcu8sb2e1CBAGpu3rAd}5)6xKNXe8Oz{tbA(93QSW)LB4-$
z`ls_7gq54vzL71?ZWf&7K#UF3XDik$#V{KpPR72U#Ow+*>wIk1^|0?(j`lmg>0cN2
zheOa~J==GkBb8UN3{t59U#pt2`o4bU5Qafu*24Y=TZ;+e$s*4Sj$;wlY~7Zb9Gp2O
zNl5hj(tcMT_{7L)vZCO!(-Jd}zLV&1;OBM8yLU5Svli$!!7B_*OO;?px-)}QsBkk`(@lP9^skPgd96uo}Vo-RMjF4
z7!EhrbzN1Ddny_tc^k6HCo!0GDdjC$9ikt+0rJ+hjvpL|AE6!$q8Zl%xD>bB#hV83I{8A-)xKx-E(=Lo=HFcGi6xqCb}qt0OcS`<5@@pb*fEg?q|=m-P;-t>?Hu>2)4LY{I2&;%
z>)gNAxLbMDFYSV^$9{8cbF1&oG|!R!X7$a=FGHm7Vr;kF;<*|u^jXB|R3sFeehfq$
zn}$K>UIPmzyL2+vn_IARzg+Hdgd*%^o5LtrifYi3R=BNr*w@V&;%pkdrz7m+X!{C3
z7KLgr4j8Hf5fIq;Xu~+B0@iTVPZQ^wx9#NrXzad!Bp@1Ne736x&@?e$F1XI%41Qh~
z2GmP=}^(DJ>MM`I%
z^}A>ix+n1cb{>pth$du?L=NK%9cN*7}U`z
z*pZ(b`IblkD{prl@2kl~?^r#u_}A+ixe7iBG=6;VF>G#Qj5T){8*WG2d^@owOA?;D
zJx%jZGmhhna_@KevgG3!Y1!k!(&;;lW8hJA>-6mFtwGRO^aWvExa*O2o(_!F^=7X?
znMXAtuHe)>o*)OBqpQtV;WrOu^}W31V_}1&*fTwzbak4u!_eBN^g}=DZCmWa=KKN&^8vCn9lb;EnX+_sBoy7+2Eeo4_jsCgp
z_lQ5BISi!R5id30TmK?_H3!m&NOx@ZEn%=wH-Uea^W$BVpj8ye5f|q8l?0k#=Q}RC
zTnws2cz0)fAK+j%%a}Vc9*O2efmFAikwc~fbcD-9yRhd9HPO93a(;273KEcff!mIU$N1W#iEzdJx
z;dj3xvBSKkp;#e1ExxANpe6fqigTs9ct%#dCM;JhLEb=}%S87^rYQ(@qzNV#@E@e{
zBTI?Lct$+#p={|w5%!q27o%N>4$jNf%4B27$qRkV38;0fzq{<_m1mwQb=6JQ%e!)*
zVe;XR7b1&4xFe>(K`FGCR*=cbt$3rAnTrmwm!@Ka1Nrq|lcz7BHWX1oTn98cs}gh-sCE<@?fS10P|R!)rRv!qm1
zThilev6UmYHHmzH>HQp!<0MRw~Vyg(jKuoFSwen7M2
zPo!@;yOcTZ!YE(feebmTzr4eFO*!+dzqqBUPsDNc)A2ps6l(`%
z5BVNSPxw>$($Tyr^41kOTA8vExW}~YBY$UA{RLjI8W-l%tzbxebEb=pKMj>0-M(}5
z5*#!~Vs=Cvt0l|Rs#y>RQQ&>Wd7zPZ=O7&EiT44pC;f%<{#&0Jzd`(6-z=AG)g)id
znDcj}>s4ojb>q~X^#yx*Jt5Qs`H))je}ZL>jIaBV=4Ig?(|T4S3cfc4c4Q`
z`dE4b;k9tRpGFgdhC_
z{C<9bp8M;2C*b!}&I&(R>SrBd`B00Y2AC(+1?~Q?v`+mLO_U#Ec0jeeLD~kzr_i-z
zFJJG;=X0^g$Nni-peZCa*8HNeR~gvH*vO<1qqE5foGa^F)yFT$
z%OdXt!Xo3`7Ji5PMw@v*M3a8()BuDY((7~I=2^78HsvJS_Z^Zo?vJ0Pl*?!T^V)BJ
z3#5*{HIXO#OVghbUJ8mEp*nT2Z~x%0tk|e(tM3YY82NQ%TTowLHr4=S+FzdE^UBG<
zK}8goNMHAgpY6wuk>q*$ylk!Y_aDz4)L{qLS;X@(jH0t3TY=Vjf(%6*`DI_7>-r6v
zbMJ5zei$(K(+VPmzWg`j`gParBhM=0v!auO{DVe?s?zSY$y>n;pg+uTwi|X+TFO7;7?Cx^;r=6p^CVfj12gUle=c5&|dquJK(qQGGjd
zek;EGvRyL$E?$(2Z(YnM;P>&2Bc2sKfT#R$l~)G*!n?X5)fwnOe9(Tcn*Te#AOFMv
z5Yk^S3i)lgoE)P~uOg2G`Qs=yzkLa)i;$}K%01SI&v3wJ5Vw7ZVTQbbRp))KLXA8z
zGb2(&agd*)$Sp=c(Z9?^{BSmr$8*62;LT-bBY$PTy{$)yG~4XRy$k%{Hz!594eEEr
zg{;mfg1qA3^sVv8i|V0HT%ok0Isb)094nuZm!vbv%4zlwSJ96t>MJP6vUmCI>Ybr-
z>L#nt{jK3hD?+{VgCi%Vi^(VCmes@~
z_YP?8&+D7#D1iRkGr-!A&l>@s*;e;lhrmZ3I*&YbfzyC{D4IY;0M&?3eae(qrgZtJ
ze)F&Ta#MsnRX;+tai@B9IA7`Oh-dk}VfeHV>`
zn6}>ILxf_xFzod3=HR_@A8R6V)VtHZ9%y9YFYcIxEQo&nPVf2*A_$}NZyjACj!&(~kkMCiNfHiXs^*f3Vy0n=Vl|l)1CT<^Usvo=k!v{J0
zAdw#o$-1i~$h(zSqr8e9&3TP~(QQwg
z#fo>ax|yi?%ddZXF%1P+xXV}^vqy|HaO{b`x5lH{ErC6XEw#_kznQDf&oN2j+g$O5
z*?XE|l7pJZL`t1J9%5q;apBzgkEd5C&tM#&N$XVD$M%Y+wzT$Vh`2tjJ;__ESFReO
zo@5YfHu7(f=eqKB;C##eG@0_&{|FgRG1P?OKLvcV2TVuxWy3sErL_p6>h#ufj+}U2
zsMlt9T@va|<9qo-uiqDBOca%hJ-aWogd9Qkf70n?eDOdyo8KC!QbNaJ>-c~RnaCg9
z*X`MLKODBRB$RINPtPre+`@IOwJxkWO39Ir){A?5p<^6(Hy9oldQE%fK8zD8#c{@9
zD73y0{8=M!Wl?cm1$Ms7!#F}QIxpS5fjUWUM>T^ST{=bRDdY=9m*|eM$jABp8GK=M
zP0D7I8`|qY+;L6u@c9%k#5|vL2BUFY`^2{WHkujTbICR%jq7N=u=A0QiLs;O-GqFZ
z(~L3C<3ZOLZ0&t3ENJofZ*Lu7O;}XDo=5kr?#hruT+TUlZm9-!8Wl;CetjA=2eBsb
zJe%Kl$e1B9cC-j?3Z%2g=i`_}aYhabt!v<3OW)gQ6FDoyJ$;W}F
z>E&bM|SwfhtwMbXMK8q&V`fl4KZSq(toeETsHp?+Y-ZAP)Azb^d7i~U&d0nvf
zzSjQxpPp~h0I@h863LIrPO;oXD->_9So`Groiy7mMdx0UBtZnAP${CAGfeadyn-$T
z7AKFjetcdZaZGL!d_0M|K6V|z30^0^`!~Nlmj?L_EkBt>0J>NDQ@hqa%2Psd|F1Mo
zf2y_KWZbn+V~j`rb-6GOKWB%&PrTj3m{_(r;~2T%WqXHt8$Hj%lBL2KCS~v57U!6p
z2NyFNU@{Hewa^F#oYvq(o?s1M(CXkYc}l8uYn)Wgoxm
zPj|1k6L!)0!MaEXm-L<)BIEu~dtMj|oi962l#TYJ|G9Rm=U3E&-a=iQKH_nx0(DzY
zH|{SzYP64qBvG^3%488?r^SA@50YwD4HG<3bGi!d)+too4)_g;`cJ6GYabBDIVA)C
zo<8Rr)C@7a&Ht#c9Q5Uc%*su)m7mjp-RqJIfRm`Vrd;Vi9zk>$3uF}lrTy|~D}DB=
zgvO1$_fKTM;X)noD6?>~1#~+796Zp2h;KtyYo`}H-}x|Q7202fTWTg#m>*n^ci@So
z>h!?r{K-XL)9XiDU0*str-TlCAHU-yf6{AcCT%f-0b@3l`CbkVC{x(n+4@KPI*%L~tV&;?WA
zoRL1TR=@DYTuCO8M#!(#tC=eDm)4(_N)3tn=>E)GBEHbPsfN~qUTD5?8+&nuuinhM
zPpt+1Y*TF?Up^)YTD{lN-j+9Rth)cpPXPv%8UhG@2AxzJk3aqj!Y0Ir@+KBIz_!FipvW
z;mY2&H7|hgKR{oR%YzPn!X4>lrQsBZmV)2R68oyNgE`d-0J;c<2HwPQmxSYN@|LMaM#&E$uoCVe#?WR%u9?G7-Ms7FDJ`lR=I!DBEbLUC&}iD~A@)E(%v;Y{
zs5@u1%-?6lRspvnuC>%Vr!%aBj72hrn6rrK3C8!SBf5qWYb)dfWfv&n<^EuR@
z-sUqxYBX3&2XzhEO`jSiz%h1I#=;ohkK@iS>j+~9uLjmK{)Qc*WW%^h0c%&{jzc%i
z4&~1`DKWusHvlL6XltY7kF35(rAq%Y9KhE=o
zz&tQDM{`U*F&fT;W9lzs_U8R#5~#)CR9}9e*Xr`U+5hHGboEf~v@P~N-`r*gwMDzS
z)R(`exRXJqR~VxnVO%KN#{d(3Zyv4|@B>@TU~H4|hV_Yj-tyZI)&BT^ywZHY*WGrj
zapdU_Hb_|pa3;zB7=N*?_N^<6eYZ9nUOWoUQLyE~u=5U(LLBIJrQr$SFQ^q4cn-oB
z8^hdzYm%=+_J*}Tu+C+r@Z8E~fd8_ga7*_cM|))b#?2w0f^!aY9k-|enxFXz^EY7f
zPuIk#{R^;DSX%+H4(P0XWNh((J^NMPNb__0h1)-4z5ZSI;$&9%RpSUrQW=IYb_Q)q
zd*PSImIpG)UcYYj6j%P@hvss(*yWsCT5R#(IaN`JRer;1Hr%gvErzhJ@g44f?v8-F
zcl+=VI0RqYm&^b!poPD3Iida0H}+-N#!LJfFLIB^Hh%W8(i8k;$Y6clcijfW0U)$v
ziIe;c#)9wP@l5v*bu18n2;rk!qTp+6*w<`*>OLPCF%(B)==VD~*R~JW;8*Ps6Vuo8
zn1!eAQ_qWUSO@*!d#m{=j=!&K`R)9tAF&wm2#7_tt%emSCSw;V?e`r&pidQGC0xN5W;%otzH_?z0N!wYmwdjuOXQ*&m`g39ETcs
z%gupC)YEQshDQ%;(||UL;A_*y@cruH7Z}i~E5s65a|(0@0Db){CKDgi(`o`~O)@R?
z#i1-G_brzwed7VdSl-MAp#8VHq8yNIHWbdBLCa3ARbTwXrC>bVmmeQoxqG^AW2YZ=
zb^H?zfo(|~g&_E+q=k7e)Pw;c;ynq1SnlwiHR1xWg}rQkTew>)jOR8$dsL$L))wXO
zVlm6@t+yPR6t^ulNF*NMT{opI^8;{F0)7YmH~_ptyc6$;8Qh#O0$SRkmnuQcVjRb3
zHli9cvf7x8g34Fe7+`v1NwC&GBH8faAYtQ-y(~$@vWHmZ2KVeKU^kdsg8f{)&di<{
zL!NgrEy
zrte}Ro%w9C^&PC&hFJQ^ZCYEMWC?SdkUwm8yyfCHjuuUHe)33W#%otUJFx_Vrdtel
z;pgle>m<;-CFti~_+71xcJHZKeqhyOB3ZrdI#ks(7M*Mc#cXtKYLsjSW?;ZLa>(i%
zh9S!w+C7?;Il)>QIB%8w;eXI}VXSD&-EaH`l9gxXHy+@O0-2H6vtDLu^Ti~v=UAcH
zTlZURM%7JNFS9-WEj)tRV$=X-ssKBW7*Q|FEujVmWpt8SDVB_kzWb`3`3lcu0}RiQ
zF0MdlrHVZQCHl-5|FA!F?ve0&eg
z2cX}1W0iol=is`}FPJ0v$}LY-VZeU$x%pCg@tB;uzUs$(ZvJs`fq0W{~p%haqnq
z!HqsO`U$|@PmagckB<&wUf0lb_b@kfvKh?_=?k*$?rMnzD
zmKab@s?nWEpu&YgsXO3Fr{b;N=Gk{JRASfFXYMYFyhQ`iDPm-MXB_t@$nkj^rXzYC
zzOWgPvo(cv^W_(Jf%`L%3TS#5|5yCqj-8R>AcKwbRZWgLM5
z(5Rgk$s$YNk0Uz=0fU|1+*raGr3sm@S$$$K#3vlH@k^ky;XVx+#uCwwvz^2@Ged
z-kqB+-W+~$yPyHOknAGQ{?ohK&Z5p*DZJGnt)jzI~
z!!l573=FN6;yMFeznr&v>>oMs?_4P4qmXBAbyAzwH$ul^tBo)fN;=}j0ydkiozVcyB$_i@~fmmdZz;~lw#BX~)4m{w2RRu`H
zXV2HR?+E6J?G@&K!7)n#Ilp@+1@#m62RZ}*DF0Kg>C3Nlj}eMN%(NvETj)N-zqHX8
zmr?tT)6YT-b8ukok{`a3u3+p7324I`Gko&0t@)`jhkFCetOv)^_r9pls#I>{E`&%l
zZ5o$J2ge}ceFi^u*q^**cy~802FdhWRIj$Ke&Sfyyc<&ggTqB
z{$a}re`4?#;f=lum>-9=_8$$T^e{IJW4Oic!!6QNH7~av!snc(AGtn!0hx9*!DXDo
z`*q8aW-h9R#^uZ3!(0i!@t_+lT0-+P=gQ0FF)^;LIiH&23#`TDX6X&QU(So0KRJJD
zn@@{V4UnnJovy4P$z?KC9vLsEGghBmJp_Cf3KG=l^-}j`v5x;rct~Y
zzvaHq6VM*duUe?T&wXdQvBf#%nB;-ZA}}fj1O5Blcjem`SDPCw>GOWLedYqzyumyM
zm=p2__0}XH!1nUygByd4d}5q7jp&E~_ds|jz-x8+z3ux1(*%s+4D0#~0mj;ZU!9H9
zrey>L*~2k4c=h*tO^^NvLyYPd-XFJE>rxXVf;wsoe7P3GdbDr6??g=eejkpR2UFXS
ze;se-XDp&r@td17|jxi9{^{sr1G)PZonT!XzMQUUPvLV11(
z|1`G^h1MkI@cxQ50?S6WI?CVno1#Sz_YeUDxMUj!!CEw!f3}-M<@t-*9gMBMxc;t@
z3Jlc|>uGJZuRWtH
zqpc1E-ZQ$zosPodj%L~9Ckpxn*jFF)bq1Qm_bL|Fg+bkW1Uk6s%6OF9Pt2h=Pvyr4
z$cxyWT(b=t9>|2)0n3JBGPuu2tYBTf4e@3ObQNfa+mZE()_O;-(f~4XhkK58>30o`
z2G05ToHP4<4GhdxZR;THbMXw@fcq#FYfIg3I=?#ZUZvZ^+8lNhlsrr;Z^UBwmyp8wQ{-fE0^#~{M
zqec?uF%yR8moL#R8MKw_a2*02zyrJG;rStLJXbN8E5Hkf85THPp)GIk_c6ef@TV9J
zFHB1&L2L;_O8D2tXza@aV7z$vsaCIfK4(p&&YMm`;y9tux~`D~h*7lOe#hSlGIyWn
zBB?iSeBk9{5HA;wdCq2(P)G3j^_3*X{s8x#!;MM6X8?*PhcHID+AvOngyF%(Hv+ri
zbeh7P3TSwO^#^a|t5)Wnd$@h7Ju*rV&r0Aeplam4|Ll#aN|X&_N#n&hI!Zu_RK=^U
z%X{we2rkUnF`==V#$Egw_D`&~#-hiI@hIhz#NLEvy+54yjiZ-3Dh_ITwyD}=AKJSv
zUy5YZ#l_d-r=AD)^B!Qw+q}GV?YKbjwpXYk1X(0XJl4U!y^xcRupX+t2;U};QlD3Si*U3Qks6f@P|9rdNoxm
zVUt$FbfjA76>F`M-OW`>u?>>50>)=AGiclFQ42
z-|#DsT{zXdFNn!L!(3J`y?!?}ue#ayGcIY|jTW>7{n*P+OOW(P5RMyIW;i;%IZK+F
zYI;xWd2_YgX>J;l#TEg@wQ`bQYOVLT{O)AKL_0v-@dSQ6dyZVvv^!whMGEfu>Zf{0
zbDHH32M^Kw1M0s@tt)Qad^RqPqDAxha*59r@6gj}C7danqz)C8J<45gbGeS!G+zmL
z|06b%j8QhR#!JG_3A5D8o5TbGpY?$qjI`e;QGv=uEG4QJa^RZyPCl%UxjN1hSJ6p~RPsBJ?={iS#k=Rto~YaIhT2o(giPLO-t%3AnLSuQ
z|07w}W-a)?$=-^41sd7$JPIwJ*Dxke$~&;}$AZInR^(k^5R4+Zc;&@P+u^00UN<{4
z(1*|)HCs*0Mc7pqeU6p73kJ&v_E013%@$X1-py|9
zYI@M3FiW(X0lQd~aLw4wW?Xi~6#4G4^_h~ky^>KA*hCq&dj4YXdHC=|vr8xPD%baB
ze!_cqiQ+|*(LHiI+oXM9CpM;QDY{L`@5QS0wMX{x+9$2lOvkdnSukOKhI8%Z2ZgXz9S%
zzWf2pLgEJILMTmWYY`L3V(y^^IbXpdC=1yorCW>Q^&*}kQU%C%DiUg;Al;ql7p
z$NIk;H#}T$)_X^q)9m7T$qd!v64tsOm2ibKvXWy8WaxVhlv-)-d7uEDZQ
zRRD|-_6PL-P@$*OLH*;k6ER6+IsiT3C;Ah(kefePB^Xq4S4b6Jom$kJyi`1eGMLwe
zx+c!a?y?)x38!y-zb&5gYJfo%K4}+;LA^`xK_@TASy~a_TMxlYD&Lz=kTZR~i>~@L
zm$wXK0sMQfs=_7UzGHhi0)KGQUho2WNhCBL;2|NrcGx5KX`Wg;fQtt$cs<$?7?lxt
zGPF}?P)#rMXq+si3fT97
zhT*^F-XlU2IF$kga}4HI+g%u$nK-<-AibieqbUwpTld=dcmOwbzYk<)aH)bcrwG$!
zXTO5Q*}J=E%8LNpwbO8R_Eo1);0lY<@I0Tgl`_hDG%GJ)En8%NnDg6F&|KGLCmVL?
zL5>ci??2~y;moM=*bHCfnIAi_xNuW4HwIV6zo%_tyURB2UVg_zpRpDf}0EUU$YiW=3Up3%z
zF+H0%Iqz|EQQXzzkDs-#hUO|IXTH8pwG4ea{I>>v9kJjr~JOVu5
zCJoyrZAN>Kf0{4oUm)fV8y7&agDM6j`@mzF#gNk(X=7WLUU-hNT6>0TvbyN?3tu^Ljo}>d1U6eODRiv6;}LAW
zeBOWZ5+-PH?X4d9J{!#WY?ID~QaS)*Jz9SIs`ii=k(a-MIP!T9RW)K7RumYyBAWKi
zUk`w|)$#fa!8+g0FM^I5;PVS+s12`_cfP_FT+l+-?OBWJdLwwPa2;S8r@HGddh2S0
zbMz@LGC2OIzgkZHe4fJv1L{Oa2S?)vlFN1296gnKwMrScP=e3*C)WQ?fv$7Jod-*?``cU02t9V2qn%?G9INPTno=udsj#^=bc(7!5_s9HPytS3HgfB6!Lb^9Z%eu$a|V!99fav<}ByS(}6
z2r*1npuH~GvHDBXoQn(mDxH@v>!}u8afF-<
z9>Yaf$H#@-Fro@#Qr<8Z67+#?ixC6o(wgEO8fGt57sR^3_5gR>?W%Y{igKhI1==v=
zV@tSfK=vD7hgiTpcD+`!Jy;E8sg=XCeya*!b&XUG1?21y3*kD1I05*Kzu|J4uJ`{8
z|AyRlgjle2Py@iZ9YcU3@m132Qa2TVt9D2V@o7nN>1j!>2lSY|`Tx)hHNZ8*%PW99
z`Ji#wpnw4`u3YRh{O&@^Ng&v$ehvcRy7lkRacrUN>@z+r$pCvlIlwSB-5%g>4!_eA
zKwCxk46JX_IqyCgDwWcvX_d2QVUTAHPdF|cpTF-O@xlgLVw{g}EI*>(xN?`p3wf`q1!3Y2)_GW5ZL?&~=FAkN%qjw{TA2dILV+X%&-td+nL19XqDcV<+-7
zfRIbz-T{t5TMx_r`8iBc)`7J6)i;kmF;zgP!GFh}{m-!~(?hjjZ_mB~O*I;~+rpVQ
z`gA3=#nagJvf3i#TMk&5(t8&%%N|gug2Fd_pAq1?0D*!$B#5M#h1y$AnBx@yn&NX0
z_WgK3J#`s#)W1r|T0jhh8hH=>{hZ!!Q?@oCaam2#n1N1h{YX|CjYSQfcf^K{n?{fe
z0mfY>GBkREL6Udyz~ES+;u$cSqjAY~$iL5<4FNrM#U9n0RYVAg@pu8c3vke@PlF06r(!9akSdnyt8c#i&x~g=J+eY0hoax^*-V$@-d6
zE<3kgWtO&uCEgon6DvMV)tMgD{Os`%w`od;zTX}J4O2aSheM5x6`Rup_RilPnR6Ua
zQRRWpya#--`)%Dq+@i}e?5Mpb-fYMVf9UM9+oqTL77b{;x)1e4Kiku5)K(T*LnPAM
zjEYA&&vp6EyOD=Q((T!P^b3uICp}Fy^M{{BPPZf%pOAaPH83Xklih&zdkfwJq7p}-dhyZZuw|JxTkx8(``J%$`c%L4}EW-yjj1D{4KoIf>LD}d#LAdNG!&33PD
zW3-z)b&?e+CrLN)MyDsN*jlQX2;PL=NcT7XIMU9BdnoWJ_~{jpch)^7P}f7f>EA2c
zY5f@ZmhlhqpF9?~!j{lKR?de(Kr
z*pY2%E2--WU29!il$W#T>-m)7ZC;);^et!s^jE7rKA8V_k9{h_ILU=)du{So4wPkhihTrZszntAKxm-1Mr#K4ZOkL5EZBsm7(#%4i+*Vy=$F
z`gA8GYg-tD>O8wRyQ=*zJ=gNFcVM?-)A{rzRC5hAScFEBlH!Kdd#<1@u_s`xXw$4v
zha|uU1uhH+-Ot#`5E$a^oCnFxFrdqUJzyQ(P{FIl+6>y%C)PPBJ!%Wu>1}&>Sid=n
zKJPC4b-tbtzF#*)jUaJMAqdU1cJmH=yKcxZH%a<>Z$nR-(Hb>sz#Y?R4}D0K=;2I!
z=6z9jKnJ+AxT~d#IDX8;2ATyeBYHj{w&q|DVN>fUYrfBQkSe
z#Kr7QVKBx3w6*1L0hRbh@dePhyLI;-G)%{)_c-4s7!&&uOM;;=dW&}c9IGlxD7
z(Z8A(XL!^9}vz3r|*v-7r=N
zxxo*9)0bO3-Py-0=U%5xTb_ZPGG}%3=TLLGfkC;IwkoPp`k4dgCyOXt9RNO$r8!#2P9%@#8edVm5GjfY}s09S&B&jGB*`=-x%5oj$f`i?>T
zh&A1Q@{7NVF#&E5xRe7q4pl>ZL~*%Hcr^>5R&Zb(^(!7BAyZMu(-Xc7E>ozrb-P=<
zX+?K#MQZH;pCzfoITPhz}5jT4d6?|Qs=|zuS6Za
zt{Ko15{aYueW*ZsyBvfeq=Bv0@~%Z|IY7VTD}5fwzE^y-ToZFR05r
z>33Qg^#2CNV~Zuk%uRyp&m!b@VCRh=PY9Uf^ZSn5wuU$l)!pI-c?UG{xdnRn%$z**
z^2H6=^KaT%Ee5E6F&LaEQGl*?IxK9=cD
zpBUYy(KtT4J_8_>yv6~zXX8@S03DI@Rs9vK(4r9LJvJXx%g<}R1}2y{=vP7Gj{&>m
z_zlL2N``i`Z~dZz@&3f6yT?4Fj}iO$)W<|MU$G-za`^DWkzp)swmva|R<|Be^NAUV
z(bs1&K1@45Trvu*>%K`L4;A&{MloG@wBnxBfW~E=@AIl5aD!}*$VQVV1eN8FeuC|Qw>l4cevizr5MwAlv4S#<0xlJFa
z07{+54f1O3KE`+tuis%Uc-_@|Cxe#?^u-zKhTk-cTEW|s^LkgQaNNE*oS&EU1@6%e
zEkBn25QAtb>yw9D-ru$^Lw-$J$+3Nw1@=xgBKwJNKlXsHiXY4m!kxPL8!$G)7DG|)
zJ`Xr^cxIt50D}sKaUs8Od{4Zg&b4OQt9*+I6iHr=%a%{v;tnvraoYpE3g`Iz6&td(
z!IOXMpHJ~gcf6C2s{qy%@Y4>MBMP{ZW8_RL&}n^9EN`%+@%i*q+s+as_G1M&`;
z5?(IQz+3Epey`KFc7)>wZ4Jg8-`;vw%ZmZ+W7n5kOb4zf_jb`eFPxj^6Z_0BWSn>~>hIx*!+?V+E5fTDSKd!H@7^26HA3bqhc56qkS0f1ayWm1?JSZ`*jEeO$jJrYo4RfM+
zxY+EWzkkI-T$EC`I6tqjYQX~ARe{!C|6XI;JqBg-;_rUsE?^wS8+rxi;oMG=S8dY9
zfxG{`W)4IuDD;3tap3rSMIm|_yj~Cs9!}fBc=>zC4afzE5)cUByXCUO?g*A%O8VM7
z0js8l97tfP0kpO%w);pFyS5xI_Z<*^iYY6PbMFA6n81WM>f`SY+SA~is4c$LFAc*h
z4vaBv`Eeg|zXv`6pIe=GCDeaXjk2Lhv8-^~_~}-o9x%HG;0s<2^75|4Nd<0ym@m1xXTE1|6+W^&=H>(C=;8#iFg
zqV2m$CL%xS0^L3%WW5<13NAn3{TNaVbV8NJ>
zQqdF8ps=b_yH6~CFbs6JtI=Fg=4UAB_c%t_^0aRNF|p36#feZ$nzEVCBsGust!}|-
zaYM>$k<;s-7cX(=vaucpJ31OB7o*=_$2PwmTS{d)DJ{$m2L~VJjy{@=5rLKn;HS2L
zNAfMZ-kV5y0Zow|1A^%6$vjLVP7n{@BG;#IeH_b2B}MEg1JZkRotb*wkNB$!9Wc1gX5vN|;a~wDGmh}!q#9FGPrs62Lz?g`~
zvN7Zffq(%{mu>N|6qnz@*$4CH4ir=9C8~re!$hi=y%qBA^~E_A&f9aL4Pz_{80Lca
zM|GC{EsJ;y1D*Eqz5|M$llXE;r7OKAIjQT3Q6tIzdDS+&tMcFCs(>K+vj8{^Qc4lz
z>&?)poOk^Az1tSwgg$EXM-NJfHJ~qR6!^l&mQ#bA<_o`e9YMT(bY(`AW%6}hj;F_6
zm@lDUo4ZnavTL8q$3qC8_#kAa(GzL~xO2@g0I>z&V}-{_5DuqL9SpRs{nflauGfF*
zXY#fEFKK7{?Di7U|B80tXE))T{#Uf~`K&$2|0V6f&r*QA{}t_QpPiq#|1aBOs^Vy3
zCJA!Z4R+8`G$#;76UdlA8MXLV_U*=v#LgkmEYDY&;Z&6DrzFqz`KRrMK@+d@UI}_a
zQGAZ)?K=#^kU>wtxZLP7L6WRFyiL4uoSsVV;zk&~-)~26Z5wCOJ`7`$0dAqXWbR7^
zI5HRxa`QZjI#|}pvcTWw_W#w|90K_Nw0-dZ|IVPusl>{_b-Vd@cxd9S5fqh8E$=IzI20dd>I5Wf$f#T9+kOhK_FD;JxfLJkd2i
zBuv_v6A$%ptE@ac_G%9+quR9zH1e$F`Oqb!;-$+PuD5mw3W^#6xM{m}&dv2K@zr{$
zD{Amw7gb+-cGcHhW^{J!nd7qdjKR_jiky-DoL-fBZRRwnXx;2?OJB&~T77U$O!miX
zJ*U39YO6*XJ@Wg0)*6;d$4W}}91vF7<94);-H8l!2@g2b0Q%BjZ!|1`XW408i6c;_
zWh1-_p|x)3Pl6Uj6QEqN>iOsxhCYu8a0(pCRS{~|9DJMvf*cPW(5jF7g9Cc@S=`Rn
zP2wrGG6;rabo^cT`QzjaBg!G;$Y33607-?!hnZsw)y6JIr84Vo?
z#s<9o#fdCBb{%Ip7#l<=!o+<=(XrB;wlp$0nA8k#Fdn@F7Zcg!e!qjOnt*|@`r*K*
zB)e!Le?V*d%}^&Sx$iD=DLl(-JI4~Aauo{t1FpZ{WnB`-uQ3Ml9iapiNi`kl3f)_Y
zMZ9nn`J}sswhN$#wmdSQPRN>oKEZsMf4EBg%oHSO3;aM9@}^0^b)YMJ2IOM}gY(#B
z1(K|{G>T(cl%%?qpRRCF
z%YPO#D17*=oY~^QLdA$s4>U2PVGNT
zjAg!<$vHvvF0tk!hdw0~ouzH<-2sf
zyEEKwLH*NyWqAT?biVEddrF_K4akza{Un|b|NOfxhobSoBC#LT{F}&!IF3m=LG^2$
z7-K-t-ar5T>#MCYb#3vGx~`NUGLeuP309T)&%c{(&oeW-)H|O*Ln=<>B9GO1p3!=`h6#1KH4kE4D-JDg7)lE*)Uw?_t&bp#
z^eC#A1Av!x8)L9sjJ7+H_e11ZSfQnbk3|?kuz2&J6OtA}bZM>1t2`HcUJ?w&S+c{s
zK?}Z|7FvEg!}yRff^~%*FdWB;FznxN=D2%ncYxPR*V1E;GqKUud-8sYoD)@oW`cRU
z#|W{u4+_ba+KPg%E>ExSB!e~Cq3E%1Cixv)+|F|0l%CZvLT~j6p%4i^CL&)+$
zyYeqEFq~vQ{qA;_wz)bT=lJ#U)hO_*Mt%RYnw2`
z%M`C4=E`|Q(t&QzLn59u(dU=pvEWQJQkcv{!}4^F)H>cOXTjHGY%rCHviV7zsr%V@
zG5RV**7gfeQX!I%ix48$
z(xx0SV7LJb{XW(u*iT?17K74Oz%WD%AnU<#S#iy#P%{j61+DZ=$-aAnp`#l@xV?J;
zLpAF(xjc0e1(C>Gs^I`jNkJ2KtohG9-ZOB6l0Gwk|v$
z0j|7=SB|1&j-I%BiGfc#2SL{=HpHA4_6nur)j-kJK#~z1m=RF`?=hnm)*oHu&T;DD
z0@vC&uBA$7*tvO}bK7kFSBs*v-m@Bg*>EB|D?8Q#pYTFpV{&G
zXsaoxOR92!A2=Emd?w%n2b6m*5(!B-25=$EWpZJEJ(Q336osR@kYFvD-C^lpXXvw_
zEd(7yc{wv@5B8d~3B`3Wn#US>5KjBXQc@yO7Y*hR1b4cuiIku#c=6Zk#LcjVkTqaY
zPok|owZPv={;6Nh!?m?c_suNqoqvW+D#Cqraja_>b*m2w5OzgTfrXfH-b0-Nkf;PK
zA=PAJA6a;tj2Z!)GR6S&L9jG(LhJ9xLs;@SVJM2J*EN4rVA7I1<>dMvTmazNZbU3>
zuBM@qitI&E)m7k(aOaF@dSbe9G;)So9_T_|IkeJh;`Df30zO=!TYTTL6D^So&rM!yx}P$Ms?Qo!zzCB0-y}uaEz%TsqK3RYsF&2
z8AzhLVQ)(q!0aV8oP
z8FjwfJVK11+C8(!yLhDIo$#DZT>~;
z9JDz9`L(}a2Z&GYEq{>)d@iAlCP&GjZ)M}!_&X2a8i&u{13CmrJouga?YQ}Vs5!ky
zg2t)TNh8NaU{+Xt_YN#mZ9ex8Ba)RhB!E?C$=vQOzOK`>*L%q6U)VHkHzs2`3UO+j
zxs<$&rJn7JHq19c3V&9TkN?0y;|i1T!!u6Y}I^~~0WQLlvYoc7{yp+W1(me?qkgLI?s8`cA!
zgT(tO)OY(7VH-Ws_!qtH>xBLR5EIb`f7Lm>Pk0m9cnQ;=@-Eg3AiYHsEJk#t*cm8I
z`5>~gU?lS4bU{m-vc@UM@zdc*)fb4)CPo329#EUyr8
zE^!HXmW3dqFH!kcs?B|(0BI?I6`teo&d}_
z4zT~4UkMZm7q?cZDg*!aO~8
z1QA|=pPG-*PyNuD*zE!q_bq2bPc-XvdOg5prt=d%Pk}qjhKAUmE(}R0S#v_^G-tonH
z9G>z#tL|BT-S@B8_^^ibSZVr#z22t$d^SyXcrm_E`ZqdDwfWE=d$G?9F_6`+o;Z`%
zg!eVwi=C@wuYOLKvv4T|*=O~0d2us4b5DJ7n2t(t_mOsF$uULhy5;Vzb$eNUSKO}3
zYWw#`TMI!gwQvJV`tnF5wt@2+-ay$x+@d1=d9{%a?twX{R{J&^xwt6r%dLIf2DXCt
z5Ol>u_UvfH5IG0dSOa=v`T5OMeBKXA;Z;Vj%kdTR@SclKhjFzl9``Vgw)rVk%CZGn
z8jBtufM2v1t5t;Z*mG539oyO7&Bh%y0LSNMb}*H9^Yrk4a6LuM>P0Z_S+GRq<573n
zNh46;S=V$XpV#`{tWUFHYT=v?X%CqMP7F{N1jaU;CB?IXwXYfBgcCUg7{tJsaCE;5s
zP%-89mMN*4e1k`wZ|+2HZx!aeZeNu1q7p8GX^?k(*x_=fqUIGG6IQ3zFPf}>WWLuf
zJ0z%-0`PSG1YAf2FzWsiWO#3b9QYi7)Wa~sb7c>BE(Sg-*+F9PA(uDw+wN)}h^4`h
za?}7=hRi#zT0LqL$(j2)69hp6yF$4VYU90T7Wa7u=aW&+p^?1t%H~vxR%I9V!?Qi#
z^voKV`55NeUXJzS;lQPhi|ah}$x;o**`_MUffb5lKOjKnCdo|L>0-e9<~qwhE3!}*
zOp%hIQ=`r12AXkH4ykS6)!d+@Y``s|_j{LL%e!UVE=TL~;N+bLgD76FOw^%$KU&O&
zH|@2~Z>ufs7Q1y;C{NTG33VdY&saP_{9OtL^pQoMq|O@#Lz3L8jWxkQ?ht6hlFPKt
z13+W@Ybs|7(Oj@0L{?Xe0PYev5G0-l4fMLZ@OOa%M90cq@*=)=V!rcv)93ftCA4Rk
zjpX8W3!wi}cEzp8uIbG@-<0D?5y1L%9pIaJbcN4%)%XWyntNe$$$nflbm&PbRV_%@
zo=aeIrnS|(fa{^ibPtvySHdy91I}MwYR`_gM}TjXi@3}Myl+KyD66z7XZS2>7HHHQ
zl>}kOP469cxkplfv*VLI@@U4`*yTY3&GLZDiU&W{z)%%|jOEOj-kWcK6A@vrz%5rP
zVT0w>#ndjo7m5P7Anpyp3vw-Q(x`B7ZfD&^WKwxiIhjAJyfQLf4{pt43>To$!J6yA
z$|9KnOuc&6xWvNV;|lpp-_hI(q&0C_OAg8J>H_p_wZi~Byslt9-Qjvdj~K%tOu2q^
zF(3xa4C8mZ3lXh877o(!aF$#7^?JUtSC(bR?3|s?@^KD_@Yr|f1uQNchg^V+js5dQ
z>v_PJ2+C*of-=JH!>5yE%noL29HxmybFhYqJQYm+Ho!d>%$ly$a{>y=*l}a9cqBO0
z6``W#yO6&9=+NbHQI3%w4D^hoZayys;7HT}A14
z21O;UQh+Bf*#Usb&AjeRfdT5aG#zp7T~|!o>~YMlO;$D)=uO|N5v43S=m!t$dvx&A
z8Tk0%SUisD^?XQ}GrT`@Yyhr9)J5@#rZqjC1WR}bkB154ZxAn1h!;97ha^Y$Xb${0*IW+3;-5SCTreq;J(?G2-WM|Lz;V{_^UwXIA^A`P19mH~#bk@b>U-hBxoO{a=46vpW9gP2>2_P5*TK
zuYc0bU*I`@9sMo?6fC~IyR3fxf^KjA@^iDbU&X)rb^i;x{q)P|mqYyPU;gqJ?Em@&
z{rvLFyBq4ze}gNR>*V$8FO%QjKX1j6mPIPKl3x?Q?Bie1Kl|u+$M4_py+arK#h-oW
z=a0eV{-NPC_9HNG|E+xgyZ@K8i<7_owesJkAN=<;E*s<349#E4*E0+K@~?jhOTX{`
z_SesMU}x{3^MCzs&-cH79Rg2-`tXfk(62N9(!HksR}4%n{P&meRSs3Xz486ye|pl0
z|KUmdhtt)?Wf%9m-%lC&|8O{s#s9Wre$3myzBA0fzBAc7?*7mfjQU^L1@=GPg%3ph
zSqDC!$kDGpo!~n%#{Z2Y*!bT_GQs{Wh$8>rs5pvuILR>q>wjYnIQ*C2U@QVB`tMog
zKV!71pZ^~CeN2=8cV-=QbK0H7=L@3X<@QL0#}oY2K~$74$Gr>wj;ypcV9Ifh+-
z{Tx@^Nn73=je46W<>OS0#(l~kM9<@L@4%N!(8qj1K)EsKou9>yM*k$_aope!Gab>%-C5*nZKSV7T6zs8n4xXE{nsv3+@c9)?d5NBEUWUWkB
z6lR_K^|HUO1KyV%WodbaWQ^#->2E;CI&bI0wjDm(hx`0N;Q#m9|6cpwYybbfwu^PE
zc~AwbI^6TgrnliC9qVx-_40@rhpUVPhQXr(8Hl9blZXf?8DHHH-X!$QY#$;ilIWa@
zxh10VmHk|@{gY*13`-#lWg?6@A=ZEP8a@BV+y8H`!58~a
zxBuT>qldWt&yUSNUt1G5{cjkXe{*f%+W(e$_m9`Mi2iSxcmH&4pqgLpXx%K!j+anA
z2!%|Xv$-X9ag9x0_*k#tf8;s;@S9GDZ~fZW+H3|5w`7k}MJKPx1I_o|55tQ#@%y`U
z>+nWtJHIs-Vd?>R!pXTo(FSewEe%tt@jFA5c3fAPhxL$9)T?SuxyQ5BwF9siE&^M(_-r!lK9
z&9>_yR(5|`RUd9@9^SOj>D|*hRHYm3cwFK%!E6rH^!&hg?{%Tu0#&QsP9N<@B0M>C
zeJ1Kauhrm15$!mJ0T8X(G&f}hH|VkUDmf20wy5bfJTcYab7Xz$(NB*n@OzjS?wHu6
z&WcgZ4;DQ+F
z7p1E^k4wDY)W*R3V;=)=pV>PYTS24O!9685&9G1S4pDL`laLZ=1AzI~(`ZK%r=3heBV_c(d^4Y3lFb{Sa)nzD0DhcD~lQpcUA4RA2FiI=Ys4ZwF#(
zjSVQ+gw-hBBiUchD}Rzwc^cmywsvBR(_s_S1KDi&EDMS^C?$`?-f_uj00Rx=VQ{
zuHl4!6jha)PwIpI96H92P74^x1hD4HCY&RfhA&fj(Fa-dQfuhQLi;i_qW+w0Inyyh
zODMridV_`uzkYEKZ}sgpvs~uhJ3gP_J6-!$V?3~utNUaBvR@sQa;jjzQM1fH6fnik
zPi^uOFh@f5P7!c^cV3fr_@tM`E6zf9s&6sv^d6W4F|cp7_Qr)~lOeTCDjIMtHj=s_
zmorP&Jle$QE$J4@G;O4II^25G4(+3_?(FKO&-c4fTMl5oTm!iFU9T?7YifsO!@5bo
zi@{Zhxr(moizqAT{asc94(-`YO}T9rSbDC0KR-M;8oL@x^~@$8+yl&N9}eNy-cIV&
z<}53!1d(XC`0}^1L)AmhVxu}p>wwtOuap|#(q7hfT_TI?YtuC
zMyq;7;lquz)nh%2!EcRsH0unn8x-cJ{}O}aG8+4^gI(C3RLhH)xLBkth*g_^;Y
zh2uQ9uAbV*j(<{_3g$kmIOMQ=c9LB_Pj+xy-Vfy}3vdq9b`Na5Q+;?xKH`nKeQvq4
z`L_btqP-pJ-WQ|^mT$7^=Aol!Fs}+6tc`iA)hEX%6r3XiyLs3R_Wc9Ud#1a-6X$;G
z>&wfBw*K3`4YqHJ;F^1H6t`GnSHLWL^p>8^o!G)$%Xg9}Lx{~zq#BLs4!~yWCQeYh
zgwWc;6MI^PCb`p1k?8x!1Cw2S4T#V6u|Ex>y{{Sg*q&+UlZvV-~`Kgo7dO7{4BPdgT9PV;1hm)x?HaZ=g7UviK$7>J@4ya
z)}M#UWj@Bb9A@$b=O$+yH-9a>_xdxgv%VbfCH749*F^z~7Qe)#H^jMZgMTaV-?OuH
z{R6B?Xb0V{ui@9$OYAC^mX>$S=xYSso?s69T|CQ${l?RI03IILqK_pL3O#+Of{RfX
z^wEn(z?^5D=mhP$l0ko!;;<0I4dLKg;XKmX?S}+M`B3agWbSXhH
zLN~NVI3ncVf0Tl}Wo7a7*lmc~5iozIS`TbdKL!pn+i-
z#wd({-+2LF6+WTyDn;p>v_H0
zmf;~E$OUDC$9hgG_K{tI-Lb|#9Ef&Zv%kQhTeWt+h;CqIT>9HDfMBQ3Or5Y^G7tB(3Ir@rx?Z
z=FKrhehMV`cD5zh*447yhGprO_CYo|@O1@W<&;|mU9_!Nr3
zkTO}jvs36_;ykwPLRwMdGJ^FqRX_Lu_{j|T#r-=z{0{Wr+v4Em&8==x^l-zM`vV=<~btTu|WwTd}X-2iIJ9jV|uGBgaE<
z8|@!F>#g9sA
zyjl|mj_);xm;l<4Yx$aie{*yp0jA@61wcx<3jKp!576x>&|g;j{Q+DCw!p3V_gDYt
zc6_mocunh#>;mAXaKdJ{
z3*G>*QiAs501&u(
z{Khjp(A9VoH6MYwGZhwv*M0_;43A%b@E2aKP8IN+?Ls%1egypyd~w7u6V8%%z;CRr
zuT>?~|D%s8r1b&J<4>aTr&x|+fLkd}K50n}`0uZNoE`hn9gvE(K|l8CckcO=qZo=@
zysa~sdz|?4K;OK6w?h$5)tOriu{%$i7`!3RSMz9se&83zNcsy{w$#;LK}QF2qkcxn
z3qGm*L6=109qG^U1~?I5ehYA9C9vdE@kvRHUlO&bhI@#SHNZCE*XOh1{iw9qtE{vt
z1@nFnMd>l%wzUgwWosw062-F-f^}{NSm%UYT>wAdYY))L*gN&x^C~|yiawDAYQ)oh
zHXx4Kbq#&NiVmz%0@$y75}dQsA%EJs3jj*znz>ub6Rgu#AAzsfw$Yw;z%K%WkimJf
zkNd_&2eK>Qt_?F^K^qm}SY1oZ%lEknJAZ%1VO2D^Ue9I$zB;PdHaw2nFMs%ht*Y+h
zC};KDu;)_59{5lI7-k;8LJa-JynX$UqnmN3i_+UyTEOl1e!ad9mw7)BJso;&`{4CB
zbDkLW_C3&Xwt~riy#At2HJsB98&a&(9w@Tg6e4tTYSI3F$7gbX3m-Cfa==6Mi0^g28
z8`1r6kD@Ty#)n*pI(70`yY6ig&4`8^1o|lbm0u5!2qw~dwzJGVYsMVMFNkSb9JAcH
z>^|xYvJG(a)tbm{TM5Y{z>%u`)7C?O(g8S;h)hIBUo7Ehu~_I^8{sG`HBH2gFu7P#
zU2;90S#ud;Ir)_+$WKd{R>0_ZCE1xcDIDc4?~QgN)Mo$8bG;poHMMU5!dCn?1z<{8
z9P&m{vz||vlWWD>oiN6HrMSD1zaj}am0x{w*e;uK?5QFbro#I2&o@}h9`uQl?RHJr
zfVS)!s$C5SSHGfDIHRu+MU(`k3@}~FC`00&jzC
z`UJ|y>X(v4FL3iY64zEX=nj^v;6DTS7qQIxcBxCy@AY92izZmvZCUb(LoHp01)ZNV
z6#&}{IHJ@n6~I$<^B(O3)L*89t9`&nsE?@^%4^aKn7G_))O0Be0Nq-brmGbmYDSjJ
zVkb0A!n}zKSPE6!Vov4#w!7fbmt1c0*W>zlmB40%H?gQa2>5Hw{jLwUx?XR}Dq?uZ
z#a%hkwYRLIBgLSfZ_sab;60#z{1>)5(!x_IV;`_u$W4KZjcM_+MY|E?34X)1>68eB>>M@F6M7-
zL;x73dJL+disQviLdt@=nM3#h=l$YM5Zl9le>E}5J};N2G7bgsnbCk)7k7#7?Yi#1
zI12A&3w=nM^ajkY$f^&(N9!Y1UZF|fpO3mcf`8k32e%5H;COz3FLD5o0`wEJ!C3BG
zOnn}Nv{9g@u&>{@;k{j8J1%|aVB`;`wDqTwW1FRIIdHze{TJkARb7@@tGiB^-U`1P
zzuK3loRmbs;+&-^+JSSx47%sKSn}!`@!Q#;G@6;11pB-Jn5Q!>xXvhZ*Z{U5DwEd-
z8TRY>E*3%=$j3G*NL!ZAxc1dkfYSzlcNN0GgnAJT4cec|TyKDm`)hqq(+Sh^;Ph=%
z6$_wEYJ!KKBqW*jqur^~t0^m!aKf_TwE9O|U7^4qc<*^x8x^NmHs
z2O(+kZV08yq!nVT+!F&i7U;h^{OLP~Jl=D~g}P^biOG0s;JTLb@i^AVH&&7;gGw}m
z7xmt+z6q|v%E>H)(Z~wz^}eD0`ge^A>cn7O-ntKJw?{C&-O+4%;=KE3=R3X_pp>C@
zqD>s7K6ysL0yh!MB7B5yw8NuKGLa9~s@Klz;bwz9yFH{)w5`qdTpr=(ai})~f8Yep
zwA-0Zj>T0tJsNs03moN?5n#kNqZbn8Y=+Kej;jEdYOmLmPtoWa+RekS!`mbW
zy$Fjf+s*@Gfqhe<&;YPI&W~O@B$rJ1?9rU}kM8|5FM6zmcql+OF(A)y<03IlgwzqKFB0V`G}
zXR9m)A6fG0MMAd)=smuFeBCuDxH!?_i_OPdZKOUIP--kcbK
zv0AvkZ=h@T`%3V>+l4i`CPDco<4Ev7IYZ%I=S$AFcfveQKRF&$irxn-#~1OW|I(0{
z1#ZUKggp=B`U&;tVIAN`8rYvVkMX0$C^pPkXGU#ivncSbO1EvLuMxci!*CN-aqb2v=`h
zqt%X)4mUsl6!Jxoo40S|WJP)rMJ1MR;V(Qt-JskAFf8fRhvWF3(aU@0U{Qh)@Ul0%
zx+sdd*a=@;ck7Y0Crg+BqGYFx(f-FsIdl&4dEvY%m$;oW&Ew9BsF
z^v{DW1-W})fv*MYcdPf<%8my~8$i>i0c?|c(5b1cMB09j*B@>Oqh7M!@@!UDp1~n$hW^L~YF!8bOI6jwM8ODm8n`uCe
zPOa%ky*1ePbC>;d6hMolYsgXj?Y9czF0LdTKs(Jpd~RqB0@TT&{RB|R98j$^pnSVy
zPdT)n5Q}He!GW9(jC~ILe~JPBudZii*FB~^MzC!dd4zWUM}IL^^Z?dN$~#cIqNXk}
z0FQO-Ra#Mq6Y8I_2Gq%Plp=09h)pN&hci&)k7ko!tQ;%b&aX|Zc|vWBgndY#nfg!7
zYjQc*|K!_!G9~DrJ1#5U8yCOdfia?!vdTo%te+?81A>35eWv;hftFnRPQeJMt?HNx
z);R5wf`@fqe0T*EOh@6^1G{>D^AG2AKpOYh1#m}^~0kn4WRuB^2h&f_K9YBj&iefZPn
z6Q=U8Xw#+=h4_%)3qm;ft?Y4Z$I1+U>S^4PcIV
zVjN3|1?}y#OI-7s9{_$j|KIq^heqZ**23vf2!(9i{;t0}I-~@W!Pus7A;TDAM{MDfmyJLT)Ldb_
zT9u?lvafdhAWo=O?Rg;1@-*t&`g!Azku2ZuD7yoAgDvV_RKPi^f(&`#
zRqspwCm#d`J6`^$jviMiFJMqNfq~IRMKT&0Gjx0JMcAgqGMX(`KlBOaq!X|RpCBhP
zEdZnL0s7l*^m^5y{&*?x-