From 0a210d81b8a02b83029bd42ffc976c96215ce88d Mon Sep 17 00:00:00 2001
From: wbonat <wbonat@gmail.com>
Date: Tue, 29 Sep 2015 22:03:25 +0200
Subject: [PATCH] Solving push conflit
---
Examples/Examples1.R | 290 +++++++++++++++++++++++
Examples/GLMExamples.R | 207 ++++++++++++++++
R/fit_mcglm.R | 241 +++++++++----------
R/mc_anova.R | 50 ++++
R/mc_auxiliar.R | 29 ++-
R/mc_bias_correct_std.R | 27 ++-
R/mc_build_C.R | 126 +++++-----
R/mc_build_bdiag.R | 10 +-
R/mc_build_omega.R | 32 ++-
R/mc_build_sigma.R | 262 ++++++++++-----------
R/mc_build_sigmab.R | 57 +++--
R/mc_coef.R | 121 +++++-----
R/mc_confint.mcglm.R | 18 +-
R/mc_core_cross_variability.R | 10 +-
R/mc_core_pearson.R | 7 +-
R/mc_correction.R | 8 +-
R/mc_cross_sensitivity.R | 32 +--
R/mc_cross_variability.R | 20 +-
R/mc_derivative_C_rho.R | 17 +-
R/mc_derivative_cholesky.R | 21 +-
R/mc_derivative_expm.R | 20 +-
R/mc_derivative_sigma_beta.R | 32 ++-
R/mc_dexp_gold.R | 16 +-
R/mc_dexpm.R | 45 ++--
R/mc_fitted.mcglm.R | 8 +-
R/mc_getInformation.R | 33 +--
R/mc_influence.R | 151 ++++++------
R/mc_initial_values.R | 102 ++++++++
R/mc_link_function.R | 160 ++++++++-----
R/mc_list2vec.R | 40 ++--
R/mc_matrix_linear_predictor.R | 12 +-
R/mc_pearson.R | 38 ++-
R/mc_plot.mcglm.R | 67 +++---
R/mc_print.mcglm.R | 57 +++--
R/mc_qic.R | 29 +++
R/mc_qll.R | 37 +++
R/mc_quasi_score.R | 14 +-
R/mc_residuals.mcglm.R | 26 +-
R/mc_robust_std.R | 24 +-
R/mc_rw1.R | 40 ++--
R/mc_rw2.R | 42 ++--
R/mc_sensitivity.R | 27 ++-
R/mc_summary.mcglm.R | 97 ++++----
R/mc_transform_list_bdiag.R | 15 +-
R/mc_unstructured.R | 28 +--
R/mc_updatedBeta.R | 12 +-
R/mc_updatedCov.R | 45 ++--
R/mc_variability.R | 24 +-
R/mc_variance_function.R | 314 ++++++++++++-------------
R/mc_vcov.R | 10 +-
R/mcglm.R | 119 ++++++----
data/data.rda | Bin 0 -> 25969 bytes
data/mcglm.rda | Bin 0 -> 271560 bytes
man/anova.mcglm.Rd | 15 ++
man/coef.mcglm.Rd | 4 +-
man/mc_bias_corrected_std.Rd | 23 ++
man/mc_build_C.Rd | 8 +
man/mc_build_sigma.Rd | 2 +
man/mc_cross_sensitivity.Rd | 2 +
man/mc_derivative_C_rho.Rd | 4 +-
man/mc_influence.Rd | 23 ++
man/mc_initial_values.Rd | 39 +++
man/mc_link_function.Rd | 4 +-
man/mc_qll.Rd | 24 ++
man/mc_quasi_score.Rd | 2 +-
man/mc_robust_std.Rd | 23 ++
man/mc_rw2.Rd | 20 ++
man/mc_unstructured.Rd | 18 ++
man/mc_variance_function.Rd | 4 +-
man/mcglm.Rd | 9 +-
man/print.mcglm.Rd | 3 +-
man/qic.mcglm.Rd | 21 ++
man/residuals.mcglm.Rd | 4 +-
tests/testthat/test_mc_build_sigma.R | 26 +-
tests/testthat/test_mc_link_function.R | 2 +-
75 files changed, 2245 insertions(+), 1304 deletions(-)
create mode 100755 Examples/Examples1.R
create mode 100755 Examples/GLMExamples.R
create mode 100644 R/mc_anova.R
create mode 100644 R/mc_initial_values.R
create mode 100644 R/mc_qic.R
create mode 100644 R/mc_qll.R
create mode 100644 data/data.rda
create mode 100644 data/mcglm.rda
create mode 100644 man/anova.mcglm.Rd
create mode 100644 man/mc_bias_corrected_std.Rd
create mode 100644 man/mc_influence.Rd
create mode 100644 man/mc_initial_values.Rd
create mode 100644 man/mc_qll.Rd
create mode 100644 man/mc_robust_std.Rd
create mode 100644 man/mc_rw2.Rd
create mode 100644 man/mc_unstructured.Rd
create mode 100644 man/qic.mcglm.Rd
diff --git a/Examples/Examples1.R b/Examples/Examples1.R
new file mode 100755
index 0000000..d1b6803
--- /dev/null
+++ b/Examples/Examples1.R
@@ -0,0 +1,290 @@
+# Set of examples 1 - Simulated univariate models ------------------------------
+# Author: Wagner Hugo Bonat LEG/IMADA ------------------------------------------
+# Date: 07/08/2015 -------------------------------------------------------------
+# Lastest updated: 28/08/2015 --------------------------------------------------
+#-------------------------------------------------------------------------------
+rm(list=ls())
+
+# Loading extra package --------------------------------------------------------
+require(mcglm)
+require(tweedie)
+require(dplyr)
+require(mvtnorm)
+
+# Setting the seed -------------------------------------------------------------
+set.seed(2503)
+
+# Case 1 - Linear regression model ---------------------------------------------
+covariate <- seq(-1,1, l = 100)
+X <- model.matrix(~ covariate)
+mu1 <- mc_link_function(beta = c(1,0.8), X = X, offset = NULL, link = "identity")
+y1 <- rnorm(100, mu1$mu, sd = 0.5)
+Z0 <- Diagonal(100, 1)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Linear Regression model -------------------------------------------------------
+fit1.id <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ data = data, control_algorithm = list("correct" = FALSE, verbose = FALSE))
+summary(fit1.id)
+
+# Using inverse covariance link function -----------------------------------------
+fit1.inv <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ covariance = "inverse", data = data,
+ control_algorithm = list(verbose = FALSE, "correct" = FALSE))
+summary(fit1.inv)
+
+# Using exponential-matrix covariance link function -------------------------------
+fit1.expm <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ covariance = "expm", data = data,
+ control_algorithm = list(verbose = FALSE, "correct" = FALSE))
+summary(fit1.expm)
+
+# Comparing tau estimates using diferent covariance link functions ----------------
+coef(fit1.id, type = "tau")$Estimates
+1/coef(fit1.inv, type = "tau")$Estimates
+exp(coef(fit1.expm, type = "tau")$Estimates)
+
+# Case 2 - Linear regression model with heterocedastic errors ---------------------
+covariate <- seq(-1,1, l = 100)
+X <- model.matrix(~ covariate)
+mu1 <- mc_link_function(beta = c(1,0.8), X = X, offset = NULL, link = "identity")
+Z0 <- Diagonal(100, 1)
+Z1 <- Diagonal(100, c(rep(0,50),rep(1,50)))
+Sigma <- mc_matrix_linear_predictor(tau = c(0.2, 0.15), Z = list(Z0,Z1))
+y1 <- rnorm(100, mu1$mu, sd = sqrt(diag(Sigma)))
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Fitting using identity covariance function --------------------------------------
+fit2.id <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0,Z1)), data = data)
+summary(fit2.id)
+
+# Case 3 - Longitudinal model using compound symmetry ------------------------------
+covariate <- seq(-1,1, l = 200)
+X <- model.matrix(~ covariate)
+mu1 <- mc_link_function(beta = c(1,0.8), X = X, offset = NULL, link = "identity")
+Z0 <- Diagonal(200, 1)
+Z1.temp <- Matrix(rep(1,10)%*%t(rep(1,10)))
+Z1.list <- list()
+for(i in 1:20){Z1.list[[i]] <- Z1.temp}
+Z1 <- bdiag(Z1.list)
+Sigma <- mc_matrix_linear_predictor(tau = c(0.2, 0.15), Z = list(Z0,Z1))
+y1 <- as.numeric(rmvnorm(1, mean = mu1$mu, sigma = as.matrix(Sigma)))
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Fitting using identity covariance function --------------------------------------
+fit3.id <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0,Z1)), data = data)
+summary(fit3.id)
+
+# Fitting using exponential-matrix covariance function ----------------------------
+fit3.expm <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0,Z1)),
+ covariance = "expm", data = data)
+summary(fit3.expm)
+
+# Case 4 - Logistic regression model ----------------------------------------------
+covariate <- seq(-1,1, l = 250)
+X <- model.matrix(~ covariate)
+mu1 <- mc_link_function(beta = c(1,0.8), X = X, offset = NULL, link = "logit")
+Z0 <- Diagonal(250, 1)
+y1 <- rbinom(250, prob = mu1$mu, size = 10)/10
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Logit link function -------------------------------------------------------------
+fit4.logit <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "logit", variance = "binomialP", Ntrial = list(rep(10,250)), data = data)
+summary(fit4.logit)
+
+# Probit link function -------------------------------------------------------------
+fit4.probit <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "probit", variance = "binomialP", Ntrial = list(rep(10,250)), data = data)
+summary(fit4.probit)
+
+# Cauchit link function ------------------------------------------------------------
+fit4.cauchit <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "cauchit", variance = "binomialP", Ntrial = list(rep(10,250)), data = data)
+summary(fit4.cauchit)
+
+# Cloglog link function ------------------------------------------------------------
+fit4.cloglog <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "cloglog", variance = "binomialP", Ntrial = list(rep(10,250)), data = data)
+summary(fit4.cloglog)
+
+# loglog link function --------------------------------------------------------------
+fit4.loglog <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "loglog", variance = "binomialP", Ntrial = list(rep(10,250)), data = data)
+summary(fit4.loglog)
+
+# Example 5 - Logistic regression with extra power parameter in the variance function
+fit5 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "logit", variance = "binomialP", Ntrial = list(rep(10,250)),
+ power_fixed = list(FALSE), data = data)
+
+summary(fit5)
+
+# Example 6 - Logistic regression with two extra power parameters in the variance function
+# This model can be very hard to fit and require very carefull initial values and tunning.
+fit6 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "logit", variance = "binomialPQ", Ntrial = list(rep(10,250)),
+ power_fixed = list(FALSE), data = data,
+ control_algorithm = list("method" = "chaser", "tunning" = 0.1,
+ "max_iter" = 1000, verbose = FALSE))
+summary(fit6)
+plot(fit6, type = "algorithm")
+
+# Case 7 - Gamma regression model -------------------------------------------------
+covariate <- seq(-1,1, l = 100)
+X <- model.matrix(~ covariate)
+mu1 <- mc_link_function(beta = c(1,0.8), X = X, offset = NULL, link = "log")
+Z0 <- Diagonal(100, 1)
+y1 <- rtweedie(100, mu = mu1$mu, power = 2, phi = 0.5)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Initial values -------------------------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = c(1,0))
+list_initial$power <- list("resp1" = c(2))
+list_initial$tau <- list("resp1" = c(0.1))
+list_initial$rho = 0
+
+# Power parameter fixed -------------------------------------------------------------
+fit7 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", power_fixed = list(TRUE),
+ control_initial = list_initial, data = data)
+summary(fit7)
+plot(fit7, type = "algorithm")
+
+# Estimating the power parameter ----------------------------------------------------
+fit7.power <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", power_fixed = FALSE, data = data)
+summary(fit7.power)
+
+# Case 8 - Inverse Gaussian regression model ----------------------------------------
+covariate <- seq(-2,2, l = 200)
+X <- model.matrix(~ covariate)
+mu1 <- mc_link_function(beta = c(1,0.8), X = X, offset = NULL, link = "log")
+Z0 <- Diagonal(200, 1)
+y1 <- rtweedie(200, mu = mu1$mu, power = 3, phi = 0.5)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Initial values list ----------------------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = c(1,0))
+list_initial$power <- list("resp1" = c(3))
+list_initial$tau <- list("resp1" = c(0.1))
+list_initial$rho = 0
+
+# Power parameter fixed --------------------------------------------------------------
+fit8 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", data = data, control_initial = list_initial)
+summary(fit8)
+
+# Estimating the power parameter -----------------------------------------------------
+fit8.power <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", power_fixed = list(FALSE), data = data,
+ control_initial = list_initial)
+summary(fit8.power)
+plot(fit8.power, type = "algorithm")
+
+
+# Case 9 - Poisson-Tweedie regression model -------------------------------------------
+y1 <- rtweedie(200, mu = mu1$mu, power = 1.5, phi = 0.5)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+fit9 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", power_fixed = list(FALSE), data = data,
+ control_algorithm = list("method" = "chaser", "tunning" = 1))
+summary(fit9)
+plot(fit9, type = "algorithm")
+
+# Case 10 - Poisson regression model --------------------------------------------------
+y1 <- rtweedie(200, mu = mu1$mu, power = 1, phi = 1)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+fit10 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", power_fixed = list(FALSE), data = data,
+ control_algorithm = list("method" = "rc", "tunning" = 0.1))
+summary(fit10)
+
+# Case 11 - Poisson-Tweedie regression model (Neymann-Type A) --------------------------
+# Neymann-Type A
+y1 <- rtweedie(200, mu = mu1$mu, power = 1, phi = 1)
+y1 <- rpois(200, lambda = y1)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+fit11 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "poisson_tweedie", power_fixed = list(TRUE), data = data)
+summary(fit11)
+
+# Case 12 - Poisson-Tweedie regression model (Negative Binomial) -----------------------
+y1 <- rtweedie(200, mu = mu1$mu, power = 2, phi = 1.5)
+y1 <- rpois(200, lambda = y1)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Initial values list ------------------------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = c(1,0))
+list_initial$power <- list("resp1" = c(2))
+list_initial$tau <- list("resp1" = c(1))
+list_initial$rho = 0
+
+fit12 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "poisson_tweedie", power_fixed = list(FALSE), data = data,
+ control_initial = list_initial,
+ control_algorithm = list("method" = "rc", "tunning" = 0.2))
+summary(fit12)
+
+# Case 13 - Poisson-Tweedie regression model (PIG - Poisson Inverse Gaussian)---------
+y1 <- rtweedie(200, mu = mu1$mu, power = 3, phi = 1.5)
+y1 <- rpois(200, lambda = y1)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+# Initial values list -----------------------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = c(1,0.8))
+list_initial$power <- list("resp1" = c(3))
+list_initial$tau <- list("resp1" = c(0.5))
+list_initial$rho = 0
+
+fit13 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "poisson_tweedie", data = data, control_initial = list_initial,
+ control_algorithm = list("method" = "rc", "tunning" = 0.1))
+summary(fit13)
+
+# Case 14 - Poisson-Tweedie regression model (Pólya-Aeppli) ---------------------------
+y1 <- rtweedie(200, mu = mu1$mu, power = 1.5, phi = 1.5)
+y1 <- rpois(200, lambda = y1)
+data <- data.frame("y1" = y1, "covariate" = covariate)
+
+fit14 <- mcglm(linear_pred = c(y1 ~ covariate), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "poisson_tweedie", power_fixed = FALSE, data = data)
+summary(fit14)
+
+# Methods ------------------------------------------------------------------------
+# print
+fit14
+
+# vcov
+vcov(fit14)
+
+# confint
+confint(fit14)
+
+# summary
+summary(fit14)
+
+# anova
+anova(fit14)
+
+# Plot
+plot(fit14)
+plot(fit14, type = "algorithm")
+plot(fit14, type = "partial_residuals")
+
+# Residuals
+hist(residuals(fit14, type = "raw")[,1])
+hist(residuals(fit14, type = "pearson")[,1])
+hist(residuals(fit14, type = "standardized")[,1])
+
+# Fitted values
+plot(as.numeric(fitted(fit14)) ~ data$y1)
+
+
diff --git a/Examples/GLMExamples.R b/Examples/GLMExamples.R
new file mode 100755
index 0000000..652f23a
--- /dev/null
+++ b/Examples/GLMExamples.R
@@ -0,0 +1,207 @@
+# Set of examples 2 - GLM examples ----------------------------------------------
+# Author: Wagner Hugo Bonat LEG/IMADA -------------------------------------------
+# Date: 08/08/2015 --------------------------------------------------------------
+#--------------------------------------------------------------------------------
+rm(list=ls())
+
+# Loading extra packages
+require(mcglm)
+
+# Case 1 ------------------------------------------------------------------------
+## Dobson (1990) Page 93: Randomized Controlled Trial :
+counts <- c(18,17,15,20,10,20,25,13,12)
+outcome <- gl(3,1,9)
+treatment <- gl(3,3)
+print(d.AD <- data.frame(treatment, outcome, counts))
+
+# Orthodox Poisson model
+fit.glm <- glm(counts ~ outcome + treatment, family = quasipoisson())
+summary(fit.glm)
+
+# Quasi-Poisson model via mcglm---------------------------------------------------
+Z0 <- Diagonal(dim(d.AD)[1],1)
+fit.qglm <- mcglm(linear_pred = c(counts ~ outcome + treatment), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "tweedie", data = d.AD,
+ control_algorithm = list("verbose" = FALSE, "method" = "chaser", "tunning" = 0.8))
+summary(fit.qglm)
+cbind("mcglm" = round(coef(fit.qglm, type = "beta")$Estimates,5), "glm" = round(coef(fit.glm),5))
+cbind("mcglm" = sqrt(diag(vcov(fit.qglm))), "glm" = c(sqrt(diag(vcov(fit.glm))),NA))
+plot(fit.qglm)
+plot(fit.qglm, type = "algorithm")
+
+# Poisson-Tweedie model via mcglm------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = coef(fit.glm) )
+list_initial$power <- list("resp1" = c(1))
+list_initial$tau <- list("resp1" = c(0.01))
+list_initial$rho = 0
+Z0 <- Diagonal(dim(d.AD)[1],1)
+
+fit.pt <- mcglm(linear_pred = c(counts ~ outcome + treatment), matrix_pred = list("resp1" = list(Z0)),
+ link = "log", variance = "poisson_tweedie",
+ data = d.AD, control_initial = list_initial,
+ control_algorithm = list("correct" = TRUE, tol = 1e-5,
+ max_iter = 100, method = "chaser", "tunning" = 1))
+summary(fit.pt)
+cbind("mcglm" = round(coef(fit.pt, type = "beta")$Estimates,5), "glm" = round(coef(fit.glm),5))
+cbind("mcglm" = sqrt(diag(vcov(fit.pt))), "glm" = c(sqrt(diag(vcov(fit.glm))),NA))
+
+# This model is unsuitable for this data, note that the dispersion parameter is negative, indicating
+# underdispersion. Which agrees with my quasi Poisson model, but the glm function does not
+# agree with this result. I have to understand this difference.
+
+# Case 2 ------------------------------------------------------------------------
+# an example with offsets from Venables & Ripley (2002, p.189)
+
+# Loading the data set
+utils::data(anorexia, package = "MASS")
+
+# Orthodox GLM fit --------------------------------------------------------------
+anorex.1 <- glm(Postwt ~ Prewt + Treat + offset(Prewt),
+ family = gaussian, data = anorexia)
+summary(anorex.1)
+
+# Fitting by mcglm --------------------------------------------------------------
+Z0 <- Diagonal(dim(anorexia)[1],1)
+
+fit.anorexia <- mcglm(linear_pred = c(Postwt ~ Prewt + Treat), matrix_pred = list("resp1" = list(Z0)),
+ link = "identity", variance = "constant", offset = list(anorexia$Prewt),
+ power_fixed = list(TRUE), data = anorexia,
+ control_algorithm = list("correct" = FALSE))
+summary(fit.anorexia)
+
+# Comparing the results ---------------------------------------------------------
+cbind("mcglm" = round(coef(fit.anorexia, type = "beta")$Estimates,5),
+ "glm" = round(coef(anorex.1),5))
+cbind("mcglm" = sqrt(diag(vcov(fit.anorexia))),
+ "glm" = c(sqrt(diag(vcov(anorex.1))),NA))
+
+
+# Case 3 ------------------------------------------------------------------------
+# A Gamma example, from McCullagh & Nelder (1989, pp.300-2)
+clotting <- data.frame(
+ u = c(5,10,15,20,30,40,60,80,100),
+ lot1 = c(118,58,42,35,27,25,21,19,18),
+ lot2 = c(69,35,26,21,18,16,13,12,12))
+fit.lot1 <- glm(lot1 ~ log(u), data = clotting, family = Gamma(link = "inverse"))
+fit.lot2 <- glm(lot2 ~ log(u), data = clotting, family = Gamma(link = "inverse"))
+summary(fit.lot1)
+
+# Initial values -----------------------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = coef(fit.lot1))
+list_initial$power <- list("resp1" = c(2))
+list_initial$tau <- list("resp1" = summary(fit.lot1)$dispersion)
+list_initial$rho = 0
+Z0 <- Diagonal(dim(clotting)[1],1)
+
+# Fitting ------------------------------------------------------------------------
+fit.lot1.mcglm <- mcglm(linear_pred = c(lot1 ~ log(u)), matrix_pred = list("resp1" = list(Z0)),
+ link = "inverse", variance = "tweedie",
+ data = clotting, control_initial = list_initial)
+summary(fit.lot1.mcglm)
+
+cbind("mcglm" = round(coef(fit.lot1.mcglm, type = "beta")$Estimates,5),
+ "glm" = round(coef(fit.lot1),5))
+cbind("mcglm" = sqrt(diag(vcov(fit.lot1.mcglm))),
+ "glm" = c(sqrt(diag(vcov(fit.lot1))),NA))
+
+# Initial values -----------------------------------------------------------------
+list_initial$regression <- list("resp1" = coef(fit.lot2))
+list_initial$tau <- list("resp1" = c(var(1/clotting$lot2)))
+
+# Fitting ------------------------------------------------------------------------
+fit.lot2.mcglm <- mcglm(linear_pred = c(lot2 ~ log(u)), matrix_pred = list("resp2" = list(Z0)),
+ link = "inverse", variance = "tweedie", data = clotting,
+ control_initial = list_initial)
+summary(fit.lot2.mcglm)
+
+cbind("mcglm" = round(coef(fit.lot2.mcglm, type = "beta")$Estimates,5),
+ "glm" = round(coef(fit.lot2),5))
+cbind("mcglm" = sqrt(diag(vcov(fit.lot2.mcglm))),
+ "glm" = c(sqrt(diag(vcov(fit.lot2))),NA))
+
+# Bivariate Gamma model-----------------------------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = coef(fit.lot1), "resp2" = coef(fit.lot2))
+list_initial$power <- list("resp1" = c(2), "resp2" = c(2))
+list_initial$tau <- list("resp1" = c(0.00149), "resp2" = c(0.001276))
+list_initial$rho = 0.80
+Z0 <- Diagonal(dim(clotting)[1],1)
+
+fit.joint.mcglm <- mcglm(linear_pred = c(lot1 ~ log(u), lot2 ~ log(u)), matrix_pred = list(list(Z0), list(Z0)),
+ link = c("inverse", "inverse"), variance = c("tweedie", "tweedie"),
+ data = clotting, control_initial = list_initial,
+ control_algorithm = list("correct" = TRUE, "method" = "rc", "tunning" = 0.001,
+ max_iter = 100))
+summary(fit.joint.mcglm)
+plot(fit.joint.mcglm, type = "algorithm")
+plot(fit.joint.mcglm)
+
+# Bivariate Gamma model + log link function --------------------------------------
+list_initial = list()
+list_initial$regression <- list("resp1" = c(log(mean(clotting$lot1)),0),
+ "resp2" = c(log(mean(clotting$lot2)),0))
+list_initial$power <- list("resp1" = c(2), "resp2" = c(2))
+list_initial$tau <- list("resp1" = 0.023, "resp2" = 0.024)
+list_initial$rho = 0
+Z0 <- Diagonal(dim(clotting)[1],1)
+
+fit.joint.log <- mcglm(linear_pred = c(lot1 ~ log(u), "resp2" = lot2 ~ log(u)),
+ matrix_pred = list(list(Z0),list(Z0)), link = c("log", "log"),
+ variance = c("tweedie", "tweedie"), data = clotting,
+ control_initial = list_initial)
+summary(fit.joint.log)
+plot(fit.joint.mcglm, type = "algorithm")
+plot(fit.joint.mcglm)
+
+# Case 4 - Binomial regression models ----------------------------------------
+require(MASS)
+data(menarche)
+head(menarche)
+data <- data.frame("resp" = menarche$Menarche/menarche$Total, "Ntrial" = menarche$Total,
+ "Age" = menarche$Age)
+
+# Orthodox logistic regression model ------------------------------------------
+glm.out = glm(cbind(Menarche, Total-Menarche) ~ Age, family=binomial(logit), data=menarche)
+
+# Fitting ---------------------------------------------------------------------
+Z0 <- Diagonal(dim(data)[1],1)
+
+fit.logit <- mcglm(linear_pred = c(resp ~ Age), matrix_pred = list("resp1" = list(Z0)),
+ link = "logit", variance = "binomialP", Ntrial = list(data$Ntrial), data = data)
+
+summary(fit.logit)
+plot(fit.logit, type = "algorithm")
+plot(fit.logit)
+
+# Fitting with extra power parameter -------------------------------------------
+fit.logit.power <- mcglm(linear_pred = c(resp ~ Age), matrix_pred = list(list(Z0)),
+ link = "logit", variance = "binomialP", Ntrial = list(data$Ntrial),
+ power_fixed = FALSE, data = data)
+summary(fit.logit.power)
+plot(fit.logit.power, type = "algorithm")
+plot(fit.logit.power)
+
+# All methods --------------------------------------------------------------------
+# print method
+fit.logit.power
+# coef method
+coef(fit.logit.power)
+# confint method
+confint(fit.logit.power)
+# vcov method
+vcov(fit.logit.power)
+# summary method
+summary(fit.logit.power)
+# anova method
+anova(fit.logit.power)
+# Fitted method
+fitted(fit.logit.power)
+# Residuals method
+residuals(fit.logit.power)
+# Plot method
+plot(fit.logit.power)
+plot(fit.logit.power, type = "algorithm")
+plot(fit.logit.power, type = "partial_residuals")
+# End
diff --git a/R/fit_mcglm.R b/R/fit_mcglm.R
index 9a02370..3c2d3f8 100644
--- a/R/fit_mcglm.R
+++ b/R/fit_mcglm.R
@@ -26,137 +26,116 @@
#' method. This argument control the step-length.
#' @param verbose A logical if TRUE print the values of the covariance parameters used on each iteration.
#' @return A list with estimated regression and covariance parameters.
+#' @export
-fit_mcglm <- function(list_initial, list_link, list_variance, list_covariance, list_X, list_Z,
- list_offset, list_Ntrial, list_power_fixed, list_sparse, y_vec,
- correct = TRUE, max_iter, tol = 1e-03, method = "rc", tunning = 0,
- verbose) {
- ## Transformation from list to vector
- parametros <- mc_list2vec(list_initial, list_power_fixed)
- n_resp <- length(list_initial$regression)
- if(n_resp == 1){parametros$cov_ini <- parametros$cov_ini[-1]}
- ## Getting information about the number of parameters
- inf <- mc_getInformation(list_initial, list_power_fixed, n_resp = n_resp)
- ## Creating a matrix to sote all values used in the fitting step
- solucao_beta <- matrix(NA, max_iter,length(parametros$beta_ini))
- solucao_cov <- matrix(NA, max_iter, length(parametros$cov_ini))
- score_beta_temp <- matrix(NA, max_iter, length(parametros$beta_ini))
- score_disp_temp <- matrix(NA, max_iter, length(parametros$cov_ini))
- ## Setting the initial values
- solucao_beta[1,] <- parametros$beta_ini
- solucao_cov[1,] <- parametros$cov_ini
- beta_ini <- parametros$beta_ini
- cov_ini <- parametros$cov_ini
- for(i in 2:max_iter){
- ## Step 1 - Quasi-score function
- # Step 1.1 - Computing the mean structure
- mu_list <- Map(mc_link_function, beta = list_initial$regression, offset = list_offset, X = list_X,
- link = list_link)
- mu_vec <- do.call(c,lapply(mu_list, function(x)x$mu))
- D <- bdiag(lapply(mu_list, function(x)x$D))
- # Step 1.2 - Computing the inverse of C matrix. I should improve this step.
- # I have to write a new function to compute only C or inv_C to be more efficient in this step.
- Cfeatures <- mc_build_C(list_mu = mu_list, list_Ntrial = list_Ntrial, rho = list_initial$rho,
- list_tau = list_initial$tau, list_power = list_initial$power,
- list_Z = list_Z, list_sparse = list_sparse, list_variance = list_variance,
- list_covariance = list_covariance, list_power_fixed = list_power_fixed,
- compute_C = FALSE, compute_derivative_beta = FALSE,
- compute_derivative_cov = FALSE)
- # Step 1.3 - Update the regression parameters
- beta_temp <- mc_quasi_score(D = D, inv_C = Cfeatures$inv_C, y_vec = y_vec, mu_vec = mu_vec)
- solucao_beta[i,] <- as.numeric(beta_ini - solve(beta_temp$Sensitivity, beta_temp$Score))
- score_beta_temp[i,] <- as.numeric(beta_temp$Score)
- list_initial <- mc_updateBeta(list_initial, solucao_beta[i,], information = inf, n_resp = n_resp)
- # Step 1.4 - Updated the mean structure to use in the Pearson estimating function step.
- mu_list <- Map(mc_link_function, beta = list_initial$regression, offset = list_offset, X = list_X,
- link = list_link)
- mu_vec <- do.call(c,lapply(mu_list, function(x)x$mu))
- D <- bdiag(lapply(mu_list, function(x)x$D))
- # Step 2 - Updating the covariance parameters
- Cfeatures <- mc_build_C(list_mu = mu_list, list_Ntrial = list_Ntrial, rho = list_initial$rho,
- list_tau = list_initial$tau, list_power = list_initial$power,
- list_Z = list_Z,
- list_sparse = list_sparse, list_variance = list_variance,
- list_covariance = list_covariance,
- list_power_fixed = list_power_fixed, compute_C = TRUE,
- compute_derivative_beta = FALSE)
- # Step 2.1 - Using beta(i+1)
- #beta_temp2 <- mc_quasi_score(D = D, inv_C = Cfeatures$inv_C, y_vec = y_vec, mu_vec = mu_vec)
- inv_J_beta <- solve(beta_temp$Sensitivity)
- if(method == "chaser"){
- cov_temp <- mc_pearson(y_vec = y_vec, mu_vec = mu_vec, Cfeatures = Cfeatures,
- inv_J_beta = inv_J_beta, D = D, correct = correct,
- compute_variability = TRUE)
- step <- tunning*solve(cov_temp$Sensitivity, cov_temp$Score)
- }
- if(method == "rc"){
- cov_temp <- mc_pearson(y_vec = y_vec, mu_vec = mu_vec, Cfeatures = Cfeatures,
- inv_J_beta = inv_J_beta, D = D, correct = correct,
- compute_variability = TRUE)
- step <- solve(tunning*cov_temp$Score%*%t(cov_temp$Score)%*%
- solve(cov_temp$Variability)%*%cov_temp$Sensitivity +
- cov_temp$Sensitivity)%*%cov_temp$Score
- }
- ## Step 2.2 - Updating the covariance parameters
- score_disp_temp[i,] <- cov_temp$Score
- cov_next <- as.numeric(cov_ini - step)
- list_initial <- mc_updateCov(list_initial = list_initial, list_power_fixed = list_power_fixed,
- covariance = cov_next,
- information = inf, n_resp = n_resp)
- ## print the parameters values
- if(verbose == TRUE){print(round(cov_next,4))}
- if(verbose == TRUE){print(round(as.numeric(cov_temp$Score),4))}
- ## Step 2.3 - Updating the initial values for the next step
- beta_ini <- solucao_beta[i,]
- cov_ini <- cov_next
- solucao_cov[i,] <- cov_next
- ## Checking the convergence
- tolera <- abs(c(solucao_beta[i,],solucao_cov[i,]) - c(solucao_beta[i-1,],solucao_cov[i-1,]))
- #if(verbose == TRUE){print(round(tolera, 4))}
- if(all(tolera <= tol) == TRUE)break
- }
- mu_list <- Map(mc_link_function, beta = list_initial$regression, offset = list_offset, X = list_X,
- link = list_link)
- mu_vec <- do.call(c,lapply(mu_list, function(x)x$mu))
- D <- bdiag(lapply(mu_list, function(x)x$D))
- Cfeatures <- mc_build_C(list_mu = mu_list, list_Ntrial = list_Ntrial, rho = list_initial$rho,
- list_tau = list_initial$tau, list_power = list_initial$power,
- list_Z = list_Z,
- list_sparse = list_sparse, list_variance = list_variance,
- list_covariance = list_covariance,
- list_power_fixed = list_power_fixed, compute_C = TRUE,
- compute_derivative_beta = FALSE)
- beta_temp2 <- mc_quasi_score(D = D, inv_C = Cfeatures$inv_C, y_vec = y_vec, mu_vec = mu_vec)
- inv_J_beta <- solve(beta_temp2$Sensitivity)
+fit_mcglm <- function(list_initial, list_link, list_variance, list_covariance, list_X, list_Z, list_offset, list_Ntrial,
+ list_power_fixed, list_sparse, y_vec, correct = TRUE, max_iter, tol = 0.001, method = "rc", tunning = 0, verbose) {
+ ## Transformation from list to vector
+ parametros <- mc_list2vec(list_initial, list_power_fixed)
+ n_resp <- length(list_initial$regression)
+ if (n_resp == 1) {
+ parametros$cov_ini <- parametros$cov_ini[-1]
+ }
+ ## Getting information about the number of parameters
+ inf <- mc_getInformation(list_initial, list_power_fixed, n_resp = n_resp)
+ ## Creating a matrix to sote all values used in the fitting step
+ solucao_beta <- matrix(NA, max_iter, length(parametros$beta_ini))
+ solucao_cov <- matrix(NA, max_iter, length(parametros$cov_ini))
+ score_beta_temp <- matrix(NA, max_iter, length(parametros$beta_ini))
+ score_disp_temp <- matrix(NA, max_iter, length(parametros$cov_ini))
+ ## Setting the initial values
+ solucao_beta[1, ] <- parametros$beta_ini
+ solucao_cov[1, ] <- parametros$cov_ini
+ beta_ini <- parametros$beta_ini
+ cov_ini <- parametros$cov_ini
+ for (i in 2:max_iter) {
+ ## Step 1 - Quasi-score function Step 1.1 - Computing the mean structure
+ mu_list <- Map(mc_link_function, beta = list_initial$regression, offset = list_offset, X = list_X, link = list_link)
+ mu_vec <- do.call(c, lapply(mu_list, function(x) x$mu))
+ D <- bdiag(lapply(mu_list, function(x) x$D))
+ # Step 1.2 - Computing the inverse of C matrix. I should improve this step. I have to write a new function to compute
+ # only C or inv_C to be more efficient in this step.
+ Cfeatures <- mc_build_C(list_mu = mu_list, list_Ntrial = list_Ntrial, rho = list_initial$rho, list_tau = list_initial$tau,
+ list_power = list_initial$power, list_Z = list_Z, list_sparse = list_sparse, list_variance = list_variance,
+ list_covariance = list_covariance, list_power_fixed = list_power_fixed, compute_C = FALSE, compute_derivative_beta = FALSE,
+ compute_derivative_cov = FALSE)
+ # Step 1.3 - Update the regression parameters
+ beta_temp <- mc_quasi_score(D = D, inv_C = Cfeatures$inv_C, y_vec = y_vec, mu_vec = mu_vec)
+ solucao_beta[i, ] <- as.numeric(beta_ini - solve(beta_temp$Sensitivity, beta_temp$Score))
+ score_beta_temp[i, ] <- as.numeric(beta_temp$Score)
+ list_initial <- mc_updateBeta(list_initial, solucao_beta[i, ], information = inf, n_resp = n_resp)
+ # Step 1.4 - Updated the mean structure to use in the Pearson estimating function step.
+ mu_list <- Map(mc_link_function, beta = list_initial$regression, offset = list_offset, X = list_X, link = list_link)
+ mu_vec <- do.call(c, lapply(mu_list, function(x) x$mu))
+ D <- bdiag(lapply(mu_list, function(x) x$D))
+ # Step 2 - Updating the covariance parameters
+ Cfeatures <- mc_build_C(list_mu = mu_list, list_Ntrial = list_Ntrial, rho = list_initial$rho, list_tau = list_initial$tau,
+ list_power = list_initial$power, list_Z = list_Z, list_sparse = list_sparse, list_variance = list_variance,
+ list_covariance = list_covariance, list_power_fixed = list_power_fixed, compute_C = TRUE, compute_derivative_beta = FALSE)
+ # Step 2.1 - Using beta(i+1) beta_temp2 <- mc_quasi_score(D = D, inv_C = Cfeatures$inv_C, y_vec = y_vec, mu_vec =
+ # mu_vec)
+ inv_J_beta <- solve(beta_temp$Sensitivity)
+ if (method == "chaser") {
+ cov_temp <- mc_pearson(y_vec = y_vec, mu_vec = mu_vec, Cfeatures = Cfeatures, inv_J_beta = inv_J_beta, D = D,
+ correct = correct, compute_variability = TRUE)
+ step <- tunning * solve(cov_temp$Sensitivity, cov_temp$Score)
+ }
+ if (method == "rc") {
+ cov_temp <- mc_pearson(y_vec = y_vec, mu_vec = mu_vec, Cfeatures = Cfeatures, inv_J_beta = inv_J_beta, D = D,
+ correct = correct, compute_variability = TRUE)
+ step <- solve(tunning * cov_temp$Score %*% t(cov_temp$Score) %*% solve(cov_temp$Variability) %*% cov_temp$Sensitivity +
+ cov_temp$Sensitivity) %*% cov_temp$Score
+ }
+ ## Step 2.2 - Updating the covariance parameters
+ score_disp_temp[i, ] <- cov_temp$Score
+ cov_next <- as.numeric(cov_ini - step)
+ list_initial <- mc_updateCov(list_initial = list_initial, list_power_fixed = list_power_fixed, covariance = cov_next,
+ information = inf, n_resp = n_resp)
+ ## print the parameters values
+ if (verbose == TRUE) {
+ print(round(cov_next, 4))
+ }
+ if (verbose == TRUE) {
+ print(round(as.numeric(cov_temp$Score), 4))
+ }
+ ## Step 2.3 - Updating the initial values for the next step
+ beta_ini <- solucao_beta[i, ]
+ cov_ini <- cov_next
+ solucao_cov[i, ] <- cov_next
+ ## Checking the convergence
+ tolera <- abs(c(solucao_beta[i, ], solucao_cov[i, ]) - c(solucao_beta[i - 1, ], solucao_cov[i - 1, ]))
+ # if(verbose == TRUE){print(round(tolera, 4))}
+ if (all(tolera <= tol) == TRUE)
+ break
+ }
+ mu_list <- Map(mc_link_function, beta = list_initial$regression, offset = list_offset, X = list_X, link = list_link)
+ mu_vec <- do.call(c, lapply(mu_list, function(x) x$mu))
+ D <- bdiag(lapply(mu_list, function(x) x$D))
+ Cfeatures <- mc_build_C(list_mu = mu_list, list_Ntrial = list_Ntrial, rho = list_initial$rho, list_tau = list_initial$tau,
+ list_power = list_initial$power, list_Z = list_Z, list_sparse = list_sparse, list_variance = list_variance, list_covariance = list_covariance,
+ list_power_fixed = list_power_fixed, compute_C = TRUE, compute_derivative_beta = FALSE)
+ beta_temp2 <- mc_quasi_score(D = D, inv_C = Cfeatures$inv_C, y_vec = y_vec, mu_vec = mu_vec)
+ inv_J_beta <- solve(beta_temp2$Sensitivity)
- cov_temp <- mc_pearson(y_vec = y_vec, mu_vec = mu_vec, Cfeatures = Cfeatures,
- inv_J_beta = inv_J_beta, D = D, correct = correct,
- compute_variability = TRUE)
- Product_beta <- lapply(Cfeatures$D_C_beta, mc_multiply, bord2 = Cfeatures$inv_C)
- S_cov_beta <- mc_cross_sensitivity(Product_cov = cov_temp$Extra,
- Product_beta = Product_beta,
- n_beta_effective = length(beta_temp$Score))
- res <- y_vec - mu_vec
- V_cov_beta <- mc_cross_variability(Product_cov = cov_temp$Extra, inv_C = Cfeatures$inv_C,
- res = res, D = D)
- p1 <- rbind(beta_temp2$Variability, t(V_cov_beta))
- p2 <- rbind(V_cov_beta, cov_temp$Variability)
- joint_variability <- cbind(p1,p2)
- inv_S_beta <- inv_J_beta
- inv_S_cov <- solve(cov_temp$Sensitivity)
- mat0 <- Matrix(0, ncol = dim(S_cov_beta)[1], nrow = dim(S_cov_beta)[2])
- cross_term <- -inv_S_cov%*%S_cov_beta%*%inv_S_beta
- p1 <- rbind(inv_S_beta,cross_term)
- p2 <- rbind(mat0, inv_S_cov)
- joint_inv_sensitivity <- cbind(p1,p2)
- VarCov <- joint_inv_sensitivity%*%joint_variability%*%t(joint_inv_sensitivity)
- output <- list("IterationRegression" = solucao_beta, "IterationCovariance" = solucao_cov,
- "ScoreRegression" = score_beta_temp, "ScoreCovariance" = score_disp_temp,
- "Regression" = beta_ini, "Covariance" = cov_ini, "vcov" = VarCov,
- "fitted" = mu_vec, "residuals" = res, "inv_C" = Cfeatures$inv_C,
- "C" = Cfeatures$C, "Information" = inf, "mu_list" = mu_list,
- "inv_S_beta" = inv_S_beta)
- return(output)
+ cov_temp <- mc_pearson(y_vec = y_vec, mu_vec = mu_vec, Cfeatures = Cfeatures, inv_J_beta = inv_J_beta, D = D, correct = correct,
+ compute_variability = TRUE)
+ Product_beta <- lapply(Cfeatures$D_C_beta, mc_multiply, bord2 = Cfeatures$inv_C)
+ S_cov_beta <- mc_cross_sensitivity(Product_cov = cov_temp$Extra, Product_beta = Product_beta, n_beta_effective = length(beta_temp$Score))
+ res <- y_vec - mu_vec
+ V_cov_beta <- mc_cross_variability(Product_cov = cov_temp$Extra, inv_C = Cfeatures$inv_C, res = res, D = D)
+ p1 <- rbind(beta_temp2$Variability, t(V_cov_beta))
+ p2 <- rbind(V_cov_beta, cov_temp$Variability)
+ joint_variability <- cbind(p1, p2)
+ inv_S_beta <- inv_J_beta
+ inv_S_cov <- solve(cov_temp$Sensitivity)
+ mat0 <- Matrix(0, ncol = dim(S_cov_beta)[1], nrow = dim(S_cov_beta)[2])
+ cross_term <- -inv_S_cov %*% S_cov_beta %*% inv_S_beta
+ p1 <- rbind(inv_S_beta, cross_term)
+ p2 <- rbind(mat0, inv_S_cov)
+ joint_inv_sensitivity <- cbind(p1, p2)
+ VarCov <- joint_inv_sensitivity %*% joint_variability %*% t(joint_inv_sensitivity)
+ output <- list(IterationRegression = solucao_beta, IterationCovariance = solucao_cov, ScoreRegression = score_beta_temp,
+ ScoreCovariance = score_disp_temp, Regression = beta_ini, Covariance = cov_ini, vcov = VarCov, fitted = mu_vec,
+ residuals = res, inv_C = Cfeatures$inv_C, C = Cfeatures$C, Information = inf, mu_list = mu_list, inv_S_beta = inv_S_beta)
+ return(output)
}
-
-
diff --git a/R/mc_anova.R b/R/mc_anova.R
new file mode 100644
index 0000000..b23a1d9
--- /dev/null
+++ b/R/mc_anova.R
@@ -0,0 +1,50 @@
+#' ANOVA method for McGLMs.
+#'
+#' @description ANOVA method for McGLMS.
+#'
+#' @param object an object of class mcglm, usually, a result of a call to \code{mcglm}.
+#' @export
+
+anova.mcglm <- function(object) {
+ n_resp <- length(object$mu_list)
+ n_beta <- lapply(object$list_X, ncol)
+ idx.list <- list()
+ for (i in 1:n_resp) {
+ idx.list[[i]] <- rep(i, n_beta[i])
+ }
+ vv <- vcov(object)
+ n_par <- dim(vv)[1]
+ idx.vec <- do.call(c, idx.list)
+ n_cov <- n_par - length(idx.vec)
+ idx.vec <- c(idx.vec, rep(0, n_cov))
+ temp.vcov <- list()
+ temp.beta <- list()
+ for (i in 1:n_resp) {
+ idx.id = idx.vec == i
+ temp.vcov[[i]] <- vv[idx.id, idx.id]
+ temp.beta[[i]] <- coef(object, type = "beta", response = i)$Estimates
+ }
+ saida <- list()
+ for (i in 1:n_resp) {
+ idx <- attr(object$list_X[[i]], "assign")
+ names <- colnames(object$list_X[[i]])
+ if (names[1] == "(Intercept)") {
+ idx <- idx[-1]
+ names <- names[-1]
+ temp.beta[[i]] <- temp.beta[[i]][-1]
+ temp.vcov[[i]] <- temp.vcov[[i]][-1, -1]
+ }
+ n_terms <- length(unique(idx))
+ X2.resp <- list()
+ for (j in 1:n_terms) {
+ idx.TF <- idx == j
+ temp <- as.numeric(t(temp.beta[[i]][idx.TF]) %*% solve(as.matrix(temp.vcov[[i]])[idx.TF, idx.TF]) %*% temp.beta[[i]][idx.TF])
+ nbeta.test <- length(temp.beta[[i]][idx.TF])
+ X2.resp[[j]] <- data.frame(Covariate = names[idx.TF][1], Chi.Square = round(temp, 4), Df = nbeta.test, p.value = round(pchisq(temp,
+ nbeta.test, lower.tail = FALSE), 4))
+ }
+ saida[[i]] <- do.call(rbind, X2.resp)
+ }
+ cat("Wald test for fixed effects", "\n")
+ return(saida)
+}
diff --git a/R/mc_auxiliar.R b/R/mc_auxiliar.R
index e00b203..ab99821 100644
--- a/R/mc_auxiliar.R
+++ b/R/mc_auxiliar.R
@@ -15,30 +15,35 @@
#' mc_sandwich(middle = M, bord1 = X1, bord2 = X1)
#' mc_sandwich_negative(middle = M, bord1 = X1, bord2 = X1)
#' @export
+
## Auxiliar function to multiply matrices
-mc_sandwich <- function(middle, bord1, bord2){bord1%*%middle%*%bord2}
+mc_sandwich <- function(middle, bord1, bord2) {
+ bord1 %*% middle %*% bord2
+}
#' @rdname mc_sandwich
-mc_sandwich_negative <- function(middle, bord1, bord2){-bord1%*%middle%*%bord2}
+mc_sandwich_negative <- function(middle, bord1, bord2) {
+ -bord1 %*% middle %*% bord2
+}
#' @rdname mc_sandwich
-mc_sandwich_power <- function(middle, bord1, bord2){
- temp1 <- mc_sandwich(middle = middle, bord1 = bord1, bord2 = bord2)
- return(temp1 + t(temp1))
+mc_sandwich_power <- function(middle, bord1, bord2) {
+ temp1 <- mc_sandwich(middle = middle, bord1 = bord1, bord2 = bord2)
+ return(temp1 + t(temp1))
}
#' @rdname mc_sandwich
-mc_sandwich_cholesky <- function(bord1, middle, bord2){
- p1 <- bord1%*%middle%*%bord2
- return(p1 + t(p1))
+mc_sandwich_cholesky <- function(bord1, middle, bord2) {
+ p1 <- bord1 %*% middle %*% bord2
+ return(p1 + t(p1))
}
#' @rdname mc_sandwich
-mc_multiply <- function(bord1, bord2){
- return(bord2%*%bord1)
+mc_multiply <- function(bord1, bord2) {
+ return(bord2 %*% bord1)
}
#' @rdname mc_sandwich
-mc_multiply2 <- function(bord1, bord2){
- return(bord1%*%bord2)
+mc_multiply2 <- function(bord1, bord2) {
+ return(bord1 %*% bord2)
}
diff --git a/R/mc_bias_correct_std.R b/R/mc_bias_correct_std.R
index 82d757f..cfe7f9f 100644
--- a/R/mc_bias_correct_std.R
+++ b/R/mc_bias_correct_std.R
@@ -4,23 +4,24 @@
#' of clustered observations. It is also robust and has improved finite sample properties.
#'
#' @param object An object of mcglm class.
-#' @param id a vector which identifies the clusters. The length and order of ‘id’ should be the
+#' @param id a vector which identifies the clusters. The length and order of id should be the
#' same as the number of observations. Data are assumed to be sorted so that observations on a cluster
#' are contiguous rows for all entities in the formula.
#' @return A matrix. Note that the function assumes that the data are in the correct order.
#' @export
mc_bias_corrected_std <- function(object, id) {
- inv_M <- object$inv_S_beta
- temp_data <- data.frame("res" = object$residuals, id)
- temp_data_group <- split(temp_data, temp_data$id)
- D <- bdiag(lapply(object$mu_list, function(x)x$D))
- r_rT <- bdiag(lapply(temp_data_group, function(x){tcrossprod(x[,1])}))
- tD_invC <- t(D)%*%object$inv_C
- H <- Matrix(D%*%inv_M%*%tD_invC, sparse = TRUE)
- IH <- Diagonal(object$n_obs,1) - H
- inv_IH <- solve(IH)
- output <- sqrt(diag(inv_M%*%tD_invC%*%inv_IH%*%r_rT%*%inv_IH%*%t(tD_invC)%*%inv_M))
- return(output)
+ inv_M <- object$inv_S_beta
+ temp_data <- data.frame(res = object$residuals, id)
+ temp_data_group <- split(temp_data, temp_data$id)
+ D <- bdiag(lapply(object$mu_list, function(x) x$D))
+ r_rT <- bdiag(lapply(temp_data_group, function(x) {
+ tcrossprod(x[, 1])
+ }))
+ tD_invC <- t(D) %*% object$inv_C
+ H <- Matrix(D %*% inv_M %*% tD_invC, sparse = TRUE)
+ IH <- Diagonal(object$n_obs, 1) - H
+ inv_IH <- solve(IH)
+ output <- sqrt(diag(inv_M %*% tD_invC %*% inv_IH %*% r_rT %*% inv_IH %*% t(tD_invC) %*% inv_M))
+ return(output)
}
-
diff --git a/R/mc_build_C.R b/R/mc_build_C.R
index ee30555..d436e3d 100644
--- a/R/mc_build_C.R
+++ b/R/mc_build_C.R
@@ -5,6 +5,7 @@
#'
#'@param list_mu A list with values of the mean.
#'@param list_Ntrial A list with the number of trials. Usefull only for binomial responses.
+#'@param rho Vector of correlation parameters.
#'@param list_tau A list with values for the tau parameters.
#'@param list_power A list with values for the power parameters.
#'@param list_Z A list of matrix to be used in the matrix linear predictor.
@@ -12,77 +13,68 @@
#'@param list_variance A list specifying the variance function to be used for each response variable.
#'@param list_covariance A list specifying the covariance function to be used for each response variable.
#'@param list_power_fixed A list of Logical specifying if the power parameters are fixed or not.
+#'@param compute_C Logical. Compute or not the C matrix.
+#'@param compute_derivative_beta Logical. Compute or not the derivative of C with respect to regression parameters.
+#'@param compute_derivative_cov Logical. Compute or not the derivative of C with respect the covariance parameters.
#'
#'@return A list with the inverse of the C matrix and the derivatives of the C matrix with respect to
#'rho, power and tau parameters.
#'@export
-mc_build_C <- function(list_mu, list_Ntrial, rho, list_tau, list_power, list_Z, list_sparse,
- list_variance, list_covariance, list_power_fixed, compute_C = FALSE,
- compute_derivative_beta = FALSE,
+mc_build_C <- function(list_mu, list_Ntrial, rho, list_tau, list_power, list_Z, list_sparse, list_variance,
+ list_covariance, list_power_fixed, compute_C = FALSE, compute_derivative_beta = FALSE,
compute_derivative_cov = TRUE) {
- n_resp <- length(list_mu)
- n_obs <- length(list_mu[[1]][[1]])
- n_rho <- n_resp*(n_resp - 1)/2
- if(n_resp != 1){assert_that(n_rho == length(rho))}
- list_Sigma_within = suppressWarnings(Map(mc_build_sigma, mu = list_mu, Ntrial = list_Ntrial,
- tau = list_tau,
- power = list_power,Z = list_Z, sparse = list_sparse,
- variance = list_variance, covariance = list_covariance,
- power_fixed = list_power_fixed,
- compute_derivative_beta = compute_derivative_beta))
- list_Sigma_chol <- lapply(list_Sigma_within, function(x)x$Sigma_chol)
- list_Sigma_inv_chol <- lapply(list_Sigma_within, function(x)x$Sigma_chol_inv)
- Sigma_between <- mc_build_sigma_between(rho = rho, n_resp = n_resp)
- II <- Diagonal(n_obs, 1)
- nucleo <- kronecker(Sigma_between$Sigmab,II)
- Bdiag_chol_Sigma_within <- bdiag(list_Sigma_chol)
- t_Bdiag_chol_Sigma_within <- t(Bdiag_chol_Sigma_within)
- Bdiag_inv_chol_Sigma <- bdiag(list_Sigma_inv_chol)
- inv_C <- Bdiag_inv_chol_Sigma%*%kronecker(solve(Sigma_between$Sigmab),II)%*%t(Bdiag_inv_chol_Sigma)
- output <- list("inv_C" = inv_C)
- if(compute_derivative_cov == TRUE){
- list_D_Sigma <- lapply(list_Sigma_within, function(x)x$D_Sigma)
- ## Derivatives of C with respect to power and tau parameters
- list_D_chol_Sigma <- Map(mc_derivative_cholesky, derivada = list_D_Sigma,
- inv_chol_Sigma = list_Sigma_inv_chol, chol_Sigma = list_Sigma_chol)
- mat_zero <- mc_build_bdiag(n_resp = n_resp, n_obs = n_obs)
- Bdiag_D_chol_Sigma <- mapply(mc_transform_list_bdiag, list_mat = list_D_chol_Sigma,
- response_number = 1:n_resp, MoreArgs = list(mat_zero = mat_zero))
- Bdiag_D_chol_Sigma <- do.call(c, Bdiag_D_chol_Sigma)
- D_C = lapply(Bdiag_D_chol_Sigma, mc_sandwich_cholesky, middle = nucleo,
- bord2 = t_Bdiag_chol_Sigma_within)
- ## Finish the derivatives with respect to power and tau parameters
- if(n_resp > 1){
- D_C_rho <- mc_derivative_C_rho(D_Sigmab = Sigma_between$D_Sigmab,
- Bdiag_chol_Sigma_within = Bdiag_chol_Sigma_within,
- t_Bdiag_chol_Sigma_within = t_Bdiag_chol_Sigma_within,
- II = II)
- D_C <- c(D_C_rho, D_C)
- }
- output$D_C <- D_C
- }
- if(compute_C == TRUE){
- C = t_Bdiag_chol_Sigma_within%*%kronecker(Sigma_between$Sigmab,II)%*%Bdiag_chol_Sigma_within
- output$C <- C
- }
- if(compute_derivative_beta == TRUE){
- list_D_Sigma_beta <- lapply(list_Sigma_within, function(x)x$D_Sigma_beta)
- list_D_chol_Sigma_beta <- Map(mc_derivative_cholesky, derivada = list_D_Sigma_beta,
- inv_chol_Sigma = list_Sigma_inv_chol, chol_Sigma = list_Sigma_chol)
- mat_zero <- mc_build_bdiag(n_resp = n_resp, n_obs = n_obs)
- Bdiag_D_chol_Sigma_beta <- mapply(mc_transform_list_bdiag, list_mat = list_D_chol_Sigma_beta,
- response_number = 1:n_resp, MoreArgs = list(mat_zero = mat_zero))
- Bdiag_D_chol_Sigma_beta <- do.call(c, Bdiag_D_chol_Sigma_beta)
- D_C_beta = lapply(Bdiag_D_chol_Sigma_beta, mc_sandwich_cholesky, middle = nucleo,
- bord2 = t_Bdiag_chol_Sigma_within)
- output$D_C_beta <- D_C_beta
- }
- return(output)
+ n_resp <- length(list_mu)
+ n_obs <- length(list_mu[[1]][[1]])
+ n_rho <- n_resp * (n_resp - 1)/2
+ if (n_resp != 1) {
+ assert_that(n_rho == length(rho))
+ }
+ list_Sigma_within = suppressWarnings(Map(mc_build_sigma, mu = list_mu, Ntrial = list_Ntrial, tau = list_tau, power = list_power,
+ Z = list_Z, sparse = list_sparse, variance = list_variance, covariance = list_covariance, power_fixed = list_power_fixed,
+ compute_derivative_beta = compute_derivative_beta))
+ list_Sigma_chol <- lapply(list_Sigma_within, function(x) x$Sigma_chol)
+ list_Sigma_inv_chol <- lapply(list_Sigma_within, function(x) x$Sigma_chol_inv)
+ Sigma_between <- mc_build_sigma_between(rho = rho, n_resp = n_resp)
+ II <- Diagonal(n_obs, 1)
+ nucleo <- kronecker(Sigma_between$Sigmab, II)
+ Bdiag_chol_Sigma_within <- bdiag(list_Sigma_chol)
+ t_Bdiag_chol_Sigma_within <- t(Bdiag_chol_Sigma_within)
+ Bdiag_inv_chol_Sigma <- bdiag(list_Sigma_inv_chol)
+ inv_C <- Bdiag_inv_chol_Sigma %*% kronecker(solve(Sigma_between$Sigmab), II) %*% t(Bdiag_inv_chol_Sigma)
+ output <- list(inv_C = inv_C)
+ if (compute_derivative_cov == TRUE) {
+ list_D_Sigma <- lapply(list_Sigma_within, function(x) x$D_Sigma)
+ ## Derivatives of C with respect to power and tau parameters
+ list_D_chol_Sigma <- Map(mc_derivative_cholesky, derivada = list_D_Sigma, inv_chol_Sigma = list_Sigma_inv_chol,
+ chol_Sigma = list_Sigma_chol)
+ mat_zero <- mc_build_bdiag(n_resp = n_resp, n_obs = n_obs)
+ Bdiag_D_chol_Sigma <- mapply(mc_transform_list_bdiag, list_mat = list_D_chol_Sigma, response_number = 1:n_resp,
+ MoreArgs = list(mat_zero = mat_zero))
+ Bdiag_D_chol_Sigma <- do.call(c, Bdiag_D_chol_Sigma)
+ D_C = lapply(Bdiag_D_chol_Sigma, mc_sandwich_cholesky, middle = nucleo, bord2 = t_Bdiag_chol_Sigma_within)
+ ## Finish the derivatives with respect to power and tau parameters
+ if (n_resp > 1) {
+ D_C_rho <- mc_derivative_C_rho(D_Sigmab = Sigma_between$D_Sigmab, Bdiag_chol_Sigma_within = Bdiag_chol_Sigma_within,
+ t_Bdiag_chol_Sigma_within = t_Bdiag_chol_Sigma_within, II = II)
+ D_C <- c(D_C_rho, D_C)
+ }
+ output$D_C <- D_C
+ }
+ if (compute_C == TRUE) {
+ C = t_Bdiag_chol_Sigma_within %*% kronecker(Sigma_between$Sigmab, II) %*% Bdiag_chol_Sigma_within
+ output$C <- C
+ }
+ if (compute_derivative_beta == TRUE) {
+ list_D_Sigma_beta <- lapply(list_Sigma_within, function(x) x$D_Sigma_beta)
+ list_D_chol_Sigma_beta <- Map(mc_derivative_cholesky, derivada = list_D_Sigma_beta, inv_chol_Sigma = list_Sigma_inv_chol,
+ chol_Sigma = list_Sigma_chol)
+ mat_zero <- mc_build_bdiag(n_resp = n_resp, n_obs = n_obs)
+ Bdiag_D_chol_Sigma_beta <- mapply(mc_transform_list_bdiag, list_mat = list_D_chol_Sigma_beta, response_number = 1:n_resp,
+ MoreArgs = list(mat_zero = mat_zero))
+ Bdiag_D_chol_Sigma_beta <- do.call(c, Bdiag_D_chol_Sigma_beta)
+ D_C_beta = lapply(Bdiag_D_chol_Sigma_beta, mc_sandwich_cholesky, middle = nucleo, bord2 = t_Bdiag_chol_Sigma_within)
+ output$D_C_beta <- D_C_beta
+ }
+ return(output)
}
-
-
-
-
-
-
diff --git a/R/mc_build_bdiag.R b/R/mc_build_bdiag.R
index 0f197ef..c41c93e 100644
--- a/R/mc_build_bdiag.R
+++ b/R/mc_build_bdiag.R
@@ -10,7 +10,9 @@
#' @export
mc_build_bdiag <- function(n_resp, n_obs) {
- list_zero <- list()
- for(i in 1:n_resp){list_zero[[i]] <- Matrix(0, n_obs, n_obs, sparse = TRUE)}
- return(list_zero)
-}
+ list_zero <- list()
+ for (i in 1:n_resp) {
+ list_zero[[i]] <- Matrix(0, n_obs, n_obs, sparse = TRUE)
+ }
+ return(list_zero)
+}
diff --git a/R/mc_build_omega.R b/R/mc_build_omega.R
index b6624ed..80b1b3a 100644
--- a/R/mc_build_omega.R
+++ b/R/mc_build_omega.R
@@ -9,21 +9,19 @@
#' @return A list with the \eqn{\Omega} matrix its inverse and derivatives with respect to \eqn{\tau}.
#' @export
mc_build_omega <- function(tau, Z, covariance_link, sparse = FALSE) {
- if(covariance_link == "identity") {
- Omega <- mc_matrix_linear_predictor(tau = tau, Z = Z)
- output <- list("Omega" = Omega, "D_Omega" = Z)
- }
- if(covariance_link == "expm") {
- U <- mc_matrix_linear_predictor(tau = tau, Z = Z)
- temp <- mc_expm(U = U, inverse = FALSE, sparse = sparse)
- D_Omega <- lapply(Z, mc_derivative_expm, UU = temp$UU, inv_UU = temp$inv_UU,
- Q = temp$Q, sparse = sparse)
- output <- list("Omega" = forceSymmetric(temp$Omega), "D_Omega" = D_Omega)
+ if (covariance_link == "identity") {
+ Omega <- mc_matrix_linear_predictor(tau = tau, Z = Z)
+ output <- list(Omega = Omega, D_Omega = Z)
}
- if(covariance_link == "inverse") {
- inv_Omega <- mc_matrix_linear_predictor(tau = tau, Z = Z)
- output <- list("inv_Omega" = inv_Omega, "D_inv_Omega" = Z)
- }
- return(output)
-}
-
+ if (covariance_link == "expm") {
+ U <- mc_matrix_linear_predictor(tau = tau, Z = Z)
+ temp <- mc_expm(U = U, inverse = FALSE, sparse = sparse)
+ D_Omega <- lapply(Z, mc_derivative_expm, UU = temp$UU, inv_UU = temp$inv_UU, Q = temp$Q, sparse = sparse)
+ output <- list(Omega = forceSymmetric(temp$Omega), D_Omega = D_Omega)
+ }
+ if (covariance_link == "inverse") {
+ inv_Omega <- mc_matrix_linear_predictor(tau = tau, Z = Z)
+ output <- list(inv_Omega = inv_Omega, D_inv_Omega = Z)
+ }
+ return(output)
+}
diff --git a/R/mc_build_sigma.R b/R/mc_build_sigma.R
index ed951e8..bdf47ba 100644
--- a/R/mc_build_sigma.R
+++ b/R/mc_build_sigma.R
@@ -17,163 +17,137 @@
#'@param covariance String specifing the covariance function: identity, inverse or expm.
#'@param power_fixed Logical if the power parameter is fixed at initial value (TRUE). In the case
#'power_fixed = FALSE the power parameter will be estimated.
+#'@param compute_derivative_beta Logical. Compute or not the derivative with respect to regression parameters.
#'@return A list with the Cholesky decomposition of \eqn{\Sigma}, \eqn{\Sigma^{-1}} and the derivative
#'of \eqn{\Sigma} with respect to the power and tau parameters.
#'@seealso \code{\link{mc_link_function}}, \code{\link{mc_variance_function}},
#'\code{\link{mc_build_omega}}.
#'@export
-mc_build_sigma <- function(mu, Ntrial = 1, tau, power, Z, sparse, variance,
- covariance, power_fixed, compute_derivative_beta = FALSE) {
- if(variance == "constant") {
- if(covariance == "identity" | covariance == "expm") {
- Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = covariance, sparse = sparse)
- chol_Sigma <- chol(Omega$Omega)
- inv_chol_Sigma <- solve(chol_Sigma)
- output <- list("Sigma_chol" = chol_Sigma, "Sigma_chol_inv" = inv_chol_Sigma,
- "D_Sigma" = Omega$D_Omega)
- }
- if(covariance == "inverse") {
- inv_Sigma <- mc_build_omega(tau = tau, Z = Z, covariance_link = "inverse", sparse = sparse)
- chol_inv_Sigma <- chol(inv_Sigma$inv_Omega)
- chol_Sigma <- solve(chol_inv_Sigma)
- Sigma <- (chol_Sigma)%*%t(chol_Sigma) # Because a compute the inverse of chol_inv_Omega
- D_Sigma <- lapply(inv_Sigma$D_inv_Omega, mc_sandwich_negative, bord1 = Sigma, bord2 = Sigma)
- output <- list("Sigma_chol" = t(chol_Sigma), "Sigma_chol_inv" = t(chol_inv_Sigma),
- "D_Sigma" = D_Sigma)
- }
- }
-
- if(variance == "tweedie" | variance == "binomialP" | variance == "binomialPQ"){
- if(variance == "tweedie"){variance = "power"}
- if(covariance == "identity" | covariance == "expm"){
- Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = covariance, sparse = sparse)
- V_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = variance,
- inverse = FALSE, derivative_power = !power_fixed,
- derivative_mu = compute_derivative_beta)
- Sigma <- forceSymmetric(V_sqrt$V_sqrt%*%Omega$Omega%*%V_sqrt$V_sqrt)
- chol_Sigma <- chol(Sigma)
- inv_chol_Sigma <- solve(chol_Sigma)
- D_Sigma <- lapply(Omega$D_Omega, mc_sandwich, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$V_sqrt)
- if(power_fixed == FALSE){
- if(variance == "power" | variance == "binomialP"){
- D_Sigma_power <- mc_sandwich_power(middle = Omega$Omega,
- bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_p)
- D_Sigma <- c("D_Sigma_power" = D_Sigma_power, "D_Sigma_tau" = D_Sigma)
+mc_build_sigma <- function(mu, Ntrial = 1, tau, power, Z, sparse, variance, covariance, power_fixed,
+ compute_derivative_beta = FALSE) {
+ if (variance == "constant") {
+ if (covariance == "identity" | covariance == "expm") {
+ Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = covariance, sparse = sparse)
+ chol_Sigma <- chol(Omega$Omega)
+ inv_chol_Sigma <- solve(chol_Sigma)
+ output <- list(Sigma_chol = chol_Sigma, Sigma_chol_inv = inv_chol_Sigma, D_Sigma = Omega$D_Omega)
}
- if(variance == "binomialPQ"){
- D_Sigma_p <- mc_sandwich_power(middle = Omega$Omega,
- bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_p)
- D_Sigma_q <- mc_sandwich_power(middle = Omega$Omega,
- bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_q)
- D_Sigma <- c(D_Sigma_p, D_Sigma_q,D_Sigma)
+ if (covariance == "inverse") {
+ inv_Sigma <- mc_build_omega(tau = tau, Z = Z, covariance_link = "inverse", sparse = sparse)
+ chol_inv_Sigma <- chol(inv_Sigma$inv_Omega)
+ chol_Sigma <- solve(chol_inv_Sigma)
+ Sigma <- (chol_Sigma) %*% t(chol_Sigma) # Because a compute the inverse of chol_inv_Omega
+ D_Sigma <- lapply(inv_Sigma$D_inv_Omega, mc_sandwich_negative, bord1 = Sigma, bord2 = Sigma)
+ output <- list(Sigma_chol = t(chol_Sigma), Sigma_chol_inv = t(chol_inv_Sigma), D_Sigma = D_Sigma)
}
- }
- output <- list("Sigma_chol" = chol_Sigma, "Sigma_chol_inv" = inv_chol_Sigma,
- "D_Sigma" = D_Sigma)
- if(compute_derivative_beta == TRUE){
- D_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_sqrt$D_V_sqrt_mu,
- Omega = Omega$Omega, V_sqrt = V_sqrt$V_sqrt, variance = variance)
- output$D_Sigma_beta <-D_Sigma_beta
- }
}
- if(covariance == "inverse"){
- inv_Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = "inverse", sparse = sparse)
- V_inv_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = variance,
- inverse = TRUE, derivative_power = !power_fixed,
- derivative_mu = compute_derivative_beta)
- inv_Sigma <- forceSymmetric(V_inv_sqrt$V_inv_sqrt%*%inv_Omega$inv_Omega%*%V_inv_sqrt$V_inv_sqrt)
- inv_chol_Sigma <- chol(inv_Sigma)
- chol_Sigma <- solve(inv_chol_Sigma)
- Sigma <- chol_Sigma%*%t(chol_Sigma)
- D_inv_Sigma <- lapply(inv_Omega$D_inv_Omega, mc_sandwich,
- bord1 = V_inv_sqrt$V_inv_sqrt, bord2 = V_inv_sqrt$V_inv_sqrt)
- D_Sigma <- lapply(D_inv_Sigma, mc_sandwich_negative, bord1 = Sigma, bord2 = Sigma)
- if(power_fixed == FALSE){
- if(variance == "power" | variance == "binomialP"){
- D_Omega_p <- mc_sandwich_power(middle = inv_Omega$inv_Omega,
- bord1 = V_inv_sqrt$V_inv_sqrt,
- bord2 = V_inv_sqrt$D_V_inv_sqrt_power)
- D_Sigma_p <- mc_sandwich_negative(middle = D_Omega_p, bord1 = Sigma, bord2 = Sigma)
- D_Sigma <- c(D_Sigma_p, D_Sigma)
+
+ if (variance == "tweedie" | variance == "binomialP" | variance == "binomialPQ") {
+ if (variance == "tweedie") {
+ variance = "power"
}
- if(variance == "binomialPQ"){
- D_Omega_p <- mc_sandwich_power(middle = inv_Omega$inv_Omega,
- bord1 = V_inv_sqrt$V_inv_sqrt,
- bord2 = V_inv_sqrt$D_V_inv_sqrt_p)
- D_Sigma_p <- mc_sandwich_negative(middle = D_Omega_p, bord1 = Sigma, bord2 = Sigma)
- D_Omega_q <- mc_sandwich_power(middle = inv_Omega$inv_Omega,
- bord1 = V_inv_sqrt$V_inv_sqrt,
- bord2 = V_inv_sqrt$D_V_inv_sqrt_q)
- D_Sigma_q <- mc_sandwich_negative(middle = D_Omega_q, bord1 = Sigma, bord2 = Sigma)
- D_Sigma <- c(D_Sigma_p, D_Sigma_q, D_Sigma)
+ if (covariance == "identity" | covariance == "expm") {
+ Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = covariance, sparse = sparse)
+ V_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = variance, inverse = FALSE,
+ derivative_power = !power_fixed, derivative_mu = compute_derivative_beta)
+ Sigma <- forceSymmetric(V_sqrt$V_sqrt %*% Omega$Omega %*% V_sqrt$V_sqrt)
+ chol_Sigma <- chol(Sigma)
+ inv_chol_Sigma <- solve(chol_Sigma)
+ D_Sigma <- lapply(Omega$D_Omega, mc_sandwich, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$V_sqrt)
+ if (power_fixed == FALSE) {
+ if (variance == "power" | variance == "binomialP") {
+ D_Sigma_power <- mc_sandwich_power(middle = Omega$Omega, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_p)
+ D_Sigma <- c(D_Sigma_power = D_Sigma_power, D_Sigma_tau = D_Sigma)
+ }
+ if (variance == "binomialPQ") {
+ D_Sigma_p <- mc_sandwich_power(middle = Omega$Omega, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_p)
+ D_Sigma_q <- mc_sandwich_power(middle = Omega$Omega, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_q)
+ D_Sigma <- c(D_Sigma_p, D_Sigma_q, D_Sigma)
+ }
+ }
+ output <- list(Sigma_chol = chol_Sigma, Sigma_chol_inv = inv_chol_Sigma, D_Sigma = D_Sigma)
+ if (compute_derivative_beta == TRUE) {
+ D_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_sqrt$D_V_sqrt_mu, Omega = Omega$Omega,
+ V_sqrt = V_sqrt$V_sqrt, variance = variance)
+ output$D_Sigma_beta <- D_Sigma_beta
+ }
+ }
+ if (covariance == "inverse") {
+ inv_Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = "inverse", sparse = sparse)
+ V_inv_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = variance, inverse = TRUE,
+ derivative_power = !power_fixed, derivative_mu = compute_derivative_beta)
+ inv_Sigma <- forceSymmetric(V_inv_sqrt$V_inv_sqrt %*% inv_Omega$inv_Omega %*% V_inv_sqrt$V_inv_sqrt)
+ inv_chol_Sigma <- chol(inv_Sigma)
+ chol_Sigma <- solve(inv_chol_Sigma)
+ Sigma <- chol_Sigma %*% t(chol_Sigma)
+ D_inv_Sigma <- lapply(inv_Omega$D_inv_Omega, mc_sandwich, bord1 = V_inv_sqrt$V_inv_sqrt, bord2 = V_inv_sqrt$V_inv_sqrt)
+ D_Sigma <- lapply(D_inv_Sigma, mc_sandwich_negative, bord1 = Sigma, bord2 = Sigma)
+ if (power_fixed == FALSE) {
+ if (variance == "power" | variance == "binomialP") {
+ D_Omega_p <- mc_sandwich_power(middle = inv_Omega$inv_Omega, bord1 = V_inv_sqrt$V_inv_sqrt, bord2 = V_inv_sqrt$D_V_inv_sqrt_power)
+ D_Sigma_p <- mc_sandwich_negative(middle = D_Omega_p, bord1 = Sigma, bord2 = Sigma)
+ D_Sigma <- c(D_Sigma_p, D_Sigma)
+ }
+ if (variance == "binomialPQ") {
+ D_Omega_p <- mc_sandwich_power(middle = inv_Omega$inv_Omega, bord1 = V_inv_sqrt$V_inv_sqrt, bord2 = V_inv_sqrt$D_V_inv_sqrt_p)
+ D_Sigma_p <- mc_sandwich_negative(middle = D_Omega_p, bord1 = Sigma, bord2 = Sigma)
+ D_Omega_q <- mc_sandwich_power(middle = inv_Omega$inv_Omega, bord1 = V_inv_sqrt$V_inv_sqrt, bord2 = V_inv_sqrt$D_V_inv_sqrt_q)
+ D_Sigma_q <- mc_sandwich_negative(middle = D_Omega_q, bord1 = Sigma, bord2 = Sigma)
+ D_Sigma <- c(D_Sigma_p, D_Sigma_q, D_Sigma)
+ }
+ }
+ output <- list(Sigma_chol = t(chol_Sigma), Sigma_chol_inv = t(inv_chol_Sigma), D_Sigma = D_Sigma)
+ if (compute_derivative_beta == TRUE) {
+ D_inv_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_inv_sqrt$D_V_inv_sqrt_mu, Omega = inv_Omega$inv_Omega,
+ V_sqrt = V_inv_sqrt$V_inv_sqrt, variance = variance)
+ D_Sigma_beta <- lapply(D_inv_Sigma_beta, mc_sandwich_negative, bord1 = Sigma, bord2 = Sigma)
+ output$D_Sigma_beta <- D_Sigma_beta
+ }
}
- }
- output <- list("Sigma_chol" = t(chol_Sigma), "Sigma_chol_inv" = t(inv_chol_Sigma),
- "D_Sigma" = D_Sigma)
- if(compute_derivative_beta == TRUE){
- D_inv_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D,
- D_V_sqrt_mu = V_inv_sqrt$D_V_inv_sqrt_mu,
- Omega = inv_Omega$inv_Omega,
- V_sqrt = V_inv_sqrt$V_inv_sqrt,
- variance = variance)
- D_Sigma_beta <- lapply(D_inv_Sigma_beta, mc_sandwich_negative, bord1 = Sigma, bord2 = Sigma)
- output$D_Sigma_beta <- D_Sigma_beta
- }
- }
}
- if(variance == "poisson_tweedie"){
- if(covariance == "identity" | covariance == "expm"){
- Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = covariance, sparse = sparse)
- V_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = "power",
- inverse = FALSE, derivative_power = !power_fixed,
- derivative_mu = compute_derivative_beta)
- Sigma <- Diagonal(length(mu$mu), mu$mu) +
- V_sqrt$V_sqrt%*%Omega$Omega%*%V_sqrt$V_sqrt
- chol_Sigma <- chol(Sigma)
- inv_chol_Sigma <- solve(chol_Sigma)
- D_Sigma <- lapply(Omega$D_Omega, mc_sandwich, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$V_sqrt)
- if(power_fixed == FALSE){
- D_Sigma_power <- mc_sandwich_power(middle = Omega$Omega,
- bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_p)
- D_Sigma <- c("D_Sigma_power" = D_Sigma_power, "D_Sigma_tau" = D_Sigma)
- }
- output <- list("Sigma_chol" = chol_Sigma, "Sigma_chol_inv" = inv_chol_Sigma,
- "D_Sigma" = D_Sigma)
- if(compute_derivative_beta == TRUE){
- D_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_sqrt$D_V_sqrt_mu,
- Omega$Omega, V_sqrt = V_sqrt$V_sqrt,
- variance = variance)
- output$D_Sigma_beta <-D_Sigma_beta
+ if (variance == "poisson_tweedie") {
+ if (covariance == "identity" | covariance == "expm") {
+ Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = covariance, sparse = sparse)
+ V_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = "power", inverse = FALSE,
+ derivative_power = !power_fixed, derivative_mu = compute_derivative_beta)
+ Sigma <- forceSymmetric(Diagonal(length(mu$mu), mu$mu) + V_sqrt$V_sqrt %*% Omega$Omega %*% V_sqrt$V_sqrt)
+ chol_Sigma <- chol(Sigma)
+ inv_chol_Sigma <- solve(chol_Sigma)
+ D_Sigma <- lapply(Omega$D_Omega, mc_sandwich, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$V_sqrt)
+ if (power_fixed == FALSE) {
+ D_Sigma_power <- mc_sandwich_power(middle = Omega$Omega, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_p)
+ D_Sigma <- c(D_Sigma_power = D_Sigma_power, D_Sigma_tau = D_Sigma)
+ }
+ output <- list(Sigma_chol = chol_Sigma, Sigma_chol_inv = inv_chol_Sigma, D_Sigma = D_Sigma)
+ if (compute_derivative_beta == TRUE) {
+ D_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_sqrt$D_V_sqrt_mu, Omega$Omega, V_sqrt = V_sqrt$V_sqrt,
+ variance = variance)
+ output$D_Sigma_beta <- D_Sigma_beta
+ }
+ }
+ if (covariance == "inverse") {
+ inv_Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = "inverse", sparse = sparse)
+ Omega <- chol2inv(chol(inv_Omega$inv_Omega))
+ V_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = "power", inverse = FALSE,
+ derivative_power = !power_fixed, derivative_mu = compute_derivative_beta)
+ D_Omega <- lapply(inv_Omega$D_inv_Omega, mc_sandwich_negative, bord1 = Omega, bord2 = Omega)
+ D_Sigma <- lapply(D_Omega, mc_sandwich, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$V_sqrt)
+ Sigma <- forceSymmetric(Diagonal(length(mu$mu), mu$mu) + V_sqrt$V_sqrt %*% Omega %*% V_sqrt$V_sqrt)
+ chol_Sigma <- chol(Sigma)
+ inv_chol_Sigma <- solve(chol_Sigma)
+ if (power_fixed == FALSE) {
+ D_Sigma_p <- mc_sandwich_power(middle = Omega, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_power)
+ D_Sigma <- c(D_Sigma_p, D_Sigma)
+ }
+ output <- list(Sigma_chol = chol_Sigma, Sigma_chol_inv = inv_chol_Sigma, D_Sigma = D_Sigma)
+ if (compute_derivative_beta == TRUE) {
+ D_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_sqrt$D_V_sqrt_mu, Omega = Omega, V_sqrt = V_sqrt$V_sqrt,
+ variance = variance)
+ output$D_Sigma_beta <- D_Sigma_beta
+ }
}
}
- if(covariance == "inverse"){
- inv_Omega <- mc_build_omega(tau = tau, Z = Z, covariance_link = "inverse", sparse = sparse)
- Omega <- chol2inv(chol(inv_Omega$inv_Omega))
- V_sqrt <- mc_variance_function(mu = mu$mu, power = power, Ntrial = Ntrial, variance = "power",
- inverse = FALSE, derivative_power = !power_fixed,
- derivative_mu = compute_derivative_beta)
- D_Omega <- lapply(inv_Omega$D_inv_Omega, mc_sandwich_negative, bord1 = Omega, bord2 = Omega)
- D_Sigma <- lapply(D_Omega, mc_sandwich, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$V_sqrt)
- Sigma <- forceSymmetric(Diagonal(length(mu$mu), mu$mu) +
- V_sqrt$V_sqrt%*%Omega%*%V_sqrt$V_sqrt)
- chol_Sigma <- chol(Sigma)
- inv_chol_Sigma <- solve(chol_Sigma)
- if(power_fixed == FALSE){
- D_Sigma_p <- mc_sandwich_power(middle = Omega, bord1 = V_sqrt$V_sqrt, bord2 = V_sqrt$D_V_sqrt_power)
- D_Sigma <- c(D_Sigma_p, D_Sigma)
- }
- output <- list("Sigma_chol" = chol_Sigma, "Sigma_chol_inv" = inv_chol_Sigma,
- "D_Sigma" = D_Sigma)
- if(compute_derivative_beta == TRUE){
- D_Sigma_beta <- mc_derivative_sigma_beta(D = mu$D, D_V_sqrt_mu = V_sqrt$D_V_sqrt_mu,
- Omega = Omega, V_sqrt = V_sqrt$V_sqrt,
- variance = variance)
- output$D_Sigma_beta <- D_Sigma_beta
- }
- }
- }
- return(output)
+ return(output)
}
-
diff --git a/R/mc_build_sigmab.R b/R/mc_build_sigmab.R
index f5554c6..446df48 100644
--- a/R/mc_build_sigmab.R
+++ b/R/mc_build_sigmab.R
@@ -8,35 +8,34 @@
#' @return A list with sigmab and its derivatives with respect to rho.
#' @export
-mc_build_sigma_between <- function(rho, n_resp, inverse = FALSE){
- output <- list("Sigmab" = 1, "D_Sigmab" = 1)
- if(n_resp > 1){
- Sigmab <- Diagonal(n_resp, 1)
- Sigmab[lower.tri(Sigmab)] <- rho
- Sigmab <- forceSymmetric(t(Sigmab))
- D_Sigmab <- mc_derivative_sigma_between(n_resp = n_resp)
- if(inverse == FALSE){
- output <- list("Sigmab" = Sigmab, "D_Sigmab" = D_Sigmab)
- }
- if(inverse == TRUE){
- inv_Sigmab <- solve(Sigmab)
- D_inv_Sigmab <- lapply(D_Sigmab, mc_sandwich_negative, bord1 = inv_Sigmab, bord2 = inv_Sigmab)
- output <- list("inv_Sigmab" = inv_Sigmab, "D_inv_Sigmab" = D_inv_Sigmab)
- }
- }
- return(output)
+mc_build_sigma_between <- function(rho, n_resp, inverse = FALSE) {
+ output <- list(Sigmab = 1, D_Sigmab = 1)
+ if (n_resp > 1) {
+ Sigmab <- Diagonal(n_resp, 1)
+ Sigmab[lower.tri(Sigmab)] <- rho
+ Sigmab <- forceSymmetric(t(Sigmab))
+ D_Sigmab <- mc_derivative_sigma_between(n_resp = n_resp)
+ if (inverse == FALSE) {
+ output <- list(Sigmab = Sigmab, D_Sigmab = D_Sigmab)
+ }
+ if (inverse == TRUE) {
+ inv_Sigmab <- solve(Sigmab)
+ D_inv_Sigmab <- lapply(D_Sigmab, mc_sandwich_negative, bord1 = inv_Sigmab, bord2 = inv_Sigmab)
+ output <- list(inv_Sigmab = inv_Sigmab, D_inv_Sigmab = D_inv_Sigmab)
+ }
+ }
+ return(output)
}
#' @rdname mc_build_sigma_between
-mc_derivative_sigma_between <- function(n_resp){
- position <- combn(n_resp,2)
- list.Derivative <- list()
- n_par <- n_resp*(n_resp-1)/2
- for(i in 1:n_par){
- Derivative <- Matrix(0, ncol = n_resp, nrow = n_resp)
- Derivative[position[1,i],position[2,i]] <- Derivative[position[2,i],position[1,i]] <- 1
- list.Derivative[i][[1]] <- Derivative}
- return(list.Derivative)
-}
-
-
+mc_derivative_sigma_between <- function(n_resp) {
+ position <- combn(n_resp, 2)
+ list.Derivative <- list()
+ n_par <- n_resp * (n_resp - 1)/2
+ for (i in 1:n_par) {
+ Derivative <- Matrix(0, ncol = n_resp, nrow = n_resp)
+ Derivative[position[1, i], position[2, i]] <- Derivative[position[2, i], position[1, i]] <- 1
+ list.Derivative[i][[1]] <- Derivative
+ }
+ return(list.Derivative)
+}
diff --git a/R/mc_coef.R b/R/mc_coef.R
index a9c1a3e..6f7a5b4 100644
--- a/R/mc_coef.R
+++ b/R/mc_coef.R
@@ -2,77 +2,70 @@
#'
#' coef.mcglm is a function which extracts model coefficients from objects of mcglm class.
#' @param object An object of mcglm class.
+#' @param std.error Logical. Returns or not the standard errors.
#' @param response A numeric or vector specyfing for which response variables the coefficients
#' should be returned.
#' @param type A string or string vector specyfing which coefficients should be returned.
-#' Options are "beta", "tau", "power", "tau" and "correlation".
+#' Options are 'beta', 'tau', 'power', 'tau' and 'correlation'.
#' @return A data.frame with estimates, parameters names, response number and parameters type.
-#' @exportMethod
-coef.mcglm <- function(object, std.error = FALSE, response = c(NA,1:length(object$beta_names)),
- type = c("beta","tau","power","correlation")) {
- n_resp <- length(object$beta_names)
- cod_beta <- list()
- cod_power <- list()
- cod_tau <- list()
- type_beta <- list()
- type_power <- list()
- type_tau <- list()
- resp_beta <- list()
- resp_power <- list()
- resp_tau <- list()
- response_for <- 1:n_resp
- for(i in response_for) {
- cod_beta[[i]] <- paste(paste("beta", i, sep = ""), 0:c(object$Information$n_betas[[i]]-1), sep = "")
- type_beta[[i]] <- rep("beta", length(cod_beta[[i]]))
- resp_beta[[i]] <- rep(response_for[i], length(cod_beta[[i]]))
- if(object$Information$n_power[[i]] != 0 | object$power_fixed[[i]] == FALSE){
- cod_power[[i]] <- paste(paste("power", i, sep = ""), 1:object$Information$n_power[[i]], sep = "")
- type_power[[i]] <- rep("power", length(cod_power[[i]]))
- resp_power[[i]] <- rep(response_for[i], length(cod_power[[i]]))
+#' @export
+coef.mcglm <- function(object, std.error = FALSE, response = c(NA, 1:length(object$beta_names)), type = c("beta", "tau",
+ "power", "correlation")) {
+ n_resp <- length(object$beta_names)
+ cod_beta <- list()
+ cod_power <- list()
+ cod_tau <- list()
+ type_beta <- list()
+ type_power <- list()
+ type_tau <- list()
+ resp_beta <- list()
+ resp_power <- list()
+ resp_tau <- list()
+ response_for <- 1:n_resp
+ for (i in response_for) {
+ cod_beta[[i]] <- paste(paste("beta", i, sep = ""), 0:c(object$Information$n_betas[[i]] - 1), sep = "")
+ type_beta[[i]] <- rep("beta", length(cod_beta[[i]]))
+ resp_beta[[i]] <- rep(response_for[i], length(cod_beta[[i]]))
+ if (object$Information$n_power[[i]] != 0 | object$power_fixed[[i]] == FALSE) {
+ cod_power[[i]] <- paste(paste("power", i, sep = ""), 1:object$Information$n_power[[i]], sep = "")
+ type_power[[i]] <- rep("power", length(cod_power[[i]]))
+ resp_power[[i]] <- rep(response_for[i], length(cod_power[[i]]))
+ }
+ if (object$Information$n_power[[i]] == 0) {
+ cod_power[[i]] <- rep(1, 0)
+ type_power[[i]] <- rep(1, 0)
+ resp_power[[i]] <- rep(1, 0)
+ }
+ cod_tau[[i]] <- paste(paste("tau", i, sep = ""), 1:object$Information$n_tau[[i]], sep = "")
+ type_tau[[i]] <- rep("tau", length(cod_tau[[i]]))
+ resp_tau[[i]] <- rep(response_for[i], length(cod_tau[[i]]))
}
- if(object$Information$n_power[[i]] == 0){
- cod_power[[i]] <- rep(1, 0)
- type_power[[i]] <- rep(1, 0)
- resp_power[[i]] <- rep(1,0)
+ rho_names <- c()
+ if (n_resp != 1) {
+ combination <- combn(n_resp, 2)
+ for (i in 1:dim(combination)[2]) {
+ rho_names[i] <- paste(paste("rho", combination[1, i], sep = ""), combination[2, i], sep = "")
+ }
}
- cod_tau[[i]] <- paste(paste("tau", i, sep = ""), 1:object$Information$n_tau[[i]], sep = "")
- type_tau[[i]] <- rep("tau", length(cod_tau[[i]]))
- resp_tau[[i]] <- rep(response_for[i], length(cod_tau[[i]]))
- }
- rho_names <- c()
- if(n_resp != 1) {
- combination <- combn(n_resp,2)
- for(i in 1:dim(combination)[2]) {
- rho_names[i] <- paste(paste("rho", combination[1,i], sep = ""), combination[2,i], sep = "")
- }
- }
- type_rho <- rep("correlation", length(rho_names))
- resp_rho <- rep(NA, length(rho_names))
- cod <- c(do.call(c,cod_beta),rho_names,
- do.call(c,Map(c, cod_tau)))
- type_cod <- c(do.call(c,type_beta),type_rho,
- do.call(c,Map(c, type_tau)))
- response_cod <- c(do.call(c,resp_beta),resp_rho,
- do.call(c,Map(c, resp_tau)))
+ type_rho <- rep("correlation", length(rho_names))
+ resp_rho <- rep(NA, length(rho_names))
+ cod <- c(do.call(c, cod_beta), rho_names, do.call(c, Map(c, cod_tau)))
+ type_cod <- c(do.call(c, type_beta), type_rho, do.call(c, Map(c, type_tau)))
+ response_cod <- c(do.call(c, resp_beta), resp_rho, do.call(c, Map(c, resp_tau)))
- if(length(cod_power) != 0) {
- cod <- c(do.call(c,cod_beta),rho_names,
- do.call(c,Map(c,cod_power, cod_tau)))
- type_cod <- c(do.call(c,type_beta),type_rho,
- do.call(c,Map(c,type_power, type_tau)))
- response_cod <- c(do.call(c,resp_beta),resp_rho,
- do.call(c,Map(c,resp_power, resp_tau)))
- }
+ if (length(cod_power) != 0) {
+ cod <- c(do.call(c, cod_beta), rho_names, do.call(c, Map(c, cod_power, cod_tau)))
+ type_cod <- c(do.call(c, type_beta), type_rho, do.call(c, Map(c, type_power, type_tau)))
+ response_cod <- c(do.call(c, resp_beta), resp_rho, do.call(c, Map(c, resp_power, resp_tau)))
+ }
- Estimates <- c(object$Regression, object$Covariance)
- coef_temp <- data.frame("Estimates" = Estimates, "Parameters" = cod,
- "Type" = type_cod, "Response" = response_cod)
- if(std.error == TRUE) {
- coef_temp <- data.frame("Estimates" = Estimates, "Std.error" = sqrt(diag(object$vcov)),
- "Parameters" = cod,
- "Type" = type_cod, "Response" = response_cod)
- }
+ Estimates <- c(object$Regression, object$Covariance)
+ coef_temp <- data.frame(Estimates = Estimates, Parameters = cod, Type = type_cod, Response = response_cod)
+ if (std.error == TRUE) {
+ coef_temp <- data.frame(Estimates = Estimates, Std.error = sqrt(diag(object$vcov)), Parameters = cod, Type = type_cod,
+ Response = response_cod)
+ }
- output <- coef_temp[which(coef_temp$Response %in% response & coef_temp$Type %in% type),]
- return(output)
+ output <- coef_temp[which(coef_temp$Response %in% response & coef_temp$Type %in% type), ]
+ return(output)
}
diff --git a/R/mc_confint.mcglm.R b/R/mc_confint.mcglm.R
index e6f4453..457b52a 100644
--- a/R/mc_confint.mcglm.R
+++ b/R/mc_confint.mcglm.R
@@ -8,12 +8,12 @@
#' response number and parameters type.
#' @export
confint.mcglm <- function(object, level = 0.95) {
- temp <- coef(object,std.error = TRUE)
- a <- (1 - level)/2
- a <- c(a, 1 - a)
- fac <- qnorm(a)
- ci <- temp$Estimates + temp$Std.error%o%fac
- colnames(ci) <- paste(format(a,2),"%", sep = "")
- rownames(ci) <- temp$Parameters
- return(ci)
-}
+ temp <- coef(object, std.error = TRUE)
+ a <- (1 - level)/2
+ a <- c(a, 1 - a)
+ fac <- qnorm(a)
+ ci <- temp$Estimates + temp$Std.error %o% fac
+ colnames(ci) <- paste(format(a, 2), "%", sep = "")
+ rownames(ci) <- temp$Parameters
+ return(ci)
+}
diff --git a/R/mc_core_cross_variability.R b/R/mc_core_cross_variability.R
index 99837aa..610bff7 100644
--- a/R/mc_core_cross_variability.R
+++ b/R/mc_core_cross_variability.R
@@ -6,8 +6,8 @@
#' @param A A matrix.
#' @param res A vector of residuals.
#' @param W A matrix of weights.
-covprod <- function(A, res, W){
- res =as.numeric(res)
- saida <- (res%*%W%*%res)%*%(t(res)%*%A)
- return(as.numeric(saida))
-}
+covprod <- function(A, res, W) {
+ res = as.numeric(res)
+ saida <- (res %*% W %*% res) %*% (t(res) %*% A)
+ return(as.numeric(saida))
+}
diff --git a/R/mc_core_pearson.R b/R/mc_core_pearson.R
index 9533451..82e1693 100644
--- a/R/mc_core_pearson.R
+++ b/R/mc_core_pearson.R
@@ -9,7 +9,6 @@
#' @export
mc_core_pearson <- function(product, inv_C, res) {
- output <- t(res)%*%product%*%(inv_C%*%res) - sum(diag(product))
- return(as.numeric(output))
-}
-
+ output <- t(res) %*% product %*% (inv_C %*% res) - sum(diag(product))
+ return(as.numeric(output))
+}
diff --git a/R/mc_correction.R b/R/mc_correction.R
index 5212d9f..b57edce 100644
--- a/R/mc_correction.R
+++ b/R/mc_correction.R
@@ -12,7 +12,7 @@
#' @export
mc_correction <- function(D_C, inv_J_beta, D, inv_C) {
- term1 <- lapply(D_C, mc_sandwich, bord1 = t(D)%*%inv_C, bord2 = inv_C%*%D)
- output <- lapply(term1, function(x, inv_J_beta) sum(x*inv_J_beta), inv_J_beta = inv_J_beta)
- return(unlist(output))
-}
+ term1 <- lapply(D_C, mc_sandwich, bord1 = t(D) %*% inv_C, bord2 = inv_C %*% D)
+ output <- lapply(term1, function(x, inv_J_beta) sum(x * inv_J_beta), inv_J_beta = inv_J_beta)
+ return(unlist(output))
+}
diff --git a/R/mc_cross_sensitivity.R b/R/mc_cross_sensitivity.R
index 86092ad..23a3bf2 100644
--- a/R/mc_cross_sensitivity.R
+++ b/R/mc_cross_sensitivity.R
@@ -4,23 +4,23 @@
#' Equation 10 of Bonat and Jorgensen (2015).
#' @param Product_cov A list of matrices.
#' @param Product_beta A list of matrices.
+#' @param n_beta_effective Numeric. Effective number of regression parameters.
#' @return The cross-sensitivity matrix. Equation (10) of Bonat and Jorgensen (2015).
#' @export
-mc_cross_sensitivity <- function(Product_cov, Product_beta,
- n_beta_effective = length(Product_beta)) {
- n_beta <- length(Product_beta)
- n_cov <- length(Product_cov)
- if(n_beta == 0) {
- cross_sensitivity <- Matrix(0, ncol = n_beta_effective, nrow = n_cov)
- }
- if(n_beta !=0) {
- cross_sensitivity <- Matrix(NA, nrow = n_cov, ncol = n_beta)
- for(i in 1:n_cov) {
- for(j in 1:n_beta) {
- cross_sensitivity[i,j] <- -sum(Product_cov[[i]]*Product_beta[[j]])
+
+mc_cross_sensitivity <- function(Product_cov, Product_beta, n_beta_effective = length(Product_beta)) {
+ n_beta <- length(Product_beta)
+ n_cov <- length(Product_cov)
+ if (n_beta == 0) {
+ cross_sensitivity <- Matrix(0, ncol = n_beta_effective, nrow = n_cov)
+ }
+ if (n_beta != 0) {
+ cross_sensitivity <- Matrix(NA, nrow = n_cov, ncol = n_beta)
+ for (i in 1:n_cov) {
+ for (j in 1:n_beta) {
+ cross_sensitivity[i, j] <- -sum(Product_cov[[i]] * Product_beta[[j]])
+ }
+ }
}
- }
- }
- return(cross_sensitivity)
+ return(cross_sensitivity)
}
-
diff --git a/R/mc_cross_variability.R b/R/mc_cross_variability.R
index 6f9013f..de7663a 100644
--- a/R/mc_cross_variability.R
+++ b/R/mc_cross_variability.R
@@ -9,15 +9,15 @@
#' @return The cross-variability matrix between regression and covariance parameters.
#' @export
mc_cross_variability <- function(Product_cov, inv_C, res, D) {
- Wlist <- lapply(Product_cov, mc_multiply2, bord2 = inv_C)
- A = t(D)%*%inv_C
- n_beta <- dim(A)[1]
- n_cov <- length(Product_cov)
- cross_variability <- Matrix(NA, ncol = n_cov, nrow = n_beta)
- for(j in 1:n_beta) {
- for(i in 1:n_cov) {
- cross_variability[j,i] <- covprod(A[j,], Wlist[[i]], r = res)
+ Wlist <- lapply(Product_cov, mc_multiply2, bord2 = inv_C)
+ A = t(D) %*% inv_C
+ n_beta <- dim(A)[1]
+ n_cov <- length(Product_cov)
+ cross_variability <- Matrix(NA, ncol = n_cov, nrow = n_beta)
+ for (j in 1:n_beta) {
+ for (i in 1:n_cov) {
+ cross_variability[j, i] <- covprod(A[j, ], Wlist[[i]], res = res)
+ }
}
- }
- return(cross_variability)
+ return(cross_variability)
}
diff --git a/R/mc_derivative_C_rho.R b/R/mc_derivative_C_rho.R
index 04b4fa2..c31bd0a 100644
--- a/R/mc_derivative_C_rho.R
+++ b/R/mc_derivative_C_rho.R
@@ -4,19 +4,14 @@
#'
#'@param D_Sigmab A matrix.
#'@param Bdiag_chol_Sigma_within A block-diagonal matrix.
-#'@param t_Bdiag_chol_sigma_within A block-diagonal matrix.
+#'@param t_Bdiag_chol_Sigma_within A block-diagonal matrix.
#'@param II A diagonal matrix.
#'@return A matrix.
#'@details It is an internal function used to build the derivatives of the C matrix.
#'@export
-mc_derivative_C_rho <- function(D_Sigmab, Bdiag_chol_Sigma_within, t_Bdiag_chol_Sigma_within, II){
- output <- lapply(D_Sigmab, function(x){
- t_Bdiag_chol_Sigma_within%*%kronecker(x,II)%*%Bdiag_chol_Sigma_within
- }
- )
- return(output)
+mc_derivative_C_rho <- function(D_Sigmab, Bdiag_chol_Sigma_within, t_Bdiag_chol_Sigma_within, II) {
+ output <- lapply(D_Sigmab, function(x) {
+ t_Bdiag_chol_Sigma_within %*% kronecker(x, II) %*% Bdiag_chol_Sigma_within
+ })
+ return(output)
}
-
-
-
-
diff --git a/R/mc_derivative_cholesky.R b/R/mc_derivative_cholesky.R
index 971e5c4..5dd3d3e 100644
--- a/R/mc_derivative_cholesky.R
+++ b/R/mc_derivative_cholesky.R
@@ -8,14 +8,13 @@
#' @details It is an internal function.
#' @export
mc_derivative_cholesky <- function(derivada, inv_chol_Sigma, chol_Sigma) {
- faux <- function(derivada,inv_chol_Sigma, chol_Sigma){
- t1 <- inv_chol_Sigma%*%derivada%*%t(inv_chol_Sigma)
- t1 <- tril(t1)
- diag(t1)<- diag(t1)/2
- output <- chol_Sigma%*%t1
- return(output)
- }
- list_D_chol <- lapply(derivada, faux, inv_chol_Sigma = inv_chol_Sigma, chol_Sigma = chol_Sigma)
- return(list_D_chol)
-}
-
+ faux <- function(derivada, inv_chol_Sigma, chol_Sigma) {
+ t1 <- inv_chol_Sigma %*% derivada %*% t(inv_chol_Sigma)
+ t1 <- tril(t1)
+ diag(t1) <- diag(t1)/2
+ output <- chol_Sigma %*% t1
+ return(output)
+ }
+ list_D_chol <- lapply(derivada, faux, inv_chol_Sigma = inv_chol_Sigma, chol_Sigma = chol_Sigma)
+ return(list_D_chol)
+}
diff --git a/R/mc_derivative_expm.R b/R/mc_derivative_expm.R
index 6037510..dad9c66 100644
--- a/R/mc_derivative_expm.R
+++ b/R/mc_derivative_expm.R
@@ -15,13 +15,13 @@
#' The argument dU is the derivative of the U matrix with respect to the models parameters. It should
#' be computed by the user.
#' @export
-mc_derivative_expm <- function(dU, UU, inv_UU, Q, n = dim(UU)[1], sparse = FALSE){
- H = inv_UU%*%dU%*%UU
- P <- Matrix(0, ncol=n,nrow=n)
- diag(P) <- diag(H)*exp(Q)
- P[upper.tri(P)] <- H[upper.tri(H)]*c(dist(exp(Q))/dist(Q))
- P[is.na(P)] <- 0
- P <- forceSymmetric(P)
- D_Omega = Matrix(UU%*%P%*%inv_UU,sparse = FALSE)
- return(D_Omega)
-}
+mc_derivative_expm <- function(dU, UU, inv_UU, Q, n = dim(UU)[1], sparse = FALSE) {
+ H = inv_UU %*% dU %*% UU
+ P <- Matrix(0, ncol = n, nrow = n)
+ diag(P) <- diag(H) * exp(Q)
+ P[upper.tri(P)] <- H[upper.tri(H)] * c(dist(exp(Q))/dist(Q))
+ P[is.na(P)] <- 0
+ P <- forceSymmetric(P)
+ D_Omega = Matrix(UU %*% P %*% inv_UU, sparse = FALSE)
+ return(D_Omega)
+}
diff --git a/R/mc_derivative_sigma_beta.R b/R/mc_derivative_sigma_beta.R
index e539fe3..5757aa1 100644
--- a/R/mc_derivative_sigma_beta.R
+++ b/R/mc_derivative_sigma_beta.R
@@ -11,22 +11,20 @@
#' parameters.
#' @export
mc_derivative_sigma_beta <- function(D, D_V_sqrt_mu, Omega, V_sqrt, variance) {
- n_beta <- dim(D)[2]
- n_obs <- dim(D)[1]
- output <- list()
- if(variance == "power" | variance == "binomialP" | variance == "binomialPQ"){
- for(i in 1:n_beta) {
- D_V_sqrt_beta <- Diagonal(n_obs, D_V_sqrt_mu*D[,i])
- output[[i]] <- mc_sandwich_power(middle = Omega,
- bord1 = V_sqrt, bord2 = D_V_sqrt_beta)
+ n_beta <- dim(D)[2]
+ n_obs <- dim(D)[1]
+ output <- list()
+ if (variance == "power" | variance == "binomialP" | variance == "binomialPQ") {
+ for (i in 1:n_beta) {
+ D_V_sqrt_beta <- Diagonal(n_obs, D_V_sqrt_mu * D[, i])
+ output[[i]] <- mc_sandwich_power(middle = Omega, bord1 = V_sqrt, bord2 = D_V_sqrt_beta)
+ }
}
- }
- if(variance == "poisson_tweedie") {
- for(i in 1:n_beta){
- D_V_sqrt_beta <- Diagonal(n_obs, D_V_sqrt_mu*D[,i])
- output[[i]] <- Diagonal(n_obs, D[,i]) + mc_sandwich_power(middle = Omega,
- bord1 = V_sqrt, bord2 = D_V_sqrt_beta)
- }
+ if (variance == "poisson_tweedie") {
+ for (i in 1:n_beta) {
+ D_V_sqrt_beta <- Diagonal(n_obs, D_V_sqrt_mu * D[, i])
+ output[[i]] <- Diagonal(n_obs, D[, i]) + mc_sandwich_power(middle = Omega, bord1 = V_sqrt, bord2 = D_V_sqrt_beta)
+ }
}
- return(output)
-}
+ return(output)
+}
diff --git a/R/mc_dexp_gold.R b/R/mc_dexp_gold.R
index 7bf77d2..4f241e5 100644
--- a/R/mc_dexp_gold.R
+++ b/R/mc_dexp_gold.R
@@ -14,11 +14,11 @@
#' dM <- matrix(c(0,1,1,0),2,2)
#' mc_dexp_gold(M = M, dM = dM)
#' @export
-mc_dexp_gold <- function(M,dM) {
- N = dim(M)
- AM = rbind(cbind(M,matrix(0,N[1],N[2])),cbind(dM,M))
- PSI = expm(AM)
- F = PSI[1:N[1],1:N[1]]
- dF = PSI[c(N[1]+1):c(2*N[1]), 1:N[1]]
- return(list(F,dF))
-}
+mc_dexp_gold <- function(M, dM) {
+ N = dim(M)
+ AM = rbind(cbind(M, matrix(0, N[1], N[2])), cbind(dM, M))
+ PSI = expm(AM)
+ F = PSI[1:N[1], 1:N[1]]
+ dF = PSI[c(N[1] + 1):c(2 * N[1]), 1:N[1]]
+ return(list(F, dF))
+}
diff --git a/R/mc_dexpm.R b/R/mc_dexpm.R
index 83d396d..8623496 100644
--- a/R/mc_dexpm.R
+++ b/R/mc_dexpm.R
@@ -5,33 +5,32 @@
#'its derivatives. This function is based on the eigen-value decomposition it means that it is
#'very slow.
#'
-#' @param U A matrix.
-#' @param n A number specifing the dimension of the matrix U. Default \code{n = dim(U)[1]}.
-#' @param sparse Logical defining the class of the output matrix. If \code{sparse = TRUE} the output
-#' class will be 'dgCMatrix' if \code{sparse = FALSE} the class will be 'dgMatrix'.
-#' @param inverse Logical defining if the inverse will be computed or not.
-#'
-#' @return A list with \eqn{\Omega = expm(U)} its inverse (if \code{inverse = TRUE}) and
+#'@param U A matrix.
+#'@param n A number specifing the dimension of the matrix U. Default \code{n = dim(U)[1]}.
+#'@param sparse Logical defining the class of the output matrix. If \code{sparse = TRUE} the output
+#'class will be 'dgCMatrix' if \code{sparse = FALSE} the class will be 'dgMatrix'.
+#'@param inverse Logical defining if the inverse will be computed or not.
+#'@return A list with \eqn{\Omega = expm(U)} its inverse (if \code{inverse = TRUE}) and
#' auxiliares matrices to compute the derivatives.
#'
#'@seealso \code{\link[Matrix]{expm}}, \code{\link[base]{eigen}},
#'\code{link[mcglm]{mc_dexp_gold}}.
#'@export
-mc_expm <- function(U, n = dim(U)[1], sparse = FALSE, inverse = FALSE){
- tt = eigen(U, symmetric = TRUE)
- UU = tt$vectors
- Q = tt$values
- eQr = Diagonal(n,exp(tt$values))
- inv_UU <- t(UU)
- Omega = Matrix(UU%*%eQr%*%inv_UU, sparse = sparse)
- if(inverse == TRUE){
- eQr_INV = Diagonal(n,exp(-tt$values))
- inv_Omega <- Matrix(UU%*%eQr_INV%*%inv_UU, sparse = sparse)
- saida <- list("Omega" = Omega, "inv_Omega" = inv_Omega, "UU" = UU, "Q" = Q, "inv_UU" = inv_UU)
- }
- if(inverse == FALSE){
- saida <- list("Omega" = Omega, "UU" = UU, "Q" = Q, "inv_UU" = inv_UU)
- }
- return(saida)
+mc_expm <- function(U, n = dim(U)[1], sparse = FALSE, inverse = FALSE) {
+ tt = eigen(U, symmetric = TRUE)
+ UU = tt$vectors
+ Q = tt$values
+ eQr = Diagonal(n, exp(tt$values))
+ inv_UU <- t(UU)
+ Omega = Matrix(UU %*% eQr %*% inv_UU, sparse = sparse)
+ if (inverse == TRUE) {
+ eQr_INV = Diagonal(n, exp(-tt$values))
+ inv_Omega <- Matrix(UU %*% eQr_INV %*% inv_UU, sparse = sparse)
+ saida <- list(Omega = Omega, inv_Omega = inv_Omega, UU = UU, Q = Q, inv_UU = inv_UU)
+ }
+ if (inverse == FALSE) {
+ saida <- list(Omega = Omega, UU = UU, Q = Q, inv_UU = inv_UU)
+ }
+ return(saida)
}
diff --git a/R/mc_fitted.mcglm.R b/R/mc_fitted.mcglm.R
index 1370036..2c212c9 100644
--- a/R/mc_fitted.mcglm.R
+++ b/R/mc_fitted.mcglm.R
@@ -5,9 +5,9 @@
#' @param object An object of mcglm class.
#' @return Depending on the number of response variable the function \code{fitted.mcglm} returns
#' a vector (univariate models) or a matrix (multivariate models) of fitted values.
-#' @exportMethod
+#' @export
fitted.mcglm <- function(object) {
- n_resp <- length(object$beta_names)
- output <- Matrix(object$fitted, ncol = n_resp, nrow = object$n_obs)
- return(output)
+ n_resp <- length(object$beta_names)
+ output <- Matrix(object$fitted, ncol = n_resp, nrow = object$n_obs)
+ return(output)
}
diff --git a/R/mc_getInformation.R b/R/mc_getInformation.R
index 9eb9145..49cf2d0 100644
--- a/R/mc_getInformation.R
+++ b/R/mc_getInformation.R
@@ -7,17 +7,22 @@
#' @param n_resp A number specyfing the nmber of response variables.
#' @return The number of \eqn{\beta}'s, \eqn{\tau}'s, power and correlation parameters.
#' @export
-mc_getInformation <- function(list_initial, list_power_fixed, n_resp){
- n_betas <- lapply(list_initial$regression, length)
- n_taus <- lapply(list_initial$tau, length)
- n_power <- lapply(list_initial$power, length)
- for(i in 1:n_resp) {
- if(list_power_fixed[[i]] == TRUE){n_power[i] = 0 }
- }
- if(n_resp == 1){n_rho <- 0}
- if(n_resp != 1){n_rho <- length(list_initial$rho)}
- n_cov <- sum(do.call(c,n_power))+n_rho+sum(do.call(c,n_taus))
- saida <- list("n_betas" = n_betas, "n_taus" = n_taus, "n_power" = n_power, "n_rho" = n_rho,
- "n_cov" = n_cov)
- return(saida)
-}
+mc_getInformation <- function(list_initial, list_power_fixed, n_resp) {
+ n_betas <- lapply(list_initial$regression, length)
+ n_taus <- lapply(list_initial$tau, length)
+ n_power <- lapply(list_initial$power, length)
+ for (i in 1:n_resp) {
+ if (list_power_fixed[[i]] == TRUE) {
+ n_power[i] = 0
+ }
+ }
+ if (n_resp == 1) {
+ n_rho <- 0
+ }
+ if (n_resp != 1) {
+ n_rho <- length(list_initial$rho)
+ }
+ n_cov <- sum(do.call(c, n_power)) + n_rho + sum(do.call(c, n_taus))
+ saida <- list(n_betas = n_betas, n_taus = n_taus, n_power = n_power, n_rho = n_rho, n_cov = n_cov)
+ return(saida)
+}
diff --git a/R/mc_influence.R b/R/mc_influence.R
index 102943b..42db2c3 100644
--- a/R/mc_influence.R
+++ b/R/mc_influence.R
@@ -4,84 +4,89 @@
#' Leverage, DFBETA and Cook's distance for unit sample and observations.
#'
#' @param object An object of mcglm class.
-#' @param id a vector which identifies the clusters. The length and order of ‘id’ should be the
+#' @param id a vector which identifies the clusters. The length and order of id should be the
#' same as the number of observations. Data are assumed to be sorted so that observations on a cluster
#' are contiguous rows for all entities in the formula.
#' @return A matrix. Note that the function assumes that the data are in the correct order.
#' @export
mc_influence <- function(object, id) {
- inv_M <- -object$inv_S_beta
- M <- solve(inv_M)
- temp_data <- data.frame("res" = object$residuals, id)
- temp_data_group <- split(temp_data, temp_data$id)
- D <- bdiag(lapply(object$mu_list, function(x)x$D))
- tD_invC <- t(D)%*%object$inv_C
- H <- D%*%inv_M%*%tD_invC
- leverage_obs <- diag(H)
- leverage_group <- tapply(leverage_obs, id, sum)
- I <- Diagonal(object$n_obs, 1)
- n_group <- length(temp_data_group)
- indexes <- matrix(NA, ncol = 2, nrow = n_group)
- n_obs_group <- table(id)
- indexes[1,] <- c(1,as.numeric(n_obs_group[1]))
- DFBETA_clust <- list()
- D_temp <- D[c(indexes[1,1]:indexes[1,2]),]
- inv_C_temp <- object$inv_C[c(indexes[1,1]:indexes[1,2]),c(indexes[1,1]:indexes[1,2])]
- C_temp <- object$C[c(indexes[1,1]:indexes[1,2]),c(indexes[1,1]:indexes[1,2])]
- H_temp <- H[c(indexes[1,1]:indexes[1,2]),c(indexes[1,1]:indexes[1,2])]
- res_temp <- object$residuals[indexes[1,1]:indexes[1,2]]
- DFBETAOij <- list()
- padroniza <- function(x, M){
- as.numeric((t(as.numeric(x))%*%M%*%as.numeric(x))/dim(M)[1])
- }
- dfbetaOij <- function(D_temp, C_temp, inv_C_temp, res_temp){
- DFBETA_temp <- matrix(NA, ncol = dim(D_temp)[2], nrow = dim(D_temp)[1])
- k = 1
- for(j in 1:length(res_temp)){
- Dij <- D_temp[j,k] - C_temp[k,-j]%*%inv_C_temp[-k,-j]%*%D_temp[-j,]
- rij <- res_temp[j] - C_temp[k,-j]%*%inv_C_temp[-k,-j]%*%res_temp[-j]
- Vij <- C_temp[k,j] - C_temp[k,-j]%*%inv_C_temp[-k,-j]%*%C_temp[-j,k]
- Hij <- try(Dij%*%object$inv_S_beta%*%t(Dij)%*%solve(Vij), silent = TRUE)
- if(class(Hij) == "try-error"){DFBETA_temp[j,] <- NA}
- if(class(Hij) != "try-error"){
- DFBETA_temp[j,] <- as.numeric(t(object$inv_S_beta%*%t(Dij)%*%(rij/(Vij - (1- Hij)))))
- }
- }
- return(DFBETA_temp)
- }
- DFBETA_clust[[1]] <- inv_M%*%t(D_temp)%*%inv_C_temp%*%res_temp
- DFBETAOij[[1]] <- dfbetaOij(D_temp, C_temp, inv_C_temp, res_temp)
- for(i in 2:n_group){
- indexes[i,] <- c(indexes[i-1,][2] + 1, n_obs_group[i] + indexes[i-1,][2])
- D_temp <- D[c(indexes[i,1]:indexes[i,2]),]
- inv_C_temp <- object$inv_C[c(indexes[i,1]:indexes[i,2]),c(indexes[i,1]:indexes[i,2])]
- C_temp <- object$C[c(indexes[i,1]:indexes[i,2]),c(indexes[i,1]:indexes[i,2])]
- H_temp <- H[c(indexes[i,1]:indexes[i,2]),c(indexes[i,1]:indexes[i,2])]
- res_temp <- object$residuals[indexes[i,1]:indexes[i,2]]
- D_temp <- matrix(D_temp, nrow = n_obs_group[i])
- DFBETA_clust[[i]] <- inv_M%*%t(D_temp)%*%inv_C_temp%*%res_temp
- if(n_obs_group[i] == 1){DFBETAOij[[i]] <- t(as.matrix(DFBETA_clust[[i]]))}
- if(n_obs_group[i] != 1){
- DFBETAOij[[i]] <- dfbetaOij(D_temp, C_temp, inv_C_temp, res_temp)
+ inv_M <- -object$inv_S_beta
+ M <- solve(inv_M)
+ temp_data <- data.frame(res = object$residuals, id)
+ temp_data_group <- split(temp_data, temp_data$id)
+ D <- bdiag(lapply(object$mu_list, function(x) x$D))
+ tD_invC <- t(D) %*% object$inv_C
+ H <- D %*% inv_M %*% tD_invC
+ leverage_obs <- diag(H)
+ leverage_group <- tapply(leverage_obs, id, sum)
+ I <- Diagonal(object$n_obs, 1)
+ n_group <- length(temp_data_group)
+ indexes <- matrix(NA, ncol = 2, nrow = n_group)
+ n_obs_group <- table(id)
+ indexes[1, ] <- c(1, as.numeric(n_obs_group[1]))
+ DFBETA_clust <- list()
+ D_temp <- D[c(indexes[1, 1]:indexes[1, 2]), ]
+ inv_C_temp <- object$inv_C[c(indexes[1, 1]:indexes[1, 2]), c(indexes[1, 1]:indexes[1, 2])]
+ C_temp <- object$C[c(indexes[1, 1]:indexes[1, 2]), c(indexes[1, 1]:indexes[1, 2])]
+ H_temp <- H[c(indexes[1, 1]:indexes[1, 2]), c(indexes[1, 1]:indexes[1, 2])]
+ res_temp <- object$residuals[indexes[1, 1]:indexes[1, 2]]
+ DFBETAOij <- list()
+ padroniza <- function(x, M) {
+ as.numeric((t(as.numeric(x)) %*% M %*% as.numeric(x))/dim(M)[1])
}
- }
- DFBETA <- lapply(DFBETA_clust, as.matrix)
- DFBETA <- lapply(DFBETA, t)
- DFBETA <- ldply(DFBETA, data.frame)
- names(DFBETA) <- object$beta_names[[1]]
- DFBETAOij <- ldply(DFBETAOij, data.frame)
- names(DFBETAOij) <- object$beta_names[[1]]
- DCLSi <- apply(as.matrix(DFBETA), MARGIN = 1, FUN = padroniza, M = M)
- DCLOij <- apply(as.matrix(DFBETAOij), MARGIN = 1, FUN = padroniza, M = M)
- std.error <- coef(object, std.error = TRUE, type = "beta")$Std.error
- DFBETA_temp <- t(apply(DFBETA, MARGIN = 1, FUN = function(x, std.error){as.numeric(x/std.error)},
- std.error = std.error))
- DFBETAOij_temp <- t(apply(DFBETAOij, MARGIN = 1, FUN = function(x, std.error){as.numeric(x/std.error)},
- std.error = std.error))
- output_clust <- data.frame("Leverage" = leverage_group, "DFBETA" = DFBETA_temp, "Cook" = DCLSi)
- output_obs <- data.frame("Leverage" = leverage_obs, "DFBETA" = DFBETAOij_temp, "Cook" = DCLOij)
- output <- list("Id" = output_clust, "Observations" = output_obs)
- return(output)
+ dfbetaOij <- function(D_temp, C_temp, inv_C_temp, res_temp) {
+ DFBETA_temp <- matrix(NA, ncol = dim(D_temp)[2], nrow = dim(D_temp)[1])
+ k = 1
+ for (j in 1:length(res_temp)) {
+ Dij <- D_temp[j, k] - C_temp[k, -j] %*% inv_C_temp[-k, -j] %*% D_temp[-j, ]
+ rij <- res_temp[j] - C_temp[k, -j] %*% inv_C_temp[-k, -j] %*% res_temp[-j]
+ Vij <- C_temp[k, j] - C_temp[k, -j] %*% inv_C_temp[-k, -j] %*% C_temp[-j, k]
+ Hij <- try(Dij %*% object$inv_S_beta %*% t(Dij) %*% solve(Vij), silent = TRUE)
+ if (class(Hij) == "try-error") {
+ DFBETA_temp[j, ] <- NA
+ }
+ if (class(Hij) != "try-error") {
+ DFBETA_temp[j, ] <- as.numeric(t(object$inv_S_beta %*% t(Dij) %*% (rij/(Vij - (1 - Hij)))))
+ }
+ }
+ return(DFBETA_temp)
+ }
+ DFBETA_clust[[1]] <- inv_M %*% t(D_temp) %*% inv_C_temp %*% res_temp
+ DFBETAOij[[1]] <- dfbetaOij(D_temp, C_temp, inv_C_temp, res_temp)
+ for (i in 2:n_group) {
+ indexes[i, ] <- c(indexes[i - 1, ][2] + 1, n_obs_group[i] + indexes[i - 1, ][2])
+ D_temp <- D[c(indexes[i, 1]:indexes[i, 2]), ]
+ inv_C_temp <- object$inv_C[c(indexes[i, 1]:indexes[i, 2]), c(indexes[i, 1]:indexes[i, 2])]
+ C_temp <- object$C[c(indexes[i, 1]:indexes[i, 2]), c(indexes[i, 1]:indexes[i, 2])]
+ H_temp <- H[c(indexes[i, 1]:indexes[i, 2]), c(indexes[i, 1]:indexes[i, 2])]
+ res_temp <- object$residuals[indexes[i, 1]:indexes[i, 2]]
+ D_temp <- matrix(D_temp, nrow = n_obs_group[i])
+ DFBETA_clust[[i]] <- inv_M %*% t(D_temp) %*% inv_C_temp %*% res_temp
+ if (n_obs_group[i] == 1) {
+ DFBETAOij[[i]] <- t(as.matrix(DFBETA_clust[[i]]))
+ }
+ if (n_obs_group[i] != 1) {
+ DFBETAOij[[i]] <- dfbetaOij(D_temp, C_temp, inv_C_temp, res_temp)
+ }
+ }
+ DFBETA <- lapply(DFBETA_clust, as.matrix)
+ DFBETA <- lapply(DFBETA, t)
+ DFBETA <- ldply(DFBETA, data.frame)
+ names(DFBETA) <- object$beta_names[[1]]
+ DFBETAOij <- ldply(DFBETAOij, data.frame)
+ names(DFBETAOij) <- object$beta_names[[1]]
+ DCLSi <- apply(as.matrix(DFBETA), MARGIN = 1, FUN = padroniza, M = M)
+ DCLOij <- apply(as.matrix(DFBETAOij), MARGIN = 1, FUN = padroniza, M = M)
+ std.error <- coef(object, std.error = TRUE, type = "beta")$Std.error
+ DFBETA_temp <- t(apply(DFBETA, MARGIN = 1, FUN = function(x, std.error) {
+ as.numeric(x/std.error)
+ }, std.error = std.error))
+ DFBETAOij_temp <- t(apply(DFBETAOij, MARGIN = 1, FUN = function(x, std.error) {
+ as.numeric(x/std.error)
+ }, std.error = std.error))
+ output_clust <- data.frame(Leverage = leverage_group, DFBETA = DFBETA_temp, Cook = DCLSi)
+ output_obs <- data.frame(Leverage = leverage_obs, DFBETA = DFBETAOij_temp, Cook = DCLOij)
+ output <- list(Id = output_clust, Observations = output_obs)
+ return(output)
}
-
diff --git a/R/mc_initial_values.R b/R/mc_initial_values.R
new file mode 100644
index 0000000..1e59bc2
--- /dev/null
+++ b/R/mc_initial_values.R
@@ -0,0 +1,102 @@
+#' Automatic initial values for McGLMs.
+#'
+#' Return a list of initial values for McGLMs.
+#' @param linear_pred A list of formula see \code{\link[stats]{formula}} for details.
+#' @param matrix_pred A list of known matrices to be used on the matrix linear predictor. Details
+#' can be obtained on \code{\link[mcglm]{mc_matrix_linear_predictor}}.
+#' @param link A list of link functions names, see \code{\link[mcglm]{mc_link_function}} for details.
+#' @param variance A list of variance functions names, see \code{\link[mcglm]{mc_variance_function}}
+#' for details.
+#' @param covariance A list of covariance link functions names, current options are: identity, inverse
+#' and exponential-matrix (expm).
+#' @param offset A list with values of offset values if any.
+#' @param Ntrial A list with values of the number of trials on Bernoulli experiments. It is useful only
+#' for binomialP and binomialPQ variance functions.
+#' @param contrasts List of contrasts to be used in the \code{\link[stats]{model.matrix}}.
+#' @param data A data frame.
+#' @return Return a list of initial values to be used while fitting McGLMs.
+#' @export
+
+mc_initial_values <- function(linear_pred, matrix_pred, link, variance, covariance, offset, Ntrial, contrasts = NULL, data) {
+ n_resp <- length(linear_pred)
+ if (!is.null(contrasts)) {
+ list_X <- list()
+ for (i in 1:n_resp) {
+ list_X[[i]] <- model.matrix(linear_pred[[i]], contrasts = contrasts[[i]], data = data)
+ }
+ } else {
+ list_X <- lapply(linear_pred, model.matrix, data = data)
+ }
+ list_models <- list()
+ power_initial <- list()
+ for (i in 1:n_resp) {
+ if (variance[[i]] == "constant") {
+ power_initial[[i]] <- 0
+ if (!is.null(offset[[i]])) {
+ data_temp <- data
+ data_temp$offset <- offset[[i]]
+ list_models[[i]] <- glm(linear_pred[[i]], family = quasi(link = link[[i]], variance = "constant"),
+ offset = offset, data = data_temp)
+ } else {
+ list_models[[i]] <- glm(linear_pred[[i]], family = quasi(link = link[[i]], variance = "constant"), data = data)
+ }
+ }
+ if (variance[[i]] == "tweedie" | variance[[i]] == "poisson_tweedie") {
+ power_initial[[i]] <- 1
+ if (!is.null(offset[[i]])) {
+ data_temp <- data
+ data_temp$offset <- offset[[i]]
+ list_models[[i]] <- glm(linear_pred[[i]], family = quasi(link = link[[i]], variance = "mu"),
+ offset = offset, data = data_temp)
+ } else {
+ list_models[[i]] <- glm(linear_pred[[i]], family = quasi(link = link[[i]], variance = "mu"), data = data)
+ }
+ }
+ if (variance[[i]] == "binomialP" | variance[[i]] == "binomialPQ") {
+ power_initial[[i]] <- c(1)
+ if (variance[[i]] == "binomialPQ") {
+ power_initial[[i]] <- c(1, 1)
+ }
+ if (!is.null(Ntrial[[i]])) {
+ temp <- model.frame(linear_pred[[i]], data = data)
+ Y <- model.response(temp) * Ntrial[[i]]
+ resp <- cbind(Y, Ntrial[[i]] - Y)
+ X <- model.matrix(linear_pred[[i]], data = data)
+ link_temp <- link[[i]]
+ if (link_temp == "loglog") {
+ link_temp <- "cloglog"
+ }
+ list_models[[i]] <- glm(resp ~ X - 1, family = binomial(link = link_temp), data = data)
+ } else {
+ link_temp <- link[[i]]
+ if (link_temp == "loglog") {
+ link_temp <- "cloglog"
+ }
+ list_models[[i]] <- glm(linear_pred[[i]], family = quasi(link = link_temp, variance = "mu(1-mu)"), data = data)
+ }
+ }
+ }
+ list_initial <- list()
+ list_initial$regression <- lapply(list_models, coef)
+ list_initial$power <- power_initial
+ tau0_initial <- lapply(list_models, function(x) summary(x)$dispersion)
+ tau_extra <- lapply(matrix_pred, length)
+ list_initial$tau <- list()
+ for (i in 1:n_resp) {
+ if (covariance == "identity") {
+ list_initial$tau[[i]] <- as.numeric(c(tau0_initial[[i]], rep(0, c(tau_extra[[i]] - 1))))
+ }
+ if (covariance == "inverse") {
+ list_initial$tau[[i]] <- as.numeric(c(1/tau0_initial[[i]], rep(0, c(tau_extra[[i]] - 1))))
+ }
+ if (covariance == "expm") {
+ list_initial$tau[[i]] <- as.numeric(c(exp(tau0_initial[[i]]), rep(0.1, c(tau_extra[[i]] - 1))))
+ }
+ }
+ if (n_resp == 1) {
+ list_initial$rho = 0
+ } else {
+ list_initial$rho <- rep(0, n_resp * (n_resp - 1)/2)
+ }
+ return(list_initial)
+}
diff --git a/R/mc_link_function.R b/R/mc_link_function.R
index d83ae7a..660b254 100644
--- a/R/mc_link_function.R
+++ b/R/mc_link_function.R
@@ -28,120 +28,150 @@
#' @examples
#' x1 <- seq(-1, 1, l = 5)
#' X <- model.matrix(~ x1)
-#' mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = "log")
-#' mc_link_function(beta = c(1,0.5), X = X, offset = rep(10,5), link = "identity")
+#' mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = 'log')
+#' mc_link_function(beta = c(1,0.5), X = X, offset = rep(10,5), link = 'identity')
#' @export
# Generic link function ---------------------------
mc_link_function <- function(beta, X, offset, link) {
- assert_that(noNA(beta))
- assert_that(noNA(X))
- if(!is.null(offset)) assert_that(noNA(offset))
- switch(link,
- "logit" = { output <- mc_logit(beta = beta, X = X, offset = offset) },
- "probit" = { output <- mc_probit(beta = beta, X = X, offset = offset) },
- "cauchit" = { output <- mc_cauchit(beta = beta, X = X, offset = offset) },
- "cloglog" = { output <- mc_cloglog(beta = beta, X = X, offset = offset) },
- "loglog" = { output <- mc_loglog(beta = beta, X = X, offset = offset) },
- "identity" = { output <- mc_identity(beta = beta, X = X, offset = offset) },
- "log" = { output <- mc_log(beta = beta, X = X, offset = offset) },
- "sqrt" = { output <- mc_sqrt(beta = beta, X = X, offset = offset) },
- "1/mu^2" = { output <- mc_invmu2(beta = beta, X = X, offset = offset) },
- "inverse" = { output <- mc_inverse(beta = beta, X = X, offset = offset) },
- stop(gettextf("%s link not recognised", sQuote(link)),domain = NA))
- return(output)
+ assert_that(noNA(beta))
+ assert_that(noNA(X))
+ if (!is.null(offset))
+ assert_that(noNA(offset))
+ switch(link, logit = {
+ output <- mc_logit(beta = beta, X = X, offset = offset)
+ }, probit = {
+ output <- mc_probit(beta = beta, X = X, offset = offset)
+ }, cauchit = {
+ output <- mc_cauchit(beta = beta, X = X, offset = offset)
+ }, cloglog = {
+ output <- mc_cloglog(beta = beta, X = X, offset = offset)
+ }, loglog = {
+ output <- mc_loglog(beta = beta, X = X, offset = offset)
+ }, identity = {
+ output <- mc_identity(beta = beta, X = X, offset = offset)
+ }, log = {
+ output <- mc_log(beta = beta, X = X, offset = offset)
+ }, sqrt = {
+ output <- mc_sqrt(beta = beta, X = X, offset = offset)
+ }, `1/mu^2` = {
+ output <- mc_invmu2(beta = beta, X = X, offset = offset)
+ }, inverse = {
+ output <- mc_inverse(beta = beta, X = X, offset = offset)
+ }, stop(gettextf("%s link not recognised", sQuote(link)), domain = NA))
+ return(output)
}
#' @rdname mc_link_function
# Logit link function ---------------------------
mc_logit <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu <- make.link("logit")$linkinv(eta = eta)
- return(list("mu" = mu, "D" = X*(mu*(1-mu))))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu <- make.link("logit")$linkinv(eta = eta)
+ return(list(mu = mu, D = X * (mu * (1 - mu))))
}
#' @rdname mc_link_function
# Probit link function ---------------------------
mc_probit <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu <- make.link("probit")$linkinv(eta = eta)
- Deri <- make.link("probit")$mu.eta(eta = eta)
- return(list("mu" = mu, "D" = X*Deri))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu <- make.link("probit")$linkinv(eta = eta)
+ Deri <- make.link("probit")$mu.eta(eta = eta)
+ return(list(mu = mu, D = X * Deri))
}
#' @rdname mc_link_function
# Cauchit link function ---------------------------
mc_cauchit <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu = make.link("cauchit")$linkinv(eta = eta)
- Deri <- make.link("cauchit")$mu.eta(eta = eta)
- return(list("mu" = mu, "D" = X*Deri))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu = make.link("cauchit")$linkinv(eta = eta)
+ Deri <- make.link("cauchit")$mu.eta(eta = eta)
+ return(list(mu = mu, D = X * Deri))
}
#' @rdname mc_link_function
# Complement log-log link function ---------------------------
mc_cloglog <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu = make.link("cloglog")$linkinv(eta = eta)
- Deri <- make.link("cloglog")$mu.eta(eta = eta)
- return(list("mu" = mu, "D" = X*Deri))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu = make.link("cloglog")$linkinv(eta = eta)
+ Deri <- make.link("cloglog")$mu.eta(eta = eta)
+ return(list(mu = mu, D = X * Deri))
}
#' @rdname mc_link_function
## Log-log link function ---------------------------
mc_loglog <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu <- exp(-exp(-eta))
- Deri <- exp(-exp(-eta)-eta)
- return(list("mu" = mu, "D" = X*Deri))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu <- exp(-exp(-eta))
+ Deri <- exp(-exp(-eta) - eta)
+ return(list(mu = mu, D = X * Deri))
}
#' @rdname mc_link_function
## Identity link function ---------------------------
mc_identity <- function(beta, X, offset) {
- eta <- X%*%beta
- if (!is.null(offset)) {eta <- eta + offset}
- return(list("mu" = as.numeric(eta), "D" = X))
+ eta <- X %*% beta
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ return(list(mu = as.numeric(eta), D = X))
}
#' @rdname mc_link_function
## Log link function ---------------------------
mc_log <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu = make.link("log")$linkinv(eta = eta)
- return(list("mu" = mu, "D" = X*mu))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu = make.link("log")$linkinv(eta = eta)
+ return(list(mu = mu, D = X * mu))
}
#' @rdname mc_link_function
## Square-root link function ---------------------------
mc_sqrt <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu = make.link("sqrt")$linkinv(eta = eta)
- return(list("mu" = mu, "D" = X*(2*as.numeric(eta))))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu = make.link("sqrt")$linkinv(eta = eta)
+ return(list(mu = mu, D = X * (2 * as.numeric(eta))))
}
#' @rdname mc_link_function
## Inverse mu square link function ---------------------------
mc_invmu2 <- function(beta, X, offset) {
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu = make.link("1/mu^2")$linkinv(eta = eta)
- Deri = make.link("1/mu^2")$mu.eta(eta = eta)
- return(list("mu" = mu, "D" = X*Deri))
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu = make.link("1/mu^2")$linkinv(eta = eta)
+ Deri = make.link("1/mu^2")$mu.eta(eta = eta)
+ return(list(mu = mu, D = X * Deri))
}
#' @rdname mc_link_function
## Inverse link function ---------------------------
-mc_inverse <- function(beta, X, offset){
- eta <- as.numeric(X%*%beta)
- if (!is.null(offset)) {eta <- eta + offset}
- mu = make.link("inverse")$linkinv(eta = eta)
- Deri = make.link("inverse")$mu.eta(eta = eta)
- return(list("mu" = mu, "D" = X*Deri))
-}
+mc_inverse <- function(beta, X, offset) {
+ eta <- as.numeric(X %*% beta)
+ if (!is.null(offset)) {
+ eta <- eta + offset
+ }
+ mu = make.link("inverse")$linkinv(eta = eta)
+ Deri = make.link("inverse")$mu.eta(eta = eta)
+ return(list(mu = mu, D = X * Deri))
+}
diff --git a/R/mc_list2vec.R b/R/mc_list2vec.R
index b9a2692..8584e39 100644
--- a/R/mc_list2vec.R
+++ b/R/mc_list2vec.R
@@ -9,25 +9,25 @@
#' It will be useful, only if the user wants to implement a different variance-covariance matrix.
#' @export
mc_list2vec <- function(list_initial, list_power_fixed) {
- cov_ini <- do.call(c,Map(c, list_initial$power, list_initial$tau))
- n_resp <- length(list_initial$regression)
- indicadora <- list()
- for(i in 1:n_resp){
- indicadora[[i]] <- rep(FALSE, length(list_initial$tau[[i]]))
- }
- indicadora_power <- list()
- for(i in 1:n_resp){
- if(list_power_fixed[[i]] == FALSE){
- indicadora_power[[i]] <- rep(FALSE, length(list_initial$power[[i]]))
+ cov_ini <- do.call(c, Map(c, list_initial$power, list_initial$tau))
+ n_resp <- length(list_initial$regression)
+ indicadora <- list()
+ for (i in 1:n_resp) {
+ indicadora[[i]] <- rep(FALSE, length(list_initial$tau[[i]]))
}
- if(list_power_fixed[[i]] == TRUE){
- indicadora_power[[i]] <- rep(TRUE, length(list_initial$power[[i]]))
+ indicadora_power <- list()
+ for (i in 1:n_resp) {
+ if (list_power_fixed[[i]] == FALSE) {
+ indicadora_power[[i]] <- rep(FALSE, length(list_initial$power[[i]]))
+ }
+ if (list_power_fixed[[i]] == TRUE) {
+ indicadora_power[[i]] <- rep(TRUE, length(list_initial$power[[i]]))
+ }
}
- }
- index <- do.call(c,Map(c, indicadora_power, indicadora))
- cov_par <- data.frame(cov_ini,index)
- cov_ini <- cov_par[which(cov_par$index == FALSE),]$cov_ini
- beta_ini <- do.call(c,list_initial$regression)
- cov_ini <- c("rho" = list_initial$rho, cov_ini)
- return(list("beta_ini" = as.numeric(beta_ini), "cov_ini" = as.numeric(cov_ini)))
-}
+ index <- do.call(c, Map(c, indicadora_power, indicadora))
+ cov_par <- data.frame(cov_ini, index)
+ cov_ini <- cov_par[which(cov_par$index == FALSE), ]$cov_ini
+ beta_ini <- do.call(c, list_initial$regression)
+ cov_ini <- c(rho = list_initial$rho, cov_ini)
+ return(list(beta_ini = as.numeric(beta_ini), cov_ini = as.numeric(cov_ini)))
+}
diff --git a/R/mc_matrix_linear_predictor.R b/R/mc_matrix_linear_predictor.R
index ad68375..b50d157 100644
--- a/R/mc_matrix_linear_predictor.R
+++ b/R/mc_matrix_linear_predictor.R
@@ -16,8 +16,10 @@
# Matrix linear predictor -------------------------------
mc_matrix_linear_predictor <- function(tau, Z) {
- if (length(Z) != length(tau)) {stop("Incorrect number of parameters")}
- output <- mapply('*', Z, tau, SIMPLIFY = FALSE) ## Multiply each matrix by the parameter tau
- output <- Reduce("+",output) ## Sum all matrices inside the list
- return(output)
-}
+ if (length(Z) != length(tau)) {
+ stop("Incorrect number of parameters")
+ }
+ output <- mapply("*", Z, tau, SIMPLIFY = FALSE) ## Multiply each matrix by the parameter tau
+ output <- Reduce("+", output) ## Sum all matrices inside the list
+ return(output)
+}
diff --git a/R/mc_pearson.R b/R/mc_pearson.R
index 13fa211..1c07483 100644
--- a/R/mc_pearson.R
+++ b/R/mc_pearson.R
@@ -14,24 +14,20 @@
#' @details Compute the Pearson estimating function its sensitivity and variability matrices.
#' For more details see Bonat and Jorgensen (2015) equations 6, 7 and 8.
#' @export
-mc_pearson <- function(y_vec, mu_vec, Cfeatures, inv_J_beta = NULL, D = NULL,
- correct = FALSE, compute_variability = FALSE) {
- product <- lapply(Cfeatures$D_C, mc_multiply, bord2 = Cfeatures$inv_C)
- res <- y_vec - mu_vec
- pearson_score <- unlist(lapply(product, mc_core_pearson, inv_C = Cfeatures$inv_C,
- res = res))
- sensitivity <- mc_sensitivity(product)
- output <- list("Score" = pearson_score, "Sensitivity" = sensitivity, "Extra" = product)
- if(correct == TRUE) {
- correction <- mc_correction(D_C = Cfeatures$D_C, inv_J_beta = inv_J_beta,
- D = D, inv_C = Cfeatures$inv_C)
- output <- list("Score" = pearson_score + correction, "Sensitivity" = sensitivity,
- "Extra" = product)
- }
- if(compute_variability == TRUE) {
- variability <- mc_variability(sensitivity = sensitivity, product = product,
- inv_C = Cfeatures$inv_C, C = Cfeatures$C, res = res)
- output$Variability <- variability
- }
- return(output)
-}
+mc_pearson <- function(y_vec, mu_vec, Cfeatures, inv_J_beta = NULL, D = NULL, correct = FALSE, compute_variability = FALSE) {
+ product <- lapply(Cfeatures$D_C, mc_multiply, bord2 = Cfeatures$inv_C)
+ res <- y_vec - mu_vec
+ pearson_score <- unlist(lapply(product, mc_core_pearson, inv_C = Cfeatures$inv_C, res = res))
+ sensitivity <- mc_sensitivity(product)
+ output <- list(Score = pearson_score, Sensitivity = sensitivity, Extra = product)
+ if (correct == TRUE) {
+ correction <- mc_correction(D_C = Cfeatures$D_C, inv_J_beta = inv_J_beta, D = D, inv_C = Cfeatures$inv_C)
+ output <- list(Score = pearson_score + correction, Sensitivity = sensitivity, Extra = product)
+ }
+ if (compute_variability == TRUE) {
+ variability <- mc_variability(sensitivity = sensitivity, product = product, inv_C = Cfeatures$inv_C, C = Cfeatures$C,
+ res = res)
+ output$Variability <- variability
+ }
+ return(output)
+}
diff --git a/R/mc_plot.mcglm.R b/R/mc_plot.mcglm.R
index f790179..7fb6442 100644
--- a/R/mc_plot.mcglm.R
+++ b/R/mc_plot.mcglm.R
@@ -6,39 +6,44 @@
#' @param object a fitted mcglm object as produced by \code{mcglm()}.
#' @param type Specify which graphical analysis will be performed, options are: residuals, influence
#' and algorithm.
-#' @exportMethod
+#' @export
plot.mcglm <- function(object, type = "residuals") {
- n_resp <- length(object$beta_names)
- if(type == "residuals") {
- par(mar=c(2.6, 2.5, 0.1, 0.1), mgp = c(1.6, 0.6, 0), mfrow = c(1,n_resp))
- for(i in 1:n_resp){
- plot(residuals(object, type = "pearson")[,i] ~ fitted(object)[,i],
- ylab = "Pearson residuals", xlab = "Fitted values")
- temp <- loess.smooth(fitted(object)[,i],residuals(object, type = "pearson")[,i])
- lines(temp$x,temp$y)
+ n_resp <- length(object$beta_names)
+ if (type == "residuals") {
+ par(mar = c(2.6, 2.5, 0.1, 0.1), mgp = c(1.6, 0.6, 0), mfrow = c(2, n_resp))
+ for (i in 1:n_resp) {
+ res <- residuals(object, type = "pearson")[, i]
+ fit_values <- fitted(object)[, i]
+ plot(res ~ fit_values, ylab = "Pearson residuals", xlab = "Fitted values")
+ temp <- loess.smooth(fitted(object)[, i], residuals(object, type = "pearson")[, i])
+ lines(temp$x, temp$y)
+ qqnorm(res)
+ qqline(res)
+ }
+ }
+ if (type == "algorithm") {
+ n_iter <- length(na.exclude(object$IterationCovariance[, 1]))
+ par(mar = c(2.6, 2.5, 0.1, 0.1), mgp = c(1.6, 0.6, 0), mfrow = c(2, 2))
+ matplot(object$IterationRegression[1:c(n_iter + 5), ], type = "l", lty = 2, ylab = "Regression", xlab = "Iterations")
+ matplot(object$IterationCovariance[1:c(n_iter + 5), ], type = "l", lty = 2, ylab = "Covariance", xlab = "Iterations")
+ matplot(object$ScoreRegression[1:c(n_iter + 5), ], type = "l", lty = 2, ylab = "Quasi-score Regression", xlab = "Iterations")
+ matplot(object$ScoreCovariance[1:c(n_iter + 5), ], type = "l", lty = 2, ylab = "Quasi-score Covariance", xlab = "Iterations")
}
- }
- if(type == "partial_residuals"){
- par(mar=c(2.6, 2.5, 0.1, 0.1), mgp = c(1.6, 0.6, 0), mfrow = c(n_resp,c(n_cov-1)))
- for(i in 1:n_resp){
- n_cov <- dim(object$list_X[[i]])[2]
- for(j in 2:n_cov){
- plot(residuals(object, type = "pearson")[,i] ~ as.numeric(object$list_X[[i]][,j]))
- temp <- loess.smooth(as.numeric(object$list_X[[i]][,j]),residuals(object, type = "pearson")[,i])
- lines(temp$x,temp$y)
+ if (type == "partial_residuals") {
+ list_beta <- mc_updateBeta(list_initial = object$list_initial, betas = object$Regression, n_resp = n_resp, information = object$Information)
+ comp_X <- list()
+ for (i in 1:n_resp) {
+ comp_X[[i]] <- as.matrix(object$list_X[[i]]) * as.numeric(list_beta$regression[[i]])
+ }
+ for (i in 1:n_resp) {
+ res <- residuals(object, type = "pearson")[, i]
+ dev.new()
+ n_cov <- dim(comp_X[[i]])[2]
+ par(mar = c(2.6, 2.5, 0.5, 0.5), mgp = c(1.6, 0.6, 0), mfrow = c(1, c(n_cov - 1)))
+ for (j in 2:n_cov) {
+ p1 <- comp_X[[i]][, j] + res
+ plot(p1 ~ object$list_X[[i]][, j], xlab = object$beta_names[[i]][j], ylab = "Partial residuals ")
+ }
}
}
-}
- if(type == "algorithm") {
- n_iter <- length(na.exclude(object$IterationCovariance[,1]))
- par(mar=c(2.6, 2.5, 0.1, 0.1), mgp = c(1.6, 0.6, 0), mfrow = c(2,2))
- matplot(object$IterationRegression[1:c(n_iter+5),], type = "l", lty = 2, ylab = "Regression",
- xlab = "Iterations")
- matplot(object$IterationCovariance[1:c(n_iter+5),], type = "l", lty = 2, ylab = "Covariance",
- xlab = "Iterations")
- matplot(object$ScoreRegression[1:c(n_iter+5),], type = "l", lty = 2, ylab = "Quasi-score Regression",
- xlab = "Iterations")
- matplot(object$ScoreCovariance[1:c(n_iter+5),], type = "l", lty = 2, ylab = "Quasi-score Covariance",
- xlab = "Iterations")
- }
}
diff --git a/R/mc_print.mcglm.R b/R/mc_print.mcglm.R
index 1610986..7cc8ae9 100644
--- a/R/mc_print.mcglm.R
+++ b/R/mc_print.mcglm.R
@@ -3,31 +3,36 @@
#' @description The default print method for a mcglm object.
#'
#' @param object fitted model objects of class mcglm as produced by mcglm().
-#' @export print mcglm
-#' @aliases print print.mcglm
+#' @export
+
print.mcglm <- function(object) {
- n_resp <- length(object$beta_names)
- regression <- mc_updateBeta(list_initial = list(), betas = object$Regression,
- information = object$Information, n_resp = n_resp)
- for(i in 1:n_resp) {
- cat("Call: ")
- print(object$linear_pred[[i]])
- cat("\n")
- cat("Link function:", object$link[[i]])
- cat("\n")
- cat("Variance function:", object$variance[[i]])
- cat("\n")
- cat("Covariance function:", object$covariance[[i]])
- cat("\n")
- names(regression[[1]][[i]]) <- object$beta_names[[i]]
- cat("Regression:\n")
- print(regression[[1]][[i]])
- cat("\n")
- cat("tau:\n")
- print(coef(object, response = i, type = "tau")$Estimate)
- cat("\n")
- cat("power:\n")
- print(coef(object, response = i, type = "power")$Estimate)
- cat("\n")
- }
+ n_resp <- length(object$beta_names)
+ regression <- mc_updateBeta(list_initial = list(), betas = object$Regression, information = object$Information, n_resp = n_resp)
+ for (i in 1:n_resp) {
+ cat("Call: ")
+ print(object$linear_pred[[i]])
+ cat("\n")
+ cat("Link function:", object$link[[i]])
+ cat("\n")
+ cat("Variance function:", object$variance[[i]])
+ cat("\n")
+ cat("Covariance function:", object$covariance[[i]])
+ cat("\n")
+ names(regression[[1]][[i]]) <- object$beta_names[[i]]
+ cat("Regression:\n")
+ print(regression[[1]][[i]])
+ cat("\n")
+ cat("Dispersion:\n")
+ tau_temp <- coef(object, response = i, type = "tau")$Estimate
+ names(tau_temp) <- rep("", length(tau_temp))
+ print(tau_temp)
+ cat("\n")
+ power_temp <- coef(object, response = i, type = "power")$Estimate
+ if (length(power_temp) != 0) {
+ names(power_temp) <- ""
+ cat("Power:\n")
+ print(power_temp)
+ cat("\n")
+ }
+ }
}
diff --git a/R/mc_qic.R b/R/mc_qic.R
new file mode 100644
index 0000000..b20d58b
--- /dev/null
+++ b/R/mc_qic.R
@@ -0,0 +1,29 @@
+#' Compute Quasi Information Criterion (QIC) for McGLMs.
+#'
+#' qic.mcglm is a function which computes the QIC for McGLMs.
+#' @param object An object of mcglm class.
+#' @param object.iid An object of mcglm class contained the model fitted using independent
+#' covariance structure.
+#' @return The QIC value.
+#' @export
+
+qic.mcglm <- function(object, object.iid) {
+ mu <- fitted(object)
+ obs <- object$observed
+ n_resp <- dim(mu)[2]
+ Q <- matrix(NA, ncol = dim(mu)[2], nrow = dim(mu)[1])
+ for (i in 1:n_resp) {
+ if (object$power_fixed[[i]] == FALSE) {
+ power <- coef(object, type = "power", response = i)$Estimate
+ }
+ if (object$power_fixed[[i]] == TRUE) {
+ power = object$list_initial$power[[i]]
+ }
+ Q[, i] <- mc_qll(y = obs[, i], mu = mu[, i], variance = object$variance[[i]], power = power)
+ }
+ Vbeta <- -object$inv_S_beta
+ Vnull <- solve(-object.iid$inv_S_beta)
+ t1 <- -2 * sum(Q)
+ qic <- t1 + 2 * sum(diag(Vnull %*% Vbeta))
+ return(list(Q = t1, qic = qic))
+}
diff --git a/R/mc_qll.R b/R/mc_qll.R
new file mode 100644
index 0000000..e9e4c14
--- /dev/null
+++ b/R/mc_qll.R
@@ -0,0 +1,37 @@
+#' Compute quasi-likelihood function.
+#'
+#' Given a variance function mc_qll function computes the quasi-likelihood values.
+#' @param y A vector of observed values.
+#' @param mu A vector of fitted values.
+#' @param variance Variance function (constant, tweedie, poisson_tweedie, binomial).
+#' @param power Power parameter value.
+#' @return The quasi-likelihood values.
+#' @export
+
+mc_qll <- function(y, mu, variance, power) {
+ if (variance == "constant") {
+ qll <- -((y - mu)^2)/2 # Gaussian case
+ }
+ if (variance == "tweedie" & power == 1) {
+ qll <- y * log(mu) - mu # Poisson case
+ }
+ if (variance == "tweedie" & power == 2) {
+ -y/mu - log(mu) # Gamma case
+ }
+ if (variance == "tweedie" & power != 1 & power != 2) {
+ qll <- (mu^-power) * ((mu * y)/(1 - power) - (mu^2)/(2 - power)) # General Tweedie case
+ }
+ if (variance == "poisson_tweedie" & power == 1) {
+ qll <- (y * log(mu) - mu) + (y * log(mu) - mu)
+ }
+ if (variance == "poisson_tweedie" & power == 2) {
+ qll <- (y * log(mu) - mu) + (-y/mu - log(mu))
+ }
+ if (variance == "poisson_tweedie" & power != 1 & power != 2) {
+ qll <- (y * log(mu) - mu) + (mu^-power) * ((mu * y)/(1 - power) - (mu^2)/(2 - power))
+ }
+ if (variance == "binomial") {
+ qll <- y * log(mu/(1 - mu)) + log(1 - mu) # Binomial case
+ }
+ return(qll)
+}
diff --git a/R/mc_quasi_score.R b/R/mc_quasi_score.R
index 99cf710..1f398bc 100644
--- a/R/mc_quasi_score.R
+++ b/R/mc_quasi_score.R
@@ -2,17 +2,17 @@
#'
#' @description Compute the quasi-score function, its sensitivy and variability matrix.
#'
-#' @param D A matrix. In general the output from \code{\link[mcglm]{mc_lnk_function}}.
+#' @param D A matrix. In general the output from \code{\link[mcglm]{mc_link_function}}.
#' @param inv_C A matrix. In general the output from \code{\link[mcglm]{mc_build_C}}.
#' @param y_vec A vector.
#' @param mu_vec A vector.
#' @return The quasi-score vector, the Sensivity and variability matrices.
#' @export
mc_quasi_score <- function(D, inv_C, y_vec, mu_vec) {
- res <- y_vec - mu_vec
- t_D <- t(D)
- score <-t_D%*%(inv_C%*%res)
- sensitivity <- -t_D%*%inv_C%*%D
- output <- list("Score" = score, "Sensitivity" = sensitivity, "Variability" = -sensitivity)
- return(output)
+ res <- y_vec - mu_vec
+ t_D <- t(D)
+ score <- t_D %*% (inv_C %*% res)
+ sensitivity <- -t_D %*% inv_C %*% D
+ output <- list(Score = score, Sensitivity = sensitivity, Variability = -sensitivity)
+ return(output)
}
diff --git a/R/mc_residuals.mcglm.R b/R/mc_residuals.mcglm.R
index e0a023f..0a6b1b7 100644
--- a/R/mc_residuals.mcglm.R
+++ b/R/mc_residuals.mcglm.R
@@ -3,21 +3,19 @@
#' @description Compute residuals based on fitting mcglm models.
#'
#' @param object An of class mcglm, typically the result of a call to \code{mcglm}.
-#' @param type the type of residuals which should be returned. The alternatives are: "raw"
-#' (default), "pearson" and "standardized".
+#' @param type the type of residuals which should be returned. The alternatives are: 'raw'
+#' (default), 'pearson' and 'standardized'.
#' @return Depending on the number of response variable the function \code{residuals.mcglm} returns
#' a vector (univariate models) or a matrix (multivariate models) of residuals values.
#' @export
residuals.mcglm <- function(object, type = "raw") {
- n_resp <- length(object$beta_names)
- output <- Matrix(object$residuals, ncol = n_resp, nrow = object$n_obs)
- if(type == "standardized") {
- output <- Matrix(as.numeric(object$residuals%*%chol(object$inv_C)), ncol = n_resp,
- nrow = object$n_obs)
- }
- if(type == "pearson") {
- output <- Matrix(as.numeric(object$residuals/sqrt(diag(object$C))),
- ncol = n_resp, nrow = object$n_obs)
- }
- return(output)
-}
+ n_resp <- length(object$beta_names)
+ output <- Matrix(object$residuals, ncol = n_resp, nrow = object$n_obs)
+ if (type == "standardized") {
+ output <- Matrix(as.numeric(object$residuals %*% chol(object$inv_C)), ncol = n_resp, nrow = object$n_obs)
+ }
+ if (type == "pearson") {
+ output <- Matrix(as.numeric(object$residuals/sqrt(diag(object$C))), ncol = n_resp, nrow = object$n_obs)
+ }
+ return(output)
+}
diff --git a/R/mc_robust_std.R b/R/mc_robust_std.R
index 0179196..f881d80 100644
--- a/R/mc_robust_std.R
+++ b/R/mc_robust_std.R
@@ -4,20 +4,22 @@
#' observations.
#'
#' @param object An object of mcglm class.
-#' @param id a vector which identifies the clusters. The length and order of ‘id’ should be the
+#' @param id a vector which identifies the clusters. The length and order of id should be the
#' same as the number of observations. Data are assumed to be sorted so that observations on a cluster
#' are contiguous rows for all entities in the formula.
#' @return A matrix. Note that the function assumes that the data are in the correct order.
#' @export
mc_robust_std <- function(object, id) {
- inv_M <- object$inv_S_beta
- temp_data <- data.frame("res" = object$residuals, id)
- temp_data_group <- split(temp_data, temp_data$id)
- r_rT <- bdiag(lapply(temp_data_group, function(x){tcrossprod(x[,1])}))
- D <- bdiag(lapply(object$mu_list, function(x)x$D))
- p1 <- object$inv_C%*%D
- V_robust <- inv_M%*%(t(p1)%*%r_rT%*%p1)%*%inv_M
- output <- sqrt(diag(V_robust))
- return(output)
- }
+ inv_M <- object$inv_S_beta
+ temp_data <- data.frame(res = object$residuals, id)
+ temp_data_group <- split(temp_data, temp_data$id)
+ r_rT <- bdiag(lapply(temp_data_group, function(x) {
+ tcrossprod(x[, 1])
+ }))
+ D <- bdiag(lapply(object$mu_list, function(x) x$D))
+ p1 <- object$inv_C %*% D
+ V_robust <- inv_M %*% (t(p1) %*% r_rT %*% p1) %*% inv_M
+ output <- sqrt(diag(V_robust))
+ return(output)
+}
diff --git a/R/mc_rw1.R b/R/mc_rw1.R
index 7ee8087..519ca71 100644
--- a/R/mc_rw1.R
+++ b/R/mc_rw1.R
@@ -7,23 +7,23 @@
#' @return A matrix. Note that the function assumes that the data are in the correct order.
#' @export
mc_rw1 <- function(n_time, intrinsic = TRUE) {
- R <- Matrix(0,nrow = n_time, ncol = n_time, sparse = TRUE)
- ## Border restriction
- ncol = n_time
- R[1,c(1,2)] <- c(1,-1)
- R[ncol,c(ncol-1,ncol)] <- c(-1,1)
- ## Body of matrix
- n <- ncol-1
- for(i in 2:n) {
- R[i,c(i-1,i,i+1)] <- c(-1,2,-1)
- }
- if(intrinsic == TRUE) {output <- list(R)}
- if(intrinsic == FALSE) {
- R1 <- Diagonal(n_time, diag(R))
- diag(R) <- 0
- output <- list("Z1" = R1, "Z2" = R)
- }
- return(output)
-}
-
-
+ R <- Matrix(0, nrow = n_time, ncol = n_time, sparse = TRUE)
+ ## Border restriction
+ ncol = n_time
+ R[1, c(1, 2)] <- c(1, -1)
+ R[ncol, c(ncol - 1, ncol)] <- c(-1, 1)
+ ## Body of matrix
+ n <- ncol - 1
+ for (i in 2:n) {
+ R[i, c(i - 1, i, i + 1)] <- c(-1, 2, -1)
+ }
+ if (intrinsic == TRUE) {
+ output <- list(R)
+ }
+ if (intrinsic == FALSE) {
+ R1 <- Diagonal(n_time, diag(R))
+ diag(R) <- 0
+ output <- list(Z1 = R1, Z2 = R)
+ }
+ return(output)
+}
diff --git a/R/mc_rw2.R b/R/mc_rw2.R
index d786811..6788c1c 100644
--- a/R/mc_rw2.R
+++ b/R/mc_rw2.R
@@ -7,23 +7,25 @@
#' @return A matrix. Note that the function assumes that the data are in the correct order.
#' @export
mc_rw2 <- function(n_time, intrinsic = TRUE) {
- R <- Matrix(0,nrow = n_time, ncol = n_time, sparse = TRUE)
- ## Border restriction
- ncol = n_time
- R[1,c(1,2,3)] <- c(1,-2,1)
- R[ncol,c(c(ncol-2):ncol)] <- c(1,-2,1)
- R[2,c(1:4)] <- c(-2,5,-4,1)
- R[c(ncol-1),c(c(ncol-3):c(ncol))] <- c(1,-4,5,-2)
- ## Body of matrix
- n <- ncol-2
- for(i in 3:n) {
- R[i,c(i-2,i-1, i, i+1, i+2)] <- c(1,-4,6,-4,1)
- }
- if(intrinsic == TRUE) {output <- list(R)}
- if(intrinsic == FALSE) {
- R1 <- Diagonal(n_time, diag(R))
- diag(R) <- 0
- output <- list("Z1" = R1, "Z2" = R)
- }
- return(output)
-}
+ R <- Matrix(0, nrow = n_time, ncol = n_time, sparse = TRUE)
+ ## Border restriction
+ ncol = n_time
+ R[1, c(1, 2, 3)] <- c(1, -2, 1)
+ R[ncol, c(c(ncol - 2):ncol)] <- c(1, -2, 1)
+ R[2, c(1:4)] <- c(-2, 5, -4, 1)
+ R[c(ncol - 1), c(c(ncol - 3):c(ncol))] <- c(1, -4, 5, -2)
+ ## Body of matrix
+ n <- ncol - 2
+ for (i in 3:n) {
+ R[i, c(i - 2, i - 1, i, i + 1, i + 2)] <- c(1, -4, 6, -4, 1)
+ }
+ if (intrinsic == TRUE) {
+ output <- list(R)
+ }
+ if (intrinsic == FALSE) {
+ R1 <- Diagonal(n_time, diag(R))
+ diag(R) <- 0
+ output <- list(Z1 = R1, Z2 = R)
+ }
+ return(output)
+}
diff --git a/R/mc_sensitivity.R b/R/mc_sensitivity.R
index 540dcdd..8b2e8e7 100644
--- a/R/mc_sensitivity.R
+++ b/R/mc_sensitivity.R
@@ -7,17 +7,18 @@
#' @details This function implements the equation 7 of Bonat and Jorgensen (2015).
#' @export
mc_sensitivity <- function(product) {
- n_par <- length(product)
- Sensitivity <- matrix(0, n_par, n_par)
- Sensitivity_temp <- matrix(0, n_par, n_par)
- for(i in 1:n_par) {
- for(j in 1:n_par) {
- #Sensitivity[i,j] <- -sum(diag(product[[i]]%*%product[[j]]))
- Sensitivity[i,j] <- -sum(t(product[[i]])*product[[j]])
+ n_par <- length(product)
+ Sensitivity <- matrix(0, n_par, n_par)
+ Sensitivity_temp <- matrix(0, n_par, n_par)
+ Sensitivity1 <- matrix(0, n_par, n_par)
+ for (i in 1:n_par) {
+ for (j in 1:n_par) {
+ # Sensitivity_temp[i,j] <- -sum(diag(product[[i]]%*%product[[j]]))
+ Sensitivity[i, j] <- -sum(t(product[[i]]) * product[[j]])
+ # Sensitivity1[i,j] <- -sum(product[[i]]*product[[j]])
+ }
}
- }
- #print(forceSymmetric(Sensitivity))
- #print(forceSymmetric(Sensitivity_temp))
- #print(all.equal(Sensitivity, Sensitivity_temp))
- return(Sensitivity)
-}
+ # print(forceSymmetric(Sensitivity)) print(forceSymmetric(Sensitivity_temp)) print(forceSymmetric(Sensitivity1))
+ # print(all.equal(Sensitivity1, Sensitivity))
+ return(Sensitivity)
+}
diff --git a/R/mc_summary.mcglm.R b/R/mc_summary.mcglm.R
index acd3c21..acf9ed9 100644
--- a/R/mc_summary.mcglm.R
+++ b/R/mc_summary.mcglm.R
@@ -3,54 +3,59 @@
#' @description Summary for mcglm objects.
#'
#' @param object an object of class mcglm, usually, a result of a call to \code{mcglm}.
-#' @exportMethod
+#' @export
summary.mcglm <- function(object) {
- n_resp <- length(object$beta_names)
- output <- list()
- for(i in 1:n_resp) {
- cat("Call: ")
- print(object$linear_pred[[i]])
+ n_resp <- length(object$beta_names)
+ output <- list()
+ for (i in 1:n_resp) {
+ cat("Call: ")
+ print(object$linear_pred[[i]])
+ cat("\n")
+ cat("Link function:", object$link[[i]])
+ cat("\n")
+ cat("Variance function:", object$variance[[i]])
+ cat("\n")
+ cat("Covariance function:", object$covariance[[i]])
+ cat("\n")
+ cat("Regression:\n")
+ tab_beta <- coef(object, std.error = TRUE, response = i, type = "beta")[, 1:2]
+ tab_beta$"Z value" <- tab_beta[, 1]/tab_beta[, 2]
+ rownames(tab_beta) <- object$beta_names[[i]]
+ output[i][[1]]$Regression <- tab_beta
+ print(tab_beta)
+ cat("\n")
+ tab_power <- coef(object, std.error = TRUE, response = i, type = "power")[, 1:2]
+ tab_power$"Z value" <- tab_power[, 1]/tab_power[, 2]
+ rownames(tab_power) <- NULL
+ if (dim(tab_power)[1] != 0) {
+ cat("Power:\n")
+ print(tab_power)
+ output[i][[1]]$Power <- tab_power
+ cat("\n")
+ }
+ cat("Dispersion:\n")
+ tab_tau <- coef(object, std.error = TRUE, response = i, type = "tau")[, 1:2]
+ tab_tau$"Z value" <- tab_tau[, 1]/tab_tau[, 2]
+ rownames(tab_tau) <- NULL
+ output[i][[1]]$tau <- tab_tau
+ print(tab_tau)
+ cat("\n")
+ }
+ tab_rho <- coef(object, std.error = TRUE, response = NA, type = "correlation")[, c(3, 1, 2)]
+ tab_rho$"Z value" <- tab_rho[, 2]/tab_rho[, 3]
+ if (dim(tab_rho)[1] != 0) {
+ cat("Correlation matrix:\n")
+ print(tab_rho)
+ cat("\n")
+ }
+ names(object$con$correct) <- ""
+ iteration_cov <- length(na.exclude(object$IterationCovariance[, 1]))
+ names(iteration_cov) <- ""
+ names(object$con$method) <- ""
+ cat("Algorithm:", object$con$method)
cat("\n")
- cat("Link function:", object$link[[i]])
+ cat("Correction:", object$con$correct)
cat("\n")
- cat("Variance function:", object$variance[[i]])
- cat("\n")
- cat("Covariance function:", object$covariance[[i]])
- cat("\n")
- cat("Regression:\n")
- tab_beta <- coef(object, std.error = TRUE, response = i, type = "beta")[,1:2]
- tab_beta$"z value" <- tab_beta[,1]/tab_beta[,2]
- rownames(tab_beta) <- object$beta_names[[i]]
- output[i][[1]]$Regression <- tab_beta
- print(tab_beta)
- cat("\n")
- cat("Power:\n")
- tab_power <- coef(object, std.error = TRUE, response = i, type = "power")[,1:2]
- tab_power$'z value' <- tab_power[,1]/tab_power[,2]
- rownames(tab_power) <- NULL
- output[i][[1]]$Power <- tab_power
- print(tab_power)
- cat("\n")
- cat("tau:\n")
- tab_tau <- coef(object, std.error = TRUE, response = i, type = "tau")[,1:2]
- tab_tau$'z value' <- tab_tau[,1]/tab_tau[,2]
- rownames(tab_tau) <- NULL
- output[i][[1]]$tau <- tab_tau
- print(tab_tau)
- cat("\n")
- }
- cat("Correlation matrix:\n")
- tab_rho <- coef(object, std.error = TRUE, response = NA, type = "correlation")[,c(3,1,2)]
- tab_rho$'z value' <- tab_rho[,2]/tab_rho[,3]
- print(tab_rho)
- cat("\n")
- cat("Algorithm:")
- print(object$con$method)
- cat("Correction:")
- print(object$con$correct)
- cat("Number iterations:")
- iteration_cov <- length(na.exclude(object$IterationCovariance[,1]))
- print(iteration_cov)
+ cat("Number iterations:", iteration_cov)
}
-
diff --git a/R/mc_transform_list_bdiag.R b/R/mc_transform_list_bdiag.R
index bea0ca1..d6c82cd 100644
--- a/R/mc_transform_list_bdiag.R
+++ b/R/mc_transform_list_bdiag.R
@@ -10,11 +10,10 @@
#' @return A list of block-diagonal matrices.
#' @export
mc_transform_list_bdiag <- function(list_mat, mat_zero, response_number) {
- aux.f <- function(x, mat_zero, response_number) {
- mat_zero[[response_number]] <- x
- return(bdiag(mat_zero))
- }
- output <- lapply(list_mat, aux.f, mat_zero = mat_zero, response_number = response_number)
- return(output)
-}
-
+ aux.f <- function(x, mat_zero, response_number) {
+ mat_zero[[response_number]] <- x
+ return(bdiag(mat_zero))
+ }
+ output <- lapply(list_mat, aux.f, mat_zero = mat_zero, response_number = response_number)
+ return(output)
+}
diff --git a/R/mc_unstructured.R b/R/mc_unstructured.R
index 6c229d1..afc67eb 100644
--- a/R/mc_unstructured.R
+++ b/R/mc_unstructured.R
@@ -2,21 +2,21 @@
#'
#' @description Builds a unstructured model matrix.
#'
-#' @param n_time Number of observation per unit sample.
+#' @param n_time Number of observations per unit sample.
#' @return A matrix. Note that the function assumes that the data are in the correct order.
#' @export
-mc_unstructured <- function(n_time){
- mat.temp <- Matrix(0, ncol = n_time, nrow = n_time, sparse = TRUE)
- non.diagonal.terms <- list()
- non.diagonal <- t(combn(n_time, 2))
- n.cor.par <- dim(non.diagonal)[1]
- ID <- Diagonal(n_replication, 1)
- ## Covariance elementary matrices
- for(i in 1:n.cor.par){
- non.diagonal.terms[i][[1]] <- mat.temp
- non.diagonal.terms[i][[1]][non.diagonal[i,1],non.diagonal[i,2]] <-
- non.diagonal.terms[i][[1]][non.diagonal[i,2],non.diagonal[i,1]]<- 1}
- ## Output
- return(non.diagonal.terms)
+mc_unstructured <- function(n_time) {
+ mat.temp <- Matrix(0, ncol = n_time, nrow = n_time, sparse = TRUE)
+ non.diagonal.terms <- list()
+ non.diagonal <- t(combn(n_time, 2))
+ n.cor.par <- dim(non.diagonal)[1]
+ ## Covariance elementary matrices
+ for (i in 1:n.cor.par) {
+ non.diagonal.terms[i][[1]] <- mat.temp
+ non.diagonal.terms[i][[1]][non.diagonal[i, 1], non.diagonal[i, 2]] <- non.diagonal.terms[i][[1]][non.diagonal[i,
+ 2], non.diagonal[i, 1]] <- 1
+ }
+ ## Output
+ return(non.diagonal.terms)
}
diff --git a/R/mc_updatedBeta.R b/R/mc_updatedBeta.R
index 8f4ec1f..e71f14c 100644
--- a/R/mc_updatedBeta.R
+++ b/R/mc_updatedBeta.R
@@ -11,10 +11,10 @@
#' @return A list with updated values of the regression parameters.
#' @export
mc_updateBeta <- function(list_initial, betas, information, n_resp) {
- cod <- rep(1:n_resp,information$n_betas)
- temp <- data.frame("beta" = betas,cod)
- for(k in 1:n_resp) {
- list_initial$regression[[k]] <- temp[which(temp$cod == k),]$beta
+ cod <- rep(1:n_resp, information$n_betas)
+ temp <- data.frame(beta = betas, cod)
+ for (k in 1:n_resp) {
+ list_initial$regression[[k]] <- temp[which(temp$cod == k), ]$beta
}
- return(list_initial)
-}
+ return(list_initial)
+}
diff --git a/R/mc_updatedCov.R b/R/mc_updatedCov.R
index 47009b8..a28ae7f 100644
--- a/R/mc_updatedCov.R
+++ b/R/mc_updatedCov.R
@@ -13,27 +13,26 @@
#' @return A list with updated values of the covariance parameters.
#' @export
mc_updateCov <- function(list_initial, covariance, list_power_fixed, information, n_resp) {
- rho_cod <- rep("rho", information$n_rho)
- tau_cod <- list()
- power_cod <- list()
- for(i in 1:n_resp) {
- power_cod[[i]] <- rep(paste("power", i, sep=""), information$n_power[[i]])
- tau_cod[[i]] <- rep(paste("tau",i, sep = ""), information$n_tau[[i]])
- }
- temp <- data.frame("values" = covariance,
- "cod" = c(rho_cod, do.call(c, Map(c, power_cod, tau_cod))))
- cod.tau <- paste("tau", 1:n_resp, sep = "")
- for(i in 1:n_resp) {
- list_initial$tau[[i]] <- temp[which(temp$cod == cod.tau[i]),]$values
- }
- cod.power <- paste("power",1:n_resp, sep="")
- for(i in 1:n_resp) {
- if(list_power_fixed[[i]] == FALSE) {
- list_initial$power[[i]] <- temp[which(temp$cod == cod.power[i]),]$values
+ rho_cod <- rep("rho", information$n_rho)
+ tau_cod <- list()
+ power_cod <- list()
+ for (i in 1:n_resp) {
+ power_cod[[i]] <- rep(paste("power", i, sep = ""), information$n_power[[i]])
+ tau_cod[[i]] <- rep(paste("tau", i, sep = ""), information$n_tau[[i]])
}
- }
- if(length(information$n_betas) != 1) {
- list_initial$rho <- temp[which(temp$cod == "rho"),]$values
- }
- return(list_initial)
-}
+ temp <- data.frame(values = covariance, cod = c(rho_cod, do.call(c, Map(c, power_cod, tau_cod))))
+ cod.tau <- paste("tau", 1:n_resp, sep = "")
+ for (i in 1:n_resp) {
+ list_initial$tau[[i]] <- temp[which(temp$cod == cod.tau[i]), ]$values
+ }
+ cod.power <- paste("power", 1:n_resp, sep = "")
+ for (i in 1:n_resp) {
+ if (list_power_fixed[[i]] == FALSE) {
+ list_initial$power[[i]] <- temp[which(temp$cod == cod.power[i]), ]$values
+ }
+ }
+ if (length(information$n_betas) != 1) {
+ list_initial$rho <- temp[which(temp$cod == "rho"), ]$values
+ }
+ return(list_initial)
+}
diff --git a/R/mc_variability.R b/R/mc_variability.R
index d45fa89..6775da3 100644
--- a/R/mc_variability.R
+++ b/R/mc_variability.R
@@ -10,16 +10,16 @@
#' @return The variability matrix associated witht the Pearson estimating function.
#' @details This function implements the equation 8 of Bonat and Jorgensen (2015).
#' @export
-mc_variability <- function(sensitivity, product, inv_C, C, res){
- W <- lapply(product, mc_multiply2, bord2 = inv_C)
- n_par <- length(product)
- Variability <- matrix(NA, ncol = n_par, nrow = n_par) # Take care here, I have found
- #problems when using Matrix classes in this step. I do not know why yet!!
- k4 <- res^4 - 3*diag(C)^2
- for(i in 1:n_par) {
- for(j in 1:n_par) {
- Variability[i,j] <- as.numeric(-2*sensitivity[i,j] + sum(k4*diag(W[[i]])*diag(W[[j]])))
+mc_variability <- function(sensitivity, product, inv_C, C, res) {
+ W <- lapply(product, mc_multiply2, bord2 = inv_C)
+ n_par <- length(product)
+ Variability <- matrix(NA, ncol = n_par, nrow = n_par) # Take care here, I have found
+ # problems when using Matrix classes in this step. I do not know why yet!!
+ k4 <- res^4 - 3 * diag(C)^2
+ for (i in 1:n_par) {
+ for (j in 1:n_par) {
+ Variability[i, j] <- as.numeric(-2 * sensitivity[i, j] + sum(k4 * diag(W[[i]]) * diag(W[[j]])))
+ }
}
- }
- return(Variability)
-}
+ return(Variability)
+}
diff --git a/R/mc_variance_function.R b/R/mc_variance_function.R
index b861263..f6fb9a6 100644
--- a/R/mc_variance_function.R
+++ b/R/mc_variance_function.R
@@ -22,193 +22,169 @@
#' For example, if \code{inverse = FALSE}, \code{derivative_power = TRUE} and
#' \code{derivative_mu = TRUE}. The output will be a list, with three
#' elements: V_sqrt, D_V_sqrt_power and D_V_sqrt_mu.
+#' @export
#' @examples
#' x1 <- seq(-1,1,l = 5)
#' X <- model.matrix(~ x1)
-#' mu <- mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = "logit")
-#' mc_variance_function(mu = mu$mu, power = c(2,1), Ntrial = 1, variance = "binomialPQ",
+#' mu <- mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = 'logit')
+#' mc_variance_function(mu = mu$mu, power = c(2,1), Ntrial = 1, variance = 'binomialPQ',
#' inverse = FALSE, derivative_power = TRUE, derivative_mu = TRUE)
# Generic variance function ---------------------------
-mc_variance_function <- function(mu, power, Ntrial, variance, inverse,
- derivative_power, derivative_mu) {
- assert_that(is.logical(inverse))
- assert_that(is.logical(derivative_power))
- assert_that(is.logical(derivative_mu))
- switch(variance,
- "power" = { output <- mc_power(mu = mu, power = power, inverse = inverse,
- derivative_power = derivative_power,
- derivative_mu = derivative_mu) },
- "binomialP" = { output <- mc_binomialP(mu = mu, power = power, Ntrial = Ntrial,
- inverse = inverse,
- derivative_power = derivative_power,
- derivative_mu = derivative_mu) },
- "binomialPQ" = { output <- mc_binomialPQ(mu = mu, power = power, Ntrial = Ntrial,
- inverse = inverse,
- derivative_power = derivative_power,
- derivative_mu = derivative_mu) },
- stop(gettextf("%s variance function not recognised", sQuote(variance)),domain = NA))
- return(output)
+mc_variance_function <- function(mu, power, Ntrial, variance, inverse, derivative_power, derivative_mu) {
+ assert_that(is.logical(inverse))
+ assert_that(is.logical(derivative_power))
+ assert_that(is.logical(derivative_mu))
+ switch(variance, power = {
+ output <- mc_power(mu = mu, power = power, inverse = inverse, derivative_power = derivative_power, derivative_mu = derivative_mu)
+ }, binomialP = {
+ output <- mc_binomialP(mu = mu, power = power, Ntrial = Ntrial, inverse = inverse, derivative_power = derivative_power,
+ derivative_mu = derivative_mu)
+ }, binomialPQ = {
+ output <- mc_binomialPQ(mu = mu, power = power, Ntrial = Ntrial, inverse = inverse, derivative_power = derivative_power,
+ derivative_mu = derivative_mu)
+ }, stop(gettextf("%s variance function not recognised", sQuote(variance)), domain = NA))
+ return(output)
}
#' @rdname mc_variance_function
# Power variance function ---------------------------
mc_power <- function(mu, power, inverse, derivative_power, derivative_mu) {
- assert_that(all(mu > 0)) # The observed value can be zero, but not the expected value.
- assert_that(is.number(power))
- mu.power <- mu^power
- sqrt.mu.power <- sqrt(mu.power)
- n <- length(mu)
- if (inverse == TRUE & derivative_power == TRUE & derivative_mu == FALSE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu.power),
- "D_V_inv_sqrt_power" = Diagonal(n=n, -(mu.power*log(mu))/(2*(mu.power)^(1.5))))
- }
- if (inverse == TRUE & derivative_power == FALSE & derivative_mu == FALSE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu.power))
- }
- if (inverse == FALSE & derivative_power == TRUE & derivative_mu == FALSE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu.power),
- "D_V_sqrt_power" = Diagonal(n = n, (mu.power*log(mu))/(2*sqrt.mu.power)))
- }
- if (inverse == FALSE & derivative_power == FALSE & derivative_mu == FALSE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu.power))
- }
- if (inverse == TRUE & derivative_power == TRUE & derivative_mu == TRUE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu.power),
- "D_V_inv_sqrt_power" = Diagonal(n=n, -(mu.power*log(mu))/(2*(mu.power)^(1.5))),
- "D_V_inv_sqrt_mu" = -(mu^(power-1)*power)/(2*(mu.power)^(1.5)))
- }
- if (inverse == TRUE & derivative_power == FALSE & derivative_mu == TRUE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu.power),
- "D_V_inv_sqrt_mu" = -(mu^(power-1)*power)/(2*(mu.power)^(1.5)))
- }
- if (inverse == FALSE & derivative_power == TRUE & derivative_mu == TRUE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu.power),
- "D_V_sqrt_power" = Diagonal(n = n, (mu.power*log(mu))/(2*sqrt.mu.power)),
- "D_V_sqrt_mu" = (mu^(power-1)*power)/(2*sqrt.mu.power))
- }
- if (inverse == FALSE & derivative_power == FALSE & derivative_mu == TRUE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu.power),
- "D_V_sqrt_mu" = (mu^(power-1)*power)/(2*sqrt.mu.power))
- }
- return(output)
+ assert_that(all(mu > 0)) # The observed value can be zero, but not the expected value.
+ assert_that(is.number(power))
+ mu.power <- mu^power
+ sqrt.mu.power <- sqrt(mu.power)
+ n <- length(mu)
+ if (inverse == TRUE & derivative_power == TRUE & derivative_mu == FALSE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu.power), D_V_inv_sqrt_power = Diagonal(n = n, -(mu.power *
+ log(mu))/(2 * (mu.power)^(1.5))))
+ }
+ if (inverse == TRUE & derivative_power == FALSE & derivative_mu == FALSE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu.power))
+ }
+ if (inverse == FALSE & derivative_power == TRUE & derivative_mu == FALSE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu.power), D_V_sqrt_power = Diagonal(n = n, (mu.power * log(mu))/(2 *
+ sqrt.mu.power)))
+ }
+ if (inverse == FALSE & derivative_power == FALSE & derivative_mu == FALSE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu.power))
+ }
+ if (inverse == TRUE & derivative_power == TRUE & derivative_mu == TRUE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu.power), D_V_inv_sqrt_power = Diagonal(n = n, -(mu.power *
+ log(mu))/(2 * (mu.power)^(1.5))), D_V_inv_sqrt_mu = -(mu^(power - 1) * power)/(2 * (mu.power)^(1.5)))
+ }
+ if (inverse == TRUE & derivative_power == FALSE & derivative_mu == TRUE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu.power), D_V_inv_sqrt_mu = -(mu^(power - 1) * power)/(2 * (mu.power)^(1.5)))
+ }
+ if (inverse == FALSE & derivative_power == TRUE & derivative_mu == TRUE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu.power), D_V_sqrt_power = Diagonal(n = n, (mu.power * log(mu))/(2 *
+ sqrt.mu.power)), D_V_sqrt_mu = (mu^(power - 1) * power)/(2 * sqrt.mu.power))
+ }
+ if (inverse == FALSE & derivative_power == FALSE & derivative_mu == TRUE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu.power), D_V_sqrt_mu = (mu^(power - 1) * power)/(2 * sqrt.mu.power))
+ }
+ return(output)
}
#' @rdname mc_variance_function
# BinomialP variance function ---------------------------
mc_binomialP <- function(mu, power, inverse, Ntrial, derivative_power, derivative_mu) {
- assert_that(all(mu > 0)) # The observed value can be 0 and 1, but not the expected value
- assert_that(all(mu < 1))
- assert_that(is.number(power))
- assert_that(all(Ntrial > 0))
- constant <- (1/Ntrial)
- mu.power <- mu^power
- mu.power1 <- (1-mu)^power
- mu1mu <- constant*(mu.power*mu.power1)
- sqrt.mu1mu <- sqrt(mu1mu)
- n <- length(mu)
- if (inverse == TRUE & derivative_power == TRUE & derivative_mu == FALSE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu),
- "D_V_inv_sqrt_power" = Diagonal(n=n, -(log(1-mu)*mu1mu +
- log(mu)*mu1mu)/(2*(mu1mu^(1.5))) ) )
- }
- if (inverse == TRUE & derivative_power == FALSE & derivative_mu == FALSE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu ) )
- }
- if (inverse == FALSE & derivative_power == TRUE & derivative_mu == FALSE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu ),
- "D_V_sqrt_power" = Diagonal(n = n, (log(1-mu)*mu1mu + log(mu)*mu1mu)/
- (2*sqrt.mu1mu) ) )
- }
- if (inverse == FALSE & derivative_power == FALSE & derivative_mu == FALSE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu))
- }
- if (inverse == TRUE & derivative_power == TRUE & derivative_mu == TRUE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu),
- "D_V_inv_sqrt_power" = Diagonal(n=n, -(log(1-mu)*mu1mu + log(mu)*mu1mu)/
- (2*(mu1mu^(1.5)))) ,
- "D_V_inv_sqrt_mu" = -(constant*(mu.power1*(mu^(power-1))*power) -
- constant*(((1-mu)^(power-1))*mu.power*power))/(2*(mu1mu^(1.5))))
- }
- if (inverse == TRUE & derivative_power == FALSE & derivative_mu == TRUE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu),
- "D_V_inv_sqrt_mu" = -(constant*(mu.power1*(mu^(power-1))*power) -
- constant*(((1-mu)^(power-1))*mu.power*power))/(2*(mu1mu^(1.5))))
- }
- if (inverse == FALSE & derivative_power == TRUE & derivative_mu == TRUE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu),
- "D_V_sqrt_power" = Diagonal(n = n, (log(1-mu)*mu1mu + log(mu)*mu1mu)/
- (2*sqrt.mu1mu) ),
- "D_V_sqrt_mu" = (constant*(mu.power1*(mu^(power-1))*power) -
- constant*(((1-mu)^(power-1))*mu.power*power))/(2*sqrt.mu1mu))
- }
- if (inverse == FALSE & derivative_power == FALSE & derivative_mu == TRUE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu),
- "D_V_sqrt_mu" = (constant*(mu.power1*(mu^(power-1))*power) -
- constant*(((1-mu)^(power-1))*mu.power*power))/(2*sqrt.mu1mu))
- }
- return(output)
+ assert_that(all(mu > 0)) # The observed value can be 0 and 1, but not the expected value
+ assert_that(all(mu < 1))
+ assert_that(is.number(power))
+ assert_that(all(Ntrial > 0))
+ constant <- (1/Ntrial)
+ mu.power <- mu^power
+ mu.power1 <- (1 - mu)^power
+ mu1mu <- constant * (mu.power * mu.power1)
+ sqrt.mu1mu <- sqrt(mu1mu)
+ n <- length(mu)
+ if (inverse == TRUE & derivative_power == TRUE & derivative_mu == FALSE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu), D_V_inv_sqrt_power = Diagonal(n = n, -(log(1 - mu) *
+ mu1mu + log(mu) * mu1mu)/(2 * (mu1mu^(1.5)))))
+ }
+ if (inverse == TRUE & derivative_power == FALSE & derivative_mu == FALSE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu))
+ }
+ if (inverse == FALSE & derivative_power == TRUE & derivative_mu == FALSE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu), D_V_sqrt_power = Diagonal(n = n, (log(1 - mu) * mu1mu + log(mu) *
+ mu1mu)/(2 * sqrt.mu1mu)))
+ }
+ if (inverse == FALSE & derivative_power == FALSE & derivative_mu == FALSE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu))
+ }
+ if (inverse == TRUE & derivative_power == TRUE & derivative_mu == TRUE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu), D_V_inv_sqrt_power = Diagonal(n = n, -(log(1 - mu) *
+ mu1mu + log(mu) * mu1mu)/(2 * (mu1mu^(1.5)))), D_V_inv_sqrt_mu = -(constant * (mu.power1 * (mu^(power - 1)) *
+ power) - constant * (((1 - mu)^(power - 1)) * mu.power * power))/(2 * (mu1mu^(1.5))))
+ }
+ if (inverse == TRUE & derivative_power == FALSE & derivative_mu == TRUE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu), D_V_inv_sqrt_mu = -(constant * (mu.power1 * (mu^(power -
+ 1)) * power) - constant * (((1 - mu)^(power - 1)) * mu.power * power))/(2 * (mu1mu^(1.5))))
+ }
+ if (inverse == FALSE & derivative_power == TRUE & derivative_mu == TRUE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu), D_V_sqrt_power = Diagonal(n = n, (log(1 - mu) * mu1mu + log(mu) *
+ mu1mu)/(2 * sqrt.mu1mu)), D_V_sqrt_mu = (constant * (mu.power1 * (mu^(power - 1)) * power) - constant * (((1 -
+ mu)^(power - 1)) * mu.power * power))/(2 * sqrt.mu1mu))
+ }
+ if (inverse == FALSE & derivative_power == FALSE & derivative_mu == TRUE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu), D_V_sqrt_mu = (constant * (mu.power1 * (mu^(power - 1)) * power) -
+ constant * (((1 - mu)^(power - 1)) * mu.power * power))/(2 * sqrt.mu1mu))
+ }
+ return(output)
}
#' @rdname mc_variance_function
# BinomialPQ variance function ---------------------------
mc_binomialPQ <- function(mu, power, inverse, Ntrial, derivative_power, derivative_mu) {
- assert_that(all(mu > 0)) # The observed value can be 0 and 1, but not the expected value
- assert_that(all(mu < 1))
- assert_that(length(power) == 2)
- assert_that(all(Ntrial > 0))
- constant <- (1/Ntrial)
- p <- power[1]
- q <- power[2]
- mu.p <- mu^p
- mu1.q <- (1-mu)^q
- mu.p.mu.q <- mu.p*mu1.q
- mu1mu <- mu.p.mu.q*constant
- sqrt.mu1mu <- sqrt(mu1mu)
- n <- length(mu)
- if (inverse == TRUE & derivative_power == TRUE & derivative_mu == FALSE) {
- denominator <- (2* (mu1mu^1.5) * Ntrial)
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu),
- "D_V_inv_sqrt_p" = Diagonal(n=n, - (mu.p.mu.q*log(mu))/denominator),
- "D_V_inv_sqrt_q" = Diagonal(n=n, - mu.p.mu.q*log(1-mu)/denominator))
- }
- if (inverse == TRUE & derivative_power == FALSE & derivative_mu == FALSE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu ) )
- }
- if (inverse == FALSE & derivative_power == TRUE & derivative_mu == FALSE) {
- denominator <- 2*sqrt.mu1mu*Ntrial
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu ),
- "D_V_sqrt_p" = Diagonal(n = n, (mu.p.mu.q*log(mu))/denominator),
- "D_V_sqrt_q" = Diagonal(n = n, (mu.p.mu.q*log(1-mu))/denominator))
- }
- if (inverse == FALSE & derivative_power == FALSE & derivative_mu == FALSE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu))
- }
- if (inverse == TRUE & derivative_power == TRUE & derivative_mu == TRUE) {
- denominator <- (2* (mu1mu^1.5) * Ntrial)
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu),
- "D_V_inv_sqrt_p" = Diagonal(n=n, - (mu.p.mu.q*log(mu))/denominator),
- "D_V_inv_sqrt_q" = Diagonal(n=n, - mu.p.mu.q*log(1-mu)/denominator),
- "D_V_inv_sqrt_mu" = - (constant*(mu1.q* (mu^(p-1))*p) -
- constant*(((1-mu)^(q-1))*mu.p*q))/ (2*(mu1mu^1.5)))
- }
- if (inverse == TRUE & derivative_power == FALSE & derivative_mu == TRUE) {
- output <- list("V_inv_sqrt" = Diagonal(n = n, 1/sqrt.mu1mu),
- "D_V_inv_sqrt_mu" = - (constant*(mu1.q* (mu^(p-1))*p) -
- constant*(((1-mu)^(q-1))*mu.p*q))/ (2*(mu1mu^1.5)))
- }
- if (inverse == FALSE & derivative_power == TRUE & derivative_mu == TRUE) {
- denominator1 <- 2*sqrt.mu1mu
- denominator2 <- denominator1*Ntrial
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu),
- "D_V_sqrt_p" = Diagonal(n = n, (mu.p.mu.q*log(mu))/denominator2),
- "D_V_sqrt_q" = Diagonal(n = n, (mu.p.mu.q*log(1-mu))/denominator2),
- "D_V_sqrt_mu" = (constant*(mu1.q*(mu^(p-1))*p) -
- constant*(((1-mu)^(q-1))*mu.p*q))/denominator1 )
- }
- if (inverse == FALSE & derivative_power == FALSE & derivative_mu == TRUE) {
- output <- list("V_sqrt" = Diagonal(n = n, sqrt.mu1mu),
- "D_V_sqrt_mu" = (constant*(mu1.q*(mu^(p-1))*p) -
- constant*(((1-mu)^(q-1))*mu.p*q))/(2*sqrt.mu1mu))
- }
- return(output)
+ assert_that(all(mu > 0)) # The observed value can be 0 and 1, but not the expected value
+ assert_that(all(mu < 1))
+ assert_that(length(power) == 2)
+ assert_that(all(Ntrial > 0))
+ constant <- (1/Ntrial)
+ p <- power[1]
+ q <- power[2]
+ mu.p <- mu^p
+ mu1.q <- (1 - mu)^q
+ mu.p.mu.q <- mu.p * mu1.q
+ mu1mu <- mu.p.mu.q * constant
+ sqrt.mu1mu <- sqrt(mu1mu)
+ n <- length(mu)
+ if (inverse == TRUE & derivative_power == TRUE & derivative_mu == FALSE) {
+ denominator <- (2 * (mu1mu^1.5) * Ntrial)
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu), D_V_inv_sqrt_p = Diagonal(n = n, -(mu.p.mu.q * log(mu))/denominator),
+ D_V_inv_sqrt_q = Diagonal(n = n, -mu.p.mu.q * log(1 - mu)/denominator))
+ }
+ if (inverse == TRUE & derivative_power == FALSE & derivative_mu == FALSE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu))
+ }
+ if (inverse == FALSE & derivative_power == TRUE & derivative_mu == FALSE) {
+ denominator <- 2 * sqrt.mu1mu * Ntrial
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu), D_V_sqrt_p = Diagonal(n = n, (mu.p.mu.q * log(mu))/denominator),
+ D_V_sqrt_q = Diagonal(n = n, (mu.p.mu.q * log(1 - mu))/denominator))
+ }
+ if (inverse == FALSE & derivative_power == FALSE & derivative_mu == FALSE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu))
+ }
+ if (inverse == TRUE & derivative_power == TRUE & derivative_mu == TRUE) {
+ denominator <- (2 * (mu1mu^1.5) * Ntrial)
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu), D_V_inv_sqrt_p = Diagonal(n = n, -(mu.p.mu.q * log(mu))/denominator),
+ D_V_inv_sqrt_q = Diagonal(n = n, -mu.p.mu.q * log(1 - mu)/denominator), D_V_inv_sqrt_mu = -(constant * (mu1.q *
+ (mu^(p - 1)) * p) - constant * (((1 - mu)^(q - 1)) * mu.p * q))/(2 * (mu1mu^1.5)))
+ }
+ if (inverse == TRUE & derivative_power == FALSE & derivative_mu == TRUE) {
+ output <- list(V_inv_sqrt = Diagonal(n = n, 1/sqrt.mu1mu), D_V_inv_sqrt_mu = -(constant * (mu1.q * (mu^(p - 1)) *
+ p) - constant * (((1 - mu)^(q - 1)) * mu.p * q))/(2 * (mu1mu^1.5)))
+ }
+ if (inverse == FALSE & derivative_power == TRUE & derivative_mu == TRUE) {
+ denominator1 <- 2 * sqrt.mu1mu
+ denominator2 <- denominator1 * Ntrial
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu), D_V_sqrt_p = Diagonal(n = n, (mu.p.mu.q * log(mu))/denominator2),
+ D_V_sqrt_q = Diagonal(n = n, (mu.p.mu.q * log(1 - mu))/denominator2), D_V_sqrt_mu = (constant * (mu1.q * (mu^(p -
+ 1)) * p) - constant * (((1 - mu)^(q - 1)) * mu.p * q))/denominator1)
+ }
+ if (inverse == FALSE & derivative_power == FALSE & derivative_mu == TRUE) {
+ output <- list(V_sqrt = Diagonal(n = n, sqrt.mu1mu), D_V_sqrt_mu = (constant * (mu1.q * (mu^(p - 1)) * p) - constant *
+ (((1 - mu)^(q - 1)) * mu.p * q))/(2 * sqrt.mu1mu))
+ }
+ return(output)
}
diff --git a/R/mc_vcov.R b/R/mc_vcov.R
index ae3c6c8..f3ea538 100644
--- a/R/mc_vcov.R
+++ b/R/mc_vcov.R
@@ -7,8 +7,8 @@
#' @export
vcov.mcglm <- function(object) {
- cod <- coef(object)$Parameters
- colnames(object$vcov) <- cod
- rownames(object$vcov) <- cod
- return(object$vcov)
-}
+ cod <- coef(object)$Parameters
+ colnames(object$vcov) <- cod
+ rownames(object$vcov) <- cod
+ return(object$vcov)
+}
diff --git a/R/mcglm.R b/R/mcglm.R
index a5d6e7f..8af8530 100644
--- a/R/mcglm.R
+++ b/R/mcglm.R
@@ -23,45 +23,84 @@
#' @param control_initial A list of initial values for the fitting algorithm. See details below.
#' @param control_algorithm A list of arguments to be passed for the fitting algorithm. See
#' \code{\link[mcglm]{fit_mcglm}} for details.
-#' @return mcglm returns an object of class "mcglm".
+#' @param contrasts Extra arguments to passed to \code{\link[stats]{model.matrix}}.
+#' @param data A dta frame.
+#' @return mcglm returns an object of class 'mcglm'.
#' @export
-mcglm <- function(linear_pred, matrix_pred,
- link, variance, covariance,
- offset, Ntrial, power_fixed,
- data,
- control_initial,
- control_algorithm = list()) {
- con <- list("correct" = TRUE, "max_iter" = 20, "tol" = 1e-03,
- "method" = "chaser", "tunning" = 1, "verbose" = TRUE)
- con[(namc <- names(control_algorithm))] <- control_algorithm
- list_X <- lapply(linear_pred, model.matrix, data = data)
- list_model_frame <- lapply(linear_pred, model.frame, data = data)
- list_Y <- lapply(list_model_frame, model.response)
- y_vec <- as.numeric(do.call(c, list_Y))
- sparse <- lapply(matrix_pred, function(x) {
- if(class(x) == "dgeMatrix"){FALSE} else TRUE})
- model_fit <- fit_mcglm(list_initial = control_initial,
- list_link = link, list_variance = variance, list_covariance = covariance,
- list_X = list_X, list_Z = matrix_pred, list_offset = offset,
- list_Ntrial = Ntrial, list_power_fixed = power_fixed,
- list_sparse = sparse, y_vec = y_vec,
- correct = con$correct,
- max_iter = con$max_iter,
- tol = con$tol,
- method = con$method,
- tunning = con$tunning,
- verbose = con$verbose)
- model_fit$beta_names <- lapply(list_X, colnames)
- model_fit$power_fixed <- power_fixed
- model_fit$list_initial <- list_initial
- model_fit$n_obs <- dim(data)[1]
- model_fit$link <- link
- model_fit$variance <- variance
- model_fit$covariance <- covariance
- model_fit$linear_pred <- linear_pred
- model_fit$con <- con
- model_fit$observed <- Matrix(y_vec, ncol = length(list_Y), nrow = dim(data)[1])
- model_fit$list_X <- list_X
- class(model_fit) <- "mcglm"
- return(model_fit)
+mcglm <- function(linear_pred, matrix_pred, link, variance, covariance, offset, Ntrial,
+ power_fixed, data, control_initial = "automatic",
+ contrasts = NULL, control_algorithm = list()) {
+ n_resp <- length(linear_pred)
+ linear_pred <- as.list(linear_pred)
+ matrix_pred <- as.list(matrix_pred)
+ if (missing(link)) {
+ link = rep("identity", n_resp)
+ }
+ if (missing(variance)) {
+ variance = rep("constant", n_resp)
+ }
+ if (missing(covariance)) {
+ covariance = rep("identity", n_resp)
+ }
+ if (missing(offset)) {
+ offset = rep(list(NULL), n_resp)
+ }
+ if (missing(Ntrial)) {
+ Ntrial = rep(list(rep(1, dim(data)[1])),n_resp)
+ }
+ if (missing(power_fixed)) {
+ power_fixed <- rep(TRUE, n_resp)
+ }
+ if (missing(contrasts)) {
+ constrasts = NULL
+ }
+ link <- as.list(link)
+ variance <- as.list(variance)
+ covariance <- as.list(covariance)
+ offset <- as.list(offset)
+ Ntrial <- as.list(Ntrial)
+ power_fixed = as.list(power_fixed)
+ if (class(control_initial) != "list") {
+ control_initial <- mc_initial_values(linear_pred = linear_pred, matrix_pred = matrix_pred, link = link, variance = variance,
+ covariance = covariance, offset = offset, Ntrial = Ntrial, contrasts = contrasts, data = data)
+ cat("Automatic initial values selected.")
+ }
+ con <- list(correct = TRUE, max_iter = 20, tol = 1e-04, method = "chaser", tunning = 1, verbose = FALSE)
+ con[(namc <- names(control_algorithm))] <- control_algorithm
+ if (!is.null(contrasts)) {
+ list_X <- list()
+ for (i in 1:n_resp) {
+ list_X[[i]] <- model.matrix(linear_pred[[i]], contrasts = contrasts[[i]], data = data)
+ }
+ } else {
+ list_X <- lapply(linear_pred, model.matrix, data = data)
+ }
+
+ list_model_frame <- lapply(linear_pred, model.frame, data = data)
+ list_Y <- lapply(list_model_frame, model.response)
+ y_vec <- as.numeric(do.call(c, list_Y))
+ sparse <- lapply(matrix_pred, function(x) {
+ if (class(x) == "dgeMatrix") {
+ FALSE
+ } else TRUE
+ })
+ model_fit <- try(fit_mcglm(list_initial = control_initial, list_link = link, list_variance = variance, list_covariance = covariance,
+ list_X = list_X, list_Z = matrix_pred, list_offset = offset, list_Ntrial = Ntrial, list_power_fixed = power_fixed,
+ list_sparse = sparse, y_vec = y_vec, correct = con$correct, max_iter = con$max_iter, tol = con$tol, method = con$method,
+ tunning = con$tunning, verbose = con$verbose))
+ if (class(model_fit) != "try-error") {
+ model_fit$beta_names <- lapply(list_X, colnames)
+ model_fit$power_fixed <- power_fixed
+ model_fit$list_initial <- control_initial
+ model_fit$n_obs <- dim(data)[1]
+ model_fit$link <- link
+ model_fit$variance <- variance
+ model_fit$covariance <- covariance
+ model_fit$linear_pred <- linear_pred
+ model_fit$con <- con
+ model_fit$observed <- Matrix(y_vec, ncol = length(list_Y), nrow = dim(data)[1])
+ model_fit$list_X <- list_X
+ class(model_fit) <- "mcglm"
+ }
+ return(model_fit)
}
diff --git a/data/data.rda b/data/data.rda
new file mode 100644
index 0000000000000000000000000000000000000000..45849863b8a4166508f3bdd6a41d2467afdbc2a8
GIT binary patch
literal 25969
zcmZ>Y%CIzaj8qGbY}IJ2VDkL@pCEW}uD*tWfqj<g0tVArx7Zp?Zi|>OaPhB=x$V`!
z00$Zj3=9{*yx3V_3dG^LmUSy2BLU0@5s`2W+)j(Pm8RV~<GhVQk%5OppFwD05Cc;K
z1ET{7F)%#<DTiU<h6y01gQ#O=;|d1mgbSh|b9mS^mo1b^65PfVq+(;1z_cDD$RHlU
zz`zPp&A{j&0%b8UfQ$wUf=vgxm4U(Hy?jIi!vqmV_Y<cVWH2x~UeK3iV9rUKae<+L
z;YPp$h7DE>4D1ZD3=Ba_Rc?7cF}eN!ua?)cuc7u|1Ai@B{?b1n|Eja#@)f)*q!<_&
z7cNN3U{+l0*1*8*VC(F_Ah&?6;RHkG0)`80GiLEL*gF6J|35S|)JZio`1Vz=y1)N_
zd94fyH2LdfcJFKORnN)7!itL(k{B*Kn3$Y6$T-2!#Gr?Bfy69k&W;Nv%nl4Dyv%CO
zoC_xlJG(XenlQULToz<f<a}w?aa2lfrG3zhdrLZBRR^VfWuKCA;VZ*e69#5CUn%of
z3o>#APFyfy>koWoWb61!Zh@_{gUPW4UwKTLg?-r<@-8{qVe9Y6vHZ#3kR_c0kPu=6
zha&?+280VCr5G15FmT!i1$j*MorH)`h$Je>$mJ;z%u(?*_`;IEZi|EEuJTO2lJP$z
z%b?ffP};Ityu3zIk|rQ6APYdoGDtGmIxsLu@v_zbTDAPL{hN+ood(9n#$v|83QQB4
z4uyB#xqIgHju$V>lA|@kR&15k6?=A3v|Cg<FUs6(i`&JGt3`M34!2zOGU8hA&M0Na
zD9es_kF&UT@1FJ4@4=k1JL`&z!+j;r?T(pVZnC@F<E8bAj+04-xo0=#o=tjoW9y+K
zdG9pLOwS!Io_X__xz?TIdP_ZhW^Puo+?Dt4_3Dn6*t;@26>jNRo-3W46L6Bz&`&1s
zZRy<$cXqkFyX(p*u4nYzbmz&gcQ4Oblv}-fc=x(&VY%P#-E($Zp59sJ9$TJNddbPm
zCwlkpot!2YJhqn<&w6h1ZhGv&^1K@>@9v0s(b@6N>QPa;_3lOG*<047mlv-qk1e?~
z!T03dJ8vGndc8e*>D|+-C(qq^$Y^@G^W1hP**j;7b_l%-FDs9E?)Sj6^yb{A$-4_L
z7SEYjUfjibZa437J?q_b@0LA#Fz0#R+-VE57jhg_G74nZvSXd(+2A4Oozm+qZm77R
zL$)Q@K`7PHEbB`1A};5(7cyoqp5`jZeYMT<>4D3k?mC;ECC*hlcVQC8o}ESs2U%xj
zZ&YsCcI;fj5hlJ{yO|tS1xzle%rFZu`B>TFtk9(AbajsOt<ORnVoJh}Emy?!q666i
zHU&8<^ewsXRrBthQSjZ>A8ha1mE9}ZcedzbbL{l#S!EhcoCblaOdcAG#aC|i)pWd;
zxNeWZO0Au@9iw{PVm_OeW*og%#48b==b^PKMat0OAg^*)vDktWlX(|bZ_L(~HVbZX
zdZV-~ic^@4<Hmw3cR$WjXFJ!y^D^d4OQf6ZG?qoK9UPnVt2ZTSotkrH(}}dS-e)XZ
zSw$E(X*%dS=bSmMv1p+%OK<neHb(=kfHtq)Nq$GKc_kcF`mN^2<S<jTvogX|?WD?K
zvm8qUPKB91N0%6RupUU=65t{j9%ZB&e7Ys&m|ql&w32{H!S|^qYaT_I`8VBLcX;Qt
zBG-eNZfA@qn@1YFzS=7#uFn#Cr->uAjU)GD=Bue*Ez%*kf~)Q=+o7Vfn8h$M^Ks>$
zN^v8Og~gRdw;b9$T~{5uwQZe{g_s4~(g{8Sr#PhchBPPZHoaT9&{Mfsc+QkZzIP8Q
zwCAi`>eR8VX#MOPM^mLvl>6LXwSS^QR^al?X=YKSCv#45^+%lRXWw|~dEu{#-#MlP
zxC%_LpXqBR)#9Y86MN^Ka_sqkb$5SQJ^kT&xK8n1_rX&<Q<<l<{V+e2b$Z>Ui(i*L
zdbq+-#j~S@d$qx`l`rP4;Mu)C?8>L+X<KBzoz`jUd%+u}#y2N>>PopYw>_sCO3pac
zd-g=P^)CH&=U(3}ig{o5IwEz=)2y|N4ykR5+jDv~&-1nXTD?nVDitjWX_|Xsncz;7
zJ0hL|2}i|px9t6XKkvt`_)M7z`_lKmeEaQZ;k>6G4hE<$(rLARmgKT1Z=KpISDoo8
zotZ&fml$*@oqUwd5#*?1<nn0q((1^?Q#llJ9&K6UbL&RY%ec9MJ2z|-ajcejkR6tH
za092!1#V5R>n3N<Owc%RG^BU??sM`yQFn~jNv7S}`#Uu36c5i<6{8ag6P)xuuPr$1
zdG65l*KZrF`~57$6v_(ol%rS3oZ(dIvUQWp(At)zx$U;|DYLnIbFPQwmUwOn4R)24
zRA>}R*?DNDOUz<V;UdX4mq3obGy{h;L(d7JW*oXgorYJ6yOKKGZoE!*o#=MtRN|Ti
zGrM#|Zd4eEyiG~VRN@Lu6Hu7etnA9f_UFPz6~U!0nP*N#rTR_jRG6UgEGXD0@5ZVH
zuOmuwjorm}^1U$HR66f??zFxurAB6QtL}@I@|&AI%nNxMI!!n@b%CSI(i<L|lFC^h
zWGBDglW<%}>AkF9l7|b+1lL#=j)sF5^}L&s_)3oXtlp&{w%k1DB1h*GT`fTuCWkad
z$CHgZ+KKl%zxSS>AGi0??`t2q-`?G`E5A_V$Z8+!oCj%FKKM=8A1-~;HGk_ex8p{o
zQ5uqa*_*v3Bn}Dg>`B%Md~#@UV_xs9+hK}x&aKV8^X-dz>}7TL)!i4ReWl$sctb+N
zXFvLJU+l-7=NjGb-yId~KFHm&adC4rGgr}~wPhCfE}e6AS7MtxV`cfGmD4U6Jnfo2
zbIp}Eg-fHOgQm&&nZ|5X`t-iY);)8Ij@W}F+4dEU2O7>W<%M35;kYqj(T~^9^S%d`
z>^r;f-ka(7_Ux%%zfm}g&)mGC;*yN7kmp5r)sG&Fbc!lueB1n@q)EL6Pa-L^H>
zYnLf2=UAlN=-JvQZ0aGz+2FDDLP}1#A+wswU5#U!eKNrhnv|XyOr7?I(}v^Ll&QO_
zRkBSJRbN}o*xP2<x302Oa=Z0%vxKKQB}G@)XwKwXw(Eh=b>a51j$D>Q%%02#G<Ggr
zqEj}zJ4br0_nTQy>vnbvZhCNtMTo&hz$M~cLQY<33g^zEmU)~-J96@h^4xkY=Wbb<
z)3f(ucI38MhZ}k#75KOqjHM@r1(#f!#4cpIHSNda@NL<=j~n!p7amrKIdD>3<)DG?
z6qRO;nTBj3K5Io49^UvM9el1yFiCEj$Hq<ruch40i8D63xC(80yqfRIrG;DGZ#nks
zFwcVyHIFIFThlT=7;*?OD_vbKVa;yvBWHp2q<}>hdq4FY=Wa1bGP|^q?PcT*X{p7m
zLM?LzlazYD8y(A86BJ?QU8EL0N$iA$$|TNP6E5y@PRww)>9C4JiDd?7vJji4VsewY
ziI9*RulI@B#yVy*6J}M-+`87v!c6I;>oMux9u1|X&MHzZo0>dbPbqo?cs6M~d30)}
zjZz-R*NHkSHzz6;NwHL2DDF$KS(moz#B;Wy1K~Vd)!YO;SXlj694KnIsUYPdJ<;mP
ziz8DU*-moLn7yz$If+B1SHR$kCx<{&E1Uk&xT>vYds$pgsj~#`oe-kh@HS%N<RVXn
ziE2#Z1{RO?CN6ImSjv^PZSGaB8;f%_k|#=Ex$m_^vOp{#>~v46D-#>HK8I2&1LvzV
zC)Io=ADH&`ror{bIo+GBd~V*_Y9!&%mBY#9kUJ|jPv__YAxE)k5}A&gE3ZhZyL^e+
znUI#YGwoQ<t<vDA$*M+2Zn_F)v#_eH*%~H#sCB8s@-n3(m$-V4T-j^6t#|2(ok<do
zGn;Z}OmAKsRrBlDx9{gPznATsCwX_L>|NH6WvZ32%7P2uJ>`4#{rvpapP!%Ky{GY@
z;q>kT;a55B9E%dpn4L(Iay+Q^ZC!^EU!R+~?Pk}2gEJG`Zf?oE<#1;)-}~+7)|(%n
zTCjfE%dJr!9G5OExWsT_L4ZQDm$pa1QU+BeMu)Y`4i}@ERvE5|&Rv{k+RY@wHDy6S
zufmq4+z!SWSCa!ndvn=VOC7t~8`ib0J1{Wm*wu|&a|O2@6<nKiwL7afIbyZcu@zcZ
zqNM}3>23*IbhUSJ#I=o!-Ht7~xUF}^ww|P`y{k;KbXSWeheWLH4T&<l7-lM+W$L(M
zZExV(=BRGhz-*~wi;i|(?YP<<?RL$qJKHs*Ytc5--jJjfYwg~xiw?;R+^)NGRqvt|
zw>GX8+bXtVaY#;dl&;jG8;dRmMQ3cu33XMC(sk>Vx^e7cmaf~zkfg;gS8Z38mg;r8
zWxC+Tny8B_ZU;u(6waA-G}5fs&D3q1cEsXn(WH%Ar8;hhBx!UidoH??)nzIwW}0x#
zw_8*?T1w+uw^-9H-R6vqt8{WQZVPr?4cwB=x^b=8jqaq4i`|Y2W+_ML8fmOe&Y9LL
zcI@b4*PM)NhATRRU8T);ZN0gQSM5UJqOR|Sxht-Dr`$MSS{%4aw>d-CZG)?)l<B50
zvqr^f4JIB^hTU)0MdyTZyLOpri}t!*$#q+k>$+$|?(}od7K?o@P|h-y(tTPip<%fF
z@~n-k7Drsu?agI%P0q=6GYP-5LsV)@w5RA|SE**#jf-7!M0z^ROr&%JBW?!<xaQiu
zd+M6B;+DrXlce5m!KAHfCmlECN^xkK{fap#7?^3+D`jTdEunE^QIp`e*Tp`<DOa+z
zE=H^t+mvxl<050=rtHA2q7t`VC9WhLX-#%pvEjgatzA8nu6PROIBxLSxJlxSmbT#h
zyz9bJUdOa0I4)$kw1s*8$lkayTJ2<%mZ?j&N^a7{u44y7LT22!l4Tf}a+@V^i<pv6
zP?mCj_nXh2+q|VYCmHs5CTtXSOiNBWkkYY<MdG5TphLnkHI-#<#}rPQ*nWNYHF>sc
zz^utB7n6j#I5ie<YOowNnr=1arpM9Zc+r<@l5-br+}gEB+3kLD)UzwOu1%L%CDyXC
zN=qjl3|!KaV!)=sEwP{{;@D!LsGf+$YC3A%Y#NIiE@m86<`(mHITMvs8MDIR!~|xS
zXqCv1H@Z9)DJ|+r_VLira${V+d-C0_rzJF6gk5`GW~*IhnXIU0(spd4+a>{trcUKx
zgM^8i9?l1EZ}Jq}6q0mHJH#c*B&j#yV#YO7X~9hyR|FSlni)RZIXQ5Zj+n<q7uO~&
z&)FVR1wA8l+%86Hn4Z0PdTGjWvpXA;S<XH;GYMQWIVWRdOWKMqlce4VUSZ9dyi#1j
z2}g1|1tuJ;Ezb#*&=u_Pl8Bg`()H=3C?^~Hxogifx{VDkBymZ758h?swou8mIVgYs
z4jnc_l?ft&lNBYrd_q_hYyXy+gz+i}ZelY?=+p}Ga5>4y*2OwOM50A-lB+69NBZ^X
z_1p6|pZxN|`|R#o_2kz%x5|3mvSkzd_4=aB+Ou{2Eds7IYblHMibR-9I~uqpq1Sao
zT<F=2y_<EVdR=C7ibOfx<W+S$*1KShqg$$r#8Jl#Nm15bMdPFyxr=!=Wh~<I=+d0g
zE#}oN<t!kkEp2o%*oZTk$8Zs&N4MRs?WR6mno1`pd9!gFY9+X?ax|HevS7wC!zN~i
zMY${?fot+N=N<gKF5{x4m%;*1mqiy&^ho)%3ZzAe=!68Zm%FG6vWR)OOlHwZOZ`?H
z_w#;BO|D4JY(^0=NnYM1UWQ3#UM!atOb$(&>6o%u#8sL{Wx)c624&4QVFRf|wj}|V
zIc5s7{QO?NO=Xj&+p(EaLLHkN7Yil!nk<;O#N{k=fJ9otML_}2Nut7<6Q(Tq^=|jx
z`?jA%&m>LX{p9`JJ^QMc+rEA@@9no=FYfz=pDW(++<JG^r3WgFhOWYnMxCz4EYp^D
zDp+tV4A<Vhd-vYk=k>zl-nlK>c0_GU#-ZeF#ZXblZQOwwTO2ct+*T<1WSq>{WVE1Z
z!9~rQShGdkfdWZhQ+gSbI#!j3o7S;C37j0vJLCGfmy1j!TOv%2*=x%_C@nEN)al78
z;&8T0%|mtPX`fp@N~+SX9S2=pTtzfmRSi~L$Wde1)WfCJ(J9E=#3Lns@B8+fdsZCl
zZMMA~b6aJ?tRopsCYSeKeKturQJEv6C!(XrvnwGmiJM_kt6I;DO^&J)7o7>5AM^aY
z*yr6_l6qVhY221<OiPxyu~1^NMrX4@%Ec**jYQ9BI6E?&Rxn`clGre%Q{-5uQpbfU
z<)u<C=khMUJTS${S5UKMQkJq;l!s@lV#^dBCylrg@5k!1Pu>;uKcu&PqAMG(6jyI5
z?~;pNDNP1xjLAuws+?LXf?67BY93RZQo1xW7JF_ualz3kD8-r0DD#Ka@A>k=&TOVF
z87COhjC7o4dU9~;{Vwl)x6^ROLYEn-GmYAo1WV3)`*AOiSHGSj$61X9sx_~Rnhcj5
zjJ>@`bCN_a&y-^cikmfsOuATZ8gX<yXPM=~vt_ABPO?{0z`~?}Oc6IpMFXdnMkh~Y
z!NqQ7LM9uHS$6M;`(5+9=GXoB{A%@GzrMS?doK6h=H<Tfs`6d$my1tccd_Q(`bE!+
zdY?TlufBcvT^!%b>+@`HSu7RFI>F{*Bq-3~A~sQ|d)CY)Jl*-av(?s}RGqH7Ti<u*
zwRNGJRc2a!E}5&>edqP=z53ji<@;jA-o2}~GgqJ6etv!UzVgeLOsif`Jfh@isJ2u%
zBFw16VL_*af{<v_>?!KkYft^QvzqtryLw#3ch|+C$;qx8x0&^BNx77@sM~F0Ca-H(
zZ<=e;qTl9b-h$mpIzH>W?ldfE_GFb*I+$eIBQ17dn$krFm&Qp7KUTM7X?S#tDfKdW
zI;PGvGt<=+6}B^XRSLS~BEig}rll&>vFOaxGZ`EQm)YJ~kTxUPaiPkB8%o-DYpZK&
z%ft5RxM>->_4alg>s6T$#+%e}vg>3=i%FzJXT0w&7r_l8t}5S)g}eop%vM=p%Bjhe
zG2d$Mm!F@X-g|pC<>W>;4W-GPDpPAMUBz59PHxr}aboMz)Ho<DWpb$}pwnbZisY1z
znTtHt47=0>X1EA7tEEghG36pp(<H;Trc|Sx4kbe)x3-P|rzzY*#~4l@UZgT#rg|Os
z%lC8dTRk^ZQb<zV*zJ;`&2~srV&=?dqo9dxUMiDR4Wwo7F1j&E$t&1nsllWag-*B6
zFDs2wBo({@H_dR|$S5*RMBpTkkEgM7Wucdou!crjyj{)v&UnE^y;4lgo*Pc_95b9O
zrEzG2NylMh7q5T>hh-X@+yZ?JXSi{6C3~qkGYK_y1qx}*oW|8~LG75w5w~L#mO1^A
z{rtRUgG-z8^6QnKrajyA=FH68>E^{nS~;y}-XA$J({5Mw8Ucx$W<5F5iV@wF2fE!1
zrL<H|C^Mx=98_u(^6YWzIWtL2a={T+&A<R-wk8z@ozC8Or6o=$8{FE2Ig-kY1vw;@
zmAnKlni;i83Ql%4Xqj?QXzz9I<7J10a=(9%dpC1gBUj{v%R(tGzD837W=s{Ba9}}?
zM;pV1MT+s(Q%<BPrOYeo_}n>VW?QCFlCoOnq?E;(YLnEsSruERI5msR+xz0T`T6qv
z-8*}`G*wP|8_Y^lNoX)~Ny#|H)X_5W-LnI36A~3Y9GtmQj``(XRBhG?ob2NhxUylg
z(3DLnPDyS7NwZSe+NSbOPzjo?AmFK~<Z;AP)!+mdr^`tNg-o?xmM)XY0ve1O8#Z%k
zCTAK5D(~KMGN4VQB~dM<L!z@`q98X{TWg0iW16wCk0YC!j>Ih`pJa}ngOUp;dHG1Q
zHTkHrE|}8DtKhI?L8F#hG^>J=Ye$Zfqe!&Z(pOWaaH`x`bV|%@CdUqqLz5a@jrVNM
zjaxG<XLIXr-IvvG&y~L`vO0H}+n`x9vCo$&RmGseg~{>i(Hk?GqSoitPJO=Y=bdnq
zI{w&Ny~@Lv>?|w3$5zjIuWvr}<mc7%te@|+y&Jh&SaNSf$4PbLe9x2D&9>bvGxI(B
zNwrTd_UnDQm)DQpRKN5`^XuJ|ncYQSPhR_+d&lLp`Ra>BdXJuc+qvOlvs=sQ3^Bbq
zCv&Fs%J}RJ5)bX{7T;%m>)r3)d!JTGq>0aYcem!fz9);Is-nh*4iB>hr&6ax^N26%
zGzn0cWfJLYz>&+w<GH!ZBrRxCfTQpAbGKjLt+8Ib%eH*`x}A6SSyevBUGUzoL)g`8
zQhwJY4ihdZMWZz{8-$%6obr9zy>m}Tc#xaV<o#?C3cZ1q*W|o++)ACmQ_Ug}&cR^H
z<jD4vTT)@dB>~^t35Oge^h$W^eOj4yu0o(=mP?>;Wy`^k4_O&)0jevFT&Eg|bFaR-
z)S-CYyYig9YO%3aCvWc#R{EMUXWG4YSM&XTt>S1FND+xzHuZF(l!ZRG*tFhB2bly7
zu34ySiU^5iE-?x==wuY={m@W0H}!%@(mE%ewg81mT^AW=<T|U(=5z4XniAP6Et1$g
zZI1TJnR7$BocWGUV2}N__wMWN)BE0i`rMkO)U{~K)MU4=MZAtHR$R%NEavMfxMAXo
zjx>{Gw?!8>25vZTU+PJi&byZ^Z5Q1P7fxBY$jMOTz{M0J7X>yQiHl1NT}~-0ii#Oz
z=mc)#RN8wt;f#S#Pnv&us;iREB$tKDrU;o#-@S37-n%W+n3LTWoz?Q_nQ=tfKxCHT
zM$Ji`I)NQNIWw0SDY>}nn_lW&Van2!vLzs5Wy>U^NeoApatc@P**<Z>0!3~Cw<8%R
zITr{pt0eJwN_x4?xUBT)g_0V#Ou5^}RVEo*k_>y9#JVN##`r8$n{HvSAk!#0Wr-K3
zhp*6-09BSIlMd(Nk_#!DRC<_p?noAy)V7r4n3+l28IHpuk&}(*9PCmST$HLZOH-&t
zAycE7VdpN7nV>?u*mIR(O2bh%wFsrO9u|SLR;P@UEoY8-r6x-#dDvMvMJ$t4^SR`!
zn3UPpv0}yNm2Lvci6$L0&6+f4PIFWX=!t9;6k5zH#OiQ~=aiDO%lGxxd!~6v1|$SE
zrB37&-}NZrjI&dqq(Z8iOGb+16dxr?mm?>9jRG#rQdT%LiBn~wTf-#*FD3&6wv?GY
zlASyXEJ-IPICYt<+k5%k&(C|`-naRqefjnJNuRCW?c00iyH)h5pSE{vX3suU@viUZ
z>%GT(%41z6m1~N6JZ4O2F}aj%WXRCDBwl)=+mSRkgDimsjv6Ntm5e%DR7{kNCKv5;
zS?u?I@{KkF!5PaKT$LP6d_)w5Qap`ZCkZg}q<9>OXlx0bl5u(WntkPN5ggsB4326B
z6V*IKnpE0c)L1xzRTX+gRK%PD!gt4ZH1wQt^4z^sMWQDtgV{s1Nx>^p!eO$+r5u-;
z8x~7keD|x$uJH5q>R0dQ?PuS9$xBy_WwM0Pq)AFanks_A5<V##CiZr2WOF&z=DE;F
zO(?<D=SYf&r|(Q*B}0Q?mK2A{20oqKVjhPwu1{A=ljOOm5@fyG!(gGyfut;rnVUE!
zi(EJ+Eg<M-tgwiS<$|!}<cy{urbz+^COJrPDtTPo)FCisvX+3V(9A9_70JaOuBy^f
zU5SC79)cdCf)dWIGfcl|U*EImkG$oZ^PBIzU-z!4>igdF+xKx8RPHm6kKLCrRY#d+
z!W7lAv!Rv>u`Jw9oFb7bIj6Tpr=@+|8<}!C=j^mJB_rdceDmVz+UBZHPG6mQb@q1Q
z?6sSXo-EwF?(^b^MK`!qlsGP=MI~yABnh09G@QApS7qYzhwCm0xda>(-<{_W5arP(
zEcC04u~}lG!5Jq{o(rA~tU>}uy{BGSaLkjXSCeHbOIk$3nHd}*7abK9J(spzKQ7TC
zso3-D*Nf{jRc5kHQjp*{GfB``$(6%fSXHfcaiWm3ucXRCjwFp^?{@Ec<Fka9MY6}u
zSGSQ-Ehj;eS#kl(OeTk<78Wg!Wj<#PPdEF$`|aoXmTm$kxujIh45wU7G?=*Hl9y9s
zlVO`#rpJ_NoGq+^9ga#;N*rA-oi0t26g{GRJiJer@c43M&T`_I6fD@(VWuPE*7@%B
z=~Y1!E^y2+tKP-w=F5>~%xRLwlPc25Fi|L&>BN)(H<eC<&RG-LI2NS|IH_f}sqh4a
zaCFQ%rWR~$=qRY*rYPVs(?Q~5i;0?oF3V|=9K}i6FL&>~Z@2q?%)5QI@1CB%_x9=S
zd*3@RykTkcxmas4gGEV_qs=Fhx6v!;hyllhJ?myAI!mmalsI#uE32x)B%Vu>X+g6@
zIyfATrF1F<Y6>ZPYnV!DWgM85nAXwIqcqF%pjeZGqCs;oi;F{2%AySlNll516g=Cy
zbUIoL4l)Tj_`F-^<&`LH;%uU7loBZEkm$iCp~yCcb7qT9j+Bx&L#nZ|m(y%xm#Lgm
zU50^8i#iR0etzD4UoL1e8=D{}$B95UL7hXUY{71aHEvE}QZqJVI?1!7O^xTYTV*)g
z37?aS23-l#BGN(=0wOkZE=$}Jz$>h_D3VE2O;VI$29JQFv6Qo?XUCD0quuvjUy>9Q
zmhfabsgvd;<hUUqQLR;q=Wxm_p=60plWC?LQJRxnW@#KsSSg}&WFt>+$H5LmR!$=g
zUMY@a!D1p861b&wGS2j9PME|w;gYgb1gFUfVa0C6fFQvoD`zOM1xlP;oG|5DlaiOx
zCdWk#SA;T5g**ZmwskGwnZ>b6Lzr=yOT-N!pM(ZeX))(ywKFGOW-BGDxaAdoet!S1
z+|RFf)sN5PEBjgd@Ach}Rxjh`{QABr@1k0MbVl0u;%f8FEA>)l8r|M`>(wKMDJ^HZ
zv*s$S;Zi&HYRQC-)VQdvoqWXw&pCGL>ugF>-7Gb;(`?(<H2v_U=We~c`m)^IEa%Sb
zxo38Y-by_ay}fGFn`zH8({i3}N|#<)Z0>EGCcgU0Uca5@mX@<m?k<YFa(dH=>0z&9
zUxp>BEjxR0Ntd{JiIx{5)74pOy-G6`tySlwZZGrMX<U?ZbNZ@9>S3kHX^iH}-iGmt
z2rLoU_VR#1YUJI9Dc+vkXPOT3I=V7k))CiBILLa~%ONzd(B!PiJ?<i|biD#6m89nC
z&>vZgzAW$P&Smu1(rCI{b9T`#twW|MS%;Tx=1q%K59WQDU*Gznd>-$%xVO$qC(dp+
zKbV)v60^uR&@bmsSFnP&!@;vtLb`mX=lK{^EEP)9ILOQ9lATrPCU|PgB;je1ir33c
z&sT5C?=C-iXWqNG?{RV8_q<skF{xvb$byy@%~c)~5)KMYFlglRdwJ*QjhE$nUtBNS
zz2oKQweMa(dXlm?I%aR~-WZ>$J1w`w9y@b(#<nLXjm+)|=T6>kZD_9d_Uv5GYu@E~
z+a{;Unj8#Q3^L;8nsTPii`6VpVf$XI&!tcA>^=AO^SyYByZv(=KVMfnqqpRB?%AU^
z%f1?=WxaXt_H^#{>c~&8ckeB7zWeb*RSmz@^SJo*1CwSQWI36k6QnL4u5|S7CQsEV
zYfd-KbZttQ)F!nxwCBR4FyToan?!OwE}5C#+`IE!Twc+=C2M!?madFE<2l))bm`3M
z+*{AfKC5^eeJ(ezTyl5&wl|Yi=4_ul>B-5x+k<wOm3Zd)oIGbfan7l2$Apq5=}gdO
zz2j??GF`_|QB?786Yo+_FHbYCE`ycVL~l*<U3Y`$c7cLJgUS(y4goU(20@nIRiQH;
zZOhse*R8Qh<<Ml~)jkI&C7oOybnAQ46W6IHr{v{k>hwvS6}PZ4TeIy-@_D!3l5=4_
zC#=$g!@RARzE<hk`*-K2lS=nu&PIyGPFK3KQ{+&G$00^RpH)*-PYd~6c4PH$*>=f8
z)z`zJQMDylExI#i2FFECPoIN|3Ja#VdI?<!)Kv2HTqfLFwbE0vg;OEcA;m<Ajnh
zVYTK~<CNt>>zbyzNOEqGPHAIe)s|SqEF!`la%Yv9Ucg2Z*Y#emy-`aKhtHhiD&h1^
ze7e)xtm?BJGee{p8crvyR=BCvbCT72iqj-bk9+xE8xw9T8@-C1+8=8d)6mG^!ocI=
zB4GaXG;_p3-TZwTvsUK0?$}Zxx$?oe4Bn6Bi?_LKa4XbL4UEj2q@w9>Y`Neilh1_S
zQ%0U%5xk2oD=xY0+vMDIkWXE*xTt)(V!3|M`_z*&He377oK%>0^!DuV#O}1{&E{$A
z)HEj-r|i94TxmXgdhVLJJ~v{|mK&a#cCTpDRo~m2zHOe{IXU-v_NFtlw--IhRu8sb
zd0I_!X=T!(GZ$PIH5GWLuuag|DeSj?>peYPZ63qQ7;Ps8Cev9K=dQ?QDhO_LcM7<F
zHAiY;Y47rs&dJPfqN-Kp8Ct#}YGo!5PB=6q9lPu!<-6Qysh;W4g{fgiLc6q9uCsRF
zH+woq>e^0b<L#`+ljqzK-sG`7B8bgDQ+%qb#7+*sfbtC6;PxVM&L=JpV!Qp`+DVp6
z2bf;Xd?hHr!Yd)bAh<*FVc3mj-|xtUZ|yxRk?Zle#%k;A33Gd9PuLbwT9Z&%9o3jE
zzAtWh&C&I99;9w~T%4ryxrS?Jdzr-HnQN0BoWG|1e0V6}y4Gf6vyDm>6HcFeI8XO+
zq<_11cAgSLBg+v+LEeJ94~p2UOCN>5Z%=cN3y<C(>vrv)mS<W|p$WfiQmVkO4Uc|C
z?s~F0$fWq{#_JDnPkhYceA`2*piYd-H*xWe^=aSPH*aNFccJS(Q`JB3(|2#)$$k3P
zXm5Yhs}EeA@9NfmsOQ|ICC~k~%1%0w^+)w1+v~e8an<;Kt*`vrtE7G{zw6iaYhOC#
zW_<h4fB&8B_M^Y+dtSQmSJxK5-Y>mWzVFBX3BSG_QosIw;&1konc{EEPyShc>ZSA3
z#c$&EKTC<<k5B$Nf9fUnn|9f6_NU!+fAc@%=l(5kR=%tKQoGuH@!NXK1^nA;-k-1C
zp1{AocK`E#4>z#iseAfG-g(pf;$Ql^YOg2Qz5QN)>VKe;{mMVvzs6rZ<Noyg`M>F3
z*D3#c|N8%iKSvsW+P|sa^zU4fy^;Ivy6n$64u9@{n}6$nu9Cgw&--ux@A&imyZznz
z@{WI>f8Rg<zt+Hhzu`ajANCI|{_}s4e^lSC{PzE`|MKtD82&5&nSbG5Tk;o|`l<io
z?`-^+{&WBH`uX-H^%wrT|5cYStH1m|_;0xU>-zQoj^Fe6pZ<6KxBu?Sf6srfzxThc
z{=xtDfBKjF|I{yyU-s8ZzWTrVpZ%Xb{?C8E|Hc1h^{@WAzp4*a{#sxEYyCB!`s?5S
zf2+U!zx;K5?5{r#|M!2B|510${h$5M`se?b^dGR_^mqPS`>+3R|F6Gw|3f|dxBGwU
z?thp6^Y1zH_xRoaU;p3v|MUO*hyMSn=l}4(;(&ni_x*eSet*9Bi~Rn;A4xA7m=^xu
z!QdGDJ^t(W|KGm<_y6)g^lN<ZA3KZ3eIWF|fPsPe!=Bf-_u45xvQvHpT5!a`|G<Iy
zlLMHEL@@sAuRd2k_r2x0k8jVF?|EMZmSg;p_wnsL+ug^vSA2UfclYt_J=Nv?c^~-u
z_q?yXSG*Uj#-0OAArpKc%OCguSC^kJ%2LrNz}&&YyOiZoV8v8El{m$U)<?n}%$*gF
zkNo|Avv}`=TKRqQ4euQo7;5VZ<Xs#e2~@PmDYAD!7|MH|$K82vd4Yk!mVrTD{p|t<
ze(*XtB<q;Letq0$@oVC*`o*CP788VD?wkH=|CN99m8XCF@i+CA-+4Rx?iV$XC0md0
zGHLxjAsF5v7=D>a>m!rapBRqSM>tl0oS?K`KxzHakH=*S&&GXWf74*M-hqMP>+c^(
zc0rkp-}k<^K3BXxuXrv*<d3=az3+P%7#P?e>|kKvXJEeq5l1HL-S2(h`}+2rmmiTu
z`S0(%{HU6L0kRl`UHEf@<0BT)M<Nxie2UJU%pG#JJMY=<Jia~d?t6=Gk8k(qecsFP
z+pf5Qfx&Kn1xSzG69$H#{M`*upMr$oxZeNX_c@Pm&tc%-0vBUosC#?B&H#xIVKV;A
zE8bIG0^vYO#y=QohQWUR`>Jm*pbF$G&w+KwgGs0$jI#Sa=Xu4q=hpYi=Xmf7wF|Ke
zePk%KDV*c=vGABi;Q{&ddkhQ=zy5&D`aKV(3ze?@UVVH!Oz8Ik7z>SNe^!l43Z~Nj
z-*d}zg>TQn__*luz3(lrZ?`;GsC<N9=%eP5LYWwkLYczu<1&TEWeP9ZFfcHF1tpH}
z40bcHn!$d*d~S8|@*`L!5rX{xD+xK752~8+J9J?pl#ihJ&wQWr-un1<%X7B7@9o`v
z&vxfI+ui4E=bcaAZe`aFN*>G~5*Qfx3lLh6S&aYOVO-{~b>G+T{+m31WAU8l)mH-f
zCr)4tUFaCte#)kD&pyBZ$+_EZgR_$S@pHv<V7f@8*<V!t&%3+345Cer!vew>4CDhG
zK{NnvV?2Pe26zBmfUuIekx(K<sHHWFz3r3!%`fNvSMR;|z8r_M?28Y)PfLGOk)2=t
zv;6+s6CapXDg8ObvSFpeiY}2wyt*L^7&PDSi#Is^v^eVg6w}$s+fV*+x2w6@+W9L@
zvt!arCpMK9Pt|27e@*KCDPtWmjnVf&ecIjqZO(6eW^8+PmU}^?iizgc_ba!`uZ)~|
zb-U_fSGl+Cmu~odlToj__j&)@{2$`oyOM2q6wI``L!J~(b#Lw2taYGCOja<YML08I
zw*)^2OHPy83yp-<q}D4NI+sW-m*Kgjs^V$T@mkR;ajU=yXDvZ-jV{A-wIzW~lZ_9@
z?>c7H$9b$vWAP)Wc^p=XOzo+su6h-1oOPwBfBxC%w7XGf-NJGg>uhlpIJfh4K=!oM
zMK*?u<8(J4J!w>`=rPNH)5h1((lPDR+|Gq39-L|F6v!;DdGGXnZ|9+>mG`UCo;`TF
zcyGj2wpTfiW^TT0Bao)Ew)$J->d+;7e|?=7b@Abx-P3pZfB2}p;%n~J-(L@xTJy*K
z;cNY0)T9u-G3wsB-CE*hg&d~E8(l2p%UbW~Ea>gJH-~SLf!3LL9ZStiw^sFrE-MU6
z5E5-p6j*4tRdGqjq<kHphgmBoCNXW=IVZMD<+16j1xI<N@uZxbCnBNHe_e2K5`Xr+
z9HEFwO{puY%M9IocP!&Q__(!7t0UVll}GP;TA*opZQ1QLFX|TEjbY>6{X=o@wNED%
zPNkgFN;!M3Bjr4!(YtBd+wvT<EY_y4+RwSLNaN)sE|=Zyd{uWlZk@a&K2KPoxxt^M
zWlq#}vA(kN;j3DIPuo7>{Q9=^)^9o=(m9fL?YP<X)8N0ScH7OF(rb(7{5$_mzp-Ct
z-|D&FCKsPs<bFGLL+>#m^{a16tTiffALgDrp53!4Ws;eLQst?hEoUaHL>%6Gap}?j
zoSEXsy}xI5ZnSORZGCNS+M2YJfk$qC;Myk7H*fp8vlqVBg!c0v`*5GR@l$MKtKX(x
zfpXLBW#46IY`tEpxU$gql;*jsp;w)a!#B8=Ml=1}9ko6uNc*a>eW^{|eVMAyKkp~a
z-~I682O)RepnzZAZzg1xmj}yl`zgAkeE*qZ=j}GP6e<)?KHRr<#==K|`_pS*Gxx9B
z&^U2n=4`Qv>~jOou9o3`ly2d#SzPKq@mqXN&VzMw_iSfLKl)q!;cE4p?XNeB<*je@
zf0r+|<*kMNv8sydU9Yu%T+NG*<}QD~x+eIx#p+2uJHIHG=PvP8y=i@;?jmpW-ioO5
z*ysWu%jo2C6GN|(SGR+&SsQQr<6CX<<?)&~>#eFQd>@^T{;+J%@3OBqlB+jfufI{X
ztM|C{+a)C}zoqZ*zW%u)CE8tfZuOtF_gDSic>hq<I-9+AABtby`MfXk_k8*ARki7N
zZ$8eg+WG1C+vLvMt34`er|eoES?`-u*S|Wua^k}YTl=<Idem-xWVV;f;^*#P{~Yz5
zcfYKe&G*&${jZ&m+_#HVe*UVvsV;u)U;7K&Dr)(|`Pvh&m%d&uzv%bHb#=DaY>sXJ
zDgTh)Zr%CmK3~4+*L?W?qdNEXSH*9!i=9IHugUIg;aD&|t1*e?<g$xDCqFd#rYoY+
zGGUU0qS?89W+x3(TF)eferTQbbwi^{yGMWlx0A1TN`pu0mQ^z{85LIETpY4!o6o6$
z%xu<xB??VGHT!NRdAMD!;yA%FiFsYs>=|1#H4POXDj7^!GU>V2@m-sDF7BW7Thf#B
z_IgoO0VCO+p-d^AYoEWfEWNtx-SOk<dI6JuEtgq#b6xbqXV3CO4KFNaI%d9KQ%G^r
z)C&%EwXBmxB2~U_oY|C~Byd(EQ%SKPf6I~7Yv&y1T<h!n`tbSV7f-iqI0iVpnCc*S
zOGLhJerxkXg|Z1>6$M@1&UJPYPt!`NNItlp|8r%LKZ_E>oTZ^BT*CFIHI&U4Q9tYb
z*=V=Jo~K12MiFg_uX@Fn&6)FcO~#wsx@o1aO;ThYI66Og{OgmF<|z>eO976CvI8?U
z-#wI>)Vtes(cF-)$78o%pTSVM;YidH+bvwVZ#Ty7*`2t=dG~r2v*zrFOZeNSPg$<<
z^#NaXf>irKzN^<~hPp&OohRwgkQeaN>&_ZMm$(3h76%t5uaKgYm<3NWW_>)^Ds?tH
zI(I_mBJbmklPeV!-r8T-$t&y<y}J90K~L5nZiOZXg}GDsb_X%7JSD<DW!1_RtPKA`
zESV}&EZsbmc1>mYw!noyOm60FgMjDk+b3MPl;iPq+tyh=XH2iQxvypDkXgVgo2EE7
zqDetgSHHyCdxg*}v55=39G3XBoc!>(O2LS8=UxpJ4+UYSlN%*%RM{pOIoK>OH`fb{
zmX6=0wYU1i|LDpU*Lo{cu1veJ-85#cNX|@226vr+lfq$9bK@jV+?==KTF<J|)2r`&
zzFt<GX{DZ7=fs@hyzHq7^NOM+R(m4$X_&R&y~$AV^hL`Rk)5iduN_}`WU+iX@#E=p
zRxX(bMK5-q*ECbe4#<jiiF$E1aqZst<azE!Ov=-3^n-J)T`i&=Qdh3gt=RUh?8#ME
z`GxB@|8u&^{G;#v{QUXPfBvZXf4a_n`i0-|^S3{LEy$!$`+-4-#i2pM@o$DMM+Z;G
zw|iG4s=ie`nzLovr}joJ*Gz9krGWWKi`1HCEiG4iE_doT&+C`zVzYQWIj1t1G^(^Y
z8kp&>U|Hm;HCd^#C3Rcc<Sye`GrLrd2stuLa;b3@I@Kg}a|Pp!MdxR7G`R^idnk3d
zs90D8<kZFYGA3!R-?*LUAO8tY772z2aX~I0W6DDOjO&8q+SWJszmU4Rre5+(NT5{Q
zwp*SVhn1K#ni37Fo;k*3)w%>ScZ#@OZqvJ6>EJV^hqXm7!CGVSD+9GJ_w){{{Sh@)
zcyi~CZ{c2J)dd}?0RhXey?!mq<iXODzS+3cX1%jZH0O4sOTo*(@=Az!IWjehDewG{
z<)NNdZE0!Y6ccv6Cb!VOZ?);7S22&i-2YG>9eIE5)%K|0-k)zpie+R+UtHDZE}Z@B
z#_A;oZa0=0mc6<D(kJ|>xu5>w?ulpJuEj>z=%2bgfA5=>_it~%th^!jZlTRtRd?$T
zvr8ktDIN+qe`WKH+gAJge*XzS^0K<Z|MAx9l=v^T#bUp`e|~$XyZY~yzsnWRJyi+t
zP~v1+(xfO<HH+u7k|5&}L7@w4R`A{Ro!u31!gZ2_hQR5R$>qALA1^J^_}y{!ifdzk
zy!WqthUMkgtM3(0z4v(T)a)$n2)UIZ!biDRx;kH*!eF%XL6i2wM&6PnX(y}cVUCfj
z#e*8UuQR1&1u&do2w4}d*PF?nxzKNAsotXYDUBY*?O~1+C#`XFpB{BRT-i=j_33FP
zE3K2uOy_C)_&T!AaLH+nb5b^Hnenr*T}{@TVS=!;n;^qe4W?W!g{2%KTco{Kt=GJD
z=vGZoSHmlpEBnovCj@YMymD7zpB}d*AUF8Y&S+P+Ym$}^&Mv(roYiXN)AD-i6_Fsd
zOS2sveyA{K@Ori7HYJ*#&e=L6+H2R`G|dUwhNYGzUOV|+`mLwkwfgwn@3z91L#gM&
zR|N+Ma>m}ZnAO=R{q|7mj8{r#N-Q%J8cKKVS$%r`hTnWA#FZ@L=KIGlzx?uJRju9p
z@XIg1e5<OqEwNS={IM!+#do8?a_>2o7th38y?e1c>Dj~LqRFy%9mCDF=bG^@IrnaM
z(X(^Ao}Lx=p1eEhNXoY7<+0cEro`M{D$y-w!n&@wJUr>xu>*~POSZ6NF+E+k?Dm`$
z+spI9HP(e$WLTbhwzDTX#ZV+K)=X1zworyvOK<Y5?%C6xY3RL^d1YQ6c{gTfd0F?Q
z=&3u|bx*r>?aaJWUaqn(;q~skm_<9cm**udFMIcDr$@NnyU8<--n|ra*UVSSaNWB_
zG7pXfu6U@o>w@ZH6_H}UU6zL|ij%E-PduLVcBgoGrg9;}v1dDG`JOF$*H^qDTzgh8
z<LTKuFP`msndf@C*zfJx6O)YIb-kPP!aTNcr_9Epa#y!^@7}Fmb|mz$%)4`QN)K&U
zFNx*7^IC70^5nT29iF?SmtHQGSy_7b?#{W#-b|kMP>MO`*xZ}tVZLS0V~SUq3q3tN
z<J`NX?oN?F!Lz&PE_UnfJ)D=Nu|#)gUdW{-N8VjCo^A20yY24XLhb2>I~PwqSJd&`
z_r!IRcMlhYm+9n9J?(CCO7_m$vr==OTW1Me_b!e+x9zFu+3ac0cE;XSTs>#y-JL}{
zRnHaejCiKEd#&1aQ@5Uw^fKL|qdU*cc8qy>_vEUl>vB)Wl*c^XdG|=sy6AwuLn;T~
zIr2*;ubK7Z`qU61_L&?94ldbyGF8Er<7;JXV>8oR=baoZYYyu)m|Reqsi3(rVWMoB
zN|yqs@2015IWuH9cPUKrFbSN!W3kXxc9)Gw|4dXb|5)mIi6ddbj!RQY+yztIUiEfa
zmE81rlyYTqa#p-txN=wD;qE%!uP?%GteL;$lgyv{*)^eW%QVjYjoG;WqXqBZ(w+a*
z+aAl-_+Os%dTD?7%}1>Tr#78io3*gWFa5`Iv8S4cl+}99?e=r?^IUVD-H+X(%-C$}
zd4WtO1%?G)ewD4Z0cxHPC)8|^nqYa%t9Zi6tkrX>PnZUonD07myo}jgiS-oAhu0;(
zzhmE}Y)k4?5M~klF_S-t<+<L4wXB@(A<Mj`T-vfxYtE~J+YBF%b*vY=QR;VzaqIlN
zt7p5k!;h^Csht|D>Rk52^x^j%vO=mhCo<$_#wW1{yz)=0j1`_cL-CwOl)sJU!)q=V
z=9(?@$WoY^wD6<rHoN=AA8yvz$@BcNljpPNZ$G^G;fEh(HP%T?tK{sXSy=t!zU9YV
ze7E_5>1^lLxGi?tVf)k`aeQ0I#Ze}FsMRThKWCw*qe%YO&9gP+g*VJ>alFa6?d2(z
zC33HB{di|_Z@%=;b6-@iD;(SE|GLZcpu3#JPc`EfyWr2!Yd$_%-Prr@^vCzvHy@O~
z+*x+|D&O_`aQzeBGjIF9nE!a+ysyt=qGT*uJYH!uYgtCPyWOmsx%J75bAel)H2hvE
z>%AhTwXF5Zt`;kWYqv8O>&!ZB+Nw}9Ma?^A)k&dk&B3po0vcB*`7X6|$zAMrW0i9H
zuIr|vD_Bl4oiUD3^S3$W9Tp^(K0U&HWxMN(FGb(}^{21D_~J{&U%7tw_4i+VDXX#$
z6%^$cU*BGq)}buo;C!RO;jhSo=F~LHY8B)Da|RZ7!*sijD!kdlb6-9(V^*Zb-^$7_
zTUU!tR|~zhZKeN_()!h<S2zFZy7~8M6aPo;$<u-!9)4+lw$Q$Je>A@h`$eamJTWPC
z%k<_PtDgzGT8$WtH2JfH6*Mbv=87cVUMqE5u&9XPWZ5R(1zeSfo@HEiTTts!6dpXe
zC*{c##TIr8qqr5}K_zLtxy<fvlIP>*O?~<DWxc<@?_XP6i(ge$#b2H&Feq7Ds!Z}o
zc$czeNvDEd=j<P=w%*d{Vw2S5p73{n$265H6Nz7c&av&Y&;7W5TD{_`lRKX{vAeJ&
z2%T^d^<m<h_oVlJ?0VDh=K9@BrS9y!d&y(Ml9wB%A20oMHr8@py}kSV^Uq&>{`u!0
zzy194k3awX^ZVz|0|E?EZ-i#Z%Lln+92Iz@Vi=I|Wf$A&M~dcO()V2acH3An%dPl%
znE3f_iIexPI{A8@)A#p2o0|RjX@1e#++EGNKO+8=#@pII{@?m*H*d!xYrgqs4zzl3
zYBx<j(Bdj^JY@rCho+0hf+kiKbHzysOgu%u!oK=MU%qTNZ{F-*RaK?Gs;bI<RaKRI
zdD|h#q-0XfBP^l(+G@!?Cle?3g^R!JOAiWCaZ;3U;b-fc(2{beL6YIt9A^c-f8X!l
zzrXI^zrVYG|Na~M&tHAxleYm3AKll7-L<;9f7|Ek`PnP3J6^U`QDkXoTIj{_mV1KW
zlTYcBPd<6^snTYizxw2pk3Ut~%(ed6z`#&g+ta*rl6C3BM3cn2w~;GY1pKdmd1?Fi
z@88<LfB&toudh$9udm;D|4!F;1%;Xii<*^`EIb(*G>r5wI4D)*y}o|E?sjD{zr8l+
zY}uV!`)7YVt8aNNNc-8%$F+ZdOn3Qn`*id_!;ed=^%vR($85WPV_Om1_ucD^p8gH1
z{e5fKyXlwDM!yo>GATT>SxaPl*K~IN{Izek`NZ`6jWx9R{`>m=`|r2!zyJQ>`|rPh
z|6QMb>)XHU848+@zDtyOU!6J4o>^v3wDkMawclUful?)${r=Vawcqb=VPM{U{DYJc
zh+WtF9YocBzhC=L{`tQC<G0EM(kIJKVqjo=a?zFZSIIt0zUhzlH|F^t3B98J&9lko
zdSty;yt`)81e1L``sPhJ_dj%HTkcdbC65W8UMH+N-!%EN-d^3Rt$!>ge)Z3jt*Cmn
z?bPL}XC?JBoY(d(GiUu;ySivaOJJDKR^bDuA1$o){B?7;_I~^GU+n7VKmQe9`1#NO
z?9YGx+kXD@-~RCPyYDT}70=Gyb{m9tp8G!My(NUb?e-p!RDL%|V%zQf$#b7uo-2lM
za<^NaE1vWG<lD3R+Q1sO-F<I)@B413?5$P*6$L%nMB^_n(p2i&x$pP;{p+v4{`%{0
z-TrteW7jil_V&{seLj9Wwday+ZMkK=M|boiv1$INHZT2j!Rdd;m-CPM`1e03+|KNs
z`P3}??z{EnyYJ?g@4kEZ?z````)~Y~PkvL|3v=q;_m%gcmcxY5x!bMpeV+qyEC#v%
z3CM4I-p{Lij*MpSeQtdhYTg|TO%RFg*5|&%gAXEvLFSe3d2a^QUw8jrIcWZ+8r7<K
zb(8nJufA74=ks|R6a@3e^J+xk+r57Ece!2b$KSC*OTTuPzyEe8^sU9po$uctudlDO
z{{8#g`}gl(zkmP!-TUX~=iAkM16%#${qC=?udk2WTZKeL&wX!scH5nwpTP>jlzl*7
z{Uwe+@n=8(_x=1g{>gsLKi`trisyhrNWgdke{}<YbpziA<WW-j*?WKPH(=j;04D+s
z;lp+CpS}0z{tV`O51<G%^#qrMG3xTqf(i#17nQa*fA$B*$SDJZ{Or$l;30knvERS{
z${*hQ9Uc5V{Pq8SDA<5(@=q8*7NiTCan5tg^G|Ch*e-EU0Yf$RzO8GsqaGe^yI`&q
zFh!Nq2}}f4*vRoh3X}zTE=(YpASnHQ-)}HF#0HM)dv8HW@Oymyn%lYC?%P9w0^B3a
z)%*F^_xEnQo%cOD7es;*+4gFXrrUdwOn6@g;*^84^=@!Rxcd$yl3#x9cJ4Oh2)Ort
zeLuuN5{=6H4%V>yHEdo6>d5@^>&K5DyPZpdtDr_fym9yIw%fVLacjQ&_2b8nbGH%Y
zK@_7PA@c6m?c8lQpsoW^(s#dp{P=O(?OaG`L&Azf8sr^NKtiKp+wI&f{C4YX(x=Zy
z0<Zq9fAw$ukNoR@@87A{KmF;?ekg$%)=;PHem^_{2&V@j(FO@?NQi!~&E0l82Ncv$
z%=_-wj~_p7=Wc_9H3>#RG*s^cMGVw15XJVrwx*^gciU~8M(zIeXTSHCefymeQG{T9
z`S<U?|9dca12T>n3@c!M5~G-OHH_cy9$Fya&?v~p@a@3IbJ>>X%J0{<@Phme%D9QL
zAQmJO(=1_q{<{4kc#_qw4hkF~9$~!qzWN@bq^JRvQq{Zfec$kJ8be2i0`K>G+rcd*
zB=<5yvlApmK%x;dvhvGcydB7-Qnwl8%zLj^t=fGzG&H<=)v8tROG86r->q7;`tH}z
z&{~N9!6b9_zI}FfX1BBVUJVT`-?eJhy6>T(p%+ZoW=CDyy4K<G>3+`7Xy8@j^~W#(
zR*hoQ_quJXR{aK>4IyBu`Z|g`P^?2Wpf)r#9vVep6QEIa3+@NpdO=F!-mhA<8oc-c
zteb%WRx*P9k6$lX32{M$;%1P&V7(yYtM{#1wH|p53as*g2QOUj?)UG%?05YoUylZs
z|M+L$|M)lmV-Ss8%0s<GGR1x#3Qmx$6|RB#>%R9_palkIWUMcLOiapxdmSS+kPvJb
zdO;z(zWhaPXkBYK=$tSHhMGT60AAyPnuPvxty=ZIKx;K9vOqeqBw~7{K~OGRAs18e
zVZJ^p&|4u}X9opV0;X$`GAAeq_ugN-YSliZCJ*D@`|<Mf(z#nfvY6Q(!>fDmhlWCf
z3KGGv1_=>*!Af?&TD9sIL@Ssa#M%7a@85s*FaG(z5WV63tG*wc!yf-V{qav7QfPvG
zL~R0^b$$}6YvpU-eXm=-`y=Dmub8QG_q$cA{z4B{13M6R;esQVTn!EV0@ev3V67O;
zOn_3T!Cmt1*Q!<fz?ll7pY(*+el@hN3Z8V3YGurVdiMuVg~WiUzEINfyFdR!f1sw~
z{POD$f8=gMFP>I|E3=2k`#C?Mf#=tmHy?)qC$_B^`D|rqXf2X=*}m7-$+z9kB{#U3
ztM~2aKb*Ub+}aQn-0csyq1D&8Gd)JBdhh-E!yj&wUryi%;Z@hRuFYcS*R#?3h8Dt~
z-hKpwmyaJl1@+pCAn9fxNO@@2(_Vprf$`V<@>g~B=YRgUfBxq;YV!vaoO|#0qGlB7
zyI+6QN#||@mws6D7HXP6FCcl}{rY3iyX`i-odC~LAnPzqMQ?`CJ>}JHezj^PQc=P9
z`Sa(!_v70ihDM<l4ZFWi4Gn$w`>+0`Uq62T+?PN587cv4je;8b&#S?m_S?DJV53+0
z%g=qD``q$xedW3GInTFSpZh%Ly)|^qYra2p=m|1ZwLL#SKR>Je5_|roo8q~>JU-zc
zi*wT=KYo)sv(fKy$<E^GQ}4&?RhGxhdG~(#!k4kK@5|RoPrRJ5_40J9s3)g)ue{rO
zCuZl}yVHD2w>^9Ibgt)>Nhy<39v$ZKxgD$@onCsqyyVUM@<}=R<=#7<sJR8YWk;^&
zQoi=9QK)gLgWX>(Nsd*Q>^oe%7`y_PvS_dhE~qSi-aIYMBHeASX9(Mp%|^z%i*4^L
zzI&s3x0Z7Cu5Q<z+n(->UNUWx(7U%i8CPb<KF>beU0&AP(d*^X_gZRkXvxN&2$QIt
zJ4K3~ZarOOy7SrR@O5v0h`uj-y6}3+b?u_pb7PfauRRqC*E^*uWV&;D={Z*^)8gGb
zq<6i}@2;GCySR?+US&;q?C#>~yVbk;-rW%Pobl^z<({tV<*{eq$C*s>`uTa!&Yi)`
zK6*vv)py^_y<WUay_}_W@Adrh<XtbG)K<MXS@rX}ZE@deS-o&W*Sj}jcJFxk_12!f
zI&({2mwtWsUGLYg?+ew>m6kk>y<0Fzz1;Df@ia@;(5w%~ZcA84obyddOpAH*;sIM@
zl-L9*fsl1s7gjoJ9ox3{({FkI?PZE~`u%&a`_D5yyzBYpja9{hTk_tmxL%&IbM?+U
zx2NkCix#`;y{qi{`0lP{e)s9KXA+JSJ=4&rynX52-CaI+il6WLE_>X1r;5zG^75M4
zCofMgUaq`*=ce7e!pr?$y)3Y^eP=cIuHW8_%WV=z&l$|za?^F@%*F|+)ApHN%+fO5
zxOMgWdXG@Skm#u2wT<82?>!`6etP?q^!rzL*j~3P;{9P;Soyy6>3KczhM3oPi@xuk
zRrS2ICg1zayPd9gZ<gondT#xudfxN$(sL!RcP}q#*?m{8`}OPlX}@dpH<$0QEGjS9
zH*fFlr8VDG)b+RT`gwhxexdl@>b;LEV%kg3H2t`1KJ(7YJI{;Dxn35{tG)mK+u~eN
zx1(Wook9&>84L?VUq3&kKCi^pO3OuH-}k+x*Naq7_dUHTv9Ef4*Upc5<-4n@`}SGB
z|9`itvL^QJZK-mZvUkOMo_@J6eZG2ca^6&)21nswBUXtghggITaZH$`s?g{rsOjY4
znb6kSt)8?<RiRNcFd&old+Gg{*uL=M620{6)pw6`PtcxQT=G5k>80PMudm-M-xV|S
zqM=oKta)L1@p9F>=j_?vmu9hKgf&Wq%GWepPzX|G^k^!&Jx|@}VxOc)y5E);r;?2(
zmY406RWB;~Tx`Dc=I!ZLsoS-lMQ&cL_bxv;dRFAlob6h%Cw1J<y|li4^?Av=>vin&
zo)_=DFZ-_gZskqyUC$qUZ@GNEc<bGhbMN|<JC)zQeX_c|+(#<u^`_Td@1DJT6>F8h
z?&Z_DvDf(uW35Wwy?bu^-hIxJXHVK~@9C@Py)*i~|JCZ`Q0a`@S!S-UUphLR6zB+L
zYZSWq{LEoJrr&>eUyt3-Tv-!t7kgbV_F$j0UU}@1zE@A4i|_7;+x4>Wj-B=C9nT-U
zuZq3g7rU=~>cd@@H}2k@JoRpYYFY8VcYD4U&Sfv2dhXG?7@6I?clYesTka5UUA(;f
zdinI?Hy__UzkTns-s7p4o|hcGe(u$e`?hB<DLrnz^z`DnU$2)>57%EGyCm-A=j|Il
zJ>I>0r)=qcnLUqZY`s%@_YQC6JwLt2+iI(0U-zCpxqHXXySI0}{CY2Z@~(H>cVC~7
z{d{lFF4?=y$<t%M-a9HMy}O*hyn4}2p`Z7jSIsGRciyFU-}d_CUDJw}3(tz3dVc-9
zeRf52=lfT^FW2{baz8!pPRyg)y>B0?UY8Y@EiU)FTRgpZu3x%v*}F8)yLX<rv)*kl
zn|}LV&AZ*tSG@baV|sC4dF&mX=WDHQ^Jw)2z0K;^ou$)PIdRP)W6sIe#pOky^JpI5
z`}DD#y?pQca+&Ax*73GwKaZEE+m*+~?6O|IyZi9Y&wKsDg+)*Ap1rT!sJfSPQcCfT
zcPu}z?}=@gdtJF?_qypNlaln;m%e*FWgUw$D>FxMON4{Utu`l5N0F(LiK^wXr?RbL
zYpeIX+x2eF?xlBk?_T$AXSvn%>wEVa6;F8g?rnK_$&2EN7KNs#7tcL;?kOkV+q>t^
zUKdN@-+lMp)LZ4HcdgzP-r2R2V`<6V;%%wd^UBLhzkT0(t}s5o`tJ4O-Rl;Ys}=~h
zbOdmz3e9wrZ1J%9d;aeH^4)jeeJ}i83SR2G`<>nKv%5j{5@;w2R!2eEdGN`jeaE*W
z&%w;Ae17rU^P_L?mFMm|4bt>=+wHf%x82^33N|v|(a@-S>cY^%TD%)tk3l9<%&X6R
zp7R`1r)|5n3J#uIBXK~Y`_|j7-@Y~c;xE(C(9rTZ?_b@!zoFLq>;L6n>b76}vA611
z+}^6+_Hlb_{p0r5#>efg+8r7i8XEp!=egoN|HbD#x4!os&EfMvdl2q@{^bvG35-0w
z=eZWhLNvVlToFiQ{egPnZMSo`U%8#T?JCFyFb=Og2d00&ue|qt4}6XeNr5HAJAcI>
zA^K<i$#3`jp4a`a3;!=yulVbIojia0r5`nR^4T9iB!BzihaiS1$k_O<XU{Fa-T7U=
z{+C_d&mZf)KmJ)4KmAMHfB*Y$K@+imA>jAF+j~mie=M1A_y24~MMXAv3S~NcDCOxt
zy9@4b&u=k&_)f0weER0I&pub~xtDJ~8^QSCv|yiNV?g{puBLzf$IpW&qh9_3g(`Gb
zD(3y2YEUe1yPdng{Q1lJa8O*@*~5bFyxsO&f4A07jhDX!3M&xa2??0WuXVTI)ieLF
zzr1fcWI+9^RsQAWesir#US3}Azwpg3xhvMYF~N=>zpMD>-2*9G2O?lvt_nV|NxuC;
z`QTsvS^wlPKz!4;`sdXbS@@R(226Dn28RR^vD|O2<q?h!=J{LPLCq*My!yh#%kAyQ
zjzb;i`wQem5a(NY-Q#QQ)AKi9KD5V1_aPd%W%qb5%G}_2n?HPz*uGs`2Mx8Y_x|jI
zishSw=WlYi-EaAhrtkN@?{nT)o<IM6&!6KKN`fkyT`SZeQ%Ep!f5Pm!?<>!JhMJWN
zN+w9Sd@B+Mg<1FT-e=f6YySJnbH#JsTSGHFIJJgbq9})Q&Vl8Mr$V_<D)zrUWL6Y1
zg}LoER2)Xdf)W5aeqITa4h{bxx_0gA&+}geE}JMQ+T+%BRJB_~!c3c0;-DKdOR$K^
z#KkN*xlKN2S$iClMIy8XnGSV%DRuEmC3cD|Ve{zRaAsmq%83?FmIFp7xO_R6G6{8d
z^t1>ioS4Gf=5lhPr{SU#sTxbJs5J#$W>(`;;MtUtvdJ~c;mCw0E{&*FN)huO@V<pi
z)`7|TxP5-bd!JXHyNBdkbU(a>yB@&^e`WJRp^hh1Ld4Y=GYRkBye01YT+ASfc?vQi
z9IdR`c?vne-oU^wQC9v*0#2@?bqtK=pkerWaOQ+e7?&YyfwEU)O93$;$<@Es70-EZ
zpZ47H+~?ca%0)yhBROs-$gvO%Dak-af${zw@2&5Bp9?R;pI-+pHh`t_EU+3lu^Mc6
zUh$mw)vy%-NDjTVYSrq^VDQ|Pk$tDAi$Z`vefe+u`zik(oL*ObxAc1Tdh6@4caOgN
z`mW5p>bvjV+jnffZ%@7SF7~?huJ`3heDC&Hy?a$&Jg00|U-jO3ubw15Uv~G^yS%&a
zKBX_-@$~NNJ^Zmng|D|yzGJ<vNapd*JMYWBZ}+skd-w9*cYE%hE8g|o@B6wvce*r}
zoz@VL(p<EV?T%TKhf1Q-<k!9LUcD<fpIcRZZ}z2^JN)L%tG-@(SubDfOWeEadtY~(
z?4JAHd+)oQ;+}Wjon9VZP<^+u`fkO`(>w28ES|jQ`>w~kzV9l1Z+*v~rR7~|^{Vg7
z%Vn<{uk$NCysi44__}wccQWsNz8+p!eWxpC_uO;bJFl0=zTSK9y<c3_mv_5M%f(~g
z$E}aOZoTum=$+-Ix1Q^lUVOYgzdY_eziP>y@4NTDFW*(Zx1v~W&vSp%_px!)S*%y@
zK7IP`z3F?8y}Vh@(S7>7`R+OEEB4;q6}Nkx$#ebk*moaK?|G7!ap#`ryB*bw*F8D+
z?z_<(oA-s5ckb?*dA)k~?%mtBuPiIHy?SY$tZnt4?|XOeK6~=|?Dge++xO-jy;pp<
zxOnfq;?8+xJDZoAJ})nLT6y>K?%SvD-g~`!&WrGQOyBq3ExojRcX|1o-S^h#msihC
z-}&y&M_a$V>OJ@F#on<lpSQOvPT$J=yncS^x!t>VyH}Q1UY|Gb`{b*aK0SSJdwqNN
zuFt3Vx|f&UExmhs_uL;_&wamt`#uJE^1Sk1`JCtZh;+95{k_k7-dmr8RaD@#w;g*k
zV|Dr5_txiNWiu#whdr;pR}7iw+;%&A&%gM&&#liD|M4r{`@ZsgUisYTmiNl9Tc0bR
z^Ggn8sYcx1DzHUh;?DbO$duUa+^yi;vfc7r`5f@%d@i(<%=^6OeI>Nuy$*IfXax+s
zn5e(#==_6ewNlsa6-wZ>1CWCMAIQ$~InOI$Zr%oJ*TBm2+<V{WJcrEGf@ab|K7o1-
z?5ee(rb#twxPT?V#J>M?KqUfrHNm#qxz8)_eMaPIunI7NDF34OfYcycunrV}p*!yt
z&wU21m+CDc%W0shAX)u+^}TYqN5NXaMC@~GNHqiIfr;{|AHikYulvz+e_EgK`mwJS
T)TaCY=$*%ME+K)X0{RR9BMNcs
literal 0
HcmV?d00001
diff --git a/data/mcglm.rda b/data/mcglm.rda
new file mode 100644
index 0000000000000000000000000000000000000000..30f7ffcc63641564f873dcb6308b2ea1f6303138
GIT binary patch
literal 271560
zcmZ>Y%CIzaj8qGbRDQ~<#kz?3KSA(VTaM)p0|Ns?6^LM9V9;P-Y5+007#7I<=4)`#
z;$&oC5Mf|o;AnJe;Nn{Suu6tu0!RU)00S!{16SxV9tH*mE)|9)9xH>o{P!?SsFho1
zJ!P$_tB7ss#4U&1LpXz+9XLybE^souD3R%9R(4=ukeHI~pw+<S!0y1nkYUBZaM2-v
zA!`eR2m^B(0~Z71S_a+>28|S+tqx&rZCe*`cwNZqWjJALWyN5YcSGjP1*tW=vKQWM
zJe|gQ_oT%f<=EM~cdVVEDt0V8I^x#Z4c(c&xkk6MZqAZ4OFMgeXVl#tv6?=yH=b^D
zUB;cvbzAXT*NvCk9<8{Udgr+>uay+jibv5Ik{RX+va-QGXD(=ngfEa~c=3gSVH-oH
z*Ro!Qpsu9~3=ATS3mEtvFBnSa#WXAkc3|t}U%<TZ0^_oVz$K0g1uif!h-io~T*%;B
zz;M|)gn@yrSwM=dL9Rhcmci@*LlOgn2}1xVn7B9^9XhlcxB?iqF=RP8FgqLp1tk{)
z<Dv!zo@@qY2L=XK2BroLkSH&hWMH_!a3O$Ufrb&|HU<+&+?-%s?9<l3)HaK$m#1mM
z0)`6=3=BrB3<_Ng3=F@F81_nlv@<Ypf#MAjpPI{7IxqyeGAv*cSRKLC0E$zH4U8eI
zj1CM849w}fLDT|<0Fd~q040zy0SqDxTwKsZ0TN&U`-IU!G0>^0flGsT14DoaCj%=h
z1H%Fa4I%3XrP~@9Rt6d^Vqn#1aNc*<v3JExo_)@)EF1xhF+B|oEWNx6ERsnK0zDDR
zW-YuD%4!_%@|YMJ+#-55v~6gPV3ABRI%3)+bx1Oiub0nPGD)(>bm26&oI4D)4AzR_
zO(l$F)|oSnC3Kr+m~P4NHJosPE8yB}L$9@xL5r>|ueM~+H`8d}JA+||1H)7oOM`_B
z9Sc}kJ-`lU4XtQ6oe|-{kifv4^F{LFz2htyhZxwmzV>B!QPxqC<s$I;xXW&pKF5Wx
z7k6B^>ST0MahaNrvYOyDXB}Z3w}@UIjx%RAN~XA^E;dLnd(@P=Qqmyyc9-A`DW+@<
z!HG>qM@~*w6V_34nIY)F5ahE}ErRpTUC-GrCr(b&@R{M3Hd)tEGO^t7vf8GU44ZJC
zCr6k!NX^-3d%J0`lvcxKrRv+$BA7IUmZ<FBc*G@jsw)>$=L{aMrt}?FT&fBk3l=nV
z-^=S#l$<qT_jXOCY>iH)0F@<^bzM$}oZ(e<m>P7N<Fla$SKkH(ucb=;#XD4mTwJ<!
zI2<*nEiBJtSQ+Z&sg}U+*Ra6FFCu`^ful8ORhmK2%q>DXJ)4dS>v)JoG&H5Qu<T0u
zzMZw3_3jR_sCRk0=M)=mY)ie_S@f>FJa5yXqbggcZF0+5<h$Xbv~jXhS>BnPtvxr+
zPTM%!*tjh3UEU_|<ZO*eZ=R>lJ?Ck0_w3ObW(GGiXQ$<CI%Z~eEGKtb#JMx?p36>)
zOo&YFF)w>`=4Q6!?m2JHoVlseQ<k1)W^iU>$BiRy$8vJJqS9{M-g)PA#BJ%+-80^7
zI<0fI$INikku%ae&%G<S*`7MRc<zid$4*PkmfUpo&f`s|C2pMUxta9tZn@>n#MBwR
zGjh^XKV&Qvntf4W76St#FF5uY7&sdkmNzgk7$q<;FkD=w=D^^<z}RTO&|s8tA)r%J
zfPo=%GG`kzNU@p&8yhc!QU-&vbDGB_1~-!#J{eq<2EM!oK9fxtW(J@4N-|ry{8iPg
zo-1GZH7A%VsI`F$O$LSr1_nlxj0;8vEk_QpTwE~IAY;Z%lM9T75{x`*EDR}b95Yuw
zIpn_B^Cjb|3Kz8o76t~+g%cPU7l3SKU|?WjVBj^m><i{)E?{6-zKmf8gW7`09+MO~
z7tG2~Q<yb#)x3F1^QYR}nz8(qsm;9JkiYJ;+U6xQE>Ot~`77;J?bfqI_{(3DtNvG(
z^)fK9IWU|68^*xEHj7bc5{CkVLIwi^LuP{%uT%m90}IEA0}T?F1-&u@7ppJ%YBO)v
z;!`;@%{(&vuNqa&$e5MJ#ean%H~3Pezp3x!iz_Y%oZD7XE-3hpN5{G~?$e!Zr#SuY
z>P}lX!y;wF$*q;Ht7R8YH<o#-9j?00*@tH{&yk&(M=zbdohMy(D`R4=ciW4Vb1ry2
z3p_ckY}sO7tu-MlHs@|zvnD90)9uia71N4NE38&pCA=lY#-?V^^9A!}zTjKP!Fc&~
z+UF}<rFB<ZhrKaB*LRn1{^!M)Pp6gt5qFQ?`{>wZ|5a-@oJ`Hh`M!rgRh_^0$>&{Z
znlrb>*Ui&QUpV9J#~bUv|Bv6*)NelbbZV{K^7b7!+0N(K#k*aedB|l)o_Xb&<?buv
zlXu?zKkeN2u=mlIHEd(|JKl~>UYBppZ#t{$d-JIj_N8xga@lSu$X&W(zDJqG=6-tN
z^XRP$Z%ge}jWXU=TAkTk8tY-(VSDa<g+ddDLerI%&gOUX&KR85GEenb@Rspvp^DlC
z)|CCTc0Vk!{P$sA-=ftM=XA$@%sdq)a-XZ@@ZLFEwzlv3_uo%=e&NUV8UNPRlrSAP
z?|k)3y0+2Ec_rV*OFw`ATip1n=g1A6_!CFDWe<P8dgsfFA4kJ_WkOd?S{oQTH9}x!
zsAyK#)}X6xH$ux)uC5MKjoO;9p*U=9)XLIVZGOS46L>!#ncZusQeF3Y`uR4d9p>NE
zAN?_M;r_t%_4W1jaeJ%2KCB52H;S8=_2K#QJKx-%7$yEO`0CVTE*^IG>Uo}o3^%s+
zu6)DS&*m>Ne!{i7_^Hy_4o8M2!IeI_FMn8^JRF)C(xa*v)_ru*kE6>h&WHV$ySz5f
z+^6JA{NsM{BXSqgXY;tnDfC&~*xPlwd(Gy}6=wHr&RniMF6G~JLo+fyP^6H3soK`>
zM=pHdQW#-5@yg3PC+|LCD@(R~lsj*WN^S1$!qXRzxZha$u_|*}56kxTi|-VFWct?o
zV)N~rU##La3f5jw(pc2TGuQinNp*y7a7uCU+m=ot%e9MilS0k~Nxb5&`2IDY{m8Ma
z8w#&#C6-j&nV)={JFMkVqs-2t(=1Y1&$N$Z^)4u3^UY5Be`lKW++)wK`mdR6Kfh&F
z(7e=BadsZjYj*GOEsl9<x`Nl_x+Kpr&nwQKTA$19nf3AV@;p<Q%FaNURxVMg4@UYE
z_RjFSI<c3dNhAAovf+d+I}>y^9Zp@?RPyjL^X#<?Q{*b&rQIw3P}F?jsqu!L7I~@O
z*ZoxE*+sK$+Aizp1YKPav(q9<NaTqbf9$N4L8~w9QSq3l<kHFRV8z?C)kjJ%gRib4
zmB+=fQAuQOa_S75wH3Bpf9lHowb!10+WkK6`pI{FFD9IvtmZS*h)<y7!=0LYW$eN7
z{U1O4n!M$D!MQo=T$fZs_AGf4E#p<bJXvjPpZC>p*MLlJ4?P}(OaV`JKD~7-e3puv
zvne%6_t#YHR|?k_(i6HY*yXC(G`Vw9;c@0ZB_6fST>eS_iVY_Ays@xi{N}W!H%U&=
zCEL7WllEkTTdVmM+w+1p=k4gQ4Y`xl`ufn6R}Y@LJM>>Yc#KU|&o83NqPe2DNSp7#
zo5!uZBBvF*gI~;K(z<BaJ$?Jz%rl%qit}VTRAv;)EsIsKZc~%=``*2K>)DdY-ZQ6e
z?U`F_{chF5ZEEIQmT$3{D9`qek7LcbPd1I3G39f-KTI+$b==bK*R*W;ZK0mBe9`OL
z<`q#lZYYO+xxzJ-HR$#yd2XgnF|5@k+C1i!(#38rdh_?D3QGQ#c&Iab<K&*{Y>$&m
zfBAoMJ*ToR{@@*7o%B^3RbMPJ5<MpI=);EM>>okX^LQup)J#~Av74*tZF!-a#+qF+
zXJ&;o9s4pXh`Zl-`!yepi;G=UCwD5#=;TehE<gKb{iHAU;-6GrPE<P+vZywomFfNc
zz4?s?#8^8+w3T*nJ!;)~>G2&7{R71^IviI;GvrTSlJ+$I$vc~U!qd|ilG9h-c~oh!
zaCcpBna|{V?ToW0Di*ngF-<kOQOKlZRXMR^TKLK3B^pr*PX%qoZj0ZkzA)3OdU9J<
zC7-i!ytBK1gCE!AuFW@(e6q7w(b~@4yG|r=Q;O`p8xfZcW`+40Pi))0M|ML$-&Yo^
z{J8c-8Iv}oZdVhTl(>b}(?|BCz%#)+nLQV0s#whUd#q&MKbL@y4<a^NJ-ynxi@C)l
z=emrGThu`<#w{;1vl2q*7+mG`a{O!*(z)LCxv>yi!HMD-UthF{-Qh9bY?AWj?f2aW
zSp*ze1RNRG#;$(4lQH3vsH#mW6GOd<Y1cWvzqQN4Bi5f5N!!BHdLeVMY-#HHNF$Em
zV`rD@uYa;roR{nBp443%;y)c_WH&04NRFL!QReh1KF_6W@9d;AzSQ|M96We-;i((%
zx_2aPQRLdlm)^SZ;@;ao@7vf`T)eG2`(X3I`>R#CKAPQ7@|kIra&k`G30tF-lara1
zTsnnRJSH+<RC4K9V32T7P1WE)Jdc<EXOqt<x6+RYX!dDskE{|`FMoZ|_D1`~)xy?G
z_Qam7lI>Zz#cUeOl$O6uPX5iz-IY3SH&;sis%OvAn5?1`8N08n?A_dF_p&W)6TP1a
zACZr$XNZWOb0H-~diwg++tZU1lC(G%nLT`*_U4OHZ}tQ$|JFv^L+hN`m?clRZd)xT
zYPl!xc@>{PCtF2|e&ez}r`K1vNg9bJe9{g(;w2YbS@vr7#`x~-uBtuJ&*hHK6H+N#
z_lom-MdRjWr)I~W-9KT`ztSJcJ~I!gK5a{y_4;_y#m6CcefH()pE)8vU2DY!(Uz|p
z^|j1|F137FFCQ0|S5%k1_2KK2j2*j-kGo~ZbqcMuxK$a#?s6e`?Tyood-v7L{M}GE
zqq-u??eN!o4*QzK^}H${Zt!$5iFqfHv^7P?#^%7*oV9u<q7>LK2sF6K`6ZY69S(lT
z$|m;5Sitkia<0IwN1sKpYq|!tYlNK7WmaczycXUb-#L4i_Z^QMxpe|9%Cj65dp0e}
zG(7b%PW;r37iFQVL!V6dnlLNZuq?P?USY$fUa|gFLQ4%)JSM1^DtJsdn5yzf+t70J
z3a>LqE-m>!cdNGNbDb2vxCOh;&N+HxPwK5dS05UsoZKiJ;CC$V&$FhlXL8C*oVUz5
z_iUkpTj!kl%%yel((FcO|0JI4k&nNAs<Brz_-p>5l#`Rw4g8h(UW-RCo49leUh!yR
z;gAdn*uyJjl4<jnq4N2>%IB?%ElXaW5<mB(SZ4ndpLfM9vzfkqz53a+Fk5?TjGE6(
zV-BCJ^Yc#Ce4J)*&ql)Q#L6cb4tI=`WC{{WR;t};IoQ_AC9peegXPYof;_{UYwM5H
zipbS^eYr5j`lS4lYk`uIxgYWpV~a$RVs?s&eo+;E9(C>74bilFuXeAyx#mpU&s{TQ
z4yLW{c<frZV_U_J?>|o$$UH96%VgD?YgO{{@>0FI7CVdM9YV^zxr1#F*ZivRGdJG6
zb1_F+nof6R<t+Z~X<KD=11mbTbSz|N7(IQK_49%xTXW}9Yqm8HJ8PD7rL&xg``Ga=
zylK;!T?cEQWh|ZM9GdcxpS|Zwn3RpgjDo8t8AXC!&pk>?PdIgxbFx~nh+)yzi;`@a
zVy|pUB-;$duiTVc64@J`7WK&3X8GzXt(?a;TzM3F?Vysj=;{j*?z&-HmiG!S+i>Kv
zRAO?jjm~-n6Um*59Dzz|X)YBvBRzf3Es@?nS;X5wbEemS@qkDU1p%imd)7Rte58`F
zQ!A>lG%PF0`7B?jjv0^h%+2wF58urFxb#y~@rS$zt%jEuWlv~arO1_f<m9>=(SAY)
zcqT2a?V7ia+11)IjOoJ-lajR;#ig#?*zjicA%P}k0kz7mC4GkHB6~zmq&ZL8=cTak
z*v5`Um73ETT9^48=v;W{(SuuC#8szU>Cs<wHd8+}>+8b~!{%+%wcA&}aQ8bZZhT^?
z*EG%8orfYetMQps{K!k%$>*_iX}V5Y^ZX<GIb?DkN@O0cdwRyn%Jk;S`-?QZd}@>r
z1qZMn;g1zu6QrIbX@4T=Emw1dcFbX|Z(Cj-Tu~*thWmQ!GyS-|RUh`~$L(uXGs})V
zzURreFCJG+`5&xl^LF=`u6HwV@%4Qf-<;FWzJLGJaWdDtqHAj-H>Z0%ZqW!_+<7im
zW6HFkKwI7G<#OE3ypNAJJ~7L=c`%ZvFHz?}#AfNPhr7!>c`NR|W_@+%Sk+Uj6DOk<
zbo51ME1lRpuX#cUtM8S){HaDA-~M=u6uV60ou2Zz=%?a`R0B1inRRk81yk&LP6evZ
z7F_GaaM^ci*2moi6D3x0#B9@?Z+(O_)Ul~XV5U*V^>FqnEitL0E?1tPoF&KQ-jUn5
z;IiXp^|dZ4oC$7jM!ZoS+q920vvet_ugaHuZ{rlx66EDC_)>DG;e%quvtMcoeu<n|
ztIFy%Wl7D44_!*$7G0~BDz%+)-J~d_Y%n(^qF20?@3Qtp9sP9K0|rOh)%{eZHbyw!
z`f@4hiEY5b7mm&jccZ1fBOMmpo80J_(XvweM45n->|5@OZp-G*`Mayedpg^;8<lpC
z-HM)_O)oP^4Bc!~efn7CoLIvtcZ_BlrJT5|<}*{lbY`czvopVUSU~nFgU7DReUnZV
zgawNnJw2VV{F%nwa>jt1E=5_MIWzrO&lu?*IUUWk)NRM9{4Eb08jM7>4i!A`pKDd}
zV*MhX$bQ#?FF9c{4WgP=tL9oR42e!M2r+%rt)Q8)(%Y?9q^(zqRdGSa1(TzUvF7IA
z51G|6EILzo^ikK@i*+)N3tN<11RNQfI4q*A*_PgXB<3#eQhZ#e=-6`~BT=_zHS?B>
z8}8NC&VI;YqBLbk$^G5kQ3g{xxK5t$U0YqHvwrVsvGutRPHxjZy-l}cLc-<J35U2|
zWglDC(70xrafF5<ySC+lLcVR*bE0?cx$Qi8n_PBS#<@C{w#pk)&YqiX*3^_0tDcs4
zdsuU_noqA$%E?L0kuxt9e$foNHbv?+_lb*{mm@aK5uI|>;8k+3+00F@r*E&{QWDpi
zbWunpV~KNE;*_cWhf_I{TLOJ%D0NOh7vd7mp=8*$*+={n<FY3yx@_knb29>^1r5)s
z^VI5!u68Qb4qIlsz+UX%k{3an*0rr#y>(*!l%F56Ig@_u)H)WN{vl?o>4|0zjid}a
z%bpcdQ$9@9@M>J77!he{&9sV7TBe|E<zJ}-H`8+_?EM_!nIpb^mbmu<r8y!8%ii8)
zm-;?qXJt}D)7IqK*ZmqFY2A8w)7NpW^W<|WT+MtF!lfSFO=x}){O4qNV^Q-|+p2HL
zXXf;kSD#Ec6B@QQYHQZj)gY>N-;WSQcE4HiJyqRCCyJl>GCx0i-;TqM)w1U2hpGk7
z*N5$``uh5AZeBK9;v<-*tgEZveQjFs!>;BsgLc^3NWQYY-R%z_d#`i!oxT0&Q5z%9
z_$74<j6=Sa==c<xeCHNQOB1vf%~9HTLda<*FK?jVjEZV^$Dju<8v?k?-WFZW?cBan
zz5lp?`1-g#`f+0GxesPvU*9dJ6ME(4<z*6_NjvsWui=i?6G=^Y&3!P#?EJ}-XQdwI
zdj9;xWc+Hc_-aoc6RV!6wc(-HnhqPi+n0CZ=+Wl|9Sge~j&_UbM4Z`lJizeE>3PRz
z956`SnPJ4UiVf6sI?c}6`}p{rm}z=9?)8Oh>+2Ycxz){>adqV}iL2{)yqzvQ=4Ldw
zx@LaN)GV!IyK6Ex1}$~ibeL)S(FBoYDU1I0?n!3ZcJe^l%`)CxIkmIX4_w>2HrqLx
zOGLx3ZM&#{Z)+ZJ&fEXYM~|@2`!3G&%;7VrcUocd&1a^}9{HV&YcnQFJdS_0|Lh}A
zweQy5S`HIbXL1<$yy<DTOYBRsEfGH*IHP7tqrxQpuTesa-8#1LGCNDL^-4@!ka2Oz
zO&<xi*HRbe6+1jV^`-NWr*6l@RT`$e&cRY`YojKv$VztWjxx$vyXY9NbenGY)&Qvm
zS30FuI0{@bHC&N(a-~$;GSjTBtHV~^T9a^fMbg#kP{vif9(OlQ5n(&Hv9~!bx8sJ{
zIa37>b+#U535{0AfV8TxoWo{0C-zRaohi92hCl91&!MfkpA~|Z>SoHUi`uv#Lwv>_
z{kT2dhl=@frbcb!5zIPZkaeJknWcxR;lR<vrfVz-5)ucu9Tv-(sgZK>+n$SSbZ+K&
z=-ys*$DH>}p7fSHw}(61+)9;9EDCS$x7@wl+uz2e<3^{{R>u`tQZufYrk&|^4wl*&
zb+p?oXW?cmYqiyr)H$WLDt%1tI%=RBeC%0r>l&G3UOFp-B&JF4O*_Yw7kWE<md!;L
zy@om3VLUm9yLb#v-xL;!x%~81vCD@pjk|0TN$n*!)9<IR3=&~$Sk|-R_`=qf58k^D
zHXCG>6~vwqTbsoF;#FpXf=JXJtFsfmd}sQ;c4%a>ZHf<zY~_l5kZTyzu+O$-g+&3s
z&F4kTArVo|YUiJo#z!Um{pvMiG2<R5-ON_4DZIS*)H)_TniX*OtQ+U|rpO21CTN^|
zoc{K0=d?mYrhV!HetcJZxy8N<T$z4$+8Uh=U$pp@b}Vw%n;P0y`eIwcvU6-KB?8Oq
zK67(RZQfTR>f5rkpvb5+oBjG3P0vGOcenGb{Cv)5Hox5FmA7Q>g~`kP@Ja1ou(sv#
z60yk2M-8)0g8Z5Or)#r&+k{M8<@Kt6mCec^U9Xuxetx-YJXdX2qld<sja#etEp%$-
z644N30r$^^RG!{4{id;}<I1LQd-$Tjo%RkT4p7%!=t#u%Zkx6$(f7{UXFgQ0vQ1&x
zqZq7IE<AbD^U$iF*4MZ9bQb7DD1{X*<2ln%ec*EHYg=t0C*Jq@{nK8Zo>f!3cjfKf
zb`JOJ`?gO~vCVBcbo8Z3(3XNLZ>$sNYfe}CAKjl`;4AKS(r8<ROUi-h+|^Ep-nl1l
z>`=SU+Vy_Mv~NdPUz%l$3oVzE%NO|~peyw;JLaG*3-?yttFyQJJF-nryd1S~+ohJg
zR?cXp&W=T08k#{-F2|O3tn{1Xu=1;K*LlN~og4~HSF_h7DE;nD_mF#-oiKCdGER-{
zHlNQ;R9mn@OJHBoH+{E9txg*=7EM%eb$WW{f$x0XJDou$PWnx)Tr%f*%vIw)F8IiN
zpi5)%1i{jxqE63~YLB<Bi@D~Mu58UbdH(^$*22{jw!MAD?Uc8{k~3x2)hnlrliGV5
z<|?_kOw<r8WpmnAV`#J>ieGs_tE1pa;aM9^7JTV+Jn0m&Bte=r^H$43t$?YTN*9%u
zcb!<fi0SL!fEmHN^g^atK3U<$*wn~sDej!M^Tq-RgU-Ae8}vOk=tPL=D1iz(Zl+TQ
zFN6nOh;nUAI<PTOI#TKB>FIjhO;0a#9PLs(+|GF5);dGO&u*E<4tdgNc)2HvWh;6F
z3F#+xssyOtySmEhwRVWpn^_fGYDAsBw7YP(C}}PgVk>ysyhLtK=S|ttSW||}zH^-?
zw{DX<z`I`Rns@(6??t96pSNsjF<2)7s!J+EjkV%8asQkn=dYp5_t2@ur*kI1;{=lp
z>o+lZebO>ND#I<Ov@Njzqa@F1h8>Zsby8OIM6_I6FQC$L^x3UFqK6Z6CP>A4mVJ?V
z@*^d>>e|bf8(*Y!u&b0UoRzkGYt5CkmXpcrJ^1ZhBMzPom}}IerQv<BW1B^}WI)5C
zUmj&|oVcBKDh7wMF*lxaZoX|6bnA~*z|{%5+9I7y6S}4ca58qj;!iup!pQJsf{4D)
zvShZ+b2)OGG#{vYy->>%HYm6{<5Y&ohLD3N!u~lOI2w_3M5(J};bQU3$!wh0_X{c+
zOw>tx*(0#`)+G}S&q;SDE2>VY*8Pw?$9Vq5<O+$GJuv}i40|N(+Y&5f9o+0%(nJ(F
zPRUuaF3p>$5F{9^*C@1a#-|A(2Tls9u%1p?Tdm@fw{%nH!k7zNGkSvf=4u5e9kH_B
z;KnG!e((!xzo*G{|Ae({2P{l|)Mnf?%#Ss3RX*XmV=G&Vc>03IjI)bAU0r;^VaLaW
z;0xQTIK<vIZqC;3K63ut)4l?4$4kF>CBl`wIBZL88GV&K0&}ATduF|4VOrs(lkwK{
zf)t|;$EUF4oh9AU3r+mva;l9E9$OL;_$B7zw27N8%=rA8o%8a%XZOVT>Sph3j!{jQ
zy;~KUEqFUcg;Q-;uHw~3&5P>-oE{kc)X7*|rgiett=qRb*7Ut>DL3huSnO7q%)CrJ
zZ>i$h+uMbFTy2gxd4`4OKFoQQ#j2wez@`(iFja9zW?x~N?4&!}-m0A4Z>i*T`TnWn
z%A&6?maOrO+c-`6Qu3S<8QHlLYP45(#;!>}_u9FQFYb?NJsZ!C7ah8>^R6mfNLu*t
zU~gfZ^p?&~jmKpQZ)`EleYRuv+O^ROx1PN5u1wZ1k@Ysa@JS2roJoPgSKLe${>`1Z
zQ8#Kglj0)RFwV<&PaioiK5K(>!QIxhBkOA&ZiRO<UjN2?X>!|I*KJ1+l=S*a-=0+|
z8R_V)YJO>b!P}^7n@^k#Q{Y@1oa5eg(kSO#@98=7Iyx>3u?qa>u}%>3Dr;F$)o+%u
zSE6@TdsEWx6J@RC4?7=yyX4_GD_Py}=&47~Zyy(Y`jM?6&EQdjR+ggv>2vX77q)I>
z(2_9k{NS<W`Ek#QDMtHeb12S<Q9a#O)6%FtMg5q9HH%zFRFF&WiniM#8r&yl{gCl<
zd$4!UrAsgQ5?-<&66j6clH^nNt>*V*-_n_jGPi6=S=DawRqW%<Z+{M)<bL&K!~Cyd
zYdM3?pFOa@^QzYr)f1JiX6f3p_UaQphx-|tJanAQ5r5n2aZk+JREZ@alYBlrv~o?@
z#ds}K#zIBR_*`$F+Ksu#g4RZ~i0Q2qt<++j%B;{8P**5+Oj*ryrjd=&<5ouJ32PmN
zG?ZFJH|*RI^3%NX@3|9|H`-J?iw+x{osmAz?aGO>v(0jDzGh`~zVP*-O4dZD_UHCh
zVkag%^yy@Bh*WZEeNrH%Ir+TI$1N&cH%@>5eo~`=y>T|fwo2(W%G+~fciR}nyk8^1
zx-etr)ABv%HX0p^$dK~oj}k0)GRf47*_prUYI^95>0v8YX|1R~Dv>lJdFm5hLlY~$
zx$I${ayMmnm+id0uWF|95_7h-b~oPk`S?nlF<oYO#+8jNH)8dnjXep;lF{8qA6@;v
z>EPNX<%bJ}mYx-U#_up&W~$`Em-6ePPx4$}9m0J`B0^<mLdW8kQn`jFUlq$gzUek_
zI=;c!bC$&(t{;a_$ULmBl5N*=p5qiBrM%p4u6Bv64VTkQrdN|RgIu+8kHlWuxB1e`
zNQu3To!PquCTHeyEYf$d;FDpOSaS51ayy@lMd2Y4RzVhzl*TQtTuO7rG`jufT9q6%
zKbUHj@9`yg*}56aG`lK|m%Umem+-pvP{NkYee&;hxt(q>F>Lczo;Y*n4XGxMhe9nY
z7;}H>UzwD0a<Zh4u;>qc<4;Zn8+624w%&Erz3Ui|RdU=QYw|Ly^_5%P4(za8#E^6#
z(NRT0C712rzrV}U(vsdBED|_#=FB$l%9%|DJGqYN@P614F2KUrwDm;YYl&o@KZQ2_
zYHYcU4a%I3E(<jkjSD=s=Bx}_TB`s2^O8P?vqJ4P`Bsm5ODhh?=w9S&US##;8?#`A
zT<@-}WhZ;NyS{m(t32+rFzg9F703Fp-*%zoFI%2mj*O{UcfC0}niD#n_2~v|Ibms4
zER*vzQCRHEgQZr-6t=k4t&~p@_@?`3VoHP<)74Fa*RBb^|FrF=_^rb;%#Zzh#&_cE
zi-&y^Zc3XTQDoz2-rybar0nhO?RkfHG<pP2l3b}YRi&diu)?)lzvsk>^z9qh@1COc
zhNo+C#AA&D8Ftym(>M4IZ+m#JQ|a@GmzS6OP5b23ut=ekr~Ai}=A@$yPa?#azJ8iG
z<&E{mB-w5;*2aX0g=_~lCP{42;oAL5cFSQ&r7y``iDzve1_f_;<MyVDX{A=1baZn`
zPDj#&r5-_XK0HdkMVE4}CC)lG_uZ9;2Q1hkf^-hd(J9F3zPSB>{?VVxT-865Ec%o?
zoWvA1b4*w$?7CP<eAkKIMM+!J&ZXt$<>eT*&#)9Tio6kM;wvV{cI9Nova(w-H!L<g
z7Cn$iU={9&xWpMG5$GTm!(VQDeao50yRY%LygE67!$FbJQIj*y{ZZs3hxtmiC!5t1
zravn6tX2=0vB`Pz9*<2SlTOSJG7|dclWD%>wGhWVqiT*h^M6cC*`$78zEO-&%`73k
z$ySf{l>Z6O`=9tf^xOQOa@MD6)?WTQ{jdA){9k@|4^^Mv_y1h|`~A1;5ByPSe?Rr*
zq`hUKpxL-xr@med2T#hq-dG&Br|K(65Y(N#9vZi|>b3UuPSAwi>+9=dQ>#8@Utb>=
zzrvm+-0Y{k!F~Tr-zNQ;eev~|z%5Eur$nlLNy^-hcX_IsnYHKg)!x^yOIa19mn*s`
ze7RPd`98VDCM&dx)7GE=^h?<%zVBx=aVRt_n6mbM_^$hn?d|Okmww+H_I6f5*_^7S
zHR<Yh?@AN<BYqd%`1$cqosYln>n9;oOIGSgh8Nk!hP)3B+#UHZ^0i&_{Qmsss-;n%
zjqT#Qv-U=(6bI|x&G(gGd~)W@nbUS>zG?}v5B9#8b=8K=uG{EeI@j`3_gyM>7QNS7
zJk@K*soGa@|5vwL-S;zGwk_0gx#se$-qn9&vjrpg(pWyN&WPPL?X{}HcQ4y5?eVLe
zq#b8*YkX4YNH^4GSv0|~D7pVrtYK|fwCTB(YW5GeEc|5q{p=k5gE~Ab?v;Mu_e>~r
zc7Bv|&}(0Pp<JT{yIv}tj3`>HdoX)f<m#ozs#QgGE~#4_&k6r{W?og;(P;~h%;cCW
zIpx|VzUxN0H?_~Vv-hmI7xiRrm95DPRnaYXUd0PJcD<ad*1~pjf~`@quv*K%U98ip
z%s0!+3r~nlU2tdTbGNXfm8Lu?+l<PUuYcome>7$1swKOgvtH%f6Ufo8uzuFU!UBnw
zZfSAX+eI_9o-3rBZMt=H=d^XfCIM!JYk74GpUw$2^Nrd(Z7OG;^lsgqUg=sTTRvT%
zuCz{A`&mrps;S;>wUU2Qx*mL-r|R=*%G1^}593ai-FTy|;eT7>?bT^PqDlUjW%soS
z-Mi`3zi)Hud7&jj>w<)4ZZ;IKQ}oeD`I;IxvAvp|cX~%)ZmXKFXiWGcm!(}DEuLaA
z+w;vOH>LS%a}-=M(N5iE<af~G%@b#VMJ^o^mM6J*Oc8j!#O&14DVt`5c=exN()5Yz
z@R_6MUb%6`i&VBwiwc`~X-7fG(s%M3DnceNlo0SZ_%Tp+x6z)Kb5h!xjc=wqFIH`{
zZsRQo4_PTR^~%birLMgz*DliPeH-m&Rl56`*o~X1vz{LNYQEek(&o8ozpYc{i&KX^
zex`Z8FqshOeM4!cmYe43Y*Sqm-hi`~o6Z`Z5=b<3GBM|xdeYEmtEr0bjGf-ga%b#F
z`Bm8_@tW5-Rd>qPSqUp9&JMmQ*V!)nI<+)uir}K=4Ix^Qe4Rh9zbFt9xE2zX+!@Nb
zrm~T%=90)`ZRw_otKFWdwz_;<zw!RAwVl47n!3{Sqs84H?VM;O_2t<Un`t@IXB1w(
zW8J$U_|6KhWm=bA%lbLqZr$=SuchpmeR@uBxKfT6m*iBPTuY%<t0wAh?DBCGyK8m(
z;_P>K(pT$#e4c!KXQboA(3m#O(5%~61Fqh@bacT|rCgmauQyM3KNx5<Q7lAA-fG80
zrKs6b(<VOnd_G7|^!uyO=d0#L_}hHi^ZC5?d7H-T`R9$am;1b4wfWq>YfI8&dtR^G
zT~&I0RUE&~ml+cuI!*Zfe&6d$@pU$u&*r-3ihS~qxa0Fee&M1jl@%YK+)4G>+;mmI
z-l%kn?yY5g6Sivl_+*RgC9O&h?DSuEEkdwqtJhSms4E(a)=miu^S<s_5X~Yg@|^9r
z=i=b5?lNhOrvj}afsR6tLv(JcS`=+qv|yFNk<EfmnqI=JJz5KOc|6)3LXO-{sahR+
zb#1qIZfNN0(9lBxdkdFV+49O<Ua@#?bgV^P+rC#F)6Xbhd4D-*ubJwcY)O5iI>`$%
zw@#WCyjf<_IeFO^4R<}3+T=rX>yAB_GF4uy)Gza3YNXOyg)-adbH}H>Ejh&Ce_hF{
z-9p3T`P7xl!iG#!pG|s{*vENGi_<Ck^5?g1zO&vOzaJImXRXC{YtfM%2i>a<pQsnQ
zJZIBvy|P$anVnZ!{e_;^pLg2r@KaDoRQ8z99%tVhPh~__PM4GGow9L_T;*fyYguuV
zmuE+%y#e?3+xC@8{@LR(QRU^AU%A@rPL<`{-16P)`~7<T;6+Qf2V9ycswsQ!)d@+S
z=cPM)bsx<a%&Yx==IOm6+x%e7>#@dFCaGRiwL;Eu+fTLH`g+gz0P*#4SC=3Bl>NXg
zapyrpr{;?l3m>25ohDhYyj06fYirgAi`~CBrux)d6;6&63H0L9jSF11+J9P(ob0#Y
zEkR2>B6Aj>-LCKL$s(W|wK*6xLb*%mMq<RZnP-0596c@OGd*gytj?|;qd?c)paHdN
z<HQy{eJwkubLsjTJ9ppi`L^@wjv4FUZ&G@q)TObg%T`k$DEp<@p>|cfd)H+yUH-Il
z_MTYFo4<LKw@+AitoX2@iCz1pPbLv@4_<Xzud(zlu&RmZU-v0LK<}{f#r+;$yO*xt
zcE~Vv@x@oizVTeG_cPM_)Lm74%ggsv-j1h7Rp+vH1@l~b^DSED?Y?xqZ?ZL8cYJto
zbh+RDi5G5FeB7NPec@%sXSHYXYftW4SbnB_$2GCylrtA~MT(8$<}EOv8u^;Xb+5^s
zN7sFyJz4e3eC>~toV)v82buBC{8T>aqvXn<GH;fU>5MP0ZDw9FVb(@NIV~>54<(|j
zf64sNtdBU<6>-%&^pb{a&P~<UYVU~6`rA#8YHTxNb87#S>a`_gmG-oyLQ_u6p4`zk
zt!49vuKl~WR&^aY;d{Dyx^m1`F8-YbYcF%_Y<~MB=dt_k+Z(rA@$tp^PW`bcX=it^
z+8GW%ule;s`&@Et&Tf72#W=-#W6$GswYGUjuWWkyN#oW21C8pQ_MSgOCtEe?yuWiz
zzj@lNr|;(OKK)f_@o`Jbof|F}8csDUdCgd~w`lE>-s{oj+$)-LVoa_Ht*!ao&w4J5
zO-20awvhdG@-E+2X|6sPbjsL9l3zwFXittBcaM?&`KX<K>+P7GKdsoB-@Njq<!SYm
zAuFe!ojLQ&<gTj63znXVSA20IZf}>ZXnKa*>3|i-`d?jXiP^OJWXk^~%U15pU34%}
z(6lN<H0-qFy}mqV4pmWq|7hpn)@LfSB(MIQ{=JCz$i)88*OzWYG)6T}jZ1!|c(Q4>
zv^vjWo_U6P-!{lyt%xh~{;a7xGr4GCy!(p6OZ?g=b8@G9@|-+!All1S>6*^+rL|0Z
z&CiBiD4xDhX4B7z3;AnrZ~Iluy>X^-*MW;tlU6_9b}%4ox4~7@dBO8Oih3>2<Q9B;
z%8^A&WmS{6B8z|{Tjm4@@IosNg&+=vK;IsdSMBqKTz=(mUD!LtCt0DIvwdHvTkZGf
z>N?M?HEcd#j<&Y?e9<ssr9tTPf{T}p+s<7-VZ@VNDJ)^;$a1~lk)@$#VS!5ao!BYY
zzFnzh`5b>#bGvkJTqs}j96N0fMHic=Ui0m)T+>)ueN^Op;cl*cZcnZ)v#WmIsl1T2
zY4hE}BO6wJ4CC6T_4T9c^hjN<rE}}HUcL5Id&8T(ZBi%h965Mq-IUjvGi9E$F5yv{
za@6o{dW2-&?V!VxGMDYk`*Jy9xqqXQv)bB%YOhnR8!JlIh}^ifdCvALo@f01nj6}c
z7p#5wQ_U)4Ud-8DN`Yq{uAa(e+@7ki&|UUZNRq1AVji#QuQVR43)oroR8?VFx9F6@
zi`A{Zm#o;r!Y&HU7j;=$;T1IdXj0avOG@FAv%c((c%5+3=Ciin9@qb=7X$rQ*NW|7
zs|vigZ^qG6wHlw|lZ;O%{CL<pTiv+(=&?s<yf&UJet$<lWY?-4YKOaSPdxhQW9Y}L
zHU)2U>Lxmi^-rFtEVbIwZ0fz`d-PU4^$%Yk)q3gM&fDjv>rJZR@Vo|^EV*-X{iZ6(
zUXh}nmw(=UVd;@J%emce-RyfS`dqkkk-4das;HD1XSPwrRVCwnX+6F^|4gUz2ECiU
zIDN&XS-rD7Ek!2A>|{MXUGL4esh2$_ipA{pOZ&LmqGHEwpRyfya^Ae0@k(Z2&Bf1a
zB419oow0wS(2~&FUAM0kK3DZ#8QAyhXsf8}+^DUkeXHJgTrzjM`Yk0+JX1>S{fXlp
zm4D9cUg5WDXHn?s>3T5_k8o#(t&R>_F1pEky7tG%AEV96RtBkCmum(+&wu2|>9_gR
zL77@b#fc6xQkiFY%=j7Dc}lRL|3Gx;(dj$pdb&v}bA3{BnXogWCQe(Gf5*z*&*v20
z`eB-T>FSB)=0)n_)xJefj|Zjp8n%^*>)uzBHSfxdTmH!W=%LC}9ut-Pj&1tGe&eS9
zQ|G{-r7x$8>%G2yK1A}1j_Fa!I}036J-8#j;+o3uM?bpWe>tUb<&V$K>DTAB*}N3I
znHsWj(tft!81^boF3rWCUm8q#KI`z($;aN$c|Cc;_fF%))pJhI+PqYD|18j~RfK8Q
z)z!hXwq{*j<@KvOUstOOw4C9s&hD%2tY_Y6Oc0ApwR(KzerP*ubnP3N%TrdK{u5xi
z`d88GH?KCGex-LdW6!#@Uvo}`-(nM=y<B}dn~jF>gR`^GL@Rf1F?*_eIqh`0sF;31
zw2I$c&!8=vpSWI)C}mk5Y{{&-ENP;3#$3&sIX?o0uNEGE_Tk)__{sLI_s+c7^UvG;
z>au;uUd%WXT|D{M&0l?C%l!_z+DsOI_9Ad+)MHPzMXQoBFMqzb?MQc%xWZa3i|x0|
zzZa}8o}nE#JI$A!Ytuwm;qpt@*4(-@cdAL!iDMe8x^EiwU3U1kD%^I;?do;M7KieG
z)-3!T{%YE+%w;nY+oVoD+O~b}$=8B~Tp}uOL{r`wA8vak>$`cOQ_E-f6K!*&t}b00
zwe?J8q&H8rzqQUY9*L}4=U1UyPnJyWGXEaG)GJQ=&EZAMcm1D!Cv4kh?XLlgFW=8N
z@5va=RHLKJA6U+MOy_&ef}oHJ!=jnWn^!!Vwd>Xu#jMO2(A?SOTUVPC_p82U{c^?X
z&@KJ`W3L_89PfJCQWf}6J?<yV+w<H030@YNU-t9QGC`ghkJcGR_}n+2kfgD&*0l4~
zl009>*=(Cd@2{DV*gxgoC%H>q%MS0;F%#@=zVLYS!A8L~f0LCh8@XbZeF)iBGPh)H
zX3(51HH-N(%O~1dxxa~Dms>F_c&ANCpj*gQE&raUJImYVyX`VQXYqLYMLyA^Q!9n%
z{C)jJY31cr3#Z;SJ+Y`yjA^R+^aI~6NA@i_l=QW3t}5qs-8UCw;=E32s4HJRJ9oj!
zTZ!K;yi^QQix1uC;%r;CR#5rm%9<OOs&;MqT(d18bMv=dje37~Hm#bsY|ZP!V`*8-
zZ4(L-&oc_#t9a;Sq4jl}`EBi)x7YHVx#~OBD6lo-<Kkug(ot_`xp%g;%GIs2*}D5k
zf01Qibn&w5S1%r{nf6F__s(nn8-r@u-bd`nd1%LSOj~(Dv5IzB$<Cl9t0siZ4ZZX}
zqChHh+3ZYJ?b&IwGM7z1v+m-Oh%K3skyVSia@X2e?Dm_oFi?A=&7Y$yoZqRKs%C~3
zJ=O}i*ELIBS!}+P9g~Fl)y-?8Pk5X^xZ>z;m$IlUp>k(`ZOyv6I=oX)=+-MCeVZ?z
zl+Pp_RJxvi>Cl;!D)%?jo!hQ#_KVqe<;bc_cA-hHRVRJT*zDcet2eJ?m&J|qX%@bf
zn(UVojucJ&Jbm4}k5zfgZyt=6dVHkt(YiTLyQQ^H)ppMN$K9&4I(tQC?9E<D`$#+I
zV&#jqldrUV?mm4k?&aBM56?%wN&7HYin+WqW^>e*U8iIv7fa^3gl+rLQG89J%KYu@
z$#vIfnItPO(vD8*jE+<n^yL!EF4=J9<+E3&CQnudE!}7F__Eq<UlGyyzSHzdmcLug
za!>7p^Phb#*Cteak9!hTI;o{b$7N;EQg7)kyCo;<Y0E3Qo>BKLoNL>5=g(iBDf2|0
zFAiL*`ZUp2U*w(Jw&-hIMn>Lc8{ULxMrUM6+|8C@dHLB)<@vtE<5IH)5}!Uh^se~l
z(&^KWmGko#H%Hlviq2X0j9owRk2$Z>_L~NW4TBTHk0)K*!?j9)U(DI}oQY>n?aZH*
zuBpMNm(N?c>hLk$hq5catZrR$Y|{rf_q2-CEhX2kPq*d?T=ja=!nLcJ?xZYcJNn3F
z+6AE?G4D=YJ8NTiHL;C5maBWNPu=q`a~+?@1g7<E4m$cKlT1o@*hLp7&q+PhW*OGS
zY1uAm=z3|Qr@AnAYrX2N>PXAVPImq1lX`#kYJI)0MPH9{@vpj6%5$VYbYE@nsU?Bu
zSG8Tb`g!5XiDH_sg%=#vnKdVJQt78fE0_J+rLp+bjHXtxk6$m%UGL|lzh;TYL?wrn
zYmz6wF>lv)VyVf?wVZa#&{ZT-%h9Qq<KXAE4a+yB=`eqm);PE1zVzv4saAnZ2a)}X
zA<=@8nGTz4UK?^~mxz6eSk7vF<o=_VNmiAU<z{UcUwtI?()*R?G&h9@ef)9J;PRqL
zg+=1iqIA+^TC2RfEuN>%%3LGAdal!=e`@aySp*WUt`iHoBl9M0?d|;%p9`vNrt4j?
z@t=K4*DcSm&3pOEnf!Xrp;LaQUSB#bu{U#>tH?_xr><?Mjps%4SS*vYO*+GI%%UjI
zaQm*J^&9&9f?j<WvbR}uT4%>|of%hGrLsFtH;U@HAta==cx&svr?vaEtFHu>&JXmO
zBiqp}92KmzDn~U_NM*_$&&gSwlV`eQY+_#VUinGnvOd?;9~z63jhd!vXN9bGS+w@v
zv_$V5(?+{BVc9Y(-KDgxG7fk0I(6A>mN;fOQ7PoSTd$N!#>K^My;8;?k;R4+m6G^g
z_xH=I7EfjQ`)Ji-uFYz05{Vspx^XNv)$X!0{r){X9<ttgu9(i&%%CZur`03^qtrIc
z*1dHs?3hN_+E*zvPChtzC2H#v34{DC8q3<w)v~@><vndv?C<72i<hrdB`z$Qe&(iK
zG1rv*l~#MTxelkkeC)R5ZS)q|-DU6D7`sID4^B>XVLfiL?Y3=F@#Xs)%FVWLuxqtt
zrcS%A%(Pv8x~NjY*H2Gx6mEPRzTUKTTG7lk5ywt{S()&_B;$iayx1l^b(t@5@8iop
z++7kFQ2KoRv)R@CiIdmd=x4FIvCpsO-XST$#NDjdi<WAv|E|<_^zaquJx-gqEcH4O
z?oqu{@{L{XirYtK-(Iz9!>l8jbB}%QpV!$DU8AG6=K2=zRNsG}7`}ARmaLsz_V%Br
z_iT~*VQV69zM7)Lcj>u}g_@T^;Wm+pUjj3p#uTts<@|CBDzjO==0rJXh0h6XAJrGv
zxHO&?DJg5VwrEDWO(~4%c1!xEBdzB@>+ViBF%?^vTSqUR3#wXJ^&r?}=hIWUnZ}<}
zl#~MQOe%A_^JdZFeBIN+la}39+`Vu?fBD9(w~qy#WlHNgw2I}pta-uT{kPdaE|pf=
zIyK_b{xkL|Tb7;*$-8uQX=K8)m~E@3trgUrdRWU?*I>o2wa1KuxK^!NwQ9B7DziRw
z-7Q-g1;cc5w>2v%DJi7zoH=<&fl+{=!_vKI&bl|&^McR3wq6&#{Bz7(E5B`3S;z8D
z)rp;Jy1YI6!@EBrD}M^w{J${yUu9B2#h;W#pUzI|tGhFIzy71gpWT1${QKX2-}|{$
zax<R)u$U{>?KxFa>D+r+J-*`FdFMa(+kCzyb8i}lXqfc6XY(q4^Z9R{u0G}6gD2&G
z)2?Tp`11R0*zVa2d)MC)o7E)o@}+v){m3)D_f-|YD3w*scDwOg>(Fh%T$48&lVdZM
z<vclcx9f9^Jj0pqHt(C#iXP0~^+HwTVd#_~ZL1}(7Hm7T^ZW(Nu!jo+4SyV!SS*rz
z#AIzq#jgtA7vC=lEIM32Gt+JP#vQlb8<`9I{QPIr?tj-8KC=wDQza+g`elFK{fmEB
zKK#6B@vj|W=a1L@Jt{g+>uKA@s6PgG?z<Y_sL)z?FTCwm?@>OfYbgiicbp9M_ig<;
zf5qDES9SJZ*V%MhS9_L`-?XEzox`?yH@3g_F?&{ibo1L(S<{0B2|87QVTw|Lp8HDF
z1Qs})w_L|n)a4hp@FKsd^&e@Iyrnq}+m@;KYQBE7rfch}u<tjHt1LOW_SaKZcU#t#
zMN7^7m)Gj<v#Z}Y(=vF)_mr*8nnh1fPkkBp)G+DD|EWS>XP3t-g`YMy;t4%pT>HK6
z?40rwxkqz*rfNT5dHEj8sUWWvuY|>fSzJy|nk&01LM$ptW1;ERX|JxW+7gncZW!77
z=|jeH&8Qu|wy%$!iB|;;q^t}&UX*t+b^6A+t0GfZ32B~dy4P2J&O34Pk%KFz-FLKY
zy_@~z+B89qz+5MlM_<=D1y==4{+}6j`BI`w$do^(=dYwbd^j(Q>*{L1n@g*`AF3~!
zX@1;RD`fl9+Y`mIu69p$Z)n+~{O8-+V6`b<T{ca#f9w`FOL6<^kAhjZzCG&j4N>#$
zwGS)$>GkpIln=jjrMZ6X*Ot32v4ZDQuja#3XR0MOJxyM%FIuUUvP+DeXN$(%Pky(o
zxfy;r?fqtFb)@T^^6~S>vXk!?`ApW*D@!;K|NO~0*Qs;2n0a<goM`+rb)wT3D~W)I
z^V^e3BA!1!ky@L2?DNf2l`kjjg={fiA#(Z6-8EkowtakZQE~Cn>oeYd;!oRh^_fI;
z(hLWW9+p7AuBQ@`yH=i!Xpy?5d3~KDtM}6vDx0H~pQju<@$2~k%h%ZzH_O6SH@n;G
zsc=h&-gtT1;IN@!-btGyeUkFWD()WsawfJ`THI29Wu|31Yw$0r$oy})f8{sNTH3b!
z#^P9)#m}?eX1&~C-Mzi+jrDiW+4q)RZxm8`F_EuRNO@_Xl1s+cGvFaJL2uAvm`xKM
z^R5M#e%mN<Dc3uV?GR`MO~68@ripnI|Jn6PXGxyWpQ`mgF!uGdV9VKmT=X|_wfz*%
zU+LgE=UAn8WzXi4FX4CM>l@l^ANfraK7BM&-q)o)YW<7sTlcowdh_nBpDp_O=&zlU
z?5>HkO?mJ1wcnYoxX)*r)~2MLFKX|WWG`rv;eYpO-?Pnojd>D#qtv|Co{XMo_cA8=
zyy2;-TBi^CoOL@JZSb?LX!DLv!@04ScW-$%ul{D^cZ(g_MH16utG4zYojGyGPvw`=
z?kg_@<n7wr=c}O=s_-WDW6JVLnH$S=)&wk!S$X-|^GA-M^JZ*WJ7bAQVw=Q7L!Wu3
zbKX7@FM9Iqd1?AryTd9cCp(+#9oZB6?cH>j3%Xx8dbyf@-C=%KDIUDa*NSt(v^i@P
zFQ3*8YFiz&>Uv(u-|W9$+U4sv1ST8`IwEW~-E_-wH|w_0=w}kUw%Fa8!t9=Ps-Rc3
z>)z*IyJ~sG-mS}sf2#eeRPcDm^u{%jAN1!u3_elLdUuNO-a{8&78y)UI=5iG{!FR&
z!d9ZUE3eP_uc+djer{WA?z)nneHC`^J=Jule+xF7$rbF~RV1`yd(Axes43y=a}M}3
zCaISAH{_~wB)Q$Zl$@=5a%PC3@4{}EEu8@cSMHRiAJDfpx*I7TeP^Nc<Vz=>u5DSV
zbZy6NPX!H&<L_olFK@kg!}i$L_hy^fCjIuz($0LcJbB$jr$3?QHzapnF}q=ESY7q=
zbJH<zS+$avPA|7x#?I89ShHN|Nrm=8yNY-F)x`KpZfr4(j(g&A`|9%A?$TK$xe5E9
zoq1^U`Mk~N%RcjZ%D5sfgsuD^6wV%O8oFbG*z`0P(Y71AdB5A$e0dvMI`8h|vwOm=
z=Ip8AlKLjLxBt6Vs>|i03l%q(R{oh~B_Hv3d)_PWNlTSLiSoDmihu=2Rt7Egn(}33
z(2794wJ~O|w4SWpTq%3~?MKCV*SNFibei6HI^$EAx6Mr6g(;_M18-HERhEC9vgytD
z-#JCOk*gl<{rk<QhAV3Ix5C{sYI}I=Tr?uHvi%N*EVG+ep8Y<o>RD%HN$i@bQ*|e6
ziD(2Y6y0@Jxr{Gie!`!>uB)}KR)(z&4ZL${saKg=>v?7Q_W_^e9188Ee}2Aoe*W4E
z|0HskyEmm9Pq1h$G~vE+EZTIr%wgf}FA_H|J?Hb9x5h7dsm-O0QnCKi*F@!p?y@;u
zQ$5q@#foL?j>~NIoB1w(D#xMIdfX+#Gj8qM7WI3*xnbcMBR4yq@;~2-4X(a)ow=4P
z`%dMDr2z(~@9f#D#v@@6wbiPq>$$?kd3xerB3kdy7>1hVJoU1BnQE|b*<{;0uP?3p
zCVeaW=!BnK>lbW#Vx3=ac=br;k&V%2pR0X+LP5p!D&tifa>M2?oqBev>*Vjxc9$+Z
zwnJk5#8dAK4$NKf&wb+5wx#8v`e8eRyB8O)PAjR-%eQ8gIU7}wR&=%Vy?%~*-ma71
z)l*gnEi(EY^j34T{rt9b^G($zEnD~E=u3$-Utylso>^?HzbYfoP1lK>cP~s#S|dZT
z-(BKM&&`*Lrkq)mgJV`YrEPe#Y0kD;oGZ2)_Wa_xWMlCx;^#z>gBfo{PRv*w7<i#C
zUMky+-Cy$y=ewEuek<<?`yA`<I#)Go7VmkVOR?rzg03Qgt|Ebr&Fa>h?^c$cUcKS*
zqK!+hoQ(IVo9mptVrjyhEA3NezS0+3=;w5K*8X5F(PfM}?OrLV9@}1~mCl;`RefIX
zx8Ui$8s~I&l`dJEBrluW70v5w^IhnYM{%C?E8Q-wO&KP0RzJy-$$iIHYu+#96>G4y
z>-|oj1pBb5i#7aX&rU0H&(<(3Y8CgsbjQ_2sBf~&Sy!%e+j}`hVl1xK&72ZeWo*VD
zwl+3mqww;ubpq4Ft_CMNtybN$_4PfAu%+wRw!E3k@^bBte!0o7C2nv|Uk;s4j$hXE
za+<#3G@Y!-)R)<5J_jDJSaey*RN{K{x=Lr!Wqds88#Zlya<;JPknhvKI$=6R^KQTT
zD0*+-q0;7#&q`~)oQ>L=b!}Vqbi?U-d3Sf0X5}R(9;xiu_;O28Ttd>E!s9x2lZ2jZ
z&|AuM%Q>g8c*BIcX}_6jpPE{TuMP>YTlGE5Z^?|yzKJvardS#AL@V(oB$s<SddnZH
zUFe?3DjKpiWVPgyinnRrh8KM{tn{cjJ2ys7BnY(5((l&Mnd_&ni&=8?s9wxY@B;0N
zuHcPbuhzv>z6uNS&rN#j`D*8k#|o#OFVtLS#OqzNbNAQP%~c(R2A7t4P1W$axo!8)
z2R|g9IVZ)=e$%zhfWKJNpq2NTj>L~nAr;T=S?6}SbPC1a%MD$%m8olgrDKn(n#P)$
zXQs7TADoi%i+7@+;My#YZI@Jg?F0>{y?WBlC*{^F&0BCLb@{rdQw~IL`Z%*_j*p9)
zlSaTor(V9=8<VF6T$6MLtzRh@Pdwbtn>fXy;NjtR-Wvsv4o9T&$yj8wdaq>RKXfYL
z)`o|7G?n(fF{*q%>$T#(AMPi=->>WKf1u(N;mReV;S;vrqw5;StWJ+bU7C|OdUR>J
zWE3PjOxbo={g`{)eJvSIPa(gsz>usbvoc)+PoA!u_IS^;H%q70%zn8k)$f2)<V4k!
zC)d*+#T+`dC`w&v?(Ll?uc^gd)|PI%+S(^FucUH@>>NHm&dbl$Z<a3Rwz=2e?6k^Y
zmSvVm;)#GJ-h>CIC!Jbzqv&2%(WSe4=On~zbJJY<LM$?M(@~vGm&~s2yZNN7dTy!X
zl^c>}*WSD|^_s@Q!Tt74V0c}a;?2OK`|h>16T2&o*=l~QO^m*N@19er>A8)oHXoR~
z=#tJW&(CL$nB3;?PuUh3kn`k;YnRcM<W%oDZ}00mCx@=yQ4|(#>~Hh=;;eI7Z(9=X
zZ`-%gx>nn~b#2dsXRlTr&pE5YC2`@*s?5sM+@`Nup3OSRo*5#m<s7nZ(%!0fQz|_a
znm81h;)FONK?T!>ij|@nxr^p3(@75JKWDL7ady(un=hyBiw}Bj!sqj3)4hNTt9y1n
zyj`>HmX(G=li1HW_57|a*F&@=YwZlrF4uajvc9^u(y6yov_F4p(vzQpOSZd4PQ5-S
z!|v2%=~>0Iyq90Ml3jG+z^sCXFJE6xUmsVv(0a)-w#zGK9980p*qAhL@{{hHVlyWz
zH}UivI-N3)HB8Jbja)6iV%>@5eJQUBm8|>@&P=#^ewur7)w_^oLHSXq;-Zf9Tro<y
zsv><&f5ww(`;&sEeY(BWpG{6N`}Xm;y*?R1Jep2j>?)zWoqsZx&pW7;@oA3pIbU7N
zb3yK!%Y}1`D_^ZT5NMaR%188I$+^bVV~zFs!R4pkDk>hj6W+J#=DMHV6{RwYc2{Lw
z4b;v}i2CZaXmxe};`0;3p5A$D&bG;S`pz?PAFIAzdu?)P@BS@I_lN}ca=K*YXFX4G
zespEhS=*GWS)Tik>K?8>9{4aRy5#OTrrBLy)-&6g!qwfq<D{qiDJ8ElXZK%t#oE|x
zUca@;pPie{cDHXmu;bh^-mlCz&zkwnetOG9`l({yRL%D`*#X(DS6&%gTbtisUHkne
zpR<z7L?xGw;vX?*JSHqWY_oQ=>7J_SxaoSOv(3_7GcUM`X)JL!?E4;j`})IwoZVjK
zOgG){-9Gwg-A++s3*P?5#mO6vw&Wan`Cz%<oagN_@6}A!r5DCrdHGP?Z_ZO6zOc2;
ztUMA1dtRTo@44CUW4w{m*LM*HiPo#0`5Einou_VpBJ5|>_VtNQ;haabe5cwi`syqF
zEmB9q-Q!@J-}&V7#k1GW-L-nnt=o)Yy{EU{5D7DkdA(X?aoNgKc}qijkC(i9aeUjf
zt-0s!Prlb1Whb?@G(;nyD0u2yFNc-k`M*t96s&x%6|&Q*m0dSy<K`3zjhvgCQ*)=8
zo!J-it=xX%*3-|vNNsI6JKHALTRbhWck*Ut+32MuJDwJA5kKZ@mUH{&=BO{Lv({>?
z`7+JLT2*{yYW9yGl0KE+pR8M|!nwP3SA^*#h0pV<o*XsMEGpRe^V<5!+oQOu{tJp`
z^GF%x+}Ql_hQv<ySWlg~UEQ%Qhi<23K0P)?(W^BpMMH^mOTOvaoMp~RB|USee9`!k
zW7M%;rs*&z+ug_&lQ|rmgwISmq<-W;$Hb*RY^ukrylgdJxHBE;I5AoERLQ*w%nLV9
zxN$1-#(SO2Ny|8jRaa}Bo8vxrrd3*CdV*&P=f)(zSvqNwLRyD=b%Zy6DVVQ*PNYJS
z!-dglVH5k5`7M8*@K(OaU%lXd9s8}{Z)Z!*S!Q+e@<ct8qwCK5cF4|Mq_HBcFzuMI
zGLxdi!WW<9MOwKeZC=dnDw)#w%f0dJ;>sr$E+tBDRxDL&@;cL#q;g~OCsn(ZhA&#>
zI=XVU&Nk3uX-ir7;84qiv#nB^3%VBmIHG9x_$|`}@22EuPJ$lu1n%_nzIZj^p;OE8
zjwh4Drpbi!oISkLB~e&IDy=}~<<YW<C$DzM2w%Ru#aK`<b=4vzWkHE$Eo~{BJ~z@H
zZi^9Mk5ca{e9rqbIIKwK^C?L!=Qxcs?hh)r@Y~$VxRh+~%*WQYW#NuT3nKajS%pKE
z_%w+u?^V1Y^M$u9Z&Tr>wXbEBLYB6sE_EzA(r_$s(YAv@X%D{n1v(yn_F6{Ub@9~I
zv#d5ux$c*=Y?>s?@33h3sgAB%KJ5##C%DY|)gXAHV4?>@SL2$Xq)RiDSR5btSoR&x
zvCT-dYs(Yd`t$MJ?50GKmmE9+ca<iJmd#Jup(A_P=c{yD!c=eL8S6eyw^>*CTte!D
z#sf{R8AlHY^(Fn4ahh_=={fT@aVaac7iZ3HzaCq9J=WOVck6Wab-`JmzHT_N&@brK
zdBu>5vn@~my1jJHxLTdv7oFxkT}W!{L7iO(Hy@Ygyu946WCyFWOW}0s<C89Ci(Z)Y
zv;O&i2m5(P{)KJ~WsBxIasNg8%+5crlUDI;6P@@!|9AaGB~3wp{uQQ+XS(XLF-=*T
zwX$`Fppc<TC{xjs$WX1nlkAJdi!?tM{ZW^%u~UEh-cE=s{Z1!qp{(60p-o-W?nIbr
zX==a84&`~=^+zW&|JFu!8-t#mod-L0d$T9WWH09XGi$oRB89zz0!IY}0#&%*P7!|-
zVlkzyf_a{?jo@mY8#%^iQX9^;9X3vcaJ=12&Mv!boC)Gt?aT;nQE#c*IpuST`>MH3
ze}kSbQ#`fHNh5z^ilApmXCMrFT=vqq%&T$PD`KwprTa~@=R7S@KN)Vg@yF(mIg%~Q
z-dL8$mjqwB`}CpjfwtZDef!ITkMB6UVOLCqU)N*5v)##Nu^NG=?(W}f-1s{y?fe3n
zSNAW><TAefiF3t?Nl_8qE7O<n^_aS_$iL~dWm?F}l|Cz1>zg+I^HF+!w!3hi;peXM
ztaXRC2&?p~>`e;_xohmRXPL1;*w-kTqvv@0*6f+Q;(Womb5@c4DZ36tCAX}w)_nU$
z>B<_*r4RGfj_%)5#B=_XtN2R&Q_;IKw}<@VEDAbN?)vPIYo?G$)$3dVN0v#aQrq*g
zeP3GY20b-%*vb_ocVor5B5$7PviTc%0vqNvi*33Xx>?xQdds(0@;=v?UCm^tI!dmP
z-{{I5?3u~tdDf8oX1EaJvZ<PLqKf%K!o;1HDThDb^=#U5xp#I}p)civ{VSfc{gjT7
zkq(uAuJ!am%HcaPb!Og~;TI*YFK!krKDa(`kCo_Q9}B<S=b|2qxvpu)tPNT|byDC_
zndVjLCl6lgN?N)?P3z5*<#*ant9-unWU;fNV{4+8S81`SkVVDIazU{^ZzEmqO@7P5
z?D(87C&$z-xG5LpW)PEPrq|%<F2<GF>e{vP>-1LnRl(QHUN3iQa*mvK<&^oC3#UCj
zmWa%&x@o@B#0z}hLEX$q{l@PyCE1r8J^f~y8_qK<@YeEkT6+AB=5FPMg?Hw6Oxb-b
z;%a1$(oEKO-M+JCnr(fhv1qCBrNwqfv+l95oJ!F0pB1Eb?wH|})0fs~_|KadE4a0~
zxiwX6XTah|_7mciK7I0@_9w#WW9m!cP+y6oW=<j>#NKE+a~ztyYQkEloe!4y7)VUg
zD(2IS+nTcO{j7{iqxTzbZ&_QJ^|NSGSM~BK`lTC7=H&!5w-nw~U*>br`UksG3it1z
zy9K*us?C{GF)3kvRCH?3P19_}(3&~E$1@x>t+q<+IR9~t%G$|IDG?K=ZZY&TOgApu
zdT#QSWO>t&Rp;kRs)jW)UB8fUO}P5lw8uMpGmdwg1n)0*y0g>#`}y^vR}LJ!eD}?%
z=FFWl{v^k_b=}-zbXw<epU0~@%O4Yh_n8_c2f8|Y2#L;iOWtEQX=;|xp&dJ`A5YTG
zDiRcP>3VRe%jUVx@*QPeZfBSIl(zQD%d)Rp>63LeX8Dg<KJl*pLE2thgI^wuc*n}g
zk@NhMg59Ev?W*0yMN3m2M=Wp7UR;(Vv$Jan`zFhr?blPY3vMPKe-eA*=Bt=mUtY`A
z;p=rFYi_%~zP|ob=bio92m8)}#~7Y%{^Xk<-g)d)qyF3{Pm(;I{YYPTq+guj^iT8f
zQi}^l`cH!N!@JH4r`x=H{5fg$_Ja$Z?aGtgHs?%v6ZZJ9M`Wqg#n)eM?T9*Rlw+~(
zz3#<pz3%mVb;@U2wx3K^$};`gC6f5=zM*!|65ra0Fp16U7tQV~JG#u$IJl@_(z&Ru
zcjcU2gyUBPB(-I0tFG8JRc#j6trzPryHANs-H>%%s;BbGT&sZB+L1-;Cc90Va(3S7
zSx?)7mHR9ngVD48hZ<+M-P9_*sHD~@r2JxfZ|Ko(DN#S&&mL!Isyw;p-Y;jZ2AZZm
zrMbrY(5bsgN4v!gw@Zdp`%k#uGuP#HsM~9)Eq9Oh&1U`g_R~D&-)}WH?)Wq5L2TXq
z)QdO7KHpnWDs*)DuU(xo;-}}<Ou2DmM~;E2nOjNBSrNnY!IDqSebW5ne%YB%ILDZN
z@?dD5xuHpq%FD!n8)r=}7xfsNncQn~&*aOh4f+=xSzgZHTzv81(Uli=-ncC&C|i2H
zOuXvluDh|m_Y>386ImiQf|k$S<aTYH+tSguNJ&Xh=)}>PHk&SZB(^R{oL+o$&Z1{8
z&y`*EoY}m(SYK+Dbzh6vn=31)RX(3x+?;Wk$2hdMcQTK`>5ONeuYS01P;#`iMpk6=
z;=MI7*?~ccc26IxUX7V+rf(kg?^)5YndZJGmwo3?UAA=3wQQAZn@^sbSaIF*@JsQh
zns(9>8-nDvcON;w&tH7vcJD-~qSUm&of)~i(<iE4y{9u#;&|+YRlnEm`zILg^!a>7
zpwzX)F3GZGbHmPFkFUEJcJK487*R9Ep31kpzvWjX%a(B{G^<=YQTt<}lFQcHpRzVA
z4{Z6aIcr|__w(hSuWfd2lDhi(G^De>Xm8cmf4;jeB<6nn$`4ALzs~&d&F7we;$G)k
zeY35(a`jnGFAX0(nAxnJSZ+U&IhxDeWw+#-&mX_{-K)z9t(a|-C2hQO=G564>Pg#v
zNjaTfajoBZ+0{eWG$&^sEw9Y$Tasln=jEU2GhXITI_7dE``wDb_N_HX)~uRn`P*QP
zVS2o$xy9RVpQ5y77Y}(`W?e5Sl~S1X*Qr^=I5JY#^~y?<dtujI>lNi3Q>PtV9g}7F
zJgDTQ^P07-r*`&Ac?FwQ>!#}%&6FuT4nZdMoNl?>P6n^*6g2UeDDSc-Yn@KSg;j<6
zv)pn{s`Ean`1p8J3g}FuMYE3V(ODNQrW3KTcgv~8w`!K{S@KFfYL!LA)n)H=N^a=?
z$e;32{(h!!@9`(%Os~KF*6Ta4eco5+71MO>uU_$ZurKEIlV!(_9!_l3++@7yS;oS?
zsu|}scO@HDeZTl|=bHy&HFHG69_KBa*fLi~OImW-th~TB$w`IFW*?h3HH`Hahr(Rr
zZ+Yb#3bRTzPfP)g>?*m;F@0qCD&KznrP*dirCK5DVs@4+xiiNtv&Y~_M@L7CtI~l>
z_tHMEX#1r4UD4{q^s~+yUkullMd)n(EZM&6@a0QfyCkn_rOvB-cIW;rgTp*m<C9Yk
z^BCQ--1D{i`x#lo*m!nVofq%p#ig!<%<$h?ak4aA-M9Pfwf#HiY+UAh+1KQ9`qRvn
z`KGf@ny&d&yd=)_T){8bpI-iL$8<#6Po7KLBsL-Zg^9^F!R2aUsnXNhZf~m;ij94A
zQn07=CfE91uirluEz!66)+)F*$HdL>dgj%fyQSaXisrtzU3xG$LE`wl3->qvYBRi^
z(8OWno6~;3Q%J=vX~Rd|kbv7Wm1^UU$v#zitb8Y~biMSoSI30*>)l_te(yx<*|s@b
z)mEO@4Qt|M{qC{ks#n-Sr|8E`dk$}v6MS@Q<+W*VD?(1I?ua_^=*wpB)OhuMi`N|1
zSf;A2p_g>D@}#6;R<zQLtFDfRSiQe)`Z#%c%ZHbd<uB!etiJs1nH`z%RD0{W-Qkw5
zJWsDhr|p`-x$I4*Yto<U$n$IFN^kP~loxBTVBYpaJEnL{-ZH^ye#4oL2mSJ*ITKY@
z-mg{I8MW~Al#>AxRU8VQ1&%DYO4cvwcG!F(@6u(58P?BMt+0*@d68E1T5j%Kt(UhV
zO<dbo7`zU*)M74rxZq<`)3h(w*VOB2y=hy0^O@gV_Q|O%0*=;Pm!b}~7wxL(=F(hb
zwm9;a@#Nxpd$cY^-8;UipKI+}zOANyb50k_m7FSb)ipVG+566sb>5*XT{&tpxkC&}
z_ANOy>%|hEqb4iexZKxUT#1s3{P<`eXK+)u&eGE{hI^BmmYnVOnX0*xmrdlrX2sJ+
z3wNv(7n%6xC0|+Om5U2zxNhz{wP*TDo=e$+S<m0UzxysU>2}_XkU+6I<;`o?Cf=|q
z?YnxoQ%U5q=OtxhrPU?|i!`@#+5GfYR*%$^l4<5SeZ=6ndh*V6(}k`k1v{Q@7Gs(@
z=gQ%9wUDV-EK0pads03{nXb4Mwv;=nH#=}`mbu-l!<ie-MX$Q$wzVhqY|r}aXekl*
zd8OAQZmioKXmNPuHs7^dr##+#`R&SyVmj5a-QOdaH;a`n)>sqhsJ-0x;rcmymQT~r
z;=j76OJnn_K-W->W!(3-#Z)!Tz88HuP4W8FI6n&)7e(3J1J|zF+DmlxoPC{|!q;iE
za^9Rpsf~e?PL^#FJS*_Bh^;)^?YhT|H`?56D&<S3baAY|zQR-Ox>w<o(!_^P_lD+o
zdM(lkZgrIEnq=CXbb8q|2hLL|9zq%OROf~*$uVrvUdy}ZD05SlMa0qWlI8lgN<PB8
zA+2mDmn|(!=JnIO7`i^}YIJ#KdTO>p<5Q79*D2?`D}JQy7FfhIG3VGK#n{c(E>2-v
z1*Wcbsa-!&_(_gYb>#fDq3W-VW-bj%K5w(ir}FvK#6t}oQ+m$KyZl0TyPCoL1y`@e
z+&E-wlyb7qV)11)pIN_-ovnVgRpfE8=CuiLmslv?<>)c_^y$W%UB}Yv(#(W}vX4KX
zs;N}!os*QfUH1*o?60>9CTe|s)y_J#eZnWElWzA@AH0%C@tLz%LMwW9^Wzu47V&2^
z@O7)j@-5bkyT0@4+s7-;37nLR`5BxyyCC?|&81>*&vu1x{hZJJ>8sV>EGxUSva-^1
zrMKsA_w^5awDSAa&}(78R-asSI&q`Loxbai8=rnCVtM-cTgs!iJNNqbr+fdOTlUxY
z^g;dUr=NcMS+g%5#P~4fn#eY@u+nW)r(RpK?q_(-zxnao@7=op@uzAl3aGRT-M{M9
z`>FM}4jq2uwQBv+SQv;@uP**<Tw=8GbHNlNZ^e`BoV>2C@*;~To_JEkF)z%m)n&<)
zu6<ADyxE>tRK8&S;)^QVy_a5kS+XluZ#tN<N9BFMj}=Rh!5%xM;Ge5(OSXz~i|Isc
zOgh@Fych&Ld12uD4iyt9Fn=Zg%76a(XFp)z=MS~dm3!{xo6kP`42khbO<icQ2x#BG
z`ts$=m+v?J-`Ux7F5P(MnP<h8bIzq3|C%`y#66VqDN<*m;f(WpPuk7*e^un}cI?>X
z&!2vr5K>c8Qkd}J!i6IZ@4x?k(sVg<%aQZXH{W~|em&Uf5DZLWLIaHdF7&@g2Y)u@
zm+!v&?tAs#d+*D4-+gHotsV{nu}4nY&40iA`s>X*Z{?azI&vdNc)8zPtJ?Lev)osQ
z?{EM4|HLBDB7gDQzw7hg{@wrizy2pEQj`O&2nC&SHLsvsXVYmA+E~*+@slp(;FH$>
z_1Ay>kAMBQe*M>f`>+2x0-6blSAJ4|Bj*0`^PmPJD5#HH?ES|cc|AQn^XAP%Ip3;n
z=bcaAeDlp+aCCcazWD|ubneutQ>RWpO#fPYPx$4F>YF*+qIIX8PTh1fXPbY@90<6Z
zw>^68wb!M)?&fXx@9*!Q3<LA}-51P}|C^fP_3GBGOS#*2$F2x<ZeX=yVq$U?Xkl2<
z%AiogFvpK2G-;}A-(udJzv1eUt7>abKaDCowdcng*RMCP`qtii9ew#((f=<da+?DU
zOHM!ixZ}Jw*Zt+5Ct^EpyfNXPbbi%`OR<t1KVJWPT)6g{v$NwtyLG#F@BCJH;~jV0
z=gQ_x$A!FN9W4w@cA1`w6&YH1_5b?{$bWmgZ)e7}CIv0_e{RbEKmPtB|0U<)|Hbwf
z{NL{Tp~_@C!C-Cod9OnOCU+KGXpPW2&2)E#9#cRV*Wov^35VN7Od1=x5<IvLbqKV_
z7@NPpP|f_&PTpakP~B_US$!=W4Do+#t(!L#IxmuUSpN6UmCM3fJ|C(Xrx<){tYl|+
zk>_--d+E_f-)73%-Ot@n&v?bA@hkTY>(&E4tY6F<Z*_f~Df;+>RNxkI#h-PL*=!P*
zNhi*hFF4TOkaoPlzEOM5_q8$izQi5Nu$p-B;>C;S9dGhj1hp81Gbfrju4kX{m90VY
z;Nu69(%0-??UG&as(ps(VJZ8I-?;CcsbQAWQ+_F~_=;EIMEuWe#~<7?mdoro$$NkM
z_S<i_-OgRxw?E^2!Q8WR&z_s}U%&AD1e>{gcm18)+Z)^4{7rAm?(#Fcd#`-edvShZ
z*_+w>=HA|(cXzka`?xnA<#~6{@4mBpe%YHJ=AcQXztY?DR;^k-f12m>U$a8`?eb%u
z$EM2_9J6AUIA*z6&f&|P*}YfH>i+%v_s@F18r#2r|MRz<udiRb|MJK0-#>5v4hH}A
zPklFf|7q{XANT(~kN9iy|Inwkx66M1kNqFdV1I&lbJPFz>uuy8*@{1}|Ge{2gwTO$
zEC+;}d;|4ccB<FR{Zz`y;3?!Vd6~y0mq{*jALV*(P`@%$pz49Jna@q@BkzTq{5~ol
z{JqChq2Z5kgP#S90868}%f-nHd@pKvwx|@Er`*{8>HNp#mFKJ8xBg%9WP7;(|LRLs
zu2(N!sN#2gVavYYi~OJb#^3ROA9DQ=G<LjDtR6BipnIa~fs7Kxx#!yQ&3qEvemGfw
zl6f`BLhyi`oy#-f87@zrHD_l^9b7Zz$TC$!)ssG6Nj?)D6+L~t62kYYeKF5-DN=hg
zdsE)YO53&Wf&X6qD}V8S&Pyx5$xB{ZIWCbm{xkpdpZyQ(kJs@3(f<)E{`*KdhwlNO
zJHp@GAN&&D;-!?&Tqxjw!%Ja_+LigASIT<M;5lf_Albm-z&}lYh6t-gaH~adQ^<mU
zQjhus4_UZ#Jd}x9cuHpBe};V;jRFb>`H$aM)0y?iVX}mOr+UW%jpa>M3pAde{Hw|c
zI!)@zOs7=NTS`Uso&S`YY=ljGE~*}|KT@T>WA3JclOcYJ_Deq2dwluA@P*-voK;fw
z$#d!!jEQ25|9$@p9;!cL|7-bQW3i+i>sGINv}2#--}A5cU$4KapC5f;FZcC7-@h?l
zu`J%RMR@+Ah{=b%1*Wk4l4W0hbiswH;-63EJPMh7Om6*7n<W>8U8H0=O;j%IWnU0;
zXS(j`KlTTI|9k%K|APPX{?9Af<?s3b*8j?X{_Kr=^Sw_W`gl&zBfWUxT$PzBhrB#&
z8=tBgYBrwtl5!|k)wEanGii;iEXNDk8IRnB=FB&;ns{Z_hwIs$>(qA4e-tTvz`e+N
z;vLyo$E#N+JZa}zYhn5N@NOQzgVv0*o(pUFtTbA3s?{!Frc(`Pp4X&eXVo)2(>$ze
zH!MEYDi>lj(_!vht~om`-@p6c_@DV7`yc+F`if6v>MW1_-}Arpzwf`-{~|Y9=IcJ)
zvpw<S{;yxp|4k@!ur_Tu*H&yHxzk|I9MweKwM{#iZ5Zb<1TZXgv}sKJ%>MjA-9n{B
zwi5&8IacVbi;wsC<eE~&w>)6h1>TIP8V0*3KN#vAtX1AjPV((EQ(3OIJH~IJRmv6N
z0RNY!Dp&qH@4otK@-n}fR*Bx1RUNfBzo~Ah`Iqu@@-I<Q{y%N{6OJ7#n4`{l=uklq
zZ+qy~`a6~Fe|7V}X#GF%IWGHb{kH$P|6Uf)=DE5;>mT##73rTJtPH5MX?z^!%JNa@
z$vmTqlmC<sIZT=}>Buzx{v%p;iW_gV*jVJ4F|NDLqQ6AVmgTeXKNU~albhAf%r<fr
z(wLXzI!R}~kQ+<b8}sM|U0uBAo@iw`UwqNI>VAIbNd-m@7Z%=+3~xUQupej?^H`vE
zIrIDG>1Q4mti2||aA(QQH-B!v7juj?ulis4bp7eX$KiXo+v>62HoA1KU%LCwd)v*k
z@8@rQUc1-z|Ac@3C$GNxeee6;_W`9&ty}^b0SiDYM*=|Y0}&1Fg^r<IA{zBJNB*Dx
zZnu>G`Tyky&s{1LjVyb2P52()KQn87L*>iQgFc!FMz7t!{BqmOgyVaj$83>aQ{G&-
zqvvYilV?w(e){<9^UE{eJ^8&RxLg11`*Pn$clvj3-@18&{QcML^PgE8OG`RmIb|o*
zrLolHvak8dUtE=c?d<{#7Cu)z(!WE2rKV7(s-hx#L1Fq#Ny!(6m%a+{7rlC^pCh?P
zeYNC^yOlSXJ{?kVjjCH~zy87Bz|-I2K#Rb|-_Mu%$$oymUCqzW?&qiOtNHo)`T5x&
zZS@yjj}_Yf|NSq0jribiS4-;sKj%OFmpA{y-~GG4=lq?&@$1^C&aA7eef6?#?t5OB
zv9V<SWQX~8m0ZgCe?Hs4+`Rgv^3ybD4?a1c18K}=;z1SZ%dXU4@|@%J(vG2kP0LR4
z9&_iRmHW4q>zkMJE9Hni6Svvhed)vcD)HN|-P#R){rdImcgGRCeNX0BtY3R$-|E+|
z-zEK@S6f?K`{3COBmKSXI>oQwJ^cFh>-R;!PoCSit)Twd{NMln&yp{;U4H%czrwHI
z^OwXwuv>4p@9F%?-P^DG9=kf{wV2jd+gV><U!MpWSUK`t)irwAvOK4*s;&@UlS!VM
zFIAbJocbki`~U2#3(q@pk~g_O`7<;6T<u?;uhzeR|Nd$Yu2a7M^I5t_MnLL7qjPg}
z<Cf~lU;my@zvKSo=e#fd)%V1nU%K1B>wg)`{Uz;}?sxB(u6_2Wr}j_O?Y|lI-Y-*9
z3XVMlgYqx^)knWf{QBd^y|;yr3*Q$0)jNNw{lmlV$;p!^PnvdL`S+fA`_<Lgo;>k=
z^5p!<lm9a4_;7~QEP1FV_2C<vpZWVb?iay-e+vDKYQ6C1Q<7tb!NUt&7t%N$c<>%v
z6e4G^l&8vvzeMGmYz;@~UuP>>n}WQ@LTkO)Vl<fkZdrSX=|#K~gSFxp{sYCEB5o(l
z<4qH=Klq)^VQ&!I0#nJdZmE~^-v`S{F|7Fb!brlPcOj41_v0$hu3mmw5@>2tyz|bz
zeDPU7=gy6rdns*u!RHOen|I_Md7~Cr1_GrnVeiX#pUrt6op$k7_O|J3mtD?C?C;iC
z)VbB^YSq89RbOAf&MpOm;?KV#r%gWQYw~Bw6ZPdivx<c?+BeKkHF^BwpH;Asr~0mw
zrk*O3yq8tF&wu;rulrov^X0Rj*Yr>P_|LxoqpiL5-DipQhrjH-|2gjI+ow_+|NN`B
z<tj=#pkb+a`nceqnMFPa>W_S~oV<Y1f#HSEew!&Q7bc(Fs6O#l$XtO(Rxbq#OKcf9
zohnZLc)+f1KUInM{l8B_C;A%XUu`PNI=sBlT%7q|El&gUix}1e>^~UK7`wa^ejvNP
z(fUXoUvCF{iS3bTY9TXU&H<e>{N$w0|E^E-0#7}ask5kG_HVg@#e3P`^WX28{$Kc?
z<v)j7_6zdT>;9iA^4za>X1-C)NeTal7A}W{c_e%k5;m~#893i=PfAd#Y1zOxpGWC2
z=Q(v9=D+TdFK5kazq%yT-k|2>zrrqwg<TTM)D-`P|7QI6BDww7`M<1Z3})y%urbVF
zO5j|lwqrs`%=1Mdf@iuso(atr;L`|Y>bt>i(bw|7!Tyo=<PgSlPXsI(k3XK=BYu|C
zL{Fvnk#MJ!e};wNg?WoIJ&oi9rhT61sp$DkwNv|~9G|1<VwDg63^g79Jc}HJ7$!ZL
z#IT<!W5e77LVtV}GC7xU=tWv8o@D?2;Yxxt^K70Q9{&ziTM6HL6g2q)!#DnvYV`}#
zRSr%5<G<0M<&f|j&qI-w%-4MVr<l5I`pkOLaK?cjlcjt4A`N~N9G*P=1!q#$*NTSa
z>Z%Dvo!nf~LU!#(n1$p&RH-^&Si<~Iy*SCY)4t_@=YQcO#wW@TMc9n*`6O!3lm5T(
zk^P%p`l1&rj!XZY{QvpBf_u`3tKZlDzwrM<{jGZQ{{i)%>L27uw>|tXVt=dtPW`d^
zV|w*_zt69{`DVJg;@^MSXDb{3C;s=bKf(V-{hM3^Tfu|}!bj#RxJ)XUykyqHZ=Pa)
zf=Vans!uUqAeORKtzzb*AcpjlbJV}}zh(b6|E;b1rn{}Tzs-Lq|Jzbs^~s*;lRm|H
zpRCjeGbIgpW(ceK%#`i?&hWdzhPB60iSrvnC&MpB8<(|e6^$VtLK&QMd0rGcZBbyh
zW1O$%q3I!GpsHSDk-{DDyP>{8_0W_?J)wvCGZ`K=)bEO!|LpJk)Hn9uCa?Mb&i_^Y
zn)!d`f4}_u{qL1OY^OXk=PdX%dCFhSFZ-=(-k+I&y{_cv{h8MKKUaRSH>x@R`{lpK
z|7)$zN4{F<=X;as!t6sy9ehVtDEJF{s&a_3Y)~lZJfzgn!X&`?>EDlkpNb;RX&;l(
zGl-6z`|9iRxUDZrbZfO1TYV3^Yb7cl;bm9W#BD!6^OR@qwj1Yp{<iV69F>2wvbT6^
zO3$Y2Uv`<_%QrV$8D{pG>GbC}SL(`S>y|q&T)XShb%mdEUrR47y|%lP_4)Ve4Lkd_
z6S8yOJhQ!)z15EYoY>Ubx-w<``_rGDI`zOv_w1~nJLcNPuVS=sx$hOZudO#}&2;PL
zgAe|{|GVt}nyv3_`Zk@moVshC{<{?yC7wTgEms|HP<KVx*Z7x@!C^CAlWVc@{?)&`
z)Q;Aj-#q17w$aT$f3H4f{GV-FwKi;ZAY-|={)(wrvMj4+ym;~A#I_kfR?VHAR;ARP
z?8@-JWm5E&_ocJE#B=Pfes;gU!*JD#HFo~jw(ahcjk&*d`oHI@!G@9CGu~f0c%|Al
z`%+D>${p**xbJ&q;+&8C&Np^FdGK|eUi#fH-!{~n%|A2U`{t#v$xHJyI;-<n1Rj6%
zc}jd-NL;>_b*w6vh(=$)V%3uY3*D7IHXYFjSm<tA&3ipyVe_1YZia4*PJQycE=H~V
z5$x$XPJM2CDrUiwlk5&ZEEaB=6XrHgrG90k;gPWY^B)Q>(K&fjIZ&13n#YSETOkYP
zb3VUIHY-~3?psv8UZp`j-^b+$zjdH>3fp!e{Zo^CdRVv)Wpqdz1q#S!aF!?-zu$TG
zbj#UwFAg0OnZ|PCDJP$&`2UUa#VlupV?6nE>VCB^c2-(xl=CQW%|l7ua+MwG39}Z(
zr=MhPW@2Cy-v2WH{Iw%R#hf3mzW+Jb%v*f#uZIP@^h#?#uYNxN+4rp?AM_+N_J%$>
zb-H75K6m7_w}rNwb98>k_D6lY_V)PAH%mXx*xPpa>g=P>uV0@&eZBVgX>1&C+Bd6K
z&V2X!?StAd^LaJeQ*^IRZO^^x&E3-4di#ad4w=5%L!#e(xxRCA?>zHCF|kZwheky7
zsV_P@ddn|=-}}D$-pY@)@8qkSo9{hemL1;~eZMNXU-{s#diOiWcC5R0t@-WR<-4S#
z!NAtyw#mxy^_%L>ovoca@7}p{r+59fwzjsmdbjoK*RNkcZN2(U*wm(e`lo;K$L`#@
z6<&St<dObAm%jbhvi$8o=bxGK#l@Q+#I0xEnqn>;c=cFHgmn4MJ@e(m&+b~})XEnr
z7Aa-7_WtUtuRiSBVS35tl-fo63-h1&^Gqr<x4-%PZ^5492keTj=3daQn}1pT_sQAO
zbL(!OD}A0{d9S!AuUvs|4pX6_blkE}>$zv0YTmp)+~&yZudl)d?>?%S;XZxl3!Bf<
z!X0kS$6Y^uT>r_Tw)5fyll<GfY&A2#bCo{f|J(kz?9_{Y|Nn0~Qob&xT!k%h8^gQr
zUhMVj7#e3a)jwxDuzo{i&3*@k-fzv_4BhMt*nD#<jit8vnrSdG7oU0fVv~h-o$U9V
z*QH()OW(UY2O2NCwebJ`9n<VfXMXP!cba`>Tdw?kDdC+V)-rzkc_tTZoqFo?JJUCI
z{yKNOr<L|(O4^3Zz2!46Y|_N5`6oSBdatV8?YT4aMZU;fDL<ZVdY@-Bm6hE;SfDV^
zDzd%%cdgpSOeL3jm+CV`g3q0Lnf7H@<i2a>x1OCl6aHbx^zWYx4L)3dxn^JX)aA>S
z{rwI8+A>{v;c;PAUg?_T60P*-&zGN{Jb&Hf+g~d0o!{{L_?fRyi*A+wv6FWbpK<2h
z_o?}NJ{R>Dd)52)X3xI;V|{A3>+1Czw|5?Sl=P#<$F*ux{`&NfJI>yIZL?BR=h30<
zw;6do&)sKbE0}#L_vBW^wpxYC-&1#3+9-ao>&RzsKh)oTxcu-E0i(VAi{vkDsP}yz
z{3AS-A^!CgPS(Pd_J6kpna?yzKH;o>+E&&ruOu^9Bj-zz!h0oNzq1Y>t{zfzpE+^;
zM5e8)udvS5*L-0eo3nPVTIm1H%JH7-j-6mv7nAz?xjn2+*pjoCBjGmN$JKMhSb{u~
z90hlN=<>YvYNDsjk!IzahDQq2PRz+ul<_EAGGz&8P~<ALHH}kL<u;zO6W8&uS#{v6
ztFYlE-lRGEr9(be`W`>MY~P8ZGVwom{F$fth5j|lc@Z%wX6AvttewUwYiDv>+BL*J
zYc}Iz-dT0?qtgQYL<YbAe&;6%Zf=aM-FH8K`spV}zq{7$k1w(5_;-w*x$&DH`{Epd
zBi}zc-+%x8WSX<#44xJK7nmJvo0=G^KAhZqHT?Ct%dXBZcpt<w&Ee-gpu-oCZ5uRc
zQ&FXd=+OhWtgS2kH%z<n*y6^y=W=@t7k2+uzn^$GR6_Q{tRmhdj|CUfCFaE?ZnthY
z8CAI{#)9n~ca8U_dqRu0SIz$PvD`&G%swO2?3a#Mvf-waUpE<FvS0nFmY2(V|Kum$
zw_cXnU){O+*S5~jdv-<Ho$H^y__6P7&{Ys>K2v)Y7tdRob+fW`b=bGnRid*%hjZro
z-OakX%XD?%!l-R)Gq>*gRs7}4mn}7a-hjY<=d<U2-k-kl#`phHd)1$=yxIPz&Mta(
zZ&}!bwNW~1tgk~B&)$9a`NJi*Z(2&OtC0~~#msv0`iH1@=lP}PB!|`a-gud%z3^#q
ziT;HDpO>sRZkf+t-(LFg^Sp0)>(;H?Rej{Y<jvTJM=!H?AKevy*UYbX%R1TjKA&Co
zt$X8Hl*IALr&6EgPeEso@=pCHw-_v0+ZZP5&C$|(!4hcnq=4ZqLn4#L{6o=`cQCrW
zjdGu{V0UJ5#iY!W7gg^t|1NvqP<HaS`j2^!OefyxZOKtil;*hMw#l?Pfc?jOBWK|+
zJ}WI$KCvZV^SfaE(X(95VOq+&iHWu<+f@r~80%-KwJ6_juy^?@+~%I7sqQmXQTUh2
zgk-f7Z+tsVRo)35n6C0(=*YZ+@aO6+GCxy(34WQz@<sT9zspL_^9-{XelcVpnDv2s
z0&fF*i7JD-S!jvN!drglzBf;jsD2R7pkjT({{0WR^yVi22=;&cfBbi_{4f@hlbkWj
zJGbF~@4x-_^PlK(H$B`@|EMOP;o0~8-}iqt_s9hN_ptvkZ$?&o3E%oarVb@G2d0lw
zji=T${$M!S=B@hJ@!=ZPolIM%sBm4X_mJWJXkI-<s8WgT-kQehJrm68Kbl_eFIj&2
z_2aj(3^EQz2Lu?zW42zK<bB`SNv_Fp@&><y3MX%>r_9X?*ub!dJ!WoEz2|484aL8H
zJN=S!xYhbc+c`(7U@g~y4{ZY03bGUUPHLWb=C_l1o!>@g0l9=t6Tf`jwdK>Q9WHBT
zs<Wx7Of-}d*u00KdV>oigM+LfONomjha-Q7mCB3B7Q%~sKJK%By7*0hO6^DS#b4U5
zsoj74Q?>a5^F?MiVFg*e-cDm7mR_EXdI#cW7`NFd9h_q2c!@vLPUUWL>@()&|EHZ`
z;jWr~;)oybw=l`v>)RJjpIgyX{&)9>^((`6UDzg{e{PlZuHzrx2UVL_>mQHZ+jeuC
z-s#?>UQ=&<-+TCdTEXfY*0ZFfxoY-XKRbNx*%gy(7P`;tCF}Qp@A=#sx-_(Z#$)E~
z*%S6XzZTeX^Pov=-uBR#_3X8`Io_Uq`q3gs_8p%xZ#mz$Gu9XVcI^4QzWmPJNY4qJ
zvlpwr?Z1EHP4(aR{swXX>)B3+&%d(6@Z<HOyAf|@iS1vva&h<3N9zkKW{1g6``&q^
z_fet69ADn~#;0%Iym_;7^XAR3ezyAW{J7)Gwz~b#->%&JVda}QR~GjFjEmH|x_0f_
zwd?(}v$M0;f311DU)R#Qy1F{AY}>YNWi|iJpT1SU`s{V-HV~NmHa9m{@kOrE!)@2D
zUAuO9=kDm}XgPiH?(oM?D+<1><Li0#XP!Lw-Amf@EmvIF>7+J$c475Z-)}~fEo0)O
zV{2~3Jyn&z`{Sb^?}vHYUc8V=PcEPM;l1_h-K?jdY|&d<BqqkalAC+c!L9S&PHZjO
zC(VES;Jtj72a`Qdm|f0bwD3wQ(<x{PnJ{0d#=-Ca&ynoosR#6!Ki~eeam!sHLG=|f
zhoo30r#znc{F?gol&OCcAI7oPn}7b*J)L3QO7@!h54KeQ^SUSB<yqIU#_r(Jq++H|
zb-x=7%B*6}?e_|HJFclYt+$`A(3G{9&ro8<hLj&G&I@~RNggz7>1Fug9?B)6ak{O4
z`HSsJ^Vr1MZ_ht8>E*^SPbSy94(dDBeYGmjW3`Lcng9Lsr9ZDX$@Tu`>s!yU|8ULq
z>^&E&Ym;Laxo#`nsmJu_R>yj=+~=7=2RrxVKl!}=Q{9~PNzR&kyrz3^JrcI<T*$*+
zGpzqU->7->!tMg|d20?Ut3QpdD1RwdZ4%9`pa1uNSw2ry$2Yw?zR!`5S{v4fe)&~<
zb6euqnxqAbcJB@e3O)QkHna3j+&usLp)L96-FJs3KK+q+{oVhPJ#Sa#vUJ~f4GS**
zecki--dlYS3W|BXsscCEyv>{S{e*qnyEoU4q~$vOYnOUjc3$DF)9<2cz5SUf@2$%p
zw!O`_mhIafy!iAB?YAGc-a660YL>^sn+Auub=KWEcRYXo@@<DD`9E%OPx>ut%hCP#
z$Ln>wzUderyK=VM_P%#_dwX?padF|fy|=gLJ-V~=(uP^JCYOCh0@rle&bt+_{H}KK
zzw6h2+owF`pM1wNSIH=6!>nM(uNAenPaksK({s`+m-#r`e74vooeMj!&4`H0{`xuO
zHa7^|I$N-D!iL>pYx~YNiHV7B{rc$FZf<UFu8i0oJ-+|kTsyhBxo&fF9ogQ}`e6U_
zwI>&pslS+0xg*+u|DwIi!^v$Parv5G`j-@Cv-{Y+F7g#L6ivxc7n#t~(^Ko*-Kch0
z!QIA5<-btHzx~A@`rkA%u@<SF{M<grfBG?I5#bl#PwwzH;bQrzvW7LO*jQe03FDI4
zfB*mS|1EIJ%|`7)Kg;Dw3R7nFtyKJcH+_;eLzP;$>yf)ToNrVM?Ok?W5?1e46;wXi
z{Jd6siRbaBYMck|veno-eozXEKUu$T@&u3LPVs${{~vMeQ++b|fq()JgLuOpW)60h
z{v%;)zwSTc^~qwA|HJ-WlXNsV&TW`nsgysfaSgMps*dNrPAmN-CoLwc`2Lh{kpI+Z
z>L6+G+;@^u#vFN{HxCt_b6R+FA2i&&uuS>db@sUlv*(?k%pbmtN#42R%%xvHd+hJY
zWk0XK^j!0ssp#C$-|OBBH_Qz5`CWM>N4IE|e0;*wIVRP7x@W?@b*J4pv+qs%%0hmd
zw^g6L*I#<{*t_=g<Dxlzd&Ao0`)+1$+_fup#<f>>OLM=Q+@IU*E2um5dcEzr=x@1~
zcbFtb&fM$s=FdED@7Z(8z8^jNJ1ItLd!C-P694kbuH40o5Bb|(C{4YY`Es%K%$n_9
zq1!VxmTljCH(`J7q>`B3n{KjevL18Waqo4WT$5n=ZqxZ6c79(z@8Z|@Mt`2~`(D|+
z*XhsW$?uJI&-kP)kg}9%dKr-@a5=krZr<}^X=~Gp39t4opM2!F^|3G8^$qhb2HO7)
zK67hb>Gq2@d1lr1`|fePKizUq^>snH_1Q-pxwRK>&*VMZ_B?H&qNH@?h3x4ywx12z
z+1c4^?;qUkP?Vscpr9`xtHHg6O)|5LiRtyNUAr>%J1Z*gGVd39Kc&1pF}D2Hdg~`v
zdS`R8*IjR$+jqS2+|$P;ckQ0%Z@>K}{!!uNJO9M2oqn6kYpxIIUaZcne$id^tK+tR
zFRZTo|Ng4;;eFnQlT)3m`jlF^gbFRYL^N(`^am_-k?Q|DC-D4%ld6;MRhX~8c|>sg
z-*Sdoe`~T=<@g%fq?}9EE}wmNT6Cwj_SU+C?L9M!GB&;cw&Yn!qw0zf))jG`6BrJs
z+U&JECMGm->&oTDGq3B+I%tsiY%AXlC-=17D!(r#6!6SZxH#$Sq0||j&CSOO^Df?E
zSUUG(gVD#$>rzg%xhpZ=s^T+!Guz1{B5ZTBrI`aSBa`<}<@)Egc^h|coWRlGE9(0F
z_oKO{)3QWd@7~>Jk@@a^Zpt^MkM5Ow{=SvdFR?k<n`sjMNoTL%yD8JxIqIwIxtCa%
z@_w$0pL@b2v6im0LVteix$iyV?=#nc_ojiy=7Ux~*H+DDY&ei$AR7KZ@7>>zwHL~3
zT=&Gul^^+(<FSUh%6(BKr-*BU==bR5e|(PDy}YQtO*XvAeE(f`i=d46<vEY$p4kzb
z@@ro0hTI0dmfYi$<=3j8nWy2gs66zEXXzKOyYs*0R_gSc9hZOE_~q}vpE7e!Pdk5p
z^W3k`%C;}}zG8QM>WSM`S%2pp?~!|8qIi}=SFZp4;~AB|zwD`)FXqu8cDicU{1(6M
zYVzyLKY#pv^~&E5<;!@lOtxKrsrlQUZ69ZzX5L`+gjr7FY?|W#_0_-LKl)$4|JrB!
zX(yH)+E%J!u}OBnckQyx_4^m}^}RGbv*Bt*?aY7~yz{=*zc-AQd%LGMyeT~X;-O!w
zKitZAyKd|AK<>NmE);+CpYy5sP|<mg?*ZT2lyAgpZsdLY=Q=xIa=%e=d|>B=4<9Z{
zvhEDHDfEq_NoFRWg)Z}b7qw;cjc?6<#r5@!iAjH4{?*!z-}iply8V3pe*f}3-}nDK
zBLsex)r)ss^?#{seCD6t%RZ^_E0-@%RQyu8SFUHiluyRpEc=Id|JxrIp7{U#`o*s<
zOxIs~Kt}rC#pgdJ&z+cdG&gBaoQ|%+#)nDf`FE0Q+ax~MY?DZMX2jEWSVSV}vxKQZ
z!oh|)v7O?#;`eW@`}_O*`};R;nEk%}{d?7&T{3&^;^JcaR{MpO*6s^l`{~K7z|w*(
zrow-t`kj~YeVuyr<(c@@@8*oJYoC076j?Jd;LL<genD|Iv9}8!ANv~m?b?cr$`HXm
z;}4ThJm3DNXHLY`Zn2_-pRt`B|61aneXY2q_54+;`Go3uCX3gv%lDsLbHTGJX~FLu
zb9$BP?*INh|Hk?E=`|<T@9*rWH+et%<L>>7+Dhi{`+5Cz+wY_9@Bh}XEX=P{jP#Cf
z6Fd69p>pH?wZ12f{c4SpJ^9=g+-Z%-I`KkgZ_d?-P<CszH-CgAbYj2Ed2sb}gVguC
z0jbKTuFsgAaJzuHaiTiIj0FDgJ0>#Df6w-6#_UbT!cvS~?}eVsVDg!)<=P-n{i*Pz
zm%Edf&=aO9(-Rh}p0rC@I_+zK;DhO#Dr#(U`=po23G8ApZR>v8`e-u$$HQDNn{Dsg
zB<OyvTYg?<;?G_6&+F|9%{qT86lDbD-(vpK=)(Vn!GGIq9tXFo8F_Q6HhgJWXnjEJ
zfOG=i2{&O0#;*(|40jo_8dRL-cy>0{POMR^zoY8K@as{6>Hpsij?LfG+TV5x)}B0&
zk!bw4wSLnS{%@abLJw@Hda=J``{v8C@zTrJbA5O(7{;Eu^5y?q360N#E1KSJWq8(5
z?4W%>Uy;A^dNFgH-$%>D%M<rqF1CMV@UhkQhyKBnDxn+|Cz-rous1LtU}Ro)ne~g&
zp9D^YrAt^9&oANn6#xFAp7`&hcPt*3>?(Y>)uvr*di9R*b`9e>?S0-yKc9Uu{qw!f
zILS+eDooRBq!+AbT+e+s^@CE!6Msb|&uOX;W(Mrsz<<{#eVK`I!Nl!nFX}Iusa7%l
z!~RKs_{(oUKKt$G?j09cl764m>1<-SC|n_C82whGwoocs<Gezo{gomP7KVfc9&OvW
zmEK<4w)x?QjTeo6?^yddR<3dJ%ln%@7Tw)(TPAF6@kaToN9@AuJ#Xy#nrOQr?3=mk
zov@2$LXo%363db#tYmNY)f+Wf*6j6T@A!9e&bioiH)Qu+{<r?A&h_N*<samEb?YbF
zI)9qB=~JwR%9EE1e#r1l&7G#cf9rw;bA1kNdaU}<x1#9k{}pCa_vWyNFJ->?;!A~%
z!Yi}pt3UUBz2{Va{Ez>_Pg(!_Bkkt@_$OKaIDDnv{2%}RpTAf8{N8Id^JllWC;Kkp
z+^FQdH#xuJOq?=r%b`Tc{?3nuKQ8%Q3i8}6S{ov(wQt^UMq@*s=ViwW46JnAGA}k>
zHZM<lW4=w>nvF+7E8%OxgM-a1J`%S1lk5Kei;Ii8C(qX&JW<Zu_T87d_wwTWY#+XF
z-(H@Vr(YPDy<<<DzNl<}@W=e_)I0I3w9GARLw*^Co!Wo%YufybNl%`u%cVR^QxNru
zH=O6$cO_QTY4-KmXYViDx$NbSv}K!Krfrt8T)pSYWqqTpw4_**J^Mb#HWpkgTDLDY
zt?kdj?IEiRCnVm<c&4UiCM#|e71`OVbIe>->PTLZ#qUQ?jvUrGa^%R7!#O;M-{ff}
zB{#3yzn8hm=jxs_=O=r3wSGA=^#W@`^kf(Brs}u%l{+PW^-e6QpLa^AO+Z1dUibFS
zbD1{2Eps={ys+1^ZD+B8*QUagufE1Ta<+4o%{A2hl5%-cm$H{&W8;Bo-20lF@37n4
zdtCOhyq9OW+1J}MGTG)d>bsOo<2$K(lJ~lriRAA8vt;V7bNOcYv|VpM_gDBC!^1U=
z$L=y~_(Ynq6ciK`+!PLcG@bokfj{d7<-ZT#vpqP(USi_1ncpFlT|u2G$2h0<%v4S@
zog1F>TE8#j5C47VrvCJE{rpd_m0bH_y+7W5-mD$-_Q-MU%MY$w-~EhVeSUS>*TgU4
zS#us#*NQhBac@^XJuhnaU1bO9zY5J4yg3>oe2#AqI^bVubngY1Q}NcwqO49C@1`mt
z0S@Ow+R_(ZBq`s!$9&{-Luc!figVvniVx09=v#fXzRK>5TD$W__w(;uRv&1qI;G!q
zPhQ}Bv)HeyBf_nHit|s49!P!WXKh`)p;|J0!-rt@DVB3r`G)cRob>d}f1}?E{9E#h
zZdd)Siq1)^XI9;l(RyKXjnbM5@2x)qzPCM)zyJM4jY0b6Gxt86y^xf5dr4sQ>bKuI
z`Cj>&7jL}r`sLT4&YM$nm&#VR-Cz9tySVL7IcDR${L;I7RzB8u`+q_1pV+?dU$*6)
zHapv!Yk5)0rJ1?@-s|sYuK$(U@Bh5^JzIX}mXLqHGCaTTeZ2qj+luPH|Gxc<_7!;b
zCMQmA$KTbR%Vv2f<Z|tkcR##zzyINfOP{9wsMt13=C9EnyYS|N4?a}Q{#QTu<jL&C
z7mpZxD9et?Q8)X2I{o_Q_%GKjcJ!{VUoW*&Qrw}foN2MQrP6Mljo-LqzD@G1Q_|CT
zky5+wE-$lBPD;$XO^k<X7rmZZu`lJuoMp*5x&Q3L_HC0A7drCn>G}Sv#TyGhe*6`C
zrOx+!@#+5l{ylc?o8LrPL``T`m*IbPuk-rrUo}TA*}VI+<kQ4Ib0ZFwc<OoIR5;}B
z$Ee)?m_<FtEV8J2fwKkY1m}2Rw{+*|l5?^3QVw^wOkB|OzezlOb!zY1Zzty|KeTqK
zP-~n0$=}iV$x~qoccC=Ty=vcrr<d=_HqDim`Fdl5%kkQiz3MSyIvvIA1}9ROK6w=B
z{C#4TsCoQzi$^Tuky!zqYKsFr&I)hPnBJp)evR8rvus_NHK#2l9~d_7;?!}UG0%YS
z1j`LJO-sv07M}nXK}Ap7{kt?Ls#)n*`DEG^Kl$+6NaFjPM&%Q0>{<#{Cz`5EP+J}1
zBDX2_q~S#+?vr{~s=O@C?p3dLw95}Fn$eWlcx7@*pVHaKnQ|Q0uDESVRy%Qhhw?TZ
zb!V{~!Z%LlB;|WrPZ5beQ@3FDEbYTjLw2zrD(Sq$eT#2;+2z1GwM7+^WUasc=v;qf
z&v%aI@{0B=ZSO<g2!EI`^W8)17HbDzMia3k=M9!WYOiH~Gk=rS<VkWO(ucw(Kl7QN
z87bAV&c*VDQm2TZhq23<cB?velR1}X8{Ipf^6&83XZt_y{~G_NzxUti|2_NecCY-`
zck1%~^gnj?zwAGLwURt@C%-6|L)38M42hL#+w7K?Wli4q@891$@8A5o=<sjp(jPXz
z<fo-B*PMKGGxKE`jeK)iy=jZ5FKFkA+InXBcS+|{r@w7nV;+6=Z0LFMmY17eJ}jP@
zBa@T9=y%8Mc(Y3GZ|k3~Y&~}WHIt3y>K&pF=CY^Wi%3x08?pD?-TV0)ZkO$~T@|q}
zW&i!dHyf`vPe`d?RDAjU<FYu(O>buY$Y*-K>etcgLoar2_qq4?@zrD7=6(J3fj^?+
z-J^MR)30}bJ>fWi+OF(%dvZVh|Cb{B?|<RclkdI->D*3#Vxbeqd?@nBzWOWQ(m#JG
zntILFHl}WI?amk;?rV?ix@YUnZk3Pyzxl&ufw|%LK5zf~WM7%dZa3TQ{5eN%CpJCa
z{>r@I+@3mn>E0uc?45Uh(ckc4!_J7f5*_^{@wCHSUpd)X?lwMr6_b!qR=^u|^ym)`
z4i27fWxvMf<>x9M?8&|-XSnXiz1h1rZ{ECJ>{0Kl#mDECe+^1H%{}Geu7@&a|7zds
zx6VlWXtpO;X5IFuPtR`2JEgz#eB8t@YN!3;4W`eUR{mbiegDU%Gx@no@Bf^`*S}a|
z|BiphZ@(5_pPQ}nDSGPiy`rzXHal!o+xF+F=vAN1j{FLnjXNXvMjkDT*f`Ny=~$cX
zk{Y?K7gM)y{vRimb>b_Zzoz7g4?f!-^BVgI=r>N?$GjnrX}hcX=DW-@d=JJ<=JVgF
zP`M*>^-IZEj~Ns8|Ck<H{(3`rZn4X$YL71xo5Fv8yE%{RT#lbw_TyPM-dI<Q>3C}V
zs6Oz&S@hI|GfB_YMIO(2b+Tvf8}F|HS&9pnJ-oj=|I?hlge=a3vv};@y^_%ramu}?
zVeu=$kKu;s!mi?@Uusy&R8D&^d$4-2waonUKPd5C@?O`Q>hXSjE|(i`RVo!W8&6-V
zl2LT<mAOh&f2#2P>SM`zclTG=-Ceh^_jat)>pNF=yWVwr_2}+<wtxLDraOIRq&#o@
z)Bjy@!rQ_h4`c$WdG+(RPiFYCKX#AegyTDyHXP6FeEgwxqwL`$jPHN3JudK?6!OI8
zUiJ!wtL)kDckh~b@q&OabInKTPxk^&ws|n0lIx9L>!GlcpXG<sujeN&Hh*p9;}LCN
zR$}V#Mdm_P246-@o3X@K57`UXSU1Ep7nrkcu)b&TiOomPxq0e?56<tT7R=x`C}Zw^
z#%B<0$jBjAIz!KfJvn~*@2@YE-yif6k6!odhT_Mcug;jhuD0SYT<~*kO>fS3*Q0mL
z&SibpmC)hlD17iI`<}_V_}QhqGcpdGU3K>F=Lx_5*500Y;>@4lm&>AR?H}i>il@#{
zU+wziLiM+MoThhc#5PP{eJwjr;QvqaJgG9-|5n2B8v@tIx0HV>$X;D(xBKtg_<tdb
zDr1hMq_+#2%bPE}a3=2i-jFG)BPDE8`pd%QFJ4pg`MLk#r4O3rTXwQ#iDiW?^L|!5
z@%n?HgWoJ~%DLUQpMU<z=AB0({r{H!G`(^uD%tA9;>D(CtK~la=`hRb+-kLEntsWf
z{hpq$pH=?(zHgTAtLI$i@t<zzZoH+-cXAoW-=7?R<RbPn-~FEXWyQ}sckV@-RU|~b
ztGc)Ki^Uq-JGXCt+IIKVn>SDX)ttZev*O?N`%@n#eT&;Wsc+-GW19cd{X6G;d|kQA
z=GoWBjpug-e=Qcac)qHqeSiAthJE2T&XwkwXHEufddxYMa&nXAE+fv#GG}J<%rt#@
z)a;>A`jb0znDW>SjWT$GQcmvwR;)GK!?thxKlRD1m(G4WDaY{1&HBCeB0q^bk5{kL
zBsqTkRo<Sx)rHYTp2_7A6K8+=p}k)Y9%h$e+`nJ=Xz5R(*B5VpV|wR)Lgzr)ET%;s
z%qzU27<eWcEBu{)|IN;pli%5Lr-}X+d#l`|yhqPUOziwIMgeYS`^{=+X1u*<?s9*p
z7t@)V8y;;N^foDO=HSdJ-@x?3_P-a8AWH#<lR%3id*8{=n~Q!_i)yjU?Z15Lj@nOi
z&u>3C?E1I(+)jC@a)I?)dUJqOQ<kaVt){r8PG9(H-fZ`o%5rkDn&->c?l)gdauD@f
zVU)-saO3BjmM=4s4RuT-?|39R{}(<acVP;DhC{}eXrEFKHr2WFxg5e3JeRg?<54as
zoO8HOx?w$6z&)m-eCI>%8{SQxCDM5R+WWbscduWHd#2j-&L<*vqtxmhENb7Yd0C{^
zrM|B(+ifPe@&B^rGasD)<rJh~eemIjU!ISSmwCQa`Di7}z_!SwA-Qv#nn|zAv83AS
z*cB>@3{&}U&6KQ|u!5h#>ec*1diuY_tZT13bN%CN61%nh<)3JuZPV`WXRiDG=l%8j
zrs`?)(qp;A4I0;9Zx6Cvzjx;9GtO}*en_P@eC%g$vq_0o$lWFRptbS(+H<FVHm|i1
zsz05+-u-flfWo_3w`Qmn{W!XEK36^8O$*bS?1PtVFI`-xe@E(=k9c68vGuGgtgma&
zEWWz$S<<8TieZxeu?%Vp*z){$pZ)PC_Gae4HMj0Ob6tK^((JXQr0K<r$9^R*&YL*r
zU%l6?lnT?Czs|0?`~K*o6)Qf4Wd13cZCe)X8F<s$`|XMqpC0U)Fhlyk_0`^^hnIg&
z|FvV^I^)y)uiIDjbsYSXR`_Y6zD|`!%|GJ>Z(eP^_Ub{Ur0%@a8rRuk($}v|=stcv
zd;L{cCtEu^lPzCkYwhyIgeCV+I+j!Z<IXP2|E>H+PCRelu)ELPcGNR-vfJKL!Gz>H
zi!aQ-dieNLtBZg3s`c3i%Zog3eQ;m7uyEIjE0q>EWK5!;#53vXsp<F5KIM8}X2Q{r
zN<VgrX{=bWVw(58^F7ZwIXO9Zw>US~*4*7^^XE=gocyux<qg+<U#Ne3_4f5;ue`Hv
zC06|1#u@kDFSh!s=jEdx?tNVGr|*8=n(AKhy+XP=mYZJceGI+xEaQk=P00Ck5nIH)
z*rm9*xLDVgvN8*OdCSVo%9e1K(>9h}?4@IA@q&F90<yRK;NqDbd}az?w7R3iPe;cM
z3(5~97HyF!a%ggs`zbO}A!DD4$773+mlhS>6>^w*<iN@UO+|qhm(_2-Bb=?<$$rnY
zR6*f(Vn!2#^=iF`^?aK8zjp74J!~cZ;oSH5xgOWtCV9M<Ik`^#$lOf@o|Dwi%>7d`
zW8dHB`|G|xx^dG_T41YvN&gJ_gIAqC)ISsN=#qQFpuhLy-w)eablV^B-CgjQ{oLfg
zk{h^p#J;<*G2zC|#TSGRtX(iQB{Q>K#pcTkzs9d`>)!nDy=5jZ|M-#HN7byP9SiUB
zDX`!BE+Kxvv4+?7i~fG4>Wl+M#Vj@NZYaFf&1Vh}WlX$zl|>;XDc|u`34f@EUtCka
zo^17lz}*+yr|3U?#y%-tzhQ%I$wMu<lP(8O@dwxyFg?g`KP>(&TZij4_eZu3^`(hn
zCegQ?F8veW|Knn4`0pJ5A;#ICoZocU#I5|)`G!Ma6Q8j3vp39p<ONc`7*1B7aFZeV
zo4?(GV_UwS4DGw|=&+bX#({UQOXtn}uxi-{(@5X;{V^UJt}-hc@7ynGF1N>mkwL2b
zSHszTkL<NXY^|+XW_Y%4{a`e&TYFD{!`+g&m&d~Df(=|)93pj#rao<UpLBwK)w}Hs
z0s>NM4%aQ%__rSB+2F7xEH%9D*0vM(MeF@lFHL`$QgWBC>6l-uzx%=|N7!vw8-I9i
zy8F51m8gAL`|=GB?J=)iE*md(zOuJz<(sVzFBtm|?)=)uc6q^k=IT8!A1+jLHax(y
ztue1R$2@>hO5*s3-%NiQ5)V&4$>Wj8@r!pMhwOy|8k=U_yk1!Ldo7PMfBtpZ2P=0!
zDiNIecX?)HWo5CA`*HPUm-k*y(NK-noU1qGR{vl7hdBkcOFHV8E$cfq|MZPFnJYJT
z+vs1p)86+_UGT)RlKlOJ-?r4h{S!a`cl|;B<Nxj+zkg?6{<)twf4<DDyEd)<%+I<0
zH-Fr^o*njh`ZJ3cDf#bj-o2In++&ZY%CR@!;wIfwy#GuiJ8hAU<igFp75{217s~FA
z;J)J@sJB~iN!g<`t6IDH{>v}F{IbDTu5P#b`AJ7loVdU7;`KjmJNIe5sh(5%_sxxn
z<0chzm4&~=#N9U*FYUhLy8C(3Xa4Fn*Z6*%78k$K-Cfsx_vq1=R_6i(1vW}rbhO*5
zRkL-@Y<XAyUh~h3qM{vk^Zy-)-x+JZrSE&Ezt|`4J)bQN`@&DmntS&1r3=gNhs1~L
z*z5jH&p2AXN8L&LPXF0lX=`h4>p63eo;$Z8T|L$G^V*;9cwYW$n0sY!WoY%&FBdC<
z5}lpoe{Ny!lXPe~{!zDi?fo}rjvYJpNPgpz+5mt5^i6&Yw(nmTInFy?;^b^=UU&8O
zWc6jdk|s-*DnEPpZo-`VGj%=hrxz=2D*u%7Oz?|F^DOt{@qgD$y?TNrwaoeS-cKtU
zIV@yWYMm_OxbUH+(IR5an`6Bj+0%TddVMr!*gsdrH!Qbdo{+Dailsqg5EoPUQPEe*
zZZ9+r3m#LIVEe+v!^C1fRq>d*xA@|_``Sz_9`0N)NtI=yLBe$vw+*$Elq1r2x}0xH
zGMbpuY17$xvh2dyE$JKQc@-sfKD=3Fk^8SFeX;b5MWt$Y<}G4<EdAg(L*d6n*}ISL
zI$w5vd3dse&;R8I_2k69+g4=%T2lDxSLpu7vr3G2={qp{$O}yRF<)!?-#_KkcQIVO
z{7KmT;j9BMJuYtj^5XFRBbT;i8vLkTTpoOA{pHO(zWi~`mrqXk^7#4z4&Q>T1<IE4
zRy=iPZ?3;hN^&^zhHt^n<~M2&^pCH95FHg+<=(cq`Tq0xA8~)<xBa|5?P>d`YfP4Z
zKN)|0`#GmL{+EgA{<9gCP9_GeiFGy~|1^K!KOtDWdG8wKccnXR4S%UrUi|5p5q&|v
z|99Q|9pAG~9FsU@ntFfxL<z2*bDW#!c~02EGsCp`rCGx-8HXI64lnisJO0Ld9Ma#G
zy*Yd2#g8w)H09^+fBxtHyrlEH|Li~g{?%ubyb}e}pY~7xkY775dzoGH-CkZ^9?&Y9
z=a)Wjj(Ya&pUs}xZx(A^cAtE7_0*-8E|<)cS+MgTU-Q|k-|tVTIraAAJ{>!I=Owcy
zhxCapytJ@->89&D%YvR6%z5wqCN{O)o4sUdXrRmI3C?#u^;taL%4c<Lc{P7C|LRv8
z1AE#eKJDGFY4?2Iyz}ihk9@I`<z>@6t$1!%`VW~k;o4Jso+eqwtlNI_)b^c%>yNEi
z_%U{mMeE!XTU75{*Gt{~_usW^s}{#6vVXDS*;x@HYE!p%)uwy<@@~k4v2WyfcXV!N
zlIQZy14oV=S=zZPDJdx_@8;PjHa?!g=c6h+FRD!Sh}$`9!@iHTigoXp+0HJCy=!)5
z*15DdcJX#A@|eAZs(H)03*V^gSiaj?WHHgYy`}%j>35uq_m`w>zf;z~+1t~u;e5xz
z$)z_ByVrWkuDbX2<dvJ=f{sd6Kjju=JDyuG-6>_hKxN32d7Eb{ykYsNy7PX(EBDIe
z|8FZo4GkKcof9j9?#$>6D%j5Qy3BK<Lcz9)`ieqd<Th0FHg~^xmmj{Y=BfRyQ(0Nr
z>!*d>TDe>QR^F_svu>`QyGv*G?Xz#6JpJ}LFL-xDQM>AGBhSk7{0cK8cldJ@shj;<
zAHJ%My^?p^?i8B?cQ?-MQCX?MtT<sh-<b?Ew*CpS6LQrR6b?^1d(v^)qc$V+_T>z(
zO<3OXpLUyn>IZN29qlOIDbDAopRWD%(_~VV-R6kAXla`ne(S?KmwZYp*{2ZkWQqFK
z_7ubKD+KuEoc+)6r5m}ovmM{6AYf^^KY5#Gp7+1#(kIVP9-rU9za=3@X{#-7@C}pp
z-R16q?|$xnF!8v?g@E*@DmHqFiJLD}8*f&e!Tt80S`_=in@cW~Ps*^?t@dF5u<1mK
z)$ShG!?O9&Ys-EVyfIzKBXaWgq<qf}$70hw7pE%;oc`-Phj03KgCzdHAu_w~=J34V
zSYvgy;motGMj{Mbe{t?`xSP9YZF1eR<xw9mKieGpbKUgi%7?!eE|8B}xwSTA(qFOu
z`1P@`8^0d3*_yHF?7zCs`RZ>C`!`s$rM&$6fA4qC&oz?ww!cfaJ$!Pm|K-I->YvoV
z*+-wh_Q`JFv-*Q)nj4Qhif7K8$nSpeL0)O?{r6?jdMays*Z1ZLufBfKdhL;-U$N7?
z`qkUJPVTKND}McPedXGFZ~y+Bc<|e=?-Qyk^YW{!Z{P2~k~r_z{ofi-zP$VLa?|CK
zLVk^ke0_a=%i~kG&%G^uZs}w0>*eJo2P}7O-Zk^8^y>N9`R9YCY22(n9NWJ6`lJK0
z)1SZFG5e_H+tszVPi~vref0O<eYfx4U2pvOT&8d6j7`T2xBlZbw!8lJ){e`DUR!K4
z>_1uduM6Ec?df*YooBWg{$29!pO=)4eEIFUVfWTY9*zz>d?n6jt=8^)Q{zrXZOytX
z{NmX8Ua!NKtJHOGd?-?$5jUaSaV5)s3zb8En(y#Dc|3WC58vH83JGzJ&!wb3)YkT!
zl>ZUB*!X{G<<DJ0ZVc=frfsiFG)XDqsZsFNIpJ)_Ud(zkx}0TBz#bN}AJ&ZbRXy~D
z1Vr_pdu~u#efHVAC;vEPB^z0qbLwVDna!K|qRFtio9T~1hUbELt;%wIe%0K@wrf;b
z{-pj9Inp$>L1t0?gWRTBzN?<TGI{q+Vs}z9N59}R9#2Dyd9@Pf1HbO6^61sQ>mtiG
ztJ&{u<-&!#pXLkM|7yG5s<1x#|Ac6vA7U!|*pmB|Z_Rq2=J-IPG^BD?&AF7F(^({)
zYJRVMyWo?E;-%;6AF_=YnO=V8Xa2m?yjo5rxKYJ9>887z5bKYhx8$5Buekh)qxtAZ
zN81Yy@1znEXHN`W8v5i?CR6AFDU-&Q4qmk{CJXo)wHGtwy{K>yb(>t`(%rzKkacIC
zoGHWeOD{9aSIIYvb{!4-vhLs)-V=7;x7~lgN9k(JeBP|TFZWjM-BopO`{u@#t*48N
z4E!Hn<Ku`u^j_9@f6en|8?_*Pz8}}Veeb?svT;Jsf>$zU=j&gP{rjiqL`<tdV0512
zqypm#S$XB&EpvER#QyxZ-lp$zYsH@MXZv1nn-JD#7PH_xQ$ilUf%!%@;mSrHhBD@j
zOcDyqUGH8Ediz7J<6`9A0~PE4sop5}@lRHqcmL9lo8kuVc>kO;yB=Nosvvb&oW=3o
zt9jqot?SqSSfp~Y_K4nhcXxN!d+`VB|MO2cuCBH<`}U1#^~;05)a-Xn|M30iuX(@r
z{V%Lre&FfH-Me?s-}qbl{_Jn%`Eh%rFJ86(b#wZkU3XHJKUuhc*2?1<Rq3+ILi}6T
zt~qTzd-kl~X45S<JUo1I@4~!4(%;_{efsd}=bSxz`=YG=?Afy~T<^QCt<YY*xTDLT
zWuN#Z_x|ZW{xBovh4=E$f32?SxisgV`%asME+zTm+!bG<6E6PcUHdEU&g_V#Hsb%*
zP7}`lbN|V!&qli>Znj1Jh+o_ym-f0W?amLqyKz_UpF5vBvGM26$*v`(*B&j`yS02F
z|F)o|eyrtn%k!u7Ps!M=8GO-b>gq)=^|$Z;`KCX2c8{mYFY&$BYd0FadGe{`WZ2TT
z?>^4#oXdA7g`L6h(&5_N9JYBkbW##kU%5<pa^`jCTmG`gMeZE6AvQKPoEn{b54~u7
zl@*k>dadc1<4tuxOi$fCx$0wEP($C56HiPwtT}S)@e*c9De()f349S=)y)h(Ip*~%
z7&Ya7%-y(Ivs0nKd_vuWx203moZoTD9GB@TcFdG;-(TVIwBzT56WQ}KOlKds$G*LH
z*^3ej^RVS|%Zr!VUw&GUvaDLRb3zqgqBWzQPlSQ%F4vbzsY?rVSQ7S5-l-vE5R|fD
zM$?N6<z}1i9&0%NUE}j#18&W49e;LO*5u4{Fp}NYW__am%&Of*yLG3m?^lv$$huMX
z@TBs_!`WLGY~JJdB_fu2ub43Z^dg_4L&;g54;UC4&c8F7_RvjAq5RpC{b7lxS#5f|
zKA0TzyqLUj-~YUK&rHkaAASA$-!I*VzOQzB?EU-s&x;es&%I~j)Q^#W_bhPZx#p+O
zlQ(2k&JC<RRkzStOssdto;baI-pyMx-?}-y*`{*iLFZm&PxF=rWlfEfw^jc9l(ZCo
z^CVtMN1nNM(%t(>@9*WU@L^`iKD1Bl+1m^8M=PATC(e(Hd+?#c#PHtqJ9qitGdDad
z`dOX6&LXv3CN{$Qc*1e_EnoBEHk~^wkYN;YroPkr;c1)j-|vM(AH7I=Z<#OpbAKmq
z#gq3XRhR8If33}ui<|$_S!}+S>Z7QwSslk7l-;%Zw)ESptAFGl+SRNK+W6FKs^*mW
zN%Qhz7jtZyzI|!esl?iQGdRPo4{ve0w#fAA=apApWpDM3()!zHZgG(9EMv=?ulYZ6
z?ec53NvvIOulD6debmQ84-3|v-u?F5$IPF<SDvV=%RiB@P5=BY@#k-I->>@f=fmIX
zv-{I3&#zyez5dS9<>$1YUcC6ZxcT7S`SSAClIh!yZxi3X_1)^*x07$*z5Dj<+ryiy
zFId~|J!PK!M`o>j@2<C-CQfL%V)t`N!T(Il*wd@)=105k{rddtw>y8!8EcEY{u{?O
z*8V)M`$ccod-<dN#+ROYe3MvUa<Z3|?ZErR{dWyxm45#Jef53x$9*?n@A&I?%-GwG
z%cA|6^CIbc9{G(578Ol1n4ihoZ%_^SzU!#{-`WF9434Z)t6++o!+3t0yRD>RMdRWr
z<qz+fI4Z{7k-A}Iy5(CFdjijq8R=Y;lC~=@QBBR~(=lslQ=PPI?~D}Bo4LF%mDJVC
zChzg-G=6s8r%=ttW%s1_ii%73v>&x|nD=UZrQts5b*tpxFK6L<xP$XA-^Ka%^ONVQ
zMx=i{Fn?mxi!`SIgURbYC<+PHJ!1R%USRWj=WWj%E5D1r6Se60aOk#0+oBuhiaUAr
zrn@iVe37uLxxwGOIl<qo)%wbh^hzm#$IXR?3>SDA_s?dyKPi20jKiT8j`Bq7g-!1Y
z4z|Z0`oumdvcslPPI5+HK^9{`LGwI09bx|x<6HX!(v^2az7s#pb!E<k^O1WWnzwXJ
z<uH5p;B)03Uqhq7K;Ogze2i8ptJEj2au*5^KHqhFC+iinClhx~F7U}SnOCu?sytHW
zqoV$j$HEu){16JSS<c16pq4%R%!8WdGRYYoIn~i72JSHl1t#0KeEe+tdUb`7>T6-u
z2baJ7t&3NEU-e*$zI5BW<KKIG)<yH^N9}sCTmMkxotEaNrVa1z3)`jF_RpL%JO64G
z@B5=4_T2s%XIiMI_bZQ0>7UGwjvh|S*%N;2wO!8qulG@I)78e-8^Icn3+F!Dt8n=6
z+Di|ZtM@)LfA&36H|~M#y*+YKHNsET({tZ_&#aL8Z!UYZy7T#Q`JHd`_CEQ3E1~Yc
z@R!>4vx@j`?e+X7yZ_LyAmdtpgVz<)kKWmtwYnrTXt7(b{M{9oo4NUZ9bOrIEUPMD
z=CnUr*9w0g{Q24TwfW{{7yU0h{`jkFQN;c9DepdOeVO<>cW#mN>b~CvS2tvQ_?GW>
z_Vb%ZPb4oqfByXWyWMv#^`2jS{=BljZC7sU<&QTK+f4sv{Lue+|LN1GFTVf&ZDa4e
z@ce<|pKombirddW|NQk<UFkiWL<{%%9j{`_3TK}=Ga%yP6Q9?cM7#`krwG@7Tl=~6
zSNG$(g;SF?zRrx8wWBWXk(KhE=PL{UuW~ybe=ql9Y}LjW^B+I{{vl_5PT{-g9G{w+
zn(Aw}Y_Cnu-NaVj-h0?5Og(;2?kl#L6|>&x==&+3c_Vl};KLu8U_ZMvvz|?OnQPX!
zP})l-tYgFX)`v@M<nv=U+eQ6*;Way{I>A%ENiEuQ<r%@xzOVUSw@tEAwb`0~f6+P}
ztp>%YM~B&Dz8*J9ZxHM3XQ<+-x^iCgqfJ)ziqgyOcW<owP~O|AZ$7{MdiaUiXIl2#
zOjsCO)65&UUXRQCXU`5#R+&33?SAnK51;y?s<-M_OijG}QqvjNKS+CAzo{H)6<jIP
zf5BLKZ^#o4w)K^UQ*%%K?YPIgKw`VyzW>wrsb@cu$l_wVaM-j>qVda2j&~77KXcE2
zJow~6fJM`k2aQh;9**weP1_v0at^QWeK+Yh+0#rcAC&mVi$7dEg-d#c!iP2klg0y=
zUfM7w`06s6Cu;cJ3z>Un*-ukN8P92IZ)UGwsJHjem6n}$4NrXIy{%j(B``30ZtA$1
zGxZkx>G(9s1N)*+H5>WsO-~gP|NYE=-qv4>quw+vJD0Y9(%JQ&i>m(lK9648<RyA=
z*&DfP*{SXN+|^I)?CR`8`}cpnv~-{QVgt#2dD&;5C30WzVwbVGVzK_u<Hm-ZsN2tG
zYN>wv?Ym6Je0yn13fsM}GP}iGA0?kV)y{b|<BE~Pw_BEVW>Hb`&HI1Yv)n)U*`9s+
z>A!Q<hVEaUa8~EcnKPXt`tRRgiP@QV@6D+xSAxFQUj9>5&we3S{Kk9J;y<(h|M_@&
zM!L7|q{pgL(oY6er3$|{i#nUOIB><4%rC$4%~!`Ch}_h2rTWM2n6k@UW#668>|A^F
z=+?u#4|})eTsF6Pd-hyk+Ouc(u3g`mot>Q-zPCP5c8-nN9+~>BAMgIlKRx?rV#!%s
zcH39;_Sn6WZHo%?dwJ1!XQ8|C;;Q@`-+uqJ<o|W>--7#b$xLnQ*OxB;IP*Y-?%ZE-
zbKd{kl5n>=Zr)a%NaKyGKgg<{dG03PaONdH|Ni##$=|JS6n(w5-fmM_K#N+8x%k_e
zwh!GKA5<`x$lNigVSd-_BH6nod!D=eb#>E+b9?09-l>t~dy~?9MV;@Z{EB1k>hpZo
zuzj68PiDdAum7f1biNTg^HWFBcHyi29~=L-msRqH`~IGFLWcM8fdYo)B1?&$32fU7
z4t<U1ku8{U=wZ2q!zpHq?5h0=U-OLK?(IFtqZs`yV7K+3HS3lgI%Lvq{#vZ+g-q}s
zcHsq2q_hRzMz+`oFObfwnNV;zcO}aQ#;G%lPF=E1J(PWb>&KEGrJr_29+9$rp)se*
zx^`d0rlR7>LjN9~&Yv-}_B3<;>Z`l{Z#o}o`F3)LLcy%}O}uT-lZD-`{krM6;f7Fz
zRlvV^jk1k*EzV_HE?Ky06+`aZZ*QF=1mB;T`S#oI?F)DroLek1rB67h&dB&(A^rdU
zI=M@Ae;JviUd~y<=D&j5<74!z?s&iZ+{_!FZjQH9_|T)br10jOxu%s~{>T0^-+Yj|
zA+BCSpZ$u}gXNJ^Y7S;s$X3`EXSi+K^Xt>pzK=U%Rdc_#PvfgSubc8tIO?7ZlMi!1
z7IW!x-s={|mFi(j4|tv5aP;T9b-yP#@x7XB7|}ARFTlK_i|@ha{dOrCdAGA<T=b5$
zwPyD^UD?$7aMQ-P-=RCNCT`wY{m$xhPyFt)-xg1*mip@cecg-d^Y_lVELl>L5|_C8
zYD?0WX~}D*vraMDcRP8*q36GLi}B}YhptIj-t)5PXLN{#ZPsnQ>2o(7Uikm!yEQM9
z%HMqd{rBJh`nK?m;?-4Z-<Ixu3j%kh#jd?lUS1w{*KgiCzc*>KtE;n~6i&BTqhWld
zI?Ces&F<r>vb=mR-=`b$_BPF)IdkUa)!}t_EDF}&x>3Dtz4?V%&)188Hf)KR@L2Q3
zvA9)#&Gv|1o%Ad1xkc@IpEof!Hg>DzHTgkcNqc*HeEX?9neB;}+Ae?UoS|MRwqV+N
zi3c~>KAYZUIP&M@A&&J5rN62^IJ^`;RLn1KBXzQjZRMm=&))jOZI=Iwg)5H;`3re}
zDOS^%60s@0a{Boc29JU*7g8p2R9?tE!X}}lu*fKL!G{kEFR3#)3;Yd?eztu1nP+>|
z1E$Qb{<62w|5=S;lm8jl?^~*swU5{~zTwL^DtJ+q*z<C8-3n>l4qL+&cGYFScCD*D
z_@=-ApYv70wgYz>ESfvCxaJ@HQnM*k{)zj9Bl|K<`h}wG7OKoEaCx4vz_aOqy^!~R
z7D4@%y)QZ$&pbRRymIPgL8t2d2FwPszC6<l1U%bA)R|sh&pZ3%<R&!_i-Z*-yIRx^
zi0oySmM~<uuw3_WDZ{h&%Kr_A7wz^db9yhhw_z&Z+LQou-IkNKN21gi=2RANHZU-h
z{0dF~-*5Wx$NZZ)>ooRrWPI8%H%Hvhw48zOV@$6e>y|Elb^XxOPOnctdAQ+Vc$mzF
zS+l;MI{jhk(pkH2X4u4tJoo?ie4S76Q=Lcg@22p&T$}WF#pRaVKQ;;7ru>Er6%x$u
z_43Jd%-%isXu+2nzr9zF&dbXAyzIA8@9Q%X>um0w`)FQPKl@d|`yRdDlGX2<<BraI
z^vURF&CVY;?`>~hw&UmGAiMhgf9$L0{<^;VKYR73^5{Q0`@epwuI;Q(`OIT*y3hFU
z`Rb=1{>}P&Ipp)#ty{Oo=Viak%d3{IGrQJze);p~={awvU%YYY#tr}E<mBX}ua9eH
z-z<9{zUSM=*tpi0-`AJ!ir_f9=g)cPwIXj9UMT%8T%0Ydu%Y;8-M=oc%(ag%ANxNw
z+4`;M<jbM+f4ME*^7h83kL5xyzQ-rNdoO+b@y8#3>TLY7UZ0oddGkN9`QrgLjwH3V
z!$)kQR8D6ZR_~qL`0{(h{s!Au&I>1;jhBv`xv^*Ap|>uR_53Cz-;G#(GS1B@^+Ujs
zV}gJG&FE(O^>NL5^@=Gm!t?xhnQ!EHkkhZbfBM_njo(hb|JeEdOXn>AvKjBbwm($J
zy(W1&EJ}Unk&ac|QkznoS;Y6eaCfx(6rZ%<(9fD1>=OD5YuFcD6?|dI($#mbz1mbj
zaMmmH8F6Ryq&cMb#n>HvAje!GBGa|<kAK{|r)6heSN+(@ejwu*^8-8Ph)So`B7zL_
zqxc_QJjPIZzcTZq#k~9p!g~wkmNM8ryvlc?<nNEVo+TXxzZ|X!&GGOkSbmwoj3MEG
zL6xq67yG|U=4{3Mo6fRuntL<NH@(1f`^*CK{};7pJv8_*VeZx5gtzCjS$hk9JKVH>
zbjP4iK&5&0$q6kivpalR!@ZAs>WO`w@Jr1;FKW4F$=PRjo2vRA8#l6Fn8~-jef?f_
zcT?xJXLnWom%jU~xN_Is?08Gd^BdIKSRIvk0v_Hz-S)-MaUw^-1a^jk=H5dO%eIT`
z<viEde)JHZK(bS4+47hFQ!8rhe^yp={C}|Me*5`#cjd11XEHb(l)C>i`sq*aKlhKk
zfA{Nt`MU#O_nO`L-+SfW*&i0SZtY6X6u<rb!}|B@F3ncciP&gAfBCV;L4NsrPwUSp
z{TlhVR`~JLM_S_i+wRBm?K)}wVSV@R-fg$Dx8BNCdi1eSP<lzhu3F>O_N!)<eEGd*
z{@vQOD_1VQ`|ifWFF&;N-)ryPyVtwA@_u%8b#;~5w@;?!d-n(BJ$v@d?%8ws{$2M!
z{`gb9k2~@B<rhExly%={?LE(*a`SLZLwePz?9!bt-Z|gTEZaZr?2^=rRc~x0Vog8a
zt-83ju>Pd^&wRbct5>gY?mBF>YuEerek*oV-&t>?{_tD<vABf?UY@m3KE#}-aPmi^
zK;@n47KJ=N9!86=hZ@<1Zxxh%+<2f&V%LV|x+7&ZiDe4mEz>RsR;*Do{UW8>oV`(X
z(nDh{rV}>XmdsV3VP?>{QTeA5bDCR|ukZ}t3k3&%m?`e&afzNk|3ePb|HTY*dKLw$
z{9on&YjQ)tiTU=YzjJ++HO)TwQM$BE+up-Zy`$b~eV&T$6|*f9&PZ)CaL!iQas66s
z+9tQ4zWJ-gS6n%gmGoeqGrRN}j<Xph3p!8r96D8OdGof;l{q?dANif`R9W7;_}1H0
z)ezU6%#)|Y-g_r$8J^_dR=xUW@IuFhr&>HxQ=E^=@b<MV;SH@(=9Cc2c>2L5ne}A>
z`-f8AyZ1K<Iqq(nckQ&@(of9qSyWG&I5AWj9N3queCb=|KIRkhl~F%89Q)|>_8wc6
zMPADd!)()tw_&UQXPVv7cYXBX|Ki1U@mwGNE_na_?|+ZWzAf?{b_ePdKD3|VQvBhZ
z_@nNHEBCdWo16dTH~(I_om>8VdZc~$%rl=}7g<jE_26tv(08^YhQc1Z!(Z<8bJt(F
ze9tqEgVWih)cEZ4*LKaAF{P-hGrgr?#*hEyh0AZafBSECY}>Rp|EGt+U?o2{H+Rt2
zr&;H<wPrnZS%3QepM6g2TKG@CSAJfnANl-X=hKoIyXSHImEn5Vtt@Z(e+TdE9k%PP
zXM@g!jAqL_@6Nh>?d#twWLCx&@0NRgJM22k!}D{um&-c0wd?DwX|AlS%=q-9#_-Df
z@9&GiAn)G0@6~h4e*F0H&7x|bFMsGpo0?~HdAWR!b>0u3-FfBS)c8%ZbI#Apj&$A>
zcRtyCr%hU#dg-3|`S*Ws6#kGNR-`=5y6VHalL?FUVz?O@4)1PgXlOW{ykS+AmW|Gw
z;>M_p2cF)r(7PM4wz7VO<iRx$PPGc^EAie7p5Q4fA>*^7@n-ynNhcc5ytSIjX1>UB
z!6NHl4K@#Jzx5q%D4#uB`NAgUT^VO?Zpz%8xp4EQM8l>B8P4pobzckWKh?TtUl3CF
zoa`sI@4y|wIl8&b3A`+4CMGIY+}=>eH$UBF(I0^scf59AVQ~M%oXPZ8q-e&n4u+3{
zDe@h2>q?q9O@E5<9?0!HlYg;5MRdDM)4tk*;#ty940z6DTb`C*&KNHgX}z8C?dy`|
zj17Gc(@KB+IBIOM@@{v@>dLuIaT-<|-ey>}yqaw1H?RI*n9PUg+MMMxJ3HU$>}fQt
z)(K8dEABLQk?L4uCZ%G%@S-~76t(A>&at7*9jbpOJ^4OyvYxM}2D5=@Q(AsXirS|i
zJ`Xt<SHGOV(Acc67yEMK`j?Wj<*(W%Hg`zO>s_{WbwB@h=KGHyzpnVaaQ~OevZ`<8
z=dbN&aKD$kdiCnnSN}Z!b9dkT!|y(ae%Y%OcJRl(6XE>Le@a*UjCNcmzhbYg<^KH}
z4=ydLsQ1tiod0M3<+UGI{tN$K^>qI7U{=rii{HzV{kBg1{`qs?&dW>QUY@UeF1B}d
zT}$2Dmi_vD*PmF(pFe!hOn>X2ng5T5+%NuTU&_7JzR#@I%WBu8jDT16QEMuXnSWkY
zX{($%nfvPNU-!1PFP;BJ>-ySfUJ}PLLFa>S@HM$~T4&RvY0*VD#b;&CcUd?6t?$!^
zzoN@m-MY(`djG-9?}59r*0rA7_F>&^)e_m~ao6^*y;yLgBwtp>+<o!I7hg&$=7q~I
zZoX-nmsxx1kB!fK;pQV}UOKOOTL0*DtA(qOk1hYLf;NVy3|>#aM{hV;wtmUm66Q6h
zE7|<E^nIBxvR?eq%vW6L+8Z+;SMM&<f6S&))ycLm{qVC7Z~ong<xc#}eDH|ix0(mq
zXI~UI`7qCuX5cdj77$-Kr%Lf6huwyP=CZSExO{un4OGlOmNLBG$o+%uc=B&&d;8NO
z<weUr?q5)1yV1evYx~ZRM+9Hi#P6`Z_>AYlpZ;HQ)9s)0ywi}oEl@n^=>ErsJwFa~
z98H?HNPgEYiOwGe%g!=8Dm6B{aGP&PoO$N`!kvnHySlC%Z1d}3R-SlYaLy#b1A?m;
z+?Ks~^oRgQbM;>KZ#rqqbbDLGa?2LH{PHBW%zEAL0{M@xTKAvVT=DfmQ@vvIvO}Q;
zt&cqmx8Bm3`Rdc&kcFo2ZhoA7GWyc9*2l5WYCl<Re|39yxzt?0$8rLSZW0qJzl3Bx
zajI3wzwli$g#D`5geT7VQZ@>gc;5V;tDwcF%)bBhZr%&&dY30OtS*!K_wM%e8&@6%
zePVv!`$9N_JLR!0)BED;JBxA>at@sE&0V%^)-v<W@$rvKr~JO(esb}L-`~m&^Cr*w
z*2_PWPlkg(?mxq~ySvLCZqDAj=l|A|CqLMWtXt;);@AFHAHB_NHlDC@{VtQRGhq5w
zjj3t#Z^`_SejB!0TX9me;d-g;bJ`~RA0Novb=Rh@K6T=`El*;8US4<CS8!7JZfWPE
z&nk94`+7Tf-L4;&m6es1HFh$!|Jj@KFJAmmUTJA)QfX;@_w9c3{{C}Ei#N=9qqOJF
z94}8x+mKl~XLww{8@BuIaG(A9Pp)zQs**YVJ+{-umZs(jyh@L{VH&%4&aCj|UH8h<
z<F7q%|6pABYsQZs-(|07+kBUom!F!`|1v}Q@mIr-ZvFjtjs|b~ej%>O+`UbUEuiys
zhlAUX!`J>L^z&tS%5ZycGLF{umtV%4n|pA@`B%ml-fv;#->N)6#X@G{>n-vvRTj<v
z`y0%2yzkEVHT&lMmhMo;&!TBn8NdGX1V2r;Q?H2lVZHD>SL`wK%(Y+G119u6tozLR
z$MDp5t_NFMCQiKOkX@<5R>Zu}e3nMNLf&CsZ9c)uJs0(_Uy563xz9wJA^S;@M#rXP
z*;Q(DTie%P-_QE%ZyjeR&jHOU_64lwZ~s(y-di8OvevKMdYS*F0|`<|pHsdSoP2WS
z%vtZVKIToEQllHUFm8+2o^nyKrgTPD1mCAWz7++I&S~OH)~h$y-`)Q$@^1Xk%Ms;8
z4!&iMn<uwFO#Bf&bA?V+=_7%bUY>6UznF>KlI8H3WL=x#t>u*NnQ;2U*VY4$F>LpH
zq!X^TFTW(U`$XBhC0)Ni6iIyJ|LY<b{kXxxEwk*M#G#DBM$fHB4<~&8$Gc9L<-GnI
zuI%jnxq9!#OMd0d-2TPDW$D9kTbF~&ZoWCxp2{LEckh{riFw1GtYv}0du>w8-uwOG
z{qu~iFt4bnD5+Tb#Brm?^)HW|{>DD#q{U|amD9}Q++J!gy_^2sdfq(w;N_Rrzf`|H
zyIAPv;vIDrXO|^wPRclbcjJ#gb@rYut{NYOTYk&0{#x_?>aYFbul~x;U-;AB^>2N{
z(>;&-wtn8{w)|_&+&($B=&bKMLLWTXe$aC6(PfIOFJ5{1BR%ND`-&RdfB)9kYj3>&
z{`>ELnROfY%$e)w=Xdd6xI6^NUHq|sZ}{?WN;Usm{P(|qzxDE--E499Z*4xl@o(Zh
zy~|VQ@aXTIH1+tVE5D|F`*q>huit9l{r8+yXy>>W|0hH`;fT?rQ;i0GK4w$)iewja
ze2u6O3&?8B>)=~r_r2BP!w;XQ|NJHR6mRiKUk~Szjrde#{*YIO_5HQB&gp7yN%Aa*
zPl-q=Sk|dFTlqz#&p4#MwcFFl*QbDIh60Dihu)UHrzt1pq|ZL`iPKPXG$>)_^D=Ng
z_rBd`Laf0NmPzuzK7H^Lh+DiZ!?U#G(V^*kHoPlPxTz5R`pFXOmQyS~Ug~p1)K0lv
znqkBDX4c!AG8f*ya86yMkuH$`;o9x=+}qJ?0i`=_I|On+G#``d)kt`o`fdB47t=1z
zQM?h+yz!2Fi?z(WSMOxPR5WKS{@t>&x0CxsxcWm~<IQU#PMwlA^YN-J$leemdg4LE
z^Hb_f!6EGojb1;myKoC+hUuKlJ+UdFBXwuvcE%nS1_lQOwv7q`jMMiUp6h&hk6Cg`
z<<tJ9V*K1+FWWt^sTW%s`0Mb`d71o2ugLp<K3ur-U;R(9?S<cMSI56^m+!GSBrZ}Z
z#hoXhmCJ9YW97K+#rI~Be_}efmgW6hAG#{ikiVU0g>%yf%aCl%ZHWu3t{r9DskT98
zb)JdhyNk?aF_y9Wmen<{y%E_E+8(OZx;XUmxmA-wA6c&D+aIxi!IiQVrm}C{s{^{d
z7ag~FKI?t$&o%q}xc>H^xSRI-?|XK^>)Y=i{JQ(~>lx*l$3v&@lC5jsd+PM7U%&2z
zUNWEkb$?HNdHLGkVr!FM{3_i2eP8^S+C>#l3sxlVvt4`NJUjW;g#3y5`|rR1ek$;O
z`R=KA-@Pd;uH17kIZc@7<o+i~pK98y*UR@FJN9wMIo^W1>-MTVpLK1M|HatTk*_c0
z>|W1z|Gem<uM4l<eOEE@p6)Zj!mnQ|zbE{90XliKueP?l_LxG!NAqXQysu(cI`A@{
zJM+$!{lK>`oPmMYzLhpSI3gS5(r7U^eQjPeH|Mz*A?%``*ZsWlu&8F?<dc=o8hrV7
z-uEZi%$o2?z+%%vLr(iAJS!E~aHcxf^EI&eXjCR`tzjr<=sM13c3M<Q$WVHdi`|dI
zeD1&ZNnTJYiCr+6=iIF=W{!%}lh_MR9`i_fDwn+X-H{sZOFyh91%1A&w(EPf>O0r-
z%U1?@7MTi%dUzE@3P(ttXS88#+MqHeP|WL4Yk~H8cBb?-GgmT)9NFx~J1v^|<W&Zr
zGe=pr*;g^PPTPM!b@^jvC!R(f-?tyHFiv60yp>|D(#LM3&HKUonIp?-y9CD9xm#Hz
z=f&$En7^Kl*HC%Jt+vaGs~$aJGMF6q$W1uCQGI_HH>cJ=_cxlRbBq#~7c6O06EE;`
z@)Htj%wcGh-v9me*Mn<HmmT+KRcpV=c!+C{orJN&`X<2;hWK4+S~cJOXI?xISbXTs
z?XHin!`Lln*wozC)ZdZqJVAUy(Vfg2yLY^s);i5U?lkwDrFSoF%`@`6{Ej*Q#qG_?
zX@&VI$2L?PN|5=J@;<FE`OewQ%@UHCH-5?gs_vZly~1{`oB3<2qB?c$ymsk#7gql=
z{#?9W=lS#Gy?<`(JKoPF_l`@a&QJXMvU7WGm#vu}{N^A3v)>=TS<m;DuCjmrcbk4$
z{<S~$r%LtNyT5A2?D-wG`o~9`U4Qdp<6r!{bmePEpv$lHd5M*`wcNgKEi*H-d3|r`
z+x)mM_wBFtS8IKJ8xeXx?)cuSZ|;+yZ~L?^=lA`3>+e?g1^0y~U#vcM{nJMwF}rE+
zs?FP#xzAO9ab6q#jO+X6Pk|4G^s>d)<n5hSr+3}6&1~J|M2E(&SIX<Z1}{wA_Aj+)
zAHV(Nkbkw?UZ2R}*>m!G(v!;zUb1~&xIgdHVbAFj(L8b8J<@sw2@kkt&)degJixZ0
z{^LFqDdWBR9<m-kE*aX07~Jb_7o6bz>tTgH>w|ZzL>Jzw=hOaS@T9PD?cua7>;H=X
zEq*Sxx%eEX?7hG5%^x(+XIJ2j$UGoupEiGU<(~Ua8}y!>Ej!V4X6k~kCcoL&?m6-}
zDY0_Go(<=B9yzopWzNAZa~_2FtX`W@`eB83LD2!;2>~D6UOBvElApe*f+3S3$m`6u
zAJzhm@=~Xy)qm{&`@eC9$=}o&dL8z;`Mck}s9wBTaCbr98m`j^Jln1qK3vMHRr2G?
z%b({G7#I!eK9_%a|BAVox$41M=0>G?zq0cdZuGR-x}W!d`gSiNk=;Kfdp2nC&G7qR
z)!4uuen&X(Nhr%Q`|CS7P9*HQn-lM>eEs&%TAv?mZUQCiZy7zem}@gxuUzrXiwBOo
zUw)~G{r2_n##-L3oNCP(xvW|hi@QW}-t*5}IAQUU+fE5ToA1AAllOjElRxqFMeBrr
z{dX>3PB6%L5KvxJzu*3w^!f5nD_s1IVnC<B_VsAkJ6C@zx2?@<%H!Yrb)T)>Gxa^~
z_om-S{<2#>`%O`G9RJ)OzwG-Lc*maeU;FQU#j5zDf43`EiF2QoJ5|dm@Fmgp!h;8^
z59B?cw4I&lj*Oh$zi;2(y?giWeevq&Szq<zzP-D<yXaxK$i}B;AJ@g~d{W;YY5n#8
zr1Lla^Z)aDlk#8uUx?iQm;dc0tNvcjV>-^Rx4o@7a_iFckFnd%@NA!dy!g1Or`5%S
z%j0f1r~NFyCzkp3d&sX3cROlpAMV??@8!gO`}Tc$_VeC?&t_i>N)#X6F?hJ^{JsNk
z92{a5^bKTJ*h$YfFp}PIpRJn3H`$@M;$V47(xZzi?@k>1lf1>qwm-kd@Z^-xciuAl
zvkT_?nf=em?t9+6FlN@g*#%ZA|8KId+iYl9vEfehrHAHT75y`^7L=5nTTolfetVDk
zvPGY_9c^1Mb3#KUe+}Q`G!el(i{%zOcCmk}Ntf-JyyM1=8GIsk3^N%ngv4(?$lTD^
z)?Ur=qRd`w>%+%RJnxs(a{T#H#=_Uv%ii%Vq3+;w2cw<ty(f)wdBPX${%A3&<&j5h
z${`<vBR(Dp{kuNw{2p-l;pXsy6`2eQLVlAo6VE+ozu3KQhK$u4?M=eFrDDZ&EEo@P
zueyAr^02yS*>9WQU$(c}On9=KXGVgyZR`&_^{-pG#X6)OY_;5QnI-1v`SA899FJVJ
z_zE)C7To?dzumYx@$*kT$CR&EM8tQtnD1RQ_tU(jDdy>OwB{7oOS3FrwNypV_Nu>h
z-R`xQUj{$_{M~2o&BKaahl}^v`TAYhl50F$^~gE@zhx0#Hg~1P!v*5PJ=gWe9=KiQ
zAHS%g-`hE2Hrvl%`y*}U-~apnzEk5fiO@CK(!bw-|Nb`r{hRmSmGa-eU%P+*)6Dn3
zb<4khG5`Mk*YDrIf4>p`7p^_^)YIIZG3%s1{Rvz8ao5i$AO9X)?EK`$chefyIG#T~
z?U}b{@6L`}n?GyWl?z)xe9?*dWcd>eURC`+{rZ=6NKNg+@7unXJqY|SaN+`={*Om<
zpWTgL_)F?2zx<p&k;8)54c-NDC_LNB?|qG(t&)koC!;k}QZ6#$<We^EX$nu-rKWUN
zBz!4;7NUKSb4}OBrq1WzProop|8czMMa;g1JO?y17Q1ib<vjJ2spIk?f41_5htEY8
z%(D#YeGy+_;`Q<4=3=jpKmP2$eNgw)q^h`S`{%#mw_g%IsbPU_&`;GWCiR88GR|qf
zUd!;kM%|~!wxDK#hx7-ooEa%^7$^N}`^Uk=qfnv$l;P-QcF8|B3zyk@g~TkLeDd(V
z2TpuDc&}>-`3XpT?-nm)ZZTERI6C3w&mT<_SYMu!v3kg&;Gn=D(7201QT6lfHS->~
zd-BQnH7K>QZkU+)>HoPp9pQJq-{&qVENpx3+P>#@yFw9r?78QK60!mNq7TMNN?JYS
z5HOwpHNmCz<pP_E?*|GGe`iW(JifK|=P%`lImP?tC*IiGf2pDDo7nla?Uz2;o-fMZ
z`7S}>^PTU;y8{gk4K4B&E>52*Cg%3;{*K$4!QwY^+fSZ6Rq#~5+PL(KM8>`IN`6fK
z^ADc;7_el@YcHcOOU<i$&i`9f)qU%pjOpHx+D)hR7Mi`expwu^`AUyY2Z#FWT6)e-
zwDf$XHpw%iGBaCoUK9U?w+r{4d|>@m-l%SyU)<{QX(yAdey?&Z-g&F5Q&n3_%XFXg
zw)Rb%j5i-|=v;f`|9i3O*hl}bPK-ADo63{Cdmr!p;2GuR<$cFa`YgCDytG3^s-d_<
z;Dz~l)~;sOKaK5gOS?C*CmG)7@`;^bV0hhR|L<n)sjEwhR1LBVHxx27^xc!OW1r_H
z{6ir3FVpq~mVX}9Cs<f6Xm)(hyiIuF#-qZX>uOH6-ThQl(736t?$jjZm;<*&BbNMr
zaIL*DNq2gO;UV7(cbR1~UiqJqZJYf-<Z<-&d$vc<%I@w;)`~1Qe<yFd?yhQU!_)VD
z`c@{cpWQ?D_r9%6X16JKd~@Q=UW4!zcIL5r3Leat>GS+_@>}N~<v&Uu>a#?JCr;^k
zam-=DkFqBMj~dx@&+xUg95?jv;c$P`*ZbCoS&ZlAyUf!9EQ$hZ>Pu!DoE4tJ#*qI-
z$j?IM^86WUOP;(i)nO=<HIP$aWO-rS&8PZvm2~HV(4!@%7QcSV*>I|CUHhTJChnxa
z9Mv~PWgM6o7&&+w9JqY0-9Ewhf2ksWH($JVe$SG`y^i634MZef{@Zk7okHfzAB^nZ
zd7mwwBfa-}b@{V7Q$vJJ3mB(NT6a3@XV7Iyfkrif8BW>E=_Oa#=Fj0Q51Ko*Fi_F#
z%rRlb4fB*gzfYJtz5Vvt=hZ7dn4VqHeJ4#}S9HtUT?W0m7hi1p=`&fB@0L?{?Sx~A
z&kHJaSik+Q-4{1!r{~lai!Ll*l(*na!Gj0?ymR)SuG=}kz5kKL&;8zyk9_<4%KoN(
zl7ATM>d;j$Uc5N*IWhb7kAJqe_eOrtTeE7Ku&>IbHB-LKQ`J?O=yuC)ex^u@QkTZ!
z*NeU$vg=ur{_@k$xc>*gF5BB0{Coed%J^`m`Ny8TwLDQ)ANzEBb>!2C+4*xt@9uA7
zpS$wHURz6BTib$rvawMzcW%wj_sh%itLZ+<FWu$uz4DIN#_)TKY-T=YuD|dxShim5
z!Yyg%*WGpU?$1q>4H*i4xt6H&z3SWB-~8_6gLfakR_#ta`tBFwyH6MHpAxe3pXhxz
z&Ml2|Ve=ko#?J@L9P53I;`{!uljmkX-&2#&#gJZHp!7#KH!nN%_Kg7jMtchjAKn9>
z?Bs1!N?BK^CmxW`<hvq}JKy-hE53y{RV-|boDVhs+hV%l%<pr%*rnBuN#6_pdwU&U
z@6FwxZvWSxKJUE#r=K-@)`j%NZ2y;fVRp|FMuP(m3=H=e7+8Fi1X?y!F5U6fq2+h~
zCF>QE*LlnI|FgI6-u2)^{THi0*CmU0o&CT6%b)+t^Tbx`@1A8LqVV)(=)H4{4`+D1
zf4ld^&X+2m8dG_v@|LyVx_b1pLw4-O37=NqSo8I9rMj>%xA6MT&a(fv{;m7>d-|#5
zjTgoL{)>4&`|4v&we{1wn;B1>-m@noHd!Rz{Qc|ek=s`0Uj6vAHeB0y*_Ev??p8%U
z-MV+*a`Cm-GW3>ReYqhjYnI*4<GxmL6OTJsZ^{n#HQ)Jo_VdiymrO35H5a(-dku7N
z#YGUYUB7ts>+<6dHTn-n`LA92dGo4R<6oD5ZK_}Y^wsv@U;jjVpG>Woe*I<Lnz_RN
zW$&NteEaTyd-3moZI25(kAGC1E)t$lzA8ld?Tf0Zd-@#ie9QT=_|9M5x4Qe9Zwv2>
z-MaLUgv?8p+n>%oUbblJ;%;rb8-JGUythCv@M+FJ{iCn$hS<qJkJ($EFrz-A%GPn0
z`WpF!^8IfnnjLt#SXCnDP>hkn*C_|tL$9{4)Zn|Dt1n{t;Pa_{RZbJ1pWm|Q?w{KU
zUa|qRzOtQSG28Xu-=FWx`;%&uV_&kF8GO>puc?3d(7ZO`p8dy)HMIp-pWaE>|3zG_
zXU_V0;*Q^DIK7GgX}jXyg>T!ser&&fXH!M34cnsk5rt0YE5qxb3r~IbsJ1Tk#q1TI
zA6I(x*TnpNcv+suT;<?PzW4*42bkDDa5z<&|1I44+xPFqcklY9$liamVpZI~>xUND
z<uSS+D>&TBCpkI)%(A3Q%Mau<oiK~NE6pNMe?v<$eD;0zxZ_Hae*@orkK4~Cue^M|
z+AK8>qsCL$*jCS+kZ?D7N%Knad55b%y#874CKn=Ln76ZVm06ao@QL}0`unZd8{6A0
z`56%NU8dH4`?p^fH*W0RKligv>E8bC-+Ax9y^Q%@_t1Xl#QQhiPX2Cr>Fa*>m+AX=
zTzy~sUS;~i>0xK*{IaWmA2;!rIrsk{UvHh|-=eZgO}1>=EO~zSlGfYP-o>mf%vVia
zv}V()RjXY?zg!7)5z+Kp;=lMymF~Xfx8BA+`}}X-dGnbc?es3U|L=a0>vMSf^XLEm
zzkgu9z2#Ez1oPbUWlefsv)A1{|N0*L`m4L2RK33L>sGz4yS?P`)29MT9M5Dn2B`2v
zFaBBZu*70o<GSFK5Mc*l-q=qIUVhlb=>5n%tmSfNy4VE4smp5Xr_aed#MWXJUU#|0
zSWjHXkFR-w;j8kt2cQesw!YZ8S^SYmjIxk>6;FXjmFty~NP81K`#zhC+g?h(;TQPZ
z;8&Q-6Z*2X^owr58<}e#e&om)EmvooGv{2>6{cq=x@#g9aJPlM<STx;RNeN{S4Ye9
zO@gc%&yJhf^i3&I`X%?;^5e|YmY2RTE@pAaXwYjAXJ5j8iRm@hljup;OKn!%U9qc4
zdedzW+X@BlT}%azr`}#?dUpLJlV_0EBsVkp4!)HP(%+X`8T8NLZaFbmS+1mS^>Ptb
z1wEFP=QiFlD9{kjRo>n|$><Zq)dgIUR)N>{$-GwevYh5*&$Ec(;?tA=xqtsl(wN!M
zb;Ybq@MIIi(!lzbmJ7vuY@VbrF)}nTx~ynX$Z3?m^Fej{5qa-#|88Uz)vKRju8mzk
zMJVq7yKVh9&!0cbuGzBX;H89{GT#!vF~6$Kjz1xN;oKXG9i1+*1=ljl5)Y@IaP>KK
z=yg<c_;qfbQ$Mn{Dt_R(k$h;!<k-4B$NFNOT4PV7)t;aFrEco`OE1$aOKQK^+MVCO
zc=ojA`G?le?rbl5y6Wre4U4TpRtD8wlFzbV`)ytR<A2{avz&WX_eSpchrIZUziRJK
zIQmLL`PRm@Yge~BseN^MUbFs4>YS*<YggXhxNM_b)oGoRp`oFnMWLagY1(?1R(%nD
zWf~e<^<k=QXk|iCQTgsmKkomkv624$zy9~8?{y*;x$IVbN%`Ri?jL+nYR4n<^IXjQ
znELh7H=bNK^vt+edMio%w{5>$tI@F|vI)O!tfp*t&Tf1yeq8!-@ouq0l0WCJ=B|6z
zdPv;Mvqy_Z<Hk*`n*KY^m(xE@T>1U)`xAcGeT{F;wVS)~xLob5ihR56^##E^?DaD~
z@7sR;Wq-^1ht(B!VXD3TehMGIm+$+qf4TVV>!<!G{tAEh;>X_Hg;hVU_y0Ikli&C}
z{Mk*uY{%z0><pGHKbYsYi2b<5x9h>PW(&KWLOe#X47c**X4O1=&Nd-4jeFT-Zx-Rk
zscmJv^({6Fr|sG0W!NTI#q>j_PxN(zN3KPlv!CSL{q|FNb_sZ>c%FUg&SECax%+O0
z0+R!ilthQ(MuP{k7PEKcEGnyA`Qxtb%ishS?R>4t#Wi<?>@Ryf`(AspeD|M9m4i(Q
ziVQ6b`xrQa>-K+YIq;{a`Dp#?h5w)be|o3?N3C;O%#(ln-|ns0aQ*Duyt@~61=L^O
zyZ%+?8eXGXEh*U>_ElDq-&3k{b~n9XQ|Xs_dgt&9r&lElo?m>iaQ1$o+}+*h&sK4K
zRh+?jcH6Y_lZ_ufth%~7Y)-zKjoO-Ujg6lprG8)g9TELM(|q;rZ>wz%E|yL||M1ga
z{q47o{WATh|ETug`sC&hr+&z#<wp72$-JESD}M5)nmw~4#b!Oa^6gjZw8+$HkpbGK
z8k0QFZGYmiF>mkg$}LgR+n+>l&%4tW_V(Acx3|wN%zt0|f959(3-i9ackkSJm77y$
zYX4CDwZ(t^7q?#4WZ1l#&+^Z#{I=QOM<pdC-dBp#i$UPcx5B^k&c$~>dHPZ{K)>0^
zmDRD2Z)*I7&5c)mREpU7x9wP=-)Qk>!3yh#%g!b?n+m>giZzPMy2Ox{Rm}QbJ+h<y
z?R_EH=UOrXuiMTqUs_nm!M?IDPQcj0H0A%niM&%JjTj!>J;K>!Z|~3Yz~}y~%>~AP
zZT%hkCLf9Gy!LqWM4O1|C#GL#C}6+0BLC~&fU4Kb@3>@t&TnMBTGOZECSvV)@bfMf
ze+To1T^pFRjEn>nEo-(uit~Kb`|ROw@$KRw{Qr`Qnu|PU6gUd6;8?sss3raFm0zyi
z8WuGV_SfIzuMq#XxP6b{zS`{h20AQxc>y>0XUOcGzl0%PAhY>11JCW-e;52pP~bVb
zrtSVk-f0Iia^Bqi9k*XrH1rW)!k<PP`wjN5<TnXB|7Y0GQV`IVqrG>*+m}-=CiUNm
z`y2iDo8$`f#K+!e-xpTZUUgVuy=Us#JnP-*#|}JSvE=XkPnY=n{s`URf1`ipcf5@0
zZ-of~zt;X+-}|+m{o2i&e<WwloRa1<UCky}Jvw)`$?4OlPaZq^u59nwopH-=-TfVJ
z{os#$>5QMddj84T@kf8(zU`Zt*1erCv{ud3_>fa4q48wl@|nw*Pv3d*Ro%bSp?or*
zi$nY``l?K-sZy^_m&#nW@0C>MvRRqSD%~Vsx;4uz=fD0-`TmZU>#GtU=KuAZ5}WsD
z+Uc?<{}()~yzFI^QoTp6THb#C`Kw8mZ*BDI?dMHSH@;pJ>w6>W@3~2VQ;gcCO;2_x
zD2PulJXk-W?#1levhs`EYxSN9ybD>cq9@1tCy{yb^A^qG?UU~v>(HC9x9|TW{>l$5
z^*(mJ=8qorACD1#_a^S@jT;%elYigs`ulX-1K*1*jO#yrs#e(bZ$pj1`sY6)nT2g-
zEGoj-8)Oo`zx&7k;bR?x^H&zW4;9QF6H9m2z3>&<_x^PAjXOVFo_=8e-hOEM7p9bg
zov&&v8Uku8HaaQP*td(Ph_Mw^U8!zjus!hS?*s-0hFuK|%v~-8>hDV*?s;<L{^5o<
ze`kKZbMnXgn4H7keIyneKYYShlKZ`I$?{tkZKs`87qMm?);yGPokc<Ou7T@a#W~OQ
zOy=^LGpq7R)HZ(k+P2}J{%uX|zD~144__x%?dUFE9>4Wn)%D&jn)mPS{%m<mtIqh_
zCKdO7$s^W_)IKly*q>ki`&7!wt$osQ`FhTVqQ{fEY8}tKRJ+jAw=#=Avgi0q-=!Cx
zfByMrnYsRkqfuM4beDZw^X1R7sdKNbjoJnVY1_7yZQG)ocg$_ed1>M9PVxQ9KmY#!
zQTOS{UthQD|KuIN=Rau;TVGu<;n?cyiSM~S+wI@9ws^Ly>GV9Q_UY|UKi%9}U1!hi
zxT|krK$GMix%86bGQThNuZogCz9TYz?<L-!&(4~f&;IIhU|!iwn{(gSeq-BjdE9o%
z?Bn<JwNITsxbVq`FNdeUExccQ_e=GT@9!UeEIqI>kDc#Dy~G~<z80DP$4lb6{kX20
z9{$q)wY7eMSqR_qbLs^(i|o|Z73Mg0$iHoUcl>Rsdq}Cr=PU10lLIcxwO>|W`KoH}
zz4XX4{W^IIK5I!ErEgd=QT4@)gf`WupHsx+#8uz^{90Zpbb0IJz{TY(bN<!aKCnr=
zc-DqRAd7kVfp6>zhJW4mSrl3`vfXc-#J1pKHuL&7Toy%)j6Cj#Px1&lZfNIIC}esr
z+q~_N&E<Igl%uup)`~_S(_eOhS#j5$2)E{A_bQBR0uHi-T``vXJHu>}r|DuQ!`Phm
ze@}nQ8hom+Tu}Y!*;8K0TY3Nf<i5HuSALn<`H}gr`I`Py&U<HH`yTNr&~2Icl*jyh
zpZ79vHuKT=`0%}7%E!lN)<?gn<&y2SUs(C2rq-UnmQ&vS#QtBlzcy}saen^&iO#1#
z^MCoiMZILRO~2-sowbWR)o-dybD89+GRafrjLN)&Dp$9^jNF`77Hzxkx@q>-UANx3
z2T$I5|9$$s{Fq3qIro$E_ub!g@#o_`A4>w<gB*=M{`hsyUUsb!=e>rs`tXEyd#{Fn
z6HjdTXFS1x?NCGMQMb7_CO&yqaB|bVe0{s7G!@3pb9ZnGB<_D^`8mpWPl6xc{k*BV
z4twk0|D9A(QGNRQkIMNCQ#gOwrTt;rb=OYe!`ENR7CHS5e8(0)sE{izth~P><n`Z=
zhL`yC-U`;OcR%{GxxrGC<)6Sl{<>}3yAK_kvE`^v?AkK^eQ)<x?8@<Y_x##<8`b?=
zwq#bVmlLRbvPHLf^9*G^3$y(4{y2uOpP36J51e2Tv|-r&fJeZ@uH2dV4u|A}Q!QdO
zOx#s+Z+CLAe0#Y;_u(Tq0mtP_n&)LC?8|k&nJc(;x8-LQh34!<H*#av-E*t`cYEh;
zo=Lv*XS4m?cy-;^uQO-=e)d^?NdpJN^2g!1*Ear`myz80Lb+1TI(nC^pkd@=A>Ds}
zTy0kWl$`th$u67EoSN58PcGDKcu|@pWf8vK`E@{HH^0$bzAYND=Idg<shc=I{hMO9
zF8=aw0o&jIeWy=-t^G~x%!8O)^EW*CcK6J$nZg+p)$g7CaqrugOTWInyu94=%)cu)
z+a_K*W2U%lR`z3Glgqxx7Ix+PT&uHO_843=?_ztI{Mz5-a&hlvGs|VOAK%#fBc)FN
zD)Y)8_UE;K?TWECe)oU&Tz%zt|5pC2_g{W#7URUF%P+*{D^F^6+<2~{`jkLy^XH|l
zT09vwo0K%t#GCyO=ih#1QL`@G@|3ZROPk@HzQz9}Cs}pW^Pl>*-A(?+m)wd%6ITC+
z_hnbgFuoJdKJnl~-~C5>>ueg2-{$+WFLM4N*?;pzXZ_!%@U8jQhIxF(#r16)%6Gq;
zTl(SQoqtI?ckb<I{MD@gwXnFOIZ6NcpS#of*cU(Nlefs=yYXd*;k112`(6<%FZ@o9
zp859cmHdA}Rr9B$^<8+&b5Nv8d4^86@-}(?@K5ZAcUdpCnR`+A7$e(<j<STuum1OU
z&9v{|IDyY!dU98Y$lVV)Pm7vjTc%IQoPKhh4V#K--|2NhyR9rdb$;K|_V1c|hq1Xz
z-qps9$MkpN!zsNSNq!SIOrMm{S*bQ>UXZGK#-ts67Zp6e<o7nkci%Yh_H%(%>ANd@
zm7VSSywi^56k5#jXOsG0zc0_6_rvvlJYRa)*15T{xw%(vocPdMOxgF!^^)TJjDp=~
zA7w6keeRN{)=S|xXO4Ub&ENd~n(yQ2(67s1WMpJyRLv54e&&4hlV!6qf1ddHcT3>i
zycyfg=bn2W_i*R=By;Pz=bmma+;+P($~44dX>dl;UY(C?_U6kx?{9wGy783t={)Nz
zGj7fLZ|)zwQ-8;0b^V=p-WmU??K!_9Y(ejxn9Qi;_5YiBJb!&CX|LWJym02X_o=5|
zod4F+&3N+USI1AYlD>)afBkQuApQIGhnj+C|K6F|ef$2s$S<1zdkuT6I<JkLeEIdK
z_c-hifAi$uu+*>Lq_W;vWIo^5`8n~7ht5kMy*c%c<lbjHf9+S8`zTU<`PVAB<xifx
z<M+Pwr{%rbG^wBawy%BGFP>AjzP9e5bnc`a`*5xD_3@v$j(l9j%nV-NoyFFxuFUh|
zj$hsn`L8y%J1uuCxX#F6zHm4H%7>*7)19;UY#+SypK&99@1y*(A3yj^1+08IlAY(5
zgr2OpuugdDwTWkzzp~Y1OM4|+vtKV>bG_?SmYXh@_dHTm*=cCV;^KE-`}V@{i{=-b
z*-rPnyPNA?ov*c&OU`oD;=Mlc2S4PvA9p_c@Okx4d*P>-o~J!Y)ZZL?c|EWBq$O@&
zly-tH61``ZeCN!WGv_9ImAt&X+)vi3<kscoesk@Xxms1<djI~@m&zYJ&y8XaK1<Tf
z-)cPV<(ua}<XYtT+16{89z6YiV|0I2c<(+>z2A;A9vQytyMIQ3r`toz(Cnw{)R0G$
z-XHodcsxI$k8|_sBkJ?~!Y{r0SUHI$V#im`{zISq?z2A@wy9ij=f8n@rOKA_kJ-7o
zx#_Rkk48@MRp&F@-{2Tyr?G9i`13yxcklkOSN@z#e2MgjI>sAktDbM((*IO#gT#^?
zo0QiHJ#zQB{!chO$L9UoS7l$>=cbg`@a@_6zV@Pm#N-XHZ~r|P*tY1_0?v%8y&vM<
z-KdQbJ9L7@Oo#nGM_jx&&+`b@ojWEpFdx5=dirZ~h7iZW{Wrh8e78+@uHRzet1n+R
zHhxSoTiEZve5;bxhE=7!4oTk>rI)<XdX&~P|I3WUC6&%|AI|^Nuf+YynV<Dy*>cIZ
z`{fsjeW;GRzGnOB&(`@rYc?{@->QD=n_S`6J<oMRW45NX{JgZtrBm2SM#A!c^`&=Z
zyY70-yMN?sVo9C2?YF0P`P<)C{qcM1$9K+o?bLwohrjHtpVz$aU%Ggv%PN6iQ}#xg
zWIPF2D81ZS4RpP&NMM|$;Yqbfl}WxuwL;m#ucJ3#+?;&%+UAKDbAMmh@#Dwo$KQ7R
zIbpr${_LCm-9K&YXUyZjIr(R8_59Lh-s&ZG^Gp4C-8&BiKKM{x8!Iax8q<FBQq=bM
zPhb4MF)z4U-ksb2c-sAP#`VjuDd$IT6L0vkm-qX_6=HE&KkCig_vxSgX4|vx_151f
zZ2uhkRkUxud3#CR!Q<SY>&|F||2Np@ASK5h^Jw?;Qx>WlKeV=gW|f=YJp0yPrJqc9
zUU>=~ls<Buy*tKu@3sf>HWA^Ml^;E>d4Ej&pZK8&xwa2w`>d9H{Jn6`O0G_aA7$T~
z4}SdeH{m(^f8H%C7(X;Wd|XrUAX}a#CY~jxo+VI?`_2!g?K!b87In$KSoFT8u_iAu
z<&R;?pRPsE_ODZ$uY7usl0d<$e@8av{jS}`>HgzWiQVCj{{mG9r_VBd6!$0Oqk_YC
zoBzRAuDt&>Zy8_1yXwfp7A31}RE});uz}Oz{&|rS7QrpX6*iHl%zxJY7w&Ib<sV!(
z+w<~rzuAnxzyDpYfBxIw{rTGWzVEZ`@4t89(f6L{!#7jk3so-r!li!ud*}A7E4#ze
zzO7EpD%tU;a<-QD)GLj%7YB!jhld_pw|?TP*C5unb?es3#>VP8nQ~mY^89t}`eX0Q
z_x{eV(f?Un|6=v`g}M))m(Pkla#myyd%XVb?+-WK{wlQMu7I46P@>}$9zGGr54xY5
z^X&>JC_I)Cl)ha1Z{^)?)}G_3lMMdLzuT1Gv2LNmLhI8;eU{uB3l~W5`dH$nepOXY
zZhQ2a|KAQ@d=m1m<X6U*{}MaD)#p^&H<a#{S@_|YxK5e9#GA*(ui2vRNyH2E{I1=1
zzgg~u^5XT%Deot!{{BCmjqm>~siY+SyG-8v#|1Roc2p%EchG0gu=sj-U48V~KS`hZ
zp6z>kf65z!-G86U+6&C7HO$y0+%Mi;{QZ;kyXxI?=N>Gn-XnKn@BQ-gAM!a*y}Q}@
zDJ{<8$MOSw$AhPMFFqX}Fms;V;SvYchj;IvIwjp;W9`qF=k?z2>-YHPzJ2;%-73!i
zc^}u4r1sNUIQq$h`}a#drwDaP*EJv1(K*y+!@T~Dv8}7Wo0DPEk!IG<&(FvHkXfB~
zH#qczWpv2%_-7w$>elz$-~3ikYN~T{rSQZPPYNymRP$?JJMsR+cO~a`t&o+GMN5;s
zrfP+(j60{dZT_?HzxpA+`z)fCe)!A3_UC@{)j#)}Kl^k4{l9hBJI+6u_@VsY_V9*h
zhBA6<HZ$|~tPXn1ruX<2t5n=#rq*{HGrq_$UJ6q<6SLtP!_w&~bwa12|1ACIP!x1^
ztz)<Uxr1pR?#$n7RX)xC^y<~epFd~K<NWFQa`I*Q$!*_1Pv(+|Yw*vt*mY-ff=@&J
zfoFc-l}p!fJ~&<8U>QDhLIdBo^x0=0eXKt7m#v_I;e8#K=<@3ecMH|Mdu5+o@%Ba2
zZRUE1nh9q<eERu-SK!CPd$Sn-IG#Su_rCsMvxe#Khxhq!h`%e<Z_!Fv%p?Ev@X70k
zKOX+v{e=Cw_z8|hbL38JaN#aHnqzM%?>u?xy~$dVL51@f7xXtCQ)ihKaE(1@j=<YV
zIkPV@2%mg%y=Y$XlgG}_i=-L9#r%)g=|6Y<r(s>`fvw(w`cuRt-m5ylSa3u9llSW^
z@fY5&ST}6^Wm5a&%NOI<zkQ1Sy$Ed4KU(;1-~9P6Dvh5X>lWR$?CsXDwR`{6?X3I1
z|GN9q_n&^sn@c}FH`^@d+S%Dgf8zC}wYAFUe0;p@?Xq9RX=i6&pKX?Na?-D&rQk>l
znt$rT|EKex{P$n~#s2f3+kNlZEp9sYUa6S8W^(K<)2z~o760B!_nfPin16HQb=@~*
z7JF0wHYnW{Nc*t%QB#chjhaK#R=B>HQ0DA;Mxoo;^{iOUo`ZiRGwiNbeLL}u-)G*J
zXIZ-6%1ZT;*pu^jtys1}KEihL|7&vFG924y$uIx%y#4L+Um<qM3#u}I?0)b;bnj~e
z{Z+d2C(h8{Wi>x1?cYUve(w(-c#;a(_C>te@_}#KdcN!7mghr{J&rS<WLi|)V0E|e
zc&_^Wmz-8&yO|qGnlHMWe>*wx2IIQ#7t}-MzE*i()P75;s6c~7;YmiH{l{C<g=}Z<
zM3p4Izu#B>`F=szU;DFPg}=S|U(x?@HGhxdpUlIre1$px*7i-k7~nnid;7{~`Iq;X
zhW+2_n={jJ(V@DVv(0j58hC#DCC`89#jluqIePJnE4?3Py}FyZ)ZX1){NbhJzyId9
zulo9W)y~@Y^=Dh(9%_DCZL%_Gsn=A!(^erF7Z*EkcdGey^?(2Goo0JZf7}1<Q}?eh
zT{-r;dOQ7t4Ig|fEKXPE=j9&nU07XtJE!b+?zS6$4PTbcs@=9YZyV!1o@<vg*QD)#
zns@i`>c`2+;a0~gWKO(gUfdP-V9h^6kvD6dw!IAC_pq-pv9CU7ZgAOG|3k~pqLZ&C
z8h>Dmh~`(h+P^6D{Q>cSRX>>z>v%j?{l0I{@B0$ba^=bIs}^ovaf$8NaY4Hl4&Ns{
ziGOnM^UR9bBRwyE)xlZUb|uX>n3K(t&l{-w(DbmI#DX`m28oYex3M-|?~{1=Z{GAD
z^~)~JV1D&H;lr~JYoEN!%gZxOojUb!|GVnBcJpujn$OO6d6{YTy=GC%t)<0Ff)=|U
z`1dEpc6mr0^G-p9UnUv>3$HD7YPs=wqw;U#<$iOm%BNeOnrl^(d3j}a!Mfk8>Xtvw
z<jKifJ$>Qrb?<|BW?tU-XqQx9xbuC>_eVeflob{KrK>kTbBmSKi_?8-x4!ul?7DV%
z{Ws@r`huQKlJZF&skY34Ehne@==z@6^2Ba}thw5?@JmcvWuAB66>6&q@YYvl5xU7R
zXZ!kBQ?<fZYn`y?W0?7l<=y<VfxZdBrz1mao~~u`SirSNc5-vQxj>%T{rmR=XU{))
z;7@9pB8P*BfbenthRl?%8LL11VBh)2#bW;auR=xfKkNFv55BnZC1<nrq2o~}Uj4kk
zXvG}GFQNW(-zs^|<EW`ww_SLb)!({3e|A2(vrvETx!CIqj~`wewA5?z<D=cIzrUZG
zYgzR4aJyXi{*$jW)?Iw|qC9lxnb+UH6#c%u`^TjRcD2Tfb$__^GBJOD|J-4&RmscC
z%g)X=%efg`dt&xPd-YR)Eq{g2efw|rlUlP`w~v;azmEudvV3>A*PhgOc9W(u{P?)?
z8n^I9#npy=zru|tZT4BoTX^u!1F_USGYh{dI9}ynT+kmsp(3rkHt|-dOq2KeHAmTg
zzp40Kwc=_<nf00Is`G!hy?^v){yJ?l)?@nX687!bC$RKyWk$NM^G&WfQ;c@2y}iI6
z>mPo32Ui%Y-HOc)2?kqaRjVRbzPSATag}|J-NL!Hd_SrmzHgitzDN6R?cAp3mmL<D
zxn)XSG8}miJig7FF{gI>!5uO+HTE_k7A!B4-m_QlJ*L~pz}7$g*OlM)r;X3QXv$q_
zr+@F?(>-@4zAe^#`{a~t{<eqxt3PZiJZ5lAVq;-pecxxM{=<F|?=Pu+kLy?@(Ye`u
z_VVge(|78he_r`<>pbg%ySvkLA~qa;r}fV0%IB?r-u~WGqoboEy))8h+0&<~H5C(P
zZByLx^YcxI`Sw@m+tvKMyxe!TS$5>+v@<h}(yz3BEUMr9Un%1D+VA$}Z~u5|cxr0u
zw#3Bid9L2dvtMC<Pw3e_=d;Ro$Jkx-yr@?2bPI>Xq3{5!Nk6y7Y;oMS*kayImEX&v
z{>?LN*)Pn#!|H0^KYoKv`Mw9_@9y-T@^G=$PERMF#)GfDANpV1{N=54JAaazPc+kw
zZ$~PcE@ZH8|L7_a^!M-j+q}h&4>GyqeWI&7-)K6UsB`VeI?d6z@N47pW9+*xv~Euf
zh|s^_5cTi;+6`&P9>!(lFuJ}Ac*$T<xb@ya`{$*<{!X)<=bqku@r97NXoI4)a(#GC
zZr}CS|Mq`2jT6o8k|<(b|KLsAs`do6J)TA{cK?rmzRTcRvHR@ZC4q~x!sl!{am2%J
z|N7b2^8fBrS1sQ4PEqlOxiL3e*`JDwp|0JcF&+W(=2gLQ&%2L4TK(w2<BG*`uPUR0
zes|p3`!wh5XC2L@Zkt-ZPoJ9VHFdJ)lQlao?b#V|W75)!PoTQWW{n(Qx32&9mvvm*
z>hAWhm;YtI{^$P0_?+_}3zznOt$KIsap#On|CkwH8Ts56+x9Oh+jwU|!wbpVZRN5|
z8=lUQ`X@G_z{Aq<CjW~qITv~U=1%y$q>XLsF?-J$dq384J9+rL{%i0<ezRiU%KbhX
z+MjN2K4Zjn-!F1jqRH*K{&hP_t5)89%HhwnBYW-o`Q`D~BJ<-D%$QH?yf|l9__n%?
z)qgHb+`e^(0{fkp5hkT<`#Ada4#gHfzPpwE_}}FG`ljP^KX3S7zGZIx(@&0}%E#lM
z)~uTow@;hr+~a#s_8rdgtBEu`-7mYPTrXxr!ol`(Rh2ibCoZoJJ3CwbSM8rv+jrl$
zzy0y%+G3-Wokc>Cv*Wg=Us@fuHp(?IYHMz9ul3A1$L|07uYdH1UvYi!%Wog1*4#9x
zU;e+l{@JES+imhM{*rncx3O<(+rj!>;YWLH-|}ZX`(j*d{kgcfIR9V9=}VV>x$KWv
zF{@&}iDhvzkLv8to)gkrY7Ty8>7QYgRdYyOJo=eXpsKJ;h36xQ$2xxb0);lWW3{6k
zmBT%lLzcg{{V;z*-19dx65AwSJ^OX_uutTp4}TR){;pQEf6=A8;EQwNiaoAvQet}S
z&bKb)mmj?K|BMO)uYsJ1KuN#!z4NkWRi}-o9r_viDQv^Lb2fQ;2GhSk`c`|N&*6IQ
z=U)r`-)%iDzklnw;Ln~{WH)|zUpdv~r}p$0X=$mi^!4-Gx+?nb=kLlmJJUz5=560I
zIX*r<wq^B2`Jq{!6E&{PJ5|!LZqdEi-DPiYndjZzU0Qp4?Yq03p|!WShlT7edV2ok
z>Tmo1{C__GV0QYGTK<J!3}@AHN1ol4yITHzn)J*!)iUoFCh_&gulfJFQ1}J+gUZ;t
zHG<(AUw6N#UAK3&?&|AJ0!|#6UZS&g{?*0hBoqedT9tlwT)gM=0nS%eS2=IZa&$0h
zSReeX<oDNo@uztF^+F2*A|6{zILln*`dIqm0^?6beqZ`60-6#wyBr8)5Rd+QroKj?
zj{giZqu)I9>f>K;ws0{0`}fWK`McWV4lgHIay4FMakf%Ox#S|q@%Z7~$_HL6Seh38
zUHI?AF<yrPxd$nH3iGVgzf|djl*n1!FFyM?aPg@n$Fd~~9xi4rzT-FDyL0aBOQ#nu
z*uMGZ)lbs7@owAYw=+K8wt2bl^oz4QA@>UxL#Ua2DaS9VIm|TnkACKN@5}$k_VGUp
z3O4Lo5fK}A^O%^JSheZOc{7sVop@sU>(7;7C*F>Tl-;e466`k$yquXBetbTccwzdE
zK>z<0m0i9@c}k8}Y8PAn9(?u7txA5zrc}1xx!X#7ZP#c7FE9F@`?C1!&s$cCn*@?$
z+>`|CGTa`X`MY59!%U;EuXZs$WQ_SXVd`bR8KHZ+GRxv;v9(nxdU&p2<~Ok25cKBO
z^r{1o;*JNz+D$ubYq0mz`rMp9>$RVjGC$=mzj;MFR7gY4{0{SUi{wwg7C%cDXsG2~
z-@0SE`9${gudDkvSsiKpn{s4fXM5{{wO9Y|ws!jUA&&hz)51L4h2QP$eHXmtmAmlt
zQr0>DnYW5>y|tDP$gq7VS;C&b&HC=`-Fv&^VxFIyYgur2b=ca-&1q+6o0U}5ta~qP
zI&I%Nk^RPJK3j9!?E1Ru>T17v^Nr7TmWipUDbASDUubA(XlS4oyIo&cm{(0sPEJm$
z#>P~lGXCHH^~bE2mhP^v4sGF2f1he($-Q0m(%tW^X)mt7G(XsU;QQ@K;XBX0IW?tb
z7k5!#M#hSdXY^8{w^*vGEt6Pc_;dcZmdO`17PY^dmo{@>)6{tj*GT<)^e=p#nf}w?
z4_E!oQ+aQ6jaP<iR)ym7<=M|$x9pXlC)W|RA#}FAja@~okC^cU*J}(f8^1O>R^0m`
zU&>|XIIo@Qj@|}y@AkzP13ILC)xCSJkpH%63V)fTja_q9tnb$3h*HycUrnmYcFmHE
zRA`zRYg=r(D{hVPU*8D@zy6;LQg^>_?AYO@b^1r|zxn)W*J6K*h(|SY^6EcdbuA9K
zul<rI%UN(olqJho_0EsWPVY8#FW!Fbz@FL0S-Jn4vNemIoa`07JYCOr-}0dF_CqIy
zYwnw$`L1*3>RY=H>6IcGKaHRJzV1^IUtfE7W>AWIw^Yp!J@q|vt=G1<v+2q4Nk&8j
z8q7=7yK=DN*YeLLb@$A_99enir(u-!x{2HO?)|y{{Og-P_a|5TE}Oq&&ziUh``u^R
zxI>oxKUyt5(`d1D_F4XGhYTA|D5b}Bp4sDc&04U3<&k%HUVmKFzE3vdkmkam=vo8)
z$n%%^mt{HKW?mXSP3Y!}+ua=zyKb#%DKic!cHa8MfwwpR&7%HGe0mB@jVZsj`LO)!
zao;6eqVgh5H~Pdc_NLpr_cIthZ)E11oj;4`bo8^f3px94o?OH5KRk%@x52iutu6}O
zl8nXgYR`Jd3-6Bpd;efiFT*^^<;I^XW!dZn4<CM)_WGmo-MBK=6_$*GF(UsN-Wjnk
zUTALiQM#2!&tCJ$%Wvj?bls+&m(TA{e0@fDcFLQwU-Qmyn$Kp=K5zPr_^IvIkNl=>
zF~4gb<+rR%%?7*^;F-_N3p0&kb{0MTZmU`J^z?N73)A&L+|$!P3;ipLzx}@_fBn<T
zpVt20|8rLL{c`Kaa~~?7;c?IWteI%699<gN`qAC_>bn^+E6#V@+&ET$;9{TL<6W#T
z7f4R@o_|Ao*1jwDlUYx$ILT?HaN}{^-}_I^e$Onp+<LvbHGQ9#Z^Fr0>LGLE|NhUJ
zao0jtH$p@8uS472&>SX*P2QC%&CCA9ojqD-pdxZG-*xHT*}JV{9Hi^E-Mf%`nr&hF
zc?J2b=^MZQ%fJ8o$L2a6-j6?S9~IZMYEeko&wpr@>7Rb(o1YZc{qw*6Q^0MHU{`uo
zdi9?9dl%oi_`JZQt%bc`c6;7U<@-!ce_rjm_fEa}p!UJ3#pY&j&v?w3ygb2I?<%O`
zjnBHeJS<0ZW75&CGi#FSo`bHJYVTxCKmYT;{q3Lo?Vo?#&CCDYUj4xz`EX<7-R=2r
z-q!r}`1ttSmTL^39b<aKA8mhEJ-<vg$+|xI4fAT2h}B8!^iRJzloo%oO1L$?rrDk)
zeZSKAX|8p5V`ZmFDNAJZ-$?V7mdTs3$ahiR<fp=m+^U`zoqzi9sqB{khqX2_H`Sz8
zRB|l-9(2-u^XK<9-`A`yesF_LeDCLPN4H7tTA)}d#{B6~Wzg+Ed+hK3+QIy&yn52F
zrtqQ{yX4vL1|D5wXx+84*Ejy?pU-v~sScBBe|>RRs8r71oUdH9)zWKz$(=1RYoqpk
zuiks_gnxB))p6Io78ic)&8x8}PE%`++gtVZ_4V~uYooSmg{<_Ov#WYjpKoi!???Zh
zM+YCa{xb7@)xU2o|En*&eYEP||Ih!qqWywWo*X*)v2@)F6UPgO?sGMV35hH5&1v`$
zW+CnI_Nq7whfSh*iiP*bZPTYXZoVnwbHO+3+KUwq-K=)soxj(AIU@d=f0o|y{D&1A
z%X@z=*mU)fZSu~aJ(nFBU)1cM!5RGMw`1ADgZ2l+4tgD2AH$rLUlXrldtuW5m<R7~
zzrQ_|`^I)IfxGQ1gqiF#rC(ZHl$kltUoD;0_#EfVnRDx1jB4I2cS}oOaD_$XySMa|
zk0pl|PFuW8JmKX}jnYV+X)ZBBHS3$MziDT9uKZ)^ueufw?muVN9Z%kP<Bea_&-)9W
zuIA&o*S>J=trN$yH@}GgxF<C0M8CYAil)Is(EI_oi?$5hQ!78Re_v=N2pGCPecXTV
zb=~a!J3qf)urcoJ<v2f?f4fh$%IS4IdgI&2?<!h6)p04m-1Djs@qjzme>W^N*c+mu
zw_y8Pfu+)Ow?#T8?p^h0n|O|WvbN5_)%SU-%9m%<ot(Ds)Xft=EZ4EsAC%)bpT5&^
z{=6?N{7e%X|DWP5G;@E~5^DD*-}oKVgb*VEiKGh$Y!hbv^kEA5rCqYuu<of<>AINR
z`o;G4*Cwv-{piI#e}``W-Q(Pc{>iR?`1!C~#j0P{^#vUB7_{p9%T_41oL_xaHt<{f
z<HZ;LT+ciAuCzU(wDnTm-0XwDIkWk8S{%yZzxI98ZsVf)c4cpGo8{bG7qc^uzj)&v
z^X#2rpXW|{d-<i1ipNAHmrfyJkBM?NHCA7<LY7+}{`_fw>5WTgf-YUUbdD$J-rJ~o
z-IJaL82D%yzj@Qt#9>ils_Lj9D|L{sBVDY;EMeybX93X;u0wK_skJ_f0}fcm-*#-a
z|6!0ldD2avh>b~qFG`2M6Mp{q@#8ZyKmV(1eX(Fe9ABYbrOBks%!HSsQ&vBVkg~GZ
zm{w}Y`u&8wLe@%ohC9qfYyH&njBb~${kiSe!7nxse!k#-ev@NCN`ryb1A|El%pHs5
zmn44sw;|ZZUVig0iy}eOC+G76X1qB*|I5DZ5B@pD#VjjN-+WKF(=6k8)!i(|LslE)
z_Y^&SJw<eN`1#kBMpk<xcS?#@?e#Mc{g+=RtiDAfpz2X8mx%iL`FcJxIsU(Y_kX(b
z?|=T=pZxw;esZqk`On!_uj1n3O!jnER#sLld;2zq>-yS%8jH@Je3w&gclvzNogg34
zg1nf6E6WwC8jmZz*|dIJXRX<-pDvq(zZw2D())7M(`AaUe*1^J&(ich^xU~KapU^<
z#dE*3Uy@}Al3G4hVTrZlRo?0sd5#{Md}0|bypl#&@)|$AimkjSJMH!Hsi$6kYAHYV
z>3!%if#u)oPQ|VN+@IDlcZJ#pj`;h9JMQEe@x0f!xNt`!^U&_ehkJJ)?6z+Xl5XjW
zJ=@O3b8~Z=_jE2uU7{Bga=lhQ)qeBVcb0#m>p%TFcR&0|)YjlL|9cLt>yU_lH2sKO
z)yuCs4|jYxoj<wNCa=bs|K9S_%qMC$x&>@%U#a%3UCeqy#z{x^;Pd{qHH9<fBe@^i
zt8qP;b<p;q%)A$ePky(Qdvd9`ZSR^ryDyx#=rS&5+n15aC%13kz9p*RqVkWW<+cmQ
zMSToPnWQ1(Cp&+hUBxkZn;Kgdm$N(#zWZkeOH^KbyuQz9ckQ<O>kT~DDSoq@?Ife4
z`0K(QJ$Zq72Sxg?y39F!+dF1^-pR>g+Skjr?ELeu-hO^%_1@iL%H?^Vs#Oo?v+;27
zKR(_sXOlW{?TzbyLu~mU|MmH@URVG9|Ih#S*K9Vd-G9GuSGsLW&B0qG93srh7QJB-
zA5K>H?ODJja4yz;vSW!*F?(j{!~LJvMcQ5X-u|=x-1Acv$L~ui?vu+C(XU8%YV8%P
zQ?~i@=g-4;J^5vOCap=}FIGu=nb%`tuHtv+eAI@lpSSMaIC5DkNo<{+-QL`XO{UjA
zHGKSaxO?u0V+^0a9Z)^k|6yB#|B*Rjw!dl;zX!ej*Z)ez;zVwrJ5#&7+D1FRowl3e
zC(g80s#f`Y@{n&*eo@u>`PMNn&$PYudsw!+joqG|>7Q)Xm&s){Gfo_U+%LR;%G011
zm!dr<<;)aPiGA_^`2U{zAHU6371z#BzxlRQ>{opHv%33XpFX#rJft)4#DBrP7fq+%
zOS^K(WQ%zHjk_JK|98e6_{J@FbX{bh*NI2{<yMvV4w`*V&giL>6;$(Ie?gKdP2%~>
zbUmK0C7(ZkuzjR)@XC44uX~-clX!joyEpiD{9v56?_lKW>L;sipH_YJ$BD)2-O|r@
z>V(8vm-`=iRJ2{?ZBk6>1o@wP?GC589BEYPRPlHqkTOv@QJQ57|D!@Pi!kkn4<A1C
zJ$(4Eczd-`s&>@TuadzpVm2~`ga<5i0?ji9YD!#gH52D<J(OrL!$(b+WBaxCbAewH
z+}PN+eAqqd_O<25v=={iOH&O<E4G~eZOv_MZ86iY({kfO*ZyCnxB6>dT+Xk#j`ca3
zUpsqzSX~a}7`<<K*sOg?Zu(NwGeVVDPWBi-_*P^x_rk@C7xM-8^**?G{*u}>x9PvG
zDEu&HH~(e7Af4x8|DX8H|5MM|mv3HVzKUy>(^6l>klNcbuI5{A{wXZ}ccSr&$tP5q
z65E^8(=AhHv>z<p@Njb)chOFFnZo1Mb^XVWl$9lZw$0c$Gvvgv>w$}WZ*M+*`gAiJ
zkAy+O!E!eK<!n5B*C#TuFZVwF;P>A!JN=_S_KWs!pZ)FUd-sz+Z<VBbF1Nn@qwz9_
zJ&#q$!p|)Ec}wl`63#y3OE~lPV4YBi*p5<3pEnQsgl!J&U-BdPe%myY4<|&ebYA{&
z>d2B3t&uvjl;>@P)GoeLSNC7Ndazhzjcn`$({<S?O{+iZA5i<h`l{ahO1&xnw>!1G
zn=;eNzR2&}m)%@-PbSX%bm)`0bWP;ee5E||oRgD9Yxc-|?Y5XR&tHA=$wN;zZDpTc
z`ot%{aNpI-d3Sf0z3l(<?d)uu8t3bKm+!lLeSO^Co9V^J{>l3AY|w0G^W%{)&~5V8
zv3-1c^Mn^)%BTL~-}jWQNA=I<R>xM}(B6XqX<r-neHKZ(xOTtt?${km&Su}|JjQW>
z<HEcZZ#`b0oUHBc?v``UcH{gRetPp|f|UBb<+J@+va>V0J1@U_n_H#$vdVFZut`%#
z<kk;N>%&)f=kTa-@*FT;#mvOb(agld_*BY>UBiL>^zL^zpY-bcf19}c)ys<_?hoFc
zouzjA&$Se;y8*vvY_XL8EWN3)YNCv3%E4wf?$FiY`wyPgZwz}n+f4h*ln*Owvk%^Q
zQ~K*~UHW9c#z3V2(17LV=jY~H@n8K~8+||g>g{6ltE&#pzdkYTjNbIHr<?xQEZWBY
zyX34n&!c_Q=SuvaEq6_&D1&|d&)4|^UvJ;9dc4VpV`I{Zt!<`;PrS6gefutd>A=yW
zOQucLiV0Ak+-1}s|1#}a*y+>jb*glWO&D#w_X!#wt2ib3OGVouql!iI!FlxszDb$A
z28RPymi!lUP>_56^c<(Dt6=eOPw)8(?*BOA7S(Tg6Ui&dTkFMreSO$kC(W&&EAFH#
zMaw<TZOBT?&d$zB$jQmcIZ&=%&3lJqmQGvf`<m|?s`INa=cgF0T2fdr`Figz-W!SS
z%R=~fuzk4sr}yA$i(bdKsd9cMk38Djw^;ALd8C&AhXDKWu((5qJE!qp&7N7&TE9Uf
z@_)U=y#LFjpZ<E3KgVs!@(+LWBjeW?>qN}baXNGHYg1F?c8?j4WoqK;F8EwMzSF{L
z-9LrZudbf|GOKCssi=d=OFCE!mN2FrsGgP-Ex^6_UqispyZxy%rt7Z}br#eADa%^z
zGf(KmU&Z<AQ@6}jx8{=Ni`O%|Kcl*=nkU*U=j!L@=jYo6{QP`<eca&{FAML5FTI-m
z?0@>kJ(KL&tEIFro&2l3@afD`(U*+Bd7a=aS~c6FTjdwaqrhv@YcCW(uRPfBxqZEi
z%2tht7EP{c+bUB&dp;9Bt7dS4=~sTVR;2g?^8*L>sqrqDSDR_!EX@Bl?8l8{p6dUX
z{GUd@Om59Ayv1X8;O~r;N{{~9%kA4$_SRPDm~hT^E0ZMWnvajmPTg1&wl-{Dudc?e
zr~ujHmpWcvZZ}*UH*LLN`0A^#zC>Sr_2u5}u%rK8*1u8z^(^+h(6eU`o;_=?E}0P7
zT;DY>|9I@jv+p!aA0~HBTk2!of9c86^QFn|U++f5bKJf=(Kg<6|BtSpEuykk%k8CN
z-klO%aGJ+XnBSNwzhm--Ta(;drpW#9Z1Xv@dI$49LDxx{5{!#J_8ABAIUmdxGfNA-
zchFd3(nq_3#yJl<PjJqElHd3Lbmias8|?noKmY#6<-@{X^NSbsrk~%*?|lCm@3DyY
zX|^ZMoOz@lHshVU@<esvteVfsS61rHm#x`(Udz_;Iy*bdiwh4Aw_6p17W{yl=sW(i
zUzzwj{=+wJ$G-5{&wlRX-+bg{uDi^vhf*({TDMPVuT46#yzj-<qPrzO177+HyKIaJ
zU8jGvTTF|4TYT)EJ$p{5KKJ{2Q*Ywli}K5BTQ5r_PVn#ZGEO^t_{guWM$R_v3WH-3
z!78;oog5}8_k1$|ts%%wU-|9-Jf|CsKk6@@;Ox)2x~@4Tc<JT+7wz5M-QAs9q>CRM
ztou2;Y{Qv$J{gOGhlfE_S$+G_zuSJB?6#d3p1%3!H{0{YmtMvd+}X<Xe)b)qzxRKY
zdugw+czxFVtxWiasqg#upS-?7??}bxCmx@g>V1Nr=C`@qoc*v){qtV0^1WKJC)dgN
z_6Z+yu-T;f+G9(ShFX}N>bw;vPAs0@;J4xx`;ic@jGo((JH$l<n*Z(R{!&xl{o-H$
z2Aev2`4ita?A3pAeuiJmPUnckKNZfi-?I0!ObK3g)#B}z*(EQJ%s=*6<?hPMEq^b&
z<y->09pVLzfQ2zXJ<IF$lF#SA%l=^_8@GS=sr)lX&a6CtRrA2RpYdlhV;vqi@;!=p
zq|dTt!-9Tm0T;&E(>Rkwuk=m0bX2WIb+%hn#G_oNzm<E%JlWqcwXc5v>#)A=kFSNb
zK^qE+0~Rh_I=>|SS^D))Q9t&?X=!cwy6#m~-TY<e+GgbV?mKYgMta9P_cGfVjEYva
z?|)3tx+`07T4~06#R`Te_srE;e@*bcc=g@4+p+PiQ|0s*9)ED}!M%I;?mW1C|Ne_F
z>8@2@?-k#;xY(^*Oe<oeO7!-;>A(JZ%>VvR<KJ=qjc32v%H_-dEUUSr^Jrbn&f*K_
z9!<TdGtIWJ@rCro->ko$7O-snw{zElkHND$4)px3(a||`zS~);&6j7D_>oinN0wfG
z7c_I0OT&-GPtBKqEDV3tmX_Y4#iGn$vdHD~k9wYz7g1^xO}-iwvMn^Y#Kd=Tj#cuD
zcYoy<um4fhV7cS?)vH&pIu}2wm&h@*XnvGhWGJ0wyW-()mp^Cb?w#kZq+8zib3@$L
zmTe*&0rMrfL^RyN4V(DazwFOnyqEsiWPNS;p0dU66RS5yJI?2MQ@>(`wn4(<=4X;j
zybtWd3wfG7?w>i`?);|WK=92E&U-XWzv&%n?hr6!vT@0*)|j?&mBa-Hrn$R!2DmYu
zKhbX}^z@<7)r)bwd)ZU;|MxqccRzFNb^Ym6m)+-^Xiw(aA}7C6^x4tuy64AvCrt6(
zUA4f|?2v-xI?+Qr!g(@7n;)&d^rIm{u}`*Fu-}*0h54-WagU&b{WcCF!Y?12|GjPh
z;a~mp-~ShXa!IzxIKSB_f|F(8i>P}7%^`I&8f=#AHJ5XmuRJw)#oMm0tEwhcO);OL
z70b0X>+FgrmSwM;TRuTMp|Tzmm7FVuWIW=|+yC0~zl{Cgm%VfC+uuAC5qK|e(`TE`
z$*fR(y^UwrpSo~{{{HYgmMk6GK1$v5Pn=Adb?dui?F@6r)acJ^e?2PSJE_`A<y-B<
zH-D1cPhR*I?IcjW^^5V|o(1!L;+H+;St~0m@?~$}tKL{&!<AfOYtK&5d)*!Js6*mm
zQv>J69rla0qazG|-+CM!VJEj<P9<>dGT$8=C608i-}v6>_lC!XI~G3|pY-%}+1uM&
zGENqqE%mcBkS;3xX*18aTk!Ox**6Y7+8@j%qQR?pYRavUWtWzEg=>YZ=G~)F|Nhz^
z8SA*aQn6VtIy#p)*4|f{cYC7twlCgqSFDc@J1lo$bGp3I-JDey*TtNW?XGlL@$P81
z_}+wXZ=#=Ww`%(LWBQT|@9qNDKOYR#YeQ3*pGv&F`f0*KhuuG?_wPAWCnXda>9^Et
zN}>8AA5c5l=Vs4wwb!%vGz*DMd-iPFvu6*#===|RcJZ>G*QGjp%d`3Rwdb$D-f2AB
z+|tHgU2xNZPtKDROF11YbpkhfS4>!1_>`0BXt(g{>+9oor5w&Loh>mfH&06UM7Gk^
zi)C+)B{NK}uk+;aH>|$Dbn4VeHV?l9JxXC&{W$r)d`}FgknXypJ~2L9b;Q?3ot&)Z
zGwDq4pZhz1{HrfI?Z5uP<3E4)$;nCVTN#}CdHueVQF%vBt#>>;@yq#t$^pM`t4@8w
z_{YS@kpI+qt6H77x%|pXJO8<RTzy*Vd3K{sUd-e|rbtfl=!uF|Zr@raGs<vh9SoQA
z+#{jj9i@MyvBo_m$ccf`>?(ub{>xt^R)4QKuj_H(2;ZT5{X0JMUemNy^}YSHXLp|a
z+uEFC-B(ryv`rH^bG%n#*52an8N18g?u3C|-(PfWKKSVK-+AoJjR)?1uR3w@!oGcW
z0UK>i9~$)knj(JTm0!-O<`}v6Q~$b6ec%7L#@P3&#S+2a^MCFwt<GXk<4@nrce$~#
zY)1da8*j?2rFNLHPqKQftSr6mz4H3*>*_HxQ%*}L&Jblw>y*uy{922lChPC{sPGHL
zUw<lzRh*Zb^XK_`hYP%I-HvOQw;c^y>anOx^XvTA*%|xl{_i?9JF@ESZM%%0p!SSr
z?eA%?UL}4lySDSrb62+K#VudnTHRR5`)}9$&W(=Uoa!z|9a}hgcWhd2G2_FFcDFry
zPJTKmz01D)U)#TP3Oe&<{I|*eId^@V&6&*xA5I_kR{OP~zqYN>XhWgVrnyO#PV1+<
z@2>N#UX{0gb?^7?-pb?q_wWBz7kx%JrEB&7=SSR6zrQd4MdvU3W_~k{O2NxhtrX8S
zr)C8@sx9-IY-?xq_4W02O&1r7^-7ui`FXeg^|Lzpi(hT}^-q8LF)x1l)8{{n{%o)3
ze{{Bag8CHyZMG#J&uYJ&<{zU{|9AdvU9p>QZy)EgS+y^q`~knVZPoWZMz^mvGDU3U
zUK(_AYW1SM)x2+_ULTfCirFRM*IC7&@JfNv)_3>A56MU0KKjFVcju+^SyHP{ok}qk
z?w`1NwrJPZtPu6+Y-1y%r8jeYK%JCV(%bXy?ylss4}br{{B`M;s=ImHqt{-0@ZI+R
z?%CHq{juG#NnCIG#e~28?96tYxha!ez6;Abmi_VOt~}K8g3CGM(Loi?l9QgQM!{YT
zcUIWX_uKIIpM1~BtM1G<^9qiL*&cfmWnF${7OT?h>K%Ggo8#67m3-g4cz^BWzKM&~
z)`Ev~udI*TTP=G0+W+5Ye#LM6ufx6h`}R$re%2=4_MI(g|3AOFb%RDSm(CBrWrgtu
z=bGeN-v{1bvB4%njJ5IcQC2|%4<VK&mlozF94rzSWn1R#Yj6>M(I`H(-s+$4KSl!s
z5A`!V9-4<!W(Z7V`S|fKkDFre5B*uGpB1N{j?r8#Quy|E%+k<HS2bNN@671|fll7r
z+w<1$E_*vI+j-^Hf9vbM|GV4&=(n8wr@QCw-aXU3d-s`#vz-b(_})kV=G8u+^(&3H
z@bDc?#z&TXni;=3R=(sm`{lGY(rNX+2^F9H_B&)aTzL49XW{+Zt1Jq#3!0xsr5UXM
ze$tv*!%$*%VQ32HVrl7rmJ9;Q6K_g6?#Mf>rT-_oXF}NoSDx1;o2R!*-n@QGiYHWC
z(<|4Ed!fw9X146e&8FRQCHnFw{`WtBc;JGr?w|8hKc$>(<z6VD>ary8K*Q>izt=uK
zKC5Eq{gF@aMz56RMwdE6B~GEIlOpBqCdjnDXHZJ~X?5;+_tCFce(W!PW5Ipcd`qiB
za@y7F3b%zHG}W9A&RLeRRn$9dtzqWvjcR+VPI?_ZcXxN$+uPfoJ;^=W$NGN7<c8ji
zjPklWdDgK@-tF>JDtUSO{lN#H5?c?3B?_>u5j20y+it{n;o*XN;om>~%o7cJ+86MY
z&v#E`%!C8++_$^e3*F>epKsjw{raq8GmnR^P0?zS22(ps4K^(7+~nbu`s}m4o5Gt%
z{o9}Ztc!D4@rFIM?ab#Mo*#CuyR*YI0!5Z;UYR<(eb#StqqO2SpIFdP_j14KRwXa<
zmiPbO6aV|A|KpEeUoW1y_1_1sy}ggCRdo+6UA#{->)31FN17>Z?0ubks+-;yvwleE
zvs-b(=kL>pM;C10vucZ<XzqdSaeJ(5HK*m=-g%Do?&@xBv48)1KWZj1Osh0F@YAs1
zgj4J@hi^3%&K!xnkIdg}{l2d`@dgiT;p9naSxdbp`qb>%b4FIydusG%P!@Z8dwbqJ
z6~Fy=|K-a!=3lz_Z~xWxH(!4FRckkGzyGo+_vWAe7`65CGBwrddNEEVET?*A8m;u3
zYj$NhUvbWLzqv+MYongzyu6wkdbB89MPB+7gQZ>lQTGc!Hm_gH<GAcZj9l&NaKl0=
z?t@iVS1sidR%rWsE;FK{qN3u>rwCAtZO^*|rt1GZ`|tnx&-(CShI985@7lG$dl?jb
zuh~#5B48FrbfUj9f1lZgnQ!?F)jlrsW(`{EwdSKpaF51uktwH_CroYrb?{gIW|kJU
zX(@TW3-4^x?M}L}O}DK!&FR4mk(U2kYoy!mbYDFwwl->O*4C^ynic!K8UILVOPXa|
zJlrg!QSi8X^_!)S3S}Nw+5Y8sEetq*^UdF4uQJh^{AIkd=U>{IW!_v<84y{t%>JqH
z6wk9iohN8mFLY}DJJH=Dmv`Y7t$AwPufDvzo@g{<b=(~9L$Rup>NHtcKHch8)JV8t
zXj}C^GvoTU{hUdwUR_-kGVf*KdxL3f_a+={=CN(Mc>H61Fw<=QKij{ZeK{wnW=T=Y
zg6G{w^N+@CPO-hE{Pg*zn&|7B=IyQLNbgAMXy-ks#LklPVeZ<S;)h@Sl@WaLy_O@m
zdBW2b#X(EH^D!4X?$TJ)6*VF5YTVvg+raDJHixZ^+M4<5c~(|drp=B&>sK;_{<z-o
z@OS-{s>1<K@}_Mpa+=%Au(!@pbG@>pQOc=aEsF{{&cIHMNu4%5E*ET7dl^+(WmY__
zv$K{jUvVU_>aw`Y+Q{2(-PhKJEfx7@usW<eYHPRXL#Niz+uPTJ%FOp)>W<5O*IwGm
z-)LKLY<=H*71ydC9(8|1JUBW8lvPtExM_(AM~QH%ytug7eVVGb&<9V4-g^QOFZNwa
z(Fn-U+bE(DkSqtvGs{+6m1=2UUmv%hUA_3bo|jfb`l7TcJ2$r8s!~3bC>nqB+e&{w
zu8S*wTNhk>^8a)7)GrIZc5GB`?t6Bq&0VHmV9m}gebO=wseKM><6HtFnXlaSsPYls
zEOS`liu_ZDiAJ_UcdJ;09H;+(Z4+UUUZcCXYoGm3FW*m-c2xcp4cq$l-o1y#ZZezF
zw(_v~`T6-}uUT{9u}8B@v$VavcGs2Gj11vxU-z6hzWQsn(Gs?|R@pnc{8yfTocUsJ
z_0d~vuCBVaHfn1YXqvS0&i0sRd)}V55AtCbkT0IC<+F0TrP9%Eh0xjO5)L>tn(;Nv
z-7V0SBveqN;xR?#w9sM4#T9@0n>QNx2(h+(ZtXG&UA5SCb12BLWxZlo!u8{}=G&Ls
zmv6fJ?{nE#=K0rGCoJlkq3~<3a+l(UGsaJ+1ny_@3%5v~-ccEIS;@3WN$#GCkEc=6
z(QY}LiXfMXO;VE*<r_}K)#x=}^5bo*wVTS&%jlMJ)bsL!NmivGYu3hb^~CuvEmZlb
zb4p*Pr?8{z!U3n&4(}z!P2xMW*;%VKSiC$~_9_2<$oRLAwRO#LmkjGKbqW&&JkR#1
zOf0%@o0fFb>!Qz0#ZOOH25k&lx;kuabUO%4s}u8>Un8xy&u>T1x|p4&k5BtZ7*sv6
zEpk_>`uOVqzCTmGC>-E>{c~P_w)OYVD-8Hr{@wYe)j!j4*Xw6z7a1m-<(z!^LVfx2
z<>>|%(tCONs!wRJn(@0VZIxW_$?CD^d*=3)mMXCiZZfm~y%%_}TYplb`SCv*|26kl
zEK2F`w>ZE4oX(Wy^my%+5+*bKrk?pIwPer3D1o(+6V}G8x?9z->+5UY(_%~u6Re(S
z#iyMDt+)$X>NV%g_WpNqamFEM=e&9I&P?5P%2EFWFNY7M@|&hiJacEuL$~NN|5rp_
zT(x__89ueC?3$f>DpEfP7u6M=`n>b#toAwUHQ91MOW5`wms<40di8N%V+HTkVe177
zS*yO~L~Kr6dnsy%beQVKl_kxO?-ehsi9N7r#Q{!}i$yX!52iJN91*hmP2K(M&D&44
z+2klz%=72vuF|reKh;~#(BMK6+y4!York~A5jr6HqSP&Fer4SEoANPVPX7P1;gH%-
z)zeST?Z0hS^Uu=a`NbWY(IQ9P{^xB!?Wp7!saex4rem`;Q6%t5hR#phbI;ti7glOL
z{CM|}c%8L<ruwhT#dE~@FB&hpdsTVH!Wm-!Evsv9-a4^s;f1RP=j4PB-;db;N@Jy5
z#nX_V?$Q@NPx$1pn5|t$Yq3l&%c7}Lg=QT;rrJCTl{t0LHDQS+=b>)T0FSb3mr`CV
z>d#(%;z^OEO!u2LB{y`s{}**VW_eT4BX;D&jjBF%9qZVK*UI^n{&`r-UQ@O$IHw+9
zC9ZVy;u4|irbla!i<i_JG2PvLqo(rHNxPClt6Ih<XT&o4r@qtpE)r6l5xF@fYQsXS
zl#^@Ky;cS-^)vhS?RM^8xwLAt?5IRBw>gDpem#3zx%B^)^%jNQyvozRpSme8keB|S
z^ETgHtCCmB9{MJ3JsB6Lx(RbCcrkg@=83D^x0n6k$15CsVnX3r_4V<b-bqKh+jr-b
z@mv>8IC<i@W+CgbV!iE6kC>#+{(N7qwj{E!<c5;#f{b_eJ)6vBIB!ecuH&8T8Do-f
zzI2h#r8J&{DXDCRPfna<67vvfm~NIm|3{6y{?nBOYaJY<Ryr<w_UzWykXIX5T@2Zp
zb+x-j^5)HZOLFDcW~?}99RJN^CF3_)VXve&c@Ji#9$BTiwCwHSos8bcdq25(8hL8;
zWVH9oaI8NPamXb7oTrK>Z|uKsvfaWg2Mw~Htc!N@I@-;j*QIg3<*HK4YcZXOm0@~1
z{gKVA9W72@R$pK1TJ@RZKo7@(sci0>E-)1-{eIgtamF2Uk=bS*h3Dh8=e_j;CC2$l
zNBdRfA8rXzNdFOPvR<R+%dx^Q+$sy_9x-y?6KVT!@p+@4GgA)oeXi?V6IXsl_2Ab(
z{~7l_THVzCb+4X!&8B*W|NIX_JPmW~=gadXb~`PVnyM9&9MT*dwByQ^A7xDib1XeF
zmu=j;H>XQu@#JjbnP;N@)^XMv?Vk~+cHMO8gq;hemV}+1d9HcF)!o%S9DbT};%|Jm
zwz_boRjc;H*UQ&W>N&1H>N`~`@)T%8PSDEL`NnhpX{>Z#y7BXs?aGqQzciEGw)b^z
zxU=<KM`pHQkI?@mi{|papO83L^vs?k8atkOYIW=v@|m>p>yLfx>^|MXX+Fv386P+p
zm@0%D569(P{AxVq`p1ils}x^$c}>*}nyMw56&AD<TyFloBP06lQ<!pV>t~T|$;IE)
z9T!;nJDfOnsZ#jaN9lmjZJS=IeDnzQiEiS3IV{`vov3KyxhzmU>gdg=RxqVG$V
zyM36PX|<$L*{k0sXlX=nolFQra<v7o)<mhVuED-0m$R?Otort9s#eI3ANHn;X7d-e
zZ_A9&$*zqH+~KaO+viZ=qpq{6?dYtF8(nvmM9Ny0q@8qVU|_k{z%KOh<Ie&GmrgFB
zB_|&=@udr@H=hrm>ToP+)s%J3!he^k3)KsXvMD6(O7#iTyy)HaK>eNTESY%=4$B7G
zR9m)A>iGTNdsBYGx&s;(mA~s;Ry<NuRDHyGd3j-`@be&+nmd7##nyAry;Z$laHD&s
zKX1;uo!%XKdu*DfN>n#>uG?ciLsoNoT${PVql>E|Lv?4Zi83#GdV2nzH;k;T`9G6y
zfBWJhp`i-ikK-mM|F9!C@WZ3cCd-+$f>{KGum4;&<x1I)Ws*1jRzFtg4*TNFE~oC`
z*xcm4c*@Gt+f(O+y1Kf|D0p~~|K~~-`T6a$g7)x5TQo11)!oXLV$JHNHpky*?%cTx
z7Az37X=i!;#xUaU<s~PDRKDI7nNnvtlkL%*AVa4%$DpsTg$^s`{Qq!Z&gq`!Jz<@`
zT~}EIoY!~US(<$P$M0oR_o=4-J~nlR>Q^6QU6!Z?2T!D2_?F8YRPXB}cB&y@s;soc
ztuCVu#njrRdM@Gbiw-|Z*Kt~`QM*(8@C=dvLESH+`nyk6FOUn|Zc@wRc>QDlSEaAp
znKypA5_jmyD{J-ji$9xvZ4Zp>>`hM2+jcwWnDn~sja)AacEs>p&A<PvVoCpG{=ztS
z-S2{TZp6%8xps4A&9~Ug#l1a6Y8TG=#n@zi(9vN1H#y2gX<~13)SXgIPg%XikAg0C
z^+&QT$tpcz%6P-9K4s^O$}?_Br6*4wHl8<cN$LHrhQ&Qs+3zgX2`JIfU^D*a*voft
z#*d}poIfYCym|BXa;Kw%(j7SwmEUi(i$YwQ->+UAw3GMwt0sqQMK3HrypjIYs}-^`
zta9c$E46mVnfoi|C`<WGUleq9#f5i)m29GI>*q?Bdsp6-_<C5r_vmFGFV4-iCoeo>
zsoDJG!`^^`D~ow_t&(2#9+BBM)#9$?&WrprSANVpd%h=jz1Y8FmFb;VR+{TI`A&+=
z>OCv5cE;S8%f*wV`#jb6G#5{kTQ6x5BV2q^uD;l@;*-YK6#u~Mv!^$n{FBKoRhKBb
z;QD8esZwcQp0;YTgeF~HZguqWmXfP!&pC>tL)N|KJrx-2HCKE|p6g2S-T&E7+leNx
z3;TaZ(j>#qMZ#gH{a+T1*bs|<(_@swHb?g=B{s9Tr8fqiG+b7l(<Qk$KtU<gWtWx7
zTeZ_xLj6lKO{!{E<!m!=mYDPKxp%9P%N9-L!^`eyhIRj6{cvTbTl*cCyL<VjzI$b;
zv7-0xN*m?8{h=C7b}xm@&aMmX2rS*b(o<#5t%$!Tq7Q0kWy?;Kzp-h;r|2T#Pp(z5
z+9ytFbN+pJ{otW*YuzH2x2}DW*c`ts=hI$cxzy>RqS6fu?_LY!dSF&LcS69{I}wvi
z3)W3BWomx1_v32LX5GC>(=<X0E`+sKUBC45;iNv}>Qw!-2Rg=Dx(jooRtH~fTXS3N
zsNTLKlM?R46lw)G{`>VvB>1&YjPr)*-_ghayQ}eioBShamWJ#ZP5&M_Q`QIC7hYH@
z|6ZUhe^%t+wGFE+gGH0NZ~VHlqV>OHhFx18$BEqL9r9T&dw9I3$6EQw|9bm*rLV}G
zkGoS&e-C9eDE*KTGGCzd*c+bu^D&ywKcyHIvUS>JJ~CR9XfM2SP5Xqc6Ed2=eYvu<
z*Uaf&>nCIN-!_pal3o5jF-ZKTyDWS|)n@IC{>6^P!t*}yyeN}xKkI8S(Jd<Fz(HfS
zn_XXaPn?#gu_r?$&Si#Y>D%6ObJAA?yPlq}#o_uQ<BO2f*_TGwZZBB+Oz%kZk|kF)
zlU@C0++C6;Ca!Vf`LcZz_Fe1yXytuSGG4+nFWK1r`qxt$A9pW(J1c0SbNL@$?b$vK
zcTLV2o_*AG?B2^SXQy-;Fm3&~vwwTLjgJgh+(xe>NtJPt-t!kq*YyT9PJg|^aObt{
zY#UP^DBWHbWOS!0NiXY)V%_$@$6KduF6a9BciVTb{grla&prJtX=C~1XUJmi?Y(`0
zBBvKz+s#_#_C`sV{j}|$m4038|LxH@Az3ajy&>(A{%MIVxpLZ<8=H!c=3K2ZTXXu!
z;&c7iY>$}4)T`YId&oUkalX-)AO2pCRad`Vs6E-8tE<BxFnhMk`hXv5;T?}#ot-2!
zO4lxoVt;<Ze`eOND5ft_*2y~Wv{atmWm)a$sgl6S%5+XM=FrW!E9;o~*EzmOJZyBA
zMKHECam555NlmM->PJ@<+}&#U*XsOy-xZ3}_xF0$1e`c6aH#HlbbY(Fm&?Wb&hMve
zOb+$tl(T!f_5Tv1jQ5}ZubQ$j=gJLUVeUEB@#PN;_Hbu>Uw{3;!I~elYoi~(i)@Sf
zT2zy9^M3OUqn~d7?>t&rzNcf)C%ehqb+&K6W(4m0aQo>R(FwaN*2TB|f8ltieV-Rw
z+O}h9^Uh2CYn~V&pEo%{LR(n=iguDwDC6{9(`M``&ibFFuhei*dF8jMY#I7{`{S1W
z%zt<I&6X{^Vsn~v#r|`i{*d~+S}FCP>-%h*$ZLIvRk!j;%4=@P`nEVKb)~HMfmxU2
zJcV8A?IlF?*MHBD6TheUwqx(XSHC;#Uwr*5v*3Tk_e+%-Y{5!-rOj*J2FuU2o+v)c
zcy8HGZyAGMQ)D~77Mw6&^n&l$fe)_x1m2&l?bR@OT9nRu?C9NTKce@{zn!}ASfT!Z
z)5LHG3z>c;2~Cr;E5kR0J=Wovxz}-r?6R-y8nW$^?thwdVD{Cok}v<e-Sg{Scap_X
zx~A!#?6V*LOIQ}~dV8V3=}Ou6tbIS6p7}4<)p?evGIQpi6)W;)UYYsZ$0c$8<foZZ
z$Fk3y=-3*y?)0<Lwfe>%|92ge6co9Cmy<=T^WFBKq?-kXt}#hXXFjSfkH2vG!fLs{
zO?tkqx3hnTo!hc1>C8OA$QKv1@5XJJP=8`&(}MYi`h`}_nkvVh|L-pf{**HR@85SU
zDNFTU9GRae#`WXOtEqD$Pq6r2b$jJ(Qvdgt5YJPkI*}WX#qJ9BJ`*_B>{L_q@ac&a
ztfwB{(oAdRo2mQF?Rk#oqvh(^<|<rsUlmw9doJTX{Z5(qoF^X*{I0#d5`FY|;}R*i
zX{Wau#z}lV{IdJbp;MRa(^RX4mg%=Yei*s*-O0r-8cof12lnV5INHZ^w2$YF_n9p}
zdRDak4!^XrY}&%xuLN@>Vr}CO-CMYkZ}F=>FUgmsuXK|o|EKV`E-id}CdpZD`pcth
zTGKrv<5XX)Nf+<@Re9!IROjN!mzPDGE;p1`OIoZ_AU5s$Dx>+Lv*s2})wpqW)uj#d
z4rFiIxXhp9+{+i@`+ixtG&A@4mmTXlmTZ&Eyy5RZ)@##!mCqSX)BGlXYpTNI5UtAw
z^Oik!oo%Od<c-CjGEY69C4Zui8Me5@tejbTcYVs*#EGdsC-y1M{A6}-MuFvZ0X9FD
z=RJiTHi4a*^6#Th_1qVY`1kEr`>*@<zwclE^6kr;`OEocdoFyvN}=*%0COJ$!%fAg
zwnyhz1nur^Ki2Us?l;GT{Nf#@+8ZvGbH>(M2a0kmxUE!d{cPt}fydwfp1-X#rD}6c
zsm_k6d5+U1{>7gUnj^dQs!)pH(!C3JYJ7e=_2R4+-4!eM6lX;ic7M)&6~58=sLR3&
zrl+*$FFrf%%L?~JsdxWneU(-%kej<`!`vl7j0c#fPnRj+^)<AzU-ltnYt!jpf2M3_
z_I#mmv^%c7XEDF?vrD3a{^I*>9vl(wvv|OG@W#SIm$};%O4n%Lm%2AS@x$!RPnT!R
zp0z?STl{p`pEiEJb#DzA)HPZd8X6kr+;Ek%c;O*?xa(P#mDGoYn@rAKdh4^o?Q}@T
zm)a+9c5NzPm*FxH_71eSXw4JIN_YQpXw$#tzPeITJ<?}So2>lxEmfEA$T}yVrz_LW
zM=d|O==Oy~kwq_DzP!tA`LLbYcB;3jQ^muy9_K@AiubKq&K~7>T&D2AgD+{qi-os-
z5?Os@m5JDc^?r{s*A~VUFy1|;?j9{|u#{oRm4948Zp`btmL7XtI!mxjuJ`4-D;xG?
zhp0W1idrSn`{q`h+UfN<kG&&*{BDkUSt<9h<6!?!d*%x-G|bLTRgt~xn*K5P_<B`w
zJu7a8o${w{JSaTQuwtJ?Yn{P^f>;g<8TljM6BfvSI5^2qR`b7JP~=jJiJ4z7oefD*
zS^xEOp84<F@7LQm=5Fd*l(V4a|IFR@udNrA`Y5TgW2p***+s{I@2{rRU;6y&x7qJi
zGIpYT>GPd`PY<yy7qNf%@2UJMne3JQTo00Kv#nNMc-+UlXtkkP<g;!jBR0)KJE?lr
z7fr>}cFX&OAFQ^tJNRe<gRaZt1_m}$u8gx0QCHTm{63*`Xz#4e8Nx?o_+lOjyWV<U
z7UWoT=4r^vC2XnN7kTHY+}XJ|!tv5u&YjKOO24II4zVaGayW8!x_9-gnEj&Vp>L%7
zmFpT2yf$rVqDO66KKG=|U3j|t;_aWa@8@x8&3^w;*K*2w{+#bS=bT!8Hu>hfqsxzO
z=AFVQz`B!5p!(SFE5ZG8hm{U(<$Tk{d0~A-{G*$PA}6bMMpy|nEuZ;1WK+U)&*NG=
zGwd^63>N&IDPQ{JQI50QE9>fwM{c$6eqwpw^2oW?@*~w#p9?YA`EPG}79iXqF?-_v
z^mA$qdix6>e(O^gkgj1)KWs74vHf>HdsCsMV!V&eFXv(j-QWFrFQ@KG|EYa2?zp4>
z@}SdQ{jqEr`8zbMcG~GQ?HB!0t`<4(&(-%_Tfd(y{62F^!QZmK%3cdB?HZd^e=K|<
zrLn&5%re%zKBiO4pH0h>HSN(<y>dMx@bJRY#--WpGR;!?)BG+(yg0w_!lulYSG()n
zO$1t4Chb|=8v7zEqR}jfGs$(om$t+gE0e`NUTW7LOxf0SJJ@QeOwNZDvn_jzCkQOM
zeD!6qP-pX$+vy^!SnIDfMTBqrv&QC%Q^L~%@xr-@(WYtoyFOh1V14smkfuTPo%IPz
z9TYPrxO08IzHx8OM9$`?+pDy{^D$gI@<%FcW7$_*Rc%4_?Qf^-ENuHKsMlZA_fDaB
zvEb5JTg&i$0oBGw%Wlkg>8u#^|9sHx523<rr`i9lj6bHo-6wU=ouh1ruGdZfy(XvG
zqil(p^5cVVB=ZwP)!x2&dpj=W!;NQOS4ApzTwk#Jnck$<^Klm!9Lf6n^y0LV>@!=f
znZvhST5gm6D*aklEhpQ4@9dzP+`LUZ6BfOm9K7D#;PHwZ>W0hBtqkW(nKNz2ojsgB
zb8o&@Kk4k>+;V9*)1qX98iS}We>b1tb=bY>w9106eapj?+gN2p9#pbTjVaZZuu9T7
zvi;`l*&83n1ZS_ebIId(+^nqU+M0bS!t7?5-Gr_a0i`Q9%`-Q5-1SaW)lN*NYf|9u
z-p86YO7A`?x?~B9KD=S@`@}+)_pWDGrnOGL)zsm3Ut#J=uca?l<JK;d-dnGjkXOpQ
z_WNc24dz-$p7K5W`0&Cd@dZ_zKXe55gx~gQ-I^Z~xqNqz>1$EtbBWH<v%_Y3y)xYY
z)$&c|-;{{PQ?)B*?szM@eRJ_knfI4wu6Xt=Ah)b@!jah%S#JKgr@z|8%dX<;;jlwX
ze({9&<ezKil4}lKcRS*COW*9vJEpfCEW4pqp3{9#_H1jQXZ!t*8OJijt7;~4I5`E2
zN=?0B9IKS_k}>&K=B%_IB|<N}ZPv_OJfG?1(MahFXN*@&`4qr8W!J@E`{h%&l*~PQ
zR!8Q?vTL%<lACrKY+4=d;l%6XJU{C3+YSe#-7|fbB!;On-^y|8IQekht?ypZiTno~
zx5%_7HCH=tU8J3ua%{_K<p)xGJ>}wE{{Nlh5YY4HuqfLJi^3*}3lCqnFJ5ag)1mNT
zzw3`jJ}vVl3g>@3a;}+MuWfsH#uXpgqwj;3)(Gr)nRjURjKbzWlir@5J>?O{wW`<0
zr|Zlx-Ts$p%Ea1t0SRdwb5eID`ku-7`e(|GqUDeJTmDNQl&tO1&O4@@yG!`7U}VpH
zUNiB_#Y)oWKEBM#TKFf;!_I1N)#OEsQ{zOU_b5v#yph;%r{P%k;d#imq~9AH=SXFL
z;!Qa6X=T?==~;qj-ULg}Gq*`pUNkv+@{Ewpha(q<T-}@+R_3D-HftgO_MMqoCz~_Z
zJ9>z177o7T7tW!hc}dgSu1=k6;hMCVGab8R%ZukHuQm5M(Xlqx?dGm|ziwv7R|z`F
zs#X<!D_0i1eObh)@5tUbF|U`(A&1t=c<nFr+LN@axqM3U&(fXp{4*8P1?pb@zBqBZ
z!J#_!?yB}Te_i6c7k{7L$RoA>?dDV8q61Q+jA~;GUftZHuG1!b%&}b}w<*`)_(O3!
z<ugb0geL6?3V)^RvRL$i!q(R1{|(Ez*FO89__(KgM&3sSyGT*K31>RG5@*TV^f<re
z={>#hw_rw3f>rQF?LZ#h$lkAkhxINL^e@m^aN57X-_A8l^r2ADv;!+658CZGx#(%b
zw52s7AJo=8$@Ax%wtsQ-%s?gqw+mU~b)rvY1j^TB2DIs~Nd3Pd(zdbo>CD~tUrO8%
z5e(Zs(Y3|l_=^yY7ACu+)70-ybQZTxTD!Kj;6_X|&n^2D?Pu&stxq}2pInUmF!7Y5
z^U_;u_pCk8Vk!Gx+xOEMHU8B56TR(p_kWsM&|1%;d;jL`7yY+%4lFQCx5|rasjQ5f
zclzBtH9qNmLDPB)>@qJbI)BhG`P!QdcgweC$9d`)#SBXAy3~}lE@VBEHECp;aKNSG
z*$V!0hiA<djm#U$mYusPZc$<7C*!K{v})np<hS=HNT?Ujzcei(-Nt3h2Kx<Nziiem
zKOVHN*SxQ@w_5D&FUIT(v*#FU6;;1Em1caX+kT-h<7<wEJkMOinGRigTC#TcF{P5c
zR8#NG&p*$dZ(wls-P3|)&p3jgtrBO8FmU3GlQnfOw(=1Rm+%Tc^iWgvmG6mU@5SCW
zk*^PODEmZ3ZhaIdcEv--^o0dunT_p{+y|`>UfG-xKf3PSj?jWO<<B3(w<HRREjd5S
zUQ}}W-@DuHf3KeTF8)`{o&9fat$v&r|L(cbsvpt$Ka7iZrDr|ZoOCGRMIJ{;)n^Hv
znLVYAx7sgM7YXY=)0-}Qdzwq@xi87n56_tU#-~_jV*livDy7@pkzZZ!T@~XytW{H_
z!|*>pYp0pG{jV>)hpP8IS9jPpA>rIT*Sn|XwpohZIrQSoEwial%5Uw+&Ph=|{p+TB
z<?p3}lMnydn|5V$GUEzg6XoKilCQ%$D}$~xNqH4tN<VwfB|TN$bFR4sQ_9=-v-=&q
z)K@=U8O|l+C+jVu@lj)%zVni*#}ZQll1;h-Tjn}(JihnbqPF_NC-La*g$gsYzW+6v
zw#V=LO0%9j+q&0JnRfQ(&RsIs4su<o*(t80QvPD5XK+>dlDG{`buF^j*0;HGl!iTx
zG83);A06Frd7_n7WSLgDVe|bZcOU-U&t877yr$lJWkn3@Z^!n?rE6GBXBS_4b@udu
z%@gLDZn^W}D6jtfYlprp5oTlhzCCD*>{|AirAb<KSA2F~_^4n1B`NS_jGd}>!}0Gg
zOr4K?*k_)wwQP!QN>Rd_%|DO+k#hPVW`6Yh<&^dnt<`UO;ttFAPkCT3{YEL_PG03w
zt->?kY@4PQOuezlbKlkzi*3Iz-~D3t%!_C9h5M_`jh$lNU;j{Ic+%Z?<u=W5?)<pv
zuMW=_=+<8OQp~F1?a$K!l6O?UwSG*s>xwwbZnOC71lf59Q||G`yf3rMH=ZZqa-Y}U
zmg`_PKZEPOn-1!hS5u$1mh6m)3Ot@1&s&t!QYGg!@zRB!f6t<>8wu_$?|Z+d^|AF9
zyDwcGdF#0LUglc$H8=0};bjINT5pwk*Nbjg!ut84*D0gOe@7xa4!`N>Zb*!LQ{1al
z@X2VywI?UsFXa3bFE+0{tloB?{kvvRm)=3kHI6$wrOq9m7ke|Uk*D0OntSHXCUM_M
zv3bv)8>#sx+t1JHm{M!G>(YWt%&s$bygc_~f#W9bw;Q@+KYX|mcTR=Duh4b(o<frY
z&yUE8eR!15l+H8rUCVEw5A0pZdydb&vFnreCeib6QZLoqdF7hBI{13J!8ezKU!qd4
ztaI|exiTntn(j|G)%~ZpA2r&~V|P*To_6)?BT1ViYm=0(Jv+F<GWf~j8K;x3tYF(b
z$>GAf!>1n>#Pys1J$uzd!|C?!Yb!lPr@T;Zt>$g|Smn`jF#lF|$cnaCs~-e)sFZO`
zd=vF7$7{9cKFJ>`fo$0Zo%@e&DXLg}R51T|>J=rCucFM`v)m(3%AIMGDVN>8;dC|E
zM9cI|TV4O|t+W^EKD$Bg=t}V^iE-xtjJ>b!>vwh+oDt?wDf?XK{q~waJe!z*PQ9)f
zz`O0M{*BLfCssF^&+A;X%cS_b^M@0goOV}rp1T`mdgtZ$mT1P8wH?!b{<6)!K5y5Y
z8CE=Z_vFopco*{C;_;hm`$o2mu;-K3cjQHKT*_P!&HZ9_9edr=?<wJl?H#L&5({?y
z`COCdzxip(RfA={|BO?X23W>FHQ%LtUofL>{#lnxSyHA_#jicKG%c^{Tz+p$W=Z-{
zrAvL`b)8G+FaCIMSFli@tm#9~hrWBgvXbvk4C1OTQZ+SpH8nh<v{>X?#*{5<MYb%z
zDj3+6H&tjFmsa?ukR^)^8z-q5Y>;``^0;tgb57CR8D>tU7S_V6?r?wk`A6ye|Eb^K
zpR#b6^ybo@3$MS-ycG~B!+SR><YDh=#g28}dOde;b9{R6JU_Ux?9R6Xtr|DZe80Ib
zF@EySefu}AxHZ$xV%nLkX)XyBbDKW5KAc#x==Ebc$?{y$8_6Fzwk@9gGtkO!;o_}d
z&mU5eVE7rv@h@cFlV(G=QZ5#W3%v^k+xO3tlI1*dfqCl7L+xK{9UBdjb+`=FUfO*=
zYp||-B3qe$mEPSCw&q743)<Ota_x8~uKA;k_eRYnwfEL*#AGjZSkLmG>UAXG-^<v2
z<|azhZ*7SBbfrn6KI5y+!kq?bbu9-o&Yqj%Gw;mz+I3<*Gc2QdyrwoR2s!=Ye3I)D
z*|M)f-Fg?NpPW6@@YXA#X`VZm%f#-gh~8Y(QE4RHInnCT&-N|VO^2TJ{fidT@e(zA
zI^9-tr;E1s7I6;aZ7K&3>^rwIviqs&QE!Wi(=Rq3H`SjS%bPg8!0O#b!JEfUUp%h+
zqWZQ=wXa$A&N{hsEi=rmbXQG%k@03i2y3<YottlxgxO~{{tQ{`JoS#k<<@H6ZNJLv
zeg0jtw6Tw!zG>-s{yftUH9flaMadEpU+nrHe0w~p&wS?BmEk@1&-Vsx<^A{9?)~$x
zxqX%2jStH{Q(qWVdr;iR@XjkW<Cz)zZpo{-){5V~rB>88@86f+rGXL#)$czBBp-bL
zN^vU3>O}4X^(T-2>$aHKa67=h*73-$rr$E|zOz2XEj&4aH{d~@wOaCw@F)BCUhXsA
zn|VZOW$}WQCWmLHTd_~s;S@KI^?~8?!2bKY^Ul8VnPDtq+ccHq*jeA@1|==g6Q#L2
zJ#U)i{+xV%Vlnf&ql<E?)%w_Dia7Qbu3hT#^ZANB&v)*<YG@D;^zO>O8&3N*_N=$a
zD`hNydF}!4X-D5*$E<cm+BJmi-+$!%PhT#Job$8nejZK#;my%qC(Jsbsj5iZVIQ~k
z4TjTi-F|favG^ETprgHhe_7t_|0_*b1iSsZ|57usN1Rb3pM9UwK`uQjHGZ~>&D-bS
za?O|U_Y~%C%~*eE-ji+)gL_?SQg!_gt4^<9uEw7>|69<-n!8&J`>)w6D{S(Vl;4o%
z=9qGV|9{N!ovf2(v{U;pFa3YHs84Qk{gO;iSG%86UQYkt)U`+}V@9e~dRMOZ@24VH
zLjD)?Ecf~7H?z7`)FtG9GS9NBXH87xL^A%1&ud&-|M=q$KaKyDUoxh7DfNGsbn8r6
zBwuGN7pOCZX<LYG6nC^pfnBwP)i$wopTxz+mlV!$DH-1D2>7yhnzMmQP8rjD!5g<d
z-9B61&@6LmYPIZ0x6$I+Y4_mBk5v7m4j*2uO;567)+}9Sz2KbN@AVbs(#(YqZJjgE
zzg_9{zMk(uYr^Ljw%3Z+Xa}Xt`>-)|=MnW|r&oTPvQx0&#_j71bUVehjOzuTP2ZdG
zVh{7piAk)RGddrZ`^vP7e0vh=yyDKHfd9G&#l-j6@9AbStbe5@9Cxwj%<QR5>vm0k
z@O1BPp@cgrRZI0k+`nJ2ebnQ>ZLyZwyhFL!XU{s=Et8Plc%+Vv&C^n+Ke3nH;=@H-
zk>s9-k1Us*(7(BsL-qHGTJa5r2j}1V?Ogd+f7ciP(7PwUsRXe6np)K+67QB>cTn<r
zUrTqcNUk7v+ETvx1(!M`t6O%n*_EH(`u5g(Z>5wEz0LbI^ICtqOjFCtWnW_AnH0Ui
zIfr}8?b)*fZ~86WoR*h=Wae&vx5>X3sBvrEJLt8=KT$aTfZdmFH61mjfa_b5v(KK5
zxRQ8jx>TOK#Mwj1C2h~%l+9jp^xN{2;$JMDcj?64d6&9VX_sQ+G-<v|g%%IX_kPWi
zt-Zm*#%<Hn7?60d`Qv^qE;WP5_aA1mMHtw;uM9|IHhd;{XwiMe2~jH@8qcWzWzy5p
z<A{~C>kufM`T3yZUxVj*v32PR$J*5jT=}N_)vbEPUC{iGV^L~&%>ixC|2r#O*f{sz
z^$f^6TyCYfmuXS#w{zRpPB?h8G$=%n_4y7*hVae*OpkkV%rLjEJv8OT&G%MEHJaS(
zI^~0|>n9(c=Y4-xy`<QITzeD0+dda>-{~)Nz0J41u5;yrCC}q72LHcSw%p0tLw6oS
zw5nFf%h=8n0W)o6%hQ&wkly?K-Of{!w+p7Ux!ZJxo-_%)I=4|YVa>TkncI0qi=Xp!
zf0FP{Qct#bzh?Vz$K%F>_3G>sd1jgKV`GSs2$SQy_wLyfneVGp96!kYvrPG|7<9eH
zP0K-eC0qUe*VXo~{}^6sVE-1WX;9z&f3w$H?mA($ljizPAMY+ZACZ^tHvhR-ui;KV
z<<pD)R(d{oAbfB}y!QdcF!RmxF6*w*i?j^i^P+)Qa`NGyJN{<HE;v~GOZ3t73$BNx
z|IhibdP2t9?ebe(%O&n#6_(za-{g9^=l;9S%T>DKIfd^yb{xKTU#W7s+EZ=k26@jq
zb8Xhx^34C|8Q6Cw#ZPv9{5<5^=FmrSD_hq#SUf%TMX73a?t9j&Z{~UEozMR%=qx=m
zYbTr7>*oveTaB|)K3oiZ8M5cYn@5S?+MZnqES64v$Gw8Z;Pdm9muF9DHcXkpyK?7N
zVfD@ImSPPvG!<@Z7B@Xyd~szEfBMbEK@CC7wyl%T@8&ZVI{$pBFzdX<|EH{Wub=Gm
zX7Rm=<%`^w3hmpwC?vVq$okHvyzgm`xWj9vnXx^uv3{GC^850c=~Wu%86s>lZ-}e7
zb~s%vd9(NY>{*7<@$>tmZ$?aJ|9K(n%0{L4Gv14=OZvIR<<RYDfz~a>a=S`@O)f4`
zx1REU+vcKcYA&z7UB724Xpr+__fhqu;(1Lqsry1in*QHft~$lOpnI;Y+)d%~fQ$Zq
z^$SHEUT({nD!j#%KYb;q`{S8ELN+CR{#@D>W&PmHo6<8nH+Foh?V9RwE@z_A`J#2#
z4jxxpEyw<>vvq1q*}l9J-I=!Y#a6Yc-ragv@#&3dpM6J{Ug7?q(D7y8s(sg|v}S5d
z$y)o(E4r{gi*JLL%ZqL7j2btKq*XOo&u29{?Oxy6*1RJn`22(mJ$JmD6%7`tOyAin
zx8&r*8FM~e*`ySASMfgY5ov*Pt<5(CbB-NP>HN6y#NNQKOQH|HHWi*X)jZ>C+O?V)
zHvbMC$dd>S<Z%}HE+Vzzi^v?_zh8d~Ds2wWQI+|9f6}xY5tS?DBzG!g{u0@3Uw!gX
z#M6#LpI;t0Bt6%o)vWpH!ui})U-t=#9L}`<vFXpDm41%zx6Yq9d2wOl#gzrF--{Q<
zt=O?7g0bP1|Gb#ruIX0QlXMT9%bRE^SGt(F*4%RAm3(34BX)DTFUiZl>wHyqR*L)M
z&LAzV&-?p(<6@&(uYU-+wAu0fe2YVeWAq-$1YYZ&X8JCmbb4sbl<K*^_;%>GKe~`m
z*A;$*MfH75fyIOWL6LKUEuIyBP!bfhNqfCJRaNSh%AKd*(o6~@4!_|k{o%ZCx`u&U
z>+0%=R1X`kyiV_1B6GSj)#O$r`mJcYXZX!It>E!h#gJ1MjPF}6(P>NP&2c(xR<dQi
z&;E%v1{V^p%oqN(MXY(#%^Uo@{mwhonhdieG<N-8+NzRkm)LxOw|U=eMJ<zk7t8v!
zz90OdzOLciX?8ow)7LbM(_j9c{H$hqtC9Tg;(u;a_5}VeJsfdp@>)X`Blmx;VlMhY
zjv8-X>~Ga>xLJ3yb@${yp~We3G5Q<brJtr&zT`-0zsUVFwVP|Nj@;+JN$T#|U9Ycf
zS#)`;cWvX&I?<?>?Wt!^1+rw>+Uye4H+K{2o809j@M>j=&eBA`Wr1(@`^Tkxc(niD
z%AoFNovaVEub<mAN!KPQdq&MA(T+9y754HReaI$|o-SN@clOMOXRoXMo^5WqP_e#W
zz0ujIPyFh%Ss@?ZUo$ekbYhLP%lUu>LM<!K{XF&LJa^@X<;9F8yJ9UhP8VtD9W0E|
z4!Nb<&^)Vaw(LZSC#!`cE9M&pZnkN3Nxze-Q?g5rY1)T^8W+(;ZkwK%`&7v+>0R3A
zBi9>$YH5`?Z{Tj(IbZb;^31aF4tLqVVWHsvH@{17H3_k@ZI){NXZpf)Wp}g4N%u=0
zyL|7PeW*Vd(|S9@&g9D`m(+wqbvOT~Km2CC{>I6RpOgNmY8hr-PjqLUnxeId?dgFu
zJC~Em=jEJt3ofm&-+O9P$)2*F@7(Vb1GumEI@zQJOkJSxmi_O}gUgk6aIKfluoDXw
z)lhh+A8CEXc|QNdjLbHE)-wj1nL}&89=$tj(?>`48trY<_)@wz%e9{Gke_=>djS`x
z<sP1299w@^vRmuEyd%=}`|KnhIkpEM^1~~6e>L8ma?0bujwz<<NtgQjg~~;YTNdq)
zYgP~obrWj)xK_Z;HnQYszHzX18lzxa+0C+J8~bNYJ9gjEG3okTJI;yy+e%F(eH^=@
z77K5Cd}->RfEn9v_3x6sICZ=KtK}Lx{R$?_Z2C`!xjyQOb^o%|$6kJ?u+tYG=F579
zy6UX|kK5jOJhR9uS)Sufjfzaen_Vg54^zJ>oaop#SFg);W_9Vl+VqXS;lU4Vg4z%2
zw=6A_X}cV@bhglc6`#cyKKmzDMLapUa`8jWJeS_g-J<h&lmfrF{hG2%;roibhc{W2
z0_VI`2)e>MS8})N+`j7KM{&EdcE>a|_%<!|%7}H`Ryf!C$>Ld4{g&?GFJgNaI_s@H
zU!aS++oSCrOT9EcEqn7=FQC&^mi44pjqOD9bjkTGUM40|rkWqVTrR7fl@r5q@ZtH2
zuHdHXoV+vX=Fc2_<aC}WKH4nFvGJ(s`uW=}`CU$T{t-M8!x7w;Wb^v>?X&Awq`tSY
z(~_ESu=3Mvi}ilh$6^={eA>^dJLA307V*Z-w|%&*<DWd=q4xBH^V9DZ9ls~K6{Khs
zA9N_<tF2!Upte|MNAADn-S=Vz_@2snJF&C*8vZ%-X?M+Lg<sst?x$u%zT#2ixO0fx
zdb<4Cy>()je%`Wq6Z*y=Zrj?Ut8dD`IL%=q+r|B1-?#XtX-7(b`x#BS`1*y3)550j
zTg$#}_*&pSZ~ep>*BzqD`lrpCV-Wn0cX{~AC6aC;`^pl%W@J4$*2?WWd)l`R-4U8k
z_blxaD!5w_xPRK+q=V<2a@Lk)m-I%f$|t!_KK_Y0$Hef~RHHb~sdc6nhKDxgTAjai
zG%-=kF!{;!VEv6^zt~iF?)0A`@oLwyx7U}v+|1T9EyL!=-?ecrr&m^#P4_l9c;NJQ
zi7eUw4|P@?e)h7IS6jvHtB+uduG)c%#<%qEY~(oP&81u8nXTG?Bx3vD+92I^r~c{A
z`z>_p=!H3}ug(fcp7i;)($-lKpQ^Ja7f%(vW<P)Sl)vj#ADQG?a(5)`cqiQ(zSCdi
zw1|;x7x#u;l~Gr9B3_j3^P2jgCfs^LVBvr1nGxC#z2@$8KlIz==%S4OTfRJOntwXi
zYOn7VNjH)E;ZaNdjODJIiRj7SUR8Zs=?IU0>iVD?{Htx}aT)Jcdp?aPMDv!8OKqNd
zkhUfFwe`Q#HvHP8n_J|{dDYPLh8fGGZ@Xi7(%+ul5}tQWM|akZsU`DMAL+$@y!6VH
zvFWYq_SEXYxKt~fpW=>+(}nuXkA3lC*)T0?8~0a{0|)h!ZmCv%y*@LLvo2)g_00i6
zZ8z?9ZuB>fHQCxLm3i{OLCb|XkM^3cG3rr!zO3i7+LVibW?xjB5qyiWtM-&@^>aS1
zIEI3yhiaVz=D(a4ywH38Z?~x$I{zj!CnkT5XLh~*(r#13okJbf{e7?HCrWLW)Z;(q
zdbzax;Nt6s`q#HR73CLId`PZ-X69aW`f7H8-NKaj9Fy|RH~rm}TAkkC{Ux)xRlnqa
zz!k$CPp9@7)?Q{!?0)yDS8K!BLaS8m9gkx38r|8K=Uz?N<uuj$((#9vrW&*R`}#;|
zZacdyd}R?^psT+B<(0uaA|eZ&&cFQNA2Kn))npFSREtYRH8TQFK9@eRM#3U^vw`6c
z;lBq~9(z?iDNwFm%YaAgG;7(`d2NTdXO}GBrf85?To|xjw<B`-`r>`lc)0|QhnJoT
zebm3wF;`^P*@mMVUzBRzt*@%`w7-07@}qYhK|#D-T>jfvDc!u#S#2t5u=#JB;w#21
zvttW9t0k^YuGbMea(Z4x)UGoI36b}EZ=au$nkD`Jb<EN_p>HQty1$>lx3Z`?_2d1v
zyE6_QW8!;kwI_vHxWU|jDc<yB+2>u4bT{nJ{eAG|=A;M#$9vazA3k6fJ9B5f=(T<4
zKkiI8diulP4>DJu3s#)!@w=?&Vat=Y(L2m=wu?k!&8_S!8yA%?cU;4L{iBk{)kEhQ
zm}gnG9NaW>W>AgKO}ou(7GIx4PF#Cq!Skj^uNQeNX#M>B^S6E9zpQ(y!NsVjGqEf&
zdeco8EkpalZL9+RGv@RCk<ylsStp&d-D~MK>DX^yH16qha!;3>^X`ep*_k~%@<V5I
zPivaD<jL)NuR{qpFGlw3ww65GqoUie;ZzZ$ySPSjk?hUSVFs3!vbEQ;rtyR(Z*_Fd
zyM4%Mz0A^g8#F%byDX-7Yu`3|Wxi#bejEQu)cEjP|M05B`2F>~To*$B@7S=I`Ra@<
zd{6D<n%E7P{(kuLDuU7bT3%CIRNb6_ILA}X^H_P#G9H}HqQ8~<p2&80l`}Q@*ZeQ8
z+qc}EujCgChnb?NU%<uIbN}V%ncX||N$g$}$JCpj)Z4!=?$YDYetz|XV%(8+r}ad<
zfB0k`-c{Ql$~-giah-`)-16PCuknWG&*tm7T60A=VSn|_1&coVC2yCk{37Ydeejsx
z-B3RL-mUhnDvR5>b~jx-n)Te;U}<Bx*9?=aVsCCHUwcKXqRd`fc~5U4qp3Vx+C{QV
zq8X=xQVz9UlQ_|ncaptDG5eRLElbv{^IFB0abBihcNiW@jQ;swm77bu^4}bdqu<pG
z8gJX(Ue$QoF#Tqh$Ofj_=3BUQ_p3*>W{a&myoSr_?93j81HVt`CW%?EpW4?O8UNPr
z`Nl)nV;^3a)VJNYLtdME>g&HLdQ+mmWlY=iyShm)r9`$%P0wWZ@4GRF0)L&kw9>}%
zeW<A*@0km`>OXwj$}a3DF+1@};yJYiCL0fLyD-N`!Ep20f9rcI9vAm*agcC$aR0UM
zQX?LQ9eXRUPPDSAh<vKy?77x^--J8cdqp?uJd|_j+K{m*K(mx--TxIqM|fp^^gb2)
z$#t(#N9#jgGN1YaTS>-SPIJ5z-kj`LUGT`i@p|nM->0jY#CM<h+$$$A%kJFapru(t
z+UZkPx0b$s5_IGBcjbxel9Hr5O1|r8#qir)dh^bidtS&sqbARHkpj<7ZM(gEMW*ga
z`;)<w*>7K0SjpY~V&01OCV_12$!g)Xoi`=lpZU`s$k8#^xa92JLqfbZCexKglASvZ
zcW#@%EMns>yM*16`!c5rT#YO3ieNndQNUY#pOx={`G-Q9MB;8%y=s+_bbS<Z+Hdxu
zZT$}|CU%@!+w$%9rRRysH)XzdHG8V;|1!O8rkDQ5QV!Lul>W8#T$`>MOjSHS`=7PS
zl-ivuRt2_rh8}d<E2VmCm7;;$wFg(WUtVO-8Qh+9$5C{>m6`NX@5_4M7w<8WSu(L=
zM$|Db4v`AS)Xck!-}0<&D)zpxGJ)mpv>h6AUuCp?jGJRML*@{V*h$rwyk}obx8C+B
z+GNVn%^Azju2{zHJF{%QOR7l3i;HqAH~fA)shf8Z_k@Ul^}9Q^o;=M_eCPSze;XgI
z{BfqEYs0x6XFprE)py@uRh5<fYZrO?<wY@#wx~1jQx|E!OjH$EDwtaTRBNL6J4XrK
z@{PZ@yV`$o2z3ANv@Tv-|KZGviiL}gCksrSrJ1iQy_BzeX=mGdo2M&E(qiTI`!(}O
zrynfmZA)~DD17mIJ{SMHIWsi3-2FB6KtzRx7SppDiIDzFW^a@x%=yokIQPK0PwAVs
zDH)t!^RK3D<K$y?GbeW4RVtr;^RjUAjEM^kFMn^3h~;_Skm@4om}bJGSAI~k)~Ivg
z%~|2AcOE`%vSo_#+IPPuX)azdKlHT0+-A?qnF$ZhO<8Du;oX(p6U|RK-fW$`sa0-%
zEXz9EK!dnasZy<j{*M;yY28**u(|*4+S2cga!)Fy7e40b+Ho#(j<UJW$G$~^f>voS
zXRo`asqyDU(B0+RCC;BVtbfKQ8QqbmbB<~5?`MetE8Hu?|7a#1Rkvz(F}ZR=-R<=r
zugc{nsXRN^h##!sk~TMb%l*7Cv&^p8AkOgblL*22%U%b}IAvA+-`-hgO7ZllQ0<K6
z&jsRIr&lz643~Qmnqm_9dF#YTuI%L(ESHI?&v99uXe%}AYip;rbep-gqqFy|z<XJo
zF6+uemImqnG?DtS>sCkT`uh)qI<~5Z1pn3CQ_iy4SxKqn>}B)94iBr>A|F-0u&L`W
z)4R4<^zOsoDltDgI<%aYn~I#-dGoH}REcLRcTUxM7=3oZJgzU#%RgMM4D?+5eB<va
zYY)p_Uob)R&GL)vYTGA&^C(XKef`daeb0`ob(HSEv1o#@p{=sl<G$^`ZpklswUm2G
zx2oduQ^pH_bgL-@ZIfSKI(x&}<g<$|&s);MJbT3or;FPbEV6y*_+8S@I;$_yD`Um)
zFRO#pebOv17Y7;Lc~o(GGxH6>luuGtGq?7<?!T~W!t&+vR%KUNQY#YeXI}bj@Y^QP
zr+m7tSxHMC!_UcUE(uO6ZTZ)K`~2bahm|f_8JW~>NZ54xo)dfSlE2$3-d0`G5!f_S
z)MLBSwA%&-3Cm}nSwE5WNY@$f3pt1S?oY~*OIAJSpZzHKdRNEiEsqY(y7W&k^-iM4
zvAig*3%hOI&sSXu&u+2&Ebl4kV>~IH=iFybwm*d@1QZ;<EWcU3E!w_!&)VbPGynX#
z{C9tRwdHIJ{p+9q#{Rf-<lkO_55Lb|UnE!l^=0S$MV()ceylYS++t#V|8K|lB!ML6
zue0U9`MVV4l{{xpwLG|UqP1UhL32UV{a1}_IYw4Bo&iONYd;7)X@0-LvBzwu<IRn8
z{Ew>t4A9$fvU1%ao(1y{@8oP+a4P88twU`u{;(%kwV&9sKH!wWL8twi@zxCA_^0x4
zN1hL}EtC+>iF359WLeB;TN^Z2Fjv5p>+8Bp*EXDNc9*L3p3eV&8h54iWto33?p*CX
zuQ6xp%$+|vHdRW!e|d86YLV~LQWQhFw!HZH$mofL|3$YW&Tk)2zwlnE_{$OHXm_*O
z-AqR7678i|&*Xow*Fz*D_VVNA3D2Eb`x|pPWVe=FSW_dcsU+ttWb;jQ`}!V%=*s6+
zuE+L25R5(^-1E|@d%DMs)0g&(>}Yu&r*@%aPWIN^8LRg^Sr<OBsQp{qNjG7?reESq
zKFF<Y4pJynDp9@g+V5F<v+<l#E7L`u9jP4W1m`GQznsS=vQ<@P?pyU;vv>B`{yDxf
zCTaI0!&ko6JIjjg6jLqaqRxEwI?es9NAfWLX#=D7|DTOEb+v@2_s^d7|JmGsO%pym
zEkCd2arDVXM>p-whP(L=c0O5f=;z1pc~f32Tr)GM{x0v*N{y{G8ahmu>%VS#S|YH#
z??YJm<!R{)*0?}&m(Lr^6cx#cOi$Lr?kB<0tBfVrh#A^EdZ(7hvw|(0XFi|Xx1_IC
zhKVXi&)XE8UiznF?PBxz=|Kl0d(NapsflNIu9ism@_&@{OX5$JeD>FAUbZ_cRn^=0
zZ#lMdeO2FHz4cDs^yNwW*Q{xJKKEMM8@<Tr^2-kKJ%y?7G;Xj?{9xVqyl4N)ruc7*
zA|hvLu8-WL`5}Dk3?JDm^WS=HslBkjb*JIyk00Y))o)776^%&|JaqMY_^UMnt=AuN
zGRZfGgh&JwhA1xP{@>udOlpSUne=nY9GMDk+d3Yl9jxt-|9Rq+iGo2&!i4yy&vWE2
z`LfkG&fTgK<oWs7MuCp*f;88g*+rFmW}n+V*Z7v?%y}C-Ci*<EJ1Wuka^}nz*QW&q
zRxI_4tDkE)F|%jhk|Un!2Ft`PVjg-aJ+Zu)o76Dp$)DG|1k#S5OYApW@^xvNUGV2q
zs-_JFRcC|5=Qc^5QPTD1jtniyo$R$t<GhP*N?jq>4dp))j%PQ!IEA%5W|I8>;Il;W
zVJ5-U$bVbQ{g;*OUGB>MlEXlc`)j%CZOO_jPSy7>v$3AZEI*&ovFUN(@x@%CKP~iZ
zt9aismPO|~{|qW|`m3AhbmN!7*Z#xv=SW=9t2%w)_mt(V+aK&=<xc;8cyqnHaEtxN
zX|pf>59qcqnpN-EI!*F?y^e^!@6SW0?5F=Sc@kH#?Nza()Bf34`~xFW6Q5i2yp!Km
zQYPfxUjE8fRC~eo-!~c-R_+&)SNLQ7OtbZYE?37U?s#FPmZ$8hRUD76hsSfM9=p5J
zQ~ajoRvkNwjK#?TQz!Ob+LEbeuwJ(E{;4<0w}l>REbCc&kyRkZM2pwutwuBZ{m}QP
ztzT8YPv7{j*GVa|p-YLiTY6QH*95QNSHJb@zaOZvdi|`XrMElckdl_<Y~SvvduL7@
z=i6$tnrrd#=@w4DnmuL@j@{H&yFWwysob3no#$tU_b<A-Ea=irp8pq9B*fmX++$lN
zH904`t9lcsyL;)EDO+w7)_+xFW1arWyDH#s@5a>S%N3?nOp)c9@}T?T?2VDNtuxYN
z>J-JQdw0t{`LwC9weU){ul9kDUng9bIN_fE@yoes44GeoK+6@HUw-I&Xra2Q+Mq@{
zkoRiF3g6a5$)FAM&nCWn_)PhJ_JnzI-G8H|<s|JmVVLat;+aTEQ+x0<`9GJA{SJ9K
zi$&_c@xPO^v)aoKo}cb_am9)iJQ89&LDnmF&Mba#&(k#R=>?y2D&NlBxzMX$GRyy;
zx$z;Jed{jmI`>Lr$G6veXKdW{U@~9TCJm?G?V-yOV}4AEe$V5)+5htv#msD}EpH4|
zQYBmdew)@MXTJ84q?O>knO<+y?NTHdL&FaT7Kqv{-BWc|hkb3K4}-$>w>Fl>xr(pS
zyXPl7x^$_H+i&3x)~d@-jNZO`Tzg;RE?@qk6;j{C<o7T1>k-IT|8(@gl9&1i7dX1t
zynXON*5TJ#rBkJTYEKL&HuqnUab4gqWijQsFdO%o@2e#4>V=z0#5~iMGjzX_;_0sZ
zPT?Nw$((yjyU%mjoa9`hy*;QdT2P{5Ui<Ilh+4s!f}-}{e}?S3Hhup8)uBB`N*c>b
z((0N8)s0Sg7*swvm_6ld>ug2W6LKr|L`+pUGNaB)bVXQTG*j#3rt|0Q6notc{|;Ix
zbFA&4cYx6^odbc>FFd@k@~GPS6B1>?(?d*;g=@ZASQ2(+-ojTCdyLLj+-5uwc2(w_
zR=BjT((Zd2>lBXfNa(ORGrxzI(|mi1t6R@P1OGpB7JHo({;%1%z+Bneo1w<}m6eFb
zqR%hl+GaL=KRCsPS<@;<;GuS+w{P|Lhle`!|NAa@C|MT8UGaPBhK-)~U9I!CEEYK|
z_~kcOm{R@nW%8>zR{WUCd@)hfU1&kknq0^8|F`JqDY-4PiZ7Z{ek^3(w=MOb`hpLh
z>RWJH`^x2h*=TN!iAz%-oIZJqPc`8D@12JKB-3nc4*Z+CVS{Gs^j5Qqb<7;Mmg;b<
z+Y!kscGd7=+r76F-cAT}TqgKo`E0$S&nZ2|^X#k)c{;kM80&o(KcD>ZZRO>b^eZ!G
z&b6F3;hg4dpP$bsa?JfaVM1=f%aX-Qtc;du1?`yr!S(lL-hA&09`4{jA$uz&7w0~B
z_mSn>f>+#&WixxtJL&{W9<Q_O*q!xAoN?+0>!;c~PJMj$(Dn8un|Xi!URoota?)b~
zcfr4BnI<2)z0hULAv;-tUG@jLo*2|7sd#R0cTqThc|lN%`Y!Jsf#o0erRX^9$uNEu
z*0k86b57N(H`Pn*Ocu>D5dNloATi>(Pt;+#1s@VPl>U6}@-u&ZO0lA3@xi&S^Ajt=
zzV{!T*Yta{*-^(M0UYuLWp%H8KAqVf{(JJaZ8KhfGwfsA_0%udLE=Keo=-k^QluWP
zYq}&SaP^C?mVwi<1tyV~&l{|G9Nw>RQB2eGSlY}sxw+=HtW~FK<;^YqSIp7;_|oZ`
zW=6}u%&&T%S}Jy#L|zl0==M1yDpvm0`Fyv$j3b?M;(q^}rz-EKel_S0`)`vSUgp{?
z?@OkwJX~}|+kVT#GPUOuY<`DEz2UyeVmn9j+Ap@}lhj}Q*}M1Mg^)M{sni4WF0?52
zZ23B~IWbX~>AUMerIK2+yH!&z=Kc-3oolYU==T3fw`Z9fEAt+G8@@?@g}~7~zDtkt
z;#*F>`~GF|YLV*me`CB%qf?{>w*C#9^yXZO)0@1Nhm_v<{|@nZR`J=svZKW%a`!1A
zPFb(oP3zqBEra(=-<Y#q(m}6y%PJ2`t<_yq&!k0kyk38{s$N}ShVs$%K{YP-KZY#o
z__AW+7hk>&ca}<gD0}ze6N{glv&4%}o8H~~Jh|myYM-N(Sikk(8-Zcr9VN@F-|l#?
z%gQ)WFjVrI<i51S&n8~6zjNZ)Y_Iv>HqS6@KQ!ae1D@dca|?FosozTVG5URX(R4kP
zyzN}!iG@D}b!VIHTrRNaak**H>ZOz2Z>N4+8RPnS|6xtP?(!|2wz=8UA9CqDe{`p1
zQo@JKsa_Q>|7=9H9Od7;EIns>T98XG-@Y`@a^~d6sU5-R`x7k!IZpjMROgqj&d*z~
zaAVfI2^}lhvX!fOB;&U3nf1d|MtEIjYLH9(@}oz-pYZRkZ!F;wyZPyJe}?;&!VJet
z;hf^9*+YF>WCd4mToIZ(El6(H`3p1cG*@qW5prgm(<J%r-gnj<e$L@6^>cH;?y$2O
z{l&Jqet)?HPWq)!S?=ntIyq>GR9&gi)E6cEp7RBNRtY)2xbwGE@AdxO-RCvWnk?ar
zsWw=gdTz$vCYkHKxoh<VYipSuUs%<xo_rzQd9txpi1>$(AM_gnR(D(ICtkX5<#Xod
zn>Eko+0WLUeE4kr<(E4SZ~3M>``wLs*=^w!6B(O?CC-F!9A;IsyyJM}Fq@Q5(O%`w
zX0`~&^Iy&>pRnKG<CpQsx^*5y@yzDV?`&=7>z&IUG%UE^An=f}L}p!yi|w=T%bv3Y
zNy#nv?tVr2M3tYfAkTLlmx}Wt=~oP&2C45BQIWP1`K$JvC*3^F(!28x*Dk#|rzX^e
z8ctbxq9^A?py)08W0h>@B(qn%=Lk3=?IO*v{@C2Q$tya1g*UuA75?$)#U;5LKRt;q
zT)N)ne(tSj49<PC!hdJF99%T_$CYh|{CL^D)XJ)-YA$Zy?lO}lv%mh<QF)zBuU}b*
zXl)E^`yaBpb$xkT&_<Dn7BAhyF^f-@{WVPY>d5O^y|1-u;-nq<eix&bS4?SHyyMwJ
zUTJRk=5M+WolD{lZ~On@G@t(7B;|`AS?u2|`kFg=@$R74wl<bac=e8*ew-({(SPwh
zcU8$-@}XWezE2WaxxO9tpL{$wX|mg5<pn{zyRScJ`ee9w7f->7e1W(l8r%DJ>ol*L
z5?{Ldk<&_%zwL2Li$v{<RDF-`T=+>w@Jr{E|Ej|5-|PNeS*P^y|Aau+>cG$b&Z<&7
z@>W{1C%#(3)NpjyuVNPCOA}7sm_2QQ&)WWwKZ_E&TRWEK3BJ4_;9hd=HFuw%V%@rK
zpPq>!cLU8b_j*>!8Ht+w`EYVw$4(~Ai^eCeXa;w$TW@9+W*kzm*j_EK%shYj%}bNr
z@7pg6YSAuLdy^KmSwNWmW#mf9X`<nOH%VUAw3)V$V}p|W|5*2Dw;Pk&|C(=DUV5=s
zC8BWZ<-2YNr$5@y>$q2VuIHaM-dA~bU60!7_w(m^Ju0dFIxlLPQNX4bQ?)f`d)Yf{
z{?ESRq!Y0GN^t+eeUZU^1^1R&#fnF{HcRLFZ!zt8$}MD?_Qmg^`tOPvr(_a7_3X^s
z9@G-I;+fUE1r=wXow{S;_O0aS`Pp-oUb$XwecNw4)95t+lY5;PRsQc5dR_B(#n+~1
zkuOs&zDTQoG*wL^z2f+qw2iwqPBS|6);z<tRl<MPGplKDyW=dDeS7ikmd29Eht0R{
zxB1R&y3wUvrFKw#+eWJ>9*1{68>MS^3DzuTw=5Dn_OZLi$nEp<8`}iMUc^6%bD36}
zx=Z46mHMK--xuC~_kHsB{^C8tEY-UYZ%hn7cCBC2yPW+{OIVZn`8GrTH=-Fehc4XH
zPyblvrS&E7otNz4?Wf-KJ<QO!Z@A{Ec=T;qrc(=V{kYKeM`zdDh|kiibv?fGzGwIx
zUALh-@?!n{_`iF*pJ#8B)6u?}ew^Xn<PBcFHM8&j<`4M(yGo7sS^T-6n-RV8+g48B
zzI~#SuEg#d^TfURO5uUeXI~HARl5J8SWi)%Sws=fMbSG2yf3Dg%v*Za^-hRTztHx(
zL8o|^$K+kzl%f<^*)n;3s=r30(aC8ZtGmO+*S@Lx`#!#KZN=V$lYagBx<l+j?u9+Y
zOW*99CTcKy{{}Ac>iz9I>(suzJJhxNF3(c!g0m0w9GwpBxFh!9U)pyI*Sa&jSJ$UW
zh;P`IckgcR&uwNOwLcq9TKBwMY;Wn434Zs4*Oh%p`p9u~m1M{4&*Giw)AXWOD9tec
zS`(^0V}F4Kx1XxS)Gc;fy6^1h>1h?;u;tV9xF_u~x$|vI9E%s)ZRt*T)17yz@Aj-e
zzwEbv-q8_be?3-Q!tQe3n%8SegXJ6p4EFsmvDqpZwfz5$BAIsqH7E4{PFHg+5$|8g
zCZYXPcS(+p-ZZz*ncshe$nNG6`1z^YRiCXO@q6&WrsC384%vghet6h3>1|1DGv2<1
zN8*06(7b=DoPvMvz59OR?U{|q+|P<v@7lLeDPwm}(02(J6%XYXYYVs7e3guD-})w9
zyzrW$Qu=NW&0hCHc9T;kwNkakUN4r4)b-zIXU;6!$0De);_xE#x>|*T#s6Z9twknG
zl;4@crxL6iaoay;dVZ0U%g*Go*Up7bE={t>rhNOX;q9bbaXaZ^)P;#se|^QXegEfh
zx*h5&^!{3tv~y2D!kl{%EqsqMHwyR()*mbqdT)HNncv=F`s^qn<u7-SoqPFbW&-Q>
z{hQ*x6><vmE?yXS>qbi_OO5_dsroaEbU5bx(>=XoyP{IS?->pYe_Y*Gr~bKt`J&bb
z)+d6DPwsL)+_<aKE8w{O&yUw*Kgm5P5AU;HDpL~O;$Hgp{RZW`T`}hil4tSMH^1L|
z@ZK5i4cqf3BsU#;SAIkNMY&n}cAbV5Yd&7LeARblZR&#Svzu<NYCS1`&UMRXDfyYM
zJM{9mm)-gKOe!*ee!#u9RGE!`SLP-N+U~VTxKP`>Tl-f2kE*t5`+OyOrZj5m-f0w1
z<p}=2O0(&5?%jDcFQY>C=}ddRIx};^x(zLMvIYsup3U}IA9Sy0CENUuw|Tg^1N**A
z*w)@A;XPY8z5a3rYmw-x1IHK)U&S@P<EwK{6F9*0)Gu|`uTN4JSFijfne&>h=H32A
z$*dWEhyEBFE_uDkcjilPshhW@^}UY8h+i!BlwmA;r?v8>XY=+u3LCih&ENj4I`4s_
zb~<C(8QC>ow*+rD@_owt#_qxY#$!i$_wHXM-QKnNlHn$vs=SA<*H}!u?!Dn~o^+ts
zo0r!rzbsvm@Wz_!;f`ESvkmuF{OnnIRR2KpY_XaY8ODjf1X5gYzjSm{@_c5y^0(^}
z`=65)C-31qXlAk`)#FgT{P%0l&+H!xpX)H2(99Zs?$*+3{r5daBJu8roIdw+n8Ytj
z$(bwu@7%RR|D^jgBDepl*w(zs>?7aZf467Ec14MuD(E(_*mpljLY(=2>9zko-EVgF
z=`U`VUF^2Id_iLU+SguUoxj)kwdD0~dMk9lc+Z*GJuC-H_hz&$Y}w%}@Mg;T@Y!90
zCi&a{G$fqtesDo~`(oc`Kf3<tJ$fUMp?6#Gz4nH8FD=X3-n?ZMPnc9RWy!()OFo6{
zd&$)GWTmImEy2}Bb2r~xD7GfZa|w^gG?C_~;VaH{$Nrlb$@A^s0gdzd1<obwT@I@J
zV^dA#-BY6$o;&|eOZ35pciGNt{F%+ts#RNl$hOaNZIySsv_a$kf0-3;MPE#PY5)4j
z&N$01tGcoWKGHUBFTL9vB<^l}baP(5nT1<~-1<{aIsym&hZ`JR{cro>lJ(jA;U7MF
zY&y)jZGWqzy|JF%(JFOL_0NK~vt$A;&rJGrwn?z!G4Hdx%<QEfo*&=0zEu70Z)+o6
zfkR)vHM-v4aD(^IUSX%ne^b1-e%`%!^D%cdzPdjjS`J-(^Yw7Z(#>B#c(`_m#p+!#
z$gSMFp>2K6dC9t46BAkQ^UG-5JbIh`>WgZV?#-P|a}RJO>+_bbaBUP@mgCHJ=yHty
zt8YI(&&n_9X3%q8$6>m+vrSkrTPXX{srxnaUOw-1n?C>3m919}+rC?sq8+_gN$173
zva$;Y_x*XB{5pJXdh+$JEs+;b{Qgm}==j~uvQ>M9?o7V2?Z)2q_k(<-=h*iwjnlvX
zh__>ZnP<~=c@x)u;W<l>XheU1D;0W6jDP(c{nW)@79D2eJ6v${Yt`$=@^MygX1qy$
z?EdDX-|uCH5BpQh{bLo6EGpU89Y5igPw?jDeM{boZ?~V)vFoAtt)Mp6ZGWm`+-h5&
ztUa*yqe*z)k<&+7tZhve98^D<GquX4@%loKzm{M086qlUS53DtyWZ;&c)IND2_1<a
zL2Aj5-@MS@`sM4PS#J}B4tsjI1iJcIO*pWX(=gSbDOe;UYR*M*l@CW;x+IiBwi)e}
zo@_L=JM;{XySU83O^SyMoMohr@h$l)x6HR;>lC?0e+Bm5UkAV6dp9NF(e*N~pw_;f
ztOea`K*!0f{J8#-b?(zcr^^4G)nMOhzA>mr?av9dZreBwjWVvfOXt`1`N;pWo9g+o
z`u2iJ#`l}}T&!6?2QTl5d(Eo1)Kw$8Zt0z4YfcAxX_k6Tid1T;_iRx+dVHCI<)T?W
z@+RK{mhKZjx0H2Da(#%*i8Ai_p_4a7zV46K*p<e0c2d`dZRfPN#02a6>*-as`kdCP
z{CL#!QcldR<Yz*?HmeU!<hGMPwbVkRB17ZQ!~J!+5A!EgPrN#>qFsN!x$)r}|NbfN
zS6IK-=6!~g)Q;!5KF@459=zqBmKo;US=6hoVbhVoR{UXu5cj{zPlspQMsnqL?p!aj
zbxZbjaUOv=3l_B>S*$Ac?Z>KEGmgOPD~@na+pC$;zfQ|=i(&d{>B%l}`a&@iHe1;j
zl}yMfE>pHRv^B@}ik9NyuL99*=U4c!ay~rXQ>LP;`|PQz_`gM3OrlrSlj7#g|C3rJ
zzW&{W(luKCvO>KrU)Lv!#m2;CxjcKc=T3)G^gfZ+`nAWSPP*T^J>z*qRNQ~9))s+(
zSA-vPZ}7SM<dbq&*(Q-0i;HKUtg;X~B5<Vac!Qq3`p=nBNB4L7*t&@xKIpO`QQ`5g
zico`&Raw`H)x=ztc=i`LRZJ|___f)~raG&fSMx;X+nM|3Y)aw_*Eu_f>7+tkT;jy9
z-s(5GuK#-XI?Y7m!TFnUnZe9Qo>@H?b8F7NYCLsLq_O+vJ?YoJb1QvVTdk&a@~D*M
zo$qZ^&vJ%K8U?1zu*mnxu8Y(^RO<83KdiOpP{z753bVwvO|skWoh_jF;FRv0kAfFD
zI_4XLW|4fp@<AuUmQ^vITfF8+gqP*1k~ap|-1I9x3+eR4ZDUjZ$T#b&-)tSr?z#g<
z_eUh}yE5M>rtpyAyYI^n>7QFX%V=v~(XFf66$d_dXgTf8<eIa`^zgit_RKe888M$v
z=pBr`n_jMe)3N@~iFq>n+;0Z$>@m;EdmZs#Y3JonmO%pdr9_^7e3BY+sIGY1S?A@4
z{x5TT@q3{YkLdp5^w2lU0!~dyGx({h^Wwq9RHq(0)19nG_L)=*Ul2@c`BuM;%TR>(
zv$)WYGgZ%?|6p6Uzy93Kb_3^JzsI}U9!%Sux-eCZPs7|~Ig`4{fg2{54BTY<66c&t
z+_^yD@w1IPC(W5JoW6IDr^dHJ%U10V2d+tP(Rr~;*QMS})5hlK-dj@-1V8V+BdwEI
zbY_v=dj+R%9Xpd3x+$q{hw3@+UO4>Y_>CLwUnl11Ga44Kajmc1aP{T$bMG!4Es<vX
zxQj_Y^P%2i1_}MY!b#e9`m+xGohHkz+1Jyy{E4o{{1^K|HIo<jcX6rI`X12{EZC6r
zILPDC!@Zl;j~!e6G0IFt>U#HrgsGpkoKl|cI2G{Vv(1ITYqh?utJAc<PY9T0y*lQ=
z!83KbtHPIN@hs-@TYW!Bar1@OPQCdBcZ*&JWL7RdHnBw6`j=YrQrn5*Zf3K2B?1qw
z%D-&BXiBBbqLXL2ZFToqGj9LTt$s*;l^^4ZtZCj#I%}+Jb5BNw8kFsPx~}r7&iWHe
zd^ens(Wzv=VaoOYO{7oL%R6dQZj`;g&epnL=hw4qUfSyVpQnTr#=f<x`}p<s^sio8
z8`d;2T#Q@!e0C&5Y0}9M?bSbD>x=IH5~<z((f;ATJ8Qoy=H1FHU2MPZf9HDpfbRag
z4yM0#CuAI~)xDzqbBB9fW}NAtvgwNrs$!RZFeuusc7N*YJ1_4=%e=3*&w8`BP<-Eq
z8&3)ooAWp2GVSd?xnV!+4VPtkSKb~mt*gv+O+UCHXogvkQKQQ|>*Cie(MKd-T-zxt
z^oz}(zv}$&C^ND8Llyg{aYl=J_{{mxvy^Xr>P{|m`|8BJ$fj>8Vy%l;Z9B^&wZCy|
z)|Y)#P8lq{_4QTMMEClAty23Mw?DnQPk*CdPq&$Ab>7m_s?D=pSBT!)`yk_9Mcv0e
zUlz|luQTVt_W3j4Y?S}^H|XS7ckS>Q`RN&(qJ&f*)$Xhe`Cx6m{KA!qH6M7+K3ZWA
zr**qLeAVC2nn}L-O-qYJuV-%vEPFM(EPC~$*QNFai<WX-I<qohr>e+}`ipBFGFD!+
zUAQ%<ZL7Dc!^U{8MIQq_Q&z|wyj5elJ9kHPkyy0goC9x~a(2|e6u<rOu#D}t8zxcN
z(M!WW3iN%ubojMZ+VuTX<a0m$kkFrX?cUXQNhKfN_La|Hys`Xhc}VK=n(V?K8xC&1
zZq^ugEX7@xJAU0x-lA=yv*I@kZC-n0{vOu{aqgWtTiZC_Z{7UrTbkgrsHn5)E_c_h
z>GD3@@9QODvw!YJ(N{SF4;3>?e+hojeX=&0E&s?hl^aRPS=Om5%}+Qw_;Yo>^a{IZ
zYi}bGe%@q~h<AuK$J*Y;7|Y3rzWblh^cKvo<C>CKx$vn|KiiiZ-5YA0-k&`;$93Y0
z{b34!mQC;6a$#Zh=SS6TixQ&qX4ie2&~>8V#qvWvX(u;$FU-x~YvCg&y!X>ofz2=9
zH@fb2zrV0vT=2*3^@>gH4L7q3>&~_^dCYwM(R9aB$zM#5BL2JxaWW75xia4Uy<SA1
z@~7OVIr3uSi;LbVdq4dX|J%aLCTH11uPY{>uNN$OeI!RmTJi7eEp~G{-R9j}E!dj-
zac%Ecla|oK-gAG2y*hMizlyznr_81vxpiF?2EJdT1><HPG+HFSUbt&%r#SzsM>b|=
z_m`dC`uXd;9l|$yHvW4O_VNC|mov`Io;M>e-2c3jWdDs#S?&C#=4JEp+@Jo*W_@wa
zBhvH8?wiM^n)psFkrJ0E3U<G7&iw~pSe&KE!W&0J+YarI^v;_e;94aU{Y^7pM|}zV
z`h8nA{hDF^WeQ_c?$!-gU(VH>DzWs+E0Og~r#`OW(z}-XWLodXtN!zSTrYF)Z<`w*
zmRHL3{>PF!k4<5m|Gxa|S`fb@((Xmk@B5;fH-&UPzqQxpJO6NBw(7q5k1mVa`Py%1
zg+B28JA2P!(}0dCe*aEQF#7*cGLDs<m-R=&t`zZvH{!PCk9oH82$gzGlDty4yf4w?
zgMzAj@V`k@IjWaCb_wiFk>Gv5(jn65W5<rF#3?sYota+ln?Lu#Mw7|v8$M>Q+v0NY
zbp|h2Pw-z2-N%j}6&kF{>*jp9=)J`;`L>qLIlsgyfklbiUC!u!Qg6G%_;62?;PYb>
z)yrK{4O{-KOY*Uq^Rf2bQQMtjI%ZSDo*n96v(wzSR%!k5eJU9@W|UnCV3y#XUA45S
zdWq2nVWaO8rzFPh>G>r3A?DrRh``0$%)dtTe|Zuuxa?yRZ?bVKPlnGnvwJH!wkkjV
zxHYl))zxiJ_NDi4N-X7+uPL6GawUKHt2g#tr>8!N=}XtyzW9-#oz3Z6Yu~nJ{&vfk
z_*<--U-4sqsgL*T{OuQ#e!n*RwISZvEwJ_T@#S)px&B_icYNLA*sFC%q8Ps=Y*2J#
z&bVQFEc#)^NA4yA>3oK?nHAbM{`%~&t&eJnIQZmY-ofT~ceZov@Y~OF=+`AXj^_qO
zt3Qf}?Jj?reMyfcZs#n|4?A@Fc63SnHt5=6Ci?vLp6xeoz46&vw!)eFs-Z#ep+v?W
z^#?azv*cWy%5m_`E4RYdmTQFq>kkKB$hop3IkD|vtArNYD>3dCU6b$=8<PBmS4Qu6
zH2;WcTIdz~h(ldJ9;W?!U=eR&_vQ8+dyz+b4vYRtx>=yt`u(=+)IaM|Zu7pp&vn1~
z{a?E~zZZ9xwEo<?@iyzG0}qUKuAh6ZJK4PMYUcY}vZ}w{E0#TV|MYj?zufiJ^CR!A
zeHmT0a`yI5681CC3GD3RtKaRkf4fYq@}(CX^JL`@b=;}0TIf-6{`I9<UydDp`Q`b6
zFC9BI9)&6`;hFqGGUHg+bSE3JwA1hP=j?oFeLm^-$ND9P%~QV`a>Pr=HFwR}VVI-O
z=JJ%iZ^~DvdF7vW2=%M~ba)Ya-8wP+Cx<Izen;1l^I7SaqXSK!rIl)|$eVe7KBq+6
z;rVm^-MA+A<xEHS&3~Cy<x!VbY<{FIk>V2m*wU@=%sck~f>-QUoU@$&>xlP${*Ge?
z-Di!J+K)Lou6%b{xo(r;fmK#T?>NfLjiugZ-s{%Sx->Jo)m(LX&{LMRUfdN?Ywze4
z{(m-mLXWTD);W`=7|L(y5z*H_+*)0CTUg@E%LgZ_OjMneGB1R)mZ|6NsnhF{v@-ry
z*{b>A|361Ral>V;=7!v7moHwR){`&GW)NDN>FiMC{c3x-q^CqgOscNK)<=^)t@4VJ
z<eJ%bG3{=Wd~@)V<>z86<<rkL%53>-^55M_FHZ2z=@Uo)xdpmQ&0iffNwa>*(MP>!
z0(SrY_4VVa83M;@(=TT8o=phb|84(O>u_J0?R?SV4ol{|`pTo}@mC{BD(JxTuaawB
zHf3|k>veAND&rF8+w-}cxj04n^oN`kn(8ZFw8}DWS@2+@gMRLZDT}(aM6`s9S^}0b
zuI<X<;$1xL=EjsLk+s=km(@;6ZQ;$Bc4w{BzKN-tGU-umQ?`UWmiV+VWrdb!LC&F-
zP0i<fqPtQjO$ggJT_`v`H`ao$PuDPh%kFKFUhIGFrmeKm*~#)|#o1E}U6*xjNzGx=
z*mQi^B6FLAa`Ey14W{NSnZJ0UNcrJ6r;k11Sk(K~{LE5;4D*-0%gv3N{_owRyD<JB
z*ZhK4+ti+gxKF>9yL`p-6)RZ9<H{fXdRDV6ko~Ur15FQ+I<I6?x2Fqs{GP$ObE<}7
z&wqZ_escpSn|+u6xz1lQZ}CErvUbLxgospusxSR3Z4B)$C1v!UVOaEA<66(!gjVru
z73{2&zG%l!{M&gp?=0KG$wpteX8)gkHkd2KdPTBS(1b^kGp(LZy0NwF=4P)FE|sP!
z+0%RjR_1LA%e}EJ`wN%mjw8FDN(FDcUYGEJWB<lyQo(}pDF+pF-fIdg{i%2ye}8^a
zAj`SG>I=_(+x;kT-`fcRt&3k5M*ND)ag5EFZeQ`|^wLR_`og28YgM#qze{h@Udf)<
zuCsBkrpSjU?<`xj6vaR8<D1!fIensv!MyglPm}Yf9eF(ArKQk9N%@Pc72%bABHJH1
z&RzJ&XTFrw++WjlivKVFwB=0An?>h2ro~;F)EC!&J?NjU_KwNR5;k+E7Okz(-FAvm
z+bl1ZIpxW^Zm~oI^GjPbqglnDH5JTRB>F%1YI;}3b%|?LRiWJ}&K4T>-=<EDd{}RD
ztfFVLk9uNo;IXMwI_2F}@BCHU&Re?6S2R^)`ZXuV=%{IdSv^rs)|tMF{yrtD`96hH
z9-O{wYpNwEtt*o{Bl@(9r;hr<hn3w|>RhKSp5SbJYHG&j&jE>tb{9WBQ#Xxo*Xq8M
zA7^h=WQA|6`fR*9;P|;q=S+TdKMa_}m)3nE=##Z{#ndMoVx6vi|IU8s`P&QJTNmyu
z=bc>r&i3V!#Sb<-3)v%=wIR4-jkv+4pSNE>%$uAbemimJcHYU&Z{~eZo6Pdv`m1I<
z>vX&6yP1m`;%s6jJ?Yw0rxs$*rDoYGb#?JfOWll*6&LN^CB55e|KL}$vh19cFILjR
zBF4{6)_6qtS}fi6&toZHGxO(_9->y9%npaXe_S#BnefT*CqE3h-s{C!pMCYby?WK2
zf3Ln9{pW9gl|513{72vOe|GgfzTc%DZ?O8)U|i86<#0OxNS4#?3TyK%0vp6!{CVz(
zZ`H9l?9#G5&90{P`L(%K1#WZM)Qjtcrt^2(Sgo1Oa$s{&MfjnI83_^*%br!3+aCUr
zGG*iC?Jt8q{FtgVz3N~)_q10V`CFHsK44^Y_MC^Azv|cMJ)DOUF63_Gid0PB=DoIZ
zr<0z2UTa=zmb9~$-$J3zR;jDUXC~`yoAz;`i0sj8y_N62{GS%7squHoa#Ow=NqsHP
zy`DJTG<iBbH|GP-@?(b?*9T3~HZyWN+Il5%qW&56g(n}m_fJgYc_MJ3zBNMkR*A%c
zb(b=p*>^S@{s}M92xt$i{X1p7)8~a3)_TiZ*JOSaKCqBO`fC}lcHH8^<(mrPcw0By
zYM9k-dDPaKa+3W;{o^}-5*iqs>@`{x85vl@_AGN=HvQ48;FSW-M~-tb+{xeQs$nkF
zRhZkEa-Mzd`Mtg0Z;BM0Uoz#=n$GfPkvWmlTyHk>dV~bt*ID+K$8hCQ%}(AL*SmAV
zKg_cTbe4F@)nC%mylbg9AFE{7?UlE?YFb4;UKD>?`od?0Thej8r9R5fQg}{tpWnL3
zE76v3|I6)HJpRqDwNx>-_l#xTf5)+y>+RvYv;Nx1ves**{P>$2|IzE6-v8~gSrI=E
z6i@!Ryd}FWs%1*L-2YQy57R&26{$b}Vku9&r_DY=v6r0duT+0bx#Du#M^*lB<e8Nr
z9rJj*J}29r6J6u|GF#lmwb&^~C;fPDlZp@bQl-@Et~2MUirC!OSXw3gOQka?&yTC;
z&!e-N&Qpz!6rGsqCXy3gYpkh%dCu8?H&1h)b9y;hWRJ+psB@zBpSB;XzpN~|>&xeR
zD-BCe$jw;g5p>%2W~#i>5`n~3tS4JDZIh;(?W=8(V|wT($@P=fPxrEbgXm2;k<UM7
z*veT7A6~HJv8I3ba>om6ndQy?RC%6y@%fQ`rIUy5{b{{kGu;17l1cu`v4kuCAsdU{
z^6#JH_KKx!b7-lp+m?4Nbyw~M{}cb`U)MboY~858<;3yALLXk=8%J#qd`yjT`d(y}
zwC`Niu@L2*o3;ck7qQ>^p-x9%>`SL&%NOhTo7MeyH+4J_c(I>pfA{=`W(Lt){>PHP
zOem|Ce{=F<_Q9wtE03lv_S?!+^LqIm#^W-*c1n+RW{3oQ`TJ^G^Yw*d_ZgcKqU4#J
zAEu-%5PSV|XFcz}U*CQ9ESjaaS1r-&OwgCk+_~nOt^L1kb_HGJ%Q!7L-(_K<+tDX!
zQQiwJ&v49ZjhEv-WWHJdI8RGc(@FD9D|&Ku-Qz-3)ECby*|m>TlwrEG%~FYwLkD?g
zbKO)FW;;|RV^jOh{Kx5yClo(_Z!Hjs|Hvkwy)E))|HXr^ZSP9j+4iUwO+6F(+w$x3
z<mYJ{(_MUWemL6y`FkpEyQe{-jT&cSe*B&ELpKu^i0x#1nvmCiEzQr<M|>Xtv86d8
z^Djj`=RFme`0M<dhq3POl724SufNj8M9siTms9jj&b}|QGs3@BMNBf=5b$%7=K3i+
zquWH^IINkpGSekLa-qWK)tBRUtubufF;8T9N9Zx@%(_tHVvpd*mbG7|tYGQ+GV$ex
z%|BW^bJQ0;ww!Ty(}O-szRxz{7b`9(oM0{cykOd26FIhq&WE3#SYOrQI8Zl#(PhKW
z+g;`gYA-*;`PZaWL-ia_>A|Le7{TB@d^a;W+>XqiZgB5cPQ;^t)aX|ctDhE}e;lE}
zt<)a&BYMsX=Hg|Kos0i|Ie*X5TYN3sM9+`T-<R&2ac7p!=AWV~H9sCVn!k5ZQbgd2
z`jb<FK72?puyAznyX(i<;bEV#vgUQ?;aNppYg%iJFNliv{Ww)nZXS2l%VPOEMgG>-
zJ9h&0a?}@|I2ZN!ePpF2i*-(CZ0qLxj;&m8)vA)eO{lGvU7gX&Q7*x|YU&N&r*@(%
zlAnuJIM+zbdt%9|=d5dEDEFJ$>B`?S|JF@A9dte{dnRDcIWxym!+8Dam8pAnEB1cB
zq0w~m`-k(bJKP!)^w^#=m^)whfACb~V&56Rm3G>{w(Q=Q=f}O=y)Q;Gv`26%yGv(p
zqTBSxJ`1OPljiRfdU0O4%C4qkq2v+cmb90%9cK!PGn_aU_G7NCDrXO0!I9X;pRGsY
zXPhdj-e|v6qgnQh8<%pr?Gqh|6~FWnUGp!$Pz>vCD;4&?v3-Wp<^J`*UH&YIu(ox*
z>AN*Uv$MH3|5wPGWBiXVYsR^JFX%kImtXAb`DuMV@<+~DO^b;7|8aMA>LW4BnVUbq
z<p}VS?<(8Gahm5)o!hLZ{4f9Ckkt}P{doJ_RW`1?!i}LX=6#%I5NBQYHs6`6=Ajr*
z`2Sg{`bPik`yGQfr&(`iv$S{IA=+KvY1VvXs(p^AeCtGoKd)?_s=FQjYhHL}BG<DM
z$M4-a5%c!3-(21cJ?nhb-_9@TD4(MkYAy2nlG5pZv#L81I$O?OTl0?Jr*KxZ%MRD=
z%au~E+n8=rVdIpVb|QT77q-A-tM>El<v1rPnB%tS7^~*Jis;9y>Uw_czK|@KcGv0X
zc9)sf7f-!Uyb{h;#r0w8PtO!r9^aKKc|`52Gs44NEsCQ;PEPH9pVhH@dV8~yL&C&o
z+nn~w+oa0)eXAEbB~kp5V`t#n$A!Gw^9@q#H!kzF{#MPf;rpNdo~)UxB(^o|{NwP?
zdfB9ZyJqJ8?tc2eZtLB2Tgz*&qNF1K{`<ae`^~z2Q{&V*^ybYma(wu3FRPBl)jt&l
zixnSbO<2^+c9}sW>3^|=VDfs$9D#lBvO=D&vi*PL*u|&yyPRsW*H=x|<#_(rMP|d3
zG7jqvZzSHSH%d&YE!^Q1b^K^~_vH;+l6D{aEz)%$zkaJvd+3B8{1S)S55~r1T5P^@
zhTFFHm4+w#>fDba+ZNgP7jo&CO*^7};dR{+uOG^(_g<`@)Y;6QHAiHPMzDW&qm87<
zi_On2>ZnUSed5x;Wmo=|%bZOqlBv2AS_RdD<}!*3Y<m%q+jP6&;l<YT^Gdq;ENXV~
ze2970l=A!Y@6LB>YLA_-w!i4<ev`jp+U+9MB&EWrj?31<w%vjf=|^ANhHad(tkCDo
z`M*iZ&kCy7gr-&r@j0zkE710FIW60m<K-CpH+O%eYMs}-4}Xo6?(!8MYm*R*kNls1
z*K_KE`M(0voc6c;`nmn*PTPOl2aRg{UR_8!bm_0b3g>@I6q@}jFKk%JW_`%f!f0-r
zg~sAnJt?ki6IC<LcLg`u>?u0+*Lq&{o##3`rZ_SFzL+5q*eZ9BRbuh034!ei3oagP
z*nP`G-L32*%Y<FB;_l!78(;WPz%fHv>3#HCk7GqVKXhCY=l}h(S8e0;WFsa0>kmKt
zd>(aE{o1d4yQL2Iq$`~NXDB$mMMX-=<#*OjVd)sdF58(sTQ*DcW&GXrwLPKg&<*jJ
zpq7Wd@xmwn{a;qL^f`x{&1>Vgr&&zeoArfkGpF9j&Z~PgZ9+oQE0NV~^+KHuo3rOM
zJo(*n{GHzy*8{S9{|fC2d0zGH_?iW3TzkdmEG?IqUwkLdTHb#8k+<2}$-;{_-~6f-
zx;XIWzf}>8%Wo^J6_E~+zIZP8^6a&?3&m~p|9w)oelg!VI>vHlai<>7&ivKUHLGJU
z?RwR4@KK3hi1>z(B~EWxT{)IL)M9E?<DL+7fm`pU!t37ChU@+Wo%YL1IJ$m|cld<7
z(nB`8<*r9s&r0{HtK^aRKQ-!`|Awt69lrP1oRqYAe0P;tgxONlFUNyieXLJv%za*V
z@lGo5*@VwO^K~=3esSb%GB=sOcTSPMP3`AE>n<+a;Ojd+oNQir-YaN<z&g&z!}cNF
zAr?_V)90n|etLH0ho9OL@tunvF0z=`aOAY{s-q7J9j@>CyCVO%r{eLsCJo|D*TX-j
zDsz3B?Y`%=-jkCLAC?wPS;PE~tGIe{(Jx!uPf@RX&t$Cr7o7XzkJw&ot0L`)tpyV`
z*XLdp?}-u8e0b(P|K_;NZHJz1eD?HANLcpA=+#XBdnE5K4dyv(excTw>t>liiA<l{
zear28Cm-JMdqVo5Pyd&hO{?}%I4t{TpU0gpgLB>2f89DJSafHvwZv(w_?bP4f+GEE
zIoU3iYNn<0gg-lzv^DLSL}|5e*6)}it%-NtPieYKyj@?iv{+pJq_EBxOZ|HFKb2oL
zm9MucJhAT3&u6O^O__UDx@c~t@S^IR84~Udl9M&G%2rB*&Uu;1%K4MEkzek^x<}eC
z@4LwyJMq40$|>>O+p1ff>;iTMZna%-bB=}kwhmKK^{vvE)xShd+!o(5r@?mB#fTb#
zmpLh$4*bzpl3X?MkT2_op0zpZZ?=EiQF1p@*nMUG>nmFdn}4N>JLcV3);-zWYQlfj
z_1lg-^9oS78C##-n5l6)zP?pARp70qs_4P|-E)6m|8$^<x7&KrdzDFDbIxB_x>{1s
zHLS2zdu?jW;-V!{*^?PBNKZ3cH{bq3=YIKc%gnbIgxwcg&kfqm_5Sqzpr0>V>t8JD
z(D%NPB6nKchQmDc_tQszpTx7i2(9<BD)jxkEPb81MdN?zkB^SL^furyU-vId+iK^m
zSDyl7&!vfMT3k8H`qHmnCmoLKQE~^9qUUbz^HRL7aGbTj>{+h=WaeLA_*tuNdskoJ
zi2M5J_f{#3-}$*^&u<Dk$S(^$X0ZJ2M(NrUHEsU9wP_2zdctn+x~hBP_?>E1J>G;9
zZ>Fleo1GDIc(&Qk=i=o;A<u%s|1STl;kmf+lj^o54`h<H?X53)f7mTQFPL@DU#ETC
z2Y()oP%b)IHNjMT^C{lMsqc4RU9M-e{IZxql)Qmear=JTkifiW{?num>zvy^Lq6%V
zYKq<AdYiHujjGd)8VBF(KdRsK$Y{RO()}EF1pf2nZ7o~)Nis*}pNCd!+3_Pze`j;)
zuC%n6v_>WTJl7w^of|~&D^*lZ{@ZroSCEX(!JR+)4zKomuaZ~Lx_;xQbyr&!omO{Q
zdO~3K2Qf7tD}hToJ|}vNy0wa8HnSYGUT$${bHC5B!pR}JGd4vl>P4nZQ++&BFS6*c
z(o+#tFP%v(>ntZN74uXwnZxJ*>9oc}z30bG*6(`bo02)ng#Wzs()K@V1%GgBO%E<O
zs8g}~R+{K}Tci0oC$hwYBV89<`OK|uWBXTm*#X&Qyf@EjF!y{av|i}wXZZJWr<eGN
z{nh^%8J!se-)gbyE^J<=@nD1d-Qy2*{C<2qZnJ4cLG5Ffht5v|4Au!>)?ASCm3z6)
ztP8%gBGo$g%beWoJ<Gg3&FS2kJ6C_#w@+TTlY9E4ucw~m#klTzy!J-1%!6GOHXr<|
z)~`}(sM^c4wwT-c^LO7*zx)zTKD=JEkbAl8{J05U)VC|`?VkRuvuOL4=r!lxD_!E&
zJ*5_ZKjUMG+OvRY^USIL+jl9Q=4Wo|oW5lF^LtC~h6=O0&Uzg(S90ye-80jR%bG)O
zy6@RBp_}2P+V2LpupIU8zkeG1^`7Q&`bI_Cf(C=Y3g7wRyY@6ryT490Uh#y%iuaFQ
zXNNqEteU}-QSN(TP2@!VO}0gvO@9wO(9Eq=tUbHR@&7LSiErOCNZf1R-O{hOx>l*^
z*dyI=-AOjJsTW!=oyZ72wrk?ly)A#wPwwT<|G)38&;H0(j#|k-O4B!MxO^AMa+81I
z-S+3sZuWQcaxbY~Sl7^;T2eE8s)d(UZS&mdZcn#e%L|rzwo3R1?@vr=|99faq2BnE
zTdmFFT#NLL?45J=#Tpqt=WJ>&nj+48l5fRn(;GD_UG$&d5S{Un|EZRvP5n+kj%#1e
zcu1am@h5SE(V1`YyJFp*{6Dm#@4y%5|1x$xe<~Iqdvm@?(`Uo&*bmRdm*@(tH+ubP
zn})UF^V=mEVZE)lWag@PPn(o(U-~;z;@iGVp4bVmg>G87g{KLBkC~EDZ|)tyGgtKD
zTCb>`FBWUgeCO;EvpRouZ)>fNhr;czQ@@$FL|;2&R&lvy(YD)fR~+Sv-FC5c<7ZQ)
z>YO9<aw-{)$=$kRbnWnZJqCrFwqoyvwQ}1;{yoyI4{Vo+OYLmU4danGSk8N2cZt;o
z#~FI@yC+!LXzpz0e!X<>W6|k6;)@mgpSNYF_jL31{yq6@g;H|ZQ6*j$&)dlyw*vm1
zNmg)ZWP5%6S5<Fzv6`FBEs@i&i`fh%m5+$Lzx(I);%teKMb)R@hRiS$-n(g<Kt_^M
z_+I|3Z-JifudM~WDx%9ct$Xc%xb}(J7rwcka`Ds)|GP>z`{t*0E<U^^(JJ}6U|I6%
zTS{qrr}byQH8)b~o4<H_sB?_uYme5+tVb=q9vL0oFmL_!y9o!n4ykp_nj$`7v%&gc
zJ0Iy;Z&po@RdKqg6)TxHg~#`?S!vQ^rSk8uIbAN=tWamV@Z2aS{bb|cSM05ytM6)N
zXC2v{wPah!>w{ghi}SOUYlY%Z8_ZCPp5Vi6{W<YP%@3E0+;<&*tls{+E8eu=%=fG*
zI=2!Xd(`jTt~&7~DcM%C<)HBjM+2SzcO#$K`SAZ;At(6v-(2tB9GU6=wW1|0C$Q<)
zO|INJv9?qDmr|g9TT?{jy?^e%e%2rBk?VIkwEuzIYT+}y)g?>raPGa7(bzrz(0i**
z+R3MuUY9I*f0%iB$bn7UR!Dr<Uv~8SO|Hvtm#cNXcV*3p^550_C@_5SdCjtr=I3kI
z3Q7yij>xwU;|YB>cXh;9|AeJNoo6m_Jj~sGv7};(?OcZ!k#(`3+kP5^o=cmuv**X6
z^DE|v*mC*bl1mZT{OMe=SYDD#q~-6N37gG~XI|1d)Ge3jnh~o%<t_h&R6i3b5s{?s
z{sloB#jHd`9%k>Am{c6FQQqdKf&a$~{;c!<?yY`)lw+2jycBms^t|Th6;o>Wu0OK$
zpS$?~oi0CCycC-BqT+zl^H1}ZJ<CgOiF5Y6agy8YrxjoSn~xk_J^5R*G@2g%iO+l5
zm00EN5;0@WKI0WeFBkIU_bm#DncndBzgOv7jeRCgfv5g@=swKZ^5w4ZlY{#I6{d)V
zHT{0P&qwf|ieS^@ukvR<<u)`+pW~`!$y<`JWqFwYj>moMEG#mGCpsO{4bGH*yvWfp
z&$#kJOS9z?e`V37h9!#)<4?)C?Ay|pYNPk7U)IZD`S%Oqt<`IP%GWP!^OgM?eqw38
zz@H`Vx${kw9G@DyEd3D9VHGOSbh}<EeP!E}+hV0W<we|{;a?RCUFY*3i<+o$WbMxX
zAH|xjbZSy14(za(m-=Q<aCXniyNmd|oobZy)7?yLOr;gJpU<9L?PZa6Z^|p}l8ZZ!
zp8Ws*{EO#*^FA*)le$@c|L+T_pT5V`Nxkieee;|D%Us3Q{k%=3=7)^-8q7cNd*+WC
z#@RW}HwA0e7fZ;k?A(&a?!gmP>boUg?(Zh+*IvJ-d&=1J-hZ^}Y|G+RbFQt--n3YH
zbKpkddr^L8%q<_j{^!~LaK0CZ>!Syjo{DLoN<tsi9kPm$z0>t1ze{_!(3RVv9EMrD
zw{_-;EwysiJye){<ACEuW%G!oTiX_Gb~$qTSc%%j7kdsEN+lP)x+N&lnkrfTd8hBJ
zb4+~itd4JxW1Pyc?_Gvi!(J!71iv35r<bru+P==XFY^2GM~@Y)(;h9&JjfQiM{n*&
z|1$dx-IsT9-aXVGaLLxeLfA_FrGBJ_k;fc2gQC4(9_yE{?MS$zUFN|hvqz6-bLk3!
z6lU?WBH2H3<(mbjRDZPp>G{y|SK^iWw+B5W=3Y{+<OnP}e}Bs)&*neRpX!UGT-cK@
zrC6t{wa|La6cHVbP5;-HP4Cp1Wpz5P*DGSxFEi)3^I?fqM}pXje_z#%VXMA*O*5r^
z`_#}~8|EGR`Z3ja{snjYXB|Dk#nPebt^T@97lbFW`v01ge)jZ)X6v^nt{&eJ{_M@K
zht{jCc1mrZ9uTKE?||eL$(i$JuINfDUH#i&hT=Wp`e3her)7%1Do18Mma2aFhU;^y
z{<Ue2uQ}E;<wvYmY>k|`D1WQ;Nu!Dkv#{*X+s>Q*zr<}9ve>0n`<ctKL~Gyok!4$T
zbh+H7ZuMR>`?Y$%vxUa)?#!!PTzeLU$-i;^bwr-!k*+q^_7``z{#d+eVy=h#%BpV>
z>3?cDg%5AsdXd-IW}*4!%Ny1+9S><{5)<wJp75t*>bjG5mjy#D9AB?3&2yZ3{D)xu
zi{@~L?zltiKh<2@A7wA~>q|?m^VO%O|J4OP{<jEzc4cCrRWH-ixw`tle{WfT=y*_5
zi}tsw_O+X}Tz0MhF*SXK(%au2&7aqWc_fCt`jxgn%xc!HD=qaOUuPD^Ph2f~-um7W
z_A9pcrl0ulxTxO$(B|d4_kDDgJty^N*RfYVYut*rZ;nm&S{rK-w(mag{Y3AjTSc$U
z@iuzuwySI+XHQkv$7Hc7OM<$Xk0o%Nn%nY5#`?$(HJRiaGd2is&I;J*TfWlnaOXSm
z$SvZ=uXZyT-8Puplpmb8X5E@w$L4(2OgB)v-*jdf&$pJiCyqZK2Ilc-r#|kue%-n6
zX0=GWt5S-iO`q6ZCKgA<j%=OOIvbs~Ifhe(+#D`QoZHbJz%_Yg@#Q)d@8qS*>Ms&i
z6S=?d>tFNjG3(xY8`d{--(0Y}we?Q^TK^-RZ6%9Ouh~~y5ux@V&+PP$ZE-dtd*<Eg
z{?KFd`<8Au*Vh--Z|gqU&%Se{r9S)mrfL6{-&Lx}*DvYb{A(TaL%%!m+wU&ss)_%d
zBK}~*`ny|y{+OtrH>LY-?yAr&`?(8i_Z42(RJv~-yXe}kX#d?GGvi&)KXJ_F`un(A
zt7zXb$ynC@ua1{>Z+u%4x9%9@&ONJjeYE0k=Usnsd}GA|5BFnLHQUyoIM{n`v&#he
zoNrsU_Wz8H6h3@m|CP0mAMN5&*~|Uk&dWvr!vq8Eb%*w88GX8x`%Lti>XvnPOZS{S
zbo;7u$fi@vrp}Ju{G;j3>yF)wSw_3~{621UO6D$R?a$t+9rwc7bnC>a-5d||=KogA
zQ<<{V@R)S}<i1wUwO{;8zkOHTd$0ZcuKR-1OAeKBRzC{~y`A^9a3^<Z+N$Xy`W5>&
z8n)k==pwx)Ywg=s;kSFXum9{ZS9%W5gQt>f%Jg+DU&u2)UNC3jO!iehAJ@t=FDX8J
zi+Mu&nMN+Z*|R?D_2&IJ`Y_#b8!!LeKcdDHpLv{CI#i__X526IK0Za;@Y;E=$)?5M
z*SY+7TRlbAB2K+@XZcG0DG$?M`5!gjZ2V7ok&E&1RcbBgReyf<P3G7$`;!pgF{?t;
zaOFi>N)ob$y|$g*c_trSkG$?!e(cieTS03i;?I@l`_#Nzt~jYD$J}jaYuEbEOl-E<
z5iBf=8NS~1Tr??kXa4n1Iv&fGhwj=iw^eL0^P;VJdY?HjhTFRBx_fL--<L!C-^{vs
zY--DsUq@{YU#VQFa&i62D=R~}&Ye_>JA1le`EHdtnmesbca>cKroL8JP&7zn=IZ71
z?)Xn;oj;|iL3jSob)A9kWde^R#Wt>*(QIvO$>WkZXa4h)YsLa;F0LD{z0r{UBdhL~
z^=MW!*JrWPWBQSoi`8^&);K+$dC+K;{HM*wgSK*?|9vFrAZOhzmE#NV-TS&&s81#T
zuIEE(>wha<1b?iUxaqA}L~_wLJsFkv>z98LjCbp9>0C6^HsdQx>FN1#t}|qG7OkDl
zt^4p%bbVu+U;kuxx1~*z_FPWWESx_4O;y^>tAAl-@x&RIB-2>UFCGkfe$?Nqb6u6Z
z*_S<Md|V!!P31k4aO~1@_eV2>W~Q;XuW<SLYIPaE`-Vx{M%QYSuFDy8F;;C~b$dog
zR1I_G`Zt{yS1!7x>Bsr|-BK@w*@<;QS|1Bvi0VnmvbA(Q)A?$4R5zhJ|E0**MNik;
zp4!WKqwtr<<Qb8I%UjnkF<0Q(u6q8xt)=LNN^ukaee3o&eO>oJw@A5Q$G&jB{Iz8o
z&cz2E7jW-*aNm6U+P(|Uifv0a-aWB*OGii7k2Kf+g{HpAN;BV|IkCH0S#DXSz2-JI
zOC8VNvnRf>t~yW@p1pQc`lpv6{eHW52OZ~H?<;tB+9W}RuP<y>wHS*(1nWE1%@h^b
z_Qzw+)Y&n|ItstVE$fQZc>nB%X7Qo$s^$%%x6XP^d2pemW3oc8?z`$(jTt6)uXTDa
z7XBuCYMM!A_JvrMwNAV$Cu@R=7A$}Al4p~Xsp*c~8=<l7N>dhx<eifdPAvJJX{?!Y
z^4+r_qeVSEwV|ptF`Gm=cD*V-74^|Ib5+nyuAU^-{an}eq-B>cixrIM<D9y_WqshL
zu&pm|Y0bP|`^zceRBSm{+}bNKU%952#cX)Z_)q8ioBt6{UcA?zyX#kz!~5)F-Cv9Q
zKAiTydb@6`*nxNK@2lN@ep##k>aXCoXHzP}YI9cWPb$0eR`A~LRY8IG-|RZ7{o-Kx
zzE#r<qW-o>T-e3;Pb8`$;%v#Jgzf)0R9ScYpZE9B#bPBUS-u-5)|uUX|GBX7y2rZP
zPlAG!w5Bao__h8eSJ9HAY6(Z@p7^qKwn+Q)DR)dI_9<2g8^7?C*r&Ru-e<X@-nG2g
zqnp3`mK-jQF`DAJ@ZOxH7e`j!n_5u1g=fA&;=N5zwtYBw$m!XaUniUeP7Ae}JPR(k
z$R=3IS|fG1Zp)MNhbN~mlM@nMz5HOW*A3A|?#^k816gu5)lb#3xXN0;^5KG&BEPHx
z7bd1x{w?B~Sn^U$G4r;EUzZ?XxxcxiTwc>VI}_by%M+TpQokk_sg_85Pw1`p)tmfe
z?b_0_l9RLTuV}AWap|t$=f`H^8_u5ml`pXGfL+L?`(DbNfq%*;r}E#LcJb`Ow8t;k
zHM-2Ty->#$v|qFNa-jCfDTh8wPt@K}k!TvFcxlC<+UsfJS8i4YFWP-BFg0)SA*bC1
zc_)mI#p$zsh+Oz|_exz)-ka0if4Wa!=F=K`A@W{+d6}_O)gl4DX506h)ef!o`o8Af
z_h9w*khWHa%`bHiI!*I7_7;k0o-k+5+nH8nE{Z*WVpglq@Nv!0U8YnL694<E_K$hn
z7R(dRj<;A^BAFfHurM!b#l&|3yKlNKO*r>^`+)=NU&S7ex)Gc8e#=D8#i8-bf>w*n
zeH0Wrd#cm=MZQNZFE$CdExPKol{3C9eb$$qugfd*)sNneXu2b=d*RuK^F>LbKaN$Y
zcg&Kjewi%c;yT4xNvWb|&c*xvoI+EMPLnvsw&IIwy@c_sSf95He**S(R<@{h_jSE9
z(OIsjFP0bMy8oST%i-b#gTIH0_N;u-Jyq!W&d{&fFCs+_w%65N%sseP^i=T7T2X_`
zn~p4E?^#n{(VViA=~%zXYp)e`>U?*KUMu^Y6F4duBORwM_8@YR#AMCvuU%`zR04XM
z?R$SN5@y;z^V4Rg3-6LHiagqyy<VDa>WjS(=4C9Cy?McDuNjx{igtljQbCzNwwMZM
zWEmg6?KJPLk#f&=p}hhZB`ddXadj}xJ34uThd_lH>w&G6(-z(``Vg+T%4plq@QL47
ztb80jBQD5kon_dMN$X=8o?cjy9c~b)v$lS%&&tP%s+ZP^Xo`hD`Y)9lB>!S!vG!%9
z88#=5*89}mDLi7aI_$AcbBOjnuPwUPi3#U6$!XL*3!Zbg;M_{lvUZ0>2eR#VZP6>-
zu<E@0*)U$gsp)yiX%c_)Ev=`t#&;}Oe>kCiLhF=^Ul*T=wXA5dan|4b=Bc-tsoraD
z8|P_8-|fGwJik%f*!);bM{!5&t?wQYANxPA=lJ<|?TZKhr0Z;ypBQa8_Ez?Qh?~&j
z@BcpC=AZiLoRyWRUXfel5url9z=fVSPfLpm<z0#h>peV8B!0eRk63JLV_w9jgAd*=
zcD?Id=WES%>u#lumH9EI_Zf$FZ|hUM^6F()u37UOr~Pkq+1)mA8FGJ|y8cFA_{DqH
zR?$VWFQ+?K%!qD$v)in?W!kjc2X<YKw-&uP|BDmj``E2#b{nkAEU4c6DRa*g_2cua
zu4}rA_v`%-(Z0U#k;z4&;`N@Fe{<NGekge}FEL7O;hXp2%Zw^=3S~rfJm!C%;NRnR
zSZ3)Osdf1xB6{kULhi=gsXBYx_}o7pV`Ht4dRO$~(4X*ke;2*|Z{4@Z)4@z!H)v|d
zieEyDuH;4VeA{b$HQkTHQEK+`ElrP9U*)AVdwy2f%)Rts`rVh#Kbnt=)|!|k&MCGI
z7Ubwyx2%|_$N%!ucU2RPe)lO?n)Bst9eeXN&-1%ntg?8Ore26yDUi*yS3P|1k1HK{
z`f^oM45xh8U~8Ok@{Gcw#|n!D7QLzJc_hfEX8$<oRQGjh5xcO6$&=V_?{$=L>fUiX
z?LtT4db>qaBNo4Wacu6imM0GmRlYD-BoeJR`{uJ>{Ob*W-j)Bw^{eWhPn))dy~?W(
zDXFE2_BmfaZTSCtx?ba5$uHkpl#|+f=D&T=vvP+2{U6%|zK9!TURhWW;`48vV-x=;
zhQ{0Z!Ca@iYU+1%ebf2=%W-|8$k}zF-f~v!Ca>hO%v-xNwf$i0y{~ne;$Hfa^-DCK
ze|y_JFXgYOxxa8>{Z`}n=wj}T+Y(rBad^lpmVZ3ck@-hNLSgs*x~tsI6*t%&qi0n6
zJ@_}f|J^l%j_#cF*YAJGpI7)^@zF*ld9wY212ZmUp1Tq&b5C>1hs!~^A&Yuaf4y5e
zRqXa9HQ5*C?t3kcEcz_{E~p~xF7KVm^0(Gr_$K*C;J$oOrR5}}a~tOGsQnpHYF@YG
z<VCSNO@GAZr`BEF=Og)F_-N3prycQ9OQUtST}V#clzTMk*`vzq`F(*V{xurvPM7|F
z6<_t!z<#~MhwYVePSaL<snv2#mfX9yS>kVXW$VPcj`w1b2Ht-zcyn97st<`}QOan&
zx}YX3`+HX`Z^hNGT2puqzh>;Q=iDoF;>eK$smp{}&c9_ju=!VgSpDA?UxCdT@<R4^
zUrlGaVOf7M=-SEqA`&;QZa*3+nK5Di%9Mlp=Nxr(zCYjaF3JC?;^af})0!A|zJ6Bw
zC~(@v*elB(9rL$Z>|zy@mHWca(y~CYprg(}anA|4$*0Wi4b@L?E1g`s=+kkX<j0#{
z-3WWqG~Hv)RL4pmv6Ed>brwxZeVcsjc<gy&E<5L(pI7bQ38uIyHC3jw?YYge`_XHw
zUrZOb1pMyOV7mEt-Nm|PaV?3BS68Lqym-g4J3Q{jcD4L8@%vZ)_Vu<W*73D2coRSW
ze*Zx|mbX8?Z*2M!_-WVU1&I~&&cFWnd&a+04fADq7jY?Us<?H&@>9a8j}uavE`8jR
zFVuHldu7mzm93{Q1pK&F#pz#dJ2UU-fn)4PZB8h@e(S&Cxb$3!sWaaH(mZoA%{?aB
z^h!yO-b#h#$Cq83up>wO?yjZLmUA~w-?>(-dM>l^;<Vg#w!y!Pv_8)L_4)KF`N(aY
zb&uZP|GnGue=~dC674$%@l0F^Yt(`g{G6sL*G;YYwMduq0b8=tL564#$F~w&%-pTc
zynETYA%@e;eqz<n-(9lBToSXFIU4Bf(YyO7&@-y(kyv%*bDQw}?HMnsIhXLYUi|s~
z{;Fe=mP}<Eo*m;>v-P|1@<yD#{?FKzzkYtI2r2W>JaoS1qt4-<r_cYa=VN$0(Sx0N
zzRE`%y=W6br}N+KJ|9`UYwFdN0byP1dtC!|{qL-J>Yo*{@wLRx7=I^@n7Ydsem#47
z#%tM^yvcJu#vEUi>>R%^u}YKG-fY&cMJEkn>O{ScnO`vaT-fyZ#`~MSGRb)kAJ0Gb
zSi^37Y<=v6$I=PY`(n%|n{3V%-nCUyV7tHR^*x6>E?wTR#j-}*t$gZgzSfD??dA5=
zaQ3JdJwM^b-PGUq{XE+t&6MYw-)GF?Qk#}j-q|ekiYFoR<hE0Ocbt5#&+;<aupuih
zrO>@FP{w(O+Wb|vnHtAt_%_}Y&oMk^c!#<8oRGnqL*56c&B*6zPuwfpl-1cRSDTo1
zgq?f6#Z|cyt$n*T>}u<d|F$yb^!l*W<u`fdx3{Ke*fQq@yqW*!#M<UbuitjBkoue>
z^r`zpW`-5p%AYOqGw-*m&pYTQ_Vr8L{e%2Aw*#vFuxYkFU%N0V<)<C{gOndLUnooy
z+x?;U<USvXzePL+k!SyJ_FvQDD`fgYx%FiI$;nfl#O$xH)z@u4@vl@zcg{xfzaeMC
zx2Sv(4OrxHW$w$R-p=p5)p)~g7X@|cR9J;-GQE9Mc`8Q7VgJWa{lCh^N9Rf%&etj|
zo~Y_qG-=JJvXHP<Mmy~I&nZR-$EMve4Lp@u@Ia@i!=QYcXiQ-KfmP9K&vM!`+4E=C
z+D!KDRG5%d{z5-VZ}YLB$ESsNeD(MkCG=A4kcj6iE~%=QEk9<ia`+Lo-fFSiPXl{S
z=0lUuURYVP^-%Suko|sJ)UN5;D1Tle*eRtV#B}EKQtyXOY8_D`{9Ws|vTS&tJpY*Y
zrZ+9owd;0yvzz!BZ`V>3f0urR+4k?XX)i5$xkNYw6b^j3DSIbL@!`uYxi<g*TBq%u
zk@v_o|NG||v+jqmKUS{lOW0?u`+sFV?+@<Prq|6<lJBnb@!c$zEasQC(arkI!XK#(
zFFDp6@1E>)E}W@+(Qbithi6PJHU72#-pU#&f05Hi4)gZKx~*~jFuN@I%FS)!Cw9$f
zIav5<*HK-CWAz*who1ef2z&ng<*Oq(sZ$FzKAMWI3TD~2m1UvYqvimOBF#-7d(Q^E
z+Whirb=q?QkD1q>n{NC0DPfn7qn)_m?6Uin`vX|FNs7MQe=20=3!WDd9WJh}lb=tS
zV)Sa>kvGXw+zYRYZ<Ujr^59yu{ZfrZi=!K5qjfsU$_^L(()jw)yUX0dS^Up3<DUAP
zMNBd8e&$x45=waay!NW*!OW}Do6k&*UDf6*`t9|HhX(8K_(yWBiMDDHyO{esr)^_@
z44+PSk9XOsM1zd?H^1r!l{hX5v=Y1f*|%y#<+2-_JMSgwto(YJ^WotK^`E9mr0;to
z!Pxfl8Q<1^=UFo{PK6hrkzd=Z*t1f`<mKiiskMprs~+A-{1YXlUmWsbU#1){ciAP|
z8T<P=H9K!Bce?+Sy}88flGnc3S2AY4_Tg)teqz5~#GNIZ$|A43?HBZsytSqyaOsMd
zuCw;8win|$DSD%1@*ao9+Sa=TuD<7ZUh?7P*T9}V{?U8CwEuK{!0xoY;oI!%5-u-K
z-bfexJ5yYswyL%w@N7`pcc#kzH+{S0yIeE0{>);yc<$9j`HCx=V%NHKw<gI7Y;P~p
zN#&kobmnK&?q^dnF8nWMJ&>K(^)~iGOGMDx^D>Jq*~&XPT3q~Q?n-+(-?!CeL8Ty1
zPx|Vh-S(3WySkF(t&+apkoZ4$<sI+a&p1{bm)mc#)KB&M&4=%eoAZ@5uZewL@cdQ?
z({_)}P^oaeul(z0Jo#){lWT5z`H`j9oMYkJkKFwB_4wuRDKFla{wmd}xKO?_yx-jL
zu>C>D5c$sH_sZ|W9@XBNB>G+Opl|4&o^?9v&n@Qi*cD$so4B%Yjc)1`!;_PyE1k{D
zkvQv>YjtSp?wN_2jpq`5)-wlJuM28lFZDWT*~%HSmQ86~`Sr&2kj)F?f<9zh3#78I
zRGpGgopDe{+BK~6#?6POO#gL?e!S^9^UKwxB|XRZqM<-X=VfE#>srdo@@BrC*%cIc
z=T&fR*#g7llqA!MUB4EK+_)^)^756p;uFE({?E*3Rwl6Cp6j?oxK7{D+d;kdiuV%1
zRdLUkOzO|`o9g*c{-Jq6&;8#9!tC}f&qLNR3!azWQ5NoW#^^^?_Db$6f#;XaOO%Y<
zFu(ihRp(uawar@YZHsah5*ELi?XW(<z%6uf_W80^5g(Q3Ij$44`DdeXy;gn04e1#l
z`E{3^KmNgE<E)QspNE?J+(~Tgj}N%xbn!~&qUSofuA++$-E(vkyK-8z>}t1><B5z<
zTy7?}N+cU=^$sXpS@}}?>{*Z2#N_SWQLam_@@`a2o1VLj&GKLAme=;~+dF1Vxb;p?
zXG&$nc~)-*gMU&9@7120X;r7JTxr#KT{83QF|F3?v9~pYdZ({;6N##kT_u{a=0g@!
z*Saak0&*;`NDFeeZ<~6_ZQ=YmtJnWoX>@CLq(QFimn$2ZS+_VgNz7gPrK;zr@srg2
z#Pn*nTs^g?2i&JOeXZsWXYqeFcZ+@1ALsZr@?ANL1*h$7`Nw|ju=5*(|1+mQzPV4k
zK1?+`F|1i~-rTL(tnpc?U!JpTzWL3g^>NGZ-&>V7NOB$Z`7Ug~!{}xof8?cBu5a~~
z%HG_bdc``GnIE>yT7O#iPw8rxHL4{^)*2T-W_p^eo3Ny^d4|d_nTt8E-kj1|*LYNW
zTN&%mlxd>JP8^^4W3A$$0|JvR3|4OFa|^oRf9}-6_Me&d^X~gTy|Tqo{nidav4GfD
ztD4zvN8MU_{_1OghD(dQ_ryJW;<|dfoyNKzKj~BtHr>6mQ`U%XNekCKaj3wCP2`~a
zj?YuN?pfxZTK0bLva2&o5B-0e)6`m3e|X}7*4@)B9cqsJh_bz-tr)%O!yb+;ak8h5
z{nMR3+wW(=@s|@t3Nv3-uF;V=QpYjNd>!APa49qKd7Hex%CGstvGsj#`hu0KQ@^LA
z{z`29zxA&~1mnWrc1QK~N)mo=^)^epYuDMzW?p@;SDmF!=D_9~zoHYCKKNp9C(ZmZ
z)?D1E#bRC-&xKh1?1PtnB#HT-PjTa4v(wo&az1az)U>qpkE@hs&YZhwkxhC|mJpZs
ztH`f2n2x@5+5Y{sOzonlQ#BT9C+vF}FD%-9@u^g};ZkF+BFSI_dx1h8@x|pSQ))Nv
zbWl8T^}cD)N%w_$*EB1cH|$k7+FG?~x;c-=CI7EE8cEfj3!()Z41;;@ESBI+O442S
zRH0k#=981#(iUx7y~@dPTiVHdrUgMYGoR=>Z)q))QvX-A>9%0u_ol_4*Yj|PE_)@n
zCM7cX^uAn%Y+ZiuLdEX+vRQTEn~WAF+Asd|e!~4jQ<j&xFJ7f^+-<I1cBW*A!Teo6
zuY7nLWu*7<qWZHNZ5&Y*j?7zURfI+g>v){}`+e)VU+LXJ|4V*cdbB4(aOLa5Mz`ee
z$68pl3u<b$qSZvC3f|=Xd%DH(2JgIo6TZk6%<r4eZLILfHkhIP?`vn9(53gBu5T^h
zbEJ8ebafR&uMt;<_y&P@(F^OeF8%&=x%bi1df~nedx|$j8Jv0jIqJiuviG-*Pd1CZ
z;<*v?uTS^6VzWR=od0LZnUgZBWc7+2|Ll3a#3lZ>?cZOr?*A&EFPb?~Ft_-B{DRj}
zu4_Mk%Sq<wSi$r4S&f@g-kCG!&xiPxoS$r5bKWWa#jXw)mzT4b<;~oc{&RO{P~iDD
zmrq_VZPQyX(RoH<&$l+qzlOc4KaOWlHJo^49?!1$3tjK@cV~s_KG^eiLPFEGcpc5}
z|Gzi|3P1T0vXEo%DvrkNI**Y28=<ah9Q(RfItkX92r}KZ*N6#!ZFl`-v-k^zqkQo_
z(*paY+>4IQOmY8ceR^f`k?+^~i`t%E2(RtDe%JTZ#J%5yS3jCPXTtY?N4~6k_ViAs
zmhibp(sO^zUwr)M_pDjJ?2oO<jJ5i7`*X$KPa=WoiTi^lX|MBq9zJ{Oq4yV*D~>%c
zvR;!_dE;_%l=hn7**|*Er9|y5S@5>^yiLX2-EH4;r}_8D+8zFL|9|z&iOOeU^*p96
z*ri&KUUphD;iK;-UrmjfM~#IxrYw;@ud(HbjL!xCO1JD=g-Xq()1=Q{dF03{v*P>o
zb1FByFZTZXSEzsB#)EUuBNM~pW3!buNS^(EW2u7Ey#?p_)!+V|vhC2<y^c3ozwXtT
zUjFCA<1;7cABydmv(8KUE}M7qIyIjEx0<Ht<)>He`(!htS+L&!P;*|-eno}XJ6p~M
zN6gq-a?D0ovd2#5=;@wh<C%7b#j@Y41Fa_}1~>hD<+GD<+PC+w)LUMEUJ<gko+bC%
z;cmwntDG{@zUwVz=-_eYs0d#fd)mQEvu3W>6vd9a&!#o@&x>PnjQf}9wa0s%qk;CV
zKY~45j4oe#Ah4!YE|F!!`jhMSI{%i;s4RNiYg%#e$idxihxRVd`~8La+QxUYH>R-f
zkG=d!@A38%@1^}t@T=@(|HI>cC*{Krt&druO7CB)e!5nhDE1}Rw)WHAu6G;}FTXFo
z9_QB4v))Hz%HDikh2VUC3D-HId*^PgXr0BP{X6Q*C$H;E|Lkg#4*Xhu!|veHd1B{1
zS$}AMyI-Z#{YA1;VrA9rhU@aXo3wvcpK>-6*FV1TY5vjEIo>j&FK%x0ob%9W>!W2C
z6V_iV`mmF+^;J>$A88}^E8;iop9KECZNEI%N9)Sx__`^zpYMAYt!ens{e8bOmtBwg
zyL+`0WPbh1f4zQIvwHjcy3?9r-TJ01N{f|tbn_SM?JU~b=%nMWe8Nm^rR$VCkEUd9
zh&y@dW$gRkyGwPfx{e!MbKR;r;peHN(=zAIOAKD}`~Cg4W3#L?BN-Q5x0Rbvvc=|h
zjwHwPYqkH{A1jrFgj9NkDjl|(TFSm&=?>ri@Q|&lejG8s4vRAWvVXn4!|LzO4=1)e
z2Y$&5FN$|i-1WLV^Q+i`;~R_?uM4`&|NG{GUroEes_;}-f0ZoUxAE)z%=?1VPPV+!
z_J7o^&3CtG|K(SDKN8EYB<y!RB6nk3@~aMWiE|h4WOEhTx8I*)Zt>2z@J+)f?(KE%
zO}2dY*Y2&HAh$Q)|BIAu<K)BT6MtS=&F22fr_bfvl8A|u-fr@qq%%Qr(x09b=bv($
zT+Z}l6n{&q{&X(J_FIhqW_^d6iEhtc{<JPh%)6ZP`}@Zo(~svWE7&cq|NKcWxp?`P
z*~?SdZ?9$kmp;km#XH?!N3}NQzZ1H@*I4N;kLay?JDBb<iAkS5(YIIRr)2JpK#kSk
zjacOl{87@&D6F2R_0hYFqbAi>*q$@){ncNUNBI@<g7XTa7oNV%x*$t1UL)+G_0hUZ
zZ%;ht*dx6(QajWmt?A3uD^;<IvAc^dw^o_CPD;1R2wT;;-8$mHd3lAnhvm_E_ZDn#
zJ}&MTG<E;q0>v`n>y}5fXU8>d)K=(~D!nw>c>3pj;mXSQeG~HfxBYKDBfe;f#-X!M
z^6wrstg-2gV68GYIB7gNF|P1S@U&ld8BW~EpWb>w{zdG8?~C}@`0L(rwtT95zrRNM
z^|_zhr#V!L{=MHmN&5D^w|k8h?538Ve{4L{SiOB;ouz2R(dw6TF3A1qRxq6Vtg|?E
z5sQw^Eyt(p7tUAKQWNnxKhakq<x6^(o?+6zx!XU*N9^xgU)`VgzQTF6<iGEBt^2bt
zRIZ%ZWRlaoR4q-p{^;!a^R|Cgt}3@%via!ym3$lYKi>TscdTL`f52Xb^=_|csui1a
z{Z$S$mu)%Sqken#zU0W6wkPc@Pc%<F`{&%V8?jk=^XqF&D$Yq4<}`d*t@%?TaQ4A3
zYtL+AKC@Q%|KGVvB}F&qGEEIQTb?=V6@Q)UVG-$dzM0-u9R5#v{M&yN-<jjM>52a(
zr6+Cjrd=^7QyqEFJ>TMcu_(>Da8laW*0Vj!YS{v`Z`9oWGe>7R^IUy@8N>Fe?>J7+
zPqb!Z%`^{)3hwW^Rd(s6_jbD<iLASoTl25}o3MjNUjLzDt#1_%Q^^Lm-qMmKU*hxb
zKR70Cd{F+!&ohmZ1vmT-R{t%mct1}qUVWaVUAyj_n4iB|PxHAcyxS6UdzYE$MO*Xx
zJNoigU(fAUu*<A>e^T1KYsG)*2b-9$>s9`~_j`#^Mdhu%K~ppIK5UIz{GMaW<M8tj
zr`?I#;LQ6|;$Zvxxoqncu3nhS=s#Vd?BN<oe*c&GrKcynWjns_x?(YJ74M{+1-;Un
zOP2h+v)#M;>h%9*i>Fzqo$1+nMm^zL-M4by9Pj8&M^E2eeIUuRDbAF=N-Xv8!Rwbg
z{O4I8x2i4@;&~o2XP(oaWchAG>l+fUHGJYLn<ef9_Of58ymZxmuh9~*)ANIt8iW^3
z`zp(D^n?0^onrfUT5K^?e!Ry!zhcRY=_fv1S;RT}-B*`|M$Zko>#jeOi7Ygm#Oc!C
zpSP65`L05n_>zx`sqsIQggR#TJ9-Jsd;j~Xa8vrlJg3K|zlyuHRIfQ|yj<76{(8Oq
zmw!5zlD>)7T$A(bKR*^`Unrfh<VW|HAKHTb^Z0{#rqu7`)P7lE=yufZSCP?`4f78F
z6aKLOWpuR6$6b{!i({vKnX=+YZcpj^{3k(d&X+?^Y?#|Bu6+N|=S!W<kM{TmEfo6n
z!XPp3x5@o|nl3Yzs)?BBZ*be_;3(Ib%zZm9<M*Li#*<xN$ur)N2)w+z_SeNyKW6hc
zEqUvD9%ne4RDb)fQ}cW33;9F!i3el~_;#Fp=;+xVv{+=9%i7JhH$^vO|K6?o-1|I>
znqT?hq}{iej9%Xew5&Z6@o0KNWs^Qff%UtKwM|bNdHb$U^i|B=$1l83>0O!4hC7Ep
zu$YH*bngkudq1N^KR{mX(zfsWxV;~o@3zZyW6}Lz+h0|(&9|pS_=dmIvHn2c_vib5
z9FzZYMEvou2}gVUnSPx6JW=~Ycz!oa_UTXQRwusuv*p&Q3s<!Ep5{q3UTU-R+tQuV
zGuS`Ry1RQ;V)W(BwZ9u41ef)mGh#lr?&D99!?kz6<v!TNd_~Xm@A3<0yMzy4yZBct
zG}--&nz&m{k@ka2^>4LadcEU_Gyk&qLVVaaw~!o-S^n!%zD*0Q)!h?#$H=qLNGs5l
z;dV^LyE_rlTb{|+FO=!^%iPM-@FHyMPZ6iO(`vWX9LlyXs9%0Y(8@}4uGN%)|D4KC
z&mHM>b7#G_HZrnFd`)IgNByhCpOyIb)qF__VNUtpv-e=xgZ9LqWiL1CUYwqOGq*kc
zR$goKrtdDPw>QkY_UYafkJjCpg}?V_E@gXB6~bknSRd#Aw>r_T{Ce_Saj~wa+AMKv
zKjia2-92&I!j?CZ^MY>fYZv-IWqaJ*<2;<n|Ba8yuZn%}PqB~9e_vTPx7zIM_fPU)
z*ZgPBa)RGs&*V?$hbQgNzx~{Z`JV3IzsCa`?ku#k{&T%*X_U_Yl(WyyKUF%#^ZfqY
zn3w5}Jnw^&E#%MqFj7lTc3ZHW^L5=-J@%jCRb~^8pDq3nyK}=@;iBJvCj8g<{q5o=
z<~QO0?3JF_$8A5DpLBwMW3>9S&n|xbOy`dVtrDv@ihS(1pj*U1X7A2i?&&OMd)j^!
zd;ibyI9+q7^K|*0Esg$rjFcYsU3b5=Zq9avw&ng;)Gll{|9#hpMN(Jqi>1Zi3rQE&
zOmWP9c=5=ssF(He`De<TtgN2RTV&bUmN}Kp?Y8cc<rn^&{L(kS?EGt@@51vQmw&w9
z66gP}X#Uyh=QPjfm3CPa&Uo8r^X4`eZ~f<xDqiCYIp(fB8^m>YzEe6hW!f<tJMDRA
zo*C9Wvwrl_{a8|S>aFRD*9GQFuldVuH@7geX}4xVuic-STP}XFv$<s}F|VcA`>@{R
z=VuGNZIn88)vesuo)DYxj&1d(x6#F06diP1Z><Xta_-@n&~30LKWTz%<BKGl>eF5!
zcgrhJo~ysbCI29HuUzTA_3L8vgU?@H9vA-brFv?ivW_+9J@L=wuhyqtD?Cx0ba~@5
zf!%C-`I#nfoACbg<)c?jHtf$8u4bF6>ss&mEq;T9)R}#Z1<~8x-z5M3@^{e_Yaxk)
zolo~{EB&`J?$GYXYXtX*Kj{}Uo9;hpcHpu9uNYU$EKC37oB8$WUg?g-Z{Aj#7v|`8
zRrT-RBs=&0q`Zq0WdBz1pV-Ieefh`jV=3%KQOiH>WKwrey!HE3=;CcZF4uS2d=&eT
zyi@OHb<$lXN&mPB71?>G1g8pfsTY3yZp7yGul{Bx_x3w^`;_kR2gLu`@4544>U`N3
z@2&n>XId_j{&+IS;?17x6Ylrfoquv~WzO9?Sugjg(c*$yMMaNZ`d^Ixxz1Kz(?8Hp
zr?#$i&t1ig;-D&)gJK^NC(l1Naq8Qax9@wsSL~a5{JLVE$gf>fbhQKAuNltXQrp(c
zwX4#na_#wyO20U#MBcgk6F*yr=ly@UU`~YZ|7lC>>z4$);Cy{nSie3q!D+Y42Dg{-
zlPx{vUQYUyZdjRXaXQm9YF_88#ZS3E%d!NXerUNoal;1BGj@NsRNa_=cKb7jim6ZH
zC(e>NC!TlLTfzF!Mf-PgyOP!0C$HS08OG~AX{I0NkCvDhN2+#;o)q&CxKNj7l=!D#
zd18khcjJqWosMmPK5;keX($H2Jm0lN)_Fo{=nLoY4;NcL$(^fYJYl-bq0=>~bn6zG
z$rlP`{<1Ha8Nem$CF+=W|9HXIXALi6`$}S_8<(Gc!mW9aFXm_UOUV))^@L}|VKsAX
zpB~8B&OTY!vRFNj&FpX4L)$r$dB@F+X4xz*v{)qU_WJ4FO;_&-+~ZxJeqv#e=?jBS
zDL<+9*AW}vsAw+sk@@}Ee&X#jj4vw^<o%X>`W+cBpK*Rcctz&S*9oF463g4Cv@ia*
zXUE^7WydrU_~I`tN%T7JDBqRrx5(s{iT1I|=&3UDujd6H|FCD~cjZGf|88$RbBEob
zvLhw+!iKp`&(9@q3;k%YeUHPb`#ZHlPso3XG5mc*{_Ogn<tm3Z`|th!wt4?k`)${Z
zn3u<8{uEh!VBZhkhntvp>t6hGJh<4XqPYC^uBjP^esI=5;s2_1pJUIp(BjsqES7RP
zM_<^VsF7)LocN00EjF@j`A4qA%fI(5uC2eaFK~jC{AZW@`|l|(lgghjcdqz>;#5if
z^%AYevoB;_GMN{pR^)SYSG!kL`=%e(QQF@Z9z5P4bA6BHk9U*!<Ik%vm2#2)p?>`9
z8;2sEi4zo$2A8JKxnTWIfIY$JeQ#3pz3dC0Gp$?OXU?0a+wIN%KG6JdX3&-J*FWFL
z?OQ)_`P)-+B?bJNw<b>0ZFwWdyItRWTIaIq^KS6`d!FK<c4fh?);VWNZ_D*K-N~J_
zda-qRK!;^rYEtyBM%x<0s)i%poxOU0j_1FdaU<irNcgP2@P}HC8(!D6Y6b59@Uv88
z(dR{roL<lRD|vLAOTT;Q4@DENr#oFWlH$TY2C8^}f4Z=4ZC&KisY2`4J#^hY!=H1`
z_O2y?uD>S*8tC>vJrVdcBL2|PwKnswtISaOXU8>#Y4MQ{SAx6jCpWZyF#VY>KBIHT
z1TPz3@7KHc_E-FgNd&i7uD_pg%T@o-<E>8b+JvVplWsE-aXGT!`VoCDHC4`!!p}S8
zlI7c`>sEU<UoYPmbeW^x#-!rw+rBp}7yqWM61Pe-&G@p&Xhovg@!$pC7hNnh3%7fy
zb?vAS)w;+Lwrpmc&EDmnJTuHL-~BbEvHno_nl!uGGc3*SfwO`Z37cprwN|q)Hk{oS
z%j;Ikb;Rx6gs|2+xm`u_7tZbYntGp?bxqAh8}=ND(9Q?@{#IOz`q}nsCgVkY&Y!Xc
zAJS{YE^K1HwLa_r$GS_BHm2pXzJAeQFQ_w^`%yRbVo8dzv*VGMkrj!yLf1~85ls4i
zZpzH*e$$>m^`FdMS#o4$%+cp}Jc^%x5zJu|zP2T<Vdn?u1Ai+G@9F=}Q`XVuQsd9Q
za6;T);a`wmMa6yp3Hv-N6&aE%Cnc>BIbD=`f%$=5@f)7Th@@SMWG`sx%XZ!RW|}p(
zva#_m7uy{3RV%lvM7_;?%dw$Z`u;=FM+cfD*I#-w_urGM<#Q9WZR@@+y~A6t_=<n)
z{HqUob}tjYXkE{}t5Yj%>ZOG@swH{m{+b`T?@Df4a^A}Ac8qE6Y!dh1<*s`d5oPiu
zzCOZII{ECz6y;O8%pX5B*M}D8SDfIz`tJOMQxi?nOn$a%O|(3zHruD%N&VQ2h-Fi@
zENGkZf~_~t%`JFUz`DOzW=?GB(0cLpt4Pf9;O6-U{lmk~8)bd{vf*F)f|N`DCdMrM
z5yLOazhYA1Q48Z!UegjjJ;|3>dGd6Vd8XD$=CZ!X?P?As+D`vl4;+uZd){Cn-^1I6
zPmOn83Y<CRv)A?gPo!t$IDbqQ;(K*@Y0#c-zfi@5{xa8cL6bPGyIyZNV)t*BnNZL9
zz49WP+?TqK;w?tAf8TmMA=#0^Zsl=)tA~yUS!DK~TC8w_wQD2i11%l*)Z!PDg?8;}
z-c-d_SN`zl8k4ZYHwEL`qQ72P*RsB$Exob8fA{0NGtTu)|9NM_CT8FLJNur_o4>#<
z`kL^Cuk+W;&58Z8GIf1sU|RK>Fzd}L9&HaXie0v;v9$fgv^y>x)^qC@ygS&csbu%B
zrreri?e$H}cl4hAlzZ`RXRdTnPQ&-tZvS@cu6>|rXMW*(R^YKkN%32rFHyR{YZ<X@
z#vG}<l4DcM-k0`jdkQn#XA7^)Ij5WW*7M@RhL|_6Wj7@Ts~-IR{jj`q$SpO8Qti9;
zN;gbC?GpQRrs1(p_>UbduLS1F`aQd{yYk_k`B~~EH5)U37s>26bLz_GZHczRCa0$e
zCf&3ON{jvy(HB)S!(#ect*K|?euuQ%RZr68IQ)0}(IcWS1fKn>6}Y$V+nJutYc!_a
zZ?{&6S(pBMVk|4WH&^%)K86OlDt%@*iI8hwRxp0hpEQ56!epM-87&J$g`axe`oeUL
zx3aN$S<4E`f42@z`WM6V^G(MFXWLr|*IB|pIII4Rom{z9cH+|+QT*CJGBep0zq!|U
zcMkVS$@~a^msk7Fns^3n;hV~zd|=<Du2YY<i}h@tHiOr(&iRSt{>xT(x9^MG$h}Uy
z{(JC->e@|;9F8BeBM-cBpONf-yJ(HR?Rv*e9+%c{)Ozykjr#NS73Kf^=bdXlzNb*V
zeolS+lCbYa8xtL)53cMs=h9iyE>ip3KGLgb>Q85pjm=Mu5{rDR<T%y*ows{btUs&!
zZr{D74zItI+*$JU+Q<FNpQ!x3zV9@Hfts~}OW%Yi?a?pX)2_CB-fiaGf6DoO{<ljO
zA~8E+rkrXkmRY_1@a<_w>L$8NGhW|aomh18uH8S8sOayx7dJ8QUoZJnTyf+4XVT|3
zF>lpf`Lk;0GJoEA%0f%S4CR_q4b`6u#<82fND})IVc;x&HmKjRmPb)CaKoa+pKDwi
zGs~=BrgUBIeN;6ew^>P2Gkx;k_$jgspIM~0DE?n}iQ!aRP*TEW$$~Nqd7bZVjs{h#
zYfh}&V88Zxgd9)f)Twh<*`MsoUY}pEtF7+2PF0^4-|_ClDn=(P4*b$D?%;D|OW{cP
zdtnyu$>N)aUuHb~e&d>W%k_nigUX%xOKoi;u1zn!|1M<3-K8tm-~DA_qpujb?{JYt
z_p}eSYKDKm3v8L)Bq?~m{C8Tdj{3rPl2UP-)RyV1YEE?9Zmwo_$h3O*Y-^<}+#HYI
z?%ch5ePZ#;{Mvsxu@Bn}<86&z&MP^3Jm<;`|9Mw6<nG0tycl>Ph%a-ZA!}T?*}l?`
z^OrqxRd+WPpBB^*<n>gnw|}k;hu%$|VBO@>!?D*yzNo}}Xg^cD`jnW4!1|YGlA3?m
zPUBKow`=}CrR5^G_HMfG)ge%kD&@BS`Bw}1rqF8_PAE7Zw^KaMv}nQRn3qeQty{0~
zd|u=H!%ly$N$vi4=Ll;j`|mH>XT>%M?5?}^<%yk~&oP(edcM%sKB@iho>*?H*wtio
zY{R_vWzU|xNP3a#t+UvB^)pSAzgypHy6L-nPcG-rPL?=Un)6ob+V=I?64$ofySBBb
z(y%8vz9+fDe8=QWi_16XoO8P89q^g$x9YMp&x<2-u56oVA|leOR_^`zujjVcCYPNn
z5B_LplXvdj{j=l$zDcbICE4Z_L@2BLmB_2vs(NsB;N~5_|D|pd`}KA0*0?7T(pw(g
zkvlv~N?=PakKOz!2csu0b}~C=IH_sVLa$Z7;wG-MH*G8YuCrr_;r3N8%f+N^_dPoN
z*4{NaX!p_io5i}XyyywF==pj1()_?xzXG{b_6P25Kk@qOPu_2>oj-ob8Qqbuz5XY5
zMd5+R4OWf^9uyvU5WhLC>-YOz^SC9BMg38(d!F%o>i&4${CC?A#_H_3FFNI`LFFE~
zLYJ<bnC3%;bD}$(1k}3XU3xb9Gwf1f7g|)%A}lYEwBYTJy`0yd2OO!`!?f${W6=kb
zEKYxJ(oA7Vv58aKc>c$gRUbnYSI?O#ddxNH1lKuZrH!Irlitg!eN<Md?=<fS+E9KZ
z`P|}(K5EWN7H$a>g-@5S)=W`d!gXnu+a!-mMa4{0+tx%ldA#g2UiZ_tO5pFz%+^0y
zPaX(Q`?G6C{zcEr%Z-JFg#}zCJG(rRo-A2%>c~+ko%s9h`P))Yw0pjM&wl=Rq~WLk
zUTbe3t30}ySxf)u9hd2!#R_|c3`0LQ)mC2Ydfb{)vtdzUgoQ#_;fDJUrT6Yw^)JR-
zBdF%=nF~u30(2O^TbvPFq{H`UmG+-y!N00@9BrD~WD>!2<*|WR-U$aL27wg}3@uEe
zZ8{ro?)3P-%lJvE{i3N&4(@^qN@hxn?myExm{z3bk+{~EUnNr0IxtXOM_0jtAtr5|
zHlskkdHdZhY=%AgH+zz$Y>pZB$|Or1`z3MAaNe0@iSuoSJ^3$tk|pmPGn^ZH#x}if
zW0102t9-)8;y<@UrA#v1Gm<5a-IF+GdVfjsuB@wnrcJHW4gIGQ;<0>dc-Gpe`O|N|
zS6LgiJG5+l(Eih*-C{NS_E$!RW{XAL+UtG#^uNS^!p6oHb5fUClq4FIIm?Tgl}+3~
zX|wm~8m0L=<G;NwTz_?|*9n)HZ<@D0mRHU*yZ6xP*5&i;UN`f<ZQu0c`n8Xl_Ttz5
zx&Gc`V_L%}CnuwS_UDEj<v%pdS`W5=-)Qhr{pgVgNB`Xa{^$IE({Flm;(s6gt3TVR
z5M5DPvO_?vo87GQzxVmeae_*dU0OtCQoLO1g8naBI&;R18D^W}Qi3-4a5MPGr1AXd
zI&IooY|Pu3mby)JnsEEsNZ+n7&xI+RYhN<7C@^p_Feox~e-?9!W4Uu-na6fz^P=jX
z)A{v&r7Eo1n7ng#<>9qU9;k?iCY2m*IdkUZ>opGzw8PJE^lCeFJ$y9JFDYQfX3l5D
z`Q{o`va1d6$EoXiEq1rlIh%QOgYcX*m&eJrnsE^<ToaU|r-*J?XK}jtF~eM^rh6Md
zKdzbLu+TF!U;Ain@0*B&FCHh#xm$&$7oA$Qptic!ZdV&a(1v{+kMFbg$elB0zOeVg
zzL2TEmPOS2UuN2N__pYW85ZY6A0<>;{rOP%B2u>7VClP+FF$+<(4Qb+Agqz7t?1gg
zY(wOs4Nq@pX9nF4|L3X|b<j)nxa#WOb-MMx4pe!H{hHoU{GBbPTc^gQ;Y3)Nr$<?}
zcHSgS_O%MiIpq&c8D>5-=)30P=c_C;%i{MV+wV1tr*9YZ7P2m4WN1-f70J!%NllXY
zY3Uf9u}v_xPq>}y#OpGff`bYgk@MV-Gsa|WmR!=l=m>|9#FvYu{Vyk18BDwIu}Cv)
z`t%u+y~;EF3`^@G&p$bI>`GAm%!diPJ-Zn)&Lw{h^%DP>A7#PPoh!AKeVt^{hJ_jn
zUU^1FMx6~?sqM6+M`6X|xTzc~F09mxxasoL`e)vXkXhfS=6n-P$vPPtCwxsaLdSRB
z!5mh0$<5c479PJAcGyHi^WxTtH#08ZdOG=c$gHg{C!XKDp}%$6adpf7o9u7e%r^Zw
zDH3&KE7PNc6;m`eR<mXH1w=e~#dg<Z{fAZig%<5O7^ielL5D{u>7e=_QL$4YqT!E1
z!Z&8QguQ8bYJb#dJNp*pD4#%24RufTDIqJqJ-c?(Y02srTz&EGi$Z0-IvqP5rnFr{
zJ#^WW)qd4krb}5xCN-7h)Xb?$z8xJMEZ#KNCX$0ik%6nBMUjDn#i`vP^!lN0_IvBr
zsw{l-?b_oVmG5fOpS_-N>c)@V+tiZd*2&)Ro>LG$=LCa;^R~&lAxjQTny<AgEakLV
zfooxB!iLt&)Ad=mlWw2!=~?Bvd++}*H7loUpIUV5NT^@T*_+xgb+`Jch#hWtd~U|H
z?UFlwPITF-rt7FvcyPMKU!#RWYh*NTt`5yv$v0C;>(i;#Tc+tg$~u;H=7w8>pE}dw
z>9a&RHfRZNxzsiFrrcrO#QJH*QO_Q~2#gKT%2`vlY1ilL&sIOr>A7$2Xb{HscSg5p
z$J~@v@pTKnZOK^LwJJ1aW90YG;i}%L_h;(2|9&LK^JK##r<BKL-W{8@h1Z4Ns;f%h
zxY$)EH0ZRL=%zfYtc1%&pPP1_O;Osn#BX!Zs;r3b;tZR_?M~f}-Pu|c5V|sK%`TCZ
z;Y>a!{U^(9$vkI&{?DqQlfIJ{pI~qjXy|1S;BZj7&CK-Sq+GH=>y_o~Q=WLNI^re}
zxH#F%qu8*g#kaJ0X-Q8{OIH}{)q{?_x%@dt6<T>&`oD&@AL?prUH{PT(ednT^B;-R
z-~Ty!F6z|o>n~?Jr)Q_>Sd=DN&TO<Q<#}@F;4hsDAFgXG53IiHXiKJE^Y)&vcRfS*
z&dT|R|9Q6V-9KfL@6w7W#>5*e^CyZ;Roe6@`K!KO)eVQaB~h)Ck(v>UwyS@ue>=^a
z=gN_J$N#+ickJQwW!G=m@n!6v8_b>+z4hsxMg|$9_$g85`tul>Sj5Yw6iw|oQq3Ip
z?1lO_zE#B~VyCYruPo>9?akwT`~BCIM3(m#ubf?Z**Z8lFvum;)LukagMpoSs>|a(
ziwAr)0-5`av+GQD?LDLHSa%~XQ0q&~npbbX{`yxu>$&Cn=}&8F_N+f0YGkhUHZ;_i
zd8%8U<ke88))R+n^7zVLEsR=W*pnPAam*;@=B_>2XYJ}22j=GQz4z+ppQp#H=6Gyj
zH+bf@E@FoLamo3!Y+lzNw^{zSvZm>aGT-+F$3v7~7W<y5uWxR+^DzGhOJeNBb;rz-
zG%q}s3A!hs;Bd0rt#6{l!~GvRKCn%@HO0T!`sSbS{Pws1h#4%gR1^Q8#$DWV`k{1v
zdEbQPmxC{hEx!9>nV$6R@Q*7^_a05rQPENTbx^`SxUS&s4XY4~>UGk5je83gGPJ5m
zUU214iIus_zTxGKM%`=d*ItGHiceoR{jOJw?S+5W#meO#iHkjF7vWu}%Kk-DCYB}O
zo9o517pnQ@_8Z2;uZxZ4U|>-YIKjZdz@lEpfBW^f>VOs>^X|$L24^J(f%z-%?+^A3
z3<}w*zyG{l@BbC^zfXP|KI4UDOBr9Ra=*;CX!Gz`9@}ugjgda5Z|chLy}c<s?)HlB
zvazhW^PlNF{A_VX*h%NUq1;_r!N_?rzhn2^ujKjm^G~>>dD96A<vu>a;~T%WRIOY0
zH1mjx+o_&+MuH39L`~&hy#3()CH2SG{P^<onytR&`H)w6-RmnN^ZJ^u$WHF#4(`t@
z7W;hrhD?i9`{JX8ty#$(c7Cq)>XT2}R!_e8<m;KcSGK<vmA`*DwJ@#v>HN>nKfafn
z{`~pRJ<?yF?-bWzUy|<_QaJ5r^UIf?<GMa|vtQEqvg`k(>tFXReqFInq-VX;DgJcH
z`eRCVft6Lix_1b)FbQ-7FmgCBD5M1NuYdhcZhN%uY~5?Dla@{Ve`~)-pZr}OHP5$V
z2i90zUw@+i()1@`+rl;pFjzlL6Zm`de~lbJm&VuUj8Q*#^xuE^O~f}e>58bjWY&_b
z`g^|)T?<WYReG{w9aD23f8+6mor02#U42u8&rgvwInLNKBWsF=@c95ORjt+z+viGo
z&-m@XhJMm|bn3*TwYr{MM_#diTa*;D;`P!uYQEO`HoA`QEL*dmm`oHlX)T|)F@Bfw
z-x`~+segaXbUnp=EZuP0>1P6MS>dLSuP`!5Z@A;WH8k?iO+gq)DXzXAam=W-Cs}e!
zYJH-Z-o=T==h&J1*g57DFg$LK>#a`e`E7dZ>rLA&&Yizh9Ak2>zn&Dg<LzSS)fZlE
z^=<jLRDRZj!b2;sW&OLq^_Kpl<v%v0<mP{vd?ob6tlLRJdix%1zV><Yx*h*suiq7=
z6214{w~eOHUnl)7nYwY`Qj<4t!wya~W4GOK{dL%PhN~*pbsHs2+)kLhTy^P{$*H@3
z>qB<k_m6(PF57?oRsHX)ufDnB`>(ju&g5CLU+$iY%$nO9?iEWMGrw`nu=lB9PjU@U
zvc&Dp630v;dy=J0jv4m;GVDqAopH>#|HmEQr~kVTeE9!PfA!(#+rC6s?VWczJ3oH;
z<!}FLulfA?Tep8N|L47r-v8foZ+-syo3DPxz26|8VIR36rvAi}!>0;2#u+x-FWAxY
z?&af$A1b)okKbPW@s|AO_kZ4{>{I_=b8pL3=bKAjJf5vlZL%<z=RF654CC<*b_PC%
zmMgD!{W}`_xNPq9RJHob?4$QTmY)A*{FhZ=m6$_F+3A$lm6kopMjkVlw6NN-2JKU=
zQ}vprt1HAAvNHVU%~l5AIbj)zIYGsSONCvR1+?<aS{NNSLG;61iyNL7<+(TSjPThh
z6{#W;<e%BK_`rg&8IO+kD4#QWQ*hk(gvT^DQ-dWj@3|hh%cPz%{P*YgCf4^gf0!l~
z|L1z(B$Il=uz5<$1bs0+mmt0JBL$z7_Q*5$aWMBEI>2%3z^$p1{QSfOZcOP&@cz);
z{==*8d|GeMDs2l*%hIwPNsF|Tx+TRWG}olG=6%%Ce)z%qfBh33L$@vejvYI8T3%UR
z!TRghQx=c&>)8!F>h`IgR>|}U-QZKWp|kbNpQ=Z~ljSrgt7(axoUG<kl>#N(uN}EI
zW6gsZk4>-j33I%2V&q^^e8IuO#K=*_EOE?;Y09FDTp7!|Zmg8}Y2|4n`fA&@>4IK+
z=9yY-GCSIx;X3Eal^YI9YNsxra$r7hyH)aIf#>3wCCSq!sd`TB{L1xOen<ErlgEZf
zyDr`A`IRhrOlI4QiB+csdAYQ`y%)aF^qM&F&Akf!5EB!@7b_pGTzNBP$>J1YwIx2Y
zCK~;ka*B1;cNINGmBMwOoOBHjtNxxe>FaV^n^}e4Z+vtZ)e2__bM$mP*x+WO*^_MK
zI_sAD&6gEl_BgqEo;-Qdb6v(aUY+2A_UB1_X)Rst`k886GOH86Snl1+z4&396rafU
zl8g)eSB@F>G9*i!i8rpSo0oB-V#SOd>$lJTD*3zj@4x>%a@UVMI$PE(y`jF(xWuD7
zCE=K>>oK=G^5W9}Wd1*z=l(3a{>qgrCubUZwxkHVPu+9nvAnX9K?(ztK!cmd@7tz+
z^%rwzfB5mby6ejNkjI7<YXY}gl&$D=&g!@3*VvdCDO8ZzWqWmF45x9=BL>EriJD$c
zOeMF2Q~494bY_*jPXF5<@3+t~XvX`Z**9ipI`r6IlG(A?smJ`1LSev{uR<3lW`y3W
zyBg|t;BG|zeB*?xbIbbrrs@eQ$E*xH<nTED-(sigNsojV^GmGQ@o3>Mk7KUK<}Gvc
zyZ`o&-R7S$I=Z2U&hWNNu5yXJdnVIZ;+V<Io@5gj>G+V@qMBbO{{w~M$&(sEdlN&}
zL`rWzbVpWr)~pHJGz@#vP8yfo4_O(oeB-U=ZJv>eUqqj!GMy28E;-94a%NBd&YoOW
z&$5h(8am7E=YitcFKDWAP2l3ii4PYo?spCMo3}3N=KVJx7Ht#SY8spL=gpcETjNEz
z8{A~F_at=tYWz>N+q7!m@@p&paYZx<Ff=g8y09<_G&p%oZ4r^F=2LNrbZ2ccT-=i^
zIrYl5yHPLASo}J7`^{bIZ{*#};Kby@;8ePVD=?`?+j1uRyEE6Ue^k}$|K9yKVfLo*
zQtzVX-{E1}4^@hll@cdSn0RgKJa^8F$sfHwecXI1L1w!0i@@J|cU7F4Z5b9Sci(S)
z_?~;&<@@<VXKSf-t@<C?IrIPeul4^nCT#Lj|CcuVZ|sEdzoM7_M{uZQ*2b&OIUD8T
z`ggzfsi{vZ6}gI}R^$XGCnvd|n3!L9pv7J*_0*z+w{Q2f`1HiYG6&sXd}TuDyCW~I
zpPpTR@O#*TPO%l2*K>C$wzvn_<T+_5Z9i&t<fw~l*m2kJ>-*OpJaM8=Z)=QW;+2)+
z40mMiT-+{ur0#i7=-X|)VRbq``NBUw{%Ww1BU1fY)>a3H^lAlx1_s712L=lnzS8qg
zR~lVUQc}7fYxwR;z?b(&U+qe-nm*O)eD(?tRdyMHrN7^VvpA~ifB(MBEm<x*@~v6^
zrt5oO)W&bW{B^=g%dV7uW19(loJV`!oNMh#miYecd+pnKd6|m2_t#!>b$8`+nzZ`u
z9+URHp;5(&w{LCOmz;RJ=CJR-FMW!tuU~|2SZejL!nvy0uqXAlaPYI*OFQ>}%eS9*
ze#!UWdmrsN^GWC6JIjd^xy|1kGwa?Id?rO#{_(x4klVb+zi;{U?$hgiZwtcZWA3+h
zZtl}<ykc?IR#EUJcXZ^ouw!!X?H7b}TwLh>)~0gXy-kbVdbun&AIqP6LCYlL!o5oO
zooB3UtZFmw&YdUw_dMSh!M?ur!k;g{>EHap>1yL#84f?02M+?aX=xuh(LX6IV|#Vx
z&Qrnh;?pNBo?+KC=~h=)TU%4}tI6AQgKRAtm;^XjwryU#{$sR~lLG?-Q%(Ax`@*I9
zzH{Rvcdh-m@W7AS$5%YECM{(TUT;3de0i|s6b^>U*NZ$pUrU^(9dYJsuBy)DZ~Ojf
zzd4h1ii^+bM9cq!J;_E#W+l3;*|j_Ws!z)<|08QoJgN`N5l#y`?)YM%ae|J;m-@yX
z(^pT^`nyf)v7~cRyXNPg^Mwwat~+7;q^P$$v47^)BPk9H-I=u=)1^OIl*nYpPhOJ<
zX=4aWzBXFTt|t=x&GzQ^&6>;3JQtF&D0ukdHb494s|#l@4PCqb->bFdl2hxq&Yd4N
z{dZMp-nY7CvAGZ5GAX?Xbzza-y?xc1S;q_~%s6H^ZJS}w^eO3UZTW86=C9kh{n_U5
z|A%&coPFlmlK;2mzp64Y%w1$4vumxvPU*a?PS-cpT2pqMs+ty8bm0=mPm?bDmz&fQ
zjVgPV1s3g$vvTU5>aqB;<>bDz0jpO3FziV=f48&y%U5p8cU5uwyv_eU`y98g*l_0A
zH09dbXvdpXJLS&ad-jySeCeeZwZ2pL#+rz_U5K@D4t?^eZ=;&voYje6teG2~o8Qkk
zX1GtyaQfNLhCQuKSMF85{P|ye{_)2@EB@5!&+k9}_~*|X3hV3^oq7CJT;S83ocBN6
zU)ny`uH}E#)^c~3l%dx_7afK_h132tyFU6<cKXiSzphHpCxmltzpa0M=My<8g^yh(
z^$*{dEB|SI{=-JT>F0|-k9M4z6Ty6-aQ!_-*_3%(c0N|{v3&CpG;Vg*?^&sWK!XEQ
zh0-3*$CH=)tM0sW$H7zYM~bRz@9{5siZaFCo=yoXW-Vp2v$F3B7g-t9UT`vK2P1=J
zss)F~z5B(#Qnzm1x^%_kn*q!Htu^K^zIWKm*h_Dnez_>W^Uh04m;AjQc<zUg%Ek>p
zqE_3cxn0WHY<E{o-yugjJG#emUV-PaC|{4o?iv~wf3iB&E1dC~!|m<mn)Pc*eX;$+
zS<|P>a0dsydbRm_cqaQzT`kR^lzpn6pRHLN_SM=|7=Ngbs{elFEH4i)=b3|*VMdcO
zk4o7uSezH3C47qiX!wbku&-KPC*+N{-g;tnwau_6>$Ai$lg^%G!ziinr`o2!r)|2n
zdh_I<wLWWBt$N0}e*T5XIilPRnwA?k^8fogqkO*n#Jm+37m7SH?8y|CIA(U^m}&4#
zf5TlIT|a7zf1a4AxX*4@obbyp)2B_C_V@HD?Wt44R?oclTJ!nm%j#3M1ZJvc+_e5X
zUwWCM#Zdu<7DdiZ2M0wC2PGDEX@MWg>n9x&D>#vIN;7FEt5b%>qAuGHJ(E(79`-!>
z*IQFL^Q+F$^#{TuL%I_g7^IyIlo%MQek*1eO`5heqBUW0Ow=#Y0#!a+p@}O4Km3}S
ze(0U1czx0i`NchFe+mB3-l(_XgsOYUjg(1kBA*J6dh|+(lr8T+{4gu*wukr3#lF72
zTGJwRb>;fMYk!g1dUKlgx128v&zu(3(Xk3y8FWj2nO{e!cYUjr)j_wtT4r52N7FwY
z37__H)6!&a=M5Dm7S_hRYnUcI5?t)m>-bnPQz0bd;zIYNid$~(qMOp}l7D{aoe{G7
zueR2z#IoXtHadO5N>7BAo0%!*PBrXFn=!dYH0sv3eX=V#GexGX+`o9rgvGPBn9RR^
z`r-T8PWP_P{V#2GYexO0%;P4LzN&<7@F{)tp<Pueai*_X^e>Z)z_W6iYnFBON||U}
zmV3v*!o<kc(89pL<glPY+t)|^_+xFmwNblWr`1b}-m)_}vB{G0dcsu8wQIg@RTK5R
zB6?8DMB^>LoJ-Hbyap#hB?ga)LeBj{JJwHm87*3|wOi-owXCJnuSMyd_4e(SZd#ba
znl8CCpkt+AMu+^VZf_6msRbF!7C&=OX#Ta=I4iiq`d>cl%Fp%vN{i0b<=B;}B^1pU
zSD0rl+U;CE)vzboYsN9t$DdN?&wSs1{PBm>S!uJ<vom7e`IfWmT{$^rrrErdXPq<q
zGAmB%U2BbI|2!uq;Mt-xT35AopGui%JUTBKYP8g?S4yNXy=y5K*vt3$r7cc9TI|-#
zouc{a!>gyMsaJ1m&xp*}lopxVl_{z%YUF#(r#tjggxkKb13QXixK4d--u$N~E$2+a
z+wZ$?JLeVYhTRRf)7$p^aMo7+MHiC_j$M8sz2!sN>=SAdk(`BBmh4%hbt6W0=9z_O
zwGDfc-DYvLFeq}cFHjI*;9%b%uTlBT<#0lF`$<dFNuRzOO+VSD?W)tm8<rz<$>>n!
ztJWPVepy-0vkpsys0W0L88LuDxMP$0ykrTkpeY-V8SWHhwK?X~v9bN=6`4*U>GKaR
zb=Ys)F*R$g*sQRp^*fiAZ3^4u75bunRc>moPOfUy@^u-k$2l`B95z~<EmJZx6k0ml
zNR(C5)6!{SZ`hg@lj>UPy+p-%wf9ZaJ+=SOhRs6P7VWz4r~5FiaQd3#{pX&)?+!jb
zQ>1&U(hi~Ll0`9bk1N-w7EfO#(mipJ`HqWmtnHFjCZRWupB8g-ipg2}&Na8qBmD5M
z!VRAu8GZWn^zKX+jyLs>ReTaRZ0^=_$+)n%*Cs{9UbP@RX=BQ&le_J&E;+Ji)6x6=
z-@oZhFKsU84DPv|p=8!{?_P4a#4+QTnfnC)ozL0+E9&?|!w)@Edt!e2)U=hZo%k)|
zqP*~84krc%!KDlW3=EuFs~_dgx^?T;EoRM$d%9*z9(3vo30!}Zch=F1<@2_RiEjPL
zw2|rX`tRQ+C<uz>dOl?m|CA}57`WJ}rt(ouO{1%j@<q3KW*e`yX7AFO5wIw+kZb;V
zzSN+pr|<vDiFh|zxrB4^tmzlFo}6*aa1BqgWT)Yj{r~m}uey3G#yjfpQRj65r_OW>
z$E?`evhY~e6v2#sN$pn7Cr|mOZjR8L9QfzyTmPxo@BCfCyDTgHNDc4mtmX18b3I-K
z#V+;QCdQuhF{31&nYSmY>Av7%r#_dD{_1Mr0wJT~#<wM2m*4wr+Wb&cS6BDnmdFo{
zv2&)+6?4~kSj5^6DiHcWX-wjnaZs_xF^^lzet4g*QMD8Lbm{f{`Tb8KX8n6~O<=i+
ziQ*woiIn<jPxC?TCg!6dT}hx8lqq}flGd4dpT1>_`S^JMS^7U-bLaf7H}8+-M9M5-
zVBlckTr|f)k;94gj;`z0t-@tbjv4l9Bukvl^8NIqrd*@^bGBOAvx*wasV7e^YmCmC
zEHr`Rim;M`gY9y*K(RvMrERH3l{4L5dQS10vP#@CKYsFCmvx0g`fE+~ca{9^{-xiT
zTUZ~xOyit+XW|}~jW?EkYFe~tkCx>A_D>PppR5dOT{&s#R*4y3!<KqY)pD8Eeoo!`
zr?so~$2l{mO`G{+ZNQ)EMQb<q#?IVwiZ6ZI)*m7#-fw305@t8wzVJdSkJ$vHqeX7L
zT&C?^N4vQ|h21qdZ805Cb8&H4#i?VfL9A^N@^gD<p1*a%XHCS$q@&$p+B0XY-?Z+w
zk-yJ*?l8WLg2a@<IUyXq1$Xn7ng2LuxbEDs&Azre>mp+w?Y5RbBC_epo0<n3lT}+%
zbe4EM%4CfEF-fi{Zz<26ozMT;q}S^5np$ps7W2sHCjXb~dw!OzUVHYeZ_x8yUjj|)
zeO`gan;m|vyL$E0`Seqk1<!hQC;Z_2bm{Tos3_aa$k$=otm(T70vAsc$^Nz|v*M$_
z<u9>A0umP(#XipXY{b^{M9TQff}kl&1Ck|<O*HIDlX$GWMr(i5&h=N<zl_=c`q8@e
z4}T>rIAqkV9d$5`<MXV5g}Gsgp&#B&oBRE)_iBk#%eMbBUAxVAb+Y7@npZb`^Jg;a
z`rRt@eLi>napSG8tpEL2{}B3X_S6e!{-`}&tM>Hi|4YBMBP%OwzZX7A+*D=opzz4)
zJ{4U*3k!7_u5-@aU4lEi|BC(A*O!;ev2xkJssEV$tdPw`my1201-zcLE@nzykmu=_
zMv_gvle(1_&YAjSs*ry}*NKWrTO4aryEP3B>@8>REIRjRd26%qXBi7^i!)ls%8c*o
z#n1V3VM)w%(E~r{q@HB_dFIT?X=kkxJ$9yPH>`Oa_s>b&E?(*BnZ|&R6?+yj2(&23
zHE=LE35ZRQ$b6M^YNC{h#tnW+Z*QY*HM;6*uhUom6I-`;<;wj{@9T=c7@zaG^jXc@
zsMv6sdUqKE1N)P9M(KTQ(d@$O+7)-@@R+A+b!kkL>k9a|Bd`FJA>HPDob0a`R#Nsx
zS^x7b-!tC6cWR!8ORQ6C?cKjchhfGuvld;WV}?DarBY5kSu)N0^dFT!ocb^9g44ZE
z`P7I?KYsA8{P(lyjDnvc>ll_q&z@*ebtYNjjGJN4gBe%MxUT-T-*mGy+kUmi@9g=r
zH4k3iYLuzQyL^V_!n-#OBQC`*?b^oWS5V+Qd7{Bht$))#^h$};GaqB~<5+Q~&0w*z
zyi$mEc1GaQd4>}{y1MmBiJcaH`tV=8W@Kk&{gh2dyRNPdS^YcnQ_}pY3&h%4oy=D0
za<`hl2>D}|yd?Pc^nhf)bAjKN@&BmH)4g?lpO~oG7jCyp{}t`s^lCQ;T4eH++&Jgu
z$LA-SpR<u|=Wzx;#@hl;jK^gd5<cW7+OTcxNl`PLBx!NX@EV`xluf#4m(ITBb$0cv
znPGP`Gd8a{_^Ie$Hz$MStaEBCYaenH9&xA(jp%OLnwrnaa<VIWP4}c^osd1+Ys!)i
z2`yLgY87hcKj(E}&5DZ)C4wc6nI!gPzg*;#9G~q~UHSa*TB)@>av@JVCT!g|D~>NJ
zF5FR4TdUWr<kL2zNvq!Ih1~MG5fS00>9??}=uk;ap_aRl7`wEvj?&`KdNG_6>e~L-
z8!qKZmS$8*tqm;;3%YZ)Xw}cZ?%Ap*JLVtQXyMh};3AW@qERqNH#cpU=ZxUPM?{rO
zG6I*`q%3^IUAF1Z@tY^SJywS&C+q6|ik;OmRm^*B#$6|i$4vo`8!b;ns~oA%oIG`-
zZAoq9dQPEb8rR(ab4`tZnKYYwir&d^Z7$Q!E20lhXtO<aT4<ZK$8DErU}Wh3g{N;S
zaxge3b=?+JU=U^GU}0jY38+!yb~pYg#$$Bw5YyqF!)e;8VNqgnkvT!KX4;7<h0|WS
zPGI0ncYZ(9uwio63a%qNwnY|u_{i*SJ>fI0Q(5O^*r9~C+ir4sHcK2k<R+1N&3h`y
zA2ax-_4I1X_LlZ`e?9!ERk$lp<mS3t=Wlf@cTFs0RhQg!!F8Tb$Stky+rz3(Mj3uv
z{pZZZ+43J#6F<D=Ps#kjC2&->`QXCr7Q>!oEyJGF9iMA-Bh%EBMZH%}57zfvTa=al
z$;xe+ugb+O^O<Lfp1m>ksOr_ou!E83^fXEY+oWB~TK6n=Yn3{DH0fxoTd$O0M6$$5
z?H<Fa*IGS#rNrhXzLsB_7N!F(j@43^sp+oGUHbIiOtqzHt*c(@C~K$c#3)Yg|8&>m
zg;=e+nopLMmHnKXA791)npr<j+xqvK&lS4YZlphdo}S(>@?v?naoFCnpj>+qh6X1A
z0R;sC4krem-si?CYM%OOVV>ECA1=SNaN3SD&q`;ljmv*K=j?Oic`@%cN7kROOyB(T
z&A-K`7i9A4egCxX%=AeD9Rh-kjEDX$m1Sf0-R#iF=FBb9&N;<vn#O6zzt*|iqI^Ox
zOWgacd#vuUx2C3isNBDr9^Hf;n~eV?@f?*Y7G0W?7(L0{MKnjL&p0<?U7cUdpFLs6
zZXQ!y-?dfr+K!2d+{`_@Weg2I8mdn1?%dF0Ty(eK`iUEy{V%u$TxAMe=2XjTRN2p4
z@3Z{qQ8!!Zjm+0nB4fQgC7BOBjO{&|wmERkwXCnb53`Lun<t8|{Uurbvd?dQ`05xL
zEncT1QpK5>&NJ2+_N1OL&QD0%{`>8}f5*g*s&jG8s4A1G-<eps=Uz_Zdfl_;(-+5O
z_vNmytkHd{`fSJQ#7m$FX09zPiVOh)EeZk+RhJjW&6Arj_4e-=O>w!S1qys-d^cPw
znOhD?TzM9&$Wi%R?#PSU!f)rqu73Uc_3O*|_Y=gG!XLJ_uI*c~Hu10J+^NF;D~=gX
z`rZ3@+xE`2OI9vlzSi@#|7)2&EqA7tW<Ec2=GIx)c_RC6-gi0kX^U3vS5f=N@r;4j
zyN^bk`TeNAHFN6v)|qSU;{MitxE^1y!@K(Lqzj2s8zUa2N`2y*m~h90PwI<DpHoG=
zpP8Fa*~I56YbQ=zKhdw%|JfJqHAYH@M33v8toyih@?<~Fpko|=R9Fl4b2DxAn6p1<
z)|Qn^wcP_&`qiw?Jh!25gJ@S6zvkTeebaXQnK|=+>$l%a*^b49MJKO6zV_Y}l_XB*
z87uite6;9f^EhTWGxsd3EW2Q`xY|;ys*;zXr*y=4-!^ny=k}R-R{ee5(ZCDaw{O^J
ztK@oaX1LXJxvFP(S`;}PI7L_(m=ze91Q@JVXurMeIj<|~(50|YFBzlHvO8T8TPN!j
zE56>p>F~*ClcP?_O;mcaFkGsk!T-QQ0T%6z&wTqQ{(Vx#z9^$3w9s;|Q#DV<Nzb=!
zp8f08P4;oDe|_}hTjQ58)Bh>;=*iEYKfgaMZi2*1ZNcIn7ygFU`2VQ<y?7_fbN$@Q
zT_xW`o+kK5`@a3QGN{ux{nE1|2UFG^>o$;7P%w#;5PT#p`fNhlhWSs_4Es4ID|T9P
z9Okp+aO!$>bwdyL7smX62?iI^Ol(~5Y}jhBQBdylEs0w<n`CZ(e#qj!N=fgKj$Sj<
z>j~{|p2^AIUnQj4RCHj!<P<qRW7iG+yBB;3+48<BZsyAQTCCd4FE+3@Jl&OdJY{{!
z`bSf>cP!;kytzoA!AU^u0s|uhgWLt1sm~@&2wLh;bb<4vm5uwP7jki}O6Cir_%lz|
z&du8ySDU}+lski<QnA*;<4W8~F8e>%TzWQhwnxFwl!l^b?>BCHdVR(h%O&mn(>;=J
znCxx$I~iMT=jb(kou>1qqPYDx94`gk2%EO=(H7a}X}j*$oy?uL+i9xqhEKbW7xlmB
zRlk2r@BZn+Lqdi#vlSX=Z)o58{k!$99l4Xz=3U}WpW^qYdj7>rau1Hl2U^&=7gg3@
zjFms}%$Or;FWb`3zw7S3-@f@S|2vQU&E+!E2G-nq^CVoqN&Z+qONx)<`sTPquM!ax
z6%)gM*CGXX9XkB;^qbGWYgP06dcTS@Hy&8`r}ul`s^5Q~EKz=`JgGYCRonZDKEXZe
z!u#q&uB`Gu`>1J7)SaXw2UUt*{Yn&lpD|bd-*-<n+f#uiOBeR~s%vU%E#<$qXhEdp
zN|(qK_ckX+h89H*ExT2zA`6QOxi>6*wZM|;;$yXWK>^#BF8JByX2~eB$*E-dx<L0c
zzc+g@GKOypzg5}5e>K-`@2pghEt9@HN}HC_bw*2ch32!VVQbDl%l$6$(!W(@uj1Rc
zMu+{@CB>{>6ZFX5D}Q3dnWH+vOJ6Qidv^Z7#!J1QW|{wp{1w&MlbAG5V`B*C<^Rr`
z?GLAINYqnjxp?!=nFmi8#J!7R%3Xy+J(s^r-@keO$(y%!e(C<+T{t26a7mY~dgP>e
zsePB#_DY^i^b6aVayIcz+2_luMMXtLv&^gZ@9$W=ZTrHV3q6yceBw^d73j_tn6s>`
zGnDhEPKR%IbkB_$YJ9Qr(*)zT2`|%0eS3A@Q>#^liM_6g#jb@m?H}*I>sB_o(5-V~
zuCb`G`477{4}_AoZki<Mon+*-Y)yBy7dNZ2%7LzEM$6QxEoU~qi|^Jya<bRd*wxWK
zr>(4dn@xj)Knnu{SLCnt`R5wB%H`vK?Ku^*ZvL;1!{w1t!Doy5Kcu+rIXmTrl$GT#
z<+GP2Ngt~3m5gy{-0$EJ7oTjf{v*r#=In6Ml0OTbN+QzwT#S{vilWMj%5r($zN(0n
zT;dX$dgP_(q|3tWTeh?=*nDcvB`w#btaC1%%yFFNbB5d4X`@(aQ|ac@b5AdP%gd;t
zbMoB5iAz?Vy0qzP`1ju*oefQmta9ov^|zW;zPVQ~KjEBtXp^~TpjD*LmQ9<l>n+Su
zXPJ26dbDG#q~nDb0z1QJ1vv#>+v{5Q)3wX>k5uD=<m<~dj;Ob5KHJ>cY5L-N$D_IG
z>2eDl-xO|KdFsiNCr{I|ST0XGEgHFMbJ(PF-G6j@w>9ama(S8USy`F<$&#T_Y`a9E
zfre#i-38aG3w~XJnG!3$s}@b0w9Mq}k}FG29no4Y!mVtwqFZa>!wC%x4hp<W)-GML
zN?cuFH&^T<<_WrTW?kE-weHnrnXx2!vY>Z{y5|YcTb}dFG8GwAExIg@&tPc(f9teo
zYFWpYX_KALW-tHuKRjyhQq4fm(4Ae{hbHc3*?8h={nOST+CR9RTz4;7ID08y%+%nt
z@FP7*oK8NEwgqzCzwvaZw`8=b*Mw7##P|yv>;COo@SXXPPr!ymmD8$<N;7sYJ*k+u
z-f!N4)0-wuO8vER^XlJm`MEstDNTnbNPAm*D?NyD3i0E1-ulCD#yWk^?v{xs{1bd4
zwOKVU-<&tum*v2dKQ;SrMZOYr+AR^S(tYM_QET>Oi_f>7cr7~VG}F@glIhm1TaV5+
zmCjn|)M;(1c`3#}fM5Lc$1Pg({)Ghm-!o4&t>tUgzI~hHJ_{z#U`>zBdv)ZXMv?Wd
zg&P7j<ASehL~5<|l6DPSwCBu<h{nL1c8UxP4U9h>oCKO0IGh9+r1;q;ElF}HQaSms
zt7@hych;1JQ){M*Uh`y3<W%ls|J<MAr`^C1!otSMbS=#D+VUy3T9+6s3$*$)laCyf
zD(sa|+TA5SV|R8>sp5;(i&e6N)tuVWO$|;)WkziYo_M+Gd{uz9V?^PB8IN!0wuEc5
z`iM(UUbogZQ;C!DpkB*`&gVy`e3)x-tV1xqNn(cY78|)YU&5{y83~DY1l;=kY-y_S
zs(yE4P6J)>`x)C=GLN%F9^ZbAV`FGfib?AhpUT&s0`6Q7LS>SjIhz(dmJ)ESd{g_~
z&@j~O`Jy$Az5V^&POUClEA7wBEq^FGeO^MP?~*^<QC|#>{C=s;eYUtmq%@Sj^}Jrv
z0<HscpZhqR`}aF@ti99YBYMR5=)u>X0yca{Knv=)S8zBmC~`PBG(47C;3J?eaWzTl
zk<z|z1~ux>`uB*--4ycB-eMYie_kuM0VAXFulz5mM`D~@5|zJg){eH%iE;i_c<`3R
zcliU|(l>U8AGu?9KyJ%K7LIi+n#W8SX2_p2oZsU5!?k{rIe%&M@g)sUL}i=rn6|v(
z60v_Frg8sC{wGE0@VyyP`UQ{b7+Evs@E=!<trhfS)4#C(!-uKM*Qz-3Xg5BxIMyW0
z>UhlfMaRb@;dhVCTHNjP|5#<M%j7eE-?Xf3aBADVZ1&mE$cSfsR&pkRE)~n4$nn>G
zsjaDNj@=l@T*Fb^vt2@2*i$0<=Y)S;2J<_L?&>MN(o)vEaPp34*JJB|%`KjHtd8hx
zJ7(CEl`L`0+~t_zjQhTdzVeFl`HGeM<5pSO&Uo?SRmU-wk0+0*?p^rFbguZWssNb+
zc2-3ThQb5Soi9|sd8@G1DWZ9u)v4pzJJux2$}gJHYbj>h<EDJuBh1Z7pwU<Od56vC
z6G9=E4Ba<vGScr5S*ls9X4td!qEUTxQva;^vu2lWJ*lZ)DalpwQ>&`{n%DAc6Yo0u
z1WyxNx9+1!O7F!-Ci?>qw-g>be6r(&(j=Xo8--MNuJln~>Z2U2{6#A`ao^HOOJ4Fv
z9Sg{|7Wtp_<i@j*Gq$G|&hY;(t}xYd?tC_@TQ|54&GeVAz4xJ_v~qS%o7{_sGvAf`
zmoht<-+%4Fzo&DmeP*59S(z1ivg+i`6&4yZIxJW}q)(swxX^>K^{STZrK1^}I_zv+
zISOw)S((bSh4Y+Ie8q<!-9@U=6Zij}?4>^W@8Zzb5FMQjN~aFalq(XGWV&+V3EMI+
zkJ{`ar&*pJ5BBabOh}z97O>3XdaKiJrDDgoQj@;ypLuc0D=#mvOY{GGhH%ed<ZxnO
z`fO_%d2qsG!xYwsB9>1pCg17wzrpczF-x-YlkVL{B1gqzKP7#=dBVZ|SD-?s%D$!E
zYIBmed}_MKdt6VpN#e&JVZ&&_mIZS}*>zR=r1y(FJvHOTyQ29;qN`7Fe7upvGi}Dn
zpO;EauKdrNRFd%jT8Qp~InS>CH1y|~X&EyyL8Z$mD9Cl<<OyO76RjmUC2E$|_w=sN
zaS_ix=xS+L7qKodYtp$xhd*X!a&q+Oiu@}1kiZokF8WqAcHM=JRlmajtiSoar8s$W
zr(DjHSG$#0dWf4})>8MGqc^#!eSH-h8~-JP3kHjJ?C~)4nts84N4y!&N<pP&9UT#a
zgZ~uUq_nRYPPmwKN}jooKky6>-{Uu}6>XXy`G3wRPETBC7Zh1>ze{8P_e-00|Bd|f
zcQNOzZr^3S42&EL;uT+B40`-8b?w*GZ~FO5<}LlYb^F!ajQXglZ~jNEe;j@C&$Vq^
zrYhffbGhr0h{;pY^AcZw#}>Y+`}gSAm#@FIJ}9s-yeoS3KBIPyeffgrd%w+#J)869
zn||H*gNf{u``DXjOs`UU|9<)R>0BJ~a`LkM7p}Ne?|W8t^6}$1$?1zeWgJib@Zf=h
z%)$eScMl%!WfT;2Ik@ltrQe6Wa_T3z`2H#lx+k@?OzolhpSp0Kom0H$9{RYm_F{az
zOMg#d+^yc}^%HFR+>@7Hc~fR}&u)6~^2Z<boWI7lPWyhv>){@=kbf)FZkvfJ#UJjo
zkYVVsU?@Dd@cQq3Rl$2VZA#xx{4w?Jrgh(}_xiJ&?-l80ZHh~rX}mx8^8P35H>{rj
zVacZ-HH%C6J{~t`WMp2kQG|mh%IW1+@wJBj?-#sZs(08}SMc7R`TMp+3JYp24Ow$0
z>q684_pMtPnDc^HbxT~jb;>1Vg-zRj?+w1^?iX(jd-Zz$#H~4w8_jHgslWWc?Vr(4
zIr~Wy7X8%}satmR^u$eD{{HR$xZ@x1rdu)QCvKf@Jbhn(`kQr}lJ%+4oRcS=pIdt+
zONc+?nBnw2&lVjI46NU8@k*yGT6>!JtJ8sf^|oKmivE87>33~!%`VfMK~k@;S?<5f
zeqHyY>L-@fi3~;YiPulxeXH>6OtNI<hYR`hZb(SzXCFQyd_*WB-pAroRKzN&IREv=
z#uqF9+I9SyZ~5ollCLjYTT5HpH*Gt-Q-Ohzf!~3_Nr2->!}@amL%-httf);)jNa(b
z%f`;7oOH})d-?u(t}njKeYWAp)coh$1!JqkHYhMW{k2|ReTs&MM@nbaGts3Nt=0CI
zol>!h`K`_M=jW$i)0V#|zM`V>|GtlHVCJI5t^BL3FU6OZUkUh|nv==<bk@wxTTCC=
zw#TizcX7#%Kg+fGm%p68{oSS8O542s1G0L;H}a}#`82V4ruoE$d|_e8JR75YVSDnI
z3;_j)#=R@0>@Jx&G89hS&(VG3>1H;s59!~R|FE#1!^@&0<RQgT7<aRdb&i6cvG|Uf
zeRI^O?3&{Lm~m30aO~bidt`%q_I%QR@Be=3qii{MnXfNv8cg^%zKM`jaGfB}XRO{)
zYt6rH$G+?R8<xaWiXONnb8(lo@`Mv<9S({d3|kr)8W@}w81|&8u-+;8o7dNOuJ7FI
zU-C&2;_e@7(-SH@PuqStGuLQ$N$k<v*WbJ<led2+CcS`Be8!BHmRsu#<&KtU+0I@u
z`{tWFaS;+iPv&0Y{cNW^xkNmFZt~&nJN|eZ)VOMWKjB*aMElx&@%HtdFD4%7km+6E
zx8+&X*Sq~5cb;r5eI@_-t^NK#RekGvd^@!^7$kiSPQ9gnNK2zIWADfN&DE1V_!zYK
z^758S$CR%Bu=>jW`TVcH-dQm3!nM$)t0w>Zzvs%AD_`#Qf4Wfp|Nh&$H7)G2B9ru3
z?<Y@lXk=nqx2E=A=lAwaIrT?9q*(-PXFtDJQk-_ea*6xjl;^hds^_o&CM|vQ&(GG*
zbF#N8t7P04@xED8!2e;wyu*)d=e{YwFIN+vWf0T&<5?Sj#7l<;CxMpw+WN1OD}UV*
zV!06-Y9-Yg<?0ox8O61={?56Sh6_$iEDSCz*=KAQC*_|`xuD?iaGKk{yWj6uep0Ia
ze!oKYc|*s%S@}PtzjT;A`(nGR@cZpPtB-$eUlq;17pUa>N66u>v`X&eOMmyq2Uhal
ze>Zg_`(KZVW-oW&dG|frNZ{8c5#2|%)%}~_$En92l~WhLW%K8~OZbe+Ll$uv{RZ!H
zr*?d?`p=ln&v{bnud?UsCEun_y0rZL*Rr2A6N64J^H(p9EmrC^by>0H?)Riy(vy4F
z+~MfE5u#zZ<C$S^=?2r|QO1q?9)vg_{$h1Es&3VD_06Zm7Svi!oWv8qFkxYLduyP`
z-(PmqEv+rHo$Y`BQ0B5&TfBH7=ZXRb=01UffEi3LpYA`c!8Ci?&WVD8;xn3pH>^@s
z|EYgw)?VA)8!n{Wsot4vXlSVNZf4`t&JPPpFTQYzYkm3hg@2Xh-J3UjOUuOtZp=A)
z%W}5;d=J~^IDVUNmCY}#uItrK_q)91@7Z%d^-n(h66I`AEG)d>%WuEDrG*}{6FnxL
ziMjK9x9egL*+P$rXR6-y)h^l<W6gj5^Dogow->B$UU@A34gcxqU;j;a`u6JUAJdDu
zbvCOX9=16BSMcH2D%qs<$N&5`Qrn}~x?v7qx{NTp{`$|y_g{Qla;iRh<<&Xnk_A?W
zcJ!~Y&D$QmCivexzLhb5_~sk^ySub~-=7aU5&SbYvtM{@@R#lR_PZa%86R%_SnJ#K
zsFwZk?VsK3A`A^I3@!}~EKI@&!W`bssa6WysbBvoq&q&)_R@i{Ll+*G?Q2s0B0vA~
z3-zz7*}s1J)y>H;cizb($~J<t>X!e_d8=~yh;k;+!84hm++LZML0^;q&V8S%zWif!
zU~%rp#OblqGV(L?vRpsyId8VbtnkO5I(zwkjk<YDa}F_{x_vz2?bh>JMRl2if`WcO
z3KkYTw6Q&$yvO9(4qM}cr8DMFcscK3ZT_*&#-(CxUV$kb$~`ANEXq$<1#L2N>{bzx
z5s_S&u#&ZLa+6NOmtqy+FPsSv65I_Y=EuJ_yC3w==3T!{;Nl9AjQgw)zy64?)_*ir
ze@oDo{k-$nF~luMXxIIm9sA@#9Y<YTT&=G1?;raQI{x`rmnOPLRmn}E#^P27qs0Cj
zeSv=hrfd9)pMLBS6T{ir1_}ZV{!7HqcsM9d5a3`@R1|PJa^$f5*S2mk>)IdPItFtK
zXTOdOoL725-u3s2sry8Z8TOoVV1Im0@%g-Kh4Zg1T)5!jW+|={dDAT`Zx-HsrjYP(
ziLrgM&GPjkp{umL=ZB~MoND@U->aKUxz$q`{v{-|+pqDO9&t6SwdZsx>xxxtq_%vj
zT6Khnvv5~Rj^a;?8}|+RR!wmgzY*|+clvj!>2GEKtTB%D-I|%Bon!m;Yy0hM2Hj$_
zrIex$T5Sr8+T`<l>kZMa%e4``mqL6(Pc>crm-XtJ<gwhCvpJVithPBsSiL<gEVHsF
zS?b9#!|6+U!_}iUMx9KZf9ieBnk~DY=6reacX#c<!)O24mOh*9_{*U!Yf{DLZU6e`
zFS&kThSmG>nUecVt}CqJGM1DP-G43btjxL7iT`W1{#t+i`PcIQhHKldWxx0Jao~N(
z;lL!&!QsxEB4zPhZ!%x#h6d&&KSu7g$?thQH@<B$5;N_({&Pu5aa-~~>G;)dd<j40
zVs5=_d$;zUT>Ia<%f2jhYTXwRZ+U3TjR*Hq-WzXurfWIRd+ObN|KI<vm$(0)KX=|T
znMnb<ABAL}>czBvdtSY7>-$&V&lSyCuJ$Io?XMc2!%hB%nRmB6w_A5`m&=B`OXnXb
zx^=%ZdZ~5&(QQ}qW`^&+pBu*i=JE0VSz&+D<+C^6s>w;apT9lM`}@zTy?is@=$Sg{
zlsvBCTHKQtu{mVUQQyGOLuN<CbqeO}^Virb-L<bUpJ$<mcFp4LIdA?r%~vb7bw9Q+
za%NWT-%nrvTn$w{b|$pS>+l=Pl{06}SblrXSzlk@faKsS|GBR-C~`1(cxAon?pk8L
zq1@uZ@j0iiyD66~ldAH2^Hjg(OQ^_(QnvOC2Cn!Lm+POOe%adpGv>mo8B>4Mt&^DH
zz_H)6AnUhSh{L%vmo{d!Fzhz{dNJ;O(YBQi4_Cau+ZuXi`S%CKcbRU*|Gu*K#^<?z
ztJS=xbI(eaI9BB|RVem$-^q(fY2HUA*C^c&SzZ2bz3jz{7tdal`YayfI)9z=wbx&}
zLShv~A~woB-@hYh(IxJT8?!adJr4emR>}=D`)#}F(ocJ}=S9cz%B|1-`ycTp+xL`@
zeB&m$gLe0KXKq{jtJFSg?Owm5(au}F{=cmg4GsTeT5c_Wf7PE*Z^O{qonh(u+rzb|
zd}Z^UZ56xx^N;$d6aLxhm%TTpsXG<^mow`vGFX+tJS)tV<MAw$T)9vw%itgLC;pjI
zv48f-h>0~v?>#MhvZDU`ymxymo-T~v{kwI7#yMkNj*kCs-N)Yv-`KPG>1P@4HyV+4
zX&)+W<@%TRU;gasfBqZS<Mbub%twlag~al1_BPmj{{4Jjvh=!SiDR!Mju|`O+;gS=
z{`8a2pFa6iIX^xtzGi2B^WFHl-!vybfA+IG|C5s1qit%k;%8mv9dfw;jq6K>as7q#
z_*%J+pzTYJMorDJaZfB*QE)F<;+W~eo@6PGV}{eG%({O3Ur*1MW54#rolpN{|5X0d
zoNpU~ey?=r<@Xhd^4!_EvpJ5*=ef9~qW|-s>e}q<U;q7ncSD7`xVpUk(_PE2{`*#C
zBV2N3s<pSY?6&{;KmKI@kp9iv+y7{4jm6~{^%F(HlUNi47&;89B1JMk^55w7Qt{K6
z(01gp(#?)HkC$*w|E~PoCu{BXJwclq3_f@r3Se+Jr@rBhOa0R?syC;6$?#^Wd_Ggg
zuuX3BgU~JZo|YCrEj2B*yY3a;w|`wLHer(G#5RwlC&F&xkqWCf&GY#^Gxbi$#^&z|
zFYD42m~>dARC>j$f3dJSR@eTKe$OMg@eP-{?YWF=%1Sla=W`0>)sL^te|$c9lFZ#%
zzVa70<-HZ2`_@`?!kNC-hu2DFS6i}Q+gEy2TU{fP@ycVPB9?0Z+C`frCr=B{yPCW=
z`sH8q=?`xiuFLtj?)mGU1MkAq%M+Kp`u+EPl+Jp|*7CkDQxg9rm0y~+@cN>f=unL<
zmo8tP+F4cpD=@xMeyWM6W!dY04_^v!FgU2azWZs#@)b@s@z?);y*Gbn&jW7O-j4hE
zYs)j3jrY%a%E<65TxFJBV30W@<A?gYUWJAnw|@LTZLA`~BkDNIKhf<|8dvcwxy*ZA
z7bY%zvO4jv#cgv(`QYm-R<G|D=WlDN{nOp*W9wORVd8|1X~~&O#3uZkQ;^<g-|D4w
z{;mA??Z<yF`82OP*z(^zqtYY7$%Vc)I*ne12WC9U+4e~L-o>R057%kD^mq05e*bQk
z)3OKI&+HE!u9+cL5K><5(Yh?;m|?G9vc$1|!=B`-o@B{s2B$S`=gnMszSLFK=gE^N
zN1hz@Ida$3*IT_Pf0ubcSaEX7fwQN5eWyp~wrB0Cz83fTkFcm`+NPgAlfK1kMh5j5
z+n+h_e$I)x%=<~_x7Wwg9j_`fv@pd^VB~O6FkIsCXYzv(Z&&%~PfGvJl=~iH@T#7%
zf%)RGgc9-N&%=+t*d^Dgey51z*yAbQ5;g3-$M!mYFf`dxa?0C$cKFW|O6Cuqm-mOg
zTcYAQH!+IcxM$xhrLF~TiyyrRV~?8NzI=IzzC8QEn3U7q3RZIptmafpujpR9dS+B%
zr{Jc=)~i~ozJ6nMdiUym+_UQKzBeB}yzymky5HuFtG?V9Ox=~uy23PFGCxG`cY1DN
zyr9LdDP>nnQ_Cw?N&A^S2r}B)v2JHaZ^F(GZ&|(@@BO=cW7+q8U8>I$cd=$3+hy0i
zJ@e*`b??3|nppfxyV0+3(^n^@7fSoXY`eD4I^I6()RFk@<qu<$XNnmtvA8j-QEJty
zqRrn{99^_&-nw=Dd)n6D57U0L?_(-ganJ5YQt|GO&qQ@aHG1rn(QbTgx$MTX?H?GJ
z7#Z(A`SJO@;iQen<vpX{ZQs89$GVeVUdF21zg_8OoWkv)VR5GJ;gre$G_^Asx>lun
z1+cE0)wEe(=Fyi&^IjcU$Zce>RCJ~Dq#3qJd-lu-RI{v;(~>heBUbk+;<02@W>kJr
zi&3}bqBA%1bRFhAy?*8W(w_(Zw`xT?ic2)ca2NMvJlu71ZiU3cqW-?$%O|gJu856Y
zrggvI>BR=oTNN+)*{cM3FM0WSY}ypGDK#N3Soz1qiO%|}2D;PF7xeonxOq7m=Ivnc
z(eQpJy?w{C#8$CNjg$PPCS5X2uijr-UDB@C!?5n};^bGF4+S2mTwbQN;%E_TLIHom
zW@9z6h{qFo3N{Bh{I@UJa3V%LJcVn!NYRTUU2p2|xtU#a<Zfx1erMIY(ArbGCGG8v
z^=9(2%w^d2?B65H{X6<^y5HHJ!=Wjn8l?Y(fk{AS0wV_tgVj5QVw*{t+A}Rx%s*ed
zESmkH=*Cs9M>BR^`5*Ek|Ltj&Q=0;liZ+-|VB8@fFts&3eX>}Jx{{<~+T{Al$&*(*
z-P^TWEp*r7>25yZNhwMDjXE!E@Ni$Qo4#K(BJ|eDox7{EQ%pktH|8D)@7TUv`h=#%
zu~p)XmPcZa_Bvm!O$?E*j`r3F+0?F=*sT$|(h~$+%ulSCap}~GaD_0j4{OBADjidO
ziWK|AvQF(fY4XV<COo_@HTsv#E+eJg5wl;-Z&qtPw9{o_T7sD4)Ur~aFHbH_F}wOj
z>GrgBf*0L8R{1U1o~+<!?y4T!<`cQE^YxTdE1m0OW7lZikvzR<L3sV9pbeKcbp#x8
z@|qg5Wy{K(VDE0zhY`=4|CAZ+blI3%?{!ow{>7hZAM!NBt)KD<33n<L9#kj_F=F1O
z$iU&i!SRzrSn);3mlWQo(GJISHmQj*Otc6((bJMsQt7|4-zVf!k+Z}eA*DTA19TTK
zs(kwykaO^XWcJ0vPeuhrX|XG(ScKN9hKMhfTI(5dHvGEJwd;x-X3UjdCbXpBdfS^-
zEF1y~8K)Mz|8ncF|9Cyl@O1F>6vYaoOLq;9o%pSv>avqdweQ<y7s<xP#s(#yLx+D#
zyr|fJF=EE9MMnz*Z{!<2+HvmEPR>xT;J;Jtbx*6;ik&qo%=j(4YM0`5`Jnog54|F~
z29F=Nw||`SQ{u*?OD8-km1pU5|MY5UZs-#{>dx_Cx6?~O*;_rm_9rG5zY2`$6h5_6
zC?Z05-^Rd;|7RTjwKDv%*RIh29~1Yx%T@F5*m6<bJ!n~f|8m>5Wy>^HubH^$K$!o=
z%*TBx6CPIo{rCCu=iq+Rq=Tz|hjojYuTwhrO6kag*yJaPr*3pA|6t@`@KoVrXGmmt
zz{k4uXXw=q&#T|qZ+X30@aAXRt%dIAbwd09=*Rw<9j(4jC&uzpkWf)kk^3e_jwACM
zzZ-pg!WMEg(U^N<x!!Td<I5gS|F%_4OH^CREwV6S&&iHW?j3i3F4DQ5xOSCq^tVm6
zQ<a_ux)#4spIUfjQy1TwmNma(zVK&z?GFgOw0_3w3+Zx|*<SOuym}ciGqW^({c-nj
z(b6W12(F`2p@GjOCk9;mtC#xvY5#`lAx8hJ6Mz1&b2`|&MP@^!OxphkeR)^hl6Ku#
zvFxo5*ETVAl}jt7EnkMFMsL}&K<h?;%(m@0#-FEMQ9XDuFj_o!+uO5Ki=W#5{aG)&
zV)@0tb@L3LZJqch>iV){Om$fed$XqmKb`aMPy3;r%R&No`|$OJl_k|x{5krlAbtP-
zi|1?hU2zTnzU{*g-^x9cQletg*<YKd#zj=0(q=mAd!zSh(PzEDJ!g8KX)v@XGUcVe
zHvekC$iWf7Akf0F*OcX;ANS4Evm*;7ceSKmQ)YS8x-#I*nKN-*lV7b(47+qH)hp=h
zzSeXm$(f#B3{_&1!P#M^T$lD%Ec$i&^t8{$UqmcC&(Dl{9_7ma{p~mXyu!s_LSlFB
zsrk6<j`iY&v)|vDvd`cB_wJ%vesR-(e{WPh3@n%ZrQFE(@a)+w&h~|$-it2w|JU@!
z`OvgpvGsqS6n(0n6=UODr@CJC?3<`Baf}!0`V#s6Jb%@f^2hFIzOGC54c|H0cNzYz
zNdNB|weaml(TwX4AN|v4`T2R_$Lq|FciwerulaE|!*5rcTDoxOLc14_gqwD5o+l`p
zq^zg(be`n@y+?We9{KX;`a_lL`98L3VSKA=H-34R(ynv-kw)rs``wy%+Z6Mob0*|%
zX0G41eeRsecK@ykJ54Q0aee$w_tJ3#-@nhej+*D@TU%BAx^rRi@AJ{MKEJw`>DX82
zt&9>+I}%>^*!t?Mf1;H_{++J-3L?_^66egCt<ClaG-J<ipTFn*Rc@vy5=+E?zA#Q*
zp^$$1!I6sUySI$L*T0w8p}a^iWt&Wg=bU7(j=Snd545mNVNrRob<Mn4GiJ?oUKbNL
zTVs6!YgNqCJ2G2tX|8NBZ<`|Y*!<hV*;=lrG&Nt`IAVM9bYjL{Pmb+sg1_c$WNx;|
zJM!ra$Htr7ozIQVD$LmYv&Mg~s{0p>wW~yT2T%JjUo`9K(tR51C(C;lUDRAWRa>~C
zo^|T{B}z3jHZQlF>z*63YN<y|*!7ey?Uk)%EKhHGybk%csUt>IYm)xwhsJ85H?`||
zo-!>l`}Fr<_L(yr&+=Gu{s|r3o#6aU!&^<cJjlG1-NY+mRm7&Ve5|>pfv#mI!YkGF
zR=&8QIW?r6X>C-|zpH07w7DDVj!vBYN2#hY&8OI8$*f9OE`c4o>uyd759WGnd3sf%
z$jU&0)<a_F9JWoG^r~yO`lnxhXIDEPS7cylyu0;72ZI1t0|$$uv~kxdJ<-K=oX$Pj
ziqh*V^42{!T3NsOLzI?tU36?vU~G`i0@o)jLE5K6Rt7DPYx^rNb0h4l*$S^i6%pLh
zT8+X_j!tuU=6|%R{!q({=xf)ryR^41+O1Obw(!z)nMkKC5n4f!D?BTT))}l?IA``$
zXV+h8BIWBZ^!{>jjajs4$4sV1amm2wDVCE?I;~cnv?9Mdd+Srt8-cMQMQ@ubO1niB
zoq2gSta_#<CjM95NN0Ly;@Q^BV;MpA{xfIi@#u81?f?2KC}wW?ud=#RX<0{Q*_CP{
zDvS>4WV5_$S>o;`^Lv$jSJU0b$YYs*x>j<FwnlB*JXP!BRIc@BHq6&4*|JsPvY&7F
zYAYYziLJ*@Bs_hy(EHjYEpf@t%%&ID3$#vYe^n|u)bcH;{XvwY#~f}xuboLs+vfPK
zpAxK{Q0)@Pz|g>Lzgdh`@dX1z3xnOc<6nYHmM?dk6#vmE;@|tjJMWk&9^5WymYBCr
za;jPI(%DCRyR!p1O%08596MZJf4SJ&*yH!$knV?P{)d%#6_z}9>F8O}u~IOqUvj4B
zOwpoG?=M|@Ddri;CmHj8%hD~I7W}y)Z+!5bamo9NS*P-6Nj_b<uiqf=`q72^SS}vh
z>(g;>;W4SOG`YBhUq5pamL`PhFy<A`5aZa8(rWuGEa+h1XKj0?sZAjfjGo65HVP&L
znF)R~zp0x3@5%nAJ=}aTF%jBgIuT(hA@z%T`?p4xSw}POYtvUu+8Ft2)8(W+hXr&U
z)e6^3ZWKH9aq8r$XI6w%?X(Px<8+=buI(1QGJ3hTk@lgYo~c1a_w<F7EBL?0%qn@l
zS2IZaj8W4S(d9A%>T=h)41x?+cGtbJ+!8(GnAx8O1_swHEQ%ZqtPaT%r!%-?`Ig$u
znUwY9>m{>qZ|YvEzRdcWn6xXLyF<<5fluLE;bdi5g&FJgR2T{!+{HW|pSL-k#H#7F
zT`W^0Qd4~UpIMh8Ii)KiUc7kvB2pqTWtT^HNA#Su84nW@bGB<fGdgy^@@v?~EbZ*z
zaL>OflT~+Z`uj$=S2tNlBl;=p9#uXI9Y5Qx2HVXO4<7%NqCS6WRI(Af#HH3K)06!#
zx+DtM)UKA;vHp%^aEL;$n99i~Pgc)d9B8_A)26FZyY$lXPM!XqX>4q??)axwCo7xU
z!Wtu2EYWCPEU9JT5K?$BuJC)yqeqXsJ2an}>8k0Rkkn`8+LB$dqfo@)Fz>5Q>8cec
z4n10^z1Zt`V%{>pL&t1|v$Ks&E7!b!r=h1V=I49XXXUEk)Pg++U6yrm+}#;>V@q-<
zpNL9dUY>LLM8lrk8;l$b&KDRtSX45f&#UH1T(a?6Sk_Wio$0Bnvf;(r)5^?}(pC#*
zu|J-9^TL*s62}auA5c1P!N70xZE>&0R(Y9OK|u#id-Il7Iz3r4rDCJWjcyJtOJyao
zPGN~->D4AvvexgNx_4Hqz)4ZhRbmc4#mp=nD>qsBvfVmXcADwL#*d%pu7CJubNKq#
ztaHm7r!RhaJwEtEc4~aEjp3jDUyuH6zyIi4V4aM<u+@3~^RZhlE#zJQJN14M1H<8G
zZztw{oF(&!g~i3i#l^MfQtsVk{W)(es?!tXB{$x4HT%=9yfgN?;zzYvYB}vz8<RdS
zo>Ey_srl~T#dMGJUZDq;TFS@GeidGKdck8M-IqV-9sc+%=Ge5!8@+7X{jTcGKI$yY
zJgsFuyH5L=sQAk6*I6s_-?20IF)7FtSY~W^^de9xaN&Q0^>cPFe`0AfY3Z|h+Odh&
z*;ihdSv_C7cI{fj3Df%)398j^dFkD7@>{00Wu8{7+R{6fB7JX`rY)-!vE{Gl)485w
zocnrC4ntj^(d!KzUtOEtWN-JmJ4JTh`JK1_TYi~yG5yp1hq)!C-=_B}DoRKE`qjGd
zxQd>lMsf0@rK#TnSBtF7bJYuE-mv!9^B3ITe_iZ*sq`#4u-*2=>G_L(tzGx{)u))_
zLHDPd=y314<+AAMH{q4G_g-fwf3<e&H>|mo<=R=<d@C^iSLeO{_xGk>s^faEU|%_Z
z!}jfOC$O(r!xOhqOnRdAN$H}^r>g5`sd*V3o82ca`fSCc{w43-roU4QR8W?cR|;8{
z{d-E(R;QAdSh3xGH8In-giZVFq*<9%QsZ7#|E^m0kMhyXJRMj40KW<CZ`^<Jzx=Yx
zd%>&R3j!PrYz{2SjmKR3k7T)Raen0UC^<GKW5UhCr?<})?CP|U<7W>%a`;fcoK4H0
znY_oPE-SwN(R^86RZ-SfQS!yDDrIHmWcBMaR7LvL-+Fx2eP#D?Y4(A=GYaG!tuIO{
zv){1i?&7!@75$rIX}9na6}=R@<qtlDf4$Ic=(TOPSYpik&U+R4>gS$(y}EVJ+E+(@
zXWPH3-&OLze$On|H=BO_w)+2;XaDh!1`%B@pNb}R3H(i0cUSLSJ4LfjtYqG8ZWfK)
z*Ewou6U44=SG}FQZpZzH?Gj7;X8iZoiMCJey{V%$Eq=$k^`|=4*S9<^TCF=xZr}dS
zU(-JREtEcU=ED6M5l_2xf)15z*|KHJ#^2`Nj$0!2%zyv6#4uHxS^R`++Zn%zRhK7i
z_`PIPW~H%IqQ{Q)re0I3)HOG47Pi+C3=CSZ(&yT%z^o?WxswmB2&k&o42+8I5cK{i
zy#IqhgOke%*NCGtS`-BqFflMNt1*<V=F`0>q<*LU*7iR+YHz>&u4sOhw0iwtACKBD
zj{-HL&of^C`kuU3OwXR{K-;-ve*SOf`DQAGbhnF6sCZ`BQ=7T;@eR$Vsj2_Eeonns
zv?^!Hdd>XVa!cm!FMi5Ub8g~hR;I)Q|5=_0ZYaCZdNiTYV~N)i*9<;K&GfW|fj@Nk
zSPSgkG=tVfN`$Mhyg0Gq#EH+>!>)bUF<tqs_L7uby_CFf+g6HCpFVT?8vZ#O=l<|M
zeXdSaW78Jze+w3fJ(af%x>foAYH6Ilkm!onT9(Bge<b#YrF9;+Oit~wI2mXey6VcR
z%|CA3(^;7+xz@F`)}!nD8^4fc;or5Ye*M!w6n*~K3-gtgji)wjT%)m3M|he}#Ky$m
z+9#sA@5wyNj*g58TpH%xCzv~N)|Er27Wi$Q60kDxd-q|bMax4L<efev<|_Kz(R$vj
z;LjShH<wtmC^EEcc3@EiA0*ZnJMnf`Vp^bDr;WChw5eG0$~|XHr%C0$N`H{z^D2O8
z(Ry8jMm?@@p$1QpdYhRho0o5WH9<R0V@;v-tItz116eYUC2tW-3{ZDJcKrAX+sfH%
zwj4cdcj?i!wI=h`E{bfK)fkgAQS88Ei!+?<Yd(Cvp>s)lNu%qr9exuwPTJ65`_EV4
zrurt)JrO2Hr!G<Pv-_^3bW%cTXS?uz(U1FkUViR)x6yUOj$YBi`zPby|M=Od*%+I<
zS?t1Qi?f+6QA=E2yIlP?&He7?$fe8fpZv6`Y+6}Z{A!h8u76BBZme-#>Cx@uFjrV<
zu2=DulPl$9ALc&)lfLZN!zmm5EM0nqQay`Grp%e}G$r`Bn5WZfZDswP5gd1SJC#nE
zy#1@?TRz1k?KnwBS?e2zW9Clr`O6S;a!I@CtORbRDXVYDo?zr)5In)b!VD?`yhVfp
z4KKQHRP~yw+xabU)tf6T168}d%lxKoi%fX*+HR}J{`2go|1R*1o}G|zVMgIAm%JxD
zMTa^WLhG*i_$W@>bEcb5u&`bxUHWmtjWhMF3zeOUmQ1xuOxbbw>9aJh^*_5ePc`~`
zFTw528o#As<`%{??=U7T@!H^;b;K`h(Xl9>Z$4MQmWWC$jkeo!Bl-O9ct!p|-jGH6
zH5NK`Jn7*+eMbBG<!5Wuo<7UY%Skhw!<n*b*|OHE{g0lyZ1`Fq^RC?T>#v2;-PNyV
zo%nqD_<;kb<<?cUJ#?IAv90&L!GdT;$LmjdbX<3AiWAva-6-vHX=!{Mt5Unj0@p^1
z6^}JS3zKR}Rvq*@@kCi(zD`?l^-<rJ1v9=nT@quLT<WlE-o7gZUXp@YM{SnIEDTa=
zoYG=+HTRfXXyB(yTfd2E`OloTQB3SNV>ZX*o_h-_5*Sz*v=|f_8csE&W>59H;Wa<m
zVoA_l3761s_DhW=T)4S1y@mc?+TUm?^)UR&`L0acJddLcn<h`u)|oS_(>pKNTIf65
zM(xx5Jjtp0O;eRwGhL0kg?38GZde|(YD=hA?b|sbod;8jLU${Q?liRYE}E*w*0SKA
z`Ay}X1P{${`4;a4b^D_SA`Z-ZvRlEt<Hm_gH$zuU(9&3`>Lse?)TVUrTAbZZF8R0f
zzdy*>#oy7RE3|3(KHZZSXIo^yE4p#>NZrTH>q0D_%zUlXm0d0Ln73<#=E?n?Z&muw
z8@JAwBbJ_=(sxa7y_&{)kIsvNfgb~xKYy}hZPy{O;|AB~bR3k^eq)%VBj!|+l9E%}
z`789r#{+xz*iUnEa${JrcJImOZ}S2c?&x0r)5m?Gv$<xWR%YbQhnKbAUKO<suit;y
zY%zPom9W_Xr{%P!sL49{x;Qd4IB<)wC`d90FmOamI_lZnxO!=BvUcZ&9riwFxLh7g
zvUo7XGPfwewfKU1<r7=^ZJ%$?^l%7U{A0~u$v28#VJ+=)2@kuuWgEJM3aux~%sXiL
z%k`yTyuWHv?33m8ze3`^o0Vo=RdnvHoHK1!om1<+utLu@8$KUiuwv14&D=WMdrvNY
zykY<RlFPLj_dB-jc-U){{deKn&1RchFRNVk7ReKhntHJKf$r(xqW`Sv5BOu^rWo9e
zS)u!Jjf`t^Sy`o;R=|qxX@925%seOm`_r72KRAz@9t@bVI<qf&%H&PDtL8t~O<nnE
zyS?K6-Ssl3)gOJT*yHhUMc#uAuMfK)3Dchbb842##(y&R{OamV-L^I_elqu`{-VQ=
z3MEp{$T724?>}@uUuXI*$yZWb^6}U1&B_11<L~@E&XFNZnt~G)s><zq*Z;n+`rgS?
zb-&Ewte3qL58Tf+n{7LN$J#IT-}i1yaeO_qxAACw)b#%J8Gl80%h>OIw>ee(sp86r
z3tLydd%gA6*ZjS09_(Av?2X&rZ<_J`?~GI0wRcPVcFuZpEj#stk;k{w(e88754~S$
zw=eT$TVjU7MDv>#kNd189`{-H3EtdM^XJW-H*fCT-+$)D)8{KM-}=Dz{rk0xQ&%os
zdT{08g4P4J?n*AL@*j>F_C7M~N#~!qH;zAl^WOvCtM}HO?_YoQkJ9hD^?RoLKezsC
zloS_Nl~kSdmd7ReHn0E9mN;e{^|-;}`654t#~uO=PI8s+^xBV3KmWY<bLEqo|DyNR
zC!gE-Ipc44`RunZw7s33i;b@Lau?L}xpcO)D5=cee2RTu<%@p9p4$TX`prk56#pul
zHc8c2>Ftx48J}udx1SVzq-}NUb+&c(EwKgFmO?5T_k5Bit}Qd{IrSoa-{H&8Pk#BO
zW#4=H@a;nq;wKpMC;#|!>e7V^_bxXVRJlChS66oHRQEV$*z+n`;+$|=b@BZ9Z*2d~
zy(JkZH!uCV{JjUPGJAMVMny@9M@dg!n8$f}qLNeZgJXuhamf;APoDXI^Yzzx^Y7hz
zDhI;XMeJ<4TlOrPwAEzG4pWOr#RC4u6B8BPJ|s&V`yp}6xOL<F-?nv|FaN%M|LEJd
z?;m{I`LCu@oWK42pRi@qrn4>Q4$))$2lfI_vc$c^(moCf42+-?v;@A(H@<&2ci!pS
zzaKw6{P^+ngSEwfetpxETfbB?YF1Q?)EqmF80JSsDjtD*XErpj&nsYj{+-|Evqk7E
z*1OlY{AT*|H{j2XZTtT0{8(Qj_S(Pyz=~Uk&iHW)FFJc!n1%6d_Isz_4BvhQzuB`b
zz9ZgB^!MSE$uB4Uk$;<%SZgrR<Ki`^Yhl$@cf7Oa@g4vE`+9m#-OV01&UNq0wr>4a
zZQHBM^S|p_S^d+)*7mpGoIC&f?~T}v3(k}oz5gaQt!;i|(WRL>T=7e%Y*Ll)vUw}4
zEBoo^+<j}=OVgVDcxPL>Sf}$&%Vc5^_uXVtdg3DU!(&&UxC9>S{u9&jN7-2L=n1zc
zZWo)6G-#dLw_n8c<D|l)c6NKBh0NTiRQx<B^YChZxN_vDi+rBum!l${W`#sb*7og*
zKInL7+OCZg`qod-Q&mmrnk?RbJR#=ae8oi7#CWx;xsN8sawzxlX`X)l`sa?%635DW
zjEtsfENy0BP_#P1!J@!$P%(X?=BL19amjL#(9l_n?x*Ov@^p1=I$8WQR3lV7I`+E9
z&cCnsmif*MdjDb0KYPWg-|Fn;^5vhO{p%)jdEOp%8-wX<oV1_GTlYCMGS}UdvVUaa
zs8YD|Vv@!c?!QS3OJ=?en|d$vLFE~lCya5SUw?gUoBB^j^P4v7M&;C#?sI<$_FA2D
z{$Vz0lJbO|!5UGgxTN?yr{8$%y2t)_j?^vpZwEha(UVn~dUj#=fd>Y9Y5o76ZrXY0
zS;<%9dO@$N_eDSSTCS{f3DpSI2wO5~QkPHIvSqAiE5C_N%Ubh4{fw74*9$F+2Y=`6
z+qZvhpJ7jK%f!|`pH!6^XYHpGC;t7Tx^(F@mDm~KAGezM#>%SSo^1QcRaQ8^@Z`H6
zf8taUx6j&F()4k9Q2X+x6N?-gnafs6S*T<*`8+nb#j&_v+GB-cYgg+O9hVx9y`W<l
z=eo%)Vwh~q%<Qo&z29c}Yh~6oC$dhNdnHONUak1%_}^49i)}^h&z!CtW#Im_GgW@^
z4Le<>I}*yXPIt=0Z)fe%lT=fiESB`@Yxe62$4^v$HrXiTzD;GH+td=Ja~ZLJd1oKI
zCBpOKk#aImSXt@U-OBeKFHT;Z)4!#4i|B=ZiwFIdwY7D>W0EEAcs=Hpe^Qn&dN|>;
zs`8Hs6L<XeP>Bs(5-6ARq`v>qh081Eg#YrCf3x#vRrP-T@QD&ZtG~Tw`Z~SYdC$Qo
z8!`kG3Jcu&45bv-S={P&6e(-IUb;nEJY&V8V+TF@4Vz~)$e-G`U%)`nP|(%xoJzEi
zS)G@K{fYiP@1L%*I2;wk=XS|RX8(@=(_cSJ{nwMQgU#a|cbZF4rT(*k9ZRNG`Yi0;
zcy@!H!+ib)Yy11#ZDl|031^<gqR23fgMmYX!%2W~VcC&g?d?-k(>>>R#2TGDx5iJo
z^4+4?c*ig1FMm}X>9JUr{L4~EB{i+_ugM+8wRtT!CNpkU+G3a;XmC+XO6>O3C*2cw
zOT5^es_<gj_W8O`R%ND4l;c<LKA!*hQ^m#hS?*^~ai0_Y%cu9XWX}SJ#^V{iGEOFr
z0fp=KFBB<FzVhZ8|IbTO5*zg|E<TlTJh6{`vw?vV_Y)lzL;tm{Vpn>1OzI21ryNke
zZy8U*D*HoQKgTrd@6-MoAD8+~qcNAK@$NzX3teJsGP7Ruop@dpskG?ifxakxeWsWr
zaS0KUi(IE@@udoU?^qaLuOpxGszdLJVNljXSFH%g73JT1eEs)+*myD2>72lG8A0}f
zA3f3>!4k(FL`y`PFV1D>Ff^F(fBoAtq37lMCYCwr?K8L~5oNjm)J}~rfp+o=KmOI(
ztM?py=Kt&s>wV*+pd)@#7#LhU7#f_|li6<^T;R~SKWCzfljG~l9J}Z82L>5lv{iK2
zf5|~0HJO2B4f~WO9sf2TmgKSXh@3Wa7MINrM``6*nSURzT=Bh6O4|Ir{yryXhdU{Z
zIfV!F3L{U6KiareMxuV3yg=C2rAsm%o&SCy_;^z7k27m@y<D~niCs)9Ts3{_cE6fh
zmM?#KnPdJbN(5YeQuv_d*A~m)@9I7^GO_JxGAZ=9$hz^^@gp3I|MkDR|IYB*b!(*)
zwlREDn>`woxc2QAG7J|CRJGF*i54=;_p+K{C}F%?Z|=m^H7jLOL`)15k{2&})wBO;
zlQ=`Owam?b5>^#fMUN&~UU8bSo41+i@y!o?8#eH-f4-DMxQ|15y12Npew$(Mk!pht
zLHs*!zDTtDDr)+n{?+>Wy-IJ&pJ_d~G5-fc154}X8G-Sd|NbmuU{G>lQD9_he!SrC
z_ucp8nG8M^R{v4h^D6nmhjrqo8wG<;*8c3{+a&x@rbx;>_R}J#Qay9&Nu@hyyw{ID
za%tHtDS?+)3q>rh9*K^g@?-ylttwe%8XFxrj>s}|vlwb91|)^K9+@V!&GrAQ)VF6M
zwzp{Bp0;aY`^wfex*t_ce0O#?9^9Z;&A)W}qW166^QIUrmyz5&U0mEKa<as+<1Ui&
zuXua{+C+2;oVmGq`GOxlQart`bBTyyU`Xr9U9UHmO8u?7Hus^@vqQ@1Vkzr`cXNw|
zi3V}}`+M(s_1Akdeq3D4duDligM{u}Ejc-P%LxvR%&jk_ET&{MpDUcL$P##<$Am@V
z@56-;yM^C$i!I}eV(8HN^~Yh4gQLxq|2ZDpRC4MxK0S0v?CbdF$P<1_r||#jilp0@
zuDH9+VrgJ-P-ZvZ6)tdl?wQ;p{H2?BZHe*n$gXnM+Z^%5N4fU2wU3=N-}UF|&2vI~
z1bC$U-5xC$3t4$oBf__Ivgq~XgKxHPebe>5dh_w6{2PxmA3WXN-Th&iVb8&L1{>^p
zWGxT;X8zmM^7YVRm!Ksv|LT=%pPkBIA-d|@{nUD?r=^<7zps_WueP@0{F3-AFuR}o
zUX?9hJMV<U3Jy)Lx488llv1d$xUqjhm&L(~x=QgCQW8S@osRP@Z2Hh(ry3vMCQ*?;
z#nJy)_0f&W$sgSV-wW1T2|d4Izy1G)yRTO7eY$2!V7x-=-)9Oj%yMOsn=e?()Hy4~
zc=_COn{)L4{%5Qwnjd67TCgyo-|f7gcH=V3xy5VjpT9ePVa-d0ghZaOvNG0l7yo;B
zb+vaSUusc2&cb-QyL;BGVCgju3XBY2SQHsJ>h&&sOw1O)e)hrsg)9C$9Q<FQeIZjk
z>wTe>aO(XnCDy;*iS*3B=ccIR{w)3L&;PX<FDAAxYyK{MDW8Rdp@V_x-F3eHrEMn@
zC&bRaxVXx#L*bl}`9cHhImv+>ODhDIFD=tjR{STDnYps-D5J;kAAhRvzY<)WXT@aa
zWVhsEg~*|kCHuP@;`@w)C-krUtPrNdV{80!!-g-9rZ!1LFyAS<BNjJf!<8$Kr+934
zUD#xI<n#pA*O!BPZFF3$xBhyw{ceMYrG}vu6C0P9$nVtw(cB?|H=EhtgcnpR%K!Yu
zu%<1kPV;>J`495W@4l?RP>^QgBD?+lN0zoldUuLNzRM)tUOrK5_ZfwqEQV*MeAx1|
ztD{q9m+PK=TzCHLoUY03e|ecyMaQYCo_%qdt>ps$rW6V)mHlXNP*jjlP*C7;66gpJ
zXklW_j?s+j<>v}hea1DZ>;DtKeIX0mifqn2WYXbZp5Y^|v~yvSLWsyd^;bgkSmJHl
z7996*^jN#hcGJoB?@q~~KVIwX71hYP64qCDbJmTWk~1r(D<{pFxc#z3^vjl4%TH&`
zPX6dUq3O<&ry7Rdt3N&!eeuxBlBKAB<Hnqa3Oe$a*B|3#j5%2AE)el^%CQt<o`-%N
zv*t{iW4qI5{w<4Rp_<>PPcL$smy}XC?{|HrcXINMIcx2L4<;If&g)mWyP?#%ft6jF
zRY#4rUE=OSK|ke^$&%r-mS?)^Z|az&J!M*?u0Y%R?`%50MT;(R^%*-&6#c5S%-3tm
zEB~pJri!eQ50uVdv5{rwu`88FGM{Rm-g!eqtkY7Jbw%*&kZ{?mJ?*y^E#x-V+Vb+O
zzV>$!bI-s6uH>FWD_o0w+?yN}8RkE=-w4`JugD<%J806nQ~l1%%CB}kR%%RPWp_EM
zm=ciK=+6?i*yzyog+Bt#*(JNWT~5oah&s^of}_dBk1cuTIe`h<!lD~Cd7OxzvfA6*
z@wd^Xs{tnp?Y^tpW>4Fqm8w^tRi!C#Z>i|X*{d3Dc(-4Uoc;CZKH1-vuSK`k&p)l5
za$8EE;NM;SFAfaiJ?eW!*WNw!r|Bhsx1p8CwEpFClUJX0Eb3%h5TsHn)&F|+L;tGE
zN^PkdL3|6vk7Q;Q@UtHICb#QpadlI}_nE6ce-krUY1#KVNTYp*lBM9e7rX8nZGG~~
zZ+~68;pHtaZ<f5#ef;meZPRA6uSd679XH#0xl65X#g)%{fAvjI_@-{|H#5S*!GVd1
zg@u8EZ^wH5;LaH<f3qI>_;{h&{rKu#uk7a5f3N&8x!*}Fa^H1FuO}ar3M?Ef#jb9w
zcl2JcQAtU6+WxDhJ;@e6$&&S7ZX7e57rE}&)~D}&PIsPCIBlPko!tM5)yEvGWWH6L
zR6k?$|IEXPJ!bRY&%gZq!OmZ0*Y=v<k5##Jb5Y5(jCr0vrMzzy&*5HV%U?D1@fl<3
z!?|f61<g)IN1eL6SU7d_6VL%Gd0*JSXjIAa*}G4?{i4KfimcwPN8iK0Ui&-q{O<c=
z>3R)!o97iy`<;C5`B(MRmlnS6ay?}j!?*K#^)Bh-7RL;Gj~Mo($!YE~p4&bD@3)=5
zKU?jMTg&fTcz)G0Yc0vW@28xdz<A)<iiW}i43_`3Cq2;HBpRt=(C+JVM!)IE#K#q{
zBYLF-=NtB<F+VvkpdeTI(&f-jts^0_Q&;J2W#tvWx%sK2>z7|Uze-l;=#|7iOfQ@#
zHesIS+Dxvig00JPQkN}lRdVr|&yy^9_(2wbdhyOX3l}c<@!fj9+^ZY4n){nsKb@@K
z*Y-1d<@9N0L7#57`K0c@cuq`AT=ndd(+U6BVmOr7mv0HbX$)D8nzpb<bBQo0R02=S
z85;|$y7fwl)eC-{&~-&SP+PM;FED9g)cl~OK|z^UjB09&m9CaF2-f#a(PO!I_-}uJ
zZs2{7Y~^aZc>T4U8yBh+{bQPta@q6N)HZ+Z-sNAbv-cHG){f`D|2QLd*{{vnoBrh1
z%iBNtyLs}o{qgJf+`d1fx$fq^Z(ltaYj%~bFET6Nv#}`d>*2@MvV8KpHt;FFoKtvQ
zrf|sv29A`@!`EI<Qx}drGU4o$vy*a`xZW(3_~WN!d$CZcit&hV+*)_1&4<#Rrks)q
z4-KB<;=fU)B>uhFwp}(AwtY@4DQSJN6L(%NwV&d$$x~L0f8mRr=eN(<Y0u{~Q<?3=
z4`t`d_*2vW2G08atT=tqfi%Vc^N(M@At5rM+$#QI@~xR0*Tx^d)^qfP_O|QO^!D(Y
z&0<fS9DT`ZNfOuQ-pWSlufHu$GNq^g|Fh@ad!`TFTzmGnl$s0P+r<7O{rIYfQMpxO
z7pg4I+7~viu6Z$a>Qrg%i*r}qkBj&D^OHj-@vmjxYcHj}@1?e14?OUoKx0);vgEci
z7exO3I-w!8!u;0?4wE(4-zRKiRXomWczk0J|0U(uzjgm^mHmGwKh?hf?qee}hOb>0
zYm=Tld2;fF`8x6If?vK~x~!BD7oX8=e364i!8PX8p+9FF7?>{za5ykA7+yPet@@R}
zbIjEkJ;pbdU4C=i=1)GQ?Kt81l)TJ$UZ=Yzem{Bg%EF{HMFID&hYsC=M<-89oT%lH
zY^h{5(PyUZ$;yy>_5H{0zvb8;HSt~hE$z&AdylHE+86hFecJBzX6`?IoAbNmwO<~1
z*UvO(rIW|wX1&7i`z88&Z3+YKFJ9$6cUPtDJ2~f@Pg5oylem|0Yss@2`@OpZ9AvlL
z5;1U1`!93BaFH^XlBv?6m)zg``&O@7HF0%ROmcEE$6^DQIW37UZZZ*z>LyN{Y8YCY
zacP5r0gL1GMSM<nALBdgI+{D1omx{*?K!h`>z1zGsaslVvR-?{%$<MXPRH$Ix2HSZ
zm9cJFD5I&l_^9Z~vl}n|zjWf*iR06L{rIjF>;3rlotDm*h7s!+j?Ori)}i^#NN~O3
z+zti?=L-TYiVWX6l;7xB?O*?<rDgsnSDX2t-27Zkr!LkO<FRT__}F(y@9s*W$9m@;
z_jn00O_o&9n161!&F{TOla>Fr^mcAgpTKCtQ@L@5r?sZp=WG3kA09RANy`eFm#io;
zS>jmwDx>*ZR_^Bt(Fj%#{xNmSmQ5R{ZrqY|kcr{Ib}hkm8`tZpy%e&Tx}aBN#!N<+
z70+IOS+(edM%+~0Uyku7R)>}a3v=^(wjO%uw&R$w_CkNd2|WvYl8xPF<t=}1oNR2E
zdVAIE`ugdSy1J&321lDXlQ%@Hd(mSX<5(b`QFv69)oHTGl0a|SvVWngb<U>j*;!y*
z9ISfGGxU`2)p?VJJCF3->`BX<G%s0E%30!A`V>>H>a3HQFK@nl`SaKL`d{1Dtuw1L
z?LKb1<BwT&U7Tsgv(uli6!5Z7IKIE$w_=Ys&-HI9B`=j;pW4RdcQ9eXi`6|^X;Z}o
zdZk2)-MhM;vbqa6FflN>bab7o>`7hznETDFqOIr5PoMs;rlzK5FKlf2@{Vfhz1+5A
zMI}-E)1+_s%3L%zGz(Yq3xDJl#N2i1F|*PwImWgxk*4xL!s7K;XC-~iF2DcD@y=<^
zQ^6~Omi%7yLQ`~&{|8r>O-&mltA7NqUYe&?a(u^W=NnV>dJ@-*Ub9RT`E;o@*;r(w
zU$9p3sl$ev>q9nN+8x=rVb)wZu9>Hc{^@G4Ydp9-<&=iKR$5JF__8To^THp6t=F2U
zXBp%s5K_4LRJ-|{dxkfpLjUCz9e=N%)6KDGe`~i%%bqP;cJAle&vZ@bx$&BaHLKs*
zZ*-GC+&g2FU`oI6+WJYGiqbcDd3vcEx0yfC?wow<l=MSQ6J14}GI_oLpU_*2*FR2D
zGCicMq$@Vb^wKu5u=>{cr%R8SnHlsLvv{6xQ*lW=+NI}a6cl9rXRcKEBY_r0#@5yr
zll1kE_*@cvRBLW=Nll*dWben;6Ib7@=IT^&>HlfcBw`|X_x{|qOBZ#X%*@P8&0go=
zWNfj+d<*BF!$umq7ZP0ky}nlOEwbb+vwzZDIZ5!6C+8_y`%68`UtOQ1JlSN<lA?;o
zX37)f&n%g7YQ<FP4Q6vDKDFp7d8B!HxpwC);VDmEp3C)1o;^{@<>JdFulK9DEp&gf
z!$#?`ZDZRus|{<OB~8-26lCPNq(9TG>%b<z(5*`|Qd4g|eaAnef5!BU8#hL3t-i8S
zJFBvBs>8g`E*d$~IeL25uLUFG=eTCN{hs_ebCXjW<HLKm|J1fG&wo_6*-rQE%b)M>
zeE0a#V`VfUK5UV&#^IYTZ#VBx{;4ssE>MxJxTkZE>%L_YE9(vK&stU{IycIB^7|zd
zZ7O=4`j#eN{yOD#ckPeK_ql^-6)-&bvGZReb05PY1xAjEN-my?0Ub_vS%p;G*B_Oy
z?>>-VApZYJNwxUvsDF>cSryq1@=rV3`Ms<}JJfxzJ|lnT-x&*>lIyL6?ygQs{uDcB
zGN+))x;>p2!&fX?$R0f7`hu3knG3m3n~RAoce>{E`(b}lXGdqJv4r!SOI<4xqqC0)
zvM8RIt-v~U{U_z0_08Imv-Yc6e`e)mWve!^PkpAOB*^XHIOoD7mdXhM2B~f`pC@j2
z=yopk=#AceGP>v1zu<irLjAY4Z{A}cayX<eU<>QPvz&6R|35Q^g(NZ+nLTgJ$otyy
zZDr=EJ88Fq=DuCJboqAMx&J?g7J9ySx*%UXOQ^o>(Ic-VD-JDqmR8mHu#PX2D<dl7
zg!3g`r7q1&fr5b(+c#f1dL?A$*<wbCz&FRrypE)aZ<;Ie@7G_6s{Gg;J6Ra7U+Xwn
zp}@ev!jqDcIf*6nOwY!}bt^6BNH0vB=rpg$#gx@*k&)W+Z`D>7A@`e28jPA47+S@S
zlr)R0t2bylEVY!fPF(mcxZLZ=T>(uenaA~44jnj<95mxte0;r$_5Vk5+8_O$1Nr&3
z@UDq_)fMb<MNMo&y`{9S!~Ubk61z)11f6a@4p8IIJl#F-;`Wta>bYuaT#r;`w3r43
zIUSJuJVWfho8dnrjunSAo-HX<7UUHa2)!;9=q0jT>Lr&YPh;;_!=8QnS+dh3B4*5r
zs^5Q)@$X+f&ptUVjr%+HANV!fGe=YM5u5qj-`#red&M;VUAx%&X+8U$@QJnTucg1P
z)8qRoW?1;~&@RE_iZQ(sCsv#|w3yp?#>FGue{6g>16w0cpLlR=-Kupu+wvM17(84x
zFCRN5=OiGgz$DP(s{Joy&G8M-@5cPxeMijOJHI*O_qEHvZBE+7-n`-yexf2VyMK4&
z|JR4>s-L$%jJu)#w{>~=HOr_6N(D?zF@8(`ZF@WYvgo~u-+#5=lY6-FhTziX{fETl
zrpW~9xpb_5$dah=(EHo6-&~?A98)=rdrtcptXp^GXvgh+J+}8I{*z<){4mM2angiH
zCV{6njMwj7%Oy}}blHdDVC`Om)XdmsG57!Jn*`4I`1m;fG8Pt(NqC^}^=VSlS8m}i
zJnU;uPnjUH?dFUdH%@0oWh~v*vE=Zb{Ja|`N98TjVlQSYt#qILtFLeQWz*vu*1ddm
zsVDvQoH)Vt6H_fO>1pY2s@L7ULU;d-6NfhIXH+~*=W6bq`l~Nm=T*jJi4)<w*!o(w
z)%W%NJ{}!<ynfPs@x7l!)BHug=oh{Zb)6Wj_3KAY>fOfnJ_iN`j|B`00$*4dn8X&{
z@$NCx`xd<N;fDXKCVzBYDg8}qVZn~pOP8iZymU!=`Od3y?eyh6+tS<;)E^udU<kbZ
zPc%kxZ%laj^;LF{j-AinZY-|#pm=HWcHO-p?>;@a-lL!`E7zM_6>$1!+<(#7l2uF7
z+_r7ovR$L{Y4i@QDX#=C%#;dQ9e(Qa@%xHOF_P0Si@Z8%-SYML`t|GArple4J!$sb
zxf@>k`}_Hx^}Qax?}*Qa`n9_3rl#-z{Z&3c@9*!{{mLemj~=}BGTRiC`rm)9!=jqN
zb-sKh>U?pwRSApD%}rud?oPOQ=8tB#>#s$k$Lt$6d=}%i+4Erw%g56`#mT01d(Mkh
zKC8&TJo(VJ(|XD$r;BASo;i~#!H@m(#NRJnA|J8Eg&AZ{b@LIGIXfrmu-7?1-kAL_
zGSV_wm#FbssQcZ^m)gEr&9LWWT(-5e<;|1#&n|iM%VqCwE8T^Y83Y;_x6i#R{{8Z=
z$X|AK@jYEu5-0XQX{o=-p|*I^)~$+?CpZKcEaSU3*1z6-_2xuD(eUXnHi>tui7)uh
zXdl<NR`Jlp+j`S962G0A@JmEK{Z&0{fO7hV08NFd8hV0BjXVB$b%kzT_v-S|Y4J}D
zEG07~IhnnEAEiCzZQ5g*u%x%?Lugl}e#2t7Inf2}ajDr;M6ISab?Rw4x-VTSmJm=l
zJ(w%gRmk;Zz0MWUZ1#2Xn_87W^!?})a*Kbclv;LsOW3l5+cw4v&t`AF*SlufpTNZ(
zGBPu-w|tt!I%UTM8P%7??gg2e&OJ2~{%HSGx*^oAl;o#>{)7J0Ltp+Gr0#WHGo{NY
zfBMguof~}4Y~RV8up#jA&;B1>R?80_JRvjr_oo)E{3SD%ddq31oZIW+enR}Yq>9Ck
zng8}?hF4x&dpEhX`!WM_^ZV~-ELt8ESl$r-F}MHsk7cVYI#OB%8<O_b2QHiVi?3F@
z^<?C9aqZKef8Fdqclnj3&c_{>*6YREspc^x{^;g(X#Dc?`I+@E4Rp0vhIJj<WAyaj
z-P)vyN`|iYX0O#V{ys(W>;D&<%8l<H`e9vv@!G}m-yfH~i>a)se;gmJH~p$wQpQ5b
zDZ&|oO7FO1C9g->Jzbdb`|iU_y3h7KP0?eQ2x|ZFSnm&8Q`4=Q9a&cL8^fk;s!#g6
z|Jq^R2is(3D}`kli>UP(z4>s#e|v!Fthz3n)!y+J|9h<vjooE9=iPVr(|11g<@KMv
ztGj*k#n<1YO-nLP1<p2|^^;fp(v85y+hp$Fc)kAK?ROF9#ajxvEBF13^zL{kvha?@
z=`GIg-+WbK{L6Q{gf3lW`1ae2W5qjf%UqUv6;~?xB}84~x8S_Rzqu|1$aMB~)Ln^>
zpXmSYkDcnziha%JKmKGt`}hlARsZED`o(`e^Iw`(Zr!Z^z^vN#fjsl}hQPyn`08I3
z`LF(8u^~=hV5#>5k-d-C&wcm3VQ&Bf6MK{KJ{1Ll6$~wkrrfjIHk*t4-_vyJd8{AP
zU-#Dfo?iUE-M6(aUFG=pyY}Kg#rJ#@J2uK*Hfdz6Io;&h9QfVu-GA%D?)TnvH~r9)
zy{ED64Qu_*dvdpR4=nU<T&U%;GQ<1(_ILcnyFZD{<F026pRPAuHKEe7`R$)-{pQz?
z#afT9GTN|S_j~uf4?oy=*;|?9<=Z6oXRL1AzQ2ymb)v@kmtPN^d;CuJfB)}?`qP#z
zny-^9&D{5UPj0a4iXTh=D6jWi@j)bd57RE5=hBQWYB`H!xWqUgox5<m=Pmz@z_Q=d
zp8lz`k-hXX>d&;NY23v#@0?3dPk1}6djCE7{k6{z#J=6TzhU2c_7(YM|Msg$uKjyu
z-}%hh&3lTY;~Q%B>FZBfW2Yzder4Rx_0J0*{$2X%>4v$||Dxxf-~1#`L1gb=b;Vph
zuA2D2)h+*xRXREr#>dyMs#X2!F>^X|AG=k{qrX3|y!uku)SF+TqsUb_Ly48g$0ab}
zPZUeQk&e5Htp!W#H8wN7;Y^HCWcj^ZPGpZnT)^R)B-hHxEapWUA9VBc74|mENIYb@
z@Z9O(!{gk1yELZ%cX}i{d8Sgpl2*f>WD%~Awq3X8z4%rlye~~kOiJ{h7!x0t5O0wj
zmq65^1JBa31-EQ&Zq+-TwD_sYPI>YF%FP|KXUz;ZH8nOh+9K~4GOgZ)h3!pxqS(^e
zr9F-dCKkP3bWhdWDs;h%IjO4@<<$4BSzS6;?t8$wTDu2J6IP2cypl=dujqe%^t^Ul
zzgtV@nPf@z9f#lS{4Tm^{mmO+f4^Fpm9;9XKD+e(tB)VHZz@{l+mf<F{_4?yhnM0u
zyGHXIJkA#L_}$Dz`MPMmI4gq{A{zHv65<zd-%w!W;5ot2qTr~|&}6u(bl0Wd&$hf{
zJ^E;)k22^2si|z@&%={8gvOiR>hSPd5hM3#m0*4@=XU7@`2oV2JDQCBUp;%~aQcVZ
zwZ4n%&dzXbsB4mG@@f7VcJRaFrVWp8{CuR}cb+xcjlZcNOr}$LK^ME<#WV9~U9Y+<
z`Hu1JpJR1(^*a1df9=?rvo3T(yic;^{11tT&W7!iZ2h$E;IBh`TwHv7j|B7b_SqDR
zelll$cgp;fmo$&)b92+`W6L(}Td_toPJjK&f~zaCH8uZU`Cp&9f2l<Et5?g-jwqOB
z#Z4@T_&D8#_t%k(2Z5a*l<Xt)Hzq$+bmh9Dv7Y_Ln{%mL?=Kwwzpz3z)mTR|aqs->
z_qUXL1-k6i=7dXGGhN_&F4-pHI*-APiBDk7uSWt6PA*P|k2oD()VZWh^jdturWVHI
zER4rDJMbxdE3<KEU+|?)KVox(YL~TK?BVD2H($rw*57F2ZP?|0BJ|V3w%*>J20Q&y
zUfH*I=1o0a|L(!e<5Qj}wXsAV-c)$tRpHxyC4F`~g)FaLp<l{-J6x}Cc@biFKKb-L
zj_>=<F#lma_%i-?fZwy0@>a2M@sDEDABy*Hep$%2M(Td?i-Z3XS%tpeam`jTpJXeO
zwtQmcuf6+3Gllpc{Q394&{qFNtU`SIQ}(~2r>m=z*6x$d<&LwC$S!nw$ff&`tMH&m
z;hQgqi{t$hf*-AZyT(^GGDZ8~cgxD`S4Q)8y?<G-wcNBWKUV%%Rr<%UX{**9J6f`J
zdF!NKc1M0NnA}&Mm><36_Y>K^%de!DNoqw1>#|0kc<2{h0Xj!uBLf4Yn6B83&u5Yy
zB#yaX)A|zs%g9Kv?5m*1zEf9kmbR6}ch#<uUUkVrsOzI5OO>wE<sX(?kJvV4F!z`&
zy1S<CqeGoSz|nO<OFpf0QY>(ib&P)_)3#uPYE)B`+~GaSYZRNN&63HR?p-Lx5tzck
z+;i-Q&b-GH_Xq8tDsBDazwCbb=W8p!SKN>|aczc)`Lqd5V$-@DGyWdv-1Xe}pkv<F
zTDwaU8c~U7V<qLfbn=@Hd?~(q`o#fTd%mSR^Mj`UsLh=9tF^f54fpK(b_+jkd$GN)
z#>4z{V+#MpWBISTjdz~-Kb@b|Eo-A}ctTL=xtupKWnMpT{<STeA6tK7b3mxL!G=<g
z!gY!vdOH6X=Kc-iT<K>d^!wD!_e-VcUx_u<_4W1A*!7!n(zei?_$zxh9NnI{cGa1y
zOvNEnR3;X^)E973WSHu~(B|MI(6B?$sp?9=$&mGK)4$Bs*!p@&ve(7PNR2til9O_l
z8M9iRS+;7jHmB4j(=(DeqH}V*0u7?)BpvUW(=unG{}g|X>5+e(O%KizKK!BRSIA>=
zd1a}v>HG;NPyBYx<=+zCG~-#>7u^qShCN4>vNBd>WmI~4Wy+rNYne6iMD#VE)2nPl
zL(kS$&OUsvI65Zc_me5dr*;^3ubx?Y<M&cu^X;eO!}WG;{++$|+rNKe`&kk{ZhDw`
zlY#NLyj1t?^7Bsw(l_(1$gkS6GTvv4z;op~%Y5Bw9wyHWd(sa}oY?m?w)?rO!|w;L
z5|ti3d{jSC?C;W_>?av_Z!K2jGCM9a|C??F^Blc{t-^_X22WdDFT}*{l9<z{`SAX-
z{@8O00vz7y<#*b`){5!Qk#(JT<5|X}zo(*?-|&qo4t~~hJ?HS5?;qth{{0geFXs{=
z?O?hzotN=ATf!5azm18)Gp#vePffh=J!$7s{X6T{eQjQ~=8}y#!}nwB?tQZFDBx*$
zqwCLZvUTc>rR8gcOZqaa%c|C{d8PPTp|q|}dK=r6@Y#Jy>0AqqXI@`eBIaM~|6$$N
z(0<jpgjF?7wQm&nY+}A~PvTYUqNNvnzgV2`4hZ`Dw&i~DtOr^PR`2$oIzj$*>8ckU
zjr(`p?oNoFlh);!`hCu6Z|{Kkr=j9|)^Q&UQfKu%{rcOZ{XRwif2}gvc<#%2*OU4z
z7VA}4NpxiNAN22<(51g`)v7f+tA4!9y6@Fu*_O~RF7iB1ru)5S*2|VF%OyFpye^0O
z7=7}}*4q33)XJzWU9Xq9B{SUA2~2(4ZPdK-${DU*Tbm`zqkFvhoEVu_G6)DPU|>+Q
z`Mo4?34f1B-B)eVvTdiUel2Nk2+>K*E1W%r>r?T>+joAQ>XorroT_55q%Dvyq^(PE
zGV>(QITgAmxkQYZIt{OTYH|v#-s5$xYNOG<aLK6shhpRGKbu>2?r^HQpf1*FJp1-(
zwV;(2Tniq}i)xy-YJsmnanIhMQ@3t+y;}9MW>V@_<IjsD4cljG?LXg=vu;H}AX~Oi
zag5iNqZgO#ix$lhmPwh~tg<q6XO!@2!H`u4_WK{5bLHmjh#Rv*XKsx=b4pd!>(<0=
zM^-0%E8Mt><)B>Qv?pyrenA1t|Ey8A51f`!cPi;<gDPJ_w2#@Ops?2L!cFpTCKX0I
z3uzw|D0|)4_EI5P;!MD?&A)4(JPtmt{_liJz&wwlmCKhtS<1g=+O$QBjVnB5u3gtK
z@=_I<+MPKmFFP;0pybIv$$#?vPtTu<Y|>z8aFyb@mh3q5eMX6<5F-O4Th5!yk6$oY
zrucX)2@BsXeQ%+Zp;JZ5Im4vZOH6D!+YAevCgzERGwpe+8hHA}1e=ZxVLdV|w$ZPb
zmS|}lfAujYrB2LYo8=^tXD%L!E|K!?Qr~%*8JXX=yK=I}Y4^-DOyN6tW&*pBlH+y3
zl#BhT-?-Qo@xFa2>LVI(eB=7^h?nR0&g}fN-5`G9wf_sXUhST_O-%Hfn1tvyVS^9e
z?^~MX!llH-CmcGwD(29eYi++7l-4I%$H`sKW#4Wux;~%7XpVrpp7t`qxF+G1xepdD
z>@Qm0T(aHlX<_x2z}RNtmGkGllsg~S{@qG!k=!DI{VzVxtNrTSeC4p`*ZcAUhn>T?
z?9YGF*ueFGwZfp`pPuh^W>#PRAKE_zS{T^4QvTU|<$0S@(;F4rtm9*U<>2eCuW~IM
zM=sCTQqk@9t=i(Tg=0dC=P}oBf}I&(MRn3wJ-?ChgOkzgAfFE7OUpbnDV2?O`5K>s
zbU)?oFaD_(Q>Wv<Oeaa%R>H$Xy!FEqrFAjA@>5*G&a626s#x~7l=7FZf744+CZ<{m
zaj9`Be6u)jfBRONL6_xH@fFE?3+EkL%n>|s&yjB9v>opMZq;(#NZbBs&R<4eW{zr`
z5|xsi5HGQYR?CGd3WsVHmAHB=k1OZ3inTr8DB;0jv|Vx=%W?;&szh7&&EIPM_14QR
z(Pd;PcJjA4VDDJf)g$lccK5kh+xrRKSC2ngdo1^p$otEW-@pD=Z6m&oZ_>?&+z$1A
z86JNO3IqyL%M}D#7(B!dO;ndin3e6X@#SaI1iN<0(=QCRde>i?m->nEMK?p`vzU0M
zTfgn*^DhmZ{`sZ2sCZla?98|qn@k+1UP)}{78VwY=%~n8G1<Z8R`+C`fV&G9E$nV^
zxhNDit;#X|$6QhWPhu%6lf0(tI4td4+SB=6o?lePFi42GmjB_Ih+BcndjAEB9u&Q2
zuvl7HEOz3oMO`I6xhA=qUrtWuy80}`RVU#1G56zB+kRVU9Af_YXUP)VtyO|1hI4#o
zq-d^e`@`AiT(?wGQY+=Co0xn0#N)g+EhU}d5&xQ|DDs(Q@s#Zm*ElpUSy4W-CrwN~
za^c<yFINVuP3Z5RxOBp|>XqV~bk5pc`gr_>#VP+@&0tRr|G(xpT@Ktc@^X8()N9g%
zizb@d-c}RS1JyO(O|qY~ZToAjTX!a`{jV~4xnE5NgMfY6{Xzvsjt~|_rns1*2X~6a
zTp7>E?=M=kea?#b|J8-{F-nQiF+TQNgX%Zkm6*At>!J4VZGI^$z8u;cygXS~@t}h4
zi!8Q(b6q9|w>AQXe?!ckRBXPB&(_uCqxAu6@#7mD44XC><Uytyb4q<+RaqqgsmS
zPWv6(*?&A&`*Cn$-^7VJGIec9VL>Z-rss85Rh)4QDHdGY7A&5t5_avyM=!y2$w?YT
zmD+zi{S7DddiJD=NeH_IA3J{a`0-=Mj#s~0qGJ1r`$k*Zj72Zl3<?V`ZdKi}<LAjR
zseSjKnD<<1xchkVlhe&E7vsXFRa{J;ux{d%_4Xy(w#$WX-RT&;KJfdbNp^-I3y!VY
zCbPp+=IEpeB400Mo_h50pL8|X$LPEU!R&qsi5oA4KIQG-=q9*H$i8CpNAGJ!LTs5P
z9d}N8@)&YB2~02e&!xT3Bj??kZ>es)>WZ#%nrqfPIepIe`iqa|v+A$>Q+ug(pZB}x
z_CK#byyKa--<17SWm)=~D|4BDabK2MaQ84r;g&lO9nbAsuGy0uDRIp7+i$t5PSGRk
zi~r;Zuh`tM>B+q*>oX>;YYmJ{=`)zaE-RzMl^wDwSW`2m-?w~Y-^PHN+qw(xt*>v~
zv+DFpJys!C)tsD_V%Dr-uh#hQU%U2(Xl`CW*rzP*{apV4ZbtHD6%~|O1lf1X@mj12
z*qjhHt<o`l%DP~kTp`_p^)Xu0bxvz&$Vh1x&e(ZtwU$9pagxbn?caC3rb>9p?CG^Q
z<>+`ye5zu+cFde#CtqfsJZ#7^^Q@Jop`>^JNq>`BU7JnIw8KhzR~7_miL4O4qL*|^
z)Lll@UGAuo(mX$(owadpK0DjbzYVH1z8!LMWl*yHvelD~PP#TNXkilQU9ceP(q^88
zA#vViR!c&}|7bdTp18B<NbOg-PxEqH*6%qJKfB-m$ytU&Ctie}I`Zg*3;X^5PKUZI
z+}7ExSw73o(Q&5pm+hYFYSj#3GKcy^>ca&-vbu>LSB~iwQ?ocIDYfg;yu3R#jWhN~
z9yz*xon?m#Uy|DoBg0PytZwp?CvI`<mWsN#uA{G`ug`vNe%q|6oB#KE7ytQu@q%CE
zox}S7cF+CJwwaC5>n>YC3p>vnF8xWD+*LRPRO8%w?@GVl^Hb37_rLQxzXFR^)NEHi
zRdi={%Kj7Tsy}~o>eQtr$K`w9T(*DW%}-vp`nD%;$T5$+wtQXQ(H7Af*3{I}BCGG-
zg;~3-xC`q2%oQ4(7!oEha`;U2iWPWlxXe7{nBg3&Wp9gvE`QtmcU60sPkH&;_c2v(
z!?^>>{8Ca9e~Vr`dpROT>gwX(`)dAeUDKPpYiDTDibIk2jwfzVlv=lUO4PS7wYya+
zu^}GL%SwMQDU;`I^RW3Wam;+nF(W<A@Txm?(?911US9FC)FR}nY~1|N`mOIu#isoh
z%;&K@^Y8uhfOWxB7ERUFSjwH;#~*n7?%(FJuQL1JF533(zimgr|9X*o#vxU&Tw7<|
zh;;r?IPGuqkMGqhx%)FO@X6cHyzs*J?S=oWjg$R!8V_4suM1fJ@m%_&!k;nwV!WjL
zdN^L~KBFV({a&LkEult#<?kzo_jW|6r$(0Uej?<!&yhz&L`1ME=ifp%aRGz66Ura7
zFE|%-z7{sBa|yRrzM;UtBK<{rqeF`#PeTg>TUdNx5Yxr?_f|Mx?{_$++~=j4aL${P
zv$gTi0mUe<ozFF<z6h5&7gImEp`pIisa@(VC+mw-irI<D+cnl!9)I}aOO@B7^Up*Q
zEnj~Azxnw3^$MRkc2v6NzbLvtOa1(dZSNLFdF}kJG0h@eu`lM!;eA!{Viy*6Z2wkm
zTh5WMeed8m_v1XtS`*uUe|r%s<+Od5jg3b&+spn5@;Q0m--U5YAG*HmQvKzZQ-81Y
z{=O&9Qe2I{Ci2#_4NRu-_RB7rXXa}NC^$48+$v>mlF@YK4Y%XX7SAZ}H%q>AzP@WH
z?sq(8jo;qts2Jf*;%C%%giGyubVYXZ9yu0Wy)-d#!_BWEzs<Fdop(vS?(&AuH?Hrz
ze%$!|+yX@ghA%7%0umR*H+p*6PIpfakLlCQXKR$=iM_ud&T(5h=iPOey%&4F6ny*U
z)D2nVo;@#=p1r)OJGt=fm$q}?c00d|@m!!MyX@TF`x2*Gm42`Ldgt!)=?m{fCeGt}
z_3M}Rflo!Bvv+*HbGJ%0u58unl_z&Eto-p&%Pw%^CL<vuZO8XdRvu1hXvsNIESMIP
z|09G0<VDuYCcW~VQVKt1D?IOrc*iQr&lNoxUVmBa?*BcH<YndOgiB^0onqatchT}Y
z_npFbTp@c8ddVK1bno}{L-*HDe|SSB`ufvd>5qJNJd0brJtMX%?_J*SH%9lq-c@*g
zx8-@c-R0dozlSYYAl$sD%wGPc=*8E3*SECps=4d8diM6j%kN(PE%VrSaPI3VRqy(i
zK5OJ~*sZ`Mz`v`^r*!q|P2$P7GQRfL-@Ellyi)hW)aD0D2hR%c<NNb{`{WN10+qsz
z51th;#jopo8Yjc>z|VI5y7ap9jCbW(tGB$^F_G!*ua3nx7cJIqoV|ak_4;Kczpdmh
z{Jm3O`EHNmKZgU|m$!FL{*}<(V>$2Ng1knz;;CPGuPGce68I^~CB<kEIVZW^<={t)
zW19sf4~l)i_$NBXCtvMt`bCwWeu3q7AsPWcwEo2D&lM7}%Z}fF`}&&`zy4kGj(a~<
z<F0`>+ZS%WD{<4fUmf1u?Y>%TX*auQS9I=rlPycOY}vAP>eRiHJXdy3^tp9m=BA2o
zhmAa^25nVZ)z;P)YnGb-zkU*b@%6OW&W&DOCCC42txNe6WdAE-9b;1AjF65sp<1y&
zr%jy_5!(GlM1?WmtgNVYMzEn?%&qrw&!>7DwXU6%8nbfh+4KB?=51TH%sqGG*n}1a
zrB_UhoOKT77p7VVUDmN&>Yg^O|J>B@-7;2xtL}cjlNd6!EZSpj{LU!PrIVG8U2$7_
zWcBnVx@<`)n%8<37wWAGE%sfOC%Wxc=)xmzoYz+ArD&c%x5bOwWVO%Ai0QMQoHCks
zzcJG1xLWW8jioz(Oc&}7x)=5!sBnfz$G*M$L^quFZx!Qjc)<Gh)01=jrvqHw{<8g9
zakKO^YrKBf5nZ2Y6R)dHU%gc}X*FxAe(`>H_m4K;ay-ubuwkBgU}C^#<G(GTb)wV0
z>uh@bAvdomCa}JK<I%LZgee-!|NZlxW~Q1L<rMQ+w}vI;^h&+xrju!*L7N_Je)w~%
z?r(LgMcGfUIGs6dY-o@=`SauMAO7dW(*OQh_;%I@q2mS?XZcTY@jnUUtO@<p+Ft+r
zSH$IiD^#Yd+b{cE6>U;^SFYW4;lH(Qi_-6W`NqO0$iTqJV3{7E)VVk2J?JEWtm9T+
zrQbU|Zk_3NsH3CDkKf}&<+qYro=V*ZAL^fbFKhk$p{_7CtzJx_&f@g^#s}dacBnSi
zHHmchH|>z$yhCR492prQ?u$LihS@xPch2;OZ~A((^1*JKwQ=*cS9gW6t(w$jH8Jz8
zY~9<k-L7^!-pF3CZ`yXdPwv{fb$ixmS1DdhGkUymZKc%P!xye+-aqC$cb0GCi$9sO
z>L2aNGG+SobIr9p!|*Ur-$}f6o3O#ht~c*K*y_0^UpH)R&6mICVR=rVFMEx?ZF~Fl
z^yE)#7rvL7U7k=RT01czWwl_+Lcx?(;&$0fS$!ivyw}%zn$BK4JyOf%U{K+d-5eW3
zIIo&a>DxQ`qus0cQpXv!=gXcZb0yD+c*zp6?@+0(uIcs<->mD8e_C>E&6?lS+`p?v
zgm_-QBWBRTB+$a?-I?iJy>w~&^l;mVP<zRZ&O2tW)h~+c$$uD8`gQdy)7W=kGou;A
z@Aq%}UR(43-qRzuBL8l7o#s62dcmZlE2l2Z(d)9``>aWzB`=}SrSo8aRwEM&hk&4<
zsM}e$vzMif8}}{BDVw+MX!O?qX8UF<@lKp^NhbWO=K8Czzn;3=cR2FbCWS*YdmN9}
zTg=)kx87mxwaiae%Uq;_JEQ-pZr`^)cfHLGXPbPRH#)B+jv4<+6JTKORc9zLUHd!g
zn&EUc!=9Oy=})WvY?(i;`+tV<{dq-k@s=Mv7wu0f`ae%CZ?o_!pKbBcGo!jH6S%5=
z25saIZ||9#{V+7NUd^!0_+n48^pRu<(5^PqhZjzV9)0qys&0Sz=GR|;SN^^CzWnyx
zZ~yPzU*CWA)weAnzrJ)Pf6v-Gb@i&UtKnI{=dO`Ix##oxA34bq=OufRC7*BXNw(Q$
z*pv3M&2W8j+VgXN>P+m;r*Gc<>TA`ky>YAOhcCbS{AZP}O<ldW3;W_{=KonuJoCO^
z_{@3mv2{}Zj2ZHLd*kLEeShHBpWw$!7A#3vFk$_}4}X6!tA4cmc~!HazS*s>ez*OP
z*>mUq{j)D#>fZg^@82w)oVc&`<n?PVhhKE}zMC%ocwcpWU1p?7P-M@(!^VN3_fP*Y
z_FcPGa*dvt!;ioH{ZU$_^G@%zEw^t_Qc}|T^QcdIFL$4Xj1z;!1BQ+VpOe}RpINFc
zSm7<U(CcSll}-G8#hEjh4E*>dPj}ifKRNYuao0Hqr}NW#&&W+ZSmL%%!cF4!@>>ew
zN=mLm5rxql#XVgziT^eVKK**6-c{-8iFGWU$6|i)1o5v4yf`UvvBT7?w*omPp4aD3
ze|2ER#k)u4E<CA^-5#=g#@!1`y|0UGSZ;dki1ReFg?9>5&g;cK?71ayyws|d!)VHi
zjrETW56Bi&t8gaQ+Xqf;P44V`>f)l5IMb-0^iAlLyU_;gjXD2EEcVHndsl7#;#XFo
zd7C3YtQU?y>NU4tROQ+qPSc3#Rm&P0*UBtXT@kdnrMX!%D=~|kTTsM){TziQef_oC
z`%gY{d;0SAvres!{QY->e}54XNxHHo=fE@DZ(my{%-Jz}j=4B<`<=hOEfM8xo02xH
zxcHY<X6K0~%WRg;UOPu*WzC~$OJ3j2jSbAM+xPIU?#J4vIsBEk`y3cK9PTV&U@-8O
zIOe`<hOw;Yt;aXDyep<g7>gQD?i73WDp2>luiv>X^#WJ@!~5LI-_~jwsK5XFZ{sH0
zjVI?S9!;NbEpq-)k(Nn$pnFu2<<H+jUtegrYDG@;wkZ0`%gULN{lZ;hd(S%UfRE?i
zmj^yvwfaiEvXK9`d2Zpy{iUCs`E#02C*twJq%~>Z&y;5Wo*FpE*x=u>?Io)Vg1?=%
zSo-ixp*zpsnu*%NOm`ZyizY4Bu3jhNEwW1I*DNm1l{;T<@vOdI_rLw{y$|In+Sk@(
z{%^X;`K(GuPF?i*zuD2Xdo!)m_TP|QJAeK=Q7O9zi#J-z=kLsm`zl`%v~;t`D!#qn
zu4?P-uit0!bpP*dVF%_~9(xh6UDK;OFfTD-s}$?xEy9;yt4&l{y#3vG>&+SU^ZR~0
z&iPZ{^zYBVe)kzSelN~rtGFTgcw=y|@lv<Gg22_5QYTLyI(POwSIFk3X(F%pm^|HC
zU;Kr+Kl=D8od$sx<tzpz7WU+gM_E5jS+ViNnVOY%Hs8qOT6!uWed$E+^HKgWixQHO
zlKF4WbYM8<)&8xjB<*#hi^@4|t;Y0#m4W_KV|+r{LNvuiInUWJCTRUKdG@$S&wW~F
z_PmhHU7Dxj)(1G+zg}SEwdAPGdaoy&jke9_Uw`fSpRV_f2Odc4^_~_KS=ON4!gyx2
z!dkCcVpsAv8gg;%nHsS&Qb${I8lRfpgs8}C+??Ulw>*2a+E07Mw548sXG2zcr|PWO
zHZ|yRrm2CZZn|3co7G=`{jKZYZh!I1uUfll+EcES94>!*=`Z(}HS1QcH8=X*RQ!H6
zqf4=S`$4ymjNgm;DxTE-`(Ck+?|(zk()U~upEua?{jb??&)!nZoqw9^hH(0V3mTt7
zL*I&bovaId5^a1ae);czm(TvIxo7_LMSWfMrDZ<X&z(9`^ImJytIhW$&)it~^U?XJ
z3w*bZKQs${viY&Aqm!}yF2g<lma{IM*An5iXYU-drfW4Zf?7<B98(!NLaKG`E-amY
z_04H>$CYAIq6a<Q-7Vs6zR$n;`qQ8P^J-ps`On``KKE7b&6-eAx&4LxySL@+Y1`9a
zt|oh{(o$)Or;5g<IsLZtcTYcge~QS;5~b+}itN?g`}R~n^E8>LU2)o^{J2q8KWAuM
z>o-G{O3r7sb?x(?Rms)Q`QmYL3IE!zjpx@H32pWfS?EwWV+C7NQC-9G^Zw_j6}8pm
zy}rI$X`_Me$*C@n3WM#fG?u?GVBLE5<jIpamM?TI-*Yc2a(mDVub7&R(K&pnq56OH
z@12Xt+q`*OYMghTxwvq1p}Um@TUq>%)vvoJr_Q-~uIFjbheiJ*&rH8=|8~XFWy`(?
zRqnN0Kfkx=_R$bazQ5@``}EWoZM$*j_tpE|*G)e@c)D(X?SmB;zaN$PdE&*246ff!
ze>8*sy6P9a;lF<Ho__J+d$L{kqa@ZgezuNWqfve;Xwmg@A063^ixw_=`@Q<VJ(KFz
zt5?6?-BqHhch@ue*}3G|bLZ+@*?ZtbBL~a11xzeQ{(PK%?fFrjH$OJ`UGBR%<NBiY
zx@+%<zqJ1#vzO<?BKMEK#JXR)-~aG)Pkg+5qk7x(Ki0gWTPyR9e7Jr2^5ywu{G4Zf
zQ)Oar+-6(z$7DhM`X{%4-+%G{^UoW9ul+yxdk1&D`R9KtKK>8sw%EQyq9TjU-`zH&
zz2c#2$9kiVj*gD5-8zfclpHg8@ZW~zP(?fU!H;`PKh*fk9}nLhFa6=4q@zwnM)Fb1
z{)aX?MaT7Oqc%RB_w!<%S6Bb*ufHW*)c7pa&Lz1U8Sl?Ld1=1AxxGp1hLFMoYo4C|
zvtipR%iQ!!7MtZGJChd8>Yw#vX1#ihPL3G!Ij3n38*krzaiU(u{kh~uku^e0mM1Ps
zOkS9@=$MtBMxg3X?%;5b@{Fk6tgpX3YTqb+Oe|%>W6@)sa!33>1R5Rf_*$uSF6fV!
z&U)q(kBiu(r=F5a*PdZBFJ2`);^EQGJ1tBK4h$ZJwp}v5{JeY-E%Abuc@klV6dvmy
z>OP~ReKRR>;Zd&8l@i59!s1zmX502ME%6l6xHP9vO?AqKZR?w+>V*W|n$veZS|TFH
zhvRG5yK75RxPxch<Puw|8K|<=n0Mvr=NmdIbadISAD7O~&dx6AZ0ziupw$z!L4KKW
zlagDv>B3_!r-h=uI~MltnD9gF-PXEho!s>YA`U4$@e#k+*VEUh^-3jB^;`Ac!|(ch
zOtXdgL}ME^?R4$tjL6sQ2{Ao%QZ`;LXxXIo`af+Y76vWbv$1O`m$Ycs=}(gv%$YZL
zURbAY&)HrOSi9!e)j5ABJbt@pTI-evH-G1xJ##4WNAa`6=VZ3pJ3EDBJa3ilJn=?n
zP4}HSn?+V;wCW08U!^5&bg}Hk@r3PHrfDYx${u|*MZ+s-(PKZ!t{ZO?-(+^{=Ng%4
ztyO=kp%J<$`kjBd_1;q^wmbj5c=G1;@6=VFe(ssZHdR5OajFA@_;MLr$?pY@{ficF
z(^3g@>gsq>5pDTs@#1xYQ42*^8PC~WyF2N-_2ieeooYenyVz}}?7erxCN#9}r)o!?
zQpG>x^+g}zcdU2LUMr=hbtG@e36CYuwyJwBN|>c{{MD260Pk}$Cnok?*%NvG!A91N
zC#F9;KKaV)qEov%MHePMta>bbafjs1-+_BJ-c>ubZ{L2a=+2)TcPyAM68yz@Y1xAJ
z_I7sP_i0S}nX%_yzWLPeKcCHyYPwLxUU~bPjO|_db=RNd`hNQ|F>w89zN>-WXQoP7
z9(1=Uviv0_bTQx}SI8EpX*I9UU+e1FA}+ge*V^~#TsOp)=as8+2TfVAJ)BL9Z^e1v
zy61WIdGkd)HJ;Z`skbt*y`{h6vFg4rt}{MY?z?B4{H}M0{LKx<%aa!E+xXP2*J80x
zk)@^OrIR}IL{^4&g{H?oU%GGU{MoZ)s_xlUtZn>cE!(HzGi9l3|4d`im3JP0J)+3r
zkmBGJH9I~#%ieS4>;CDNTD8=~%*+C28G0S6<GFaEV@2e<=aRx^KUUS>5ShTXIQP%%
znP(2aTBfshdr^n1d5Wj|tKBo7RzBbM`hE57C3l0YX0P4<YvS(bhDilsI;;t8KWD56
z?s_$q>r_N)H`nGD8v^#({^eTx`;|sQLh9+;Gv1yLEPF9kY5%<qJIuM)-i%V3vYIRC
z?h~${MSjtTl4l982#BAx)Tz~LDeLB6r@D2H>{_%*^>){}^%v@c?E|wn&Yh`s>;Him
zj{7%7IwDk08@x0!mgii!cb%^4bfY&jPj5CBdOB@ph(?;)($o6?W{DjWkr&)KyIuI&
z<6~K7IUzNdx%9VR^{l)6#o6Y`h6?$}i+khLTGy<3sdgsBSn248Uw%`z*hfA(ckwSP
z@7eP=_UC;s2>8B7KIlfxoBiRsVq(G~U$*@3-}U_T<05N*|J%>v<L}g8J2q{9LRb~g
z`KOlg?dP9qnPeyR-FW`j;?nf1U;c))^HnNdoxA$V`gQTaUzYnW@p>Pcar{@!eY1Vm
zee8_F%~z8p{@eaCN}c&A+Pr0tq+F5PgJ1XN_+NW_d9PZ`-Z@X!)Ow#<{Zywx;!OL6
zqam4GdnP<f-4c@O`Ylex_TKBr`f2kEV?O=(eKg7Yfjocx9{YOxj!2syKa19`|G~PV
zcS_PkZuWxxb>ADR@BU}FCoVSW!pyRUr?>3){H&XM@5aK9vRfR#eadZJ+sJNFJ-he7
zylEcQy`0CQPpk;6_*0zvF`(jD`_)G>?dy-3&)(k<b?V&dyM_T5yJf{cZwvZfecS2t
zzPlHlAAGF$7Jc@iOWu6;vG`9}w{G2!J!<N+%4@%Az?RC_wy9quBb;wYaec3QT;*H0
z&HCS$sy$EL`d?}MS+4x4(#G|3#b2hW`W-=k&+V_=b1&#;tz6@=wR_J@-Jro&T%u9V
z6||`MpG1U)`$5-7AAZ=xp8Ex={XWby^x9L)IAQX$;!?&tPkc&PXDnDC6LL<kVgIKq
zE)Uby(~tJ3_qiOjF}@jj^U=CRu4XY`*OjEJFK>_ibeUhUH$b3;dlS?8_3KZI{OoL;
zF@L^DwC~cBxtr?izF)ij_O5H^(MPu=XTA%Vvh<b6(wEs@LEDnmd~Vwn`@H3441d1D
zmFKL`$B1SAf))Q~aUH9D6Y*l-;crjA9r?fAsApyP3f7k10;P#ue;m`)AGy^jT7Uxn
zyku$M>CN*4IZxhfw5w_=*}TPs>Dqy&lJxX<E?;*ZoHu``*y=1s*L&|Si0^g0FJ*D>
z#v{G&SwdSjt@od5%YP>|W2vk7{p}`EmAAT9t;(vZs*m4wtA5wc&<AUz<iD0L*yrDW
z?%0XDr?1{md}8|fNZG$j*KVx1c$>>!@x-5N_c#5roAzj`?vACY*35YqU+XWsbtmK6
zoxiR*{DF6-y>*ujjk_JXQRQFe@ogXf>e(-SaqC{Elu5^pe%-5~pG3nApR8K9@84GY
zhi7j3zYE!Wx@O115C0Xt*`KtDMfZPkj1>CMrhfFV-1nIQTYv7gO9Cam1HFF_dbBq0
ziF;pHQoZ-*M^Fa+GqbGY>8*43zig?K`?2t2#U9r6J3n%#eB^Lq7Ge^3!u;S}`H`>t
ze*CbpQ?&nfL1S;-e8JjsyMuo(y>@+iGJjrJ>DTv`>AUn7JD*-45*vTNWb55uyKWs{
zK5N&$Aa8!thCS75G`-(!-MVF~=?(jj*WGsSc&2H)r8QyNeBWHA{k+>BBxi4&zGBvs
zIlh;dA6+zeQq}Y0k3apg<*4~?*YA3M`neX(SDCAuelJ++oZ7Vh-~uHjC8Y(gLaO{`
zeCp%Lmo$n=Ze8@nmfv<>?fDCjwg1$ae}2BHE#dQoqhFqR>u_vJ>x=naZ`}It(b_+*
zE{|Owf0)^#&S#;~m*nnZ5_wBX>5b&tR@r^KdiLJ`ZN6EXBS5CWPv&UrY|&T3)A!#x
z`RL*+4UJP>{nu|lm|1VEq?04Gc(2p6l+8t7m>qY|5v{Hg^g8%OaP>Sf$-}GOe!u+r
z*uwb>MYrBA1f6s{-LibDsj^k(%M}NsV*`{;4V3jtjvP60=9^lpP~4999SjZ!6_|`y
zXkXG1TC^%sOMQulPHTuszxHEYMIDV^4W(s1?saADks4YmSGrWzr2Lw)x+YQL^5Q?S
z6DRDdZ56to;UmS|Tl%Y~CH$vhetf2)$Ij292Y$(<o?q}kVWHjA)(O9RUfqh`Q@LT{
z@sPhu>Xw4C_P^!IKC@z0e$_mx73;9{r_9aC|7*O(zy0R^b@yJt+U`lt-CCmOXHNLv
z7vDee@ATh}Uqw&4+Fx3)nk@gRRasSE!RXVbM^iUiY;C{s=Icb~XF*fN%5=_tN>4wZ
z_Cw6c%crg=-hauIHS2qYxuvBws(6An%$swo#ci+0jD`O!&Rksl^2_bxQ@&iPH(C~&
zX6_o8v}enV6qU>Um%Ji_KodT5^0tS4m|*C2sH!-&yX?Q%I*%s{X1wuAo}`lH6De}C
z&Fz$%$GWhs+gf8{Hc5NRzc{sQ#*AsxXH0k5es}u&=;&|1r(V5ZWp(Z@)B9QNAFQ7|
zVe<LCH=b!i149*q`iECl!3zz|r)M!I?Nhh^@Z!a(Ua1R<=UB=g>MGg0OTJ;Lx0_~0
z-~LFs9npU8OQKHjb?013%i=Y>Ror5KR5Rd}2G7yo)+W{R2ln&d<yc!^DiV@ez5j}L
zj_o$bn||{jzN=^#b$<A6+Qns;|2{X9toi*^a!$s%{r=|zMWw6+-A#%tcY)g)T_IbX
zrW+OaWp!QIwk7u9v9*s2?{3!fm#uYbYRRZBNvf5+BD?R#w->*YI-5Dn56rg~pB4B>
zYxND@Y@zwPeB}!le>9WUd!GXCOR;usy#3km>hc?_`exkJ-{~hKqoe5`E+`XP^HcDu
zo^nWJq-)Z~m?x*_RxjD5_MDyV!W2WVL$#R`rXG8;l+$^W{_*@k6MOM5{}v>?OL%aq
z_wPTQjrCiVE8^WXA~jVG>n)7g+Owx-mFoWK@62C*S^V?+@Bj0j%F6tDzo`6`wd8s8
zRV~-{-SRwNIL~nVx~U6yIyea!zF>H6`RljSV)qP<Q$>nf8g-Y8T?kZJBvR}izal-7
z+uq$xuRD9QYbNXCtyYtoLjrqJs+JrR6R<j$<#x%hIAN-g70=&GUUjj&JbO7FslCYM
zHoyP3eD||GJGRBD`_}ESoOZ72!HSFLyX+-T{FiU!vbJuFiaO18Rq%S=^ZXr=zhhoJ
zy7sYl)7Eu9d@L){(tFx+o^hU)>)BO#+c0Lqf(2Tiq8Mk)ogu}g#khD;y2#q%Ls!Jw
zGm2IDEWJXaG&h@{k)2~Z>&RwdvF0;#pM6fxP+xx0VB^M3Eg!qWgrE2FbzEOrETVSK
z*yO_5I(2ognm1o==47ZJdueZ>eLDECeX4=+|Cv93{`~oi#m_u?b%ndc(nLF*>q&Q~
zip55rbSik`^YEmQsn4`yX(j1t&vU$<+S@0E9zA(+&fK|@J5TP?TKWFTFYVQnCnq=a
z-r6hZ6?JIJuK72;LIR!yG;lCDD{${wuvNokW{Bn#5q*CzuVq_nBDMF=Zkm}HaLQ}7
zglISGs-#CTtFu2Q-E(_s=+WBbv1vo*(GaGCd-Zfx_suLVcq0Bphjss|FpbnOZQYV}
z5n<CWXi9CJ-)Nh+BYX8TwtaaSff4?-r!5mx%)U)aI(v7Kxb&JITc^(ZXqBSD*i}}U
z80LSTOY-N>OXmaqyQH+VjlS#H)M=g1-+ue;&%atL|8DvBzJF=cq|4WI>Q+Zb8CQI`
z_BJ=|{{{<L)14dagtykkt4~?K|G4|V!}V_;pN;o^_cchy%B9_<tYqHYNw@3ohf4D&
zKV2*E>Adrzz+<7|+x|Bj{NuYIcl4gtoH(7DS9|2@kIz;&=xpunZ#i>#`C^x#=?8+Q
z2Y=S|>V6is^75>V3=MDRJM%<r?>${vGJk)=rXPRiZ`!1Kx&NsDk&vH%n6AJ4a%yUB
zT}TYa{Rbi)5gOh1=D+VR>)vL*_2`nlhxNXli~U(^-TLl_^`iE}4?c)JKKTgLLf!bN
zaL1vl`x!Oz^E^)doYyLT?^eRTng1;=1pPc7?|WPN>f=MYbxr~;3`T6-?`M38bDn9l
zFIZH_USi?rpKrULeE#6dzx?tG-Tliy#>WTxzhAY)Gto23w)5A|OFwn(INzKVdR(<@
z??3+Czx!Ig{jNOf*7ru^7rXMo-&<2(Jeqz~EB?gNZ*z9O`Bv5aYRM7_(>=N;ld9TH
z_sy!f`@U-4_NlKHU$i(ep~_=kXaD`n_Tl}n7H_EJ^s}G5YJFf|X!sg?G1c%V^LtNU
z__IIX>HV+0LMEBaCzIBc{Ww#a@y9f<%FO%E<u8XSB1ChS{F}1N=H7%N?vAHk%ZehZ
zC-1%YmRCHlP`J3cIU_(~X`+2xifZ9Qk%e4kVGeJF)H<6^sZEx+8m!gTm*o_*>CexJ
zd-NAi)7f(TPWs>a>iqA@e{Byca>$*}e|`7Ww<_C^`|sT+&6;&?*Xrk=e*7u@8M7|+
z-b+Q7wST{Ftlj$P$o={|tLLxRxYHM|Jvsf)u2}u=zrR<1`D+&%7$RVQs-bCtkrtQy
z<0AqpigrcpEB_v9{$p0VJSJTIx7S>ifJ>hLFaOKcn)%B<KCABWpD+Laq%8?9T_1m+
ztLU#@VODAD|MhQQE3!6fJ4&qoqyOP%$B+2?KJWd1vH#<n`rP>Q{S|X+?0@IX&}LkF
zdnxnk-9ICX*te~a>rP1jpL+AnKYtmM-{#lT$|Zi)6th*{)bi1+dHFm(?#25}a?H%4
z-3-6w^7W1#WZc5((0q*dT&$1K=e3bKPlcp7Ca?=w+sN^4X<nk+qu%%5oB8^@qjMK^
ze$|=2%<s&MdG{BeJiqgs^!=?;Pxsvl-ITa%wq{B%cU6^bciL6max<oD(pDxnPc04Z
zmUwcIk*Q#Q?fjhlbrY^|<XXBsSyaR+z&X#GWr7nE{}evK_SijU4o{wnUO7AYnt%Gw
zz56xQ1Lrm`cr&kHk%(@Q^~u>2Pgrt!Zcv%b$E*?HY$)^M#+OFX4+0&U7JPy&jE`si
z;rMg>j!*P60Z$W-1j)J47iM%FaA~-cX(6QOk|=!0&>}a{;4-IE;*@nO-2=@cJ5m|h
zpGsep@!u#S=x2IBYmyabir~pqCP|J9M=t6mM@h!BJZ4F=SD517$@x`SNOexxMJu_|
zV;qyXbWR?RZTVrORM2AK-+n`0;iA$W<-XM!Rx`pm5}z|KI54ejNkBuFK)1rm1sNrC
zN-RQ_JZ^jH8TE>B1<Rk0jEtU)lPCO~8onrK;Ur60r6b8b0uOgGG)!#}P(QLgw#{$H
zMVUEmegTJ!gibMXeku^q@NIGYBy!|+EZ@H42^xO>D_fGBofq&_Xlc!|Xjm=c8D`L@
zXr$usW5#jCEp~yN7ha#8ASZP4gP+Yq4i1ioE*2~dvw1IEkh5rw7Cqkiz&?zfd0nuE
z`@AZS!n=aEKKBb02`axo6yCC8^Qwmv_#LHgC@RZz3u(0dsMC&3=*s6vxT92XPf-2v
zqB51w*HnHo_59uX;kdTv$qUy54=Ji;_chuqa`w^bOiE%<ZjWX@l5g`uNW`kA?S^B_
zvKMlU3P!xV&H<adk|g(Cm$<NSgIguj1cmh`8m=Zx63H)~xUn58p8fblLD%QgRc{~O
zu&MU_@VxPY&RZMXd4GSsz4g$F=a7W}du*ylLZ(3F<_*mnw?1!}Gq0}0^hF!bEJgjO
zqVm_aQo)indVafSl)w5UA9tT+-|^tn8~25;cXxRbdwKafhkJI#hu7;ZcaJrn7%e4R
z+w<+>_tY&`4@Cs56q)!Y8Tct}&?sp>_-%*dGkf0k+Wjv1Yu|o-`t;cc8`Hg;*Rg$g
zv5jl-vuB>KE7yJ3cdDE#n6yPXN<>6WsD-KIxun3vlUn_DDaV^9aYi~Me%o<?`C^}p
z!nM}IDVGykdCp#!Epl3;AaR_vR>*XL;c<aE%7x1WIMxYYl#+V9!0N%)gz<a1rD
z+5Ibw*xcDS7pdn*TixE7_agA{oA~@==L{+zKi<7G@wNKK*<a0h(`O%=_2t&jJ(=t4
z^Y_V4d@BF5Un@y7>JY~Yt&dYVG!`<5s5r7JGb%W+P7sZ7;S^OC>~S(w5t*1M$k5FY
z(38R<s^H3TL`BRX;y^OXL=Q*S2@x7@2COYjDqS9K24Wsff*l-(G+Y@ym=s0WjF=2q
zCmFUR3wUfu7Sr(PY3UNu@K|w(ktH-BY=R(zvJj_8fS}ZcKoLy_LBU8yAs0b50Y;Yz
zk^&(d0wFC56C7J43`9gG2s%s%5m96ea$#k12y;+mWC)lLATS|NC6q;lQM4tHlT9!{
zB`~lhghi26h$kRGMaYm-l$FWR!AZcwVS<Wl=tLJ4QJ0Vw)`*^z0M?!$SJntaPl2Wh
zih`~wEGm&Af+>s%6Brp<1e**5LX|nVCP*-{sDuc%1h7tYWN7ISV01GGa#T=Zbm?Re
z;ozC1BjJ)L=E}gNp~I%+>gXxj$SJ}ks3K;dA`miBB*l@nMPiXai<_g0O0vX32@aJo
zg+)eA9$GpIief5m3?70C3d#y=niM#d8X5#ylb9t~I0Pd&nm7veGq_(n1n_@8F>m(Y
z#XGKgI-be3%lt1FBNw~&anw8WvzEM8S6`K0dgxfz{mYowxcOxE<c%{ASDrcZY<bLc
zz1QXK=8EN(zItWSZkvrB-^xo*Smsq$<dEsf;mj?}dQw40g;QaZLrdF225lzk-j=B?
zDhe@+T2oZm*?AavXC!Fna&J1cLZdVE0*6u8Og4p9M+qf^kQE(A3>p}9CUG(})E`&k
zE@ohHsXyJ+q*Q;9wL(z*x6(VIm;XC&)=y4w_;W+)&jm*N%>wR!@_(M(ab`{Dzdx?}
z3T>YlTCce5>R_KMV&|s)#Wg-6fq#)xkz0ekG=tra2Dt-{!vDhmTx(`zE^TNu5d6xt
zwS21h`nT`y`hNSh?C-4oIrSF2dil)FL0`LT0w*6yGKtCwdn`6PvG-E;!^4{zOW*we
zHF-<r(^}8X6~-4gu3hUjRarY&ZLimCdxOo!|6k|d|5a60Rrhb%y!m(kuhaio|F!U3
zNRW$TpwEp(9!d?JXFD7wZS>?a2oh-Z)>V}hbC|&8$s!{7UrT|lC&cH@m&^q(Ox9&=
z^=kWWxYo)p==xm#=)7k|X>0|R<{f%^?K~!$n~QF@?ucHtP9mJU+~~7wxvsg{+jH77
zmtHQL>BG|z80r!l=__4stX<$OnY3!hif_ey7kqj`TqAjccR4R(yrp6n;US!Nx7Ow8
z5`{|w8PA{px?jLNfpy`x16seXtPB$0tp28<;;+S{F1_8y^I}ZIGbiipdp`fpJAQHT
zf5&CYJdT;U&B_j28?`36>}^lp{eQO?ml*cGziV<i;b2qd9%cDE77zOZ_ZL>~o?M<T
zmAULgf61-3tHvg09`hKSRpz;T**M_T%@yln0t2`_MHx9+Tvb#=RX8*@u_$Z^5@@={
zq8PD&>x4&xvP#3E|5yIIHK}EUFE!3d+xhzZ1hX<OxdjC;T#_|?!k$IXTo)agahFRp
zlvV8Pj<ieLjb2?=k~CPJVpcMFX6~M1V?$%zGiMHLyPdUa$s~_PPn8xW!R#Z}=PUoO
z*;XR=x&MEJVf_}{l>e95gtz|NaesSt@BN<2Is1)|v};d2St)yDd-t7+{to7xK!tDB
z0Sa7fC+4SY6Am=l`y+Dm!A)0}ojd#b`kZ^$%kS;2{QCU;*9Y0JzvstwmZ|u&9$ma<
z>(|iG$W{8bp&6@Pqe7-GTbHzIdDz;htwCo!bvR!vU)b1q?cnRu4ZrTjtq)&$<;1I$
zuSQ!1UCy;MD0z4&EmTxku%@S>X+=v*(;_1wg)Si>B}FBVMGc-xS_@v+CpC1xe|_QP
ztl-s3nIcmrl`;otOtp1S@d^rA_c(P;fy;qpjnGK1V;l*6Jl?LE8H=TOSR*BmdEPwt
z@!Yw3$KwA#)Y5_{{<4UO@Z5IjPtdpT5$4yoZ{L01an5^*&z1c>RW0JWksD*0mitUu
z8Pxf$LtVJF()`b*<190cQs#JCKb>5&h;dIK%c0Ia0m5D~pZ2W$^O!;AQG@hS#%G^b
zg_wk_^f_R;)jGPZ>*~&_XFt_W`K?qtp{k&&)=Fx^mB!`Fy}Y(|uKr3zN<|90jNK{~
z+VwIWF805bl{|Oh!p=3VXZ%%sIcz2N`N~c8$nf4dfByX1+S=OrAkaL&ea5O4xswz)
zSvt5SI+Yj|Ji0tOJhT;*HVJV|;u2zNVf17Oa?)TDVmQI&!KJLgqok;?NaOR=4h<uX
zodpv^7Py45gftu!QDKo}Oc6|RRB%a9WDsLuIuPW^kkHUGL5a(y!KKSJjGNVsX+p?E
zk*<j@O%ohD9D=kA*pyk7jwDS~5EN8W2~cDaXy6HD$>d~YiDVI-;K1Y<!YL4bQpB}B
z=FYLx+5zWewsp=*)7==d)4N8k?Lp5?XHng4moD9T^DIxh+xz~a7Z)|p-Ih8!Q(Ac8
z!>G`$sm3x>wkeA`IvSk{I~{d3!qAeVtD{5dmFF53H=PKbF0F|jE=fnb1SLO4ixqbH
z-xGAuV*I#J+fIdPL!;6pE+GXEo*5UHnH|`&a>o4bAg%<5fVG^GjOGG@?>4&Z>1w>x
zwA_s&Pl=(TEy`tr$c_e)hVB!Y4uTU?LMF1+>WVl&Z<4nbc|CFBB%O#=8=p>eVw!x%
ziSf@iM*Arz*moUt+9Sz)h<%~sw@0gAAKOu7=*Plq?|o^duSTU(rpv|lwYHhB-M?O3
zxR^i9?)+W{nIFq)Vt&3me@?CDeDi#Ml~@jW5tbhdo*#eQA@#xCNpT_T&qc2fT$m+r
zVU~d13d@hh4-cMSR#qz9F23J5X<Z7tQJSOC1Ixmqho>8621s`-EELdF3klXzoXT?Q
z)B-ji?isvigq1ia3Z7_+b>yj7c&V|U<)x$HBZ&{ra$Ipo_6A7*U}#k8n8cyvp{3KY
z>O|L?o~DMLmWEX)Jd~I`JQNx>2`M!hS>3uC!mv!wOuFmT(ZxO2BrFaFM7gZ;<nS;{
z4qh(K9UAG<aYTUc%_P|wjK7YmY<$2yd-kkv7oHUUZ`l6I^c)ZKH!&IUZ@l}jU#ZG}
z!(mr@OtIR3-p@l74$K>wr`xv68O-dQ78kMU)2CzF-_M%s{cMZZS+v%o;Nj8PW;uzQ
z>&~6w4oJAaJxg|VkZ#o0y3mzHZ+&iUH_M+VoOX70V&~>J=MOe7oPP1tuG`Zi!=r9S
z-K?xEzkW0PI|%%Y&|+F<<H;Dv<m4)vq9PzT$;DMtMN~v(q66!M35-E*j;`*}QQa?H
z6o2gFbbozQLHz!_f9;2q4o|uKU>axW4@b>bD}MtE>8{18Sv`raAszuaApt7h9&49o
zPIA>aa+UM8hr~ypH2G%6ZH*5&<}vo|a_2L9CM$5bHK6Q*hQ|(*C{HFAm7cuCdpgfg
zU3k9TuK4kH(-$^B=G?H$$x|#iVn3nlzY(L-3B#|_@qP=JOqis?@podT(uc&aRab;x
zdWb%fx7GM_`B{2s`y0_0H9Z>s6PGeSO41Kf;1AOHBN)C?@RbAmN)|f{?vDqUSF$ia
z>EK)Q=RxPS4tpC>yNZ>o1^M=fSbY#*&*1QDLGxFa(gglE4cm&p6PVXCFt>_W1h_nO
z@%tn4J8AXdrq1h4>~S1(Gnl?6w1;VYO;9}9r5~nb>(P0oOJ+y+=`Nccou`|2Z3x}b
zBrd?9r=+$-^hy`Iky5deewCKY4d#1YE0?mAELb*0e4&W5o6?H|z7kC@4*EtcC|RJ^
z6O#8}Sr3E00_SR0>kUEQm|h(9=Ls=QXk%f#)HK7unT_GW0Urs?rJ})H9Epz1VHz1O
zvK-8vB8CpGJ|~qJSXw$XHVJWXc$`?Z=%hlIPD|4&mL`qGN*Zc=9No=Qw%$25;VkEz
zvy%71xs_J%9J;mWTF|M|DJR2DwN5cQ)naX{^2otdL5BZ``=cM(aZ;C9WIP)U)Xsm(
znab|k{zN*9@u!ORe?~oX)4rJCeFxufET0hnK>6$6`q)?x^~?4R^FxAv_`7v~TK4$y
zs>xf9{z#I0%^xLSEqr_4ZT{`Iwq5)7ch`h~29}Of3WDy!%qA0EH|IzgROWH`v|RSp
z;r#Iu)M7p1)6n9;#`$D!<@Gt4*W05?cimYR9<}q-T)*W(i)UI+xH4;^$%72Pdb>IQ
zYI*+ria9+a=CED*xt_EAZfQQ}rdmZ#IWu$m&Y7j_->;llUth<h!DeEa;^5@Km=ZFP
zk+n<ERZ`Stf=G~nNF$@mBte!QH>M>`N<E3nLQDshCMhd&G;*?OxCWGdeZiEUu&B+=
z*ws<@vdr-;)+m=%yqX%(qOX?9S0yfKd9rB&U*D<&`#vV}ZM@99CdzMP!Gj#0S4CI!
zoY+KiRBpXJ<g+Ts(dw>IPRm6Ond5EGT4wPmOj)RS-P?h$$9U$HvkHC_7$476R7+Ll
zGElsdD3CI-lW}&c!({bZ14ZLrPOXO?5)G1)YCIEMCQ0TfYMfN)QR&RQt8(IKqhXf^
z$020}B?b*AO$D7ujvkeUZV6YBiD4cF9?Fa?N=F+7S)6nt6}l7-d9b83Etq@bQP{h!
z3Q0#;_cfVK6kaU&;vip-@M1>361kG_Tt?x3;Zhbp6;}DB&3l^kFEWN#3iwY_*nhxL
z_)Nlr>JRKc6gfDb9bvxT!M>CA+l^J3hRmK_?MoR-W-d7(?%DKWMw*^}{*FMENq*ke
zWp8h#ot<qRw$>~+>1em;^$nknc60CF7+dxAQS8RTiq{Xe=iTFMndGT*@%mzS#*md!
z%hQd{&XLf4_WXJD_Js=+1Q@t7859}XljGyoW+*gISIgf!<B3qnW!+O;S65B)H1@d`
zb#Hp($qJoY502b2@jA|{xy<nNQXgLKrlb{XqO`L%t%|s0z^8XEFWXN2_r=+NSn4#M
z1o%Htj<l(ddNO_c){yNRmOYx%`B0IwGSG2Pb7iRNUBM5%cW=hM+k5-jg|*C!tvCbS
ze~7WWc3*6cQa;dhNwb*oA4_ZOG@0w)wY^sS+B|FGGp~yiBwYCKsK4{@PunWV<S?;z
z{i5{?*QnaLzss2Sh5cURwT{vS@+{5|n!YZQ=MlNqRhszg@LGY_?rX)P)Z_HyI7%OU
zeek!S;sWz}@%1dP4}LuUwxckqFyUj;-U`j5oktswwu><oCe4ko$&dlVj|)nZmiOuR
zsVi|_ZK!xqw5Vv&oE3Xk*k-UtDW2}v<G0g~;=I;Sy6nj7BRhM~ymoluxU-{l@rA}~
z9bXf^Chhf5E*E;Iwj|bL)<owZsfoUp%Xux@=9D~>IXtVNWWh^!F^kP!7B3pg5*D@h
zCpE2**FR}BUn726&GJ*ay3vV=hYj95^RCPEleCtR*>>k_w9lEM?S5(I-cv<fW6s>1
zIj8T<jhJSko>wO_O77_{n6f3)X^O_8S<_Tc1{^reb5NkgLv@K%n83LvF{PQt^77JL
zTh|yYXm3%F6IDCMP!S=n^X%K*qn;1NiZ9QbC;0qY$Y%wa%H#dgt*&(|<U{0Fv1c}a
zJ^1AT`?>Q^61ev$G2Y~Bc`YEDyj|SEcQ)sW$){7B1tK_HS7~hjID64U8NNxLYT`AE
zx-=Heo8$6PX#%H!f=7e6q~Ik_%{U{Yr>CZBg)Bb*>AC`&1INkBGt3&7&GgxCu<6y+
z)sI6JdoG>Ux3}$OHi<Cz_xHJ6%5&w4ad-raf&hn^nZ<3<se$Kg*)5FcvawaqyS`S{
zH7zt$``Xmip$k5|mRs>v^Fl)8-n=_!9-rAc-B>QhQtH{c%%w{oeTvI0-FdR@rSG=e
z(-OFk>pU%1OYBUO)tff=%^p4(@yR?h5A_y{EiKq8boOSD-m3>w5>w5-rwL9trpel{
zX7cxBj`}*|D~%uZLQbB@oM@!Hr76kBBk?t>NI^m=W8dG-UHz3^=5M@@@oiuUdd7J>
zNp_o->JhmZ28)k@S2^}g;a|u3<v@+&=RaTN_8+<M{r>hFuh*@!bK0q;Iepr^xTu^<
zo{L>(ZS_={W05n-Pr#8q!^YzIvHf*vH`|=`6tkzrG@Xt))HNqC_|(F@UMaJJsI6H`
zXPXu~I5BV>VQAn`U{O$Voa7qeBBWrbq+q0Bq^-!>%qnv2+O?IU&iv}UGu}B)x^ivR
zRH41WcjtyaR|`v9zu&C<aPjKg8EJ|6p@+^GuZ}KCn0e~<jEGM;hh|+e)qU&b;-;jv
z$#?ErQI`&_YfVb4PHf0n!Eub^M8|4rl{M3PdU}+F#FPR#Bn4bPG%{MdxTrWBOj_A+
zK&xYEN9!B6McmaaaxL;4u5vjJ&(3pl7|P`YA2O8ZIKX#X!0)zbwGaCN0r}4y?JWw+
z3nnD>i*d~H;C^=Wk%L7~#YF#23hYesA@K*B+K=*g@qb_u-fX8KJ1slAI_|^%oNM1R
zYVY|ScZ_q5*18)SV^Oet)<OnFH3tr%iplpnm@mogeP1FkQ^LF_p{Y!PON8ZmN1?(j
zLtZ(vc6sOi1mg$kpNc<!{bD)I*-7*4tPjr)&pvzc;>Dd8H?j5DOkT_@k#XRquY{D9
zm6eo)gp`DZh2_s5l@=XZv;O&ZwFum431M?}+BxH?<QexXoF9GK9z6}6w(`1p#nUr$
zwilk+VYlS2`K-5Hm(x-XvpoyCv(URRBGTb`cp9%nq}Qbz9Y<%a>C)r6G)+wJSrSLe
z+qMNgCQCKiPI;{s5o}H8n3OHbqhkGgo4a_7(>CU$U2#YDG|4XQs^by3o-n_WujPgF
z+--lPn&-?q@OfTo@~k8Gjz{QehznQ-SbSh^W%#W$N1}h_%|-JV75>!oTI<+Js`MYU
zx|<)rRzgAKl&4;|$|O&<lPV8&OZHrMu~PV6(ZF|tadyYo*Vp$7^Bk%BEg)}Kd!fer
zljjXB@ks&;g)g*us(f)W<em`7a(ilC)IY1INq!0u-g@mWi?|fd_8mKx`z-feUhcm<
zHnu#zygZ2oca#|wMG_fW*aTSrJX(;GlYNCr&BHO!e1nUf%@5{>^A=x~eAQztzIAom
z(=B>lA;*$)19GA)Zs#P01`DkT&E59K*~p6Hq0hF9+sxjs^9qU3T6Q~k?@zs1tHg|V
zGfhb1sp{=G;<2j9qe!D=;*U$tkDBC#>YI%B9o*jZ%)wY@xwg{Lf}01d&+v0pI5qaN
zyf<{X>mcKyCoO87KhgZFhPaE~qP0g4`yc30YUw_eRLA3Bu|&Y4@ZsUMU-z~;=uF{E
zS;U|v>N8PO;D8s41{ceGo7tNdyql`CUNBhA=XvFvpZ`?M=NcA1K00sB(bMyCPgqRI
zb9&*_+QB2hBOtDnq_a)wbds);%e-!uz{t3!HGXpzz52;8E3-s+W$*&d2{R>^q&-{G
zrSapJ3WNO8rAsefP<CjrY0RHjcT8`Y*)p?vdh^W8+?+bpC!c(>!hyY*(U-&W@q$A&
zm!HiyY&YiF_Mx)xj^6dW)(dA$WczwT1<%H2&zw2)^2%KWE0^+2ySF)J?uxjzd(Ig5
zv3brg@0DCEpc|e0D6LpKDtgo1%SFMFhLd*e&f_^!^rS!}sY@p}#I<eVWVIbpH#V|(
zwW|JnoGjTIc73|h7Dtw;U0rK*)D*sLI+En*<i>Qf>BNo=W+H_ebvcLoSeTg-T=pez
zFo-?E_|j2cl0$`W3SW<KvA~~*JO$1@b2_HFTD)QWsKeZKl<T7Kt7ebd0Kc0m4!;yc
z7gT-_e%yUarTKx1QcdBH=Q)$)l^K5A(0qURenOeGY~Qil?_OVCZ}s8zxjl6frOOY?
zZK=9or%_X29lZL}g`bN**WC5$pTE+7QI7la%gc`ZdU&Ouz4^+kFpF3(sn0=D6HFf-
zn>F!D_wl}-C+==VX+0*DB^PH|S()n9RDO7V=FhJZ<qI=6$H%=|^xD{H@lv@dFP)c`
z&kU~QIq~$cp7^JK;tv>R{rdH_sJ4rX=|)%k{ogmgFW+x_<M-$LMzM!(&wWw;tleC9
z(`h5G?^^<j)^-F*ty!D9>4MMcpj!q@U6YhncVF`{4?CTd$GO<tXzMz$8M@P^_Qe!#
z)z(_N?b52CNk;{oLl17dIdkgHzFQh^gkDTZIMnr&J6h|e*Mw`2KBWp8>UeLNu`O}3
zsEE@sE~bMsj#>m}Na}pQ<X<`YyDJ;xyWjiXy4lF?775;VFnY$@1$RzQ>0mdKIl)}G
zYXMi`BKczBLnk)KeyR8vZ`m(am|){1q<y%9|ARn<VTD2XT1D0`585vrVJJ{ya8PUE
za8P(N{r21Gcl5=@#YOf$<}o<TqqC_??Q-j#bx+T3eY>YJ_tlfhB9mCN#8gxtzMU*E
zht;hoq~ghg+j=oOD>&5(RnBj9y4n?{ldyT})S3mHYu7q5`sZ<|>8UZd2}!JT^i1mI
zuQIscyOmW;VqJ%7NUCPhj5`NIR(m@}tXjLOSwpAnV#10k37bk<R1<H_y~)4l?~6n7
z>wa8lGA}-&Qx@p_E-B`ycSYyv4*6_`ijDkRy9$-s-xxB#^iY2@U;V^Xj(u;+Kg8#$
zZ{>AkeC>aNLv9Q6wXgGL2HPuY{Ict7Pu^H_m|JI3VW-;9NLLZR*l*>hR3=4Fez-mF
z@29wZDziT%ObK*#Rhbk%DbQ8O=jP_MHxFiAwGrEpuzL6Ge5<Q#dAXx3Otx&<vZaNI
zk;Ac>>Gjt=sR9Ozm@ESXPo61uwp0zaT&WSJkZ$soZ_cBKTT5;|yv-}i+iFy#ecdwm
z>9)Hwisw9g_a#QAXG7-XsZl{!EV4SU26Io{-gRcv?SjLd-J+o<Lxh#3MJB#YY;5Re
z-pgQTp_UU}rWwWO)8@-yl;#s~ykW<LYaF{X9vw7HQEJ@Y)O#hokgr92KJ$D5ITz<n
zIb~5BB|$&=1sBBM^tY(TcG&2NvOaXN+rL)B{<KK^EG>Cvk;`BI*4o<I+Na-q|NZa&
z{N<Ogf8RCNE$61**^&T97J=B7NuKI2RT(u50ux07?)Fta_nK08S&hFa=jIx}xz^rZ
z&z^fPT)KGu(xsJ)L7;hYOPA*Q_3Oj-?PX;sR#2)j`m_02*XC#QR!--im1q|{W4q{?
zXoEL7HxKVutt?y0-X_-Wb~GwC_pq_Ko<&A_xays<UW=vud@fxQs|9<)jxCsGa5OZ7
zW1)_l*~JMNhc>vi2)QV(Q)J;evF);v<iu+X4H9n*ANCv;tJ5*NDE5zYm4Vp@=@*@)
z!bb#!7t20zuq=#w%w8}0&d8uq#kpJL(Lwe^1;)?6<nqr)xc^|{KQ8_#=}%yMkmh@b
z?+bn}nea!VTDER~{Q7s*^Pca#`n~q9e|7Nv?_a-uIcM?gxrODy<1&Tk!)vO)p11jY
zt|KLw$H=lxQnNBmBC$<uuQHEe)m~j$S=s8>uV26U^X3f?4-XFut8Z==M^lo6!->K(
zrLOT!`+fwqNykh}oc8<C9iKxR1x~!|-OqByXUe@cCF|C0Q|@TbcvftlXKpIJ+q>lE
z$&EL1G`<zaan4Lry7nq{+hzmp8P8awB4<l)JTu4b;gPdv&Kwr|{@ij;*OkQ**EI_c
zx-|7}IkY2g+2=WY;=x>|CnsHIG%(0np>;edP-g4eX}2>HXU<SLb2976rYQ%xdQT;6
zyBW#kuyVqJBPToFsQ7fKu5mol@z6)ZQ%Sx$+rj%q+=1vF>uxEOB+ZrhqEMo+-H@+F
zu0-~Y`%H%4(+oOl&$u5Fc;;yFhR;~w_l!MZa)KP~VhVGdjO3QYbFeUeXpv7+{CvPh
zX#)SuxozB^4q7-kv`D%zq{>d6R(QPNgQ?B8x94q79=Ec*$9lo>7l$Gnm%u`WJuXR$
zB}I}K^~}_G_h6=>NWi>W<`;pBYL#Ekt4la;x8lrM<6_&!7cX9{%Um}5{-xs%jMuMU
zzkY2U*W%@(!Vc$EX7MUyvQ7MQMpww-SCXXQ+gY0#wYV217HzwdBOaGD&Gck!cbZ6Y
zs7RWT=rkqg=2ZO>zWWd6H>EItP-z!_-h3}fZH4Ovl?xfq9m*!0>72=G_+Xg|?-7Q)
z+nd^36!NP=gqJez>TmIzb|^&SpP|hUeJh6IMeSzVC%a!c%E<_~3)p5bFkEPBF%~mZ
zXcm`lmyIbcDhdw`y}3DUvupV0@SK~e(<0NewZ4Xi1iFd@sd#j0ifK(;>=tB|xopbL
zpo!C34)$cvDV0(-xh!t_v4v4({ldnWa|h2JbQHOM{rXCVursViJ(E0*Vy3ICSv$30
zY3t$-Q~X=>d?h+&h6Wugb6R!eKw{V7T^UD?aD^RFGdLr%u`$WOqCxKUF$eoihsr(7
z4J$ubethm)IWf8;#xPG$L-mNm6UTA^<`3y?1xfu^4%!4Uit_L5%75@@gPnzZvG@TN
zh93!)pAYEpGaP>0RVLro@q?40@c~CoPJCPq4}X(9GxLLX|KPiqFaIt10s>q1$Im~$
ze3{`s>F<vl3=hfIsd!9Ok`&o=T4SY9?y2o{={%S6Cm#t_$`tZTG=Ftf+w9l3Z{N+$
z{{8#+@4vCJF*`f^-qy#DR<~)aT)TGt+iM~rXBt(0JYQ84{AluHtt~&ETddlnnZ7D>
zGt-N^XY<n3j+-t`la%Sa)$a88a9?y=VrP<XXlBzLnG<WTCT+Xz=Qd4YxuRQ8$om-`
zeL_u|S|Ng5L5DU<Y+BXKbyOsR$4xDV)$l=^ijR@U+Kw3pF&yF_eFGJ&7Bl|7!EW1-
zm&7l{cr-!!n7)uk!Ho|KDw^22DknTNeCXcFYB!_(vcLmZyNLNZ_0!z!PYC$;HqDzO
z9(?86%U4-hS=CYV?r&ebxLKpJGHsXP!ZwNIW>Do>^W!j&!QN)Bu8^d~qH?xBg_Mu?
z`R+8zxpDI9vtnz@n0Mvg^CwQ67%eX^FZb4w`RJpM72C9%7EX9#(W~jz#4@i$k@xQT
zz3RJ#wB~(uzH2O!%B?FJ8qpJ)rZv&5Wzk}0*U3URQ?)r3DI9q>VNp+q(m@H1MMYDR
z1m}8rCOO|-qWC?;P`RD`C+DAT>G~%b@h;Lk?ry9KwEwnn*8*9Ny&U2?;Rg)tCw8Ch
zJlnBcN%?n=pv?*UNKXEhU8N^H6kpWxGp*ar{@i2XniHkB*kijt9<h4EeEo-pZ5e<6
ztydMl_W5n!zxG4vKe_$!kAoWzty;BudzQxH=jSK;nmll5m=rMSz^uZ>ZvIR&MDB-k
zU4Fh|*B))1O{edii?(ljdRFz_!z-6}uPl0cbanWeRoiq{rh11pFgP%%NgQrpq{yI^
z>FRRj@)h<~vZf0q)kAxQL<JXmd2BnKnd};r7If!y^t9+n(-P9$g1${SbY||%9&xS}
zDX9W$riSg&6FE3j(ps(i)@Ri{J0#3z7$kZBNI2Qmr4wPuI<be1{lffTDIC$Rm!~vj
z_`J|_b7eGCUc=)4H7W7z5n;Z>qxp~G4@&P~d%yUi*Mz*po!#dff;0*jAN7z43wsb?
zFBZyZcZ?yp(edX4dzJTw!Ven0C%*sng?;nQH*a<u-<mr=`*h&VxijCLJ7aD5{rNP`
zry2!Q{De9zxE^Tu&x??ekzuPmT5x4WIzMl+MfNp=sa+4W8x}=u+`oT+{KjK%LcbUO
z_*$8Gqw&wVGnI8wwSUeX{~Y;y`Ca~9wVM}T{B$MO@x{T{=k`?Hi$2x2J$iNM_n^ir
zpB`MXFmyl29%MD~NZ*p9&E4#NRt1HP0X6&L^`~u6>aFOI{JAUgT-Be2o^~bd)3=GZ
zG<n}_yvcBVH~&mo4&}M*rV42X<?l8<_P^WMFH@o1E@RRr<FL0}gz0PdzgI2Z$Ls={
z%Z|jQFg!Vbvn8SKg8M|~i6<7wbTGbf`O~R<r2BA}>_-kK|4qG;Djba)RQ%Uz&9`BY
z;b^oFWD<*HP&gJY8(LHub~@!`YG|O_D~p1MhicyKW%t`Cq^L6Kh0@<o{PQbpS`-x8
zU0DQPC{%vpvw4~BGo#${^Iy)-&*v5XvwXt%gw?9{-n6NAyLT@AYx3p&>-+z1RajYB
zR8(16+1Lay%D?<;zx>OZ1#J<IA>{&rXD_Usm(f-+O>vQcf}6`yCg(keS>8;!=^3$T
zLgc-326is$2L*~0Bn1Vwbl*<OQ;?Os=n;LOrjhlT!!yTcj-Ob#I}}eeeUvCvShwTg
zb%}*v3_mYg_{r5ytx7<Alc!+q4)uldsq@Y--&8Q{5VCuo%EJFn{!E*l6^BCQbD6@k
zZrSxQ$N6{)67}X<ZO?stI&7`|)~qWUVFlN_<Mvd2-St4ogJEsd)*R!shijv@^7>n_
zNO&+LG#qkBRN(u<Gb8S>8t)VC6LAiE_c}V>xN+mg8jXdCEtB6@Z+`#x&-Z2RyEWbK
zb+>j-oU+m-tBd>A48x;>(~=Fc9o5s=Y%_I^um~nByg%vO3*VlpiV@sQ%Dp@p5*K0%
z3uZ|~JrXHM5dR$G%+$EE!+hhOro2V!_l{oW5D$1U!AjtT`<bS8)vorreCk3SA7<FK
z#B&%tIPl=0yqEe#{>3jAeU;$9DDd^+6Avl5^PAs12Z7)BzVChi`{x&)|7|lrKRTat
zW1CIU9kHj)JjN&4c<$EkW@32IC!)B)-XVo2IpOirIb{!@pPyHF#n0kZ$&E9%EdtKg
z*4EaEvTQ=0B7v?Vfo^9faxgeBCAL{HFfp>&&p09c;^)5HU+Ne6mq=;LFH`6Z-Lkcb
zi`DIBwK-eE+z*QutrBf7S){=&#^Yhg+Fmk6h>7{j1$QUK`7EcJr?RUtPGm5f$iKHu
z$Ni^d*AX8H#y>5NcaHFtIL~FD%lJ&ia^gRMj^%3VA6Vke6n}4Wkzb|%=HEkCyBX~7
z1mA7^Vf(7?k%R5`tBdBX+x~Fvt#9@G*&n`r-Fxo%?zQjBs%5)pNF`j)xw&uOzTK-=
z?eSHa<f$XD(8<?gqHx!wz{d|fRVGgB(pc2^BFK4hY3cIi%a>NqpI<+}asK?)*4Fv+
z+c=unZd||q-E~oi^G*CGKh@04KT(skU*#u{r<;MtM8>Yp>nRJFSi&w@9W=hAA!+bu
z24~;Vqe+i8e@d|SNIGz|yKAD4OIL!0$brTN&wn>e0uH=4k>OZaEgYfHdPQ%6EWhsp
z8II<H<I_cIBMP5qJ`V0XU~xhGR_CFl`6<lr68cUEG#=<+RGmC^$|O&fNxnO7FPoJ&
z+4u30dwVKAny*|oheN<(!SbU=c_Xq`t;slbBGfWyvR_i>`{GrLrerLc6!LPZzoFLj
z%bq$<Hh;e8%3^ZCIa+%8<rAWe*REY#;p(iWB~^LzmfwslGflP0QQM-6joxMMY&ALT
zZT32M=hSV_^iGLvJDQ&NOQI}AW8qE3yOaDUPWUm2%|=+Z$zsFNghvirn<RLc{wd@p
z>|<ztAnmxQc=5S8FS-gI?&<n=gPpDH=AUjByZ!6hFSFV=G}IJ7>VG5r@JB|CJ@bc<
zvKhZO{;v9U@A$2CJguSo7OQz_T$Wv=vDJ0sj{y5Ubs_?>5&{Yi4VBO56dva-JSB15
zpt#`N9K#C>tkfhDd!%RO74Ce)+pDXqtN8L|e?Z!6L;a*gi57+=0S5iGERG&^vxFr(
zb*Id~>)mtchU!uUCw)g&my;4IT?-7Av?jSW>2ypK<7sGEl+JrU`@-K1=YHOO{CREO
zuUQcpc~{$SpI@cB`d@rtkwvlE?f6q)cg@?jBB_ccOZcU5aA$w;;u{X9Re$ZhyknP7
z<=h3crzw2OsAdf|?%g@@RR2avoh7Lo@=V(fdHO50Nl%zk!Fyxlu6Zw3sTQSZ`#tT|
zlnUJ~xP)bul4EwV)SWA*%N6Eq)MV(r)AVC{i=zDDgpke7=^Nx*I+!>E1U2U}KT=zA
zG{DQcWd&!AtLlUe-Y$L{XB`&SO-gFzKETUuTbR0xd9H@Oc$~G&&JRCd2PTW}R0|e9
z^zy>`1;!pLAHUGsE&M&zudeme8<9VK%C>Kx?9!AqKdQ0Fy#7W?Mb_dryKBss61z^a
zdau*E)$Fojg$Un{tq}?0+gNP$mj&PI7H{b|;ImEo_IK6dwcNWxzF#u(*c6vluyc00
zPsH+7-P40-Dm^wRRAHE@{VKrdx{HlN&6>1nyv>b5(~1*3IVVJTe^_wl0i)O#sfi^s
z0<TA=o^Q@eW)}<o!n3A(!qi|fFIF|lJjHn{eeF&+98xsotlINJ->*_*)m<kBC*c$7
z5v7M;@F+HOO%yOM(ce|=(7bK;im<OGw{*A7-@Z_7uZTfGdG)+<n+(2f`@P#woX*ty
zSGKdQbX_mo8%du>#f4FE8Qr2n&z@cozkXOGk*Qfxp<$9`bmn)dd+HvC4ZnrusY~7#
z$em~PPe1ENLE^`x3gM=}wnJed2i5ByzhG6hQRz|X%BXa-6!x81b@1!2e_k8c+&*{y
z{L{~oMw1s?r3Q+W{Qv#rN#WAOW0U{gj184}$UMn%v8M4&{$#(Ik1VeV+>`0ezC8b7
z%9%^a*-dAbm$IwIw>gwaxS2>v{Csj*K>zU+ccl;a?K*-&!d^^Uvcz!urm|PR?Pq;i
zUOc~L&%OxlnM-#cJIE_#S|9AKtmyT%lB4ie)a>1T(O<s*w2Jr?XL=%hm(s=c3npqm
zNss*?(4z3nKGb->V#x0swcle!80SQLvFqmD3llr9E|TJ>lKw+I@q6NZUCG)XuXeeH
zN-*dBnRblf_WMVtJujW&tGMjCd+QdpNgK{GddTd!siWF&v-pY>Q}2Su`wmH;xTR<|
z%{g!LU9;ZVDZVN0M-Jvo&)Rsl#?Jb+ms`J{vz*$Dvy!<g@3lO=i=&gX@|K9ieQIDX
zm}8u`Wy<{Ewdy<Xy`A}e@t#XUFRzIn=MX!X+@L3Oe9oQES~d?EO=m@xd?{kId-pT6
zjD@db|5=@Vvu7(sWSJU!bVjK%r8EZbKlbv(e#ewQa}NjkKU)#x`ykNsv`vP<W*^pV
zOq(BYE2uE-m3he0!0Qz@i6dMoNaI2StCIAAoW>9Ny4#nqct38*db)qIrh?CezYGNd
z7j|k+V85?W=y{^gb(Jft<ei*nTz8!gbv-_;&G_h)TE&VZrA!%(ho=?Fn^_#=xP0LA
z0)}TjxgDabEB~$KJ?r<sp(;!x@+iXzXZ5R^OgCCJ_EqKtr~JAaFzM16_cSq$-i{i!
z2M)Jx<(h3{W@y;P&=jU5?ZxA`U3u%xZO=EqyrtX59nZ6Um7wkk=Iq#S-~HF`-)uf<
z<(F#<BNMlL^IF5g#F+S*(dV(VbB}51sZ`^#x23xcd=o_VI2IhR$YIcN=3{13Wx97u
zbF<mDX>w(kzRb>wZY?f+eq+@Q#n+l)C0lKMXIQ*Dn2>a%LDQR|!NBP@=P{w3GMp`<
zhvt}B8p*JzS6zB8=kRIY!PaI8|6IEZ>cuCX-}tlP&mr~aCv7%mD>ckI{7OFWLHl)m
zW%ZvEHubQ+Zc|qfnjbv(W`_*h{9Nk|@`s8%D&`hR$j`Hyop+w$hQBduqG^lCp-n<R
zDtB6!y<XxtYu0qdGNW(Y9HoaI_XsRppgcw8nDDa))07!GM6xVHOdbV9F)2t+->S<X
zB66HZfPYWeM*r+SA>qbYQUBzNO@~DeoCs)M*QqD{?Z;=?qwH5yy_lXvDeaJ{sM$Jc
zisqCFM;0<M1Ts$W4*S`m%J7mkUqF#LwD3gZ3DHT4r}-L`PVsL%B^Gc=gyoB1OHy{4
zNI=jj2_^o01q%vZOG!!?oH}spMW7j{gYYXEb`hJVmR6O`{c174)w+T<3cFll`~B*`
zrOBd-XDt1l=3P8I>vw9gSHH)CiGl~31b3)dalT;P);js&=_|+128yM;+8iU1X_p}B
z;OD-HO~C)S!X2G5n-{yb$qF44S>UcwV6jp3IEP)%oRtp8<)gwonUADPu4c66yvgU(
zDLcWz!H&b?f%qiB&Lg$I3K`T4d}rPK6eN@+E3I|-VV@U2yV43}4hMgGr<*Fu`|_;x
z`F1qfDtNN!1|K*x!{LdwoQT6J=2&5wMRmps!m6?%GY+x4MhNlv@)}C`?Q#iNWVgUf
zmpjR)gY)@f;Uy^#Cm0#=#b_*ManMW_-RvaT&2*jNfL;QBvY0|%L)cPd^UQ1w4Gm3=
zsEZaEffD&;Wo2<KcW#;;`+4`B*~6^9uij$oXUh9_h%~yF@ST(Cy51Zg5LF;zBw_G@
zEvcyTl#x>5Nxc9D#jb^V&aDap0vs0_C6)>uY;b7oJit7CN{8A(s{pQIqoSBC?yAg<
zEJ{VUeRzafE=&|yHu;<rgFwImRq3{;g4dhY$e4IczHRk8ZYw{3@r`-c-EAjtPgIlL
zZ)WHCL^kDsMhHvmx;y@{Qcsh!-yBlR-uKnJ_HLv3<if4nwl8PwvvqmIsbD0X?$9R3
zsA_fX(w@&9*9E*;mU=y9QFuJTGVjgWy_&v<xpLMletzg+=4;<cZ`aN}{49#M%(U|M
z{^Yv6Lr14wnqqcg`(4I&*Y1h&o(u@EdGNXEiuu~~YrL1!?(P#_Z(`lH;9s!Fd-LP-
zGb|k5o!?e+e&aiqc_pgLm;KhWQ2kQ7DA`D^FWAkl{(?Vq`?U{mZ-^edbd2%DY8MGP
z2Au;p-dkwO3P&jJXW(I%xzpp{kjC8g@ajby)vhWjhc~YqbS0D;EL3(pwP{>v%(C%q
zzQ?)*E{=Immp`wKR$5WJP;KGSiZ=c04{sJvncQg0zD(4*_k}U5{p1UNEN^eTovWPp
zZrbg=d)rpLUAscmY_`(vRNsGJHof(q@Q6!aN@J3SkisMlRoi|8?dq_rt8?zBZOof@
zl-u{S(~+F(_b!XvJlVq=$Y}S-E2L0{;p9QHXh)5tNYfuzCfu)@`8_+<NLf8gOIvnT
zhg|LFnjpc+GDTfJk;Z{BlY<<m3a7EE9co#yd}52Mn8dQol?tqtUfHKV#}r(MKegzE
z9BW{i_3jPqy$40=)kMxZ&Qp)==lk&frElcZ3l$Y_-aqAx+&z!6_>|7qnGv$5FTT!8
z&nSz%eB|xz4eUKOmDAjBEbaW|`qpZziIB1=7pLd678liO1LN(s#s_WunezIlT9jYj
z$)RHGBkaEU<peV!w)Xth=T1((;Ie-0gTx1N2BkkP1u7pgXW^1pU{IKLqlc$MM&Utr
z3bTTeq-Kj}+`&{c*V>Ca*KTQh{c6ii>D+r~Zp|(dY~0$BCwyjo){}N8_8VXKR^M#;
z6?y+{tog<J2ZX2j7oVwh3clPT>Am78+rBfu=1CsDv3Xm<Bc{XGzwWYJ&L=mYVeX!D
zdp=ZGCO+QuvijV;x*ZbfmQ#N%zE(Nk;cQLc`4%<7GskzU)yuxU9amAjL-=OOq|)>L
zb?@A^-`cocU-gW9-{U4fUV-P=KhI0f-F&lpcQ$X;QM2d@hnwE9n-g>uZt7~V^Q^fL
zXQTbQ?5AAudVv~?)K71jfBikzU@&pQ7VSNE6@67b&Pe`>KlZMwTF&F!_Rs0X<?gpM
z((Y+<Ow#6E<lv$n(s3e)sq&PRNQG&_To1({$-@;k0XElXt=n(V`^w!b<VmCf1A{u3
zN<@fYi@983)J>L@C5at}3MMR4Qg-5UTP%?#BrsX?nvH+AMV`0tTBjl<?*_x9DSV0+
zFI$T^mtAt6sQ8pYqN;USZ&#?`o%s@{L)@Du-1G4Gx+m(U-~wUhYuk@(*}9-A`m&6}
z)@O^fBi1q1v7~$uVSKUsQ@921Vy9-U<~VP&@~csA_ncjuo?^pM`O(W|@=YUW$DlXu
zN|VpI@p+Z@uU5IV(ZlMxbHjl?3Dwz07PQ}aI9;o?H<hWa+~!}s<aZ<XHO~9P=NDS1
z-g?_6W`29O7?;SxEm}GGydR&>Uae%UEgkzc?CY&F492I;J)eKGElaTd?a#As%C|o@
zmt>DM2AO$Xesw7@>7e1}d0Rf%Zu};ne<sfI`NKOOf+t^J^7ZD<XN%;{PrJ%@|A__r
z`Wv6m_4zaVGbE;P=`l)5__{UtiIjN@G`cDBu4Cjq;NUvrL&*aJ^@V%XEw5jBAV2rK
zoFl)yean~f=eoPEv5CyIxmH(fx8B{FYu??@?=0pow{br5xniEqx1$B0@1}3{n)lpp
z{hG>YhU$~&?%Td-a{Bh|{juwp+x%l|`7n9Gl+S;QCrM<!{2e>_mHE!wHM{4<W-AxW
zkKuf*ktD%TC}SYnz^Rfa5-cW^?7uBifJr9M!b!p)r;M}fNdO~LNl%BF>Op2krUHS2
zmYYf`Ck}<EHauG?t&uFaGxl}d9$RjQr8D{MCf$wQFk4|Fx9?H6`{laA_hdQVF!lwV
zcb=hUBh%s6-VxE6ba3rFz5u^7b1V)ev@`o2G+8hu$-wh;VwbBq?-u^{7tC$4f^od+
z%nQN{WjX@Ym$(bR{#$(0itk^A+<dv6@9d7xnyGL|^_gJ%Y-6=$l7`c{>a8!^w)-8w
zYOBcborCMwqx|VzmELb+XVxu}`FQt~M#u8CP7^=Oo4_+yzf&h{+ZS%R(q(59#Z&@5
zOKj`;>J;s^*KBK^S*qL3`-diK+VRf&o%dF8`_>1ADlMHH>t3lUDuyn*wBG1>o|oIk
zTk+Ax=dV9szC3W=k#qONs`#F7DZf-L7rWp3)6#9_6=k2RS>6O6_+)SJB|(yZ^JLy6
z!}hC79&Z1f5m#ii!|f_t(R8(&ySZd%RJ}Ayd?2x0*>-;Xn{S%WJ$gUCZ+K&Wx@J?$
zvIo5ii=5hZX0o`;v}{?|yr?hFlEY)N<l2nI{h|jYzu&36-aCO?qT;yVp~%dnNea`t
zlX<=s`^Qzw?R@!WP2lT|x9)vX`(4*`%l!7QZB=!5C(Wx*-`_vusm(Fr?H~JD#59(O
zbgQ`V{i`|Xq0{cb?&Q$%>vrQED`_#AQxaZgT0s}5@hnzq&e=5Ww(d4D&V)(RLb<MJ
zI?a==+kc&BVPd55*+(TlzkB9gd}zH|;_oi$r<>*-{JO_>Pi?Pl^S-YaD<>Z{*EyfQ
zpFdHb-*3kJ_|H{s%NxIbc+9dQ;ZDoX18)>v)*qiFQ>ZwF&!zFhB2{l6i-!V&9+DQx
z)}^nDJ#F~)W4>i9O?=++*=WIrua^?JL&OpcUh=lxojC2#;a8ia&Dw8$`ZdkC!DG|H
z1)K`E#e(1IxGQK0^*w2sB5{(JFK~zB&9pxbU-%a;+;?!-yz^-sTV}MkPA}d+-Az7y
z@%{+~iTBu_Du}4hE_U?Do#}MS?^#{;Y^Q(W?ei~W-g%*-<0_`Qlzr>eE|X0u5-A<p
z>N~THmHaN>`u4l(LCPNC4bB(e-q@UxyVu*>IU_e$`^~nsp;tC6+^pN2$*_@O`|M5b
zX$|FmOXjKAJn!!a$f$KzxcvFc>_63-ODBBYF(;&vS*hck##F0yGVMIA(mroz9NrYU
zRQC$Q1Cv7Ejd?epOs+1SAky;Ni9K;f(WS(U4V63Cs%{=-F<APp|JqsS&3l*J{P|*9
z<)S=u)rv-^>v_@>wK<cfq`kSD`|M0h@{3bW*ZdR|Cb3v5^(;`A>EQ`EE!}M{zdDfj
zgvcR|<Drs%QWF^*k2O4KSMd)I58df-tBK!6V?~3L(RP734ojB1bWG}fwa4IH+3Sib
zAH$b>-lfUDjCu3ixNi6UG||Z?&U}16p}PJww?eC4?WVa-<zN4}+zK$5p>r_qSg_3!
ztM^~oKkZ*NKjPlAf^VBoRn$7qv&+2V!gcZ}lXL7$nLGSD)SvTyPwKkxp`tm(p+Lc0
zS?5Bkr_ZhGnf(jC@e7}Ht5VvuNtr+S^!_;c>36<0AHE*UrTn<f?d$!MH6<_p^6Jb_
zbEw{6Q``OK$2V5a2W}ZG0*7S_zZh&{e)yB+$#i+mfJN?&M>L8*btve|n$<eDS3bP_
zwA$q_^UfQ#%;l%1ys$m<M19U1X+M#2-FGb$3RCC2fA@4_VpnGWP9<xe#+^2t2UKS6
zIH_kXc%glhRE5Ct&$fHN*=zjjJpTG+Tl7AgXBj`AcU5ef9jVMD*!=R=+uSE>mQI>>
zZMN>*#VOw>r$(&H=~id>v+w=vH|bAHzshksUJF)TTAjMPILFy`ZR1mUvo~K4pL?Fc
zd9g;=_+aZi#_Lx;p7(Q^XH>g6XH#|m=3SN>x8JJz_~f^3ZQicubrL^o^YynMI49L-
zVsm0+6pN7n+b))X$L~$QF=;ND7}3hvzdEiW$f@86Z_Q$ke3Lm%#vbzvPkgxJ(eK<<
z&?KR7h^0Xyo9C>0t!?77edYa?X)|POY)n3VnYZEa@_6pom3!tj%{jPJU~ZGr(d@2k
z0(HX4+oaz+h2PIHiSX-an75{OldRE%vNsh!-QpkHZFv)a+IjYl-g)9?GFv_`YUhi7
zYPWsooDc6`s$c!!clbet`dXWu#@5>N%6^CZQ{FprzDQy;@f4J43p#V9ImnX5nN5k=
zz=cUuWO9P@{@WYt=g0^B`e$42n?C1j%jeu_Z?Z2NR2JlPUb8sXtiTW{=pfwkc$(so
z0vWH{bEZF=^2M3Sij%MNf`U>(qo8}4k5+M?wA#!jm(K^DX-^92Jvw8Z@}!x9p&WvO
zGaCXpTX_!Dg=GE|c&0dENvcTYgA)!`e|h&Zf0~tXn&0i)Nq@N;8?DPehyK=j@HGD1
zix15Y@69{RwsG(Mylww_%wN3gZ`meo=6ilk;2Kfm^O1X-lXLf*eK>E($&=u*pe*U=
z;v_>c7Ota{T2iET>hEV*l)HVut;eFMjSgR{wlFgPvcGcD(bayHl4*U|y0yJ&$EPk`
zDz?nXEZO5owPM0b=1l=kC5;B6jeo!VZx%3pBr%bBBP0JN?fjn??Jd7M$Q7GSpJKYG
z{i$&N*@cZl?-+i6u4Js8`QG9CqiRE&^KBjM5sL)oy1Py&_L6JlQr=$IKX>2nbM*&r
zy!x?l#oJQvmp`~)OjWwK-fdmGUd~<3+a_~%e6j4{b&@Rze~~kR;Y&k6`x53$azEBD
zUe4y%w&APN(XK67QifkkzS!&Z<o$2|74x&S%Kp6j;y(p{Q(f-5I&WCW;QX1PD?q8q
zzP<SKN(X0YaelsyT@s6#R1CXBCT<H{%OZ8G>9XMI=)I-OzLd7QOBs4PTStUxY*hT7
z%r!~JjrEO`ql&u?$At@AHb2idM@ehQyu8WKq8L@2@WMd0px<JWcHE<m7!`)6$F~UW
zWcf3L?-TQ}$3c7IOs?{CJ7&B0KEK=@)&5YEbt3z-#*eSnPX^QkUTpu$Y_c~y@zSMB
zm)0m`9OQp#vV1qA$(9w&f7Pb-^<TZduDvMppvk}Y%U`Zh{&Jvh2K#N6D;@1$e!bxS
z`5|NB<%Ze`!UrEdEZ8;2n?c5(U3~emmmlAB$l9N3KG}S^@p8K#^W`SiBTkPVJ$kh1
z%l|Ho;&eaFvrYAa!gE<JH@rM}`SRr>YZx*eOSY`wePgpR+|MdPLexSkDk5~Qt}NfY
z*Jo{~T2A0(m>_kL%f~@{pW55a9@T5Li!YdDsxIaFyZ==D?S1}29lDZQT8EwA{Ez<r
z$1BKCaLXpAt&>ue4+pKE+!MgN;cNb~S25h2lB>769<mjfFvr=XUGvnXDZc|GCx4jB
zRNyyJr0iJslGz<ut6bU)xU#wAKA5*%y1;+$()Eh?yx6U^TceF~ub*{Dd(t(bMJmc@
z$rfF~RE;AGI`%(gS=gh>EwZs?%Yq(}%rK5aJ%ZB=7KQOhFsF*@s%mFz7-%r*X8XTd
z$r5};VYSs<zxmNC6ROTfwl4HcYtxH(IhjW$Tx8>#!mia?JHxK)NgB_JS?lY1I_$P!
z&e648(~gSWTJk<k@Q}vFl(=ij_onD_8BFXCU3g`##j=?kNB(V?SgLc;<l$mrcei;0
zZMWF^{%GvG;P~V5tO@29g+C^=>4fa>h<Di65of?}r@mFh=Edh5>(?FSPm-JB{OS?6
zb2H07A;l*g`;IX0IDF&F-@fCU!{2jERLZY>&T;$c`x9SOPQJQ1t<`Ah>FMD$FPHLu
z+J0(Y=uemX_jTrcS~t~aUDc}9%=4%3tv|K1=t=3GOevwT=ws8{1-x~(I~DC{SX<-8
z!d}svq$wEF!*ylFjp@Zpr~2<I{q_BB-Zu7V?q7fF*>87$d(6RZ?((o;75}5!3EErh
zLiv@>a{Z}u3yeJNah=mGr{VR1RRYs0tF61D!tJ$H3=W-Qkxo%iWLqBAZ?@0OZ|c9)
z)*Wx(JdQg4Y}%UD%}1+l?|;%`uJQA3uFfA*Pwx+3#HNL5GfO4!O<d)Ebn`3TLwr9O
zn8Y1~r!ShZN!2~Ouj%OGN!KcR>StRA235<1Jf3_n`gYS_=ZAMDZKzz|)8KxHAxt7?
ze@xN5ET^BDOji!p`W>}h-7>f4u9!^9WW_yC<%GA!ZTy*dy2QrixvRudCd+4PcQ>eO
zhK5x3W+k=879QBS$~YvZBiTol@9qLF;ScW~<}2ywvv6^UOw77=LBse!{W-0LU#0T-
z1QP_?9aftvCTv+_{MbD<D@^;xrW`x(o?gozs%NXZyJa*VYX-fVcrPfZtFt-du}Z-+
ziMvOxzj>@(@qgjNqI=Ky3M;a5AMJ{|%v#E2G3BiZ`*K#}Fp+7&OHU=(@AWmC)z_J~
z{K>wu8)**(_q09wy|nD}@}S+5l-|3@Y45wb`r)peH&;WVR;Oy&?wQNBG~>|8L!YiM
zow2Ra@P_K=g4;|pE-o!u5h}Z@ed-P7ZlzPK)(b0VE;L%l*Oao<ELPBYx%>X>)3yHY
z{r!uNCpzxlet{MSp`{#~Ry>)+a%RtG-B&FNI+HEE_<c03R%Tt%?%;Ac%-AWbzG~0y
zJiYH(8XxjkaIn7-WR=-}clGywCfYaZdBrA%MjXst8*M$)sEyS}YsXs6>a;a(R(Fr4
zr#%%orWKu;pI)@Zcjs}VM~gDU^PWlfy^A?+C^h3?a8mhF-dxR8t`)1ar8<@<O1L?B
zzBriq+JxapLcybIN0IzTvpts1;OR)(cp&Ll!L6g_ALZ@N@_z9TVhz6D@!){r4aMEl
zh3_`a`z`!3{V>btf8}n=TRe>)$@kO<eA%PuV6Rc<_pWV|M^KRL)+qG}%J-|!vhTLu
zX#4J|_=is-(L1Cqvi^i*Ijbo;Omb3wG-cuhJ5P_q3k9qAXD!tH?Eb8=k@5F~?@IR$
z=q)Nyuo9>{p}YG=qo<ML_oIi~Z-4u^%3H<aQuFENV!0GPABSBpF6uqJ(3mAIvbfZ(
zk)v4AW6nvZf2?m^@7uN8+2?GKwP-b2?k4KISg3KSl$&5n;?+eWy;n6(uXGdWoGQFE
zNQ6o7bng>qqpi|6rf*o$yikhu?OV2mdwV%f-wG1==p0qCH040|!{%n@qcH_(JY45^
z7CC-6x~QVy#cn~Hh>EFvYWnTcQac{~u#w~G-L;@<*1xh}8~A5bE~!b5&<s?aw$;&e
z1+Q;kLG4QC6iLmEp5FTx&aBjAsb6w<=e@FXr}jE|l}>2SXLzDkpkwW8(dH*RGa_oY
z)~q+(T+{eBXD6)4X%twxBByt<*sgFsnQ3m`?w>mXHZ?wYz#%Db<lH*7acSB$502$8
zSKoXS7U1&o<;#~-=Ucwq>$b|1nJ=-tyWnBktA|cElDEA*oqMt^<-)0Lr;CjZwiliH
zx?$P2Rk3G|3Um7?t692i-26>-O4stTurfj6-;dnCN`&%VX7>KQC_vn>K{9~vHq$+Z
zWM|RFPLad?_cj_ROuzJtBdb}Eugk?Rr+nX1hOd)2->TUB*t5v~FoW%k{`KPbIP5M9
zK04ej`0#*i`nAPh9P8uv-~3iP{oS|PRr_Al?$27a>SSn?*7Y;t`#V$@TD-9EJ@05K
z(eT#fa_wVI^Ox^6xrDkjcdgo2(psUawQALB`_l{$mFksF2Q78$pR`b<Bj~8L<W%9+
zp)ywr7IkrCE?m~ck<G-P$$0OmbC8DDE<vG}F12-qK^#{crfG<tbkbP4_Q7&Ft`+I&
znWk0iy60V3VOJ3-x$E05&Yp*rGjq?Jo|$GM%~le4=VY<Lwe8!|J-(i`vfRyQBI8q`
z?d_*D?Z(2Bi{47T+Nu>1vwF(nAg)J7ccP-WS`TcSCO+9saiQ=Dfm{V4%ZWw{`<aWS
zK6yB2cKf(|yCHu-R^<1lt73+(E<N+E{P%irBqWmg%PsD|-Y((48lywHl??yQidg*5
zBB;JcShQ=6r;=9Z+x<HQ>=!5<s_XLn^Zd=XHBL(Zji$H?xc|(Ze`?BdrdHb<)svRS
zEnn_@x${c@YrZuCACLI0sqN^z(lzgl`*IPh!m7gfBJOuCRt3yYI`+Nc^{;At?^64s
z`c_G-#!JUbmliLT?~@Af)u>qT^02?Z-(tVavl8|x;%=OWy4Q+4JaGNe0=6TJj~yy5
zm^?h*xt!TgZ_ee8mq*N$^p@~?SWPfhnDehz@M33ivx7;(mlBP=FJ=iaFJ-Zr;PZp6
zu9jh6L!E(%Qi6-|f*FFvi<{Y511vjq1RQSMSk$!4EzUV0vS?1*<{8@`&S&vGoZZS*
zaN^)|-Dl78-o4v#M`!2k*>mpP^W)boIj4E5=*gLbQ@Prt*>^`C<Ja916n%5Xl@l`W
z_dYt>6=h(wa19Ut*GIRrTg4_GPh6*B_&{QVkG2TM*(C-E**&p_-LaEclsVkQuW%XM
zzQ`r?D#)!w`NCX|;s>7}MerY7=;o$$G}JXDNPfwwx~W<ktK}~(sJvsPo4<*Dae#fb
zU{+6hi^7jcmX%8s740lP9)3J^CYycE!q*2h)*WH^W;;vnykA3>bMWm2ZA)CfE3Xv&
zp748vz~wHvFAR1TjEe<-C^5c&5%5ECA%B~+Ag@IGahV@&ImL;e4jL!Oo7m;gYK!^#
z{5-oJ_jKl$9Zwx^bfh})%wS4BET_)L&Zp03@zJNET}XK&gONgt<Ku%hJ7u~i{+lV{
z)6*rp;-~UdnH5hRceaJ_toV}j)Iny2x)<Z6mjOI0*g|+#u)bkd)1D~PVSj+{o(3C}
zhAX3N?_a_4fRHtdcb)z7inrR|*lGL1-2C4`8~9ELbg?MOUtgVNxFRY@B<b3;Y12R3
zEAHr!I>BhKKL1ml|LKPu4nJ*!pUr$H{>09+SfA&(?tG0r|9}`llXS1oGiS~%t}JWb
z^?CBzGZwiPam<I;Y0O?LwW56K>%6qIG(EAKeyy9{NtLY%TBY$gGgP9y#d?9Jm`0C?
zs4Hubd}+(yk|V4lqMkoGTK$jkzuGJqVjrq?$Zv|>w^bf3bJ?cYB-<pjt!ov0zJ!18
zbki4`xat<$2NX{J@;1=<?*}7M5ytRZ=6fnr{fxalDuW|6=5$<ZY2VE>@rSYpOWTG+
zlN`P;m@=vAjHdiL`53O{4?e{4Z;aaWVENCBGghry^*`@wXlQ6?{M%bvD^{&qwQAA-
zyo({BfgrKa5YdoOk-*T9P|?uP(2&rsRjV}iAI@5}YSoHG>mP0nbq%an4h#(q|F%{n
zwEK5h*Q!-oE42PegLT&{Ypq(fF(mZA)>RRZSzW8NR;^mqwXuGF)~Z!2_CLL~sB86)
z+|ba_!0=yUU5h|QuUe&b6k>0vYiKCQPXB`-%T{%Pw60qJ8LS*+Slw)pb3y|_%EYu*
zELyex{nk*?@IS{^E$SA3xg|6-@?Y*Mjn$%HRUi$a@%OS;i`LJX8X8#73ijoyRV!Al
z((2m3IAhhqU%H{8p|0`Eu7Yd_nH?Jba%%|4-+^_rLtRBfoqw6F>e_#CYiKx#vHUXF
zRjXEYt<qZmyE-H=Bs7#YG_-5==Ua=q^j}7;TC`&Iudr2Jt3^ZW4_yWM<-hHUMIifD
zt^N`Pw!LfBDlRQhFoMEFYt^dtzj#4@>HgQc3KSIj7p{VqFNcPPg2E2ukO*+_ty=Zp
zQ)|_#MXOeIe~!}X()x1^9A*1q4qCMG)2&6TR;^n3>6X@t{jb1wf$Rh6=vuWJWLS{K
zf~rZ2j*98bUv?$*r&QPet67Uyb*)<cH*D1^{ja$Zt5#|KUzxRHmB#A7*<h!Kfs%*#
zS6z|N(7@0ztrZZ5tY39CG_d|DC`LlRMuFq!@9WU;D^o+`_gxJQ^?wtk)%`w7YsIS7
ze`kk=x(46N(ps^9($xsC<5#WvGYu4a`~I%dTCpM2&O6k#YtgFipAw>xcG01(N-OkF
zfh=_l`I#Fkny^ai&$h0OA)$eFZx?WF_`O0jRMhi(P^gE-s#U92+r3@IwMuKn>hDpj
zx)yaUShZ-?fxl&u3Kk_xCJA|XERYr8{k|^mZ@$mxmh~cK<|YD8r@eEh?_u6)_xaxj
z&fC9h_<5ZCO8jHjgneDox*_^>)Wa874ou9xxa!`{cdBPKM2%*Y_$_8_s+a!8aPna3
z_8le8F1Hf`GPVY(Z*$?Dzui=e_vUuZEh}z^Zd?3;wPlO0fAP9cw`HcE%~gMJhVOXt
zPqlAvN-sXC<o((obE1knC7^G2pGqN1#{pOK1l~?QjRGf6g-5fUJ6IArj(PT{$g(Ik
z_k2-qOBHnDaFAxQc*7(qJn>C416!aRlkfy)CgCQ5p2sVKIM@}B7|58NX*+(x@PMvv
zp57|KQ&V}5%$ub6;^PI=Ir8T95np>hsJ_0C^Y3u=oOzoSpC0-Asj^7U&fW3t{vhF-
zqO4s?>iN$q*eAa9`dJ~(Zy>i{hWQ=Gs!F+e$D^Oy)ZDYVJEw8(noqWQ-_M%gIDDr3
zv2c4wAm6|KBAyu~UkW~a{#5z>lJUk{4mRfPc|}Ivp0_Waeo(nrY)1A4&uMboXQ!O5
z75SjH>|D>@-xaf8OWRgHeagQ|PCs9-qEgOolc1km!1J|FUw*w;AX+WFEbm%<P_ewg
zvL_4omq!Pu>-sL<)sqovb}&CnWqY>HU6BXUi|*}mv5#B2w(P1|m-M!ax^s6bN@uWE
zh$zi0+Hl}&$2<l{LCdd-4G)9L-$lfv3pEL}KJ42e^@MlQZz+{_CxP1IDxEK~p14N5
z4VRMVH(7W~?DyNJDc>>zH9WT7bWVSEdu1lq3A3%=?kxygxlNf-!CycngHLCo0(;`5
zW16bWoQ#`7968vIC4cG>bd)@-{&+#h#5)-&YJxM|cC6D0Hk6zgR53|VCgPDmkIQ8x
z^8*X`T-uZlF<y^oVLY~K%Gyak8z#l{+>q~kFzIFc#Q+QDPlBstW~yH2GJbeK$z3S$
zK$1ocXW)$%mJWtPN-NWoW(x>7uauj8x%S#BPuuS+u3oO{t=h2Z^sd{N)*eea-I<=X
zd2{b>qvEoA>CtD;`P}o>Jf(1!YtQ2QYVW;TpIU$4`0U;D*zbAi$<7r;&kA0Zy>5@?
z+7W%jecHM6QDu&8d#85Y(uz*p$(b-|v(O}i9+e2gRs4P$<<D!o)AVZ_qz;Oh8^#7E
ze82nJ;lkW+YAx)K6Dpb<8s~}_JxD!%M9^4ulOS^?`;lYHBEElupLBjy(NFk&`COAt
z$L|gPW$cStz9-#1_DWVlmW_X+>S@7$J@IMGx0|1K6(+Dx(|ph{?+eR&m+FgrU!3>a
zY3zNm{HX9=!S}8|6=YvDKJ9+mRP&MlnC>6;j~((M0lyZg9~AhtZ<Y3sgZ%3hDwEno
zm_IcoAG9c7c=#`b;gO^KF{biIb29d1ESKV!(kOqFboBiJ;fd^f8Qwd{y@`9{^6ud2
zj(rj9t2DkRDc@}nH)Pn^eY@j}(lZBfIi-{%FI{~-`n?qUS$22KwP2pha=N*e@oeK<
z)(agL0(@5j4tA9)>|LRs#gXK~cST%``&QRuSLY8MGd{R~=&DTEyMpspL+a1+=`s;P
zJp2v!LjzwPn(%d{v$4ogFSn4itizlx38%jGDD-cSJ8;A{cFoZ(3;H@Yv~KC-TH!Q>
zkHg`J)@B<YPj=4P8hxjO8JyJ_N^CAWvwizBO~HJ|+Q5pCIh6~SuyIw%sTJvH$Q$u0
zEUKN{q2gJ2K*i=k7o*5dr>cFzli%&;d=$o7>0|n#aQR2I6YF+cvuGMAW}onCcr4J<
zmg4VeDRjRvmXk*&h&drpO}Z^m(Y@tC#bW=-!93~{56GAPJlyhGjNzq|hnewuo5ve`
z`i`r%$KE*IFL*rJd;0=`tG`>`JYHJY^<%QgGN-bO3vVtLyso&ZttrId^`>;arIL*&
za}5G=yqq(SREfv)%{rpxBz&kz=eV>bm-ksUr>fS!b6h8VJ`_-Jvd2^FNrZ3G4!Kq4
zOuh$BM67eqsB4v(w0UCO!hj2|z8VhKpUiUFzip@HiYv~2?!0{OSx&Vd`BSBR#616^
zORAvyBPD^liaQe(CyE%AEt{CQ-!kZtZfDdYkBM2SJjD|=H?E2}@absJiB+FB%=Gl<
zo98XJQp;sk3CCvDSxh}?PL)4S{8mp<ay9pqD-k}|9ejbKZ-0)0;{ln9GZsP#o-B`*
zZ9*o_7Mti3+Ww^Tz=wr>jaRB%b_5AIv~jqKEfj1jUBK4hc!D9MKymUqEmIkFx$GvX
zple=Y8$PcQ-4JoIQ)rT6XxKKH-x>>M+H!^+N^lZ!bn@f47$L@RwZJmmYx0I8BE}zC
z-Tpl4NMc^UhKG++Tgd04Y`oz8F6C9q3;TC4?)oF5A`!~d$*?%bG`Q=c>^F9wGo8*#
zHciS4FKS-2`jBdTqE%u}rSqH6s+|ixRjnnvKF4vXELc(5Hz8c&vJ*?=HK(4}3oi*S
zJv=v1^m5NHN0A$;4ncqZocCXDW4Gjsq&4I1w?7l7U2R(Asy^#N`p121*Rr2m<8|eh
zOLl(8Y!}s)54UhfKL2-R`yKD_%#>SG7PvG>HYu^`H0qf&3i@4T@2k2VCAWUzyk4=s
zDpQWxQd7UFY=7j_ZpHa)l}VY&hJe+c-wrg1c<{b_&9SzSPf|oxV3FX=X5XM0Qm*c6
zm7XRrthu^IP>G?*Yr-<&!}B+A9B~wFVXQp(fmL>;!W6b0fq_Shm_?t4y<UIOX`<x1
zQ^^NxbQ}dG*qm~8R$I+FK4+2e<_jl$*G{%#Nk6evf{o*$lww1Ytj^mQzJ2O*ZaMd;
zFLrmhpz%<lNLZuulnI;SUayBXsyy0@Rz75YV6D1BQS#i5Cwv^6pNBeVp9?-8(&5w~
zAk@gr@~}&g&$`WHfuMzK;epxT6bjf?*)>kAdU!F^xid1}S@yZ^IVUgHjah3d?B$Q;
ze2?4PwY~b(^7VV~w4LBDGUb?;A}zhRaku%GbsEc$K3&k&{`j7C`=d)!)`nNCjsBR$
z_xr~dudYs&MW#M((?TQELX#2=LfL{Moh~uoYrLEAlS%f)_Xyd0Z+X{PZ>;Aii>Ur>
zmBS-`ZQd-`O9^U)%#RXZ8U9e1E>LUx)mTApNokY%T)|U~pPcV-EO%4*z2@8UU2=OC
zd(3(|Aw)Z3fv;1CuaJmSx(IumeVj|DkdVhj`PIrR)z7{vwVLaB*VTAe?A>)8R;3+`
zo7}`_M@L5oDC9ae>+{&#`>Cl|Sv1BpX!I_!^qAn_p}@2L?z=nT?><{jRX5@<OMm__
z^~v>1yLP?!Qf0=xyVU8G?TxuN%nRNf-_5*R<ikPz2km9-WsJ=%?+&s(kXR_l@brT7
z7g327^D0>%Pq^IG@a@0)#h+%33_s>is!2bYd-9X?e$7vQ%aXJ6PYYb$uQtOZTJF!%
zy+xTqueUzi75C#)!CB$NWv&&54^I~x=;j{gGBa6q>0VmY6rD}2n{NA@kqqt7y0tMn
z!%$K<voOrKb24Y5HmeZJxsKEg@+?M4@`nYFHsn0qQ!mybVZg7jL3*3^Pj0(s)*bD7
zGCPW9d<w2~ePR&)ru`-#kA<Pl4S6pIJNL&KHHAMG{3`sRRQ=Ia?&3dr3G0reM{^2q
zvdd2SeAe=v;+b<6&+b|0ef%Et<AgwyL__n%FAQl8zs-z4e_qS;^5~q1kBPG_Ztlt2
zrjYjRviY}VImM~}-~TFpUHtjwy}5jC_0`GWUq)WIaN)wfeH&|ee*a33f1VH~Xj#Cq
zsAz@evm*ig=NL4SB#Hu07S6hHw&a9SRM)Jakigc1Z*S}F-8SveT*K$aYb`6!ZJ%>|
z^5n^8L3x*MMVA|Ey6uR{Y;)QX?BlFv-L-SVnbVn#D{^GarbeyJ%t*S$t#xsktHF_i
zEX!7nj$TKXNedTiUsxn5!lisSp**5bnc;B3i-pzfdJ1t4^#M(99+(^CCCRx=H;k)h
zpU+;({61lxLQ`?Vc7dlGKQFS$=zi{C-_ZHxfd8Gu15NTQ{7v@E0zV|$_cOdU<d0+D
zb$~yKDbv}Df3wK9Bq;;-IPS%wwI>#?7yO<u?~Q};gP9-LpENvosEg=7%Ig34=OU>O
zvK7pqj+7{+A3fhuD-dw5>#tz^BBfg47ajZ;1%4?oP8X=_X#C=GviZvaVI{?{&Pwdp
zh4%}bXi9M~Qds69+%Ee20pnzrliia=jf{-6ij-tI=Iq!hDD+6XpIwftJW0-Cr-0Bi
zhGd7r1$-tB#~Mq;UYQ4auULAeFrj4Odf_-e8@5jp7j51y-1Wrg-cs$U5svoN($BV7
zNIsuc^L6uV`JHd%rXEZ&Yu&p1S-5do7}xCtuix=iB6oJrow2w6$hG!6X*$e3Y>#$0
zI+)8%ZAhrRes9u>OSg7gec+3cX|nz0;IQzU<Q;D2b5(ZB*S>sV{p<6g^k#`w+b*y9
zQd0JD!lk-`yjJ1fLWT)RGCLgNp3h-YT50BZJK~T!$IRs{PdVaNe0$2ovS5L`3oDa>
zPvxX0292DOP5p*VEY@=#RXLn#NHUod@#%0!VWaAus|QXN=1EOZ&yI~OIC8zAMbSne
zW`VhSr0iu6yUrBr=kuH||Gam6&R<TO-L>cQdgUT!au;1?FgUSDN#x6N(XEG~FM1rg
zv_yHz(=&5e9&!mR<UHY|klZK0=UUXYA}F|(@sSdnW7gvn(>N@SXdYCPR<h`1d8)y~
zz3Hmf7wJYBkpR|~uI7oXfnF>uZ=&S7jvN&}t;p+=zn$+$A`esYaVf4-6^nuzEgqQ9
z5j@EGgE^&9Ah>Otc9U=P=f>j>>Y7JH&R=3$rs(ouvd@!4e1>glDFt^L4=vq&;mDHd
zN|$WXoMviXnys-QhiRIXt;GV3ITn7VpNf`Em{+yTV-0VDNYf;aOU%nX)ckD{J)T-E
zcec(^=j#c)VDXhfN#mU}!^@5riPz3u2w@AeSjKtaq@jG_f(b0_2EpvB?v)rYDKzK?
za6P%CdGoUQBM+v}E>kC1nyNLVi_My})Q3gkY3jPJREDAmk3OHO&V`%O4~Se9dC$c)
zYdY5|Q|*pNPFC%SCscY})K4;N$2M_jxEL8PP+{e<yE02l`5ePb*OnWBDMvM!dDxR1
zG?F+HV+?g%70rKVXM6ju+@{OB@s*IS)}1LU%}sRwTF0Ag+m>^Dae>d2$4Xm|HT8=X
zc(4j|%$+9`Gl50nMS{qMN33R|%q@;r6q-6#asFKAwJgg?q*K;uQ`w>xcGVq^I2Br1
zv^c+~99mHN^Sb&|cPpg{%ALl`r83waF%`7D<!<48&5_W&Aizr~S%;~yp{Hr7h<o~z
zy{ZQ}e~ELf6f)eI^nlOC&xx<7&8zEW`_{_EXC|?K3t+xr-1D+Ugvt3$*`HDlz6Guh
zRd20&{bOJ6E!rNtkfV0%Db<ik#$6{Cw|l2f-!HatIx9;_(JI>r?OE&CBxmN`y1I0k
zgc7H?#1Vs}u$N6PXSyyemGp?vRuIc*yxd#4()5^s;yQy^i45m+(?0l~a^PL^$k6$i
zK+T7Sf*Ep#3VBKr1^?GM8bq*tp6sgs@T22TA;xDePtMO$e*1f^!Y`9|LWaSwntn7X
zmL2$-c(M73x?3yX{k-nCmEN)v8<W0R$rxFroP6j~aO%N>2Nt0kA?uer+<mn3U-_H_
zg%yDe_gAl-82@HN?fx=b$ES~i**CiU-f<^t-@E?l8>1(md~$^^B>Kkm4Np(+XyXx{
z$<x`|c>ej*;*Wo}wWlxkJL7l6DtM#tMzK%oN%~1;%C)Tfx@=sNJXjC7`mf<~QDhWJ
zVGLwR>JUsi;p(x7Nl5ExgHFQ&4`wAUCoV2014j-HBY_|$g9#hBIocRRnC53LfB0{P
z>HI^|yLdH>7>|g2_VSowd@OStQ|JEH8BtrLSS94z=fB()w{cOs-jyvL(U&K`pPjXK
z!tHHtXA{3^oA9i<vQ{MI)6cT28hjb9u4<t&-)*-Q&v|UwW7f;Gx9Z!?!^WyRB>H@2
znuo1>7rl3P-m|>-0k4f7R=r;Ab53Goh3d^sUk^?3z4OX)%AGT3&YawpnWuIuWy4~N
z)UCU74)ZQpnj5s>#vId-a~6p%r=`_3(|BUv9TwBt9Clm$FjLC0P)mk0%}*s#5A8Y8
zyg|>^ywE3hDf2P=g-&iB#!~8=yWKB(yi3@3TSg+QC4Osn!-NBkRqkiH?#9+043Mjl
zYl&-lVIc26-{nVA@q*&zxrqY(+s>yk7V<?iRGg`}lf}QEpTmho_M!m8<gk^+Y7b3z
z&Mmesn{Rwor2fqRE~^U*b@}9*R8P<TUb)0$qSDioK|9yQ%r5SGdrCwjW~aN~R6VPZ
zsaMtphpp|oryClov9)CC<tLGGsjIKATU+(&bTOk|?GfYonwNy9@D?0oT&uz&=##b1
zH$0N_9#f8NY||c(fEyYDQ=AHy1j;tLt$grO`<R@<HEHvTIo!9*eNXs18wahi`0#My
zghUJ1wfuYfT(jFAdu{Su+-sgPqn$bML*8U}bA_FGJ5N=7Hm*8y_@E*0huhWp!W-vJ
z@8U4Lymw7fB3qjZV}eRBo71!Yg-r!rr8oIp1?E>i>6Q{+D)H{bt>=lWk35{Md?=-C
zj$lNefP#a~FE$+y*6z1kc?B4ca~gbDKI`cvr_QX8J0%313N(x-n#w8~%<}WFJMyjf
zo6ISh=f>Ck{>^omH%<L@ee&OThyPyR+Tb~J{=3aaSDZ2ucBc8vG~zmGYpD3TvNrs6
z)lHE;Yl-Eod*j+(g`D$h4fx{S<C1z(z-dKE*R%C*1}Zxq?7YCPvG9e?Esi%`f}ZiY
zd7Q78U05f!RA4<@h@{<`<yR-AZ(=fLD^k+YJnx>Le%wF#=7%$<SDs6rFmDnM=fg=(
z#^)G6RHsgO(<oo4(K7paR;0JLnRn2xT&c4=-cGvUezB>eY}*%xxD76WcDG7by|HcH
z@b|3Tk%Ul5PL(E!9YIqLt7;rqOIo@<<iKOI+ktBqd^s%DH20=zN|m4e2gd)gEq|BU
zuTEccW!;Vqe6yJY+ZGFce`MG4{qeTzT>Nh))Tgnz@TN)JPoEjYF5&U2eZKo=f!^PB
zQ-U8R$*VX&op(mQPJN~A3f(*2J8U)TdYmO18T<t$e*C#}`}glR+yAWn*7I{=f!Txn
zJr9q4`#yi~+`YWtdo}+le=>M`;P&ljcRF`IbydzY?Q3xFmtR=Xef7Q9-m27>)7SgW
zO+L>s&$Wp);#4g2SKW}UA+Oa;_RXAC{&Rn8`TO?!%zSJzbw85ah2(4Oettjj`{DQF
z@B4fNs`hO4*L4V}O!U3ts^R#>=7hpUw_Ou{EIzvXMT0G8)SieF3p`ZUC@`OG-PD{H
zpxwXNp_cg}SD1W`U{gb7bzIcMU=M{Avq~;m<tDt`a>Z{khucQ4D^nKTmNwXQ_0Xw+
zwOvy}6Lv)Goz~OfCZs4Sa$%OR72g5pGwq*_tEo9NpYO9Ne6ahP-X%T}mVL=JN9<La
zVmI!-f6?^F1f}mc=9j!U`)87yOJ0!rJ#iDKM>RV9MOBk0TJre)iL(jr7O`6<AIx{|
zhTRK8E=45~&hyWY8Ss_8do$Tn%`@fBl!!Z<B4!#F<=md#$>Dd=Mb=C8Zs+yJ>mM)t
zT3C2(@zv&UuD=z&e6y%uAKZQP<JXF7$HPjNhsg#pd2$L&S*EqrRCDv5CB+{Z1Qlgt
z9GNU<b}Z_$yk`FS^IO$L(<XUlDfI`hymIp7-*62E+gjhd+TudHmSl=P51DS3n5KMo
zV`k`5Pk&d(tyvydg06pierv7s!w%6$?GZ{!R$msa+ErCmRT{=$`byTVl~X-HV@hVo
z)D^qK&Mca>?&`@&GZMR=t(Bg*(OPlUhm9qR+zK=<E=_h3NoKWH^PJeSDa88wp+o_D
ziMlV-zXi#A?VsdfP_q2VLGFrF=g7vXygyvaCKxHSNeh^Ll$+w$?=<^`?eV$Jb1g6a
z3o4H{Yd0%aDtdUTG9=zhpHIfT-Ql6@L&x6|e<%L7JEi>gcisK;%{RYqx|VV>^v;5-
z>~E(x{k^;-VbSMakw7>8RvAv!MwJw0HIFAJC#x+idU|T=?82F6o~5Nec$#wOb%~W^
ztZet{)vH&pSu@XZd+#j<7xwcUd6Txid0abbud1l_^;z>SPR&TWaq4q6`-`xfH<kKM
z?Ykgs6!zr$tz3;u*E*weQ@3_SPkSnKUOzQ}HA&inkHtu#m~)qMw(QRBX1g=CaXLR@
zIvVQLuu}JXs;^^oUb&i7e!1<koI1UgFKxb6Ss(m->6ex-|I;%;(swqlUnknTsyTg2
zsRnO0qnC7NfOKkS<m!A!HV@G@f%dJhWx561I+VC2nOK^x2`(w~SRxh1#i6*Vt4VU^
z%=w&O`}{*TeDx1GtZH)T#j|y1cBk$(N?66`w&8-~Y**KWjHV@vb5~9Yh|#r<lIj-H
zb`(-tk*%;ngzrRXf~M&E{4)72O(pH*Ps@DxPc~RjW1r1%aP}d0w#GXg@f-L5ujP_c
zY`&Kye(9U)9pArK`5v>M>6rb)=K8GStT%?XKkWWg$bI&o;raH6B@e>~7CCA0H{bR!
z2W(_vWZ+)2E$Fn)rngc$S3~1ItXaJ^-L_Tn_xD#COEdze99ZLH7#?MHNj!S5{@${f
zb6;<7FA<CC=hX^X7jx#J-m0rvpxKQJ!S1fQy0Ri|0?#J;YDsit9CVwybAil`NTr}T
zn?fpgNxwI~z#Z~xg31@WqE23kGUu9qb?LKzEjCtk<v04_CHtgWuDa&=`JUUXA76Yh
zm~AUJ^YEq5lY+biS+;0=(&R7^S8HY3xVd58n(!-0@0b`AcU@ng(AHO1JvTJ+ZMo~R
zt2(oMg?0vV1RPe?n3)vu&~4A!oieV;jj2&VZERA39T^<~oA+L06BT*0&|9#yO?$S`
zrBnl%iXRP~S2!M@KRHwIM4yZQ4y8obgWa`EzZfqj%;kN`&r!U0;{C7>EWxvdXB~XH
zfiHaHcUH&K?PvLeiXPZ5P?<cHt7xU?spOMc{0}(hZex7s`snxG{@rhX*xlbA{qD!V
z`;(*JUMm%5KHe|qQT454#RU;xmq+TNCX0`Is_b5=YL>;}chIipory-`q9&0GXI_?U
zdiU<@#(b-pd%b^yz`uEH%cik$XcivUXyrWKF2orkqT=WtrNBM4e{$fYCp&$#7RoGI
z)vUQTG;8g&V^VAK)?U_}@w`Tzw`5cIG_j>8uWr4pSZEX?BcauKC0)`bxzlIv%_0pW
z@r0FpZ|v*r{=YlLX4;cAHJ?*y*(IAFGM{W-e48P2jiYR)(kG=Ku78`i230uOXt1+P
zQqSzDU{XJJknaZvhd^pa&jN)eGZ%;54<s$-_D)Pvz5e=}kf+@BRPGmxx^x4ads}8M
z`KrUYoJ&aW?xK$(@d<U4JXNN+Bz;^gD=EY;cV+M5XB+pla<TVZ{cN&jOMdbD&7VIQ
z7#9~8eE9HTb8(@F`M!R>sS3rHKQ27R&nqIXq|~|ocOZk)nm2c@+hiZMJ)eE$N`}!!
ztw~<jZb<~qG}tsVcc<}M)3v99gZDlYy>lwHNG0gAwX}OV_l%1=4vPg>ZM!JO)wM9Q
zvvpgZq`0}vNnfX~g<%@!UTPe6RF!b}Gb?~|jlsF**}6$NDyB!er!s5V&*pu@&fAn<
zbB6cV)aK8M&HU0l9ISsBc1;g@B`d*rCcU(Al}cQUoQw0tIT^N_-1vnU-yBd7;6EWS
zzmY{ifzeu!f$6A%6WbB(i4)d&)!x}El5*1ZO4e1kvr~Lk?kcpmzP`RbZvBbB|0D$x
z7ByZ}a<Seo@$VOhL3^X*iStfvwZiA^4c<R`|6|X~FYmXSnwoAXC^Oiy;mbA?!$KF~
z8HIe}`F$Y*9bsHS8?FRcy0!}N_V&3tX}m}eJX<`|_-3##%ZfU;wrLt`0tHTS-kdGy
zD=nooF){1XG$S*!h``WDBi=(zYl2pEx;89K6<X*vVab~VOG0=&?c8cKq|Wa-e@{eI
zH2t1~f&R<Qe`fKq9O&GUxG&-DBXNVBjl}|$#*_EF-71ro&BJe0m>{2G+ElPSXA(ba
z=NjQxbC1b~T;u$DHsN<h`OH;4{wq5lbLd}}UnS(RKal5yfUV+UR}tS$y0;fKzUwc0
z#k{@Xq0E=ZeG<nk3Legg3|#EGJ?~C;QA*U-tScuc>-uiZs`)LV5wX~<w<|(3Y^is>
zaoy6ctyvMP7!(=8+Z0rp<uvCqGMFwqr={85-SKF~l9{|lUW?PJpYn->&L~N_%yabY
z+M8{&7vIl38M^V2*^C&O&XzNw7GYC1EK}uhRZ;3)61?TyWpOz(!x!HoEI#;nBxP-x
zz`I1~m-$2+#T|>wWf^v7K4fBVO@7GFaWQTR-;~%t{##p5i#OVzU}*TP5aY-v^IP((
z!Ply{!Z+%=IOLT@{xNBPbP}}F@Z=D4<C*$tQ&Exo)TvVpdkfbUr#~)8p3cWp^klp1
z?y|`0{Q0?3e}A)V@R2T9duwaf)mbsymfo6~cI3728o%|iRi>+h*ETKnYGPnm$HLIy
zpun4(rNrFX*^@M*&(Ud>r<YfR*PON`Yo=(2tGOIH^WY&{%+sx9w^I#Hp0j<I8SQ%8
zZ_b^|MUTEF`}sUqXBOYM^RQ!a(bvVh(xhVLq)dG*bI)+8n;RP&9FN}A?!D|wcI#!+
zt7XCBzE>S%Gy;plZgg!}J$da;AtjyMO(NbNeJgGSNdI6g&-}hsJkYP*spD;4hRu#o
zH|FnPG26%|&E9NYST}?H<gfnYA5Hta*bg1}oWQNo$gSYQ5%EY_rjennwdox1YunWp
z;am6j?9|?>G0kO?w^7vA>^;|6cbDC{xjF3|`?mQvH>dH=y?eVb|MvElnJSa>((bOj
zyxecI-O(-ygUHP@w{D$TT|M*3nbQXjx>`;Ud8D$W?nDRQ{_3~|p+-h_n(zDfYo2eb
z30C>Bbh2T(VKH~C+1_uNlUrvNZkx8PRrh(!)L?Ic#V4L=yDau`yVfJ9xJ6a;mgNiy
zT}g#dCdVz!PDj=$itBItc4_4r$=<CgA6sO?EPAZ_beS{To%s?S%mel9zV$xYoqET8
zzWIl04SzFrHN~TyH&h;!S<HX3GFIOI)uLKW=I2dpixgKaI&v%|qG7hArKDDBs>&oE
zwTmjVZSrpJ;=R4#VPA$tpyRuDBJum{8b56Bf97N)bLYfD?U^EKiB{#3U-%a~JZ?Lx
zthp|8tB=8%Kc6;tT|B_xzm9<;`>3dqk)cex!tt!d$y%MUMKf=nyLKz%8Gp_7>_rlq
zlTL-^hFv@6;^L!qXK83i%wfI6^b{>FC7q38td)}<%w0KoYis1q3Ev;c&(=wN`0$#H
z{#gzASx0mq%v0>|NZxi?Q}vAS)O(qy=gd4YC*|G+zuWs3l$~(xs4e&<G2fc`kiE9c
zml|F>H+%lXA6dUA{#fbo#$-{Ai0zNxu~8MzBO2T%3u|i&Yfe-B61mxtMOvYW{pa!t
z3X;qf7PSKGM{E^eGO!6mGV#1<zj^St*6BGq`{ZJC_US*JIx*()la!ROCr6*0=~p$X
zmrL_ES37RSZY|gOWY%RhS<VZMlP5c?FP&UoUOu^UVs&-(y_+4gFQ2=+`<1{#ud`Zz
z9$e|Z&Y|SSk#GJqVdd}amUovHUt^eJclh0cOCAjoYn_xd*RFL5IL;v?eE<9B`}1Y&
zB(MBf`s3L_3sv37qls;-$F8$0Pt<JMBd}QOrnkvbuiNFihmE;RBSpJ;k`$M+H67U+
zbZXPWh-Gu`&R)bNEZ}z2$3R+Uq06-9`^-k$nyxIn7_O`1sLL(E6diUuP<WT@A|LH7
z3mM-T?>JK36QkM6Vir&p_~KQo>}j)G3tg^CmN<E@Xj|dBv#E|nrCMKa&I*0mekr+f
z!9~9v`Q$%|{0ovlao)S#e^YJXn+^O68)^kOm{_=0_buv5d9rMlr^?yI&gTx4v9rs(
zyu2dd?z-aZd;F(6u`v4@9DVeGr%1)a<m&YqmV2y{=UItp1YEwlbXBS`&x*vhbjf*X
zZEX!4P7H!S{>yTz*nQvoZr+R4?@oSSZ1sI**O4<zS0|?G9Nv7h@vNxm6K!u634y;i
z-~DGf)6>OfEG(wn$h-C{!#Rn$%an2-*{tUX@PEhP^5&dR#Xc4d{%aHLR?N4t%V(%y
zUw6%anr#vbuZMt%;sgPm2?|ckeQ&;nzb)JCUUu%COyM3&?m6Xrk*_zWH9qIL%&qsc
zd#~xIXR~r7)MjOF^zA+0aM{-^akGz6&Ck!zr^O28UfS+=ch}6BGpFC2IqUBqMfMqv
z9!jemBpC#3_Wn;=;pBYxNB!Op{yje=KOU<o_PjTHbxE+yI&aOy>s$|_qH{K<sYhxH
zsflb22s*Sy&^q<tQn|NmDG{rb#XN-sR!vdqFj6sAlAJ8rm}If>$Th>nfNX9Cv+G~>
zdu*Rrt*!KS#g_%K?hCnR-bji$^8E1p^BKi^ZWS**zT{V^Oy^GnK570llHc}cyV&K=
z@Uw7Hw9M0&VdiKIOkxWQ(wKL)M`Nq&=CrdL7Bq2Km6#{Xy*z*4PU3-mLynku!Z&5N
zUMY>oJQw?18heZTq&}aM73Nuzmh`$O`TTivr~mHm?C$*T{O;oJlCB-{DM1b$Cyp_=
zz6jVTF)_(iqE{%AsWq(0Xqv(_r&B?U8iEQPhjdsII2jTQR3aQb7O^Nac{DkxEM;W0
zC}3El;o_>$#I|p*?jz<|0wHO6d1(#lyL6cZHv38jc?ua!?H21^nOzjH&S+YXv#5B~
z(PGgv+a9(WDa<gttr?oy8`5f-Iro<E#ONF?-^4<v2}@g-F4gLCQweh6yWudAM<da3
zYk%fj69c7+i2H|A^%p+zkiPp@@Okm1#vSt*o_x~!r1pPmxE8;pLPhW&g~C<ATRyCi
zkzZ@KYW~XkA^c05-x}mNCs{vWaC2nf?TB=>oEjP$a`?<M=H;GB-N$RrS5B(?tN#9?
z8;?kU&FAy|GM6jZwL)eJEt!=#na8M{<r&X{SfQ6eYIChh-rnAxcX!X5H?J#i9_oua
zAj(@?rPbNr-}hAF$fWSoUI(XhEHYeLotAJrm+NAv=~NzXmxWHfp+-w|F6u1tP2cw<
zv|pxT#zyNL0~<#d?%lgoukaq>7cdZd^r-ET%`};MrTkSE!L0`$T&Z2izqS9W{F8)Q
z;l`IWR~0|H-*2kUsjbNPb<67Cx&2$p+#GKz%S8wKX70Ok*}LGubd}qsd3X1S-QT}8
z>l#Pc2VpTym0MetTvrDzjoNIKI&+75#Qe4+zfNCV8?`Mp>+0(CuyYIo98Rp)k1z-{
zuvTr&Yu{~q^ZQd_^}WXT)DF+Qby^~6uZYXZByYK0F*0dtE;}MqH5SW#YZrCP4H0DN
zJhaT~Vyn>6O_Nqjw(*2mtF7Le>eQ)maY}~CQe{o+g$=6(LmLc5(-hw))NSEA^RdNW
zVxGa`Rvrz`a<+H8cT8PXZ!?EVvgIkn`t?`|2$$O^2<i#h?0-_p>OPl2%!@^&OT9o~
zp~E6}RmNLPE)^{%E_MI&^CzBYKmPb=_wimSlQeFfO<h@6R|m}&32uA)cE|Rb*-s{U
z>c79=%(5u)@b<jByFb3Yb#r^(mD@3I_UxV;FPD~<{&4o}>L2Op=?Tvh(;oi#@KJ!}
z*?|U`6)jz>96cf?MI7RCI>GA@$~qxZsew!D#EOHW6CGGZCW<O7*^sCj<~oU6>=2{9
z_@g}$3^NysxSO6S)(-PI*f=ZTOtP~{OIFCn!=jrv`pkLsIy)w0*|Tt;14WDP=2~g*
zPS-nQcsp-q(@d!qO6$H9y?J`}?wxxD&($tBHOvXu)!LPE;$f@AMA6)wEv`v#Wwu6&
z1?f)Tt69{iwRTz*f675gi3!$zd3nh{Ob_dwiw^qHlyoL?TZ8!JlCxXnPD`JtJ1nxN
zsqQfM+PBvowfXfb=dUPw#ol?O?)-&p?iL2d@7W6&-ZXMBPGpch#Nhqidd7;!%6)xq
zR_y$m>*C@yk6T(#KYXfOd-2c8xeGqBulKJh{(juv<ZuPM&>!cdX@RbulLEs}&x(0;
zTCZ2=^vApD&z?PdmR4sWIdk4VF6G}U&n{`K-{csygzKz!K>gAU#%#{IhlJ{<I5h9k
zR}uJM!Nc1Tvc&P(^#y_)^HwrVb!%FXdPMwVofd=e)sHpo2Xw!>oLSDim%mGvS1MY^
zYjIAh$+Cce14~au=ZXX=h$WalvRG4?RDQSRPt-&M#Z{XO3{v;pGvxctaFuth=zH5I
z3yKpj&$)Ro`Cj6SmL2V?b2l`eoHNgV%@vzER{l-3^G^gm`1R$%hoHt=iC=3&f`1AK
zI5Jkautka+#O^Lk>kY}d>S|yY#MB>WTlLZ8a@{OXCr_0?CzIq29{n;e**#@e#fbV@
zyb>{xKiF(+Y-}i!wWv#D@xp}*nxvwvY%~fMuUuU(z542_uSL6d*2ZiVeWf7aazX!>
ziKa(&Yo@1a(1b~$A&XzPZwwA_NSf%;BEWuftt*r4^YiH!e+&K*oc^JXv+Gmc!GtHi
zsR#L9P1nv(u9A~JJX0>XZR)wd`GL=h5{omIPFrmi)i3DE8nSD>S`1gS;I!i%37r<N
zHbgAb^maPb5v8Qlv1&!flJt}rXB_9-|46?$(Rjhzgfyf3doCXDGZ8*y`)zh+Ypai#
z=JuV7<Z2J`8#G;xHoYX!+a>J9(iO;15I%F?`}+FY&)0ANzF@Pj-cH7@-oCud<j#^$
zAB=owWiFe?_V!BQzFC>enxv~Q2ziR+&y{%bc%EfYyGHU1!=B3}B_?`XOiYaTRQ~)S
zB_;Xu-*mIwhizJJ6L<7-9Th#^=xKj#Q?y^Fk<Ogkm;d!M8|lnd_sBH5$?z}3e(u*d
zDsPW2R`_vwOVSY|`xRzvy_1i4YMeN6Qf=P)rS<(gf0oR#&%3WX=fj7&EDYN@xRxE;
zwl(SkBhM$c?nTn=9+whBV%O`Rm)o`N@(R<J+cW-b_T`ykk@q}$_v}5Z-QGT3Rx~?L
z+Sn{*2eaUVr?1V;%T0FnPJ43T;=LY;qo#o&olROt#aC8Gw{BRpA}EQ=MahljLg##e
z(+y7?s?X>qu(5Fe$az@$fQ!E&`F=vx2CGM|RR(v9Gu}Ne<cs?^r@Z&ribpnH>ij>r
z`1xckY7Hz5e%<_fc*EglKK{*K4sV>_{w@+dd0^EH1A}W%PxtA$?RclOYQ^I3FBbK&
zEIizP*HdHTr2a-rO_iQ~CX*h`nb*p4dHw24t<KO@Q|EBL$dF#Vc=4hxjhBm8f(9X1
zuE@x0VBp>&G3ow|AAIa)GCfT^dIH8Q@(ZRM=HXdV<kcdYaGFcu_v1d>usgq(v9y2K
z$;O{|x9oCZ?(uTHkJl&t{mlA4=_@~PLup39GoyPxFQolMoOdOarmDKKZ&*4hqFKVE
zTBvqXp6~Ormk;k;FgW?AS|?%R<ock_Ov9uCC&8EYo5BJw-xP5bj*n~7o8icKep%-d
zhnurpH69y_oLdxpUT>#v!o%&o+lp@8E!_R3;Bx0$$&(j1x0XG7xpU^Sq<1@l+SJ)p
z9xdj1db8`qvDH&^rLU!#@~pj^5xpyzy|id&RLaQ$tu=0jqDclvLz50Hy6tAPE5uQZ
zX=@<g%g|84jwTferA0jrD>T$EG=51x_vZ|M_oMxXS&E)r-*>s%XyH`eJw-+u$%f8s
zs(<v8D=RBQug`rbwDilpdh5oBr$sAIHlJw!Ajtft`_<v+we|6<KmN9j?ayDkapTqR
zf4|J!@OJgLj;-01yJBqgVwTO`GO6lI{;~}v5(a)2FD9Ie`TY37gp$d97VpxHy=O1|
z@%iVks#{f2QBmQuEPsCAw{Gh%ZKr02mX=N?jWx-FIsDB$9;=izdO}<x*0?wXhK7V`
z9rWm&bXt+g$%&<Bib&lRCZ-u1&!p?B>l`}BBbgpMF+<4dKdVYZ?f=#f>U$+}zCS-~
zay78?!;v3VdvbSu+_rn$-McT&ob&U&xy7<A%}UMd$-?l7udZ00E-o)Nb~$>}+bmW_
zGho@7%+=m8Qe77fBUUG;F7fJ24yoQ@WXvtJ?RKDR7`MVg>rGxQK@#sa?3De=b?wR@
zw-6^b9Udp?i5xo&edidMx;3vldnIq_m8%@KPdduw+75-xGn{k%;<GBz_C&>#od?>v
zTMw?V^~ma}Yu|OzdQJQ++iUzM95zXFhKi(|Y|C1;I&76$&MZF}l_Tt)yM4u1Eb1~)
zP-f+b@i@}a__4pQ@{QMBx0EA+PMM$2+x*=o=qj}}(A`ob+}qvVwI%S1t87Z(y55%T
zrwVpYzARwr`n+lZU+AT^NgNXz|11!lm?0qY#vzA=_ixg>5I2p+jn{9iV7$b!NZ8xk
zH@oD9@G@Da&DsLWyhmEKh0f%qr6r{K%xp5s3C~#ACOO^M;L62aX>8mP)4IB^YHmBW
z?dnpUO;fUSH+I~1yPca9xb0f2ZpO{fh}Kn20udYB3|qH^=>#@4i*~G<*lBIBA#{?P
zib_i&!zA5Uw^FN2ch`%{PL&_Za4}x+n%j8JljZXcn=#%i)|e^i@8KagDfV*x%gg5)
z<W%M<d|mNzFIPo#tAAR(;;TTria#IgGHNdTOz&W5V91`>-@kt1gX`C?r<^SFnYQcV
z;$yqBx98n^d);Q;^xdUzZ?C?)ySV91=we;j(9lrNEcLqw7L^DsnyNDCjHk+^vuADo
z{Q2CzaDlC?;7gw5t1)kL3LU>4vg<GqxTJJc!KG=z>o)-k?8iU;Nk49v`23W-iHTjp
zfsI;IbJuQDc)D~)nn+k$U{=J%KnW$KpsvL-CwH2-?0p$<BqS`<$uU<uXa#S`rsJme
z=~hd&9m~jW<lSr-!TLpCVtaXszl7rphvx^LyWIKovu4-vg(vS{QD>^1eD>GBUDJg3
zxTi4eR?$&naNs#1z-H0Gcafo+Q)QZ+ot3A`B#%XBXX%(`b!ixM>7B1U_VwlSc{LKs
zEk_J{3X_g@*CbC!T*H$*E9bHAe4EQWOE127QBwT*^XJd+i;Ihk3yOY9Xh_dk{a)hJ
zB#n=mTv}UiNPLv>I+`^x@LE&%#EBDwq9t7>rJ9}0`(!5Ja&Xbqj3d^H2}%lcFEt%(
za%BxkW_Wi<?)9(#T|x`EET)(R@^JLuJH3Ce*;QFJ-Yaj^9xu3a=ZEvc)DMk%i8m6T
z9eM8Xk9o5H9ZL~Qn_2cd)U_vvi_Y86*M8l$LXmSLlaFV&0JB1iz$u$&k;!R>TVy}>
z9=Nm9<c9b?n|)s{$J^EX{MhDjyP4(m_lNf-j#)BGJiUK!wo%T_?a$?Hk{9I)=iirj
zwc+6Dd6my!C1t)&dinD8lP4UrimE63YPmIw`q#D}7r(0VJm`SnX8t+aQ^XdrC~SS*
zD$pSw{Jq^%%Xp@xvO>=B#!Vi=ANQBGw4dSO;J^I2=6YT6OS_p_hhNU@2{c#hI=;G(
z)9svo?#zj^4U$>IG`Iy>T@ntYq=dLQCT>)93FJI7X_JAX<Z|5^(wx$kTlkN*g+$j>
zS-f0hBKKRkQ23ZZxj@a2ier@*<ao;G`)97Si~WBxsHXVwj}_uAS0Wm;=Vs3Z^?!EX
z-hF%bot?#!^Nf9JYD)afRSuojSRv4vb7PXH>P3}FzN^E^-Z-z1pQhU_`ugN#r7bf}
zmll1xG286e)2gp;E;BGPuqJXaX*||pVGj@MS|{KVFeBndglmn%k;I;-T*)G#J)+4S
z9UKjvE=RUq(bd%AiP@OYq1mSG&F$zYB<bZUojUbQ!P}Y9ts6pPx366$y0K#w)3&mV
z6NdcXJer#g)L63bzh3I}y=;4^evaC$vgm1@=iAMS%S(1NpL%lH_G#fN^AE4;4mz(m
z-fm$d$M2JE?^EYdWA|3RN$yW=&!0zK`=`Xkev7P3ofhb7yDy%3v8?2Q+>iSoJb19L
z#iHQcelgAOkL7BbpPZaj`P^q;<>d|jHlJT6ENc?gzk2WNokUsIF9D7$rU6bHzxJ<B
zI>6z;#Mk2B9;ekfckb-POBZAcd8T`4AI!P2Wy{7C<;4pZ&W$*e`*Jc*hs4FBWhJ~V
zGkD5Ay+0;r_-^ZbX|YeNPi#&vS++u8$GZzpmwbLOXIkNdUom>(g%7UBPiV+-4qJ3Y
zY)1zJTMC<-(@smR)zen3xtw;WH>Bja(z&@>i~J?3s;@7%dsy@HGy8mwA29(J93J~l
zRJfd~XZYUNn&oq!#p9Kimla;+HCeRz^T`7T4xIh??BTO#zxy;+DycNG32Dhj$Xzh_
zaqt%x<5vz<E~esh4>dMgKGA4lX;$z*@=ryP`Rc15|13W}JU_)Q^3yD}4KHqOIdk;T
zDJ5-g7m1U-;d9?H9d#ABQEHuNyv57a`m_B}oj(qL&RkJHaj0sMe~5pOpQ~&8n#RzE
zMe0X|78&@rE;{_c`Bf^5>*PNQUEY%Kme~FC=2+Cyt@!JO{CcAU5*=L{6ID14K5(?P
zYMOmR_e{u_8=sT+w{|MDH2iQ_=saD+dD?^}Pd8Q;TI&Q)7TYQ0DPoagQ7E)xjz&9|
z`g-wo*C#BP#H{$2ZPmXkUOgfIt)f<~|GY-WzCk;y-s(cn1XlhshIwy7_{%uj!<5*=
z6zu*8RBdFj<5=`Z!10Gt*Q|?zd}~<v+qBrjn08%c{qbOhyo>;UpVHn42Kku4stv5K
zjxhXu&>(LU%Dj?=`CC^*tzl@@guqvhj;0GZe;sxI+Qn42A-G*syq(3BZ;2y!#|pkB
zffg57OdfGBXK=XG(0H<=sYsz|R-g#z_Ey)*0tU-M(MuicrnvBSC|DJU`zd{1)P1gD
z-I37BkB*ifS)LzYxYEECbI{}wSBV0{%L5LV8`jMUDS5Pdxd_9aF2zcQkV{Rg*R$Mc
zSXeI-_D^H!mjK?5t}l<cBm^C;3Rw&in08HIG+4kit3c4;fl{4;3$KTU#e_gpg;iOu
zF}41aubptL5}A-56|kypZm#EbCe?F0FRq+vHG4A;r-qT|Fa0!uL!yi1j`QD`s#N2s
zBe20iqkv7><$?z5s$&bgS8iIMCH3oz1M~7-moBXixKfp?$gPvPrPJircJ^&Ao?JRT
zneE1m)43b`L|HaQi}AGAeNR@uC}8>HMC1pf>J>Q~R!ALi(0yLm@7bR_y|~eytz>!U
z<kCm{d-Sb*uTDPO^jl`xyn6LXpZq!G1U4RTZ>iC$x++sLr($_&lfb+Q4BwTNlR+SM
z*7x;hyS^U1y*<yRvheNc>3T6ci=Lj5iBLY?ujJAxq<s9K(%YNisdvvtrf#_!HdQxN
z^YXHJNugR{tES~ftx@W-;!tSf$a*E;!~Ch`I}^i~nqx8-Qk|q)yjTM>6S`;q`|GYy
zFR6a6vBe?j@J|lIAHOA1{u*m|tzhxkvc5bZkj1y&sf&Bz!iDprQ!-REYA((E>|J(i
zTXSk_5|_x^rA8r{6)PvrlsKrh(OPABiYe;^H`BFR%wI|-dwXq=oIfdmXJ+OuFN@37
z8)nRyG4+nUz=hzs_o9l^-Ys0x{&R(Qhuk9P$tRaDKBmFG@Z}cA8y$Q<7z7k595@&+
zbO_AhXgK_?dhWbSU79CW&E!hiu((uaOJLh7Gv8L3okdSi^-U?;)#Vx3nLe*3u+Jv(
z`{JJDsWU7mvze*$BugB#)z#HLefq4f_FL({va&L&qTe~HmpDFid@7=O|LZ;{kwp^n
zis!mNDxB{-SN!GZ;SVtfHN8Tnv}b2rjk=i{u+Zc|k}D@iePg5ItDklL{a5*?Ri1Zv
zwD*xjilEQYbJvUR9yUC3a!RmrQ|iOE<XsjiOGQsQZ81~3yH)OIpU7e*j>j%(YjwpA
zZ$09Y#xPMZOzZOA9hWq#qcqwKB?VOlIU1eX_rL%B)4rj<^5lsXzKIdi9nzKEvd=b%
znA?6A-*>IBP(yt0x8}-UtQ^h}9^a)`nHxNJ{kHRprT_AqSs&D;Z+tMFr}Ca5P-nsd
zZ<hBA5u2PE)&*)$SoK8v%)=ITewLe?(|+!L;XmK5-ZtE8cj)8urZ)xUG%r*vetxRr
z0Ozv#jz+?r+b?v@$!jVx6rOx9c(G2K=#kZytvgn%eZq9*>Weh(4mFO|p?$5Vr|V_>
zIJL1uSGCaO>XjF24op?M_A>ZKukF8f;>3y9bw8seD=V{l@7P>W@o1Uzp^5)6yRmr-
zPveCH%)eRf!*wQIm@dq-Xzno;g^O%^9E6jTPX;Mv{M<F$cJ}hid~;owIBg9}OM2P%
za<<H+Nj-0#N#`XtC(SHeE$Z@Ug^`4zM5kzg!xN2DsvTE_yuH0vyPi*;o#|ls@%~BO
zg4b=Q%<~^zURqij^oGTGuJD`*Hi>5{=N#Lv+n%8CamSU<f0A!@UgS{WUgN;LNNR$#
zlY-g0yOB9@pKiQb-M(kjxtp7-E<Wlqa}^16y(aFyyR7oLeVv6w!6BO;?GJmU%szfn
zk+TfyHH+DqCTX(ilve93lbnc+@Bb%%-yfnNzvsWtf7_iABH>J}C;#dgIPho(u#0K^
ze5jo=;cwTJO^&WMvqeH;a&mg|%c?ILYhL-{t#5QoA}!_2smS!)D{rpZ2x^=P-9AHe
z+U>%lr^Hk?oyoYc;mOGu-Cn^h5o@$rCf{5B{{NxBXQplbwP8#1FJmr-=zqE9*T3xR
z{rdIWtKa*kNpF(XyrlA8mG5)Wl+XJcz8pU$XSz;i<%xDBxySP1tq#Bbt*YZ>+M~oG
z?&2_a0%M7bl~`0x`K`prMW=N(w{EzzWApZjVUe$|%I`nS#KQjmN}z|Ab33m`r?R9-
zvb(FKQN?kco0HX~@?_Lzw@o;{wry?HRG+gKF6=*l{@{W0hggaPB-U#@;bnN#WO!6t
zg!LB_SGChhmNhQff@(EiZT|dPZK55vHhQ<i29*wmpXz%B-Szct&lg9%oO5BVO16k1
zSJlNCDTl5v7b{507G_;^ko&@+sWFyPiyrg__+5E(Mnb2XzhC#%iFFbmHum~Nlus6G
zJ0({!@z&vE;vrTBekFWqpC^9b{k=XpB%*CaR<oZ@#lHH)ui5(7#eXo#>+nx;;oQh@
z#U*)hSC_?$CrW)mmK8-8Q@YJEb~rv^zMd#vC}8vX{(dW#iHGzg>zGdWg#DGY-o#^Y
z_OkD6i>ogJB#wJcdMxXDeBaLJ`b!FL&6@pue*eme6O`md0_`rc^dCL<lS@>oTI<?!
zgXU#7CSQ5I&^sfKb<3&r!!sux_+ZU0<m#H>96YhLM<jryXa1>__Q=gb#@hrpOn9fb
zCPK~RI$OTG%cGc}#)Ji}0#}5TR<2*aK8SOHK!B;K*8hukGk+VeSv6(r&TgM+=WpJ<
zoh^1O{1U(6Qp>4NPraO3vMlw+^x0D%&WxRNcipGFX>WbbsND|j=grAjI>l$xm2G_y
z+fJOkzuJ1M!sG=_wf6p(?dH#a_QE(q@u|PEzsV#vlcViZ)p&%N^xpmIUt9KnNjB%+
z^%oLf+H7wO?@>H8C!^3pM1EuE*T1>qFY3DFt@-<&>FMcd-8l7l)jY3^ovz}&QCpvT
zs?1q>>aEFT|9R|I8>Y6cu}cl@SyP()Ie6lXCieLzKaZcYt2{32@pv<vU0%-R^XJdo
zyp<K+vi-{u1%W~a0Y%MQ4hk=M9A0S29O4jCY3T@Hb8X@Yb90~X)U`lz=g!Ey=opT4
zQxTtOk*|1G8*Gc}IkZS|)x}w9w~NhXc1};d(XlRC;+eK-rnjJ%l+N4*r;hQ5g=uZ}
zI4vmR;;|uAH0gbIM@Pq^Rg2b~UCQBX#>&$salS&c=diR?+wDj$mz(cmY;0_-)qWmY
z*x0ygYRTeN#z#D2Px%HkJKR6?R{f6W%fNRVo*$lj?%CCX^T~(U<(51u_)%Knl78*4
zuiRzk7Y)W~vof~1w#?b-d3l-aZeQ~a5uoX}llFC8K29yO0tH<qK0ZF)f4)^yQ+21I
z%0&0w&b!@rEw#!@(^wjE+1KXJmr0wK9+|dmQPdi>sa0Rk+nu~K?GW>~nV;;h`#(%v
z`EPF31Op3^DW>K*WfCX+p9*;x*ekg3KWFBB{z-6=UbMkNr%zL|nM9a$l*BGNdj8cs
z`RSrSsXhDUAL?(~-*jqjlh|{oYT3MNr)6ely<Kzg&aF$c65T=rG9~#MHOg))HcbuM
zt>w9VsoTm8l3&W0v}dv&oT1tDV0Y9po^n4nR_AZc=CN|Uj>YHiUzsz#@4LL!#zXBl
zCjLyS5#De>^z@v&tG8Frx%K1LuOC0Y>!sXzmUrs$(rKYCb9NRzJ>(~&GRafrD~r6H
zQyhnJ`Oig4O&p4zoF3_lW?2&cR;`RZ$r5MJnp`&Ly>!W(Z$X0-&xEe!ZyecYJiqMl
zEct@eL!~SGcC=PqHND01U*2iLN@j;-#uo}DPOGQfwmSTkGgf7NRNzEGfAz^H16E~P
zSsts%WHFw!)6ev0ceT%pqa|l%?M_R)8o1!e(`U)KWjnni)O^L+B%*U!Iy+QXdxxfS
zhpcsUb#%Mwr8Ti*+QqHgZs!WPL|ii5=*SWh@wQdWQIMr;5vPEwgyhdJ<;q46GYxMf
zI~Tp%ueSD?E@-jt{`3GRr*EvxheR@tBud^*+}m7y-Mja1%F|CjON-)sEe#(3nqPf1
zTaL-`(VjTn<oD?foQFF9G+O)cZ+@LI-}H2^lpc5#wftJi%kaFryJO2YhSvQJ4Goie
ztzYz{b;+7FJF`};+Pr(0mA2Nhq^Tk+gO;tDbMxlSlg~eDESB<m`El*))vH%+IvpN$
zv`uVlW@=}|(tYpmyNDi`q^P8LNnwUuDx<c8gWu-mSL`{~Kh#(Jt=O{pLz`3Fsp)Rd
zvTo}gXM4J3sjKOtpuW}7lQqNje2jHA^GymC&)OnZ(B~2xdR6zx5$oKU5|>UXCg;q$
z=qluOf|GBHm&5B|f4|l(#t+zbxRwPgNiA%fa;S5a(`GIP&BLD>_L?+3JvUQl(h1A;
z4cw-y1;T9lGJ}?6vRrBqVsS|A)`@74OkpqHxqkin^%EzSy}h$~x}I^~rqepRTi>mF
z_xAW|yU5hgrX?nmJXIDi>ayYVaTR^OWKzgVRbP>ctg9=ghD<tl*4;i*q@<+8-{$kz
zQd3itPen#WC4Y048YwiEWiDh{V<{rAFv4l-3Rdp`UN*IOVXIZE6}Dd3aBZcWipYfc
za~ztcFV`}<yId)ibq`#<>_*hr*VmS6t@2cy*}m3TcZzAKo2Y1*A#dn3nP#Rlz83%d
zHe{J*WnSEV!;2w3J!g+|w3Vg6pCZFo$NxknANbH`?Nd=@ESLQF`{awqKD5bB*z?{z
z-}CD(TfTPNttS*5l@BmTG%$52u?jSBv@jq4zQ8>1?(W)h+2;~w`f+<}Kfk_sfB)X9
zudlB!I=WZNB;(fiCms_)&3BdT*C!{hU%%ey?a^r$qNZwvOj<J4DpNI3Z+Y;_EBp6f
zxpL*up+5y*O*V$gJ9B?uAuE?!v}{74Nm7%v)U8`r8%$!Jc`0o<zuNUu$<1Y_O_nBk
z2pKw*ZR=X3w6^8KOZD@ge_MY18EG#ge1E1vVREaCX6qUW{(0L=&xUpsRocvR`~FeK
zsYvnFp7P}!vEt{ByuGkH&VZqD!K63esx_6mELIq_+1XjKf4|`>mLK4_IqhtJBD?y$
z>dk#Mnad<OHl5Zp$vAzyS7OD-Qx6|b^3<MecIV~k+Y5H8O!Bmxw&}mj`QwZMXW1_l
zDLTclCUCV+)D+tksuAL~MA<cR_VJ|-8h<W27=GNiDQ3Eg4%>x=fm+vIPFI*YsWnx{
zS#;^4ZLeb1rKzQc>Q0-oVv$OcvcTKa=A>1uK2BmO*?~vWHVAWxrM_W4regMYTk93x
z^+xjy%=t4{zO`M}F5iF5d~5UT#_#sU55&*>iu&(k*JQ`g@-st*qb7jA`N4wPHFoR|
z7XR+35viK>#cw79LnGtl$sJ7`Mz)il@34G%&gOB4K%ceXO}-NX)(Xi-7!1xha#yjl
zF&>;HCE{?Qd!gWkz9Yd&8ih4Bb$0XS&DWlK>ioHzbyi<^UdY_6lzEY!eWCc5<@fN=
zUj{D@oNK*SYW<Pt$MgxtpI@kcIo~+n^kL)r`9BjMzJ8Q(YsQWpJ9dd)mHop!<(<FB
z*P62*HJ^(y-<xmz{L=YtucEK7z9>6usfXAfzKz@JVs|>{Z;79DtubFK*=iEEceL)T
zxih&1Zg%~8D8pyTVYF$@9ifHH0<%8oUS62uS=duJc~*x^z>Sm_g%>XL?@CKyoOGkb
zwNz1S&W1;($7NF@)+TiEo{-ah{BEB_3m>zJp!0?c4+I@8o^W0G{oz1`JqJgyypy2P
z@d*qqzR&)@`pp&8_G{N{-)R@_2yik?G20rpWW|Q+h0$BtPMJ;F!j!DK!>RGW;|9*3
z&y_T;If?IBCc2(oMT0%x(V|Y`pkR{fH@~CvHk>=J;-7y$y?<RRheyRfD+d1?zwG)C
z+o#F&*e50QKjacG&DnO_GugMtH!N_sckb#*=Th!;9n@Q^>1~}@ommj#ba<mA)A@t;
ztiR+Jf8^c%$h23DEicvBti&i@ASHKo(xc1be2p)hAI(#k5q!jxqw<DbY08g%yHD+v
zDA}XwKl|$Pf=QgEBF{F~{ojA<+t$7N^WOR1k&OCT_3MXz+@8C;tGCC*79U;rK0fyG
z@#%T{R{j0`{r#7z4aX*(4vb7)d^OHDGOTpzGoP8uvwb69t#%C!3G`Z(zjB@;Lqq5h
zhYmp|0VV~B#?ztiTppifSQ6I4_3G%7AcyRb31R|sw%lk*5^@TRI+_$9=-c&MT3Ymm
zN7X+5rhO|HR4onpDz{+ksau-KI;Js#EA+f&deVBlR_KKE1>F{PU9h4(kSB?Q(`@P_
zo`B6F=Q`)E6R{7IEZBPdcot7?Vn(w6&Cj#4d2AS*g)F33)$_cWenEfLLj8owxdqnD
zrH_O!Y%n+2*?eZs=5Lw5?I#3U2ETv1<;m|Uj}0s){#IxIn(_PJU)~ORmbfe7TUqS+
z`x-y~{a@>T|I@cJyZqVje_8DJmw$h=wvd0FMo|Le!NwcEtwjqK)n2rFp)RC+xIccq
z1K%FzUdG1`j~%N%i1)Kx>uQhFD1FejM||z+e|ma)R)4b{^B-S(9%y;twS(0~zAYjz
z{MNWHzOrlAu642&B@e_4gWhSD^Li|<6u8_rN7njntq-rQt&Mfy;!_!)cG>>>5OUA2
zohxG@C)<%J9TN{6w!O7b(W2pN+hf^l>F=MXZ+tmH`9b+RpZ9xr9&Y%aB^xQed&<U>
z5hl5o6S_pa8j@JwrmUI!M7ZfxnB<O*B~q97Kh9e@>GahrbBtOOr}zq`czk(ZzGgxd
z-`1!MC98}!wt^QHB{vxv9NL)|uLyBCS$SGecZo=ltSO^p&x8+kk32V8tl_#Qn6PBh
z5#dDw{}p~GzN$(TRE}6L%5w97z~w6!U+zkbUv@>-jL+aq40l-Tl*PJ9N(w!RlMZAi
zoz*!he>Iu6U|DXJcBeR-gz%^G-&%c*ylX7$469ES{&>0j@y&lhh0o;*f7}UMA8M=j
z`SsR*hF>xEUK@TK;3!d&5iELodd7|GhF1bzMSNeIMA>H*p1r%n<gmEmbIt4-iEU0T
ziLRnjCK(emj`d2+c=tkpWr9$q*Ydoi*Hi6kejfSxXq)RMrkhOMDXg3H7SEJ9areNP
znUlWp#l^A+Pe>OM;$#xg5J_lF*nU4nMd#EKyF33Y|J^WR5dQvOw!!(qkJ|fH`TNba
z{wTDzAClBswOEv^G;3DGq}5g#J=;XXb+>E{bLpL7x<H6KMdL|wcH1t~jWc_BGJC#9
zCA%EG&XwKr&UnXR(YX$HUbm^{c!*!WXZCCR+?!RQ>anam94X~R7x_LkIc`W`Sro|1
z#d7!V-Cikk^V#4EvbgHgQm?6A;hL*subq*(ySung;@HRb<pw|hJ^pyW<T8&@p{vJ)
zklAH#PjAo5xow<tbMwaO)2B}_zw|OSHLhX#(uo=?XSCSaUJ`iV#dv_z%Vef5M-!`;
z#A=U(eDyT{Ega2vZpixgq;M+<UYJ(F<LY79aKv*`r1~v2L9Sj_T~CijZ&j`v(Ttwj
zQB2A?E3P{nXq~h`;hDy&RZ2M<dR3IoYFktqRi^ZC@JKy;DI&G@hNoS8i0%F;c^+He
zxmGW%`}4YrGu!LUH}0SPDB$grBHnFsa&gvt1rB))4tIGIwH-3DC791-9TH&W;d;>U
z(5a#MLBH~PetxOa4WIa2=5|ybVxMc9^LfdNh!8hN2ZwJAM()4=KC+dwe;9D!*SzcV
zdA!?7{Cm<jzjj`>x1Z@&#N&ilLDv#Lc51O5ikRJ+?7$(|C3wqZNy|3gUHM|^a~dsq
z>?e429PJTsaS6(?o_zkHt!$dvOnwF{{`AlYC1pXud2)97pWhyd&`}Z7nPOCG=AY5U
z)oIQ4a?{%>(u|w^Ud+~LyO87NA+)RSj+5O3qZ|pI$w4lT6FUS2Jq|2RFb{jp$uYS?
z?XFH^TJ%DXEj<kZ<r}Pp8YWvb9kThr8eg>c^W1;SV;^xoU~RT^nDYKkdv$yDxwYmS
z&$ON^zt{8Hbg$$x&WoIH<u+*?Q}OWUe9rf|CCwwF^v%;2&Zl2j{oHfqZg#=tV*i8}
zD(254mM?$Rz3STfWfg*<pK2Lx)brMC=wW1GVNr|-RN`T6@i?);B|@P~>D}3{d;0!-
z{&hUz$>EA!nKSp*NGNtJD)Hl9FxgTd>ac6~YQC=EP6mNPFXr>|3iL25v4}*?d40Nl
zw!Gq<n(K$#eLlW__SrW*mMuLk-TIr=>jj^KFH{S#F}`JD-|Aucw(Dd0E!9nK3o_Q-
zys6^E$=DTni_2jhBZCO&HqjeLHgL_dd{?VBd&zZ`40j(LQK>|y$zDOLSqzI~)+JAS
zJY{L9;I_8Vt($JTZd;RO$dVN5wra;zfs`$KGLGv?bTn*IY7kLY<dAyo?C;mNA)6uR
zuGEQV&JiBVTnzst{#d7XZ)fY9m94>#B;Qt7)C(1C2s-?%e4gs-$d9~l--fQRU+Q$%
zWMWab!tvue!Xm8#bA?2^oZ8g8;(1QavD1_M{3HFthYuevFW+<c`T6-7%8X({UeW@O
zTsno63X_g{X&FxFYv~ly@;rC-;>@I!vzD7nzI<Ji^rocL)Ku@?m-qT6?@PYv@7G}n
zKghuM#zExNk>f=(`>xMt2-xf`wIqFiV|H3-_q1DTPKUZBmMnQAyFuGZrB8j={Jz?o
z^^tGdzb$dQ$KFxpYj!9zc+1^@=X>5|2UR#9(kk4ec-{HN$0GvrkJOkH=lkpHFJd~O
zz}nm0!Td%lJ>}f>jCFE1T}2xfZ4htP*Ew`^Q%a<(spdAOE{)}>B7v?BjSMqX4ynwM
zPMsFVBeYZ`?9QccZ*On)y1IJxnmzlEeb(3et@mG>hgtpOVa6;C`Cb)uzd)}ms}^Xy
zerI6hr*fTR6NhsB;Wn<MrH}kp8?z^R^enlv;Y8pz5tm5|RV579P0_i?TyW>D1dri_
zSqDAp7)?3vEOFvH+`}z(=t)=mgM>umEjw%$M!K>uS<JCOP2uL>a*2r@Pp67@7`PSA
z<6Kjssh)6n@ob(4=11=GwjJVWZA{O-dsl{CJ9klgqD}IC*T+wce9~PMjtS0vd{<lL
z@X|(A4uwOCHyq6-b|xwYwTP?{)D`JgP<7bXrKvPqb)DQyzXhdMt+NZ(Mc!P$Zktro
z4F?^&mBPn3u5`(5n0CtAiFNAD%TL#=<vm+%w(L~LwGOsK|MNMQj51C$-M#5EVIk+T
zH%kKqyF69|WK0QEaClrZ@AD4sUjcr8*Y5J}2=lI#s9Y+L<Pj~RlXJZ+=QWF#!>U$J
zwhcG6x#Vpe%$Dw2H1WaZvMZgA94n5v-0v4JkNN(@J2_*vr>j}|VXLj82iY!h=<Z?E
z66I?Dx>j`qr_I^-_j4n17QL^}PCggAVykx2#@v_eGgX2Xv957g<C^uQW>U_=mn_U2
z;;+S37@1i9swQ-D+FQ)+-_-cRGh*Q#hu<n60u`1B^mKG{C~|a6QrdQ}ckiaf4BS=u
zbLVYeKjGtdmT7WHjjxZ^rz(W;ghb^g=dj;low_tQk9nu#&9A<brd^o2*LTu3!3Vjn
zcg$v&UyNP5#Nn2b=!!-W%iB*m4vO`!^LG<-_3UObR@m9Z>U?xk+2r)iZ=6q0nDKq8
z-MWYg!PSv3f)dW%^>h@G)%j-?#>(2{sq#o5W%h=K+A@~}4HwZ4o~9)UYr;I31cR=Y
z?N*507|?Nh-WuKmAtGPR%ch<-u@Kl4x?Al`+X>(K3xgueo>qC}d9BvyT$`MHagCB(
zdH9F(x8~p0|Mqk)v*zNTNf$Ts%=Dg_t6Q)|Kx5TvA*pVq?n@!ob#MNzKm6nH@{j3{
zZY~e9`z-4r_sVxovCNuP^TXL1>-Xiq%9lQV!)FsGOS`Y49)r`yQ_La@3_Jo}3M>-6
zSn>JD*Vo5)|Ni~{`=1z*g;zSu{`u|WzP>*0;D%dYpY!R*?Ok5g8hyba`_SU3cUy0K
zTz7Y(l1uk2qb<d0*)jV3(c+7{C(bXwvFA;h{%+a6<H_8g-+ZgK*gK~%_>y3QV3LbU
z!<tUU#5b~n3w|ACd3G^trQ_!-UOPWzbt=?JnuzRZ5^Fl3y^(uiyBn)d56k4LCDt5L
z&M7)v_Dyvyn-6v=O!#Q^Szh<x`$WrKS*gz?jg-_|1U9MMh^ji^d_n&0ep^?;J^WrJ
z=e<%47CWw58N8zQb&hL@?t}?krvoz%<?J#IbI?*&R_H9tX;?aOo!r0NO`;3>UovVh
zZO+h^n#Stjs?;#+76@EpU|)Id)&88ni(67P|Lde|H<m8!h-{lbbNOvv^R?3HUbh|_
z9NL-V)^pM?O>*XJz3T<nE=_N}J;mpg&eAFGCLB_<N)=75P;}Gw_6kx~5-9nTbkQVv
zA;Y<u*RFhAINM8px5Js+KMtJw*yiTuy38v1qIpI1tG(qPxv#%i`@((cUExdYANia2
zC#eX}w4AVGI?vP1)8%JJ@4oY{`rh~K4?pxxs+FR7yeGVKK54x5zFlqU@7L?^uUaEv
z5ZcNanz3R@%H>P`rkX`xcBGy4_9}U6`77hXqArbRN?jZZZSCc*vbriIUoJYYzolU6
zz@k`sSip&Iz5BwDB##XWM?WVlnY3$qsGDC&n47PJXY)VbNnsq#BEmTb9NE&d51c(G
z70VK{ans>-3!`2%b98WUsx%a6i@I%CeuBqA_acLGfcmlna}WO&5}Gvmosf~t_eUpw
zFXfuzWnjU6Fz&-rvoN!%tmXpWTwU%QPGywhWbl$(SINNh#8!b(X2K)ct||7mN{1Hs
zb3V80ep@#AT4UpjR7My77)ha9s;)=7H8WaS8Z6rw=l<!PH$P=IH%C}jRL%^`Fqciz
zGhK9VXXP?Yxg!$Am}juLK+N*rj8sQqtxK1yXFOl9EOAGzN<)I*{`+yeUUNRT-M~Ea
z)85jA`_hd!dQ=m)iAh~KwqaiEq>k1XI{%BGCf?;cP*TM?^I^)BnUdl{%j?(FoBw(F
zXG4;pvOw>iWSz*qL@((K>mZ|}4|%61ESX@|$LjpaNmFp5K#-Ecgts=fwm0V<xpd{a
z>Asy)X6VmcJvZ*I(YN#fqi%&trenhFtQ{I9Dkmmys6FqfVlemIasU22i>!*joBdz>
zcE2tuD$4w3!>Vb=4tZ*6iz+o7abt1l<WG~&@KkkeY2sn@V2M<T*u>E4=;GR-aFkIn
zuxm|eQ_fuZPxCTe*6r(AyuC}Saw4P6U0&u(*#X8kLVO=<TM3&tAO72Iyi{2*e9h^i
zbi<s@1v#mvv8y&sXA$0&XmDrS?XcUD8gp-I8>~3SwMwC7X-7+!D^uRr!*x^Z7B0GQ
zjU&^bctO?XOJ6g$&s)B6;@<4!gbN4TXO@<}+jsQ(A^n4y-}lv@J^N#R^WDv_{%rZ_
zR4ifBrY_rf?6F0IBuh$)!oDEO%Ir<9B2jyDZf-H$o_FWk(`k`;*DDgUr|qnqu+Zrs
zznO~D=_N5ur>Dea7Mm=)I@hXf%FElur*$@+KBKqPt5qa?gVwhW1w%t!!Oq^Tn)O>%
z)I29M+<wQ$rP29qh5V;H7l{eQD<qYcd!?K)(Oe*@)3G6b6K9#>?nVbE!xc^wTqG8A
zDFyZPMYxK!1<4*$vGT0kqd2jR)lFv7Lhoa%%}#_zFqh0^T5df_EFsfcV1kpxin`g#
z{4f7hFEPticblO3VDf><9tToVZu$IkE$QJnUAV<Kg7>FbkK4^D2`n-K3ui6Nic&f;
zZ(#(t!FExugMFXbwrVPFS~GiwuJ)hmsTpExgfqH%&-essU7MXSby06%b|Z@=!;*;|
z2XY;}W_>hkYTTwh;e~jJ|CBYC3j($)YAZ9XTE*v~t5?P7;l&x9raz-2EcS`d3)YQX
z?_TG8S#x7!%q@E_k7*7EW+|+cdoX{hy4{77sl5K_b-VN~mo;38U{p=?2yx}+33OVf
zBzJ%H$3#W`$q(LsI6uk8?!fvL`qm{i4Rh@0J3TlU66&>L>((iorU-SYrzJei-LiDC
zYn1i@-5F~mmTGE=g)9o(sIh9JuTx;85yz^P%^E!k^Eg7f78&t``b_Iq5o*g_mKrv9
z+Dtpaj3*naHE((ew4MIRoZkPNe@D``x@&Bo5)Xgl-}<+i`5X6c{^bY$Hhn%{TYi1>
z%d)uU3tTJ#3vGgbnY%bNWiklsZ@kcVD%*PchuEB(TaSHLOqkyke0^i`>-`Iz^7e6G
zG)TD7zfn2aF6_wO-p<*=;&V4|Ix0JNR<-4<ZP~A%r>8x8_9{I+c5}(=?WH??^TMO2
zolf1lvu0m><czD&Qau$^y{})k^z>aV;?HV6*Qwg4E9FMO#NAik&S#mRe`n_Tue0yG
ztK4(`>DTLS4(c}djOV81rOkdQ*etQgpeCJDFvVe~7tc(|!{5&Gu$Z6JPc->-m}kPf
z0NqWF0)nlR+;oKd5?|N~SRdYcsWhn4%kx~^UD?#9bC;x-pWZ)zfBxCR=dP6xzArub
z)JA!3*e%~}7jITwZMl`pW%efQYR9ys)~efDD>d_-PBT2us=QQIZG5p<M|aXDgP<PP
zhPiue4Cel_lF$*6xieEtNyjr<<co(`;DiTddpzcKIebi;XH)&-ncUXp6Q<9+Cfmuc
zv5HMh@YBavFMPL&1x;%AdmdTgAFXN?(jI-|)MJ~tik~|l&oe%v<m=5KnCV?~>FC~B
zkGIIqa9l9$*xkqs)}?DVb9IzvaGM`L%zI0&sm<R$FYU+7hv$V?nFX8**!C>oE)
z!e<Yaef?c7Jdg3N^W25Ew+P-!Ss@{5@#KZmvC1EE@{exX)a=UDiPY^(nsd8q&AZU;
zn>dcHP!ed|A|<l5BUe!&{^U}Jby1xz9BVf_Y;wD(RW^aiZz-RIQsis}qqFTU^OzM|
z6dRHwR~>OwIG_+AZMHDJ&^&n)?+NaVKMw?=X8n7hyF%=dzR;o8^4hFPtvlX2mnN)f
z|7oPb<*3>a$;s8lED#~MqAikNbDpB&%U9V8(k=eAf8sfCAi~<yvF+et-EaGL@A&k2
zy{5_6AF@2<*CZ+o?!EARm61^1)9u3dCFh;&0<MD|)7MyA)~#J<@IcIIA%7O*pZj;t
zOYmk|J(}0Tc5B=B2P}VmUq)~Ed9+|nH1qZl*&P!%Svxcd^i1gC5IdRijd!kdmh_6A
zq{c+=Y>iyDeb)~3rrxP+5Df9$xg|o(`HkvE5#9OOjPCQN$FuhQ7C8KuWnzR<L4(Ys
zlzHcSz6)s!1Zgcgro^Gd@!l-|#)fZyn^vxKcvY3mP<185>*7{(vo*IDXZ2mXE3hK>
z`<oS?yUqL$?GgQIw(y44cjoOqIg^?h0{B-&b8O%`shnrbo$x#D)sH48;|cSgnq|zp
zeSGt~gd|75!-^NZ16ZCixd;fp@yk2=#_ocX0^24}+s+4U6Q(cV+rs^b^SgO%!`yvK
zni_oPRxWwe(J9)=DXHQqvS9g=Dy?a|CzdN#T*%RQ`(WW_hW05Zm?yOuU#?}}l)v>&
z*>B@*>{m}sZ$6(|Qm;R0LB?yvuReQg+mG(P@`%k}F7vv88_(?fKF|9m?~i@<oNvZ=
z^{SKBlQzlaXslS{IqkM(!HZ(2!#XDS{%&UFx_J5E*$<ZeIm^tpHSMl^%HVN=^PzB=
zmg55lDYLFX7p?^py4n{DNT+Pv6|&OowyATUi(SU50wMN!3sXE)8x^V^`CBo4P}=_D
z0P|VT$R;t@=I5>xv?nsIWn^y1Ysg?Uk1iB=!sBALd{x!%1;JIW0n=7Uw_UOM``BfL
z*5k`*?<`lEyLz;}SYht6A}PW7mSNDmn;}b^RtX(hBcc@@5S&u^$xV!D+X)>mmU2M>
z+5Fpo;=3Ib|L%Wx{_8W&zbET+^q))1MdzxnUOHv7_$pD+tFCKfwTm@7o}7BJjE&PE
zRjboQB4}fjVbY3bAt#rOPA*E}{qGKMWVp(ayC@_7<-U)+yOYe-MLq_nY(1L&Iy-CD
zA-8H{=M6eu6&cGN`g|lUStrab5^=C`;hfQQ=gyJ0W#6`K+qP}@?PzOjottk=GA{jn
zb~ftuw*8fR|MvE_NAumieB(^*#`-k2v%9u0%Kl;cxI#ZO&2O5|>B~9k{#%`DUx}}e
z+q*tEKK@{}wJl>hhl8S;iom0^n$Rn2C4OY6ax773Rnk<@jMy0BBC(MrL6*^O<98Kv
zE!p2&vJ#kF`mJ0S1Wmr2P_sHv#MWhfL}RAV4f%!+MvmT;b<-9mTD@;?xN^eHQ`fti
zg+EUq>dw8e-nETf*R8^)X?Hq$9PxiR`I`Df0rQ6|R63d^^n2(1D?P;G#JZdP4p)@2
z2j{y>QzzYNow9kK!&K(?Tel~z<UL_JX_q0Bpn$SqlB1h}n*bYkz`F?z8n4SW80Lwa
z+Zf+E^3(a*jOv7MNB4wF&RLh1ce~`KZ`!fYLpRS9eb(7}(rap9=p3!5V&-#h&zzP#
zyV&Gtuz^P8s^W0ACs`3w+jyL&h9nC-)>^d2Q&MF!%PKQP4@D&p>$54wn>VkqW7A9F
zId0V7Frh8QY-i^5z(029+`b>PxMoo8%W%MB_V4%4{%?~%6@T-~`uDfiuiv-f;lH)N
z>yK~wuD?RO@TKw6-&?E<Ul>&Umrp$J_IT=CtCCIY?%7nVzP|bDoWH*|h&HpDJk(nP
zT6?7O>RO<;a7a+g&(F`#2YT5|$yn_*RV!p=*2ia&D~|@IZVo+n?p$bSNlA#r<^v2)
z3PR!mIvW{!JcPNGITjo^*~6`wptF!gqq%X0TB<OcfZc9}2#50tA>0c%ws^5-SiW1@
z%CW^Hcuu6u2TsAsE8Cx*R7_aG!lQC_0)vPyqf}S$^oT7|Q=C7stiF_-y_q3LAw*g^
zV}Zy5v9kg%ZwG{O)!ml*I>9JOuWs?IzrxRFp8PU<+tsL`bCT+pEm!Z%-DA#dxp>da
zHa^C~9kW7PdoH+Yh6XAoXXj4I?abH^v3i>AW|2?<LyZk?flC(aS_|@he)=<w{n+7;
znteYm-P(TZdyLk${~Ip7d%d_g>iL{I=6v6FzOw16I{0UKhs}!z+;v_~&70nc#w_qk
z(cn_inBt{y|NedBGcOB{c0YPiVr{1xV`KEpC(7!yN73S?vdboAs!ZYFaC8z9_?aXx
z^5fzRi#P12rj*TFa&vR>S5uQ0(ihIZc>Vlt`j^d{H?J?=yuXKsx8Ks_{Z+3(4mTqe
zl@<<#3EyV0u}Jtx=&&Rz@dPYY)>dR{6#8?_|CzCZ+%m@HoJo5#HW@T4H7Y8dXjbyj
zNoW?*VV&U8bU@;eju^wnlpu`_K{|6@FEB`tb*|giP~=h8rQ@v-u|UPyBXp6fS9gk1
z5u?0IW@esZ24jk?mFUunrnY_i_U*lTwI$}c<)lfB`Nu!cSMg&C*L9VjBB?RoBJxDm
zvaVaYt2P}AEy?tj2+?%SIOZx6>cZ0Urje&Zwr1Xns-1>6W}By7Jlwi&L-gG*yTWYw
z+B2Uo&CC8_{Vw`M#rX!EpLhI=Z}V?l|A>#3f0H_E=hvgZmS6cBBDZ|KZ{70EZ(hFq
zVQPEV)=q5w{w0P-vJzXj=iON}XXmz8w|D0xX#Nb~(pa=`s?c4Di>r?HO0CTlQhsfy
z=C@sc`sriGj@~$Snos7rl1T;Q;U#OTE}9=?zR=41FEPR4gUZbfuh(69{i&aMqm_D{
zogz<)MU-@lG4BVCKQ)`=FE~tICTS(en#|_HCLo%bkh*~TB~#l3ZjqC=RtFsvSsWz}
z2JCAWYzcCRoDg+UL}I}~j{_T+rii)Bo#d(V<>c@G$EI^y3w}Jl_<Pw@weO~$tSW7b
zZNl~QMYVr@;X0jmf^nKzgo7dz*Kf_h7@_-X_ue(Gc7L0-ILhzC>@Bg!x64k5sJ(eE
ze%Go=yY#|jqrEC0hGqrlYAy77@^9JVg2qkvDp&4aeJ0r~EJ%G)&{DPto(QHx8un{c
zHV95sW1Y;x;>_cG;<F;ZzVIFg`-ExFS2S;HSgF$Gu|eTLC&z*l%?d0%+-B0X3r^>L
z`?)l3dc_=to3&d~T{XNetQFe3;_$h(pPOc%`|@Ss_CM3FGupc5y3`tOzu)lg0+%aS
z&MoO9y4h_fv>cT)9F}huEa95+_3VtCskZ{X#rEyh+!ya3z_n_v*wzGFsjn}+s$60}
zG@s&BVBVe_x1s97G>xTOWH0W$%Xe$#Z~sq1DZjoj_pJYX>s9hM^&2lITh4#<dVBiH
z*S9#bbywWG`}z7w8-5>^`>Xy1H}d-^t4!XocvlkFjqi_)7!_DIOk=9Os3NdMnuqf(
zQ`rag@3jp&v%e{oR8QeL+0iX1*risR#hSaJZCmuY;@5n|9V?2IjZE7^cK^D)=gpoe
zQ`fOAdC9|S$t81SQ;?Hk%73nfjmnCDHc8K9n!9&Lb;c$a=Sv5sCbKwsG_i;>CLK|6
zX;jiUv1mmTi-L=ygn)zXvxULcvGNh-j@hpq61gKi95`4wdUfx6hcLCx-F$$*VC&cE
zXPwU`ew(Q>d6^Rzqk{etd;7QLUq2gP=Q(pY-mw1kbF1ws24~)zrR52KKELw*ok<lx
z?G6<$ViE96U?@NLK|p)mCEm?152^$?*ye6g=~{g2YNWunHCwbrleVzk4vpBDvcan-
z#l<G&kQ{4+z$PyF9l4&4UXI)qD{{3Pl(RIBHKuNjy<@Ig^<qJ#lG+c2l~-R~y_#LS
z(D{1RL(L`w!?3Muy}BZfB#XLUJF{45RmZWgt6AdP-dvipXicQT)KyD$PAyUJF#M~2
zmH*_cuUT7P)j3~z6>gKv@0OeRv1qHgRh{66w#k`u&i>1udCys0{lca{^`W*Y-{fb$
zWZ%Kwqp-$-U4fB->EgwU;?7QXe#gR&#l^+G*|YUV+S!g^X}^5w^wc)l-M1@_J-&PV
z9`D^71|{1ya*y6V*SpQ9ccpdE(wq%GC#M!qNSI}1vQpb`ruyvdv$xmYc~@>-5fR5@
z_e|bgK;(1F)3uBXl?2u0S6vDUTELJzsVMv31Fub)76$X@s!R}=EqATSgG)U+c=P2?
zHD;_lch1T(T+H5ZkymJ{P^)UMm*K|Ft`pCaXHI*1V&?Sa%0{YYr-ClJM)gihp1O9M
zS?`oLXPRXhHg8rH6*J^$5prJAknv*6#p_&Yhf|r~c|0;Wa^z;^`@{ox;^Lyd>}I|t
zv~G&|zhvKUo3hn4EqUv-4=#UpI6~pihk7@;`OE$N-`@CIcYpi!-_@^Ht?DzHa=PT?
zB>$TqOk8DnZ+NQAwenQ&wDtV_B+!xhyxqj-N4rHHFMgLcuV%CFojLi8%Vzt{Tz185
z<+54vfjl2MoS0Y`S2QFsv~V;stUbQmt21%EMvv0x7c!G?rL19{x`uTLW08ZEWg81y
zf#+j}J%0{GByQNl)LJxAgM~wX-jXbdLmHYJf4#{G3n}`vH|*!$;3soyCKen`=Fd~z
zb~`ZaaHOHMN2|+vmMv4mI9e;4j=bCSx9!ulZ{O?Q?7XwicjeZ{g*Fa(ujfu_Z&hEs
z=hwdV%P;->^V)xE<EI0)H)_xDb}n6V@VHFjah^)$f^6My&eCQ%xB4rRO_HTI`x-VU
z7_waUZQo<2tXI<9W*L5NN#XH;gKn){?12Y89<RJ?@vtXZuHlDC-%Ez-1jZ!_0uB9(
zUnp~#1%*m)S^o9^%5PhEo0{A_I$Gk4SGX(^i;i&r5ENphu(BY+a;nzEgzO0k8rJvi
z@3d+vsQAuVvq@;0V!{MfFVCCiT3L(Pw7z9-)N%9pw&5bfl_?J<MWk4qG)Yl&U{UgD
z^ANbf%&^LB(XP6ASuWFFv^zF3Ic0HN-x?&AVtt=qGl5UB$<vxsC!P6Elk=bW+1nOQ
zidt5YEE?c!yk(wGgie!@yfc@;6844BX}4QvhjhDYCg?tywp)QuCp-6a>Z*I!uH5Y1
z9J$A#wrlFPJvXnoxhxd;A@FDO?%m$%b$^$+$UR9v-RRmfe>LY$qu#arpBb$b0_;ti
zt_7Yk^;T^<A;xy$_O&f_r>?9Bw0q=Wy?Bcji`eFGE8RJw`_xTaYNcP?Jood`*K0|K
z(t>V9i8G(vaoAQ#=B!gW-)6oRzFiLw+(=!x=WuXXfi92Jn_Dtz8xqt-ZJwlQsl0gA
z)bfZc!&pez*!cFm2Lj&|N?h#jZ*w&hR!!qbXiacQIHcwl{5qmGe@TF}a)jBws@vCB
z1xasWdnaMK?I6?9q>i^IrYO9-A-Z(7X{d!rZ<1-K?}7Q2Ti1U3WnFN0_M`)5Lep+B
z8ESe21TuO@XeK(gL~PBQHo@sc=q|B4>l6=4Xk3gnQT#bgKIKcL(DgGvCmSsER9U3+
z!HPriahGgD$L7_qO8r*HycY6Q<zi(MnN+1L+2*p?Cx)AgO{-&v5@)ECl3`NU5tgLF
z39p?$o%}6+uI}i|bMt*~mAqDb<}3N~hqJwT?(E4g)N_yAt5>Yu|1)3qKl_ymS=sxJ
z4FRHzNvupvmBNQ=Ud|OPZ(w?QX3k8>$jHb@8*#mu4|lW0va7eo>@3>6+)i{_#Ygp(
z$`2bRS$~+|d~l`Ia+Z>mtP^K1t!g>*=GG09lUftMoIi8sjNMZ2z20Yiy<gwqy=k?X
z(Q$QM2yf0z)|xBE+7W9`#)T*`=5{emo7~~~#o!9_5>+wloV`IlDYH&4yvp~S?^m0%
zSnU~)xm_XdyLr?mwVv52Ctda8Le(@238Pag#~eD0SRQ(sa`>o7HvTM9{^cl;+gD`O
z!fUoGV_Ra>h4oXM7P&0D;B-Z4n+sRdhRzi(EfXXTIm#7_G<di$Y<0TUyP9dU+!F7#
z3zG{Lzdd?fsX<#TO@g~eyEANZw+Zhe7Vlr|f|^g2-f(<ObIHr+`PNp>Y&&z_g7eHA
zvS0oFCCD>;KQi&4z%9jPlDur5=GqqRZ|++w{#_Qi_d)&2o%8$ccYOKxNw(CoUHIr6
z`>?y)-(UI}S-toEh46$;b-Q%<CdpQ=H@<e}*FM)*1=`z#7PC$1KQ1)wpxj=!%Wr<H
zQ1v*(I3Z6nBYZ(rgw~WE5$Qs&uYzme$vwF#k}O}PB%;E}Foh*IY~JtBRo!t4hnO6@
zSQ>2W?z3s<tT0@x(=wMWD1v1ri_Y#%{3p&Y^g0zi)yQzGqs9$}{$*Z9JFbN-oqB0?
z$d4;4a+uym1iuKncw2Odpv2^_nh#$y2z+AmbiQ^>Oe~~hLgNGW_52rKO-Ykqc;{oy
z$|=sbH7jH`xA3O?4X*HQh@CJ^RsEV!!lAv=W|l62l3%%b_%Gg*I$UWk?zG^-3`L)w
zf3FxT)eRgE_?(C+FL|TbW1#tUBS&K~N2;^I;f-xl25qwv<nI~pZ^%@V%XCk#&)65-
z&8_y*L$1^Mn&7%T^D?(BeDb^r7X!WS_{J`uw5#*h+p4=KZ)+bA{(UfI^UWzfI=Rvj
zx>FX0@i>V(H8M`}&`dk+EqZIs1-|<aH{b9&y6#ZzRr!*-XZAT?9zH)`RQl}pG1;=k
zE8hK@zxm4jh_7)O#eYB8pAo)MTk`YpjS|}jUpH^d_;D!BpeXu#cbKzXZvCb8)%X7F
z{`bHB{`)`j%(mBN`9{4kTP>tkw|Q^Q+q359r{6ty{@R_(-4D7KOP#4b^tJeD&gp3-
zC#!|kvV5QYG^+f&@9z2h?HAsb?T*#nz3=MN=h<8L_#3gbwS>EgFW_0@p5h_KBABvj
zrqq;E$FIK1`N_~EC9lJ!!Yi;zP-0`p#1@OQCvQne=1ksokSl6y(y<WcpN%Wo{^;Bk
z3ixL#Xl-!wU(S?(_Tp8+TgyJAoX%Qb#_A|$Jmug)7d56zIlIjNs+`Z9irCm4on(77
zdOj;1ENfMXXl7|>*ys`dg(a9xP*L!7vfw0t_YnVg(>Kmy?C=Wsq!A?1B=qujmcZwQ
z`q_#pEE6Yj7$r0HG;^FgVcW+fIscmB{`yDOC8dFHb&h?J=lfo%VR&`QsZb5Y6N-@w
z@6Sm)RwOi)&87X7QO>bYM-R=^HET4!S(a~ouesCmYP*Z)1)0CTHea_jdvUa#S-WM+
zI}L}QX2IE;?0CO+%~~~eO~`593^NHA-3*iMu@iD0ZQH-M&g`V@vMUo*LOeN^Fj|{2
zbTWu=6rTv3RN%C`e$Hpxj{@Av3ep}c6*hF4oeh|%bgAN|qNJlkdq;7CV1KiMPD4oF
z#ao?0^AeOZxMMejr)Y>CGrQ0tDQ4uvqO6;wRjQ;qkzv~+-Mj~v?Benbw<NImOWrWx
zWWMPk)i7Ie;pUZTH`g^quMVxcmvZ;o?4s8F$Gnts&aQmGbaTx@v#Ey!J49bx2r%>3
zlAW<tW2@3H`+0MFUnqRJwm@05SCzrrqa^apLy^Nwh8zaHiYYp68!Vsvn19~x1p77#
zr=U|CS~)l$9&Bll<7w3G=njfD&SG%dA@<32(W2SX?{_U=uU!76)|&GR`vaa|Ki0U4
zm6gg&>(c2FtxBDtb!(exmlBi6tuQTD)-W!PMVvy;+cRoU{Mr1mmG8c+9Sd(kP{q>E
z2mCLywB^OcIiE=~Xn)GkP{L_0F`@TpAICQfHSU!y{0gdq`;=H!8a<T+8Q!{DaV0-~
zytuE;;ZtSO!5jC}&M}-ZIBa&-w2rxNHDiTAMY^-ikJa<#>{V(t1)l^>RyPV+ymW3@
z<}Kmp{Cy2CE-uP!ofaFr-G7N9^ZD7!lqF97;w_bVX1dTz<ovQ)&-gb%$2=ceABnGB
zpDilI-#KOCOpOYGDJ?cijBQi?TrlvMG=u#pQ+;I553c~1CfC5E`O8!T1(G<FRvbw<
z!Q^$g&Z6Xy#Ke|=J0|;oy71Um<)_Md&ELwB7Oi3mzm+xyt0j3VoM?ViD=;yng;PXB
zLv(_iWXcM*OLv)GDRo9^<f`RrFmCU1NeEJ!@?!p$Lyi$TfdL*9J!H>au>Ja>`ay_v
z*^amOEtWD(30Ua6RbZ8BNQToC2CtTeKh6Px{*#|<kUq$C?Si~bF4GRtw~VsK9yC-w
z(CubC-f>`}C&z&vMn~b}KPRZZO;L3|kmcE|$l{R1awKDtiHbJUA(g2TXC?}nELw1k
zTWMO7E~l%4YGBlqTN-UYvsoK%F1zcUbuu{8+tq1@a->)QzwqgEejnd|-Ffoin)jO%
zSHD!`W?ikA@?g=16Q||xdzih5HeAwoAjR9kTiZ;bb*Yf&g~{scl@;FzOxUp7iMJ)U
zW403avk9?B9_hu-J)ycYp@X-<>5>rNnLCSSFY9}p*LS0Q_A0TEHOl!3y&bzhb~#<|
zvbW4m&WX^~J)-NocE^&fxo(#)SRZ9<s%D9D3cJm5hGFwP_O}OS%$ZRfeq-K=10oFz
zHcYB)3t>s%k=Q8YIzcdkiAQN(MyF!f#XDN_YD761CW>S@E@TLrDwY(^bkWuSc3czR
zw7$N%L7SHdUgmT%xV>eeV3=}~v8Ybc+SL&^j;)O@mf2yhyJgXeum@ibt=r(*qomNJ
zwP;b(M1eP#JipY+_ZWX{{{3>(#g!MfpE>i+?%sy<D9N3ccGKh!Ka%~s_|?AB+aJ^q
zmnOXMo+k7A(2+R~%zEcj<F41{J^HaGW@q1RrJUnknvS#0a%O6rbrs3EX>xhN%T|%V
zZ)qHoT`yOy0o^Vdk?gkR((Bthm!^l!wJN=~Vr^(~=(%&jwysw5`j&e%sR^w0<@k7c
znn?D-2rd^T>s2CMSv(siZ;$m6np9w~Rn~Y=@z-40`ugg*ENs%>0_?9hIk@~?Ara|u
znD>=m$NG|ZMd7vl^82P#a)d7nZO;nqG~N)hOj01iQ1;Ejl+P31CN6xj{l&JlWojaI
zQkjh^OFn$q)XX`-nQ>BOr>DX5*5wm>Y~ANg7O?m7@F<wTH~EZuR@@VJ^{t_qk{r{7
zUagjP-pgj7{A!Wiu`}DaZ}INkvsL&-mvu`=zR=c{`<9mN4xL=fJB9ZV>)N%`Z}R^8
z{q5FU{;gZ&+k}`*e5c)XQ+tvl?RCO;hO$FY`&{8?>s9wJx-`L-#pXcPZ5OQqRR)WR
zWrrdJ#T-5h#jiP$u2DN#@ml#gon^_bEFDoi8uCRaoeyZ8RZ5)3)X`Q|{&vqk>%Go9
z43;j=5&a*cQRW)+yyw?G+2_;kj^CE>6g$u1@S&?&iSwK7Y32i;f)8CwpTFI&>gn>G
z_BFqb`5P;~c2xK;@<;8@?4&CpkAC^CU!=r3QDD(kp_3vkcfReJ-#%%F^2^d+d!tHk
zy?Na#evVtHC22xSw;+qvQod_Wxtl-y=d`a%F|`+{Q}X`(@J;af&xbRYuaexqsn*g`
zcT-(r^VbjV)zwMg*!P^{s`G!gG(7*}$9+q)r`d^lo=<+r@mbxl^IJ^$t8c&mJ+Hre
z<9&JcZrQv2{f`@U4|fXBHF$gN_1-&o*G~U!U{+R^c({G@=Blr+FXelkdvas8dD53h
z>66#H$4#wDK6B>D?wHj}g%>Ysz4|jh>)y3%*RQW%zdr5M-`U$(m>3isG#{{(9dUPR
z6u)9~DkQd5%8+S-qoQA_R=mBd_93x#=ZbeQw){ALh>_!r{i)+C{Eqy+dGhZTYo?X!
zAAfrlU{xD8KWx>>D+x}ePJ5*$iFz_IzI6+hyY@8G+vL>K<JyJ22FrvE!a94mUAmg2
zb<1@{RM!R<C+*ZV0%=V_l9E~*L$@pp&|0+0P@=2ro9<NB^U2@TFV27eY^uqYZS&7f
zj*m%;yRQASTw+Ir>;HGR#V^-a+<l||;g>A8zGLrs>nkb>9RF8vHaBzqn11r;Lzg#`
zE^>?I2<)_#bSr&&>cLE-be5BIgtPKXw@-_FbV+X2&rtdObuAq1vK)^W9&XEd%+r@r
z@zrNis<N85!O1q?*$WpW&R_4maN)wz`SZ)?R?e-STRV5|o-dp#6BJcn^#5a02(kN`
zGGT-I`>R~rRJ{~*6la8SPE@E5J#omDWd>)Ej;KPz5%rxp0S7XUXhg7f=?Jn2X)-#U
zFzFN(Wb9g@tRKv*%sErlW#fcjXD*b^ZVriLS*q2YBiPZ<ta3t;>!gwHYNKf`tfH>W
zLT!p1N3?o3GIjU}v9FL+V>HW?XfHb9If?PkWyb2#%-q%{k%l4<kGhhELjrrH=Fj=M
zw~ITLdG49*TYR}5goMW2)QrB&f9=>A?T(4|UJhrBZ){aMcV$Y1f#8%g6IM<QYGg1g
zyIV1vqji4jmhSf^S}T~gF0J}Hwdnq{X^od|UU2H(q3CqeyNWxaCu#SOX_r=6x>qJo
zHn}2vNo|wQA}xi0blb43=t=jsGM7aMY8yE{TYERHDdk92q0=p1rIwoA^UQv6yBaq5
z>Lu)T^S@a0KZPlxHdV;mEXbA3b)~;~ZQw4AJ&(&+dsj$T^1Gazy6?)eWlJ4r$J9hl
zJ|1%T`iV;6PfMf|#2VJPGE8x7aERn&(~wr$J8QL}^V{1ESNDtB*V|nY+Q2a_MDXLM
zv#xFmj)KbS*Bd-j(k@N7snMBr+vQd+$Ay3uPFMXsr2C)P9sjxc#fqCjcW&pYgxs0!
zRJlT|@9eKDW`F%Hx0H9Zol@Mb)AFLza6*@VZcCS8gw}=-No7t!0Y{I8LOd+2Nep!b
zAsq#woQz^TBH0Q7+9J0ble#9?2(Af@SdttZ7`upTwzP=g&97B^qt1uEX74p{Xa8ww
zesfNoi)o<y;X^Z&&DO5Y&Sgo~cA6HY+O=t#R&#=rCX23<W;WZF;I+&NRkNAH-~D^{
z-)-Y!{tn|Z*_LzXiaxzHzt8pf(fbXTV|0GWz5c|<e{9})rq6qBO`pJA;KJc}^Jf3%
zJ8!@1XWuz~{P^*S%4$xhb+)vvjaqYi&*66QjJx+LKE2hRF28?g5!Wq)g*sdQyttaY
zR>|q+&8ca@vo@_dKi{5fN{FXQ;8$CD58DL_9E$v9({H9G9sRZRUel}krJU-9Z%!4?
zcQAhY`{y^S3i-q<?X&D$r@xhxvC-Uobe_I&FvFX}bALX6DZ5`HmD#fST8!bzrEg2$
zR-IFRUe);5?$4VK`&N8yecD#UaQX1zC9hLu*cY#CH~uT{?>?dIdEKFAKcP1ZE#{tE
zKI`1>;v?J2@4LLXJIj2{UcWE=UnhLHyz$_Uxr@&}`LuY#y6@VhtsngL&C>Peyz)=@
z@OE0Mjht`$21mt&x#yE_ob2ea>Y3g?omY_c^M=I%FN}-){d@M#o4lgK_{RI4R$elt
zG90fo!nWMuHRX8i{O~}g(3ul=yF}L{s8p`)7QEII__Fyb-`#h`yGl203Oi!umu|dV
z)34X5<L2JYR_4O{t1XwOAK{NuIhn%6BzEm-2J7ZDFNU`d(>RM<a*p1L){2f6K61kO
zqQPFalNBX?mlE07ncwrT(C@S}Qewz){cFLIJUw^kjY-n{996DgI;uIm8m@UOsWb)(
zZu58F>Q``Pj!>)*!#TkTs}4LkA!Ee&NLH@I)HO-v>)FQ7!W?y<=PvoWe(|}3U!IuU
z*kUwQJ|bnF;~kgJ4z>ktqBiv>nPiz9y5u6<s!SRySYEfx6PSF=^KxodcXB$@5^n+L
zY=M?|gWD@3T(@rx5?lN?Kh0<E$*o`CZ@II&=Sbs|LX{2?3%^fJ{yt}%UnKlWosh(E
zP|V?hBlE;5idI(loY>{xY*6(7S)$3RP;};(TSUU4j)`iM1kbnpsawB#h1?}(CH6TB
z#2miex0%JLsl%(oCm|ze>hIacs4Cga&~Rijo0pc8kdsn$s2AJii}#rRW^d)r=E&Rj
zlsT=(oJDqai0^vcTWca1!#*9{ye5!~b?R-`n~^$ZHa|`rl0Krg?$)&41sA%y^7q#5
zpL{Zg<>i;JHvbsHU$#yNt=+Kn7)R?g)i7h0P&eaEQXCsY+}TW3!&EmN>X_ElwdU!G
z#6Tg#_Kd^Ex`6@{8q-9AoCI1p*3F-wS!l0jS$|*Xw!YrJ9k2U#Jj!m*s4qLY{qV9K
z3GZGV*thiiw%_x<-rxSBK<}^p+TCT&MM@0ne)q*4)ZZjIpE$9%Jw85uf92xSn`S>d
zJ>}ui)7?6oPABeU<B^`Q&}pY{MCoJKP>+vWgeNL3b=myd*FEYkmxxB<e5)c)l}Rx(
zJC}#8jaoU2L2tplx_9|Y&adC2Fn<oy$`J1lY?sbDKAYAhwsukJy_wf<Pv3VXBvIwU
zi_J4wo*mw5An<qo8B6Phrk1zTg_E+?OMbV_WAl4?%iHNeV)GU5w(s}fd6wO~5x3>|
zdcy<v%62?UuQpy;e8@O`Rzt-m4QCbS$$Q%_&GN1`T{}DN&Z<+<b-NWe9o@W%TXlCP
z>ohaZi1kqs+k6^rWo><)d|u{jUBTqXvVg<0vQb3LV^KG&(utn-MbWAclAm(Oo?!p_
zL_q!Jr4v5P$JM^tR9(_`e3B;XxOGxaM&;q{ZZqemJ^q_hx#vkl^u5xDH3$5Sem*dK
zKC8@od!}D$-!*3b#LH`ssBiRsAU5}WVzgEA>qPa=xjZc0hmJmDz0G{isMLIC`-1AN
z4zDMx&gge(yz{1iPVM7sGuvBIRE2fhdj#IND4wwut!d|uoSgD{@82Bti*uJKRLH(E
z+3-%qK!x8X#g^ejK|@oI=z~cO3P)5tCKakSWFA(uc@m`gjH6SaCoo|HLz8Pym_R3|
z=p?2iDIyBTgml=HdM2uL3MeTVxXiOlJ{N!B;iTs?4jW$P?$5vR#;H#D$dd=2&m4B1
zWcmDN{>jJon|?%98s~5))UBER)#bUS&O;tO$=CS{FHh{!y#KR3cCY!CH7_ix6c699
z(tOVUnT55ncII0q*SWPZCpr{ocrkW-VpF)3sPew<%%PjQ>R0AnELm`|gCqCMls%v4
zO=16*`OHyh!xzP`ljIJ`9o1l3>9kfcf@{g#;&n3S=BC$M+SuGuHqP7l`ap_pflX&*
zO7gdkBf@tsegEKo_|VU<Yu@;8ex!A4t98m97H+W?ZRw`_=M*=~?`+6dSm~0<BJIe#
zh{eTW#Trf5hS?@wCNB3J8MZm;DsM8IusAevO+%DIg!#^Be{Y*0l_t*Ih3{Bw=5OEB
zU@WC^vcZ*Wl1da$(}_(<5tC+4I2p+_$)I5aL$b2bQN;*Xg-I*ctZei;aY9R>n=xtL
z326lnRnI^{71u~s?FG9_x4YJy{k|=ILRGenLWLWP$AnJNxu;|l8J3+d<DI+1+jo)m
z@fMA1vbUx^U~M{@m4CwHigOBwoAV?|b@f~Gl4q+g`puvg**Hz*xJLEN-9Klr^GjM+
zRLf7VQE8v?X@f?8zVi0qQuESjJ9HaEx0y=^d2%X5YAZObiPox4m@wfa^C_NL^IRAG
z3(Qz%bHa&z^2wbmS*~2Tc<~a~FVE<dDU$@XT9ll4y0khDrUXh%>|*zsba)BhO`ZS3
zPvkGIpC50<o~pio%jpMdVZoEz8?(fJ9p1D5etERq_A7G7%O5j*-nq|U%d^Qd`64;x
z;(Max40dGLtgmkmPo8|^)3<87|LLbLF57P$wmotCvPZnv^@{ZpDn331pS8r@_V)J1
zb;tXqZByTs7rcIb_Tp~4nt$Ks?VR%U>(;lCM)w-CGuAF$x@6tFd9!9^zRLQ#x5(rG
zi^35`h9m(c27y2)Wu*g2n*<w9xC%y0)XbQ4a6@NigvTaVHVu^tOExtabgkg#a(=bY
ziLo(9;fR!ol1avD7ljrMX6I$Kzb{|(a*+5vfkTjK=X37Jm(y0~Tv=5*?O-C)*^tnz
z+Wa{x3Tu*Mv}yy*I$s~YEU;Z$fT``otqtoxI(7<YYbz(#sTnDFsLCgBX>5^tlKZyB
zLAQPOWyR@U`QrChMc$jO_FbLL(Za?5sa(vQ_i9GU3J%}+6K<S#6v{U*zPlh$kzr%Z
zcR{BuL1(2{g&5Xfi;_1@5is+ZD!t=+a!`IofQWBz7teh2i=Y0=#}(BIT{xH7q<&FT
zC-ItR(;6oMm1HGR*Gy>xVQ;U#vfQd>abBC`+=v5vcJC^6y!C43yE$*qZaiIC_kY80
z8Gi4p-u})LkFR{>Sa+<B=eJVqG;7agvnp<!RA`dPF_>(Uyd~>m^X|203z8>26pPH?
z?p>?Ks$qV{_)xCzpPM3?SFX%7O5B`wcJ}IdDL<c1Ka^_pnQLXhqRlQ03TqU^C7Jjm
z7~N7@8@)~j7^`ej7V>QI@Mc=Y7_eZfPGFNkV24Qilik}j>?`NrNvQi0{`vf?0M$cR
zCPq}R)UoKeDZ$?KV{eMxCkb}0y_@PYPUN>LNeQ;3CFX82?CH`lROXO$+P3YC*raVi
zX%Q+;SE6sksw}iQD162<fyE}}bGbEpcH?q|dPU~F`)4XBENF;iIw%+8Hc`_{$ty}@
z+P0%>_!D#)I>OC-baF2U&rAr;P+iLs6{Zy8xcOGQzy$u|KmN-<{wpWMe*D#sPgVUP
z9@Ez9#;-SuoM<R;YFVgDnwZwqu1oh!D$}|a9belT!MV=hN8Rz=?eE=t+WT);AFSQ{
z;!Dn+Cs&0FK3DF!zDnj?QoiM!vlCzMDewQRfA6}*<~Ai3{u2t;4;VYwDVZK|V6b+T
zYcJ#pK7ahU#k=J5u`G)6Oj7TZclNz__`}cg^TVEkg!ODZ5^9O<d?LycEAk54JdLx>
zQVy5+-nnwyYOhs}T<x=E%a;EA`)?WBw5h)qFVH+N#nxg;M_}h(1t$%g0Cz`|tW)#(
zXElUx32|BL;$M>dhGAmVy+()d+alA<roLp{!WFi1LC0(XmLL1{+rO)?ur-|HBe7tk
z!B2I)qdld<?o6H6$}Y{)TCB8mT0)AP<B#PB>t0nHJ-o2F*@y4)Drx&qNA_Q34=Hso
zZCoY(Yhi8ez52Iu=NDS*r6jrvn4CKC_IK?jmLCnm&(Hm0;+?4V!RzVj$E(i9*_i9a
z9KN^f)8p%_o$K3JI0QrwGkO>#9PqRVSeVextD>~D)nwwNeko5?bFM1Ri;I^onU(QX
zzYuh?#jdJ2UmG4CUR_-s8yg#2xn1JIoUV<hR(!GLTdS$PNUBSCp*P1gWtFxs0*^9g
z%gvhq;@xhQ$t?#b9Bxn8&@SPTxGkYBkwww(iJQ?Rp+l^5m)w%=Xq&`;#x<x>WPzJe
zLqrnO|A;2dzc%}#mN37VuYYp#llF5p8+Z2Z6nL@g&CdBQ+tp%Ybyh4`a>unIy3920
z+SydGejaz8lLpR$K~vopJW|gJ(>l0bQ-UGmi2Ww}3&qCTmX`XjrrNKncZ^-9{!IM~
z{|o8rKm0Gm-hYw(d-DAK&(e&iKJ{-s`Q+F897`L!-}iIRZ&Z9C;K=aC#ad9}`ugO1
zd#k=a^I0Fa=j!U{TW`NU+__u%;DhVc>wbRQx4-U<hCo=$rqfz;>PnudtC~DjnG_r}
zXUbnK5e-$hE)B(*Gb*1yn?I-W7rX4zzXo5e+TZWhJaCL5Bc_xoXd>6YLI<u;q3)<D
z8#dqhyYArOHwI6)F<yTrpFKmyO@8{9%5$GoVjD_kC~R#I6^k*LzpU}Ok<&RYJ7<<J
zk++KH6(?Tjvwt(?g5P3J)n>NxT&)Fx4TqY7*w?klPCC_Am3~AuYx<)t0WAy~bEA_Q
zL>(P%_QmT@+rYT=velZb$p(KIe;X`+!)#!&B5Q4QdCY4*=@_%L!pM>x-CJ&5F1o$5
zY1@;Vd*)1A_vr1;&Y25$q%33ONthb5+G$F3?ow^H8_PnyTElp}jznD44b04pxE&fA
zHBD7e^0!3NqN8Gtp)MXLFA2Z=Wmo@qC;$A#PcJMxZ~U$3wOW{S|3hc-ul3=zzbxt;
zKe5U2&#d2|`km?1r}vL`|9rQwA-`+>Hb#?1mW_fsE}o~3oZ3Bo%C>l8ojqG_I;S<e
zNFC29E_-`U;@FeqvbQ(B&zJf0c;|wH3u=s~+vx;9cIyrKmub3H^_9=eqv^A%I20Ox
z&bg<38Z?J`A@KDEC*~LH4W9&P3W=_E5v*UXc9Qw)^VgB@{Y$>Leox&i`R?e_7nwOP
z7R)%ZYSpD<s|*&pwy}x?O>s#&njTP{>5>x0c%$>5#@{cd_Lu!`>zA_Z^R8X}!+dvM
z>Cp}E(&Z|gOP!ln$d)ax{WP!RPgV2t>b=u0-}%t#ciMJFrpu-Fm9H+?N4#|R_G+7A
zvhbIKy@k{S+Z!*BKbq(9%D+whseD&W@JasumCi3aY6^cmnq~0G_0pwfQ|u~U`0_7n
zpOkUrLgSPENmhZMT>h;nRCsy7<l%DW{mem96YMK5Hx>#9nF|L8wLbAHx0+RG@*wlL
zAK#qIjimxF&h->OtKs?a@56=Wh1@qXoSno!?4KoYp#8l;@<J1Z#77sNH!W}zu-6fh
zv}ZUfzcphi(@icRMQ?5WXEP;|Vh$!ws-2>x6zH}wWXrU?r=5R&793#spZuxLN0Gm+
zrhWAz?Xpxm%Pq@S)ofNhwaIwxq1$=Usna5wG;esjxXcX9J*+Fy$Kz%cvn`;9<LFv}
zH3xhPnlcPcB%~f17i+Fl-OJM<CwS=4$Ayn&bocR>?$7(aXYacUA3F*xYV73azn*9E
zTfyPY&jmkv3l!RJG_mz0CnswynrW$(syS_v&pegMA%U)-xpk|O<|X_UA3o2Q%3LOq
z7|5wIDN|U@XQoj~@MJZgxtUg=ZEv@3?b?#@>eZ!XiVX7<3`3lcKmNBiH;=a^#+heQ
zuVGBkHqDNZplMt&x~Jr;64K6f>zm!)IdkUBtwm}sH$HbyJ$+o}ta+Nr*2qm4c{^_U
zX>Sem5p^?soLwxgIYsDBkCskiZno~KuuhjrCss|cdc9<x!L%9T*SA&*eLpdK_r7UK
z*?cdr@h*Gx@ad=5wrxdw6!*PZn&<QQmx!FqU!nRI3jxh5{_~&ytP6kov#S5>XS4e%
zCR;M3jNR=1$k*BZE0u4OpAvt{p2fG9%U;Ap{p6=$Cwm>{r)@QM8GkOE>H2fwM&FU*
zWqG**XRhB^{ibpDp~Z*hPw{zCs5n<nr+D$P=A$z&8ZVNmnBNog;-A4&*BdiGoNueK
zoGs#Rr&c0YbFTC0%!{AgllW?wUWo@5D_&~%lGXUKE2JuTsa1f{yB9gmMrpo1sa$#r
zYMd<lZ^@@5su@J+2zDHeIB380lw^*^9EmWU|J50b4g}19V?X)B&BeFMmMVI5r8XEE
zy{-DK$N&EG+ibUjy}8@2XKzkS;QHr(`_5fQnWDfgKKx5M#g&g--_NmxotKZD^VnUt
zNm(7YKQl!XUTSGPurN1mc9F7T=I4t#)220b@38s1?%J)B-)>l&u5elvbhc`1ou;?Q
z{gC;tVcT|p3cuOZ{OqSj=jG(ra;vY}n(etRv1LttTAje=HIcW)4Yt0?WmlaUuCBFm
z(#2Qn5@t?nEm$PMF=bUx+SSlQ7d9zg*vQjyEKwrJ$di-9Q^fz_s|$alO)i!g-+8q8
z*Us5mI~cbeQuFg~;mOHKN)}KP@c7a@XE*aQ=XVoUr)<qQweIxiT-*2g@ALDW?zQgi
z?d@46S`>Qfy=OGH^>jV0Wuh^w)})3U^|xc;;4FNxY*yjo%A2Q8yzn@@QhR#&_1(`(
zH|@Ifug=Xpe0k~8OP6-<zI4CKraWYU*8Xcsi!v+i)jaLv@7lfJ6>gk<e@6Ae>*}$x
zvQp`Lbnf<@%DH>mscUn%n$Pv9)3=RoPdu&^B69I|?#6A$wniq2_$JNH-R|{hSHz*t
ztfz6I%Op}nf>ZlMHl5yfJL^W)!?v_b+pgs9jyaxjX~NP_)4(YU0@pIG7n`Ws*v7}&
z{p)&J<#lnfPft%zKgFmQK3ApFnyFVvVG^TjJBP!Mm<9azf6rzTSYTiC#lO<QaBpp+
zt?0!PxjhD(PHVP3-H{jQmwS7w&}sfPHyWASnEn_nbZX_gV!iFr(e5nMZXp$qiAq~D
z!?mx^Hp{u1c6Pnho_|>)8u7c1t}1zXX=&Wvss#!H4Xo?$U4Qp4^hjY$u#AyRx$MTc
zRY8*`O_|AC^(j2HY)A0iow={QBX>J5H#Sbbz3o|ALUFmV-_n`qVnWU=b6?T->hi3z
z=(M!dGv|uumYyqJm!G}u+iBU|GtU&<>l0tObLO-(8LQ5fhdzX!*3#g+v1(Ud%!_Te
z#l2IWNhC<DIn;UJYF^TU(?K1_Ryr9d@OmaI?Q?TA{OUC0+7*KW0|Wg%isvft&-h%U
zDOu)_v~07MaW%`khMKw;%)8iL+PKI(a#Kmam;C+DxhwM4MRHGk&z?Q9LH};rVz*wA
ztXY{mKx^Q|&rhuD;o$zD$jN-(CRyURM%t_$&!uwDwUwn_%l9^4diUI(J=gc1+jnmN
zxpM~(9&!l(JB=lwVb%x#79rR9EI-l@+HrpHZ_DF3IAv$=PRldf&Z#r*F56?4d0ju|
z&RM?Y=Aw&=RiDkiTJ*K=RP3gglf}9Dramp%p%<NbM)%Ou6D!+9j##V{Ny^c(nc~Gg
zZEDz}J4QPsB3q}NnYuMJvNPLDGib@mV}U%=jQV-noSYn5y38U2`<bR+zqK&(hF`z)
z++&YzLT-Juh%esR@b$*twv!wTk|tlPo8E91{n}?(o1;1{(Cmi#&O7QN^Nc(bmT{*=
zyYb|tGcMa@u=RA!&(BxiM%`=96XOv$siwQT!a`+ApsSC{%**LfThq?YHq#0}eth5Y
zJpmzCKUMrGoim}qN6X@!b>9<{g+d`-SEn0U&WbKAi(Yn$bKC9QZHrP)dK_IL#T~Dr
zG4+X(hnW<cdZygv&dzkc6RA%osDElWd?>oH@x1Cs_6C6p24(HjJ#TbXyRN+4^<IGQ
z15>}n<eGgSn3A&8IyqNyRBEKkPAiL6b2l=1+`g;y^If^>!{?+QKD6Szu=D!EPJz#h
zU9C(?zx#h!c-ByIjY5-%`HMxp#-EM(&sZ(ITDjr5OyMo{On!fVe|~;<XC4v$|2I~x
zpM3JkXKlU+N8`+e3$CR#E|B7#s5VVcY-ZO?UBAT8i#;2|RFXA46;%{&uA2I+{ZU(s
zRLAq@sZZ*3+r@>2xw)SS@X9>B7ok3l(P-B`0au+5olhB_oeyZ*JblT^2dZ2moQ?r6
zCWWlXPTn}NSY5TX=xS(K*Q%SLF%|`j`sP}CUg9VZbp51%e}58_kIEzsg_h!TlM0VZ
zrCe4ol*-&PLspsR`HIx29mU1ICTISr&0KnP`3gmb)<*&@YECTn>c7L}ei+rn^VPJS
zpYoME>srmGGnpBuBDA<vbsRPeZdabxS=1CXrB$nIV`oBHVyNrfH5(jNR8%GjNG*2c
zP~^FDWSN?poNUtPgz8tZU$#{<7oV%VbMN)t^6k=$h4-A#JkHy3M>p};@4QU~wRX4H
zu&@78x!~dY&YXV9oOz|wLS3b{25wHd>}xXj^B&jNSK^hrzX&wdXt*Di@tE*%w!n=&
zF^`{19M_9n=4)_RL(;96>1E&|Uul~!Z*Fc5J3Cv{!GTGj>3|D^efN+3-9PrO6Zro3
zefzz$^Lc)#A7T?J$Xe*K;pj=HHgD@FSC<1q9h;plO?`9Gz{O3q>(nZ(MTS96d!h}s
zB~JGK6_<T)w#`t+oooGPfg8O0?dN^FSA4cQ(aiV7_SUy9oiEunSS$*T^2x|O*y&(;
zL}8asbW@pn-jNyVFPBVKi_F>i$ICR(ReF1%yO!po(9qD(2+hwcW_1;~_pRC#I{Dm|
zlXHYa!>U$Y^qZ=(GU)m9prxR#Fn()Y1kT$0`TY6Zg$oxzJ=43+Aq-9e3>zLEUUs~P
z@l(xw_mi=cpU-#M{IX3%P_jwu!L8=0xjIWFt!}SaD-pW%qEWzslij&)E{<BtLZLz7
zTNka`=w>MD8W`)<DcH11p;M)0lBdcfPYzGjt}SVYGIQpv&S@7G=I7@<{IN|=OepU|
zLB8dA%cpyfUwHEIkK?1d^TE4*G&g_Qw>|d*`-T55Kj(LEeD_3m%BCB6F*~lN&Q6R>
z*9$Fey}do}!;g7)YJNSwe!kzP=E6d$96!fCo8;sfiIbL!g`7;8<n!{*<)3SBFV5@T
z9XWI6w6#%N=dMx^;NHa~$iyJfGS6fC2A-Ihn8==#NfDV|mO2xHP9~jobJ2}#UA*|o
zguAjbe?AxeaZ&il*tVIOv3{C^-o)0Xusf=Z51E?Vg*|n8l#d)at}MMNOs+FkxVyW%
z%g@YQ(o$0F?d{Xs^X~4>ogKM3^=u!v-rbv<+s@85P>9@|cEnmtX=9XarjW$7sjH4{
zJvnu?-`eP?UMZ`R<sD9p91e$h3_eTNE7Z-{5n-{zWX+=;EfXg+3Yn<1NO(p~Y3Z5J
z5WYUCvo$(Kjo)+EuF|7<^LDlEI6RlV>~-6Y?Dn3T#Dbr~)vKfy{o+@ic;fR<{tkZg
zBQo&<WeOE8ast8^llGM98@<@0pm3zX$<pJCV9H4mzw9G7H{9MXam=tMf4<%Oxjc{i
zWIQHHBz#lxaPs2twu{+(ex^~<pW=WymCr9ODpk&u%G|0Yy|{XDb@js9h4l*;E^J-c
zz`=N%O^I>+(;L_Ch08KD+Qs)j{&CO$PN&7ip9!s;u2oyYgxR{Ibvw=-P;=rmXbm%9
z-psAcpy_kAU|Di{MAU{4A2!U<P1w8ln#i7AFMi6Em>%CX`~L5R3=-l;@^gFFD@?0V
zsAO5sGWGt_C0;IiQ~rIBePF(Q@$vp=@ALEX%gU$bEpF$NvG6syyJYh9l`8+cm^__Y
zCV47La$H<IU*d9#RAN#0VS{75^X;ldcuj7V^qTBBb1L<L(b}NX+V&2hU=L_<XmDZ>
zH~P^%G41~TkH3GO{CK`cXAwtFddO5ktEEk{vxU4qopt4K*l<e4n^#JpQF!0u<nri!
z{cg{m#(PZv{#PmfzQOK8-Z3mZZ-P~g{BrI^&#b+`GEsS+$H$!pWe;=T9_f5m**$NU
zjK|_5lbicWvbU$%R0R5asmu`IY-X7eknnDO)|H5D;=M`7O=p{>6&r0cs@ZNS^Z)$(
z^T$4YG5lpKQ&Tt1_MYvHMQv6S0zh5s=_g~q|NUQm_j~P&665=fEU#uxGg%aLWbWi`
zhb2@yrX}zlSaa&&f&Kg6Z!=?0Wcs}0+r8!uvfE3=^w)DS%->)4d0E3I*Dn41+pkO7
z5>uzeXY9LKd+uSxjwO@)?A#p9-rh>hoV!{q=jMiIhnMMuSA5#uSbppL*98uZE7^mC
z{r&w}|9VQznl)?I<dQ8LGB3_iVB~HyWR=(3fBo?q@d)!9;$K%k*rCZ~(30q>A|@Mi
z&~@p?oWz_^uTa+rWvyFVa#A-stzDmyI8#$3FeHg%)oq={g3BXk<~mI~bVGJt>ZJ~4
zrrRP?Tc%Cpw4CA+A;25(dzZPLz%85q4D#7$KCm0-7M_0k>1WM8|4)1RmFktRiz~1z
za9A?12nvgE$Rr3<9Eja#z)*2_A5U`OS(D4Y2FKbY_uXe>*W*24o@P_=;rM#Fgo>}P
zFWpb53*it@5b{y-?C^UTcuLFDRn(X>d&XWV?Mq8N(q_*{oa$q6nCId};|LZeMh3lo
zat(}OTw2$Alo~i*2+X{wS|QRLcXr<Gcw@uUtKa9ZpY%Jm;>%X&^i?+R1#IFOSR(kU
zd|4#xu7!DQDl)P-ASxu=x9Z>X9s5&R1ROqJaAXm86^Z&DZI;uVe0#gi=R0!@*L{6)
z*0gprPp6QgtG>*n+}klTx4pSJZ|~}`mG7g^Y!7;G^y#y4w7I#lF*`S->$Oj>U;orz
zmKK+nJabmMkY_+Bw?Ov6Yz{Lm{-dXhigJw%H-7y1@lM|79lNSmmb|NHpZ<OK{=H9}
zn)H`3aW{z=mL{&zJ6^%UE4l2LexXgA$mJ6!e0%0hHQKak)0D(XD;6D2l?XHwOniJ~
z-lERe3raNHoh_2PMbc*F%rsJQ?@MfxNNk-JZ#c#4+_`hRcb`3b_Uzv4*FU|!X7q5n
z&z!O~%^E$E9C}uHnM`TUFjy33)X`&h#g#RDb<%{sSS}|w53x+2cD8GO_wRlloy~4O
z^Td-fIr}>w%WL%4zy8Niw_;c90|BdohX;18xw$Dt<;ZEB%f8194jXx$JgIhfTi9Fu
zyt~`Q=2+Y+PTO1FZ#jL_zf&K=ThnAKDzsg@#dKnQJ#H0#`SRrC<fTiO>c#b(oNSVI
z|I+-{r_86<uV3F{5fRLmAmJw}%)4SSha0QVlu47O?o5*?TR6c$Amq(OF|HFSDLPwM
z^csB7e|>M;a_3{>=4QX+R^2I$TYaX|^VE}{!pr`>``-DxL2vPPr_!tM8X6BSdd$c<
zEm(Hb-$OT!>C0?7t*^1EEA4F44a42VJU0`P6Cd2$D}D9Y9tnel89SG~cawPHA@Ss)
zRphc6DK9T~X;p@=m=-NLbC#5pl=#h?HzaRfe^wy%>({Se-;6gGXPq&gIVXL~LZ?-s
zUQVl)a*2wX26<eN`lj<NH6`>RPmYYp!JE+~($k*Ln_qp?FC;%mV8;=G`0gI&9_Ahg
zfrZ&!60<oMtm~X@Xyj}1sc`avWnMExQ)^RanS@5aTJ`kw>{qK+ty;C5hgCt7)m8C`
zugT+Ob7p^<xUco+=0LZapP!$fpKn+Bd`{-FS$UbM&!0~_I4jC6^ZwIss~a+J9NCz3
zf$KsC19t+8P^g1j@03LcUA5NcdMT}u)?IXB70+8Qt&Pb?78!_n9CFa{64SYP=JctP
z+4t|?`(jpZ^LM-U)FY1y=84~V_iz3ko9oZ2B3SrXc&CVmu5wB{uM=_Tt@Q1uavC3d
zO$&RH<r&^@%PBm^X4sP-C-vpBugSf{2b}^B`x!5#^jEx`%s0#M+(Da7FT|Z%KfMf6
zo4Tvy<*lvd)2C0aUA=ns`p0K{_UzfS>&h9UClibwES<sSZI<nIF3dGI*I(Q1!}^BS
zfXJEiTIB+`?ACILsC!LT3EaCPoR!64g`?0T4yWe33~#^fULbpvWk=(V*$XF6sf&v|
z(`UKh*p<uCvohE9lqJ27o0a$CfrGq>hV~&V$%}l4EIxjJ!jmIwyzk(d!V3#ZMPCLj
z_L{itS<nOhhOMur1V-Nc{`2Oo2idP*zkZ~bmzS59J$u%cRVExRTuV6w8*X$&Y=}I4
zG$$!+|NDC-xoh`y-iy<({B2>qUShVtguL|a+JDz~Upjg0ZG{8#AqLqad;aax*wCf%
zUGUwbC6hd!ES}QLGr@7;ahc4uXA_boju{?kWH{8$v!iudWa_k-MN1#*>^wT%nODlf
zOWR=1>9SY1w`bei+uK|GnLqhc`X0{r)%U*rzHIxW`lDoXa#U2<wDz=~s3j^b+yXjB
zoE(f2Dp{iK9N(!qtiC2t8M@VV`O;0ghM6yZ-Msba+?mCW1tO=Ib}3BN{=&?_!{GDA
zbC(8Fl}<hb2Wx@9>wSjW`wTpLwtTvjGTAKC<XOh7+{wO&@?v%t&C0ucKk!xc`}-@;
zEShlHPbXrdK=;0$76GT2ok0_x&1e!1ut=6w+_<mu=kkpQ4n;+2yfj?CeAy&Vl`b<2
z7X?s3#3JBo;xdEXp}~WLuOTT(z)iK=GdW^|$FYus4ee>3$I4P{`nnm|XFi_fvG(>B
zU0=h>U6DVxYMh^`WWIh2BbUHJ?%S;l{=3f4vk@@Oy1HuanUKrJJ$0r8o;u~&<})#^
z@Qhq_n?UJ3*IOF4j_fZKS=kg1o19oC5z}<~!Su5&m(z>+XZ1;ite8`3yjioSRI9M(
zFwcbx&n{ln*&t9{T%5)I@JH?UhwX=S#JgI1e>7d0u~zcTDvrb98QX48jPVI@jWAxb
zF<dYwCxtc8h1JC+gJM1%d}=}-CJEB;q~`gFa$``p5(a=#1axjrqIbKdn^w(-~R
z+5ej#e6CKqA=Fj6>u=qz%mq#C^D3Y6R7x^S?sT30C?&D2Vz$KnBab^zPy8^aoWJuI
zha#5{heGA^<1(Fcn(8uzhb4{~_RTfyH%MM(%wuraCQet_!ST`m>?fe>n&!*Aw+(vv
zs(Wc5uXgU%kmU3%$*#5Bktw3tQPZY|b>3XC#EE5sGLvDKneqPP|C*=&Wj%SaC}!T{
zsZ)DzU;VOWOG$}-hvwPIm-?k77COD!Tcz&wc58%h%gxnpGhWD5{)~)_K2f~r;$pYi
zcmDi&C%s};-rNs+o>$GvpH%l(gR^*9mta6o6Nf^RFrQU9=jEU0A1$f;*?wh3;c?mX
z^XC=bQp<Sp;>GkOar4d1RxV(C#yxxX-RIAr+dSbBVzS{EzmzB9VzN;<>@-g}+oLCm
zDLz@#Hgjj^YFwHa%D3z7?HB(Pe*W*U?_e-Kq3`byb3pRnw0r}$gcTRQs{7m(*?TY1
zxlQ-p9tqoYJ4H5+oei>_d*3~?eSPD{_f1B6Pwt-l2+Fs2cbBGfRmfTex_8<5n%v*-
zsrRYi;gU&#K|Pa#R8%GfUGh|!v|v)mN|i~WHlbQ-voaPRJ-m3~qG?^8s;gH`@>G4+
zJ%9fEXFvbt_tiYVXjkd{VrF;4x~pfTm(A2nHH?`NqId0fsE_BZ#>>o`IluVNfAsq6
z{)6_hJfG4lQzT#8Z1B01I5Ca&c1nES*;!vRCvN2uv3@WmWA7}_mYD_<%584Uc)4s=
zWQJm^i2q|pcD<N~?x4=nxw7JeJ$IBPKJ;l$^g0yN)HPLQ(bCguiN?9N=jYwsWs>nB
zXsLI@GVT?=QCoAPwq_qa!oXv=iZSczyw=r9wHNJ#rF(TY9b2Upn&`6Nh?|q)#<Z=I
zrs-@e{CMwoGCTV*wbR?)ysO-E?eT%nmecv(_kG;4-b;o(>6PWT4;}ns2EE1`Z|L9A
zSk$Gx6GWca-Y`e{Pg{Yh!C~nK2FLw2NM&E|m$SWh&*bjKK8Z(sI?le&`V!kDbq=dZ
zL_M7r8Il>bd1l(#+4k&@|J3(?{Ab_(@Q3aDmLE#WADKV;%uTt>_A+bZv2B-ivkSJR
zY&n>-wRh3hEej(oR&`5@xdtR!OtA92sq&*v@<W}$pQNWrj~+dIm}mc~v9azS@BN;d
z#M06W)rq%j&eg3b-s<%D!qdVZckHfSj1gYlRDO@4?CtG6+j6F)ouBEpHYCv1;9=zB
z<J~$x(_iasVO_9ho#?b)$&A-JTUG{5ovzPyIW(+j^VJ1wMKx7Dbu^cUbIm-ZzF=17
z@_n<a|JL-MKKbNL8VuC#U-Tv?r?hH)@bVS1>8rlp@>JOnG*!fKrqqL0fz5rZR?QMS
zk~=NVWP+qZh^LC8#I%H^a~m$7KT}kbl#={7XJe_!7S^za=8Ln9uRLYsuyhgl>7Z~>
z_@QmtdzFYsuVdeAEKRe{THUp3)f25{!JI}ZCnx#LNSovvl+&fL=%q-YtN05=h9By4
z3XjW5JnoyUQq-;&+2(85`E0gZ=dxLu%VxJEOZ@1&YVqoL;-Nq_MyG8`M>0yR7%#B1
zKjSf0o~$}uFln8Lcel$%mQyPd10zKoyH*%+d0squdyDbz^Ob97&z?DP({<^$%e|L;
z*%kFbmalmCGpj@OclEgYji;V^TKoFs?*qzR=2<2Yi#Tc)2rNvo5kB5jp?2rA&i2y7
z-P0|9-m@+*eR_L)-htg^pf<|J$E#MYIyF@*DD%vWSv9W$?uABqz1P_qHP!N6@v~>o
zo~@m;>6Vm?%&#abE3KEAnVCxcR-W0}+1W}>VwOLD{``5-=8_;wi^C)qhbBWXuwtE7
zcjQvG$WmT|W!lXKG1FE(xEML@R-nP@XqU7p0fB<7o{DT|=lc3<WL*E$k+#z=u*k^1
zU0GgO@W}rCdv|}%iz_$W@lbsGA@xI5H#cj~dtSV0eWu^7+AY-`i~8)WI8vT`PAx80
zJ8L%Ql*ZzAnR6kYx3|nw_{t*S#Q6M>o#aKnF2iFchq-k&oqq3c)Yv-xP-0u!vuWXc
zjj7Ln3Vf=xnHd?GR+eov<C5gDVwbflk}9h_E>7u6TG|m4DKS%!W7Sd5(9j9oMxAM@
z5>s8dIa$v&Po4bl#5ALrnwoHNe-3UAhK%&IcLq$)1p?;3IX7L!QO$j;kcy9jz;A(8
z7RMIm?dvp{YJSJ~v#!%&U~FKJG!8$o>Fn*#+d{+nM2nuf$}an8zH-@LlLyN=O%^Hj
z+Y~?Xm?)DxDW7F$(W&Wrk;a^}d~R;eIs5$i{QNnyLONAuP1O?ew7f5W{O|hyAN#xS
zecva)*G}^Nk;U)Ntn5y;<LQ&$cISAW*r~RcTQ{UmyBg{;chZRobNeoyTsv{XW7Xw_
zM?bcg9^AL*_Lp^e(eouL4lmxb<0I?AM||#`kE`sX&+S%BefqRpTYI)h#g099j;9o-
z6e~`teAAy+c-GJ2IlIh@$8&pXub-c9SM#Ih@H2@Ye%&JPB%aSHy!7<+GmDqUWeU&a
z%g?ZCmN;W@nCFtQzrW8}|J|Sd?4SIp&i~{`_LHynt)Eu2srcD^i_C?sJGNcd3o+J6
z;L}QtIq9rDHBBp+=U~WzmTs?3#Sldyjq6>JpV;~*<QCp3=exbMwCJ-vdyMJf7`>q9
zbKX@-9b6-$erkR?d*dpr1@hq+U)IXIU;MmF^xiYuJE<G5MW#+YecIPf<<_$KMjZ+-
z+0%>VD&OC?ulvWs!7h82=kswfeZh%mo|(_6IXoxl=d%EvQz98tLakzKEN8ZbtcXm#
zdAOZ#-MX~}8(0LtFW)0M<!Hj9je?y5%=as=PIzm@wAExw!0F@8(udev7~5OC%T9=V
zn6%b)%gRcJfajBHKKXrWSnw`V+4*R~xredaxb-FW6x}d5^y!q!v^l)GTet7KGbv{0
z*0jB8@^;s1*X~`tn>$Kw+ZK;mw=Qm7e7s-IB;$A0>Q}E`ow*7Ei&t-7J#%$n>6g~U
z%U90Rb>KC)BFN&@IklSS>M6^T-R^0Rm(5D?xY#mjm6wOh#$Xq3LumtThtg)oWh-J!
z40k`6`@yBK`axoXK>2~YO7}j9tu|hMmubQxFBabF1>PzvU4)E|PM@YTr_X`uv~Eo?
z&q;&BJTX^~-Q4v1+0LTf%iVSv9B;GL*?#_1L!hfj=%ed8JB_?GR<=wwEn0qFt!-yW
zm&P)#RbiQtH;<-2fBt;(!hF{@i5o{hHB{Qn^H)nqJ9C$FNyaRlCCajDtCqC%sw@=~
z*Us6dF=f@!o(*m)Jx)og)n-y`eSKYN&!SVKU$uMxQ?*Gx|NIC0%)-J4d~yH#K33%Y
zf0}<ed-`1lHm4Mo-?D3L=bg%2e#c+b<l|(2uFPfgL`4E!4R!{)ic~G?-XG{=5!Tt7
zaq4)O)25Wwnvq$n*0VfKUvX4qlB*ocs)e(fB=2x6>YFy5O-?>|<&{^ZRtZwKZgDPM
zxNzYz$!D_!6FEYiT0C?sTB{{@if*zO7d?9PDo;*spt<<>bFV%x`Qh;5i}1%vclrL~
z;=v`XECC626>5j}s3isZ<&=hoD~5)KhUH8NT=Z&N&dtqfr&c|TGd0lo=)Owhr~6A)
zHXe~g1?$H>cFBDQ^<p+ZooU>abD8Ja!Xs90Ga^f~o<D#7`~fHopTExFT)TGW+?n$i
z8s9n4yx`E3g<e6eipmnRCah9Y78Gz^=DeY7!-2#OhZZ@59<8gZ!?MG~wzah{3^tm+
z+C3vA<4sr1+s2ZoE$j1i4^}WJ*j2o}J)PIOBRN^F@~5Sx`ct{eH|GysT?M*O(YmaO
zqgbhjC1D{K*Y67p<@Q9ZaO;(>WD%IG^0?39<?3^XZNBl_7q$eRb(Q&2Vs0*U`Ld^O
zY)GK1w5;Fq;OOY+XdhFxWy{0;<_5gl=&7PIEkh(_lBdG?yB)VKUCg!Gmbx~?MKERV
zh4-0}%&Qq*Wjs%M@T1{Y{JmVO*yoFWIK1C;p5e3ggP54gJN*nY4NM)sH+-JJ@^Z(l
z&8eGjAA47BX7J`sPRg5{oP_kvn>TN)`aHL`yh1K!$JbX~c7YZG!mbh-N1XnoBp>f<
z>EuvoSs8oz{aO3HHlNSiygA%=IBezf`SX`AZ(hE<efjby$F3Z|a^)XyZ7naauB@!q
zFUDn4UA)br&60i1o$W2c{;V!@jR}4In%m$1(nYb_xByNe+luL|d=FSO0v0;Fwk~^l
zd18fw!xO)lEytae75>bb)7Y<(oN{u)u?zRu&dx}l@!Z1S_L&HOUFCDB#KQ8$Zd<c#
zX9}r=gBC2D7E_DZn9SQHq7ksrX{(e;cG%jetx*Ttd2?-bZDX?ziKeMcQVBZ3v?pk9
zS43x;<`bFXZ4y^E%#lslm%qNAJ;c=M^FIdj@~A~6mVf@;F8;oE{_X>Fj&GRT-g)R@
z?HRuN3e4MiZ>ybGS69``*%Fx2Z|pMZnXkEC%<{8mK0fM?`)ucFa=G!$tb&Kj=9^?(
zTwGkZX3Ns0r>E;hX`FSgWj=2;^XANES+B)%8OP^ER;j)eUlvyZ&c0=7b$xxcS)1ny
zd3eStgg0*Jii+H_rKYiahU3g<Oe|JQ)QqxUABtFiMZKXRQgxEfqEk6LmUQ1`QTu6r
z;+c&(-(*)g8;f9HlS#*2Mf`0itHoa4u*K$aUx8syvN`v|J~m+=W1fhOmCA01+uRM8
z%}6=SZB&?gL1$CfGrhG@TfNV&icFopATqVx+{|q0Y_qjdqU!>FEI%$ZWsTFSMXlVS
zUgmo?E9t2^ZkyK8(`o(f&AIRO57>{sufF$wx83)5ey_XUS?AqO$XmB=cfRL~FVgc5
z?)+@Qd$?dW<AFc21$ugVdNzrAG5Z2|W+qM06jJW8*y$i`6S})p*nj5j?Rl4dU**qN
z`)hL9*Dh74XOgDMl$AkCJv}vxg=S58d3mC%h(>@@zp0c)(9*Mhvn*#$`EX@<pp3?|
zXZOD9uirKO7iiQLglGI}d~)H!h4bfD+m=Ke_FU+xa7@HHDm7#ym)5EkS9P~t(fF*>
zb>N^ND{Ip#Cg;t3Cyjr2)N`=^=wR>QbMl>3*)Z!=$);<mPah|&-~GFx>3N0F*`&%R
zyy?EhT8>ssr^Es_6%|!yeHVSF*?nN0)v6WC!h$-NoK=~0xP8w<SCK%I`-wF_T}1*t
zGMBdVl>E5l%~dQkW0J?i`OaJ>6PI;qD748u{_sMfuJcxuyX?vpW$V_hn>TOX%qO2J
zjZC&|*;Qo~Yhsn7?W)!4<e6r|bXH`uw^g0v<(%8QDu2G85Oe0t@!TiQCnPE=s^>N3
z{hTf{ee#jtHviYhuy;4a3O82wx?J&%G)nF5?Y-k`larH|a>-j;TN8A}TA(94pNvI8
z#W7=FlY0`(62~l>yU!F}QYn^7*~~L5<Wjod+`Ox+L)HdA+dOmT%(SImGZ~+JN__b6
z@xw=qJ&!l64Rl=PG0BKW+CVby=&W+N$LTvBJ$i7@yu7@2$)4LKyL`{v&p-R@PK^Dj
zpL6zwYksM_^4IafzPAM(a@IxnMb`Axlo@{f_|`C6>+CdL!`!Xg-rgv9c+iH`>fLVL
zrqgvX9IVrJ3vzX(WE^u(vZ;6`ad3^dZs#(I#5O~pdt5WyW`|t;+-A%8Vv47|B*XXT
zhd<N^-~YkzNW(JP!8`fG`KFYEJV%$@Ib(R{jMuv}Gj}~KJdyQOIy$vmFKBTw&zZ?F
zD`!rebUW*yK+7bTo{ec|L&N__`?PQParAJ(92vf8w_eXZSL1l@BV%i8lVfAWsk5`?
z7}#z2ellzhTN_l~P<@Yi4~P2MS=I{*CYf_emp2$)@|?z(GHc1?k~u0*)jh2}RVKMH
z83@??$UYLM6QV3HXH$Isfg_8Wug%NL%l+=pGd;K1<8hzW+R)F%{x*NCm?ge^Ilgj5
z=9+B&=(eL}I$2pcIXP8-ZFOapeO{SJ3AAb|x&?V&=3DH#vNM&(a!2&(yk~iNv(0%n
zUAft&-MZD?C41|RP#<nhZJw}a5<6$M7Tu0IBI@cY>f&bTG&zYi!`rB*cZS-Ih%={8
z9V=Y>p}=gfdiEp!y?6HLp1=O(&wRDtPu}R(9$+*Uw%N09zXE$~=(W(&kieQ$!z(){
zO$!V1R5{6(y{L<QyTxN)lgqv)mlI}XZsawsI3_N$>B+Rn)U$Wyn(RF4uT}HAL?d!@
z(@}3xorq`g@BR1wy1%aW=l5>brnNdt)*DTGZen}k`_fsP*KN=}aKOjlvhSW_cbuhk
zPx=`43r$^?mbQA1dG51dxwN#*B2|gc&OAo9UAHXRleTs>*R59F<gAN}bVRzkIywx!
zRyn*mV%oGOHqdZ=sk6BFp6_#{|E}-wzR&S)cfR+wZ?ji}#zLRC+D~A<Gg~I`vFje4
zI0q5+@2!(hFjn|8-d<=~+_<@hOXKP3sWK~GUS29v)&2a$Vn&gQg{pN*nM-9gPR37|
zRg?-^w7=!@zBylB1o*WJX)Nk8&00LGudIkealx_9=PzHr<Vn6Nx9#KAa~n!$TQD$4
z=qx&LQA)>=VPEl6(Ynsbb7aD`(~=(COj#@uoxA4DwufzcUJ^T^jJFz>-7eP6irG1X
z<6uZ=lj>sW+-*lZRw=VOiY93-SjEw_%Fxxr!&8OZkZBLoy4RUgdmdRWY;S(Vo__Yp
zo<6tR{UVBQ?CaDQd^|m|Wbe8c%xX@l*G?Q2FYsKpYwnTjfjXs^JY}{7rkyQ`T5yx?
z`Mk>K^9qm4SXl5FDNan9xnIg~;^S$NsnhFnE<G2w*|g%K#-crvJExbImseI+mhY@A
z{}~}C&p-L-w3!w^8>c;eX4&d8$zh>eE1PJ}b}3^n{cwYvEg=hCLR%UIHnM1&oH@1r
zZgrn_*TV!g<|lraE6X2n?>(No=Vi&d$+K@(>(4x0Kk?}sb^Yl-Y|0PF<i{V7@qc1J
z@xiW_2fx?tU*Vt6x%AlH*VFYZc05h@n-<N+BlSV=U~`#>s^UsPkBx~({B%MdtIf(>
zuB+MM>y&eI`MJ3kpU*F<{BquIsq-t@kk@Yt7B5~@`73?nine9j)~#E4dR_X_4}aW`
zOH?h@@$sse9G!ba=8VCaZ2=A)W(uMa=gyotb>Y;BL)E@d@10(_^V#RhIq8WO`<*}k
zoX`J3?telg)3Feb(ty{c{mC1zTRgK?pI7;OeqS|*;o$>kS)>-+<l((x@pyXT<0Enl
z+Bu&)wOFLh%3L-hWwGjp#5RdkW*(!tRyk6cGnYOW&AZpTN9urBeEp*z{52tUN{3oQ
z?KGxIY-?)G3|OUVaOFg^k)cj9tBB~^lqFqaN)gxC|6Jrv<hmheu=VL~JA3XoAGx>u
zo<IBPPdoo!N7L*F4BsM+J}@M8b2LjIVvt?W!4&t%)9}Ij6^pu3G$vVkEb7ux0*!PC
zpRYVqCVbwG^Z9%sHNlVW5+6>QTuzfVxp{fHpI*(+ry`+JW^YqVHtm-_efrFv-Ea2n
z*&`#bK3QH~-dN$JlGlnUlhRCzrKUL<8f}Z&7Clq!vG=4ix&?dE3?o)4woSVwwpM4G
zYm<_ZD8pNU%}iS+38YA}1$I1ra`Whs+=313_4U5Zl{>ukLqm&5#>Mvy%e@7>C#%^M
zy7!7{EG(1X|7MPI>oKOI*@qaeE=ZX@C3D%VvyEF5a%b}-n_SO3oA{dDa4*9P9hoB*
z1rHC~d|oV5(#ex|Dk)y`bJ|9WpPz3msQatjXKifoxbN;+n?Ikw@EMk|uds~`e9co)
zFfHQL3kk;0=jZ%<8X7h&(rJ-skXDnUddjMft`$9RP9`)lxv+5^?N8Y#DABP>C)80w
zrSbPh#ifgVWo2dlPJi?~n=5VS&RvJ6_U}FNvAcc$o;|xi@9sCB_VMyhK8;0147&>!
zaz5TumNSzxX3f%~Pn$M8U9;JAT6*g`>9y~66g-|BA0Hq4Y~Q{;9I|h$rKJUx;<~Hj
z<KyGwLe^H+R1~~<cx&Ots&BGAre$Wv>z?1euygU^Vy|b@CY{z=Cw1!jyYHZhfk&q$
zHZFVUsnXJ?<iv7-L#Qbv&O-146SF`Q$CPDB4_Y&}UEbz1bH$WkLDy|l7I+G=OfpiL
z`q;4}a=ov(yWKhO?v95EZ;dNjc5uy?SP>v{kmVrD$tKRGM<VCu%((DJK{mX$EVAO!
zUeHbh@HLw2CzriF-7_sRb+*pkbgrGD;hxrIRs3-aiod^}SNZ%%r;yUeq@y-Jj?3(s
zdd@0c<m!2ghqFBnbE~rk6?>}PYu#OQNFZT--qyxQZY71V1`htN#05ReR!`<RboI`i
zGiOZB6y%29%9g%uaLp=OCp1~aCCNa{Q#I4c>8QY30i7evT6_QguickFZ|b%4yY76u
z-KoTBr2K&4>w3<wANXb;WKP@8a&y)qR+h_~r%s+YL8YZn3)E3kh<rZp>gwRzX3}NP
z@-6f$FE8rSe5tZejfF$(?alV2W87xRHz%f@S#4q9bHI0X*oupb-C5VJk<hxj?O0ja
zLs!v7UcaiUs<PK7-7Vr2T_~6ncJu<n5&oXdYGzxKwnc_UPB?i}tD{3_g35#m6ICX$
z&d{`emlkKoU7!7kulwSQw|^W~oR0|l8*$;`!I$;|_BHBfKK|V&!1thSf5Q8McPEeY
z7EibNyvfbXr)|gg@8>O^>&5JpxU2W(@w_>Ohnrc8ea#fx+<JC)pES7{wB6k(nR|}m
z>z6OfzkK=fe#_UdrlzK*hNY&ahFdn7e%`!U=CR8pPZM9JRY3tuedn8atrgx@Vx@ZV
zq^A#KyGz=NRhl<jLz9la3hHKcO;DP<Vv0mx&v~6^sjr)-PhMknuw&NY?)L46_uY+)
zz4N)|T)M%NQcDJfuXA_3XRt20yS#$I==TA}f{0g-S1g*Tm71zKZBfsaz>?d`9)kKS
zt+$uV@O<@B<*x?Ef?bB<aYllR6`D8{by);#KEDXC=i*Q}r15O=+`0SbO>6R0T{g)x
zTV+z1(59)YJ=Ph{Vrb}RuyJqL$k1fPBG{oFvHZcNvt650dj$<#j9J3D)OHG7=Dnl4
zEV@my)5XAfv*#ps5st~zuc|9|m5ck@1qwY$5f<j(!zRi0Q!GASaO$o{%QLvPYI&<{
zRMytkR_|bpyEJ#-t(QuDc1ADOP4bXxb`gk`Q<>!5&6Gdc_xG2}5B#JGm0Y}Zzg+g!
zGz#*ROrLM^@Z5Zx5*rggi-$dzi#PY0+!Zd_vZeR3$=%?({Mmo%AAhad7w}X4M9oL_
zbCQ>h#rn@ncJ@op?Bko2=U#SHx3g<js7sox8=KZ77KY6XI%<;_-zog<r`_=&d+C1f
zg@5{;xBO>5x8vT=6!Epz3u0dgFK~+0;9mP$Xlqo7hOY|W3nokDLl3%zx88d2VB*1p
z2lp2}uAct><L<qcuffB;Y;SLG&%b}YUp7-nG&C~n{%q6g=^DC@p_zJ{PKT`68oIf-
zI5g1Yi;0QQLw{eDt))SpS}&ElG;VLtXL#GOQGiLJP4WDqwAAgBjg5_sQcg||owDtA
zaoFkNw3wAXE=5<Cnk);~KGWT}d)qaUii8<f-BS<UxKzY5HDU$3w`sGpy|vZlW16#P
z_xE>qcXzh8{-4VnVNhOmu(6_i&$+S<MPHlZ7{uclxIz|sio|3XC~)Tl$nC9A>zm7Y
zVC&aA*{^q9FME6UdU>49?y|SHHzXWvW{n6~=;%MwqVq_{B#q;a%*S<#o_^>_o^Mz4
zS^fOHSf(zmnUOc|_tn0u{rdG=?Z3LZIx#V^Jn_l=*23|7o<4mlly0&t+TCz%XjRav
zpxYUjYlE#N<Cyajr#?8ibj89$mqm9@H`eo+(x<d#Q5aXZm(i&pkBJR@8y<xnDA13#
zd4Ir$@qygQ)6L&ES{lSnDRkbU?)!TI@2#gDM!T}QcQHkHbDCcE=1(=*oukyIr^S&x
z?=Xi#lZ^V-tjg!}b}#o(XnL{VmO~)nfX(BDb4r4F^c+<lOxIc2>U${A<>HI9)YR0H
zP3zu$+O%o?)~y>S%N{L~Nmo3c9X4^oOx~(Xp>du{PtWa&JtxCEH|gt@SBZUJ%z5;@
zm&vT^`*f$U=yX{w+wQK%%`**>;$#lbIeX^W+tQ~Q#naoaKJ?<3l1iMi$s#6mSxUiC
zpR~J1yYnq4o(T)KDEiE`-sI5D!_8B7cO}kV)E1q)wI@tRCOUWHq6n^)Nj=IO?n=$~
zigY&?M*H3Ftebwa>aAYod;6b!_JWV{E84kR5B_yf;O=nZQ1S@awVZRpCc|^$8Rzcj
z*PMIC6U=k$n89Hl<8o%h2iXnZp3kdeK3`FMeYs2-^LabP=kFg(Xp)go^O@e1az;4q
z>`bGaYy38k+l+nhUc7j5|HX?3FPuMo{>YIV7jAqmK5Uoz^yG|dS*tu%&S#4*X`B=I
zZ0eq~Gj|H?>L~8sb~{&RIS*^s44#M-t`)0UlGdMeI;rjxbj0Vg-SOOyt?SH;dmbkz
zzUuw^@7ezlee=26v**<tJ}Zj1KO!d}uxoFlZI$}o<*Pa5bkrmMBrv>u0h*~QQ#<sC
zf%W$Gz;e!=z9!QvK0ZF)xoq~A%f1$wM>=^e=47v0wJCJctjser=3GhlbL0{}Z{IYH
z`Mk~V7Yc24b#<kGK5yZ>xUX*B#9!(2moHydw(_`a%mHQ&bpbbygJOCOsWC<tPwSEy
zPfjijk(D^T+GN|6%X*TZcW=9Mn8)Jj_SRE(7cZT9>8@vyiOH@DUft_vB+dvqx=l3Z
z(AO;&J1%B!><j17II?Qd0hLLH3IZ)mEZ%Y|?@T}53A^{`@sYO{e}c+8ocI~UPG-OA
z`p01Xpk<z4BZt({Y@Ytu-n)12J<q&vKks{S;kRecb5$l-Bug%vk=SOMb>&=ASAq9D
zb{kfaPro1CpAy2fU~0>zM-O#2oz{Ew_|t`J*UrDc^W@HxCr_Sy*|_fGo>~8n>8Xht
zo(W^Kb{7i`);Ogy@6EZnp-Dy327!G{-A1=Flek?3LVX0oxXL_p7aehNi8vbO>e``j
zu<=-6XumR-q0WQyS=X=op51y}^T}Hq$CFP!#a{j@A9g5jpMLwp?=|NO?OQIDv`)~e
zsi`kGm^{f-#~?EG;VEOCJ==H;j_q;%D3QGAw9fXW)BgVSztPCVm^v-cRp_9O=cGU#
zLz79l=eIkx?k-z-Yk9VD(nr(c;@tG~^t6rnc6WD|*LSgh_){;wuZI1=e!-m8wt*}u
z-tJe80uCiD4-YelIHA!aVPxQOqU^^${Wj6mr;j5K9oYYojrH%}zt<a_OE2iv7rrsK
z`B(o=efF8k%<ICFU;cX_lv3_sy+Gpp4hJL2rIRO33-j|g=2&*vB_OMrV~1<I-&}*Y
zxA}!URVFpEwh8+w3kLU!YTPNjJo(s?ot}luO)XFQ&$ZNisWOQpEluH&#-e?-|K`n`
zw{6*e|K*ooe)&{YYd1rxYL>9)C6!6OMpn6+DyvxCB;uE>&1~~+TCwQJrnPUU9@NyB
zWEd&BIjVCdbKvE*v*)gP_UtvAe@~&@KQXaKpFX|n74)CIMQ_Q;q^YMEwKvMd334gy
z4qy?G5U~Hy(4f!x`k=M&cHLXX&-PsV`uh5Mi=t;=pXc4ZyW`uV^H&=dO_<Wo>v}cN
zRmJ1c<RepdW}I=#xni;0&#va@v-1XrEi^BiT+RwHS{uh~AiZ!w<}SCFFJI0`TiF(7
z;KXpSdF`i~|B|2lC)I3zcCz{NS)DVdbCS7*18(gOS*n#A;>y!xsnH|jbvhtOprx5X
z$6+(m8OCKBvYBR_@A0wG4V04n`O5g@{Mj>~Og1)N<v7J}q2bA%tIIj{?(X_B?e1lj
zpPN1RcuqR06=?ZWbJ?mc<Fr|wOB)Zjd@z__`Ko`u+FuI+`TbrjEC~mjx(ps3wr<kA
z*k@dEhsR4VW@q;4?RnSUu8z>Sz1i>Vtf^Wd>%;ct<rXmT`uYl)ntFhmC1JH6et$YK
zqv!WVodYhOlS&#Vb{KC8GH^`j772_rU9rOX>;8k<4`XeQhr|Y&y4?C6XWvmhyPu&#
zpzeF=ryBQzf7lPc&t4W5u!;4mN5tO1{tb2q!4twgY0O1Mz9yF=H)}Jd3VG^zmK<F^
zr)>K6yK#al-@7yx?EsB?oS7B%c}KWe=bVt2m-P)6t-89(FH_}t^!crs{dwDWzrWvL
zd-MJInvKaC-QGF=O({#Xbfdik=dQf0*s*BUDrG^I0Eda2MXD<!!wffWEUfnZ`?+?N
z?XsVH@7&uy`(3-$!yP-GRyd!3{PB<CLpk>~KOCP_{Rw&W!{I~!C;7}TwfqPFFn-!w
zp8V*?KE3VHYwqsm-0<@9lFNy%^84%l{{H;H-nK(SuEw~}%HX(?OLNMsoXx4HK@;Oo
zr|U(kOtvcBdNO8`=gOHg)6UM?R-AM5{0#;%?q3OFIk#U#ayvK(IO(Jx5L+=N{ix8Y
zgB^j^LCuVeEN&dD4oYwwk?`9c;e3DYv7>#{AN+lP??=PV`Y+9!c0MoO&+vhV_e>2#
z@$Q;??h`HkZmf=dz+T>Z^<Bg61^R`P)tnwP@GRX}7~*R($x~&bt60aP+l}IUGIG_c
z-}&3#-uyFePer?y{zW%_9hHghy%H}@OjP)M-Xi(wBb#69^L_q>Xf3?F_sZY(ziRE~
zPyY1sQBqRLuZkPDX3gr5^VQ{f^WE`sW~i%++H6&!<|N&Yz=kP}3$D(6%=72YwcGmY
zs>k2$m>vD?cj2X%B^ljGVLNs|;Gb`naee~Jd<BLCfz<471@Cgv$43u$7t71brhXEO
z*Vc@Ct)p{u>0M8?ljb=$9?!9mIA+-U;W)!(p0=oG5{Z+13{Fnw(b?4XaOuS}yRN@X
z`?9&XxHvz5{iaVFOuv5oxG5&5Sk04JbEd1R%ECBLw}562v1Ct0jvEWEZ;o84oz2=)
zF8*=rJKOE^ZG1R1a=KIEmK|8(C-QQ#nvcl2<qb?BD~(h<Cdz5-4oFB^q5j@Tc6Z5*
zZ&GsAVf^{WDxc3gd_}<FV{b=thmCT-oK3~Yvu2jdPEN|%TyS=_S<1!PRwXZw`Omle
z%f%L`v1sMg8DUqQUQG!rjJ#p|`P7B28yFNeriUE}7Z#lNXZb-p&Rx3=&PXrz4Y`@O
zOR{s)S<#}kclx4|6q=Hvf-X(*=+EJ9kMt36WAe80JHdYNgY)mn@6TVewwd|l{Q0x*
zm{Q74Fs<Rcq5AdJZ`lc6o4<S7Z{l=nKKbdXm-3I_a|;|(RdnKFQVfqzFzna4vuDq~
zu$bMu*H%_W&&vDoGmhu_k$tLvcL?Z4tUcYP`r9n|#>BR>UEg&4<|ftr{QMzuVcOYQ
zWyZz^H&@23t+^w4LnkAGkxlg8t<B$`@3nm~rS_*^@B=%^rF&)La<`q?xGU%6nVzGf
zOFM%!j%X+!)>_#;?UdJwB|>Kkb}x)R#(5_%=kvFI?cOav=TCl8J3sx>m-(J|D}Ua4
zI{omW#MZB#ucNnSZ_K*7+Cn$!=CRGaX~(!FO<!MMcWxWk*VSQbo3>`1&kn2mw{7dM
ztE<D-Ms^yNZruQyM#?hDc(68VTiOPO_^TJMFkBJq*wVU6iCIWj|K3b)<H9W+!CQ~|
zOpBP+eQeuxpIf(A?oQrxnMGp`kEXbEMs#*YPpCv!)(R!pq*Z>WrYWfQ2D~<(x=Y;L
z=HKi;jpuIuS3LXI?ct7>pLW`s-uqOkuU^2?vE;-Z27xl=BW)T633~*rO1ClbZeS=;
zcXIKW(~&H3HZSeuZkzRT63N&1i67lx_h-?9DIpqhOuzrW`%~EGUl({>_Wt|xPu^ej
znyO`zadA<vO6KAjU0;RVO=X@GfBt;Ik)KK9)uP_4tE-iIt6nTly7aOl*oVP;QTqdX
zo_e0YooQ0@Hya&#buTUUdO3Gmim&;Hr9GynE5rJp-pQQ3GrGC&bXu>;(nYVLZso1L
z%oDb1$}$O6J^gIg+rdfQG8x;pHA|^1Qk|mdW;A7~goDBwhK-Eb{LyPy$v$~i^2h1L
zcbT{QTjDvAeq1i>Y~Q|r|Ni_-fAw!a`CPr-=D+jgLmwE*tRHbCHT?a((D<9-j&I*S
z6%`egnVIQtzL9Y`u}*jG8tJ@m-^$9$$`W$kygA&?pPr_6CPq3Ml$Vs;j?BKkH2d}I
zv$HmxTD4qOhg)aWs^{rx&sNN<s*>Vd)F-!W*|vFh^XB=@lDc*4)-1oq%c&}tGelNQ
zF*$YAH|xk+hwH`q4|q%-<VEK`*)dyRW7}#Y%VLA4!mVP{y7(Sld}Z03>9S(AAWu<3
zD4T)nq-@2{Rm(rHKT`QoFR**jrAxuy)<!c{wrVNOR0(+a>%+GrzkV>Vhn}9UbW=EN
zQD#MCc;ULcN1rB7S5|s@X3f%&K+t|hk&?sP^RAXnJ~GMmmw|$V6X)~!0+0Lp3XjVa
zUQRn>kv7|@m6^w2SINt1fuUJf?`m6HEb7u&)WX31^1A4RO-j0*?FSO0v-*z|=N2WT
z_3nwT)5x4~FS=YeW2TWrc<#1qSIz`$tr0N{Q#!CwDQm^Xo@Dj3Fj*J1<muDzmI-f@
zw!7Wg+T7e&zU`0nzEc&4kIBsQTP3r6>dD8y<WJN-5O{dt?Oid42I~~&i?#yR?;5Oy
zr~PyDni=Vpx@_JrPrbXlQ=%6=WY-HQsd~QGy7BYl^ERI!&K9__Oe`|{)KsmI#i|d?
z8J_nfFS7l-+1Qvp=*RI{)<16^mTPNk*X#Z4d&T5#vM(d+w=f4Dg$p`v-2x67B^(NU
z?XrEVYPKaLe)bBD*;uryLu-+eBTJ+h$C0iT8@set^>j)YX^UJ75lre@$^7(>%a4Ed
zM>rTJmAj^GTd%=Zu(`Nc^M`{*$<a5bL}m)^3RuVaq49g<i+P>bYZ&j>u-?g1>y=`5
zH3|vzd(A6llCe5$)wbM5lOI>g-iideN(k^bvomohigKKAcjVzX?aF+9w#m!mGJA#3
zTg_ay&gRaYTU&jNt*tH2{5|xYN8e|!udn~^AOGbK$jhIs4Sv{u!eD9tDzo^2%#DJ=
zE@wJ>a%EOMcycIE&CPUcbl{?u*3l+{!U-1&MN>mIzMX6s+QTtr5~G5r>q3bZZ-J~+
zo6|loyzqh1YpP)2T<d*I?;DC3yq_@4)L&W=SQQ|@Px#$3Kc~bS3ZKuvnOAsRBAs2c
zEoDZ!X3ozjlZ?wX*Ut8uT-S@)`BdVp>Bsf=_tz!1Nu(G(5(zZyygMOeUZm@^u!lBF
zh4b#~rLLMeZOzTiY3H@R9!)#3x*<$Vp~OR>fk~(P{$=IWY?)VQPcJ+@BP=E7sF<4T
zw8dekHC;j-g_I>N6C${lco?d5%#^sh?Rf6S=H`b&zvq0OS9W{qsfy=!KFYjPQGdnY
zpuoKH2*bl2yVl<3ko+bf&cI;o5LxM-EL*RvkbE+}vRJ^e^kUv!*6T&xTLK)uUT9|H
zJT7A~k<H@RtenZIQhT<Zn|1K=^YbQ~UQCOuY`t>kjArd+p7ZC<-7$7J^kqx;Wu0x`
zPYFbCbKNMy;K<U^#KF`OqA|fuXLU~RYWFqMB!xU3CjXt#IDc-uzJ5rsva<57Y4;S4
z__(}l6Z_2kq5LjiyUp{@Hus+>e_|+obhqn3lZwN;A_m*)l1M(5hs)xk)=UkZZE`v5
z;v(PtWwT2zm(Be?{odZt&;)HpR|}CqSCODUokEkRYHDW8H+Z?(&$_7Q^9otFUctT}
z@2^-q?mK@z*&$it%a>z@eH-52m*^5?5y)Kj#h2Nki9;blvBcx!qli__Nvp+-@BEGp
zc(1z7OnH@Nd(-q=3wK=QK7BS|*WF9^M1JnIUFb9Yp?H+RVmFaQXO7x^KbXJ2|9ICb
zPwlOq&ZXbcy?HS^9-7Qq@=`owkE{5Jg$zH`588=nILkO5=`gNHo~f`?NJUhA%A%d#
zdL=I}cRm07*;udWmn>gf+_u!K*Zh3%-g9$v{kCqMnVE5KZ_iZ~iR29nyhV&fHuXkJ
zOWHmcl{Dx{TFkd1poiN;<D^O>qt4^U%dC>4v(^`GIdt%kjNEkR;)i_s=A~zziHq|$
z*f2hj-}!r<xpeQ#!u=EKCw#TAk(aQ~s9@rjJ?`Rgb+KdW)29ng>wJE{@NQ0Xiu8hq
z2aXx;k!L*au&?s@{KH~8!H<r9aO-Vua(HrbQex|wnMNs_(xPS?rFMONeSNLhbls|-
zQ(s?QKRe#qdTQ0IS?lj@zc;^jex2gG)2ib4wgkAd3U+Fzs<fm`5)#yQ%{fviEw(=M
z=+Prr`2{2<-h9(FRj_LE9|obLI&sGv7z<e<48jFUE(xDo#mCaXV4>jfaC)|4@p+p+
zJvlkaAJ3nNS6BIbPFnYKpT)x+rYD!}GP&$q^Yh~j?uNKGN*i2sW?WnwH8JZNXV8X(
z?Rj@z7r(u*(CNPJ>abNsw^t{2e}CN_9UWonnxM@sEWqlf_}<I!=+T{Sw+&@9lZ;mM
z=y(YkG>V-K%S=9|=iHmleQewHxpMu^iDotT(vAODu<JDaZ0BU>v|%vW9J|%;g7lH)
zdt@)G3i*E&$QMe{6yn_FvFF5{9e;K_dpaYrb&F=(?X{;0f+uyU-qe>9^3?CF^ZLN0
zu_9G}w`9?&gX)ZvP9AK^x%6Dro{d>EaY?Q*&x*uSue0Yg7Ij^`sO_<kL2qAf0(aS(
zFI)mq9rHyJQ{_XJ9@IIkacj%T$=nmYSh);VIb=k(a=3IvbR9dF+8xdx5r0^i`Iz`V
zgYD0L9G_Wu`ss4ndvDZ5KeB!AKI*^oSKa=VU-tFye(`g@5QAyPnK%ZXklia6^>fQ`
zE<0=1W16)nbBFJ#Olwb-35R+__**$%q`O$@1h-iV9y@s4J1OzAnq1-KV^+*dJxW#v
zEeSR9T$#6;<$1b+r%D6gv18XYjf*xFr3Iu}u5t3@nAnmOnQ^n%Yn7XltSTd8GD~Qf
zM~~m7mQ_<!5>zG`d7LPp#TQ<7dDr7a3-Rf>_q`Wy+od^4uX_UP2`2H)LLBq=iu>+P
zG!vI-@ZGKSm?29o(stg(yZdgvUe~3ucv6>T#-e5y1Esqnfv&go4Q#8rANKw@&XRWc
zpumk~$4n0MOiE++o!!iMJWy6~TIN=*5WTfgt<z^kroIqf<i{gdyISl`B!g&UG?O8R
zf{T)vSJ$GotGZOu!&Wq2ZI^P`$Y4|?6Jui@Yw+Rz!9xdY67JuR%iHQ4Vx3to8xVIa
zVdXu6^<3#;E)n^E&;4jT(L8~*ORbl+zaX6T7{i4~&4kmwCb!hCs-2BmGP#^jMyY#A
z?4r+1{u63`ettaveS#B@@cFh9=RS!iQkk22O>QnbXK<M3pg-f%>EUd7dCdCL;`%yW
zGews!30wO6``by5DXkktlsE*sJ60XBFQ39A=*p7f8YnTz+bo-P`Y}6$3noIh?TzKR
zx&D3o`fhi<M9D5&>DklgpZz0mKK;b!9d!cix8DDjY2Z5|!6MG`AmZ)q<i{%~PVAeN
z`Ls}Dv5CL!=H%J49v*Ivkw{)N_kKlW{(QTdU&mE{NaWsX&$`-exaQ_&wX-uTRL>mD
z%UvD(=H}+Kxz?65V|7;VSu<aLGLK<QOxDJY8(Y@LwKmHvnX+VxLcpScNnd7Y9Milu
zbE3vlorPIb77MhB1Uf8YN)?~H&nGA32g{ykfeUXj<=mK^(z(qe`=L|Kmwx9<wjbVn
z)VaAi?P%Y%^I1oKOYGSCarOPRb^k)E-b8G;cC`EJocrsm_Evp;eZ4hKwt8Ytc)#4<
z?60q{@3)Ss{Uu_$;{IB%Y1_7K_08Ivb@l1frx(RmY;_KCT%p&H#3<366#U+;!1Z*w
z*kztwS&2n=Z#<heThD8e;TAWsyE9BOG-gK2h#ccc5lZb^6Va1)WYwZc47G3M_a58W
z+8p=m>#nNzu`hFey?@o0ZXRiQ@@&1~pCds}Pfb0*s1cNWT&Ss=C*ZeVi|jcD#>Ywh
zazAfMvc>Pq{r%?V?3(g(1;x|#{E7<h_FR_Rd#Co{hXW>;6AzoII42$L%1KV1l`~WI
z)Rd5&sls{p^No!cuAL19t7kXQT)q0>@#>i~`I4tS-5MGi6u7}{)shvP&KUT2S#P@~
zc01GCEMeKx?b}3UVkJIwBrTF;7T`PKcVGO+;r)C68P(oQ-y$dQp|iQ-t=@Ls>zP8Y
zKa|fCcx!Vn|65<We&>;j81<Ha{Cw;R9M%O1yYu_*J(!iMK5f3Ro}StqFSSK;V%A-Y
zOkW>Xv+06ZWN4Dsn$Sn9rA$(l<Gq|bPEG0knX%YZB+#l?c>VmDGiT16QTa9N)74d=
z{WDTAd;OOCON%f6RcbZU)@N<M+<srHUAE?duXc5Y1_f$*YEA0p>NNK;`8#v+j9pQx
zvy_g^oLp>d%yn^j4j<<fmsD=?ttYz<_Vl>E)L6n`vS007T<S76F)=>=9=+T9CEos$
zz4m;^drSYqM_KK>HGDBUiyo)0HuKqZw0!5C$kb_(shc)Ui*%iqv1!$Vf`^CuR6_&Z
z*T*gT$s+%LF&h{AVUrtevvP`4#G67Eb<WIZ;||eCSnSrjI&5vo%B*X-Yu>z${LlZp
z^Jnl6`{u9D@7jKx_c1!$aMQd$Wmam>uD#qEa4Yuv;_ci#vqR^U+_`fqEjh}l(alYT
zrDziOkAHe;er|5sot^F5?v~%1?!ENVJ9SUnc|W)_qMc@M(3`U}Xesv|#(NAR^Uh3U
z35f9KFbos^e%|HdgL#~f8&&Imr=2qM-Cg}s<>loia~Hke^^rxuk;5S2;C?Q_H;Mv(
zI24+MmdZCJ3H#4B%elGP-J-bHw96!8R<5$I$z`8&#x`epd3kx6PV4M-lSu3?b8-2x
z@YC_f-4f5%+C`*A=bun`a+o<dI^STsTAID?q7LItUuXJVJ$rZVj<*LN?v^=u%$(i0
zGI}%5(a?^sQc}^OtXyA|)~?+&oyQ}lt4lD+rA5h;h1uD1<Mj5-mVXiwd^_T-Z4=+x
z>Zb6VomjPV%FCeHHB3LaZ{HC8$kJHUP(AVS;r2VN%=;KVKG+ofG-k)i)R}8Goz_{k
zZqCl4r{W%q_w9ZnZ)f}SgM!?j9|8(aCYQszMGD&tXXPyRow4WUvh`MF%qs%sMV5Fk
ziA<dyvgy=$ZJR|K5)S;jv6eMKp-GqPLZrCc<cBjNu1=gXt0s86Wp3=YXDJVRZcM4%
zb}KhrI`i;Mp>Af`+%uaV-A*=Hog1k&wUg^aHb;%7o&1Bw>wAr4dJ}l&%#M4NY%!yv
zc6()W!~5H>uh&+URN1n8v0*B*ka~Hj`T-;B=jVGru&;Od`ud<z(ZfScj+>Kj&%3+3
z?DdYqW0kKLd@NtF=&yu;U{jBXqCA&~aGzzxMt8}g#Zy9NhpxXq(`fI{%WEPwy7hL2
ztO;K3H;?DcoY{N2PoF-Uy+&B{h{j_J*QBl5K~q+(TInu$BwMxVaMYH$Oqo+&_r;aw
z?)B^2RBCW;ee~sg*5~)9oY=GW;f)uSa?(1!b8}*5Wv|YPIJMzGl48S!ERNP~n~V;2
zEK!k^2x}7St?P|ic;bmoKgVp1B6%(u^FucBTVL<4<NqWZzH0u7x#4Gj?dyNy|M2Ji
zxj*9F?|zRj|IVNL?XIqFu5q8mW#80>!@OB(%-ovS*z@j+yuHt8p}W66?;dx-+asO3
zh3@XS_V{@Jwmq}Uo)(?nnwIr>xnJDf*=CtHHYP3gnyM4Aaqis2()+KtHmu<mI>A~P
zE~fb2Z`IeGxe-rg4tvj9vzpylm%BNX<@?i{E$e3HZrgEr)sy8sFQQ%+1v9HYTb=fF
ztGlX}(y807B3rl4j?oo#>RO|_R9iw<cg5}8r7m5{5{@hpE^aEKi5v&t1x*t2@aSZh
zyL?Rc*uP)jL!xUppLu%!2LscexgRW_ykj^tm-kR$u3pR{j|uMf9*hDDI9Xb4pBC-(
z|LOnREq_*?;j$+wB7v?#=kD$<dw*Z^pkB;Q)vMd{-rw4obX@nhSyIG$vAxw7pLkro
zdUdLHSls&dsFK&;Ute1rH8<y6f<VHdUB4D_-EvZ3+Od&?$3a0`zOMG?lI^-TqPM+S
zs%K%cG&a5}a=E!)%%Lee5`}$_3N4x;7J0K}L$rXm#W^!G)8&QMo$D-i%*Zc}wcaFX
zy>3HNQhDq*KK2G)UM7A9z7j)2_7~#s^WS&Y?(_Y`G2ObMoMC60tyjrXujzX6<}(df
z$Elq#Kh&kMtL*J9i?}5}zlr|{3G^t~aJZdMs!;J^pvz@{t}cy5lb<h}Cu5WOEZE<e
z*JaY`S(749raxc0pifpMz*D4RwQ8)kgtx#Ej}C<<?)fSk&P2AKNijI=yYSGNuz4$&
zWnN2ta>|k`ZN}|Wm(IA(*k+M&VcWLb*Te1_n?)~ZR(16~x>oALVJ?NvuAL1R!$Q3j
zQ<YWjPpkRA_Sa>LnKOReH7fZK7Ok*l`vLX_uD?cXCY<*e4BP&3TCucpKD1uI_kwYg
z>Td4hn4QN<LqXv5>C>kza*nsI4%{Yj%y75F*+Yi2S?(AdZ>gwMKHg_3;b-wieO}4u
zz9&oS1bdPtjxWgEWo2a*<u)U6)~#E&?ml4jKl#&MeT8lCRomdFGt#1Sw?%cfNlv>x
z-*DG<`8k&7O40*Y*Oe<h*%GwsTE{7M*G`w7z#CIt3a*~cT^lr2XKT=^DNa+S1qya_
zi*acS8G5XEcKxr+cFhxiocBFWPI|Dzg75Z@>C?+=MP69nxVWp}SHGO~hX(VM6FwUn
zT4s5wc1`k3et5!P)#TIVcgrrC^$KU*G`UptRNYf$l78Hx-`rm&c?ktvaO6FF-sX?c
zhckyJdHQ>*Ogc0zEy+~C*_3bYfB)qxSGM0ha^&T!tgK3#nG5Hg`Fg(img%ZhQ<hB1
zlyaH6(%0nYIlm>LnbA1`M`McC1v6`!h8z}I=xsJ5>ekkkNeZiuc6+K!a%h<(WH7Od
zt*@`Icg~-`h1X@?oSmd|F?!+L*|UzTD>Uz6bF-h|6K)}{XT9?d!?Y@u7n69au5emR
zOq-NxdGHX!$;oPwm64uCp0hknv-W;|zo2>X`rVd~k2gzSmY4YZa{`aSi5l@ovui+S
zoU}>gyu1urdUtVA>c5!7#k?!>HI8NOQk%UjZ{EyZsh=J`+n_(KYUxkrtG2=IpoI~)
zb$_NkY`t@Ir?F_s#j6@OrmR{qB_L(;(?nCX-`7vfb2;QAa$x_>yA{XZ=b7y(52^T5
z-h1fYUwi+=PwF#2+8O-V$KT-dzJXb1-#y<CJX)IfgB4W*ocHvJ>^?i&%y(CwaB-{4
zWzW}F=BrHdRO-?=X7l-n%EE(sF<a7<x->4hiU=(CWzmuiTk7;``t1F4=fqW3NwJD9
z{<j`<ChuYX`Um`7)gOL1Kb$5RmiSS!EoPUZPtSo(soEJ+{5G|@g%xB>m3A|f5}RXi
zhP%7<GLOm8B28a&k)@JeK?}N7dXswCw2kg2{GXjW?fbmXQ>VQ7_$RruedoRpte<!9
zu3;>{^lq=-tCCticlp1qkM{LneO0qx`gL{wtskr>{S|HIlo?(=>8-7)X7uSsEL(@E
zV@rTzyT|Ub?(OM&Y+ql#FIV}&zoqid^LYhNq*X5aPRm?2E1j3ehV7EZnnPV0i@G${
z3T-KPnfdV9vn?K-qB}qOcV*7Z-FE4UkI7Q^tHv@tj$%?aUfXUZ&5=0#WP$94)!RCb
zM7T_JlsG0B+S7P*3QJGos}06EuiM-l`|sr(-ZTCBZQq1F*Lbg&?yib?SFS(tR;5gx
zx%e@eFd2Q01s&D_H46Oq?LIKDId9KvdEm2Q)n%@G{7n1U^$K=H-a2<J^3`drSp^H#
zIW;m}-1kh0Sg}}<Mb*pU<=n^T=9O}Zu)CDB`_1^wemu1Ke6lmszyHfOUOx2l;lr1Y
zG9M)+Wjy-tYBSI0z?qrvk^_9%E!TuCY0=&BbeUKmM?_7h@zbTrTW`-uoHV0*(goKr
zuEJRF++{rrOG9)Fj=DyQiKsd@#_iK6`!DiepvL}hg16GAl9Cc}hA-P)10v^5<-E*l
zJjZd9qSt>02KM*Uh4-!Y*_iT8pX<Z4=xwuV&pj_R`55dfFQoBfkIzh%vnsQeOiqdZ
zxWkcM-Aa0U-VZjzJ{HEFd&Rycms7-5Hl5bXx%Bh2&c3{?)M>GC!aU~<&N7?$7@W&X
zYR$XRExOS7>3o?p#k$+qoZf8g73`O$yFGVT&E#@FeP>CzsZpoywVYb3oVqolrz5g4
zU`c}uHxJL9Gp8;4Y<qkY4qW3+*!ynhZl|xBL85vAB0uk)S@G}gK^GB6r&i_-9~@RR
zUS4RC(`RJy<HzxJhdiDx>ayV#@>GejD0s*s;Q086++Neh{Yfm02h9XGoHaPHOyb}s
zUz3|fPp2Ct277JaeBNP=N!N0J&QQ)(o)%szlN<~l9tvP|5m9pSU=UJr-DPcFbUSzU
z%uH{d+qvRbc-EOXJ(wBV-SsfE^30vHS1y+pY-@SobY>w#vdo(~K3>08x3ax^zcly!
zht~6^meXbGtdIXT-y`~Yx!)YKEEU%2lPBId+*09Zk@k3YJ6~A9S<}hNz9xsIx97#l
zel!2Dr{cq*`{M0wJi<H%hn3w9%Umw#Njp0uo#nBwQMbrSNuR@7l3^=br^Pj<K1_f3
zyx~nw+Oo(g&vxczJDaE`g{@T$H4+U~bqs85^f?1MFK^0$!$!BgGon3nr%s)knYBpE
zWWyHOX)`!Z^g4B{j?T;oWmQhrP*zyeq8Pc@H6^8MW+qQfP5wdMQn%lm<_T<lZed~m
zMEc0idkWm^?kKD~y@T_P!aXMY9N{uWMJE-Zymx!=-o1ObnZ4|vQJ2QNpv+~H&C1VN
zZoGcJsPbH&<>SMBHP#>aEFLd%X)Ev2d@fTm_wdf?7LWU49<Ss#C}qoB-S3br_wm;9
ztp4c8;??VR8}+JeEMakMU}HZct#w4`(4kcwn@lo{dLBD-i;8I+6Rj&)8ra9RLdnai
z%g|-*TCY{B6b^>+BrTm=c<JK#jT>f4{P^+CZbi->$M*L2*4BsIw~sHr_(iVy-1DLz
zGBYmL9$!?YP{L;z%{eLJ$IBFzX~CQuPO8lCHMwI^@OaNtlN*7rx(xpQ{-)sYxM@+r
z<<8@xDuEM~ET5m8RQY_q$v1(A3tB}iBF=a!UR`C_Z<BoS;=zj-58t?P<LsG}GY<yO
zvs{-E5;AF}#DrCogjh~(*&;b%$3_vuXxG(ZCOSf^dbPMX8hUcK-OOrQZ5E(0VWaUQ
z=H!If#+|;6H{|*bw7W0f_V3aB#hdSmB)Z;d`TP5$?bkcova6>sNE-{=?F-n(I{g&m
z^bL#~w0DB`E9-2!1G?@*=WXh=$Or)ihel=|2~E%nfmxZ~rv$pHeNS6y>G?|Ct;-_t
z%$YNXrpa8H_&$Bfv@WZNpet9dxXR|H1g?$RnsxlT-C-W7Os69T?i{@07retQb9b^_
zF1uYAvAQvj(=;|aceVG-oEb@aX4^HQyFLZ;hQ3^5nJX)<v013cGgU-k!il=P+BtXb
z+}X2cZJ|-bll_xY>h$D<cOSYg!~Obf;od#Fc0S1FJ+b73KyH_f-gS-UDGmaz+td$L
zPP#vZ!SrVAu1k;IdL?$3?!C^Zx3lPJcHiCIk&ln(U2uHGVeiDkA;8Ys?Awv!(e{xi
zO;V%i%#3uN$G+w*$!83^rfP*~t#nN<6h0Cd73DIyg~2K5>N*A$m#!PT7?fH9b}5PX
zb~x*c8$5e^=ja^0H@9Zw-g-H?FS<f`>h{K&tFPUhx@onhcf!(+xgv=wjZ)fCno(!Y
zU6Xn9=1xz$_~(NC=AY}$KNSQQ)>qFCo633D?gxiDSIW5`96{S$l{6L|Y3=23WC?F&
z{QY~;sj>&4(}(9vy9x&SaP2Bt)h|*t`^lw)Y&;SghunlXTsm6fR03Tk!!LPdg+D(x
zr^`emz^+$#B~RtdwkwCj7EaB&x;kW1ap$?MS=ah_B(xbKHg@KOvF7Nu?#$KQcGtkb
z;Bs#6#u<rOE813Hj@`1V+3iS)#-bI6SzSC79XF;+9zArhpdk9%x37QXCEosL53#bY
zZl5%#_d54a=1uDJE^kjNTCtPC{s4p3p9X~|49AZPc`Rtt;!tRF^YiobXY&gS+!>j&
zVQ1>rtmvel(+v*u7{2V0$`NBb;@oHPcnMEY*Hpcbbxg~I&9YfVLjKi1o<4p0#;HH+
zBBx)!_3Ouv&x$9CV$v0~Sv5mdwX%aeqXO3YY_Uo?@xbuWszpT{&k~*`?2LBi-4b28
zYSm(HuPNCso2IyQr)aYVIEE^<rY5?DE{Rl`<mQ+(<NcGzw;nxu^yoF$Yq1Ll^S?GP
zEk37oM@EKwPac!W0?*B!dWM!MO@8qmjnBJtUq4?TxA#`<(W;-H-}csKSj{T<vhmB8
zPew*2Teg~f`Eo1s<;#~FSFT*XeA##3(r+aXUS>KU6!y~V<JFNW+OlQqvzM8fF?a6#
z__KY7`1F+x57?L+L_JleifCL7nWC~}VsB_nT#RA25|`Ast%8n8h8&wTRTwriE;y*a
zC#lpqBr<yT<>>x97UeR#e@}Z8vF6m;xC_eh2N*tkSSK{qJp1f<iz!ldQplvjlUgAw
zRbDp<Y!veImR@pu+vVK0Gv*0%?_Qdr+a}S-aF4OSPV?fTxl)G5d5WH%HaNN2$Ecd^
z!a|3`JV{fPdG44t_S@UsJ<H3+m+ag7H~4G+wQJX|t&Q6Dw@f1W`Okm(%W93E@h+2h
zS$g%j{^gm5mmW5{%$%}z+td@Pw}PfcTPp=DabCpowrIgZCxcbIxw<0Pxi;S|{1a{Q
z!|&9;f6qg#Wcl(<ckHOm3*eY1z&D5e{gKaWo_*MRW@4O}$L`pr)>)c&-WBTiEXq`!
zv`a&w+4+3d?i2w>76FIv=jYqmJYH~6MkB!2<m_XfzFnVuLAyE3GMCMExa_+m_gY(9
zU0v$icVVIbwrz8>?t8RT&oWSJ(v+Ec-a%nYHNC<#T`x_!QCRl%;2DGKt2fWs!#jPk
zi<Z`muB(DuPF+1SV@~Xr0N*2SihZ}XZ9B4U+SRQ~yLvaRijG>dHZmhrYK>!$8#7Cy
z$|NI2mW^y@xOv>g#dmzWbwfh@`t`fqbJrX9?+l1roOX8B)8CS^=bjg@zAf1FmeKsw
z!J7wfX>Z@<$K-k9_}vqCb@uGpv-_DypzrMH?RmFnm%X2y_tX5r6i=Iry0;JJmrU|h
znYlUb*4%7nv6WF<{mxE(^Xz?ke4XX=pEIO>?uh#Jv$8LE!l9X!xf)kPmIehb%`}#h
z>@Aa8<Fx3=)6B{)Z?4{5t74|}7~Gk5yUk>a&_t!VGx_WP&;Nb-qw^={?Z4s=a5FS$
z=5qB)nG`StiZZg?T*O&-{2j|o^Hw#NjtlW$|0u{#%W0k3YLoI|<>Y(eyEd)466h*Y
zFv)LHN%xjOSMfidj!i9~EyJxG6@J|!k9QQDbmTNDIOP;2tR`>H#i7t7@<VWiVNc~@
z$?lCym&DD?F*sry&BoCf&HdqYli4<T%~px2hGM@K)oeVT&U5y3cB`JFuv^ZZ5D~wu
zo2R#3n%%l^@ziZoBYGuHH7!j0_T!)Z?qB!U-e+e$uzUA4+0$!ewmklExO3;PJ&vFL
z%vb)Q-@&1N{}I3a;~#sRvMYRDTO6+MtnhmDXw|A!8q1=5)GjVI%em;vBE3DYiNn(7
zai2wwEX%__#bs_ifhs4q=iS{o?b+?QR(F1fXasC_IxA(CnHzO`Wzb#kQsaW??fF0K
zn;-Jm{z$(+@5{1F^I5(>t5JEeqd@1>&e_wR6`rmVwwks#`a+sd)z-(;X5|{LU3O%-
zR@Zvft$LxmCl;;ZDRAp;ow6V_XlkNMGn-CFv%rK2lMIxVCh;&V2<`9BV0fHdwlO!q
zTKDU>bcLq5_7+ZjAGqAtN=fWEz@+@ofX|TK_r_eyJhe5uIl3j1<%@UjZI_mK`bpeb
zlZ8LFB+#`4Jc8%F;3NC`07upbdR-g}?My=aPY-K~r1{Kf6A5tY;>cHM_1dZxA{%V%
zV{EhO^hM34(^nH#D<(8?iK!_x3rNhDd^KNYr?J`XZC`)S-kol6U~k0EaHc8cVqDx4
zGBb5V-5g%3sPDFW^FQ3#Q7zNx<1d$EN7Xb94fS_$oIb0tIc>AnR8F7y8H_iXYx>eA
zrR*zIc6DUox0iVM#PCpATwKI4v+Oy=$<o{Jn_TzqeSC0ll?~hM0~Qs07LPZ`YkYqC
z;6NMa=MO5ALM|@%l{vz{GT?I7Rb!Jgk9iCZn|wK~f0u!gp@qG!G;99D|ND);eR@B)
zd{LL_>(J$1EiX6yS!oy2`&01e#Yf*0^XqT_{`>yjm;bxp*6F|gcW=7jT4b<}ZCzLM
z=~t`YT|E*Q`rt!_jT}FF^U)u#Z6;s*vhRJ#nKwH6&hyVb+x<<>>Tli}mA+|dKWp~I
z>rX$=pLhGM`{!R3lT<ug4ka4Q@KJ58YL)KYcRPFQyw*?u>*VxJ_dh-xn!o8lf`J4N
z8&jW%iyL25`~S!r*Z=LGdiCD=TKDG{e?I^G%KFl#O{=DS4_Pr~$`oHsO-;$C-S1po
zj{K}H3=AnOEX@DiIdeYCl`GfhUwm3#{`0?e?e^s>GV*oBwnz6IGupJz<4TdGOrN{*
z#1l{I_Fp&sCcpmtsYkP~zu6Tl2Lc8&+`j!$6SA{sZR$VpAo=n8@|~46DIwwPU~vEU
z|N8fT|LQ`4I_R8~pPyfsy}i9X@9ysUi+}#V`&a+$U*gwae@pA%|KD2v`}^;|KVE5N
zW!L`uUu_Kow*UT}Lk8#1JS#Q<tvm!T`TbdW6AJFPZI51i?RDv{yLsEAUw{4h<7=2`
zY0;!flji)ZS$^%__4kD*o)lTi^yrCmx0ZeX{`>F0|MmL{kC%fP`$Lnn@4YWy_q}@W
z{i$=$Jx@tVNvW9Qb=%tw2K-||yHIx)J^vs7^Z(_yfA_9k0$LUXo{m}{x3>v&<hx6!
zkc!9rtAFFK{*C|gfB)aT^ZMNvUwrYU%2ux5y|B7=MZ@vrjW^zuS<ChvPu_UH^5uK`
z`RAYS_+M{dfBx&Ow`IFy*Im!vdh4yH=eG?I@Vj>3{ru^tpMF}g^8YgpZIx5m+WWR?
zBy_ay|4#_~ku<-+z`*dWR`pOduLuJJ!-W+L%&*_?Yd_Y&dWHc6-cPVS0HS{}NcYF9
zO}}*SPJ6O!@lx;7=$J{X97C0^-ELr<dFuh=s#PMf_sjdXI{bbwS0V3K`@Qzz`2dDh
zi`vuot~`Ba)@>!GMpw^mtPU3!>TD{P`}k6+$x%V(e8JuK-g`x!Hf*oX-krGUgyf^p
z+W{ck7BH|fFa$L)IxsLW3NV0$7#KLJ9{aLip8QDE$wAe9f$?MpQPo*{6&TyA@5{2g
zZdTO|W=XtxbWfh)_l-I4eSSZcU}cf1T=O_+H(wcFn=7l{0mIOsotyd$g74+>`@Qux
zcb;n0Fo{#K>ePg_E3a;6d@mbu(|DJE`QBF^uVXLUKAg~TE5B^lp>Ns&v)&!m4O<#C
z<!#!$jrwNG+7vuB19LL1-d%hxu2b4B92j;h>hUGj>e-oIwRak9tLK}on;GsKYp%1(
zSK?dVjp_F;Fn(cRkhb7;aFl3ZO3Jt(%rHZ7!67NOUK5sM%X-y25(Ew~np{wmN=Zsk
z_F?uFW;t=uAY=K9i=7KDySe$ATvkY#=hMkD(U8NJk%@_yZOTWJnNmEyF1gBAzVZkB
z4fa>z<(b96z<fnYDkF2*Y>kV{S1_<OE1Zxr`3ee3h75)bh6$b=Y7-?ciJX<nn!jNA
ztC_Zv{(&uD?L#hHkYao#x6GGq7K7XZ6BAy>1z;^uLW+UG^VO84J}ch_E->OXTTtmM
zF)PEw1mtqY1x6W?Y`u&NjEszf6+13mR$P!_!tB71!OLKBf!T3cPXkY{iIEhODu)6K
zgAWU{=fcTqjf<*WZBE*<d}Y(heB-z{(AHmS`4&_EtobWn&F^YpVBm#X3t>1-5@=#j
zFj>CX*H^vSbJ-=6Q?@IbeU}*W^fEA*WH2yfz*Iw7j1xjx1^!0<5C6M=?l*(~k1Z@j
zgFo%}fBWzLy*D!LnwLuD<kdInUgz2vp`Fk4O}}NsVQxhpe@oHQBk^orQeLOmKKpV`
zD_wS?q~c@|W)0(MzdMSH)`}aJ?CD&n;(h1tp9D4*578|pr7X57mruu5?7HlCa4(A@
zheOV(*BjjSbRS){uz1_v_V5+cFBH#t@XE`!QL{GkMC9+E{C6Twtr06txLn<{Wz)el
zr6BuKV`Hv~8**-LPCL6`+s2xoM}OI_W(|2^)mlHR;}HwPhR8<~c^Z@Ugm6X_T|DA)
zc!KKFd&k%<&&jS@viGl%e7jJ&PG7jId3f~1z0t?`x_3+!`4HzH^U;!P*M==qO{Qur
zUA~7qRa9Y4TtmQ;z;kDAy?*)n_q~4~tsn3(On7_MPJVZjh-k5(UidA+przY)H^@4#
z-f?=fp7i#ev8nwz?1_vo-W_?g{=Lw)^NMfv)K%y0+P5v~xeX6%=(M&;@5^_cIn$|r
z#pF@w+eadQExdi1-FG${RL*jH)2bE1w$j76^Fdl9Tit}z=+&v-_hXg6Tk6bjkeGC^
zV?#ejU{{v;52IKvPW|6ZX$CF|i`7!*Op;-H;v`V8)w|90@n5H;d0CZC0<oW~czOP|
zp8L#X`1{W3nm21dv$G4f<~N4OzEes{unOCAEaup|GpU;Aisvmmsu^=^!qZ5DkUf)+
z=;}T661MAk_vho$&h<VAt$Hk{TK~S5YA8M_o4;hogX%N7r~3KZ->3N4wU@m!^beWZ
z$MtPLuY`Nnx2>x_+s3_|aXziDG$SLw`@UFWTkP2tGW%ETKhRU|d#fzRw&iHorYTpx
znT3CiXti7V?{#fZT)9KW<{#ew%EcE=E0wo4@0s2)t=RtMx4P_8HMdTwUI?DL=X(2-
zmAXYAvNZ4AH4NR+S(D>=tNoMKr;SC$?>63)Q9ZU@deYL5&!laaetg%vv@lNh)pN`6
zuwzVNzovv=zkd4K>gu)c^+dJr=x+XOzT%^+&fzQ4tl{pJ$9MuXn0W$%nz>Y07H-eF
zBlt+{WT*R&`(Jb4&8lYW*pRm2+pBJM*UnI<ed6jfPNu1=%=5h&cI%^Y`~T2_%+;Ue
z_>v8q8db9nc|O_nHT<UVS2e?_zD_4v4tne7uE}-m+J4<V%WuxH;J>l+R&rK7i?PYS
zW4A^#e(#ET?+kDB92Gsj_dtkuPwAy;p&xWCPi1-^3w?AcQs?NZwq3V1)f@NTTRz#)
znqTglUQ*KAw$lmfE3-aas_0tHy)P+k(r=e{D-WLxUbS-dhqS0E*Hvb9tvTa6b9b?Y
z=hUEUf5m=X7d|JnF?CslxbS7x8t$%>eBo>NEM?8~+O)E6!J?>bQbB*dJLXj~?$B<!
zoE)#L!XYoRdfo4xbzhfn*ZlL<FgLU}M)1*h%QN$wq-J=%z86|KJNviH*_YE|Vw7&r
zk1m*;bp7sCyW@`vCe1h%BYT`v=%>;`D|MkWhJjhfulj@>?6#h<HPN82z-RZ2*1~!7
z%4e-Rc=_~&)ODc}zG6yEdfoC}+aBmgsQ+VLy?DCmmM=eIlIK|Kz1vbgt2%g3dU?y%
z$LWRJWpr~}=6F3daJXK5w!ybEOgAw&Bw4ClsI}`?s#iet(K5!cEbq#~l&77}!Bcsh
zjrX(5&9IbiQ)~Jkn?HZU%!9$g>a|BMJX^i-W=>z;^@njbIq&<+PlTrC|EziEZ~Uav
zvgiG8$qj573zo87IQN5Lrcr$xi-CrwmmZq|n`T#eZ1UQ*7H9A5=6k*F)v<Fw4u*(0
zD=k>ie~g7Cl=<+>vY)?;|3CbzvAj;UBP%_;Y3sY7sek2ftPfqc{)$4?g}?T^_xp93
z9F`_3O$%!4pY*$O{j{6@ol^U{g{`kiYreV?Kd+^&+FI8?p37Qwzg?rpBxR)&AA|KQ
zyynjSE2u4ScPV>So@2?mbqimz2ry?XVAEyl<>=rz)9fsybVP(fWr4P~kAt#Dw?or`
zF0HQiD?j72U)1e?{`dUeum6L;#~S=o`1oI=Ua#Jy-m31)zm6NcC+`0`d|Tcia~XKS
z#<a-PX`l;qL^J{x9Nfvo-KH+wEf#ydJIK?=<oEqAKGUK$D?Uo=eDc(=w94j+uUfX_
ztTV}<@1^w}`dtyE<?7XPQ7DJCVXtuTVzmjgg0fa{N3OBVE<LV${*q<hZ02dNH(F+I
zOV_=f<-5`B$sM6HvnOWk%%6QI-{HR6k!yAbEmFRDzUBVLa_^-b+tY)eI@=@?95@&m
zJvM4QTb8E!+cwsw`q%#S*GK9*n|_v?UUm8G;QY~@Lr`gUmy*W3xW2BS`O(WB3Ix2b
zTeT+!4x+wfUs7TF{eB+^<?oMkV(E==m}DQbf7!qF-_#HObDMQ^(X^v4H~j9goiTUQ
zVIH%X8HuU)g`>BrFHKFIelxMG!gt<r)_GUcYGT8CS6$88R{4Wv$E!Qh9cD{hJ<T6Y
znZsD1zD7@F8+WIL(i1;Ht<Kk78*H1ZRd$DRZG0>IFQz;Co$$R~;+-FA%qQ6Y;M0&7
zke852uye2ruy|0!@TBoG<0;0Ej2sLslDBN`Wvfm)p1g5J%}e(sH%qU_R#kS|_#V%D
zd3lYWR>9;J>kkZCay*5{<#w;z{l)r662q#Gb^OKOHobj&>t?~>pZ`CqO`hB!Gf5|6
zT}sE(ok>S4e%L9nu!u~wk^Cw5M#3;OjQRQbxmFi;6g)oo@b`b#I@=F5zhdTe9O8J=
zD8Luc;l#ugem%DM>(snNr<Pr>*X{OQzWmjv%Tta>FL!WOG17PYI?Kqe=1+?v2c!O$
zIRD9Y`y=k>Z@<mUyNg-JVES$^?b~a?>tWK)_Rl!~tmgd8$F}nmzsaXhcw))w?EbXM
z0tTeFmpyB0Dm-(ebQX((f}y%sQqx5Pm%X!<pP!#q^YHU4^)JCwLY&why%uQ@Xg1N#
z%zXJK2M%)mB8#3LZs(H$9g5K|CLE};@XC>G2X`M7*F2afu07?)JmEtXew_y`^jylV
zjb~SzF|bsqt*GviEqQ|sP9LArFK07t+P1TfIlFU~s7+g@YN&gpPxf}Jh0YuVC}3#O
zUcF)ZD)+5QRwt}Z>#XCgZe$eK&=XrSX;N><q=u;~C7!1yS@36Yf;JpIKHe{9Q^v@l
zyFI()&b_wRrM2b_vuDrr37eU>)zoDE)`OsJx&4~01~tDHOl6rZa-{5?h(y2vMFtke
zMy{__t)_A+AYZi|+7gr}nL0aqx{>!0<Mf=}<(>=<e9q6$&$avTK!7p3xp+x$WMr_X
zqnw}b9`iGs4)d-rSbuJ=RmqirM==vSB_=)m{QUC9mPA7}m2QC+mg#~?N4q22vyMGY
z^iPy?S+ZpDbiJ6Z8gDO8*NZvSSX%RQL(SpmdC%qdR*Sd_BpjSAsQ!q-LAJ}#(XnFY
z!h{%A&8Lc<Zat4g)ZR{?Xz}s!@xHru^J;#}t~)>V%@#pj-djtj3M>p1Ql9B3WRfv)
z#gj`;?1hWn7A@Ly@l@BXn~`B}Z?D?SJICMTvTxR=?n&0|e6rT1fhwA^{`OlZc}&~r
zafgY!cHd83vCjqKRwq6_KHj#M=gDKAO{WbO>|I**^<8rG2WH7j>s6VLCd_>KVCEkQ
z6DEUs^OlQ>2p(kUPu5UXa_LySRAg@C?y^8ji_<zz0vrzyfB56ep7(<Jk^uX*-?jTb
zm$%8Tx))k{J+}0L!_n+7s_&L{ns_YIC|Vh3GfDVJt&Qgc4aEQv5AjYxm%!BG4g0JD
zJw0cpy)n<bySrU<`|ZVR^R~xMzrkGNYCVzhRcm87%LbtcvyEPwE)EhR8+A*=Bs8NE
zc`H^Qov4(RxSVIh!^0i-#5>n~VsW1P`c|Qzyj|6e4GQ*-3DJqsUJO}6U(6!{7HX~W
zY`vb?dAixK);`E^O6D?Vt*OiXmcF<b$Yo{Zz#@30uw_rVK%7md4k#nGB_59P_GX)v
z<-+iaWyu|ZKP^@I=bMh3q`&k!QysKfO2b0wv|!u99XhKWPYbKelwQ6>wPO+gb;~=)
z13eBI2sV9M6r$xAS-t1J%Tf);GgFq82~@7@QsGooQEhHuzJ0aY_}!V0m%6mpOuewl
zF?Au^+s6U|Ml7$L_U|lsC~(&-GIimPb2nn!KKbYz))o@j6DPBCM${BVzI#hcuSd0R
z?(LsF|7c}v(JPGxIkT>>uaow2?_*_rQKcKU>i3KY*M?0eyiS*Q-zf{4C7n5oXJ5vq
zMIlxtZ(3BHw<lF3NmfkRH$60-FK*%P`^KjW9<aTYU9@8|Ptl8-6`msU_M2CJet&tF
z*6KT-6O(Qn7hN&QZN+})+^MV5?x<)i>MYsydfi^tn8`bJ+Pe8_rdNb@taf~SeT{9k
ziu)O@^d~F^EQifixn6gczqI}l5VKb0?(QnC_kO~2r<{znoTA(|MZ4YaY24nF#`fMf
z3&XlDESjKlEi^Rx<hJy4pVDWqxa=n-)Yz^el+dEc<+5zkm5-O+UJVUROI1}qGtD{E
z?af*#gMzzf!W-Bd*gy4{grA?Uae1H6`{+Ylv!%-(9zH%_mUZErH@Cia{w-hqp?*j9
zrE@jCdiT^D55yS$QZGHP%E!86)nRU3HV>x;Nkz$Z+>)=`gte|sK9jRVVAtC_ZK4d(
zVtIFMK12ol58Iqw-+ub@nd2{13oRZQ{Zc+!wfB1YfA;zN_r5(nYg2r6v(4JrT<NPv
z@74$yN8SGL@bCO5ALgl^RX8?l_U!6f{}~<kKbcJW@qD@qf8%YlZ(pu|Dm`vmQy7)i
z_2$a2G^Oz6>k7VYpDt(j?Y2w)@`Ia9uWrBHcH8C5`&ypv3$gq6J-yw2cWa52=(?s0
zKW^@NK2zrB-N^Xrf^N4ZZO3-qyb(2DP~B|n+UsBISKiqAakJ91%_}ovzHR<v^Ra7n
z<8j{m^=G$u7N6awn_ag&C$2j_LPRe@c7FcV=Z~_w9`6hJY1aQwf42XnaPQBSFW5}>
z{ye*^Vuj7${}<mkx3W$6HS4wEvN?C(bg;QzF}JxVdWz5D@ZR_0T(d4)96xFp&S&*J
zI@<YN6z9&O?a%9v<ZfL%jrr$|b=PjCZ7Z+cHhr>3gVofs)usRDue)6Rkd5!shi{K%
zIBp#MFiZN&9qx#h*=~=|U28~{pLp~`QU>?q`m<a64|R$?+0)jyEZS5|@xu1ab%&B{
z7d*1OoV4(xnb?%O`GRvgHtY|)A9m$Z(+QJ5d-M6FZpyF!FC1%VmLj|DvUl{S{E!ue
znxX&1-;_L@e>K_RV!D4gvv;F=eQtv5;!pcG+*_5Md1{LA&#I~%8>Y&%zim^weWnZ6
zpT6#NVV(8o{o>_+YK?*px=G~={>z>i>d)4@O!wuM4DU(Cb^D6<?)-SNzc0pylSj#N
z!)D$@<w94{hBp`W)6>(>#Ibj7ven=b-yKt_sxZkzYp>Vq1N*dp{OA4t{&MyA&;K)j
z&c5--{EJ+5`OCjiAAhU1DNLL2{h9XsKd)Y0f4TNp(jPsG(q}erG*y>8SpMbGzxU_%
z9k?~!bxqOB9nk`T+q|@=U5i-%<_IhE8jny`ffg1O4}pcLf}SE1B3zb=1-J@Waj#(%
zU>0pL6YoEeCw&yO=ZF2vlf_ydE5A>&nHUgS;^4sXLZ@BbZovh=Ho*msdzu9ftoKT4
z{gI-`{r<g0b+_+-?&gExt;UBRe)v9PzrE|C&E+olI=j@$db8qw<$c@NH#>2|rHbtl
z+kMh@oxd`DV%Mok$^TB-`cI$kzu&&}t7%odd2a9Z%Xwd;vga|ouDJ6v;?#Qn>c|bd
zn1psLWoT&FAuKRMXpRU&Cx5SKU_e8&pnz!omH&tTy;{T~{=epM`KI?>wtEwIFM4fi
zzjNDTC&S5RUe>Rsf0JUoe`|AAX;rJ_!+FoER{r^#ye#qMm)a}SKLw`!*}S9t&E{j}
zUsN_IICYBU`DvZ7OUbOdcdE}e=&#K7E$nB+FUmH5-pOiM6EZdCd;X0pwW=51yW1Ro
zwWCld{@bE6h5YPY#}9G|)|GPae^z)Q;?(hprXMd}d$Y`L_1p{lty+^4cRx5K{P1R{
z@ue>pB`&r8YS!WN%|EmL)}-6~AN2MF8aB=d>eW4SNJ;N$<7dxPn(vnG+>!BU{iS!O
zLS6E=vo)O%wn)D7taqbk-9+OLSCzJGO;~*AhmSVXr$2Sh%GZDU9@Vqh?OgFFz~=s+
zQ@Q^Ry>mZh`0s1boOOz>7pf(7HdV|!@JOP#cAb^?gut|)#XHIw=Ol0M)c*b<@$bHU
zb!w~p>}6aYR%d1Zp7iEV|M5L`U$#0o)ykIzi?dBy{k`hC?b$Qi<`>+ZQT+YCmG}O@
zH0#F)_Zk+ra_l#1J@LzYi<r^ICsF(Q!(JvDcHO)&J-bdOIZywnVR5I{+K<yJ;+`{{
zZ!(l)I8t{{(9cSK;c*?MyD#qger!Eu|IjOF%2Lfp!yZQK!&4rg%Bg!bY005kGwwY4
zJVTL#bDm*r*G{b|qG_$k8j>5DpWV?CxcB6OTI#`?mabS<DeqkV+~xLdo4K~_N<Yi%
z^lXl}S;O=Bb3LQ@R;-S#czsxJ{cd0Lvlm1D<`x}a`o4a5=8@8WzqWo#zSx#Ld-t`>
zX~w^ux0aXYmdozm9eX@9_IPyc@vN(N7#J9;zgzuL|6TLn{>Hw=m)no=-1X5bT_ze?
z%qi$&{GFNSxy9kVrzI-l1lycdl9JlhUR7oW$9m_Bs0mv>&f|(``nlTFOIcH5y<PJE
z$5Su7x$r6=eBu3!$NM>=_gwj*wkBnzR7a_XZo-Zgg^M|C?|V)+ecQAoc*9-R7}1Z;
z1*PHZXByWniZXq(bffs)n};W~^DbO-GV`;0%EQEX!)nJE?x|lMepq!?CbYFmE}Z9F
zOk3W&C)+(c@}4ZMliXm@A=vOdQEWm?zRsIDI%>P;&f*pctz;K)FMTq{k2PKI-o-gP
zd5)F~o6T6ZQ)K;-e1%6EuJR8qU0d|r;CRZzgrDaxG~GI~>A4V#UF6Xf+V}p(y_~a?
z=cKXft7VFtRka_APbd^n%ARPlMd;rJzmta!_UpfKOh0;h{xQcSwm-$46Z(~H^CPBh
zINts8qtEFJudILCXeH0y5Oi4J|KyH`g1-f?6-1x6cr!;w;%~}?2&UkzlB~*kx7u|I
zEf0ny25ZHv^sb#?+wnrAQCvSt>PoXts?o9XrzYPe;{LugNVJf;=eWp2@WE-n`@uH3
zH@!{NcDM^4<g<yfI`4hp>fAY(xOTkDF!9ip=s);i;feW%YD{b2-rid|*NSh^p3WUU
z!HHgBEZNE(3x&1?OzVj+PAKyelTeo~@#k~nEj<)(eq@`L`+MJh$0EgJA0Fi>rtOMl
z_SzKEdvy0s$E41AorfNBr<O}iF21Sd94zm|Jylq~@yG$E>vPX8+}GwcF@4eP1^X1{
z{PHtc$nE#{-PtoKE2XEEozQwXFY_=Xx7o8FYzYePzcyy*C<>i<q~JRJ^F^aOZdzxT
zYgUIVx}I^U6YPGgT^(~rm3#Hijb0J{bFa&?#w5S8ULww9Sn>J$?!&?!Hg7t=?JMlu
zB4O|H+y23kN4(9Cf8RaXxmE7({eT|U&Q3p@$MgCWJEu$P{djy`(~`Mw;}5IH!gBU$
z{QVtCc`bJ)|80LgWucbq*Ox~oxYd8#WNK=VJ8{LeyIvE&Z~uECTq(V+`t6?k2lJa!
z4vRjS{X~C~R*%h0<HFNl`Xt-=0*)*cuTS~+d{X*iyZkwe`j42~>@8GRij))F;;QGU
zyz9KL`mswQ8SUzVb3AK4PWo_8pOL$~iY-0bYMW(c$+=mjPM`7v)~#$y>#~$Qv`6g3
zasET1EB@-8ejL61uiy1h)*R!n(JtZ|s@s-@ImOJ}@+jZ!&!<+NZIYMY?71{!Uy|1z
zy=Ap+xqSIYmVTKgYxee#d;hs0A+dc=d?J~HPkp!$)p&H@4HnLQU9(xzr;ESebN{CI
zT<L0=3Ws-Jv{D4;#-vxesee*D`TW<u&SlR|_RiUU`{m=C7iLA~aT$AP7v^odzF<pJ
z)11c^rg`Vgm7jj*pZ}@HU&G+p?>lilzdY=lt7dJ~a?4uR%Gze5bMaVD;wE?gqx(<U
zXd4R_+&${MJ1Up?(vp;gsR40c;y$Shw0D1%32%+QS5$MkdY#wAXMgIVxbJkQM|U)x
z<g9N#!Xdq{vTcHP_Bo?hrSl)$-~DJw%_Q3^VhR&K&hg;i|Kri4^ZTDpo3QohqYCx%
zpUOYhROQ<&J@epLa%CAO$A`_Q4#r$MB^JNfZ>hh@o12P-`wp%6s(opj8L#Skmnlm;
zH7C1Ho*c6;=3$6r>)IDOE9Sh(yb$SF{bt_XEJJzj_g`w*)*SicTK}$Q?vo?$UEhDS
znRe#L^Ir3>H7<8<c1ncblHa?=P~Cm$cX`h)LuDSfmVZ+wcFvSN>-gz^WM`XL{&}{J
zBOi}NalWWw7FN?ylJm0MuvE_^!%*EOZ&LHVue_5rbCYF{Sw21%ke<nF6m+xeIJ=#F
z*@Ee>W|sxMIey!w`0V2;!WF`fngvboFIhe>Ud4T6<EB=f9edvTRR>#M&D6hqP5x>z
z=k-~J-ckY|>|A&_T(8%y^?J5_MMPwTR%Jx-t~(05-45|fvQL;9zv1XkuKhg^FHB4-
zj1$d|Jp03UcSYQ(#<z?3KAO*V&IpQIT9?@LbPCV^Pke_mH!VA~WGA2ayn}udj>WR)
z%{e=>n>$D$^@1!*+NWvWX|6M;Waaz`t~=j-Ho*RqK>5k-HECB>_Agz(;b7fYW}6h*
zmff#rDS1D0_&r1Kp~Klle1~Gvdc7vSf3@q4___t@PaJP}uXYYvvZ-WF8u#xNlD{vM
z38!XcKl1+KdCMS1t8vyB*_PE0CDym{<va?>`q2K~=-50>FSDJ8dmgx5nmniQQSR9T
zn)!DP7wl1}eJCk@i*+sQic^KXLN=SbHWzk<uM`#ts7yKQY^JOF?Zv)t+)p)B|CaDn
zxU65aI<nblVeQT6T8YA!$v-WFmIS`?GkYjvTP<9$PmH<f4kL$?0E7HW{s#=pIG8L>
zaII{p4pU%`pS5WTN7A3tDV~bmFLG9@@C2Qj(AfT9QwWnp(WWITM}KlCYOFA5V6;wP
zNXR<qrV_lsYZ*7!{|Au*7Z??+BE=QiA_7|)9NGdSQUcl>(p;Px*#)L1i77-(ND}A}
zQ*ltabu@A!ufR->1gDfPmyW270nfRftmamlva0QXfI$4w-<P^NzU|o`ee?Hze*5!>
zu5AxpoBzyy)z-BiYO9O?2YSyJ7ChpnocphD{ml=5G;Wvw-F#rn{AY^4Wez{K{@?NA
z)cYeJt55SCk}U19S*dFMd0Xh+S@t#gA8N{@pV++iy|GR8=#g`}Tecm^+g`hP_2mE8
zSz_#uyLTnsJ%1-We6E`||J~{c=Z8P7mDe6PePi*)8GraY&gSyJwMn+<`KE19)8G00
z$2|3#We)pAr%xznowBU#xx=j3jq}TQ<_d|&Kezh*Sbx#kw27ylyzV<Jd3odGng6QW
z54$xtznlL%?e2RegHOHMQyl(hIEkIE_q5(E{4-Ae_r;3(a`&>$dNNjpRX;?X1ot}$
zyq?mcbh@omhS}lY-~3ehzn@Rdjemdk$HaHlaTg!d#J9g`e^Y<>pY8rN|DGQ`V-uUV
z`G>RK-p8jV1#c)jwfp<U_h0r;{`y<~<nyS!<EDCcmo5HP?U%B%Gmm)n|DJuOOka4!
z$9dvk?A5=&n(J2S{_lVOgTL$6e<<A_w?9(;xc*Q3w>SRzKRkQp?}|wc&!4!}|9T_8
z@aK!Z?an_xPW%H}DRKMDU)#KGy&JySJ<j;C=U$Y>pZDk2Uh~=UOv=T7#h?9)|2$oI
z|7^`~tIVe&HG6qGIi8>TCZc_J@|{Jiex24bPxIP6`CV!JvbQ{Iciwq7`M|4heosny
zcBja^tD3sPezyAJ&yzW8AAT<pVb93rSv~jQ`OH_53-7&7&D`m|?RN6pZQs(%?2WbF
zt$BVYF8ji}b6=`HU3v5Us`Tq+cjJ!#`EI+vP<dshdc;W%PL>~Avy)vHYJwY)AVSw%
z<onT}dFqZ_s@i->$@i>lt|)zad;ie1!!uG=pRvDw=GV@S$DispZ_9pZ?fl#*s;y#T
z@HBV5>{nKi^R1N*oJ?n_=8<FH;eP(aVcBV(Hf52sZ2tECE9LUlWj33}`sZSX?(xX2
z5B_xXxv~8{9<=Re)4>PtPsMN8zfPQIrohQs8xAq{jgM=#Khadolx^8y7wj={y4BfD
z%ik1kww*8kEy!Is|M<ZVQ+^a$%yU08Q{d$P<BiOl>f7%<DhycCx+E*;z%?nC-ok6S
zhwOG=yHjR6Kh{%j+u=Xo{vWlipPJv?H8aBHp6u_xAHV;8`|K;1*IMme-ppyq+{Ve%
zZ_AxKU!2joY}&0wmos<WDOoGh|ElcI{_v9JKW{%QsLq_#<9qOd!>jpUzv$YBXqs%f
z{%9KejAUNL9-GJY`~PH?^ew!-`Ki2n*~OnL4jHep(|s*hK4;oqzWFbgy{htB{o<9W
zmCH$s@S5K6r)#cdnR-Q+%KD#=KM)*x?PgAe=eF~cJt_ojg<qOR^+t(oTWh}0o4G1W
z_T>_t+a-O=cVBr>_Vrak_j;XuvlAaL4*ZvE+n;EpIBCTPg{<>SFTbi~(NxS{w$xJA
zaKV?#Tj$C{{g$@#$NpV<R#?5w<e2wc$FH_>eJiET^=)}O>vP|Q%QJN|`&U2zC$%Jd
z>#WP$vS*aF-$~iJ%5L|&=d~|>|J?I=TXsZQ+nSTvb87Rtk3Lt9J^Q8a=32*FQioO7
zMt-e-wmqgdZFOAP)m@Lb*b1wzJNbYA`|aNs9o=Z;o^#u$Hf(X|Dz|cvZK0C;>=J&M
zE#0b7W~z3sF!V!7rdQBhtDq@MRygtXn`(LmKUUTBs@t+u)BnhnrJCP`rY!wdY30j4
zckQ)TCVzY5cE>(`Z4`BXaqO4CqOJDkTFo=pT`k&rXJw>keSiPD>)Fd}{oIaxS^UiO
z;9SeUw&lxLxiw$9nYs1lm!G%wf8X@ox4W<W@g>ucwfcuTQ<hp?S*q#ps_B(9|AJT0
z`eUs^Q+}W9Kk@0`x#K6#Ud`F-+jly8?U$d{{P)v;f7-U&|M~iRmpt>$=4?KH@bSD#
z5!n=loSvDv6aVV_C%;_izx?^l#V>x@O@9Bc|8eOx@9*;#e!Euf{OI9lFS8Y~|0VwK
zyZ`#}LjT7_mB-H8x0u)+*8lQ%zL(wFz9;{Z(~3?19JSKS{^fpq+kgJ8o9}%6_pZmS
zCiZah#oTkb`-9FGpAU*tcrClzQh09f&a=B-z58Ch)Nse0?6S&NvfakB%{P49cVE&i
zGIj6jEz6_h=YOq@ecrzHR<7AxHnX{PJ3sbY*J+vUyZ)^-yJFtcSyk7bWm;%1mH%>b
z`fZu#hvm{zjkT9A-Td{tb?>={yVIoe-h6lOJb15o;=#xN-%sw9U$(~L_u(Vbht>R-
z6whB*AuyTyyWe;ApYOllzJLDTm+Lb6&MoBvWzU;Tp84=@JokHD{qgJ_2TOl&Ro(8r
zc%yuq^~P<6Z8uKY+?w-Vn}6G~+xg%BZWVj`&i?#ayZul1SkI6a-Y@N}`tQbnp*iIN
zr7zT%txmOKzft>Jc9MZb>(&=v^}`qMy8OA{@cWy(C-I`!<P&fF`}RAo?P}P*x94sO
zR7Jmi_j|$ejT^t&_Qx++zIpe<d3B3Rstgwzwq5-1{7a_x|NqOz#y^FPjsNSMIrDE(
zT3Y>+nKS2KK6B>$%VlZlUp8xc<#~Dq%{de_W%;xzOHJDxFY0{cR&*8pHO2n&*K~nh
zuUGA7b^f|>XN%shSL-JQ{H@vlqxE`h=%*}s-J`z(K+-h}f8S~I6^cI3T)*#a{tp>G
zw#y*K4q4s(#*KUbnD72Kb-8=vrMD#;FI~^){`~0C-`L!x%g$@=c>DIPW>xh1^S?H4
zep|NRKDc=2o!oEVb<f>jesS|{&62-W_Ue!3&n|yEU%XTOL6w^6^DQ@T<pjRDqMOmX
zciEF)+m`>iBLC@5y~e6KYxQu^A9LRAalTbqD|hboqGkQnH?8@<Jhz*_afS1jU+MR<
z?Uk4Ad-MEj{rg1T%X{oTSMITj{&KGNy?;g9XXcvH)HUai-}|yX<6{1A_TBkb)%RwV
z$K)B_J$!f0|LOkAIx_dkS6;d=tNWLE{`pIlFY@MU{@eR)#>@N{GXLtmm$bdSKmYvY
zmwnUzRi5Af#pFcZ{3HL`AOD!E`-|P)*>{C=$;<OUe*K!fbpN8_?{~Xe%vZ_ny!!6n
z=eOHV_1zNPmBoMdTdmwF$;39cnK}kP{(LR@^CkBF%k$;Ww||LM`SAbw^Nly&)cUeN
ze4k_ctX5NZ<yqml;@|T>-*LY3EB!_FZs9Zh1*c9{wSW69tzXI;{CoMg(_8hX|NA9>
zTz~4x+tu@zpM3Mb`ki*~^Y4GZ*ndiOKav?4`|(%(eD-A>Ym=<I`}h8zo0$||@Z08W
z$oAx$A3p!g{CjQlpGD2PKh~d-|E~UDw(?%tbnokz{@bp*Huc<OPcij_7cRXtsht_C
z{&4B%E%WjwyDy$%zu>?2zw>ieEKu67{cXwrN&Sm7t|@+66cR9z!=ZjbyeapQ1#cPt
zny`NonAY(8`x?jF5uY~qJZ<9%{Bwr6^6oOtdwgN}{UP#S9-L2;t2Lhf=ihAaQ_R&0
zzhl<_T%XeW%I5M`X^*+@c$@OucXROkUubpL;N`W<E%(06WN+eAoBzCG!<E9*;+xo0
znuHpEsXfj9qO~_8>}0m6l$P0|Yh2q>dxN#qA~m%_euc}pZ0-KI@8_}K3Nj8UTiEIZ
zOi$iqDOdP#g-tNN^!dD=Mrqr-y!?wL=gs}~`u*9hTTRUSo%O%ez1N%fSVxlYjKq;p
zmV>cpnB(~FzxaPYd7@C+vrwZcoer0ZR1W#S5t21YX?!wySJaM%V`h^R>OcPUU-YBl
z-JONcO<50LHo01LvM~L~L?O$iJ0dRFdoi!~Pj{7&?W)h~%X#!U`dd!rrhnmQ>X;Aq
z)J-`h{A*LeGbh81dX20%7w!4%ap=3iHnwA(r!RcE$H%dA!;`;-o;?b8XQu{#bhp31
zO-pcE^poPm-pNwaO7`5DyTRky@wE-#E%;vW&)@J#X>Wz4Lzb28zll8oe#vtfA3M*U
zk#LCB?m%u`@7W2xP4B<<KI4}4JpN9%;Q-H)G&auLGww9kt8G8J`KwiOciP{Iwf}Q_
z=Nz!CdVMo+mE6X6iNd_+&ZyXY;-1vKs{YNq2dkaxJI{WSI3Cb-@LY}i%{AYPqjOVi
z+bZ)X%guXHU8eRssbyVF>9J=!_Zj@M+w^U%Wc7dhU2#9pRc0^yT5{_}cD|X`$r%s#
zd*07_*<LrzP|CE;*vztU*VZ}jcycTX)-F6|ZSdB|IQy~K&HiMg)APzAO@4ZRn!0d5
zPu+jkAD;~8*$NkDPAxurP(1YDET36Qr(H59sa(w7Tcr1W;fuWd<Mqdnv7MF4y_Iw7
zR_r_DElJUK*UY9SZn1h$X8OY1<Y}R1$d8B9=g8kplB$vUvu~FH@2$=ycjodl-~9Bw
zrk&&AgCE?|eZCbtt9-v}Y_~q#F;`|`;`YE-505oJf6imfyS;hR!gJO7^PKlBv1OF|
zukRIC;P+;$%Y_nEPmYh5LoZ*S9#c{F`Tn#!C1Mux$Kw{f|9;PU^N%|wSzpSRpL!)F
zaAL=UqrtP2cHi#L(5yR^-|D&7*&^3l;7Q}&+wb}qF9k2y>O1-D)1cH8rO2lFw<?e8
zteE}l$q6ITdaeGa-3{5C%pY3V-5=zneMmgW_Hxg=_6$K~pF%}O$26(lj6#lu_wP&J
zJ(S<LA-^zshq+L0ep_&DlU}CW`uaT^JNzE7rB==5KitdtCh3b38-K#N>z<N5p4)!N
z$@avTdG7sNJMrmKqn0~wYQN4~TKV!-b8m^{y^1GwCgyA>1p3_1<-9U7%sRB@wnN73
zFulc`-eD#dX_LBmHP)_}sL<#pvhsvU!pGhPY$mgkxNdnbdiCeB*XhiFfJ8PY_UY-V
z5ly}ckx4U^`j(tpa=^4jdH;$~qxY{KybCbQ{?^D_#GRwS`)(Ga$b?T`imO`9YidOe
zELS!g1)4wFs_?Gj$ALX>8Xmr>Hjulye!je9OwoPTrf<(vKioVLCG*+x?h&oT2Mb~)
zA1(;;uuhj0eB8Eci${rSpUtiuYnwS+t@&2X+i*5Xa&pkIZnrN5{(1aMl`no-a3HwC
zK=E)-X=Ue!&mDXIocmthxaDfQf#anYi;k$39JY))x@tP>5pQ#cm$&D1_$eDbeRXr=
z@sgh$agMtK^S)^A65;sSaIG<SU)Pt3t1ejXlbCcY?x?xwF%F;GygLLNRxOm!II&={
znO2=+W$wuvMv@O14tYqhP1HIca;?imSTo_UZ&QwB_tGtH50+}QO?EhXILhQe&sN6V
z72H=J$=%LevPSGk+OiMlB9==p^*YuYnW~(5yx@`y<MNU_R({j09QP%?l6+xibaU;5
zhig_W_PS*#VUn{Yf!Fp-))LNxvpwE!Nn1U2OUd&?A1-g6spO*Nq^Tgfu-Qk7ePM)o
za7I|;B00Bn?T!}7Q;y^%WNh`Bar(96EQ3SNJmwc_<P6w%sOWQfY@crVOE+Qi)8<-X
zbpcb6Bd4W!ycOF#u88ogG;uNGEpD5^#&B_4Z&SCTOn~p&8Q#;TG;T`qSvO<BkuygX
zOg|WNdEQcZP_g>q%bpn>laezSmgHo<_ugtPWt_WWmE?j(p++u)7fW(ZxZQ|s^z^&&
zSZrF+(;0K5iuQgtIW{|U)qJT3T(0fAl{oaIIyas<Jln-zcgeDg69l9m*|cnL;uo0F
zt5u$Kn3=_R@7mMn+D_(5*f_SOiOdo*Q8zqea^z+=&o&9kx!E(XiMgFqNN|~4a5Qzn
zVZE>iXWRrXuWZfS5YjBk<k5C1;f9NcWFylIwM8AuE{$&zqu2O6V!VE`GxURlpUJG6
zR`D>IBPmbSR;_!S%kt2eEzmD|nqRZ|feR%cjbHpfn!tT%cIAoBeJ}FbrNb7)#1^=`
z2za_S%rZboZeE|1gzq_lJI6g{^7-?NCnh-Ld9D)bekrs4=}w_bBH3-6<;F9_y-bqs
zymy(x^h2egtbMxwj17(F9iGkpbSR43*&uwslM(;ax3}kYSjsk@yHryW&-$o1yey2n
zB;D>ghw*RwNsD=8-(K+!V`F3UIb+<X$m1O>oOjJA^!c>x2}_$Uy*&HHxa+2Lhpges
zf<yePf9t;0?7UDN^KhPZbjR8`{JU1GDjUCArE?_d)a^5S^_nKKwcD0mTw1>Q_WKGJ
zwcJ_ic?VDKXnJ67vsl}&lH=OG-_N%7%*;J;RXuv=!PL`-X9d_uB<`B+Gj*z2t@Yis
z&0F4=9yuwWQ2RqS_iXPBF^jWDd17kSyL>&bG3{8x^Z0|*xyXXE^M1DGXsa&XkYRXw
z`FlUZMG4BgQ+jV~D|gr^bvv$S*FB4iGnQ>Gmhd|uSwGkGXjyT(u5a0+n>Px(Y{OL!
zPn}#H75ZK3!}`U$cLqvvb-b*W<dDgU`X*s+?03|9^VA~;i_2#Q{4z2Nv5L{Wu+iD=
zEFX__<YCQ>H&?gtY}s%vkgIT(v-qSLm%Srr6<He_GY5U#z4N*{tFl?_*=m{c?0~aA
zg46waS&tk`c-R;+JyGmYlylM4DGtw>SMRCNVvn*64(NQy_cc_1`U#gKZ>R0iow{Se
zj>+l)Vh0zqB`K(JN{UTWy78dnc)`U@I!+miZQ|Fz{LN?2e0-tm|J%=1b4+?lKG)2h
zH!t4a<9tl+zKPq@*L&Rd^R)T9WKDN@c-6m$g$|8OZ}%Jw+_+@n;+<D}^I4Zaw|H^!
z_2%l0Ggj3~DI0F{<*7P){dCutSNs0Ie#Za!{7w7A(#6eBj`29UiK}^br}j#o%lVaC
zy|T%7hRxl|&;1uS?mu}wEaQ6f_sq-N*Ez4fzxd1Kv={d>f46OXJm2W)rRo{ez7*H&
zk&B;M@a5-^O1CvNzjo{_{9M$}dZ%>L<&EMOrhPYm8r#Qac08`8cD6t3?bD%s@8-4F
z`~3+w(d#iTdH;P!n#GNvW3^RPCm!9mGx_fQ`c3ZYQ-7s{ode5W%JuW@VTo6sXJhZV
z{ob6tu_bqo^&h(Z!@k#F?B(+drr!%Tow_G4^0!mtLt?^{BWf}mUVQ#Cr6_Ud>o=d%
z`%E_d@ZFOv_cU6zww*t$@xy1QGAHFn57#@fy*d1IPFslegnMb5WoJyY*s$erB#&;q
zy=Z0ikr&rrUp_stUoYzMRl9uu>}nhN&g9*(e;($`O<(pw;K2d^`I8SgarG4j6+8Xd
zm3Ghab>!`k{wI4Uo)$l}``fz>@$Fgq8*BC!ZamW0eJ?vce8RWsTo0J5IrteGmYkR{
zQ=>)kjH$evoXzyhJUf}YCmwpx)@bsonyomy;>}iD-FtkhB}Z3&I<QxrQ7GF?@V<D%
zlg2+f=?c?#>O03dhoq=jB`Ow$%}VymJzLrL?cv><V%C2%ly|c|3_coNQvQ3R9E(I$
zp>*8C-q#m)RDVjDef4a_tT`JRP53qAb|3l|>9s4;OYgpA=j{{u($(MIo;$v1aq)uu
zmu#E*{g<__{kzxtuU-Cx&-z9t9FOJd`j}Sk@7{TRQtZwGNtV-YTQWJ-ZLaSrn!>Zw
z)K7Fd<I=OW&#rtgx>It#-?yMD=F`sWQ_pZ;5`UE*A2;uCm8H9I;M)_nrI&(d-Mb@J
zc>05!Q|qnJew~-%CRetvzw@}5JMrrqW=}Pv@*?ADy!}@B&HVjbH@?gHRlS?NTYEy5
z+%etbaY`F<)fb(v<Gk#;@})sxpLMn8=6v4hudC+Fp3I>4;-p`%-oNtpY_ru5dA{AQ
zOyR5UXRnvsWm=u7+WGR>y{FmtN*>({nOJ=KY|V=M+bn)Me|o#H*Swpz>CcJ4nJ=HN
zvyS%Y3rjOzUT*L>_v)P|rpFT7`pv3ys<ap6$h|PgJ1DUGuhXWq4@EWN7wit6Ir??K
zc*pb<)7MY*dE7ouTqNgpO>lM8)hdqUD{=xxf;SHH38$C@8cncge<*2oY1t#^phdIW
z9~&AsbefpPrXBq9%rV*f%o)F)63xtjEqm9Btt|A<GJ0g4+4xWP!Gs`@H3xGh2r#PJ
z`hB~Ru{K}3ph7%xUqv4`=Uw4~FBASg-h1su=6*42rkDuvOU2KuexCWSp3OV&e)a7V
zovIz;56Z0BCovnz70-{0S>AR$dc)yz#mJIhp(|=_FJu&4_j+_=ac%J0@R}0`<t8h+
z@6unpZRxLXMoV@-4(qPH9XahyQTA@WM=iFSyANq;XguusTvBU&(q_*t#;bh0IHyZ7
z7hm6Whg+TR%Qf>C@#-sV9eyhxD=U+fQYm{C=JcsiYu1kLt<42&-wvrpwa%UT@mA+`
z5B?Q776wlmw!OQ(@ZIbm8iEgI7T5?m-pV?1Be3^m=Zekk44DgO`At@Pk#^+Fg7EXP
zb?+t5AHAP&*CG1!(cev`-sO1*FXpY;dQ|qUy0Jxo?&o{FnU{_>OO{6|rT5-R-`6OR
z<tnttf8*!B$4`Ivali0JZDQxNri;ek>^c)yBzCYZ-y9+EQoO{Pvwv~^F0-ds@BZP>
z`r?1Qx<pvPX7ksx;itE6=III5?CRY&+q{n<&2#3KtKD|nvgd8+nA~@5$Ab&^T3%m$
z-K@*g%Rl!p?>hdtuV43xi65;N`>oYkS=c*$@k+z+J9`^h9lIBCG4kCEu3!#g``zN-
zA$;W0*V^5Ji?2?;wnQrK)osBKUNf$3(%kB`z`(<{ka3B~R-L1H$2i>jixO%LTTi|*
zKNe#+XX}RxTUr)ov{W8bU81IBXfR3c)a3q4H{0EQ7V|8eDu18rYgDU>(RZ6`ON{+>
zY%6Z@w64|Ezsn^ayj8noO6yn6+2Q?Z=U%RPA|_?_RJ!K+vR5@c;bPzAU)3=E+N3{?
zfA{mr^PBk(oLSSh*7!SbqCDHbxO=e?`J3k#-`-~`X7*rqi`|C2W0v06ucgWT?0#z=
zxHbFWg6r94f;?v?l$1N!e3O~%H_yuM9@GDhbK9T3P`Lf#+&-HfA1_b&d-1@_7hh{7
zZZF`I5qo~|Tg~#_?|kQ+FS7V{OXo<r<+rt`7f)z=_)<x?=7&k`?_l#u&!%h+TwJc9
z|Eko?;`Thdt!vwt>G#|)o^In-pcxhO$f8eGnIWy|)wCIpm>oVm4SMFeSy$z<+IA}&
zjlc(54ey`*+`ac(9naH8?%!X(O*1^c-+Hltm7Q<lt1r=R*6;qzzy9*P^YSINd-lxi
zR=;tT)h*}Enz9)i#lphE<?b%t?B2h(I-hsGd~n+wcd=bxZSTA7>HHCqc79rG$0ON4
zORPU!w|bVW%=^ynsN)m%4%rZW)h(BcwO5HVCls<XD+UR^njN^VsddpLrHX$}U6Mj)
zr)_=5yY|-~k4Z;Y&n{L7bMy45etWni_{o<C(wc{+^>?_-+4Ju|f8z5;m5a+an=a#@
zIsg6#xwCiQNJKG923FtDk<EMbY^UAqb1u`|r^R#5`56~-j`@sk+<U$_`=+B?-@lu!
z((&zhgp-5sselRg^X~Js@TvHmW$X-^!f<rSmT4=po_^W3b62H#-o08Kz2kddJbC}C
zJ>l6h&*^IOw8g``We$~;K0D#OvCU_0P~Ja<w+&?%8O>VxHCUJyTyYi9>YBHqlVL^f
z?#3huAE96e2Zzjti%U-^J2)gboN#1iRH#V_xio#{voF)niFu}7bTv-8FlFLYhR?4Y
z%aSHGoLCW&d8TVaq}b++U|)@s8#JbQdKqjBT5`*}bn0Es*j=&Ft1?5vi!^gf+k#_@
z12#;$8I=*$Te`EzH!Amn((2X8QcG?ZM;B;%Tzl)g^{iH4xK&oF%*CxGT_&@d1N_e2
zSe%?aZJku_f{h_LIUZ7`XNyjmwVLMM%3iwlY@TM&qTA`&yB<!8^;o@bX~wpDxzp}$
zO<mF{wcu=2)Y)kP9g8HzE;sY6o*|*_*t=xsR+DGt(RqrqvP-WAC3x*@OH`Vv5Hy9)
zLn&~bYp-U=o9o%CGsQ0JEC^4NQe#n5VhK;>Vws)mwpCg)Eu+~qFZ9)=NqddimYHo_
zb8TJdlUq6`mrTmOI_qAy&yuYnHv(Fm9UOEuRGgbML#~*LX5QYFby>AzwqZE8%iFn=
zGc~U1Zrr+STj;&Jcgv<*m$cn4FHAnS<6i73_Q^XnE~o3QDSaN7xzR9mg~_c}g@9D0
z-j0-}Cf*hgiPIDKHl69nQ5SNJD`1#ow@Pk<@aByUhT9l^*Z$4rKG~;gFw46&R;4wC
zY3JJmo*($uW^?{$J#@BBBl)ePOI9DlR9m@k-nN1}5}O-RxWZibZRhE#4K8<a`)rfH
zE!S<Id6>NEAsLyH*^YPOxY(<h9<m(t6;Dhk6jtqxEZr?MVaY;HX@diqj!8-UtP6Er
zgpCa3JR)0687HpI_BtYzu;a~JXG@ODTMtZZU3)wu;<lI3fi2R7D_&SwEozmN^psgF
zBHh!(p(fd~%tMmRVr@3harKBJj>}$_`-VroSJ*yry;kIwPk~o&?Ow9f<Zj-JRlD9z
z-SsJvJ;nFdBQJvmZmgxN9&0LZ*ydvP;O*9_LK+_KlClZMA~S8??mqYK^qkMdGyToW
zONu@gOYGTcx$|DddVaw@{^sF)`K*r)%uU<!#7uEPVselKFKb&`xb^n#ohA3OUQG+S
zdRHx9g(L8$i$}*~E-hY*%Mu5t8mX%;HMynQ<RoDr6YOQO@Zye1Z}s)=`Rp##7caV=
za_(cL>aNVIcde%0?%rNH)#bplcJ;uuDJCadXXNrRWF2;ES6z~{vg9&brzTJ6Rgohm
z2@$E=dK!HhoKg&1cIiD;)!uqa+iK74(qzqTX(v`mN^X6gx5R8&rp{VjowKLQf@5Rm
zY`C+?cUJDQ+@(PgXO4I&uFRNdw#4JI+TvVZk+sDU7qe$w*L7QaIn89|&v$pzmZmJ~
zb`|4vWy*3iI>_dDYO?Oct6nocy(~N8yYADcPf49(j$0YUPUa*XeYA7E^|tN3QIA^1
zywbLKM4EX{)iUPc=*`JGnXPGbvRibKWVFi4NX->Fy(-hDHD?`J>2qRfhm*?8@Tf~0
z*BqBw)Kqv~d$EDB$+6TY)<-L!N4mJ0RUTcJ+Z*`gwb18vEzdprt?tD>J!!+MGHa>M
zMo*q2zG+;uR5Zj?l-6!cjw`(?n6GmvUuu{1Ns~hoK9QEH0ha=#&e$@AXS=jrI+olx
z!zXyw#!OxhvBYw}i#v9&db-^v>ZQ$Fw`X^*bG=^oWY)8*(z(Skx5~N9Zu3V7U3BY^
zY-wYiB+zT|BwvC@LZ|3HZ`P(2cdqpcDzY5UJvDpH7EhzQ<-(i`5<+w$dK1MhRIAEW
z=arX-KA)Q$!7}xEcCVS27hlXJ&AiZ?r?*62k9An^>Fq8aFUjEI^Lod6cdu@loVxVx
z+ABV~hK!2A!84b$P4@|vlngpD<+|t9>sCAO%q@y|wnudL*_KdI(Gv&a7*b4H1ZK4w
zrwg9cFplUJa6HD4^qP0ApT;4MpPzQH9n<zYVkD}nz$T!)nZ-oKM6p$1m0_~7sHI8f
z#op8DQL2YDjm24RoXu!v^0#p*S-G-AxMlV1#I~tQ$wCY|DiakqNarN?KV5pr=9tH=
zf>w_krG|!L6ECt{in@^Uq_Ji35k8MjuXPS7%#L1?OuQ08Pr7?D&v1Deg<MwN^}Os#
zj@zX#E7#oe=*xE8^=Z;$nRb;E7u2S)9CvB%Dhy~4J*gVY@W#Pf%Ew7euq~3comr?M
zE}%P*;X;tYr4=SyrmS#TtCqTEZ@603)@vUp25wktTG_GlR_w{e)7E7@6%24lPGB~c
zkWt79kI1;@xw7QMRX@iSJPAV1VGIU49XbtGBr}>xC9TklRCQ>(tklS|+U1O)L4<Q)
zr>e3iBhTSAM|4$}o#oA&wVQ9(vrQq<+Dot9-FkHCX~WX#rN?eAO;L54*l~z)vvQcu
z3!Q{!4reWGmPIvl|2?nX=XOI{I-&f~?&5a7AnpjZckfJ#E=_oNW}3mFFGfpz*j<F8
z4c0pgGPQm<5H8cXbyH^H%pe}Os}UD9yS*)b{QrJ+LwUo2_c1+bZ(3cHp59Hop1H5Y
z{mQqdJtnhExHH;T`F?lq(wOGQ>-qLUgMfyb+mir+?(N(%eG4XSotkphcbeL?EvK?I
z*T?OM)IIPzws@Cn>8+MC*}G%3XBT^Ao2{O<PK?9GV3y)4=g`H@QXi#svsNq?@8UU|
z(YsevHOuSSt#VbiW7l-IURkp0u9TXOo0F7=`pIMums5w%wa!$^JXtuY>5SnLgIm6t
zS2B7sKKo2uaoxJ~y2y>I=icmQmNpZcoU*gzl63KsnWATMR{LhJzFry^`n;gwgv`;6
zZEUG$SQ=CdGUd7#<?cD{>$`i`>SbG{%vP_~c3CH6-77gYC|9CuV{Y%3aNSc+4|$#3
z>ULFgsm{!0UKwYm`Mk=VaawR`#@P*%Ti!_V&Q)VEad~YsX;y^Bs@}wd>$I4&(`$T$
zBAT?COs<;k71qgUc1yXX*d35*$g^$9#fmddSz^AOD;H*^PJLAxo8>-v^~$X;O`pf^
zIB-ejq%TKyNZ2LS=wwk(!R*ZjMmow82Xi+b+AOZBdFjfmt!uc|&YUo0Ip*x@@z_$$
z%hu|tXjSaVxLvO&`ONm2=yD<HQpg&;>AZRGQibwXnQPqInO%DJzu)$>YoFI;y0Rwu
z1g30Ey0|P#Yv*=f&5aE+V|k0cwTy$z&7D%*dYyWOgLge!79M+L%Z1AZLB3kvo*KT%
z;ig~KFKyW?*zt;gTgbiy#a81Z_HuHud{IV@X)9t|Te<WvW<Rr4&Sy?N@o;6J$HC7W
z>4tZ5H21yc`!?AjJ6I@B<yD*VDTiEHM|ZYo9ZM<`W_Jpo^5b#T%#!(;T{WFoRCvB|
z@5!~iFLoD{9dl7SwlTC^^w^4JOP57W&bs<WHO%SSq$7b@LN{kE>7C@8<a6MvQrOWg
z)3`d0x<=em4vCyl_VKz-#*=QAfK8#!0%x}!JJwjT%r7P6ZEE$GkMH%F1Yi6(ki%Ve
z>Qq!y<nIN%4KrTe4fA5I+rDkL?c}I)A!oe{^MjYy=Wpa(uF~g}x+`o8qxRf8><yDN
z3VWDB-EJDOB^FdI^4hfUct<zO(W&k8%)A#CJmKkVnwdIraxJ&07`K6uMvAz|3Ta2t
zb*Jt`w6y);TyitHeUs|8Ba;t3Y;s>N{ZviULDDBT+JYgMUB|IFy6|~nd3bEW&LVxz
zOUq`7wX`&<u1E_x*rd|3RJ}y8<cOo7CbvjcR-3g@ritUi%PfxtLWD!LM6Rq{xolJJ
zleeo&UzM{*3O5xL6ugLVk~3Q@#jIdt;yF<$;bW4c(j-f^XAKsb%`%mRk8bD9pK^3E
zi-_YXFD|~G*w{I#%yFmIlm&bV=zBfc?u>fl)tGllpQO4EPIBhGeCx%WzV%Fu_nSV*
z@#zb&9cfiDa8c^f>7A^q6wG$w(ku;Ksl_5bK^HZ9C!O55CRAg|(rg|xt&@vWCYqTY
zT-3WbTf}H_*qJL+JcF`kF7lZb&@Ggs>dLvXE9g+xG-FQRmdQSh3@2rq%09oC_&n#u
zx$4{H8xA}am-?W|u|#m`;#JENQcoU|I{56(tQ+k`X+6K#b2tn87y^&T@QH>vaWpPD
zUlymNeCkK7+k!o{yFTmim_)5<<m4#-)_!{H<Iu*>FE|!^@k>qE?3uZT@u1{)4NcGD
z?GkOLyPhrx+}F!wWR&Jxb+(x2xS6M6)w1>S>7mVqZPH;8yfHElxsF|tX=XhA`h6|K
zM+beU4pwdFwl=+uZ_m~1<yn`#Uw8Y+i+fxJOC+C(D_3pc$nk2EZtuGOGJ9U0RQ}1e
z*>8@wIjx;@e|9*B-i7?<cZ<K<zB|6_1QR>E>_iWjs+rqbmXtMTINZFJ(h_vxecI=S
zUpFK<9bHAXTza2#L?|W0MC4uWhS?s~9YGU~SSkgcOQf(W2Y6kQ-q7PD>d<vi>#786
z*2ZwgV~!3j!a;1IjBdI;8@XqyZArP5wngHhO-|pLY%`OCs@e{|4sDkgOejlZU68<T
zc2w4_hle>pJDTr8#KAPjZAUgPvdQT4+&Xolx(s7`xzmHvC+01MF*f=Fht;;8o2#)>
z?vjzDP<KSe^pus||B|LzFVXB<W*on2*1NoK+xMKTxcjWI{8!!XGj&VWulw-k<B$1n
zqF-)Zy{~ZFS<a$(Tl4PF+ag6V+jnR0`!AQ@dhcz!ep9Wt`HO82D%D(LvirkcRNVIs
zPkJlv&8l9!dVMADrjjIOH;Y4C{H`BfC*}M08f%)^^OC5SQ}$RGZZW&@t+VKngmEPw
zJDc3Uvh006;_kvsA2qIg{Ql_3E!*`@S7!Sf+_mYO*?)mUK}r2dXh0yRr-*0>$5c@P
z4aX2oFUO43kSV#68r#yAtjV11cIDLKoQzvX#dO_PY~0$r&9o<5tXs@fYKcY^)67{S
z8j7JV?Ei!qf9+sk{dcAR`~Q#YYyZE!;u7LKm({>S!&oO!Xn)Sttz|cTIL!58>SAOM
z1Q%#7T)1$-0#i*F@BH9xcNla+LPC?iAIqE2$Po8^{@)Z2AqL(|KhfyxZRK;{zAHEC
z3RoqiB{XTJ%d9n1rE_L`uH+5enyi|s(HS~nYKBMGlBDqD6lFJ4i8VRNxy?ytreceg
zy?UliIFr@1WOJ@?Fw<EPx3nuhD|!>#Hs&VlY`c^doED`q>9%LD=4NS=Et{5hX1a=L
zMQ+y4xHM%kr?1D7#Ufg6Cb!*`15^U1crLrGnjC22mZ@`QS(@Q2sZ@n44a;~9W^9x=
zYIe$KT4rvqsL&<eWC4jY*CuCd+@h+uY}N@iwZN6ZQ3lIAwj2``?C}(wER>m&sitz`
z$m~p^K*N?^o&#yK)U>!&W(qhnsazCPnap)ca;k)nw8Sw@-&UO|Q`9a@<=Mo^GRaF-
zB5e}OWUmOf&V*|Rj<qaHNmaYCXj4dFq)wF1v9mfWy}52DXDj8b(DIqcld;r9GbGq|
z$;zO!UXwFZHBFA3U2rsHW@K`ps7Kf&jboS8CUXidj<}}OrIXB>aWN=U(_1i5)N^s<
zmMybZWNf*mYSz`7tYf6=nKA9iHO+0AB3UNO7I~~PnH>_ev~}9Xt&@GaJr`XJT)X5}
zYT%M9Q$nX)a$V{>EpyhXWmC5FZn~DX*61J;!=!*P1<lY!4NHYKOQrT|W@-gTnf0c5
zn`$+wZOqn8*|y?#l%j|4nyg;ekX0$GL)EU$TsmtJ?}`+!v#BdhuAH0|?U89@v|*-E
z&myxVN#4mB8y%BPa?(t+GN)bhU9nbd&9p^VCMzuR+{$((%%xl6GN<XHYhGrPJ=`{i
z9M#P*%N1Qa>z2=^HHXej_H@yhF)bh^aMps0TUEVsb-ks0b5`kgL~>2J7|gZkQnb-V
zscn&4b3CP1+*%{L+395FvRhhaw?%SisckjRxTX{BnbB!9?W9lVCf}9b!69xEEKAZN
zCrk=hnkmuk8tif9SkyJ`z?Gd_jZAW*g_h>rPENSmwPbOm8!zY9mMgbZUEFSZtev!I
z(amYcjG`nwIkQZ1C%c_=nd)YeyRCOpui8viot`DL)0W*b$@KCx+~VTub6a9#R8&IZ
znu!@k$%Z;7O&2Y>679NZO|H(ZfNPvuT)jbAx{_0s&9tR1ig_nTY>8a9I564dLR+&!
z#?fw{E6WU`G^3R+1w?6^m@T?#HsQ>%ZM?xo8Ea;mbvftUNIRu+Y@u<anJ%mAVo{gV
z&KirHT3jYDYjg=MO<C!=E$XHh*Ji^_%Vv5>u9=Z7b=p*mLp3eLb7GR>%7CSs6IW>}
zUCJ^E+cYgrIcl-9RA6$@v|DCJGcK-HU9g(1OSjooP*pW@$;Mn$&7e-#NjlR~LMMBq
zoVl`1bJ=CDQ%)<oJ(nikIGwX;ipiN&uCS@1LbIb(R!W^U^DuPFIx8&}DxJ1vsVh%#
z%F?MmM?FJAva}AWMlHE@Et<1gI^~Llz(rS8*RbF$Gf|bL*(OoGOVbuzGg%U?8W=h2
zmdKfFGc^O2%m{1A2;pLlnC-S!YOA)-$yJ?7(y1G_oVal+a^}R<nq5lELLE1{oSl|!
zqOo+A$823`pPN$-n@;2ESu(X}R<F0Ikk6Tw87bR5H+za~x)hn|tt&BQ+LepbvOK-K
z3{|>9)vh_M^mbL<)|R<ttB$AE%#E(if}YM=rfydjtE|k}964pH)FrpYu03aR%_atD
z8qDelRa~5=x#-5R1i{8mN3Y(nS+`ZAvksXB+%k*G$(BwQoXnQ&ofhaB<iT_3wrNs^
zj<?tC;INZk(>hmmthkvRxKPzMVs*}$#nFMQj&9l#>gwX^8I^W3TC0~k*~fQM(BU9M
zsbJnL;b>p2%*L56T_-g;SDg0L$++T@vUYOjOqG>coy(3z<(g$JxfZ%qD_|+d60?aS
zno_5tCi_mA!ZKmfq@K>SsH{+x*4bLAlN_a$bG2GjHA9lUBo&=zMLBj%X>e;4X`C`a
zL)4|wVPe3PgiuXKj#)~A8mUXBaHs__L<Y7^6LmY*)$5ui)Tw25BBXQEwYEzqT}4j2
z9NltlcEsh7MHj_74TM%{99^k8i)YcbuxqO$ZhFm}nVr_lyKQD@=2o%EA*YjkRW5rg
z^GKO2I~jC2RLH>1>*A80O>U`X7dK5vb@UXPZOq}TYvPra)*LwDNKWL6(=JOhIksgh
ztqIGLGUZlV*6o{_rLo29%t}GFV}}G3loWkTReB6PHZq;uHf2^+mbWsOq)En-#Tv7O
zO%~*I&DK^H^3`11?Q*HdR5k5#P!x-|w$5o&wce%D+L}H}v(2WenhLcoj$FBD=FCZx
zA{I^B6msL{#LYgJjMRKZrUe{bIOSB&)|RENs>`N%^+ahk9S&BTEM;&es@F7Z&9a<`
zMM}A|U6oc{&Dq$sIVak*M>04~amh*7-Zfd14sM8$2yv5~F)bkHv~$RcEw^T+dQD>r
zx@zp1maXh9Hkr%l(oId@O|zWSCIwt#&Rn)R=+dQ|S-FO$8Iyx1Ejy_dsTA04D7EC;
zqQ$1uQdOs2o2fZXXr|Y-Yi$}!RC#iBx6W8O+pw|OQE*w-iLFc1mijIW@tt)|ZQ;sH
zPnDDdDoZA~WjU)#u}u*PRGXR+6&jgwanVV)r6LC|2YL1yhfH0SA{;o&)gfD%bJ1ka
zi6$wRJeMpJ*yfdTcImX%m8laqu2`z-JImW_i_h9^*OHozPWGsIdM25LT=F@w=)}d8
z*$LjFTB$55TXN1U?Gc$~X4dN|YPxbsa<<FDRa?_;MJ(HFsyXGN(o{E%$Sub*x7=Kr
zv8dBqOX5nJ<WhztHJ6rImu?2CvMTkM<(f|2x^1RzZ|e-EiN>7GIw{w3R;cw(5jI%T
zb4YczYAVCgiM+FdPGvS6NYye5oD$5fbLE7~DNd1D9*U|n0=+|SE;MM`?5yH;>ssfE
zo0|i>SvXw{Q<bKrd2aUB==N&4m7V5r<<d+plY?zRCo&C}a<XN)G&%a*4!g=^$SO3;
zC)L-aLvpiNui{cukDOVF+R7>uHyfNtQx;T=^mV)F?3I$K>=L+5<4WX2DV-@|QcPV^
zPP14tL^${UPh<Adc*3=SskB*Xxrd;|<VI85Nz&0wOmnoaJ?gmi$l`G!tBBu*jW_rd
zojaK;gFAxPIa)nrp5o)`A}p57xU5?2<fI*sJCr!5d8}CI?ZmZcc_ar%L+6qHSZ5=l
zs5E7Xg&YpOq4(afZgDYi=Fn^j7HQ;ExFXR}eJ@iy>H7{Q27?(q$%aOgJf##lQl!)b
z;=Z5S8|0uG=(AAl(vD~zqlXQv`gmNY->-|&7jZk&v|O@4wJVE-V~1f?N9I`tL7%R~
z6rYJYT*k&-doQ$>98g?eeP)4|gk$dIdA~a}+SPa;mP$R&GrueP#5Krk)uQ!12~Ip}
z5h=|3411F0jl7kdjEs#84y-#^#+7+CH0isCW1tuVch6a#36T?XMKZ%y3H@2dU?SPZ
zFwID^C)s;%b_+wWf}!gA>WjbFoWu?;DrC#8ejCCmazNm~*T5hZp_!=$C#P)C4sm(b
z!r;lpz#7SMa1YbVb!!@xY;~r-i_#M9*E-YPP~&7iKQCeBi6sxKiUJB%s@L2+mL2`T
zQ2dCDz44EB;pxw%*FUTi{C)n>jq2oC-u9f|D(BDWo5$|AbgjbA*%2uh%J)qDd3oKx
zwy!()Z$A^0`{rrQa^ugtrPllWIoJE?sEL)YdUE{H))>RnJF_bj=jY2cy1wdC>0;?n
z>Ee8o!u3~ob=3DIvL~KBvk{(h>hq6zZTF8anaZ|yYQDy%P3$S<r3)9#uHAKBo_BM<
z%>A?FX=VE#SN<>AzqqmI>8IMVvMGDt&HnO!$6o%)Y5UU8J~e!Mejfj&UAf<0RqggE
zk@;HtT`pv)eZb!%H#W-jPjk^T$c}Q`Wx96jySF!F?wai~mEy1Q+m_Qid&@b&<Mn%1
zs~6pzJ>lV$XBHhfE#J0mOkFXf?AFuHb9ZtN^93^){@CRx#o6Nb*uZ$hi)=@YlQ!o9
zSIhjnW|37KT@e+%^6|EJB8Sb{!awg3_@$S0d$peT;`ghiJ+xC7E>VmwlkGd4VK!ZL
zW@pO$2U#mJpB~iM&b4e)dhv1LNoO@P{w>sAEY{8@_2KR1n{&!Hw)5|Fzj&!f!18R7
z--*pPPe*J&{Vvz_MppS(?TG^B9gRhqd(A6f<&^4HKQ5d%<42sFcyrFnCwnb_<^4`L
zoZWTQ_4C>Ja@pJc=hyB?K3?qVrM1&|i}B>llxw-U*@4OD!)BK2o_K!gbKA+!IlCsk
z+4gdw*n?ZzM#~j5Z@ygfXJ-;av0TWvb-b@1?eV|(r^Y~cW3F!hQ9I9v4wsUjoV;~k
zwWsWO(MEahgL{1xH?LiO#O9c}!MD?;9O@Ap_ip>Jr~Fs!M!(gQSI^j;viI6%TQ^xA
z>u2ZLs%GhI+|YS!fq>p~`{3{^=jOT`DSf(IZu*1BaOaIr0@r-2=IyO4OjWq5^{w{C
zgSM-l7jkbsE}33)a)}<hjP+*Lz27#-XuG<sjSikG?Jo2<=UC}emwhqTQ5Rot+vfV^
z>S=>Rr@7}g9XZIeqvveWoHq0F)N_@oOLh6oDqh?QiYm3JaF&UDvh~fI$##+R4&<Gi
z9d5__^>xh5hKrA7cDl{?p03OLN_y_h$GgL<G}>(R_<fF7Zwm_DqWsAG-JD5#c5gW&
zbo@x6@`~NNieCJhd1RO6w({bnySD2potLkBGxMYHbc=*#u7|Ii#Yng3#rCDJ-4!T%
zoLed1Jx427%}jA}h_B4s)skMTgC_jkclyYOJD>V{ZyuWaS#(}AZ~n}L%WTQ%?7|hg
z2Xt1>%AK!NQxmH15ZzpKVdJh3Nha>^D}V0RTl4nBb${cTU7W|}IhAgoA8~B^)lY#t
z>m=9vca_V=E?HZ_^>~)Fk@XFk?cC}rinHrvuUJkK=K0yf%kjJI(m4|=(+9URw`*Q+
zit?X&{NVbsTgp)j9v{p<a4T}AF|XES{_~D^Ph0=vskV6&WA%H_*X=VNab+yj$c$*1
ztD88(*~oo+wUg6+o;w@%Bp=#t9qw)RAulA9-z+>y=hdB<BXcC`#AN&4`z(?;c{eFi
zW}9s0^lQK54CgI=bxU{4LAHCV3e^%G9Ntto^>t;V&YWo#8`^GmR4W};`*hs>Cfk4Y
z{>I#U+n!F9){Q>8W51Ge&$X6EYaHB#6#JYxZ|J?e_oMh2_b-!Ai!ZL9EnT3gbb5ZA
zg4Ol4i+ulDNnPE%YRi%s(*(ta!zGf_dKi6zyqQevr%paS!&0lf%3`^z<jY+3-T!^Q
z?|f9deW8`tCX3lK#ocu7XEko*z9b>t8yv(niFe^^4e_m48kv@IC39|%b~qHxy7@du
z;Prh`0mmas-)7D1aI@|7C|3>=y^?aR(C1@`h~>Bb_O&N3CtF`u^!xd)`cvt#gWU=4
z?du=b{440QPn}ouDWgBg@7T6}u1enTXJx&@w_Ilvsj+zX?$z^|kJFBBDo&I9)2_<J
zy~O6D^x<#MtR{AQw)VMhn)&rH$J?`i4{lq&ns2jAd%n6ZU&+?lYRC5<HWj}kbv<of
z+-cdDtm}{dOlmuNtb9W{@6Mv_sYg?HCa1l7bn4wg-&ybWe)3aXan3EKaBs_G{yy$!
zJ9F>n<zN3h^=<E7StIG?=M^@-eD(XM%#CU9n~u(XENs@e?aQ?rY5H?jYdp?&Snr5*
zUemILIp6H)WUio3ikFw3-P$2_Vdnj`wNJ~#y;dHXmRptmR54Sr)lI$jg;{dYI_YkU
zH7Cx#%+5E<RX2<{`0S$P+SDDhlV)Zgo4IMa?$jO0x%ZvU8S0oeo|<01>4=Qmj67c5
zmRH;?GtWqUvgql`$k{KuHH$yEZ-?pEhtik2rWr3STD!Sd*>de#sh9_5wr{RmAAf9q
z=lMlb+r#gpCtpcibLzObW}(!Ccgcw=yT6;Rk8uq>@L*%&Y?dPH_m&2w7PEAvXYVoG
z`Yv@tdywOX9rtT4`q^sC{wYw`-aYS*v2VHTW2+}Y2^Qb7`h)r^3)4Cd*u>uR=)8RI
zSIzpHySEnHJ1#v(z0c<To#!0VbEk8p2c8q1_IXCP!I^XKc+ZO8kNYQ8w0YyR`Le$|
zFX+#gpLVSB@~v3e%~cD}SKUhc{k8Y!?UFL{x7#cgKRuG0>~y#~T*gd(R`xO5xqE&u
z>zH>yeLsusj(JbMgrEN@-jo-&>!4kPsz9B`FDY&BYa89o_s=)Gc&awx)-A!OAExXx
z()B)b@Vc(Ob<g{>T-$gtgH?8B)#|er+*T8oTlMYN&Vu8sch5FE#J1UX*5k=`!nY5b
zZ`5R*w8TrN?c2-CZ@*R>m++s9?~9!GJ@1XbaQ!OwYwt{>65b_;e`a3GBcT;6xl@IS
zQDJ)R=Chi|4&;R#l2QAgCh|(iu;EYZg1b3>Dt9J-+8-s~qwqIf?B9zhD<dtHgvO#^
zk4*u}E26*Wzb-IbSZBe1jmu@l-m~nHf3ugSHdZJKHZ^E+ae71*`uk0}my_@=O@K{g
zpP)nAl9KzAICeebFS^dNO;=^p{+r?l9IYxQR@uhneRv&_m-jij!us{|e3NesY4#;K
z%&Yv~>nwZw^^nX@&O${-4z;4WRcGcmR~0>0pHQ#wIOV;5QTb*|J)5LWhA+$>munm=
zR!=qWF4(#0hI^WR#g7l0i`MVH<Y6LnVy(hP1qPo7j+%Sejw!tf&KE0rQ#ax5p+f)5
z-0cPTzUJp?ZsGg9S?J6O>6hP~>lHpdnV@=j!f#WZkGkcrX1!TacEfgC@+%Y5maX@d
z41PPU6}hsk)Zw1=^`K*di*y&NKDuI{%F;V4H!)vq+ni{(HrY)dWRlqopO|UfTs7N9
zd$Zx=HC&o{U(2(;vF?4LkkYuwx?EM?o9(r3QgqU^6Ojvh#6Gy+{h23ra@LEmhpCoU
zpB_6_PTT5#`-y&4(xOQ_t`x0t-F{>HoXnM0CHJpuroCRLY4fe_$Va7%ryb`MYwFHB
zc&sxtw%+va>Z#9s?`S=rvgLYLcaOs2f(SilH=WtXmOfn_9eBlfuE|{0PcD^;b-RUB
zzBm@22)&l1JJqn`NR2>1$caO@b5>o<-fyw2?HMb}+fzRq+hRn{etLU%Z(QeHP3`;o
zcke!FjK3Uwbz0g@IlrvKJDCM-3nDhY%+@I|3A^=%JN2~6r`(yArmAaJWav%Qn{;KP
z-@9c`Q`;33*>u)_n4i`Vw1tDW?yQ!573a6q5Lbn&RZ8=Vzh!BDTd2}d`})rQcC(!O
z6Bf-clHblU!Cu|?+Z6ZZyXp=`zgMhSn!uLP7OS{ARA86l9})ME=ttZQoGm=2OT}()
z-1T~uVvoDT{x*i65BMHGdmt~?%4M+mD$@*!L+$4j@9>^)>MOE4_DW*%sTP-ml_A23
zmppIIaB~sRNfDi7lys0up?$JNj><<}siIdY>*n0Pd3U$AndZZ_S8sft`+WB5ok`0s
z`t3I5IPocO+1}c@rR(0CP0yL6VrVks;ld0{N#B-hS4_k?(wGw$ExPt4C)(%6rrAN!
zQ%~Gjuqo{2+Ox51!&gR}4SIL>wsn!1*0EjLQ>JnBEX;cwrn_c##?48g-3iwg>WWUf
zxW#8-Mw;a1hF(5juC|VjC}*_}vs<PTYPsFp-kn@?dDgA#TF(~eb}U*wJ9d`S&aJ0*
zyj!Ka-Bj;&U!=l?3kGL;do5N4_FC#>ZqWDHar_LQ>+v45(q%y`SJo|%c&+o?sPvk~
zE$*V%yDrMyxo)mEN!l}IdAXTHL(B~(Locnh315qNE?nKF*>c;Xz<ccmo=w{iTs^})
zQQ^WBpNng~%{Pay+P23mtn19hNsCslT$#2?a@(=3H(y>nlUU3<x3hFxyN<D6$D}Nc
zSxb#_)Q;qC580JDeeJfd|JG0YB9yc&Fk{Osr7dg+ybM&5jSK`uHaU9;8g4Pzu$DiG
zWg?5zAs2ywj1*SKgmn(J3j>%N^$zz~KHHKVrX+l!c@|%1(*wcc!Y0WB9wyFw6Qtj6
z+bZZDyy2~tyR$Zf*W_c2dzl4SoGEHkn$Xc_uAEXfTkMKx)vwZ2lNGblCYvld=yp3Z
z+b44B5|6YqH#IIUyXfh*#Nm|3qA44VCT;M_T^5p)*{i##J4-6<n#^wjB_~C9S$&nB
zjOOWa{L&BdJU7U?Np-edie`Ja<D3N_%Ouaq!QF)tA7}ZTwd`(9+_ybQWSQ=QtR)lV
zZ5&n@SqLVaJ9Icz(1$H?$BlE1f<g^5b|e|3v1zRGU7M-G!*L;#O^}hfN~&H&AxGvG
zb0N1o!__CclWtlCe%u)vF>PL4nBf{}BcFq1iAEEcd?Meq<R~0xGF-CcwblbgC+#x^
z8mcBTJDzXcY~9?Vx}|N(1(m(tn#$aZ%L*^Z>@1ZITl;G2v};F2rx-4}7?iWkO)|+O
z+vLKcL$hXF;tDw7qMV$wLUYNotTeCC+fCQCZoax8J82th=23p3C?ya53a&6-_Q=A+
znp0;wmI^aW^|@elt9Sirr!D9DBlzz(HSM+e;`&D*<QdBoF)oGm_kP#ct!P^O?Njal
zyc1fQKaA6VPTABX#mK?Yqrjl}@=`#Euc*BDE2+{=U5)o#zb5eKDmPtZ-}YUAMZl4f
zLBN5*L6Al6^bTgm7Y`VC*Jf?7H5FvM^!M`H+Dj_tCP88P4{dMBEqms)_UWhI*yUe-
ztTC>=uNBX2++_K`;n2bxhh<VWT70?k#k=<^pXRG$t{+*PCdM6cv?=iOU{_)K@kxM*
ziGkC3BHKdE4Mh`$43u|qY93tl@QkZl*iVhtKbj><ho0~qTOxL2!%yYM->sFCOIj4f
z&)!`pU9tG{&qb$yY`cDMQDxQ3%{vM%7xSN8w$E6%*W`|?mF<H?C4Oy8QFqN_FJEeC
zZx^V$vEq}0;>$Vix(C!xc&N>_D#_UI#CgBui7$t`uc-CK;tf6%mk6E}VBN&f;4ZIx
zBjcXl{L9x5K0p8S&1qMI>RGDuYvzR2`RMTlM%@XVTBNBZ8r5;7<kZ7ijGWqfdlaTV
z+rTnsS>}WTUp05+UfO70Xq=o=V5>FZ`kUyZ7d|s6s3(2=%CD-ek$-E+_SfINW4reE
z%IwSHFRN&2y1xHaY<Ya-go!Fwwz|yR>T8%2*1I|@P{UD!VMa(p3**EIp-o}gDwl4#
zYKEmInOxa2JJ{@4*ZqQ3aZfcirWiGDYf^3CW+<pSzDMMdO6MWl37_+AmK(`PMDoNO
zFY39nR`O=YkH2}x=d*JK@3ap1X4^YOqMEN~C%^vMZTah}|73rPe%@_S`t>3||HS^V
zUuAZdayFK?Zaw`mt(!N!cF{4zx5rQNv-SHNF`xVM_13waDQ7=TKRfNp+~C@I>l4>s
zJ-c;wLakJ1x8mH%vR5o$-MO`C&di8(oyj$q&Gi?2Z8qKgSL4y8qr$l!SJ(1c%e!t)
zmo%;{u$-A|+!I}AmAYg9#jNk^w<J#9^XXD$@asJ{4GyyzCndKnzjC(H{?hrL_4}T(
zM;u;zOy@-7{pqFaj;Ae)nfK_GyV=$`db?h|YBPNyoA1B)2z!2!r;^D+6)BU9UGwg3
zyc1U{%W=EhL*1%y#<WF~ghYz0z1}K@nLG3-{eAu8Q*w>%yIEdQ*W3dpvUTVc@OL*X
z=>0gMT#-p<cHeaF_l7#Nq%GZp$}B>@EwEG5U;AKv^AQDc{tO4R0}E^>eqv$1#;;qP
z_UYMvgT6lkckP2OTDz^)whoVdy8Y(6zkk0T?a9~u{C(fv%(b=a7F_AxFWla?$HvCS
zMxDo{WTNZD7Fh!q-*?H&=W{rPYC69WSi09@PqRwL%Cv&{*>`@eH|O1Z?6=h)mX;q}
zadG~KA8WT8PI$ad^<cS?_llqp%d1c1EyAD751T$yo$J{z*)=PcNipY^xX+$8*Vfj&
z<okw~vfniNHJ2~hn$o<bNJMmH>CX3C?;N|E^;eR2$rHVYQ+6!gW#j)|Kc>dk{Q}Pk
zn;07#8yV(p3H;X^l8;Z8mdSJ9wKXZRO>*OL8DquscAkr^-n*DDTzKByx8$~+;uIy>
z5)BEl_QTIOd>(yJaoo)DWPbWN0nhp8b}`L(HDTr?tA#htxUQJsvMk1NQeD_nBPqo@
zJZgL2i{F~Y=5e!oL*m!1?>$&5<SW01-?up2|M{kNqI@7j|NQ$ZifnJDt>5?O$5{o%
z)~prJujQXPu=jw6r|cx37xDQ&I5*wgmdthS;kMghwK<2XPA|P3U-~X=W3=b8u9TBo
zRt9c~%=)x@=k9e^&t`S4T_(0B)8y2)o!gdZ&D2u8-7u3YV4>Ki12KwTff~+EGr7Lm
zuINfryn3}i`1b8Xt17alF(xMF=^Sx!E1vUQDbdJ%>9o$4TO}_qmpr}J_;A9~RYEOF
zYz+&Q7A(%VEgHynD3iBqMR$s>bcnQc#_{ITD@9L^y<YqKSDXJv1t%d(mB9EFj=|a=
zYYztgn99ohd<9e8j;6fX8G34=di+*O%&s=qgy&3Ap5wc7<DGT){_bOp-?Yj=VUbr$
z*C8*4BIf4|EbkXcwK%%k_P=xfp1tKHhey`@+{If2*jM`fpJ8l#@WF=?ySy(;7ES8v
z<ThBWvZRwoVd+b|HoY`ix9s=3lNKzfa$5Yw`BU}(7D1Pt4y-R(Sh+7qaei4U%p>#X
z=$8hAnu_qplQ%y7zIFfJ(wVcn{LflHJgj(C_3<a>pN7vrpE*DOoT84wo^zIW8jCBH
zqAf1mwLGC}s`#yc-u#wAhpHp{->@rL^sv2XG;s2iYW&g0ruriBySVlpj|bunF-5AA
zrWAF*?>+pRtvtwPN8<%P|D2Yo@nRd@+HMrjw-4O6-i7nx>o1>w{E468si>^b=5uCQ
znh>Lcz#`2qLBSxmlq-RPBEmi%HT*g3ha6I`NT)~N4qKA`_}_^YN7zLLgKKLVy3{@>
z&2>^|Vm#KM@aqNJ6!z9f`BAbP`ql<aPGWn`a`)%n_envW*A}!fq_OhcO{q;`Uf}XP
zfaRk`*A(6*A*&WFNZIqUajJgX_GNQ5nN!#1^WD8Z=iryda~%zzv=dntr}0iZ`6X=o
z<`)lbY|0WOXQtS?EqEEbDOjdJ&XRLdhEMInz`v*02=blN=-@S9q976ezSYHhf^?sQ
zL%?H)B<37Ji}WS?BsKY>l<GFzk?MVsU%+Z!u{P_bsYXhi%ZkFwUsp|Qda}E}fd8Rm
z0EbeCL?DA}vfu^*rq(DS3Dp#?83_`N3T&bi96gO2<}xtlzFpBZw@!t98=F>BBjZE%
z4WS|In>J)5r7~M|bV!{$$zIbmt8J6Z&3)Dfw=h^w47(_*D-vGBv2RV2j7Oz^$htpO
zUnOPVZsOv5xc6mZ*^Hp2^DZ^0r`Vn7&{*Ce=yI1O(7Mv%ii=OFcK>P-Es;mxp878l
zyyRf*EZlQui+p%?4&U)ijmwj_o;-Q7J-NpGQ1^kZxk?<W3swqvbj&=wfcdor6H~ht
zcPRG-DG3dRy$)UK9lQrVGQQI2nA0c9%6xXmuX$To1VpB^2xbKI=dmrR&atSd$US`c
zaN{FER*{!1fu@U=1TZEoSWx{)?T4X*@B_K?4R(jB^If%7KJI6mlq9hrMWe~lL|Lf+
zT&@4<((9Mk?wc5QTYr`A4H?k~uXi^;ddt$&5-jkktG{iLgstM)m62u(oupseBslXd
z+{idvZC=*f-12q1vyZb(P%jYl`Qog@B`3|)WxHVAA-SKG4br;~E^(CgTeYk`cJ1w(
zoBlS*9&}fbo$#*gf?UH+2jNp%irdO|%Dc021>NDfn~>pLwN}M!xx?h#*}s2(zh9s0
zAk1Q=(KN^3ONn{*6Ry5bO3Zfy7^@BlzUyG(`@zt#Z%0G>J5g5VRtCpS8>|=nIZ%0g
zr<<Molf`ys$q%iR&oL<}En|{&o0!n>y=eKCxfO@Y)^wFis4eNx=(zLzZQT0WBRs<A
zolj`o_e*{LY`wYuNsD=|8m_FGY_5V%4q}2@6Ba0}PpxTdc4$%$ejmX1;SY;zz}G3l
ziAN3|VU&+_30b6^F7knCeLa7{@%A|W<oc$4|2)>Wwt1(qpL!-cGd0%Nl!J+Xht}+_
zRcZ$W0>17{QA(P`)n_EUeo+Mr`@wl1Y$Vrauef37^KpTt%-jji25reJ0`#;vyks9}
z{95vE-fEtQ9gdHJ1aGciwa0VP*V9i<g<Lk~7M@^xC_Qg?&E0dB>$4^}xis^tsAoKP
zTp+dAed4;ex14-5c9p!myxfiD@lwO$PaI8Mxd}JhE}oT^u6$N@g^P``%_HHiG4Bnd
z*5vjDg>4Ui^egVG+uL1OyC*Pj<NdQg@1A}!VOmEklZ#}+6itVgiLG1>E9NsO*+)O-
z?O>{2==}0eCDXFULIM+8oE6kA@-g<@Tp!!c_$!gQG(5)lr-Gz*%Ti5;m*MBaUoy?O
zxaf)OM8)aa`(r-lW=SvTV1KA#zxxtLgZmd9|A;>e{_MNq;_)HHRIqa4`gpUmb8b5<
z<H%*2A7Ll{;1^@@G9j7H<-Vm+*RHPi<NW%?e}?844j!eDkX+S{YgZ&z2S%s4U7NH`
z%WI?Prp3D378&+#ot$-j)72l`_o6-JGtRNHR4LBXU2|pD^qV0f{4ESMkq-W|9YaoJ
zbPB|ramhU9tS;1C@Ms;&6PbJGmoKoes|^k*=i9=-_h?CjlUGyY*&jz%A2JUyWEM0v
zG<c`Ii=kPA!7BXd_qn%4o=sXkU2oHA(^I=k9=2Gtx<$A=dyt^k&XiejIlxs=nae8t
z-a!MklQ$B>57fzT=igJyr`^nvc13Pqin6Ux?sWgv*PUd}Sc*G8X0tCS%3b}Gr>i?*
z$5w-*Tdr;L)3Z)oV-_ED(A>@K^i?yjkQw2E-wv=Xi{A05T4tKc!p`<Xj7#_3DAI0z
zUbU*EvU*>Xrp7nh&(_Ute+q@m{Mfk!lY@;V-#D!nd|<<M?18Y+<~zk&ulXg|6O=b4
zvv0a|QOLz5;GN9m&2#t3t~A@a`@ub>C5u+gQl1}swM$KCO6oU_YgW8jd@}JI7hIhW
zRK(u<)Lc;!^H9u1NAjV-q(fFm6dxwtR8a7p<-r!Zte}efp}fT0ClB}*Sb9$7P1*F*
zp?ShVzTeyqH$u)nc4<Dw5|zvo620ZA(KM!5rK3{?-rdfQDcsng+q-1RHa&sq<qOn~
z@km@=>Cx~k<H<n_w&vwcE^_<rjP>@dnqi%5P-oo2@snxU^4Uj@op|!zQgPdTzY||7
zW-)OtYe-Pjc+Ac?@x%!)Ckd{k1x*^g-(J7cJDfbxSGH}@&Sjp)2`_55PqI{-DZ_S}
zt<t83-#kdi`b@25Z1=loiN~@xb?h`&T-$$K-70%>@b_D{j928u+<36+$jl9<(sKfM
z8s5xci<ew<JG@1Aanx1ij*s^w6J|Lv39)c?EN!sWdr?*7@3+2bA9q#V{N=Ux(^-sM
zr<Mv8eqd!@?wdWItLlT@KIWI_mMgO*G;VF4I77~tPkDV~s<u>*+BH2EM}sR;XF8PY
z!<k(S)vVK&sIeFE-<U94I=``5AX{C>(5OjDq*c@HSR>ztKINpT5z|bw)~u9{_7<9z
zuwb6ml`x(tN$J9wYq~VA9^1Om$irYco7n+1E?&c%jUN`QmNj~mA*JmRap6Y4PnyHw
zyJCESn@ffE&9S)<A;b}v#^Wv!DA{WjFu_krViGfBbw*2rB>R)kQ=ahs4{S@=ac<A)
z;wcPzRUT7Z`5FxaFF2gFnHtu5fbXD?t5w&*6M>T)o@KM1G5B`PnrFALi>7zv!cdL1
z-2vXbK^^Vk55HfIs4no8PfBxs)0QavWa3^;W&UczU|tRBWwTkdR5Jo5Y<Ljq^5v1<
zW~)ULy5z2!M|1{pY!tL`e{tiZ$4p+%MS(ug`vNvJ%>Ec)bU>=B)&16%nGBwXKRgcE
z6JCAiM7mhl-QYPVg&#Id?077=GOv`&uxQzY2I(UKS_}BEC~_>2UHI^b-xEs{nG^i>
z3tR(&or1U~GPrOBX<q##TkD~teSNm<7Ip9CJ#EX+IzN`5(-9sZcR_;D>ulC+2g~ml
zG_TH$RzA6pam$mNb2A!+1X>bUg{NkpbGXDl{b{05Dzn^+Z!eCnDD`C9GIfbYmDiWp
zJMPOi7Ve3Ca`g1<iBGSwXUw`M!?9qS@snAe8Yg+CicjIx-gcsWQGl>-fTO63gQ|{4
zfPsrcBg@LOb@dy>nv`e0>bnr4dHL|(-Ib>T58N}EC35*};g+aL(ydiTY~r+dH6MMf
z*mFwtfkS-DHEV`1CypFt-?1otp-HpiyFz9a4o^<cmPyl6jri9bU=a+-|GK{B<>%v1
z?-ufv_wN0yd0%|0)+zI=SCzSvrU*6ocC>AH?iaW$`moPolZDH~yq-_hGSm2@zD`4}
zM1^Vf6q6<q2_=W$X)T%sw?1ex)+;GRoN1Qg3BUW0H!pAB`|YI%gzmkN%f4Of9DA={
zRw|%nk;b%$m03$wjB`zvoZjtw?#h%)3ui1<J<zyo+ocx?3f3Q%#CnvOv`o!QD0e&>
zD5aI1B(uuAm5qOiPVh|;Ew<-tElzJPP}|qYv%yt%f`<n?|Fr87jIvi)zDG$p+~ad|
zGh+BtIp?{qZH{<&0SDtrzDWWq5)%|cd<2fT6ms2nc=SB}JbU*!Sv94kyVv9;J}kWP
z`BUtJ+s85s_%$8|vK&-tyrpm2E7p8Rgm*#N#+?co76ClL*SWLG5A{A4IB0Ld{C<C|
z4&T;F79XbdSI%|rFnw_?u((<Nknm?VS@rhkiZ48lBse}&7ur5Ypx$Q3c~yr?zxI{v
z6mn3BR5saj#QLQXyRO4YlZ!^?N>dCuCHf{RSMFFb^Qc$vuR}>*#yhLSZfCy^O6+y@
z42p8|bg0^t>Xw|)*)(D9mn9)flyBVKx_afdi(9?kYII!XWfTlF@!c|U!ymilsXwK+
z-&Jqa<c||K<IGRK@y+z0>^eywp>oqTO0ot^BwkBRXt?lDS(jsyPT<zhzGCYf#q-L0
z6M_WUnbcBdPTcZP+hf-XE#J-3T3ZgZF@y>wXXLNd`*P;JzxJxIbnVzo&CYd728W^}
zdPN)8xUc0l?aD1;>0*7Pc}=SM(n4o}JPDSF=@*LHE;+UZzTe%$&=Y3SxTi<MmU&lW
zm)=LA2UdwQ9kpb?*sXZJLVBZA{DV8;9a+mJ%H3M%T>Wa5h-T`7_5)m=5<Tn+oIw{%
zy4Wt?RL`!6pQ6ICdwz_#c};brhnMOt);COIuNhlo&-y1@miABS7SX&Lc)-?qZWYIY
zHRjd~yUHg>H+XdgOnH4{-};l6?k(8Nu&?laP<7n}`;$Kyn?>WZR#oo{k9W~m+?E{6
zCAzeBR<8Tqm2ZC8=-$^j?903L<;>tUS*{GZ#aAT?cqDG8d8S)N8E$y?Zf55qW9O~|
zookspo0sr(OnT(R%~3ezW@X{kywB#9MZw8zj|vh@T?IA#v&7ntx#VooY1d9WmYtih
zX1Rcgh{zP1xoLmSe3hU1=0Wk%9kyFnoC!L*WU{r{k4f*gNtC6WX_MU%apNjm$JGZR
zZ*+q@efGX6YY+FFCY}9q>3`lemz}pf-SGHk)}~okrOke?-IL(<yKk${q*E!%WfQWk
zZanwMn;~^i_DuBNTW@tcCTd*sGMKPLdAi>$xf&Uvpn#~`ezTG^+jTPDX(YESdwANs
z&3cC7*)^6gKc2h1&w0Vg)mzPltkiBZP2~LYe9M7IC8ZV--g&WWF6@5Qc6Ot5+S9%>
zTQV<aUb>~rXt1!0A!Y7Y!>IcSceQ(FO7YBj&$`UO-J*PjPV&c1Q-2&d_x(XaR(*Zk
z%0gLZo)Q~(!(_EUeFLBF1<gv{x6HI$k~g)TlU952Mb$(pgV$Q>kyt~b#L>oz&#fEp
zF8lC`_kp<R4+aI%yzd8#Irg5OqVwU=q_)5-BK*e%A`U26=tMEKv<nz;b6EQbue=_h
zdH3P18~oh|k44VOj_SPJ=CUy$#8Xr6^|n+~?cMLorw7L_pFZbr%~`R`%C$kwQ(Adu
zS@iU5Qs&{(bcnF<ReNyY5x+j$Ie$)xkmidHwb@VgRrtL44{)DlSDh-nZQ+k&ALBhA
z>v2mmUeU0<s6L@=vZ=7b!MDYBjJF?JO;BPq2-d8Ua(f%z@gqFHhe@l`oOu>s=$yls
zdyZ)s%l7BTYW91Y)Ko-GQeM=Mw?6OmqtDkjYAUUgVrp{I)8CwTRJcB8N$b*powt)(
z_Fb67neoK1@FoN2L<zoAONI4?wTi^&Wjn@Qo3O`le?t+&`+J2mRJ^RVa=5gyun0|2
zZf#jPNo1Xq+Kn5#u4R>^sXj5gwW~>R?#0Kvb2~oXzO!C$uCupC*~&m?4#$%*e2aun
zupDB)An`rPviXsV<*5K`o6>c!4VBgg1x*ooX?fMCV%tHpL%px86&5aFG&XhE@<?s>
z&0D|h^bAb4B|K|AyjOX}B_Ea)g9e`M*Bs0y81u<5E-TI5I+gba|LjE<`abjL8)uxC
z2s#j%l&~@(al%Qrgk?JucHC~A?snFW<;P>~wpnVwv$oxEXK&|w&bIpAp`9`fx6iRu
zvP?K&d)TR;DM>@Bu~Xn{3j3@IQA4XE!7n`d=l2^_IsFV`_#)w79I9*J84?u|6}D`1
zu3J>p?yW`oIu3l!5xfj)dylgGJ#Zvux##7zA>|xBCmK5n-<o)(mHlm*?Hb72;j*gY
z;oY~#&&<7cFG*}gLTFLb3??QenaWE-60Fy|*v|V}sZWd)d>!J&e}-Ru(zM_;PZqS5
z+3B0kyA%7YPLH!?iIb$#6t0xZAuhr`f(oUbbEcoFX5O@LTdut&+mSDqo*YvPsLK@O
za1}baK+o^zy^IrXo(4v-%neL(kB`3-y*5$4>Wa_@Hw%?pfm%}n>R6xEP83?+db&eb
zq4?4?pCv|x3{E2KYws|~HhhT_6jd*o+UUgL*!ofN#eVsN{ppGtK6cu*-nU=ZyYgpv
z2&o$^$ZS$lmY8GUcuXMm;a2q(WxH74Y3OiGD6yWd;ks~ziO3N#*PTyHc|y6m6C5~v
z3Op9cKJncy^MLn+RPjUy1&Q4|P1si6Ty}kr$(^6Fx7Y3A-O|(LG0`>WR+P`N2~(Hq
zR!1|QcXKeWIq-G46Z`3$r1j^$eVqMP^!7$v%sCyblVvvP;;OWbi$ju=j$~}uw(EhG
z+pFK944am|X*#lpqjO8wh6yj60uC*Bp}wt&lYe!_Rd)Ns=`SzOF9_Ltc%$n1DK`$U
zJjl*9%}kj=WOH<g*0DpZcih4^ubFSNL%zCdX|R^^xdyozi@61Ud<o>3w~UWviD;m(
zPso(Xl7_L@Gs0_aA4PU@GBi6MX<)pypZ}(@(<g~Xf%YLYL>gk#E^&EGGt%eCY?xlA
zAr@o6pwFk_s?1^FzGJFhzz^9ILhO~x9$!+srMJJx(&_n(x1O)p&pKJ5Uu&ykP^0-@
zdLy&hd!fAW)}Fg8&W>jp4G-*WkYhZ)X;JyouvUeZ1qv%dmqsbvPrVSc$L3Xy@FV;6
z7Wa24Ry_?@GnG8B{&=;g!tH0X_I`Y`NS5n?@&)S`)(JglE*J6?D<5d`@i1xKeL?)~
z+_{W%V@-nIPy4h;?(k=))4M-0c`&?6PqChEFEoMY*NKhhyuZTMANc!OG;R*V<-kAY
zaz^_Dt7e<5TYt6ee)*=49z42IK@+ANmU?}1rP%JvpHp3gBR1}JHM`L<NA5^3qt_SV
zhQ=S00zPjvVLu@u&2ai>ft;b!>EAI)RXq+X6Y`EEYdUW*N@ek5jc(q@ekRCdnpgi!
zr^W(1L(NZ3ZpWw8EH&Zj*v#SMvc~b=RA(keEv8wi5)(ZX1irs(^|!v$-eLLcgCS3<
zB7=f}lF5}fml`*m_{3gw@)mO=pQ+FbMx~}Q1%;(*6RK-_a&B&28yWqGQ`hi|`hjcn
zE}k@q+xg`F0p{t~g^it9OgxJkK3E@OEPCENu}WbQPeV-Rn@PgETf~ZXryR82VXSUo
zlr?Ju$0X*J1%7Qx3%fdM6qig9St4TW!mXj)wOIO2c#*X0KL0K1vpjx2e38C4zW&Ld
z(|_e|O-yj{SQ6R)T<&M^%uqh%Z})qaF=Ux%OpaR{`gH-@jE$*Fzjpd0YVa#^{a($T
z?fv15XhvM4@oVW<(^L<{?7i?rXhD6=;uWROHK&};TzBl(Y=tRGCvI%r7kBl#nvtWk
z5|5zb`?MxrmYEW|5|946vny#_?u<0`77U+zebK9FT{-QJn^IiYbgN`k-z($Xy60U_
zqIllBY}?aoPkq*ieUkZn&W??XvpqQ-bLSm!oIPVru0m4&VgW8D)ewbNm7IC?Nh}j=
zI{JR`ZV0k?I=81Wi`T<m&AIpG?^5YkRcGc(PKmlPt!I)$W6r8|+h5$Ddqn@J(gm+n
z1KZ+fy9DfuWfvP<I62{xr7Ztp$J!?nl?<0BKU%}TYH^>gNPy2|AL}+(u0KEIN}8^?
zZ)4xV>dwye;7~_zf)MxZB}Yy=E!!sc>tmkAyv$$X#iyz_?{UBG6ZGlBuf(FFMN!&3
zv8k^@PH-JraqfqI<Egew*P3m=OcUu0RV?za^hh{oK2hLv>(?dnZp^HU8CH9|yz-*i
zWr}6SoPC!Tg_`Y}p8dde?NLRK5&?^a@0|m0B#3`G{eFAiof1u^M^|n;AIW}n?`@uD
zlIuN9cfp=B2b7X|VmLL<OzJ62y%))1cDN$uXp8NZu3pg<>cvhBDU!ER&Kf&y)ev0V
z`R2#DcsH+vh;MCjUX?cUu20ts`Lyr8wr4We6|19cyaz2ym#<nW6#L`IJ`s!Cy|eCJ
zn8221!YJ9ihx@ZoGQ-WL=FBHuxh$(alf9WUB=STrDE8<p25boEV_C6OZi49NgAPVA
z9zV{-wNE@?_`;9*h(p0O#TiN47sySBVEFp$;0hn%C!R_R(lppP9E=(t2(`v63Xa^-
zQ0!vD`1HQ*xy^^looj5?bSu7=(veg8{Kff1+`kvUU;g0vS==sYbHsvs)l1vU@0mHm
z9SirW&shAUU}eabNbczd@tbl=o0V>q9h@bypDELB_3kMq(`O1RB+2C^?Ku>-EhD1g
zhx@7LEb@UOXHPzTIqyMh7gOiVpPa$mlkTm4XICNeAcd)=gS$VyDC<lS=Y~0~6|Sbu
z9i7h=&sLub_06w6(G<u0lTAM2KwfwUr_aK>L0|dYTQYL@=xeaOFcZJs_JCpLA?-Of
zQImyb+7tyoalYT)zwvKiRg=TfRh~g-Q*Up~Uia#?X=!=MuBp2sZiQvdSezQ`n`)xc
zvWV9xNZ^a~7A>14^Om1qyCe9BwNo+Uou;sZEuVQ=dORGC)I9v|@R=iN;Z$>RV~^v3
zdymHJPfhrFx5h0h)OLURdH>=&+nSXaUm0D!+8E|A<Il;}UTq5|20z~YB!bc5Npge)
zPol;QCYRD!ZQ1hJ>FZp@C+UUO7UrGdyno{%uRwcEvq)0rj5e2BWtVqln9f))G9fWB
z)#RLmd_3<$ogbGs9d7)%q2sCA`7`T8IL|!h(Th%dvg^#Lbw!da!PXBKU3B4Vl0Fu*
zRc?vIJs16s83&&!trxO8eM2sBqnK0gq>kBE$C|{PPZq6;na1=@l)FQBHsj?tBCO@^
zzM>b(eC(21c<#mq%G;Yf6KrM6Fj!?4JnIYJ%DC-oLv0K?1zLGI1Xl|vg<1>nTi+05
zxRa$-*)hA@Yl@DSReSGL$pZ_7&G=?4^kP%_Qg!6g9HzHBH$_ORd|o=KsY_GosTTi6
z<&zF7u{8%IzP$Q;Hj#;^B%`Tqg>ulUv<1S>DZ*M?zq;S`nXfa=<xIwp>l$i4PY)^!
z=q+}2SaoVyVoB4>@XasYp1m@`yjy1zliIR_M|uvJ8@_*$adAPKNb3e4cabg5jSn1t
zxtKd#TI}MamDI)9!&)RDth|O{gJeqg{nvIyjDBBiWJ_P~KGEO1@|EI=ju~FXJj%}t
zPCw+EQoGr#RU>%SoUh(n0~U)O{1txDAn@QVrX><)%q_0vtyLXUABucitRU)M*_{6B
z$UFWud*j2@cxU>g8uO^WpK2<mRrzm=iff{z$btni4S$WgW~4fNf6y@Z-4iaKDY}fE
z6QsH1os}I`O+tCLG#>os`dj~W<K3FS@4hn5oN4Id<d!U8A@aPKL5Pk0uF48W+xdN8
z?g$vL_Y_Do=4P*%lc0N@{gKouv(72E7EP|L2+q>#m^pbe6EEkaDKk857DP<(XqdsS
zbV%$#u!^AsGgAhiqNHcA$lE}Er4L#KvYBgrrg3cx7Ik{!c8Jk|XIDF0;{Nn`%Xjz9
z5mY>ILh{(A1%@SoLCF^PT{+WoB>UEg-Q-L*;5e>!M(~8=wQVkC97|s>Xs%l3q?Tl(
z!k+S%b$^m-&kEjCYyuOCXAAl+>+TYJVO_GM!&9n)bIOLsf64aqMV@!MHSzdbx~hdU
zK4ystut|4ZYvz)|*utJ4x!XJAsMxX@Cv$Q=C-p5;^)0h~YAYEK#nfQX&f%aLGf|KE
z)9-ogtIK!$vv$vCO>aJUcJjT2w|K6w*v(23e0SlvO3e|=jM6#ms<j$?J~yX4^i<gp
z_s4L7;iMOdZMT$V6t0HkN&HTqWYP9ex?hl`kHuz_%Gqxh(wHVH?z*r|Ov5ZtP%c*d
z;V+5Vg`N72c5>=xMFNCb_A-U<nxQBDdtH5KV+2p*&sSYB8nRpoOb<UqpX<+hyZC0a
zg!{RUujw+4Prk>v{JF7s%S7HMmnS@4diHF!<o5}N$G)GdYTP|zCx4WZ$LgodTiN$n
z&#iKL<x(JWK34J5J~b(=S!egI)7fIZJZZY@>UA4dO8Tannu<*?5BIz7Tgt*09Oxx;
z^y9BN{0n}^ELqg#wV=hrK(Dj$fHkx1J0-47Su89r27(3?Cnod0-Rl*~Q+(<A$Fr%u
z$Did*T5iR{{8V0i-X4}~DWZ1e#wV0mA1PMes^loI`u+J%^?l7JjRC6~4hiZ0K3Dg#
z>hi7I4C&i;Gsw85uw;~L>iIYvwD;V5ukmO1!rSI+Ei-rS){u4*-*&9<NX?6bLTn3G
z%JL-66caF!v~YUhu(WoSO!~<U5{*KevwS=bo#4?rt;o~(tV2cZ2a^&*mPly2r_0;!
zX|G~7F3b$puGV`KKIP@qtGCu2$<*`~TeLg#T7+jX)8z8-v<;V2U6dwFYFf}^z3||Y
za4+_gZ0yb*XU@zyVPN2JGtj@SQ%S9YxB0F4Mh2b6Td(g+Eo|*s72dL-lU2BV_Z%H1
z!}n7RV%iV3$!sWCGNJQuRB=>eVC|ZHXZ$r3_b9hGOUU$k-Ik77_;%l=MZs6x75KQ4
zZ#vj8<{j~RE~YKvv9YC7QbXX1`y8L;IntA>C0+i^z0L4FLHU*A%&sSMn)k?hpAoL<
z<O%q(A|*<4r-PJwse{E7FOTK=d@scxnKRrq^A?#R?5O#~^2hBZzu7x#!#Bt+ao61z
znzg%V;|iTwyjxBzNIw&_Cg9rgdCYN!!B0OYbtX6(h`2B76$$1NI;y`!VaMh*1|FRq
zDUBO0d2xvf9F@CZ&r!#p_@~Cou^`xRbxOnB3lH8XB>m~t33%?iyhPxvz#W5d)hKzH
z#Uiad7ll=gjWcKEZe;QbOpZ96`?~$)0=*Nh9x4WlH?grDE1$+#A+&KCOPhZ7{V%s;
z^Y<;Yt^Pmf+dglPFYO2Jmjt#H{47aUb+h`A5!-Sj`~Jmwg1VErOw9Jpx;EQpWBHoS
z<lH&muO;?LXDRij&3Bw~<D~1;SR++)%aRGHM%!kINu(utJ`X%SEnAxHx?kF($6_Dm
zh0Dz{Nq3klb?T>aTY~hK?B2r7a{hT@jHT{dr>t9}_MW$VwKwmyCGvb~Z<T+)Ty^;L
zqDu|w#<P+ZT3lmWG_`a4{d~8Dl}C>yq^tOI9`u&b%z3rmitn}H(Lfdr$yevS`PgqI
zoDo+k+FbH=R;<D+<#Qa*O#FOaUc2jKx?$z9Yubm`NT;VQc-m3&tU%dr<*_@GYctAg
zSI^rl&MuLbefPo{{`Q%NMLBP(r~8N;xf?sT#mulp;?dc+H%^_i=#I!g$?kI0sbPwX
z(Ubhi?n|OhmtDA0#3{bOFYpPQVvO=Nr|*jkw92dho_Wn1W$LQ9GKOh0TVug{KJG;f
zCvL^MZquJQvvoC_X3qi<uZfHo3|<|Umd+9=6LUUfDtu8ZVdwoS*I6rFCLIo%a;ZY_
z@LJ1>vK?|PzLK(W4_z%db^M-jL<YU)o*=Ge9=Ppo23xQXZ>i`pg&;rUOmBvrEJr@(
z>B_htVSDhtDcayJLxO>2!}{8H-<PVabXs=%!7M}e0uR=6x`8rWcRJ);dz#q$xvB&|
zEw~W4esWl&%YE@8!2~HTWmn#p3%D69EZUAKHfe9WVVUw|Vw#12&B_3=mMjMmPF5oi
zDXa2{^_nH~LU>L*+;Z^LHMM^0rAY_4?mg8?Xq6GY5X$5!A)u)yKY@My`|gP!LeKnI
zkm+_#Q#+(+!dAho2FLeXN-{6>%{iYp-TX?zq6sDjTPt|1k114e99H<ABX)BUd%_IP
zA}?mi(lwd~dM+e7?r40gGWWMa@}*wA9@7Bv)J>|sy6WB+j`l8Dv+Q7_S7Nql%X;pz
zkb_~I3nwOuXeISJE<5|PBh6-`m`ahPM`PZcMN;MtB`a1uo~Y01xY0O_mC3|7YN=J6
zw?(#_8kg=}hgEl;Y?C`(;n<*d;GV$E_e)t%zU*es-f&atki=rAz;}=B#C*TMGI+jI
z*yPg9d2bF+3^&W0(x?&CpCUB#vfZ{RC(W1K=KA(5!#MiX+dZyjL7#k+UNTs2navV1
zBePI&LL2w8pJ$6>xp$fT;cYEB8K1Co)nQK4oDMaqEv^$ZH<>L--o$=DKJWIsyu5Gf
z&I`U3ZZH-N+MK9xPF|P$T;LHs8Rrj;H(nHka(B0hEZoqPu%dK>?Ly`qfgx)yO3&D#
z;^63Ia9|Q!Z?E!=3l|+<=c=xpn(XT>)pce~)R|k|k&!8@mW3@{n(E@j!`h<A;rK>h
zK}zub{g-jduV(W4J=i-zZrvSC*}K!9RKE_9`B<c)p~>MfGr?otolUaY*D9DFn^`+<
zlW6Vr+*N&fSItRg&!@FAZs%=OmiK1+II?XqVA{Ck$h3c(v^AB2xX<l5a)`Ovhv&Aj
z>;A;Y-y4#>LJnOpS<2+7=s0Pnu!{3N_KSsv{3+F)cb(LhCL8f8TwapQBvYo@&(jt(
z^HSKy=O?rtUA?xONn(f8Iy2eHw}i`-tlA8?9D^;(1r!~5ds_`t{4V+wI+`c!6j&<r
zaG6MU;wHu})@T1}15dJXIZDoKd~zUVNyw@Mt1d%h!=vH0{j*oidbKfTTg0neY3>!r
zWK5S_p47&&YnHHrhu4L~NvA>_MOsZN7+#CCT*^{W;t=ABS@DFi^5?NBmM^E2T(a2F
zYqXfxN1CJY=0o<r*F|r=oR9q}5IekWSpd_gzpn}=Ok6Qd#7((qN_55amG8nhbN&Qh
zWSpiYrl$3NUg|duS)sS8hi0D9Xyx!uDADWUb<A0{JMxKwo4BY$W7B#D*MJ+%fmdW1
zF9>Gull%B%<-5J_s?S!R@4P>YO^mI9Dbh)6b1;ui*5zPTVO_;&O|hHHj;R=(n5Gt>
zFd>q|iG%%2t@8ytrJc%*Uo8$aNGDx=Y8ZDoG<b)(_|Z(!c>W_of^3Exs+%(WJc>C|
z12#GvE^R%(o5|Bez1Z+c|C+N30m<(U90+Hx^65~JWZ!?o`YG=PtK2<l1r94B63?Y>
z<FaINVkkfNBj*ms^!8&HYLl569v2_b*w&Tk*67l;u>Evmfa{MWRiQQJ3qMs=byRa+
zcHbQJwX{faj!N*y1yd9bG*0%AdCSzPvzX&hkK*g~Ykt2M`V-0TeMMq1i+?uba^Z|l
z=K1wGr;aoyWbWR*JulTNCerNWvE#xjGZYl2m^@x%Y%G=*rJ*Jv&@OnOidmNRNQmu)
zuj;$5FPW^(QK?{=#<1Stx4TZJ*i5e=&8aS5K3g34(5B#`vv>Em3Bot`1WPboS@>$}
z&B$FJKNc3oX)ep1crDZS{Onr}+C^L{>}L%8EUFf$gx$T=u%2Pb3hm}y&-&)N{no#|
zKD@knwW1e`^&AP0jo<s4qnylco6Jq=2=wY~UKH%O<fA52UdxITZI9O*ymb855gI7a
z$Yk_<^AeNy@861UpQyGyE9~h_VV*=yN2Tqimy<*GOxg7M-F5Z02wt9r*&00hr4^>L
z(yxCmke>78zOjJCUF+>3dM*CpXB(y}#fY2;S{P=tVr~@sp*Gita`{pxt}j>G(9o`y
zG>7w$bNhiQnK#!f?RAj4x8pjq+>r$-O)32$mFrm@-Z3qxy~FQdyGKB1f{w8N#Th3j
z&e*YUjijJNV9Hg==DtnIFBQJaI^MDFV|kl(b&2$P)wQm>*q&_3<t;cM%4~ke!rIR6
z&MzabsAaipCj^P_Yq;}(Ic2%-I<b@>!;}d-Z)NSd`*m&E(Fu$k4*a>HwUgg|d;H}$
z*KD1%`7=)V9R1tZrGMem<9!)_u0GGZq8Js-z53%{N$-~jL*89i^An0af7`lOsN8z)
zx%;`FWKQ3n>@0rSJ1S+W-9^{6f$_^l&$ei5tqw`+(aPTQWRk(kSBq9G`N6(?O+d?{
zOYc)A9SE9!<QSKy$l1_{rJhQY6STCFbcH5Ku9w)vT5|OAk!T^#sI0Y}A}ca0=Wwb7
zNls!*HCul4y}1WN!~s=_V}{}$EGs0tW>#%oHDN+igU_8w!oey2)%Wx!OyarYRKMcM
z%Q??9b~JeGSXQ*Lqhq1UgaFQ25zALFGHg2K$ZTjh!^tB+qlbC-IaNhJwR90h7VXfe
zMXusp<^6v|76cu=qHX(DTcl;B>E(NCnGQvCD}>C*jyteU*nF!?=ae&NJh*iA%~pEK
zH%$)po#a~WXIgsfq_UjOZMDt4liY4E%DVfGQ!>{<EUf4um(oY^B@-4Yop4%OVshrl
zl6AcPp_L-5W1n4XQ4yJ?=(Klj=9EdA#-VXSUP}6$%v)StpK>}2eGabjR9IO#V`axg
zuN|hL%UqnTX8Rm{sUpaHb*+Jh(#O<-w65;EDgxI71GAlkcBSocVd&WVDwV-&hwBs<
z*GFF(G_4OueHGT&l{wpmX{m9+*0q^ax|fEAty-wEYQ3=%t9W48icS`##HH`e7b@)9
z`#y6Dqqtw_Dn$j^b@JareSP0e2#$SU6RIM#oMW4a>-%6&g|>{nlem{1KH}uqaA4jl
ztwK+hH7vJ2R?a$liAUQ`?V{0ypsZCLo(er2t=Z9{=1ZE+2IqtZdF)yhw#Q9fV#k{9
z*n5SVnp>l;2AVqUSQsz1`n2!HoLHCLXU^PQ_-K}rt}ws8v9L#1^{rKr#(TZ=cm*@I
z+`Ha2TXE^rTS^x*T1z5r&bl0NSsi)rTlZO?v#+Mke6=d;VPJNsi|-YaJ!k%!zKHLy
zjHpTWoSHUs%eka2dsnT}yr(5;bL(>D%)o*vP8}WHEhernq?EpC`nW!xvod?1e)XZM
z${B~xKD%65a=EJiBA2ddwvS85Ikj1<HfBeC4K@z-nYC(JzRA^)cOM1Cr@5Rt6Zh5U
zYG}+#-?`5Pv!iY-6xkrYSEHjRX;qTa<TEyI2AuM3Cp_{uJo0i>5sfK&vi^Zx3jb~|
zzwa-LZ})$*Yvua4%*Q_Z$LYQI_th0ni7xs*|Jcv<yK8cPAGug;d+qo4f=d=Jq-$b+
z-Z$Ov@@AggxwN#jwA5`k^U`mH)voT{GvBJu*0AK!ulkO!ytV7stF+~Rn*C?}qdopY
ztqF`A21gWa+J(+NI(2ey|Kkjf&9^?S+4Sks+xbr;BU7h*h_ReitQU50bJ|(+%+;&j
zF)G+vZ`miP<kBhZ#Nl&CLQw8dZF?&h=LrrD#T8QyN6G1y9^|+*Q!9akQMc^atjwje
zk`(ylOnvi~7d%$uU|4fF@uPY7*Y~QA>tf^9{d@KD{L^ROs-xevfBnBp=3Ke@{hxJy
UYE13_kKUQ+C!)gE|LW)g0FTq@^#A|>
literal 0
HcmV?d00001
diff --git a/man/anova.mcglm.Rd b/man/anova.mcglm.Rd
new file mode 100644
index 0000000..3ba5b6b
--- /dev/null
+++ b/man/anova.mcglm.Rd
@@ -0,0 +1,15 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_anova.R
+\name{anova.mcglm}
+\alias{anova.mcglm}
+\title{ANOVA method for McGLMs.}
+\usage{
+\method{anova}{mcglm}(object)
+}
+\arguments{
+\item{object}{an object of class mcglm, usually, a result of a call to \code{mcglm}.}
+}
+\description{
+ANOVA method for McGLMS.
+}
+
diff --git a/man/coef.mcglm.Rd b/man/coef.mcglm.Rd
index 140d8f0..ad9b83d 100644
--- a/man/coef.mcglm.Rd
+++ b/man/coef.mcglm.Rd
@@ -11,11 +11,13 @@
\arguments{
\item{object}{An object of mcglm class.}
+\item{std.error}{Logical. Returns or not the standard errors.}
+
\item{response}{A numeric or vector specyfing for which response variables the coefficients
should be returned.}
\item{type}{A string or string vector specyfing which coefficients should be returned.
-Options are "beta", "tau", "power", "tau" and "correlation".}
+Options are 'beta', 'tau', 'power', 'tau' and 'correlation'.}
}
\value{
A data.frame with estimates, parameters names, response number and parameters type.
diff --git a/man/mc_bias_corrected_std.Rd b/man/mc_bias_corrected_std.Rd
new file mode 100644
index 0000000..0a1b6d2
--- /dev/null
+++ b/man/mc_bias_corrected_std.Rd
@@ -0,0 +1,23 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_bias_correct_std.R
+\name{mc_bias_corrected_std}
+\alias{mc_bias_corrected_std}
+\title{Bias-corrected standard error for regression parameters}
+\usage{
+mc_bias_corrected_std(object, id)
+}
+\arguments{
+\item{object}{An object of mcglm class.}
+
+\item{id}{a vector which identifies the clusters. The length and order of id should be the
+same as the number of observations. Data are assumed to be sorted so that observations on a cluster
+are contiguous rows for all entities in the formula.}
+}
+\value{
+A matrix. Note that the function assumes that the data are in the correct order.
+}
+\description{
+Compute bias-corrected standard error for regression parameters in the context
+of clustered observations. It is also robust and has improved finite sample properties.
+}
+
diff --git a/man/mc_build_C.Rd b/man/mc_build_C.Rd
index c5bfa9b..ba76088 100644
--- a/man/mc_build_C.Rd
+++ b/man/mc_build_C.Rd
@@ -13,6 +13,8 @@ mc_build_C(list_mu, list_Ntrial, rho, list_tau, list_power, list_Z, list_sparse,
\item{list_Ntrial}{A list with the number of trials. Usefull only for binomial responses.}
+\item{rho}{Vector of correlation parameters.}
+
\item{list_tau}{A list with values for the tau parameters.}
\item{list_power}{A list with values for the power parameters.}
@@ -26,6 +28,12 @@ mc_build_C(list_mu, list_Ntrial, rho, list_tau, list_power, list_Z, list_sparse,
\item{list_covariance}{A list specifying the covariance function to be used for each response variable.}
\item{list_power_fixed}{A list of Logical specifying if the power parameters are fixed or not.}
+
+\item{compute_C}{Logical. Compute or not the C matrix.}
+
+\item{compute_derivative_beta}{Logical. Compute or not the derivative of C with respect to regression parameters.}
+
+\item{compute_derivative_cov}{Logical. Compute or not the derivative of C with respect the covariance parameters.}
}
\value{
A list with the inverse of the C matrix and the derivatives of the C matrix with respect to
diff --git a/man/mc_build_sigma.Rd b/man/mc_build_sigma.Rd
index a6dfb2a..7178e91 100644
--- a/man/mc_build_sigma.Rd
+++ b/man/mc_build_sigma.Rd
@@ -30,6 +30,8 @@ binomialP or binomialPQ.}
\item{power_fixed}{Logical if the power parameter is fixed at initial value (TRUE). In the case
power_fixed = FALSE the power parameter will be estimated.}
+
+\item{compute_derivative_beta}{Logical. Compute or not the derivative with respect to regression parameters.}
}
\value{
A list with the Cholesky decomposition of \eqn{\Sigma}, \eqn{\Sigma^{-1}} and the derivative
diff --git a/man/mc_cross_sensitivity.Rd b/man/mc_cross_sensitivity.Rd
index b853313..68f9247 100644
--- a/man/mc_cross_sensitivity.Rd
+++ b/man/mc_cross_sensitivity.Rd
@@ -11,6 +11,8 @@ mc_cross_sensitivity(Product_cov, Product_beta,
\item{Product_cov}{A list of matrices.}
\item{Product_beta}{A list of matrices.}
+
+\item{n_beta_effective}{Numeric. Effective number of regression parameters.}
}
\value{
The cross-sensitivity matrix. Equation (10) of Bonat and Jorgensen (2015).
diff --git a/man/mc_derivative_C_rho.Rd b/man/mc_derivative_C_rho.Rd
index 7acb281..b683cf6 100644
--- a/man/mc_derivative_C_rho.Rd
+++ b/man/mc_derivative_C_rho.Rd
@@ -12,9 +12,9 @@ mc_derivative_C_rho(D_Sigmab, Bdiag_chol_Sigma_within,
\item{Bdiag_chol_Sigma_within}{A block-diagonal matrix.}
-\item{II}{A diagonal matrix.}
+\item{t_Bdiag_chol_Sigma_within}{A block-diagonal matrix.}
-\item{t_Bdiag_chol_sigma_within}{A block-diagonal matrix.}
+\item{II}{A diagonal matrix.}
}
\value{
A matrix.
diff --git a/man/mc_influence.Rd b/man/mc_influence.Rd
new file mode 100644
index 0000000..2886d73
--- /dev/null
+++ b/man/mc_influence.Rd
@@ -0,0 +1,23 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_influence.R
+\name{mc_influence}
+\alias{mc_influence}
+\title{Influence measures}
+\usage{
+mc_influence(object, id)
+}
+\arguments{
+\item{object}{An object of mcglm class.}
+
+\item{id}{a vector which identifies the clusters. The length and order of id should be the
+same as the number of observations. Data are assumed to be sorted so that observations on a cluster
+are contiguous rows for all entities in the formula.}
+}
+\value{
+A matrix. Note that the function assumes that the data are in the correct order.
+}
+\description{
+Compute influence measures for multivariate covariance generalized linear models.
+Leverage, DFBETA and Cook's distance for unit sample and observations.
+}
+
diff --git a/man/mc_initial_values.Rd b/man/mc_initial_values.Rd
new file mode 100644
index 0000000..ff2ef4b
--- /dev/null
+++ b/man/mc_initial_values.Rd
@@ -0,0 +1,39 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_initial_values.R
+\name{mc_initial_values}
+\alias{mc_initial_values}
+\title{Automatic initial values for McGLMs.}
+\usage{
+mc_initial_values(linear_pred, matrix_pred, link, variance, covariance, offset,
+ Ntrial, contrasts = NULL, data)
+}
+\arguments{
+\item{linear_pred}{A list of formula see \code{\link[stats]{formula}} for details.}
+
+\item{matrix_pred}{A list of known matrices to be used on the matrix linear predictor. Details
+can be obtained on \code{\link[mcglm]{mc_matrix_linear_predictor}}.}
+
+\item{link}{A list of link functions names, see \code{\link[mcglm]{mc_link_function}} for details.}
+
+\item{variance}{A list of variance functions names, see \code{\link[mcglm]{mc_variance_function}}
+for details.}
+
+\item{covariance}{A list of covariance link functions names, current options are: identity, inverse
+and exponential-matrix (expm).}
+
+\item{offset}{A list with values of offset values if any.}
+
+\item{Ntrial}{A list with values of the number of trials on Bernoulli experiments. It is useful only
+for binomialP and binomialPQ variance functions.}
+
+\item{contrasts}{List of contrasts to be used in the \code{\link[stats]{model.matrix}}.}
+
+\item{data}{A data frame.}
+}
+\value{
+Return a list of initial values to be used while fitting McGLMs.
+}
+\description{
+Return a list of initial values for McGLMs.
+}
+
diff --git a/man/mc_link_function.Rd b/man/mc_link_function.Rd
index 8449236..1434211 100644
--- a/man/mc_link_function.Rd
+++ b/man/mc_link_function.Rd
@@ -73,8 +73,8 @@ a list where the first element is mu (n x 1) vector and the second D (n x p) mat
\examples{
x1 <- seq(-1, 1, l = 5)
X <- model.matrix(~ x1)
-mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = "log")
-mc_link_function(beta = c(1,0.5), X = X, offset = rep(10,5), link = "identity")
+mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = 'log')
+mc_link_function(beta = c(1,0.5), X = X, offset = rep(10,5), link = 'identity')
}
\seealso{
\code{\link[stats]{model.matrix}}, \code{\link[stats]{make.link}}.
diff --git a/man/mc_qll.Rd b/man/mc_qll.Rd
new file mode 100644
index 0000000..7de1f9f
--- /dev/null
+++ b/man/mc_qll.Rd
@@ -0,0 +1,24 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_qll.R
+\name{mc_qll}
+\alias{mc_qll}
+\title{Compute quasi-likelihood function.}
+\usage{
+mc_qll(y, mu, variance, power)
+}
+\arguments{
+\item{y}{A vector of observed values.}
+
+\item{mu}{A vector of fitted values.}
+
+\item{variance}{Variance function (constant, tweedie, poisson_tweedie, binomial).}
+
+\item{power}{Power parameter value.}
+}
+\value{
+The quasi-likelihood values.
+}
+\description{
+Given a variance function mc_qll function computes the quasi-likelihood values.
+}
+
diff --git a/man/mc_quasi_score.Rd b/man/mc_quasi_score.Rd
index efd6549..bd3c077 100644
--- a/man/mc_quasi_score.Rd
+++ b/man/mc_quasi_score.Rd
@@ -7,7 +7,7 @@
mc_quasi_score(D, inv_C, y_vec, mu_vec)
}
\arguments{
-\item{D}{A matrix. In general the output from \code{\link[mcglm]{mc_lnk_function}}.}
+\item{D}{A matrix. In general the output from \code{\link[mcglm]{mc_link_function}}.}
\item{inv_C}{A matrix. In general the output from \code{\link[mcglm]{mc_build_C}}.}
diff --git a/man/mc_robust_std.Rd b/man/mc_robust_std.Rd
new file mode 100644
index 0000000..e8eef9d
--- /dev/null
+++ b/man/mc_robust_std.Rd
@@ -0,0 +1,23 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_robust_std.R
+\name{mc_robust_std}
+\alias{mc_robust_std}
+\title{Robust standard error for regression parameters}
+\usage{
+mc_robust_std(object, id)
+}
+\arguments{
+\item{object}{An object of mcglm class.}
+
+\item{id}{a vector which identifies the clusters. The length and order of id should be the
+same as the number of observations. Data are assumed to be sorted so that observations on a cluster
+are contiguous rows for all entities in the formula.}
+}
+\value{
+A matrix. Note that the function assumes that the data are in the correct order.
+}
+\description{
+Compute robust standard error for regression parameters in the context of clustered
+observations.
+}
+
diff --git a/man/mc_rw2.Rd b/man/mc_rw2.Rd
new file mode 100644
index 0000000..b8c7ee1
--- /dev/null
+++ b/man/mc_rw2.Rd
@@ -0,0 +1,20 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_rw2.R
+\name{mc_rw2}
+\alias{mc_rw2}
+\title{Random walk second order model}
+\usage{
+mc_rw2(n_time, intrinsic = TRUE)
+}
+\arguments{
+\item{n_time}{Number observations time.}
+
+\item{intrinsic}{Logical indicating if the models is intrinsic (rho = 1) or not.}
+}
+\value{
+A matrix. Note that the function assumes that the data are in the correct order.
+}
+\description{
+Builds a random walk second order model matrix.
+}
+
diff --git a/man/mc_unstructured.Rd b/man/mc_unstructured.Rd
new file mode 100644
index 0000000..52e1298
--- /dev/null
+++ b/man/mc_unstructured.Rd
@@ -0,0 +1,18 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_unstructured.R
+\name{mc_unstructured}
+\alias{mc_unstructured}
+\title{Unstructured model}
+\usage{
+mc_unstructured(n_time)
+}
+\arguments{
+\item{n_time}{Number of observations per unit sample.}
+}
+\value{
+A matrix. Note that the function assumes that the data are in the correct order.
+}
+\description{
+Builds a unstructured model matrix.
+}
+
diff --git a/man/mc_variance_function.Rd b/man/mc_variance_function.Rd
index b2a9c87..afffdbb 100644
--- a/man/mc_variance_function.Rd
+++ b/man/mc_variance_function.Rd
@@ -53,8 +53,8 @@ elements: V_sqrt, D_V_sqrt_power and D_V_sqrt_mu.
\examples{
x1 <- seq(-1,1,l = 5)
X <- model.matrix(~ x1)
-mu <- mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = "logit")
-mc_variance_function(mu = mu$mu, power = c(2,1), Ntrial = 1, variance = "binomialPQ",
+mu <- mc_link_function(beta = c(1,0.5), X = X, offset = NULL, link = 'logit')
+mc_variance_function(mu = mu$mu, power = c(2,1), Ntrial = 1, variance = 'binomialPQ',
inverse = FALSE, derivative_power = TRUE, derivative_mu = TRUE)
}
\seealso{
diff --git a/man/mcglm.Rd b/man/mcglm.Rd
index 502aad9..c61917f 100644
--- a/man/mcglm.Rd
+++ b/man/mcglm.Rd
@@ -5,7 +5,8 @@
\title{Fitting Multivariate Covariance Generalized Linear Models (McGLM)}
\usage{
mcglm(linear_pred, matrix_pred, link, variance, covariance, offset, Ntrial,
- power_fixed, data, control_initial, control_algorithm = list())
+ power_fixed, data, control_initial = "automatic", contrasts = NULL,
+ control_algorithm = list())
}
\arguments{
\item{linear_pred}{A list of formula see \code{\link[stats]{formula}} for details.}
@@ -29,13 +30,17 @@ for binomialP and binomialPQ variance functions.}
\item{power_fixed}{A list of logicals indicating if the values of the power parameter should be
estimated or not.}
+\item{data}{A dta frame.}
+
\item{control_initial}{A list of initial values for the fitting algorithm. See details below.}
+\item{contrasts}{Extra arguments to passed to \code{\link[stats]{model.matrix}}.}
+
\item{control_algorithm}{A list of arguments to be passed for the fitting algorithm. See
\code{\link[mcglm]{fit_mcglm}} for details.}
}
\value{
-mcglm returns an object of class "mcglm".
+mcglm returns an object of class 'mcglm'.
}
\description{
\code{mcglm} is used to fit multivariate covariance generalized linear models.
diff --git a/man/print.mcglm.Rd b/man/print.mcglm.Rd
index 01bd18b..46218fa 100644
--- a/man/print.mcglm.Rd
+++ b/man/print.mcglm.Rd
@@ -1,11 +1,10 @@
% Generated by roxygen2 (4.1.1): do not edit by hand
% Please edit documentation in R/mc_print.mcglm.R
\name{print.mcglm}
-\alias{print}
\alias{print.mcglm}
\title{Print a Multivariate Covariance Generalized Linear Model}
\usage{
-print.mcglm(object)
+print.mcglm(object, ...)
}
\arguments{
\item{object}{fitted model objects of class mcglm as produced by mcglm().}
diff --git a/man/qic.mcglm.Rd b/man/qic.mcglm.Rd
new file mode 100644
index 0000000..a984346
--- /dev/null
+++ b/man/qic.mcglm.Rd
@@ -0,0 +1,21 @@
+% Generated by roxygen2 (4.1.1): do not edit by hand
+% Please edit documentation in R/mc_qic.R
+\name{qic.mcglm}
+\alias{qic.mcglm}
+\title{Compute Quasi Information Criterion (QIC) for McGLMs.}
+\usage{
+qic.mcglm(object, object.iid)
+}
+\arguments{
+\item{object}{An object of mcglm class.}
+
+\item{object.iid}{An object of mcglm class contained the model fitted using independent
+covariance structure.}
+}
+\value{
+The QIC value.
+}
+\description{
+qic.mcglm is a function which computes the QIC for McGLMs.
+}
+
diff --git a/man/residuals.mcglm.Rd b/man/residuals.mcglm.Rd
index 8d291b4..1545596 100644
--- a/man/residuals.mcglm.Rd
+++ b/man/residuals.mcglm.Rd
@@ -9,8 +9,8 @@
\arguments{
\item{object}{An of class mcglm, typically the result of a call to \code{mcglm}.}
-\item{type}{the type of residuals which should be returned. The alternatives are: "raw"
-(default), "pearson" and "standardized".}
+\item{type}{the type of residuals which should be returned. The alternatives are: 'raw'
+(default), 'pearson' and 'standardized'.}
}
\value{
Depending on the number of response variable the function \code{residuals.mcglm} returns
diff --git a/tests/testthat/test_mc_build_sigma.R b/tests/testthat/test_mc_build_sigma.R
index 6d862cd..b19e0b0 100644
--- a/tests/testthat/test_mc_build_sigma.R
+++ b/tests/testthat/test_mc_build_sigma.R
@@ -426,7 +426,7 @@ test_that(
expect1 <- 3
expect2 <- length(Z) + 1
actual <- mc_build_sigma(mu = mu, tau = c(2,0.8,0.5), power = c(2), Z = Z, sparse = FALSE,
- Ntrial=1, variance = "poisson_tweedie", covariance = "identity",
+ Ntrial=NULL, variance = "poisson_tweedie", covariance = "identity",
power_fixed = FALSE)
actual2 <- mc_build_sigma(mu = mu, tau = c(2,0.8,0.5), power = c(2), Z = Z, sparse = FALSE,
Ntrial = 1, variance = "poisson_tweedie", covariance = "identity",
@@ -527,16 +527,16 @@ test_that(
#######################################################################################
## Computing the derivatives with respect to beta #####################################
#######################################################################################
-x1 <- seq(0,1,l=10)
-X <- model.matrix(~ x1)
-mu <- mc_link_function(beta = c(1,0.4), X = X, offset = NULL, link = "logit")
-Z0 <- Diagonal(10,1)
-Z1 <- Matrix(tcrossprod(rep(1,10)))
-Z2 <- Matrix(c(rep(0,5),rep(1,5))%*%t(c(rep(0,5),rep(1,5))))
-Z <- list(Z0,Z1,Z2)
+#x1 <- seq(0,1,l=10)
+#X <- model.matrix(~ x1)
+#mu <- mc_link_function(beta = c(1,0.4), X = X, offset = NULL, link = "logit")
+#Z0 <- Diagonal(10,1)
+#Z1 <- Matrix(tcrossprod(rep(1,10)))
+#Z2 <- Matrix(c(rep(0,5),rep(1,5))%*%t(c(rep(0,5),rep(1,5))))
+#Z <- list(Z0,Z1,Z2)
-actual2 <- mc_build_sigma(mu = mu, tau = c(2,0.8,0.5), power = c(2,1), Z = Z, sparse = FALSE,
- Ntrial = 1, variance = "binomialPQ", covariance = "expm",
- power_fixed = FALSE, compute_derivative_beta = TRUE)
-names(actual2)
-length(actual2$D_Sigma_beta)
+#actual2 <- mc_build_sigma(mu = mu, tau = c(2,0.8,0.5), power = c(2,1), Z = Z, sparse = FALSE,
+# Ntrial = 1, variance = "binomialPQ", covariance = "expm",
+# power_fixed = FALSE, compute_derivative_beta = TRUE)
+#names(actual2)
+#length(actual2$D_Sigma_beta)
diff --git a/tests/testthat/test_mc_link_function.R b/tests/testthat/test_mc_link_function.R
index 10a252e..60f514d 100644
--- a/tests/testthat/test_mc_link_function.R
+++ b/tests/testthat/test_mc_link_function.R
@@ -31,7 +31,7 @@ test_that(
mc_link_function(beta = c(0.1,0.2,0.3), X = X, offset = NULL, link = x)$mu),length)
nomes <- unlist(list.link)
output <- which(lapply(actual, function(x) x - expected) != 0)
- if(length(ouptut) != 0)message(
+ if(length(output) != 0)message(
paste("Error: Problems on length of mu vector output:",nomes[output]))
}
--
GitLab