From cd33c41df46a3cafc93ac19076efc1043df7ef1a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Franti=C5=A1ek=20Barto=C5=A1?= Date: Sun, 19 Apr 2026 13:47:43 +0200 Subject: [PATCH] Add effect/heterogeneity and normalized reports Implement Phase 1 of the effect/heterogeneity extension: add public effect and heterogeneity plumbing to the two-group/t-test family, centralize study-level draws, and attach a normalized reporting block to Phase 1 simulations. Many generators (.sim.compscore, .sim.covariates, sim.*Hack wrappers, sim.cutoffHack, sim.compscoreHack, sim.covhack, etc.) now accept effect and heterogeneity and use a per-iteration theta drawn from .draw.study.effect; two-group data generation shifted accordingly. Introduced and used standardized reporting helpers (.report.twogroup, .report.association, .report.group_lm, .report.multicat), changed p-value/analysis selection to use .selectanalysis, and unified result combining with .combine.phase1.results so older legacy fields are preserved. Also updated NAMESPACE imports, bumped RoxygenNote in DESCRIPTION, added man pages and a planning doc (effect_heterogeneity_extension_plan.md), and adjusted tests and Shiny docs to reflect the new API and reporting contract. Backward compatibility is preserved when effect = 0 and heterogeneity = 0. --- phackR/DESCRIPTION | 2 +- phackR/NAMESPACE | 5 +- phackR/R/combinedStrategies_ttest.R | 84 ++-- phackR/R/compositeScores.R | 89 ++-- phackR/R/exploitCovariates.R | 59 +-- phackR/R/exploitCutoffs.R | 72 ++- phackR/R/favorableImputation.R | 115 +++-- phackR/R/helpers.R | 456 +++++++++++++++++- phackR/R/incorrectRounding.R | 57 +-- phackR/R/optionalStopping.R | 80 ++- phackR/R/outlierExclusion.R | 206 ++------ phackR/R/plotsShiny.R | 2 +- phackR/R/selectiveReportingDV.R | 82 ++-- phackR/R/selectiveReportingIV.R | 133 ++--- phackR/R/statAnalysis.R | 70 ++- phackR/R/subgroupAnalysis.R | 122 ++--- phackR/R/variableTransformation.R | 55 +-- .../ShinyPHack/data/startplots.rds | Bin 4193488 -> 19095997 bytes .../ShinyPHack/mddoc/01_CompScores.md | 2 +- .../ShinyPHack/mddoc/02_ExploitCovariates.md | 2 +- .../ShinyPHack/mddoc/03_ExploitCutoffs.md | 2 +- .../mddoc/04_FavorableImputation.md | 2 +- .../ShinyPHack/mddoc/05_IncorrectRounding.md | 4 +- .../ShinyPHack/mddoc/06_OptionalStopping.md | 4 +- .../ShinyPHack/mddoc/07_OutlierExclusion.md | 2 +- .../mddoc/09_SelectiveReportingDV.md | 4 +- .../mddoc/10_SelectiveReportingIV.md | 4 +- .../ShinyPHack/mddoc/11_StatAnalysis.md | 4 +- .../ShinyPHack/mddoc/12_SubgroupAnalysis.md | 4 +- .../mddoc/13_VariableTransformation.md | 2 +- .../ShinyPHack/mddoc/landingPage.md | 12 +- phackR/inst/shiny-phack/ShinyPHack/server.R | 70 +-- phackR/inst/shiny-phack/ShinyPHack/ui.R | 142 +++--- phackR/inst/sim_startplots_Shiny.R | 68 +-- phackR/man/dot-combine.phase1.results.Rd | 16 + phackR/man/dot-compCohensDData.Rd | 16 + phackR/man/dot-compCohensDSE.Rd | 18 + phackR/man/dot-compCohensDStat.Rd | 18 + phackR/man/dot-compFisherZSE.Rd | 14 + phackR/man/dot-draw.study.effect.Rd | 16 + phackR/man/dot-fisherz_to_r.Rd | 14 + phackR/man/dot-multDVhack.Rd | 2 +- phackR/man/dot-multIVhack_reg.Rd | 2 +- phackR/man/dot-multIVhack_ttest.Rd | 2 +- phackR/man/dot-onesided_from_twosided.Rd | 18 + phackR/man/dot-optstop.Rd | 2 +- phackR/man/dot-r_to_fisherz.Rd | 14 + phackR/man/dot-report.association.Rd | 31 ++ phackR/man/dot-report.group_lm.Rd | 20 + phackR/man/dot-report.multicat.Rd | 18 + phackR/man/dot-report.twogroup.Rd | 31 ++ phackR/man/dot-roundhack.Rd | 2 +- phackR/man/dot-selectanalysis.Rd | 18 + phackR/man/dot-sim.association.Rd | 28 ++ phackR/man/dot-sim.compscore.Rd | 6 +- phackR/man/dot-sim.covariates.Rd | 17 +- phackR/man/dot-sim.data.Rd | 8 +- phackR/man/dot-sim.multDV.Rd | 6 +- phackR/man/dot-sim.multIV.Rd | 13 +- phackR/man/dot-sim.multregression.Rd | 24 + phackR/man/dot-sim.subgroup.Rd | 6 +- phackR/man/dot-statAnalysisHack.Rd | 2 +- phackR/man/dot-subgroupHack.Rd | 2 +- phackR/man/sim.compscoreHack.Rd | 12 +- phackR/man/sim.covhack.Rd | 12 +- phackR/man/sim.cutoffHack.Rd | 12 +- phackR/man/sim.impHack.Rd | 8 +- phackR/man/sim.multDVhack.Rd | 10 +- phackR/man/sim.multIVhack.Rd | 10 +- phackR/man/sim.optstop.Rd | 12 +- phackR/man/sim.outHack.Rd | 8 +- phackR/man/sim.roundhack.Rd | 10 +- phackR/man/sim.statAnalysisHack.Rd | 10 +- phackR/man/sim.subgroupHack.Rd | 14 +- phackR/man/sim.varTransHack.Rd | 12 +- phackR/tests/testthat/test-simfunctions.R | 353 +++++++++++++- phackR/vignettes/phackR_vignette.Rmd | 73 +-- 77 files changed, 1991 insertions(+), 966 deletions(-) create mode 100644 phackR/man/dot-combine.phase1.results.Rd create mode 100644 phackR/man/dot-compCohensDData.Rd create mode 100644 phackR/man/dot-compCohensDSE.Rd create mode 100644 phackR/man/dot-compCohensDStat.Rd create mode 100644 phackR/man/dot-compFisherZSE.Rd create mode 100644 phackR/man/dot-draw.study.effect.Rd create mode 100644 phackR/man/dot-fisherz_to_r.Rd create mode 100644 phackR/man/dot-onesided_from_twosided.Rd create mode 100644 phackR/man/dot-r_to_fisherz.Rd create mode 100644 phackR/man/dot-report.association.Rd create mode 100644 phackR/man/dot-report.group_lm.Rd create mode 100644 phackR/man/dot-report.multicat.Rd create mode 100644 phackR/man/dot-report.twogroup.Rd create mode 100644 phackR/man/dot-selectanalysis.Rd create mode 100644 phackR/man/dot-sim.association.Rd create mode 100644 phackR/man/dot-sim.multregression.Rd diff --git a/phackR/DESCRIPTION b/phackR/DESCRIPTION index 4a10d8d..71447fa 100644 --- a/phackR/DESCRIPTION +++ b/phackR/DESCRIPTION @@ -12,5 +12,5 @@ Suggests: testthat, knitr, rmarkdown -RoxygenNote: 7.2.1 +RoxygenNote: 7.3.3 VignetteBuilder: knitr diff --git a/phackR/NAMESPACE b/phackR/NAMESPACE index 30abd96..59e4b43 100644 --- a/phackR/NAMESPACE +++ b/phackR/NAMESPACE @@ -17,10 +17,7 @@ importFrom(R.devices,suppressGraphics) importFrom(WRS2,yuen) importFrom(aplpack,stem.leaf) importFrom(car,Anova) -importFrom(dplyr,"%>%") importFrom(dplyr,all_of) -importFrom(dplyr,do) -importFrom(dplyr,group_by_at) importFrom(dplyr,mutate) importFrom(ggplot2,aes) importFrom(ggplot2,annotation_custom) @@ -46,6 +43,7 @@ importFrom(grid,gpar) importFrom(grid,grobTree) importFrom(grid,textGrob) importFrom(magrittr,"%$%") +importFrom(magrittr,"%>%") importFrom(mice,complete) importFrom(mvoutlier,uni.plot) importFrom(pbapply,pblapply) @@ -73,4 +71,3 @@ importFrom(stats,sd) importFrom(stats,t.test) importFrom(stats,wilcox.test) importFrom(utils,capture.output) -importFrom(utils,tail) diff --git a/phackR/R/combinedStrategies_ttest.R b/phackR/R/combinedStrategies_ttest.R index a4add34..6111522 100644 --- a/phackR/R/combinedStrategies_ttest.R +++ b/phackR/R/combinedStrategies_ttest.R @@ -48,15 +48,18 @@ .combined.t.hack <- function(df, roundinglevel = 0.051, alternative = "two.sided", strategy = "firstsig", alpha = 0.05){ ####################### (1) Original p-value ################### + + control.orig <- df[df$group == 1, "DV1"] + treatment.orig <- df[df$group == 2, "DV1"] # Original p-value and effect sizes - test.orig <- stats::t.test(DV1 ~ group, - data = df, - var.equal = TRUE, - alternative = alternative) - p.orig <- test.orig$p.value - r2.orig <- .compR2t(df[df$group == 1, "DV1"], df[df$group == 2, "DV1"]) - d.orig <- .compCohensD(unname(test.orig$statistic), nrow(df)/2) + report.orig <- .report.twogroup(control = control.orig, + treatment = treatment.orig, + method = "t.equal", + alternative = alternative) + p.orig <- report.orig[["p"]] + r2.orig <- .compR2t(control.orig, treatment.orig) + d.orig <- report.orig[["effect"]] # If original p-value is significant stop and return original p-value if(p.orig <= alpha) return(list(p.final = p.orig, @@ -80,22 +83,26 @@ ########### (2) Exploit statistical analysis options ################# # Welch test - p.welch <- stats::t.test(DV1 ~ group, - data = df, - var.equal = FALSE, - alternative = alternative)$p.value + p.welch <- .report.twogroup(control = control.orig, + treatment = treatment.orig, + method = "t.welch", + alternative = alternative)[["p"]] # Mann-Whitney / Wilcoxon test - p.wilcox <- stats::wilcox.test(DV1 ~ group, - alternative = alternative, - data = df)$p.value + p.wilcox <- .report.twogroup(control = control.orig, + treatment = treatment.orig, + method = "wilcox", + alternative = alternative)[["p"]] # Yuen test with different levels of trimming p.yuen <- rep(NA, 4) trim <- c(0.1, 0.15, 0.2, 0.25) for(i in 1:4) { - p.yuen[i] <- WRS2::yuen(DV1 ~ group, tr = trim[i], - data = df)$p.value + p.yuen[i] <- .report.twogroup(control = control.orig, + treatment = treatment.orig, + method = "yuen", + trim = trim[i], + alternative = alternative)[["p"]] } ps <- c(p.orig, p.welch, p.wilcox, p.yuen) @@ -120,19 +127,20 @@ r2s <- NULL for(i in 1:sum(grepl("DV", colnames(df)))){ - - mod[[i]] <- stats::t.test(df[,paste0("DV", i)] ~ df$group, - var.equal = TRUE, - alternative = alternative) + control.current <- df[df$group == 1, paste0("DV", i)] + treatment.current <- df[df$group == 2, paste0("DV", i)] + mod[[i]] <- .report.twogroup(control = control.current, + treatment = treatment.current, + method = "t.equal", + alternative = alternative) - r2s[i] <- .compR2t(df[df$group == 1, paste0("DV", i)], - df[df$group == 2, paste0("DV", i)]) + r2s[i] <- .compR2t(control.current, treatment.current) } - ps <- unlist(simplify2array(mod)["p.value", ]) + ps <- vapply(mod, function(x) x[["p"]], numeric(1)) - ds <- .compCohensD(unlist(simplify2array(mod)["statistic", ]), nrow(df)/2) + ds <- vapply(mod, function(x) x[["effect"]], numeric(1)) # Select final p-hacked p-value based on strategy p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = roundinglevel) @@ -208,16 +216,24 @@ # Compute t-test for each subgroup of the subgroup variables for(i in 1:length(whichsub)){ - - tmp <- dplyr::group_by_at(df, whichsub[i]) %>% - dplyr::do(as.data.frame(stats::t.test(.data$DV1 ~ .data$group, var.equal = TRUE, alternative = alternative)[c("p.value", "statistic")])) - tmp2 <- dplyr::group_by_at(df, whichsub[i]) %>% - dplyr::do(as.data.frame(table(.data$group))) - tmp3 <- dplyr::group_by_at(df, whichsub[i]) %>% do(as.data.frame(.compR2t(.data$DV1[.data$group == unique(.data$group)[1]], .data$DV1[.data$group == unique(.data$group)[2]]))) - - ps[[i]] <- tmp[[2]] - ds[[i]] <- c(tmp[[3]][1]*sqrt(sum(1/tmp2[[3]][1:2])), tmp[[3]][2]*sqrt(sum(1/tmp2[[3]][3:4]))) - r2s[[i]] <- tmp3[[2]] + subgroup.values <- sort(unique(df[, whichsub[i]])) + reports <- vector("list", length(subgroup.values)) + r2.current <- rep(NA_real_, length(subgroup.values)) + + for(j in 1:length(subgroup.values)){ + subset.df <- df[df[, whichsub[i]] == subgroup.values[j], , drop = FALSE] + control.sub <- subset.df[subset.df$group == 1, "DV1"] + treatment.sub <- subset.df[subset.df$group == 2, "DV1"] + reports[[j]] <- .report.twogroup(control = control.sub, + treatment = treatment.sub, + method = "t.equal", + alternative = alternative) + r2.current[j] <- .compR2t(control.sub, treatment.sub) + } + + ps[[i]] <- vapply(reports, function(x) x[["p"]], numeric(1)) + ds[[i]] <- vapply(reports, function(x) x[["effect"]], numeric(1)) + r2s[[i]] <- r2.current } diff --git a/phackR/R/compositeScores.R b/phackR/R/compositeScores.R index 7b767e8..5468cfe 100644 --- a/phackR/R/compositeScores.R +++ b/phackR/R/compositeScores.R @@ -6,12 +6,16 @@ #' @param nobs Integer giving number of observations #' @param ncompv Integer giving number of variables to build the composite score #' @param rcomp Correlation between the composite score variables +#' @param effect Mean effect size across studies on the Fisher-z scale +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale -.sim.compscore <- function(nobs, ncompv, rcomp){ - - dv <- rnorm(nobs, 0, 1) +.sim.compscore <- function(nobs, ncompv, rcomp, effect = 0, heterogeneity = 0){ iv <- .sim.multcor(nobs = nobs, nvar = ncompv, r = rcomp) + theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) + rho <- .fisherz_to_r(theta) + compscore <- scale(rowMeans(iv))[,1] + dv <- rho*compscore + sqrt(1-rho^2)*rnorm(nobs, 0, 1) res <- cbind(dv, iv) @@ -34,13 +38,12 @@ stopifnot(length(compv)-ndelete >= 2) # Compute original p-value and R^2 with full scale - modres <- summary(lm(df[, dv] ~ rowMeans(df[, compv]))) - p.orig <- modres$coefficients[2, 4] - r2.orig <- modres$r.squared + fullscale <- rowMeans(df[, compv, drop = FALSE]) + report.orig <- .report.association(x = fullscale, y = df[, dv], method = "scale.full") + analyses <- list(list(report = report.orig, + r2 = tanh(report.orig[["effect"]])^2)) # Prepare and initialize variables for p-hacking - ps <- list() - r2s <- list() compscale <- df[, compv] changescale <- df[, compv] out <- NULL @@ -55,49 +58,45 @@ out[i] <- which(colnames(compscale) %in% colnames(changescale)[which.max(performance::item_reliability(changescale)[,2])]) # Compute p-value for the new composite score - newscore <- rowMeans(compscale[, -out]) - newmodres <- summary(lm(df[, dv] ~ newscore)) - pval[1] <- newmodres$coefficients[2, 4] - r2val[1] <- newmodres$r.squared + newscore <- rowMeans(compscale[, -out, drop = FALSE]) + report.new <- .report.association(x = newscore, y = df[, dv], + method = paste0("scale.delete.", paste(out, collapse = "-"))) + analyses[[length(analyses)+1]] <- list(report = report.new, + r2 = tanh(report.new[["effect"]])^2) # Compute p-value for the item deleted from the score itemscore <- compscale[, out[i]] - newmodres2 <- summary(lm(df[, dv] ~ itemscore)) - pval[2] <- newmodres2$coefficients[2, 4] - r2val[2] <- newmodres2$r.squared - - # Compute p-value for a scale of all items deleted so far - #nonscore <- rowMeans(cbind(compscale[, out])) - #newmodres3 <- summary(lm(df[, dv] ~ nonscore)) - #pval[3] <- newmodres3$coefficients[2, 4] - #r2val[3] <- newmodres3$r.squared - - changescale <- compscale[, -out] - ps[[i]] <- pval - r2s[[i]] <- r2val + report.item <- .report.association(x = itemscore, y = df[, dv], + method = paste0("item.", out[i])) + analyses[[length(analyses)+1]] <- list(report = report.item, + r2 = tanh(report.item[["effect"]])^2) + + changescale <- compscale[, -out, drop = FALSE] } - ps <- c(p.orig, unique(unlist(ps))) - r2s <- c(r2.orig, unique(unlist(r2s))) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-hacking with composite scores -#' Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs Integer giving number of observations #' @param ncompv Integer giving number of variables to build the composite score #' @param rcomp Correlation between the composite score variables #' @param ndelete How many items should be deleted from the scale at maximum? #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies on the Fisher-z scale +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale #' @param alpha Significance level of the t-test (default: 0.05) #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE @@ -105,12 +104,13 @@ #' @importFrom shiny withProgress incProgress #' @export -sim.compscoreHack <- function(nobs, ncompv, rcomp, ndelete, strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv=FALSE){ +sim.compscoreHack <- function(nobs, ncompv, rcomp, ndelete, strategy = "firstsig", effect = 0, heterogeneity = 0, alpha = 0.05, iter = 1000, shinyEnv=FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.compscore(nobs = nobs, ncompv = ncompv, rcomp = rcomp) + dat[[i]] <- .sim.compscore(nobs = nobs, ncompv = ncompv, rcomp = rcomp, + effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset (with progress bar within or outside Shiny) @@ -135,20 +135,7 @@ sim.compscoreHack <- function(nobs, ncompv, rcomp, ndelete, strategy = "firstsig }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.orig <- NULL - r2s.hack <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig")) } diff --git a/phackR/R/exploitCovariates.R b/phackR/R/exploitCovariates.R index 87c33a8..86ab7e6 100644 --- a/phackR/R/exploitCovariates.R +++ b/phackR/R/exploitCovariates.R @@ -8,12 +8,14 @@ #' @param ncov Number of continuous covariates in the simulated data frame #' @param rcov Correlation between the covariates #' @param rcovdv Correlation between covariates and dependent variable +#' @param effect Mean effect size across studies on the standardized mean-difference scale +#' @param heterogeneity Between-study heterogeneity on the standardized mean-difference scale #' @param mu Mean of the random data #' @param sd Standard deviation of the random data #' @param missing Proportion of missing values per variable (e.g., 0.2 = 20 percent) #' @importFrom stats rnorm -.sim.covariates <- function(nobs.group, ncov, rcov, rcovdv, mu = 0, sd = 1, missing = 0){ +.sim.covariates <- function(nobs.group, ncov, rcov, rcovdv, effect = 0, heterogeneity = 0, mu = 0, sd = 1, missing = 0, theta = NULL){ # Observations per group and total observations if(length(nobs.group) == 1) nobs.group <- rep(nobs.group, 2) @@ -37,6 +39,8 @@ # create raw data from matrix multiplication of U and random noise X <- as.data.frame(t(U %*% random.normal)) + if(is.null(theta)) theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) + X[group == 2, 1] <- X[group == 2, 1] + theta # create final simulated data matrix Xfull <- cbind(group, X) @@ -73,8 +77,7 @@ colnames(df)[covs] <- paste0("CV", 1:length(covs)) df <- df[, c(dv, group, covs)] - ps <- NULL - eta2s <- NULL # partial eta^2 + analyses <- list() # Compute correlations between covariates and dependent variable and order covariates accordingly dvcors <- apply(X = df[,-group], MARGIN = 2, FUN = function(x) stats::cor(x, df$dv))[-1] @@ -105,43 +108,54 @@ res <- stats::aov(stats::as.formula(models[i]), data = df) resanc <- car::Anova(res, type = 2) - ps[i] <- resanc["group", "Pr(>F)"] - eta2s[i] <- resanc["group", "Sum Sq"]/(resanc["group", "Sum Sq"] + resanc["Residuals", "Sum Sq"]) + report <- .report.group_lm(formula = stats::as.formula(models[i]), + data = df, + groupvar = "group", + method = paste0("ancova.", gsub(" ", "", models[i]))) + analyses[[i]] <- list(report = report, + eta2 = resanc["group", "Sum Sq"]/(resanc["group", "Sum Sq"] + resanc["Residuals", "Sum Sq"])) } + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) + # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - eta2.final <- unique(eta2s[ps == p.final]) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - eta2.final = eta2.final, - eta2s = eta2s)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + eta2s.hack = analyses[[final.index]][["eta2"]], + eta2s.orig = analyses[[1]][["eta2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-Hacking with multiple covariates -#' Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs.group Vector with number of observations per group #' @param ncov Number of continuous covariates in the simulated data frame #' @param rcov Correlation between the covariates #' @param rcovdv Correlation between covariates and dependent variable #' @param interactions Should interaction terms be added to the ANCOVA models? TRUE/FALSE #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies on the standardized mean-difference scale +#' @param heterogeneity Between-study heterogeneity on the standardized mean-difference scale #' @param alpha Significance level of the t-test #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.covhack <- function(nobs.group, ncov, rcov, rcovdv, interactions = FALSE, strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.covhack <- function(nobs.group, ncov, rcov, rcovdv, interactions = FALSE, strategy = "firstsig", effect = 0, heterogeneity = 0, alpha = 0.05, iter = 1000, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.covariates(nobs.group = nobs.group, ncov = ncov, rcov = rcov, rcovdv = rcovdv) + dat[[i]] <- .sim.covariates(nobs.group = nobs.group, ncov = ncov, rcov = rcov, + rcovdv = rcovdv, effect = effect, + heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -164,21 +178,8 @@ sim.covhack <- function(nobs.group, ncov, rcov, rcovdv, interactions = FALSE, st }) } - ps.hack <- NULL - ps.orig <- NULL - eta2s.hack <- NULL - eta2s.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - eta2s.hack[i] <- res[[i]][["eta2.final"]] - eta2s.orig[i] <- res[[i]][["eta2s"]][1] - } - - res <- cbind(ps.hack, ps.orig, eta2s.hack, eta2s.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "eta2s.hack", "eta2s.orig")) } diff --git a/phackR/R/exploitCutoffs.R b/phackR/R/exploitCutoffs.R index f639e13..29199db 100644 --- a/phackR/R/exploitCutoffs.R +++ b/phackR/R/exploitCutoffs.R @@ -16,59 +16,66 @@ iv <- df[, iv] dv <- df[, dv] + analyses <- list() - mod.orig <- summary(stats::lm(dv ~ iv)) - p.orig <- mod.orig$coefficients[2, 4] - r2.orig <- mod.orig$r.squared + report.orig <- .report.association(x = iv, y = dv, method = "lm.continuous") + analyses[[1]] <- list(report = report.orig, + r2 = tanh(report.orig[["effect"]])^2) # Do the mediansplit mediansplitvar <- as.numeric(iv > stats::median(iv)) + 1 - p.mediansplit <- stats::t.test(dv[mediansplitvar == 1], dv[mediansplitvar == 2], - var.equal = TRUE, alternative = "two.sided")$p.value - r2.mediansplit <- .compR2t(dv[mediansplitvar == 1], dv[mediansplitvar == 2]) + report.mediansplit <- .report.association(x = mediansplitvar, y = dv, + method = "lm.mediansplit") + analyses[[2]] <- list(report = report.mediansplit, + r2 = .compR2t(dv[mediansplitvar == 1], dv[mediansplitvar == 2])) # Cut the middle tertiles <- as.numeric(stats::quantile(iv, probs = c(1/3, 2/3))) threecut <- cut(iv, breaks = c(-Inf, tertiles, Inf), labels = c(1,0,2)) dv2 <- dv[threecut %in% c(1,2)] threecut2 <- threecut[threecut %in% c(1, 2)] - p.cutmiddle <- stats::t.test(dv2[threecut2 == 2], dv2[threecut2 == 1], - var.equal = TRUE, alternative = "two.sided")$p.value - r2.cutmiddle <- .compR2t(dv2[threecut2 == 2], dv2[threecut2 == 1]) + report.cutmiddle <- .report.association(x = as.numeric(as.character(threecut2)), + y = dv2, + method = "lm.cutmiddle") + analyses[[3]] <- list(report = report.cutmiddle, + r2 = .compR2t(dv2[threecut2 == 2], dv2[threecut2 == 1])) # 3 Categories: Omnibus test - mod.threecat <- summary(stats::aov(dv ~ threecut)) - p.threecat <- mod.threecat[[1]][[5]][1] - r2.threecat <- mod.threecat[[1]][1,2]/sum(mod.threecat[[1]][,2]) + report.threecat <- .report.multicat(group = threecut, y = dv, + method = "anova.threecut") + analyses[[4]] <- list(report = report.threecat, + r2 = tanh(report.threecat[["effect"]])^2) - ps <- c(p.orig, p.mediansplit, p.cutmiddle, p.threecat) - r2s <- c(r2.orig, r2.mediansplit, r2.cutmiddle, r2.threecat) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-Hacking for exploiting cutoff values -#' Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs Number of observations #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies on the Fisher-z scale +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale #' @param alpha Significance level of the t-test #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.cutoffHack <- function(nobs, strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.cutoffHack <- function(nobs, strategy = "firstsig", effect = 0, heterogeneity = 0, alpha = 0.05, iter = 1000, shinyEnv = FALSE){ dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.multcor(nobs = nobs, nvar = 2, r = 0) + dat[[i]] <- .sim.association(nobs = nobs, effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -92,22 +99,7 @@ sim.cutoffHack <- function(nobs, strategy = "firstsig", alpha = 0.05, iter = 100 }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - - return(res) - - + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig")) } diff --git a/phackR/R/favorableImputation.R b/phackR/R/favorableImputation.R index 50d7ae6..663293a 100644 --- a/phackR/R/favorableImputation.R +++ b/phackR/R/favorableImputation.R @@ -54,124 +54,130 @@ # Stop if imputation methods are not defined stopifnot(any(c(1:10) %in% which)) - # Initialize result vector - ps <- rep(NA, 10) - r2s <- rep(NA, 10) + analyses <- list() # p-value when missing values are deleted if(1 %in% which){ - mod1 <- summary(stats::lm(y ~ x, na.action = "na.omit")) - ps[1] <- mod1$coefficients[2, 4] - r2s[1] <- mod1$r.squared + report <- .report.association(x = x, y = y, method = "delete.missing") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Mean imputation if(2 %in% which){ newx <- .easyimpute(x, mean, na.rm = T) newy <- .easyimpute(y, mean, na.rm = T) - mod2 <- summary(stats::lm(newy ~ newx)) - ps[2] <- mod2$coefficients[2, 4] - r2s[2] <- mod2$r.squared + report <- .report.association(x = newx, y = newy, method = "impute.mean") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Median imputation if(3 %in% which){ - newx <- .easyimpute(x, mean, na.rm = T) - newy <- .easyimpute(y, mean, na.rm = T) - mod3 <- summary(stats::lm(newy ~ newx)) - ps[3] <- mod3$coefficients[2, 4] - r2s[3] <- mod3$r.squared + newx <- .easyimpute(x, median, na.rm = T) + newy <- .easyimpute(y, median, na.rm = T) + report <- .report.association(x = newx, y = newy, method = "impute.median") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Mode imputation if(4 %in% which){ newx <- .easyimpute(x, .estimate_mode) newy <- .easyimpute(y, .estimate_mode) - mod4 <- summary(stats::lm(newy ~ newx)) - ps[4] <- mod4$coefficients[2, 4] - r2s[4] <- mod4$r.squared + report <- .report.association(x = newx, y = newy, method = "impute.mode") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Multivariate imputations by chained equations ("mice" package): predictive mean matchihng dfnew <- as.data.frame(cbind(x, y)) if(5 %in% which){ imp <- mice::mice(dfnew, m = 1, method = "pmm", silent = TRUE, print = FALSE) - mod5 <- summary(stats::lm(y ~ x, data = mice::complete(imp, 1))) - ps[5] <- mod5$coefficients[2, 4] - r2s[5] <- mod5$r.squared + dat5 <- mice::complete(imp, 1) + report <- .report.association(x = dat5$x, y = dat5$y, method = "mice.pmm") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Multivariate imputations by chained equations ("mice" package): Weighted predictive mean matching if(6 %in% which){ imp <- mice::mice(dfnew, m = 1, method = "midastouch", silent = TRUE, print = FALSE) - mod6 <- summary(stats::lm(y ~ x, data = mice::complete(imp, 1))) - ps[6] <- mod6$coefficients[2, 4] - r2s[6] <- mod6$r.squared + dat6 <- mice::complete(imp, 1) + report <- .report.association(x = dat6$x, y = dat6$y, method = "mice.midastouch") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Multivariate imputations by chained equations ("mice" package): Sample from observed values if(7 %in% which){ imp <- mice::mice(dfnew, m = 1, method = "sample", silent = TRUE, print = FALSE) - mod7 <- summary(stats::lm(y ~ x, data = mice::complete(imp, 1))) - ps[7] <- mod7$coefficients[2, 4] - r2s[7] <- mod7$r.squared + dat7 <- mice::complete(imp, 1) + report <- .report.association(x = dat7$x, y = dat7$y, method = "mice.sample") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Multivariate imputations by chained equations ("mice" package): Bayesian linear regression if(8 %in% which){ imp <- mice::mice(dfnew, m = 1, method = "norm", silent = TRUE, print = FALSE) - mod8 <- summary(stats::lm(y ~ x, data = mice::complete(imp, 1))) - ps[8] <- mod8$coefficients[2, 4] - r2s[8] <- mod8$r.squared + dat8 <- mice::complete(imp, 1) + report <- .report.association(x = dat8$x, y = dat8$y, method = "mice.norm") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Multivariate imputations by chained equations ("mice" package): Linear regression ignoring model error if(9 %in% which){ imp <- mice::mice(dfnew, m = 1, method = "norm.nob", silent = TRUE, print = FALSE) - mod9 <- summary(stats::lm(y ~ x, data = mice::complete(imp, 1))) - ps[9] <- mod9$coefficients[2, 4] - r2s[9] <- mod9$r.squared + dat9 <- mice::complete(imp, 1) + report <- .report.association(x = dat9$x, y = dat9$y, method = "mice.norm.nob") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } # Multivariate imputations by chained equations ("mice" package): Linear regression predicted values if(10 %in% which){ imp <- mice::mice(dfnew, m = 1, method = "norm.predict", silent = TRUE, print = FALSE) - mod10 <- summary(stats::lm(y ~ x, data = mice::complete(imp, 1))) - ps[10] <- mod10$coefficients[2, 4] - r2s[10] <- mod10$r.squared + dat10 <- mice::complete(imp, 1) + report <- .report.association(x = dat10$x, y = dat10$y, method = "mice.norm.predict") + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } - ps <- ps[!is.na(ps)] - r2s <- r2s[!is.na(r2s)] + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-Hacking with different sorts of outlier definition missing value imputation -#' @description Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs Integer giving number of observations #' @param missing Percentage of missing values (e.g., 0.1 for 10 percent) #' @param which Which imputation methods? Either 5 random methods are chosen ("random") or a numeric vector containing the chosen methods (1: delete missing, 2: mean imputation, 3: median imputation, 4: mode imputation, 5: predictive mean matching, 6: weighted predictive mean matching, 7: sample from observed values, 8: Bayesian linear regression, 9: linear regression ignoring model error, 10: linear regression predicted values) #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies on the Fisher-z scale +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale #' @param alpha Significance level of the t-test (default: 0.05) #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.impHack <- function(nobs, missing, which = c(1:10), strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.impHack <- function(nobs, missing, which = c(1:10), strategy = "firstsig", effect = 0, heterogeneity = 0, alpha = 0.05, iter = 1000, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.multcor(nobs = nobs, nvar = 2, r = 0, missing = missing) + dat[[i]] <- .sim.association(nobs = nobs, effect = effect, heterogeneity = heterogeneity, missing = missing) } if(any(which == "random")) which <- sample(c(1:10), 5) @@ -199,21 +205,8 @@ sim.impHack <- function(nobs, missing, which = c(1:10), strategy = "firstsig", a }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - ps.all <- list() - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig")) } diff --git a/phackR/R/helpers.R b/phackR/R/helpers.R index 84bb9d9..733c63a 100644 --- a/phackR/R/helpers.R +++ b/phackR/R/helpers.R @@ -43,16 +43,38 @@ } +#' Draw one study-level effect size +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity + +.draw.study.effect <- function(effect = 0, heterogeneity = 0){ + + stopifnot(length(effect) == 1) + stopifnot(length(heterogeneity) == 1) + stopifnot(heterogeneity >= 0) + + if(heterogeneity == 0){ + return(effect) + } + + stats::rnorm(1, mean = effect, sd = heterogeneity) + +} + #' Generic sampling function #' @description Outputs a data frame with two columns #' @param nobs.group Number of observations per group. Either a scalar or a vector with two elements. +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity +#' @param theta Study-specific true effect size #' @importFrom stats rnorm -.sim.data <- function(nobs.group){ +.sim.data <- function(nobs.group, effect = 0, heterogeneity = 0, theta = NULL){ if(length(nobs.group) == 1) nobs.group <- rep(nobs.group, 2) + if(is.null(theta)) theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) V1 <- stats::rnorm(nobs.group[1], 0, 1) - V2 <- stats::rnorm(nobs.group[2], 0, 1) + V2 <- stats::rnorm(nobs.group[2], theta, 1) group <- c(rep(1, nobs.group[1]), rep(2, nobs.group[2])) res <- cbind(group, c(V1, V2)) @@ -108,6 +130,40 @@ } +#' Select a p-hacked analysis from a vector of p-values +#' @description Takes a vector of p-values and returns the index of the selected analysis. +#' @param ps Vector of p values +#' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param alpha Significance level (default: 0.05) + +.selectanalysis <- function(ps, strategy, alpha){ + + ps.clean <- ps + ps.clean[!is.finite(ps.clean)] <- Inf + + selected <- 1 + + if(strategy == "smallest.sig"){ + sig <- which(ps.clean < alpha) + if(length(sig) > 0){ + selected <- sig[which.min(ps.clean[sig])] + } + + } else if(strategy == "firstsig") { + + sig <- which(ps.clean < alpha) + if(length(sig) > 0){ + selected <- sig[1] + } + + } else if(strategy == "smallest") { + selected <- which.min(ps.clean) + } + + return(selected) + +} + #' Select a p-value from a vector of p-hacked p-values #' @description Takes a vector of p-values and selects the smallest, first significant, or smallest significant p-value. #' @param ps Vector of p values @@ -115,34 +171,83 @@ #' @param alpha Significance level (default: 0.05) .selectpvalue <- function(ps, strategy, alpha){ + ps[.selectanalysis(ps = ps, strategy = strategy, alpha = alpha)] - p.final <- NA - p.orig <- ps[1] +} - # Select smallest significant p-value - if(strategy == "smallest.sig"){ +#' Convert a Fisher-z value to a correlation +#' @param effect Fisher-z effect size - if(min(ps) < alpha){ - p.final <- min(ps) - } else { - p.final <- p.orig - } +.fisherz_to_r <- function(effect){ + pmax(pmin(tanh(effect), 0.999999), -0.999999) +} - # Select first significant p-value - } else if (strategy == "firstsig") { +#' Convert a correlation to Fisher-z +#' @param r Correlation - if(min(ps) < alpha){ - p.final <- ps[which(ps < alpha)[1]] - } else { - p.final <- p.orig - } +.r_to_fisherz <- function(r){ + atanh(pmax(pmin(r, 0.999999), -0.999999)) +} + +#' Compute the standard error of Fisher-z +#' @param n Sample size + +.compFisherZSE <- function(n){ + + if(n <= 3){ + return(NA_real_) + } + + 1/sqrt(n-3) + +} + +#' Simulate two continuous variables with a study-level association +#' @param nobs Number of observations +#' @param effect Mean Fisher-z effect size across studies +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale +#' @param theta Study-specific true effect size +#' @param missing Proportion of missing values per variable + +.sim.association <- function(nobs, effect = 0, heterogeneity = 0, theta = NULL, missing = 0){ + + if(is.null(theta)) theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) + + .sim.multcor(nobs = nobs, nvar = 2, r = .fisherz_to_r(theta), missing = missing) + +} + +#' Simulate one criterion and multiple predictors with shared study-level association +#' @param nobs Number of observations +#' @param nvar Number of predictor variables +#' @param r Correlation between predictor variables +#' @param effect Mean Fisher-z effect size across studies +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale +#' @param theta Study-specific true effect size +#' @importFrom stats rnorm - # Select smallest p-value - } else if (strategy == "smallest") { - p.final <- min(ps) +.sim.multregression <- function(nobs, nvar, r, effect = 0, heterogeneity = 0, theta = NULL){ + + if(length(nobs) > 1) nobs <- nobs[1] + if(is.null(theta)) theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) + + rho <- .fisherz_to_r(theta) + R <- matrix(rep(r, (nvar+1)^2), nrow = nvar+1) + diag(R) <- rep(1, nvar+1) + R[1, -1] <- rep(rho, nvar) + R[-1, 1] <- R[1, -1] + + cholR <- tryCatch(t(chol(R)), error = function(e) NULL) + if(is.null(cholR)){ + R <- as.matrix(Matrix::nearPD(R, corr = TRUE)$mat) + cholR <- t(chol(R)) } - return(p.final) + random.normal <- matrix(stats::rnorm((nvar+1)*nobs), nrow = nvar+1, ncol = nobs) + X <- as.data.frame(t(cholR %*% random.normal)) + colnames(X)[1] <- "criterion" + + return(X) } @@ -151,6 +256,8 @@ #' @param y values of group 2 .compR2t <- function(x, y){ + x <- x[!is.na(x)] + y <- y[!is.na(y)] grandmean <- mean(c(x, y)) sst <- sum((c(x,y)-grandmean)^2) sse <- sum((x-mean(x))^2)+sum((y-mean(y))^2) @@ -165,3 +272,308 @@ .compCohensD <- function(t, n){ t*sqrt(2/n) } + +#' Compute Cohen's d from a test statistic and group sizes +#' @param statistic Test statistic +#' @param n1 Sample size in group 1 +#' @param n2 Sample size in group 2 + +.compCohensDStat <- function(statistic, n1, n2){ + statistic*sqrt((1/n1) + (1/n2)) +} + +#' Compute Cohen's d from the observed data +#' @param control values of the control group +#' @param treatment values of the treatment group + +.compCohensDData <- function(control, treatment){ + + control <- control[!is.na(control)] + treatment <- treatment[!is.na(treatment)] + + n1 <- length(control) + n2 <- length(treatment) + + if(n1 < 2 || n2 < 2){ + return(NA_real_) + } + + sp <- sqrt((((n1-1)*stats::var(control)) + ((n2-1)*stats::var(treatment)))/(n1+n2-2)) + + if(!is.finite(sp) || sp == 0){ + return(NA_real_) + } + + (mean(treatment)-mean(control))/sp + +} + +#' Compute the standard error of Cohen's d +#' @param d Cohen's d +#' @param n1 Sample size in group 1 +#' @param n2 Sample size in group 2 + +.compCohensDSE <- function(d, n1, n2){ + + if(!is.finite(d) || n1 < 2 || n2 < 2){ + return(NA_real_) + } + + sqrt(((n1+n2)/(n1*n2)) + ((d^2)/(2*(n1+n2-2)))) + +} + +#' Convert a two-sided p-value to a one-sided p-value +#' @param p.twosided Two-sided p-value +#' @param stat Test statistic +#' @param alternative Direction of the test + +.onesided_from_twosided <- function(p.twosided, stat, alternative = "two.sided"){ + + if(!is.finite(p.twosided) || !is.finite(stat)){ + return(NA_real_) + } + + if(alternative == "two.sided"){ + return(p.twosided) + } + + halfp <- p.twosided/2 + + if(alternative == "greater"){ + if(stat >= 0){ + return(halfp) + } + + return(1-halfp) + } + + if(alternative == "less"){ + if(stat <= 0){ + return(halfp) + } + + return(1-halfp) + } + + stop("Unsupported alternative.") + +} + +#' Build one normalized reporting entry for a two-group analysis +#' @param control values of the control group +#' @param treatment values of the treatment group +#' @param method Analysis method +#' @param alternative Direction of the test. For group-comparison analyses, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. +#' @param trim Trimming level for Yuen's test +#' @param p.override Optional p-value override + +.report.twogroup <- function(control, treatment, method = "t.equal", alternative = "two.sided", trim = NULL, p.override = NULL){ + + control <- control[!is.na(control)] + treatment <- treatment[!is.na(treatment)] + + n1 <- length(control) + n2 <- length(treatment) + effect <- .compCohensDData(control = control, treatment = treatment) + se <- .compCohensDSE(d = effect, n1 = n1, n2 = n2) + stat <- NA_real_ + pval <- NA_real_ + method.label <- method + + if(method == "t.equal"){ + mod <- stats::t.test(treatment, control, var.equal = TRUE, alternative = alternative) + stat <- unname(mod$statistic) + pval <- mod$p.value + } else if(method == "t.welch"){ + mod <- stats::t.test(treatment, control, var.equal = FALSE, alternative = alternative) + stat <- unname(mod$statistic) + pval <- mod$p.value + } else if(method == "wilcox"){ + mod <- stats::wilcox.test(treatment, control, alternative = alternative) + stat <- unname(mod$statistic) + pval <- mod$p.value + } else if(method == "yuen"){ + dftest <- data.frame(group = c(rep(1, n1), rep(2, n2)), + dv = c(control, treatment)) + mod <- WRS2::yuen(dv ~ group, tr = trim, data = dftest) + stat <- unname(mod$test) + pval <- .onesided_from_twosided(p.twosided = mod$p.value, + stat = stat, + alternative = alternative) + method.label <- paste0("yuen.", trim) + } else { + stop("Unsupported method.") + } + + if(!is.null(p.override)){ + pval <- p.override + } + + return(list(effect = effect, + se = se, + n = n1+n2, + stat = stat, + p = pval, + method = method.label)) + +} + +#' Build one normalized reporting entry for a simple association analysis +#' @param x Predictor values +#' @param y Criterion values +#' @param method Analysis method +#' @param alternative Direction of the test. For association analyses, \code{"greater"} tests a positive association. +#' @param p.override Optional p-value override +#' @param stat.override Optional statistic override + +.report.association <- function(x, y, method = "lm", alternative = "two.sided", p.override = NULL, stat.override = NULL){ + + dat <- data.frame(x = x, y = y) + dat <- dat[is.finite(dat$x) & is.finite(dat$y), , drop = FALSE] + n <- nrow(dat) + + if(n < 4 || length(unique(dat$x)) < 2 || length(unique(dat$y)) < 2){ + return(list(effect = NA_real_, + se = NA_real_, + n = n, + stat = NA_real_, + p = NA_real_, + method = method)) + } + + fit <- summary(stats::lm(y ~ x, data = dat)) + effect <- .r_to_fisherz(stats::cor(dat$x, dat$y)) + stat <- unname(fit$coefficients[2, 3]) + pval <- .onesided_from_twosided(p.twosided = fit$coefficients[2, 4], + stat = stat, + alternative = alternative) + + if(!is.null(stat.override)) stat <- stat.override + if(!is.null(p.override)) pval <- p.override + + return(list(effect = effect, + se = .compFisherZSE(n = n), + n = n, + stat = stat, + p = pval, + method = method)) + +} + +#' Build one normalized reporting entry for a multi-category analysis +#' @param group Grouping variable +#' @param y Criterion values +#' @param method Analysis method + +.report.multicat <- function(group, y, method = "anova"){ + + dat <- data.frame(group = group, y = y) + dat <- dat[!is.na(dat$group) & is.finite(dat$y), , drop = FALSE] + dat$group <- factor(dat$group) + n <- nrow(dat) + + if(n < 4 || length(unique(dat$group)) < 2){ + return(list(effect = NA_real_, + se = NA_real_, + n = n, + stat = NA_real_, + p = NA_real_, + method = method)) + } + + fit <- summary(stats::lm(y ~ group, data = dat)) + if(length(fit$fstatistic) == 0){ + return(list(effect = NA_real_, + se = NA_real_, + n = n, + stat = NA_real_, + p = NA_real_, + method = method)) + } + + fstat <- unname(fit$fstatistic["value"]) + pval <- stats::pf(fstat, + df1 = unname(fit$fstatistic["numdf"]), + df2 = unname(fit$fstatistic["dendf"]), + lower.tail = FALSE) + rsign <- sign(stats::cor(as.numeric(dat$group), dat$y)) + if(!is.finite(rsign) || rsign == 0) rsign <- 1 + + return(list(effect = .r_to_fisherz(rsign*sqrt(fit$r.squared)), + se = .compFisherZSE(n = n), + n = n, + stat = fstat, + p = pval, + method = method)) + +} + +#' Build one normalized reporting entry for a group effect in a linear model +#' @param formula Model formula +#' @param data Analysis data +#' @param groupvar Group variable name +#' @param method Analysis method + +.report.group_lm <- function(formula, data, groupvar = "group", method = "ancova"){ + + fit <- stats::lm(formula, data = data) + summary.fit <- summary(fit) + model.df <- stats::model.frame(fit) + group <- model.df[[groupvar]] + group <- group[!is.na(group)] + group.levels <- sort(unique(group)) + n1 <- sum(group == group.levels[1]) + n2 <- sum(group == group.levels[2]) + coef.row <- rownames(summary.fit$coefficients) + row.id <- grep(paste0("^", groupvar), coef.row)[1] + + if(length(group.levels) < 2 || is.na(row.id)){ + return(list(effect = NA_real_, + se = NA_real_, + n = nrow(model.df), + stat = NA_real_, + p = NA_real_, + method = method)) + } + + stat <- unname(summary.fit$coefficients[row.id, 3]) + effect <- .compCohensDStat(statistic = stat, n1 = n1, n2 = n2) + + return(list(effect = effect, + se = .compCohensDSE(d = effect, n1 = n1, n2 = n2), + n = nrow(model.df), + stat = stat, + p = summary.fit$coefficients[row.id, 4], + method = method)) + +} + +#' Combine legacy simulation output with the normalized reporting block +#' @param res List of iteration results +#' @param legacy.fields Character vector defining the legacy output columns + +.combine.phase1.results <- function(res, legacy.fields){ + + output <- list() + + for(i in 1:length(legacy.fields)){ + output[[legacy.fields[i]]] <- vapply(res, function(x) x[[legacy.fields[i]]], numeric(1)) + } + + output[["effect.initial"]] <- vapply(res, function(x) x[["report.initial"]][["effect"]], numeric(1)) + output[["effect.final"]] <- vapply(res, function(x) x[["report.final"]][["effect"]], numeric(1)) + output[["se.initial"]] <- vapply(res, function(x) x[["report.initial"]][["se"]], numeric(1)) + output[["se.final"]] <- vapply(res, function(x) x[["report.final"]][["se"]], numeric(1)) + output[["n.initial"]] <- vapply(res, function(x) x[["report.initial"]][["n"]], numeric(1)) + output[["n.final"]] <- vapply(res, function(x) x[["report.final"]][["n"]], numeric(1)) + output[["stat.initial"]] <- vapply(res, function(x) x[["report.initial"]][["stat"]], numeric(1)) + output[["stat.final"]] <- vapply(res, function(x) x[["report.final"]][["stat"]], numeric(1)) + output[["p.initial"]] <- vapply(res, function(x) x[["report.initial"]][["p"]], numeric(1)) + output[["p.final"]] <- vapply(res, function(x) x[["report.final"]][["p"]], numeric(1)) + output[["method.initial"]] <- vapply(res, function(x) x[["report.initial"]][["method"]], character(1)) + output[["method.final"]] <- vapply(res, function(x) x[["report.final"]][["method"]], character(1)) + + return(as.data.frame(output, stringsAsFactors = FALSE)) + +} diff --git a/phackR/R/incorrectRounding.R b/phackR/R/incorrectRounding.R index c40a720..716927d 100644 --- a/phackR/R/incorrectRounding.R +++ b/phackR/R/incorrectRounding.R @@ -10,48 +10,56 @@ #' @param group Scalar defining location of the group vector in the data frame #' @param dv Scalar defining location of dependent variable in the data frame #' @param roundinglevel Highest p-value that is rounded down to 0.05 -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test (default: 0.05) #' @importFrom stats t.test .roundhack <- function(df, group, dv, roundinglevel, alternative = "two.sided", alpha = 0.05){ - # Compute t-test - pval <- stats::t.test(df[,dv] ~ df[,group], - var.equal = TRUE, alternative = alternative)$p.value - r2val <- .compR2t(df[,dv][(df[,group] == unique(df[,group])[1])], - df[,dv][(df[,group] == unique(df[,group])[2])]) + control <- df[,dv][df[,group] == unique(df[,group])[1]] + treatment <- df[,dv][df[,group] == unique(df[,group])[2]] + report.initial <- .report.twogroup(control = control, + treatment = treatment, + method = "t.equal", + alternative = alternative) + r2val <- .compR2t(control, treatment) # P-hack p-value - if(pval > alpha && pval < roundinglevel){ + if(report.initial[["p"]] > alpha && report.initial[["p"]] < roundinglevel){ p.final <- alpha } else { - p.final <- pval + p.final <- report.initial[["p"]] } - ps <- c(pval, p.final) + report.final <- report.initial + report.final[["p"]] <- p.final - return(list(p.final = p.final, - ps = ps, - r2.final = r2val, - r2s = rep(r2val, 2))) + return(list(ps.hack = p.final, + ps.orig = report.initial[["p"]], + r2s.hack = r2val, + r2s.orig = r2val, + report.initial = report.initial, + report.final = report.final)) } #' Simulate p-hacking with incorrect rounding +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param roundinglevel Highest p-value that is rounded down to alpha +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity #' @param iter Number of iterations -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test (default: 0.05) #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.roundhack <- function(roundinglevel, iter = 1000, alternative = "two.sided", alpha = 0.05, shinyEnv = FALSE){ +sim.roundhack <- function(roundinglevel, effect = 0, heterogeneity = 0, iter = 1000, alternative = "two.sided", alpha = 0.05, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.data(nobs.group = 30) + dat[[i]] <- .sim.data(nobs.group = 30, effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -71,20 +79,7 @@ sim.roundhack <- function(roundinglevel, iter = 1000, alternative = "two.sided", }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig")) } diff --git a/phackR/R/optionalStopping.R b/phackR/R/optionalStopping.R index 59cfbe2..1889531 100644 --- a/phackR/R/optionalStopping.R +++ b/phackR/R/optionalStopping.R @@ -13,10 +13,9 @@ #' @param n.max Maximum sample size #' @param step Step size of the optional stopping (default is 1) #' @param peek Determines how often one peeks at the data. Overrides step argument if not NULL. -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test (default: 0.05) #' @importFrom stats t.test -#' @importFrom utils tail .optstop <- function(df, group, dv, n.min, n.max, step = 1, peek = NULL, alternative = "two.sided", alpha = 0.05){ @@ -36,49 +35,61 @@ } # Compute t-tests - mod <- sapply(peeks, FUN = function(x) {stats::t.test(g1[1:x], g2[1:x], var.equal = TRUE, alternative = alternative)}) - ps <- simplify2array(mod["p.value",]) - r2s <- sapply(peeks, FUN = function(x) {.compR2t(g1[1:x], g2[1:x])}) - ds <- .compCohensD(simplify2array(mod["statistic",]), peeks) + analyses <- lapply(peeks, function(x){ + control <- g1[1:x] + treatment <- g2[1:x] + report <- .report.twogroup(control = control, + treatment = treatment, + method = "t.equal", + alternative = alternative) + + list(report = report, + r2 = .compR2t(control, treatment), + d = .compCohensDStat(statistic = report[["stat"]], n1 = length(control), n2 = length(treatment))) + }) + + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) # Do the p-hacking if(any(ps < alpha) == FALSE){ - p.final <- utils::tail(ps, 1) - r2.final <- utils::tail(r2s, 1) - d.final <- utils::tail(ds, 1) - } else if (any(ps < alpha) == TRUE) { - p.final <- ps[which(ps < alpha)][1] - r2.final <- unique(r2s[ps == p.final]) - d.final <- unique(ds[ps == p.final]) + final.index <- length(peeks) + } else { + final.index <- which(ps < alpha)[1] } - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s, - d.final = d.final, - ds = ds)) + initial.index <- length(peeks) + + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[initial.index]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[initial.index]][["r2"]], + ds.hack = analyses[[final.index]][["d"]], + ds.orig = analyses[[initial.index]][["d"]], + report.initial = analyses[[initial.index]][["report"]], + report.final = analyses[[final.index]][["report"]])) } -#' Simulate p-hacking with incorrect rounding +#' Simulate p-hacking with optional stopping +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param n.min Minimum sample size #' @param n.max Maximum sample size #' @param step Step size of the optional stopping (default is 1) #' @param peek Determines how often one peeks at the data. Overrides step argument if not NULL. -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param iter Number of iterations #' @param alpha Significance level of the t-test (default: 0.05) #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE -#' @importFrom utils tail #' @export #' -sim.optstop <- function(n.min, n.max, step = 1, peek = NULL, alternative = "two.sided", iter = 1000, alpha = 0.05, shinyEnv = FALSE){ +sim.optstop <- function(n.min, n.max, step = 1, peek = NULL, effect = 0, heterogeneity = 0, alternative = "two.sided", iter = 1000, alpha = 0.05, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.data(nobs.group = n.max) + dat[[i]] <- .sim.data(nobs.group = n.max, effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -101,24 +112,7 @@ sim.optstop <- function(n.min, n.max, step = 1, peek = NULL, alternative = "two. }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - ds.hack <- NULL - ds.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- utils::tail(res[[i]][["ps"]], 1) - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- utils::tail(res[[i]][["r2s"]], 1) - ds.hack[i] <- res[[i]][["d.final"]] - ds.orig[i] <- utils::tail(res[[i]][["ds"]], 1) - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig, ds.hack, ds.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig", "ds.hack", "ds.orig")) } diff --git a/phackR/R/outlierExclusion.R b/phackR/R/outlierExclusion.R index 3baad71..12f5210 100644 --- a/phackR/R/outlierExclusion.R +++ b/phackR/R/outlierExclusion.R @@ -390,218 +390,123 @@ x <- df[,x] y <- df[,y] - # initialize p value list (one level for each outlier method) - ps <- vector("list", 12) - r2s <- vector("list", 12) + analyses <- list() - #### Go through each outlier detection method and calculate p values #### - - # Boxplot - if(1 %in% which){ + .report.outliers <- function(dat, label){ - dat <- .out.boxplot(x, y) - ps[[1]] <- rep(NA, length(dat)) - r2s[[1]] <- rep(NA, length(dat)) + if(is.matrix(dat)) dat <- list(dat) + out <- vector("list", length(dat)) for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[1]][i] <- mod$coefficients[2,4] - r2s[[1]][i] <- mod$r.squared + report <- .report.association(x = dat[[i]][,1], y = dat[[i]][,2], + method = paste0(label, ".", i)) + out[[i]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } + + out + } - # Stem & Leaf - if(2 %in% which){ + #### Go through each outlier detection method and calculate p values #### - dat <- .out.stemleaf(x, y) - ps[[2]] <- rep(NA, length(dat)) - r2s[[2]] <- rep(NA, length(dat)) + report.orig <- .report.association(x = x, y = y, method = "lm.original") + analyses[[1]] <- list(report = report.orig, + r2 = tanh(report.orig[["effect"]])^2) - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[2]][i] <- mod$coefficients[2,4] - r2s[[2]][i] <- mod$r.squared - } + # Boxplot + if(1 %in% which){ + analyses <- c(analyses, .report.outliers(.out.boxplot(x, y), "out.boxplot")) + } + + # Stem & Leaf + if(2 %in% which){ + analyses <- c(analyses, .report.outliers(.out.stemleaf(x, y), "out.stemleaf")) } # Standard deviation if(3 %in% which){ - - dat <- .out.sdrule(x, y) - ps[[3]] <- rep(NA, length(dat)) - r2s[[3]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[3]][i] <- mod$coefficients[2,4] - r2s[[3]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.sdrule(x, y), "out.sdrule")) } # Percentile if(4 %in% which){ - - dat <- .out.percentrule(x, y) - ps[[4]] <- rep(NA, length(dat)) - r2s[[4]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[4]][i] <- mod$coefficients[2,4] - r2s[[4]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.percentrule(x, y), "out.percentrule")) } # Studentized residuals if(5 %in% which){ - - dat <- .out.residual(x, y, type = "stud") - ps[[5]] <- rep(NA, length(dat)) - r2s[[5]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[5]][i] <- mod$coefficients[2,4] - r2s[[5]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.residual(x, y, type = "stud"), "out.residual.stud")) } # Standardized residuals if(6 %in% which){ - - dat <- .out.residual(x, y, type = "stan") - ps[[6]] <- rep(NA, length(dat)) - r2s[[6]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[6]][i] <- mod$coefficients[2,4] - r2s[[6]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.residual(x, y, type = "stan"), "out.residual.stan")) } # DFBETA if(7 %in% which){ - - dat <- .out.dfbeta(x, y) - ps[[7]] <- rep(NA, length(dat)) - r2s[[7]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[7]][i] <- mod$coefficients[2,4] - r2s[[7]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.dfbeta(x, y), "out.dfbeta")) } # DFFITS if(8 %in% which){ - - dat <- .out.dffits(x, y) - ps[[8]] <- rep(NA, length(dat)) - r2s[[8]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[8]][i] <- mod$coefficients[2,4] - r2s[[8]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.dffits(x, y), "out.dffits")) } # Cook's distance if(9 %in% which){ - - dat <- .out.cook(x, y) - ps[[9]] <- rep(NA, length(dat)) - r2s[[9]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[9]][i] <- mod$coefficients[2,4] - r2s[[9]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.cook(x, y), "out.cook")) } # Mahalanobis distance if(10 %in% which){ - - dat <- .out.mahalanobis(x, y) - ps[[10]] <- rep(NA, length(dat)) - r2s[[10]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[10]][i] <- mod$coefficients[2,4] - r2s[[10]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.mahalanobis(x, y), "out.mahalanobis")) } # Leverage levels if(11 %in% which){ - - dat <- .out.leverage(x, y) - ps[[11]] <- rep(NA, length(dat)) - r2s[[11]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[11]][i] <- mod$coefficients[2,4] - r2s[[11]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.leverage(x, y), "out.leverage")) } # Covariance ratio if(12 %in% which){ - - dat <- .out.covratio(x, y) - ps[[12]] <- rep(NA, length(dat)) - r2s[[12]] <- rep(NA, length(dat)) - - for(i in 1:length(dat)){ - mod <- summary(stats::lm(dat[[i]][,2] ~ dat[[i]][,1])) - ps[[12]][i] <- mod$coefficients[2,4] - r2s[[12]][i] <- mod$r.squared - } + analyses <- c(analyses, .report.outliers(.out.covratio(x, y), "out.covratio")) } - # Compute original p value - mod <- summary(stats::lm(y ~ x)) - p.orig <- mod$coefficients[2,4] - r2.orig <- mod$r.squared - # Combine all p values and remove NAs - ps <- c(p.orig, unlist(ps)) - ps <- ps[!is.na(ps)] - r2s <- c(r2.orig, unlist(r2s)) - r2s <- r2s[!is.na(ps)] + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-Hacking with different sorts of outlier definition -#' @description Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs Integer giving number of observations #' @param which Which outlier detection methods? Either 5 random methods are chosen ("random") or a numeric vector containing the chosen methods (1: boxplot, 2: stem&leaf, 3: standard deviation, 4: percentile, 5: studentized residuals, 6: standardized residuals, 7: DFBETA, 8: DFFITS, 9: Cook's D, 10: Mahalanobis distance, 11: Leverage values, 12: Covariance ratio) #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies on the Fisher-z scale +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale #' @param alpha Significance level of the t-test (default: 0.05) #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.outHack <- function(nobs, which = c(1:12), strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.outHack <- function(nobs, which = c(1:12), strategy = "firstsig", effect = 0, heterogeneity = 0, alpha = 0.05, iter = 1000, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.multcor(nobs = nobs, nvar = 2, r = 0) + dat[[i]] <- .sim.association(nobs = nobs, effect = effect, heterogeneity = heterogeneity) } # If which = "random @@ -624,21 +529,6 @@ sim.outHack <- function(nobs, which = c(1:12), strategy = "firstsig", alpha = 0. }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - ps.all <- list() - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - #ps.all[[i]] <- res[[i]][["ps"]] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig")) } diff --git a/phackR/R/plotsShiny.R b/phackR/R/plotsShiny.R index adc0fe7..429ccd5 100644 --- a/phackR/R/plotsShiny.R +++ b/phackR/R/plotsShiny.R @@ -8,7 +8,7 @@ #' @importFrom ggplot2 ggplot geom_histogram aes theme_light xlab ggtitle theme element_text geom_vline scale_fill_manual layer_scales ylab geom_segment geom_col scale_x_continuous scale_y_continuous waiver #' @importFrom rlang .data #' @importFrom dplyr all_of mutate -#' @importFrom magrittr "%$%" +#' @importFrom magrittr "%$%" "%>%" pplots <- function(simdat, alpha){ diff --git a/phackR/R/selectiveReportingDV.R b/phackR/R/selectiveReportingDV.R index 4d79494..0b5c5e2 100644 --- a/phackR/R/selectiveReportingDV.R +++ b/phackR/R/selectiveReportingDV.R @@ -7,8 +7,10 @@ #' @param nobs.group Vector giving number of observations per group #' @param nvar Number of dependent variables in the data frame #' @param r Desired correlation between the dependent variables (scalar) +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity -.sim.multDV <- function(nobs.group, nvar, r){ +.sim.multDV <- function(nobs.group, nvar, r, effect = 0, heterogeneity = 0){ # Observations per group if(length(nobs.group) == 1) nobs.group <- rep(nobs.group, 2) @@ -18,6 +20,8 @@ # Generate dependent variables dvs <- .sim.multcor(nobs = sum(nobs.group), nvar = nvar, r = r) + theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) + dvs[(nobs.group[1]+1):sum(nobs.group), ] <- dvs[(nobs.group[1]+1):sum(nobs.group), ] + theta # Generate data frame res <- cbind(group, dvs) @@ -31,7 +35,7 @@ #' @param dvs Vector defining the DV columns (will be checked in given order) #' @param group Scalar defining grouping column #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test #' @importFrom stats t.test @@ -40,53 +44,58 @@ # Prepare data frame dvs <- as.matrix(df[, dvs], ncol = length(dvs)) group <- df[, group] - mod <- list() - r2s <- NULL + analyses <- list() # Compute t-tests for(i in 1:ncol(dvs)){ - - mod[[i]] <- stats::t.test(dvs[, i] ~ group, - var.equal = TRUE, alternative = alternative) - r2s[i] <- .compR2t(dvs[group == unique(group)[1], i], - dvs[group == unique(group)[2], i]) + control <- dvs[group == unique(group)[1], i] + treatment <- dvs[group == unique(group)[2], i] + report <- .report.twogroup(control = control, + treatment = treatment, + method = "t.equal", + alternative = alternative) + analyses[[i]] <- list(report = report, + r2 = .compR2t(control, treatment), + d = .compCohensDStat(statistic = report[["stat"]], + n1 = length(control), + n2 = length(treatment))) } - ps <- unlist(simplify2array(mod)["p.value", ]) - ds <- .compCohensD(unlist(simplify2array(mod)["statistic", ]), length(df[, group])/2) - - # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) - d.final <- unique(ds[ps == p.final]) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s, - d.final = d.final, - ds = ds)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + ds.hack = analyses[[final.index]][["d"]], + ds.orig = analyses[[1]][["d"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-Hacking with multiple dependent variables -#' @description Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs.group Vector giving number of observations per group #' @param nvar Number of dependent variables (columns) in the data frame #' @param r Desired correlation between the dependent variables (scalar) #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity #' @param iter Number of simulation iterations -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test (default: 0.05) #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.multDVhack <- function(nobs.group, nvar, r, strategy = "firstsig", iter = 1000, alternative = "two.sided", alpha = 0.05, shinyEnv = FALSE){ +sim.multDVhack <- function(nobs.group, nvar, r, strategy = "firstsig", effect = 0, heterogeneity = 0, iter = 1000, alternative = "two.sided", alpha = 0.05, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.multDV(nobs.group = nobs.group, nvar = nvar, r = r) + dat[[i]] <- .sim.multDV(nobs.group = nobs.group, nvar = nvar, r = r, + effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -108,24 +117,7 @@ sim.multDVhack <- function(nobs.group, nvar, r, strategy = "firstsig", iter = 10 }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - ds.hack <- NULL - ds.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - ds.hack[i] <- res[[i]][["d.final"]] - ds.orig[i] <- res[[i]][["ds"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig, ds.hack, ds.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig", "ds.hack", "ds.orig")) } diff --git a/phackR/R/selectiveReportingIV.R b/phackR/R/selectiveReportingIV.R index 863ceec..c845611 100644 --- a/phackR/R/selectiveReportingIV.R +++ b/phackR/R/selectiveReportingIV.R @@ -8,22 +8,33 @@ #' @param nvar Number of independent variables in the data frame #' @param r Desired correlation between the independent variables (scalar) #' @param regression Should the simulation be conducted for a regression analysis (TRUE) or a t-test? (FALSE) +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity -.sim.multIV <- function(nobs.group, nvar, r, regression = FALSE){ +.sim.multIV <- function(nobs.group, nvar, r, regression = FALSE, effect = 0, heterogeneity = 0){ + + if(regression){ + return(.sim.multregression(nobs = nobs.group, nvar = nvar, r = r, + effect = effect, heterogeneity = heterogeneity)) + } # Observations per group if(length(nobs.group) == 1) nobs.group <- rep(nobs.group, 2) - # Simulate control group / criterion variable + # Simulate control group control <- rnorm(nobs.group[1]) - if(regression) criterion <- control # Simulate multiple experimental groups / predictor variables ivs <- .sim.multcor(nobs = nobs.group[2], nvar = nvar, r = r) + theta <- .draw.study.effect(effect = effect, heterogeneity = heterogeneity) + ivs <- ivs + theta # Generate data frame - res <- cbind(control, ivs) - if(regression) colnames(res)[1] <- "criterion" + nrows <- max(length(control), nrow(ivs)) + control.pad <- c(control, rep(NA, nrows-length(control))) + ivs.pad <- matrix(NA, nrow = nrows, ncol = ncol(ivs)) + ivs.pad[1:nrow(ivs), ] <- as.matrix(ivs) + res <- cbind(control.pad, ivs.pad) return(res) @@ -35,7 +46,7 @@ #' @param ivs Location of the independent variables (treatment groups) in the (wide) data frame #' @param control Location of the control group in the (wide) data frame #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). For the t-test path, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test (default: 0.05) #' @importFrom stats t.test @@ -45,29 +56,32 @@ control <- df[, control] # Prepare dataset - mod <- list() - r2s <- rep(NA, length(ivs)) + analyses <- list() # Compute t-tests for(i in 1:length(ivs)){ - mod[[i]] <- stats::t.test(control, treatm[,i], var.equal = TRUE, alternative = alternative) - r2s[i] <- .compR2t(control, treatm[,i]) + report <- .report.twogroup(control = control, + treatment = treatm[,i], + method = "t.equal", + alternative = alternative) + analyses[[i]] <- list(report = report, + r2 = .compR2t(control, treatm[,i]), + d = .compCohensDStat(statistic = report[["stat"]], + n1 = length(control), + n2 = length(treatm[,i]))) } - ps <- unlist(simplify2array(mod)["p.value", ]) - ds <- .compCohensD(unlist(simplify2array(mod)["statistic", ]), length(control)) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- r2s[ps == p.final] - d.final <- ds[ps == p.final] - - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s, - d.final = d.final, - ds = ds)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + ds.hack = analyses[[final.index]][["d"]], + ds.orig = analyses[[1]][["d"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } @@ -77,7 +91,7 @@ #' @param ivs Location of the independent variables (predictors) in the data frame #' @param control Location of the criterion in the data frame #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). For the regression path, \code{"greater"} tests a positive association. #' @param alpha Significance level of the t-test (default: 0.05) #' @importFrom stats t.test @@ -87,50 +101,60 @@ criterion <- df[, control] # Prepare dataset - ps <- rep(NA, length(ivs)) - r2s <- rep(NA, length(ivs)) + analyses <- list() # Compute regressions for(i in 1:length(ivs)){ - mod <- summary(stats::lm(criterion ~ predictors[,i])) - ps[i] <- mod$coefficients[2, 4] - r2s[i] <- mod$r.squared + report <- .report.association(x = predictors[,i], y = criterion, + method = paste0("lm.predictor.", i), + alternative = alternative) + analyses[[i]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } - # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- r2s[ps == p.final] - - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s)) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) + + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-Hacking with multiple independent variables -#' @description Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs.group Vector giving number of observations per group #' @param nvar Number of independent variables (columns) in the data frame #' @param r Desired correlation between the dependent variables (scalar) #' @param regression Should the simulation be conducted for a regression analysis (TRUE) or a t-test? (FALSE) #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies. For \code{regression = FALSE}, this is on the standardized mean-difference scale. For \code{regression = TRUE}, it is on the Fisher-z scale. +#' @param heterogeneity Between-study heterogeneity. For \code{regression = FALSE}, this is on the standardized mean-difference scale. For \code{regression = TRUE}, it is on the Fisher-z scale. #' @param iter Number of simulation iterations -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). For \code{regression = FALSE}, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. For \code{regression = TRUE}, it tests a positive association. #' @param alpha Significance level of the t-test (default: 0.05) #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.multIVhack <- function(nobs.group, nvar, r, regression=FALSE, strategy = "firstsig", iter = 1000, alternative = "two.sided", alpha = 0.05, shinyEnv = FALSE){ +sim.multIVhack <- function(nobs.group, nvar, r, regression=FALSE, strategy = "firstsig", effect = 0, heterogeneity = 0, iter = 1000, alternative = "two.sided", alpha = 0.05, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.multIV(nobs.group = nobs.group, nvar = nvar, r = r, regression=regression) + dat[[i]] <- .sim.multIV(nobs.group = nobs.group, nvar = nvar, r = r, + regression = regression, effect = effect, + heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset - .multIVhack <- ifelse(regression, .multIVhack_reg, .multIVhack_ttest) + if(regression){ + .multIVhack <- .multIVhack_reg + } else { + .multIVhack <- .multIVhack_ttest + } .multIVhacklist <- function(x){ .multIVhack(df = x, ivs = c(2:(nvar+1)), control = 1, @@ -153,27 +177,12 @@ sim.multIVhack <- function(nobs.group, nvar, r, regression=FALSE, strategy = "fi }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - ds.hack <- NULL - ds.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - if(!regression){ - ds.hack[i] <- res[[i]][["d.final"]] - ds.orig[i] <- res[[i]][["ds"]][1] - } + if(regression){ + return(.combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig"))) } - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - if(!regression) res <- cbind(res, ds.hack, ds.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig", "ds.hack", "ds.orig")) } diff --git a/phackR/R/statAnalysis.R b/phackR/R/statAnalysis.R index 7833a7d..b0b0183 100644 --- a/phackR/R/statAnalysis.R +++ b/phackR/R/statAnalysis.R @@ -9,62 +9,62 @@ #' @param group Location of the grouping variable in the data frame #' @param dv Location of the dependent variabl in the data frame #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test #' @importFrom stats t.test wilcox.test #' @importFrom WRS2 yuen .statAnalysisHack <- function(df, group, dv, strategy = "firstsig", alternative = "two.sided", alpha = 0.05){ - dftest <- cbind(df[, group], df[, dv]) - colnames(dftest) <- c("group", "dv") + control <- df[df[,group] == unique(df[,group])[1], dv] + treatment <- df[df[,group] == unique(df[,group])[2], dv] - # "Normal" t-test - p.orig <- stats::t.test(dv ~ group, var.equal = TRUE, alternative = alternative, - data = dftest)$p.value + analyses <- list( + .report.twogroup(control = control, treatment = treatment, + method = "t.equal", alternative = alternative), + .report.twogroup(control = control, treatment = treatment, + method = "t.welch", alternative = alternative), + .report.twogroup(control = control, treatment = treatment, + method = "wilcox", alternative = alternative) + ) - # Welch test - p.welch <- stats::t.test(dv ~ group, var.equal = FALSE, - alternative = alternative, data = dftest)$p.value - - # Mann-Whitney / Wilcoxon test - p.wilcox <- stats::wilcox.test(dv ~ group, alternative = alternative, - data = dftest)$p.value - - # Yuen test with different levels of trimming - p.yuen <- rep(NA, 4) trim <- c(0.1, 0.15, 0.2, 0.25) - for(i in 1:4) { - p.yuen[i] <- WRS2::yuen(dv ~ group, tr = trim[i], - data = as.data.frame(dftest))$p.value + for(i in 1:length(trim)){ + analyses[[length(analyses)+1]] <- .report.twogroup(control = control, + treatment = treatment, + method = "yuen", + trim = trim[i], + alternative = alternative) } - ps <- c(p.orig, p.welch, p.wilcox, p.yuen) + ps <- vapply(analyses, function(x) x[["p"]], numeric(1)) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - - return(list(p.final = p.final, - ps = ps)) + return(list(ps.hack = analyses[[final.index]][["p"]], + ps.orig = analyses[[1]][["p"]], + report.initial = analyses[[1]], + report.final = analyses[[final.index]])) } #' Simulate p-Hacking for exploiting different statistical analysis options -#' @description Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs.group Number of observations per group. Either a scalar or a vector with 2 elements. #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param alpha Significance level of the t-test #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.statAnalysisHack <- function(nobs.group, strategy = "firstsig", alternative = "two.sided", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.statAnalysisHack <- function(nobs.group, strategy = "firstsig", effect = 0, heterogeneity = 0, alternative = "two.sided", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.data(nobs.group = nobs.group) + dat[[i]] <- .sim.data(nobs.group = nobs.group, effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -89,16 +89,8 @@ sim.statAnalysisHack <- function(nobs.group, strategy = "firstsig", alternative }) } - ps.hack <- NULL - ps.orig <- NULL - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - } - - res <- cbind(ps.hack, ps.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig")) } diff --git a/phackR/R/subgroupAnalysis.R b/phackR/R/subgroupAnalysis.R index 45296d6..2917d4c 100644 --- a/phackR/R/subgroupAnalysis.R +++ b/phackR/R/subgroupAnalysis.R @@ -6,10 +6,12 @@ #' @description Outputs data frame with multiple binary variables from which subgroups can be extracted #' @param nobs.group Vector giving number of observations per group #' @param nsubvars Integer specifying number of variables for potential subgroups +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity -.sim.subgroup <- function(nobs.group, nsubvars){ +.sim.subgroup <- function(nobs.group, nsubvars, effect = 0, heterogeneity = 0){ - dat <- .sim.data(nobs.group) + dat <- .sim.data(nobs.group = nobs.group, effect = effect, heterogeneity = heterogeneity) # Observations per group and total observations if(length(nobs.group) == 1) nobs.group <- rep(nobs.group, 2) @@ -32,85 +34,78 @@ #' @param iv Integer specifying the location of the binary independent variable in the data frame #' @param dv Integer specifying the location of the dependent variable in the data frame #' @param subvars Vector specifying the location of the subgroup variables in the data frame -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" #' @param alpha Significance level of the t-test -#' @importFrom dplyr group_by_at do #' @importFrom stats t.test -#' @importFrom dplyr "%>%" -#' @importFrom rlang .data .subgroupHack <- function(df, iv, dv, subvars, alternative = "two.sided", strategy = "firstsig", alpha = 0.05){ - # Prepare data frame - ttest.df <- cbind(df[,iv], df[,dv]) - subvars.df <- cbind(df[, subvars]) - dfnew <- as.data.frame(cbind(ttest.df, subvars.df)) + group.values <- unique(df[,iv]) + control <- df[df[,iv] == group.values[1], dv] + treatment <- df[df[,iv] == group.values[2], dv] - # Compute p-values, R^2, Cohen's d - - # Not p-hacked - mod.orig <- stats::t.test(ttest.df[,2] ~ ttest.df[,1], var.equal = TRUE, alternative = alternative) - p.orig <- mod.orig$p.value - r2.orig <- .compR2t(ttest.df[ttest.df[,1] == unique(ttest.df[,1])[1],2], - ttest.df[ttest.df[,1] == unique(ttest.df[,1])[2],2]) - d.orig <- .compCohensD(unname(mod.orig$statistic), nrow(ttest.df)/2) - - - # p-hacked - ps <- list() - ds <- list() - r2s <- list() + analyses <- list(list(report = .report.twogroup(control = control, + treatment = treatment, + method = "t.equal", + alternative = alternative), + r2 = .compR2t(control, treatment))) + analyses[[1]][["d"]] <- .compCohensDStat(statistic = analyses[[1]][["report"]][["stat"]], + n1 = length(control), + n2 = length(treatment)) for(i in 1:length(subvars)){ - - tmp <- dplyr::group_by_at(dfnew, subvars[i]) %>% - dplyr::do(as.data.frame(stats::t.test(.data$V2 ~ .data$V1, var.equal = TRUE, alternative = alternative)[c("p.value", "statistic")])) - tmp2 <- dplyr::group_by_at(dfnew, subvars[i]) %>% - dplyr::do(as.data.frame(table(.data$V1))) - tmp3 <- dplyr::group_by_at(dfnew, subvars[i]) %>% do(as.data.frame(.compR2t(.data$V2[.data$V1 == unique(.data$V1)[1]], .data$V2[.data$V1 == unique(.data$V1)[2]]))) - - ps[[i]] <- tmp[[2]] - ds[[i]] <- c(tmp[[3]][1]*sqrt(sum(1/tmp2[[3]][1:2])), tmp[[3]][2]*sqrt(sum(1/tmp2[[3]][3:4]))) - r2s[[i]] <- tmp3[[2]] - + levels.current <- sort(unique(df[,subvars[i]])) + for(j in 1:length(levels.current)){ + subset.df <- df[df[,subvars[i]] == levels.current[j], , drop = FALSE] + control.sub <- subset.df[subset.df[,iv] == group.values[1], dv] + treatment.sub <- subset.df[subset.df[,iv] == group.values[2], dv] + report <- .report.twogroup(control = control.sub, + treatment = treatment.sub, + method = "t.equal", + alternative = alternative) + analyses[[length(analyses)+1]] <- list(report = report, + r2 = .compR2t(control.sub, treatment.sub), + d = .compCohensDStat(statistic = report[["stat"]], + n1 = length(control.sub), + n2 = length(treatment.sub))) + } } - ps <- c(p.orig, unlist(ps)) - r2s <- c(r2.orig, unlist(r2s)) - ds <- c(d.orig, unlist(ds)) - - # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) - d.final <- unique(ds[ps == p.final]) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s, - d.final = d.final, - ds = ds)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + ds.hack = analyses[[final.index]][["d"]], + ds.orig = analyses[[1]][["d"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-hacking with multiple subgroups -#' Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs.group Vector giving number of observations per group #' @param nsubvars Integer specifying number of variables for potential subgroups -#' @param alternative Direction of the t-test ("two.sided", "less", "greater") +#' @param effect Mean effect size across studies +#' @param heterogeneity Between-study heterogeneity +#' @param alternative Direction of the t-test ("two.sided", "less", "greater"). Here, \code{"greater"} tests whether the treatment or second group exceeds the control or first group. #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" #' @param alpha Significance level of the t-test #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.subgroupHack <- function(nobs.group, nsubvars, alternative = "two.sided", strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.subgroupHack <- function(nobs.group, nsubvars, effect = 0, heterogeneity = 0, alternative = "two.sided", strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.subgroup(nobs.group = nobs.group, nsubvars = nsubvars) + dat[[i]] <- .sim.subgroup(nobs.group = nobs.group, nsubvars = nsubvars, + effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -135,25 +130,8 @@ sim.subgroupHack <- function(nobs.group, nsubvars, alternative = "two.sided", st }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - ds.hack <- NULL - ds.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - ds.hack[i] <- res[[i]][["d.final"]] - ds.orig[i] <- res[[i]][["ds"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig, ds.hack, ds.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig", "ds.hack", "ds.orig")) } diff --git a/phackR/R/variableTransformation.R b/phackR/R/variableTransformation.R index 26d2a41..e6a2dad 100644 --- a/phackR/R/variableTransformation.R +++ b/phackR/R/variableTransformation.R @@ -30,8 +30,10 @@ Xtrans <- matrix(NA, nrow = nrow(df)) Xtrans[,1] <- x + xlabels <- "x.orig" Ytrans <- matrix(NA, nrow = nrow(df)) Ytrans[,1] <- y + ylabels <- "y.orig" if(transvar != "y" && normality == FALSE){ Xtrans <- cbind(Xtrans, @@ -39,6 +41,7 @@ sqrt(x+abs(min(x))+1e-10), # square root transformation 1/x # inverse ) + xlabels <- c(xlabels, "x.log", "x.sqrt", "x.inv") } @@ -48,52 +51,55 @@ sqrt(y+abs(min(y))+1e-10), # square root transformation 1/y # inverse ) + ylabels <- c(ylabels, "y.log", "y.sqrt", "y.inv") } # Calculate p-values for all transformed variables - ps <- matrix(NA, nrow = dim(Xtrans)[2], ncol = dim(Ytrans)[2]) - r2s <- matrix(NA, nrow = dim(Xtrans)[2], ncol = dim(Ytrans)[2]) + analyses <- list() for(i in 1:ncol(Xtrans)){ for(j in 1:ncol(Ytrans)){ - mod <- summary(stats::lm(Ytrans[,j] ~ Xtrans[,i])) - ps[i,j] <- mod$coefficients[2, 4] - r2s[i,j] <- mod$r.squared + report <- .report.association(x = Xtrans[,i], y = Ytrans[,j], + method = paste(xlabels[i], ylabels[j], sep = "_")) + analyses[[length(analyses)+1]] <- list(report = report, + r2 = tanh(report[["effect"]])^2) } } - ps <- as.vector(ps) - r2s <- as.vector(r2s) + ps <- vapply(analyses, function(x) x[["report"]][["p"]], numeric(1)) # Select final p-hacked p-value based on strategy - p.final <- .selectpvalue(ps = ps, strategy = strategy, alpha = alpha) - r2.final <- unique(r2s[ps == p.final]) + final.index <- .selectanalysis(ps = ps, strategy = strategy, alpha = alpha) - return(list(p.final = p.final, - ps = ps, - r2.final = r2.final, - r2s = r2s)) + return(list(ps.hack = analyses[[final.index]][["report"]][["p"]], + ps.orig = analyses[[1]][["report"]][["p"]], + r2s.hack = analyses[[final.index]][["r2"]], + r2s.orig = analyses[[1]][["r2"]], + report.initial = analyses[[1]][["report"]], + report.final = analyses[[final.index]][["report"]])) } #' Simulate p-hacking with variable transformations -#' Outputs a matrix containing the p-hacked p-values (\code{ps.hack}) and the original p-values (\code{ps.orig}) from all iterations +#' @description Outputs a data frame containing the p-hacked p-values (\code{ps.hack}), the original p-values (\code{ps.orig}), and a normalized reporting block from all iterations #' @param nobs Integer giving number of observations #' @param transvar Which variables should be transformed? Either "x" (for x variable), "y" (for y variable), or "xy" (for both) #' @param testnorm Should variables only be transformed after a significant test for normality of residuals? #' @param strategy String value: One out of "firstsig", "smallest", "smallest.sig" +#' @param effect Mean effect size across studies on the Fisher-z scale +#' @param heterogeneity Between-study heterogeneity on the Fisher-z scale #' @param alpha Significance level of the t-test (default: 0.05) #' @param iter Number of simulation iterations #' @param shinyEnv Is the function run in a Shiny session? TRUE/FALSE #' @export -sim.varTransHack <- function(nobs, transvar, testnorm = FALSE, strategy = "firstsig", alpha = 0.05, iter = 1000, shinyEnv = FALSE){ +sim.varTransHack <- function(nobs, transvar, testnorm = FALSE, strategy = "firstsig", effect = 0, heterogeneity = 0, alpha = 0.05, iter = 1000, shinyEnv = FALSE){ # Simulate as many datasets as desired iterations dat <- list() for(i in 1:iter){ - dat[[i]] <- .sim.multcor(nobs = nobs, nvar = 2, r = 0) + dat[[i]] <- .sim.association(nobs = nobs, effect = effect, heterogeneity = heterogeneity) } # Apply p-hacking procedure to each dataset @@ -118,20 +124,7 @@ sim.varTransHack <- function(nobs, transvar, testnorm = FALSE, strategy = "first }) } - ps.hack <- NULL - ps.orig <- NULL - r2s.hack <- NULL - r2s.orig <- NULL - - for(i in 1:iter){ - ps.hack[i] <- res[[i]][["p.final"]] - ps.orig[i] <- res[[i]][["ps"]][1] - r2s.hack[i] <- res[[i]][["r2.final"]] - r2s.orig[i] <- res[[i]][["r2s"]][1] - } - - res <- cbind(ps.hack, ps.orig, r2s.hack, r2s.orig) - - return(res) + .combine.phase1.results(res = res, + legacy.fields = c("ps.hack", "ps.orig", "r2s.hack", "r2s.orig")) } diff --git a/phackR/inst/shiny-phack/ShinyPHack/data/startplots.rds b/phackR/inst/shiny-phack/ShinyPHack/data/startplots.rds index 7b98f826421c51ce2434bb79806944261b4f04ab..beef93b822a893d67960438bfe14bdf4e32f9eb4 100644 GIT binary patch literal 19095997 zcmaHxbx<5I*RM-~LUD>qaV_r7;%>!VTHM{;wNTt$iWhfxcV}4~7I$Cv>-&Cl@67$@ zW+utZ$;p`{GiNgS8$H>Ht~oXslJepN!==J%qz*}2U$j^1UQGl7P4Rr zSrrU=IbBqDgcx!U3tWx{UaG`kTiIsYkyM9?L~5P7YyD4(5HGAvHvY{P%cN39V3t_s|&p#!)1COp27tdTvx}gsn zA*9>D%1+b-neC;3w{6dZeU72eu}LSYv0{;$Dpi~kt=2mN($1Hi5)0t6*(Rzv^(c0Wwi%W;MuUl<{{?%#%blL}=Q*D?hv?rI30d%6xMvQ6n^RXc%^JD(BuZt@1Z%#t%i~ zrnq2rPybw<+f>V|7P8)B{{_@f%)VuuedLFQ@Av_ObXYf%Lu6akw8j%tf9-4%NR=g zVk4u0{3UXFmF2AA2h}A_PPw^?4F|jmxrv9qosN~xl)&@s0jU`qE_BnW(>(Jg9MS~O7UlfPOZlp zFLq{vebXJSc~T2-o+$r`hef3~Yv zoqNX)G5RmWxPwJUKXf3I7n-m8*4Rh>q4zMteq&L)JrC}L12LgY_r=jOU zOtM-LA>>0{r=W(l!W(e;V9EHgT|UcRTLNMdUOtC`{F(Ue@GICmRV za@%sfRub-D{M=Zzg(S`?&Ry1s%LoRGzGwx8t0OIlEa3>Q%s=_|-cY2O_1>&x%B*0` zZ7#zGb$@)0Z~Qik4S!ll&{S!heo2ApV4@l*fJh2V2;eDCXT+HYN_GF9rpW*KY!h|i zou9?gYVRnt?H#3;-1pm62Tvp|pp*5NvDUtNJNNf==3Vvnm@sDHiwKgqQVN~JGG4=a zoOE|$4d?n$OLLFsPZ=b0PG59W0lDDev2(?i&$~P6QG06PY`@YqPEvGCzFt=CP1fuo z+;sE9LAFgxv>7g|j+F7Hac{#qf0)`CR{8N!oX@2M(!)@;uns!!+lIb=UbVLbpY%50Du zTnzW{nFM!Wp!y?E6s4GV;~eGHn02HI&7@W ztJj-O7%DY7QqmkD5zM)#0DY(v{3EH>+jylzDXW0qH)S2;;qTZIR5nK{Fy?JZ44eRj zxOj83X5iG0KIz%ZHbVuvy_Fg6!?zmA4%~CyyOgCuV1+{KyExv=J!uWJHTWeynf7-q z!fcE{qKcu=H#Tl=nzsw-d*;*`b|i?}LP}B?nCwFb>w;DCj!T%e zmv6mlZ5*D3xlFG3=6iE2e0%q<_y-#)ST^PgA?O`RwKTM5z^ z(?YTAsOLHM%+zHp_%|fU7JEoTx)9eo!%Rfe#tzI@l}{qhRRv=-J6G#y$X}l~Z2o#? zLvq`k)nd3{wnC*0DR(gU9AUo3EC^^ZbKmPsu?!>Z5I}T=?^og^G6~M92Hm+RhBi@m zR=$5qzcH8zb%m`=5W?rFMgm*kNgcMdI2~9H34gKZsrmc0IsmwM7|qRj&#BiW=xol( zVG`;pMbkp;y1>ViX?P4AxrfA2AteCF$pXA{>n<8&d`wRH@8p;`A5 zu0P^cg{l7(6ziKJt4v*mrW8hUbm7Yt{@)W2$ZC zgrz@Ig&xP~n6$q%7<;l|bF zmU^0-k12pLHj6diJEZ)~xU5@_ztliyhg={Lr+}L6_q}K%*D;oINl?k|Lg7{s#`+qO zBpY5xt@)1I|C1*D|Nk@@7(X?2YBz5KlgE;a12xcV9$mWfE@tliMH`j2HgVh~X|~&p zw!J=%JB)zRBFUp1IV{IwNDfJ9dak`oM8${(8a+bqHDm7EtO>;a@M?WxM^whedE3Ys zMg+=_y@f8m$a6s(NB7@dQMTr0|28&zHfwh!JQ5&3-;@)sBC*(tApw(I8n|-s564S5 zTqG4EA6@s`qUF=14L7L(2AOn=F-fWFn^vayW&tg~Ho!l;D*X8o+hW)`s3fY;S7h)G zoYR;ms8epszAcSiM~zP^kVltsCgiG>IG?dvY85EVzSL!^%dzVHS&H%Da_huev@cXi z9a+r;N_MGyI}-1Exi^@&`~ja&<=oN6f4pjvRF*nZ_B7=g)E-r7zq}%(R)vQm9evF$ z5odJ9dd*%qlLltJCj8^zkZ`x zwSzg+UO;XVkrYKzU60KxFL8D%#Cl&d{xz23jwnpwOk zY{}4QrTE%dym~PLzgm4|y}V-aX370lb(s_1OJn6BWzMItggY#EHJEWdd%3NR=QE!D z#rfe@r}2L~`-_}siahE9yVkY$LexOyA?r~EyFyljSZ(BPBvh|GrgxRd?e-Uy2 zmb~}MTM}(Doayag5hjZ|4H)M?M#8G;Z zFMj;0Cgqj+adAc;BH{m}tatyBZlRW)V%q-gSzsP9 z`wikq>3H)vr`^^9c-(vl9RA*~rJfZ;5ho2A?PDZ?QB@YNmu?(`wOgP9EdR%yj`E(4 z{GN_s?dN|oS()=AZ9twer_%B2@2uZTa$OXK<#B~H{EaQ@DI4TTv_wHkJ$cq}zlQC+ zI(88qp}K`X8^M6h$kXlbH?NpCulR$0U$1={3*=P70cc*NWjVN3j>;mPD@l#CgcQr+ zTOx0{i&xLsPhG#6%lF|-W#Ez$Uc>mK`VK^Sp2j{$>3_QGZNS?db!;0vrzBigP?5d3uAYXhc}XgXC}=( zfc5Al+f@L{JOJiWdDc*Mp6KsUK=4tIL>v?sH;*M?1oaJw_TD1?Lc0Tz^$bAT%9}5F z6uBI$uqWuA=#+-r2om;J+3V2uA7wmCaw>^LE*8a%P&EmVD6l>aYd=20nJbLH30rP8 z?M^QExWW^H7s9A;O5}Q0ThdfwX6FePvlrFOyzJLbe-qFu{tQsKJ)+~+*V|BW73zCk zpr3GR6rFF{;GSjJ9r!i{#R*C($`=@X^C1tI0;Pl(3h;wof_J#>{DkJoz|7KqK|{Aw zT2_8SiK^2h{EOZ3Q(8tZk~Dr^|AH8x`ROP;wIbag%q2VfSBbyeXpg_;9jm-6Pih$< zNYMBR&5?oEt7{>#(CU8sTEj1$J=D~fd#H&uz=sq6{G=<7lBPT?dNU87h`Yy!n>1%G^~A0)L?f@iDjSsNdBEd zS%Oy8l;7%|Al$cPOghsiHOOAZ@Au)?HD%d`Uz}hWmtUL*`2^CL4zsomI&c%BYk789 zeX=xbdFQt623o=alPZRSM8tLY*WxDh?KYRFJ|hnoJE5hIc)r}*yjk1?#!*5M~!cCAxo1QzB^7F@XHULPzq#0D^9_alwfy8dZb?DsRcKsWBU9>MSY>`1v zS!WV%(o_FWIQJOkbFzDWx_?s<{Hzf){%f0k>`d_ne*0%@d0DjA_;0vs;iip0D|YkZ zCGJ_u=iL8R9sTG2^2W%)@}#D_>_kW-P|YDtJHVB?gKA^W%w2WfK~i>4SPo=scjjuc zt8=4Uo?1G;7}M+B)(s}X5bY88sPxcK50X%OCv-f8zYs%!{E4M}HC&57B)^b~jM?~D z4*)96vxIn7ZW}xt$@h`I+MZ2BeC~?dj(|X<@+V%ON_vfQuUp#0-(+=!yq|(%#NNO9 zpE+LXvjiU*K6Jw5di35hdy<46*`0aO%J(4WM&RdUZF`{$qI0x+aI~jiPF;1PKLLEd z=RsklF3FHdLym-TNx_@bHyx6Sk6*4H=k3nE?smc58Ma4-!S=AgK+;4=SYV7;MEQ(9 zb{npTfwg^$eb9@1W_wBJ!)`j^ZjcvD1ovT|0i(J2znoV`b}7Fyk{)`*WGYFVKO^O7 z8GzsDEwf1Ae6kr+obIyzB-6^@A2Jdbd_>obvFJzhY*z(*Paj5W$vBf)3%%1}3}TK9 zRVmGaERDCZL66Cj0UmYFMn!X;-$Pyet&(QQ69`0(b{#?-k@(SDx>=k=#?|ZCIfV5r z4~<||g~Wuk<2&Y)PcM>O+oao8!YGJz3#(a;8wO!XNdC#OFU@0C!htko3!ma?ER#n`5S5Ey_uMt@Vh6ToIk|Jf zfEUDMgUj-{d&uN_o1pWh)L>t1=&p_L=Y3!yL#6f9SO6t=#PG4 z&0p)LX?o622N8#dCqqu8STv0mOX8ujC^F$@{RvnY44j)mq7rPm+Y*VQ$vCmWkq5+bQB5~g%4>sa%d8Mi>6eUAb?a$lY$%~Z7}a$W0W|FAO4 zfMFy_*Wb3}dtw2eKVYwD^6CkBbt|+6F~3JLCN4w_Q85=5hD(Houai-5CKvahdg>y% zehQkVo9!zfvwHGl*!zhis3q>5@$fmaLscRTwwhW=$zmTHDNQ(wtjFMr2y#P_K5hE} zuy4lkR2w45oy5M^9FmP@2b0S!^7d85kp2=#f%TDLv55<-ECy@l7tfKLONq6UbEitx zk}}?DNfaUFHUXmNM(BJzTD?D)wK^v{V#gr+Sb4_TQHinN)3?;L`G^Hq;Df7$CBK1p zIEX^ruIKZ(|v1gjkMSlLIKJST6Yrrq{fcQl?*9$`A{FNQS8}gaYWx zP&kGejuC6~MGMjfRhoF!`1FYL=(&#Ig$;pHW~0n`9Hq-R(_5~R1h|Lh)ST_yKP>uJ z2O}m@*i)05oTqWgp=e*d2m$QLDdW`8*MXZE&Dus*`LHIEnXCTzifeT0A*0gap3#A{ zm^(jo#g^dWq)P@#o0R(1JTAP`K1>v5`h_It;JqTz;JMBcm7LvyVQZ7ixs^XV0^og> z@LHS~%_SG}8da8P&z}_Vh&wH(>J*?6B*(>Ntr+Zq0dLr(uB5QFH83(Ev!a2{Owl8; z&o&)pk0^CO>X*^qKcFGuLJQfws(I3G*NmAS{eQgFhF&~98%O%r$YCu6%hHd3{MgGF z1+0~OkApM+ii^BDw(Ms}6laGYpoiv0aZ7YZ@n+Y4Z(PFSVFxyYj|va(y2T>cDZM*&$w3u1{ws2*%|K8J@v;>p1gVzHta0 z1p4F|@xZrzo6h@|TpE+~tjY&{s8gGQj&?d~RrC5;Ki6&<@$4mCeX5W0d~##O8w*)1**h5HAafSo84MOm4sbu) zd@gJH@}!H*x$;qYB8XOB-!0102x-7a))W0XPLTP_I8)whfkLH1?L3x71U%* z<7}-Udph$0rR^)-@{#v#NKPXO0xG!Vx9Bo0p=x{B`yWJwo}QR_g$?Ns(+zumgORh0 z*jW-iS?0W+bE;g2Jq(gjMk3VV1DKcn9vMcd18gCMesSEDVkl96_H2bgK(Odf?R27N z=Y%OtOuNvVVGx2ZI!WHe=6)3{`Ba2M21peX z9P4ZJwexcDk?TqydB#oj>3hPt4;zW=CMf>$RI>S@+pw{=EB>$Nj(03aW0q-fnehH- z1B!lHI|M`X0b^DtNoB`1t3Xde3X8LBk2~a|w@Cq(lA33LI%A<)-qK!Z1;0HhbrE39 z{)H{%i9DBu!W3QGt&*Vnk^Gbb`+=68xR?x7Sl4EpVJuWPs$))abG|DL3nZi zk2VxtyI=Cb?GC*NCx)Q-`3t|kO0yU6t>_FgqQK8ifotpZw)YvgqZr5D(;1P50uXpv zpYtP+i)zwu$||3S?#jK`I7zMCcOub&6Tatj3|&ZDKu57WY-!L;vQYu zs;NoDiU}caXV

P5rRJ4xt&vQ_Q{TQkKvgGT|qs?Q#^1i+>Oa~PZdH480>VeU{3Z8Qe<|QjkF`Ac%(&gN)ErDNfhV`40U_1xm)H%v? zuq3|Q8s1U*8ZM%`YR25*0b~H70x~&Cb|@Q~@sWpeh(o<`+}!WlM-qgCy*L6jmWMyr zuZJk{etJ>}gW)^aA4|E-&dj~CHKS7sZJsgLn_B8@z~N1!@e3x{;Mz|wZEzMGuBt7( z7wkVTJ=eyJmg}Lc2DCLM^~DFL%e>zCe*s+`wg-~@d-SKt?ZvPuu!E@615A|m@T~`f z)e4UI4yaD260WmS?Qs_sGg&XOV8XiKO7GVBp8Wq*#N zsszC%Jf=swdr)>m%kZ0n1w20a zj=|5LKdrBn|LHuJ4%)lm=+Q~?C`r?W$%LV9-8x|%ez$Mr&-s5V`x5@uzke(Gh<%{o z*yU{2JWP|Cu|tsZaTTZ7!cWF2x~=bA@7={uC|jxRI*iv5saN=N<^<#{$CtnTJksqU z*g5{@K<7fAWlN^x!PF!KMYysZOC3>pS=9un*~L%)|@AdxIONk->9yYrV~oM3<6$y(hE#%7i870?klb!n{zR zI3Hn=dP~=A08(YwK3y%qmTE9T1pB`1hJV>q-F3W=yyP8Ns76$@=W=R&*|YGmMbiqh z$smf;ucU=vQZv0X{&kJq<67yekgj-vKA?ssF+kRpvQ4|UEs5;fft_Z9kDlgk=}!FR z#Hy+!L24c}`Gg;~{`*~z%(e{Chv%1D-lc_8A)jeE_ouWAy=93Ew32O(Dg8L6!YT`Y zUI^ne6=Ug-BcauKaL#A>vdhe)dHRKd+{@uno#|`G-OLirGpdhzy4O)o3YNCzIC_?* zf?#Z1-w)jkCcGj=41{{@C3!_!<|RXItaiMKv`{ITHS;fML~ zUosAjc>_MSx<|$~@(~x5P`VH&xpm`AxN#npUsLEmwv)E3WHFo+Z)X3oT+`cJ@Sc$xq#1renu`eKTAcT}vTD8W%WDU? zsSzW$$yioI%QXZZ40+h#T`v_wVqZSo&>!%#aC6jm`8u`ur<$W!Q$F7Q=pGKv2{f zjBb-a69WE${u9bldW5GU`5s8YKBlbT$$8)W=^l~_El4;P1b4|Ev8M(TN2hzE+r}L> z^gQs0BvxLBlyM&<%*|#nF`W& z=>MwW*Wal~5TvogoEb;to$PVo$6uK0Q%{WC#XX~dAr#L(hOJ9oFcqTi8GaP&FT<_; zp3k2!4JP^M)CI;dWCcFDO7FCx?8*rsiBV|#5+v zxG_})J{XqzdU#M670VxdUO2(nTCXhcZ3Gn+xp#89txXT?4kqJd*5Fz_5s-Ur&llz; zGh3WFdp-v9U43cfCwI7qqnS(;(rp-3qg68gB5%c+@pxq8ae*m!`QlPq$K2?_Ma`#0 zRgX=P&7at;z#DyL!}yQSKlzf$*jT(e&UvcJd9>UBBN2QQzO zU?6;N`8%~ULV$z0yV!_bARkXnb&lx`JyU$uz2O#n zY04D5!SFyHT${YnM(a7C7i`11qwxeW7f!GpxZoA1Ddjq@GX<-8Zv<|+QLGP3VPCsZ zoR~faDwyhfeLu!u9EwwXB;Iesd)h>wflUAjt6pUFUDEuiJkl`$e!R91^F$z@*^u!% zxD;3Ug&zE7nQ<9>1Qgj*iu6Go4{-*dD4k4o6>@Bx2@PfFKzH@kiW(W0E@ZSbt>O=l zk7*d&3QK}}fT};s*HS&zIsMB^Fs|i^*Q}W!T;{SEyN&W(DSO@p(=j8Yyjbaw|gGzuVFUAmPlKp0agrA`7rG%I@LTIDKPY31z&~uqLjpy@u6wGwjsHV1 zv^V72LK0PE8o8_$-081KD8RzZ@B7Max_5S!P@Lp&9_Q=^ySxpZ;Ih%(FWs@J-w@VXZa9h>qeas3f=}wvzE)ZQ#;S9bIV^$z34C4A*&h$qi5vAs)-%EsEFGz9 zyHxVR->T#=_xv+hJ=)lP+gKl~_`Mkxp9n+(MV-Eg1agm6V;|0`hBxTtT<+|g9Ul4+ zR6>_$d*d3Ci+>v&wuOgIc(}j1$ZH|W@SlCOJmnt9%H=hf%&SmP)|yTkS9?bE$N^|o z;ohpSJ&(0jprj_9s#X)|r%y*Xl0b=}=QmIO<-9@}0RqLJYbhh8&Jx3eu#~fzmu|4a zoZnqtw7;d^9x3rakr{u#AAD%??Mh@xQO2o!rm=aBl55h<{&NHw67Nmgtjx1v*ZcO$ z_Qs+MuLCs~Px^WNK5zOiW>%wfcA{R#JhoyUgH@u2chkMp&8~y;2S6G{X;eP$IraO| zwocb!m|dPf6OYwk3))_)E9JxQBJgwv`o;UzOj0P_x!l`2}?24KV;TJ zBIZCuH3fFeF!e{#U!Hfx&pTf7t%1z^vLE?nZE=^C&ZsxuzL$a$dmlATE*4FnxJ|yE z7_r~Mm39j)%^vj$3J*N2VG{P_2p}a0cb*r@({<4F1isUI2><$bac?Qg#W-J3cY6K} zynq{D3n@oj%lPa8aTVYo^nwV+fy#j`z);-fT7~$d$PX~jpY8j@a`*sBnK7U1(AUTo zCp>N_#F9quCTA1w>vQ0_iap^}P}$$CeDMf%Y)+oAIQ@w{_&JRFmk!_J=iQrP=O!@I z%#e{b4?5wG@V*=VMR;PlG9YXUm+-|s?(-CR_!d9_Lf~>Mwuw)qv5999^GWUjjBOms zI+vs%sC-do}&EUc$Dd`Ujo3lW*I!Pe!wQPxDqMawe(bY7EWQcx3vF8ZFRwONyVEN$1B;E2Y0OV0SIuv zgCvhgCpd1JI$`gVmvNCDwK|=4L36^@Ko3Dtz>5QTb*Jwl*VhPjM&nAKAT)v&b~*Zq zATSxvsCG_$43VBN`lrEm9uel8*W$GOxSKo3&g5~mal}iq5hTM+!ty(Mr(JpYZzdxJIrem-kP2y!-(FoI1j)N6(4xSx%YxUsHn+=hZz?V&X zM|`OF?t?b*^z(2pP-e3m&Le3UIJzks(vj1;(F^R`%=vtg5V?I#xN`E-* z>CV9zveV=58y)s|zX7N2BADb$P#t?wfDFPB18^jZgID6AZ|E*~liJj}l$_F!V<;nO z$Xd;vFa6q_FO()AFTeggZvSAUXYf#GhuoqQnMA1@{DcBT`tyO_c%e(0D4+d6C9s-A z&nDdN-xGekrr{CT8AdcY?=Ub-FHuo0nKNwEYj$;^v>)U8DNsq(Vs8A7ug}WiNtc3X z;`f+&nP!ENWdTfiOJd4+S(be4i}nY>rA@hT*>=LreW2RVILcz&+pC^T$i^!DxTSfu zE+d|K^j_2aMzr^GGV*+XjA^O-!4F8kEhYl;vwsbz4q)n-C9?{5Iu4-MnXtK2VfHDU zGbij#NWwoL{mnnuoeqhVeBd{uvs3#|T-`jo1Xp8RG$`{B_?Ht&dzSoM(s$mk>81Nu zIzDtr=n6+Y@BYt^3(=sJenH9}2S25ENVjk<#meR)<3yK-wcM{LM#B{evrNe|dWD%> zW{Ecda2CqFEb5SCZieP_RlEu;lUwk-#2h9qILppvRuHtIeRvL6xqFIg*x;;tW!Hc1SyK#|c z=tDHbiSa+y`uR0SCz2U)^d86Byn$pXBt17DgY`U7F5;4je3>1pyim?k9kq~P~^9F=b{BZ-_(7K#^4x#x3*Jzpj*p6_+LM89rS+C`y1c^ ztX4T!ylB@8YIOP60_h4fQuT#afbB0m^Cy6_AjhrXPZ6sXnht|VsKmb0lkhJ zf_oIBwpg*w?7tlTplYF^!%GPk8Up1*zQ$aT$Ux(9!CpshxpNDS{H~$QDPGfnM^DI@G8TB_a%(xA;>d7&rgx{d4@ccoX?qPf9 z=qB#0xUi9fop$%+WM~7y&5@JWle*ZuG~)T6N2ijAy_{b7y9PX4Z&dj7 z8vA|JLmTwkL+FJ28+OZf7Q~rwhLBn2i9r8)Vb7TO#gW{OYce$&OH#ET6pUZndmEtF zp4CHZCQy7GV~#%cSGTyOXlh?RtB0OM=$9}IZZFGdRZ9*;}q=la#WXh6OYR-GLlTRWn-ERK}1`SQeg z@bBM-(iC$#FGc$vXZ;6>SLX_CCMFRZ`BQEsh_=D=%Ig@6>y5!afsJL(wQfazGh`@4&c4K*E16DyCrmVl zo3Em@lFF|_N4gJ(HP9*X-TJU2gwg%`I=+6A$3Rh{1eCAqzUsl7 z2cf?Se&96CMHkIQPY^VMb)z|W*AlsH61k-lB!yW=hn@OK$S1t+8lZG6e|P?qY(_dt z(5wqWtLZR8p0_e^3%CCWSL_2D732NDIp|eHGo3{~%{%CYoP1q9Mbm@7i%m8SH^@g{ zI5C$c=)GD4A}i#kJ-t%rK*0ohB=XYDEc8O@3z{c15)2QY8TX%2_W7#Ii>$;NP~NmM zz3@US;w`py;4>Ry51s$8jA=D=d9FGb>+vDaxFqRIC#28}@{q~VPKq#4o>P5_!yKD$cm>ZAt$pOhtGJpV3LoYrGjLF_gcD`mXtL34%DP_iYw<-9d5S0lSQGWiZQ_ z)DcJfjf!_j^|eh3Q3)V9?pLZ25p;BQ?~Lu* zHt~M{g9j<<^;xoA@{5H1$rE>O=9)%y>@Cfh?B3l?GqX2N+3Jw8eb$ml+4MfQsmg>d z;+A8(B%jrV|0B%YZ9z}m=27o@^-|Adje*6%-$t0%UY}0+FlM1?w{2m_dXC~}22^fN zb>r_ceM!$ZqMS*)5FoK#XPbv{sd^F)e44)67aR32r28ks%;ASx-~9&F@7x|-rRO^D}GQsI?Do&hLb5s9(ID01~G~f)~HabH7{)AN3Oy@!FwgAan~ju zpjxt4C-dmZWoTw5PUD(Xy46n5*pVoK?VGNoht_y>k`el8bZ+eWXIq3uDC~J8h#8lL zZLq??SGiA|ipCLf+Fh(EGLuyFvc6@1(PKJLGCXj`@j@x?NWMawF)drV*qN`NRtU6p zime#h?^a?I6x(U55gq=fQI>A{uwRky#g|`4G+bHjy-Epyew*O?_=(J@5c+5kzBQqKbJLjdyIs_uy^vp8I<_R9_Ts>BJI_o_kgm!st7Kh(%C^dmp8k(ZA zs)ynNTF-FBourYPuFK1vprsJCrH@1ycL&q`N217ixWHf60hNE<@0MR=_0TeA?A+U= z^c?%HCR9Q83H4_|CY`_Td(#rtPti=Yss&OU*7&1s}Dg_1L&V^ zG>qg1Y6IIM6_p~YeMdT;|80hP4^%lEKc#<(#f_tJno-&T%2Y~;0Ly}fjDL%#(mWa? z`<5JJ9o{DVJ$?r?}jCB?2y#rBfDpkGK_vB z_zTI;;)}vgQB1vdy)?>LIF(qDVa0?z%2`f=61|w^kV3%&>*$?5XG>}ycZz#L^*mV*rT{MKN7JILZn}LSV@EEb?oOD+$6a?8xQB8V;lGP z6jgEKVjO=81{6{%;4f?$)YrCo}W4l|^cqE*79&Hc)vHU17~T*zL164tV~NFTmX zvSwYPW6Rm961C~5Y?=AlO|Gry}CIs{%Wf6As27iPFd~K^o;Q?H(6?* zhS`?T-}yvyXtK_=zEGcHDCg71+_?weA*?(qp9OHoP1TA#4)%XOBNrTf+}ehC z1eIMCIbL0du4qy<~wWh-}uQ$cxdPy)7Wr&fm1D&%Qnh1%Xqep6IU~IT}*p zvEW|L!n~zFTdZzK;5a<;py<-iBgy~P1&W`mKO176Kvr(&rZjp*g6*vz?G;@GSs7#l zBp&7X?qX~`FvSHsW3<7zwZ)fJRcgC+=os3*GFqv;`*~oPN z@>Sg>c7;)lwA%r1P2H};(kO-JSv%C#!*u_rG|_XyT{HaK9T}v$&j--`PqUux4PbjQ zO8a#w|L=gaPut?~yS?$ly0GcrjIOMmI`5c_&~xy~<(D{+ln|*@ zdrRM~lIIr8OvljMYlnC-O8Sx5org}Tx<||Rn#3)Gt4@U6$LQrU-Nz_P<{;|+s&&OFy0MrB3}I&?g8z~ZayqZz zpGuVOMHskH_#bO9V-}}GGlO0Ox4uJxw=?WnsZbfMmbU;ZnV#5EW zl#rG672_w5qfi>9llo%IO~5=EL6buv&gd*9$Z#6-r83B$xsRY*8_OkTkNGx&;s4MC zu2S-KXSJwX;(Mb-CIdtH0ZYaX*=rVmLgvdaVvny;1Or6XIARDY?KES^>2TIgLGWC8@76R z>&<1#pTK_Diml7b`i)x`HH6?_R10=C-6UkWrfVZReBf6E8S-ag%)#3$1B%3Xeam=v zAz;{p8U>*eTW#3i)3Hv19}M$(W#>Kt6k=(9vEyQq+C6mxTMnz>Kw{?sye0N5eHJIk z_A$0UrjvXx>K7h`{?|w%n3m+$?_auOd5b%;TT`}=%bc~oMBJ@lM5NJI)1pCA*~%DB zKkYioHvDc?q!Lu#OhrKhUy#|;wgLwWjB?vfu{PrdO$>dah+rcwUn2z)=7Dm={{D~L zn9%(T1KbEFd&-N1KA?@`V~9^e>x%X!<}QhnV=X1x!4sVmdY}&q@e2U$0?|d8s+Bg% zr>L(hsz-rgJX4guc|%U0EI(+Y_-A|DJ?*~3dwC!ZMJA;% z^WuF?cEp}P)~_OG!^l%3zB18PvlepT51AF^d8S}jSc>|*iMNAvAr!C_84zPO(SpeEWR$$oA zkHo&pIKKgvd^;OczIn0@FvdZ*IDF3A-7&ECoVWT@dpi00CJXl1n7H+yM(?m^*330D zVL|VZVlEdn=|B>OXa5h}V$Mk3k^U>~sw;cIpUeO3jibZy0ycXOv)^rHbu~Sb0ns7} zIdfzIXTlvlr$*cM=V^%SH9}ZdVMOS^8rD<@%_Q)4Fa55cLo5~A)(+>I`($skD$gGzl4)t5L9x3hICbJ51LH{IAANHOan^l1k7$1@!zgpI zJ~sDyx+kD-9g^YT5T)o5ij<8^;^3Mttui`$-%9ZB01K-JjKEH*tqmkl;7-dI5 zKL0XuLkL0A*PpTy|C&=E5ZxYh37`8Z><0+_|B@?49=a(Bhue}VGHHENv@AL`L4ikk zB5QEqV0*>x_-I>i)^F>z2v(<-87l4Nd0_RgW1XpA*5i$0pT=MS{L*$YizJcU{r({f zYXl`ka*m7LydTg&PTf-H{tRD;n|Pc%C|a0KHfbFH7hthr$hZ&>Z&LpwFD{z%e2TH~ zemHQIJTT2VYEAt###`At*1mm89UEc8ut`MlWscAac6Jlyj)y~7s(`r|ii@~G-`^c; zWRTgY6gBl@t)0rfnn^E>N>R)3ucDTnZ~Z-Lo_klX#Y!_co+&TXZ5mlnSnsHl!zAa{ z;8y>L?uMh9=Bj=CAIf5hogW@lTxMp~n1>h46!t{SL&Mq38C^}VnptAkQck;>F$I3N zB{X?VU(V#M-=`vRD2LgQdG-H4#SYT0AbTj-%00=Gh737F(|6^J#*%Unr zw0gnz|Cp8`H6>CO6ZR1+S%I=tN0a`UAB36Awq~QV_*i5)3}a$ybR~?e(>b#dc;YSHu5YFDl`^p$`=PWy0!Yos)w${8_jWfbqY% z*6kwn^Vnd8_+;Wy_c2bH=yxbk)?k--JO$p46?py@`8Sb_gt@W;|-&DXOO1ZF&4X%1P!haO}LgN zr8J!zZi&-m(MJz3+wj@6cYE-;YPY8BR&)w=Y)-@VY6(zK>* zG43iKDJXM2cf)U*cdKu~h1U1JiD%Z9Z_HZb%=dXfMNiF?7MR)w)EU~Azu(i}e?<;J z`n!!X60OMHUj?aCg}8s4gE^u!X7-Ds65C&Etp+bV3s()9R)Ez)OtJEy9&yPL05qg@u@8dI6vcWelL)|soo;a3OXx1-%8>4bK(bVOyC2bc! zGxeqvoHoP+D+U#ufMM-HLDdl%Vu_RHS>8MiPWIatY`VYx!$r}QAT{ZJOkXs0-2R_% z6xLLGxz-W&>A3kaK0j{H8}Gx=c)mM8Z&tXuYU`J+t_|J6wE!)1O|iTn5IaGjIw;o? zicX}`Mb?$l9{OpHNH)ah(H#5UETO873hh616fu?_`*X*<)N^&CR)2lE57nVAYe#MW zlL1f4Ej<{MvLX^=-Z9crzM|L2+7W3W;062?AjLXt;8+S00eyvj<)%uUuF~qp`j!T` z30JT=*n*hwA(4bxJrBN={MZD~gE=cn9S_5G&J^bze6QF`zUkuPAOe80;l!fhK|MaEq{3VQ&oe9$yHxB zv;ExKbQPY1KGcM9{)%1fCf~xJ!C-pj;OY{_b(_k{T`V^HA=>T{t&d^#4`;N1`>BKG z%I|01bYtU|C)e=43)4oa&`1EFQh#bR^-|nC|J)WD?LrUZU%;89Gtx$P4%ge>)LlG^+V`Tb!}_-yt@$ z;^6`OiadIbn7@#)T2)?+d}8O-Ns~v8MNg+?CL2$7!$imj_iQyG2#hh*-o)n!x$CoUgx zWS(t|AuParn>;b52@y8clXrBjBmM$|E&Pz6(u^5xaa@XsuV>9ApZ^qUUlKcmd{!qD zmgxhiAI4;VBQ~NOmKegef72TwyS;r|a_(1wx!2~vA@=@+_5Gyw-73cNP+SR{*|WyZ zNcbVIkb=Z6iJf(S2bxXMOheMsO|0-Q)E$&Bw2f3sE1{_zs-LYv4~W#NMlfDscW-R# zG)};4bxOw~kcTI!59Myy%qqd46cn@?$4jmJ&$DgquPZ;3%roPVUw3a0Ji;snv($VD zC{~X-p~QFmgXicq6>2qufZSowzf%wHf<>xAjYGDAb>IiWf}eAFSiP z?xf7*^wshsK4De2!xL-EKKrDKq!XmV-F=qjXG*hANLFJ{!{+0USvjLPCk*Cpy86C` z!aon12TXwIsSc9PA!Y zXP+fXzd%{<$Q}Pjm9bsFj=`GT#hi2;UlwkZS0D^6I`natE!I#SX$Y=?=}-wJ>mpt{Rh0PmUIvZPh5nPfPdssZ z**JelmR1C@KJiI_CXHSkQ$XIVcI<$zAsna1zkeKxG%9;$_6`v|D`ynwSB!R0nbmRYXI?~*fvF+^0a%g`;LC~Ajfi3-O_ zW>G*5fq9sW<;UXhTq%NPEenLzI(*Y)l(aeh5_N(F-LnQFE1XYCYd>??+xVdF8@Fpa z=ghS|U`q8H!PirC3zN9jYS8wL=|c6^?CJ#!={L@g&w=SuPnMXRw3`7t1AUcgr;m)e zS7lW{3uy5ADPzMoB{nzK(FSdKr6;Mxc`Iv(Sn-w!`{9ALFnnuZAdPNk(_X+&9d}Ku zZ?-#i6)E?}oVTcdR8(t!>4|r0w42~euUUa1ZXRRuA{V`Em9ngM3mQ=4jzLo2nz0ba z_O>##%^qo~jpmx_q+0W2BR%;OpHpbd6#3TCJWi#KizQC<92JVlKF_jXK;x1Cwp~y4 zbPjl;cTn;4uxVj}+ zj&o83^mTVw)#u#4%WM#T8Ku|vK!X8o$$4|FPP^Xj6FidcyOvu`b+VlH__nCsiVeRe z^VihounBiUUhli|Y_ku_RiUrGBz-INe~*H54c02;gMSRw`s=3!3nCNx=}y;8@lpvK z^~CoJ%m%XfeZkij!>15A3Y4)F(4iyWL%>(4%ZWi;Tp@c-dtm^!pZ^&rw_@&)bhFGy zCV|Z1K@m9rig9>@aOEp?yyN@ND_9Cksqx+6Q2mxO#+i(3%Zw{j{gyk%zK<_%AsxH< z%E^Mx%G$70XMtq*xsLNac=73}re>M*>>4KH(|YOIez>KHi@EdW&{tHi@A+#E zO`rZHbdHfXH2cEH6V&gpT3d1Q8a*@my$BYCo=C^P>4!`u)si-Dk{R3CkcS>`h1<0Y zZDRPVEw?~Q9Vc0Efy&LF;ds^>n_q3fq(4#?)NOH@b+@eg0Z00o-KVmy$?WRSpwA~j zFH!4Nu!5*Bj)Eb5jZFc5%wJ%na>!UU;L^vN|2{3AdtBBjzYNCgH)jNCck79k{-KRK zh4MrT_(xoWA>#4v_&iRNk1g(1J7>MVF)v_}0X;C^*sqhc=(WJa9Ja->q{N~!#YZw8 z2oNi7FT#95QxM~Av{eh<^nP=OnA$yxJ{UV<;`qcmEuWNJD0cY^mRm()9K5#8o4s?a z3;(NHuQ43vp}yp2*D9aVV71=@uT?A@+R+C;lwy7wX}P@SIpfb4-5HHKcE&v*UA@(! zG$rSZJx|0a9$o3^q6YBv(C+7TXpC>rN#^EuX#B7(o0nH#uT~Yi*xRU9C8q)Zn()!t z?vUbgA5gp)umZKyS?pQJW{`pOnq{?HmZ$Tu-9pY(Jn)vG7#Waqbw_pcYQE{}o|j!Y z@byGbpsG|iBgxem@5x~);^Zz)6Jrf>uG#aK#^M6r?v0ZH73IOFUHF@~KCw_r{3nLk z>*wMV*>7JKs2ooW;i@=Kn(_`$!y|9ey;Xl9?fbzu=HqvYu9%3qMT}DeyYViwineTm z+Ha-%DQM0Z$O(n-u5_rFN8q0TTnX8e*$RoIo6moD9DFGs7LYlPSetcg-Nq+F-4e%w z)z^D>JwItlQu;u8^S8N)5lU2{S(%jKrb!9z@7$d=z7t_SCw8px?@+j-ALw77z>JwTJf(X9mxcRp3XfMggx>){?A_>L z145hB*UW0eZP<0ry;3wa1%uP#mInyIei`!xkb!stj9Xeko7hs~-}L_jQDLoFh-F=} zJS${5PYk--zbZv;<=GU)Gm2AMoqV~Z3jLH{;GLnr^~E|BkRs(~#rn6NH<5gYw3?dH zXLQvqZyY7x%i3qeb=T7)Mc=0n%^}e;tqwXdu|9PZB+?2<`BgbUrq+%ydc-1xQC=?( z8aj|^^0KH=XH<;E`DIms$D}pB$6l;kt)5H136(UY7S?vcca0c(`aFW>f06US!lbY!CAUFNx_-6 zvh!Ubl(X|mIA_o8=yLEKm$0$2j0O%u5=A%&@C(SaSs2gR6>CCSAI~=1mKLa8<#3CtLeHBJpaz45e>oAuf_E=;Ip0*QHIy1_JntL&a8T zjlSPU!F0x(DJhnK?^@#eYHQ{1TgGG!#R+Kw=TFz(mL~!xSBA5Mu@daGQykS$sHt0& z9k8Itf(g3*@p9OUk%_*JSixRA{Bw)JKA(GS5A{C4D8FI1RU)+$!{p!0al#Tf8(wntTJyTT7oSf6*%9 zO>&PTcHcAkcVBNEIUR~u#DuQ>S3cFvsrV0{ZJk_mAGG}dsC|A|?~5QmM`&IU%%&$4 z=v9Uojx^4nGA>KzMK8#6?2%F75vdAwzeQ5Wp)3R#n5Ref8B&C5<}FP2Z^HaC(9Kh{ zouBMuL#+C>qqp{-+at#Yh;sA{daxsFHt~4_?pg553>?|kt1Q4%93U9KWt(|fs}1l> zdNV|N?t*>ug=eff`_nIg?SA?@ov|ky9}p2k9H7c)_xr$d$Z+Jg*m8oyawKtD)Ss5o zs7tfEb10kHqnRNpF7Eg#|87O6aqw0DW$`KMv90da$OQa)YhHfoilJ=Oso#F{*VQcQ z!1QPwbAr2Q_>&Z6Mwdy}aV3}Oc7uds>1O2Kw!G=>`sq*V&4iQ3#v~n)Ako&uqTAmy zLM`o+%!7`4l$*slbI<0%7fBB|xki>wHLGujYT3Q`SV3nxgH1Mrz6(DrayWBz)u?5k zByoI{y;L2yVhsUZb$=cZ&)karSuXI5u{Z$`qet)bKF1Qfu)kG0CfsxXIm}NcJHsKbKM^~ z6Qn1y&nwouG z5IPz_74`54aY`I!bZrWfg78wNK&+FHL*vBW5VUt^+44+!^~n!@5b^vB3~1cFfq1fG zOy_=ihd(zi!WHzWWw^iN)2;DX!rGjHe#G=g zzs1XpcnjBfCOT<+yETc#ESmE%ec9 zJp3Nn!>$w7_uCTpO6NesC%t6nrDlD-r%pEdFz9XSRrg%0-)pmL6ofzL<1gOni__j^ zdElDUZ+Ud@B6AG-ro&j+QM~UOqU>&M+Pekyl#5gT(#j65EMHpdkR`y^Rt1Q-N8{YH zD7%AKO)rO_BMUEwh@)6_^IyjxE7J_hE_;Y~CxH#gCp^rT*Y#PCIL{6lylo=wQMAp4 z7Je_S?ox~ImOM-#2UA|P{f_;ceS#_5wG3vE<8|gM_DB8UD|xCVjliONU`x)FmayWM zxnG@)AM)X!@wW}Jqag-2fppM>mkUvRVXk4(HL#jQ=Tc!Et21rBsZ^%tp)tX+Xd%iM zW6@o;zI&ZE*$2Q{+)X$p1~>3`Jpkok8c(0wnMdiw{A2z;huoe_zP4FT8Fi>A@!hiS#$IE<T_d~l0IeJz z1n2+ghU!ru8V1jsz9m-dJ;f@|k&zcbC0u{^+2Nc_{u zUA&7;7bBM3u`UZsw{*LA<@Kt|$gFY!FbANQ!;Ut~x)n}h`}~tL3iit&yUkaH`(`X` zxJA?N0Cr(H93px2?PDXKezi)bd4KCW`0=5ec$Ns=X0z_6Da3XWflQz4)Xnzw>ZtZb z{(bfb#*|j-04PE-u;?PpQwQ{4fp)nOdmY8-ehQ0L1|0qtCF4BURoz3Svk771vh zp34KDb@DL^R9MiyBC`{A;?96Y9EUno+eQxy7@JVV19)n`@cNO!pYccT*@u)nJco9z z7{opnj)MArZdf5P^H>V%kDbxcYv}56g4?#-u^|p%pTJ+7Aw0BpK*#`|B8lrtZ#dR^ z{%QR*qS;#bI;=& z9I=bv3T}e5B-|0j7r3HI{F|)Jzoa1mILJqfuYe$lQqpY4*9cLQmNN+w_^(cBL(f*m z)q`+hYee%14|Es`{U;+&W@%9*duPB!t<7V;mW7W+GQqum4KE=RGo11 zvCh-CW5IlMSl=N6*oV}wNy1Id(u%X`DIVmx5PbF|y8;(;om7X`{=3-de}|~&F(qnR z|0_h*Xpc)hc_lljZ}!I>B^nc?~J zM53F5ykOO5-*m+Q59Z&rq{%P8$NyuQYW}ndUZ$>Jhv07w9qhJQ=^p|?)~EHQG3FF{ z+XbSY3E_DdiEvDeFR~{t!%PvL%;`|fqLo-TnhIhpat~Hh9`lqE55xde3+)xgk8ES7 zSKMk0GLsNQrAKEyJ_cZjv$w6ngzU7LE?1lJZYp)_0e(wj>mq^!Q7Z~;OKfYwsCl`P zIYuNKHPYbYUkxpME0BuAzNl4gG9I^xvK8}m?{f4p9n49D7PC34yRAi@-_0FZXR$>< ztyDo6r*ZoHpS6mLVW)OBA%*potgpJ#;Olfv)JfuSN6(3%A&*YENgN$Vg4?qtr!;=8ge(ZAL~kC{qIRra#^09j2N0YGIF~9qH7^SdtypuO?3H7_$PJf}MjxLJ8c z)%eSM2GCGp12&f@HnqPI3bDp*8^r&B!R4j89UEthx&?!Kq-( zbfk}Bt@+E!B9NU04I;N{|{Q0|7D55+wM0z5U<6a^}I99>lTe* z*fjVjZim%GfjR*{r&eyDT#_{=7r;&aE7vU^4h(l()3_1#9Gk12eB&ewc8CJoz5xe@ z)h|O2HPh0Hund!|P+4QY8-L}PtyJr`?v}V5x}!ge&(Cj@`qjC$*k`VbF2o9&Ds#Cc z3tvl8`cb0pGTDr#6n52a*cr|^L#IPv0cmYlf-RfzT~S=~R(y`%FxNc~OXfXIqnI65pC7bWc>r0~A@9aEtw^@UQ_PDVnD zW9qbHF*W?Kr#(7KaP9!K_Wj`ysyBn60{W)~T0Oe|;?ZHytUN(LfdzaaRmgs&l< zG<|Jv4Z%+v?rufS9;+#HwdIH{X}ON#lFH7PrLkZ`OE z&0;Eg5{o)2*IA*wR_Cfh0RXp2vh*zk{s<;%Lajj3J40mDN%?dS@v)B`{jok6NaaR& z_Kqt}LDMUt9JdS|i(ph-W6WFSxUtq?qTGxsYOZ7|iHtET(axcZ#W}sVnd&!T@B7ov z&{=ut^997(J~`HJ$$t$`JeshL*?)k|Kr!z2yf2CqI+4fV_t=yGs+s@}71)_Cb#uEJ)X`ih04lCen=OxbGvzKv zR*y^f7mGk295_ok9~jauot8#VrT8*#%nKIHGPWp%wf_!KBjp9X_XEh_Y>)vpGEBnG z8BEroFDOaQI7zK?5{*dhC^oj*W&DK{lpkfvY3CE9(}IQ zpZdoqVRR3^d%ijezG^brBjt*=HlsSoZXzB~ zlJNm6V?=Xep4p&W;~Kh2SdlCen*>F{)d#kmiOMVdTjA77TeY+3+hB*oXU-KhS%!F9I+-WcPr=}l~FziAhT|45uy z`UL;;;0fW?DoYKDyuhZ8{+`?rFy=6R_hPqLG%@2)e`kV=dGA5(NNEe09L!iT3IF%> z7GnJOX>t&EDXb+#l28TdhJ&V&gDiXXI6Px}^QR5T#Q90CJm$8G?Td)0-HF!9Ey?uS zre?*RNe4$|D0ti3t&8%>f@VZ9`5uOBWQ%~~3DV4m+XO z$|^VFn-q1Hd(uBvcw%;HQ2_D_pGY~CH*Qx?(tai7RcNNdV&2SJD61&yNSRFm0U4f0 zox+`CUZw}=px!nQ|G+LHATWHFSBzEqpK}S?1;q_z#wJn2@Eo%df^M(W6UomxzB*JB z7L6!rRJe$cFz424Az|HLx9J-Hx@a8^&ot=g;~4U2@Tj~t-HNx^LRWBqOnoCDD@Lfb zJ1U%wlGq4*e3Gv8Cinv3RhfX1c^&I3#;??^%y>AF;En*xK z+duc3mB6fic*9UmsuiYoZ?d?R z1New9803c6+uA4E*Ql3{*;<;(N7S`feH>@FjuET<8-0Lxeyw|AZwHC5o*t9a)i0z8 zQtR(}wcou)hC|D(6!_~m0T9g1*O?NU8wJNm&g#MXaCV5K70L z-v5*h$7Q6pORMuOof(6l(tP+xF22D$&iqfJ^z6^QuQa{*{v0Tb#A8r)uw@7kGOErF z!3%p**!XR7I(K4Hpn1pqmtFz8^0`g7TmU91%*8cawG1aOEh%c%=D?0!j6Fh?G0@>? ztC0sS;OK_}_Rj~E8k`^IvVY>!ib|prBUm-xX5ZK`%HqEc*-Dl)ucW!p-Peib1#Ag6 zhw@le)c@(^Y`+FP~A4Bm!N4%fT7O&DdPb9>i2r&Ee)CU~p<84=#;HczVG^^q!V36b| zw3=*#Z`7s8x>b%N8Uby{(n+do?oeo@q^&=HqcwjQ&?d!n+Jd_>xa4f~;Z*dz>i%^Y z^d(7gGgTl@Z{1ftH-%l7IDM@SSxy_zWxG8xOUYZHP7=%MKJo)7c`G8aEvAf&LwUWGGBhZ!_i^z`A7&=yYMqq4D;0FMQ@%DR2tcEN^2N8?;tgKX zN5f!VMJQ#j$byec$MpIcmx{m@Wk**h%Rt+GEkkp2|C2V}jT;u^*!MU^&Tr_BRB#ex z(4p44*Et%>4`A;M_?@aG;c_)Fb$7?*izWEnO_U3Sd~qD|(vzekJj$|qQ#@H^N_0fM z0TztaF8bJ{{s}x-$*5RYd0kr#a$){fL7W% zc=r=Z*FUi&fBaP9iy|tx2J$2mNwjlAgLbNphJWL`nfgBS zMY);3BS*DXi-%u?9Bq=G8_K(4383}53JW-*S}B=y0=K>C1<-!Mr^b47?%}Y0TwHa+ zI>ET}c%f-2K7!x>@bdoeYg0REgvy7! zyDLVA)&u@mo!q0Ih|g_#ODKP@N9*9WpGgHk@K1=2*&BVrE99xCIgwM`Rd|K1DD5K5M5<ID&;K0nV2iI+ypT|@N$H8ryv$$}C{4ZMEcYtsP=$@aCgp1p|#;`(n@63kOD z($$ic;8FQyDa=x92 zQpX);*H?bxlnPV;~4wol%hvRyiM{9H%yRcrwY`3p;v8$q8vWd*o%?nT z4eG*%$U3ewICPX=2RJ>&I@efUmOy*bJGT(5=WL^5c+xmW^lgZ*`y*Tye7)S^YTbkN z3~r@CxwPIjM7vbj-bB7+9PC=$WifzQm)&KtgCu;xVMJZXhUAu9w|08^_&p?ubjdh5 zrK_LOS}{Zx$LW-A5#Gw{<`mUyloXW!;@sDtSf-dHuc%SlbqoS??Qd za_Pwo7=ZIeTR@rqYQetQbU1XWo74=;i|#L}kLOtp*8hkD2_)`xn$0s?O9p@((|iS+ z)!}ee++u{hqX%TdClMv@6XLazf0u0OkaiwndjOTj06C=r!_Y}@BA zW%nz0^iNhVA=CpQ)DseOb6wOPQ1DAKEHW<-z!19qnldV+{W>`kcCI6=Vyg_x<*Ly* zA`~Ccw3Ua$uVmcL=IGo+%eZ(Am3iCFu4JVcSl3W;D>OZ!m`i2+$iKoWE1?panQe6T2?BzSBj)raRORg{4%pVBQuCnp3Uts ze}^i?|M$Kjscd+F?|t4+&=Bkx$A$1w*XRbumW(^W#>w*Srg}{}?3)?yuKF}7*W^^F zTMpAGo$S#d2InwvBK7wmd=+vML$Fgu;u=TlRzW%X;u?bObwHYX6u_)8zgooBj@3T^ zDh&^|X5-T!)KYwNg*b*z(0bm7+qbx{)(a4Ez*$YYcZ3$!2^G}oRQS$3WGzVE!oQxv zI5}>)s04O0z2{`PF8=VQTXEtwqn)*|ovMIIyl#nnQdS&jsn}uVxo8f(x=3}}A5m8U z4+wVIAh%X`lm>06^3jqG( z*der#ckm6>%`?UB2Cn{Ll08>gIH6^nh{2oa)?!>(D|&iWe_3L$Y@DPhT92l9k`l9i z(EDMU0Y5~23-|A7t!8Nu4vjlysl(^8Cr9dY!uQi&0_O`p87Wl$8tZYu&rXC?@|!MU zqEC`SuxHM83GqwVW)%IF@ti^U&5K-i*q40hM?_yTk`TND-W4ft@WD%*&l#dG1ED0H zq8JCiy~a|=W7F0lI#vHD=S50g0t?QS;_G*4^rd9*PMpK%e%IsK|FIL$Hj{0iV7vKS zHRaxzJgYZ!JlDe%a>m&zN&fPS{muMB$#9Sj4C9bF*%z~fY+SYWyA%)$YttP6g{cj0 zTq;WX1UsW*e9$)$y7|1&W{~s2vG)FG%wE0!IjNrq5{xZ~IY~8{KKXi>D;4R4p`AAI zV=cyxR_(dT*A5rmqVt}5J7~AYSHi8PKZMFMX@vbsUvwla$V{h+fg3Hgk9g9bQt2$@ z5&dt163vzle!vboy~^kEfA>=i6AoCod<*pi3DfmI^+LTd$vtIVuFf&(pb=8+=qpKm zWGJm~f^vADCj{c{u;+FVeQ3?4D-dI!Hs^*B6_FXITSj_UeLH+fifrbq{Z2N?pPzGXh&0@>oVFa(RED|kfu%R{bjnI^P`@ZhopgAuMFcCuyzQ9@*2pm+gs$3Y zGYo5nQh&jp*Fn=k5VXDtBEOM3nW{^WD6d1DX#F~#K&0lgN)zF;g(rIcj|Q0>^7Chg z{gOS|xu&-giV{p`-9Z|bX0*%sFCN%g+G{7+!R-#n^85gQ{3c-K*vcF)=^)2P$}QJT!_JU8|=`{=IR@fbXxD$;*qJv|4d|Vg}(8y-~ z_hcg9Zk+J;?XFnteK<2#F%u@UuPHe8aY`Ufy@%yOq*^;Ki|BF6pK*7uMf@*7qA2Uf z4C1Z2ZcVWn&Yi9*r)RX4PZuG81OkI0Y|CYwp|nj&?EXAR!0e5Y^b~wU5Zki<`aP0< zK}MrHskratRu@AdtysgAsE7!2XGbCr&R`9NynlAUn9m56K+;J>k_iXg&H?vnL&%>T zVR+?(^_#^7o_$NDq!QycdPnYi2T-ooq2Bc2+*M)m#ICRs6e5aVL5*V{c>-HKgb0*B5k5pZyQ zm{YH;@d?HCo2R|bacD7db~@8PoBgx!Lyg~v?MqcAiJ`8OgW2sbGJ|YH3kVs)3YFgq zR1lhdq73fadC2@o$+ClbOe-mXKSU{{7w#8(Y!jKY{ssijsY(5I@CFLW30IZG`5)Uz zxHW&IRy>c?rToDgUlul?!Z}XVP_y(m|};9oJ)+{2?aM&ED%J52dBo zzZb@C^OLU5Kl9QlW^|Ki+-+4Tp$Np36_Cf;MEP*?NBe^=S3pmW<)C&?ylKL_zl2f{ zZrA}II00{`j^M=w`MmtPyw$1kfb1A^xtQQ-d{7Lj3WYFbsunH)yNO?MqsM8^Ex2&3 zYG}4hm0)bOksYP@u<0nr_wgBgFu>fXIk*ALUer1~4JT}|pifw$8|M}MJpA!eHNBhU zVU7Uu!kGiP*_h+$gm$`)I<`cH5I!9;8LW{q!@IyUaP$e#*|zyEx{#cyw=D#szQ}A^ z=~!`sSB5gzX^Vwex`Y2@J$9$clzLxBtlp#J9nc*LN-sp^?`Fj!Qu!NwB(cjz6LCn@ zLY{M_nMg=5ucth+XNN)Vg}$n|KfIitiC@ zi+%n12*oK`6KM&-DpUygCdxY6vsB^19vIwPV$9z#pubK;m>1vGL40hze@6A*MS&|&z;z8c5Jv;X7ENmn> zyKeihngw>Sp#&6q&<||+t6@G}K1alyk4V)Dw%>*I5QHg)zorUQw>;7uMe$ux;6U|c z_nEI2<(ILvI|sqxNJSD*Ii$CK1G51r7-@|dLCa;ve(i0>Vo3!9UIdbFHcbM~dcF1w ztwe(tl(~xE6zDkr#IaaP%2N=}(-S6*OIv>_-}kcXA$hr08`C*>mwbcmI>GV9Q~sgWvHh zQe5M|yEr7bp_E73nTQKcA&WhuG!kI6yadae-6r7M^|7L~;=296@|-#m!EK0z3Bt&DXPQsuTuOCX5|9In6xpW!5^h;&at zh%;H?KoIE||BDQS?GcTP`^F_d?+8&~7u?7nD6MFU_vf!b&~RD!@Q>j~#CWKX{wYr~ zVkr80n8ye#?DgAVENkkAzc+PZK4tO-SYhp%DKA}Ifyi6aqxZO@_ww>-HtyINjB9#y zl~JOO&WL?Zaw69#4Z{2f1XSNHkpno7;=#X&55UFTPk&eVv-E6=Md4C7jjW~R8s)3Q;@ ztegJD3LA{b<%;$j#eKh$Fx`75E{-5EaS^s6@}fPBHr18N#GnFGjWMY}Z4GKtwW<=# z91-2>ANFX{%f~E_C=igm7^`gL1QMho`g?KkuEJa6LtKI0cTJ~1h6Fqii~7t@$m-(o zrR(9@S?ZzjV@}edcwl24+s9*Kb3|H@-J~V|l?#gBi0Nzo+(tm0h@rG1z54mo7sbA4 z^7WJrosQ`TI2sX!to)al(|*QTZjt?@Q1qLui(fDu=~7(HK56d;|V4yJEP4mAh^v=W9@~Y4!rwnSc9uG zma0p?H?|_aX*{UJuhSJetmuOMP0-5$TbsW5d-_lI;1(i$;WO(`Ir%uY%)iv~F4_rE zu?CKAcN%9JY`qb>;ALJVml1y&>gd1Vl&W0KnbkHXIB2%3Rb$dw*QZUnC!2#?_P!(c z!bPYrQ9>w{tvJVYsA%95Ef>MZsHL3Z7u2L**=@OO6KKg9TxEqnis8NHBjcvk)W>@^ zLIuaz3nz05n02HtQgwaAYR7!dpU;f0_|ndx82N-)6xT|;yGFLmK(AA0^PPV6_wsZB z%=Us-7L14~FXGH;u=#iBUonw(?|qS0Mgf7(vA^7oYRiRB2_3Pl488<@Yew*3;>S&! zQ|$-#(pI85=p37TT+F*R@+CTdKe;m}+7y;Lla8$GxgTJ>Yuqc1BK|ucD$=b%p7l|M zd9j?=c)?dk%XogD_frKa$ikL=+}0j$-+Y=IKM$C}JnwyKqRz}G&FFCfnTcZVWEG?< zMc5=zUC;Bj1OXE_=SX^N74L`l?bH; zLOjO>N)ZNj<_2=mlOf;Yh6zZ?5Kn;VR!KAFv;#l({qH9-wo5h`8;4Pv4foqEs870R zH>+Qqmwt`Y@i{%{2@bs@PJLuEvsgeaW2{--kze&guI2|{Ud$1 zL*h+4Q{GkZ2?X&;&=(*N&Fy*_cIA%__|pF<{-n@*<}mtcff4{G+I{nAACJly{y)Jb z%8WQ^-T*0DMyUkt52AK0eQ#{BT*b6p9rQM;E1_wc214T99b}&oMsrK~17*h0Yvxt< zz;!6=<8k5i82SL&N?!|m`SyxiV&dtlu*Xxg%rLE(1IiMkfAeL38EzcAKWTkGS+SHT z$WJ#Oi8AQ4DsEDfcPH)M#oT1afoO}lJ%Zpve8k-8K0So(I`JOfZcT+)L2lxWp;Kpa zs+(lSYMW2IV(vbKuRmNEJ|giyLji!SuUMW(qdYa$Mw@$N(TRL8Bo&mMJ;oGxc$aG@1-m`anK83N^7nRxn{P9;}TM0&AaV@N&WV z@z7`dY`_rr$}Y+aKF5jAK2q<$H);9Y2^K<{`s9fJayGsy3%+k)G9EjjIDW+1`aGQm zx4PCBevO8{qHY;)!GsnEJyxA@4$i=;ELjY}ZI$N~R&& zP-TOdaY817jkr$3Y?%bxFI+HkU=&Rs{*?4kMi@b@(avYQ-%xI9IuRmAScv+>CL3zE zs0|AH+VH9ISvdJhzM1fyIzUF;Ig+71^UJK;4{pr0+YU|Q4TFN81f}0Wm@v%#YHAJO zgEWJk-pb2qi#sR(cF1|51>-HGb^mi(Mcj9jwQNqFaVX>EFfPoX4Jii~X8fj(Ow zA>5!2s$ZC}_c$}ll`w6@v(Tu}K_5q~lC6tnKkS;7YjFssb;9)TMN~lo zZ^uGL3`&JZVBDPd1N=e^oNus)kDhb9$6U zdF9FD>FxTsmD9(DV#E~P@C|&#rpxI`3}-kN(PRS)4W&p=|J8Gt%v(m|440NLPnyhM zx){SIg|*DmrOfIj()Sc9cTC@f?qmESv>UxW6W4M!hkOoQV+)G%{WrK$O&`?(uHnwI ze>Pkxu%4z!RA*w_^8U+mPu5y01lNp@z|fPNR5BZ5d(C^=Xi3*56ZcMSA^QFgxi>)p z*Yx&sHLMw-Sj;tk>d%d4PmrNhrKRR2IhTex%9%*Q&d#9>zd<>u=lx+DZeD8>R)+Fl z+yC0YEB+(|mq9xFnL2mwFg+cuiev`DjX-SnBFh-bl`D$7QTPe+;)khB`MK9%gi!5J zK|5x=r>vgq1}i9B)NhSNWG?AU9$0XHSS0{xNn2Lagp`e}4Dw1ytq=E8D%E`cqPo7+ zNcUph>Jv)p6^m`Fd9TSEkcv%89WnHNc=v^TP z=_KqFIr6h?o#Yk-j-cgCp{T7q5HGIPFYq|6;@n!r8^L??V{M1XhD4h|zf-pYTsbni z9yo^zRUC)h+2ak>95Q*_iux9vX0d0nYzQ;(RM?kO+z9QP8c1Db+UTlUYU^bNyF^aG zgKp==wBD;L2HEr8nU+IR0#D6YkdTY&;aio3hJOQE>C zySG4bC@#g_9fG@4oDke0xCZ#r=XuX}-v7)wXC}$4y>`}~%p@~A>%Q*m`Za}iz7%WP zV6IGzVcZp4LJIu91CUsa6kJ(t%f2=&DguqI>eC4k97h#jSUIw2gluvb_2hg6){+vu zD)8O=hn+XHn3N9%#BD+!NGTr*?l5KAW&>H|O{XObx_L>n)%q6et85uhM zV(7LTW74`(s}GSZY+B{}N_bZu_SL{FX04+a3^XXVej!2+CFS~>T;S}sKwGP-@M6^x z>sYicTiN|LXKp}hq(gB5p5>2Vx0&dbPUWfKwlf;s#6rV48yf&2>-F z)oN<5?y6OXmU#32PbE>JBuzTq(R}Ooy7lm*>LA)Wdeu;Q)dG%}6#KU= znGW(Z_afSB?wCv1Ynnc~)aTn0o$-iyG)RjE6-C|T8;iC5Z}7UUUi!LLKw48F2K!o5 z!)nTc0I1D*j7Nr%QNQ8}_P<$wX$(^WxNZ8I)9& zPaO0PPz%1fHAjolE6Jf%)p1QkE#~6LhuGVTKdgz$Ctb0s87aK$Oh{JYi?M0ym36CA zw=c^<3ZHYyiGKeHBx2#KXv|h;eO6qw#W(wk8iD5AnREOsA5e9Noh2#{ zvi~B{)YtFHSD(gY{Lt*&iB9QQxp;h?pv0#fUy0hb=Ga`aA?7}WSSU9_yv9R@zJND{ zSB`%vNJc^Ayp+!F7?cZaN8D{60gLV2(N7ZGlQj)&DylvD41;~>ggyjF_qVIYh`XA# z#@@?hwd9n0X^9wq8eDD;mLI_5@4yDq{S%#AMv zu~s!*ILGS~z~M(*4~++zD?@|4r*Ou657a|Us4Urv5s2a^M7xCgE$~vUO#Nbrbb!N0 zlA2s~ZSnIl6{q2ROd^I5KD*zYa$Zji=45NO5f$&61w__8wF!~2GLg|G2#gnSN+R0C z%FYO<-JER4rOxI)z{GZcdeMymioE!1e@o?-^%oNLCktt*rt-hm>yH(=-;TR;8qL8U zfTjj_W)}#E`2VilLF)(0pyen9frt)K&9dEb)QG*mvyR?KQzdNnp;5M*KNeE}ReHxV zKEz0AuSrR-an-k|-VPCp?|lURhCV)LRU{4(#;@gAVtQc{(jvw*PxrJu5L>XDJ+G}t z9_@}us*@?zwlETnSMdxahwQ{5!jD3^+S6M&O+=8LT)N^if%B6vZdPKUwGPB=)!~d# z-Yg*8HJ7k|2D!t7ASxKz?1!^1wdqID!5ndid*1ba{i1KVeA5rDW1AMZ*u!_(3bW3Y z0Gi#+llLQs{cJH>dVAAG$CBf0Q?vxx?jBoyLdFI%NjCEl0h5!M@?yy<2f;s~*TtAa zm)wiMMwNbL{~eGtR<0HEN+b3XOYV2`d+y6)tm9+UpL+++X(9NRAxjIIxx3XXpQQaE zLH6eo3cyhJ>lf)*(1TyFduYOXV9g$stUV>j`uvxyJu~=!A#0CPLBdPNeZ*@7;qE#eXRa3}hb4q!kdb8_vvdDmyR z^5|UvkETg*XI!gZWy}xWDAZB+spE(MNOe?R8KV5Y) z)=k=oR4(lm8%o9g{%~|0I9yOoOZd^C{iZSJ0-&*ES0)#H*IcmU*YOwYU9|FDsfJnj zi6>swz1ah*ODUzjCPG(^K=Io1P`q~j^Vzeo5|pl(?7)Prm5nl9PiUqQu@YcL5{tBpVB|KPk+#rt6^uw z6ui2XyoLXPWh3}nbE#Bk6hz3L(&Vft{KMI9A1N;Ke1I;@@U!MB_wP!TJvGl+Vf=H9 z2ZeUG4-~sc9sRwW{Y5EBfgwU6bZM^AY3@IB=#SaWs?~ab=QMLxD*wIQe6c&xO`C#u z$shZxQeYAEQ)4y07`z;#DFEm16grEM9fLRD9|u*LrkXOgjVm4oIqnxYG;z`h5y~4f z%+IQ)dt+%cFLD84DG#iqx<|4#Zj&rKj9~3MyE|uc_miAV9T_%ED1W) zTzI}Vmd(|Dh3>ijZYt!wBSrvhdg#$}h4~xWvjYy=H>$ahHqaC+hRJ!I#L7I<+y4cH}0r&U*8XJQ$w@aKJCkOFUqyE zkVM70f{k4<2=B$(l(SQX{f&8}BYeqU5tgz0RM4NGwC#}HK8U@_-1iazqfNF_hAg3X zu)#*^U|t>i?E@|qI;E7O2`?^6R?-T*sZf%0*FyjNKu`nNj)qT_b5jXV?(4`>sWnV> z6mz$z^bQl}s|`CL^AAT==9n{{H}2WCfd@Mv5WJJGlWpYIlO#Vh?A6LA)En{rT^gMh zkaquwm}2EGm4J;$clr8ns11QHoqBn)i?;Bq*@?Op;k#KbsQ445pjvam4Bd0Eo0Y$T zqahmJmq?zH!^)K?ldr|w0Cm$bs&XoP??#p1{N4_YnUg zk;_Sm1{RE+{bkO%jJ@>PW;gIl#+iU;EPXgZg{D$Ji%>Z?Qc2;j&tnsj$>B$Ld}vsG zywm4Q`Fc6_FE6py-r&jb^E2y`L++Te3ci!K;-MFok3JYZU=$HzFhcwSa$JTjRZ-s* zQDBkndJFjA`_j7md(90y+v=vtb&um)=Z861-j=)c-eJ|^i&Z=QhMj-Hj{+aI~63#wKts_Hx$X-Tk%us!-5(5zHH*n9Q)W~*fM6g?yqctnBlSP(!EITD485a zXqHHN7fX~3oi|6vsif?tj|#GZPGZygvXKJz8@k_rw^(1j#L(8AepYxDn?({zwTr#E#nZgQC z-f0%IrGQ_yCZi9vI!jLwRos1{*=)MAhRCLaS> zT{}z~)p6izR;1d+5_0j04>*MvRo{X5FTfi={r9=LP9Gv3Q4XoPY^cDMe+L$SK!QCD z2M7uq2tZyflwqIX21m3Ap-o{N%#y!OoMVVWY05+jt4xx;+f@;*oaq-U9AK;F)o`0Z z(s^*?V0E;=EJ-p=2HNi(hLWBDy&6tSI!5=uD3Dft52fk-9|1UnUIzS%zCVlmt*(5l z)f$RL*QO6>57ufjyw9>1x^3}=LqZxje|UxjT`KO1 zHw`MUbe)jHscp00W-sl)@vGFm&Hf|-Gqv|bkgNoj>UjlXf`R8R8lJE89RWz|05~vn z-_zLn3!~>N9#OG}as0vbOcC`b4k=f$Wytcb$rnJ{xv0DXezQxk07$2*tF}=68Dva< zL!k^)d}=p;Yd6}zkW&$Z3(?6fOjLi52KU_@WkRM_K*zFV*Uwcp`Ijqw?KwSG8j;EG zSEPw(a#fUoP~_N)TR6^@5;cBg*bhVwudugGf&HD5F;1^o1CC)CM)boLnnY@kt%&%b(98g^5GU@y5)Xo{zomoEn< zm<(e&ESpq+>m|V!ULKk>s^1~Nj&HfDaTGdfi3z#5HanY*0!li?G3$3Cjq0-?{7siTg;7&vZEWWBOMtiRlS>Rtz*9(?YuIt|#)GiL=C) zaGiF;@9p|>iF2*SMqNxmiPcCyKr(20wmD# z`{7a`F~HM1dWy48sFwMSkscis>-fXPj_xOOvu4WnapqQ}jjq#~t%}o`>kq0bvw@4< z*-_Rc!IMdPWTb;y2-}=42|}` z)TMw!t9%iUCcFuL!mDR^4pD@RHK~IQQBuK(Gn@iQR|>x|u*^9sK-eT@pr1zmzg^Cv z=OL~K8lHp5PwBQSGHs(B-c1nP2$G6Pq0hd}{^O40)N<5?14-jaNAr}42KN$ge$XBIn^%Pohu)HIprz^TY*VDq=MJ`;D7LIvmjEYS zc~kaK-k{9lky<04uZwa}=4D>X#221qO8yO_=z;5drHNaCn5Nu4XqZF)FN;VCR#{1~ zN-x!2@5JU#6;(6{_^Gp-X zPU|5Z?o=Y{i96Q!q9gw9*j7Nvyo-qwgkr_w_VGN2>c6dSka(@YfcZB4@V~cbKn9Ns zxU4*zM-&#ITZ}n%v*>WWqIRhzpQX?6iVCUb7Ey`YVdv@#)G207X$^NMZE5U;F{HsKAc%4&i(9EU{F6)Ms%sPk_Lj$n6VO(^sj(m-lxVw6G2gnhcocd>v9=CisU_ z?Hq7w$bN=Ra<^vO?dn1@WBxkKIrpI>nlB=94_-U~*ajJhL&jFbWedypRG#UIao*zk z*)lnL9^1g~ajHy7K&4tzxb6gjivD~_7OD?hySu0B*-Vr5cSsrb&E+0a+ekcb zAShqukpSuB>RnB>5W%Xk>W7yvviMi^(>w`bKM^kDUz}UyYY&7(w~d%WG`QkAk7bobP+;XCbWq;=($<0>%+@VX-&D$l zzM}dmzwP`gc!7MLU1*2){kEBLj;0UKxXUSkJ8q2Dim2vXXrBCVWgtNNj`H!ip{G-4 z58v%nFzO`Ox5H?JzCFI&$UP_D%#iX=mnqM%Jx`J!>P}t6Fx=B*Vb2r#ik1mq^5*XX zez{*2HK3FCGbAaNcGk7}fOJV(`3(02P`4!E_&2qK{-5wNi)Sx^i(`o8sRG@rzlsb~ z5+5TU_Z%XE3*8z_V2&+eT%Q(8n7ZE9_4`JJU&=(J{+X_ZpB}S$#9o6ku~#-UF`Mao zDsg9|$Y(04?dRUMA4ke1=mpqhSOOTfZO~!HYR*#x*CLP>rd*$8Pv4Dt8y2Oy0GPB_ z-EUZ5dinaM#x<;6G}iKt_%uo<#})8{m!Nd6)3*?d%w%@DXZ%YdMph%J7hF+kpXZ+P z0%aG(rIJRPLl^Y01u#jP=H$}*H87@`8}M*R_o}qp6+(?@*GtPVKVpI#6NZv_xJ9n^ zYTj2;(|`<{=X>mO3BE(i$f8nHyDZy&ZU$9lOPpmf;7^6$k&Tfk)=(sZ*=7IIp}Uys zmL>I+J-e1M&Lj)ZOlBa@YOOr=oml<)j7T#IAC`Al9Z_30$+st9hicLy*yLW|Ie}WT zwa52C9Rby%|ECbt>kqB-ZGf=d%?thqWqRokzk-`MHv~2nn@IFj( z=Duy9$KQcDCMYKT|LM?7TzTsPa}6QiKD){(O*lUyXPCg3$#V*8(c!$I?hYw<>?K)dq9Ckz;X5yn?jKqH`5FG! zzFeV3i;+LZ?($FLmAPw?U5^~?)Xx1w?#zZ-%b@% zDTrJiIKSGiMxK^{esymcdVir|z7f6kg3siu3tJ?bBc$U##eRntIanLOfnm zqX*&X+qnIrOC{P5kDGnA-Q=p>gs^>yZr}l*+8bWd@;`j}S&>~A#V*H@L_hQwU}ZNc zYw#QOAI>DXTC-TTxwwA;<@Z0r3iWB_;Y?M%0V2@PtJdl>-*fGU9kP9^#g~D*-rh_I zHVnKA_Pu?B)`F|w&U!g{%+D;214>MJFIvO?^3H$;QDW{MU%Js>Jn{8-*xP}7z#$OR z>y;29*sJ9tVmRZDSu0CP2B_TJ{ybeD+z^D&_LVH^D!iuX4XfB}m@|Q(K}k&c#=3sP z?|^W|tSMzyd6_v4T5{Uf0MExDgoEVA-$*8MVJ=E^y;Sg6gc22B*oeO$gd;XmS(P!o z>oof4Ul*NYbzt?PN|gX|!i#v33_X*|znpD=QI+&SvqdI&VNKke!z1|Q6tZ{4L2=@x zn&z*UEElMM@a~V~PjXQw10~C``<&o=KmzMab+9}Zu8e|ufXs1V3fk-o;f|Ywvn~j% zwsQxf^QAkGqt4Mzg#&zZ+AA+EFuFHbS9J4E+|U&QViSL2^@M%r)waLSp_<>iahMJ2 z_N-WJDO0>-S&xubVwHdh3Hvku>5hJ=%ug>qd`I`{@v$Gx=eMJ`8e?`EgT^zMSR@l% zcOA?;Z5VlZDu9o8*KXx;Q1~_1q3BPhRd{&1c@jnf1cbzR464m$m7)1H;Usz-x1#F@ zh!TF~B63=>;ImwIsIqz2T&ZUgi)RcTLt5O$aC>(6PG+U<_NgAiU&kfA}IkW^GK5O=sl&=FEe$`ov2o>Y03(duZaz1BTTW-a* zxF$vBHSu$t+%0_;<#It>#0}Sh8{LWho?~wx>-^{}zfSqcHvHt$mhByXl1EmishHqp zSN+HlE#2j^t7>p?{A*D{%gbHVxbF$0Ba+?Q-dqy5OtM&`Nz|LfVi!Z0CIIo~Nh)UN z@!@SHS$J$Bd)J5Qb9-qWkWs<)r{+$z`K_>LpLxRlQlY7a2L?LX^W{)HAll{Ax46;f z&+JYw*Nkd(=d8#dADZ(gV&@EgT<)qm_wjBB&kU*`^7(E*B5EUK{RO^07IcA(FhwU~d% zlL)Ig=uHHjd`mRUC*`P3hexln<*!Qy{#Qe&{i(3yJ?o-O6Pm9^Y&QvdPeo`K_L&sT zEIu?013$-osGph3uloqfQJ+4eJh{EUeUQwwILsm3s=0=^_;(7>R>n*ba0*B=^W%Sx z&Q9da*;hvuE#|DPj%UpJ8U+WzGeMDBW^?l4)auVxfS0?&?GQ?wnLgz-*{jU9ry9gt zw(D>3qoToo7^B!fX8idKA_o9u`~l+(FcrgjRvF78knwv(s#MJBEzg3nb%Ov;F(+3&DQbOSA_CN_B>Iob%;0QAw&5rPP zb75aX*2YwKchnx|?z0r7e_8}a^cW(X_@nO(-X)(aG_1goelcqP&}7s0iy{SQ4@>Cl z3V&ZRlKiNuLMeMENRBPP;lqo`$~h_|fHI%p4=0^mOGObq5NzaIm2diKb2yYH$Rk#> zh*@4E*UC3*dI$m#;eB^9Z)h*xxzvnmHY@7c*JJyaueorIYI1jF*E@kV8k4SOP3H+9 z!LPs5X+H|z)V?Vqu_FcLj$S|l_rL|YR>}Yth(vPHJ(1w>zud}=@RxIVo@W+*1OCFk zP>*?rdKLI0JKI6+7B{<aEci0t#(s1H}+FBty}bT&Gtczm8y8NjFoG%Wij^(6L1vabWtqv3kpm--@L1_iwJ9o3)2qNXzJI^7S#T`Zhd!ST`6hglobTr>Q& zmzNp&4pH~{k0O7nvMZ#{ac*J+bo0d~FRWU&MBdt-`*Ha!3JT=*X|beKb1qmx#xAJG z-S^ss8Q$18{y0r7U7(NL1wRxMC~HNYKjn7q*y=_Q(gF||7qhEzgHGlBbuR4B`y7Pt z-k_C^5U{&Mhhx9)r%7h3fI`qkB1@=59?Q@uGU3RxsCe5%OVaI;?*a|OW3Ne}%{n#j zGNB1YNma_b9Wf&}+ij1d@zpM~BES$w6al;Ej00n@R57naUAo(%=l+}bi`cE2di4qD z7$jk2Kx;r25&oI|HB124qbaFh1@D@i+kMxl>yw7=Bmvkq5NgQ=i7tuy<0xw0S(jX1 zQ34%XRZGNv6SKxfeqBnU_WLk9V&1G-(pN0cEHvK!C*cL>oN9lQD^L&F3QTcrnr@7h z<_Hdgq=*tI|76%`eV<*P`#E9K9LrbOJ83$1E{H=oQ4zuP@kZE&ik}NcVVqWJ%O2*{ z?tBg1sabT51e&$;4Sxji$90h-i`}^cX{!^-5&+Z7ZPetQX#6ez<}m8(9Lj|24z|pN zOa6@&W`Js&0X>i++o758#h7$G@q2Yi`62IdfZA+-y22mZq*gK7Hg*XxHsb{MviukA zqT>T*)kAT-mX%gjAi$(5?BSHGX3h6^;CLlSF=EQ0Se;jd?qk*fd=);v^|#&DG?Pa5 zo>88!(ldIz8ktqNIw#b!wqSq5_7J43ju0FF|p-K?-}*EzN-zMLzycAtOZ4y z-c)uXt(ZcCvZBu6UBR+&@P307;xpz$F-pN#Au-RMly#%dBUym_uA@fzseh}+kC|HQ zXn^R!jgMuWw!`Gt!TIV$!XpSu(@nv+Sz#E%R=_8bX8*YqW?UYLPv_?iuxjn_A2BXl zRTrlGM1gmLZy&Zz15aI^%my0iFK355jRSNR7mdB%ji_}YVE2F3_Y6BX*-p`kYMTw9EEQusk@L<1R{u21|CcO2$s&t;3+j$jZt?Hd( zDy)Ehtg)R5Qp!7UdGHFx$8i1|BOI$2R(Fk_xGUKYpeM?%>S<}%eqhjPKpIw{eKJ zhQHi@M{cwx+!}9u%cAK_no|1+<|FTnnWV|B$J=u1Y7Rs#H0;724R;r4IJ66DA<9MH zb>W)sLSL_rxFg!|!{~^G$vQ32ue0xp8kW9n(&}3k{Iu=Ap)#o)5Om>LA3mg~yK_N9 z$u_c0ix|r~rAGmTTGe4^f$%pj7IV)$Nf%7gAvf+A{mD0%O-1eaqjY2bWc_|`wglmv z$e;}dCy1e);lqdi_l#!ex>z?~FruWn&`}x%3TrD#u_v48_w4jE6hbm{#_W`Dc={)D zoSl$=%IcsNtqWvM+0sm&6p0w5{o-LS)ERtwzYjG1cuF@vIQmY(1Y1q?Uf3*Oqtk3{cKZF($_}mgtxE8ZtfN1 z2K!sKWicn5OwYyKUDX~0VU3+S$;&HwqM@IgD;&wEZdH`03hh^}Z<2jaz-L&FPm_)N z$b0HAs7e_cvt1F}HlJ>s=-nq|Y;vWQM<`}9`*R+@^f9c z3?gG=(^z9OM`*lYPtc`_6l)) zQ^_W^0fn9DN?8$T>d*n26QW?nYnEo;PMw7cqL{@HETSb#g5Kka_UD^puun#Rc}Pwl zodz-GX$Us(C3qgKE!cbXXJr#N!zOev>l~K43^ZJ)@7{02eZito=6UatM3SsxIxP-aKUWJ7&32d|33ZGrR*=krY+gGz&iGRA- z$NmVK^rh}}6KMm9K4#2bT$vFD!0etUEAB8H+Myo263k8^xXRvlQR4>8Shf^SKuG*;>zi+=U0Ge_X8LAi~2KGC78%dajh@xzIM% zt@&-2MoGPjYy|T#x((qIUQ-}(Md%^Y=;k(9&Uyu>_}Q)*4z673d@d`Rtl&cWQJL)VSz^>P@N{23 z0q)vxE_}1u8ANAZqwas=>um4d#tuay525U$I5E{;w~c-NU<|HkO_~w#aw5DI)5Quk zpehy0OPFndo6M-X?bDsP+5(l%lJs9~C-}UHF*~AxSHYeF_FIl@L+HNG{V8eDj8B2g z;OUB=bR_Et+O@9+(-)p?3;lY*@YEqNq7%|Cn$z(|D!hxJyN|0&x3)YCTdbaa`uGw# ze0+uEU9lUrn#q(lM{O`d_!O6rXYu%U^OW)->&0ZN!UOY)`C}o(o9LN7<4!_zg_%A2 z5CVcwZi03WP5J#>w|zNs8vfi-H~djYtk5!&i}{?Y&;_2^No`ABIyM;Dk1Zur7~7YR zUSu9R42OhW=dG(6%9Gd)4p_}%acpsjkzTvlJ@fL0tHdW^b79O^ep=6U%Lwf%l;*4~v%JtQGP z@M21OtpH`?8i_g%_mzDXsLqX;tq5jKX9!6OUn*(mYV_IH$4vq&vEHNK^aoD%G zKZSB&mSj0h!UTZSGcfl>`-ZN(1AbooLN9)HgsVXB>`RB7r$?RhS{#D!`c2mre5l&I zkLy4%x!Vu)I(c3JdD|n?uLat1UIA95?|^N`td;F6Zg(3RymBtbbj{gqw_QTFOixG{<>%k+;X}hjt4*%-o4RJGb_bS8C*h2z* z;VYf<>3qMLX@%|uwMj;kP`!Trfv68v;LQ}5=JmHQw4>A58P216U#t}8C0T-g6HQuL zdzR#-vvh?Mx39ADy)$gLb{G@?O*BUZHW0Bn7nE#x1@7GW#6jU+fZ~U2J>o8@ko3q| z`8?1k`Yy<@Xf6NuQ_(I#weMAQGB>!e3f!G~g!Udb;8d~=vohVZ3<=&TJmzR61*2aB z)An8=yyij%Ey`)-2qyQ2IAU501tyZRUIdnjhnJp8xH8%k*uPrz6o{X6-%vwyoHY^q znoX1S+ud+QY^JrB^ZYBs;2C+Z*UevV{c?YUsl098u4`LxTjv}-f4WW;f`6vtJ;RqzjJnZG(&l>-XUEW{VMv z;uB{yErSB7#n6kFxam<2e))6t-8Re0v$&XLY~iaZRT%;SxOnO5>B3K3{6}4hA!Aa} zMH3IW_zSv5SJ5j%3jqHf_?)G7rLjHhqj9DR`)i$iH=x~+>ZYL(#HEl zOZ|IJK=MTTL*sL|f7pqn4`h5`Wj0ge--3CU+jp+b4D8$NaUrJZ4puzM z^Bixm6M@Z(MX*nWqB&ZY<*Gp>xT@VKR8A19)c@QO3-3Qmy7@d89vMSKaZSOk z9crTx@VV8bJp-9F)eR}d|6|f*9XfI@LoJQVm$GBI&p`63jBPT5)9fFDW?#qy%4v4a z_TUhAS;+(HznbWssR7<~wkTJW#^XaH#`S8q zFEAg0UuGqqs886WAB%M$*w+GLg0!yBWP+<)5|_t??*@>L=IDO5nB!XHOAMqrLe*)j z0lMZAZnI+Ut#r-IG(233LFoA++4z;6x^o_SFI#qpvD%lI8&c0dy3gTIe z^pjbNPW_D)++F&*1Z7K;@oJei3Qjb7>cmaVyEyy0YMD$KTLrDo^%~1wGr0L(0SHwG zC*hCZlQCvoq9lmDGO@AVdPPKf#`FCkyiRZI+CFGpI0ALXN(5J;aC^4}68se(b+@Y#}V zJ#0Y`O?$;#M+zliq&EoUKR7~zyr@yhMo0;xUUa;!eEqy1m&fLkyrPercKW|mvCLPt zGM{^^o2x?SPVKBulL;~RzdlH&hh2ObN`7^4eGnm#J2p-?O+O^q@>MrBP-9J58%S66yYY*#`6RF{N+opG-%1Mi=L}Ysf z_b&f-E#S-dEG1gJV=b;PdjrxMNEaKM=_CkZ3b0<)}NYVk!Pl_izM!Cp6`o>s+HMv zO99JnNi-<_h1P7umhMzT7|{r(y}l`fyqE`B!5gFm@`J&K50kkDg3^(hnm`zaxU85@ zMB$1i!)9kS>7H0lKWO+rUX8u)(L741N>}F+EQ|u=8vYz;RZjz9C}b)t*%CX zSW`%=pN_Nt+cmq7+zUZRA$eG}sqx$Z#P-ju?_!Zb19cq5RXzY=FX$+nGuS0{(Sose1J`coTHp{fFV_;neEMM3`!} zk{#`~JU3X^`PR#OdVcus|Ehe=`C|z!e?GLs>#_Ql^xhex-ZXaM^W`55erkwWimOrF z;#3bRO8Y<#8>&6oZw^XvTjk9|L3_CfGbo%9Jnd%T_)Gv~agxJcbB@uYoApDNY;)V- z{IJHm6;7GzR!#11DQWW+xzd+(XNN@6MF*K)HydxPj{%xaeChk0?Dv}$2Io3=q8Rvb zAfNL(w}h^@^lp93S1E~gzYTE7G<#=a*vjFS=_f`o!1$-U>SfY=r|C?X2XYv`2>b;Y zW7{bIWoKu;(3SRQAFS^Qj`K+ejhsb8o%t6Xc-0M@%%E9sg|7WwaruK&IlQ@C_H@jd8T0*yJF7kHv6vh3 z@_F?!4?{IAr6zz=_ujNBXSaf(Niur=va=xx(U3&?*`HSAlYyz*HkrGy+CE5>KRE9X z4_%X53csg6KyKUHdI?;M+j5vepA!T!1~f^*`DOf}u@5{7w&3z9E@x^DJ__;hF8o}z zr1^HLyCTdqhh=L2Fsd0k&)5 zyX6h_TqUiq455WY&EwHKAp${oD{3Jx&s1N~yWNBbyLYr#iIOvkak0@JTns{c_1HVo zSIcI3J_6I7Fk2J`blW;{*n2^|tTe2ZI1g_RT=)g`zPtz;eKHL_i1mdMIkS@Z+gl5pj$gtE?|V}}hGMLY^(n3Uj;K?JE$>{u(+iXK z?pr=+tPYl{GqWw*5E?`_q(8xBNoNEw6UkH7k_PoZAYayf4ocl!_I#+wN_9B)^dj)m ziSHhhqrbC@0v@@bvduzM^0>Xf@-6Ub0{+25?I&fp8iauv?#n|tSCQ@-A< zyhA&fo$@3ryGP%>r_|&rq1>BA`UK+ZJ=8t9D|9>VPtl6KBLX#f-ZdwZZ_<3L3pw>$9$abd?-kg|YsM3gW^JA}8do6fR&fO%0~0-lbf5_IwL5U6Y7g#PhV5qxwwX zWP>@=yt4GTf=d7=)5CO)!G))X%gQYMHT}Aq{d-|t;b@-=KEKE8EXmKhRAau4R$%ew z2(vGCyViO3GJhJCHYG7>gNw5K*O80ut5KCMBaErzjZ#9QdfYFp^w6I`ebN`%8EcBf zD{+oWBUFg9ojYtlTCkodF+F}41~$)UC~>aGtjAJCs#b@g+wP;iK#Ri0RWrg z;8juLz4}D%aSQww;6{}k6tY6+5=E7Yt49L}JRn0m`FUn77qEoo5ZD)M7~#>#@lV&! zohabV)GPAEj2q29GZ$}GpM@3Au+;flkdntnBENvgMm-hV^*e3vR@j`g?qa_4Lv-KK z7U!~EbYI<-EG99|y*AjKk7PuDPs+&3%wGSquzunJI!!Pp>8SqCqO_hH3)jA_d7lkz z1*uN#FgV6^AlN;D^G2*U*Y4{FPHq7t6H@VJT6dkYYH7<@62WcHOk~&I(Aemk;6&*-o_S*Ti zWJD?aIie^49X%oKr&9ADpF^K2hVX1a>hvK!YVPeW76$L75MV;{YWEE z4rLZRKZOmW16uJlC46SWp^sz%Z+<*Qbc4xyp7v$dm5BVW8hh}y{jQ*ZxX!%ysIlTY zh`!mF2a>^bY4cxxY7zU=1zj*Z4&vuK2CPY^+^5oQcMJwiYN<{?*k(>P^JJwmx+=#p ztMRJ8aizw%?n1t9Vj=YI=C=96x`}CQ`U~>H@&IIYGVp<5Z`~CW(wUB7B3ZKzp&rf$ z1W5YZ-~pG}09<*K-{5@5+bej;uBuu&mALRWDzS3r>Ov8n6YiOb9zg*T3_aSRea1+- z=JcYzL@kf$G1!_JIbs3yOmoRUI=O${7mY59wnn1t2aRl%_O7}5btw0zjV4%MjO&-` zW`6XKnjZW!1zmVj0hOp(C@<01TKRVrfES^V#tlSZmWlM0Sik<0$f& zsaC{XvsrK>W$+&$n!nYobXSV??6qk6Za3iMc}Mq}Mw;ru`z_>C=RBqmnm}vo3t?_E z`az}Hbpb!8^}%jX_F5cB)U+8bdv&uFHrsj5KM_&UP+Exl&p`G4h^Lyl>B7+cpU|Q& z9v4rbu91TWu&B5nZ~qhg&Hg4KTow{fiYtCLmdrxdcf}fsuiL4>nq_)Gu{`9~JU?;V zLDtmpJwT1dpFow_U4~HXNWfuEOil1IRw>uk%$&PZa|}{ZI({-cC3bjV@%h-euw<}q z7DaORDK{&^;R|V(J^gl^>^|tgv~V*k73&G9gY42QtIHdcnH9}v(?=F{hszRjE3)1Z z$283o42IGFY#D&q7`3x&_sx03g8bV?k-Y|~B?jBf+-(4jH)85oH;7KLT5A`qF8NY% zjad-p$cwQu7#G>>0~crZX$)sd596cU=aA%|zGS)lWav8RLGA>e{wez$!k3ri{IFo8 z>tKJ8M-2FDz)x^pr}2~C$LmzH=}MZlEdu2AXMZY*C$!c1ILlXgzZr@6>&3I=A+8b# zp5wsP1T+v~ENgi2Tk3at7H`+ELRgDitY$>wci*;3!4z8uR*nYg#*V-H(Xb<~xR6)= zWWf)4@CVn2w`lYq9#m4J!(`r1Dup>v`p@o$dCNjMIF+prRhHk#5rR8|2THEVeWL&B z;5~@Z`4HC{4=u4As=cz1xM0AYve-M9XN-bfTlPuG-$8;;9p&CToeUq+IsxH#VXx!S z<2S$$VsKz6W*juz80qBT47AAOFTP7%+vGR-~n zFQwvHiht{91;APQs8)G=CiT6y*lBm2!~W1ZT4g|kl%(Z@i&Vm)8+&cj!ra2T4^a>8 z%r5wjBj8o*04#|1M(cqHzz~YpKXyalMc^L26HDIlDt@nsF`S44jF_zdDS>WDOFWG} zBz3tT=-vx+n(u+ouodQ)40F2HiGQ|^LM^!3=yGRhdH?a{Y&}Ywe+uZcdDdy^(gi(= zHjCn91^)9kAl{FW|J*ddVGH+RG+b*5-B(Cm6iTug>6D*vJ5@0BewnDo=-mf12i~z< zp7#V7FztD$4_phvA%|&Wtf>jYgNHUcQNd=OudY-DB&&-TXpSE6Siu(PUAs&h-@`D^ z;aXJ&{c#{5w2So#-CibOXn}&?UK7nL*`Da_9}w%L)cY;hZ>@S26(hHrDR-xX<-hIy z@yxvGuCZyi{MV&mg<5y_pVgX{QoO?NKEnl;jWuU&6Yc4(@Rc;ApF5{PvPwZ|k6EQg z+-d-}8$Pvw$-hU?spa;64$rwL6Q0No^ev7&ie7G!hSMI+i2siLEspRPYf1yw@7Ffg z-g4+#R`qHepr$Y80@Y?(zr#HE_%xWpe^sM?wcwe>P$w;(zTj*3Wp1ky@#{j%SELsXi@)CZm~>(- zxi|$;IXvCs_X+eqHY} zIAan5!yF`>f*=2SByQ098|zlnN>QVX_ggw-;w9wJe~VLDgq;uCCLywJ{;ePM;tY8 zPa9;pG?z^%Ma*t|TS)`I?X6j&hN!b z(A(O6V|{*O40nMEsI9Gi)`kNGcyAP26{qpjkxpcpxf;(ybo3R*cz~6ex3hlld;zxwN+G>?ZyLcqtZ?^w zVYr|e*`le8QrG7l0aO6b`MU0zg!WeCq>?i)7NY}0=rc_o!R%2r?o>2>dNM(-d1rW7 zJOyLWtZDs25{>U4y{d;;acC0qr{OoPLD7N)`~M$z?;X`t_qB_@iWH>;M4CvmfFei$ z1*xG3h%^;Z5a|LUN<>gPp-54xbfg5Ni6AJwM|$r9A|+G_H4s7zDQAV>d&V92oNtW# z-yP%o{t0BQwb$C2b3V^}=Gxgik0p+oX-ogz`7B&~`0*7D~w7xSPVa9CYf3ASxkXU=pm&Py?S!%Dt#KAn1({hK+a;MBYm{*72hL2P&+e-tbbCiJm*HA7G?I~Kx=Vy^S0KUN<9Ny@nyhQ-)6@QdDK)TeZ0Xk5)q8a?2^ zLW(iHZT;l|?{?mIwfg9aB+ckr~4(!sn-yW*36ipo@S3g!llqA95 z<`1`a62UKY2wPlU5Zt7s6-6GKzkUSX|LWF!p+Mev{ieusA72=zX~G8T+svak1_bo<`VO zZt7PhpHIY)n}*Kx-xBL{oTA!hQuhvq)h1=bV+i!fB z;_#HPvlbcOYarnWud4KH=|NUyX6L`&BEiE_66r4CL-bO zZyazKS3Y5;pR4z~Br`)CVEy7N$d4heuI$^~Juw!e)0dP4axWE{qaKUN=A)i{N^6<8 zG*_Hr#O?I}5ndiu3f943GAHB4&cXmH5`{0+vo8s1IOf3$Vvs4P|ui`YzBOJ|} zl8V-?PM5K48|7}Bc;1bsA4xD~ytK^)WBMzxm0;7k-5Puc;`Keg1-8!g>Sn(8Y?+~` zeV4~Z*vaJFlqEWz+)qowwn@*^70x}qEuOmFQDPB$$>?#~OLFW36TEBn`3YGc!t+*X zo19m{e`6KT-M?;!geqt!?a}!}dpsCf$s7Is@%5Py;_HJmFXJBxOi9KNoo;u|Yj(J> zFRJ#J7JYb~X`A#&dV%Ej?TZ$n;o7}#hTO=Q6ipeMRNso2U(dM=9@t4_pS$5wkcDOo zG8(LSRd$C_He^H(Z^;o}z2)31eE8k8IX=0L*Kh4<2;1x7ma*$>MqCp?%gd4%^Ge;a zd8H$yKmT4~-y<7V^sl#I1J9z2);nM$VvM%vU&%2K_ot){TRmH>WH++PON1HrxYFbQ zb`C9g)Z>?f*Xif(cdlED>w76#xP80nb7vl%$AY~6B@f&aQ~3Nie!}&n8q4n!DD|%v z`qiHOseOkiv8jjK+|SE2SN9KWH%*^&6;n(r3K8O@m_)m3Cp~kXgjdUW+|Fsdmyj$+mIyZ_tnP?^c;_3~@?7 z{q%>q*QL$L$s|)o_#VyZ)FwB3iR`3yO0Y+`_KS3{`!fAXx4wIo+%PaRx8f>opKN{C z>lu}6J!H)ND!s1v8jaMqq${HakA!(D)-}7VBPV&4)MDP=5V^+55&6@!d`mF)a?I!} zZ0Efbb~?&5SD1Ed-+Lb`kDst#K8aD}E;47rl@u!w_p9C0t>#f5hbpfpy@GUB$hy97 zWozk|Y9HHq?R`Qv{w2IlcehVx9b+vOj(4n)+A92k)g6(2{%gW=p!9>aL5jwYi)oi% zAKp;Rc|KHeB{!#g{0?q>JhZj+2JM?cD zuXZsV+DdOhN-11=bLLumK8J=Qr;uw$i_(dcxor=*|I!XJ=xP4FP@dJhYsyr!G5I>o z^OlSI_?&=cpqFLY?HQH8nGU{xP#h8Nk00yEs@Cas9^8F!+sQd_vt6r713W~w$raqH z=QzD31izA75Q!|{w96`I;U)Bmgui~*DCl&SF1D7zwybXam&;z7*sl+b37ePH;~Jkk zKNk}#IDZ>iBvrTfu&Z_ZQa(%W4+%37q;Y)8UyFGE2agLRW7E0;55&X!{1l!eT9=&O zSAL&VEq`N~^VjL6@z=pr+3B}@J9IsW>MnI|gL~PnIj{50Pu)STZk(QmS&DXt25}tQ zzkucFIyPIQ@cY*LXPV?(bzzl)35vL?N=zQg@tL1WQ&(Qe3E?|j=Z>EgZfkT7yo#M+ zP>T7ulfe+#B?TQ9V}JOA<0NL^xx6syYO zU6-%Rr~^(3>puH7`UYiLCTlc3Zz2rKh=N;gd7^)W7{K@H)}NMN zs@!4-Dq$mvDoK6hZM?Y=_fFxG(7StiMFWPX#gST7r`?m5k{K3R8x{Lt3r)r15uuLz zJ?c(!vsu0EyvK*{*9KqJOqrY0?TC&GVIMF)hkX6QaJ z&NOc5Ge^t6&ZE1(Ft{wQi=mm!KX%E?nt$V$%5ivND9Y)H9YIU->xQ?VdHdvnUnu{s z5#LP#)|cZve_w9jJjJ>$RDOFxMx>`dcGvdZp7h%OZ1g)x)H{#B=4}N<#Mg=eIYrd* z&D;Ldke|GpUMWsYKQLU4=}T{)5geF*D~oRoUZ3V(Vtyo7k*J3MfTZ{+*e?qQC-lR>K2BokPu$|~j1_SUn28aFnf1Te$e%hHVD1+1GfnW7 z2*aQAM*&pDe&Py;w&B>u34aCEj1JIO_1PVC|H`K;4FR2J-wY4D6=9e>84zcv>=vMs z2Tvln-}K(Of1(VL$RU`=QKsguW#771PHO%uHR5tTDN`ELt!DW{kh z^gW3J<8!cJ8PM}-G)-`G9!%&w4%T}(uG|lXO1Cj z1?G)|-<2`A{|wT)egSbI-PFa@dt*Ad*;Ap=O-v$WC&8I-bi%40yvXa@T1{HjU%2$j z7`x^<^L?6Sb*iYL*{YuAG0W<=-r=J`af!{I(q^_EHw)cTR$D$>u&l0A2l9={(q-c_$nK+qfh-r=l)$MHL<}7{Qo41W%cEL0z71BwfE1*qX}Kb$M{DTs@9+A>#;ss58-{tWasUsvZSvTGHbo1zdvaf2k9QL z=`Gb>a$tSEER%gXZy^^h8ea3zrVg1-cO2z;5lh+tT2svtTaeyLKq)t)5qtq#; zT3E4g1w11A1Pcji202GSXHq6c+-~9#D&3isau0dxpcEW9GPghs5<-T%q#5Hu4<7-h znpe6xJ3x^OA-iD^A05oWw4Q%hNZ0`i3kkkBkxAKEl|QWc&sfPcb&6Y+i5zvdyqB<} zQLJDoJ_3qw@ht|C$t2A}>fg71vpzt%cfdzI3-A)>X6odL5O4dzbkc!s(~mIMY7Q4?JGFW9(QSm~4W))+n-c{E|BWA*O$=@mU}apTQk zjzAvtw`n}f{#wCG?%vzE$w6(Ns+8$BWhXzuB)=*0Y%(}Tu#gNEJ}PcGV<aw ze;ax|mRH-X>U(#9YZIur<{M4PePaxVrGqoF1ug|VVhlfof>0P^k1- z{CvX}g!tG|Gy`D$hdrN`M;&30F@!v}>aR|*xTkm%kCGp;tY88yUs8fAj;!5v1*uM1 z#ea)^V?8KAbv@-oT3=I$I%Vm|Ga9)bz$>c`M^k~9csMw?cy(>{+R@k+R~8a_)((IY zc)EfF7Eq_CdmnYZjRW8elIppMhs}qLl&e#;76F2fg3ECFW%V-Y#D9S}k+VnE{sG`) z+6HV!1r2=lW_*w{z)IXttNPMa06RdbwFH)s9hPUnd@=?{gAC09kGaM#0-(Gs0$Je5 zuwdX53UF}IqoZk4D}V<)0G^Jp_ysVff(AJO`0qdKxR?`afM;Bk0LnlHptj~ZfT)OC z0z5koBFe}O;7nN;P_Lub(S!r#PrQubJGpxz&`7IR%Qm+H*?Jn~!>={V^=UWVT<_Cu z_4ESrHd)0tWvGsq5}N*Rs&$SYxT#f z-bCSO!r-FCKh2)NTPU`9ZfcsV^c+^0DJ}k*GClF$MN7nX#>sPRQ1}_Sr408JT!;Ff zL6x!pmE->WH+3Ax|F`2*K9*~SkVSB})Aw)UDVOwoYs^O`yXEnS;(v8Xr3P!=dE<%B z=K@F6#063m$mnDgkeL|7o3D*8ctFcNS>~5Vj0RLDAirf(kt0Dyl^`eO(}I9p=|55& zVY=I*pnn($jU$%3iUtQ;jE#?(9i_bY%jXEF=H9L&RrB)DnK~+D|EO#Ia-}+j(KZRR zwQ&_V0`=5&q;J4WZr#M|VXx50%KuB87&0HRm@(iz@%bYp0nEKg?P{uO_SCuqWaan{ z;7h*IS~Dsl=*s}3rsN#~=88nfj$KU`4(LyrF>1R0k6=_NLQn@G++FPX~GIe@H z)$7N^Lv}|0E>i{M%Px@bl8Z_d&yCjh9tka~uAx?PNT5h|x{fsBFeoH$f@1AIh298A zZh0N(`v7@I4KMq^-T|1YNDCBgcu;PD*6ZJzJrB2lin+!M8doBlJtsmO0JhgVs1x$m zBsfA=K@GrvHiGFYopgg~{W9O?_@fa$?0U=pS-8!d`yb-BN~`}HaxnCN&m4zr|F7$~ zI(aA0#91qv74dYWc$`Rz#`(P;A&X2lX;$Bm+=tT1%SP0JUs5>d#KaRmal)Z%$1UdnLknnp*H}JXP&GdUp#H4&cDN zb$!mNSlW%J{_$`y@Ua9VViVrd9EFEBTow_DrjKs`EG~*EgOX zqZSVeCam%3qiMmh)apTlTnjxa`-=Xbvahnk;>CYQYBgPuu>Rj=-^Zh}FUg|_-ngDkIMa3m{IHoPe2J#;Q6`F{0%zIMl% zQZwP=oZkv&pRSB#U0rMqiSGMH%8g$RhVD@!n1vF<7Pa^5_r3FAW;5F|I|5F9>13Fuwm;!B+_O&=YO;9 zug`PVi!;o($vL)@*e(*5cgkh>%~K-uYwGXOwXdYv3K9M`^rblAZ3J}F_&^b`%b)h?35hSEvu9iCy_HNCqpYwH`NIF77P0dFub@o%*U!4BD8!h zc+*I2H$^YlpR4eN$$`$vbs^?;s~M{kbJ~xy^eSoRocwWj{*;tHmI%laE|{r4o9_7O z8f5F?amebs&x)i69t|`cDoyRKT^jRV&xz{VTggpYU4~yHZqNtZuhOv`w)4K*|tXzZ#ayC=xb&DM;W_oqNM{s>>M>=7Vc&GCMs!i2I0w zEVV6Wbj`3a-%SCR4G?Q&alq9Aq~XcUX)Z@Oa# z82oNsV2D1UeeM|RiQisWz0bGqoeuk8WEp)kCR)YsxXOn(G4LC0F~1|{Pgi$+SigPE z;kDkAMb5oXEr_|_h}3kL{@m}6Pv*{(;Tye^DQT6p*!o(U+wDzz%zvl*HrCc&q^y_3 z9shRDIk|sT_KGgYTvuu>a%|IV{m|=*kHMCQejdDInC{Y=Z~{W;Y_|df#P%uO3YYsm zc9z#jwQ9#``s}^i(bL|KbESB;m+cO;NtWxaf`x_k4kvaKv!lf1nTK^9ezP%1Cp$pL zBYi!#y2Tb#$|UX|yZ9k-f35aQUf@=D=TjP-VPi(uF+-Rb2sn| z*YQksM=zP)9ZkmA%f!A($k=i%vr2V|=IVGOOF)XaCU4EK2wBjP{JeEOV!hQB*L92k zQNL5Et-$=+?c~{|%AbMEUrcLu~tt$5dmN0Ki;?RnMt(2>W*$uqkRL9$93c5G!od^ZPRC!&95A?+8hmuZgZcp|OPE_sIJ-gO`8! z+&;rEwXFX6m%cVvbE4Tbj~MQjB1Km9&%MqeAHN;86~qcU-0rp~X7$)>Z5uiuJ2?wCnW+SaTMHVij478u|K+|9I?c z#urw|ER!E;6I|2C7qsCIw;XHl531{7$KWa5VYRs~$#$3Brpqz-yVK>*@iPIjU+Pkp zS-6tjN*|QerF>iSu*Vhm2jslfCY)2cR!1scWC-3+AD@k#&r66% z<$JUnext{zO~%~%!OULEl_x;f5?-~M=ylc&((L0pHddHeVm>8ZMNC@ z`LiG2(#4g|z^dsPPbeeLM6x-=pv&+6BQ=u0`8%AG?OCH=y9Og|6?F5?T&oCfoxP(W zJSSbHFV8}HZW_5S>aQ+>OjZbkN zLP?`dL*bg+zXZq5bFAsaEU(_LLqZ5!PA!ba@}F+W?;{JI#1&-px1+R87(~2mgmi2g z_!SRGbH+-v1IjZiNdx%)bIS)Rk^u0;=jzIpKN zDYW%i#jYRj?Wb4sDXJHLNj_7{z}nGCvp(CZjc*jAFO##3Jhj`+R(2u!nl#gGWrA=@ z*eL#w_<%q3kxC*B=gkaWE1^Jsn!9H`d8%;n#Je?|La)-p*xV=MAg#IfCj;O2gj_oQ zK^&&Tqx)8Qvq}h(aG~(&_dmy1xeH}gGGDv!^Kut5XZ((6yXO_K?ZXl&GNQj&saG10 z9#u^_?wubMNXyd>9r|78@53eUTW?=-d!TH_gO>V!+siER>*hQNccsa4gp9(qfDbu1%c9Me4$LyynS(RjX9>RBAt4?W z#geY6kK0%y^3T@xMMglhC-E39nBN3df@I)0=EMOQNI=Vwg5-MJ%UXizUmTAaeW96ml%exX^ND`5LFP7Hd3^&i=S&5N|x zPfv+n{?Z%T$ahN2l|Zy)JL9`DPyjv0M{sD|?^N-ts<2K{)$KaFVA#BW!^U4eXg1-B zPb_$XTr<`9`3JM91osN*!Cl{%_2u9FtanFe zhfS>8(J$|L4)mV&F$jE*`Y!rTG1Erly?uka6)EBS$~VsX>bJDQgDDMK(8qA=e#`9X zmUNW6=R{`yx2K+lpX)qFJF=ua-nNalyVpT*v>bS+f_^GKB=*nNr#8R2 zi;S`TNq^>2JN@i_V!ULI_xiF)IJK6HXWWujbQjXDx&PZd-X^PEt5oEfkKTcc$lXAz zzURE;>EMmkTNqIZ{QJht!cz&?$P-b}yt-;pxM6U31HPU{4WB-hk|6NdQ=PI^duU+S zo`95FB%M8{JnrVbfKOiA{5dl|+%O@;*#052adwScs-B(?j_G+JzAS(<8%!RTb;lLv zfv*Zn>-6tH(t7O(A@ow?>>?%Y?b0@Zd$!Uvfj0A5f$e>f>UO%{6;EwV_5~&L25Yu$ zt0lj?iShX3oQo){A%yl)8Xjx`8sX^l^kxc=+(#Z*{7#aUogZQo zBiWsLQ^7mcQFij2 zuHqQh13P0W>$lAA@4Qb-f5cOS?dCQCn5J1zUP&s5Rvpz{5J)E>vbAVu_SWFKPGwB>UQdbk7Mw^yGx-Mby+a)Xa}W#*v(>v{IH=vyW|DmS2uP)-N6x^=_h$$&XV2*AD0OysVyZ z!uCGdLr9&2f`Y{Yle3)=bW)Fm)$8TT#~Ie=hM< zC3YUx3-2BK$Q%_psWW1GaJKmNlTDv@ll64Zp8mSv>bhZG?z`UghELVo?NTT_U%Uw_ z6zbgi^+xR|_Nk=If(>4A=fuX?!gKX6$!l9n87gKJ!_O4b~!?H z!^58t6+ydZ|IbNzk>X^WmsxJ>lL==bEk)(mZ{6$ELxV^r2JQT#_ta5gg>A-u<*M?= zW25tj6DwEpUC<3_USn$~(p-y7>K#cNMtld?U%Fm8dHgWV`|Uq!>gvb1g5Kai#C<0x zalFs*xNkr?X#8R8%T;f*n(p;e_)pDiJ@br9{;a-&nC1wphl|Z(kR}!8w;E;s?;FG| z;#~f!8oV++yKw1hjE)q)gU%;jQYh&ivAEL}}SsEqeMo#D(_Aa62XSfo@zuT~v>D)YahM zmOrEqLqfFLT+Xg2_>M|T8+fgwJ1{EfE%M;j+HXtB_l|P<5}UV-I))3U%|vuAI}g15 zk=uEEilrP~=KUv&?K3s6p--Q{-Ss?`Ad)y~ zy?9U6B8IQKphLC{XG8xzw<7pDTUp-N--%)3vv2gC+20E3yG%Guw-b6!fA4Lp2#|zM zqF&(fqdp3dbJM&9Zpj?@VlUEXkKZ~Rt6cWD$X~_omEY>-6Z5>f#`7~P@Ty^=ZFVhSau!yzKfG68m2CC5+sQX4VjZ~(`_lR8b zB(D6Pc++rIYnLL?$pG`sWwglrQy(+tL@YOc`&LQLx#Dm3%3`nD-S0S*4Q=l4B@A7R z0Og^837zwq!BOv1mxiOYH;yWPmWK^w^s#23|`iQFrryhX4&O2Yhs#^K~r)zDvq2%N6}!Pq1x zjS|o5d|&Q0Sec^z{5jyl>nd_YWNRa=1U5FJ4@EwtK_>pkz`J zAFw3V|BvkcOH!1!M^XCVLU#{%~xPzn;pXn+;jwzXyu!%HN7s2^(*R%LCaCQ zheKWWeWJPKNZ)XlwL*%^@EcjfcQ|qc@H7)Tqxm!BCr#O~-jc1lF>{*pA|p*lg%M`5LsZ8<^>dv#%@@*c*mm_CGQ_B};e z6B(Sex+e&wQL)cC;KpZ;rd%N>p7U|KIkGkEj(Sd&MybmAL#3fwM+Vwxb(QmkVp!j4s_?h zg=pNxgc^K4d>;Isa*}!qF@p}>zlGZbHZChhY^3lAlhqS8TfmRu!8Pi_mMQ#@wgDg6 z=4|vm$cFSg5wcBq1O3)9VHo#~r2&$BaK3M9IcQyC+OlZJsCY@%2t2(RpFsW*QlpXRnayc}T)7ZY6jaI*=>FOQ1Q997bJ@HhBM#n-nDG({G!! zkGdVSjSN9ZX3758yhZ=%g}_bH;%{<~io9i6`Rg655q}-sRKbJq1^$1KkGwYNgTe7P z!u8z*u1k_p^No&!X-LNlWQf=1A+@JGT)$SrPMTc2J;RJqrr&zfM%m%7>jq9n`43L8 zxk-}Y_!%~iNjZqS&^)BZCr?Om3-NBy%jlJX*cPIc9u>CH%_?C}Y?Sc~0bLX`O^f&L zO7j#>2#|@qY;bF$)uX`)X5jb$s*Iu3q24&R~`J8~UiH$F)2l@<_1A zo`}XaA~U=$SeKy<2Mm0E24c^GScIt(#IpOXaq5Q$su1PyjdCW#LxK>5jHI3NJPxd)GVRXCV{>_?x%cKRbp+Lm-cOnPMeHd}k--{Xq(mQ(O z{l=-1cli`@_9ZaMnXG=*p7p$b&rJ_kxP8(BC6ylG2%-_zkKqaShb^>oDeiyv*Plgo zH1*pj4eedh#6C^hNAw03^I}>kt76_<%`KFEo_b!%a-<&xlC=lbU7RoGuIJjiFO0VE z+SH>-TR0SEM1@?E0#{OC*G`}d6mim!P*)Q7_$9ESBe##Jb%{j%?ggW6 zOfQqo;m1#181c}FB=1+qTlJ^edsjO;UtQ4c^w#I6(yT+`{jd1vwz;@`@Yf_)t%dzLM zOfuZsRheF62O*s`%pEXpe+AwX^C!2-%|6q7B2tFGU#kA{>!$@qT6ufFoWwRGmSqr4bNOR?(Y zQ~{gefnNpb3)h+|SRcU-@X|x*|JBPs`C&tk$&frX*I3L8KAtI$kG!=ED%vxWzB0>; zxA~Xvq}@qC$XcEvLkdRV;kCFOuUF@$Ou2;*qC7S;0QO*gm~TvmByLMrr)EP^hm^Bt)NTv@0YL@mcK#!`s;VjcIaE8ivFB+F386p;?te>HQ15TQmHwSb$9ho0A5( zsX#pK<+0-l6hX9{CbAjY=_ZlYZ`epuzKHgKY^d)AfmP}(sRKoD$R>7@s+a{M5f{_m zLOJ+MWKV1ID*MGX=)=~`>iZ*vUroN_L1utkz!zdeM_}TYnQuZ(4}2}w8keY?oCx(~ zZ_cKmE%A22+7whL2tx<&e|Fn-l-!g3^|w&rK}$9dL}3#4ASrGqc^5y$Gc{jC7jW0p zY#kRy7bv+$`0L}q-GxgkI1&pL=D7OwK`u$~$A2(UF!w;K110OZ$xu^lnS>j2JTyev zGZ^4tO_-}}#fGXR_aly7<;lezexcHinL6Ln=&Gjth!K9-1zQhMBHH1`G)vKv|mF4rui2cp_+faEaexXW7FamL2Ut2rAoVq1wXH8L<{d{%f;<+ zNO+f0sarHlR#KNhI-BJJUgfi5Q|l1>E)a5)O5Q#tAPCtCjgLY-1&#pCj+*_z);_U= z3J)wVM0+FV8!MSyW3Z-T-m<_yl|1z!RxwyCH;93YXtXrpbg%=F0^#?DOeLN-Z^Tgq zl=kgiEjpo#6qv`RK=5W<1Huk(B&BKCMVLPWe2Vh_e}Q1fe*+(qhX3fgfu~>MZttj2_1Bk);*xu zXrjBpS_X=qay3}>I^-gS;=U>EiVC*sq98qIhk=VI{oJI-EwbNi;>_uJv+ z0jysJZ@v*DJ`uK{X^vAame7WE4T&^YUa03fpmDNV-fpW|)^~;5CJilZ>A|c4`z!GH z2wOj<_~C|IG3!i*2PQu;+dRStKMf!?ZosFi=SaXD7LY7ADP5uvgfq(vd+WB zAf?vu20yE>sJrxuXEA4duP~x}p(AM9aD7jVEwNttYW*J-5psdaz6Rjq8{rKnBu!y@ zUOq;iS%C6oOE$Z?g~=)0KD2Z3;i$2tY$~D<3XG@^R&%Jp!7)Me?tL#A56@ySc&}u0 zl{=SE-5*VP;eE0Uzav_;ln5Q^cjWP*v6k`G71*4Sr%BtlXm5hsB@H14?lySruk!d@ zwO&DI!Z!(V_mB;A`ZpHx6#kR$BvRBIJI~g6P$1}c0eGZdz!GeNqDy~dbiMh zk9|P40&@i>#Qv=r8ES7^6V^TnM|m4acHdWET8ESx_$9#6MOH~TEcZ~Re&4O0iT>Dd zpm@uXdE7wz|7u|Nd?iaY;Cjl~qg)Fr&7gZRX`)Fq*<_!NK&za7zOCKWG7!)Lc6k04 z9UjU(NZX(fPjd^R=XCmhlr-T~iz7E>n}0H7!x0E!E-@e$e8L;}w>aA2 zyiJu58k*=xcaeEYG`)n{o*A$-7d0DlNfP*5#(vu}c&Vo83#Cd)&#Wffe@-P1XY1*o*xiC_WC>fRW6n8?y+v@NEA$ka9Wpg z`Xkv*+EW3@^F1iRyX6H!l~B)LCookD5*^weAO>ZzAP^)%p#R}LFezyzY5};?rO>-` zQ$Uk&e*xltsv8yMxp}eX5kUT)Bt+Q;I3M{Gf41d1jDCHrU1|R!sOOUK+tdPtq74|I z!Zv_gPHBWLFO~9p2QF&`Q_c+~3Q=s;U{K1zCD&rRBs|SF5@7}ALA79Rh*o_7N83Ug zK5RsQ>a4*UumtLPC=e2qfE>7dbGASdib$U;z(eM}Z4pU4HYPTRau%(R5v_6KCB0LGsClL_g%29xQ(Te+SN8h zq#T*5u^fDPa9Eh)e)_34-*@>wWCN-dlr$~E-EYi8wo?qU4$lVXdn?*fsSilf2?WfF zVA>7sq`+=lygw+tztArouAB&jj}CQ}9)aJd4HGo?-Udq|x)h&0gfm3p)PE}~1p-~H zNQ(Dhi!;hNJRdCSW4j1+5-I@XWGcvC+M0;h6x(ob=M}br;2%3X)ZVtF(!B~-YHmTP z{*w4?xNld8nrZlGRY#__-rSQB28tAyY$5m$Kb~)u7Nq$YjY% z=`K4;zq4PL9l+uAKuNH)T(+j4pv=?2w_NZMHuF3|xj*;H>aD%A!B>3v(IwQX2_cd- zc0PAUp_m-^M$(h9*%;Bq2?g={f?Ui*#Ag8ByC~4|V+aVpRVzGgs8^DAefTsmMP=w9 zmymfOZkx+CnqG&=^#kH0BkCPUB{_5liQhM`Ci^Bx>F#)O42@-w$4)F>NNm^%R}Qqg zgQnAjYv>*dsJfY(iB5!Xq%p;kWj(&Pg9Xq%7_(3=se=lrCj?~7hnjhEYLC$;NbqgT zS|Eeq_Oo2lx&*WpcoZL+KNook9JzPiM|?#9DnS{8f^&~+0L&_SBs}7`AWVWHD@_u!O=?1) z|04qIKkxd=!x6}Mfm_Kx0-ihL2CdUj{7Zi*p46E%fI2U+}t zE>o^SG;a|ViBsXz5mh0ru!Tkkz`{to*@LskN6=PO7G|TyC`UI^Qq!Uty4Zg+i+zS@ z-Pa($+E_1IC6Eq{$GRi{rBxq{D~}k&axIzy`CWkkuaBJ&`*hhfcJr$x7JD;681r6e zJS9ZY{Wp~wqX!*X4Qz-O3ynGh^^sO{6l;jN0|2^36J{sWgR;*HGBeoKfJ+gWS{}eM zed~V(CKk*%2sXtbBuW9M%$=e^^r2#w^CG~(qJh2>#`__Xln!BGrhhyrblC>YQYx;e zu!Fe0!(V4H{Cc(M+d~L0#PBC>%jd|S72az@Imo(4v|iT#2t38R38wvBsHF3C!K1b0 zN0MJQV$YGHz~jA^{kWtFwI3U3z(10N!{`R~Wd$Q&#QjLJ$nAC|S-B^zUF}+m``k5; zE@&Q}?K6n;`L*qv2DX{7HsQ)fv~Hm)?#2A^O3QC2QW}~OS9S>ygN+mQf{Pr z>~LRO6Il>&L+upBLDQlI8bp!Svhf{f~Seg(KD)A+QLY0)K zr?58M`+@5?E`DM;ReI`ZqL5ZA7O2@qiX>2T6suWeJG9&YVH=GMp*X4&knI?>9l^)X zN>qTtwHz%%&njR`2sx-nsO_bIV5el!Q*0RwwkYgCckLhOQlON*VvO%ct07nqg#>oo zwkVG@5yzUd5lNIoSrA$jUPg)|KoQE3x`3NH0TD4s30vM91BnpLwG)UrLv|=c%#c}| z&yYjH$!fudzH(O%-)o_xY_pgo#{~u&$UT7;8D7lRpU}a(J4<)sA5nuFWr=qP%}0db zwwM}PX*NnhW#Q79^@ykG0)inWuTrS*1THaev;NsDunV_Bf9>t{U=gxy@P0ChrR*+2 zuKp~QHI5AvS>NptQgQOG#Ttq?2H@3P_8!#zpJE%=x#quP-AXBib!?#S(H$|<}Fk+cK+`HcD!Q|(8+s> znr5+j&$fUiTTLY(V2DQ<;KzLlYmkE!DN1JX5*lE6L$bsVu%)5HytPf-d$!0s2$qkd zfTE30vY#@-U#PKn35+TJ@h)$FOv;mw(EuM2Oq|8{CQ-%&`JfRfl?FckHMKp@7G^Pu zl2;WmikdO6X~NZd{*1jYMqfsUQ7uc{BLw1LP_U2}Pd&3_3QBD6Pc5#1T9(gYM?!Fw z)Tk@qsn?mekPdJ0QWf)Vft+{RT@YmN9s?$@BQD=Hm`XgBK4s1Myi5(!e~ zO#oQ=Q+zCt>|{A>@PkatGs8(fEI|U?El+>gZj;EOca5)N*_2VzYCmF#IP)^0#s#W>D)@n(TW4NoL-OdC=FE?2T&1*n={H-wy8rN_gh=VMir*r5ZFqu?5)w{6;OY zR(VLiEvU8yFPbchJeZ%Cha}V@@s!{dOCQCJIYYT1MhpX@f5$G6C}BN|?!s?d2VWsh zFp-wRO!g~F$p;{8_T9J%dmzj5i$Aq^MG(j&y%rxj=5n}rw2QfFv{|SyE`h!G_45>! z-4h2>SV5l;zxO{=_R?SrkYv736$})b@af>B2?@kpBmGglQ1Z<8lc^nn;nP8&14zmX zn1)b5=~fVOFPI8!%W_@#yW6d$R}KsJwDfHVYej7jV6c0Upa`y|CY3-i0(7A1J}Pv{5*&W(2irAYzUcO0!}C*Zvhh2_{)davd~S?9RairehI2e zYDY*C)~T!dKzuDTV!dIOvq&Ix_R!(J;Xp}hh5M$0Kkeble&EBxwq_92S|4G&ZwSHx zW{a<-8kt8^z$$5>wgqF)t$ASCS$^I&BgEvYTEic(oHz;}#1Bxu0cGIAT1qPCxT6pRUv0EyV-sZx3xXdWvgK+w(Z@pPe|b53U&g?~PL94w?5_s#Jmi;NB} z(0xi2juuOQB5$)mTl@KtA3QB)RCs;s(u+t&%s{YhUlzqX2RP{Zu z4@7paxYJ5Y-7Ae8NTK0Z6qqUi zBb6|XVN~AU=BcNXEP^kt^lLzTnWs!a7wRV{JSd-p$PYCUY&vMMU~O3XH^VK4;V34` z>1=c;@fgUH%M?9`dGLlIk1(n3SD-Al!cCMUTyOcMtWMc^4KerM80N_(Z&&^+CIbdj zcybAQ6Z~~UDtVj5PiO4`ww4N*wn}%jGZu%(nxKnVM`pt8!J%d@B^ji^`p!UjDKU0d zK$Rf007)8k6GHIs%!p&mfx<;&JvU{k#2_AS9+re>T%R}PsSmPxE`V;sNVB;~9r*SK zwhz{a0boi($Uw|4c%=?<3H~LaCwS@;K`UQ)1H+c}L*MD;##(?JikTP{925jv5N3*` zeQXZC37AV>KAA0`Nuo*b8`|^LL^eWmuz(srB=y1`65rDc+}hIzMLLwSudVJLL*>St z0ELp<4G_trDkX#mTu(WlONnydya$R2%4J4WlGPBUoe8&1JiIvEe;OQ%R3otcf?1lA zVeJ^Oi9k^92PeOcqtHSc-ffss1sJAL7Jz~bc<*>i9TAZ{=moM!Lld~N0EQVs8tncT zX>S?S#v8Vewxxv@Ev0yY0>vdraVxZi;_mLnwKxPR?(Xhh+zC+J0>RzgA$Ukm`g`a6 zXWlR8oSE~5$*^JBeV*NG?@R7ybChBF#v93OG7pGWQKZfe>W0=OAO#BGx5sbJ=m?&9 zxZk5AlJ~-r(AY#IPKq%RHU$*l;l2y-#4ynH`q>0>)P}Rq+|H~2jpi9q2R{1(Vgm8h zfrbb{8VAY`i0*%vLOxQ%7eS~N=tB&B;=q6(53UaIO|t@&D#Odauq~Vkq-CO&?ROu3 z#-w^5{BKWa5qi-nhlKltyA;{tx5UVE$-e9X|EovjIXy5oxoHtNLU-(~V8fI{MD)L^ zBAVEnd=oPsVNMQvv9gB*H^ef=ngJ3yfZhAZ<9PB3niN1@Ki;8}ZEuyIsxPw{t7P(# z?0FJi2DLY>q!b)RGrWTKVdldbM~J)`HfNZ9sz=q-a*>Q3`IU-#BzBZzURD}HwUXcs zn8r_K8I}_WErL$Join6P=#`%ft2Eva(xT~W^E0&0X4Hft)GTx{=tW}xSjauE`SkfEzAi5GhrT@faR{|6sy@5I-Br;O>V zc$l_2{uBBE#v90ba>uTP64CVMyj=5sv7aqbq`FF_hbK{=*oOsT_Z{BEXycRP*nvhm z+2M!u+Onz`oBm0XkrAtdY=w65lwQ;rpui!XOl0MxhErvV#DpfwCq5)5q*IhEZ*;T% zQ*t0pQvpbQvlmF1Kljn>3MH{kXg>TDDPcX_{&U$og6CK3gUjoRN|MA3Xb7aB40Q<6oMgBnyIfBZF?Z{vzJ1*o^5KPpO zYFGx$>d<%ZMZy8HW+O&!cAkDKQit$OlGjI)b`Y3-KX~sfQwG6_tY?p|))kY095AO? z9|5G=vx~+U?atiwtT zb)q3*fE_u+UY%%!SV)|q`X3m;_y+@?l>Y+*0I`S65@1(543sdbjVx3o zkNl@rAd=9k|JjZD-TN5-Zv3VHl$Kj5asA&Z$q04#b4w(`0J;9@ED$7dyF`A4zbLy9 zS=h{l--IIB8m={>%bO2rl65Yi%xC}j2Ltqw$C&?4>Ha^R5;_B8PbA@X*dmEw;@@PQ+9 z#rQF7&Vbqx*MFULcflgnuJFJv;=C$GP1<%=j81 zOvb*nrf_eFgLHgOkXw-8Y4Ptb6ve?G^}3)4q+l6!x`LGD-62R+0xrJ=^5lEg4}99v;xeL^gDJ^ z{iELBJ7DN!F(=Yc;ut|7?RvxozPNsL%4{|z0h3- zO0Fu;hu+gY@-_BB;vqk zh7HDQW>v0y`*8~g813Kf=r zs#X&9)2CrqrB~M|+uDDTo^rg`KXq}_g?5b}hdaj;S?1~={WFuX#6a(}_O;oNx1KQ4 z$S5BKB9+5TjZRc{OhaEZ!9~>{XiGbR5po;1b&ml>B>QN+PqJ}7nFzeoi}U3XC(|Zt zgSSP$bBO$|-V5&bcc?@~FK3dGE4OrqNnGoeyXXT~etF`Mx6}$V;jDSAR2|{0tDpFF zJVrrJ1ijo4s%Vp?f}ZeEdU#gt*?#wm#ErgueQlr+#)Ww%xY%y(I=E`decgWbZyY^N z2ga4inZ8Dd!dNen>CsZhRhQo%U%`;^?4d591s}v2ZdC1dP-a{8iu;V7k;A$DnyP1X z?~iADeebC8{`4mJ9{!cUzCP)MuWwKnUa0%#VO3E7(rvV#>KTf1*hKF_c!0V7C315t z1f#Iy;jCitvcEjXD;>b%QgM8RPD{4Ijr=zH#txp^>UZ_xH054s(R$`%Lj--p$zHt#2RQ;b2)uwM0{Sz8^F{-2M@vI+5UqDN!>gvBkxG~FdRWf?2;-l##DkE zsr0MD4S*3Z=JdyUW0UH*no1y4=@)m+y^NfyW>$T5$ED@tZ`xe8a_@^zGZ##iqqTDO zaaWyhuQ)=B5Vu9gmr3XQMeyOeA634%;7^a?!}(&vg%x<>b-eyHH8Jnwl7^1*rG;*P zwZL@6rQyjnV}FJ}v{m5ebSNEcUAo|chHLaSu_gd^Q=0yid(}C)oa|R+c$(j`67Rz3 zt&&F?)9atdoZN0l(9+uVocMhTyPH{PQm6N5GTk-13NrKDiDX=O=iq`AQ2A6~Faec|?%H5#A0Dq;cqa^ZA0Sz*&$U~mXisjI*% zxy34jUcRrwR` zd~`MrPPvAknd%)t2aYlu;5Hp+7Mc@r=1xDLARjc5&ln`907qc0CXZoaMny(rlM4t4 z^KuAscVCX4=%ln5{xlH;L@-wG2ev;+Q_lVPa$be~RD#-3s_?E-@aJJN4?+*$%PZ!kpOs{~W{j zOf$b9@Xk6F;p68O7EU|Rm}tleIGiI^AUqK2FlB%%g%m5yuDg~6zj#=X9hW|r_|6I% zY)JN5w!_4aveiOpu-1&oxwRxEIxspdcla}3n1d>d(YR-?oubov;4Vp{qEYFVfzYT7omk^7;lJLHGK z`>YqSe+Z~Qgy&i(vCKDqp;fh#71Qj8X5_KvAKxt?+oY>f75tb;M~(#Fvo(nEE9_58 zkoxAI#QpjHV8?760r-m~OEKh6|KgwB zYtQaaSD!ExHx?k&`AT7jQR6G+v(>sStcyk(%pa;IM#q}$X;j(x-lr`k>oQv1RBJSt z%{^oe)dA75CMq=TYzPK(83gO=EO+OBHX=^h8Gwh)TrL8AW#v#Q0gJ@U`b9bPy>L~A zDvwz3HyaIl0qd{UN?S!gD){7T)IE%yz+XbEA8VuVqbR4sOv`7rhVaQ3s+>OT^vU41A9i?3%NUR^UJ<(>uMruY6gD~ni`?|8xee$8O^ zCEmd7vH;Fj=j7G`C*_Ik4*)~>J%~%)8Vl^9%<&2CD0h*5E%ui&aK-2rpS6gKpRw&`y{v-SAzPEPU^RbiG z+$b|AjAztNkgYIEZjGd$5C5l6z{PqwAK9=&OtvquD#D*dMtWQvr;JAj81;{ zb>j+skcvcO&cJ`Cy0qtVv-53_a}7kkWyr0ysURVr@lQ-I7S@ObwGFuk}y8a<9JH zNXgxgx~oibE5BTrP3CxKE#!h8;EuBazpu`I^Sou5_KI|lF`dz^_E8OX-D9f>|qhl}E;1cMRPvrh0Gh#!5__!E75Q<32?Bg4*RQ zwXOC+B3#tSdj61qQN#<x$`@Ta2d~t)2V@k0StQb}Kq6w8pf|uSX(EZ?tWGD^;{@M@<#5 zmM-I_!D`Yh?dAoOR!~SW5eI*_Y*A+{r%cO^atc^;AF0Vi+t{x33@Jm~PR!mzfOxq<)ThB{du>I``kkFCP z(Hjj2d{TN2Qvjx$EVi&JGhs&U{K4;E&(a2d_RzO;l&{2m7w5WtgeMBgn(W3CXrM_r z84&N!wVR5zm%T_4E&4lI#gX`ML6g0Tl?$YW6%9yWR_*+mCv3ho7|dYiWP5c!Y~ON8 z0q74NaZ(5>_3WSX;3_5R%GNRNYzhNg|RhwZ= z7z!khP%WmMfX~A>{Epo_N?yX{pg34$S0v$`oLSGdwiYs65+6#dR@$SK#3aTTV+Qt) zf@3dhL7y%yYt?jPse*bdc->v>yYOnQTLXwsisKr(?YeT})Rkz~YMxz?=ueO>y**pl z*t%?AmdTiuE}<>p)$Frf~AU_7!9<^Aseo#FHenx|crGl0qM3Q6V05 zsKH_;2Qb4H&dLfWD>&(wS&UHf*A7IhW24TEI=Rh}y2qLOd!}QR4+~!99Ly)UZ!{8g zQY`m9ngoBw#JPvi_wX3Hg>jm7RL4UTK zfY2TeB00)yVw9qZDwQP8hp7~5?$RuAEw9okidK`D;Nl`$_0>NWsa`ks*pfal zbJZ!gc?Zp? z&+I*uw4jSs*V9|5Qeh+ZznPpv8Tx!>6oakq`NjQ{-(mm*rlm7|5o%55?mQdO(}Fmd`KNpR zEBOdjc?oAE#)iCb2#iKqhV7(=WA#}s%gx{)tHy|UdXj zyq}2JAZ51>mzuU_saq&WY!N;6b>z0xcjB@fV%daJCyD23-+2zs{NN?2RG?-dsf_6Z zSZ^83&x?3%tjp<^$S3qvT*JA;dr9uo-Z71C%g&8G$!D<~5H2YVr1yBec}RoOwXn!} zV_vhESb4yOzr%}q;}Lkvz|w2A${sBC{ZJCQumH$953>J_X5mMXWdg%y^+&#~MItV# zI(A_pMQ+p!9dGuS0BeKV0Nuvt&ztC)ZP;d>f0ipT{<0P6v4K57L=i@&ol-RI=zPrE z4ZfrKpe~FjNMYv|Ff|`#7D9Kr48a`rY@zC=kerC-%*TR-H`tJoSofo*jkJsWY^hiv3JVk`|S33 z?+q4qHc@D%bZIe_l6V?$KF{XIio^t1qGekc1U2_x>u+XGQ$I6IYmIv947!g^dx-G^ z#mtcRJ9!3A*M9{_-t9V-c-gzz3)*M9NbcD|uFHNtwQcS`?D>T1$I$6y**|zAHXzAx zjZuQ`B8enK=vC5H3#~snPxY`2|62>Vfxp@xI-dvk`{?TM&5>e}uc2NZRM3=yQ{Y7Z zT+UqcGyao~UJtHaAHZ(-gCE2btj3<}{dX`0Xe@7SNA%addJU$m%_1s7$hz~e2F;EY z$?bkTzvbQ3%hP;toa4~wd^OO5S4@%VdKip+2N6dDruvU@x0Y;p`&SW3h)dz1ad>t5PbBHf;;l4qVB$}&XMS}u_o;jRGTYcNF)AVGj_qj z&mVi@m$>LLeqS-O%><^vuydEVi%1!ko(L9NW6v5&!pWzl0hiA+nYqtB0|VETromoc zQ+WCYlUbi*5CsWp2RX7z@Tuoy^b#(sq+HW>DuSGBK!7shmc@jcTq^ezW1}Yip0~C^6 zJPuO*FDi{5bT+$fcam!|8Fl()eB8->eOAk%{B?xCTEO8~uUkdk+xh~^1)H!+2Fg`S zFbk6yUHO96<_vrkh0vO`y;ob@aNNAGN!-pPyV_o{Cj@+@M~H6tsAw*QsbtfzmWg!j ziA&ELN^8X7^^mKHwt2r_u{E=8%3HFL0yI_b<1Drb_kzE#x^iT$`fWU>Mm*7PImq|V zp%X{vN1!l!Yf8@#uo%kMLNjZBdxL0>b@#Ci&&(`@$Dq;{TMvP?udh*t$A|&ic0KmD z^dMRC@b#}CiNv+!Zu+ru(gQ{)=!JXswF!dt5b@#gN$~JVFX7Sua%AR~?RaH0lYq!~gYA*#n zvFtSEou1=oG4ewxT__kwNsTp{+RX zF!AY`bOm)gg;_YQceUd!-ZX^_;e$^p;mvodV`=u|k?-?93HP|u_eY8n(H??&OwLgH z!w|?Tei9{7O^(7VmK3>t5iLdic54vU#TliK%)RMjC3Dn(=7+8=2k0X<^wDK!4&>_L zBNFVh8t4Om_5>5&-P%iB%~H<34kkaTeu)t2{{t+5HQwN`XVD3Q5Fa<)aUQ&R`mds; zn&_(3T5U2_{cN6mMgwk3bL2V*w;p_h2a_K0c9P3+I65iB5ft69$B)C;>^n;v0||E2 z-Y>7Bdi=BHbB?pcU=}+|77DCz%&99$5h2q`)DAzd?28FXX%cy2$Q@r&KbQ&!AU|9s zk)_`pVbT74>NN-_9QMdjbRbTPP*=!e?!V@u1!q%xr$Zip++l6if>ipLe-!&rJ{?I< z+!NDfNJW++B076Ps;l}way|5fPWr6d?){tcgB#!mLR`Dd1)*P+tnzQ=#o%IK z3)9fEhN`>ciz5E-&2(e1DTdrN=~qPqdDYJRdi>X1R=cZ)aq>md!$(gY0$Ev`JMzo0 zl;6%mxB%S)jmKtdUPkXXJpu>vhU+5h%-J zTPZ9~amMZAk!%xtWEnLbIHPC=*5KRZ7O0aP9Av=#s?MmS=R)qpJpYiW{Ml9pRa656 zyu5yzjDFpFc&PtWXa%rfhA79EeE5b(IwjV2-pdy`!7AwOsp*kcFn74|RXe`;5Q?p? zw*B(uj5VTj6X-%c=c21 z6rD|X;gK_kM?OpLojRP@R`)k#@m+4XT4dGI?U!sm_tuhz#1+kWx_~-Ljv;PszQ;nk z_~B=QOP{k2P=^ARemRf~@C~mRj;fwMu*tV9)!e-H5T_ZY8_^h=^|`ZxOQl{77AD`f zW@Q>PlUIH1vpmZ@cqyEjhDfNVGfpOpkI-#pHd&hvyHF^7^vk=9TGGb|i@)L+$_9^6 z^s`8>F#0?y4%cqRS>)C>5u)G5^_J=jVGDmbn>jEV|Fm2Ufc&zrEC8tCX>p@Bg4*>X z`$(Fl6UsC0V05`YI%;IhvN$Pp#%=|(q2@>o%$eIQ|?4N?v|x6_n0Uzyh~->KA)-n zse$70JV+t4CV{#Lcc(%^GLyNYRVU_*4(FUKN45Qk;FcqgEN!ZKtEp|Q>?q@or4gjk z18)llyszD_{{@e8YEwRa9DS30FZC<-oe`y`G8XIs#DR5peTEs&au~P0_FkUsEd41`P@W9Sh~BBajq*lS9~FN>}Z7 zEzMqcFUJ$liJ-ol_fqf zsi~uvgJd&4G*YgmJ1pE}!;ZRIH3#YYF0PwU+yz#GY2m&J`O+!#3Oh&s8jTKOD9@DA zc!r1X1=uoJxLw}~@CIS!jKJM0%*oigvnadE>I0EIgf7V~!SGGbdPADH<=VL)J>DJv zwR~D~(KR+{Q<=VF_ajqO@q8~={Y%Oz@3>RzX-NlR40SRh_l0~viD;t;ks_SlYlwm} zEUykeRdWs#PRvNiK>o}E2_0lFALKGSl_9PN&MmNFn`k+zHmerX zCUI9gQ8a|-XT-q-c|``ZzzaT14&JH6Z#NyoRd)DJdT;NJmB%@Fq#r<1bKTaGAgz|A z-%pH+j|5gNi$Pljh_sgn#r#D06z3hY-8`ZzpNlB&C?@WHpG+JrU-Zsg3C6-*;(e?i z*oM3nXv&@D%87nO%>c=7P<7EnYPUrvZ6cvZjCMR_8aq_TrKG=4;FNEUB#x~rCE0v@ zdlvh;C5kpCJTW}Z2ETRzQ!%A@gLcQ7n%^YNZuUu{GIi|zTIOck2-gjRv-SFumcdQ( zX_PXpYHvm+-K`iylSO6Av~kIcUHS|w4eN-@_}|94R9B)M3`L+>Cd#KbP^zglV}{_4 zUFf3(3#}FF!eaxyp?f`9w}){CjM|%bTf6X)5X>8C8Z5`VuWBxi2EP6)AfN`oH#P@v zv#vC{+kZfW^aK*3HLUDJ*Ooe89MFYgUHs~gsLojle6_z99VoIC)uPcC*>`2AerxhV-FxpCZa5l7vpX^1NkRiqyGU|-*! z0TkE_FC}=&(^QM-?IZE~aOzCpF)0~S_4^>yeKM3KGxrp^2l&48z|sabP|~Gny-7J58NuxfSA&$XlzqqKq zuK5|r_?F>>1j$YW`v-;@(HS~)NPg8G9&!l~Q zl;&nEbC-4tOb-*r?IPEC09CAL)Z3lX#S-KxolBYH#DVigU9N+ft2fODLS(V4hLPmL z=R$cSipCHDfHoE*k%ufFJMXf{XSeKwns#1(-u&aI;9ILoC@}g|pIls6f3$hD3MJc) zl0Syy+azUVK6WN=e~^`p&3Z~!<1a<-zUEV4N1LDmc_Gi+HC|&T@->8FgTK*XBi;k&g|vyzo6I! zUUp<`%TDu=7n3WaRfomStyexf#l-T`*Pb_7rb5egx6E|XVB}8$2qC_3;C1pUd6duB z`RcW8VBV_UscHML%DGGZ>dnrY+1ORdSK+5&7`@<1aJE{vwlNzE$_dr-npAY}SI^h( z(|NEF`H@((`g{YQX)2k&ZUzB{r2T6VhP!oae@lYD6H)xmpq&dyE;KP!8Dgi@Er8PH88r;zi%a*|v}0WtSh>ss zm2`eG2|DzaaIkrpx)ZQng4f5n2GUt=Wvg9cZ|pNjJYjEaF2agcO8__r>dSn`$0%K( z=GVOSEgCv61xvkO^a56E0A+U1e5%4j=~|58g<7;f4%s#8^6AIQh(y%6p^Ep`tW>7I<;m_phDEDt-b2Y_@&SrWUC)VuFXsh78^ z?9U@{ruNWf5q%#oxarLmEc`4{xn8&9?J3wjL!at0V0_`t!@+jJCdwJW>o96C5ym$1 zq7rlQt(zCe%Fc`77e7=KvlTa_!0>k$RKi??1FG0>-;I!b_JXT@vb%;kTgb%jnPHih5(^yO8CJ`cXRfRq;Y$e)93L- zK~{lK0C`L51e$D5kgWzqt6_Qd5HaS=8{qyYYgj9VAUhGLhFqAPsPDUw?>lZs4A0jz zfv3|qullp})iy~zeKgaa83P7*f@(B{*)JZwjC&D*Z^A^tb386j{G$CIh!=h_>YxpN zGz+rF=lNZV86zqk|D44uEm^S=1W7x~ugW;Fcd^w^A z+-7(vwaU#l67fP%Sqlq%!{^pF)B*n1jHBg_icy4@qS^{sPuFaW`3&s~#iG3y8felR zNjdLrHDxyVi6YN5xKjGVv9P!*|6VoU zHJ)l3dRr%_2UmX*#ypbgr)k@dh-LT?np$ z=>xOF-LNdy6oXO{|kC*Ta7|XX~fbGk8 z+wix5o9m??r$Qkcl&I_kEExTHogch#fA4;tEj7S>xgC6|TjlRSP!Dc05T4|%esMiK z;EJKb)}J4UOGItk-)#GxoP~&j<%I$T+An*Wa0&wI94w2j_ERyq>|l2Ejv9lvVeSWu zO~ta`(eI9$;H3Rl$)9v7fPJ34-E7zxLP0eBvbn)Az5ZQ^mpNk7i~cGO{)Y2rl6!wy z|Jonc3OB!%A$!&MDvd8XmxUuUT>y)bW5h^fCE6jWkksMsd%!4-SvOZ~o6>`Gt%&8h z>W7pcqT%X&+uFgGlI6)L@hklCRxgCpr8n+JX*HLZ^NKXbJy0VFQmT3g_|h3z2Ki|* z>`SY+%pePn2v1b+YK7YCTb!y*8Ium{C&GIRxdsOBHy=C-U8XsoQ131HK4_hxZbbc=M+HOXhjqEeB18K`n^rz~^DoRBm3IPH+QTaS* zTPWrJ9u+PTq{~f459>ZP_bEj-`=|#XxdYDMz11UK)yn5b{r0``m-}4j`?!Jx#vUy) z`gpCrwU?~9YdeE>nxVhM?w^5jKU93<2NxS%zqEiQ`~Z_cn_7DvhM*N+dC1KI1_Mgi zG@aKXHc(^0YTM62t{QGiS*Mp}=EHeMC2lu4BFcgjw0=G<*WX}GIr>8gMyNEbTHN){ zyYg%HM}(+q!gp@v2SNf(#j}y#6j7WpjpSKJrjtib_R;G3TUqogrpr!AZvsKDwSpx< zrep84^NOa&4wQ31qb=`R@N>jw2EqiYpic|L^wxkoG2>cI1k=IjydBSpbLGREXhizEF5SEnx&M#bNTmipYk z?^8y>PF@Y-Y;e&0$gXYIe-7JA{XnWf9=oGMh+=EbzdC|$< zilGVF3L#$#&+7|vyKF5uyF8!K%GgbiuH+1@f}N>7+zbSEW??yn-K#d#L(~%)pUxJb zSX)?Wc!4$OYo(18Te4d`eTriyl&KcvwTDzgWlgSk$crfDd`GAZ?E^I$(E^E2Z0BDLwgMP6^BhF+b1DF9qzR#namQG z-)d79uFuv}88w9D8E-%6o-T$aAeDZ#D+_#3iX$7&D~84o_QK~Z$AT@^KNG@(!Jnj( z(`a{6!FQ9Vw4@?o09KILb>*>u)w=`j%iI;>deXcU9}`i``=5m3I>Pk=)S~Vsk;Qbx z<4sG=AA}-sPRtw%tqfoI1l~x0Pe7rguy#_iA@pu?mRYxIQ+=&WG%t5HO*vh@&$VKP z%JRaDBd5W*lUnJ)&X|n{)#NQdQB3->_X05rLCLxsTlWi8A$9}Yo_DC*;;o-_nN9zU zzF;DHUyk=|QK&7hn-R6|Fcea^_`-@ffr?cM%#V76%gynz$wN9t_O&+VpUiGK)N~?+ ztWzTc$hGZDWkFBkb2s`?Rt(^{F+~qQiq7BlX93=r^VmoJH^vpi_t$GcFZ3xv-Gv+7 zyC?S3?v!C4aL*cFr%BWE<&shLKNI}i^R^RoqK|WVQ16%9z_4y0{zJ7|cN%kjxRtS6AZ-o;fMLUXLqbNy4MC?;QhH8rfM1uI~A>l~_lUxxZW; z(91uaU6agK0!G?86pxm%`NBLvbAR;oCImXY$irO@`ZJ@uggj|72l84((y|NgCF07J z4m`IB(C>e)#_eAJ=#)wTpC_w~YsKLVoZs6Ncs$8)ud3w1`HtOa&5PjMH@Z+Ai|WiYM6PF)WEDG=}Tt zT^LbF&_2^&Nq)|YU-=yHyoKw{$F|6RrA7xe?G1$(p=R2>Ip6?(r#iJ>uv-WEA})k% zYy``A6=H)OBZAb~vx}G3fqQZ(F`f4oG>m8sgaLmX$D=N@pAsMHO?_<4Iv7n!$#lNf z5?q+|ro-I!aJ{KFD?2>}TI@vIS6sno`fIya;Ovx@KsUNgvQ7j{u~SF~A@Jh+3mcxz z4xbcQ>$IwjtBQOYJ0M7rE(Xz*PkyHKkil}pbnkLy+54!fK&#}|Dph4TlY7c=rb+($ z1`-bPpKH3I>_okIj$aL1t2*96@PE~+j^W4#c%rEs40}bmEdi9<>u9O;){8-HWbWt* zRo^@#M)Fk#$z!}CnwFRDx^AyS?7n!ty-@6xc)M28N9xX@-gb!5L!rjnCgicQ5hG=S z7B%cl!F~kz4E8E{>1BN&NtabHK5t~iQ=(6RsP$~<~o zEwO5)ScJux61?==!A7Bb$bf(E)eXVChV$-E(#s`3x8fM@HuQ%S3Dt?cx4?C@Mxsof zg@6MRg-+cXoUOt-I%jeV!o%};scfrHPZRg=Aw$1P@Tz6yRdbK$@+)}C>@!=QUppdK zaibW9+$LrrcqoyLco}&~_rQdUmkYogD6Ckp73<=J=#)>4Ku_)Atq9@M>cN$qEUw3a z#6q-tecj*|Z_pT(a4w81DSpuE*_KeCSI+WUpzT*i9labKWq0C!FH*&p%ecir2W&%~ z+mGaZ&Eq*<#y@QX)*PTtatEe|%ta1#()sG0N(vyi`&L4Cj}sG1l|he?A&=*#9XO^! zEGEcC5HvRUW9(T#)>>O&1>tueK?+l{ceSep{^_rj8Jj5;;BYS0>E)Xj2m0Ada<0-I zT6}x&Gd)w|g<~vJwy?*Z!uU=nma9|hurpiUIgR9Yq^M%IQg+~78q^zN3vf{MEko1u zVBvh*YLc$HRv6GyY>!gt`$CbJfZd-g9*41xa~ANa4l|+qCJB74Q{BtiwjJG1J2nb8 zm*dM{0XkfVP#!6d_}V>CZF|U@jiRHC?GgXQlxi<#0p}jkYJN_Bw<@6E(YX@4%Z-f7 z^hWo^q2&ABAxpKPuvp&VPZVND?55>R@TA`tM-tg$XVfEinP=1}A`S^9Ik%;L(Y!|Q zQ?VVOMH9DWk$s7M>eHSxAB%Y}sn=q$3z;@wD{STy8%cVKt(t<~wRX4uK&Q&e(@L?y3*AFO+4Bl{lnqlyxp82qAns zb@PLdMzBr6A~|O&p=slgpRhqba@EJuE-}NH-opR+ZRd2TTzY_+f#@bmAquam;=)9Z|z+^88l+kc{D) z5vE{Fa;+cac%t4tDpI+ruw0LU|M+e=xjm`~oot4KOi}9)a7dgePH)^Lm;64ly55O9 zux{#HH<^v%P$~rnnk0t)2r5>Ai~I9HT&d%$7vuCf^h-v$4(=g;uk~1L(BK%w7{^Lsze{MWSuf)n~UrgE;MT26ToL^b{Q zNq%#;c04qM@UG#`V^E)0aBA;r04kf@FFrUo_#~gF=Rr%g`abH;?~GX_3x@Xr#^J~7 z*dXgG{7%n^%_*-Ddq>?i zWs4p#-M2SVve$XY%qGpcY`o~RjPBtd#GVy#DU18siLliV^w0g)D}i=AO}yVI5u~I3oF2)n zou;Ct# zQf^QRyLZdO@(rBU=sR8DJ(IZ(*%(;r6!iP=CM+TyOjm-Khi&WE{?%NCNdrH}m+i$1 z=nsKpL3jsM=;zfb%R9rcIQ>NXV3FbMNWuU&*)DSh$!KSS4y`Yh&*?EmqUCP=CYu;C z-W(?QsA~k+$6r4C{Gc$Qc{#=hf`?&yA+;@cvo6?&x&xV-44<4_hPyVht8g4y3x1k6nRx^e%Qb%HADP+?^oPAR*m+5xFIPvPt?T{1!er%L7k+*#(G`H1Me*;q!M-Cym!%(=v{( z_KRaGH=YAUh2Qs)NiF*GwY-P*mz7F!xNDw3dZvosG2+K(l*!R+ZQ zec4vwQlo{2v1K5`$s(RGf~yyyBIv&BfGAAZW9@Yvh^I!}T&&O}8Na+UZ@lM@+1HYeG?;YK3_!$4Bn>ISQk3@8gIk2GxSKP;G3}`MPerqfGKA6&8<~2%-}i7rSuvxs)Naq zS(iDOp!B?)(&3K9^eqqO95^oJOP!O~Y~)>)*+=u4A6dUJ)hH)!Jy1!Bx`wf{O_%%) z(!H5zt14+W>mvOuTFHU*XSDL8(H^Cl!8Oh|QZ4|w;*{lN&eu`y^LV0cdY8i4;ruz6_& z>l_`m+T`34wKu$zK`u3k^Mw!EMgyiE-8vR#$tehTL-qx^=#NEt!5&}e2?T!*OpWTs z*>lG6ksP^y6OvVvAA8y_+qozQ&!5>fbqNr|I$y0a_&!vKn4C)xe0;#eT($9|d)ZF7 z;my9@fhpXzE{`kRyH5Q=IPAROmGJO-{F`-@7eDN>4bL@ygZq5J9Cv~N04lH0Zj-{; zkqhGI`VL*{xSh6s)f|JUF>bK?2+-26c*3G~nL?1WNlGfcpCtSwM8Co$uVmBhWey_VbkDJ6^ugxnhAk;+>`o z&Ps#C((^ExRP4ZQAVXL9oX9+9W?}AYJ(0f4yFh8U>**i5HNPLGP|<1pc~1p#&q?`k z^zpa=Z6sMohQN->UlkD#@#Bwfi>B)B55%rTLO#O$NT(y^U@N~3Emi;IsptFOzYYVU zUOwmf6VRom9}Me`YQb1Bh~U6j38gq+jPUxHX|s<51B))~F|6$d-@hrVi)5Tn9p4gq zhkNFe#7*M)zv5#k+AY$rm_Dw%zFui$U-_|c+B7&D=~ew*7n-Bz5sR~LQscMA#0*Qy z^VucsX=f1c99)rblz?+2IPnd)yx95Fx!#DrlE$Xr;owmAeGtWTdux~?y;975HdpXS z)2EvkNPSq0xfRq&6Z!v^9+MD8JU6u{zO?&lC%#%BpMvA>&=qq=r}DqEV+WB-$`TV; z!=w?J)tAfMIpo!9xf}^vWy?E6KKcRo->*L`+g`ixZ3?Yl@cmrxpnHBs(5rxroeM(? zd{i{GVN&u=YA%64%O+(p45r@x{U=y;!j|HDQ+n!B<-SWgWW=T%-FELxc|-e3ve?z& z&WF}mWY}#v{-Qywe~W(WnY8}wyBi1T`FLjUVrIiNyEmSMtzYNcG}GicpOhAxeR7$_ zo_rmh`bTEd**;Wgh@}e>bO-zVg}sEDvQ^z4inpRaRexDE7g_*?KZpc-h(vpU2zP=S zu;|s-{{C_0ymiF%Oh>ofK!bhU+$&$-^ubmS!6`yd+mVEh%9fGa_z~lj^pB4uJ<4kj>EzL4 zQg%E-O*)#RM)S#(c87LA>3bV`7~H17FH3Ldf}b7=E&G!Nxl)wZ^C`4b6BT&&>x?VK z4u`4Z)K6G1xPA`#;VU@^_t!P;^bBc214t6-KR&&@Mw<%k)ISXLj(z>8g!$-*+el&R z^YRzmV|p^5JbDhDOel;$&)F}^-Zg|P)2`cD3^5t@uxa{e69UA!Fdhah*_k>Bj&D9A zkIr9n&CyTE_I4`Bp$d42w+|>OcC!l<;&N9rHK56dO_snHOocqK4$ay00c}Vw-~;X zhzvmPG`BZ)Bp0p+uP~C?BY6x+Hf{}Lf%^TQR|dBc=)>`S&7J+*_f7?N${rMNK!!Ru z$z>_Zg&Rj*eXOq?5I+xzNwg9FcWUhCX}1eS5pUwb70(@>W63GiR>QJuN~_u^`}UUP zp#JJr8S_QQEPl4os@Rz-)@AwUhqz)Mn}d42{I&lHi+1CH>uoy=?sQLed#U`p4 zGy(NiCEdW(F#WRwd=dh?7Gd*!(Er8WUH?V-FZ$kB5l|#VKtKdUq)WO%TDm)i6c7ZY zdjzDro1sHWx@YL_Mp9sap*v@Wd%oYb_C9OvI{VybpYz!F{aC-uUvN!*uIKf7^HGc= zh(|~f!b$K&>#wUsMc(azOw1DU21X>E^Eo;@B$cQYhT8%6R)3+ZlUCZFxpk-Zs%@U< z14JAUc$PoTQ`5O_4>~&BE)ZAnsC*vptNG0=ae>6AJ1Z^$r{JC_o*3(FevQppP)st7 zx8-y)2vu+*jAzZgrnS}{Of22MM&$qjId=ehSfX?@FqLS$^!$8aIddvMi%es9IrD}o zGgj&}{k-Q?W#?_==S&Ju{5)ypUtF@pWY&2U0`a_C5d9yt+VF*r7pb%~hgvlS%NCbP zQ?N6`g{eL0utR+juhZ@Ry!)08k5enqie4~=pIyLFU`jNDzm}ELalc~!Qw`_Xe3sMNndxZeb|_J`Y`PRSz^uV6U}e5(256BiEW~2o@bk|cVD5P z-?LKm-zK6vM;@kP%z4B>CO$_C;)7M`b?rTX! z7rF5odiZ{e(ak5NXz4)5>#=a&3uwGWnYoQ9Tzf8&V>?!!?aW13uun(GAUJGhMR4k= zw*dD&gSB!ms~|0Di?eP&xP)jUN9_tt3QFDpc}YmOZ4oC5N*0aKPioeLU*;uP^w3O+ zTiCN{_|t(sUweOBDV^%ckZlA2^7q#CWC3NgwYG$|K!({2Za4hsj~oyj!}vs8!b z78oc-04e!V={6Ui(Danlm%v}nTF;eRWfnGY>z>(%g?U-bP*vJ0+dB5l95KtQ3$(O5 z(`0VPo!08Ew+(7mpqSMKc7{Y2ztkXs)1GIm<$R|+hMJO|IQ_on+>8D%8S4)H%Sn=l zF9xk&AM$NxXomO_B(%sfHlWY#Sa#7l!)M1~aJ7IN7 z)=CM>+NHqsJz#{d(lDAzb8+U^D>PM@|847RtVrM@S<}6X^v$DN&q#_a`-SXaww??X zpA4bzjl2R=X%G?SBOvPNjeVTAW$IM&p>=3R{NlQjL`%ZgQfeHUBw|pCYAI23U)8vjK9dUWP7qxo)XiKTk%Kq)z4xgnQ62I8aR;)Ac5R z(@@rYwRHbhi+vm2Z%hQr>{OF~U+PRgoOE8A*rGUJ`J&FgDyj4fpVDomY>e-Z0M(O+ zt2X4`1k;LUw|@5zBbDcI+eXQi4TQ%CPzw{VzCyhCB}~m~0k^g=)omTbwu$|?%Q3Y) z1>y4SM@Yq~YhIWptN6v1%45V&MzF!cK8wWy#-kI8(3m}|M@>3S^b1;4=_AhrJzm+V zjq>D56yAJV22Q6fD4Z-afLjdAK;nHEE0ySWW&Ifg`o=j}b z`X2Pif_>;{J!ES#7jhn#ax&nSOq|}yzkW{x2&>uJTmKHd4x880ltk!#W{~+>)SBs- zG-$voc5fdhHzKaA8BJURt8Muc6e|ontrzN{je; zsf-KbR#3WFwsL;nC_B=A!aUKvR{`xacgQD=(Sra-JI6$wU+P2 zf8C7gMZ|tl^R9{gt zx(D$b-BT<>gd)7bDPL0!F?&QAGP*yJxa@SZ^BCCJp!J>CmWNj}^t#;dEkpfI`+>h+ zBG)6z-MuXn+u(z@JB;$%s1JPKO2@f?vq4?FUv$+J1}X!)9X_8qcEu*c@hVcby*9KO z)HG`vXKZpy6IqPtIj;KTD&y3}&lF(jzGF_oHm)$S(}IQ>ctZgEE&{l95B?5uLSBcT zW#ctUR%)VO`6N*VdEvIOi!`S%Xd)7n+*T^#TpQLUzGEvjWx*%}1~rFypeUHkZEk1L zPvqp{=kZVDcj(T%;8Vo!sY71`?fv>uF)}QJ5P0otj}L-aTbMv<@^wF)pf4mq z&Y;w>PV792mnPoNgH}TPaONFW&EvzaXo(6%`|gUVqL!2RJ{fz~o4R|trK~*6z8XRO z!TFYDC*>WFwL61y}_0CdVxKu0L}|L>ZxN&uIE!RiCJ0WpEHGv&jRJz`2_ z1wd@4r_>i)EZB5!k;;F{gi%#U3-w&F74|MeL7gYy8E3X#$;Q%0a3)$ensq%8@-eLE z3WTg4?@Cs@ppHzA#HU<8dUDms^x9kECg*el$o{biXOL7wFFQ>P=h0Cm|1{aRPOkSU zVINi_u@!_A#)sc1z2+pTeLGzJfhmCJF$GA$t?8TTN?zhcI)Y|h%KLLnbm{9i?~=H! zHsFRb)4jU;gO0zJkP$~p_CN!oM(FBYKbwR`byq@S2ghIhS7!g};UE5Mb^poAZj8P2 zpZ<&YmHXlcS9X;D@{|2sST>6Pu)uX;=geJ z*?;{EMES2wo@%$h^IwsF=f9*`4$I?6cs`^2*Gf(5c!PW=&li;c!YO5gklIc@1x~%Q z-ELmxVV`fRTH#}#he)-#yr{Q)=5_Y&PygkbiSl2IE`Rv17FLx1GW-|*tHB24zg)={ zrvCI_mlizsL=7G&|8*e8{tKCn@?SiWB7lGPUw}XSR{_l*{_BkCf9Ai~a{ed&tJ3W8na$t$uMzZr z_FruQfBLUBw13}!aXs{{6FizhO7Ro{MT6V-}$dPvVZnpP_q9r|Hb&<+kYt!{Eh!o z|I>eI%m3lOKL6>z>i+a!41fABYn1RI{!7^6ul@`9@-O~N=*vI(FXEhk@?UJ?_n(D*Q2vX}=nwxjSMpc?mDP#zUw~ee z|7!L8!+#N>{1=_$zvsV%H2*#SRr0U>R}nADe=RrXD>pjt5^np`f6=-Bo&Qok`_q3B{FDFs@u&ZaZ~L$FUv{eh=loao|EB+1 z;`{IFzeKD4@LzVK|HFR~|N3|Q*UW!M|0U%45C8Sw!heAS|6~8f^RNBavhn|-|BBZA z8~-Kb`xpO3{U7_U_P_OCt3rS0zm%y^{!5waFaE2j_;37IQORHY*VOc%{tFcGcm6BO z_ExpD2AZdVR5{i!b4F#@JPIF#P@+dLhvHN`bX#cjV8ro zS@)LD8m-_8O$z;@kM$=vkt8iTwP!~KStF%tzYXQ&CJYpQss*o3!8I9058lZ$@OH`U zbI&0rn|>r7d!NaFO*)huNmv4Q>lxJgjdGzJ7CWu@atGE<2zOenC_<`q~_6gn^6|sbw9bLafnnZnt*V1(^Au1(@~8)HsM%X^(5fmt>3i z`TO%2q6aR|hO~pM1>3%~c$2DoPmW5F*g_5inbHsYvj@HIFvp;*dOY;dyW~eWA;DkH z3IRr(+{LLqgJ9Csmo?dp{`@*Q{%9&hjIGDq#5a?VNWPc!ye}uAye^Sh%I=Jgt2JYN z7%lJ;P>wa6E1nMQ;nL+5`6VjkYHEJ&eBYSN!(xX9x6cE9(C+Q;xh}-62 zorz1C0D`8V zxdziNfDY^{1QRVf8vO34nWR!dB8UwZi9P8hf2LU_q!g-7LcsnqjA@RrAaX)UF~K12 z%CsXrnp44MMx3M=Q z@1b|J$JM~2=<@1xE#dTxXy+$eVZ_816A<)d=jh3;jjGt4;+X%(d(9Ny@uPNd#OF=@ zJb&V+IU@&J&XswT6ob0d8-|u^hS3~c#l|Q<8;pCt=m$IK^)B?2Wm-=SF}at;K-6Vp zfB8||A!W6q1F}pS8uGOL0BzAK~^4N<@W{%k0t+%xcuA`7yFzG_Oh-~?gMAZS zph$NQyGU^tavo3D49%AEJw7b3tQzD!x zUeg$kebaVRt6;&gHRY3Pl*Gn`Go@fXkN<=-)!Lzq66!WJS3DU>6*>kE_Y@!z{lfgt z4?cg%D4*@}>J{BUeE;jhrve|^&|Z-$M1M=Vb|{D>YWKZOvQI96Jp2kb2n0AQ^|B$t z+Ii$f^B1j~QGLX*IX<0nsMR&pT1cY0oBhB-;f3!>M7oX4B zTXQt1qO-{bl)3aj1E;HG2@vhC7KSd9(mOrteJZr-Up*=*T9aB<^Ek374`(AI&%X`J zE$`k#Jed@FWW9N6F^<>Ypn={%(C_{$)#zbRf?b2_tIfj~T%&lGj`klAi8+huIF3%b zZ*YEAZq4A|cs@p`HfI~KnEHB3>QVI6e#lfvq!)g^olJgnY#w_0o9QiO7>lTU)nMO1 zFl7=CrJBav5c?8Gex-`ovVH-a1B~I)_6kvEA@HVYE)A;W(wmUJG#kzr!rTAMYwH=M z=*<^%_=);FBqjZca7$ryxZ3Lb6Lqoce)d3-sIX>X);dnp;hG^D`6c7EczE7P=Q&O!@g?Jg9V_hcozab zA9gmW7E^w9kdNwAcw9>-mKpKDklM1h&Y4JRS$nOMy*jlCGaW`ChMQ&P7Ds3xRAr@Z z=N$F&`CNm?Ze_$dIDEK8mr-|RMC-DZ^JZ?lTdn1tPgEEv`Rk%Us!}RMl~2hqvvmLF z%g}cIjlLDQcSKoVqSp3S37Hk}eVto;R6G459hoUVm?aw2G(6AqF~ajYi_aqe zIjJolc|7h7AAtW1VUl`8-h5B|(X&Wed1x_L0$MEJQ%Si7^sC-`y|c{dACucPH%k)yX7pKD%|@0{nuevP6|Gg>{= z1&u~QTu!`p+s+h93>s?~qrT%v@}%GgVQ49-&DUev5F=9pXK}1>r~`c{z31+JQ{*0D zbi0I&&6L7NniWIR@s(xv$DZE0;&?Z_%?fQqKf&& z75c11XI`T$;p=R~P~N$n(7$5S2QO%1NYl=YwBB^ek$;azoI_!)lz1ugE@kgYB=Djs z(%eVWy&})qN6S6yozW^$7G7$j2y^i^c%|4UM3onS%D=87yR{X}Th$Z5taINyjXF28 z>l$mZW}6khU=8p{^(RRi17)w0kfQ}5jnG%D1Cv432{ezvy`S`n^ANzmt%r$y5ujde zRgs3fR-+HAn*Lhc`2 zc@uE<{<@FoUSa09&4fqW0luwdCy$b%Xk&4a+l<7u^@B)S3*n6Y$Ux@E8_(m-tFz8; zAPORT!;DD{D?R$C5@O9sNdCpW$>7jU2r<K)RWA@(_^9`$Z81eeMKAyA)dN?4nYVI_lr{97u z@O%qv$FcnprgQ0<2%|p*n#wCr(#F$7xJvVmFVmAn<$Iiz{LQaSTdr)c;p>p3AK?{<~n!|vZJk!50n+ocwaPyxb7G3q4?TcrHU-07|AS1@a3rz+TPkr-I`>bBW)+# zG7I3?*2X&H41^Wlw85(m!o`};;qolcJ?BQxrk3>5G9kRMutiqzRVrA zg9mnaPJ;79g>u*E?=xG%(JruGJ%6R?zPf{r83U5W7zQf|xut>6eE?s_?nO#S-D_V& ze`ZwfLKMMR{deJ^7S6@5UW+7?lT&*hd;p!R3Xd1vl?d^q9YDOWA%$5KuPI0mrj{5H z)D!32?;V?g4)by1wD>M_B_JcUUZ?vUhDi@Okn1;whiQ&Gy{6;2{!~O?rsku2(3WPvBJSXXi}u;DcXp{7~<)#W|yV^wl}XZLmaYLPN=kyC8^__i*N|^3LidQ;~b} zmmO1@RqvqSis2_RZ1EN!!CP#ByJ4a>uQEQfLH8}Npr$oVK&e&lmoc97vz@fyuc4nl z{pR(3i#zxav1QH1}X zyN;*}J9o(@Mj2-Vlf*Msoy0D6(v7Cep5#QVvKAk)`W;?Rg{PnKW-1c@YbQNS4_oxv z#^VN@6xWa%oDEz2uP;O=E#ZYc2&*Mp1+J#*P~WfJ^o2@SN67?t#l1ea?B?jD+1|g> z+$T1BYkzeWarIQJEasw$K6>ZN3baU5=c?t&m(VGOy_{)3MY@}xU@46Yt?(5CL16rj zVn%VqbUIi4gisQl4-RBH4lP;*-ek1|93oP9(rJ z@YvK2Erw>)<&eQbwGoO~BoicraetKY!0Vtvk$|WYa62);F$R=u=-C zk6*5v@ky&W;_%{#W$%e1naKAy^qjjIZ*NMpeiOaavv{~hEyEaKjyFO7(@%`xkxpe7=4Qk%DfCugQtuK?9G*MF_pCH;{Xhs-*tKIo=g z45EQ+MTCmZ&9&DR8>7dZ*>}U~7QWTX_1C8j%m!I=ojmZWmPpkGzI@;M{HzUosDqr* zowynedeqM4Tx#rszVvHCIM4qox?@Zz;X>B}y^?WY$#Q*j7ccc|j!RGlK%(WUxsB#C?u8a4kpfSnJUR zBR>Q-Cz6a^*g@{#+0#?oaGEh5vsV!MQ(LCn)hJ}UQgC&ndZP2phxvosDA@p?mto5D z+v!puf}xijWzzk+@XA5=rbuHQ^pM8bF^DQ#ltJ?{c^h@o;xa}H?{060_g)(49dCJw=2)(yzhK3e;DPQvIFxKZ;DDuF z^~&@0Nz|j6#I5vIIqE)>w0OrkxwZE9$4$tb4!>c<(5;N}C-~^S-|&3u{z(TxK%di< z73&acb-8F#-tQy;CZ<}dLJ8s-&M z#GZfRKA|C5$Cp8Fe($A)L8M^(tVw> zYW$1=Z{axmrm`3MriQ7LIT3;R??(1KA-Web;&|q_!4?u;Kkf~9IN!UkOS*Ci3T2U^ zp22%9%H6v-6>2LZ$ca}3j_D-h(jM90&KOtb=yD&(z39SB7-6ad^Lg~8>%-|htrSM?fWTCgxO2%uS9 zMz?wAYntIVdn2z%V!%UDcsQ>zN`<$wnHwr10MaQhY{*?7HvFx(GcyiY0K38{D&hJQ9PV|HH)T6IZ8>TAn z({{ZUN&xbkA#IYkSZM-7mG>FJuBGqO4Y#^9(i33-!m&Q-EW!7El|-O^>s#y)Lf4)4 zR|2>1q)OBik-nXf=iZZy-k%D^8z+XNmwMXW7>*Zwr-Z+|U|1P{fRiPHgUeDM4$LxV zOQ8qw4CK*JbuY`{y=TMdZ@bBSxn}w7BXAj8!Ry>QS~xJT<{_={aSfXHecQkh2U^(& zhd}+%2i0ZbI%W^E%WK+yr2ND;AqXfIT6~Sh^K|}Gu(<2ynpflr#c>Bw%M!NRbe4C_ z!`&yt*=IgCQ;A?T@fKC9MD+Cs1VKDk3tHA&rmqhe8!@YKZ+7E_dqzEjkA5ia{icIf zd}iMuzr(G2$GI9ry8mdEdnl199@jyWKZS%)P&w??c;WnlA{Wr*zJ>1r}% znbsCKU>Tp!eYq2}QueYqxpW9&G`tBJsy;S-6OFTp1uU4nyJhIa`F88=3vzgPRQt{c za2hHb z;z@4guE{Acp;<+k;bT}dK&^GoKW*oZPBAHmmMr1Ux~XZcNx<-(%-FOQgVcgTcg@IxsBEeeC$>t~H24pKcY0UXlEL%cUaxmXb}Rf8YD=(tYRR}j;@yn3pKg~3 z5*rDOEhTo`SAc(31*vQPw1g<3my$SX{tQ#CJg>nTkrTClWUX6zu9@JX5T!+gs+p2` zjEZM{8ODRkw_uG1IEXqc<&Z8HlkZEWwj@BB(VVaw-n3DAOp&z7pj%2Bb3YpmUJCQ2 zV)<$6j?u;3`(7TOO}(T1)0;U*g+1Z7<%DyL@U|Bo0k&+?H&o0SZb38PU=4?g+b7Mj z6o5{5iQ73&g!(P-8=%g?9mOOSfNGW3L}1|4d>F?U6YCfYi{R5S>v0yX7FE!qusd7f zSG?23VEAP&MMEAlYWLNeC_@Gm-ue&_;Ni^@)TRg9wm?dJ7YspGsQ-4{i6h-?><9qL z`0_GB;IE~E3XMk3h?=C~*McYgCLY&TFxFRH8v&OkG$ zPro=8I4`!HJBbefiXt1IGI0^NZ+l{;EQSgbuL<0hV_RBvFDZcmB|68g>7a%W?k@?r zRK>~g_yXx1-(te6t`_M8bJXB|Wdas#<~$yXj;P*CA$n>Y} zCZMys!}Vs$lVn>@I%Ms{xuo~xs;u5cw{tA>`f6;`yTZWJTL#KAw?65C^%QZe#P?r2 zH`Tf|l&qY-rwqnMb9ndVpJlvh{}Mcj%-=WFqjSFfps=@jq2UB@b6fA~B@=uG*K^?w`!L zoV*l+5U{%;YngEEu_g7v!03%Xtrum`7U zROltmh9-nls-))nb2I3LtL;RW8zRVW>{SX3;;~NaWYKK^vKPH`AT_HuN}Czi8XcS? z?Wf6HK(K^o$ra>23v68>u>N4u4HfvK`1V-~Zmj=vkpx57Wzc&y5u&t7Hd_mhdxt~d z(mxDPR=B6`;pELS!25b1-vE3+F2k?^h5Zgb{FwFG!iX}##$SV&6II!Xc{f+|Lv=O; z(tX3Pj2_%`YN1kgGIj+WtekK|eILxJGu{iRAX8l2T7^79?$|6y4Zk%KpS%zSV={ny zLWpk;?UrBOJUeDQ-iy=^&1ZwO1Dsk(}_5gypqg(2iAx@<0noloke{|o} zy*T+S>Wxgf)g!t~xh3Q)^tHd=d5>XecI*Tt*_M6QCs6L70RM^~;uL@tL3M3JSs5cV zj7a_w^Be?vDzdc0V7b#|NWB4ulQ4&(vIpUh%WB&{V+`TXzSjx=u)3#lEnAF(JKz5e0}pPs z{SwXA?1O~`W`N-!z#Dyj)VLBo#Fr?b)U5O2V=T8!^a{)8G-o*2xKiyCB<(j}G(Kgn z4zBTV5;{KF9FU(KVMHf)9(3E9zFaPhBe%8BIALw$^_mNB ztcIApO7MsuWPmI7fh^Z}yF~%$&#sPJ(4cmYczi1=iw|{mSnd*NaZPA z;{aNP<8lC}l>n9JIMk%rQ+|!=FygGyzMtN+hBQFYch0)*v1CudU(zDt$k02<(B9sN zn5G)#=QTGDEVPn}jMkQW<%y|42FGymR;W9*c1=W0C##t3=4y0>4y!p45^w2OgN^4c zpBKzt=}%vbslE+o!E)3>5LD|l9c3%uho4v4vw!qi)U^F_nzs2;pFliZ%4JYQo^nMz z2AR*WTz_piQ!7?vAJsLN0G5tazExMfnasaQW-EuCoH`tiVCFbpS|`JU_+j=_yj(KV zb>E`ahb`2Irz-qerA1=b9!49`Ek zNV1c7_|e2$cF}1C-BD!8&|U_9$*3cGx@eh}S5)@`oU~pvXHRE6LLFoz+JQ)JM$%=D zUCdeRLJZR^KcnZQFuP;Ke4^8oxh#+f8cVq*I2wZz0Y%R9wlCz45t)IF)?0M_Q&t$R4I>++(b;IG#$|&tgG~00TUK5q1?9qx9>q98unu2k3oI7W_ zdOM3`qfd2P*JwCMchU?gU^CU<(vahBomGC?pYCtV83}Rxwu8vrcs*(R^Nh3mB>?3d0wCPn%m_h&hPX3 za|)rBc%#gLwz(#TczlI(Rp(nk0hWRO4OB&EQdX+N&18Pr zfDT51;a0(8iHn_ex%;~i?x?J6ujqHOYj$2dDS3iL>oRDVcKu=H{ex(q61j`^HQ4_9 zbElx`K@EAB88IapZB4EMsAI$}-MYXYt&@eSi51<7+;`qU`;lVA5{9|W&n`pR+xt|- z+jEHOul+Ahw=IR|L-l-t9R1~rvQvcESKg`d*k=ihnN7;H)N`^}7c>p*G}rR)Bmy}0 zgaSP{bR8dzABzOmare_fo@T?S7@R8D5}hO~^hHYwXY+vl*Da^9=z)(pMGMC8ZxqEu z&U988x0NSx$GHMkicj;!;UWgjxr}V5spv47y>!_E^@OIZ6Hh4#;U%ow%2p#H$0ZQp zf$S7$7Q=@$5h=KF@XsSCG&w7nQ)oo`J{Muqx- zqi58b1ivBMi-TvW$Sc5B1}#{d3NW=PVS)d??unJgZNW2d$J|)IrD&6Y)IbS80?f~Z zbhyP6-=BVO<>s2Z;`9wYDdUSE4i_O9W=-I|9g+1w=eYLBHs(T30t0DB2$H!TI&$jh zH5ynvF-jd-ZF_Y+P;RveNZvL%LtEI$BIW1Uy}NY@FrPEPdgrb(Ie`c)Bvi##yeZfB z2|z=I%R817zmc3W1;sep;@NrFWQyXoe=*o>=;;{q2yTcdii?b4)O1Lw;oz%Z@Cgkp zH*;hV92;R)ZlVPL?)Z7rapXU4yWF{caCydie;teMHb_W0HFqCb!lv(e%E^CY9i@&y z-GN4IApAg{TlcCGuc#-<;5<7lgg~5b(S$v7|D~2F=w{uePJow_F7}ZLlC5f0wS?mo z`?^pG!vxL?*$8T|>U^Vrdzb7ix`{+{E9w8j##R3^5^aG1VHw(lzhmz*c@AQ|i#xfX>5wGS#m#wW!b^WhT*$Mm@s)a%lk7w;P^<)*5jfhk!8rVkX;L_ABsRe7}+Iu!=n(~e72}tP#5@L0U+)7HKgoC@oe)eRm^$rsg$y%M?aWBqoZ846_U0cVJ%Z|?19PH$n zjH0)b-%i|qP`<|UQYPuv2Zh%h4foH#p|@8cs7%PT{*9y3sx&~qQQ7xp&6Fz}#&_}{ zQ*=KgyVHA8%LI1^k7`>^hBEPXbBv0P^)k20BQQtZ*p7>AvqkaS?=Havo3MVbc$2d* zdDr`#ULKDP+x>o>1uEg)s}KBXS&e92^oHNJs7eu^I&9Qak0mo)-wOJ|HC6}WEv038 zWolE-;{+3vHEm>#t307C_Ko+>uEQjI4y&P45@y2^w|$37qgJ&DyfIYim|KkQZX-mf z?iJlr!88B$P?)HOD9Z&`F8o(Z^ZIV!$PRC^wfmFK6bp5D`~XGNg5sUBsqiNw*H$C1dhh9$^od%Ebx>~Ya7k<33A{yAz){#JgJyu^cXfbBFeNrL4PU@A1auQ&7*OhZ45i;xw(H+Ep$ zz$w4JdZ;h!hcXLB1m&i(#^GPL6G`1WvZmp*$b0Em`!WaTOhxT_d9Vf zfv0zAA7-uTsaMaIbKrJ%@s=Ki8-{o z0wWN8@MP7U{P1&gIho0>tLsepDf6YskF7-j?9H!!tv2J20+(DaAHpDx^j;i-LeHPYkDe??N0*h(2vyusGOBLfwl$G4ov?NhfGAD?_CZ1-H-BywC zaqC?CdAWvl1*>wnJ1X~;(K@k>@V5+@`r*8d*wk%)Jhzu1L@`7}BD+3t^oe5*Wa;%z zF?j6qBVXY{miRlrW#vPmSTNf>=P`QM&{SM6ZoV zDGw-LACmEzd;!ew5O`asJ;Q@?IMd5kp&6#|E3TrTyXYm4wZ_c1-Ju3jhv|FWebVTkM zN?#gc#zJo}q$bD4zXbV7yJ^8`pX3-Suf!_XOn-wNSs3%J??{DpI!oesZv6amE(JxaF_=~E5?rQRDgd{ z%(&gIzAxv)WdS*sNkY1YX{Im45p^t}`17`ZjLnOhVp*hD*~mz*ca_C*_pG zYDPI3mT;#7sZ7XYyBhWIo#w&O$8br$qbk1X(zP&be_FJD2NyO7R;Bq5^mfl2%Mwk??NgaT}|th_T;xiCS;isl6~UQ5ozrP^d$ zJ~j5+DWLd#U#e9=SP`U{ja9vM6|VxE;~|U?*fh?2cr zFe3+=3qCajKx&^p!Sa|1Uc}f826*YSeY1;TKiS|NbV%>mbZbUe@GXsBBmjP?qI+@V z>El{)y3jOu-{X|10T8jRWP2F>E8Y?lv)X4PV52F=yTDZ*|13$6h5Ue}A#ZAv1#pt~ z9?hPJbUxQV=qJ&M0Z0byUst>qO7(uUG!eOvN3yB%Fz1nk#C*v)3}VgLc$K=2rW1zT z)jWoe8zj_`CECT?W87M6btLMkfnU5o)gY@`ZTq$D5Bumv}glON+eIz}xu`5WiqGPn%7W z$nRx3K3uu_RdK-Ar5X_&=4m#|dANpjn}9;LP@zNK(9YAHyy!K!FV(3mKTaMo!x$k$ z^h$h^BKM734_zUv#0?v(Owx@gjWse{RWcuZp&FQlRNDmlh7P~pQ88yqTcU0VyQC+-5s~4baln~0~V}AA)67OU$-Tmnxww- zq_=40TF=Kf!{#qEFh$Di*qwTpYQA@bQvYiH?)ihMDyy($Eo=6j=V$ty*Rqpyb7d+Gq9{*Q4*8ke>4L zKH_+?mK^DbwV>E8%^{YOhCD+Z(o^T5Elh;e1%4JycA8We>9h5<3+g>{(tMv)DG6*Z z!a>YtgGzRal1}pYAAvE%oK$k@Smz(Jk#$IPX-Noj6^s%YxqNaycD~<14^G~#+CPc? z{cts2ZEJMN_-AkglZiKE;)}P8l9l&T7Q zb5I@=_T7hK^VI4xgQ}QtfHwDh^;=9*L3_-GU(?3flVJ`~%z=Gbfv4&fwX!KJ1l&)l zTM{vPli60}GmQsg3ckfLd4a_B>y1Tzlx0opT0CDIC!0m3Pxo<}xYO`q8%~pn_;0qe z6+fJfnNp-&HKPwZRo!|u-$&T=r9vbgShi;+5U8`zFMmtB7Piz^rJgG`zVJxLy3DD| z#jtu2cEo9;(ShbTJ;haOe35#K>i7#3!oiKCH|iw2s|?JFN_8f!)JCa6AW8&Sys4qPPvsyd^FiHg4K}Ou|yc3mZD89gd zs&yR^ErLo)F23C=#)^_T+M5*pdWK(jBY8hQUWMFnyUK)74n{?%yT+uhN4T2_))K3h z6$W3vrPWt#IFce{5Vty)o!fVIn(K=laj2Q>N*NzP_}XlJ|{4P z3c3mtymX_F#1TNK#$s$(lWBX67f&V2B#zK719wt`;IT@-4!1nR*~J2*S zZ|^myidd7$6|m`7bVbYpe;O7f#P{M%f^sC1^i56SG|L=is8_k)lh& zC#j~X`%3{iZ;bDw?~ZjhOqF?OIN!JVS)=(nhHmlcC)~9Sy~Ox}o6W7Ncb3ujZBkJA zgP^$dO(BVxCH78Kkhp8`*2)dXSUmArMe>aO&g~CNHxtp8XD!oR*-ZAt^^9lmWJS09 zg3}dbhtp4x-`r@h&A8o8Iul$dFZhls@JxJ*(%SfU z+@o1YHLO}D-PrixgKo_#ZoXskz+iAjj)HKA2WDVmv{id^4$;1mg{W2sd*YA)RG($s z?weHocr^9)Y?Q|U!jDYHEc3eMY1cF(qEY(U7G{sK7FlCQ$b(fcz{7GWi`#{;r3vq4vk zZKL~AhKFbCLjHCtD|RPqjALKWc27PNF-nva3>rrLzT!>+XpD6nXo1{VSI{Ux0T?S@ zFOKcRu%TOq`}5p_#e}lqyYRrI+u(~!MAA>U9Kr8x z=><1jQ)~s1d&Glpr}5}{En98do^~zg!FfEG{qYbO>EyQ@KW|!_V=Kh(nPRSVl`oHS zkz+Y4V|gn_QcYiZ?RLFLF739AqD<`nvqRSl%KsR;{(nDoO|<_6!8h4Z^>b~Tfc~!N zIZ*G$^;JUn-^NN;Erp!><9gMDxO}4uPHOtQhT^ODVz-G|&0g8Jb9Ekdy}buVSU1NG z+!H0xg@5&l95~zlxLSBF^1`>Pai2|--BMBw5`*j!ikK%!^bFei;JXC`tgkX2jrH*~ zrL@c!Ls6~Lc){i7@Q8-deOBc&jFWepZ08wTN!LbuVPh2bJG}#cH*dY>^WIJ1qKW(N zhskangQ1$y(b4kBIUyTxeXH+X{$8j%OfI=;>usKc#_fymZj(4pWz|LrLzhJgJ(XA1 ziAjRTLu|0$m6I@-oER{g=H69r30h%M`mfimktTCZj7`kUBbu8^XFd#(Mz&;c+{A)U z|2bBg)hlo9^B>eo?Q59T@vT?(2j$6lYL(R$nnTvp^9>oaE2=lH@RlHa9n~t6Op)i# zO7|b;lwzidf~9+foGU7o<<-(y8?(`mN@O)~b3@UXawMLux68HV z@5#Hstm9a5Q@&U56JkmT80f|lJl8O5jEg72+n^Di1P$%a;N=zHRP4hp96#|bPNnnv zxY`OiF{ksdh;k*6_h*oh^P_$kjK0=O{>%Nz>uM5l;c{_0L(|Mze}{_#?R>2icLN+; z-MB^G9?NRQPx1SagA4ZG89&53YJYe>jZNQbxb4x-M^cT9{PcX$WPYCTmz;a0!CS& zi(WuG2H<4rIR7UG<@lae)O`-UC5rxp zQ)SjOtcGDuR<|khIFADTt(ObkWoi|)w97wu*^$Q)U<)`p7ds~lvKChMhQcX}f79X5 zlBs!{Limte98Ng-L8xLy<87vewVWk!N=5#efm#-ie5-cTV4Z|$Jv^cm^heC}%#^w< zV^JaEILRLn67qC#`Ca0o|6W~9+pPt| zl~k>0=w{Y~HU+CT$Ky!-6=qMS##W-4Xn?h6=7XlZUzL*o4|nJJ)KnDiYnmV;peVhI z^d`N7fQU*L1nDXwy_ZlDiijXx=`DarZ_+yf={58Yp@$wyAS8k0a^}pLd*;sEZ+Gs@ zy&v}e2liTf&3b?D`#gIh^=-X{>g$UiOQvKdj*ffwxN0ngia}9%Id6-g?_pcspQ-~S zcHIgKvg9rZHgjgEKz@)ND;nI~E{>~DOI5D_bjs*dDLPmI`KU3w0WCPR`s1OSOO~I@ z<;Bzahz$dJFRYDMV?FQd{&@1k$lr_k?g15s8INpHGyAX|410dgqjQ2J5l6e_o99L? zxyn^OfArQbUmgc%8+3$pSmj5+&82tOwZwRh*mIjg^wlcv=|%ZC-pxqaKU#Pnr)%478>=yhd<+Ej`S~mAM7glV02%@3Z}NtavGH6 ze;ihBr>roowBKR9Moe56sR#b`tU0{9b?c4LHlbNS3AA>+8}V0NtJRYmUKZNJ3s{oX za^E;^yFR)sdvRo5zU{~ymxyzs`IEE%!3fxiG9!bL_)G<)FZ_RkGgXE!S#8@axy7VC%RWl_b}TEg3W z$4QNbV{g-kxp%FF_Jo>(O&$4F9_#(Uq&Dr{y*SY71&mw@Oy?{hr46)dJ%_5w27uPC zr;W>h*k^Hrob5qchwhiFeuFmUE`NT!LRKjo}< zqByO0p9lZnLz%j_a)IJ5SNMwlUm=+q>fWb9QB$tk1sE&FQH}8h;@5A?1aDyZ*wl!h zPmB+Wg_9+|QEsD6O) z!f;i2d}WRI*zF;J5&$U4R%Kn}f}R|7-AOKITQ1_sXM5Uy#bMg+VRe#X=ddK%CZ>DN zpdtEBD%!lAr^RexBm9Y1a5+*yj$0C%WVMD;1T=j~WqWFUQ=o@ew!v#C;;G(G4L22% zbqy!s7VXUL03=#Vb0+u3u*o@J6oqFH3xO#Y?zddiEFq=a_u(sRqpQb( zk4NIoNCFHPr5PGNUXL<-QKpcH33&hQRFxrsi&XkUHz-j@U^UYy4ab7W|1RrmzL6HX zw-B|?IqE+5q@DOVBGbjDg-~fF>WYW9O@g2eUVNR&)m*}h+aYS#6*@*o?`x@jDP9FU zo-0&|g={t)=?VIq_|-_`9Jmn=;=rEMhzBB-qrqn3n6reS-t6u%WrFMI=n(bH=w6M^ z#6{lc*Bd(HzcHk1@RnP0C-!cnHRQ_g6DI2Mte~Tf;VWvZm#B+7Fn#vl*hmh=#T@QE zuD9>m!Wv(Kl)7a)$gxjBsaV_;(lWe>>GUxSlEwr`4ZhcpHJ)lOezg9dqNC(UClKb* zA9`3YW>}5dT6!j4(n|UzFcR{g?rO^Ih#}2eb#?ZCQ&&TF{!?8YVg)Edyd>pe6rZo*Gh@EXYCRqHMIKscz;{ z&gshDQG6Lu<8>fA+BWiyZ(5D``<>msXL5kQkI+l3P!f-|Ox7-Q*0WfuM|w%zsF%cU z%mX5Q0VxX424S9sfR}qpi^f9KpBw81AxGa3|IwNOle(h3xOZq9P?fs)^e;8K)oALO zRWZfqh`Ey7wyKwS4vP7I%B)vKCgQfbQ>{~nL;hsEg zCGlshcxGeNQkxO!fG;<*Y`E*}KZ5=dizs;WxU?!0T}>!(0p3GqR3ledME$5tCt92=5C z{srzRy?X%OLifSw@1EsM+MB)KzcZl)dVB-N|Momn%+gaVWbDyG7D5dfuJ~yKS~Xqg z{4IdyjYWTz8-_=_O=#F7qSz&@E?J62xcRf&d0~^# zfbGrJI|2-w6HIqw9?IBQG(L%^i;3UCmc)n2H~s3RKA;-^F8RR2RmMX$+JAmvhzMV8r`aqV&(!>MjDO>=!0E`ST6*iWcCP(% z{xS8VtdA+o_-gYc+!=Zz&*@IGqIim$9z%PkHuC4}CLUu}FFDbN5K1OGaCeeLn zO}x`IJ?9~@=iH1}3hOZ8TfkuLj68U1tQv69*OEnxNga!n?Qa$K(}TjU>&_@E0O9w{ z%m!L-(uxw;RyT%II;hZJsSpN1dPFTDY~LF9h0~}qZZbnp8#(Gq-4TSdlZ4Uu1%b4Q z%&^nYcdAOp3VoEv=3*fCl(csk(z*!AsJ1a@yEeyvEDi7~g_X!;y5o5xIDtGcCX zlqS^T*5^(oA*i z0n=VS*72KNx~U_)qHmSV((4KdKF#v(%G~XJG|uf`N_sUbcO{QxJ;>!hshUG;4@^lt z&{KDq{0J&&oXC#*3a%mS&?)%2P>A};u>+NqMYPNOcWkdNeYsjCftZyjy4 zJwZOu#9b0?MYyNa{LwQ7mlEL4wGv#%s??JUvk^l2)|T88sVhMq?bED{H(q=3Upf*C z1M|Yz8rcMgr&g>}7#3TD$^nBV|{Anc}^HT1h&?{%q}*=#PKz!x2VH@u!8SnH$AlY7#~ysZD);ex6xUxYlaXvZWjRQ8sHb(WpYd*I4U0 zC>QQaD9m+kDRUt(OBE@18?yjbzM1?c#5@BD={+X*d-8YqZ zAHL*b5y#dS;al4@S^ta>qJ}MaG8CFf-5^)e5Zz%e&vR10BJhp>zH<>^|H1w{WQOal zfs9asaePq>_K^RkQ7ny_So|*Q{Q;fhj>8y23L|%#v=4+~V1~bUKavxBa#{WoNfCel zC6xAl)C@Vn5!iC@18vx+ix+#QE}`}J-#nx*uwxCoxXa9#VBXF>!t+g-^tO|gH%#X~ zpB_N$*)aa;G zqFia<2OZH3Jw7_HH|pXq^NATlCorE9n$DG-Fr6FImTOHpl(}g4E!^D-RxT_dS}9Fu zzW&}a)rZvYk@p&aytfk>eAFs7o9R@*@~PcF=TG#-&TsIlRJoqs(4#hkgW2yFyW=gb zvf=N375P&h%%*PtSS|K(jbPDGnsItBD!Mypr5vx(>b*Ipxdv|kHD)uh%s#3&dCfKW zt^}#%Nf8c|bbEas@YI@n$nHu@z)pBD!shB)`fYm<53`$q0iNZuty+2M>I7{eFZZ&M zgnasBuyxSaQN+xI*VJ*U?1QnUIZcFFl-VoWs6u3t>2GHb(UN=a@4h}N2_;0W z7P!}I-F0AoE7}XY3W+koxtUy^1!r?sn7m4drnVE8dfO_$Y&%NjUp0O7Q1oOfXFg8x z&YQ*@rM_(gfzQ;hl=~=-T!3rXfF2nR<;W|(4X#MH>5W*$^s<%xJL}YfsVi*l$L_DV zYlt;%TjJ9$=rin;jfj*RkbfSHNhGe`TT3hka<6R)S^tU$tq$a<^uw|b0U|fRL){Cy z1DkvMNe;NXvr3D}Al%*k7RBLO>pQ;eqj+t@E{D-@20f*|sbXZUBBc;jWR>Qgon6L6 zFhyc5tHX*wQ}zcQ3U!&&amM)O-A2mo%!rSC6{8GnY)WbI0WGY1n&Ucs6F-3T&u}Zn z>c@?YQzV|<3)vD3K@Y#1KlyA~`qjVLcEC$)lU%+kFTTfxq4jymHHL53%s|tB|rURuCNHF5{MwOn)6^GMKseB zkpJ@gMPmJgyrVPXl|IeucD~cRRGndJ(_h?6HL)Q*+~njW_5^>}s9ai7HhcVK8y?C1 zaFO%flyt~=(O`pjS-gH@$n7w-Gb0`ROWFaq<2H4BT$rfw>N-bUAy%{H6cp?LP&s<~ z$J}E#9bH{=c7QLwzP^Cl2l`~~M!569{9VJSB7vvBzU_ha!Q}8z1zvLv!)I|J;%J3; z>de}iJDazSo7qq78C}uV{$F73X z;}tfsw#o-1V8I>pv7S#NQEjLsfNBTj(dkF{EZ4bCFpsFxIcL=d!iIN#A2s|0Va{Eo zh;~pO*JMtZq8bAq_=C}D?{aQ$Sl1xPy*bD| z;PqR$`SH#s`$O*?po>KPkHnv-=rP^+9QN!AjbC-pECut-xTlf8wWo+T(e-M{hFlHF zKZ4JlBdW}7kYsNr0gcMGx3pv1j(4p9}K>eXTXMFLZAuCa8^1`di(><}M;QA5M_C3VNTVs@K~r(l16-+V>+c6~4>wweV_m=^# z1&e`u6(NnqX{$yJbTgUeo1QDfQ?Q*v*g+!KUJm+gxQ&?!uan-+#PE1){N&ny0k|x+oxjQpDX?OcTogm_g%GH8h%Jysd9rW z7MO{><4c7dtxEtqwg8%eUZT&Xa@mE&P{8Zuir*_P>)8jWVo!``Rn&eoU_N#NBKnr< z2InJCDOzRX0g@|xGo#7zw!W>SO;tFu<1tgpu+^iJujhAXek%x$-uy$9Yj&YoC7}kk06sdC0FMfYs@{H*mYySPf^- zu0NrsGqguenl4`z+k|x>qn)v z#>X@ac##t1i?4c{Aolet34QA}D}6X-C162ydv$P&?P^qTvLtnY)ZeV~P_@k^PA1d8 zl2z1aoGbRp#=E!Gg!BU28SM6sjhh%+EYu zQL|XNoLnoeZst|XU0+NfP4q<+=)SrVNd(0LV)GJ%zJ{^!=zpv7q2P%B6~Rz$wVP)= zWma{_NjLzfq0W&N`u5e2ELq{&`K|BOdN)rLxe+~yWlka{ij1U!r`t5NT&Z&c2<#ycG*ko#D z*-jp?{gJR%%sVrx?98LPRcU)adK}bZQ&bZn_H=Dw7PYu(*#KiUZlE?DP`9H=`{$U^ zy53^4FbcKp8)bfXuw-Luc+Rt(mJwK4SF-mL3Jy<*?*oL@WF`%rR0k98mkh{Gy=h^B zDcj0WCG+D$@9+gr_=0LZ57Wb)=gh0y8X16rr?`e+7rskG1A^imZdP|&+J#ulflt>a zCby7k_oElw%4wcEn4B@6&LVO#WVuZ)L!_L0SB4ehT7>bJr1n0ug_ktj zK?UUJ+Cd})Ck_Vh}u&>vmAhz{y;{Z ztQ1~bpPBc757=j+F7(2d?E2IKLl05E0pjoX2}jgX`J4V9{1C(NFxPD`?d&kvBS;d- zXzwVB?s{of`#T@uX5y<-+c60m=}_ujz6mij%s#f%xAu07I*#NT%3#6Zikq$XGfOwX z8;SsuI6I_4^k`q5C7W@D6f-sBwG1b&_1>w?3KgI}fjxlqY-Q+JY;G>l2pI2Qi#Z9{ z-Va-F0tbnpO6MZx>h*C{Cc2(SOWlW!Qkd#nOW5ohNQAEuNXq9a#^ zGI}|m2R3DP4SjtMJsnWUw;}H2AXE(YcoZ5-;0H-3XbhDb_qsY6>CPspbR)VXI{fTz zL&#_7h9zei2wl3%64RM|pNgIYtk+#Z03cpI2%UQ<`;y@N&@SX`-45SQzBWq2#z#Q? zuoDxU*141s0w7QvS<99?aV0_whRR4V+>zD&P|gwa$d734e(2y`Z86zQVhNJ5Z33O) zJ$pZzHGU#J%{4|+#BeC;(baRR@wzN3qV2?KL&Jv6{MfqWX{ocj7O*@ElhLJh41)!X zd=Ak4Zt2eFF$)Y!93T_ETrhU>`Kx}uAMz$}P5$@Jtx=nG2Y2VMUFDU(xS!RtGZ0h4 zI+~Pr8v{Vr`r4vu@|`kYw%9LietbKt%Xd&*bD;+nm)9KvCrBjdreQ@5(#zup{_Kaw z-5vKnPv&aD8N{`7w1PtLN+&GvZJR$M_C@1fb>10F{4yeB?n^1B*0W$(M2&L6y@h%~ zdyMg7JufBhCKB&079>56tk{BkG$|Xz#3Psb_3F~%J(_eMZ*nO`JN)hiYX9?jW4kgH ztrw>kryHYU7$zBD;R~AD=Rc0G6ad7%N*zHe)Fu0vdXO7OnAofMDEzVQ=qiqvIq0qs z`gnsgTyA6#$baInD@B;KOQua=2X0Fm8v7VMQnF}i?^qsGwT>k?1I;eOVo(z9+0xKp zud(EHx0-4s0lR`-B=8@4vU)kQN*6^3R4yC3pr|eNIm#tiA*N7mj6)gs1X5M%{k!0A z0-K;+Dp{|r^P(E%m{xqK{-TODY}7WAz#l(&Rgir1t|xa z6d|y9wkpE$Ac@@S+SXO{)2lP^r(O*&>xQCS-cs>+4@0E^33t%qNhU|t?e|2b!&TNV zS$4pYKEw#Zj>Jf*EYT&eaUe(8zHH(YFA1hLZa6}pv{63HU9Ghgz|JH* zxw?G7mk}{!2T*;EV7X}FTm_LdnNXse(>rwtZmT1q=H$D{toydD#2j0oOzH{+^KbA6 zggF4j-A+rA2QovC+WoF@cSH3T0||gu+DN8VO@j6^WxY?PN`+P$`>`+D+@3yTU*4^F zVc|yTN3f<$YP$ptBN;k-t&|c^(3VXQAPCdQEuQ-oAX`#I)1JzEmm}U!Pbr0${PAF5 z6v2fr-5^#rl$=Pcif`k@U!#bz=VaWw_P#4YF546@aR}Qh?FcI?r1!oZ%#=p??g=N+ zeBmkEl>4fb;&5)HqRRc}=OB9G@~QX-cQ2%_aING(>Q0pSa~9+RDsat)L|Jz&k>mmW zDwmE`O6~n1hBZTypl@CA4?7$14k6HIS3=JhJT_`c^jF194bZ?@Vkq&oEMZS5D(!pf zx&+uW2;4>gh6q}<&Q7vcM1HMJ5R^d}XR4yL1}6$ZSaX~d5VUo!2@*i`2|Co)oJeqD zouV{XcK3tG*Ln%AX+o{}39W5HJh4(l5ULYViq5J>S0*GlsZI!|d%@5v^pzATgfK6wru=2^VB3;;KjmP)rx?H-MvE4GMF6~(v|=Qt^7LT`}3s|Sh)246ZV z;K3S7jHM@Yag$ZkYXZ~UkU-8Lour$^xM9s>(ljk^a8|F=s^*D5ntJ`b;={Xlz8Z!5 zY7!ULbSzA?%g?ZDyIcIrO6LA2JCY9DjWLcTL8WH9ZL#INN`S3iauzcHoU-P1;rwQ(1 zcXK+hds?z*YWpsQ_iRG*Vz1f$EXl)7YVX0EPerD)Pc!G<=+f!d?iIT4G^d;xQ))jb zmCp@Lo}|g%xSR|q*t=$$;J-xcPN9EOIa=QKy4f>i*!t)w5Y$V&S3uASnxN*ZM@(Xm zXKdXAbZO;i6mNHX92QYkPPO^#S4!|tR#SHvpCOTLm+#-*--DYJvHAt6Sr*H~qGmxE z_BQ8MrB9DGa=tBvFLtH2{iWVnm+McW%o%t2sL$B$W@sE?UY}JUF8>}~MA!~npVDze zc74?kF#Gy8$R_%Gq)}4q8lipzRox*q>F1oih?Xa3%7x{pvBQz9rQ9}_CAK}z?cZX~ zUN%cyY?QH2DO0k&wI%~y%)I;CA#Y({xN0@I$87%LzE13>e@|Xk-Tw9~EYrQ2#ux?h zv&))iNrY!$?u#FK@?vz!nA@dE~rO*Sol#i61TpGPY{st~C-{LOee*y~!!i8);& z%_bm=vfjRd+9&56Wo?o!T*|1t^sYqga^Zbow_pR~VuVTZQiMs8OKR%lO9N6ock*Q@ zC5W}j{v^-pEg9L21`EOu1`U(Qpkx%sWNJrX&m0xzwp|V5`iPvIqhrYEw-m)FM876&g8?Wu#uxogM@%e6hYKg_0=U#ViCq|F9 z6YP=&W(viZ3QRLG{3fUv=VVVG#h3B*UB&GW9WsH;*z+tZCmVYBT;Kg5cI%NKE80Z8 za-dL;gdufhd)jh04A6o`2>c5UOBjl?)!45|m;citP}C|MfN8?j`d2kX`wa_}}}88-8SdvNeyaZAba%j%%hIQJW?7x%-ks z`QI6z+^^C7>0;fe!H%#-by=v^6a;h)rHhJ~UK0 zS~pu>K}kv>+(V^E*S4C2r~T_&)79eeA;{!==KN_gtJyo2$ncj3S{-7k^$#E~aCmWt z!!7lq4T5nuVn%y2fEw6i zEl_36RPAK147{B69oX$YBg_6BPt7;Q4?iXhs(9>{ZA~`g}&;Bm`5APeWb9 zYn(C;LUsu9e_ZKWN{w<~Onxm>B!r_wxLyc>+!O|&$5EWeyc;QF<)6;o3K^1_c4HVeYmG$I_*#eGGd zGe&Q`;B_`@;BuIF%PYE}&KijP?awZ8zr>5?qXEe?ogzeQmjqgZ)cfU6Qtux}1M`pFn=4z-M~2wqc{RTef7I;W4t1G#T;0mpUEG zdofIY$(|`9a0qI|(KX#L8Szr)kBG$OeH6cp_s3EVF zw?|KiXrUj^qsa~dkhje#eLsE}VbsDhGf;Jq7r%CVxpqv*7Ld$rRn0i*YV7J4?n_@0 z;f-lws%$Y~{^m{G*)9L>i}cTQ+Rx>Z($NF;OqMIr(0f->ime0R`JWAQKPv3}C-NJU z<-k02Ga)IqQ1c)*?7CgSuox)GLy@EXDFt)`L5xt7o#`7|Mj;D`T54}1A&ZwBd?ufk;DbJ-JI-n4I z|I!a>dSoKrle-}9jUICE5%=OAevlqsSkxkHP>~t_1dJ7F+2*(cE^)r?l+>_ew+8`; zP99bBdsm~F0-ose2z|!uWU9xU z%=SNdVW0ToMTQ9zD*EDm^>S)_p}yPSqa97FBJ9*dG|v|)n~)%tlc|cs1&du)OL9eO z0-f0N^UeduSG6f8E#KSox8IXVRuNldRPl1$kn#+x+K!dwT~n$tjl$S$^}uj%X_{uo zH2q59rrvIp@nzh`QeljY|J|!uC>V93<3C|LUos-cRUH6gnEcpWXLa<%!S`EfC7%>uDy;Lm`ove{YnuP*JBXO!Jo<|LOu8UFlDtm-V5 zW0dd-@fJ2D2k}`;wc~U_IEHj%_Im$(JR>JDO}XYO>uDx)&@V|MF(#yneCAKwmGn3|BreQno5R{P~XdP1vZ7 z_(k^)uaXG$U@-FZL85SG=I33XvLt^e}2(-?thW7gD~7)w1s5icyK|@ zSX>nvOIzyKYAg_J6^ViSv$;$q6(tK;0f>*W+AGaVpMIH5w3xF(Q_TTRM$IkeW&eEO zS_n+~uhuKbtBVGK& zC*j+Eu0DR=86pdlzgy(BG(I!y>lb9H$GmxVMdD^2=88E-fn*tX4$cM%o}#W_ zPgsuBIaQB=>S4Cc(d%-wmF)4sfB#ax&1DU0j*OG+R^K(_zp|R!n7+`vYV!Iv$N>nT-;r2j0~)0TGF@F)2Cib16v7? zs;fs|8swdEfPk(b1gRd8mkcteA9#{cd@{`XvSEWX%gNyYXB3?62GG$-E=X z*^KlzFFEqaSE+Nvbe!I;5tCl62ScR+_vo&zntTQK@y2^5hS>EMqRQcQN8s1K$_Z52 z+U+la*xizRuHL7!gE{SH#=FjbHG==QCq-D`C5av4h$0-dt(~!5ulS} z-LKc1Sr&nokNr)_+KjY_EF14tNPX-vuk{JR$l4EGk(a|`N$1E;gCt5sEvA0|yqvdN zZwXK7eoz9T>a6V@cbM0`gB01ei*{^Zy%5AR304}s8&E*9J9YT-GZ%FaH@=(jSdH6M zcVBrqGyxq_4Hj$fym)ig`ow9(-p!fUR(q){TC`|nS5m>Ky8erhVXdty)38Ws4W#A> zFIPUS28jm|{9orPOp9V^MFrBn4A!dVykz?Yjsvfht&$vUyLB+oF9 z=aL=%9Aw-*aicV7(p|U``xZ;UH>^rDz`mM48t&x6O#GMegHl+^d#y$fZz@D!N-p61)UF^2Onj^(ECs}B9j(2J{L=}z;~CKx>zFj z+Oyy(qCC?M{y1}TmcRPaJX`n-iXO~ApaJW<^{V@>5uUrO4h#pOmze!5Mf-jS%Um&* zmbQaXT&}o$rH?ksCo?oPwlj0q#Zr02%}21cR<}F*L>p0lK!$U->P1g%gGeS>pEMcs zF^kS^2l2zw7MfEoIUD+M<>#{i1dduA$%6s?8AoH_C0hTRiHp-Z- zAhHo1l~~{e>zfOBP8Ym=sDWDXB~PD>d6`I4x3D=|gTyW>Y9MjUaPYJWjQY-|te4y2 zPL)Uyw4eVbDZ6_kiStnx{gUc^BWscI{`?5`MgP- zD6KG_pxQ+I;nc>sm+-0`am=RIC9K1$o%gB^N(z#oroiZ+%H@7Ga54b@a zw$w6Rho(4?+Ysc)pDc1mZ#(Ez5r9=b`Rr?*o9eYzHZqoJ5nA@k zqpQ9}1ovV2^e`&J>*%s?$HB2{XgCRUUFdi#{=FsML}bc2SZm2)+k}rbg?Fm)%=-l# zr;%bUDdz%y9vD>ec}XzeBt78E_StqYXPF>Ih^;7IcySH@AIEHd@^rhlz9h&}RN`CK z47_r5v^a~ovF!&NUpD%KobJME*Civ1&e)0MVbik5qQPnyABA>7zW&K;NijOeMv;8W zSt&5P0J6rf9tlUs*a#mIkE9H;D7bZ8_IrI>4UXEK@d{dbwbbcFC=XWV3mZ1aWqIPs z*RlA^VUGDRIYjW==VM8$!I#_TNvZRHI{f2Da~eXq1a54PW=W4!C{J1TByzaXcy^R& z-QkL9OKi=xZkKXqoLz*R=yv}z7M0(aO@1${oFie)Ts8`5;3}djXDIlzKJ}HHJr)?3 zB)no~`pq{{^4?N*|pJ_Ux;tiOFoUZmq-zGSdu?{v^G48hEntt=%g=M3ba z|4s&zHe_b0P*D~WlSpfk5xK9+8r-E+W^aUk~VynLl9O_6iWvG*p8DwTX3qaXEy&2^%Eh z>b5p)t6;v!K=AK;>PoLA|J&iCwzgN&TpY7ew>QwYwxeMK9luo z@cneXb@jubBLC&gL6owTF*ERdY4xg)qLqxd=*twZrs#c{yqt2AD+d1l^+O5dVwxSr zhrDqxp7OjnFz3!$b+bE0yXt)h{Y9MY3&7+DO@;z(Y%KDjhMj?b&2uY=>i(&l1xuxF zi|ft@R=jW*O6LvNTdEulWMiEbhkhd=_shR-Thr?9TcUDrr8SB^V6-!b8dtSqxz4Az z6Eu$I)PorXV4e=N8R<&&jQO~~R}3?g>8~j4J~Ric*D_JW{frgWu?`iRoHUITJ5vV4 zIJ4lrG&K9a=D)41i*RUpb%cK(N@wBjohGW+!6Y(OChM2qw*CIAx3acm^%*;e$v%gE zqdrM`bwii)C2zA8m`%x=nT1)!zed_JGmAB=SO=V;q%_rr;YOTBmTkTL6A-GX4t$kA zod2X7>`2VQn9BKX9aiLEYIRwgJ1rleqLMtL1SGnw2mxuS#TQQ>{Soi!)40$NSaNv#p!fthOcir1Xj( zjI^em^=!sL#?MiriAfTvY5%CKbtXUtv8!XhFpl%H#0cc_By=-y%&A;!QqH(_}8GxHt>8A_n zONP+3xNNPA&Z+c`Q?2sxVk0v(lB~75w+A?klCrsjJ;~9}3qy6mnn)+Rg|g$Ef9E3$xPe)iQO(15Y0BJz9Y+xR@@ zSV$*$`!X}*tS&Z*N_MxA(v4;RGj5Ag_w%EV65f25WP@{Y*Onh@@rEQ_h z*UD=2+Z8ARC1OP%*$< z;IB0-1j~&C?0!2tqt7>`-yVl0TdS?nb)@p>ImV=vpLFrUi zqraN7;Ywa@Qn?iZJ8Ff_``oSl7Zy^|&wS+(6^4|sTG3C*0;1b84NyKFvVbw}IF+X$ zEoJJ(jcpSR72%<=w}U^eucv;uu`mTUMysG=de~$@$7tPl-2j|$%m)0;MKAi=MYqwq z`J$sOiDh(B>tTm+u^-kQI#w4K3{=sNt1842H&{z{7X{81F1eis2&lQmM$OK7m1h~q z_3!?bQ1F6Y{e7I-l7!PER% z`(p9$Gd!uD%I+UM^}9sCUiEvQJeMBB58ULeGc7wh_5?L!R_SP~if;QF4jTjJ(ZK7N zCJ1G$y5w{yBz|T|-$FUvcN2V&T453pm(u|&c~f78e_(H7g^jb0PD;(*OEq6T;MzXY zXABVLXD@(c4tnW1?b6CSjrT*{U!#v~uVMO8!H;-0*73~@cn#!Q->>ba*HVVCARK&& z9d5}F!?FZ0h;NG86=!olWA6r`cg>xUy+cZR>YM-ktKVK~KUrSN>0G9Sn!E@$g=@Uq zG8xkI{qeac@Z`qlqrd!pQTBg<0ju1{P`wYzUayeOyZML5RGMan2ZK0xk zhB(5a>q1MR7idxbKq~=ogyVJB<^4J-fj~CLbg!p#C4$Z1576lL8t)^(Eg;=wQ{tv2 zU27LhIRsYE3>mFNkWrjEV3)h-&KAlxyuwJFrF8)L&3K+wHxc71Sni&=-}e#NlT{s_ z;D9);EZFWm5`#9jg_ca4C~`%CQ(rFu?hJ+CEG_#u-*!;hLJj`h4Cs2n{<045Dk;TR zzdNuWW&qy{?X0Q;Q+Hip=mx!H65(LaGi_|uO4VEcjr3$01}0G(|4TYSyq6?M5L^4P zX_aJ-nX=uDvmL~ZUYb+G4b{~ePiiPylPM+>)IIF1kCi69&P-;y^3OZcR}QkWmkFWP z1iaogthlMPWhYjo_7o-zn&rKtT^mY};>8uJL(=fzZg&1_=++CcH+x%RNQw-Y^#@eC z?65g4@Lp#u!7DHJkgBoEhx2-CJ1kAvovehi+H3YC$3#ucYivX)(lsA~wBrl86Vi~t zNDBn-2@iGK9Rk~pVN?j&>b+cByuPyJBg34sn}U>V`aPZ?#fJcqxdUpj>|5&fDD%XCaEb zav~~|T*u_Hg1gY(a}!6ZwpZZnDnFHA=xqJ@wTH&$eQ_AbeXjo~l;8?n5pTDj2r8N= zfuhf#!#^$A8zU;#0|82lmQWX376026-xxLAs2RQ;eKZ`=C<8`_e!8BLb2Hyu4Jzv> zt$}Crbdq6&D}pD-x=vda`l#(({^{C+9CrcQxvS%Pnjoz&t`1j4L3&833%-LeHRe-X z^mkF%a+e*uFW(%`8ti8-Or)rRLkTjK+MwABO|;H)4u+mra*S8i*o|{pI^y#{wM-#k zjdLH+XI~34w?Lo0Ywo?8(_`x* zm!cSzHzk7d@vSHKCP2QNyPgm=d)#j5Tf2XA7s$DPF5N4Xziq1@ZJk;tITd{xDJ89= zs2{1bjJ%=i>vL2h4fYGO%Wlv!Rd{f|p78(^o}!8?NPr9{l3|~XOi}w|IG~<-Y}ah0 zJ9($wve0SN0l}QWFE$U$qVmzUhg;9yN1zf=^0V)kx)CHeM1%cy-Rb%f?|l?H&AUaEC5|ybrfZ<>U3>v$x*<{k?^ zMCKi#m!m55rz1bUgV1VTf6ly~)RJlO1F|;eXuNXVyl}hAyb5!J+X(96k%cE-Jbgtg z7yFr5za4phATBgVk1L)b$9b%08x6R{Gs6eRJUr3Yf~k5j)m;JVt=!ff&LW8u_<4&y z3A=T%tl(5lNY$TZ%nDMGPBt;fp9(LSg}OF>9E4LhiAn`Qlp8^mvU?t6%_^Xu0m6)d znLjC_3v4eO7cFCg6kkd&xu!$+-K_%X)HBy%u^&gL%Nms;Djy@hvoyPotUEmun7a=8 zVrW!p+P=t=T$ZWagHjvt&}w60%o#G~RXW~*Zx;ib?IHc2tQkRpo&<&LQB(NV8svgZ zS!3pPnrJ=s-|SFaew%^vyb32|riYcRG+FK+J3Imt>AqskpCdXrPp@oE(^G!K*)eA; z@Ice^N1@xttnS8#qL{Q8x1pGW-By8T!8g)6FnOG|(r>-kt3KSl98tkRSk3HE)kn>w z5(wRt+q^LOuHc2(rNErYe1*M5#?6=Ud$yZ%$a?5nYYf5zR8R~gGkf15?&-q4qBRtR zV%}V~XUXcQG?ZR_V=YES1$m@g>}-+P;i@mK``TKDvH9lV8KnMCR%3^nrz`gg2?kGs zv5O6kuJKU0v-NZTW?b-Tyz!*Y2a8F+#`No@(^a6?T#N~MjNvO3tHr~ozAyh+v znOUE5D_=R7jah9|z2!BSjR?e++YvOP4mqeVB^mdlx~3|2bz@FmF^msE$nXa%l0{Sy zq|t3EkTN<$LoG>ChZR3Q4 zCNSjzL$f2CA%v{LR(*84<0{W0;E?g9&yny-X1SVb6qdFLz@ zP-yf)iGw+C|I)u%tR}tI;Z5HF?vs_=a5>olN{b6|zMJT+vb`zjeK8=FA8$Jw4TnO+ zSd_NU_Vad%%#)o&m0U5IXOpO!vs#}`<5bvErQ`~ZFDR%_mCtx3*=99)vmZ{5WNs2M z?1WrtN#4-19JJDX6{%F`*0*kod)Eg`uS;NWvDV|~EtFs3g+!j1c;{SydeTkeP3P}= zBuCf)PN6aLWgSI2HHo~OTSzc@55@Rsn>Z^CPr4ih{Ms|wZ86BHIe2Y*ke4$pwm+^m ztYFjk}wyCV%zRpK?%qNfia6>)R+abyH_XrD%cDH z3pgRW$bJ#*0bvx`%`x%)fs7YgzzW%qR=vOqYVFY&=pt7?d^2hW36}N(58*@4nbSpw zfpW}xZhmeOPu+vA6?d%fp%aVNavF-p{%9W0;o0k9caj%GoZpJGnwjgTtbh^3LZYnjo$W`CbDLl#T^`>I*rj~en zw(pY9_+y*;782cOC~FDSAnVSmh2Y*^%Z^#UOsqu6)2fKZPmTt56W=$4>&J|Z zW|-0{N{!|oXHLGGtHJslCmoo*OW5}@qN3MOdpDm<_3frfcfJ6wi_UwivCC^kbeH_G ztL!+UuA79OgiclY`oqwFr0AyqON#Ex_8%$wkgoP^ieA6E2R@R1g}=L|Kq9LqDW&H& znpZ+^cpP!<6jB)Uoh-rrzo+OmC*K=i^s;pBb7krF>cEZ)WdhWuYVSPqErv|!jitV= ztmNNrV2;WWj+DJlgad7c9>Aw8WnZ%mClZ@Klgo(6acNUH{2%PS2UJtrw)jm`EEE9+ zkq!b%6+wCt0a3alH55hZz1L7gDS{N~y$VVR(tALu(n3dS=%I&R0wMpMc+R=+o_D|Z z-WcEbe`DM?zA?lkd+j;b%9?A=y|VXj%~i;x?>j?QOY-Rc)4fmR(kF}5lbJe??a~TT zr+z(BgG~MUp(aqv;Mh8PzxFP$y0D^rUhUF)vv$E}vJC_Uzv$?YoMf{S(~ooSP-O&L zBHuG%kN7UUwAS2+YJnuo?M6j}6Z$;tPGOV*n`m%yTWFtGtgAl9V0`~~b%r@R?@s>L z2!hzsJ8CZb`KA(|sCivcY|_J}iu}n(0>sbg9;Uy(i9fXMZzFk)==Zbqn;X-_{mPk4 zcu0Y3v}qk7VdB}*dHfC%^3D(NU04^S7(0W5mv>qWw7C$o$z^U67ZT|zeWW84n>$I? z*3xBZp|0z5#SJf|nr-?11r?(K3B3Zo{ zTWRwpKOoPTs9_Da0}HpjHc9+iK^dK+c|mVqQtwM;EFWzSqfh9muX={BIp^8lIZ_;I zebXJ^K)!plWz)1fkjUe6Ka?p$^TbFkviHYH(Jv59mR_y-Xs7|>tZpQ#+%6E&SH|5X7Q}*9A)IhG7_(Tp@ynU1 zP13`M^Ihg=MnWnA+b+c)-Gi`=W9hi^nw`=U;BGS1y)xg-`T**wGOowwb=mB8GwJWGdDr+PE z#upyqxbb$m6S^0NUs|NPg3ylQbfE?hRPKyXPtDJHb9*c9ti~OKWPP(oCvx2rRRZQrCaMsJbV@cCc*{wl-I-E7hapcldD3D3|fP zpzIQw_vok-6jB*vdd+55O%%?YS^(1g+*Tt?g; zSAW|9{J(D_ohsJen`ZqcLVl#clNyFPYEN%(?Tzwnb#Wlor2d z?^UFV{UmQk+_sxG)gjeU{Vn~NWC|5*|kV9Ok|UejC1NCmC5~*+nIR64djC|z!9$cMO27dpDDd#BF_X}z8;$xc+OeQ|QgZ%I=a@bb~w8vTCd)p|2 z&FyXlmt0!LM`KFgkLnmjzFIVzS9$;ALAwu^dSq3>{H9#Kc?Y5SMg)r%l;LFYM+AP~ zlC|PHesqexq17!3t=hVY&)MKHXN8FgpVKcIp0Q5VJs`M*FTpe;14zq}sbTpm=!EU_ zim-*!KR;;qO|OoFj#rS#t5dE_ZhHAAm1puZj$NRo40@u^|E?RHY3N%bm+rs0~sqh4vTVkeChCq zC0{lW`TaAKmkz?9clQe+TEg>nG;6*h!*hI5Y^WEGs?c{OY=Rh9Fx^snK~YJL>uc>% z_2)JlT_`zfJCgFvsZP?yM576-Qcb#lY9}vAt z2K?BHJ00)8L3CWosSqJD@Qsha+)SDe-%vpjYQ&d>)MT>m$lo!od27L>-cuJzwzUVP}Q^%_U@`WPASd$zb_MDk-1M3 za?q9&AHg+rRHT*1KLmcPH7@WXflx+n98%N>73U>Z{y8+WY94v+9fyQ|q9yZZE4$JJt7H@g+_4`MoX1`)qiH*PyuL{U zJISmsB@q{!>?J%BfC!wAqe`h1CwL6iWa`P%c1Y285Z#6ekh8={=~_6x^N>25u}i`$ z^ojYlXGz&+j%~J^lW?$omA0Wnf>o8IjXNB9BbqZUU%1Qp*=L#MLe#TlnE=!eF%^R2 zuaR8Owj2X(Kbk-8TPf#qs2IMc3jZ;k`}arfe2o^}Z4n=Pr?qN-P5BKBuyGvSldhf# zo6mu^w#i2*vKX=x3~%Kgn(@F`hZwBKTCPYjXoneElvFOJv|amMNW>0$aD{rUFdIEp8 z*qOx!l7bz2r!(wI)^md0lB7EV`#PP0jj)O3bVwuy;RAEEV64SbMsam(t{nZM{np)VHGfJw`*YaE`*^kgA zsPH&-rqF7!!3&rO;jKL*i?B#euwH)n-uiQ@lf%1Dfoyjems1l0*<$JGW&P;$>FHUB z6?x<}$m3Sec0x?}&+-utgVM6cc4NViXg7L>AGR5!3mdLcq7{S4F8faMs&B zDlyR?4e^0&i?bm@jr8<$-h9N071_PE8VYL<$s%e#*#;xV4k`rr$N?qH^#q$5WD(zv zoE(Tw;%sxU!4_x;AgZi{tF#P!859#2$hO>Q#73+L)C9B?*RYi`+vk72ugGRxoV_F_ zO0ek)=oIb%G!Q^pfVV#s^?|La(E`SC^z>Q4DOUFFyMT_8_CU7dqqH!ldm2pJG4|i^ zjV115A=@%E+$C@LPQOx%2;Qp&Gr|{rczP#S+?$_4Xhy8}TUX>sjev(V#S{5Rn$S>J zx%p`@P4R1_}2`c%a zM9X=7GxU=(uu{#o3-Ogyb9>#e4lYY2MlQxeZIVoG!jDYZP;2lBR2Ve2&HkY zk14<*7p)X?$g-I(+0Gjgpz%B+g@sUJ*K}y`b3@%k4o-qBARPYEk`2aJ{m&&@6>Pod zVO&E-%Bks}t*Z$mhegXsA*ZT_jABG%-Vuq`b<5A+{xCINcyC$Io<+Zp&)6xO@%3Z5~_ zh)7k+7pGbO_fVf2AYRxuR|@SBJ&mW5sKW6_@WUdugc=Y7~-|5(6uh$2&xHTs;U2ka`f*d)J7iP{MH z+=UeMHQGMSysyae{n_KyfyWywIX%yw`i~iuD{?L|PPax_%`1wTM%aB|EPl6SH6ls5 zmVX(2@8B+KM}I8xv6N62Q{K6}jL{xy=W6e2@K^qU$gw(Kp~bWTa+ja;Vi#^=oJnRZ z83wFZg?z(M_L9~Ksw2h(t2UCkPlK0?e{`uMo^R+>O-oL`X+jlMq*!N73oQqjy(@Lb;pY2dW*g*Z8lfw}wSFJh z6a&=+(H&#*GLQ>si&cyNvZVD;45ywRqEMS7j42&>D9ynjR31ncOdJV3Dfpv8W*Ys71co`J+Cc z!@L8)X3|nT@mkJ%))^0P$z7m}6rkQ6{#rhyS+3~-5_9bu^wvLe*V6KYr&5C3U9w@c z^>yDRT4|49N#+OAP$$N7t>Mc-*E*#|%3cazX1TQEe3yFX(qeiI)t0!EttKXlF9({wl z7Rkducz(9m(k7NyG^ow^pIF|7tUx z9vt=ZZxx^ZrTVB8dhn4QrQO2~E-Vvg7DpwogOG91kvytK-l=NFSKxK<-%82<_)43g zn5H-2_2_x9b@E@|80Z_&Gr!A6`bT5=4TQ zmX;K;Z#2|$H=Q)wwMCOOP3F!T5?j(StRB^WYGQ-+!9E|v8U$4pU@~$xt3I_=(l{6% zE;uz6Zv2GXV$`>SF~Q3}Rs)Vc06Pt~?Tt75r1m((|A5@f-!?+l> zH@b<&p6y}`rE6WR8Dw_J-pdS%oCSSP7`_&~RpKCdd5*}4df+hT`XKq4wyS<$;!IVH z`Rw|@l6@!M(vYxu_YNHtwO5hmedVX}xN(EQW3ug@6Y9gI9%s-NPh?zMEKRyF1eCYg zT)w&C$6KwY35#RjhFl%m=aOjOCy5Ln=1^Br=^f@rC_c8);;&qOYN{t7FbRtv`{@>q zNawH`w-1>?s}wvsl-jp&wb{L4Y2GLlQ^_Df3`*^&nIB8r^4>= zL6{@`$FhBsTSqB73MlAvuZQhC`bX`IS44s~be9JZ;XP~M1w(~554T@o%3L6oORQk+ zYlw!$AUAUgsXJ^me!&(tz4YkB?yBkGK5GTIEGk>cIDqWNq%&Nrovt_G+z~d-Z9e(H`h>s3E-`;|WrP!1(J*^l7c}}zZF2rV!4Q)SKe*sIHM^EN_Z{XOAkK1d zr2-PQ-J7m|*(NyPQyvyB< zq4E2-j4nY_?x~iP*`-t^O|(9R?J1b&cfcb1MkPxo%=^-cRlSXl6P_uBs4%!?Xr+#< zJ4hRRR@<|^Os*&!nH!mr5o`{2+z~z>Gh3E9c+hAw?x--FrDbx({8ZC4xjsSjfOka^{u_x{OKTI&%E;G zfkVzY(_?hwwqJT&MNA0E+ovSy1lop=!f2|q=~lV*k}db9ly?+gh#wqwG#n@B-BL(b z;L$dAZ&B>k7`N+;zdO@dV@Q$kYKenf{e2x*q}Rm;`2i=SrDJVE<+rIB4%+fgR>OAc zN9NvQjqc`mqV+n>9*6b*=r4S={XoF^-OH(EGcn(3%c`k974{$AjD8U(ifeb$o>D^3 z1ijt{>$Y3=YtCJ;!Q3+(CvGSX=`@ka1#Z}Jn-55QF1qO%C8=5_1U~a~7`6wdC^tQ1q>LYcD4(imVmmhIR2mcy4*_i+UA8$v-}5 zLya4p*!!3+G3|b8U1z^n|C(%e(ueCrNe7fzrE~|{J&xcuU>tNwUfy(e4J#`sSna2! zrPS!TX8mgZyVPz#jYHUwG|NgKQK~nW98XLB~*a11Jnjl3_Xb0XTV7a zm9%<^3_bV}3_uC_5aoLBL^M{phzswSQIR_vq%K>PD!kCHfxPGDkDcj@Fa{#> zurO7%6g#LB=MvlfeR<2#9s0LNjEd?U2!`u-S;K5{z!>dUyQatezK1-C6_WTxwU1= zj%t$2#;~J|voRpnOGXq$o3RIx%KJ_781B&XqxArJjIoyFll>-Zt9!;l@)+;o-aruc z{HDyGoC6r?-*GNM9;3YP7onCg2k7FaMDq_|3Sxm=eHh;<+y_cD(-I(s-m5!qP{vBj zh-AoP_#MhWt@wqW^H2ER#@5f$h*QJe#XErMqBXmX%GhaJ)6d4(p5U4XNXgf;b2cJd zgrZrpEQlmq--A*QJ@2JYx(}XybjEcztv=;?@sQFVrgKTSiI6uH+h0z+WX9XGm$8i1;vZ~|(^fsz;Qzz~vF0qkw9 zh*REQY2aNtv-K4D_2hb@z`KR|>6jT7=~FG&L2vGbDT)e`#awqLJIdaS6dfdqgMPnh z=y}2lhJ-ite9ZFCz7Y3Hcx8}{DW-g~@8>BMh3!atEjY^n<|1GXZ%ZYt34<-3?#m2m zJqD}-GT@{@^6Kqkvn0r*w^QW0~YUy9^OhIF2 z?2~YP*4Gd8d7Abg{h71yg0`J~bdoEe)hl6)5bIhb4JtUTbi!ImfMk4+j| z4(^oBL}Ld5Ah8^$i+n5&sObVqa_ZPhXLercn>6Z40PPDs2!bq4gdMo4do^7cA~GYH zJGMK@bq|60YNhSZbYZ=ZP;7NNI7NOBu{!3$d=%(pV&q zaR;0SH?Eb5zH`5=Oe;Hg#9Gr2>m7jEFohq8L(kn+Nte3;8%EZ7I^epHM(6}EOTPV8 zz5lPx){@9^|MvBd{=t6+)>AdFCTBHle@Vyz;nG$AkG_wH&@{c7ple@91cX5&a$FDu zjbO+Bcfg5^l5vJ$0UWZ|C2j!lK+j&5w^TwGdO$Xo)9R&&BEv}}>zFjstpZNP4U|NzMYMW#1Y3q4m~B46Hj}O{ z-kN`OQfd{grb6}Kx>rqD<>SzE&fQKCpW7%*S#WxZ4(U-DB^v8WxM9yrK~!>S)U6BE zfpfh$dCcj-!)#0m7b9Sy2M-eY0q>GpTmBA2ElRX;|6UhtTuLT|nyG)6i!I7!H)^j1={33{8z5&i80+WpMRJ8R6pXjivY% zza0^LnL)Pwi?BHT?->@|Une(*v(Upg)Az}C#!CufWo0%1~2pGbv^6Y6+hNUpBBYz48QlviPwk&KlcEJ;g4>CQ_60Ccxjto zAMFU3+LA!qeRI2)_)_c6$JKUDKNn&TKatMxqS^AG-C z;&9LYg~K)3{(D~GY!f-|y}s1pUeDPm=W|%NmV2b#Udv`d+xZkJJ#{--MxK0I5@i z>yGE|9|?9qm^Gd?!(Hx-E6}*fk(?JUrfSCmv(Pxfo+tY<3t=Ob@|ek8+*}9UKd7&hi71(B3_@?G7V1c+P<-cj9^dtEnOevv(|2w88>M#Fuqy)SCTRI&0`oFeY z;8^=pr2g}|eJv*fZ~gl=x;T8=Ui1Y|W!LFbhKA>e){Hl{dWU4!JQq_`%iL$N6HRrp zqk%H_Ou0j{&Kdl1iT*z)EbfN5BkbP9XI{+S#1JmVH%Ncr2e^932E#Vi4>uU18VP|J z%echsQ1$yF;BLiFtb(YVA=6`^lvoA?d#AuwmbzT)KqIkJH!0h7%}Lf|(za0Kg$8w}r;a0L$!$=(4!?oy8( zaWS+V_K!&m@CB3q4JR$IP?`6CPFmC?v6Gg*(BG4m?`N>kFYA#we5X+ddA3=^0O-6Sk_5yS84^-nzU;0oaYv>sB?8uPw~Eba;%TcNKNSze0Q%xcnQK4p%8vGip}J5ue=; z>UX6STyB(HS0DL2*}GaHhTOEf<^zX(s7_Kj)!d5A3;kKu0tS+@7%z4$=LqN>zbVsg zTZ8O2rW)mQ=O`0={P z%DeiLmj_0c(ZKh0SjoOq5FgmX-IHH-g2+V!{+j8ht7?;Vp!EolF??B%Xe^@LSu+6z z9m+j5YB08O;Op=RPV8eO1J77RY$>R8e5~8i`;mNI%f1K?zDA2rlKZEwVzD4$K4+rhxeH@ zI?^$hYJA|EuSI5q$ljKG*SvI36k0CzbUt3V5$F4{LSW=`a4Q zzWFAz+YhfRYbU38IKQa4!cY%nd1DynbopFM*TF|1 z0>(z7%y{QqNxc=jOUW#Zh--w5)gPirRpJ9(Jwz7luMJw9AL7lZGX9Bb`u>S(+7`AH zGdFQbde5|oEGZ_In zp84KB+jvs89(Y@{X?rY+XwdOQf;)z;x*>cZo z-+R)Lj_VOH%{6;$N~Cnp_mSqpM?Z(jQ-=0yB!$|KNM;gk#(g;CN4T4&q>fi*B5J-EOKrd&NLEc`q{IiTJS70v8_L~trrw7y=szNOodHD(oTRs%;u0bW zil8!)|FeWhdd7Bthc%(5&5HjiClYnn6L4027MXhh{XqVQQUZ1#hNP2{^~pOT+NNNJ zV!HA&-%lp5bw-)W99}2dA|BiQlDTe?l99U#8WHZ^zkF5Jz0_%}!%McW{xVYyhxRkK z-k)Rn%=Tfl_BZf$x!bG>>gaNM$DT(gxHcn|P8SlC+@B7Y~uoeXG8Vt{XP(mUPeG(t6(@hbdjQPADstYVZNwU_fk&N)wIQ z=F+2!==W)ct{N=dZsIQE`*_&y_8drY)Nt(76|Pn;cLS8se*Xc8I=6Hqo9mQ&@;sxZ zZi;Bk&|%v0<{gD*uwcuY#k+cbB_4m05tXTnQbL=ji zbI>;mJX5hJcqc916}1k|+8TACh1INV1;?e!ZiZh)g%p+{Vt% zz*c;@+2=WfR-V0np^9<->Gh8^9lD0VE_xp?$a?#_sxemVRY3Z0lJiC$JoM8@63%qh zIC}p_S|qTc?{tYPVT6&t#N2qzeQ`gHaZehAV!{!{6Zeo={VW@j^!nvb62#p_tKSZm zdb~;{I?B8k2={We_8YlwiqI`&6!S}ze(fDx+Am)priBkgIzgvg*SjK5R${y*sTPJh z6TF?AmjsU+bjA;dgr7->rL>5Ex*5qjgHMa@FFB4?c^W*FTo&eqDa4&8@~+~zGsQ+??$S}G|aS`!R#%E=(zi)_-Q^*lNBYNa?_XQFu%bCG z+7(C9ECePBNdrv;>m_-QU)mz%&q;g`p%s!tbzteVPh!Wqo|?v(mMSLp0DKX4@m7YU zaZ394Bjt|A9v}KjDMVB|c2*}=s01DU8Q=ulGrZO`f84=I$bOrCfTezSyte&X2HB>I zwF=u=ePGRmuaz@UEJeB z5wa$L%(HU_$UM;FjR730N;k)8y_(XHiUWbnA${?6rb!2Ttd*QWZh3o!-h7;bz% z&Jv!)>|+g6bW-_bH`*q}B=T$?upl#bB9ZPcCsc}PLWGLzdw+jx`xl> zYsN)?I0ExA7)ol$^9Jp|_n$JEb1_|Pue!R_)k|8hSqp%IE^Ui$trVc{#k@o)M}J@c z3$AY?^OUOaM=`*vvmXPS^267aRhh3cWpxvX7CooUJke=avC@I)QT>tfa-c{~FW=O} z?i`@jz1dHWP@>+(%_b5II;8Cs7f4TgIz~mVPd@l4_iCPX(K3E>Fi%K zy2Z0f4)u!LdFKPRg(PV{)yPlj8W+VRiayoC7@!(y5PVsTn} z8~7(GmuR#O9^L8tu=41cfU7p_o@+V)i`CYhLI7CIuJ)#Qu^rWgl`Vrybv~~Ig}Es2 zj6+C}Dv=P>o@15vi^`o>cv2PXYMF9cI^$Vc19~g&I$FWeibHoaT_2FR_TMb9ws_6* zapUb|NP)Fyj*|LYU#~Y6!i40#2H!6|!7sg%y8dNwk3pc~njAehGWhhU1syzlE*FJy z>-3Xo2)^g+Tnsn%_m&jg1s3J$)%acBGA8ihCxj-xffbRiXBx&0FvTeD zBdo@ld@nIdl&P8V%*AuAUsf8WiY?;YW;7h7x+znZu3=;c9N7oyscV#_Gp}FX(^J0} zfBL8=n$#?5!C=#wzNOkusv(Z12$7JRZcKVGHLFR+WDkJHOL$w>d8z8(oVr+MmB{eP zJ+GGu`iK!3{A8J3a%6PWxNjIG*X%g@YvZvgPZ&MH%mj8Vk<+CbuLr;f(G3vUzP3{) zF8}2a-FJH};JwfKKi1pu5qot9%y^K(R2Ls2w^QZG_Rs4+5jO7pUdh5v?A38>l+!hm zM!&*K>{VEp`v;!95B(2#a=*lHJbAzB$SG6gFFcvBap@sa`Qhm&K;P0s`))b_P$q0d zU;$-polcq6IHl`YKsmGeqJd64PQq+mmTcd*l3~SAN)T8F5Jyx$^mu-;klvPl^&5fzNvCh|8+EuIH(OOliu|i)qwT%av7!oPQ)n!?R2iYH0*XVptqg5;M(j_>{IwVV}DZt5c?#elDsuewSZj ziH7w9emAqkl4-sV(Hq<;*u_zuY9}6^z{ibP7=VDQo&1sK6L>d!ChyjjGFGN(gdtP}d~k_yjSD=sH=VvMUWDSffGF6dhxUu3qK zvhnYtW#%@XBNOd9$oBe}4uqud+!;;9T68gJR)NwjhBv$fn|NoRVIi=UGKyCVpdFW~ ztUVXf$bur6LYV!N@u~&_2&&d@XrevqG{^}w(q-aX_3A(9Du=2zcsp30gQ!l>YR3$hBNx{-ow34ZI_^^m6P7EQ++nrs1uyid)l-Fm@jvbutJk zqFv*7TM)GVPkEM=p=4VzAfB0T;Zwh6z8KWjI)-xT*zvS~OZ;1=`Dg^#nqZD%U$NUw_jyEk3k3(B)c#MN0hL zKURdMx!vWRjIOX)trPttzjB7Ty=IDpy&py1>I=OlK9t)74T)Q>*U%3wsE#Osw3yx% zSp%?unz~EOsulHLVItTyq5E^dCd_#Ia27tRNh&|>q(l9w8pDuS(tZOS#Yk9e*Aw_J zXSBGLSwC@QOXliyyQRSZf?4AGl%Drsp z7b?lgWo_#CCM(Ft&C1ND$_R0#*!*BPwP3WF)z7IXA z-XX8aTv*taypKCOAD3YPZ|0u8{!@SUVB-12-+=dm2b+S~jVY&>#K;%E89RG#;oPxe z%*13hPstlz5zoIN>&I_cPL|T(JbK9F=1(p&%juk{zTeEiqf=<+rXi!c68u&DH^{w~ zg{8eYQ9H?Fk;WEkGAiIu--gkH92R7_Z)VJOb+YD=)Ysc;>mf$dc@$d1IrDm|HU;OD zPl5$khz4JsBz)T)8YexCvlW2}kC#YuZ!YUwh>&^;y-sN4!C~FXs)tPcp;(yPG3zP1 zh8cJ8s_I(DRRHUL=j&7T6n$K@(ovboL<&tQRWJMl_O{8xH0O*c$n>DCwPd7+bA17d zWFZ;Ig(mWq2}y&lq2AAtt)@L)a09q+S(#fd$*VF@4wNJBiXx%Cpre3$h;PI4LCaN_ zY04bRlPx02%a1Z<{3UMOV}*KDGH&7c=OQQ=;+(nO|D-a;iXaeYH4Zd z$!TvVyS{41&kWTy6&6@6ytzAIK7jByoStp$jatbNgsp*nT#5b;cgy|*?oL_x4{*14 z?w^!3@+Elv$=^Wf>c9AHCKK?w#NQe0JAX0QRO=unu74)OQIG#(u-5_jJC72~y&MPC z(qp*B0PZ~~K6?5m8Xj2&J#hUSWgYqC&r))r)Z%Z=n*593Cf0>}eft|Y?)z80Mx)h? zcmBCm;yi!0dHN@6wDF&+T>+IAe`$K>U3|K(Y1ESr6OKX{C!713PTus5yL&=6YL|yF zm=f<5QDsU`sh$D62v*g*IEkY5-IV)YPc|qlK3D?1?WdTYmZ7|Ux0b|K3nQaAY0DmF z%|BbW(Xr(FaVVw<_~ki8jv?=TKSYSUp|A`7fGGEj;oaU+NP&LBgE>=(zbjEb>Cyj4Z`Y*%SM;{g*Z(oS{bT#Trni+_{x8wn zOc3<{PjCNkqPJt?jN5&OgsL=exfmJRG+dq_P4cASv!=G;&gEGXOgAsomj&o8i4Y!l zO4KT)j%R&M88@?X-Sxv&8(LC+BDiW|o{nC%>ZFS8d@kZ7yL@N?p=fTMuCpnK=o5yX z^%?(#Zzp9V5G`73Lh$%)hwJ-KP{${p9-}lg_6vTtX6Ks5L*4NVI`Rzq?0NV)4SI$d zajwo4XRa}1ELxGNB^g(Db;dV;RwwNkn4Zvhcp+jBQ2ybd`NI#xV~nh98LPdH@hgMQ zljKQI(zr!2z!NxBX?ai^wZ}gVOZmJKl#s^_I^EK zeo3F7E$X$J9Jq@343ord2H7OdPlrw3aHQ3?+ueTgrEvF)d9(6$-lKHt1aG~o-J7J0QsuG60-GeBB9rVm9e({hr-nwW^C8&6*-PSICVJyw8d$Z>SYXf2jk?5Bj zF_?6Fu3ILD(<8(*^`g`bqvk{Mp*V2!v`|i>Btg zvI$bSxU1Jr`*k&oGJ4}M?K;}$Ec{O!j0nmlmhqsy`JYN$D&JcczMHR41w(%BYca~2;H383Ra%N0^DE#;6CDYgi{_dFq8uTw<8Y?y05|q5KobCPZ(#v$tSSKD1_JJyJhfSl#JSL5jp-l6Un^LD*X#AzOIz-kIrel60LqX1 zZC>O`qE?V&=xLP_vm<=z;33w+^;+QIRxq6b<!A+Pk$>!NFr4pK@}?zz}8 z8G4%M2~-uW2B;|mO1^A8nR53@1P0t_1g@D`*0qQ8&}rc7T*ZE!KgY{!&PBP@qe`Dg zz6Z!>1{ZcKP&{YaTNfVoyKNO%{5ad zDh0noy~SW!tq{CX6*Wy=Hkr--cJ{{i3AOQqOXIdI^}MkgRom+U8&x#`yERP>`8qJC z4V?NL8o^3np%L4;KcNx3tK+uOIA~;=`UXxO3yrve0B8hT1n2{x5xc9up%H9#tN=7p zg@s1ydEZnk0MH2b^+Ij}t|u!8{aL-!SD4tZ%@&l^w9Y!4ZBjp%o0Q zZ@~H@u)bkR1iQXr9{PKI!ylesbdC+!B{*jjvkN`#+r9ILvsTRjR}eh(lg6iiTpbeQ zo9uw=gFz$^-B%@W5|&0lRk?td7uf(M1rk6BQy%+PN+lcaso(B*4Qu*pwumckQszggj8Z%Dwj{{>ybkr% znLC||OLESpYCjRR8I_-2!1hvYIXgHx?t9+4Xd?FE3oMRI?x5qrMFvD;4Ot4^{TE8z z%h%5QE~Ue52K}H$3C4+6pQR*0!qB(psNnm(xEt{+&%;w`(P8meV}DW+H-CgJzMx^z1FYb6y;8{)hU>z9O?wo30pPJut>~FC2pxY>c4aPQVtzV3>WcyUo6${42`N|K zhV(bTG^cpzAi9*-tO@4pi7cs}xL+<wc{?d80wWtyM*x@NmUj_uYou7U$MpBe`J$n;lDm z3E|0eBMlvO7O!2SV-q^^zesX5z+n4(ZMWZi)4#0c7aO<56&I8>X=LTV&p{D*qQwzi z{BkWe_DN%`J@aF&*RloVHiQsM0p}8x3F=?;ws%x#BwSLHAyVMkpm|sHVBt4N7WAXo zlwF45vzAKK6#Z%(L@6w`OM$1Wy;wKaE zSC{UdT~#rJ%PCS6GmAmK@g+mGtkUs}-L4n<#z&Zlvd*oTzmKyq{egOF%>LqK&V%{V zB@rasaYNj#AHlo7@OHn*XBwIYmnsmg^p9Nr7XFYkz>^H*cWsNSn2x*NO!uTkER1vW zCD~1}+uX5}ag2dZ)}F6A2J$u>Ai;TB53Mh~@ke-D&<|&hnAS1vu0>oVHByo39ApDe z@9f_hDm{ogk|N4roIe!sIlT;ZZRTDcdVFSf=T!URWT&e{fAJ$RroeTp&L(QedghsR z?1SwwLL2`wyLUok_i4W`Kcizi6GWURiGN1e$E8$A9~?5jwB-Qzi-=nn z%gCbShGjqT?AmVBRUmDg6q{Dl%7dz$-$b*EIe`jHoSb(MKZ|ed>c_X(C+y9nqnBdx z-|lkKXrZ|EF~k02M?hlKc+rn(08z)K z%C3im9BP}Vt?poARU~}8T-~_yk;aW&O|3UiJ;LawoyNTZU^CJyZ4WvZua1!L2eiZU z=N}9s8`Nj+F?}37jz{!^=Ag(qq)x%Np!FkQ&oeE)COa6#C*^RAR)GikJ!5I1#u2ey z{ROCcrQ^@zp6$h2!OGPgDvB-zrG_f`P*;?z8VX_`5e|4Zc{BB z)d5+z$e%cjuK=zzlrxmWX*NWfs$j(*>8|P=_+f1AF`|()%&uyLG(PqxhX{JJ*9y6u z>SXt9XB4W{209hKVQqT8QxQHDl(gLY_9G)#O&*j zU;QYkax{B**hTa5D!1j@GHBj=FQa(Aa}B%h*!sFJw&zA8Q)?#GoX^%b!22fr&-KV} zE6KuhmT&TfFSEbg%i(mV7`|!P18nGMWEUubUvE6LrSraFt9od^c(=~lXLiZp{W0-u z9VLW&wvHZRR{I4dnP*HmHu4eCEL^!>8#&*T|>%?Q{W+N;LM zd=2621Eqk_&DMQ=Q!89q1HCv`M+yPW)m?z3kYfDa?N_&Gft04XB&36jxO=GR9*RPu z6xr}4WEbe~J7FYekX(0RhP<)n_;vhzx2GAG4%LTK4e)n2E*KpWW{^`cI=-tn8kZ*U zJRGT1t51o|362Xq=fb<`5WuansH5-heoGv4+0-j=H=Mcs+Ca(C57(Oz?^!P@1ACh{ z%7Yhrzy$_6b?TK6C4BX=w{u^wF^9SI&BIya43DkN{jY*+$|AmzbagN2NBN2b7!gR} z)tw?a{f|+!MTwk6tlv0@A>4R{H*KyoQv)!UYflZkb~!rf9$f$N1%AM8BAnkd~N2EzfWT= zo@8dv+k_zE%P;FaLmpE2Bp$xP`=B{C#`em>OqNObHBA`!hj0_kx$)J5+{xlCLC05= z!EDohtI(Log;sk}e2FC*;}^VbeP3f-!}~`zJsZoyN|M_jca{KTn6KDWiNAMmy)XDd#rTp!MSQ z`_o?ti5zu1f?ZQxgtk-}vIuVn)XDxIw%$6Z&1h@m#fnRT;#Nv=cMGm95?ZV{h2k#3 z2~Y|YiWYY$ZlO2?FCN^BL(vf2ottyc%$>P&zkf25WM%LDuE>7hto1ysGTI zZBIuaI{EaYQ%uJje;9u>5bj37Ua4c3>AY3+p^Fl=jee>!LgH_!qdEOP3Buo_e<^Cy zRBOvMsp}@w;J|HQjf1EkJG!&@;mUV=cxRQa=OhCq_wRTLoHf}#P^JZK4}uP;p;OFU z_ie2W(Kj@u6beU8&W6RZ?s5Er@JWy}zwYD-uH2MFh?JLBK9or*o_;e{W>d)mZdOx= zeC;G}{eB=;p@Z0$L6uSpSWz@(&i|5YHSL*Amt8uFbaz<7r%ve1|Gx1n+kj&oAv*{x z5Bou>93ZA@USD0@WYChB$_T zLRi>`Gew=(xu`IlN;GXQ{&F0%*%>1MA!R4pea`K=f@>$Lu;}#m&`Y(B6w-=veHm|q zb<#fvGAo$OOpfc125Zg>igS{+8-NA2<3!lsV2P}Y()A6pSf!sq6S+i{H9KuSq~qzC zk~<9*==xOl)$vPE=h9@Sb9*-#IkUcpzfWo^QS=4fnf~lBcskqC4}b0;c^KEv@eQyC z->IDxO!InAm!O~iLjt`rRmOz~oeJRkV(=EZgVegQ0ShF9tcdn|%J|uuxU{&ye-U&) z8O`Zig%s9q7b6q=@>Tj-ZIA%F2GTn5pB@DgfWKu!OM=XSu=dK0>eBsMr(VK;;$XhUJhz(>z^;MZC1KGX3E~?KzRIm^5(V_ zB1^zZYAPctI51w_bAa6n3sZ0mvmR8Nf{JctgzjS;r!+DubXxdAb;~b3Zfo5I8HMd0 zrCTT5)$)R->aj*1f&U&e|eR&(z&Hq18QC6{m~E?SyW@xeE%oqZ5UJdd-|JM{6#o z0Azz_Ttnp%LTdYw!o=Wx4daPzc25Wc;SH` z2JCxrD-KL<8ZNO&Bt+zjzv z5dt}SdM*jaZdNP(msiQlICiX`3{B^4g8SRMHs7B3kVB4NNS9FN*1KedA;eWB(e1`b zbJmXXA7AZtPj_7lf|9h=$9ARMJ5U6`G70CtYy$?{S6I6t<3Gxi{8L*Ma<<_sm$Cv> zECFw599|ZD3vcqG%fcB zJUsmjT&wz54l}MxXAWNdcvc2x(Hr}%HX9S^kV3ED^71jwZ@XFVnQNdSW8j!#2NC6E zQU{U#wlA*JSp}J=T$>R-a68?hJ2+8z^7GoImpqz73|WT$ZUcyk=w{j-{Iho~qsxM{ zUmH7F&t(z3vph~~YTi~6{at|S4vLcaQP;7yEWKYbT92bKAoViu1n&8PZg=})GKLQ0 zw^m?CT-+tGG5q*jMI=)1{JtPJV8W}~AJRPYU z=KmxM*8P2s-I%(;8V8ICvPojFcr$-Lm+)@<=jT14kZ_Gr!IO+rx85<*K{h-|nb+E( zTr9ndAt0r6!2FFe9`U-2{hTS~*k_OG^Nu@@hev|3dtB}M<}{0dl~UZxvWUP7wwz3) z+*n7sb8WX{x8op4$9g}pWI_qn-l^46@%*}a8Ot3OcuvM83M_%Ik+7Fk1Dc=@u>(^S z`9VfmPj%A1ue$yHhurCOPOrQp?xh7gTeOAOo;;rvc%Cxh}3#6nQM9Y2SZw!#pEhM-63f1px*q0__ z2rw=A(>@U3*EC;Y?bl?#V~ntbRXaTzF7?~jA%D7=kgkQ+NIZ6#YmDpIkMcL}-^NMT zDK~+zUhTK#mx%7>6w0OQy`>6myXDNq#`}`x%gdrJOJ=k!eUV1V`e1l z9i8<MzbHfUMnBYg%rO;w{bIr zw5=Hir^x4U{uu=Da|YiIqFZ+GDM$5Zq(97FlqMnZ3_6#{mhLcqb(|YglD$5D%z=LP zh0st;osG~?aztoo4sCXhj!4q?6823^j*6U$SGwb2Yb-WB+R^@{K6^1v&?@})ZoapH zpcPHu-6I-MX_xqIfNR|PDYQ)9N62ix@6@eBdh?&hiYD_-)W410Z_IpK_TYI=|Js+% z#Ix5lNmBgSt@42~r3W}NI5%P!?Ot`SC^?>@i(QIL@yUL<3`@;e+%3J5Ivu<*ym%zE z0!o)pKN|#?o)qTG6u^x{>odxd`&*}%OLpfE*;xjH6sMa$#`1SeLV?M+q8Ct(y~^LJ z&H2TxX8Dj{BM}qL&K^*|ZfBPM1$x4-bGph>c?3!X7&6;WA^yu3iFsDysBmr$y+k4A zCp#R4iKN#`ellXV*SRN)Ck&cp34VMpoZuR0ovBzUO?sjX_gqxdE#_?3ZD<@uUNSVX zFRkp|OK$jK852=Ih;G~~#N6+XZ$57P+!kB@GWGEW-Aa7&x&lnBK}y*EG%H_ z@`FRL%tP|t4}pYr>>M+LME1wGe~Ftr2qD+qm!{3oCy~|h7O-$Oy-146K$SqwSIh7< z-sH(|gv=){BJo~t$3 zS#Mosy4Zu81SsRWHemeyKH zSfzO&&@bF=xpfx^G-aq8o30`5uk*}~Poe&WIh= zRp)h_Ak-`@{&6eOm~gzebW6n6n(qbRqjuj?xrxe5znrJY0_9h zUa!?j-|p6X7^dcOexeX<@p1H(RhD7!W0+{AaNoLD*^)X_J%~_lS>h{>jjj&rhwmWx z=BxxIHj*~gvc_FkG6U$1O5Hvlp)JaDi;IhLW?twlF}|691Ny%GbVghAL(2X3{GWT7 zG2eEDu{*5fVEp>P-N6y&iO*0vey_|rfFxs&M~{z1N5_n3r9npLXb)$zsGZS^o_a`% za(H*^K{u`sZkmYKr^p3A&u%~kFsVEPR_Z;+p zZr+3c(}sKw2vR!4XI_%tR+gq+MBR*^aiOhcA+AVn%c?hga~dfL&QbLTON>lReOFA4O?^Q-jlZ*U&XoS#Gx)la*>gXYSpf5*Rw$o}LrJCI;GMTm{9lw&8$#@LzCePHL?FW8uWxg2r-DW`?7v(OU z5-Wz*pxJ=qI*}|Qy{aptHeD+a(I(ib_dD8&EKaoGT`QjuR@cMY*wAxCBbTqoyl2*v z*AP9XuiIFct$a-2!!zo}^Ai2j57G^o)V_+4F2YZZtUx|wEVQxrqt4FLNWBe{qen~4Re;HDY$R{lz zZSb!|=ei^LEhSu1o*Nq-Cowwh5%c#u2jqtbfj{O`ZWv-L*h@Z3sbN3AWDO3=iHdF7 z``|o+)^x>FaZ&#lUxsTge?tFVPh)Z~A0N1xz(}uytu-b0Dsv5?HX?`W|K5TSH(A4i zP_^?dKNeVAbwn<$xRVivYbhu^pSj}RxuYTwBr}l$d4L6yf>RtbvBhc#er7c!?fz)!v!x|b?gSop^8>*!_gGKS=9fMtu(Obd#J zojUE8f4$%6tu{bR!-F?lL+o)|?3Ld#@@#Y?I=st)71UoZB4_(C>64$i+imvzt@NKx z!K|&hVcfYXG$M*%o+M;|FutoYF_Mb zmgrnfY~jufq0FRWf;njR8>PwcJ9Ep1Awb@JbYtpq>TPf+DKM_Eqzk0*lq2yFs5)P9 zaGlGa-CqEFHX?*y-YZ0V12k|t&048%j9-l3hgA1O^AB+&ztNTfbwad4M^s22CQ}_xQG^z_O zXiv_st=38oM13!xv1Bnt=Qp&2@BO+0IoCUs46-K!f^v!^vF9aUKOS;D0@m%jec%2^KAcarr*dLenIg`X0h)~6p;G9<&mRL-|ZCS;ir2YxFgBtH0FN+zPsHG za0OeqTja7xeZr`&ZgZp^PD%~@c>y*<)(l3aB{TBp;T_`v#J!JWY-3B>U}^G-HA=H1 zZxunmYyO-ZFi>aByX>Ri%`c!ct)prYzp;%$#Pl^%N%MxWEd*Y4G0xyrYfd`PSRxv(sU?2u5)w-N@R&;%s0k+Hh(dASx4@IL7=JSw5R1`F8*ex8ic?-W=)aE?}jE2A>q zboO=+$@f(JxaI55)JYY0DRsXyetxs8=uIQO?+-{{d2Eb^tMJO*;wD{NBAKznKT2dX zFRmotOTixdHG=r1rym}ZiIVPZQv72QvD#|*)pus- zeDX~2&&=dI!;-0fQ+%!|8AC3~UaFAECXv8qREL?7dhCDrx7P_If;qNr4(d;>Ct#@) z#R0O%2MH?~!oq4Awf*jXazr<#mCP^g@Hh?btpcPG?t3Wp+UxFXZG47HqXts!>~%=~ z#_8K&@CY+LlnKe6;6)oo-HTitynl~Y0a%xeG3r^wUVpcQz`5CTI_Y|iI#L(em4c0r ziWHnOVR?~K5_ZMFf}%vQ0Pkpk=pE@@+^AiNW&A6L;=v^r*L=2hT65mzj8bHX*F9YWJn_>6_NHJ&v`pk(l@xdvATz;Bk>2G0Fq>8F;?Yt1ytrvIwuvAbzA4GL zR%iu#b72h_{UI9NvM8Y73$zhxr(x_KrsD99mYE;V9%r~T^LVa@RX53C&nz{a@5kJ> zdt6pzo-QCuAn~1|spG3Z0|cHgH`*?c;`iv>l}pic-t_z1xa5et0;%ov%?pzt5Hv9| zx#Y~MCxQxV_@w9I@9%q()3>&W6+$2WxtFk*hYU7~ADtN^;t%PzsV`+_%*-9drhz0u zxt5X!G8you_c!-BMM$X&ULeisueG?lOGojTwu#)-U_hCfOZvNOteXJ48u!4=VwdeU z@5ov(VBh6{XGq`s$L;>dxm*<)OsKIzC*!F~vDxq57Ml+p`i=3&XbhPh^)Ee1_ZTK* z(&FRp2wy^R>NhbKUC<86Z*NdGLl@7#AiA74SUT1~}m`%6$fL^G>LgVc$7 z(WMpN5Kv@GfV}+dy>Oj5OEe8UeYD2oHtK=$XcXmYezV~i|0X|PDCj$p{RFD?_-_lzK0&SXDE^-JBHASXDP1|#o`>0PjM`3g|CTAdFp{VD#o#cS{>#SsY{vnfkwb$?$U+AK)DA;T~>0Ua&Bi0^!ejM~4JSkz;OXeTvpkfz z$=dIbSQi>Ic#z(GI@B##khrVNib*3k-4lA8*~TYkP!WUN>f&-U>9R7ieEd!VNgIZ= z>q@-S0Fu1;bN4Oho0c&Jt^$3dwHr~)s5Mhc%mv#lbkvi7zml>^Q+NOW`!H{vN)Qk5)yxlWDWpd@c5c^P>5DnC@NPWmdj%OiGWLF_iv);_9`RJR-uwKPJ zMYc0t#r~D#ixukceIEg@xqLcOP8V7kg&kRnV>~l_-4)He$j#MG<{b&^k4uL`**!nz zN90Gf4$8cE^zw729Br^24!ovo_kR9xT=95+1l>N${J1(wN%`l?fWo+>q{y#lh3k>& z;(1}#mjkgJ#hQs1Y3ruMU(omb(_)8zGxfycO9~vM+5cRX^+Sgf6pGsFXkU`9>tTzb z5B4@UxBg4kFqq-q6honPT#f9=hAlCv`kk7X1hK3?GZ!?ZFHAQqoRa!{9e5+=jKTTu zi`40#dEAf`(NE!+;NVp`RC^b=DfTmRQpi8z_JZ48NZgMla5D?=&ly|J z{-T%IFYo<*nFq)U7-pycwaatX4F6V+FFQA@Zv0YySqLznliesSIB!?6bWQM_{}c=4 z2WtCS(_Bd$f6jh$j9-Ed9@x$I{&PMt5qGQG{QSaW?PP|{{g1%$i~+5_-y%szP`8Yj zYiB6}E?=GVl{2pD>gIFH8rjuzrUNr++!ufuE@W@}qpQ+!^84zlt6`C?RmI25yPtPK z4f85Guf$BB$rK7cM{gg>$=wM#uL<5dl!HlB zRjip}JU=}#S3Z!TAxR!4{~>v#9Z@8DeI{v>SW+s;8l90b3Q4{T0?h>#7>`E(U`V|S z;+_dApdGD$#{2kO9d_q(vNF<^*5~r>u~P!EmUXkJjy@Kov$-Db-)9F^K(TekCU*6=6{gRds z$^GT4p<(n`=(?{Z=R%WCBm4K*s918U;A-{xTd64!l7NvC);eZmFn>EqV#2zU4AK6t zM5+Ol&F~X4T};}*wG1*fH`%RYhH8#*nw0z(Q(?a+VMLJ8ZZ}P4ld&LV=eN(x80Y*R z;j}7hab8{yOYoas#7h~P9oY>8(GzuTo!jr!Jq2C9K7h5@BQGOip&LtsoF{uqZH3;{`?8i?;q@_?_tGHjne>z$@y$i>w>h zw9V?$zyClx1=R#?Sv)e6`I?vfni1J2@wYx`Ya_kZ0loo6x_`ag+iGl9m}qR2e*%<9 zZgG1#SbF##p{j1m6+s82=HV?RQX|_Yq!Y0u=d;L$r`hvaMtS|se4_f(XQ`0~m_{HM5E6vF%AA3;KuiWVM+$26jFwP!}mR)L+6n+9MKmT27 zaJpL87naFs*yCS?WYvK0@3O{9;Y4dq#Ou3eo^1Z6;kK<&&qSSS<*u15#~S{ZDeKNjo4Z4$eVL$_8oPog*bYZAb()O++& zyj~>j778#3|22qULJ8D*vO(G9>1v>HH5p$$4!QQ^KS*1KRV{z#JAfU}g2 z2WnB=BTkrt?@rw%+Z&Dtuk>H3EQ*`+7U>helnbiH#GM4IXI#Z>w0+S?u zF+c8v%oaI2k>Hogd;tT0`$iP!IPO0k7|9RsM?_ACervb%li!mVwHG& z4XNJOZS~G<9o%fSlz9k>|IAyH9e)2+wFzlF(sNdOT(mNLE@9j=!9U{anfz0STARA* zDqOcw!cu-O7B~Mc*B5l%3>Q7WZa=yop+%l==6ar9KRQ3P^MhZp-_-hDyIUyE>o~d3 zX`IXEMA=cjjWE-@&=(>ARt*0(PELg5iauKU!WRY>o>)}ejF(tFbo3gzIs^z5hrfl4 ziac_#2A(bDJNM7mgPvw#YqOWpW6_0cZ^f3xJfnX^)gB*{355uwq6LD`MBE4?(|YI1 zeFj=I`>=8m#yG2HU-;MszE|oyN3?fqv{9ne%#Cdpx@i983hZAJn2-KYz}6PwbQnnag-O@31urL^(pdD zNYfC;&%n6pZV?d5f89x*RJmelpRQ#GOv+7(H}SY#g)CZ*iF zcobl;7MR!=2~2>E1t!%>b8;C_Bq*YV(v3Zmn+W;362ibmJ3d*v-vns9Tt3NhX;*o; zyq(ClZBt8r{I^jg6!f!Pond488gv!~lfa2k71z?e8x6EHALVV@uI>!m=8?Syno&Xf zKiLYK^w!~@B;6$cny#u1hcPFE%TD;YmZB^`C%JZadh@6`ntdQ2!HE*+669c>os`c9 z_axV=qUxdCiWj)cUm8!IV=E&Y{H7dB%Hh3*wJ_vdQQO zYhTZ0Toq=%{WL8vEz}XBEMe&gzelV?2_X&ys~0-k1RNE_A==3l2<_xtWUc30C9-$I z@Uy7hGQ=dw(Mn_)h*^2fJC@*I)J)fRdUII%G6&<}?dV!wEvE_eeS^0Pl(0k<^h8^s zxxe6Kvwm*wu!PkyvqgX3NTS#nLuQquhYSP;Q0jZ&-P9X`J~P7hW970fv6^uW~b2DtH+qw=%bi<2#RLn=|Q9 zS>@pt1+?9PR{;2;slzC_0xS|12&oF0Yw6$&#*b*V5I--qOy4)y?EpFfbxov}RS zh<^==y0kpZ_oX+A?1G4Al`qS@dYwo2z6hAE3*bN0K_{QFa)8@hWN)pGE%Z$lf<#U%y z!T2NTo`*K=Z`f+Q_Q59;vu}hcO2gsTHn(`6N8w!ADCuVz0al(L%2G0J59?!Fsif`K zwjbVC@1F|@{lGgtzy!>qyre36 zy%%1C8u8$kTEFjh6JUcf$C=Prp$R8{;Ur5Ga{~5=r)SMBal!`~ln1BRcJZAtuO+jO zUT!HE56-jb!jy4&YB&IWoiX3mpVKX z-0^E3I<7TRH6?~>~&yi0*g8Jzqhpq zZyAoNquzCNKobKjZEhRa53ZBqG|wV}Q$!xzQ00My7oXZqTQad$3C|X>Zan?yt}5wj z{M6e&P?ZEx_*TO=3av;1$Dz+HJz$!aU(-+gU5|vx;m%2*T4e$|>CbF@a<;JFSL&2I zVhj-OPmjmaP+FJ4l(EGS0~>+5d)$RN1OJJJnsd5RJWB`G*9=Q~8ttD-67tnT6T9AB ztev(XxjchWj2Rut4vmy_=2<_LA$v)0WBXGz7`|+gmSd@pZ@cZ5Aq}bveLN=f7o^Y?l$(wSS#Q!{ADlz|k=qD~j!!L`kkQw` z{k^`^aj})rhgZXjyG4Ub|F~Sv{|SfqGCDP_*B}2nWMZV5A?LLD{Mtk;?-E9Vox+BY zEs5QG%#ns{wzPQV%SC=XW1!@q}C~OWHsqUczy#2Sj3a;yo zOJi!t_X!R$JYQSnmFxsD3^wC}n&eu3)X^ljnyAhbG>A@gz(aQ*wQ^9CLF%hSo)@h~ zRM|^O2t|{k1HGC{EpiPT;Vn1x95qR?++BFO!ugvZh9>4swqj?qI015%W}z~p=vNJ@M7N>*He`QO$KKfN_Hk+Fm@Ozh?!y0S z4%nMS=c|cZf~jB04g@Td#DZcnuY`}hLjoAC z1{)&=lja|%euzj_``5c=CcAaSeBPd@GPJRLozuz--Nr_3SfWq~#*+V-Nk@^A%F)RU zbh?p7d&JOtnTHyWjuAhG5;}W(62*JNtMk^|r0JRUg60|h5@#t9!K&U^BmEcWUpvO7 z*1J#_lV?-PxTU7ujczdEbbbIm>g^Vs6D?cq?yF~@90}z?@OUdW5_Yb?U`f!t`U?Uv z9~X9&y6iSYb3ozyo{EWZjz%fYN9`Ks7V`KO<`JNDf3vmVQTJpUTVID>qHgW3#%d+T zR~I$~P>`y7-;aMD%}*ja@`F2zgXK5OGEQlgr)Kg3gQ9Z^r|RSFQjJLHfk?%P7Z>H* ztjNRJ9pmFm*iRXd)%O6d)->jvbBFR!Wvusw)9n8L+%k%_RUEr+$@1e%W3@u3qzIc> z8a~p73i)is>rwkD5*ucIf;(a+dB0TI34G@g*j-nw-ja+fOxc4~qqA_TdL{N&LJ-@? zB%B`EdOACc<)dZCm&Loqu9Oz>tH!i-QBwcS0b!fbzBFiC;YpN}<|=|dmmBGJ7WdiX z5f=V*4|wm5HsFzWQBv@1Zd?B6N= z$n~rz0R23{O=kZ&%Hj9DbRnpB%Wu?aA`kG#5NHqK(K|N^;_t}a8JS@&G{&W}e1U^yBq4sk&KiYMym8gk7QI{}q zZC*!iDP z3E%B6qqGYB%14PE>gKBdCAZe^_xlz(83`$#CpbdWYK3JW$>c^sp$TZvM9^450ORSf zm4FhAG6l0Hdt4(aW5ffQv|fXy<6HRrj?{Mca%DI1|J=H^#8saGJIg1NmV{Z+n~~z$ zp9WPgpU;ymHxxhutboqG0exa{DNQj>lUh0V>|{+13StD^GL_?E<2}=ZqCN6E;}Uu- z2-qRT+*OqGAi~KsBh|7$BIBh=MeE_}Euu7~5a0>RBvp!dN1~9JA>L16K|L!^rST(h7U^s?!dw!x`95C(z&JB{k;@>x0VCSrVuCRcW`+AgAE(AvJ^&EYGjCSdN zFN;M0mQ!3WrD$~Xbllq6^PY>qDFeJlyj!^b0^j~4UifvU<& z8-2T-DD>P#2-sS$C711Tb{JN;=ceV(^*i3Pox|=rIRW=5hD2?;!Bfk^@9TM-)SZBhT zofDe5NwR9WPS=+y5-M>16ibeqBe*3Z_|!PrMq=xav&V&F^A6F~tXWy{Er)TjWGDFSg%4_C&w!+<2h9 zppn)*RA*+TqH|W6&TG>~dsJ4}-ChxLU6L&Mqw6mYfm-sC43D<+b1C9Dt@a1YU;Sx? zHOwS`4p1sA=e@@LRs^NFanE+y zlI9ZCLhj1asGV@ye#OGTTV28oH4k}8cTU)(xy#aLR3jg5%`=1|jSXJsJK z)go)ocyHFEZnltuuM(_W+T1kCtpAFtKG0O|(o{+bOAA(ok}-PFL>PA(ONo`W-k{3~~fOm63B>tCI=kBizpe3|>>q`jR=E{S~@ zW^AY~8~bI5d*vi%j?3Ah+wxlq`*j-W@Awrb=L7U&wRhtCLeH3XU5)h@9?|U_h+W)( z*vUf%&mH>Y?DrQWa`~?J*o`p2UBrtcyx*z;jnkDbZ$m^XPkAB`xXlouRfij#!9fejSFXz|MHEO)umF&3f2R>b}xO? zB&T+H6jrOez!ZD(8r!(S0MEzG$>zmF5bOF(_O&@(hr+@Fd$HV8{SyEUPFUs$kVo@0 z5+mjEoKRRuH^*k^B+sSFunblZl-tJnz$Ge(zhvfU?GqEQYX=?}Pvk$!xJ@F(;~)9v zvky*QEcd-hqmaB7)O6g*P5;C@;r3c1sK0FaL)`+|H)ql3ugE)OfyQxBb9NFCUV(7f6c8Q(3UWuo<8EO2pycuaf zM#~b->Mvk~50uxmHj>-&AlcH!L7}U54S*xh*)jfP2u>)W43wIp)gSj*C4f-^G*fW{ z80WB?D??a+SFy+r*JgAtqWuo(^4rc?1*XzBDxyp~zLouYl7a1B2d`)bFMM*#^4+`{ z;Q|SUzmKaG}(NE?1#u~TZnlKAFjwD)EM7ta<}@9MFhhZcEGe6Z*w_WwN||-fV>nAcqbD&s}uN) z^dfxP$ht+^B4j#y#ZDBIFZ#5G5+1x#ED?#K^$X2=OE%s35}QO21Z8NKRTz8UWUsH9~TTPZOM8I z+BJB?YZl1Oq;A*Jj*|h6$>SM6{NQpa`eCW`4Y4ET+dk9!rw21w0)|7jmoO@g=Tu4! zk?hjX>;>>^ieB5h_MghrR{R+kz-Si!j9%NHllvKTRgk>Ni{n1qTIA}~s5Cm8UMo*` za@q-vMotz6SqMc=zEw_))|)@D1-*zV4&qc!)L^y$1hc*K#-ZISzOIn!7^}w5YUH><~W~a%F zYk4Vb0V*!^5$1C@`mbn{&_D}TP&@WHY>m*0vWISf5nGW0{@$dO}JSqUrzu2&V)iybV-ozl#e}f?B`w|lK;#=)c{%e zfWV~&K7Wj6&D5OvopgFp{hktdT$ZWe!hsy$@7-pZ$^SKe#{WD1yc#9Y1-YJTPGr12 zGG5{Th;M(}|Gz}~+|82oPltj$_H5aaYfBQ!A&)#oCLR!ZiVTT1oIfZDLaVtfJ>a4uAf$f7*gS2l<8~lhf<-^8+G-t^tyFdNFJPSn&K6{+~lMo$*0d z%>t1#$kIt0!h-#33Q=d-F_WGmQJFE5D76IKP3oBD$%OKrfc$y?%_-*04`%#QCXh#| zG_8Ivx0USU$9L%Vk(63@tp9m|h3wEEXf)PXe?@ZCg^9O^V7x<;32_YXIne*mWM(!-Q#C&jzOKyu}P*bV0 zN|*u~bieN{j?WYWXYomFzLW%R#uRt_9OYPrh;U4^?%xwO&u_kHYzUFy3a_UgAo=Mf zwZ7_TD-2nQ_h#QZ3~7UQ>~FS1NpoTpHk`GTgZy8|^w$x=POuBb6;CSB+A`+s*FPgY zO#5u_ZrTaYAQ1@6!B)aE=eNG1a?t;X34I&opp3zzT*r$N2f^;#n1H2Zs$EYiZ#k;! z2yNOOh+#cX~)AL+g}RRx~HMe=qr43BM+C%ezfXVRob2Da+a0g@8Z zQau8#;lA72T<2aMX19y9h1I1WNxzEL;B@;F0~=^_uJ7OtKhYE`elbPF`xDg+)y7G0CQ>9x$b z8e9rWFDbg9P=Rf1Hce}E=A5g%SZF*9&+G@>$O9Z6)P6m@OqP(6@PPK~+Lu9VIh52Y zEPhTb?<55n)NeUGwy{97E@su-M!O$|l(=ZNd8vjzLzb){nS#>=f9xMXua(c%`qB^t z*vq9S8>wWElR`BRDDNAsvjDR3`Ra^gP@NUu*XtqA7PVdTNmFm7LgTxLWT6CbyV$2W7c_%Vc3!Yuo~c0r?!wYbxq5y`AUCxi5}{G zGpe%Yd8`oc8tLA;sMS5Eh*u2Gpm40h)knc?jy9j~gD+{?{t_8V-jHB*aj%ZMraZhX zPyfuvnMmM#SD=o`+1`^^Vs>)cJlU7+)=HSN1)4@XgD&)q8aM6R{Fb(W81>!BpDh#S z63z9GWtZM+yOo>$(m`DFjzs*>8mET4b#ATF+^0=fsdWrow;`JpelrX`%M*7*?ebcC ztK=X-mqBc1Y}s>vxXXpB{3;TqEtkhCiX1(BZ(o;4>HM5O#OtkiBBOifRVwe!D@mk& zdaSdN6F;oDoz{9-G8Pjz>yBGrA^hS@L_OyDj-0%7^I-2{FFier4 z0H^%3!3izTZ$waHwb@s1#5KJ+OfgxDHM%B$9ZO2d%c`h>l?=9fC@W84dBfW?t~ku=yikmhx9iXr zjL2zaWqig*2o_|*ua%F;W9VcO!Mf0HIcN^7I=#rmb-?}_{qeCa|B`cYV!6zdjbVd{ z)vUGw`NpVT^L6b9CmrWxsRvM^aN{*J?(7125w9CV{v;Gp4&~ce zL3s^{ZZAOm#-Mw7oAj~qIjsV}!^urNKp^v(>8SQNE{5olow|(3}NsN*po`WIIWdZE#mR~hnD?(woAXYXa3I+o*R#i7$?Rew&$x*eJ=Ob z;L!^~oj4=ga|z3Nt$y3eNx>wBJG91<-d}+$Ffyk^V zW1|o3#Deq`_l{daSKK7B)^^3SrAAL_`VzY=vcVntMZ&3$9kx@iDAC#{-kICI+hd-< znjrf9Rh9?^T09XLd3=vk2@7e!m`na9hnA`K#|ur8={MfGU+@Zs2pT^6@?JbW?a8@}MIGdOtWj#bNXZdqku8oz*^>~OP8BQH!Vi1?hA}1B-_d1|N z0G-Z@-4C<0^^KI}j42SGW=d(2o@@m}GT+{f&jig*IJPL5MsJ#02-Gx*a zDq7~fIYbcO2{ec+?w=+a7s?6k8o>4H{JBhoQjaRVoZE;r-9b5~F10=so*Mms0KPy$ zzfjKrbeD7pp7H#fmEDIn1m&gqabFV8zXS5^g6s8wYd56d3H1?wgZs&7RYlOq`ehnbUUO!Nb2=wz;j0g&k<$!kx@KH37$iM=Ppu)dw}Qrl-+Bl z@!Sh(Ur*A$0n)q?(%jF|pw4LeZwC6hThzMuKpp-ItE&f9b|0F?_b|by|JO4)-4*^O z;C@uuZd9j#yC=PFFMkN>AA{%bg!JzsWqOaY`^WUV?@$dLO$Gqp>jB^KI{Yt2|G0-f zrq^K*f6M5P1?X!T=>2FLA+Nq4hx^Cj{t39BfcrCWKMD1BJKRqxyN?aa!|QJ|qb(3E;Qln={w&D!d7$wb!jEnbKdtQ8 zpn3ao$o~b%{}Yh!?-O1=4>-O^c+q{BZioM%j`N=r{vI>W8vNCHOa27#|1{wL8Myu< zxc)5g`*T3|AF6EsJOybUh3l6f?F*3h%cg7<99OX{UxjqPz;Nnv{~yZk|J(%s|Dehr zlm7|i`9)>S;D1We{A0oo^8QP}%ZtkB-Y>)bKZpBYf%|^}_by_45B# zSG28GKB-Ez)rzJWYpYbhk(WOO9RCmDMcWCwt^U_K&VQubw)$Uzzdx(L`wiu`)&CCH z{{h$k3H<&!(EV@9wTp_h)&B)$`+rrq-{H0vw$**Ab#3)!!Vhn&=&}A~xc|Cx+v;B_ zx2^uQa@)^0l-pMSQf1rf-ztAw{g!gu>bD6legAiGzfal2J7_!krpmO{h7-G=?i>5`H-nw0?^nV7Y9d$7TQB3EjbYVkAf2#g8@Ujhf z*}{1FIpv&@k?H`#5yENGYGC#y*t^cpCaJpP8DR9i3qO(Nv4i*QRI3Pc+iR%FK=-D7p>jfskyL zOw>-*CdY;w4ka%|b}+$XycSv5Ob+F?>--wmw#HZcluqoJfM!bGV@p2#__>+c0NJjb z;&+!T{`ycu?olyaG~%HOM-@ir?{)dOi5I(<jpZW93PuKJ;QP@cXRLcb&v#Gd?#H>A;#~Dg8VKzT^m1jdX_WH zjxm@Q!-PbP>4lX@-CuQ;-qd`W(S+$PwJexZpOi`z^6f zdOm_0EGVd@rzWP3k4}t0hfPsza#8RT9URtHLf@~^jGwIcqS`nZb~2dJKgRRsD9>%% zxUAjy8PcVmADtK<8}<}wtCOSpr!fi&#+oA1W?RTJw(F1zYZbMNynW#K za8_dad_$DQww*lvjd5}DmR_nl*-V!6vWxSwuNhvBIIIMD*&pZSdd|xMCw<#)ZUr}% zL3F~KUBrK#ConDm%WUOib4)NPv1uT7OY_8ba{E6;K0y*xRPqug>{Z2wREiiyc&jSb zS{BzQLcVeJHYb1EP;>Jxcj6P{_x6JPc4~foUE@I&)n9HPmugFQlwr=-FW%VKkxd#9Tc*b!@!2*Lk08VcfjP)8FToj-6HKck9LIzfOf!hdkfxqr>^+bT$6h zEB|P_xHdLE`?~3=<1GELdD2DvmGfr-;~PR8oFA_}d3I_V8tM>q{x<@JQ*aM%jpXOy z9ymyTbq?^~AHd&brr7t?W=F>-_+0*t^C)K7E2aA;RjeCZG;?9e2R2OKonN%ULY_nArbgU>921Uar|Nz2 zh-rch$J-qoNsX8C9WITJ@%QfxNTEafl!i~R)Q_ofo$2@Q%6b3pocHf>-{&}Th`)8_ zToK{tT`4OgDz3iY;kZrV^9qiD-G~NOpGQN!Ijq{xMhHFBPAyFxUL#*%KaCIm@xC zsYQ#Dt(uc6-Ud%O@7wO+?@z0Es~OLE|9H;(Csef6aoxDHRpa+WT#lK;^Uk<(x-(Di zlP>;0iRm~=_hpV}%6j4 zIzgM%=#14JUM3?um0rVTnk`wTbAd9gfHIAajg6d`nrPU9wFoY5xrKO)^=Idu_wi|4 zE2nEuMr78C>Dp5cp0<6h!Ebay<9}L3+fmWU*UuupXOv$jmNmvh|J>_+^C(N%E2aCL zD%$5{tnCYK`lcsaPe=HQ@N=8L?i25EJAC$ZrDjbbbF)>mq@wjj%6og1Wy z$bzYv0r~904>%~?2~f!B3K?4=LC@p)f*MmkkzQ0`Mbhs-=*o)^CR^vDU>g75Q&C0R z#M6JsN#AxTKIJPnG3iLB%rS{Q8|9dI!|OAX%pCkGnUZ`sS_W5&S7EE*wcN5P#&^Z# zaxa(TM_f7ZDMm<+O-(JX<|)N@$K{rpQhbjqxA{&f?qySod-BZhk4E{;n&02+rq4_% z>ZZ49Zc5?uTmhZ2x>I~g@qH!B^!JbQ`1xDCdU|tj|X^NR-34usZGv~oXW2Je8SC>S(N;w`@XO_ zypk+XIt=@Y{zC!AyO8nY6Qh$)c;pw8=MTr_nW^dL0!;Lfnq~szXzJB`3>wcKtnTdC z+O{|R&O}3Kc0Q%T9j`C16+v60sLkai0TRyuBR^6E*$3jt+D=T2pM`h5D zqmbZf-DK*ss;L^9x5fXwYE2D&P8F!3fH`eyXtAQdkeq&8y;3#wc~!3O{DPCd?e($r zDA5@GzaOC=8I2en+o@buVd-=>Y@|K@9XX199Dp2%x_VC815@+QIWuM#rtfOqA>!$W z650rzZk1J!OUfPs3^e~Y=Iv1Q;87EDtnU{qd!iP8V!p9w#NCEvKZ8+6x`_=xnZ;0A zur0MSXJ?-t9G#f0O;3)_j-Rg$j!zCwo;y>U9zQWS^W^yK3A$j{+gx^@&5a!5Z|Pp2 z1oi@HVJtG`dkpWFsPZSfE;aWJyD5P&Ftl8?uGN=0=ChS_ds#ZWBQOCE zwx-yxvR|nwBB54(9j1(gT1l5nh-^Vjz}0g}VfSKIne_cCLn=Lpc?~Io@8cB}Wj|su zM2gm~?1xiOSqltJY1Lif54~(z2)c*|4+k&sn^zU%DxNOKYV$l*tpQ~}7^ANn)copo zqbgOeslIJ&p?u5KYb)t4biG~z*(Wgd)%Wm0p7cXOn^d;`gS#yY^nG_w#I3Uljo)13 z9W?htc)vx>PXV_!?#*V20*a9Gv9*QeE6}$;un=^wgaU3>o>4a4j(G`cMPIimdklJ6 zt9mDqQk}|aU=>QI*`@5k+iBV(il*I#X?L6Z z-FUy(++UCP!*IP3t~bDSA6##O>j5IC#UQ6%DEA7eyFEm1Scf-<%d?Bt=VB(y3IO?) zmF=E_tv{930oBU7HBS3xWOhqfW;+*HX2S(#wsT=*W)?vb@M>!O`{8<rScONH%_@InIil>y(#y)_-F{bQ7&y{GVJ`GAfaU67LZm`N zzv5A4J0aw&Bg|Ln;pk8!o!MTf@9&1|J<5srbhaPj`;ZGVd}7oRZPaTz{38VIUuW zv%j(?!G6>ucYAc=oPnUS-@C!M|1Z*hPN~d-GQ;DKIS=8{_ir|Q-U`!{TQJ?5;QD}h zo_ZS|RE65lEi`?r+Kt{}>h3#ifE4QC{ zqL|zVY5G^9pJ^psjqh=|j;q&RKcmZ4WAZzr?4mi!O+TN4>m*zujO764dvCFN+DX&D z0<)T33q-%@tY+K7(06Aw5&8!gi2kn@qII@J!{e*!TTBIS!zf(FeDJpS7rf! zriGlB=~1sR-Gi?X-TezeH-U&tjd$9lKZ^I~3h`z0pS!a1xMd;eYJ9VBJrCEX)Rpyz zJ+15rQpO+lbZhmZv_GuH^h^80T1&rqf7mky|6Q%{hrLhPJ~nx6_``noYV?O`S-a;> zl=p+ti1zvT!6+Vmf5Gs7H@^Q)xPHJqzXzYcAFdyS>qWSJNR{)1d`P`A{rDB8yL%z% z_Of(#tGTB1_sp9#p6FS)ei*JFf$Mj{^?QnyB%M2a&y|(nu2+cewpWPmy{{16dlrJO z#`oPO-F=AvdkgVnE&lI)h2(IB>27+3=HBl)it78ZD?zuH;6feL zrvapaf-kXO@ol%9mmi_L%q<)*_4jYav@|^&8c@G|i|=*t&;ZTf55x6G;rbc4 zx|{bC{?QF*rlr$=TDdc%&wvarNAb8@$@B1_j4Mar;gMcP)}My!XW{yM$ufM7(f;u! z%MinpmH~rX$<}wwtFK>B<=Xuh)D>;_*DnNJjqmd+tH92Nbe-t$t1>$fjhh@RM8%L} zEk%%fx73$H-UXmVxewqL@?4937kPez6>;^~LB!Pn5Yf{;o*7sS&S`#3~*}z zZU;CBa5I5B2^<1A900^~6yOkny9wNz1IGy5AAmOzK%IFrz*_(wAn?{4cpHI-0Ujan zC;;BSg8-&~CxOQR@c!KdMi`6&Ag`|nI8FfRi{ki_&`bcdj#+#fe!(En88O1!uJAv zKY?cf@E+4}8XqR*`T+tTC-4b?9|{1r@#g?OLEuLSdLK&Y+$A%p)=5dH|@KM}yTiuag?)6jrz^-l== zX98DqTm4gjkbeIuf&W6_E`a|^0QtlFKO^wpO2Yq3;DsFc-vCG})bIR%1YY*Q531(d zD#h~`0GgM-TwwSc0^eL1InZW6x6`B}2RaEX4L~=6-m47OSST`vC2I++%Yo|v zHV_yDz>_YIz;1wH0=q4!e`}$XyWz4) zpQdXYn;xy(qHXngbpM=r*QIfN%l!WU`QJ|P@r%qe=}U8z6d-@E{=dT>%-03i_mVVk z3Z!|{oMp5zDb^EZJqe#wWjy^^&m6N+Y~$-BeW;r*&$LEEwek6b>?8%bY1OK@gq5v`#yZ=@!0ru?F4Eb#gMod zevx7z&mVI!8?B|3dI>dp0HA_29-QpZ^R~Cnk@WmZd9Jv4GM_%Ey6->d<2}MxM#^mw zDU;{?;L`^=o_06=r+ww8={wx?nS)Y$m6bW11{ZpbCr#!HnCJ9=B{q~>v%M%HF2%y7 zbo#|^ddu_qy2?&XM|f+S7|(pNxKqWH%sq}YR*EGKp8RjzE_Kq+|D#FEB2pyJoBn80 z7r~?NyIs5&I6pm6x*12emu|hr_?MOO9?tDECn#xpy(-$c-WsQ88o9>P7svBp4&U!Z zJpJ+Vzc+>_?;6q*y?SAM?S{`w@Gc>zkK&ScMP-*l+_tMzxm{WQB3HW#FIV_`hRD*_ zgY(CpK(~px3oJgpm8QRvi_RUSMk=(M=3R1#xr&9dE58q6p%<&HT41-CkkgN0S*OvA zX++o}xWd7^>Dd@d;MxP%Zi7EPp{1)(l`UN z^0<((+IW!MeEIKFw(Gc({LeQ=zTe!x_sYxN(7dEF$JUmP9q$4LE=Td?$Bt7xD1%8R z1G=Q!20(XXYCyLqF6G9SZaH8&%*)+dy$qH|fTkNDfJ?cdd} z20)j3kv86A8sAd$N(*IO4mj&i(=Izzn;k(nLeTwS`j2lZ+aC3-Q|r#ptL#x;81;u< zfmO=Ra690-h^1|ZdRt8LEP=GkAk9)pgJXcMqaI~F6Lfa3LHAb&lxhD}(3krlAJ)Id zcMRyQ1{@oSj_~)#h;CrobwC$q*}B}XQ`SRKgwK7-Moex{W?zqw&rKjV-S-VD>kZI& zABSh5GHeB2bzbdhm|=Ohc*?Yql<7N3nS6LpLp^;D!TT^g8-O%ew>zM`yWn~~;aT7B zCVc4pI|0WYf@3eFy%Eyf0BQD;G$_{taJ?DQ?uT;ekcJW-5!-G~oUWaCVnlx+oquY| z^4y??fKTwW-pv!aV z=UplNRVv!ey_fU5+D(7BIh`(74oZJbl;5^q&hJ_$edSY(kBq*oiod)Xbo1EVbZu-pojVd-4W$S;?O@E%grxts177Fk!PeDlM@;k*8sO{~(f`$YY z6W}I?PUWLT2hY%z;@<4wZX4jj+Y%Svm#|FVGWRdgdrmLZX?A4a7ae%M#0-zG1(dhP z_q;e>f8&qjf1wCp#q-Ap<RrRfRhV-U+FIuHv=cVBUsq5B=*^VS6X3_gNsbPH0r-wW&2y?ZSh5w_iM3?>~~CB zDWe1?U5KwYzg#UXcLK%+u7lC#54t;r)&VY!;8IEgh7!7nu@6C4Z32^&1o~#XG-6%L zcq~w_a+gs0;kq2I(1)d)wpPM*Kw0;`zOII=(K%z->HD>+Q2(=?rf+q7=vLYOxB|3s z8AGRg7w|}zF*<;+bxh9J5xHQUBH!JN?+O9>LpM2n)Z%{qy$d1N1jba2ca6#aO}yVg z_(IoxUcXNJC6xC@Ri+=e<=oi9a&o#m2`*gm9#nQ&p=yHZ8n!WYdsmn|)&hdNrSb^b=N?yMQ+9d+coRr(SNKhOl;<*#C;m{U>CzATm0oJ#b=NGeX2!xzje9c( zxsqdHNWKMr&3=q*0PA_PDyHXpuH@>unMXi$=?h%v7Xi-YD%|%TQ+AmUgCFT}cKWKz zqw5OnI_^x2y52pm@=QcgCGuX)a}HL(hhAtd=J9_wH7CnD!{;|QOMrizH&A=S=#P;L z_9sowkG^!J=0`>O0=w-6YGYjue6>4#Eu6d%=W>RV+Bt0m_)>iUUh%P_*$|hl%GyMH z%v{J&5v1|+7hQZlXfFtz1eOM%n?P>_`T+U?Xb1*|BVgzS-8H{5=RFO-Sm}Z5R6OJ( z#3zPYY#^`^0QuNNV30r;0N!u(CL!IA~#T$f^hCqw$XEA{IoKS7AJ?gfjkV)?u7Jj zP^N>{=?^K}`B8jJfS$&uncl?kbuxU|8HMm^LG=MXJ#jz4@bxf!%OO3*hmA)o!vK_t zttOS)=`r6`7>}`*S?$R&)(Won_`{+jMq9Jxe7>MDmnmnCkyeOKk#e%ghi1zeG+Nmv zj7&LMG()oiCb{Kwqavz{eCe&xYMne_V1bjz#vLmMU+!T(xGf=#r$j^GavYCWpF?=i zFeQfI^O&w0r^ESup3`X~bV~3W#e>dg;?zu?I3001oZt6yI*Xd)H;N|#y+LOz8eex^ zG--U!5aM`B4J^FJGA>sBm{!C`V`Qmu#tk-fE>Yu*#?KKY@WP zlo`m~Evct)1P=$3{yHDP`%EP?&YPiD`var`Fc7zmyx*j3n8`6^1CsUg%^*x|g>46+ zR(1?Zom9q?`=~O8gvR%{O1-<@Jf}e(VP=`EYA`|w+jDb{H zmW|I0pEx}_J$hocHmzs_Haen|I{?IZUCHU(+J?lvSC@+1E*CslU=UpQPLq?XY zLwyr}?+(1Ky_MA4bGD zp3P8-jZov-M^#bomL6TQC4d{PwC&1n>-qoL`xo%Ijw?S9Km7ne00hu@5PXx(Hz|^! z_|U_WCF)_zmL*xTD8E8$=myXPT55JfjRq;2Va#MRyWeIXvlDxgO*~^~W^8Bk&u%8` zjV7h1m)ID@zkro~Rag zg8%BtHY^XldZKFfA&m1O@YfX|W_+kjN~+(3y;c~N*Zt7>@(i#V?WNj`(zjl- zxp!yrB~y&isZzW?k=a9O6jz_Q;`<(y@UDyyZA*_Prj+UZ$pSHRzTYb3C?l(XM3Bt}_IlzS{y&dja! zU=Gb1Arm1|=F)DNqMJ-`pFst04NPz|MP81zGzrE0+;Wr90z?kv86593*aUpN-T$1f zUSQJ$Bk9{=x-Bg}YMmN9p4`%sbl+}T61pumJxjtDc@w-*nmj0qLnOLk6f z8=x&$B74bt1gXGi%p%uJYa)dB^7NL3L^Gz{jQGromgzmHWXnM|lSd$UEP@Ji7WA(1@q}U%+QPU0zqmSY6!*W}CSm%lID9 zr>cj^^)Fg-eE?0cW!#>W!z8}MU0Q>_smHUI}e$sM( z2zfn$yq;pbp24usVwewu9uHL;d9GEo51@RM{Aw*a2l}+?Pig=5Zp=AMQ?0Lj48uJS z`UL3$mj5Murqcg7OaGIY{&`IQk62->v@^K`X}{PCH-u?Fis4>FzCQ!OK3&55H0I&c zR@e#5&qFA?GNws)wbN9G_Tj4PlwB`>RsF0j{|bhEn8|+v)3EtgF%I#qsX4aAw}xRB z@Xqjvolt7QBqFwDbXjsa93e-iIMi|PJn$m?TR*8c+Z$2=d#XLcX@uQ2|9jq(4K z6<>`DKNk`IPh%B;{O8ZQ|0aW|2@Y0^BC{Xf$mFR(Djy8@Rt(GP>p3@ zYSsVW1;74Z(9ha@{sQBp{ujU>6MqHtxb>^8`2P&^_%Y1mSFmositfLN?*9Sv{$-X= zr7Qe%Fvqt4561f+@%}%d?EVVI|I3bY5>EWp7X8oQSsnfYPF$^$^*51HOBu<2#)c8%aYaK#{a*;;{9(J?>AVyuY->b3K#wkmaA1i z*)q!J?_oKo={m zI5xiC%In8j*%ThV0lHni4#7VD4_G~@_Wve6Be}QQNA@C^769LZAWo7De~0l?<6>*9 z%6=BZ|1ReN(v<2LC>Eb7wzzNKMv^|CsG>t{A6V+M|27mQAfC z$FfB%C7VZSJC=WKnLZbm?g)OVG2+B3bro2Y zo(6urJgVH*a&w>!Stlb0EDx;ZWBYKNPlqgj>bCF*7+2D^U@b}8OWdE1Tw1Nv7Tfku z0jap8V=Zn6q+_l`z=-W{EhgKCfPdN!UogW54gvS7cBQQP?konx~GX84#m7P?(G-%)I>-y9tG- znV+DvXcEZ%U1?dTJjEXVKFPu>w{;3r48R1A6Au|QgYC$&w@*Ci7I4t@KB(Z3hf6MG zIuHjY`IpP(1uL$3a#L_fu}*4wElTFC6t>v$cNp>0Yy&52A1-<9jFd;Po$P|}e&;CT zY?R`m`0a-1VZ3O{B%mnmfhdr3ANSu*2s%@ER%7hl3(4B#I4L+9K`y+s6!_}GK7&>6 zyHR+%9lMZXQ@L1~DXDHW<)M?^3!_S~KVbr#D7ZTE;D@K)X<}*|NE@I zZ+{vhp%#=x%{&H?qt>t4;g6f)14f}!5RO91|3pN5i23M5STpRfTZV~CW%KToVmWDq z&mXk?PnrDANAYXLFGmf0obTz#e8Ttbe4K%B@rPC0t*Wl@@h3_G_oS4-y%1g#%KyFy zf16*crY^cmrfW1R-`26uiN1!T;=KpLYYMFKwUxiQ5Xpzsl$mdy?gJg>_rX;8)kAdh z{ObAuFy-vZ|9d0yZ|6gWf5;5)wbG%><(Ug}jr!8Ug=&4qiuAA%Dfbs+D#xkQxiWt? zGV?**@{y3tZ%~<^nYTWyMe+0F%NYMrNKNHY5b||5p5*(J6*2KlZ5Rj0#f8$0KJ(p`>Jtz>ziKGzz&~j4$$Y~- z$NCP?6*lbmR4cWz%M8$!=#yO$Y5wuBLVd{K6Rj^7%<7oB>PR!HEKhZ(Ecb_%<>?ia zE#;uYc#L&WI1SbpU~sam(cYVem_1&z{epN$YOPqeo*YfS*7iuzn6SG^9# z!+*pKpJ;t81;D@T21avq3IkGC`JOk^53D1cGyLtjidlYj6(DM^lB6DAfGCOQYj*gL zn&Gj-%EL|xEGY~#< zDgRkBfkdjOx(X0A!nul4hUhW&qUj%4tDQ6a1Mdojoo%wB&FOhaCjMTh>?`2zkiB!w zQrj@q2>p@~I!w3p%Mn%^x!MAsj|7X^HP-f>m*(Y2$>kfiq^m~w{4;UkFPZ$~b`q@C zkv~rEVA1m@BK)2uKSM0_ix+irIfC1onZ5T%4XvZ?gFB8-*IzxZo6&KrtaX#wk8N2j zK=PgucjtUxfv7qR&cGXxyiQd3MKk_1s`u(wCFy3eLUdtRyBx(Qfud!8-G^Kll@ydb!r30cIxKD7;BOm%uZPb^#G*v>>BuP^?Q(MMg*R} zm9|sjGjIcvUTc0@a++6)vCaIIxO(pPE5#W!4$9!-8ol-Sj+bx$WAT%aXq+;_uhux_ zXm=l3&2*D02wRv8TD)ggOF_t>^7Ib)U!2*VNKSsd?YFjmwyb8lY5JNz|L4}fzOT~O znvC8gRqn0ee-%dEcY^<=4RtR9|JsY*?@;c25Y%X5mQvGn-F;oGq70gTyo1%E?gt#% z4>y%alk)@J&G|**yI;ru1#%xslrMi!I@HB{*@F_P$NRukhe?4utlfuPgHgzpR!(zz zz|-sh8C+CXX&=bkZ%)OiX0vyG47 z{W#uFXn#*Vp8^g6+W160*DnWYGb-gi$t00hZC3G5K~M#a81}tJhNAqd5`+4dw%7}0!I3sq^J#rym8 z^Ox!Q<9L6p{XC}*7~-a9V^;sUoA)1z4`BC0JD%5{o$-|Wi4NrW0pL2Qws%|&Ifjh` zH2tPUde=Dc!M0m(R{PWN?|2z0Sb6YlDsM$6GR<6L!LDjo;uDxKHn!?=G(V?-Y8XG| z{*XgJF5AsCQ|AvscNHPit~FMMkK1MbkSW8*Q_3)spK?E)K!&fTl;P7|mZ51+8!N;6 z?J_@Y%JBY_GK}MQiJpHrnVkL-58q8WJ!9spM(I3@_m8xn`(I3d4)5pi{yg44%E~bc z;rlsHVmZ}dR>QPaUdg8KC02GiC-Y;L3_bxS;Cx)?k(N-t&Ij#LOGbQuC(npqy-4gz z3zCPqlGbf`eA1G~c}pIDM7!tKkv!^@w->Z~f!r@bx^dz~xTVT$Z8g#@Sm7c}NLA9G z!uzLnl+Bc-3#iI0*XLNbIb=hg{_KrfH!}~?Nm$m;Q`Bz|N`t3{L z(r@1ex(b%e(fD$;wc89Dys0eCYq$56k}4!>S`8k)+fBODZwy}oJ_O`nxm80z28~TK zXjodX;-T?q25n2@T2$1^Nlm-o?@GpLWEy2hr4rc#{@JvjaU@s=v&$?*$Tk~6oyvCg z`+~OURb_h_?+v^YY`uba(+D+I6L0Y03M?bja6WFM)i6F8+|*O~dlm1dk!t2ZR6bSs z%U}+xue$SjAS$16d@^X{O5vzos=5iY@8cVp)W@%YKh#rSr#8}m1Cm^ww7=}nB5!KX zIV7%W^7c-eCVi=iop)_~u97}`yuqTO7#fA(mzZ2M4*nOJCLsE|;hQYl8bfs6_fAHx zvhB2JcL#KU5y}1_qr(a4IHRc!=oAFM*L@7>9!95;$bB!P`yz793VdRqQdn|e| zhR!m2B!=F{=!q55(~LeGLLaf{1*CI~od6qn|-~8R?RP>WqjlxnE&KJnXttv=~B5jIJc4 zYmA;k`lF2g7^BZ3{qYd`ETjJ{hW>M;|B}&v&gf6Y(0_yU-!b}c82x+?C^hHK-GW-&w=a9(#%Z&a465+_NApI(%FQ=rxg!HQx{pE!8Ym5jd{wkxd zC7@qN`WuW$4&1PlcCaac@!oF|vH2ePm{`cn6BcOFfg-7-leGdF@O)CAW2mGTRmAen!8_?YkzJ90N zO8-!Lm(q17!TiRgTIP5X{O{C|jyQ&KMltMWEIZ|6C+2%6^4yDIRC-&${LbW7M|#`A z)Jb`j_K-;&t&Ds89lHr;A z7gpeBxf?vD_N2{yHf6Z2jLsBuhrBA2Z<^tB{+O_y)$ZKjvYDQJ)Y=2Cc;LH@dp7jB z&bJbJF0y=a-!#Y8{Zi;?Jt~x4*{s zVO=nQ@XF#RuPe88Dtg?R7M~20|B%VYe%<kuAx4fr>Y+Fyu5^{ko4x zx-;qjx*;P{J!>OX9@CnD@cl&^)j3kBy9=ofBF_v2PPva9^@F4Otr;L4XtAP!GK2Y~ zy3Alc-@z~mR`{Dd{H@!x(sH`>>1uhdR=S;TYHF=O*FEL&*6e8Lc1dYXSZVQ<&U|OI z=I>@EHKfD#vFe|{r9PWBe|a;1udH7F^q7z~e?4aYzPNh%17mwFqOU*A$Hs2;bsD=e z>_s48Wq$`{*5b27?%rhieT9ebUVi=a2l^A`!~gtpe;4!7$@`o7BL-{#`DH`lV;^6> zOoAdqCw;VrR@{`Ee=xZh%#QS+CHQwR9Py>X^;&}Ov{1zFQRJ6hCF0vxRB0oOMnh<# zL)ylO(%a6+-j+gI;gU|~kA^JQc#VA<%HU5zxUxG5{`R8k5NZ%r9fB*n%1z2{FUA?b zINKO+zRf>`1 zQyAworgH}4oMf`O2mCrG+1!h6>y5FrOH|L;;iD%r={^JJn52ZCmN#E6&6XQhY-?>M z^2oFe?`@B*i{~vK*1F7NQ6&iDJ@#U8Zf>DbZ`O6gkg&4&kK%5AOU1Wq>WHT;8T zFj}iIiF4fsHZU(9CaSF@UW}Qk*P7*8GeTf}?=4?iXkL9dLP}9NvDS6$lLB9jUYCIr zajo2%;#&)?elaq1xOgXvreIaO|2%JfO(B2PRhj|#6Y{z0;^=TZ3^OPRm%Xo0;NC9r(8d@)X5 z6Djkv38H$QS>bkma>WOV4)%8TqOv<_B%cW!oo#*R&pwgzN>ux9X~Q>MV$2K}f&9d_ z7>VWjTNCb9=4L` z9j~brl3@rc7_`bjD>rsZOcTgpZPPqAgGfyi^aKkV&+zc8wb<+0xKJ8{xpARUiON{U1GUt?t@ z?R>&eS{xIJz&8afm49uOm<%RJgPdz{%R;cbwndv@x92Kn(yfjdHsZFcf-+Tr_!kpu znBfr&uL{DqOK8BtVK_sD>0yQQ20O!ERey$u?^dg}D~}z_ob_IDKv@0b zOs~>?d?o6?{fFo{Xxv zD61y2R#LsY2meptfAy@7J?qE#^bQck--yp?_Bx4o+H@FU@o7{jfZu@ChQ%1>Nj2hA z7`_kuj&>SM+g-BNrHvNVkX3`*ig}tu+JvZLT{`s~IyipO2Z<2d7S9C@h5d^h893h$>Oxal)u z#k~jrQ#sFAb#X7&g?caS7aVo*XRNxoAJcvS?+;=g-h=snuOt6T>pTm78+ybM=26hc zUr`y~hxhm6{Rxcwc$dohe>%$ge_3Vy0H*yUru{)I7-p1jk}PoW<&=qLw8iru5-oL#x#Ly@%YiJe&rSf8PzTB+LEXL$=I69Ys7B^ys z_m^B7<%Ow=y+0d>UNE9(?8UGzCKax=e-kU*F*E!hcZbjY0i|#6_jo;eX|X(2D_!zF znw)e4$lTD0w~62zG>#3J#BWOUES8%~3yw^IUppjqr}TNI_%@7V+r`e5_3^wgEA&9q zejsdkG{46gV5P9)Q#b12Shfw=j^iYG{c{d3%|<@ z@BB)&D_^_8Y(qszBlG-7TpQhE@!y+@zu$3&7aH}M*5vF$wRE-I;KL{j{gfG7&$Mi* z+3~_!a9GO?Yq38qEp;D6wbWdom+v>j|G&2X0n`7K?SIhp-)#FIg6_#y@jD!o-jT@k zOf5Goy|6n^=Pn4BbEoCy{EwP^0^dJ72I2M6jr5OqL5_C2i<0M5cjOuRp5loZdEOl% zPvw8IOZhkz$tNIJ=jT$B$`!17;dT?Khkc6NdU;i+?{78gafiJVX+J*r0vdl zo5Y1=rT8Kb-_0EjtZ{E+Y%FDd#=4lF49-eg;rk~8PYmx&me1#S_-@Om z7tA-S4qEA}ba%0QQ2M=?zT-;Uj=b{izdB@xE6-zd_IUB$MA$Jp8R|UwT$@ z>(lk++0Aq_sPQYl`&m9{|B7~+)jf;(Qh8Ks*|cv(YuU8>%)|C!*bP?LUI^|~4`4nH zXgBUu`xYVx!5p6s;r%d}n#&O;pB*S4y2zp?kHjr!xWx^CzufCO&#{-0b*06lVaoSD zydP!dSMFnYKaO{E<`H)$si&EGXLKILjJ_Mq$2n67Q7llc}BecU6xn!?jf`A|28s$1Gz#X~pW}?EjZRq4 z6-_3jxECdNhY%05hta;2#ElI|{)ZXem4K!g9Y;FB=wu9?L87?#FgnAC>f&Aqp7F`+ z!ancwJu&nUBg(_Wj2^Lw!aU08@r1P6W&L0XJ;mt5NUl7n=hU~JW%OZ2{G8DbA(8tz zq~{qu$LIwlO8cXXUSMf|oY4#5J}Z>cCif>Gc=qTg7#;U1tre!SL1`b0uf`#D{TTgJ zU*BmNybP5tmnzlLbhUh;)ZS&`iR4PPt3Ka2Qhe1In#J)ZH?EU+=D+t!sdBmfGU&T$ zLV7b_0B>CORtpL~m(9ebk2YqSm1ea}uXoVLggsC7WePGuRQpY||1?`^lxKLzdPBX! ziXrDqi}T6l!{wyT>&}y&0ovBct|N7_Mq}BJqrSS5AJsC>_e93c{|O57BX^a$Yni3e zKI?s9>>)5pRi1RPVzsy+^i6IePa^eQ;A%qo~YzY?Ye8fR36dLWlq zmm&YBBD5J5=BjRP`TvT-S*4^m*69SyBfOg=g@S!%I3MxbM*9~4>Od90H0tmXA@d@u zh%=7H;Pr5ooa&LZ48so`%YlC!HAf9@?Fq4{kvVCN42IkX+jgr{SOkMj#$c2Gkey7{ z%!a-1e8zO^GaLb^y=BliHskFpd+)$sYN?`?+J#p!Jel5QLc(2jgrjuqdj`cpU|EP3 z2Q4~}S+%nhv!d>*mUe@#rTyR^U{3#)i1;Q4Z6)ihYOzu}Nslo^_ZdBon#W`I-SE>Q zQLwAqn-1RMb_Tk+_^k#id61 z!d#<1y=V?McZcB38GE?NfvJ?zE(*Qeto~cZ=5b2tgSNjpZs=F%0K)pC|7u+Zk}I*f zbb>VmxlxbW)zWmiYRXZ6>ov)+{Zpjen~ki!7BJJiFk7CfHw@tf6{s+6{tj(qZWVz9 z`f~YdhvF=vIIltaH}{n2X%$>Oqn*tSSR|Fj)1a|a@lp0oyFjn16W;Cy|BWY}_AqFR z9(o&V6!TC3^GHG-hb&hp4jtP_J3U)4&SuQ}PFA+<;2-x$d{vnq!e<{sS*bEUY=zrm zg;U^szzVkwWn;_gFv>>l-oA=?y9@j`^wj-0n8z_rv2xrE;cw(oc}7r16X1U#MD1v# z=pvRqZT8PZb(whm%|0VO`#VuxjHhEwOapek=C~NN-nWK^$6V#I(`_dNL4}8hiISzI zQd>68(Ra?rdhyD(asER)UH;;fa+@zo#n#E!-Lf!c<^YFuFcz=z1w$vCb3TW?OCxBaa*onrbCQ$f@*sg>GNeQEJRqg0!7)t=(d-;`4BZK=}B z{+*~c=}OOfvnsZ~nr|=V{$gDCoksZVe`)*gGX39;!bGaCxboTM3}1BB`3;NDzbB$q zM2)HbVt~#csizrcqpmP<7~D4Mv?xH}vjeD_uOb@E6EE%<^R~*Vzj~e8Gp_JJ32_{?G6QANtM!>BtWv zAAamU)vwOK#u4QnW#y)C3)t3S?CSEuk6(`?U6jAQz>?rfS--lEzJujccN0%zzobrI z*26ELn;4Qo|8Rab0!H&w`5TYq7s%g#Z}Ky-GlTvq@)yG|=UMx20)9@EZ*@+mv!q$z z)Ryb?4U_zzTZYwL;H3T>HMgcTsXuRJc~ow_%%L&UZpU`NwEAZ1d}k25RJxNoAAgbD zJ0MB__4~mN=wAP&bVtBk`q>Qr_?QR(+i3^EeKr_jgPsd(oX;ShtLM8CmC?tDU8_*W zh(+;tRwG^FGlG2R%Egd%rGC3)J4^eTRR-Rk^NhF^a@1Lmwly+80*$PZ2Q8wshauWJ zd7Kf&nPB9yG~UW+dkF1fM3$Jl!T)Rs(QZa`wAwzTgJ7NvMRYiVjxb7kyqThtF?7Zv zW;MN{wdB$I<#MA@nJr%^SId{mwWhh)?%b&GjxOtx6~q}6th9TTtROl!ifdCrhfW#G zo)8rL*wKb@t7nDw_m|!E+>QENO{$>21Y$MNxr1mE{p zgGUbpKZi{lO={4hSD2I>)gY>r7aBlmp?w(R3_0R#z&J`qqv${C;5m+N;z^6w{K&n7 z;D3W@tAhs}yGJ~Bf&U7Jx);I!Og87C6p6if-;X?19!YUH!2a*UvK&M?(<@+V5M8kH zqXyHvSbpLb^e6B?f!`}kRw{2NE&eL~GmQVec)t(t_hUNuVE&aH-ed7l$LKxCc+f`1 z!+1Z7Jl@NAJc8vuh53IJ{0)i6!9O-{KgN9@%IJv}{f3pkI^ynwEd3AR{VBXZjXa*j zygiCxJ`DP`h==i+lGi_Ga#DGD9^;;4_4g5r&qpmjcAHYKKYW7a?~ma9Jl;Qr_ZKm( zPg?vhwCK|j{HI$q!}yo+ei83;$afZHJ@4S5P6&UQ@wkNd8s6)8UqD_} z$#`AH`xU&uiubE{{|xe60^M%^D6C{-Delwr{y4j_QC_H)X37_8OVz5AM+T3I(svq5 z!MleqG-gxgk>g5m6Dszct-sIn6NX2wd|I1Xu?wQsMwFeeS6uO<(vRMd^6RQc>1c~q zaGRt97a_c*hUe65!4R=2ya&um z-Wx3eyXF4rkQUdK8jX_6zEzgXesJ}8-`_e=&uNetS3bq(V0X&l_h2+ebSB16@l)=@ zk%`3i=%}QQM6{$F*sBfFf*+7Zu3)B%5n+!iaJDcd?7KfO=6t<(DAI>*l^D^z+I zJRV)BkJ*iN$|}c+ROMj0X{pgRpit+3)|cF%9_adCS!rqSwsL>6!`v52i;Lw;>Lq$L zT32fHP1%^*`B*o~)5^ma)N`_5NN`fnD?I!y+>d$hYNShi6YRBZU7hqbUoKbV=bfdP z*Y3Yh?p(@x%5^cnWcNb0DSQu-_Efd4O8%GVc`{oSKfm3yRY|*NJ-}pdZ#CT7mva0Q zxw>n+llML;M*Fiv`nmTlw2dk2YvU@^S42Pj#kEP-ua9n~OSVs~aK;Il8Jv%kjnFW- z?Z!sHw`=n4Hu7JhxZ^47a(s>IvYY8HE!YmgiO)zp`5WhHQ}3eXkDxrzziilGW$SJp5X;qg$J<-;TPOuHsjOG^O1!|J8K;ESjHkAJcq) zp8Ss{l9%5;j;>KX-P&~hcG=BzGq_fy_}&HnX|=8AU!~_$$@2Fy4}S~g@|Cqo*KY^6 zxV*o(8tGtuj`x!~o@a-CccT3H{qpWxTOS`=n{@s5cZaq`M(#-+l^-_@8EUPXapS;PC*X_a>lI4tk8y`y%KGMwI>s7(JPQo}X zOyP!hcM1-D=_}i)ykR=~hYsNlIsV6}`rG9U&^#XY9eQANe`(frG-7_yFRWgp<*!uW z|6j9(a9(FT^9HzImpF(8&!M13z;Egq^ivrfMgPsZdfV~)F}}iawKh4x!fXWh>k}j3 zcPGX3BKTjvxQ;Z6>6Ve#1O#83P-}{6-CK3Vt>C{t!Q=6BkT|~Yvf^zINw3J#BfWsf z+Z~eLK9-g%pLrE$6g8F6^%8?*sIz%6qWB z?gz8JRhS1M_-cv@t8^;WC(pLJ;ioa*k6@g~Fzov{xAe2j8ArbF0dq`x0rUGY%*V$u-ygNo;{BAr?&9jFpLEn0#r+w)U%>mPt++3; zxGKN@B%}^!G5m~Ghh-+OX_Q54T*zE;$l}$II^1ZbG0%9;Vg4!+bx2rH!8)w6_%CDp zORNqrS#?;)d@r=}_!URKzZz194XX}cvC`sW^b+{5^0?}->8L}B%U|b|*m)~M4Aa3LBsPwWBu3iG_4ghX-k@K+>qfisc;kJTu;Yww;Ie%g1 zlEo!&a`9qqaD=x@>&jL1%RoN=5uK~~a%FD58O&C1NVfW25t8I9`XwPPbsm0$5kB)z ztQCEFnQPlJINuxc%Q@H79{_4;ijDe}#h7Hx7lJb1u(J6-+2Sm)q_SHVI`I zbTj4`3COdgZ79Kl^L-&_X=duxxYC$M=mkpim(!OLV*IiJwZoTE0ltS1?QUKX+KF4(i9tHJDf2U^WVo!$#l`v@uB6 zJ;%W{XU`uu^2{XN@;edXH${9WptBmxLiXpv`1@hB|7fMM&6yq~lE%{T4x^x4sa z%LYHa0Y_>Ak8d637&!2BM~wf@2!FtF%a(^Pm;8sRzIQ=_I{56{z%P_-j?yRqLv#OY z7x`yh0dn<=7bElLmvOh5w}66NH2uHe9*h0(d(7~sqEIH+tS>}(_fFTF%{m7ny8uy| z_}-X;SyoMf`1?)&z}v>=%Vzm~w*Nshefu2QMax}`(s6CcZ<$*kP4CN~QdmySSWqvKJ>G#aPw6xfaVSaCt!m&)I zDTJLKReJDw)QrZJ>-!+GX$7>G`;G9?^Dq_uJtqINiMp9n8%;hNW(yz@EqL$Iqf0O=(RwL$NNm#1vH2EoBn^LL(Rb{)2OmNk+f`Db8wdJ0|t+Ds>25%s-M^^SovR1 zT4zr}bYJ)ogtzrpI)8mVsSZy;>mMx$1_k2WoTzN{;p2E_jdUln)0BN*^a}e2@0_gaH z5k7h?T800p$=^m(chBmgw(zl(<@=Wlfex`vhRvF zCiY+IM60N-z4W5v>-@cgdCuV6`5nwtn*8Ovn7<58Uf=Q3(a5aay(#n6+ueNqPsCSm zDEkL(aHHrRzFuW)sF@2@e6L3T$6EO{;@Y36w09CI=_D^ev*_R$x=X1n>(mzC7)Mo3k& zUO(>UK@bI4i<-4dTVN%~gMU1T*8;x&wwF*?4Q72xJ%SG>s{yYC4R^5ylE`8tLKd#P ztV9-tl(HyvQ5K3H9S<|6(~Q#!2PGkgd958_OMe{e_! z_q~*RRJ*>O3s!hJe~DI~;fi+s@{!b$RFrKHuAAGU|x)!jW(dl~A zzk9SCUgLE+cteM`2hKbrc6Ghu+{?>BsW2siNN2q6kjqKDu ztdadWFcIn_+WqbIgiz0e18Lbs^6DZ&{q?m#C{>Q9@cx2BHc{i}3y$)y)%f{QCd0MF zqH}ovcuHA*d>zU16E0ak;tOTHV9_Tb9E)7?TnQF^4E(cfRDV;m2ZeV}Vp?}u{wnUz zwdj9n(ccCAt|IyOI{fQ?nD!YxZv^vf5FZC=4Cu9$P8bun5p-m}? z;)Q)THjF6V?vzCF_NF9?x1SNo2Z%3WypXAIb`UJP9vnXsPjOX(+nvaqG5@iewroZ!( zQWSAi^W>KS*%=I&|irymWu7*%|@xVc(LBN zbiH_~+?=n^PU#fBzU)Zh!22q*#jEwDR?t!_YQ4C$Sk`RXXXsrhy;@nke%8@c=$=7Q zyn`YtSPA52sujoM{?_9QrRjR334U`+GVjak=h*`I+iu+wQ#z5T*14H_0A&U6o~eLo zD%q|1DDfEy@`+p{vZ&Lsu05WQv-(Gz__Eox^_k@VF^Cyw99{mW6&px4U$tcOawoW% zvt&a@R`R1LNnh&(5`sfquFoYqUxwPB`h8Ma=2Gip+4si5Mz8-;0DbrG6(3gl!I8OM z2Dbrd*^M%O5t6O1KZG(@^DVUot{~-MFwdMCz<6XjG-$sKRIJX*I`bGK3Se+u)U z`ra|{OrhHOI1c7{T9hW89XM&Fsm!LfInvyLX{xxpz=zHv_~#%N!9N3W(kkmd40iy- z?6=ZAgkfA|Rb{bWhmLE%2|tq>_&{=R7&KcDf8G@@0?NsobN~0G{^~pLhEAFeQpLfe57kD|zQsV+gYV^nW4@CGw zh?*RC=lrd+uL3W$nTJ}zmu&wb=)At{{6zJwVF<^DH02$EuCzPnr+^jz0z|jFQ83%x zSaiB}yZcXDc5b@r$Rp<TPmU<&6Cd+RJu3BU2XL&a{8>b{oy_9W4{@O)_;KwETx5_%2X4 z+BHtqymmdN%lEmo?Rpf}N4p*evt4fj|E->%#&&J0wCO+ZSHOg^UB`Uh&#elF^Yh!4 zeuF}*V#=*lx-{);GjzXwg{SMcpKhkB_)WIMWj&o_8wC5a{g#-$#G{rs!um75rF4S# z-wh#rd3;*J-l5C!H+lS|CapVJows=Wtz{;=Zkn_b&kZ+%j8?)ozFlY?mgKw^?XjEb zW>6Qm$E^;o)cwGmI>n7!*?``+4HJGhnvZj;6~>3;@FL0~8$+~=)X%7w(O?L<%-Fbj z8*PP*f@$71&PcV9P1Z9-lOaUr6x>v83!@zn{3Olp7~0Ede+V5!I?U)GqZOD!{L*sx zE;irFc@F=3eGSiHXq#LmSxUE;lKH2Gc7ng8JqYHfhc>b}!x*N3_ffo$fjOQlJbwu5 zvV!@Y0&O;=_Y)3TWtOMaP<42Dk>_jV z&L@-C86N)DsVw>PtC_B?((Z+Dl~(WI={_5^oYrl1GJ}1FE$N-$%STj}sJ9_Lh2`SQ zhEy(G!a1|d(xS55$~Pb!)3)#Oa^6BU`MK3dm-iua%+4H$nQs4WTixu#Jii;w$2s5TNQ=9g<%s{Pe)SH1QJ(AHzJXHf)P_^J22$S->}reFhjn?hK^+2G6wO> zZ9Nb?yk#fKP>oG_jBD-erZw=3+wuAvZmbTdUr$-?CzGgWd|5~LC!JoHrCii>7T}hS z4^OR8x=*fAy3egfy2NM9;={+7d^+l+pZ~Ji&$;EA(~>VwU#0E!`w#X9{T%!MV^$b% zET(*IY*uDoUc68(*XG(^Jm6!+NvxBb5dCFanikUXd@uOZVYRH?hvC(@u>V&{k8E z$a4$wR7*QguyB(Ym-HBVUc<6I5B}E=Rrz*-o)W)={uDpY(p6#if&cYc6=py9)8YMC zM&*AG=2PYEFs2(;-a0F<3VQ^3-^J2DhUrgX+9#0DaSVGm3wsK~p20AuA^MA~d%<5C zxgWzm;7I!+EA97Md{y4h;{6fK>!S{Sl=l1Z{(kVk-a3i(NMS#SVLpWSr?8%$M)x!5 z{xG_qMfY>){s_9C$26WqxvMsje90nfSRn!SSAgLB*TT$zb(6JJ5wL=>mK^+Xsj_&X zIbXh1zEIXy+h%@@`K_I=sB&6s6!sFw89TgX90vB+}xZE-{G!gQltdU?@`KW@f<%q54Y z^d}&?9hp{mnMWVjz? zXo|`2ikaV}JhqGd2pHG9&_VJ2ELb*(R`K-F&^)|plc+bEx&Ka+U%+mCm+6nMsOit2 zdPOup3Ax?yV@0RWP=@&Z)i9nxvFRCvEhm?}c1OyKiGH@yu%Z@?sKIY#>@isf%!2ot z{_ZaVJYtm6F6M79JD;cf+}6HG5d@peeu(Y`2Oz2!nEb-z<0`*B*oKu!T?I(hYYs+= zApdvmB3j=W3{>sm2oW&R{iCRKJQ09-nME~4!Siq=A`h*2Ur#z1MaeDcM+x%^SFE_! z*9aerY9=I`sJt9A^WvpFda+trpt-??xki23%p@tk^ySvShweFa{~de5VW2WQrB@pk zG3n!xqE6nDc`A98ZR9^?@{j$_=?QQR$|1d@)9;+x_YBSjQ>;5a#>01Se?ZS|XrCv| zo?Nbr`JjDCe_hs3NAH2@IUw!H(H>Ho{BG!Cehc89^U?kW?YCsu8?oF>IfHYjcQF5C zmirEvX0#bpgp_+2d^1J94n3;f&yst@YA@r7{?NsBUEN^)Kkf!rRd)z_YMedm16Y;wn|ks@U2p^ z3dW_-tP)$^acE-u@}C=T&*%BL759l7{1T4eA$Mt^w?|BSb2f{_uN3xyi6L5ReHPez8{h`-UHCR#;f@4 z!~0>pAHn+}ydT8-QAhrQ1BPE$M;-ZGMdOUWyBg_cP?u4B@6ztSPwr!h^5p}~v0E#T zzqcCcW)gq(!XN#*oc^3EeOEs@?nswiY1@JE1*NbrG4V|?$`6e>Q`-H7Wb*UBv@q30 z`6W5>;Dh?jV15G>a-K_Do1-%*h{#}ifDu`R zL|LD>h$>)|(KrMRrbjKBj36?M-pXh*5*bk2&v!u3ka`y*@+%^#xO+qB0Hgh28dKAm zo!qF}Ro=rE9YH$E=y(jBWOO=&?uGdCDv2Is6nAdr!;H=%J%aQo(qoJsk0I9?m8(E$ z4cXQ7J*(<0t3Q9FzMd0#fD`Dr7~fATM_T9y)L~F$5j{7z(5N@-=*-T|`LQ`EPT!!M zG7f|X@9$)f4E*Nr|I~{v5>=;Dy_WK<}NIh8l_8%V7|c_NB>E8H8qI+33cy* z{GVKIKX1De@xUClDo)3>o?p>?0e%$pd$kqkk4L1V!dzW$4+HuCjKW#nojM4045LF> zyc(_a{*)sWJ7^p;Vmj`Y%(z6w4UZ`kVdhOPGZ z*>K-=p}tsYw%VZQsSH|Z2x}h>f=z6;(kRdHkoCrFD=!#Q9sR-bF7c%8+of=R4~t6-jzNK0>IPnos+=(N3F zE1arMWmm{;MOBkmT|(tt%_0WzPAB}SSr*MshQYT9%_2{-eN;R~Ao!}Hx{qVL0{FK3 zWOu3ZF^|t50pF(oLFD-Ym?!>~v20s0?kwon9##3DbHx7``h5t?q2fFY{_B~_Z9w%d`u8ZP#;^i}z4pcj=VFx`Wg?qLjTr+XLjKZ?qU6Bs7#((p9Uo?F*(bi=9T7mE>$kF;Z#o8?A3Vk@v_ z?Il6nry`{2DxLNArfr4y&ul4MR+07_u(nlQ;H*E>YO*_Tz*?4+iQa&9SXicNp`4Fp zi8+iYNEy63>a_TdS{8NvUca}0c&lnjol%3Vk&|7R@z&k!FIRpQ=>_B4~P@ccQ-AE=Mf&iKUD$1`23 zk4QfJM0zz|P`|r+Lx+i)66cXR8@=Dzgq>BIPp)(VDLe08U_g97qVW2KtI)2{0n z_-h7=&y(OEX+`T3{0jjz$KHan-->0i-ahEosr-p&5at{``)6b3=$E;F7P7xieE(zH zoo{lton9}**TeKj$M4;pcsDJ-i<}InU$P^y3VsLo&@^ zItk3x)@qwV&hO|#y2NM9;=^^$lvnC4n{JsJraiZfuk_FyhVs(gbK7*^N}IpDnZH+7 zFMqlZ^Hnc(bvlowF$6sRo}Exf^H~Q!Z|#Slot9(`W;MhE^&wh`qG> z=&uWDtWLh3=JV<1x)z@s4>G>@BHyUFaTJk-=|)EVjD{1^CL5hXflMh2C0)R zZe>ds^^wwNY<*@U*5`Y`9DC?%Z8VQE8X@AZlPud z;>w$D;uKxpU^zm!xx9s0bNk94WcdWC)hPZU2Zsc}`V0@>EfC<_ZsyCMTFxmpeV*ar zyAAwF6*&#Cw?fhM6*y~AAl1S*$qT@4XXzHcZ9Y`y_gH172{n=Z%1E!}PXFH2dHF9+-{;=TNz}jPUj5jt-;$M}+a( zHTILXVN=+`$0XhfylLUXKH#(${DTQ8tg3^37>+<+$#*yS3(+I!E`V=zwj2CI`Zi*o z%&q>^=i3Ly=;(!*WMgy$=!YIqF;FXC38JuL=`?U`y9xW*jFrOEVd!P41J+V>{^Jxc zYUvrmkB+hpkbXTR+_1Nz-abamtl2Ol7{Fe{me4*PRij}oBO)J9NAb7i6P7Mu z4-Y2vRo$X4>|&w$;~3oNOg(^IFX&bTAm?bnF--WTS} z#ZR}p$)}5@#@v!RQl(g_6+gW=Q>vCP+;#ZKr;9V%=w-SUvR$B0PZeL7uPhcL3lg1f z@7`LyR^FRT(vdCS=A2iI9eg|G@2c71%cW{XSpew)#t}O~sv+$sQ%qA@W@#OnSfcaj z!ZM#+bSz`K{gjN+clCO^`T6XK42SjFk&xO56WkXDGhI9#Ncpo%SVH0R6B%1lI|SRvv}(w!fO4{O`(d8CQ@)U~OLwc@qK zTIuETg_&w;aq)0_ct3LeI)lKa#qwU8P)|Enb^1mu3&Q$HT?* z=PR?XzWCyPPaccMif3907}n`6ww@Bg+1~xp#d7i5+16lRtY5rdG!vm#*W%T}@5N>- zP0lkySTXRL%t|;+vk8l5hjT1<{fu2%7GIh=;FKZ%o9J%ub2yg6PFOAeH!c2wJ^lhP zg~fY)ljVd!#|b`s;Q%Mi-ogoJxA(OgH6@`15S3#-NxbTZ_ZF6L7po19v;fe1!w+XI zdaP*I)V1@kraI=%zXoQ$I*tdez~6$Ee+1564L=C}#CI>4lU|za(xf&Iz6Bknq0+P8 z#u~%86X2g(&-1h$d@DQkY!9ZbET+hkjoz|StLKliu(TlaK9=SLrdI=P_3&ZPZ$s?@ z|73o(K%ge+MNE4S_^b2#Fy9A|&mp`Y2LBw3J@|Y-XzQ>em_PL<)gtmaih0??%5)s# z9b@UL{5QeeA+Yms67Q!l{kvHnPGBBRV;mk&FEI;dj#)~Vm{A;Z+Y@KK6{MS)BC_|JHKw#Bh|4w_oDN~ z6LaKsES?0unwK^F%_+WPsUr~I+Tn^sEtlT`dJTHbeqDMCOJ@ z{vWyO+vXe9eyn%I(ya*uO~G#M+JH)bV~X_S*9a6x>lL$ASu0TK45vs(g(=WiyN zX!H05h;CY=rX*K?)j%aWX8L!udZ74>M+(!Nfs|Xt6=%B>GhPU=5n5(yz+mgGz=>MZ+p+7erUkZiud+ zB7|3vtJr%Wyy33aB85tKFGN<)Keof~>q0&`zx@&ECta*id=41l*U_Sd%G^PgIpwy_ zknd;_L-9Eb;eAEEb+m*r z%*vm|@@K$2&H5!U?M(7kdFhU?g+I;9Nk+(2-eD{+<)7KYc;)SeO$w#AS}%epumL(p z^`C|A*#ND5x%`)RGhHkfe-T9U(Nn?XHdlW9ZrWTi<7puN0}wo=z?({TVx6rmRWn|u zIOCO0F;~m7yOVNmj3}o+m0hhI+bPo2_8p)D?DP~=xkn&flN*8VHo0LfsUq!ce zqIPAT%L9=i8(oyM0+?^~Z-U^U{|9}?~^+IDN1Jt1lt8F^~!rY-go1@n6enf zE*2xL_39q*yI2P2Q>$UUs?yqp_nmm($I{H;97;Xk>kxPJq)5dwXL|96g&ffCuejy{ z$);f8shRV^?xaRsGVFMwLk>xW98a`C9Z#gDF&I@L2Q?NoTJUoUH)G~iMdb0`gK<>j zrQI1N0M+RGFe`xZ`!S8f;P;~p&JWe|BaZx~pC8&BI-D2UY&w|-W^hhPJ!kv|7?*9| zY*WmY_Pa2i?syl=)8E28Wn-w9Q4%}sjVXv^J&I)8X^)2xKX7DoLfXcN?6$XC&lT-N z;z#=L0`nOE-4HxPMA3dQkLwSk{fv$v-Noow0-8d)%c2uVcgN5PMyHX^Fd`Y=?u8@-+^s@H@}DSn|wZ+LE{mczbBZ_ z4)O5a)(RrDk>6a+bgBP`ts@c8@ac@T-f@DjCnYJzYP<{_QEz)!Bi(`(&Nu#{6gFZC zDXz3{9lerF@_G`|?G{9QiQi$1pKH~Qj0vKsQv^#uDUX2i!wk%oo8IaCA@2!g{vrPN z+T|$LOqtb;lFW19>lbu@R-XCq0e?GF%?{{5F;ebSj+BOe2&FX1B)$Vo^P3@yuZp84 z*M9`_bOytXIKmac*9jCRsVVPyjHmLh@ShHpqV(Tur60w268UaMerhWG39IbeP(DgI zR+Gw)gI2H6N`5)Th zP1Bvs@5P0;OnKtovvO{AhuwKPmI+WmZl+03CgvUG?B7cw*PyA5xf2*PwXqNDcJpl) z0D`k@SJ{)B<=AEs$z^LD*cp(mWj$lapyRSfe0Dvm%$D8FUfX{dQnkmBgSpJSr^(-l z>3=pvR$a}Pea~Px+E4J;|8;+6`*z0B$P9h;;_VF`FX?zde0W`RK4EQ>^k>p0GT1+v zKIk3s+x~fEn54F#r{FVd@;IwgXHOB6rbNa-XHeH zr#?oma6RC^-)yfN3`dY_-t8iqi;sc7PpgQ}zZLFng5bP7iti*6&BytHom6%{58uj& zz==XQyKet35A!>IcQqgd0Qa)wV-6_}Y`Y9Xy2$n7z zXf=O-7EKv?z-W%VS$ZFy%J|PV+iCWCubN{B_35mQIFY!r>_5j-!ORBkeKRjVs`H~h z^!0zJkpBt1U8qFIarIek@QXJn=ARuRC;*$a4Ty23vewl z<$CqA?=C~;Pc2)lX|uKOm8JUPQp5XXmHPhVQnjQx{!3~!UN@eu)Sj$0weKqM?^~$W zn;$DxS1p15nR;zjf37YlZu+WMp(Wr}K2AOGz@^f{LZvoGrC+L*FSj!1%N?1XyYE7M z`la$rQ!~uA8207#|D4?}`iN)SJJ0IN2-dDuwMKIQ3%%A>A*PO&2^l0$p7Y%moT37aN8LZ!1G=RH5=G(4NbPV!3#8Cpz(omJ^OmMI@ zjGQ~I_on@HMgs!~j!xRUM|qF?B5QvmnKg0stZ&&|?mdY5(;=p<(RYkJ-Ld6=*Ro$^ z<>ntdJxYG)K8Ah=@qV1b@~jV)ir1!Qs*CLp1zk58Oj!)nvlG^{yO;e89^sjb;Gu{N z)F<%1iDmv^1RwQ3vuuUK=O($+`H-K!No=@MTqPgJWTzvP%bcc0^F`@lvImudSTO(t@ zDl;8#wJLmJ%U~TD)Lt-4Gl|_mlM-t}N@u7KvPp}Iw{aP=6p!0Fxp)k-|0C>w0sjv% z-sI=S=LOvM(N;TYRIj#%zhZ03Un*7>i!=2`V`-tuICMIX^F+Y+7wi`@u7JN}L!(}b zd5L`_{h&+1%iwc+9o~Mi!y0AM_Og{MfnC>g^=f^ofxa1(d=I+wR_Okj`BI}a(=0b4JXF2FkJu`Afi2?Oixly;P{2MsBY)d%0Y-j&>;Qwu`T60zvb3WdRJ1@%td&If4 zwAg;5rMlK(QLZ3Smuk@l#5)FmJsd{(I%5ppn~ZXB{f*6wLRDN z8LeqfVDnQZ#y#LKdu;{(^v1oQHC%fgQ}Nh3d|FF@kz0Xj8oH4oo}i_~L>4UWp27GunNKhQ-^Rmtad9C>d^2FIhV7*1 zv^(69Kob-!#}n%IYK?ui0KJ^UH%)0yo@W)4rb6;v%s27r_Q?grcL(xy&A<5^+&1Lw zh42?9UBu#bJg;37nTr27fIdT5Nw?7=#&7+!PKIL_F<50KTEy_{z2hZ}6}6}>KkYTP_QOEz z@kD@$^@)%PgqCEbzE04Xi^`+@*IqYcw=!db2ZX?z4pZ>HeBC)C?tE@z- zqgU-6YmO^uPFsH3t848?fLOgTrusx{6s=LTM$sB$Dnzu#pCYX>kPB}yp&f|MO$hK( zMrBJi?`3ABELY$>yE03GYw^qc7aXGaYq9*lrRcqhn!{HTY7Wc({~Uca?<qXE91bqa31bx;y`rLkd@a|1o!5+U}1baZRN3ch*XZ>JLAQvmz z$&00Ymhl)URjc(Y7ivrDyqUnhUdH;LfV2AaxyL7Eu2yM(F#p?jn0L*vYqb>^sQ6w? zFV9+;(G+{<%1yQBhpt8LVr>tb&usZ(X{p)_X+G-c*o8)Yq1iD|ifo$AZTocr18~d?GHv|l;Y5PmOr52xR1%0@L zPth1tDz+-RHS!#54>Z@N>eDY>P@9_p`^GqU9qC|-jB}$YiQKCeN zD!@;L5*4<{k^emf4gBC;`Zjk@qrOy|rInBAr6zq8yLa)GrBb7;WqQg5!WcE^q1gnV z`z#$f{1(sQ?eTurTBwovZ>LYktwXJalGtvC>qaXK^u6`C0@yZ4e5qckiS`Pjy^8h< zqP>dt3ZlLK6l$*>Edl)7_e9ZszWu_~0-4xNRtwXwW(*rHqjp0a7C7m#ZUb!bmuZ+~^4xc++Y&Qd~Z_L)q zi?u_|;(Y0{>OC9@B6ZWfRvH;Q&0+QM?AyREuEw>(?jyC2e_+ZEm%;LqdPUmBfF?2dO?b_ZE@mt}YNCCl#ec5ruA5(nBd_o<5& z^>G9F-nKxpb_yg5BzLi30?7i&^1W?AA3-0-UC;*veFS}O8}yN*+Ck7q&<6y41by@+ z=p#q9%TeuuJ$Aw3?t(oa*dy4p4zVYYi`9QCx5Y2>Unr=s-}eCPKW2^I-u)K{FL&CP zv7N84%BGzt528GX@*v8CC=Uf#PZL+!v=j6Z^bzzC^zkl1AKA2%O*;_u2^1{uF6aY- zK7u~C3;M{WonVh(4+!=M_UKEnXZ>&5#nNqhK-+D<@fWD(_4r=lP^C6sZd966)74V# zWmyvj(Tqeh!my$lbyzUbj6^f)STlkSzYg%7zltlX8Kgo0>!Z_g((0sVZ>6tf$opBQF~$k(qk6zT_pMzK?jNGTWS|1u^wNop$f3 zG%a`8`yzt>Jw+^yKK)ob_<>69r4Y(h?{ZLTcrtohIaTi0m;LZ>vj5*?|G&llSK0r! z*#F;Q|9^}94_&XZGj?;iU6r}bEb`%qc74=ZSmO(EipI&E@hjnCdFJX&dz*}V475~V zH7=EFO`qjWi#XN6cJI zVajhVTe*DK9qx4|l=Jw%%Kbv(dD+Uo9KU6S{Vl8LYCQk$a+|_M(F=4sZvJ~#Sx+vz|2sci zvT1H^p;~XAdf)+G)sBE0E&q2TGU6IQ`!K!{#EW8#YFV~G&qvJumb+E+)`V@+50(gV~;-`WC+e&1Ou~@8VmzYSP@dDVK1|gIs&>h_=nqrcyiZX|~uS&uGYQ zS3<#dgs!MVyvuk;eK)@$EG`5|8Cg_K;7l*pCh!k0(Wd>mgTOtJ&&S<+A_2$EY{Rbd z_jE%L0OEm}Y9#4atlfKwR@{}=b(t@&}i!kTQMfmxrjcF(C^5eHT_jfiqL$}u(X>C}Yo0enCyY2$x zF%{*q#SM=oOcGcRCy z0(sKMIb!m0=Z_CH2U4mo3(-ZM0)2>|1cl|VVje3%b2mHMp`rPUUcRR5K4GeMUtS^H zlNMpu)VY1YU`>4#Cy&8gP0y*iG|~>98rZ8(&~B;q3j)lh-Uu<%3ba=jx&vE9S98Z; ziV@yGfuxNn-67fmeZTnH4CTHt7EAx_w7V_B8!;(4&6l2Ky7{@1b}!jK9u%BStW}|N z3i&_$njCzo@sZ(YgR5nOpJsJ+xvMF&)_RhCd8lAk*u8mc%K<9G`5+;Rr_>l62MPegCLM|5ou(6 z!7o|puB!cAM)^Bjwk)nAM9~H)td)MK3Fd$D>0w+`f*_5kFm+&l1c_A!oY2%-Mzfxc zpDw#~EAhhmzZ9}dq(JB5wmnAHQR=f5Hho!|Pm-F;STv;@UkO(=ywUaJLLstcbMZay z3=eX==}r%m+VGoRrkdDyDM>*o-fonu73wR3D@sQh z_6RP&neIF%bbqo`!vNvpFT(tcYio$a@$aoI#!E4|0DzFwujgda@6L#GbpeQ3}VfqHu1 zRrbl8HKFsacIXAqE1fuoQX~~`S5{iGk2xO#UF`N34vIgJEXa=Lmbzrc7pRJirX6p6 zO?o8}8BWor+lL7x6c|LJ;WaI#sk)HXMKrj>FRdOba&QpLB7V1p%Gp;ScF8z+lgCCJ4x+ zli!8F2WqkTWh!MFIlBE-!!_ECOEVEZ|Fl0f$aD6Bk7m+I`5$>cOrzG>hyGjMM03=S z%l6RMUKy9QuF%d*F~gtVl3?3O)?iOay!&2PP8n3*L>&)`|+6aXUQ4%ux#kU7|KYHk6tpwacOU)XOX`5SE#ovax0a; zE=qyyR2eQhFniUOepu3vyoR8$;-ES^%|AloiW7K7MwrzOMpQ9{I<}9H8Ya#`yj0X> z?75hmB;qY}Qqp5p0}K?Bce=+jEyHpkeHZpz_;CLjxSft)uVkdhHaT4-@nj1o=?>mWO#W+6s+?0tm z)?IVmt=Ko{-o%# zWcW^ubzN)K^j3NTV%+xu22y;-hf%o0Ev?w$v8fM66Sxq<+4t;pK9J&Uw zRmCWATX)=8a|CIAoG zP$+aC>3P9s9Ica|9=^whC5vYYBzw2*C$dw~*~#E_ERv|EOX^cA-FsxJLKd(V{0eD7 z-TEE;JoHY06zlo#dWJNSC#(!s36gACk}lbyzUWV%YV2_)GFLJ;F2ywox)!{<+t$So zq`ZR?CbMlQejQ(YSCZr}8e%f2pKYa@^zq@Id@TaY3v)%`np>JNi!*pNGHpeH45Tp( zwL=axOkmyOxF_9#$=N$`@YJj~=U2N>Q2CB?%o5?+%-&F**sa)OSsLztpAmRRgr1j)oui!N8`fuHHAOrw zB86zO$2O_|(j{raAY<}ay~Q~AsFB0aI$_=^ew|;Rd8B1{rhI{VZ2knwU_~dv_lXDe zq~WLiEE#-avNw0Reoi+oKVztmoEQ852o+|W!w!(mT3%J_n%C&^;eQmh++C5+C!?Xj zTQMUeiaRJ{5Q{=#JC@_HF$Tew!7YRxz}|qOA1a@CS?b+F7tZ3mzLt|q&CRlT5zwtc zUiJxAV2zt zNN8p3c@35}6hyr`J;-K2(27}9zkKvXmwHO6$?$uY-&50rjB*Br>K*iG*fNsLFL%|M z>po7n?~Zsvm}#m%MqY#`U zZzdFHqg*{{+cda8RvLUTpx4~lNyTT{)nb}2!IHuM6Bw;LR>T~MJ7}nvVD(51osUd2;ekJCd_9J|JO~JT?rfR#h>6{5vSCSI*jS1^Hp8c z3qQo>L-05CH^t(^=aUk_PV6`<73}??lTkbzSM(zac8}99zQ%PAHXr_vu=$+P*gIch zOQ1dL#?Pq^g%9%G*Oy-iO^~kV;KQ0x|D4J(fu3|R7L6LbJjJxJ_7Jw{b3U>wQSjRb z2Qr85_7ZX-m|cC&l8ZUQbTB7pZp!cK(kESp(F99|MA>N*yM17eonr}K>o2C!Ux`K4 zb|*U6mR_pIHHX<&GmrR2zLn?696y`)v%Y?%g86$>g8EjH21s_O{}MFU6OE$g0$eSr z<~SiL#rK2qSa1>W5#r-odG)bF>dH4!u{az>eieadHj z=lR{Hbn|+Sda9Gfhi^lrkX^G-TUDHN@y;p5W7OifSEcRNJX|6BZDd|bSiBtS^ttqh zc$#?c0my3plVO85%9&s}$?g|(+;o!$PP*X;6ixMmhi|t+-5G8%$gh)Cxyml8oDjqf zN0@HDv)9Rv*OzdA7S*yKMxDj_C~r;OC?~-;p0!LESp)xO2Y0WH7GarM`RcTBs@HXB zI{V4CwM9!1IQNMx#j%O*pzvQrvJ=4WjxHZyYc6MJUV0JN;fYGd5>PoDZC|cJ0x6h$Z`+Z$M zeP)K`rQDC#csJkA?aUxY_(l8gbluH3DOM44azA&fyj;;?&+1npdTR*|yz0K;uHqsh zhijnO0~C>{*+au-A10nH7n^We>T}yUzs4=$6EC^uo*3o#cZ>-Gm4{2knTfCQ(g-^Kw-+xZ#>fBqmb&E!dR{QkMQZzMZ%MDpsM(Z&NpMHs9mtRvF z0M;-75s&R(7|PyzL&H3~;c=z9bn*UdD{Qu0X@B(WQ}ZIJN8Gki*^*F;d(IrOstsqZ zg-6|7iC3X~kVg11nML4vMsLj={Yoqox$3_Ao`L20X~yf}eq?D8#aGrP;szy-%pQP< z#X*9C%;`N+Z0|`6-ZG2jxLZO)cExS1C{3a zDOs_EABE%7Xj+YcF^Om?L0ad}A-2~}l`e%Ut+M>hS}8xiUumS2;#!XXby1&?<}>=T zpu|u;;t|TGT0WCV=XEzVt|jyMiOofC2g65-GOsf-L@eQEG^+A;Q^U33Gq4Jx_e9T-&36bBCG z`03a4Afw*wT~QZNu%wx2J5kpzH`rMzoF^lC_qKldam8GEEzv`3V9c1;XrRnA)~{bw zLu1pj7#$zA++v$c?V?(&%y=EfT+i&sj}~py#e=0TVFZaOauH5vPK^_N z!5RW@YnH#mfT1jr?UNrV)&tH)5C~Yu!#p}t*eriiZ&=xVLH+iGi%BBWfSd8r_R6s_ z9c0`VPD5e{@MQbQ;pLq%>2No|_qh{JosJY`-E~pLA zw8Ua7Zg7LEdLcwq#~vrb=IGl9X*za5L6+KumO-6_qg7Y>f7_=K*f$X&ci+TF81N&q4HpE_1N)Ns9uX>2F>hj)ytf$)_uPd9^SK80G zIsR3jDEH?3>KD1W#k<}!Zw{bkIad)p5webDB{Vs4YB4JE-v(FoA67oxCqoc|A(R&s zA8#<(ArTl5>>bg=QwsAZ_+SX%1%>mCgUJ)lVS7yfD;4rb8ukcg<|c|u9pr)HaE@gR z&t)^(6}KPZZj^<)Z@ueC%t6>g-=xWLffZL4XR{E++eI#shd0%SW6wb-kC^p)v3up` zpK{)2NT<8K95W@{rAyeQl;dSN`8|*9%@gDSfoD8%Gks8X7mZj~GVu(Iygcr`_aVF{ zrzr-nJwnH6QZ!LW@|G&K5-bC&h9WIX?D-NdQ}2m{9V@$t2G}pC`{FLAjz1DMVP@Ss zy0l|1pGbpov=T5gpBEW@u}o35r1Wy^URgwm^VXZtLp0b?B)}M8DupA~NB!-or* zKJ%p7+&^2yE9Vth8L4phnYi_4W?K33)16pC@9Zx7f{56GQXs?O!W|HyiMVdzW;}UQ zo}#2`Gk9R3b`L(|(iACNeA>^TWX1Pgl09%vxW6#2B4)q+UUi}8XFmq=+Y#ltKOXV# zVwipH2P3I@Ua;S8dmokjL|q4ISkHh|P7_8o5>rQsVt0dL_aV8j9*$hHh*w@oqQ^XO zMRqyx4Baa(oZB0pX)bh3tNZElkc8t@Gpx+WINnuDp|p>nfmToqKlMh*g1pS`o(^NX zPN1BZ=85!~aT2^#=NhVqHkt*=R^PetBb#2_93va&w^{n!Mt?^>us3ndC1_-aa;u(QL0~XJYF6ZEcg+d270-Flk zOA9{9UVA>8Zp}2?H0j?rJ2vSvx_%lOEk@}!0P9VFG9o*;Bng)HDIy# z+uEMz!t0*3O}%_NY`iB`LbTateDcf>Z4Qh#MhcKF-#RTDl>QJZ*irLk@2x>|2kE9O z?Py{pnj97J8|-JXzDdtVq~&nWygfZ|WTv^r7x^J0q{-s&&0GP>IK-EBv2rHZ9wG*(+>AlJElMixL1RHc52-}TvMXR4)a&q^BY$aS12>^f*1I!8DhgQ~FX6Jf|hB%1*7n|`9+ZJ$ALX(j<82}{Cd90~BM z|0PcJjUMCN=kRMlb?PsORzAokfO7z@%daa z&MPy4sY(Y>c8hiNg~@iy+#FHhcLLOR=-1ny#Vy_WjoDs7It9MrAZ&l`lp%NOT5D#~ z|3X}UUsxeaq&ye}Gr#J0<$yj~T4_yHi7qDq{e;n|WAR>9&;Cp})cpOrFHtwW$OT<6 zrqroBboclKN0eTC~udR!V32Z{Bh4q$KHd8@O)|7)5o+Kq0lIEY%(@-ZsrV5 zf6pqW@Ac9H?P*Xwifl)@D`F`LpKA1XSL?fVHO4LNKtm`3lc7UXWNGOUbYuH;V>UHt z8cdq0^$fiAKQRU;W)ax%Zkk#mUY+5WwwzQh> zj&R*y@3L&Qqh{UZFSi{E)NR|sR6z9dNz}`hjUq+gekBFHNG^^QI&|TLzR=e{M!Hzu z+#3A=#%-*{;=a>g zbCdt=_v@^nTUna=HvETo)5lMX(WBA21n6liRBeJCynOUpXR!sXiDM=9L9;dAUI}o2T}{UN6#-c#J2Dn2XN> zQfc+XMWGC#8Iv$~WQLvJb(9T1kjNh-Ki!P(VeWWu7=}+K$(=@=J6x}BIyA83A@Es4ZkMJ|XHnQN$M_MPCrU(k;&HwaAbQE1S{MOU$gDy6TN*L%@ z>Bo*lh)PJE@l~$Fox5Z1u{T||R68AYE$VMJ@55J-cA&rWcQ|`xE1lT-SU2flcoBD` z8<%8p_oRau04?ONa9%W0a!aIIq)5A;MS?{Z6-a#8begBkkfexm&Yv})dbATHdlcgA z?*jP;ce+EqJTcB~Yox7IOQWn*M+=o!kSco&Lk!%u5`g8`IlKc-^*7dhJ;*3 z{QV{aO|4>8f5K_+bu`n#b(KBqe>AF&d-f?I;VpI&Ox5DzW>}K=KzwPC!_u}0(Be)b>K)RLe`IN5R%#c~V zFCA02kewdyhrT50kkt{q^q+dhaQ=#Ven8PeKmvq9KY;H8WI>>pCxOQP9~#^<$|tJ0 ztMJ)Bu7AG((#X02I&dd>8c15h<~vC-PV4VGZOKTcpH|DTyINIotP@ggR4c*I zMVR&iy=lk1$93UteZ~!2hrzT2AKa0qRhVI*b7Puem2SFj{(>p=D;?FdzV*U0go-NG zC>aj*KO{E63H%z%r(6wid@>~xF)DFjW@RQj5-~6VJOSA8PoasklVt>s^;!1%TL;q` zcYn;R9EKAIxWqp+Ul(g9c`~xAHH}u$hZWB&g;*)7@cz~+ z@nsA&qnW0}pPPk${w}9-71DWu`gb3Da#c*rRI2j=`0B29PIeZCu|FpjX|DoW1bUBAL~5_sD$P%uZ@dyRL>(vWOG- z=^LT*&CNBVEc&R=`)qU>dY6sYMzM2{UR-+^C z;ArSFuV~@gn)(XVzV57fW1n!ufUM$2AZ2z1Oyhg@={K&QxF#`{{BksPey;RisWHgSeYICEsyXar+$F9MAo7^*a5=6KK8{4_ zTW)L&vT5~@^#aqKI2xL@4vwTWc+S+9tg+~6GVCOgb& zkFu;8c_#t4>zzh(>BLS=7u{p}qr}youJ6FoV-k=}4IxjFZQnAvO(al2di#gFYUf1` zq4ve2hk3H0jqXjA`L?n2v`_J|a_mzs!nO_Mip}I2BM+}eK^2NlqDCIuy9>g<3o=fe zJ(?$x%v>hUfB|)HHj$j>Y%sx_FCK{gP3k(n!LJrdqK~+&HY?UcSM?i!@3cF ztTpUZ&wB(|H*oKfC8WVD?8>X9z&PYj*9@~qV}fj0yf_5TaiQ|1>_!%*gD zD=8o)!i@gW(^6lIT=jw7R>TGuPx z22ri+u3gX`BzSfnlFBsv;R!NauT6L@3gPiGFdvJmgov6z+7EtOBBsJ8-)q&=SD)4p zKEg7R3qyX~H`2o1-;Q93ls9I0IaS?$76NZsx+(XYG#l<1e`GDwtvYaDemIMYpJto0 z8K~}+wKFa(#Lrzs-QTv+X4W&bN-(rDVAa51e6Qc`Udf?KuMV{0!{ks=ucWGl1o5h_ zAA2sgdo@nXRAw zVVC#!`d$Us-x_h35IdBa2+QEAbDWvp~oku)lZz2BpOk+^URxVjtU{B?;| zSP<;gYRvXuYRM(Zg~`^cERsE8AgrQTBiHjx*@jHx=W3HKv!2d-JQ&vA%d;6gdA&Ru zuXH6Rwum|FKPz4?cYLq`n-jMi7+&|6rw*;$Odsd-`PSakL>S%}UbVYY^spD}tMggZ z_A`@KFKi~h9i;q<&nwJ-#U3YSt4_Bhy+x=bN^;{ut*Q*M1 z3gqs7U`^?chy+M8Hj!iPkVv=#+*wwM)yZ8Q6l3moakAfJ__VW#Y9@cfs}k4%$*f49Tz0Q4)US)j9Ujl8{&XK2SczfR`*URjGwwOAj&(2A8(>KA?(2rnjO*3 zCq@r>s_hAf7gKYSS@+`QO3^KYbW=x}vka~qJ_LjAh%uf9J(kVN*94WM9$VsPVABR4 zW>2h|`2$+EC{1>yK?wb&=u(zaR{P-WH+&y*kCh;yC}6=+t@J>q<%X7GOWM*Z4<2kRSl9fO(;(G(1JzuQl8k$Q3U?lGSOBc)^_r@+! znfFXh(i4wJP5l#(bWOZJ?%7=^B*B||z7$QXQOC3uOcY{Z#EHjOO`>Lv8E?|T+1}|E z=eKOWlnzKyJBU=JG^`d+7s|PHuu_Dl_ZCVO zhTeMy>4dTS<$+#Q^1~(H!JxE=zzSyf#+78>)Mh=SfrCp+s`1%f*2S%kRtx&KGp;WU z8&}f1oLnnnE+VTtXNQ8joL1+^CFS;9j}#F$qIL*Sik$8!IMaEzI;Tp}1gN6+xmUnw z&s#R_h$L^tc@Gw`dRj)YSiDppFb$J&{J{$`=qyH$Uzdh;j{t3my$D)5%}4&WE3#t~ zF7zac22sd^B*vW9PledH$6|5I+{Ln?w9hl02wD#{2lX0S!)j_m+s|*YC=M67>r(gb zE5Xn|@NhnOFnUkpi84GMecQi}eTUKxJM3s_=~64N1;CPC&+?E?%3;$(0LPiR-sjUM%dyhgdY9CXXN#tNf{nJfClP0> zhy#_hPFY8xMoJC(@N*eiYp&jc!aU-b&+ufX{cneD>$I3w{AqWazT}!1C}U z2=PlRMZ+L-&oERqMzI}=^sy2=3ve< z_q7r7?b8{C^NfhTH<^D-Nf56!+;(Vw<|VweZVy$NenLg2rl@%A;fA$_8@>?spkTx- z%y_BF2<8fXag<;qn|$0a@6e5=L@By3log;Ve*MAi3u@B&uPBeu#+hksB6vp9RK=nV z{7;VUC3BTT2Z<614u6mdqfm$5z;dbH>?~H(34KpwBI;|Y=<_HJNwp_fY{?^m@Hogv zxb5U5{hWt>s;_lFy6DENIH?x`mIG?=tVy!T*lQNW7nv3a-XKO_vk837X__&9buQSo zXN5#&iJJ(fM)HkPkO&csO)I!SgL8<+m7 z8bvD{8|?&^BZ0bcP;|v*fo+lTnWIR3Xt`nZU0qFg6EznDgqg+>j;Es{ZVRV>;e$!h9vU>2lHfkRF}5b!WLF7cnW%`H@2yjMPG>X{2Bq zF^cF5D~;-*JnA+S=ES!%5a#rE$!HK4OR(MK#?#u8$NYEgx2X?!2QYlLn)RM}hCfz< zyBC+=PMikZ8vnaqS>ejxWP2%2wF(N!IFTiM)AZ>pIr8!DU!PVA_^0{N4jq^$IK9?boIx0r2Z?@I=+~B46Mj^{bMcn%|dY5p!{`8aLT(G0IuJ);WV8As=HOL4iTZ zCS%1i<10^jXF__^6Q01r#xVYs#o%do0ENhsFb z-zcjI<@{iFg7**V+c-e!f5pTU}a5L3nO zB0-CtXN#q+DIJ<_S1%T75V!mWa26;PoT?{kx7afl{ne~kfe{RPrs~i|VD~qMcOuO8 z4~~}~FS%FYw%Z$Jx7(8|9#o6fo(h5e>s9yrAKH3O2+zY(EB)y`)ybAz;~9s`2h;sMXNY z(wo{9p{owqzX8;W_vHp@#j^!Db?vY=xSO_Nj)sbmvH4umN{{-$?d+^eWv4Uoqms)4 zsLi&AEDIbG8Ns@m2LPzIvK}KzT#TJ`71|~bO``u;O=D37_pIj8hF<5y0BTM%d&v6o>A&d6YjN#-FqsG4r8X$Y0~3knZ460A=fcQ{=Kx|DwoWQ)VLb!ge{an6t*AM05CFyO56_L<4(GL{ zinpAMZvVczb7wlJikoP-nva7BR}BA80N37Va}K&#@qbd_jrtYR248XhiIPPgZa8hp zUCEB3<(-#wZ ze9HwV9TWGU;tT$mIpY8`XZ-C(EnxMm-8yi6In&Bw0eY{6VTHAgDtT*TnB@&P?fc6eSRaq{sOw8Vm=NQC=^WQzu&Cmvz}kUNVij%H+Ny|77is5#bbi0XTcp2Y9C+ZBXv{ zPr$<$c(g&S2xqR5zkWojx%ZY0s&67jaR_iRG<^MbaWm|!zwt?{;>PSRXHx(V$eTJX zd5{aPv%1yUUDJB}rW#iMv7W%=O6xW&#M8Si+&EE^=Q9V%ua%M*B!cJD$-@0qe8{nU0H;C2RdO1BOq85sq)#lRx6%^m(pRsfTfy zUc01utuUEiPJ@N6-o4WuXmoTzD$r3-UWCD$A{US1*QoR}y&Fbv6VvEVBRaF?00qI+ ziOr4}I#Lh&>)Iyv<*-!@n7+VI6K3FT&|!D|#>l*qPt}1d0Y4>NB7cfk2`;rmwSR%H zN0h<|-di{z;X*}T4gr`Upal){`;b!8(%O5!c|Bi_1UY>2dTc^Wwfj0P^9mThfLoSw z(iaEnRXx}J_G}Z1G*xe^%XPwY;ADV3e?&1Xk4I;i&y0EA1T@I;6l~q&nUF&Vi1Ob^(z({hd>jAu`yn}FC!c6sho%ha~^U$~ol7eq> zuoZH`;Bgcrxv+sf04VCX^uJ$x9vbfX3i`~CP9$bV9YPIaWd1WlERi@w-g@H-1uAlb z@{;4^Jeu`3M}4n`eT1tnsrC)%Ca7U6zcoXB{~vZ6GgPd9{?1kKbh>{3b)iI^hj}Pm zfiW|TWq14kX?mGz)jmDqBq9A?#;dcey%QKP@IP??EXX5*^gk&SjxX}{PwxuBi%xVK zxBn*(J*>)qdFc5wM^w$kU%sT4BTj{DCtpS{z{|uhwCY$ znPgV(FzNmh5p}}8Y2!=$`OZck6Od^OzGI1;Aeh(X;-9wI{fEf z#*w}G=hYtY{5aDpBHR=F#!E&=p$ep|`V;3(Z_JX>|MkmokQ`1P1bXuD8OF>a!pA)B zKKrZnnUbD+Jo1Rod^gPBKBDJ4$cj4FKkp#X&)@rhZ`UL!Al(;TH4e%b65j!{DPXnU zeFzr$sPpWr`V+f6(iH4HQ2%^W5t;x<@_)UlOy24mYs14jP@#DTb&Ii*OnT0s zLdJZsA@!SW2f;*V-CLkVjEgd&g!~K2$6XImPSv79k;UKbxf`|~c}|6D%rglr3W+5B zp;K}r8}dGd0*lvWc#d_v&toIBJhZ^Nez$<(XL#Za%_Mgobjw8^+A@Qq% zM*7bW+dS02J`c7~g9!irJZK&cXhS24ju#AzXsRqKa|_o%R)^4RZpI@ceh9-gf4Tl+ z<+n=8f63ij(L!sX6-kQm$8SfV?KV5_=~3xgp#7ayUqjIEpLNOF@O-_$p*OB=(6V&) z4pc%X9!ubM?nH>d?q4u(u<~t()B=Ow&RKFJtwL(7-NhY*rO z$oki^PrrT(v=mp`!x>IOE+FZ=#?k~S8|$a@Esx={aj801I>JGf>Rge^G^cIW{364z zhN#pJzzMB3cABL{;gD6Q)Rs`m53kj~9S!D$J^*zI@T4cyR@1_Opje-)q``H&a2I0S zsNl<8d+teoq17u`E~0MkpydG;uuq7{@qWMO(0DM#8gKbNn%%BZe~qQapqxyrv{rvD zaxrg`i6>x&MFc+Yiv)4@BvoABkDBlKI+v75sBpST?M=l&0XUiJ7w^2wZVG}vD{s{k z^?zwi+A$aolvTCnx2b$%twHUsiC0Ck$p|I{R99jDV}-qRNw9^BgE&~j#X$n>>T)6m zR?N}g%d>p@0JQ4KT~6MDMO;p#d97CvCw!2WgU{yK25(qoaziLyqhISYe$V}eh_AwA z=h<^zkI&a;W;5G#Gf*qmEzm8N zt)8@v2abrq5tzUc6zWgV)epO+=Ndrum7w=reX%r}RZ!d`;9h;@NC?5e*s-Op^2hT?Sb1lL@88W2}#|889@MH*Ebi|!lQNjTj@J1Y=YGb zuIKWL7rLq}Qlp(pqr6SJcR#3%g|_atFTcE5?%AodIW@an{hV_kZ$Wk#1|TD!&mknm zuORK;DjtC3z%ZU0fMU}@aMbR>WL0YdHA&BX1 zt!(d_F_ckb`ItpBDj;e=eW$Al(qZ0nZK^kVd4BcT$;ULl510b60jA&D?Rq3tnPWk_ z1;3n#loP#SV9D40*s9cP2WtgFCU=u%+kP?f{H#w$+7?*$pP>$f5lq(4qPYC2ghMa(~}nK3G&? zEdIUI-TbSu3to2-9<4WdcsKsNgSkOAx7oEQ%<(&k4-lW4B-hV?m!q*L};_{Q#VQ+@-{( z|FB^^a^U|?@gF^yx%<{H-c}1H^!C@1T>4xemK1;23r)^DFz&wkxNucrvqS6n6!od%z%;J~Vd&|{oUo9pv8WjJs_#GjI3Mn^iH_-d;aAE_p$^5G6HC&eP!Fl9% zW5Co-g4g9`Y{G}LmOJ`mj{^7e-hL4bD0m8?C36`Zlq!TY69d5UEiW`-VqlE&gz!nfT-stWTZS8 zAy1iN-8zyaE{DCpfo`cwlS89HpKWe9-L|ylVh3y=$$f z$6-)lHz@iRlkXuo2_kiExNZ4DPQioG=)3WQD8fRH=l5O`cu@UiBK!9Z&M2oFBe*C6 zws4g8p#KHFtzvrCtL7HFQWyZ_8%q;FzO9|4Pbwafm!$dt%gTSMB6vt!2$VtWSY%Yi~WHTRrK?ERQ+CJ$!S0P z1J>ZCX2(*gi2t@w2v}8XrI~%jI=JP3=;U(n8mBsIQADx+mHHfVotqx-ymu+`1pVUb z0+H|VhMN!*#qFNE>R+VrL_a_ZbEK1}b$~IKU)H&Ph85f=L4(L=rkA9x9x0zY&_Y$> z$WK~JEPkn7w3)pHADsqT+*H>=q>vnNelz$REqvGwDe9o8S6N=4diX9ZTuvp-+yZOEijBE)e z3*L&bJ-V`?Y+V@DVOSQB9mT<#4~bfgu*CW7W+M164$d}mbm*7tdDUwU*0y7%^?IqD zUNiuKTg=*a>KYxowrc#o=Q8fJZ71!rvQl+XH-&UErvDQ}AbF+K>@)w#c4!{X2-2v? zTcp?ANeT+Z;(vO@F;C?&w~wF@=F&KCvunE5Cv4ylrJagv%^zK;jIb@{Xq-S&4+;Nih3%YPqIXF8~>YPa8)@A*7-M7$}^(@38EpNTu z)>@I$7Ww`YnR`$EKx9cs4=I?H`9r}rb_2bQ&l7Yn(0`lCY|3PZ(-7K8E%eZZBi=fg z*DwdrJ-OzPh5tlIUuupH4ZrRi_s61=i1{0!`--Lz$(D7x)fR?LltOAQxjx_%jy-=! zZV5aS4&pMNSY~(mG9k?F(w?e990ORLWErHu`JKl`_{o3v8tw4tnWtf%YLAlo}In&X%qnm<)a%rKtERlvaE zUWk_^;m}+zpH7X4K@m(6o0OW#3E?cgaGUlXZd0@(e2QES{e7un&g9QjEt}Adb>6Ax zEjTU2N~S#3Gx9zw02>i;f)FCfIgxQfPH(m z3zZm<({2fEtZ8@Od00dSL0~Wg&z{g9H7kZnnOZHQa9b>2dgvJQXPqZOT91D+KMxR_ zeH15td>$OY^#8H<7C?0^ShsG1y95so!QI`02TOq9?(Xg`!QCZj2<{pP?(XjH?t8yY za!&W@^Y?xCzptLvt*%+MSS%J4EGTNvG3NNDR@!$CHG$F{&_Il~$#-@yP23;9hTod% zCoIl0VxIL-K(lXbzYd?KW}%|Lv4#jX*$Q zi@9j}u}tG+RaB1ZOBdBWV0?r&Y){)F?I((g`qD^!5_!+?NyOj{ZJf@3X;K&bZid>X z4qBPfAcXXV^ifzbS)qhB9esO17yhbv8j#QBr>q&K!210jm4gRO&6N}W`*As`+Q#oH z!Em5Eepfl}zq`+~uMQ-*?rsa$y|lc)>kuy1ko#Yi0f!1H!pPsRe`XX>dh1%#jWVlK z+$-a)y>Rp=O=5D3cJ#I<-q${jtL00QX*@6Cs`#!xJ%vmoz_Y1MBkh#^$&9U8_PAU;6$Z4 zPY6+P6)5Efgv6vEkGJPgv^b~TSB+L+d$ZvD4k!lb|Hg!Zi3FKY&!dSL5q^L2RV4b# zBE=Sbaigz3z|TK6WPUJ&#T}>nkdVmdvaJEGL_u2qHhCnYu6KF-X9EvM{l`)cInz^u zZN8JR66Ho!H09BU#A`(zX=Jkd%_(#hw&Godlim3&cv$a_XWmOhp22xYyl~`8p^{nN z`lCJOyDwIQeEH88X)4JIM?n2O@1-x|qlHVOj;46)a$}dyGTu&B>Os5C%^3pC`l=o- z8q|xo^VQX5c6L-osb#@`<}tIn?vZCCf8{YXn#J+2FFon4mEZb?({mGWLwq;ELwmwo z5wQMTrVyRwS*6s%?pUcbrQ`XmQpvUY)*06xv?988r?&f$ZdKk=>6v}!%PjRD;ZLnL zuOAJZtNfIw3gDtG3;IsA8foSuSaJmB#rOXvl-YXb*qrnd%FNC`!}~X{>%^R)s)g=H z_1^BGpK@YG6@lWS6fCZHqi=RL14s!eVp)}-%j;Dr0AdxnG2s_VfM+4!pxO|2PTyM@ zFT4zTVMOb(QD(LjaB=TJ%`C5?=>>;(IMs$jCqzJIqrsPnIOm=t9?R`q<2v3StXLX% z(WeRh1Q0G}f+qZi$Y2J@?{A3YFP42hs)-$XPlX(?A{Gm8fW#ew9SV;uqZPZ_`oMEm z;8XovHN!eE8tw72?exNVh?~FwcpP_|k$OjYefNj&SF(1fb&dUfYG*LwgS}ak+V>x{ z)X*UV#MID17zw$ZexaaK@m^}^eYn-eTn9*VaN8ey2x3}6R54-sQV9p22j;sc5VlJ| z|4;Kdg~hP24r_NXELWp&%mj*t*mtQqq<^Ss=N#>yLg^A;kB{#)4BU^_nGh16z-LESqiN=NRXIg(J zD1ZD;P@;emlu#E+W~K255jUMF_e#5z^%Hbj8fNrt1Koq$V(!V-`3fQ|K~y$T3W;lT zSd-xlxzE^v-WFWgS%JAsGYD8DM^pruldcpb8FK^c@;tx}4;*JI8=B6=?(B#!iL`tW z=CJmrpNf>snmbq9*VKvW)@O`T=_(g;Si^E3A^CZtfW=N(usws`40-t6Gb6&C(#(LD zfj72D?c-MI&Dq02z@4N0JZ{^&0FeF{*IsN-v z79b)LP-sVB?TkJ!WQtOs&Iem|FDkwi1c4xsc03b;ujpC&Qcp=o6iz556yFWwhoKx>e-U|c* zg4M4iuaxh(CvrwfN1LN}nZoVS2koQtag&^{k0AKdI$D5!3AfEm0IHU(o-+!jA41}8 zw;Rv=2=2&n)9z=0R|Cb!%g8xtGXX1MCqfPa0Vl)`0^uj94xl}klud*}x13FcqPOG@ z0zoIBU?kr1sqpRN_%lS}HuxTZs3}Y2HkkHSq3_ZwesofrpHOUNR5v|JXU@3NLr%ZJ zK6|mxA&mbE0GUM#4K$#58zubRv9eANUZ+f%Yo@n=lThaaQQ+v?bx%W1gN#F_xGVx= z&HFy_-bY=PmdrjEw#Syp`@=|e{N!lODTf@Up zv~G_YuOAYv(CEF~hg>L<=zu?%0FOjmHzKk%ny5#x^~J9q70{O=tqXg}_trm2C*n-W zJeTs8DpbvO5Rlfd^iLcwtiGN+R%>*3xSQ=?E?(y7|05*1TJrahq_0=2>Ip9)NveMdNoM>GNggWxuY@FduK%YY$#R6t^`I1^;$7DP~6aSC;UT(65iD zf^I?M)l65s6Y#`Wyt##*AI$MQ;?9^ZLl!EmwTTM|G^?-sEo(34;|Mk@bU@wU(J;Zc z&mHsIA(Oqdw4Ze^KCR2l#0cHz$wNxmpNs_=p1ty0oJU*(3SIPmS#DUWYqaWBI3eC9 zw6?dLd&GEw7|MA4wkFLH61sD^g>@t|{MT(=X1@dS3O@cyQNf)~#Zn23AHv<)+sKQ` zCle$51{oHRtqZljlb&1r1u_IpW-f9l+nIg-lVMm8@-I}O41*A;i;*~l4C}e;(V?Wb z`6)(gD()w!E`w#D{M1`~`X{JJ?2E*V)5Q8Co29t=BReH1JCLyi+QDo+-@o1gyJ|mc zU6kUJj>QP|TK}~^3TSTZ1!LoLPWLzv?y)KThe_1i3<4kG8ec@P{H=GGT+Z*iNznzq z)g@X_(d{3F8icxk$Ne~%q2(TWRe$Vobvc9w4}>Iu+I$A_d~}l9$~XDk^Vay_`zPTB zQ4@NfP94Rq?pO0|T27EB_LR#Lrp-A)XzGX=x#D|_@@cH8MU1f7#Z{y=yxHdsR{S*b z$L^eOp*c`52l6iq51*z^x}NpCoQhxvxiF<>jv+yw-%ANI_!_9zCpo(R?fV58a?ta^ zG+^vqpcs@0l#szog)~FNfZD_A|1zsFr0`d7afY$jO*YLZmia~+@)nk1z*|7iGf&Qi= zClD51l6MFWY&uKh`zqd%IHwM+4s@92kRnE^~@p->d#@d z`BIbHJ(U|*^3|Ne+6Q!hP8br3R-e?rx_t2~DN=V;Tv8yyNJxIiq5`yXLi*mRkhlNs z8`zbtfjmy}E}HiI3(g0@ujddf;^o3018=|FTImWYmiEEFw2YXnS^_4tEdLr|GV;?* zg-tZG|6GZL0ctCMmlnsryYC^|vR0mEbTHWB&)*_s>@3E+L5F{(4ejeB@#EPRO(hpk zl@t+0T6%H~^BS|<>QSveG4~XVlZYz}|EZ~ei2KmG1w2&%GHCGUpI;y+;&|D&bP>$- zuwiKtn2cK?h|fuKCV(B@bm?fm>%rr9Zkj6-=YtygC%~IKIHZ0=qGicEKc;e0(M@LO zh2V`u2iklB^98B`q09z^*+jf&-9S`fUJ2W`UxpYfhxXVb>Z2?Xd|_?w!ft8mZ&T@b zY3bc(qv9R%2|r}f(Y%F(w!X7%^mz3Y6Rn9KQS*!8gIRTJ%a<_lhCN3;o|?}6Y@iXU zL{?0TjWL7AGO9V(*Vg$r5_w{UZ@_&COSRmtn!N1Att>UF(U_@ogHD(=7aCT&6 z@1CXtj+W{WMo*o1yq$Wh?#>z%>rShUjWxHoSin@8iXY@YUqhe|#BR<;xNokS3Kde}T{R=qWUq2$4G2&4S7 z8E9QbF7ITE*Rz|gZL8Ybrum=ci);5KUYfvVA%NjfA}AJhbP_)A zyf7d>YT@c;O5bosyh;c2<-wqEWSC>-D$ zP--nxl+Hf1k32#AAAiKTz9+ylEYu70tGHDF@v5*+lDih=P&h-fKhY6 zQ|^oTb|ax3{u#>wT3$La;7^d;;JxQ{(*M)TizQ&x{A4#X-GZNZl<$6CbEE-nw?;K3_*tyfZRNDtu!L@Q#aRZek|D(p z(+ATV8=yT%KgMZCWdln*1zNPN`8%;~odIT$YY}vdUL) zX_(S5iw{d0jJRQ3{ZJTUgrKlSP)OTe$xwi!E%wL+2b{BGR%y$9=-ZP)buPFu9hY8_g(5bq?%)Hr|$dNCdatdzqZ?Au^)=@gls;*ZnYKkd- z3-94X08~EaM+;8^YDw@olw<@4r5HTD)hddfin_w&P@Dh^P4NcGYlN^W8S zl%LC&p>2oElrccaJ6rgdhAp7%mn24$c)n^uEQZgU@DX$PUlo-hF0|&8}d5%s!?k`ZJ+X3;9{w!6{*|;-Q{pH|9@` za=Vf!8C(9J@q%z!Y#omSHhtW)`s4z9BTW}EQIWTu6)t=+=T~>fUuJRfPmNSsTK7C! zF)^)$4R|An%Uqtv?-3$6P%=jAU3niz?h#Twzz&OXWq{(j3vy|%eI^kyJ49+tTQQ%| zwz@l_)J;7{?7aD)N^zOPL z_xx=6`40(UgWwz$Cp| zO~AhIKGTcmfR?!DK4ZQzMR$b$t_Rh-Vnw6WaW1HoIy(5;~+EnY!w@+r*`RX&n4xKWD zG=u60f{DKkQnp375KgW4P>nD`)RWn>On=!t51QG8SN;(3R_wD`)Q@)MSKgD6>=5Et zmlm&|uA|Dn(P%2kgwwqlpj_iqXb%!1tFNc7=FeY{S9}t4Y#>y~59fJ@yTm;WjuUt~ zGgGv|VRXJ4a0}Z)n&sT*Z9nj2bzWbfb!F|EQT{QO(zP3HqOYpW3s1K3!NX;Ky!${Z`UH8FT(hllWy8U`EGAuBJuv3Q=n&M~a zXfrHnTT`q zPjJaMa}*Wn6z%ncO4~nhH#4024z)t+dig`Ha~2w^I%Sz$55_X*dbgrq{C3p0+>HCu>MYAP)MguvQ4jfKFg2Jnc1Dvf!p_bv;fQ z-m6H~^XOT%&va=U5^0oj|HkZJYUQF7*XGDS>qJ>r5|ajsTLQsE>gG?I7u*!*-v>5g zukXJ>8r`BC%(IJ1NjaI@Tda|cv@$R+R|emocAAmjpZR2$>C-3cv1G-YID9?-?d;$K z_Za0$+_-B>)urBoI`oAeE_UGne0?|4PWS9HRg2-02RpULLW@e!Se=IKoNFaQhUFi2 ze*~^1&&5q*2FGN`LRND$kUr!f%jvP|a4xYL_9wtT!%DHy$)5{=PHguiY49X-Q876T z5w7dlNE)si6~|M)tgtoHQ=x{`%2bm0533)i8_9pMp_b4?SdH0^?$t4#o6nJNIdlzz zV!s5t|1063I)hV*#1dVxYD>mSLNncM-pNv8O~z4X%OO;Xuh*p=S?`n}H&JR@HGet! z{So%!Uz>Xx>&_327$ao+A8g_8jNTa-rnE;&&%`L|>Z4eF`U=>c1M)tuf(sSH{||h7 z)l`sgpNjYc_mD=2_g1K>t+edMjeXaK(#M)tk9tw5MlxLgw zNo*Hh6^&G|ffTpsv6g1`&Ipd}8t+OQJ2`BI%x|j^SL`nh;Fp0- zPBoHN<{_aXaB=*!$@s(#c9x6z-KjDv1^x!SsDHThs}j|^O}{Wm}ou`9UUu9urtq+_nAlqY@E`uQUu}>`a0ah z>$?H1FP|rVv7UW=CItEtygJfeqA)lxIYqm--&0;Bo!2IR}{WJ!s($_6F#~xd}-W&S3Tm5gELjt zb7+ls1Y`6WnrXj4VB$-%q=?1Y?U-z0wQh2J6vX>~!g9e5zqQ_L^V76y-~XRzHbuGr zqS@%ihRxOr@ohGoU={Au*j8!Neo3V?BYHXHd@CgsJ(gW!*vT?u)BSI=Y;?Tcxm7^D z3bf__);@mKe|-G^`1t?v@j>&M|MBtvE_X-F|AkhpK5UuxC#~44I<=-4mb5;L zpqu%dOPYUI<$O4%DfYi;Scd+;>|`77NNMXj=`mKL8;awdep)7~>PWD6?UZ#Mjq>tm{&%b0(#9w~6fqvwcpZv!I{#(=HQYm0msJ)+*e2fwvH}`easXRh1wn=G zRBpOoD@ad9a-w~s7}RockAGT(6+D8N)Z#?UGo63C=m_xSX)#prn7OI9_(E!#!@E4b zKOb_D9(hqBHmK8WR|_H_+CT1O|M#OBP5QakljK}YF{MAW5c8B5Ri*OP#rs{sNO;a| zo6l9)XN_|^ifQ%*b6w=bb`1wc!c494`@$@B(y1AiFnN-(v~LT~5^XO*8jtCv9?ttk zJ1*VbZf1wNZM&RFp}_E^k#KMzXua8#Y2Fy1^B`Dl z&Ul?6n)dhAzN)d`dtC+HN-)O!k&JX60Sp0WA@&EUYE=-_ZbtlVUWy>x47G0_QzP&~}4;_`tJ zLZI!z4_W%a%Oh`Oh*J7*dD~+;97h5w<4-ix-VdgO-|1*sn0pmtVKu zi?x5rnlf{A&Z)XHJvq@+4H>m2$|~`n3EPJrhspi)?(I}>bWe;Jj;#U`^YYy9s0C!b zX|N2sH_sc6gt<_W6HN(GF?anvL6~b|d_56Y;#*y@XNe_g|BkpvdNF^{Spet^o3yH9 zr{EoW=8}R$UD5C|(-fqAOwA6`>QAQfb!R|I075r ze1O5&))$u>mta=aa#kg@mqcuYsBHTo0X$s4v$Ua-WuAT04!{$H`go%|~y#$en6D|@_( z4c?$fe zhKM zupr3N?^;^fo8(c~PAyC9%!H~i#x6ENtlZ6l1F>j-*An~OuZbJu3D8zA!H&^9G9aEB zdmqL4?#R-9ZIXodqO7>U=uqXE_L?r^O50Z;oKqu4;q6h3-!1JsIT~7VMS{z=S^UOr z@1&0NCIN|D?H@`?6j6VTV|;}EK(93`b{)Riq@7;C(B1D<`6Ao+z8pex>b%fvd2HK~ zz00Npw(-Kq4sZUP+W(p)SXHt(*!1IX!SP14LjULEeH{>fldy&uv|Wpc8HNNNRKiPS zAs{+2C>sA)E-z?bxLzg+%)4J@py4N2MKDhmHvVE!PgP+4|9J3b5-^r|@kf|_kpGL( zvi0j7ESjuJIo8q2GiF%K-x&-MRM;OPs4_65(7PCk~aGQEmw7rAb)k!KvN%6NZjRsMZ-VE{(Gv zNc9Vgu2O#m?(x*rbZ=475p{4h)|pkb-d>n?;Z_*5U`_5e!9D@!ADh=WBM&WJoy=># zf)IQarX5QpfdmWqvL;VtC>2BUDbI2~GJg)$kb_4}{FMEvPmr`EHt;{MN-O7-N#nMqEEm^xgq^Gd(@l`kf^iyf& z^*0-GxZQ7KS$FM|n9%I?uA+uf>{vSKXc7@1JQ1nWe#d6|E%>u4f;UPe#crKy zP3zHXr2i>V0{UZ(scZ&%46h8ScM4A+tBFouWE zsYcLvbMeuiHeDQLMvrgg?SXAw!iZVlAwN!+wJ7x43(DMSfn;(qZUDPKRMMEb$FV0Yo`h4 z1OaE=K67pt1OaP&Ff0BhEUP$u92lDfk9;8`IhD?4K=0QGH7;gt*Kn{vCfA_JCzjKg z3XV0doi4yoxy(6K=_c1jNmh$>e3!Uw*Y0uPhy1NA$1ndh>Rm>68xT+*$&9bdnx-CDk zXxneG?%iyOHmhi977vP=S`#l*aoMdum@VHMc`nn>@O~?!rNe;l_9$`~EGsQ7Q~K^` z%!E~a^E^>3JB?a>X8m@YCia3XS&xRa<@W4L*H zPNydtZO8S4KHj~a+TNf^8}!a7Th2Z#{z#>66SmPVC0+cbvy9%Z`Z&Ag?h!hG1(UnWBo0|!z(o)Fi1rz6Zx#oX?#P7}#*pet z3#5G_hjst100h2TD^Lu)1K8}ld9k?CaP;U42V$4I;FWJzi{6Wy?BMPgZ{B4D=w2LE zJG8uoaW+3kenYKykmMQ7FS0=~Ke2L6`dUeTASu-+KHQLV8vP@5(!^QKNBOoEsUj6K zaQR#Qqb4-QS4igP&+ojFh`pk_kg+#{C7!H=pI|}>*}I}*?@EZh(z~K?HYg;YAY2ol zlEIRE8>lGt%MUYA%{;;Gw_nNK!q+O)jY z1zhc}c@gaGer*vFcFhWUoD_Vd>L?cQd#wIh&VMEV?WKYCpnrz!2Xs9{_60(oMfm{1 z?!?=HBzL{--l$J;Yi$sK%{3sxod7&w)E7u*GPv6gx()Op@0|NK2v>gaC)bp-T{+^@ za=}yq@C61CVG^oo@cnfRfc0&q3Ffc*Ei-#3=g; z#1Delk*d2*=p%B&)dg+_v6G(YNo5pF8L4eD`)+ZQmRC-RFQmig;F#pX_izglbX!l3iOl5E!;f{Fr( z(|333HRKDNN-OZa*PT;*+xl@$g2iO3#_96>Hu3RQVej$_>+CkU z@m0~b1#6{ewTDNHAE(Mq4F|KBeasEloU}Pr;o*yjw?d*Ek)}bS4(iOBlo*KZ%}Z;= z9GBC~?-KN)THBK_+6(ZlU?RVrDJnr665^vObJL~<)Jat2fSo9zazoW=tbUhTW-AE_(|s@AOL6^L^uOU9LvTDt`kZV}*d@y{Yh?;Cv%aiSs}WOCF*y$wS;a z>yqK4jf|~W03C`Oak)PUZ!f-Cqt;hL6@zh<3d-p&Q{`Z7?NVu#XzeBOn1q6gp)Ez>6(5h3w&GKew)KWp>`2CjnZF<9QxEbx&ZAjd*&bwy1_a&e z{{Fj4h~UMKw)DbF1RFN+P7Dp6YXzZcrW;5&OEo)2`b$hU!J#}zvPh+IYm6u8wuSmj zI2)WB^Tr3*`!cUkU$=A`Gd>xKTCPZGkzeHFVDkaM(qB zdZ>}8o8(726}_=F%~}q+sFCE^d=;mUo@4LkfyT4RVtd{9u5hqBU#gBpsXB^E(^`+4 zjVF|71%#!(96yRAB~R6T)T6n`WaZg-mVCN}Ly^Id_7q~1FEH{iX}d&ra2uYv^CKvq ze@n6O;l3)DsVWt{F&bU1R#|L7S?jm#87^Nh)Uw*bglrqp^_nEMT zGFV7=sjKzvvbv5$?-HQa_WYFf{s;k%22F&-i)X_*7A>IvGYd&0gIyW^ywrb3S7Iz?IoQgTd0-2C8kSCYztoG<2TZD43npCx#tz@fQ}EXsSJGz2%DrN&`8FT9E&O7 z!7KuSIR40ZNKYPXj6?1oYc^`df=&L2o8ZCsKXm~kc46XxVsQ6PpLAVDOoSeU9>4bF zcI*qu2JG;9@bKZ)Osd4g5jzNPzcrW#gm$R~p)1$m@jq*KTr6v)dtl*xe<*x}&M0F` z4nAm?mC=#zALVLLQCKY%v$8u0z}Ly{9xbc|q|_*~W+n3b8_|!dlRRVnz{9YjLwh*= z^j6}Q@QzkDNzL{_m%-?{BAMH=5?|7B=&6duWAF)Vf3H~}vRXU;s1HrjdYAh1ZPw!R z(PF}9*D(W2O`PdCyk;-kw{1kdkKMH2DznK|KI#iaK-Qk1P7(Az>pkGiGSU`GQft*4 zOBnsC$RF6TuFILT$2jg+L2S(|H!wW}J(=`fnZ2Lzk8Y>hH}pKo%!G#)h)An+M2tX% zIZQ8RFaxEEYj8ASWrAvthy$XaPt3kg%;sto;QsGJEt@OJ0UBp>5_rA1JCg=w&fVIDa_5 zALydEW;nk^NP7@=IGN1Wygg&yq9G0G39;P~InU$uGFMHHv%;^4UXSGuc;GPsfgpm(I}SRo7ocx0B-V-2QK!`Q29s~Kj_u-q#2&q;&J zZ(78NeXdq&X;rgpRimG&+%|8w&+=$>+f45m1X`dx6}Z5X4t27jX%W~j6K4w_iKLr> zA0lzIH;IIEx!$)N{}5+`rdLW=yBAs3>R*lCMHyd72}u&KVW#~dO*N~6#gBQFGqy1? zLV}Jch%4=YS%X#Rk$w)vNoZQfy~=dwW_w-%D2Xwr}s;$MDC5TvG$Wp*iJe9cdY|6qY%RKJ&zNE_;^) zYrVbE)XviSE9yceU1oksiprIa@nK5KAUFsleKMHT`a_KdaF0jQTjJou zXl=8JivZzAue5;>#Tp|RRz}4pTgo#Y8*febMlmt3$1uZ3C#xb0nhLF~>O-277IjUn z=9Botdh+YxI!V;;M8Gmv3D>P_ju zIDaE5k%S%|_s!joF=s$#>+5iZ(50K%Ecr@ctiE3o)N&Fh^rYmFO&rCn_UteDy0A8Y zpY!!?KT!mAQg)EQeql(mv~TX}wK8T|;}03KjTfm#W-4pf;zhkV8A^X{DD}9a<%TyO z91lbWXyKd+!Du6dlv04+uK2Ihq!=tr(MHhybW`X^3YN56zsJgBkl~9sxN#qnZgz|lm5wj}3O^bH9Bb9C& zUZr|PZ<(&rM+{3U!S1=B>NFV6mAJe60gjv!^q-8aB=f-vg;gFOA2AbmD%VR8$uN2q zB>?gLdnpQt@|YnN{?i>tW2#W6TV}qsMtaWZzh^EluplnkM!&; za+_>Xk3=*=;L^=3WbKJzCnu{B{Op3g|so<5!RO4lruf^kGO<@| zmt3@Qe-xDsPDq*`^G(wU&es`=Xl_CRsxX5> z%WcKe&yo_}NGH3#UIYWUtns&|4FZTYStKnFqzwY_PY2*Y`?X%SuL4j&qPr|;Aj+Ng zDI1A^1LECj`_*|6Jb`Y*Y6nuJ0;QsNS$<|E z!FmJGF92|P_bl1Z%r&Qm+0C?ygsL{n%ESOrIv9d?@Eow4Q|ibhSAjv&qGhN;I5_SB zfeIGGkCB5Mu%7H#Cbj-F#cTduoSbV}+;-pX&1BL2Vtg3~*Rq&KQ|p%7N`zEB4O2!3 zib4}$%dMvXSE6jUmo>wy@E#%PGPI=p1gTeB`anPn>OJ=GS`2xS++^*VfS^`@mj=xt zPb%=5Q!66ilB;l*i-xWHVXqU(nS8h91SJy}%HCa&9)3a6Jy6d`hV?)$CB zUFnRNf-I9*paCZ!Lg`|5Q1ZY@7|X=^ooUq_Y_aUiw-T-o7rlG>)W>bw;vvbL&qZF z!#-?edW9f~20)aBddTiWz$m?}z|UUfPW=<_j5P6^`@rc6CvAKr=~|>)qz?~uBS<5P z&Y}4@o{$ubD^etCIE#+8E%{Sly{parP1&=D)HSSvS8(V44zAi$4EcF=j^mM}VvP)} zL*U|+Pp&1ahFp1)IQrGyyLDQ+Xc_-fUNVK)tU%58Z43SK&)jpb6+rTgJNh%<)h#s; z>fsUnS@Wv++2}>d(H?*F$`32{vh=0(2Tk~pP!IMsEZYz&I&3Bzsfg5*O$N%3qqGbn zwz{!s(+SH{-R$9dsG?ddtOu=yz{wekRshCPZSy;GBu&B~XsFk$Fuv;R3|^NGD6TCm z$j9GDP&ch^O5t$Wus1n=Cuhz(oth~cvX&n7esh9m6YRm?e#d7y<180^$OC!%BS@=x z*e(#Vfgvo*DirYpz2JlF=@do0yEV^1!u_%T<7@Cog`_O@0HpYgw05^bX&&JP#&{Z)y}&fe_OAWE*=PAPXO{y7!e~Yj?#Wy z_V|?iO$YOv73Mcn%#i0QugA~)mwN)z3x!#FzECz(Q|OESg(`T&_q8Mxqp>&HcR0Cf zDbtj+H3~UYN=<~+e25sCK2J6b@VJqPT;HEC4H;*O2oI*V4!)BzSG~pmL}6%eIlaVT z=p>@=$zpWB6I|mBi~8xj)?q_ja=&69(WFU|UI+y0fS}lnLqIir67=iz!rwJ`vhf>4 zVNXjBCzP^jK*ZQ(?m8w$s@J*oTZFbsZAc+MiEfuMw_@+X2X{_x*j(Ehds|j`4*ZcMNXakVpyc* z40t+w0_dYUXtyOmFdlBVJIxWj76a+ zS(qFec;dOKo(D5xmUP^Z1b<3TpB#TmkQ|yhyRnq62%Nwh18Sn<|zzb$?`+bLi$(6B>7tGx`I4Ve=fj>^i^8Ha~aT*tW z(&DR~h1u`4(7XPlkDvaydhuf>yQwjrj$QbB2iYiVJ!22@GJ=#XhsFkney1L@8;~3= zBVA}jf!c*@X=K2kx@_Ae!eXIqA-WQ0B;Fzr3eID=a{vxTMV0vmhI!Cd@a6Qb&@oehX-sj9Gt7h%oz(X zUSH?cbu(8dqL}I-m2OcMk-v06E@%lHa<{rgg$6aJg^ z`(Dh@OF1dA^*-bu8Z>(EiWf3;-nm0P<7|(9)yCg1{MZfzehzm&^jr!z(Y+Pb1bUfJu{k5Llc&()GL%R@!ayM#c3bh(27RV7?(Nez&o)9QP*=k3=Z zWHaRfu|F;#j;Cg?{<^?8)({FYc-p z8!ePw`Z-(I&2Mj@)~bAF%Hll>mB?nO%7e{JXkOF37!Phiv(2^EVk@rCu28r#!^a~l z7x7Wo{XFjw`V4c|F-sP=qfeNb!4fOjyM={_=`(Wnv@;e|M5U!Ol1Rwxjli{Poi|?+ zpOr=sr_k5QyI@ga+$~FDk;YtJTVRloTp`PUB|I-UmHk*oMv^>vav!KZZ&jbOcwE2% z0@za~24(iuq7I+!LV@Nnx=z7dC60#4is#u*jcsr_CV@I?6)@xK4v01Q% z`C+mA&z*cK@0(e=?Uq#|*W^&}gtCZTBzZ#kPZ=NnEZ&8^$z~N*bxJPVuc${jLCQ7$ zbC;tp-Sh8?Lb!2oqMCSgSZD)BsJYv(nGuLd{IWASK`}P8?JAWA%t^}aC^dW!icHaV3bIboCc^i;?X|rrj>mU>yML^ae*my=&hl)o zqDD?BDW&mui`AcHklmc!bZJY5WW$_-fH!;5Zf5xW7B(BSr3KdC25qxSb|2m#^#86( zEMeyI#mgo>{iu}+Bj+B17o}k`aPE@xyZ>`0w+hH*7O@NgNi-su;e9WaGv6Ho(s0cE z|0%ZjxF~E*R!X~Ol1Antng3*-%BECb6)nKndhSSt+c%*JYJVKISQ0o zqe({>3)hf^+{O$g)5m_dF?=|hbwZ1SDT5PHIr|TEKc!3W=$=I`S-%>SB}1ZOim0II z3cb706ADPsQS$pQ($>rR@$FO|4^=eu-+E2*xC-N2=l_Gfw+f1D(Y}6zLvRT09^Bn6 zSa5f(LhZcMTTYp>cQaBK!R7+vj}e)|I*s_hHwpqE~mX>W5X-sx`+P z<2Rvjt0~{Vs#$-;rN`d+?iBq%wobefxVNu*O{duiP)qz~BY@Te=ryRX|0QQwM#vUE z1XGX~+uLYbfh>Ty^H(_|{tSjI9HpvtO;$6KGXYmB;4;@7z zeab-NMGYh5+1wHtjKYsNKySS}vi_4D`o0?bnHXPw%&OYV;|6p!4U>?7EE{U|@ZC>` zsW#Sr@@Fb|aefDsvOcwTW&()FgK8gHN@u_ED;DC(F2szc+s zvuR12{wr$EgC%G+1c_}+@k6l}6ZKk6!UPP}ngC9&;x^jMDb>?E)L3fU%jV~yF82ER zA~8PGi?r6@dtl4G1IPG6dU7>UeSBnwxT%2^chhO-D(xme?0P6z<<%BHY^NJa3(TZR z8*7veO&|ZY8%nyk(aFs73plpf*^DO$1MS)99HHRObcM{oro&+*_$gU0bpuhe-(gEb z&}H>~aO6hN_;2`l>A}I+0O!^{W)-?}u3v`y^$J6Y?VY57`q94PNLj|wMzM%Sz+r@J zDE@oe$Aiikj2grq>u4MiC!9W!g!8%3JT3L=Ja;gQS*?<03e1T;!y=`!Rh8=}1ro;G z26eqo$k#hrz%S)hM4t~ob81>R=# z@}IW|n;k=s&yK?(x{SGh?2;jOkKlWW;A^)j+@<7zFDV$Tu!H4i_~2ET*L@l}vMzvp zc?v8}K2GWO5ZJ%#@q2ij4lFdN1$9Suzy%#b^!{RAB?EuB&o(3l~E$uw=v`82nhW(p{Rnz0wcM9#m4Aj;db0d?@nnrAl?z0V!_!3M)wvYk6rtH;L!!o9K zgWLI7J9)axXp`3UxE(H=*1o0@A!vipmA2hVlT_W}0>49Nvl27z1rE zzg^#Iw{Qc2@hli{%B4GO8hRc%F~T`A-gm|e*S)o1y8n$G<#=|3h+L&-hVg;eBdABd zDz6nJ0bY_D$H)#wjWcZflR68DA#dG0JJ-`CQK@nUI``OP-ib>84gT!LgVZB~=p%73 z5mOIx^zCxMEh?IzWe^d?nTF^iS1=K4k9^E6+&SGHGZX`aow7oo?i@ea%su4!Fj&Sh z)|b+aH%rLik)|Dkws>7gU?+)L1ULf_(?n2B8gu+qaSI?lAp#rW^v{ogoi7*kjJ9Fi zD%W`IG;jeo;D9uh7I2MEwaro7jKB+Tz+L#@T-ziFw?E$b#5XZ6>f#%4E`WGb;C-#^ zr{)EDW9qHvvF3&J*g$#NzeWW4aUdP zLgy*Kwvm+=?V4JK9k@CXzqw)rkP&xEB(VtbRDsB)HxqvKm=kqcW{Roo>*V?2J%f%s zFmm=vow}kd_#GrOm#977Y`6cFB3~KXP${e{r;}!0siFsdzy=hWAhi0A_uc(2!_h zD5Of>zBOs!l&t##+Blr*@PAE4MXV|q{}PDQ@hp~{U)2hkR*;WG3DL_m?-P(P)UJX8{d zXAhc_{M|BMp#PK|JtcK|X|;cQ~p|J0N5iu}6x2d5tI*f+!)e;OflukLe`0u8Q@b7QBP%}%kgBuy`&hc6?ok1X#J3clicQ3>p< zAP;GoqdFU$RkXf1)KV#_BZ6gP+o$BW+FDv6EbWToH*W6Ko6ut)>g9fk1ze^t^2FbT zo8QTId4L~BjZxKv2Aq8qyi#x0rhUki@F`I0;Zm7y z;!Q+KilN)-yHGOncGaydjVZ!+f<`K|)wurtU!~c4rEf&@3bMPi|R@)f}>eT+6c3l!yG02(uEKKDyP%sB1)Yd6>>C<8Mwl|4%{9ax36S_ga|>#aKFt#wu*SRWb!|qq6#Pu@(BN9O)-f(WOl$eh*Q3xv(p`CPQDyV z#`H_%!P}h=3nacBbbQFsEzD!`@_1aOFJsy2)!QTx7Ux+Af`KrLAe&#RL#CbN@94o4 z%ZjGR+O0!`q$D)*5PAKs}IK7kj>!}bhW-d+(?z5{GB?8v=Lxfe9b|Ieo=MbmfKV9HIUEn`m;6Gj9KV9HIUEn`m;6Gj9KV9HIUEn`mAWWI;rar}m`LLK7 zc+7R*X6WqdAm~K=5QC-~W(fc6jOKI8!bXvfBE?6{m@lVA2ILv?i0JNL=H)`gFf(zX z={A+SZy|97yeX5T-ND17Gu^=vGjV4v#}c}N-g%M_Jc~KDTd~(#edzVX)#M@Zc`_-g zk@1gSl#dj_BVQH-PaiRbx;^#x5FV$?h-l&J)@BiS`{IAc7* zkl4ICy%F*}UmodtdfyUDMq0u6!T31|7L)ouOSX!kUQ$Qbw}LuL_}()sap+dh;R`v%y&Dva9>T!z0YKkU3Gf>?OBA1H3GhXdEIH?vtyS~>2ie*luQ z3N<#NCzNpMyg};v)F^etL*8`}NsqRh znGn_4jd0I8lquTI67b0#Hp~Eni0(~^KD!zmn%O&Ex{UlbKoQb%fa%D_XjXi99r+I+O3xYs~!3QQDuD~ftp`r6n3(a>Yy&};I?8}8myV0gg8Z(-gBkR^ml~&;!#AT*k|qck=xn1=syY= za2!n#UtifQn)o%sl{KluV_=et&z3_F$M z6~R^0m7h-zBrfo|mS=VbDa#;sNRLu@fxKx3=h)Y~_4Qc#nOo5OB;~#bGbNf;(xq_s zCFEqh#WF)_(Sn}|Pf#Cv<>h(_x|GvJ_*agP&g%!Qbsn5=#(lBNb+hJ|Pa~2prlej_ zba>{@vwoXge#`m05|g>84g`>eG_^`&@d+*MUpoJ&Fb}0^c!ktmvO(yiDA$YX6a-mk z;PW(MVtYinB7z&TtLa(VGGa^x%MNertCp?B>wnAIv<+q`(joG@n9H188hk2&(#

JK4EM1ycaNJNu1OA~dh z-+ilKAfS@ep0~77`Fx|3eNWWw}1Uxwa1dSi5PV;r*GRE^)6NzCOV zAbJ|}6~$x4;PQXUrDn3gxw5KG!G1%8Y@QXN+zbZ5bl?Cp+rNqs=Wzc}Dfb3|tiGx= z=;ozfZ|N~c3-aXClZP)7BwxfCCH%F9{hr<8pr@W-1Z3v(Fw_1oRGOK|-;L70o_4qzh=E?l3&cI>Lye2J5-5_iG0Z51EPj-Z1^Oz7IR`D-y zz7b= zS?GV8UQ0l^&GY%A2K}k`j?Ca^uAc+xKETH-8JGT~|8UKO49Ezr4P#f?ivvq&W}4s;spvV8H@`6 zbO2t?X-rqW_&3xEnHx)!rEgs#YhoUe`1F<5_hij8t^42|bCGdrCV0}nk`|JOy*{wC zoh6X0*L#Q-Urqm-vFil!N9NF*jSsOq-&O$=`f!|Zs2t{;N`1EDw`?Xrq5{Zb{E4mH zhE@i$cOselaH{{W$$rvggWG@0PvpYw3=dP(AsGX)jlb#`_rLubuqai9j+ge4L@VI) zziO+@=d&uzzXvx64krT1%65x~d#m!_b}h{&swNj)jXbHaa7=%n;nK=g{8fTcUY?Lv zo)GF}d8c{0M3q^-!c}m3E{F%&POaModtp23m)NiE47-QIi(0VQ(bO$l!GLNcrz>FN zcA_}y9g0z)Dxg@ecFwuQEmq@9xv`B}iXg+NJj6ch9|_g^zNVq38de-)0vbIX>Zq5R z$g~A4yk8>AohB7F2!smeNc{P8eFs(vMB6UhK=a4NSpR!%va1dp{FP?^^UN{YG$@!8 zNONj0KUfnHd#f+`1Px@Gt&9GL8Z-rlS{o*gqqUzB1YGiD(PvFebh(*z1j@&obD~|G zU0iBaiZYca`B_}~^fk&aV;LB_?{TFuu$7gib`ZWoou?Ro8^h1IPsP5k-mwMIw+nWj)39RY8g3~#x}$8aTh~@TVnJ%wPw^r~g*T`@iz#6T$2BvT zt;Jzc#41}tVfD}}E{nqc1y!D6lJbYoDD4;oW&=54sxSIV9p?1e$Z{PXERs1x6ou>Z z_td16%4>VhPx1a1);7w)-vB6v8~W8Q5Vru57QC#cg4VP|{GnrAi}sG|zU^v#ZYW(? z5Qd;ZHx!>6;%~@5M%_(`S5?B?4sVEgLi<*Vo?f7sCN_$mW@*VE`?ql9F2-9Vx>mmL zCZRvDwsZ92)kN$LRC}JcDWVbeqGr}((Zt8ZQX!tWP8t=e>)QS7b+yipWAp;1^#BGAXOy0ZW)a-F7*aOnrrk%pLe_NxizW3|-b_ zuTq|Nd(q)Twur8zU1ggO+*{yIhFDzCh6fq0A|c_s3w>2c6eWtY#R+NP8#OT%qA}lz zSy$!%V9;GK?*5K1)NRYfcVY=Kqcjs8_dM+`$uCjz7T@KWz{PV_VO-V{Q(ZpRU=zhv zJM39OCUBi7x!Z$qZyk^MZizkc!F}o?81Ck-09moy&~0MJz(7}f?pyu7)5W;^X!LZ} z?^vuOLl>t~2CTeRf5(EY0e?`j({AdOQ?vR-R#R1t3f;SlyeazTy1FLKqS^Fb?UBVB znQ|U>=gZkc(4Ce&ng(q%dZSi~mQvmU6sGD@=X%5TqKw5=?N6gs+n+0JCkaNrIHGb% zyVT)!TcPIJz*q+{}G-=%2xq6^*V zyFnr#P@&xo4Csy1+K<_N?ksZF$MsgY`~hS}DwkJhHNMvhwjcMf+o#uv1?DRtgCg?< z#nJ5U84nTmI8=7(Fg)^gQfaTC zWMNQ|-nH;ON^s_98MnDeNOdV{y)WQOED7r>%0dL%(%sV{Qcvd$d4@%ZlFJVjAi0h= zTqV6Xv9~$_x58+G`awjbXU&LbWaOUnD1y=()d9D5Xo4m|ndE2aTN>v^Z?&N}^0UHP z%iKN`Ucqf8Lhvm!&S>_6H4s4(j(?9`=kOx{UttvHWr>2x7kukjUFIhDzIv|xJI{&_ z&)_X!43e6AqUafRO}AeLMeU+T?c(QhBOY}2;MI|X5%cRK|kpRqaw!JmP-J=9C+8W9XI zx0m;MLAwQp_pEC#bhm;q&xu*|Kp+FF1NiDCc;^Le>ki>`*SITF(7&Y}-255AtR3OV z4QG!5{+9%%_RXgtFYhgOuyOmAmqs%Soe^|Pf7o(0oY{VRHX>m^pN4kQ$~;DmuC*Zcq~_snpnoznBlxRbtL_R{TD{uVh3E1PI;=sse{i$6k*Y~6iq-)SsF zj+PcY9)tg3jGE4lajxfkiKORX&iZWtp&(`i(d1^o=vg7k-7~kcX)chT+!Dlz@XCpQ zyt0qsS&5TOA%~zq*uU}n{UFtRv@A1aQ|miSOG}tyKNmJy+d<6>09?TNTIc?R99yFNzE7 zz2Yutr}LO+CmDVgf~!g%U2&`HwoP{dm|nS3k0oXVZfF=*%HTTd9PDdmc!+b;plwtr zegP>i8TilB`n2+caa*Pai4)~tt5n*NzILdzCp;61dm%ry4z+tf3y6DRJ$-4{xU$$k zQM+u@+u!)9r$s!xHoI97G{nMU&A9}FvCNf^gE?t6=$Ng-269R@8vo@%CqOJtF;Zlj)yl2v8+$&id3JO>&ukYUW09(A+#;-J^yK#N+MRZ3f;D zoA7-WI8kVnwp{eK-~C#+vG{{2rCap7lYRMh_ z=AB*UmyHY;gk{IS#`cmluqUxuv&m=|AnYbS|?q--e3j&U6BtM@O8Fa7+u!qTKt)rX8;0#ht zVBZ0M08wXj$C5fsa$UmbP+sijx2p1Q6BY9O>^yYRd7A|3W)!c$LV6NB!m z-;wns2R7NWgpCP@dz+3awm-&RDhdBa-j?>i1)&7OyQ_72p`3pJ^FrRDS{mRN09P){p*yyhskR51rX&hfM{?%{_r}Eed<7{owU>?*Zw4ZaJN<&RVmRT*18pju^At2qxY{ngksRx{+aAy@OaNJve+~RtVD)|4E0fge8<+{{h32@pKm!o5nk9)4P z&EvGXB&d^;JO0ees0JTbeD$-ILW}CmMT8t_gVl9+(z1+^d~&$>SMY}&{&?-gG~L9C zR;OHsCT}|}r|i1<$tK7u!6}QrQLLD~-=(1@u%#EWyzeo!LoHLW$jnYQm&F?+^|1RYwa+_L2IJc&xeRllQ6{^0dG{9MUfy&6^FA z0v~Fu(|BokEYgbc+K*8Ua+GU4muSIR(Q2w{^oetv$lEHIQEe1SGN-q=MKj#QW)8r4 zpM+5nKH`7pI<@Us7ZQCA1M}jsby}pO@2N*hL0s;4Hx<)wi~coL|E1QoFs}~MnD9Vf zizzcxNPyZGggK4uNY)eX(=FwJu*B7%xKT(AlVbf54J#3*?*!ds8fn?!{%NhSs0Yk; zPY`yD@>mkxnXd23n^zcr7V3#A7z86R$`(cVB#*)g8<7$Wz%X`$L}@L*Nc&vwjeAea zT4h*n@aQv4l{I#2G3~l+S|$<^3M`0(w?`0K#jIc%QBDSE5V6V0qv*8GT1}^<$eHc< znhBxG5RLIft8lKHnNB-mm91H5{pFfDpa@w;JW1kywarX!JJ2IHIem!a8tS@08N}Os zysYt_z8?(lO`-SCt^LI_xZX*5L+p1=>iv--eJUo8tk7-6k4Dv;o z=9X;dav5%@Rt(a4k-ylMR_j>Rad`Q1C-o4ocWPofvF=iiNVQjU%~<-}p6hrs+?F!y zY%{!1)c*mV_U=m?u{>>9WDq`G8pX1uDH7(`<2w@s0d=m^i_R{lle5k)^HT7AB#tm- z%pG6PRhGT1Z>nw7RJXljYMqQ#TAaP+8_4lxY zi%Lb{s4}+qUnPP?hEgu^b-=Jjf%?ns=UA?xCr$}t?KP80v zyHj%j`WFPCe<6AFxcu#c;(N@ah`y#PU|&CUV!j8&8azX$|IkaKH7%dG(w)Je7y0BQ z1Pa=b#~05yE@PP|$!rPz%)|TpCYl644J2+SIrZDD=sLdb1>MB}yn(hgmEA-C+9sV- zff}luVGc7lZU>EfMK+ywsH7r{_Pk8949ZJ@gYg#>gD$8?7tn3cMx>ekLoT}DTJ*Ut z(RK)AwD9o(C>B7v;uU(aStj&fffwI=0ayl*t)P3Ac5uR*VZvXN8Gol($X1A^=(n|K z8-f)6pfmp9Gs^xOK4T)#Ql_`6x{IjF_+N+<6Uk#5NfW1kmx+FLB6)CaxVANE(^>o^ z92|wqTBsMEa6>(1fenSY;9lE-4Oo5Z&0`UvW@QKM&TuzAu8r(N~%jZ zIaG1UlBaE?bE`fPDiSKg8>+ze#)@fDV?sCW+^!FA;+s3C&)0m5WL^mk83QB%jMSS@h=p@vB&8AS2J7 zrjEk;!<@F3@qVn@BAbDg$xb}_^!CMl3omZ6b7G>JO#D4J#xhw>KSaWnwBY^*OG$aQ ziZ?TR{O2)!HR5uO*Th6odvLa(1ycRWRB@W7Kkzh^-s9T6n~-iDBsAgl#Ep@dOwP7D za6u6ym;*TB)3~*3`-HqOyNLiBVWSmh>4|5n=Y7Z2dA50_6;V_LeEwG@22Cer{a^1I z^xx9?gQo4Da!$yfNQ!9r=Z=4@Ke-FtS<#RwO{S1cbSUiiWp|u(JUJb08E00aVkwd&=4e^L%@ z+0eq>ycL%SqP>1hquTLBX0h|Ar9IE|xOH)R2F~MV$9$F|q9|a9!EFE^A6%sVoh~xu z@4l>21Hx#J(>@GU^d(u!_NkyGV3UA}NRTLN<(naf-$nJSRgbI=a80wytHwQVl7~Li zn)D9f8#22>ri4nHMY~^vT%z)2`>E}N`<$CeYRCc=;W?+L(XeyP5g+ub(r_wzy+Sc| zxSLrWt~cw>C#(19O6Vo91a94oTVP1snThK%fY)!xfMT;a;tYK*+uP9ZiTtD5>@a=t zMehv+@EmBs89=vxFIi0TB7;!F|$%*8ERiJIGG z>a+}tKOkfqqd>D{7gPG~oKt-Zh0eZW0IUY!VN@Bs9#)!1Lm2z(@vBz()OpX-vxpWl zbgK_2J-b#q!dukEG3t-HhC0>vfaB`WeHyw-9cm+@+46sf1)@#LdcBiXxE?eMPe?jB zvVF5wh`a~mjuZA2Md5&jqc2VJC$Kc2)oLK(aK9gs=&Wv#$l!z>U0z$b=guY>PmYxO z7VI}3y}KH{dxLnj)MMg(#fHOyl-=LEhg|E3{T$7`wChdSh8jDrY~r2OM?tod^!}-j(T}g`3ic7b74^P5kS$um~ce zIx`Bml|mB~&Y%#??=U8|emgV|YBKfH%oOtE?wEcODgA{1{Xux`41o*BT6CUd?Fp@h z48leKK1YZX)U@lz;`RhzxJh3a8}@?Mlj*puC(=?fM?Key{ph>q zMffPX<^}dhw$}dUS@EsAordE5tpXn4Ubp1Yg^rvYDOnmscL zXNML+J}$AWNDt%(~mh^8Iy(|1_0$Ed2>nt1WOn2TP%;lLWO`AVeZ3dqj+yu5Ki zy|*hZIWc<%LZs|yW4md(yCr)yRip|Ulhr3_U1&G)ZhXd~p;@j{~Mo z4y$nssNq+mN!zJJQ_mD)V!d-y!9)~!-T9nZ`;?84;izutX<_m#VsU2>a7YhMX|ZSB za15gJmi8>3be-!QDrv(g^?_VVKFjddL=EN}CMgY~(43`>3)s#GO9mSezOkYpa;`95 z@uhui=YR$ir*WJm>d{e7J9*TbBFf|B{)twRrNq<*wa!+|OQ>*isNd1Z6ssvG&ZM0;qQWc|<0utnc^xHLP zj+6{?ucvb+prvjig{Uqxzy#b$NFwLHL&v+&VS$c@*1;GLjHOYdMKH?z)X0av1^h$|y#FlT(qS}b|@>rhdaG$8N zPN*(@0d6V8vuch+^P#YaAmpWe=Nb{)J*$u)~Et?|F2`A9SmA^H&T+$D;?$X zp`YCjH&1mx6D59~w|Xz1T0@1z#QlMUr=Xv+l1czwE%Kv%leTu#XE#Ikha@0MmakkL z3%e0s$IL=B+QwxVf>TFUM|vVvjkS{5<5m7)Gh0J?K8GU;c}R5+JmJXUTFLPCdyVHo)#)lA5u^YF8aL zQBwen!O3f$eDEPW7tyJ?e2(5yj9`xe+e?8wT)x$7|3rP)0sFhJ?I&xcW~AWLN2Ac5 zy8?<$o2@|IG7RQ`Pe*dV{u|%tu=CxC@>gRGI3U@;z`vnYAN}M5iQ9_=Uu|d64vSzu zIlxY9yE_5%naJHEiy|xJLlx#51l+|Rtz?>2eK@18`afW3O!7Q{?Zv;CR)OusQTgH^ ztzf}YVE;{M;nh*Z=7$ID$tJtw!o3o^GV8}C3BPc;$$PtU*b&Kzp(!8lu-WL_u0F9A zaesHY*j)sdV^qar_`^#WaV93aHya(@V1W>bcdXEl`yIc`t7}kJdCWK*i7wJnz3)D# z%B=qUb&#E~h!*Ek=xf#K?g@UEi%_QZQE(z0h4U>E)fKbt=GuqgoP2z^YcYNIP|`%S`ezB7F8~A=gVm;U!|OFO+~r zlSzhDT2Vv|HD<}`yG+pv%s%=fRc@wn=fC;dY*`iyh`qTYZM32_jRAU(&Zu^<&*o1mG#N%@&-uAB9Fy ztwqvMygTOlr1#&1V|A#Kkovq3V?3fd+#f@3Ym)l#9%X!Lh^Kd^ZQ8bWy^y^QH#Fbd zm-)Ap)6LljF7%1pvk#luT~Ge;?(BL|1sHu2Vt`B&V_nAz18?g2mKB2*8&3t1brghJ)Thy`OsYVw=?Z#!d?#R6Xv(us>Z_Gi` zow4a{!X4Dqq_Y>kX}k17G*#rv*r%Vzg7psz^|Zz3LW7GDl>h$UZ6iGMCq=#h!B2W7 zAgNaP!{^SR5uoT&bixKo5UV%_{?7B{qnYpCp$-%JtFl8;gXYM1KQn6<#?e@OvSO|Z zTvunpXZ_1z^D2OGd47YzjiVXv6={2i2LMLE0KiC~zy{y{x}anc0JW3^+XLIWb$*0B9Z z(amHQw=xsIOgPlxL+&1HsfM2oSL#&|WYX^nJG?^DWF!AE97R;T3Ii2Y0(cxQ_*TMhs({{^_%|5_D#uPOiy4q_{tZk4_f-%YnwgS?vO-#Huy`hf!YcKmue3<>O?Pf`oLd|4pAl|4iIkQ8V^*}}KeBq|g*AI$X&dX~W2$0-=w z;r!g?w_D^FB<&Ke+VkM1)|hsZagractf-8UX2$BS$x3CF4DNJ$;@Q1<>EiEkzjGw* z(U~#I-EnOqn(+yMZ!0f|328^dWt1 zAqXF79v&{vlYUAIIgXy(1RHkE%XGBEnVnPk*^!hY&0<LMp=TP>&RVaTT2|F zv(D|hf*$(jcM~=9b;y$QP|Cd2Sfx$W6<^+JGeISEU70sNPddjuN6o}%SU69fdJ1Gc za>Sz+nT?{D=@&X~Dy%uLJvUw3hlf30V+~&)Eh#Oi>bDA*&Z#ZA)Hk(!T?W>l%*`&( zySpepmoxmX58W2Y({y@!bp<#n>NW`CZqy(+v{%w}?1zT0Fx))H&@Vb+k{j9nqeUWMpMGhFu;cA^|B2&=~;rQ|?J z43o|25`KyDzku!1Ea}$}ic4VR4CVJ0=$uH}wi7y(IJ{-|D;WL=PQ+&0B)hac@4ny)`? zKJ%`YdL_$=Pa@+g_9i^$_I<={ORGXrTpFMGx4BS4(cbeAW%^kM_&oPJJUV+=Om6k9 zH=3B-91BdZhx0W9-c?=YgtQvMzl9DllSxz(4d^~od$p=J`Lk;cVO5y4_BLr5f`v0A z&}`q!8m}T_oQs5nUSWJxc)gRU^It6!o8|KG;PP=6N<4kXA0loifyMLlOHdu@LaM8X zr(Z(G6Y$u{S<_e|8^`n8u|BnYN}Gt<9qA`;5$gg}zmtKBhX7Y=!|D##6a>7$iA}@H zK)|*2G60}Ug^*W+26xK<-6l3e{$sH)Ki-4*^0G#s*8hc44#SFAW^)2wW_buTU8>_esI|!;c*KM7XnKM&|jF z&%n}|O52cNNhd(lS>cxsA$xQA2g^~cV4=PJ%8gRdI&14zZk(rZq~Zjy?=onb?<((#BYyf`k3L75TORKZcYCz zEF%8#sS0@tSRX^Kwb<-Q-XSf0r-U9AQ;ATJ0)DWjgqwjWeURu(jUi(tbg&Yr;|vbn zVL;!O*>l3ugB)`UCi;jHylNFhL~_<0aI1)x$=(wgd&?~PNFJO?JRU=&yvBmU31xif zw9G0}Do@;wzqaq0Jxhbs@kD(34LS{>p5zea)Dvzo+MAE*Nii$cY1u&}x%C-p<_vR2 zs?Q5YGWL@1fTv3gZ;M?S-R{Rhf^^CtX4)Men z$7%NdWZKpX*Gy3S+m(c^7v}9RnCIP&NDC1!glFMXLF8wWQ$VOeJp~Y@?{5J$hU*p} znW8-=b^tijOt2Tz&KK^}c1Wu`zFXilMxPgXwwFZWantXDjOfpg)LTy&k{6~z55{Jz z!$Qs)g#uM;P30nw@Z{xi-x4;T#>KTjh8g;_V&7VaDg+X8?qEqFNj+%9h+w;!LE4F3 zcSSC z?g9s47hOYdnINJ)X~wGs!H+4WlT21AGvA`q>w63AQ-kWPGusQut_2rv*Fmn_LDf-H z1GTVZZ^2p4AlR*u68f|FxeOsxtXxQvDur~Umh zeWghL{mshPINFQfFgFm;*98oMRxK#4GJ zIVt$hL`6_1Xh=m?ID;q zNiJLhhdvojBgOI7i*7$#^A7d819PG$E1DON+))w2!VEyAx%#gXtC>xWvFVnubc0x{(hM?n==TpB*G zyLDYm4tzX&l{@+&{SsxtnWDj8pCmqiX6N}KH?*_(v(aYgFsSd|WaNAl+582K^v8z9 zKCsDqhnKwJQg7&$UH1B-N_?!0@j22}ttr&B<@AW>4uY&$_h#-$tRt z4gZNk{S$@yCkpj{Ckkbi!dut~SuqEirr~q1^2hZe;9n$fW$gR-1F>4bEG>h$$+h%% zK0AB*#0bpfchUrbG$+y`>#oVM{Cx)P&N*OV@tOAunO4}yL6RXLRy+5Zbib6n%5x~^KcRJxbYyg$mhz!{R4UBGCypDs`mQrj^c zx3z7%)Q`9RQE0OtFFf$=4VLQpFL&V#RcsPP=D6Wy!t@&TL|@>fN2++Q-N&SFZ-X`4a11*~Hn5nn&D0?S1Hs&O)3{&by?OM%$y z6JToWNm+U|o}KsOuI93l_5TD|GZQk6?8XcV?42=uC;?DDY5!wm-a+s zERn(I%Bxn?^s;npx>{B;1VLc|8Y;}#n!gI2zRlzM@4TAv+5iJ?S*5ck{{P`!9`;6! zfUW`c=I`3epIs^k`t1csn7~|F5zi0^K1D;3u>e(N_LZ3wNJcFZ|J`MBQK*lcTEGO( z{HOdGJ`n3h9!B+=Wh?zG`Bd9GK=@Y*>d)GlSIW=&?K**(!cRIZ_&X&_TSyP+HHstr z`=E>extw7Fkz9b6Cg3#%h5Gv8f#IcrC;~^1{(oR~NNeZs?RZYW5o%o0BHZP?h%#z7 zt@fylen_nG!A|2whG#o@{fuH9W(Gy9wzTfO3z@s1Xm0uH(OERM0B_;cyyjsh{~^(8 z;@WN3Zy$rGMQnbuQAaFS1UPXo^6FR4ZyCAbw}m{KIg>;G?ou^Ol&jm8}$Jz#YV zpsfHYs6T6GbiDto408(oHwO+Yt~2s~qmr_9y1O*?%x1sBQ2#2O|0e2WBQGR`7@PVF zejfgp@HP^)^eJlO;NNkm#l8ghwmts`dv6_7b=1G@N_Q(DARQtlf`oJ^(h5k2lt_1X zcXu~PcXtR#H%KGh-5d7#Zq(=TiFf8bzd2{lJ7=6dBe2;B{KH!7v+n!4&YZ046er~b z;%2W$pPf{u)TN4beIf`W& zCUJ56;c?s!I@A*_njdn6ing^#lT=#$-}*L;4|wu9&B}g^_tI>ii{1Hf`}&;^3O90R zL~Cy8iq&=1AC{%zdE>^~;HH0mF!M2^BCUyIZ6%03?dTFQN?I2O%AAMX4c#kf7x&V` zy7ht^%05=qvHV@CVHPomXNP0WySp)a?~0|RWKxE;P|P-ox5bhJl`yu#1quyeJmPC| zM&ffp#dX!aqckOz`t39Q{d6=reH&-?wg5q!k*Or+*_q|DdXYDw2=qJAE~8&{6*YwN zptTL2eM@R9=2x$PG@E2yafDnaiAuXex~J`2TRvlC5&b+E zz5xh4-vEsjPRm+@VD~$EA3ntHNXuDz<;7cv+f(=KHK8%BSzc02Jc+$5)ynOh)_hvC zwug$yQ@Lx`B8#HeLNW>t#q+N_Zm!&pXXV~xWxNzhb59=)7II&7zuB4Td->vcR)c=F z#!_5Y#r^C$Nln9iV}AR1#yVHEMNP$4`(*9)RZG3mB4Z3nw7~@J_|(+Y-b{e)y9Euj zX)}9En{TvCQwqCRoYMUWBsxE?4HQ=SZVpe*LBQ)Q4d_9Uykb&nlrPQTQX;EltGT?R zP_F7LK2OoP(AG`kWx@;93c+bgzWN-3PQ(zZ>S+H(0C$iA(XfnI5ABR&R!r~Zs&y*r zkD3e^H2bKLZ4wD&jv}hlHl$iX@KTH4h{qm7@Q@hxK%g@I0ChDnz&!MpSXggl znyU)j@&I0$yK2$F+Y21W-JyRMfAMyj@Q}z-%XNJ!$If%|h{%3YdVyW+M<&OQlXJmr z8H?krWi=)T)JJ`EQdNpW3{l5xrlmb(imM;PVLmIZPV00)-qju>N9aXvaVXS*GlnTs z(eZhw=2I4y4LZ}>!Ue$E8m$RuxH~5PaIE!;`B6LU(awBwJt(w{lM2rTAb-g;&&F*% z)55nmKOfGe5ud?WJoN zfQy5uM%7Wjs9D~6mr!aVaedyx=(%MlNBFxDK!N{o>`o{EgHPi~r zDBRvF%JifhCSQxaUPC(cU71#SpO!#(!bglu&ajs8b@GuJ(R2HVHqSTZ{rcuC()Y_h z=;_vS8mRJ5KeA)o(xy&eJeiGi`K-tlnjKGm1!+MULV`3wCsR#tp1T^ef`ArOE<_xR zv%v(+1%?PcffiH=G6Yz7GC=;~ zDXwp){4`HvAzO~V3M$W$#AY4||ABASr3-dTA12@19|)*NfUKs5%*Cq{gAlMmI89>m zN&V9F8lS%6#B;E6Cym#Fy?`R%_fdJoDeT$c!Qi*J_5bAIxB6+f50@uM+HRGW6i6;pVfc!XZkaC; z+Q8hGzO4{La4Qg;$y@?7z$;5YzDD#=0~&08=n4-?g#&N2g0U;tPL_a|0iWU<0cm(} zr$XH|N?e}ghFvI-4x~cu#ilzm!OZKXU-x+IG$S7|g*8$GNZAWMJmaE3dygW_<+EoU z*NP$Q6n>a~r+Jvpq^RA6mTJfwx$8=@5S{~XfcR7-0CHFa3IgA@*nrUu)zpcJYQ38O z)9bSf`>0>>y&Fi?|G86j@`)Bx2R#E4KVmL6=G5P6= zveM!RsHFSIfHG-_vuRLiL|^xHT6u=vBBwVH2uQ#Wd-RRTo|R~Ec%{4jh{6G6_mub% z`68d*b`a~Wa3qMPuw}v!t=5&7hRg^R=}WU+!7#|ev4L+ z_j;)29#GuKJx40W@T?)H_Ng>9@hth>e5BdlhDyG4+%XrC6#L@vk5E?s)JH|y(is`X zVrg`9Iq*crnad6iP23r2-84vg`tC*qMr*Y35g__q?CS}snC+=H^tkP*(Z$TogY*`K zEB4{V+|7dwjXm}P7lzy0fQ1V>HPtgi6rIO&4n(-`mELsi`P`&s5O%Y6kk>QBGe!mY zEUB3sP9$C#JdO}66Iop93-~yO&JHUgXZMWG<1949kM0Ssq@W6=;(I9D0>mdUD#fNB za-gRsV>0?pZ}B^5;uBh0X1stkjtAs=Q#VmxZjjy4g-HWwH{PSW|g)(hdRlnJM(zd zhYaiXn;F4=xpQp}2SqzBVpG_Z09MKhdX;L)@C*9#&xndq20tnbI!c(_V&l=9n{j(@ zKkN0P)$a%~+G_h=I@tmrbO!nMZ#;RX;4jJbZhsCv`ley;Dc*^3I_RP%4|EaEhGL)Q z0O*0&1KWk+fQU^6zhnjfZtU4?$#J|0SaX#3_8|(ZTV_tqG%@$rABf*d9&y^oR3dyG z!k_3Ze50A3j4fFuWbR|<%eoc!^9v}+#^Dna^OZE^ACt5|BA_^@=6IZ{MWd-v-er-qR69gLHNl`O= z+ko0J=6zK2f$bEQTG@m|(;q41`?8XAETQR+-nC5Z<|Q7E3|&oPbA9fHV;qX#5)9%u z6`&bep+r~gPJeZwL~hgncA@@uq5gKE{y*tL<%uXL0GrZ%%h zn>P1T8x~+{L-Q|Fo8npw_I5ItC08&~&ED6Ltq!C7P?`u$dG>CxB7PB$y$@u?G zY>oeTyMiU#?cOj@+EE4(4QsI+eZ<_Da!D~DGHG)sQ+DQYbAx^ zVpZ9uCPP@bsBF!u`&?+W^hJ+b|NR3<+K^Vf=;{5BQ4E+gYb=7Gj8;XXS`O>*);(AnirdWjP2hWnbddK}`%;KLO5I!SebEak5b_ z_)J@p`k4;nk5`BW5$$K48TPr;L!ug!3+PoFp4E_-Fn$OuMtxVWyJ)jGmQc~WAY)vl zD-(DyGE+}Esb|N1hV+4D z^f`@?e<#~~%ILew6>EC5z`f=8eOeh@x@9R&Rmyd=KU_KHu%K7-~iF6UNC=;DUZ~R8f58$u92eGbrifL54@-RBgG1kYo!AgdNis zHuos3ADIjS+c{!DH!sG9--N=qXe&arsl;nkS_Sixkjh6xX_#QPoYDyjQu7#eMw+V} zQvi~gJiPz8O9@iZdp;;XWvY0%hlL+ZDSE4#eLU`%H(y;4K;1U1UKt;Qi0v7}H*Qw1 z^j5EumpD(F)6ogh4vMlb!dPi=>JRlL%d;a&tyYGe!D#y}WT%A6$`K`05)mdtS97v; zG-v?k=35;gyH!~0<5Z$^>EqN*o59IAUcAXOW3YI4+<;aTgVAqw@iwre)zZH!9*~C` zu1CeW-DGt=!GUls9Q&Kz3~L{vyf5!=oJVx%cMML8oihzn7o0g{DI~dl0>Ag#>-*INyK{e_0Aid># zum!}%5%)X7K4Bh5lt+!f$#Htq5REG)-c}zZV2@*#>c!nfHXoI~aI?gVR_Q3(bi_DM z_LeCX!-G%!M^;6fD?Ui1l@H}EbE%ueuq@EvDEe@z?IO zEk+#kjSCuY5<5e^SBAEn`T}sjq5(XH-RCbKOtGox-`^Sg(ca$~KM#zgN-{r9j4@Nq z*8fI3_w16ep4B2vu>ly!eehDbu^rxpr=br*nDAiKA{?~7 zACKG`BgzXj7wF66O_JNB}JlxwM_Bo|;~4o0=A^JJ^3$XKMP3cu)fn z52m!PAlv<$N|1}DI#ujUuzt4FTUc2Q6yr-J&#awOF?+t%u~mauKE6!jvCu63yW!hc zMkvIh%yYz{3;JPJTu>;<^mr$aKkv;Ot=~3|5okmc&0Za49#`?hM_zTP?{HsNXc^w> zMXl!6u!wJmXj@3Y^81J@Fk2KW<$M*irp7XxmD}j}rh9*<`z<>fCu?~+%p7t~sIe?H zwbdH{;`i0D15e%$ILj8TUX|Y^GGCwkywNq)|8=9=0tYA#SmLRRhnBAW35_~}-3Z~P zO6rcNfir<)V(qQk=DJ_sOY@VZ!R}i7A5sX#^Yhb=uZ+RF*(b*b*|8qk`;`bbcrH{+ zFK{D@UuBbOU}#iL?Aj(oJyM6Rk^7WrZ9G}`b@#4G_WRyg`KCL44!LaoOte>#nCwS+ zY3vV{r}3h=$rge8;sqO(3)n@Zp-&^DNRw>IX+N}33WQSPV@xMh9@H0*foQ|`&Cy#= zNV|mC=jIQqubZq!`6yRZN--%XUCB{~geVVsT%H;)Rhsa~7Jdh+g!R>87uf(hj}H(S zWF5w8HY&vvf3}f@oq6VAN_Yczzw>y#BtmN(R{WN1=2sWnd?q>l4TT!owIfPXH}XYF zS=_>yf73ie)0ppvvZ6zo%C5e>8LlZ2_3n!=E1wpgX7h56lG|Fu&ZIudDx?z0c-W$% zro~0-@kBE(vq!C;7i(8518qAE(Z)Sg=)#WpGgWe@=-k)--@%o$$2i6VVv-k0nyxJo zw}4Z8>%d;~CCUItf__X9qgx;GOdi}}r99JEjA$%_+Zf!V&mk>oqjc5PX`>>?hPnc} zc$yOVq8q_mU(9D25~FO`nzBZjhzqD5O~)HLie}?&ACoexS&4e*bmXME<**Yn@W6|p zTNOY2s+yTo7TWU6b^oDW=faLoj@}t7lJ(z-pw(V<1?IY75zlb!?vW zq+z6TKl?Lld%60a*FNfB!mjeWqK00csU^60EfKML%Fz?!cdi7m2A6ar3%bAs0-&B2 zK2gXxtL~KKO)#9|rDC-=T2aWL(7mD~OXN+u1h-qY`{D@M&s%XY6SMX#(&!p{S3PQ* zK{llGsEXbnk)IZGgDmL4gXSE#Kca)G#>Ncd4flE3z;a*MP8ET3243?!Q~M>bYy<0x zm)2#)x8^~E8A_Ow8hs;59JRk9&L@4ZY^<0!iUhz#5P_y-+)K@Wy9mm)R6fkiM1xO_ zL8tS49N`l!0IzfZSe*`pijNkM{||EDC4%{%;pSKhu394dA{lzazHnN2oi<>DM~@%& zI1F-I26cs17Imut?9dqgu|q@qwL>d{?9ikcbX^u)IO8>zSk}5Zldz7fZWjO1N{Ld4b!~p$rpKFFFqTqh1{02%rVz&;#Yr>v6+uB zg5VY97n;?SiuzI-i`g@FbjyC3+&M2nBqYPF?Xe%1`x+E4XMm~trPi|HSUaAWWQ&2c zK{+H{xe&w(AeE+)qpCk=iooOncYr|g^&wtWi80`K|5zM&fhPDAbCAFb+B5aZR*c^D zXs8mX^+jYe|10<>E4jgLcf|HCvD__^DdMYFE{93>i|;1ibQqdCV2E12NT*2YiZ#V7 z$@Ux)`gHf_?#1KtKXxytk!3W7Urjl4QRN&3p6q~6Z^$oC<%5XX{W%sCYZdqgFyP4S3%fOb`A6v{u9l683YieNhggP9gC2#E@bKCqP1{e=U3dzKuj?h$eA z1$1ywtVMGaw~c@z$#9=fs@o|{7{D9yUE)R6hbLpA^?$46(yRAHML%7~sflp~7>#jv zy;aLuV%);0Ghw>CaCapD-LwI7TGUGnH!egM)QaV+=TV$Ux=+Z?xR5Km$v}YoJF+>| z_qiwyjhE6CL+w86FTKz=EcB#w4ia?CcfqkWuKoIt8tO;HAb3f6C zmF+@wOP#&&C^4psGMdTlT|e?Rl+CvR5r8NB*f&l1^WIcB;;4czN2n zm$&4QaGq56d$>~?x{*js>-IdezS750m?wfM6Uo`5PaZm2c6=O(ma#+mOzd3C2@c7zWwC+!(;$C}nm??Xy^x0AgJ z#{wID3RlW*#xKC?5U`SO=VaI3^Ka6odNJcqru;>OpKE5kBBbDp(}km#)5DGzW}o1^ zNgo}1w~W*yov%gn7BC?Va|FJ?-xXQh5osdWK?nyt`j(n_g>SY|K==BiVFX+D#`$OW9G6G^ zu!JO9mK5cO+``8pM)1S%QagmLUlY^-YNTO-71m`NHJ+<8t^~4%>1htv_4nFD@^|c* z6Elhri0-Mo(0l=Wo^sw@`i+1omysaM2)l`D?{ijxfURlzh2lsN?O5VLr%zAJ!d6GV zB1)r#s5*Lg9K{*qJwP$TXL-S^t83XPb$sTUdzQvO^?V)Jnw5EbjL?h%&e&gz=9e@K5|tjH-$Mc6=Z=4!KJiO> ze*^?Z<{jR^T{)?P@&O(mzX0@1KXA=Hj;H#G_6~9Hc<82Ns02Ft%qXyp_BgX50p)Z1Rb(H;+ zp>k;c>b|E#4uQ`8jDLh8L&z%satOCbLU-|Q{GS1Ey?d4QM*^J2V}G^mBwu0i=rGM}aIlSS)#ufBEI6aVuE@KK_N@{)ONEh2Q>#-~NT){)ONE zh2Q>#-~NT){)ONE{{w#8(A6f>C0U@G0lwt2zjc-%-pZFfj`U^_C>%_L&c4u}*wiP1 zPv3L!_~MZ}v{t?(N3rbo0?mXTyYPS;LyNodW*%*b zepa03R_H}@osvxt<#WiG*EHy4Yz#Q7wOBo!G_vn+P2hgkF|_P{eyvg!bFB0n>9zRV zsDk{{;S%7yHuhdwze7gTtqbL20V1f@fcvdNFjmUVJ;?!jWDk9@h{#CS+i^%y;$X|G z>c_vF)t=ZOvYmEypa=7{?=Tgv7F-GPSRC^N`++c-fMNi8G9O4bj``Bp&OO!D58;Fu zjLB5=fposj=Lq8nc-HC8ERfGzflmkJzc4plK9giD52Y~!07%H=67Y4n!UywtcKutJ z-pYn{luLeh5x7j|mM=V8>3~o*UlISWI1RU=p(N*8lNf<{W4N>7a^3w*Tpcz2Okp9- zH!B=BA9_PT*@Jtk+1q!P4XMrz<h71mfb5NZ9xFfVPCHn#4%}w|#0n@f*ou-p? z63PZvDKg1o%tVK@C@c~!q4OLDVpiMI%^-|{ zWg$zzhvK^q_)vV;4+Zr8#fOsUDQju;+lN9PyMrpzolU*?UkIC1B&GlG*IM@i{@U?n zh`*NF{=fNapZj;6`D&9ZDr|YOeXE}%x3))X&vAp_x^Y)8CIS5+L^%tD2vMS(ehN`m z{uH86uigt$o>RsCAw=;4{S=~X*+GOTl4<`CqDZ9uAw+>y{!@se=6`ZBB9TEcnW;IZ zM|YgUinDPm3c5hXlzb%NT*NMu^NF~q4Q>_3011MX-^JHpUh8^7`io?L!W0E?u^Xr=X(T&%4g4ecjE_@sY z^X3nTt}#3+{JBOup9GpaFG*qgS_^;MOs|7JYzJlZh@3nY zL#B?1EFHK+FQy_d9p2yaRr!uplfm^Phka2;g(k))Wur3N_PX@hq~;E^Y;?^|Bfayg zH7F>rHlAH9PdcQsCz)HfH}w(abZ@wyigpp`8d-r@C+^|t7F(FaB9ljCR49LQce*1t zMrDK)vY>^y0Tco0iU#H~LpJZ1p}5b6@&?y*4)p_C({H|?mTZnGH^-kJzq|`jZy7tT z+>2oVd#KL(_=5x+OGKUQ9I5O-aMUesI!*r$2Ig+v44S{jcHRHsDAVVLJ+d{ z#MEZ!T)z8V-7IGHlU^VPF{vh7*no`HWk0C~))-fl{n-2wlrQxuPw|=m;RPs+Epz1j zC+Xr=im%2Rdom#XEndRdg>|`WMenx*xhMbPl9Eg>W9%Aj!*GG@+5aEz zOfdYJ{zeO5Y08m!!}pju!Wm>@TCjg$aaP48t)^Tle=o8N3*O0k zDY6bk`HqENB~vO|hDqNGQq*SaP)pIr-0XMqr}|L9sYWWqF8 zhBM8Ul18z_o`FZUF|~A=qGHKp9XB7HGjdrFywM20?rHGCE8nDFn-V+oQfN>vDtt9@ z&4AT^S0Bu8dJn32N#{-sL<9?qN~_5gRWyy^jzw#MxBn$K2oK~2`Txu;ltw>LLHaM= z7iW~id3aMOn?2l*u6c?!CJ>&+4CBWHN-fE_pNT?Eeadq!eXp0AlV1Nmha1#y3k_kb z6*uraQz;0^LH+cW0LT#np+Lw(n+aY=9$x9O&YwgK;zzs7B%MDg9L&)79FCCPC5zQ6 zu`S!7&+5m;ef2}izRaqxd9OM5rJi$B_Kb307VctZ zl{M#i-=JCdT!!{^hS+y&@p9S^7pW@tfVR1c(eb!`!LK;>>u(jtlA)UuKMab7hd0v! z5Gag9(_RkrE|Emhc>3}-pC|7d1O#BmGWxI$tYIrNlMzgemX+Zxmc6?AUUj^=`a8o5 z&pO5hcEogk4ns_2R2Q2pm6R=UzE4LmeG=EuTliK%scg{Kj;mK{#~`-*ZbX~qy0efH zAzDe(77yjXH?I&o0*z*Y#u#*7jQA_|TCmTs>1Tb>=9uS589W=7(rYQK!a*a_T`7f+ zTGoN1WK2PVSdvM z!!KOCommf`Q%6yj(zi#z_HI5#jFfSY`s%o}fWrDhz4uDycp6UAnV*T!1$q>Nzkc3~ z#N+09)?=^6EJ||SE85g_lWgyZqqF*;4rl0Nd0@Dy|}z6Q|WBM0J;hbk+giq*;?wZ zsU4C(MB0xwBHp|umrKCj4AVE@uMYOv**ZH$?CgTnrQ7I5qOO6SLX?AomAVCT4RZ4c zT`tltXrO}Y&;ssxhv1R*kgJ{45Ioe}TGpl2;ncuUX`%LdED zFwv<$l{0CK{c2pgjMF(I6Gc+d0GVk^>VF@aOconwmU~=6S7#15??n3l%zzZwSdu8f%77)!#sl zA9znTNw8H2YdR?Tnv#`dY*lH9vD&=+w~c~8A7dVszDpL>!JY>+Q@?GEc>&3!Xx%>i zEA;;a%0&vxXHvWvqiH3msWWYVf!_Phv}+UXXGfW+;8#ak!zU-vjAg9z?6E4Jxu)-x%IbYJU*VhS3mXwF6{`rgI# zK5dewtIfR?g)_*f@e8oL?g@j-o%6UE-R3N+d7Ics`TLS7`UMrFQGnM|*hzJ%R{ANI z$rA%oDCkI>nLj+YY6J=e!c$k+^qPLU0X-188&|oG6d^~EX(XDP+Hr_mPy3I+9$ zLP3DZQrv42o;FHh7Vcq1RKpT>84xnbX-GVNW>pRV<7t=;cze=e*o2z0*| z*5SM-uRPh|5_4?o@>~%&%@Vq4tAh2f%p&zU&0lfbUp4N%R6?8j8X!xT`el~xitu>K zDk|Jx_U5OiT(H!7SzCbV6-s?&1&3k8-Vq{1Q~sTH@a;BYRH_aooaga_bMJv;xE;5H z#}|43G+r4k0miHR&Bk^puHOvh&j54KMgq4{qlv*hGIV|BnRqJqXwElnmPjuU)yCS% ztW)tfu=kc%%LBwU>#;WC%Z|CyXnqV8B*aO1MHMW_Y72cusS3P>tf7c+Qr~tY%>OFi zQF!tiQIY;xBLMRHMX>(`+1(81muq~C`w`CsDq_s8qrpu-!tbRfB(kpwWcS0T1GBLv zy_9f1{kdA!V>-Udh2wiQ^YM^2ru0P|nZDv<9bj#QJb=rwFko@j{SSkjD`q!h=l$bh zJ-)jgXSruz_(E)2$N#~x{W@=%W>9c#(?{MkfzrdTkobO}HY7}@cn>PiVJiM+iHf!;Gm%Td%96p*v|cH? z_(+Oh9iB%gX%xk^{_M8Jf5eq-X2P|7`|Wed#M9JaD`RrM)N09bq_GuP z@|;CE2L^YTGyq8jPglaS8H$M1ITf*`PK|tP;3~`wPyrE-m8@|Mt2fK$od4*jSqiC0 z?RR~i1b)QO60zQrJ8`3SNU^vG51s{9*SnkNCoLr)yHHn>hS#7Q(&d;fp7`T2`Dg2A z7o2mHF2DeSQQwPFFq_sLJw69s=NCMd-^eM#EC#pVLO72^q(wB(_Vor;u35nnVbf@j zyVhR*__W-k8O56E&whg?Do#IgWpPOImmxhsX`lya2k8O&0zE*|Zv;ySI5v;V#Q>{L z=Lhy~{@rl>8$<Jdrn!QXnF$oy%fQ8Ob%*1-(au2*M5 z$qA}zFGT@hD>;iZbfJ0T4K@8r8tU1npoK%BnHDdyg-5Bvn(} z&T!bLeF&l0k&87fwI%0P(QB1{guFDTY2(3Kk;1m6G~3(%FgK=uqsNpS>&n#hkX|;s z35ojhU36KP)dcO;m5)LJJ(l{pm6}6pjHL7~H)XcGGcmcQ5|4l!^f*E!J;xgET&(<% zeiOs81$Tl3wll|$u*|bKF%M$V=fsCE^(S2L=ufujHtv)7hZaB*-D>-Zs4{ zqOz=&Ny2u$^|51m(>K`VGV$ZTEUP-hz8Ebykf%I=Hkb&23UIo7LgLc>Z#fNNgR-+o zbl{8Msk7gWr?wzstW)yj~_PQ+<-)eAikmd5q zzUyrnMLb<3x_4sw>V7oiF13CCz7?=6jJOjS6Q%>;KEx5fch~WoB56*~#F71WuufM9 z0{Ff&971*o_Rz@(;yi;#^FwK{n9JvYE4w2s*{&HeIsOzbs)gh{-P;Q!eRQ4%P z-QCz<=p!JLis41b$`%r8zyBn`J_3x z|8H~uZ*%`|bN_F1|8H~ucXMBlj|Qyn;7_k$*f)B>D*8~puUbxj%UOBPVJ7u!8Yd9a zR4u`Htz_2AP|H-5GUk9%`CVpICt|Zhi6%nES_u1;HH|B9{U)e^pgk3OGv!xWb8sHb z)bLkY6E2WLYmfTpdcF%D0oL}uLBfEZYOZ47^QfGgimP|HXS;ouD zzXDhK0oEl-41Nn4b+8z6h-Scw*PMM4o*7;YKi}lNp#3YZ*{Qz%M_hA1piE7e3zoOp z3+*vk658YMVP&3Z+Cp#GbXHfyA?Lt98|EY$@NPBajePj^(jeGZmf+ z)M^v)Rm9a+3esh6gHUxmbj!dBx@E4qkY`4c%kiBKneQ%?x8h*SwZSFu;Wb~JU|2F} zeB)LL?Gx_E(u9-{)&kgcB0D5?B8B6ZAbRa?CvT;|N4dWWVLGYV6`~nVmcn!$qkTj% z?s&Yorx_V;fL%Mob1U$TT@7R>VNR1z;(nY*PfLm}Er+1<$}Ra*#ec}Pg;nhih=+X8 z8U@D*8ga6+L0IJcj26boj}i9$6|Eq&ZdK8nZ z=ad?pI0?P;0|o&5V}uWR*GT!Azie4y&mV1)juO=U-aO;#fCL(U2sHCBl~M{nZBbf? zkcR*iqIqAEkBapxvSI&*nqI%#E{Gg0qdUMvc`1eB=W+MG=_z%q(qqb*n&?&lP*obmBkoN0!J;-gZ>Jj*M zz*ecu`+RJkf}t6Dh>Ac@D)y=uv!~QA*|LDPT7(;RdscS58E4nUZ+{( zA9|e%jh`}PL0OrA&1kFo9oySzZ$l}l7hBuuytL3t>nh~*DCsfbz|no`t#QV$qk9gy zZtnfjJ%a4tkM6xPGKBHRkN4!2pEAJhyntVmY<0#y#`m?7^>H}xykQ;Qr#kCuu<|8s za^gk2qxX*-ga>OLl0bp!FGD>@TraJOKs2p@@+P+x3_HB*TP_`WoeDKm@oq4ZpY9Q8 z!(Dzf+4C6BYZOuUm)GXtd@O^3;ehpnI(B7~AKTk19#45TsV#2|vSu)Qjs@JQArE&R zkLDV^8*wBXz6U9e0QbZAFF2&%I^Rd-lNBV%kJ-L-^}_+Q!-3eg63rJ@A!puOC7(~J zWUsT1ZT+*3VZI1oyI}2e!jR`}T$=eq&iGtk)*2d`oN~2h{P#q7n^9bxG=L^PFg8bU ziMl!fI{Dt8l=zH!-ne4$jX9QY(M0<_e9^MJ3R)0#DDivI-qPK2fQ1Ea7OmCYM#uq` zFI$~BxIQ+<=EDu)%ROPBQ+~)Ufn-(IG5$bx)+qZPLQ0;~`E#l+6=T*lVyfJ39-VW4 z>Nnl|b?P7b0G#@>XCSA38E43;pNp}66R7Lk0CoNJ9iXny3DotWm<)$AK?=i*Qv|+Z zQO#q0YDjAfvmqAx%}r+Ai*LtH^ZKL-1rb!y5i%+zW~FCdP-e8ih`g7s#`!w+7)yPy z!vQtmkaT+9qDc^&x7+|xF&dzN8qTU~yCP_4qpUSI2(2|Y0mJlWKnAzK77bPpTw7es zIX)OWzVWt?RyAPHi9RO7l5l3IT-HO2C?2C*yA6|$EKs`x1yxJ(ML=H(UF?)H*in8p zbL4i&6&T%vrMD_q?bl1IvaN`}IFT}R`e0$S(H+`51^rlJ9mBEcWOFE>=q&YpMC!sI zeI8fHH)gB{j~Ugz_SAif$AwG?T8(9sF=%4%A)z#}@={?w+Cp8?iy5d3gKf?8$uA2= zu)eFa*$ThBkgST>=(`5S??XSw@6Fl2#_wQU)_KYyorGFm&m9_*QZ#mGlZUD;>s^Vp zJe_;SKhAB_6Db&{ItM(tNLJSUxt`S8*lK6b#h#b$wwJ)Fc5#n>$bCq8vc%G;_(Z+6 z`k^r}8c>tIk;SVNkj8wpes)BIB2@=>kv(ClY#Pb-TI}~!AYc2fM+*F@=7^J}xq{H; zN{Y!UkSqWu@qN?4B);Mo+UL0^1+~d9v`-ER1nomE^Be8Mt^uHZ#=b((K2rpZdA=z# z*0QkA9+OKD2LAYs_Ia-p#@MhFH3N(Vrm(~HtL8O12(}D{=T$a&V75wxy}6&IO01* zzsw^3Uk1!BpOEhMLxj`w`@a?Er8g>< zL_`QtLlTky$6%NWqODXL67^sSW@3~s?Jn&$yM8-VU}PTD9qO09JM;j4hBeCg?q!M` zYidk$^%$XJfXw>FUTowQ9IvcP5=!%rm1bPql(>@|@7cQE_NoF4y5nt`s?FX9*(&G8 z7d%CrJ39;$lgiXpm&8RyePtRpK3wa#Pdacv^cJf`{N(k4!QXGDNI4@{_%VZ+=qivu zc<1*z{0J6nbT=U#d2s#J1!@!T`z}pGqNEHdLT^6s_(Wb27Y)bi*MVutUQumZ^QtIh zAaQmC&0dH?U%T3{LV4zFWy z|B#k^{vaPc$1vz{i{JUZuV7%O#mYoy55{_JOgmPy1o zv0K72VkXC3j~&0HBm)zYJ$lJ$rG(_{cvqX4Pqw8EnfYT}(|fc#@{}s!^o4J$$(0nt z8)=2hlQEUIC<;%_UFEN>6q)Iq3s1{(3sttXvqP8lvqMWqyTwu)USChIz!vV7m5V5T z*VYZaR{Cxki%);Y><8u8drb}C4fDrw`+~(PIAj?nnvJH^yeD`I&!O~-&PO!A7M^QX zg@#mN$PwOACeJK2O0>6-$kF_0aXtKMcP=e?^nTk*;MCFnbGFFs0sg1$GlM#>YwWU# z1&LRlK7F|KI@Rh$C3fDI7nN?$HlIuc8bBvS3uMe6yi8}?#&Jvg5TPU>Q0Sl{kbhPc zwMpVxY+10D3r4aBW7&gg?MosqFuvWiP#v^vwCF9eFPT-WRVxp%MQnT)#%oU1ua?7! zF(Ily06!RVl^$Kxu}4xBiDR9gi~r&I6#*4Xd(GfVrf#zFDdEPZ#ah9dwvXIwrJQ;L z$ynJ_!Eb!}%=8CwWAo%Mn;#r}AkZ6(y=7s^?u6y0d2B7ApM&VM%Uq-E652+Xp%Z%b zzVCpU&-kiowF7sHDhnNqdz87TIjk-`dPiKqbjmo&F6ndn zuvw2e_@W$>`iy`xv8Sy9Z?wmbQ#QPQ3>R_7^E84~gcST>bryMu%nn^BMZk^9nwAsT z;~>tzQU0QY{NsL)V~PDSNT`_>W`S#|4rkPa#%@KnUcZNF@xp^o`bQR>n=8rXg(vvI zC58tHALte89mNtK@|`wt2{5|rQGP&knd(+alb(b?T9mZJhorAtKlbh}=p74yIDF{- z66&Dbm;^WPll{fmH}Os*rssrN6+9h75b(N|8V(#18^p$opH@BsAR6+78vv-j^qxa%9 zRn7sP!_Z^Q_dMM=v74!=J7u1z`;%5?xq{pNXCkleU_sOBe9+*~RCJO0}{ zCh1HYWW=6=B(N$tE7|&&Z}Z7WKTvm-EhfCwcEw9%7O3O9Z|p z*t&Aok27a35rU)YEDp{qpy1$Hh`|cU)<7?m%k5{Rd)yjUw++(-$Yjvl77x$!iI=uAcz| z+BwA9_hz^K!QHjSoyMW&4Q%OLQmV{d$RsW^M1t?O?xJhx@0-9-WHb|xuQ&xNJW>&Zd-gktFr zQ8pmE9VPDr5o_Hr#sSjVy@uu39lr-Vl3a2kFp%7V$HK~HlOBE6g_U-sbw8>! zU3gqj(3@bTjy&qPm|3gOy}qo%_D#QB{iDIklN8fvAs>2 z?U`js^y)?V%da!wZd|OF=kU)i^i0zLhi<;(5_~?tQUTxPrYp3K76z}!G22+1h@GE% z(@p1~{I1dK>F)u_MTlG~-?=Lu8B%@B4uyPI#Vwp`;HOYVQlx{@%$82mGMA=v_JVw7 zj^X}k$a+*--rY-lR#CpeSa$HmI!e9l2nTipwGi3Y^aT@unK||jXMn}()kL4 zr{hbLkyO5S(BWQG{#?+8sH1r16X6zlDg?kY&zxVxQQ7b63?u8%g%$BvLvx<$!IdYY z&U{OY=4pg{|DV7197B1x3#J}bSTw&pv=UO*Z6R!VU$~lTh*N`lUfEht(4(Gq$baGe zs7Ia@Cdn`LQID;E_HGLdbJ(a1+l?&tIfLC)p(cv}M`)z2NtM`%mugkit>@{Q4j6Nh zmPd+0S~xHp;XbEh&g9dhQ>OW*^S7 zeZBj_HO@S>i8=X2-LS!4Df-Qg2H{nE)wPW5`RYz1cA`2S;h*a?Q8k_s#%E0VtMcC1 z(%#q)&ErGQGz_zy(kD-8wkoMNbg}Plr7CI3!Qjfo8Qh}NFC~u##&;$xkS1Q+%&vbL zdE8%HYl|u9ZZ=8!aMRVMlVRu*fb`wAzQ{Boec~B%v=Wnghim_HEru+_x3;&}icfbw z^%a@^T!KX!NeCCy)Ws2|gg${0?Cu;R?x_33mtM7J`K=E=B%$a1=!c)F&_%3P>{R8{ zA7+pk(`#6!{R;Q>(<>SVpS362uMUXy+z(sc?jda5KJb^d znsZ-r5x0QL(f{a+`)H-+oqYuU}&JMZmVpjTx?g+d?!t6VC zCqGe!j}lKMKMHInz;16UAx0{@kAJm1-1P>cblX=lC)3cH&SC2EAjDM?Hz$>4PS_#M z9u~yXENXJ=E85UB<1FvdBTmMZ!=wlJ+OV;-cD<)#f@(yJw_L3)IG~2sR0Oce;Q>?) z9tdCF?ZO$7rK?I^vc()F_Bo_DWCXHw*zG#6zG0$Q&Ii4Vb?Up!MJzChuD7-|O4s zJNQqSbd7>RQ9XLryykU3Pul#7U78*KF1J4jEpr`zmtvcv{d}Q(@8yo-Dm(q_9^k>b z&wQp36f;mEhuE{&K>H2$kxL2tMiNTRB;`2>#3_AF6CBA~NKnCy%Ho~;m69p$`}$>x<49-k{AC4XzYlqMFi-_g!7+BS&Y%`@71g0Q!Hzw!(&8GG^f zjGeocAbbwxboxRrg?!!<#Q0V*m7A1RsUBrY!_#Cps8G zrr2!W5+$gm{S-&(w7mW!-X`)2LY?mdv(4j*wB&qtnfQyO;%UrWk-_sv!}h;ZrL20# zG+K*6N5GMvb^N&}PKFA^qbr4Ybg-nA=2;uxkNV#Cn5Xb$qD+yEwLs&Jgo~wENMzQM z#jYVDC_wRl!r>F4`;Oy3@+qM!Oa-ukeYLDAOF21>kqMF63-wmdy4X+;71t3{xxJEa znJJ(>7)EJ2@DkBVJD}9KhsUj>j@X-gw%)4zqtT@jMLNV2ZJ?sNejFfyb2SL#lU{k9 z;7XhNXLc(X)YD9Rf5hjp-5h@j-Vq^?tOd*a`NvlAPok@?-hYgRGhc$cyD7RO{oek6 zO=gw}Y{K%?u}3PDKF!`p6#_LM&fd6@^|@IjmNUA(42(mHH8{&L_Z$Xm=7#aNpY~7foHfkvT{`EvMlUj*J zVBs{AXzlLwoDt1pM1B1NhF)r{RwRj4HC6`aN-Zwm-PWDVB51m7^Wz8lsB#uub$lH| z>)`6JM=e9^Gwxb=5F))dG)|=F-xW~x^qy^;;4L|T-m~l&L;Y5BG0#W+cICq_F7T%? z|8WmXm}Uh=H;uG+F}(EM<9B%XO#Ky3x5kyp9)^3D`kDyxrDcNynS1s;Q|_qZBwK|B z8K>nO%l76cKGjtbX=<}UYO`fO&2Z|+^NLm@ANyghZnUqYn`^2TOpKaCcWRDqUsq<< z;x#zKxd8efCoS!Ijg2*-8Vsjj|C!;kC&<^q15Hun|99#oH2S*~{QsujPxA0* zK>64Hif16|wpvJk#YEQb_x7jzU!w;H6j#VH82{{~*u<e^G^>{2TGA(*jYMw; zrzV;iHgqUwI^|ixrJUf5n_lEua6wA#sjqd@9ycDSV^l|kRYioUi3O>NaRpNv(RBCx z=h-9Z1#-m=IjgA%ySVGNXa;;(*y@BjUm40c8m2WhzqZllS7#!=`_B!h&i6b2l#wF^64v|HpmdBRQV#$4Hir4u&YiaQ=-qHy<|yh?|mS z2F!F~iJHaH{kTYIpadlR^rT z5%~laT9efdd@Vd$dC<>+$#XHEMRCo_I(TX5K;Aluo2U~XPn#h$M{{@;XTy=V1ZfyK+`ns zbTUD-vXEE&cBZGbd4Dx42V)>+K}C+4LU^X$9%h^zVl%(?j(ytGK0j%7{F zX4MMLN=5^_!6ZwkOJCz$`*P)U$mbkzarfqE9%(f4RM_2s_+!^3aPg0$O>VO$&$kt| zt8RFqr4As{rGepke6Wf^i%)yeKo8pe$4@RIk&Wylh7@TEw&u=l(dbIW1cZ*Z6@^AFUKC|j5>KCy*l9z*ZC`MaSx z%7+aIe0bt7(WT_%FU+Q}WWynTvhnraJ#n#9NoAO9;BSQiGeU;t%j47YV^164dv1*^ zEiD8luZ72vW744WSwBVo6Vjk6PjnYF$`326P~2Esryzl_Q)%P7+2>AJ*4Xn|pV#L? z?@QEQAah^oo({M(3YuHuWelf9zS0=+D>!tly{Sa8`4T2zbm z#m9DKYQwbb3Kp3haHE-akLq`EOxMv6nkf>ScPY(M$~BJY!?Ieq5$KA)e}6k^dZ$wiV*a89#BP`Pfynx9r&*++p>8@Ma{5%nqggu%X&x^f=$bZEA6tL6XTZ$dC~@YvseDjkLP=lG_1-Rdj;VOf9>yGRIWTX9sgb2~ zTK8)LQt~KD@_6GYtsjA?QVIzBJ#jOuTr0szvJiN?{EN>IRQIwhrl1A!G~AJ95+!-- zZN(BjH>YakR_v<8%@d=I&*JGxS+G`EPNrHV>W;0w>)kmz6gFFPX6ZIHj>?CBl(wOJ zn7TX@V^Y6#9A^h#i}o(LTcpuKliertRrvch8xw%$%abNMZ5#NXf!|d&3{Drsp5e|M z!Oh0#Zb{3kUA4L$aQve{oy#W;pU|X}bxYxUgL->ipK+38?{)P}UmFCpym~L!k57_m zDvVdH|1?}YW`#6dL~N|N=fem~c`zm0ng8LYrC#sqt${#a3>;1Mn~gH%$)w#N*0t>A z3!Z)es#x|ra*^+?JuaOhAp3bUbY%D0vG$?;xPK!GMh{=N3?rks=nc77!p!FGi;2Hy z6$Mj;XOo}(AXV9@LFa}uwYF=O2-eOP@<{YAYz@*bU4WTv7usIkFmZd+Q|Nztr5MA| z)dhy9SoLr4)csgHKMC_);|-+aB4r<$O)IldTi=o9ayt^s8BEF`#pIJ>c5nU)&x+}S z6aqc;>H3!@02D}}_X;^500%@N9Pz}@iV<20MJN}1Gdbv|@4ewm3~r*nhJ9_gNcGxq z5&N~_A}*xiA^`r`;T*Cqpt^_1z=r1#71+G{h4?-@w8aY(oDKbgwDH1m)rr7D2KUe= zwtw}m%LnDs>+N*MyBkl$>2?ZLved7wlOkXh*g4p?(TnF#ah7^xYSzKO=Sz`~jK$wc zpnpK0F|WJsU?Z-XxT!-`xX{8PUEKE~3kHsg@2Y^86DqlR*T<{8yr;Mex*iVI1)jZBUmV^b7;*#-awz@nw)pN#-1T-@Sta{R zt6?4C)Mmym&z^WA4LvjsJ+apvHG^Ta6{V2A824on>H5}%u?H9-hHQSzL1YIMlg|5o z9Do7is~q#9j(OjfQS_br!_Sm_0Ala;^7X@-yw3Uw)r%!x6ByXvB6qqq#8BG!{zt|$ z75dEC?2N&egVI8<67Vd7VLe;c=wVf4kI(aGx@iE2ruTkA|E()8$qKELfYN|jcN(VY zFAp=X2l7m%xz(U+^4?kY?P2Phv3oXn@JAWQi*I(sxR&4FZ-ieFzfs8888Ohez|!-_ z(gQgG-=@xPnapG3<%cV52O!vaVonIID|`3A+GS^*sYdpn#e@UMe~e}gg*VW`WZS3H z*|UiMWq_$Uq2^PG|Q+S{%mM83LAn2-Np>u++7^`Yvg!x&Pohdp>Ahug;O&(ci6<4jcjvA zTJ?nUT8?AlQr#OM={kf_R{z5&J3BuEJ7d0HLBaeoN0nXWS4)xYCQfPj#3JV;bdr33 zqHHgp>q}Nl2Iz-I*(fTeor71XUQpQhFcps$`;hZMz*f(ukXQBFwzx3N89?5Sg`NfF z)|4C!VT{D8KB1WS4u|j=Mn36>VNv`t5{22|nEO zghQ89j^m3h^gG+TrMGIqIm||hoX>FulnXfRwcxH0m|SY+j{uj zxROUKw?(Iu3cpyXFX|y+Ygm3fmSYpqs2eetfoi6PCMC|zGa$8!CdH=CV|M{-%+KT& zmfe%^y8sGJTI?+c@q2-juROP~wLDpji5SITNs7l-qpM|i6+WvB$>X3jc?91x9VYUj zVfuO?X)y1Q^9$^>6i=j3sDgNx8`ktj4|GywlPwMKmt?#S0e2_W|j;fpQ4x0URkEWYeR;{=Qw)WP#UD zR8Qpj$w*k8{nL5T_vBIl*aP5JKfvrjLFj=ZXRtBYmdFS5&|UhTYpbRMO{jzYt%m@^ zPn%ve(bgeBCYv4=*(}~Vn_E-Ip`d+^n3!rCL5vUyE?xsx27@}A+AYCCul6_)wkxJj z+myUSk72{an_9)y!bQY7tXjP_9C6=9C=9;toyaDCw$xR2yDKy|53_Zzq(Sz_UCBoV z1y_#GTt6wze^QoCacFYteX`cFv$P_~&#O!~dN9OT)%bjUf5k7zO!;^WU)Vva)~IO2 zBq5d-CK^ZlK~nl9<-7CX+x-@AXJ?vaw|4Cu{c$Ck0v(AdcM&nvEv&XHmdCSNr2`-{ zZC+f}ub&RIj=IjvXHZNl3!iY9W|{cIA}b$_JwpXoL=_U=A_k}i0@;QD8o#dc< z^c!?YbIBS;NHfjvY_(^~heX4QpMiEMrTZIxU6&e8iIo0tTY7t}EtZ+R^7J{H#*tR~ zuVHIwyaa@0g6Ig3dC9;0-p8|Ho+{xQ^P9(^=2yfh9b|t`q>OV_wlje2j;ReT1B&nK zG1LvVO6nqF;+`H2+X?B(``H@Ovx4rL{u|}X6sA+_xt#Xztr>s+sHC)oFUFKBEutNF zv3Csali@jV2rdupJ!C07P!1cLe!c!7*90&R&YSZ#Sf8S8^*9V4&^J$UJ7}e{fC`qv ze3?Z|KB+T$Wk@|BJ->L?rg5W?0NWtQX28(wU{`b?(TDG8i8%zrR1StLEIy$u>8bXi zAXoo9Ga}EfTwmrUXopARm0a|B_)`RjFrPbk-wOU%1mOa!*>M!2LrFK7SnPJv&Qq}60ynztiOlY*@!C}lkBH^=pTp<4Zc2St)!5`om{ zX4Nv;rlRonNz8ys_@gipH^jhYEc!@>)*bnU%2r8u(3gW~ zsFK+f7WpU%f}Q67=KaP6ldFv09(#xM?0+0JErf27MX_vhjzQe7LQ^`YN)-4@E4XrxoAG`HKAk#OvCZouFQG3l<|GdFNSn~dm5bpz0O;dRtbb3uEkE9T z#ZimzUqKE8M+waF_rKB`$ld~`;H}6D3&v;n#uh6z2SFF z;P-in=ivQ>=Ieg`Kl>?8$ueYCVHAJ)&fmiiS_j4`tYs7-|FykgEwYH!Wpp!m0EX=o z8rgQ<@tow|%p|cBt&Bn>^>>m`u;S+m*?QivE~Xx|_5>%ra{?}o6e!m#>NInx*b+03 zm2n`|o}Rd|wW#|o>QCw70wT<2?B19q(;EP3((YofT`;Y2FiOKT#rxa4t%-v5$3)_p6+w*#MVZ(yz#DRZ9n{*Mv<F4Q0Y z#eb|{e+pzCfBlD^dGA87w@cbOfO8p2EV?=5MH4X;)~Xzv3Y=A%TR-MaU$M{ zc*B1fojR2rU7Yo=!2iLbX1ba5h#iV&HH#qOHFP%}gBa+wUVqPK6YB=J<~W_}4OORR z$FIfO&(`JBeQA~VXWo}jTfiA<{pT|A9qz zECo(E9YdFK+bArAXQ7wHwv^;dG+tBwl~%e2fM69qaT?yy?Hfcss}h=mpEDeAT(bV`g1(koI(?R}% zb1R_5b5n3MA8E+|_KW#pyodPKCU|fEvRWiPc-Bm2{8Bep>rR?Bge!^=*(3C_*NE35 z4wd+AUJ7amp$RWCfPsRP2E{Eyc3KOiN{}k>KD66Egkyva>|N+*&G-{dzRPKPGH>LA zajIHp7cjU@Ft0?yD6sKA1f%_F4E(#|vGUzaf3Ma1m)SK%xXZC5nx4)>4b(?cNiUkdtwyJ6Usg}jJrlpPurH-13 zFw@$JdA$)@iM=JnB&Ebf^x$EE0{SYHzVV;z7EQ>;6juc@iHB-^)*T#?DL3kACz$3_ z`$ir5R{Bb}PkymP%BfaNMj?wH^&J8G#h!sVKMwaT&UFh(Z{B=L5bmJ(jUrN4fKyu& z$SGSB{CJg=Ech7r!ekqFmrgbA`P_O1Ple(>=_A`yN(JU|e}QxNUH3DGPlua$Z?cde z<)It({Ao0FlPtAYfbjf0ox*;NJ=D{A%U+~M;NgP8Qq%hGh@W9B@+iuMq%s|FrLI-e z?V^xHuWBhnFW3C>gegC{)U?^&yIRZVmwUUSmW%i|Z=Vh}9a^GJjD#~}*0y-T$K&&Y`*(iQ!9;^$`j<`w%zuhtCc z6jtouR0Y%JaHRdKOuBo$bGXygNvAh+O5Q`_1!EK<|c6zG$1GLaucw`qV<5O+tPwIl%? zWtpLcsg&mss>bpTH5{-Yfm59 ziRB+wY;Y^=!x)}ys^&TG8eH{5yCVZpa$GZHCKY<3yMU`r!#G;lf=M%o9$=-ly>69`TatfEA7j$w5VN z;I_e2wde*hUxUD+MW(1eLg0J_iDL~#J_L{4n$QA(b~6NxJm!l&O=8hH7|iDR#J}N& zWB=oVSp+=Q@Ce-})Aqf7wH4}0(9fOtas}|D@1^ zisD*MZd>Rk2qE7)?Drwoxfk;OANc3N?iec(1wUVOutI<@0$9P{7Zwc3biaFc+8~2} zaX(iGc=>ux_V$H+0|I;|{a@URMrpbsZDBj1o)y?%BB8F7awcVh1S_GA=gB9{$xliZ zJCn&*i9~MVQX3u_`T*izoxY-(E0NZa%VC z7c{}_MZJh+m9~GjCnavk@chs}v%AS+{rPpEeO?KLD+QPsdEmY3B>G&6K{E06yOX}N zXAPx&O;5M0T@U756E*aXxbyp%!wW`IK;dA_`Mm=pDUDdxOH(oXG#HEZoaW}DTL zhD(wU(^A}-_{&-PQfpymUxQDp#Nd(dONoJuuY5sg@UuR!GyGXF@mb)y^vUc`WkIJw z!5!F=sp6p?rUltn05vj)>HT&?kCS2n(I9^>WfS8fyG}WNKx17Xvpy}SWvp6>fL95alKXc8E!iAFm14w?FbkDJjY&YgXq3Ah!x zn!c6b6BM4!%@qf64{Rv(I6-GY?Qsp*$%LOZFm3a}O`e44w+?s&!#En?cCup-BsRh066*dC4F^vPb%q475(y2qP2 z{oJdjgFd&fqg9eWDae&x_t=xuC-InDTNk_#N1iNAAqGaT@s(6x_
081HTUY8$eZe#>S@WdD1BR)i^;l={`{Y+Wp<470%Ly6;{`LnEX126Vz$3U zjFq(pixPC}`|EBx7M6`2oquJT-?(mcG{13m`d;T(K{3;1BY8BU&Bw%cl~!T2WeBsU zd17gS;qip1mEND}ybwAGVk^sMg>~oekKgl5!^BpPqlG@bgOHb9oLI+Va8}DMV$ITn zrBSBet1AYsTTF8|wM6(YE)mF6D($a3 zno8Pl0UX$gp-&BK{Ud1oDJ^AprcWAO#_5yZa^&e#6sKc~F`8=#;qQ;f+9blXyd<$U zDT6jfZz!7Nf{=g;+wl7uI&+jVy`YIQ;$IOyNGHuo$UMXf#DI zx@MC`gK%C=9QhK)pV4T{!k@9>&uDQp$skQqCWz6cdHr#kgR!yTJeOjKfFAl5f%9zqMa>M*zl4N%L(!jNlfd+l7kJ;h}bO)VT=}X z%^9^8qs=sqNX0xLGRLQi0~`ThqXx^Xkoe^ zX|`7DMY6AETE%x{ytENxG+r{m^AyI~d^fYY+lugn(}%!Rhqiuc!7gx5&-Oi*>wHAj(7R ziiAn^E`+fdScyXLGF1lgZT&ja^wIxUm=@`b>AioKX-<7PpxR~JB71u;8NCc*+F5Wzv11VlbWSO_CG zQd{HFfVS@yBtgFpfOy`S<@BsBTa*LPC1w)JM9js zF-mrR?aD3|QX6&Ve3%rizv{rYgDbM9F@9(g53w$~V|nd7So`ZnncXjBcM#&>7EEMkOX5X?En3y2&H$%*zxK zZ51XSr!vu=&Vn}K)RK6g_MRH)*$n*IEF7zo(hkAt5#^zUXX7+xV)Xt!7yGT1%6#VZ zof+D&SeJ}oq(}c|F|6IZ&JfT1y!|@mmoGeEx!z!2=df5uv~v~0-O2Oz`F7;M9}DlS zKDxUu>3Y44pz%$Q79u#@`fL4GGYs(&G-`V{VjZhKwMuU?JX`S;aXeGq5cy4znwF~;(B=0xug zl&jnQg_t9MrTsdqA0{V%zj}plhx<~p<33tg;?LA)M;EEEJov&~pMT@KW0EBN@p>dX z8l;6gb0qlt{gO+(BR&TIf%5!~ek{pIw{;KGXZudA?Yp<=X<8bZnf_WGb$N-|5Jr0K zjGCnwp92&9cHjC?#`JYe>3igSSh|z4(w-0NSzUDW-7RH`U)|I*y3GCLk6ZvnU^aYx z+osxYlRL1tgz4*9Hqu2A7{(~?;HHOLFBHP1-{C}oLWJWS=} zC6$v6ZwF?L1uMA-(_fF(?@aQ`t}cc-hllTz>l4F>_NG7YgdWTO(=k2zdX$hib3Mw) zB$fYn>XA&Rx&H1D@sjj%ww@jtLaI+fMo-iy;Y^V5xkk%xCDvA7gsw>nYwM!|IIyA5 zt(8biGn7>4y*sIM!V-w|_m8dNq%p=d+%Sy1hNG{||MnV=d94_T=R?ru>;%|50J1M6 zLdV8Og6{f462jW7mIx3c0imBEtd>|^2T!(GB40X2wsvC0V`qMl$S{-y1RZG-$dF(- z3m|+Wk$obOeI~J@Ihq7xIspX|uqxv?5-4>76L6SJf(ayG#iKF_rgj2^?<7?cOe29> zCouE(p@;fYx--|tHeh(4Ffr#|M3-9N7U4QNM#hGwhIo)&N2l#o%cd4pE7x&?EUi1z zJ|Xh(*wOiI+HfI@wM{=vDv_>KP5YRJbW^7ANM`Xx2V3-gRFfGRCz;O>T|rE!Prkj#MN^DbgYssVTNl zQ)G;5iDIOt*g{Q-9Hm5yV5FwhLX9b|C5n-nQVX^5?62(YuLJKv9sFnyuY7<+iO=OH~iDINSv4z?sa?~VR1S7Rc zEz~A4#C5~N_7tF-*blM0! z1W!X7ZD}K?6rNTRPfL53HVV^5(4nOF`(lho185_-Fdk2g(!}U3NTMf+wHd#OVr>g; z=22Z3=TC}>zoCDLKU!TaQ=4st(f-tytI_cwjE=`K;W0Q?!SQq)PsZ^SCOnPleK{Nh z6P*b-o`GXECc3IjbcuW?;&>L$znT^6{KqrVo6PiF9>RQ~=?S_snCMRKfIosRUY-!V{#xH7 z(h+SPsX%c4(Sv%4-{I>TA(cXyP&# zjNPjFUsF?)C)T!+VZ*v7pJlB#*V8mH+0wk{Y-W#M{e-Ozp|4(ohnL`CT&Kjwa3#?n z+91)=%F)s`G`B$W?}hlLrPT(R8`tSJ(w{FPo;S9|EoO@2B+5W+ewyoqgPR^-(1fv_ zX=38Bbqd{MW@ut&tVIO14UP59^fc*I-_vV@GXdx3&rwYL;}nycu=dgiPNF5~ZLc(S zs=lU?k*D)w9H%kD@_c6_wp5!w-|2QrL`UJyzI>cc{u35IhdQ7hoIT=n6s&^ z1sIaG2svmrP zv9Ue3EN4F6<_<*egyeaN_R?>)KW!KyR+w`z#^I2dUWg+fgp8eVE=Fqx+K@>&VdEsF z<;BzV!f|gL6T4=9{2f|>Xe(C%(z`-9T@f7j#o8Ra5v$pLc$_GU_xs~%h#l?#JS}Zy zJRMI<>}+W(;B6RvbxeD%N}?r-wVJ*#n>=E$v@dA$|7n~J8oIT`+eBPx7Yq!qn%k;h|Q@9_ih9L6k zj8CR>EgjiEGp|`>J8F2Qb1;g`<>X_0-sv^m9gRV((zcC7thi!2_0nl+I;t(^bZmI{ zCt012A;Pro-wl$~HH0ZNGe1J4Jcz9YAE|Wg z9S%9n!wZoEa|wQRK~8YWhxN=mJ-$k2BeyKU)8Ft;Psmdk|K)4GKWdUL&5bpvABc;I|0NId9pbdH| zm}|EQZ0{S9Q(IS$Sm1Y4G`WPacW~XsTlL=*L(;^EEgtxoM2pMmahwe}&l>n1V+DC1 z%GQ>CB|SaYBs~49rnYCRS$4I14C2VP-?2FYaq45%t%jnyQV3<(oLk~m^dPnU?0icCaf4B zJ^L+bX_p@(d9J68{ar~$X>$`7)^QciRKi-dcITcrdX);u-2ax&k%G(aXP=Uf8t>czr61t>btoe*N|4yYl)RK3nPJ6 zCu3Z6C3-h^CG=gze{5I6{E<;5QGmOzmj1DX!kU95UA3W=1omk}}kmG8)^wST)l!VM@of-O#M%%20A751yFX zJmVdX+C0<$sLeC|j~EO{f7Ip~qc+df=DC|3P@89VYV%BOo=I~ywR!&kVDsEMx3%2& z#H_euBfYYgX&zdK8=w=4$>Qpgxs+c-eOr0w30d#Xi^KSulsu-hA z6;r2*Conz7`K3-3Q>Tjm41fQ@Q^gF6{1#S-XojPttx4z{wWdj6nZ~H~=np69kSjx) z-X^;G%{*?=Go;z>w`?hd=#9aX0E=f&Q{4Jg7TlG^gBx9p;HaIi_3IfM;9jxk+?pAe zz~R%pWqV(m!R1ngsv)OK;R@gLBdLxD!L{$mLC3y-0xNGp$$=qp;CzMD`r+n!JuW*aEx!e=0+8BI9zl&1BG6M z`${k0%BX9>-P)3*N7E0%@wzovZFmO3O~3KZV-M*;kkofO<0J{-Ju}F;#8@4ySBNI< zR{aS3bZllH_PYY6r{v!!5B7)Sy`O~MtX>9|y6-cGBvyj;pkTj&H=N;U-KMeUO-RZXVLtHmaGFHQ89{J)SK)fjR|@VPIPSqGEDl%e3T8TO zNd z*I?h4IdF05=5T%i8whFOQgvTF3ofa=cMMs(8}6LU3@eD;4d>POytfct2|jM)-^@Hw z0ilj_o1&8aV6H;OTJAALpmob_-MQY|;G*G-H%=#w;8|~xXHpyU;J~o0qv1w5oW6AR z!0Y}_5Ncm^&compm_9%BIWPJ#Jh0s$pLix3dhe(?aG-QE+*tHE=SgHC9Ov7&xAKuT z9I2^3e6#;P(CYnI{%}AfSUMUf%^WQRVG~xI!IlNU$=x%zOMV&w=LJ`X?udN|Pv=)$ z%oV-~p-F4?*N)_d$a@)jbKG{pBdsZchk1+O_RB{+MGd#W%QE2Y5}hEhRJ5J2<>+#7 z_~MWWd!~Z*t_{w|mI9oYwA$kOWD&S#@4WCs?-6W!Ibhuf;q9>Eg~O4l+lIq!uY%aC z(P!YGw*Hl|qKe>T<}*=dPbN(2J#pabz?)#p+3)jWUOu=NpK?(1S~c96b18e>_tqBfBJYJlgrv;8qc0)(9N5VZnd9UkrnqX%6ZtrNGWH|b9*z=|3 zpW)Dl$wrxH9Uwxt)XDqNRyaQO%HHV%_o><}k_xI7;)H|jyUrnA4LCB{+Vyrn@_u`~Qgi{_`uqTWWfp z!4E$09=s&zl)x2OHZLq+KwJs@Jb9iho0khae<~HNL1*EnenHrs1I}=2*~s!2eNxe{@zknI3LkJ4>vD68G2s!h0E!;wx)5thAoobnxpix!L@obw}+AnOb-kB z^4>224pm*8J|;m7Y;NAqzdP|KILz@+;V~!!+ajqxk6#=DYwzXvjw#(${oEL^R%H5Z zIR0RYwCRs+uus2OS!A#dTs-NiHp1r$EUJ>4ZfoWPQztdL9vr?MT%QfscT{(WnTKBP z+8V45PgH9@?puBtEVoDK9pmK(Td(w|3uSwOWs_dQ>?^NefX&qFFXqgI`%mRoh??@i z3IDYN7CvKxt@*RrEXQ2{$4{R;9@g%+5IJ1cC)Uvo!n3Yugt<+Di*KLsn7z`1`#A?fmW`edM$^l;98ly2o3F=p zIdxUwWRXzqmEBSB>LKzB5KIF5yGu83bWsM!X&U?0U$26T)*81pp6!N9IlWe#yFUZ2 zvW>ej`R#M?i%Tr7G_(Yr;cN-RcaDeCJ~>76r*nc|wDUB_#^o^gt0(t@C+?sknsbi( zvozR!jpmNu3UJnE!H?MbfpF@S9S7U8EI87C+nWcO(%_}y>n>!r5p0x2_x+H63{NH> zMN%4Ea5{aPj>`GxV81+mF;G%iPBhR&CAoSscNsH$2gMoN;U?V$%{SZ4epw=ALA4b#O z9X|=S!rWTwrwc%a!`A*2&qUbbG-#*nW=pW1#Hly_*cRBrHn)8G-lOmg`*d`p;(735 zEB%nwcn)s3pMJ1GKMEFb?;6Ufeibf9zjEGbE*1!o+rV6MRjKE z(R47|-4G$Wt`GPR^#h~f&Tvc6VVzQpDV+Ivz0hTE6pYy)lU1I!2(G9H{}73Zh9gFA zt}Z{N45s6*M?6q|4hg-#TRAlM*E!B^YzsXGg6{nzm-#mCg3EP2%Zg542cL9a9?hC! zxMg$V`g~_Eh-?@yF#U@LT#{>=x9QDAc&aerQ25#r;Jw3hfsp2MSUvrf`k4jsaNA+M z{IifuFu9=rRzPbw#Cs25w-kK{;P7_EPlJu19TWEU;he`XAz#(I(G0dmq^R2>kmgl`iwP^O9cC$ zFQjaeT;M@52&wme1gF>ejSLEuf}2rqFL{_|f|+}gp3%)9aEbf5;)GUTc%zaYh}Ml! ztyJPNIdY+S-Y>L@I6GtjNO>eIc(RHOuG|aQWgn{tp|a|a8!!9>1IdjR>X!MiJZbbI z9iFogX3N*V!o?k=m2Qb|jhqITH;4BMne-U$DsfBf&pn}9P;lettLXrt7p5CI*_{Lf zwM!qcl#{T~;U$M~qbnRAw4@*ZIbS$?LTI#$>rFWGaZXgWR}<_Xn>k?p_DS&gaLLOx zZW&-(JYD7sml@c77@#q`sR+D0C06yiod@BIv-+$ZyC1~Po%(7iF&Ykklsv3-`>JFKUAfGeGa1&a?p(j@QVF=feyCCC8x9wa=f~u)d<6dWdI#5^!eF+}@?D=Sg}^8~ z$2!<68ZImP3ZGC{H^wU{0EvOK#WvrPv(&Q{eO5gn{5^^1Gtuwf? zxd+Q9RZW8j+t!a#oqHbEEgYi5VPgpwe@@t>lR5-^#CLk+)HJt`|0ApOUQS?*y*~4G zpA$G9th_BTzWKX8-*0D3{{T3azi{-{AIIS7tNzJiiSdXL~)~u5|=l zH1fF@+gt|sGlk>Q`K7@!vU;jKrx#q+RKC4+@C)!*eU{HXeK3T)VH+8@CKdd5To

_~p|@yW3d&eQ2oG*@Tz2R7dg8(|1@2v$AzWtqu^=iGc`g;XOn4 zy}qhonFmsd?X=8q{A%u=;roxGVgvy6%_@V?PXqws&6=PwYMSjqN@82l;ARGg<*;7H zg_8{zzh$YFQ+#e4d%jdY^et5XWr-C-dq)crxKM||uvS1{COH%8-Op|k<5Sfr&)B9= zqm{h79M9Ay4da!gjWuM4O02sdXUc@|^m;Ri(LImu)bFi92y&tx_8FFcSx)Z<_@*c~ zM_KkmS+_I|dq--d;kx6mMpSEA)YOoXsGRQ8Ee9)ks({>BmaMTtVK{ZZK`f+t(F zfE>eja~V@HI)L)fbHxIfYkqc@fCiH7 z4M;sxY%=R#3ujU>X&^G$i7a5}cV9on>dQ~Rypi!mp1zB}4vf`j!UW^W#M~|gQazzV zzz*U|N&(=_L8u?;I&MqmtyIw-?dG!8=z|~>Xaax#rYfvqMiBXg^0BFIQl8cQ@b|<* z@-PL~&l!Hffr)!D{8xzGU*$FBG7#~Ya(VN4S7k-&{^D~w0b=MP3A`-Fw?`wI-|}PM z-88VDR-wg?VdI&L-j0}X4YiE}HV#tDFSJM}t)t)D6!5Q3;V*_oWQ0eQeV+O9#~D~< z+Frcs1nHz;8*vaS?rcCXwLN)t-xhXUOP2EBp}Hy(-b{?~l@e9lF@c=+$(VPVgedT+ zVPq=Pdi%!G--AF$bwrSBpj0UmXnJPG>ODd}11dN}B|Te*R;ny$}9HP6<+=%(KsjXP!sIA1?8bz#q;M@>K^#?rVIm zi9qd$EBhG;13hXmt`-4HIEso?J7v{*j$%D^&&g~rF@-ghb$$cOh*4s}f8^&U45M2Y zD26o8Vyvx^axNrZMmm{qC8;)6t4YMZBW00D4^+Uk3Ut7z8(Fn)j?@24#l1v=v&y*y z{9L^pWE&Ns9=9B9rynqvjEZbhm)7@|2x~J5_p&{-&3xEX9=4pdYUT{*xvTF@7z`Rb zF)By?B>wMTNS(InrC*(oetwz?MR09AALqSf@Kv^&I}rPJF`98|q=!Cn2AvVrZG#(+ z?)<(lIzY5g@U1QJ%14LUJ{UP(>Tsk${|mmL#hpODu}&mn&fEnU)F;nwoUmoGZPef< zL3@cXQQhCZ2~9rsvtBSWc`t%nZ@ZPQFl1w%xBGIxadEPCU+*&qY*X`~^))f2#N~4Y zDz?FzR4)HggGe^PguZ%wm-*%rVc}x$Q~*Wzljyq5ySJ9rW>{nS=+QJHGe~FGc)=|11$p=c`V(aLfPC3~M9+3T%*+YK2%9S@r3f%M#>l)x;29$5e`=lA0 zHk0GzCl*lq+yCl)f`_Owigm}0Ke!cDg6Pp{f+~*-G(V6)dQcA0i3uH$GSinBhRsj1P{$`x7@sooXzLOe&sw4 zdIc8b7zARj44PGIV^pGvXMCOz0qcdk<=zngv&G&JPkd+ry|NU@AyO{uGfjE7l}_F= z4ZI;O`lw)>0d;(&SCQr|4+elfYF-3C{VcP1HVqi;2fhe|nKnlt{0dVrc?JOr|Dg@8 zGTCd*4V*h(3BgN58JH|3!0LqmK_xmGeEnEoBGdOS*m2zNDXK6Xpr8 z5Moc>V~TqgWR?i#L!x@Jo8gE(V?Ni5(6j752+kb!G^F4uuOU_9F?%WEa~^%BD9p3* z#cwufS5NvCr%yjebmDt*s?nrw@FP@~$w1@n4eYCuGqAjU>{I<};!wZJIGm+LHl*Sw z`YX(Z8cm1)zdf9?F;2Wdu09toi+_nY^l{=VlhP)9M0K^o^qy}0TUq}i3C%htG5>LK z%3!B`)S(w|hM2UXCgw_S65r}XjcGYBl<_odX=*xH_p^3|-)lLv`M?vZ9w8+zDsyu1 zkbA$)iQ|%g2xEWMbN==JT>(xH5*q!raxq>HLxiOVh^=E4W@w65skt&ULi(4Dq$}RS zR%t#t@B-ej#$NJ}fYz?$+{_1;f72TWNNlW3%@D>!=QaU&bV`Oqs1deEc!gmFz+9BZP72~|BI+FX7 zbd(UbL({E(dmhm*Q2vifRf}jk*=~W9rsLHr;PqZn z`we|2f^-{U+Z|j1oRtM0&up(wH&ES=WPlY0NDlZRibbO$bqAQ+ zkRIavkxxm!AQmzsnds$5eR>#34Ospj0YC9)bhQ5cc#qLR4S#6Z?S!Oz^WWr9f&O;J zljd(S9LAG6ctva!ArjdG;v(i6uN^m-sDZ~^uab0xuu((<=bWLT$^fP?n1(CZu8O#a zoj2Kyh@?Z^xvM0zFFfV3NT~mc4mpWj9)->P6CD!2eN*zwNUw*7*Qt19YpiszuZnlG z?FVbSL&(+Y6x6z&re?`p`Y$zS`Oh(rR7Z(E)BUEJ1~PJR1K9M#SzQ%EX<8w8)lQ;QGCuqg9o$ndZPT!3kzk%aUunS1z}x~4^mfRB0wkMjjMfM58g0%xXpF;? zL4AEpVjl7h_Ex}^_da;I`ALWb*vQ*ia1RiMh`udeWxvfR>KBzED}0N&cIAj%l)Sfr zYq{X66jbeen9%Nzg@L(mQ&D@-WWoleU`T1=oz-E=(1yvM4efI%qHBOmh zPqc`SxNV^s3)xR9Ds6^y7d-BJG-pJd7fQqz1wFFu`i#kv{oL(`tg6L~NkoZe4u#q! z-isx~Jtf5IK)*V5;C_{nIC+@WJ0E1%^|M?j+T_l`816LZvgpDi0W=yWQyTygp1tcaE-hupO=@AKe<@KTL{%O)(*ir z5Z&f$f7$UDlz8jS=Qd!$q|Z)gq|VM0wbWJC%PIJcl;~P5S^PO2B0zr4S9lE);acoA z^S?qM;4=L|ikIg*acOujv1=dmmLF_}ka0J*Nt(%%5o;gg;}u4TyZpq9hIF;Xo2NH; zv^&|RmSkPE8o&20acNpwg-N;<=>TY{#iVOsowWW{pIJ3YPf0I3yK+oj`p@B?p39MG z<@PCvk0mH!W9v!}SlX#TiAk4< zTbFB{=@*n4vbf={N|gQ10LO`0odNVkOk)5@ig%~lqXpdv@2mF5Jr}W0)zPBGE}P^1 zY{$x7>LHDh{zN!7+8&r?a$||6#oEIssJ`%lw!wAbI_WkI!y{6~8WPb+2(6H;nV?w! zsKys}8V$v2FZVTR2Jvy4-a^QY>5f68w%|+4;w1#^$iI3!SRg!z60-MQctgup<;|oc z2=|;J*>3pG07RnQ;9Cl}>HDY00mSL5d+V(e{8CTBA2oNmntT2Jt9<>~*}w%KV((7% z;JzCt)yl}$nHagzVX}9ufy|bv)}OT-J&LmYBxZAHW%i^^=R!Xx`(l(cZ0Vz9g<_o1 zMZXNsGppA3EKs-A_db}N{pbM6EqLS$Q7WHO3ZI}pG@>mDDX%qjfc(}V@xBzx>Lb^^D>$1_7WLFlh!lywiV{A4g5H~udU=i6z&u{OwHU*b2&pkGurAc=mh z^OyjLku%usOA3zP@FfI?cE2FsXT!ZXZTONNUP_(&z5_G+z08cgPV)NIufdo_&zK3sLkoDLfHGx=_H-#Dd_p$33x$>CI3fm>Y-KSxw8X**LbBOY z(~l$eL%j1dy!`#*AxlI>Jc z@bxdKOuLjw?4+;p@?{G@a7D^II}W!^Lfp+p-sbZpSUG?M^Tuw(2-4xKq!Tg^Q!fqz zoj+^2Voe7kt4v1Gq(khSQR4GF2BpcU`F01*qEFxbXeOLs`5k`H`?}}KIm5|211A%S zOq3(z^;C3YsRg5vDBDh?cpklEZV<8Bydh~lM%+#S^|PNOlP4w3QX#Tek4;HeH#W{5 z$lcu}fHX@S{=m=n&<;qO`NIqMmgC`%eA40kGiu7={B*X;fn9E=2FUJbr$T`*+LL9c z){WJ{nZ|)YXIf;b&xs#XpiWFLWzpzRQ~5cu>cIf(WEFy~%t~xQjqauhz`b%FinSBJvrsO@_B?C#S{Kz4$>bK7K-Ejm*t|Ccx9Wsx} z=>D{V|B6Yh1(c*;V{&{xj<#nL96?taKojFRo;Q>qK$B=eD~?cf3v9qAN#z?K`b?koK5^CrYDYz8BfU>wtIuQ!^hm+O4INffQ@Av%N;r?`A|gi<)alJQ548i1Pcx1YAu(z3g+8Fl z5YOrTMfs%-@vFTJ)Tp>_W4xow53plLBhpCX)e*BWaRr_}?2Eo%^t)#+KbqxTNK;Oc z^kck^X0{3xV$A0>39?zKEyu)-OX?swVM5t;v;*MEGg;O3Ah|a{3V?6`TW4+juNqA?OujYaox}*<_V9#n2kqL1=bn5SblEeVlztvtw6{3b%|4 z^hw`XjH~$6I-^jZJPU}|I!D&5Wo#J?la-bFF&lU>CiZE%Z)AofIxf0sD79$Fk^@Cu zP6JhEE@E>Q;F&*WGIHo|cNB>~j0q{D7;)5Fic}JaKc)eN?^ml) znr*j;SBJUuD{+f7oMx2|=9HYX(BlH{VbA!-f*!3^JwP%e7F5T@7P=9{c&lD~en^cP z^1(=)C#YLnN6#r)i_+QA(%HCpW*XxMjz6}0jtf8O$-9iy$u=paTfx|pIj%UTTl+#f zqj-wX-tRm`HnRSzW%7HnSF=@ZFASdWeMkA@g{tO*gvYsKrk_oDiEEI}3F5VRVrQ6- zsSZTZS)QZgGwx!PSV0fM2&P zekH{bg;Lx7Ef^6^weRqPp*uvipI@|*RWvJTpiTLiid7k3V>6BQEMzqLnvJr#{H$od zqMW(0V3pdLNS&i_g~_WFn`6oO3wd7k2d;$_1n%FlX97C{GHXth7JdGiK9%S2)Tlzl zZaf>1CtvRYKS90wZ_6Wcc{+IIv4Ubvt-*FEKg>b#gx&HqC6NZZ&Ui-w_C9zUCdjt6 znm=-gC~UEb*=|_vw>LhA=Os~CAhJp6;v+|`Q;a21t{6V+Z4?3fN$vO7kt&2%p7C8j z_TBjNi4boEJm>Y@1m>xeL6n?33~q$)+L;hzhl~a_x~16kU7>_8265wxJNdx3?gy%Z zo1m=^jVsE<)F*5nEgsEpW-#i;!YyZD>YqmTJ`yASl&EidApIhLE2Uf zH0JstGV^;TQ~gijQLrYowiy*;sOSXk06TnivM3=M<7u`Y1E+NDB}d1 z?#9+}jKA?hV5E1aKokCbmOxWeD8=uY{)olE$dm5=+m78J4fvu!-*5gI^*x%JuL-HD zo`rYDepnZ^J#LIa)gaV#>2fW7gds8jghPZtIE3N@IO`vsekkjdNxWsg>Issi3YxF_&}qwv6%@bE%jx8Pig|E$<*3?J zD>VJWB?kj~hF&jNC@{;^UubT_!ws0BJHiMyzKv-SZ9FD3;$6B8jeGoKW$IpYP;b;e z^*n#-6Kr+!!Tk&H@|&rt0=X@uyKP(e#3`PR_&A+q51r+!xU;mYG<(|rWaL3o#+2kK z$<%6V_wS4j@k1@;V#V_EH_!h9Hj68tR$K(yMb6zWPkQu&mwL^$2#o1RC zg8x4Wivm>SAF^LpSm-dcF*9*z0_#DpE^R37Y}9VB6C?uV7Yq0ikAbxXX~eZB7dmuf ziRSdl(tkv3fYY+g4s9&8(+p!R+-Jl4+HcG}#htXkAJgPI;Zh*!+r(@ONLjFDsVX(G zGD@RYYbWR#@(oQlg!BAQ(6kcSkkj1XxZFGcD$+o*RY?EhH+Jv7zGkleJG#Z*%`^_n z{3`PNgMuKsjyjHl`7Onq=kJ!<{yx&M$Jw;4&}*UT3LCjM2XXu#&CF`0-i2b9Z} z1?u{Bc#GQqMCx9>owi~~wG-6;kp`28Z*_L)D^NDhRW?=;-mvi&c;|)&e+v9jy@ z*>);WWgIwCP+$qFt(V5K>e%YM8@`K7R-ktW)#ApjtLpb4I04h7^hqsSzeA8FQhCea!(yH)N%*?yy zjSUn4b3qKcAPrULcb=0jl(_3;brZ5254yPzQF-jiB@7d79m(@W ztHHqqmhJA^;(I37O$s&o2Q+%RAvg82s$XQQe;gwozspZGb(Vl;V z@Ve$l-O^dJqv+4>4en15H$9g>Jr&pR;NP7)xx3Ws>R2%0%9FVM@j8DK+h7L@0BbDV znQYD`R~RstRG-5LN`KF!!LP%~${QVgXM?Mp&d53H6l}IiUx2^FV#%^kGJ+O)kR#4( zj~`{ki*Ffwzyhuca?9Y@aA1=%=j?ijMcyKN3r{%xarFEy6}qNAOpxczq!b{AE?E*K z=D<8GO%_hVciAA@y>)e2mHMxz1LgZVKtfr3{EzxNprP?Zb9LMRB(mEMwpMrweok_1 z1;_}AWvvEEB0>r@j0ZJ&`k?)6on8>Ae_<9zrsA((cTWxS&2H=c`n;I^qsp$^b13|2 z5y06`WJUt1Dd2P2p@c=&k6}PJ_8Qv{SM{1HB^6CvGHO=(ifdj{u#{XrLQytisD>(O zrS=__-b@X()X}-jF{k33<}RV4iwm7Qaw2DwEh>2IZ`c;27BtOZG^g{mK_){yFB0TY z?k>J~2P0xDXSH6Qb1cKS zr;|GR562SEtxe)E7+I5wb?$8>6i|`T^4Y! zP%LIF=K%h!YmL>fNCNiU^qYY)cW*>Y7jgdFfIuVG1*+xxj0}|wjDLtAY8I=>$s=ff zbo6AZRkumCg)VhSx_o?Sxu9OAgnOlHvHy1CW8)~}SS5!BA+QJN*lT)q;`r{%fFib; zU_XL{t?Z2L=f@A|FK+ZH+4-WRrN59UT;d9Y9c?8Nu$FleEb-QK8G4!EtWDio@tc&U z3JYkum-=+Il1dU+PS0Ig_SGfRo9Y+M(#tXuJig5~`BfpE>jRCcQ;c?CSLq$kPiC2>91 z^s!MJ{aJ&NIN88y&~&{~?cRW>r`LiJQfvtDx7d&^aVNydAoYy;t|8h}&hRV`OIhf% z2oI{K(oa%5(Obvqk+uXmQj>EU6Udvtc|PRzgExOoWl{i?p^Jn_%IM`B@l5tek#IU&UxIlj5_+g9t68t6p0 zkXSxNLTA3N-Rp`26&CwvcCxZu6APT&B7Cr|cK2w|n;GQwgAvovb{iHu6rnT1%Ir^*H z7au(72Vp;ueAZqrQ#R}iq>>}_JzyNn3_T{{FUXAX&<0&)5XhhHUsoyeq#xjHhww+j zZ0WQX{hwM#j%~j3`Pz6+I64&9H*~!ZnsGfCVu*d$1dQ#K$f0FU0_D9bUUNZ^M8dj+ zA;=LDMR?%?9z{>MvNISMaAchY1D7#V6^7AOI-V`~MXylAwn41pgW1>w8II^0eh0tajwqwK zorVyLHxETUMci-WzTMRCc+9b)0Aw>g@2?zQ0QTK z;%e_`oSAdw8}3{}W%7mOYU5jZ*hUXe!l-@$AlJipbB(9@P(tgg3TmYsy`zaw@oD@G zn#kHWOa1aZ%+gVt+?^>8?>eHZkWj5vRVcPtoJEi=C~2PXjhjMXh~;L?twu_Ky<_YR zU<4ELVwET~IBgrjZHr%`3yY>1LY0=pG&aOEo^_jrsq1+bIs4(hZp}22_bg`x)lN35 zX0voJ|JvjGV|4LoyoDCFQ$yqU(`KZLW~J~Qv8<9qR(7%k_5P-ob1dq9;j?1v%x&pwH3SEbEOhx4fWY)v;j z3;``sQIOJE!il!TSIChY4JMP;^+pJ5)2JP81+zuM>czL`C>?F*+$Z*;_F}>2mpukA z`w=bErhmkiDTT%0_%q)*BbR+zFB!XM`jhlCh`VI^36sclMQCwjuGq*Vr9(&O7v>s% zAOLX^(G5Cd8NZY9L&tJU)EPE^Baz1m;Smn%khd0iLd-c@n8qRGO8sXL1+m7I4`JH4kzZ ztr(PvHzRNhgU%|IKxf8F{9RoKsW2+s3OLj%EbxW4sIoJ}ijN$V>TtaysrHzsy!x?T!zVIJMe`#ssw6TIY+1k+N0mwdJVgGMpg(lesSBJG<{aCU3@Y~G_be)&J_-DObR>AE(0 z2o_v}I|K+C+}+*XgF|q4m*DR165NA3!QI{6T~CwDT6<>pUi;L0>iu+TK2g=_N)^?q z^#6J8`}$q=zKg&Xq}EM(OEn$G`X@X=?4TBS)`1;Av+Pl@dvSz#T&y zAPE`mc(1d&9Eu$mvCz(>>fiOq5plwqd5N}&^}tt z02y!?Pb8Hf8Az7wKKvk8%dzTLSa&$zVHyh^S9qcBNG&V+21$*4XjB2%8)}F~Y5Us^ zCRMWg_G3|^`zT5zBDMqL06wZ4vG(9C6cQR}QB%9E=G1hb>nYOPr|x7nN9}cj#Rw! z(D1^|g&rf@f3&#UW+TKup0x3qPL10BgO2Jtv;Nl>0eD?%SjL&bF2yuCI|tCDq3Xi# zJ%U9T>O)-s#jzLmtCMZ99i6zz;X- zHot7aw^PiiQQqi-D4wfx;$IInEgVByxQ2f{Q7GPARrvX4d{x-dqWC$PWg1-j8bL*9 zdc>)GGP+{>5%BZ2om1JnsyQw)eDvly!z+!LqdX#=W*RW!z8zWgWw=@@THQ5yV<>hn zWPdRfxhjs*t0@$tfCtE8|Am3PB%4hqUBkS2T%0pi=^`aBu&nGxN{^Wgp@CtiA)55+ z0(Ho(?{Qug)gRc^5gqEj7*V`&8AcQ{V(n^>fZ_X}1(C8St{E|&UjzDIzuWb{@T$Kb zE=j$54(fqCC#nEgl>>Bz_DXEt$;yNyu&W3OJ96DZ(6rEZrZ2ToV;R#zM-p=*et$`Z z7$Fa&pA7q)rq8dTvSNMJ(j>3`sT03d8N>UVteS_~8UuJ!4s^z(zJEC++Sd2uKJupi zCad&6M&48-^-~KwShHRnyX5u86VWT-L%;nMs7m7;&&e|^kdmzMy z=3>JwN3LvGu^hFmcF`oy?b(I!YZ{nOkEQ(AMU=;!6PZ4Z9DbD?mr_bge@o>&z>!|4 z)vC0WnC4Ndb%dgF-@YqkKroQmH&Hi)xpWNEXd#7 zs$~kAXLdhN=H{!@cYx7XQ%4)bk`^E%K=Tadvk69^WQ2dAc!{Lo)e$TB)QNJPlQ5dt>>Of5`u zy}8FMd9YY~>E?WM@ZIQ2j0l2!5N$hDGAr`S@`j$&L~ySa-Z>^xz(ML5#FVKST*!v1Qs0BdYym3CU}V#( zK!`}l57W(o<<0q(n69xv9?$X!o!&>*)O>Hd9Mv)TWET)Ps zUp&%sQdw_s9HOB*?~N1km4Y83-~rUkMP{_okAxoRmAU`!xZZy@P3-t8EAWq zd?iuth6-X~DnDc>%Qc7;ZJT5dF<>S$o@4=HvM~J&2WY5*yH1OA?s7DFMIy}XtNP>X z8d}CaW@)JRH$nZ`|5rm*Mq?#9b0hExkTqaAR@Um)zagjl_*QeFB^d9r<{2WLi|94o zGo4bkaZsvw9YkC!Ow&?5R|;__%1ZaZ`(N_6RKYr z&{8!?-B2#va%)At_g`PHMTX?aTlzlM-rZpub*t|V-Q4!wfL0)Aan&5n2A3ghE=6Av zG_gV}(g|iL3E@xqK^=6G(v+-7SJg@>(EwKry#q*^TE}*(({Dnfp*}TH7Fv}iVk=?9 z1!2#`0-za^M``_0wh}!zMqWSCo%zXToFRw5k!&&?{0FApEoe(0q{Ty`S}T@1Y>DaQ zH6JA~6@7Gk7eN45%dcq6bmq@m%JyVU^+7x&3o1{)+9%a%iDk8yy|-PD3dJVZm5W`i zCpdC;?PNympZfllA9_2X^nL+XfFNP$F{jB556^!Wn{eM7|9cr~YIGpIY3nMG8-lh% zRoWB(whHW2Lw;S>h`QnT^(&x@?C4sK=15}&8oI+I?s~4lmH>{F4K2Hnf9Wb@60qul zb@F|g2K9ePMLOxcWqCd+qoSrrP~Jvr0HTn=`V`Tn0CE^WDGF8rSSx^a_1~7mSa6OU z3w%#>U2oKse}y8I2-Uw)eiv9lL7E={|63-n+`e|_(;K^}|4|0*ybh_qOt`GAhfD)3 zn?`8oI}2c5sQ!DM?pNxRMXL=2_NX)-;9CBHM*_@Bt{WDdUpYuf1pu-2U(zvPf9~I2 z)NcT+Z-4E1<^SuA>Wp%+t%Y4FP<6}OLj}rk{_4A-fhG?BXQt)1g*n7Oc&q+=8G$2W zmA|4=GpIG}$Q?@?i^uC`f99$HSE|hZ|MPa6p-;kJw(9`zjQCIQ%v<1_Rtr>n+r;|= z!7}-u5iGkX^IACynuU= zK#*ITpL1WKZHVgkI*N8|wkPCjpXSx_zAU^3Q*>A@~2jh(>fsXfQO+Jc6?1puk~OrlnPeV}gn*p0Ld-P~#{xF!=l(CNSaC zHRkS7pU+F7u}J}^SGFsL#K~uHC)sHT2-saCN0V9^ z6$2V{L3ooX*By|^V!a#h3&oLZmDPdxY&qJr${jG?$G#C{m<6cO@ohbSCg`H!_~h+I zzUq@WIT&+vM)WJV?Z6(pF+v;`7$I657{-1(bPp9i4k^qtmZaQjOdWR-@N;-y5LlB~ zgS$rw;CVFxyi~8<+zjD3@!S9aR4E{t?~~JZTt3>od}2->vppYgt=4Y^`dVwF+8i*XfG*1-x&5i}aJ!szAO5lT6SXuyOfYaNdv8T^j&z;gL zec$7B4&}nBc%@_;*#%7z8s$No^@lG2Jo3^dtLiBD;Ni*5piZt>sT?(4IX?a%%+R?E zz3S%q>2Q1dW>1q^gIqICLO*ZDaO-$t;o~%x#oEyqSF6~?ZPXp*WQs7Gb(_ZZs(S}k zu+MsAQ1lsJ{D_$IO_+0@k%By%KMYAv5JXH*`Xe-nO!BeT6AgSb#}=!Fwxo{NO?X#F z*Q^P%RA@Xo5u;=NrHoR1TFk=njvd{OB3E*0;cK;|l9X5ro2drR9GFx21ugPW!gF<` zhp`Gfyd=Hg8px|nALAK~BPG%6kLip1+2Q!Uwk2z*r$L;Q{3fxC4+?@9#!^ zJW-x@+qvOhF4q9BPOfpoyco3ux^XfR;GGGn3(8p@G`v5U{SJue&2LqMejg)PSGY;#N<$ihOA#Ji;&k1=i>$xT2L1978o#0( zz(3QNKC}Lb%dC?KEA2F=!mY92E!Z;!5DMCD!iX=Y!5qEqyfFVZRG=U9N!yblJ?kDk z`}`?!XHy!=NG1?jMSHopFC*fBn;M89)bK2kq*!{gc1;_YGTI|G?%ICGlQ~q0h%I#z zd$)P-MqhAE-{1Fi#qZ;HYe~@lkUb&X=}mPG4~Yo8_47?N4yhoHS8I8r+{fL}!Fahtl|tK%W4i+2-ZWLr6Z*Mjz!Uk|o&4E#Px*29Efm=% zc6vXpp>b^nVTZBgf{n_nDbx!H$Y+W%hdT!dTB%nV)v3{fTADy;H_Mq$;g=}5-hpdt z-K97*5U^x-(mSu0d)f(V?pK*+n@8bize|?_V*{T_wNx~8OE49e?Zi^t_xKh@XOa2U z>QI4}PGYsRR36eYv?IfB7nB2d#5{C09%&b8D6Y;uH0{lGL5}|r+)XW()S`Lnt_eimiPx;e-Zod+V2y`E)y5_;gEaj;phF^%i?9fQtE)2)I>+$Txv2d`4>1 zgc|vDbxL%-S)h^$jLpc{^RV^_^%1g}*yRFP4o`ubKVNCb1O;`0Y8oZ#1R~e*jn9X_ zt*Y2iiSivdm3c=b_3R}-WGRmm6=k&FZzrczpIQ8f0~tP~(uQ2+cKHj?mpE7b1JFlR zSk6tzr2S&)7jrHf5JcyJ->KUMXG`b7(rEyhBq`M!A;_{60A;JXQHL0)Ua{D53{)S} zuuN;Qu;xJLh~JI#u0EonYi(l=qg- zy1chxwxb1OLyPp}<7yrwT;??wA;ZlO<9#0a_SF&8MH$exQ;x9QfzaH@%xHOjj3427=># zMj@AlNi)Dp)*R~OL~{30d^e6gJN1M44&x1E709M+1Wzrdtm_FelBTYlQ1+4gFOa!0 zR;b=^8NpuXB)J{_K$K|K%TfdLq8Ecdm#N8=rSjWAts7XJ5T9IunaJ-JfzrRG0IH7P zn^$Sy)%|QhuoAh0#&(w>{WhrRGnj(F1ZMKtojW%H`Tj4o733RjC3*Xsw))cbXWDAe z{Wop3{Ab##@ekUn;TLTs_+PY@;GbwKs;a-xR%Jo}+Uh4Tz0@-JFWQRnPqbATfVN7b z`%PO#ywO&!TOzLH|08XM^?ycNZT{f2{ZF)&>+#=dEA?XlZN>6NTZOa#L0bX3_I}Y; zh>X8!tLXott-k+3TUGqO(pFa3ziF$QH`=P^U$hlLX|}JW$e(GeS^#Z@4xp`wG#DZQ zw3YpzXsfw5+DZh?{Np$j)6m+?zi6vR%s1L<;TLUnA_kzX2!GL5P5+{;THa`@mL)3( z0Bxl-Zes(Wtz^dde$iHOzi2DW|DvsW!2bhn6~pvKTe1G8t@efAXsdI!|4v&m{hhY* z&YbW2H*Mwr2W^G=ztC1r|E8_J{6Sl-k6p3t_2(|d5)JR{qiXy#Qg;S9;|e-2 zrtO)DcU4GBk$wo=qhs^stsmQ?75(!hBjZIym0HBW#Gt6i(L4Nx@Yf)t3SOLxcNnRR zCl#S}BjHPromVe3ehGp%WIKKuQ0CG8Z`|rh)NGmhR-grc#;qc& zgU&U00W*+Z-+$v)=WnL1+dAKdC6-0C0P>L1+dAKdC6-0C0P>L1+dzvEUDm=LeP1H!v) z&W8aqZC*uzlbY%o={PW>8ZHswZh=s&;01WNF6tBYDoBrkGiofZ*zJa#K<<2z8zDkQk$z4 zdejNmAG*N}#Y~zs-AH3}L?*1*4$%bY{aDg;3?90(+)XN`v4$z$ddDULYZOf!`-WI{ z4QFpjtMVhaw&wx3yY%OQvOkhm~Bwi zpsXipq&6!}9NOmIQdG#lX8np&OcB`*G$&}2)#CP*OZ+zTfZ)`x;r@K^+tS%jz_=g4 zA^*Kz@pi*5QbPl=f&h@3qKNX3>;E;~-|$c(1VG4XuxfvK|1zQLcjw^0EUA?VJN*2D zaxy%PMr5mm4NVbV`g_R#Yex;ceAOF#`hm}&P5~AcKuQA}=Z3{_ZG`?gq1Jd+8G!!R z1%MfZuD2GezZF3OQ!06;j!5&sS^ZOI^-rDEKXq3B)LH#gXZ25=)&DQ+ ztnf13>_%Gb`3tq%r@HoAOQj>otCjBzicdoWfnHFR5^DN9O4SJ#zKnYY|ICA>XdRT^ zJ5PImtWiC26{=+`el0dkp;M&Q@^IhGG=|8ES5P_N%+KkB$uaTiEg#58#MiS1yv+-yKq5Un+Zlhwm=X)sD!ZjX3j8(P;uELc;>*=%U!Xg09SCoq6J2^wkH_~m$c13=e#1Ga=fF`}`KU;|B^voumDMq&-yn#T&ur*`nayeF)I zCl4TyL8Ef!oHEj*#4e!+TGyu5>kq<{BhK&+jn+A)kKp*0sT~_C%Yqq>+<^`+EuYHF zA)5LGlSikW6i0~ssp`YVF4hYv8@kZ^6ASEQABGaiSkmVZnllm+8SA1}4O$((?OS5l zO^nDrN~d5jv_&y~gR-H>v=R}`k17vC(?ZWsT{8mp$N4ODTn-qU_Tjn`d;xt@dhghz zhVqO*2$3ruV!q4;8XcTg_#HL<@ilt)hR^_caIfFDd)7*6}&D(C)gw zW6JHc10qt(94)xgvHV`V<}YL*CCn30LvdAi>zss2?M$NNUvL`?)2qSQ)9^H#*caBp z;Y*&cb;*u6*1e4|$R>%xu6jeRUCQjsfMeym?}d5Gf7gIvlM}CO$N;tgn@U9$yU8Gdx9Ut4Ttg0dtDuRQU-Ux_6z4I9-0)s3kF`lfgM4^g3OBTQf{dz? zHS;#^t9q}?kmsDbdtO!&*jJe!nq*l9YKXh)Ao|R7R?W5iI!)RXxN_BBT^$%dbpe`7 zj^?Ltj!ZGAM$1M*OH9!teCO%Z6US)ZRea;1B z;u%6YA$L}|AbTkVm^XP%)q%dpr7GE|{WSSl5SL1LV)Svhg@iS1b>Pl+#EZCx0}^fz z3EqcbIVwH_eiSR}6dpr|V+ck$(nuI5`=jQRHwS#((*>^S{WeFa$;oG%JdUM2*AIDz zJ{plZK*EE}PmzSAPZ2=y{2aK!m0;;e_a<4;mnL2`msh+PUD=O;TzdhVKU_Icc+1eo>yK@@zQ6wjb z&tY%vCDL~K{ZAV90?N7UVEdOyqSVr9vu*q_aFGwhFm%g843+(UWmRt0t+LdIohCuj znLLVDQ|uZAp1@>8o7yN`mf{_xZ9SXwe!Fa&p8P+!ADLw@fgmOxT~Wz7FWAdobXRFZ zfm{h*pMhnDV4K~^W)8JUlR3xf-D`}%ocdC(Dn(1@ zvl2dW*2LPJyzj=dz*rc?UoEqB;zT@uXDsv-IDA2K>0N|Zvx$a4!JI)n%?ROu&}e|915{s z|E%;9zV%f}P-Kxb>7$(1kIK6Iwko^v+EC+H%UEiv6xPI%qj)3n>5Cr6&YZFbS#4dH z5@y58;Pe$;R~AFr(q@Hgi?yAZ#XN{IGlj#8=hq6Y#oQHvRWZ`sN5az0!4JQT~)GGtQK?m3vmq&a@chB(f=%lVoPG$@AMG5>+A7rq&Arg@PE&} zDw9K5pGhs2@*0vbJgIj7f|;>#0}>4%WNfcmTx*jl1^w!_LB+?N;ly*di!5*NqINgf zfxYwDm$g%82ScJ*E&@t3uQ%<=PJZ*~R#IM%H;BEMgbtb_uCY$4Wvp=YbGz(9_4^S) z?U`DLUC66l(Ja;3y}99A%Kv1LE1tudcEshk3<3TGRA-30)Bi(w}ppgrs9uG|!ut@?! zjh(!J&1`9)%?}HtxRD#p`b^^66nXkGXZK`8Vu-zDXVuJ4>@UKQ?;$klxHw}#;mg0s znwwE-T3M%@wuG>NoXR7O4hprdBsk{~<2I0vZ{I8{mx7qJTHvkz3<6Z0;lHWWMtMsd zYffHG)-3vdu7G$3Z1d>h-J%XHxpubsXJ0N9quE;)3epRVET|V+nn0=w>I)}L_kkxR zw(Fmeij0qCs8|>JOXyeE_TxI|RZJ#ENp9Kw23Oq{ydFRCIbl3=OPWQ;6W0U2KQ00~4d(;weOaSVJ&0 zpM=Ttr`&Bgi_<X`KAYCu)K zpCvl!`Uz0-i4W<4xNBx=TBK1q!6?`Is+zxaQUQ>;y+UrZ(-Tt;cW9KCH&u*X2yV$D z-M7e~r-0Bd+7SFlXSbo5FN7e0nSORJisju;u^iwMiZj2`;44M-whVO<56f*w{zsws_CaRsbY z`CO%;`l?ELIWx896`aodKw=foohy3XVoUoS@SM+iiD8tU26vrQsEHl)I6f!L$eZO_2pOn@95c9p_!<`3u)z*_au@agwYJj zkS6_|qxgmioFpAvk6_8Ii_mI_+qi&UT{OVgp;#X;;?ku!-)okAp6>?FGBw%x2R|(l`8EFD<-M8~)l#T2PNnCyF zqhS6zgZ6==jAo;bZ;lCxx-HYa*~`w@g#TJ3yUQ5~D}IzK^I>}o7Efx5j{VV|cv?xl z4Z8%-?Q+Kiyah{v8)6~pVDydxG*S`X z_W(oGja%Tb8GekcQ^!Cg_n(2iB$3d=Sx@J(H%R=XMXH>y>tSYHqgQjS#{I2CUT5P+1*LGA3(vSOWbDhq@ zgNa|)pWt5W$IvZiA7p<0zt+c~t`e4W2WS7&URLn^qt0qr0+LOH*yLBl#73io>)up& z9^ABO^AKfcT@xkUSgN*c)YO4afHD-!ZomP(6tKJZ_je&qr)Ahf~pxZ zIHEM{vk4VPUZsh^5$jS`bnW9unJgzQ-K|9ll-5^BT?B0s3~+L&v}~#USI;Cr?0HtR z&>u+D@EOM5qi)&w?Y!7Xm4{Qf3-5b0Y z%GAX!%>~~Wcll6rzEG%sMPUDgf6o$HTAr0bPBF(0Dol?%P(}7GLPk92jO##M_zFop z27g$A%`Xm(i)e?mV(;wPU{LEz2f}h09d0SYl1;U-`x%~yb{5O$5)f2>_M!kpFBNLy zZW{uycxcVO3otzT(I0DILPqdi>{%w$0k~ZJPB|hgyVHGRXXykepK(9v?u8VHwyg2H zzh?SHR>J5ByXU&}%>_P1?1^&S$!NL5xb%7XrmnCb;lKN^1BS8C-^D%&yaK)4rT%KQ z3QL670BGyRoFv%9b(Y_-w%p^thr2V>`AIsbJ@rGkH$(F^bOYOGP=${*F4HYm)zoE5 zeC7iT_U3lH^eHKeunfI2{(r|0GU}ACs*%b<=fWdODT1|7(vMKGNAQo}1Zk5^C z=T&Vgu(ih@(=)G+-wiyBYMs_^jY!39HZUyRX}am6S7%0Bdj+TBDw#BwhHT$J{g2 z*rmL{(n#C{-cD4P?NK}4VIb%}1D|xPfz_YUkkbp39-d5uK4=?>U#A?QpRHu)1h%o( zOxgrDA-p`&?q5%-_}InN#S)ClKJ1$Ln+?`ZBmj=bIfL5=PnCzWfVxYA+iv4qOdaQE zMF8(b-mt$+qgV#0*GmbxZY2B!o&m>-H-#ARabVuG2VrRyu2m(uh&j1aJ7D|U4?nt_ zbi_SQfOpeY0W`~NDgPsPb#Y=p#I9cc2K>i1f4$sXI$w4v3O~Im#fC`ridAx67KF_~ zMbxdz6q­XK5GI>A~htOg0V3Ver-*7&g&=m4r+L80C`366Y15@F?DiKys#E;O2 z-t&)gmhBelNqWR|9g#8L3#ID;&r_3~r@{R-6Lx~U2)NV~4_XRAhd5y=dd9@#mT2~q z&*<#mOTO~ZzwdBWu)s11!!js~iYSUGV7T9YN~~h%nEX_hQkp*PcoX%y3chob6wcz; zoIUj_WW=vn##q!2POftl?Z4LX&TWB?JX%s+V)R5(e!^X>N%U~dmfW7nvevG)bz{fTXSo@YqG=-xg^LBfQ?zUyYvB@L(J`Q^E2E0An{I*)k715mu5 zZBAM}OZ*G&O-Xs{%U2QIuUwn-6@%OJAqip!F;!#>RikE}vs0+nwL(X<*h!@z>&{jf zQkOvP-@M1f6=;kPKP-R=L0$zX7(ZAnMz4v!h*F!rOFEQef?Zjz@#bY&ZY_i?je5e} z!q(^9bN#_x5Aw8{cTD70z2R+Im+!xpgyp)uQwZ z-I_vy=vi_aL<6dB1EFKPnC zjtM!rndWsoXd&rWAYvb-y=$HOEHb;Xrzf7a*NHk<#hPQRpBREQIc0 zE!_n`g~b*xZT9Sg$PP&H?Pkabq4&KdJ~9w0e6Ur$~#NK{FqtVoE?VERazhtHD^fbfz% zQ{uTV;aQ)K+1zfd1$Nqm?yQCqW)r|gwI`ylo@j?VWuap1*CmrF5-0o}dHw0mxphxW z*)(|1n4)Zs&ukNFk=Od9nHe*Rn{Vz6U&|rS-D#n?CGcQ&_Z9D}N=^`>_*wDG9fG9- z!aK!xRzIj}Srwj~6mm?T>c4P~UvGYAlq=74I{v_BmVs(Eu%OKM&}S#&YWB=RVnIg^>OV2_Ro;_Ez7{Yo>=6B8K#Qr`|b4vw#je_r-WHJIzvb& z?>YL`S%F7snbRf!)zNxZZV`ysF8Iq1>lp~Mn;Ut4tmv+&PmUi!_ieU-+PR4Y_GXu^ zlA%%31A$44m~SCHJHRe)K)w!zIQGIpzMiN+>V1AEgu8&C8w%}iZ9WuO(}xz?1FKPIz!950jRC6tJV3P{s4YJ#wDuDFlI)-o zqApLgcfJ95HJH7Vy=%=NXi$j7w9?D@@f&p0i_Z_l4FsX|hRXNMqS&nTG}kvc1r+!8 zQ|bJCvjk^X7N*>{4QJ*%kvFWyDL9?V6-8Y2F>7qy3!yJ}pe2(J_}ZsREe?>{uQ}HX zEhpDaXmzcV+sCr?GD3J2Pss zEvP^V^7os}^KLSh#`dM1RV_nv++ZK^)E+Hit+$4x zEPX`om&vR*!G-S^>tZe7t@EpHk(56-fOPA+!z0yWEfH9udn7jGXxhMgu}fmp;C12Q zk0_kdcXOg_1h~&uE)OME)k|5vyP$K1X9SKi1g{c)btGH@1s%pNR241J*D1T!8|xmq`LW=6=SM56Ojt8~cA#3cZNb@Y}#RIVxsJm38N*mf_Ac614 zJrw*iC=nClO|I`O$#h^N8$mU0`G7(-2;*i zK0o>_y)@L#9+Udy^9*Sb_M(s#rAhC!&7NWy(H zlDKx;Jg0Xo5)@UvkA$YFHQTItpPrrxU5K#}WQn}TI%m%s+%+7EWU~USvBS=QqnlPW zmXBVk2ry3cUN&WMsNLE>aS*e@2_>NMJ1BZgJS}6h1PZWaxv2;^h7b`}sk;dpMUI@3 z>P7y9ArNIj69d*HqsjJ<&6w#s4+qXCNS(x-AfQ4_Hw=l-Vu_S)Rr97AJ^3mo1MzEZ}xwJ~I7mD;T3kfRUm(^PqPLzDRF<(DrAl}YF*6gO> zZG4=~UTS_n`x85nXMnEHzbzrf(9*%w*LZ5!24@Vz8a{0y<>d34K^61zr|ct1U#5JF zaSU#l6R}+NEh~I?j#77auF5j=qeXhV>ZRJ9keblpFYKU9Onp2z&U!}A=3jr9vFCm@ zM>);DUT8YGzIZHd9mBMAt=7D=#oD6hJ7bk?zu8g2WITZSqYe-ztSZ*gd>H!5*(3_}_eV+$rsFD!h*(HF6d=Q+B||DTBq2AU<|t`N8eCxmB)u&>+w* zrLp>X?mmcvrM1_h@Tvw+RG7O#Fb4JN79rt$Ip8vb*mkljz=l18sF+Di=B2XslXh*r zpX>YdVA4B>X<_+^U1PNKn6oSQ@CJFd7WWSttDda-->q1B-J?FUsIkoPyrZ}F`3Z3&9q1Ot+KAn5)WBp|^} zecA*U!KHw9JrNgVZb#Y`EMxx*ZE3~TxXrDf?z>hrL+z?Dk* z_tZ!Cz%lN*_lE=Q3O&nNjatGm{mo7_?GEW6571Gq z8#=%r*SMKKFd`DX%u0lCqC}+UGA6Dw_E%vF;MpS-^0ulowwr);!(TE@P_3p7a&SEu z`>qM{ugF>-$fdhTs=@_ip-;Of%5v6kv2{!y_MVT-I&;v61GEQqy^TP^4MaRGX{}WU zHVY38iD;T&sjg4^3`o4B;o2(hlK};p(!`ujE~**LT6mxB?MG_-&J_k%x3179FS+|S z`-l!cZjyf4ASK$^+(Y{Ac*FaBHWr-uG18}`Y$N7;2$VBmyU@*gn|(_G>Z9+6YC(I_ z4yg-e;|Ket4i9sx0sG9h>;aMtu|-A7C#JR70P5wM0CD~Vi}*tzyvF0W906#x^4waI z1wJeDE{{w>tK)peC`xfvX!2IFXAn^BrExwtJO~Zz*(+;4?IlLVcyrU#RdAMif5*w_ zIuLJro;255_B=?XJgEtnGVQiYKnLw$p4|yr>$EZ^Tj*B=oFA%Xe#}St@)u`JTZj4( z)(ZEUjq{=}wElg*44$N$#;LWZx$o7?OQ%`X4srS|y6kygzg54aqOQ+m$QP#ulBMQ8 zVGdx(J*n-Znst=39@SR2D!8Duin!r6)NJsIy)-`>adE|FU;b=5+>)8h(B3{os&&yrjGDYGrVT!CKco zSzl2xRmgnUpkH5+Ih)S+po=9ZG-DGfo0BjV9iGf?*ox9W>G6u{+=mEF#Q4Ze6%dj% zyI+4xch{8`8wc6xoLaY&rnA6JVuJR3J7ftqXs5HwQvXrKQ2*g2zVV2`PTUTjpoIN< z*7kcUTO#804W3o&z$dH0F6K*b?~%XVv!NR!3;GhpdXyAa3*S;tv5#vWe8&xV*;u+6 zly!zZwT==vq9PUjsZnw(DKm_4O7{-$B1KUnW*+E`6v=@rJZuq~YP zXygLMXo8;m5kp5k#DE>l6yw$P&?;%Hp?g<-0Li^hh^PsXBMeK!A=EDM`1C0Jr4LNa z?b(Q(*Dc@e*$Ps3iUwNdA+YK9fKa2YkKg)i-@Sa8tD1S_wLn?1UOYBJNqGN)3xej{ z!iJlxDMkHdt99vO{}L20=^zt<@`RCRvr@^)HmC6P`e)(TSqg8OXccBnGqRLY$b1=k zR&4y*sBgqI37K0O9Y(9Ql_THZn$-R*l%6+N`O&o&MOtsQdD;qlCLG_srV>z+@n zDBduhW?fUUrvK%YjOxg#YV%tf-tk7dqF}$Z?PvZAgVmxRMsC}Fv&z?ZCHYQyaoNvo zGa1dUs8NU}aM||4;nN>CJ5fiQTY-m!f}XVl~}42@1~ilTJ+e{Jc%^_3Un~| zL3>x0e-Khv(gNx}%g#XoP?5P1v!HWyf}UrzZ9g*IQ*S@%Yo`Gzk^>d-QZu^IYLn>({atiFxB914y>v&a=;7Ymows4ifiz?2R%L?eAUoRfa=n54k|OTz*z^T_Kf8I`Z1%zhukoNK z!4>~^)6NVbvCg0AH%x$*ug48#%a&rtMnkQOd3g6M_z5_kw_7XOJG^e(k7p(e)@QrG zc=sLG=#czQ6D{~p<#%UhoGxIKy6y74rwm&?9CxV!M!***{>_gfY4ALl_vd#WBBYl; zS+{we`buKqg4eFa6JMQto}4~RLIN7Dwzxm~DAWDqfaghvfOW?8-px6C0^M>(zu*9O zCQi%nswjA$YU_158TpFg#0Bm2=JI#Eimq$tF~C~ghTKuajquOY#4yD;aX{)Cc%5V3 z?E=jl@@ZxeuX00MOVhGkXx?m(hA6lNyJO}L;TyXMOOL_@@>XMK6u}P4RJoJ;jhvAk zNf1+WWy}>bnNn|%!}xcg7J`6Bg<6uns38teen}#E+DA)mLjrCIOStwv8NIHBv8se+ zS(ezYW=wBSNF9>6R{#EFn}3n>AT_j`up4~?1{Jfcjj{ld`ysGNy+AnS2y#;4n>+A8 z$rXbz={HSE|It=_*fwkEI-u(%e!f$kEMv$h31 zp_5(X<1U{+vmCJzlHl7&J9MHr+W)Wvld;nE*F^z_k|9780rD62JCFmfzXzE_j7fm zRN)V$mTLPvrTX>L9U-UDL(Ajbnw_QM_0mFOCZO5z>`$7i^#YfME)}pM!sVAto+cik zDE9Fasv4<6|WDT?nCa1tMIddB~e z?t+A)@Nz+@O0inrF$#79#+hE$+B8vj0U7l@ZW*;vg+i13l8HLlxK-@acvG`NG8J>S zoXCeGfhqIvqPkX!Dfwq`D#75ilVA*l@=l+#llS!PkXw41hCn(w3IkA+^O8(i$v8ob z*P(KGTZT{-;0L`soE-4(v=uxUx6cZ2UkI>k{L&wunu9&SqCLqiGSh=I`R0J4?xlpg zZ4a_C5@=|d0(Uzth9=h568zvOSztmkgRFFf);#?SQ9;hk+H15C4CC-d6IfP}%r3zQ zHfpE6l!f;CB1BW+9`ozm9P_lq1|iSy$Kvrx@A8|2Efz(fGcs++rtwRUCpANU>;_rUZ$J{&T_^Y zxE}SDp||%a6F~a0#T@kOrvSYHzGfjHEfqD>(EkU)Kt8`VCj4_beB&}T zU%oBS_`vU1Q;Oa@Ta2!1V@^#kq<_1Le=PwntkswRK?H>%bJ(0ToBm z7~5OQEj}JqdAh9KL(1sKB?Me9z@-6P2Eg{k={@3v)AnZCf#KH-lK92aY$qae&~pGgyT7R`RBt9t|~?`tntHoOh*}T7uG-o5Hg-Se!zrs1Wq7dp@E3ub}?{ z5$tIG>P~_w&H{zBep4KurFOGi2Uf2Vl>g!ymA@}x`A@G!`Mf8g7IC||0M&+Vn#a;da*ajw)Xg_~_t^WuQnFOGwA!u+X))XH!* zJuCeFor5%Te90O77U6HYI*>=MMvV&JTy#MA7kbFFgTgoeJ;C6Pguovr@QVz-e3xTm z9MX4HI0f=W6N~WquhU<6mI@-w3NG zyDfv=N|D^OS1?*S(1&mrFTo1^3p9PV^dzz~_zi?lcLsY!@beFfm%l5hXoGn5-VkBG zlzN zx}tHMg~36ZoCBI zR3DhR@lww25y1l-a?h7$%S|5uoZa9SlFb2tztIbnNIb^l+l_;_C`EpY1_G2G@Ai1h zuY}jRpt7%~C?`~T3YJNai}i&63oSkU<0SnzKv@F;rv@i)pdf5d0LQ_`G)c=SlKj_c z8JOQs>kB!~KnP3?KSKW}qUo=;C4=G|EhdbsYG!q5v_^(60gy>9;bj4|l}cmH%jyL-L}|1hkN$gsTd1vqXA-APuK>0n7- zhZjTY@R9*;U~HJ`(CgVm>-;UI9#Q3)P9){^c-ho1X_>ltt`?(N7^lh8G{k6@#K9~x z9Ta1X@$!Ty3vZFee8or*V$|!h%&aoz&4FqnwW2Vd!n6p!eGUmCZTlqkOj!Vq9ZQ)jflJd672jd`d!gQk_Vs8 z8s21WDjz4wBR5o@OpME!T_+Pv*%t)s3t`7|Q(gTi?<25)r7f1w&7iP$7f=SFz3cDN zVv4r-J}yr$E;TPMHyf3ui`CLh;LwlAwfBtB&yF1hl=Su7*mdZ~X?6J%yA4*qhg3h) zaTEJ#2TRRnX?F2q>76pGx-Y5RuRnxoN*6kHi#)2N^RAKmtll-0zgz2UoHkpVjZ&rN z9o0gJ6E>CA`b#7WJq$GXSBwU@ry}vV34KcLd*p<_-yzom;d>_VR{(Kk+4Pp)=q#w0b%=uvTzy$dP(u{D8%D>aY^_WrlVz)wBV5n{0`w;OYem6 zT_jwJ04WL%KskF1b;EoG$Rj_D?u(H2O$eJ3;YYdbfjaLeWgUg*G2tKQBJrOA zI>`63@b{!7KKBt?9|tuTAzV39&Ee+2c*vl|EQgZ z+wMDS)CLmHcUk=rhQ#d?&>r&GneVgqcm&{l!q6U1!1GDg9#6sZEYR$6z~TQ%+T;I4 z@R9u&t;1!1mE%S3?$w#ml9}GRw_&Aa8K!Y^O2sEgn-8`B@83acuYm znb)-a`=&`bH0(BRGxZowK@%Kyz)PSd*$FJ?c|1Eog`LQf`+S&k>(*W9zoe}^B%)e5 z_c@+^SWEBx5sx1c;q7V|eW5$F`A9ecYo*^v+D4<1_4Rvo-y6@If9bi4pJ&;}^z3JO z^~Y18o%@2hNfp_nsxhIL<12Tr>TaE5o5W78?a>jtCK&rb=q z;2sgp==O?8Mt5N?7~OOX&IJ+Kt3DS)ORc=53^k0Au9kszi*fQNrPsd5dIlKj7g+ZF zBASsNNL7v87dvI72eoo?zZOmI+y!3EyolzSUrfTC(mv7T!vO?w$-s|S701;p2e>tw z^`JQrF~7Ni@sb#5_9Dky(vwB;qmHcbRb@0|LGu{IV^|o?SiVW8_QN++c*e(3TQZ=g z#q8J77Y1oR0_s=(ZL}TM@*^MzL^hfzIlwn>1Ku7@NdYo$Q1~V}4)B@VkXAlnTJ;;M z9==(9UihXj4%8=%Um+ao&kKE^P&i@JqT7&8KF5UrLAe9khxDr@H#JbDU$#?%uTwrn z)34*#DW6}9a`PgD@gzX#3mZg;r`Doe2N&(9sT8lP# zm^p;;C~=Pvhq$5tG2$NY(9yb6TJv!Mu2WlIZ9i(jy_;aZQLi<_H|rA^Tt@HC$r352 zQRf5rqdvvX=06Xa#vTa?4?RbgrTedwO^-X9-trySF6=5|OS-fkSK#zr*!fDWvb>o5 zL$J`M?c!!D|CmMjyS&-@?N$oAlTBI|&nDVzGZAeMeTTH+o#;*V7yLBQL-=7;Hzy;a z?_Wau9J&ch({5pD@g5NT;6l=0&M03H{)dP%ewf6|cicdg0{inoQ`}hYhx{nKe@&<% z@-T~^0^BY@{1=3;q)TD)X^58!z7({(gumG-ao;I?JEXl3wg>7j@uQob(?Z{bMi~2{ z9tWUYDVTo{QV&^Axe>h>Qjae{d=J3B!f4MotB(-cN}T_N@HdAg+~b6PQW5x%SUi59 zc>&^Oxl;K5uJG>^k@Y(T&(lJ`f8>mTJ}B(5eIEt5aufJJCvf*byoCD%z%N*j{)HF^Kl(7IqQH&(U5FvantH$ZJq!dbVpHa|Mn1geIzdwlg2tv|&5$ zEgsKy+yh&QY{xzF+8B->$@hH@SKn24Ugq(`k#x%aI-}d+u;DS14$gw9J_+g$i9 zFa2y+y(?zA4?o}T_XTB&SC}o;YV~Gerd+sGd8a&AxKe2@7D|PA*Q4$)D!BI-6%*7< zVmv0IdluWs5A>V~Mgg{|9dpx5wiVS9Nm_tyk_Yf-o8;QaSZ&W+ZI4S*zv#{QAm?G* z?-%0IN8g7MXs^FXGjb$+ww)hnAAJiw$RXGkI;Zo+q#R;zIyfD-*R+UgFMm|`Tv}dW zSqpmB=XlmU$^Ou_;(V=eZMjzZa{1zHwY0o^bh%ueKX(1PWD!T2g_Y%UVV*D!mr+Cj z+!+KoD=!yqLyVw=M4N4Z2U>W)h~^+{hx=PR{-Eey4^7g7heYIvo8jpXi>PMP({rjH z&>8r(VyUs9`1kcf`*lTr7!gq&UM-0Zy&q(?qDaDHw}+&vo)aa7(p2i0h-{P3@;0GY zkVyaiBC-z+g^kj|O0CkoxPYGlaMTQ!UGarCrs@T5PXwh`F=mM>yS*fqfuCSiEk;*0 z(T#)>&Pg54s&9b|d(tu=5b2uv6s!B`=(@9J&b?_?zw~lMpiKjH{~)@}RU1du`wXkc zgQ@D_ch@2Kq{aQJXSt)w1!iH!JN6uM(}tW!r9I&k{BBteNMBr%?B)rlh!^gyp;3$E z@za0@zRM{#HE_Z6Z(u(fPq04>&jpI})*C@YV}5D(vM(aNh477E($c+GXml^F72O@s zfYH#(AvZ990n~ZG!Mn)=4WtHye$xaR?w~=;u<+Y~818t-u7W|5_u1m8==`(AydhEN z^D&3qE20+l)%*voLD0;Hn*DHV4r;{_u_NwLb%*HwPJGhdwv*J^0nKbJdKs;<@8&j^*4mY5_Us=m?!v-tvtC|L@I>ufpifKgF@ioAIs|bp$43v%gNx0XG<@EwS zCxkA#l_=M*S{r2WYXAogDrH%36WAysPYS=!42$3_3h6M&=s;&0#rb65-|d_cQBDkGD3c+^=j0Gv#W>m@Gk0s?AmiwDV4 zW91RzP=3h?_n7bx(?EyD&=bUs6Nm9r#9scE#?6=>jd_m~cRz6$e}XuSD-!oWjC+tcq|ZadNkd{ZHvR;0kHxqr z!F__cxORNX;AnjgGItK#dE%bwf_srTEaN5OURj0vEO7|?E#f|(&UJh)RKIIzJyV+f z^6HFnt(|qm6IXG&B>+?-s>7k<+sqzn=ltzI)J_}JWA31KPu;vYTW@2QTOY`!K&5JYKoP>_9UkP9JOSY+) zD~F6We&+!{fq? zjg}#Ub0VVFVOa8T{-hf}N~SB$jJivLMTa%HSt-rp&iavg{ZtS3?>W;Tw-bP%oUDlKMrR*0L&^A7f=md}V%I z4DM0zf52ytNT26L^vAzAd&KpODWI{8+X=Lh^p!KkMWC&mwS5ubz6oLbg#Y~uz7Ip5 zA%IWw3<>`jt$tSS1kg&(MsWchXQOnsxd-wc23k-UYd`DMH8{f>PFFQ>!>lzHA2wS& z&EvD8`W)84AECf&E>)_YzdBh3YN?kUN5YGdgOhQME zAaFw;4IoUXxuL(GIn=F(gs#7DC4QQIgt$DoQR2oEIC^r}j(C^)Oefwp<1$UI}hpLn#*)2xm!khHRn|6idWi1iYre?CAgnH(RTuOVgr^O!ZH zz?zM|2IQ1u)_me@v^+V|^ZKiwgP^eG2YtCHSaggrL3q*6y!w)%=qOZTR)PW_$Lsb0 zx#xI1TNygS8&n*u2v9naZ7M@Cn3yp8pv3>va_!=x<;(%1?WAb zr}yW9von?297{Yb()62L#k71T;f$n%^LCPDM_jfwm6wkomT)ge*3*w_>HRyfF@8*k z_s`=xNqobk?-XIng?I;dSbB;7cr-os@%m1PXnL^Yv%<_<|2RFJ@Xql7tjH5aOt3jU zI9!tC03a;u~#>P2o&yaGYmD#zjXw*(J0{H(n%8g=ks@ z`074A{d;k=-LI$b&_FNqTS;iO`fZ_Hf0VKjMaKg=zWbVMf5m7B2ta>1yh?ujdONFSnN1ICc?*gH^%pUqATP3?QV{5b^uwdp1@84 ziyhx5#9cMMw{$YTyFIRpB3+xD6yf7q#y_CL`{Ouxr?m9CekS8j>+rg+r^#Pltd94b ziKJn0lJuaKKKDu#FYK+BuGUwYMu>`w`rK4e*QavMy%SEB_>hia;P}ah_4tHGA7@|1 zaEftHeSTg=zkIjC-(ta5d-h~;f^WU~Qftm1gj;Ge6ehF~tGR@*` zd`!eRDbrl<7($&6LKSN?<>ml?<=*1;yWtoLV~1d`>BtKp7)_N zw7ZJ6zj`{d_CW^!jFvuoJ}&*UB5JJb)D>(J8in_BDe-<@$NQ6U=|7nc@5mG9c{vzP zEt%B*wc~yJkx^lutvT3GRM4LlnuU|A%sMb^d`qdEcOD+KY(fho+`q(I%nn02xc4Z8 zzo$&BrSPQ8pCjyk5wtYIUk`Ka*A)w?MF2(*V|tU|kF**fG#VBga~RWV>o91l!#N+B z^{4|qP^MS;JxfP@wcRw}n|2EK~?f%c5x)(y`@}Xd*(;j*l%(+i*RK}-qI9!v*ISld5dJB~3C6Q0 za}F73wi;{pgTg;W*&!#M9~Szdm-h+()MNQPo(yk~53R)ktWkVO&cOc`@*(3Q_XhBx zEtU_3kE#2Ge-JOWiR_0|58DZ}+6J`8jcUoSq|PS$p44}6UW|680`m6@|Bz191sz!M zm*J;`t`nXXiN2RON7Hv}yhicXaRwXvOUQXuo(q}Zr*5mFs`o2c?*~Ot_xCpq(j0sG zU>7NvMnbG}x%<|#Tt0ijt8o!FO7e$<-g_S=xbWG_Dpw(%#mitcOTQRO|=#qxN?wNDadVQkBg+h}}``L(S-nP1;JbPmSk?*adJY3m?72jauv zVR=d6rNHYg3a6d^oseZu1}F4#FC!d>e6W5_1Uo}s_@fX<@p%=V$-8hvHmSlkq-0mw zhFt}7JDA?NO6Yn{1Up^EcU=~YYFjL<+_e+P{)f-qEEnNAuni^QpH=?z=zk^=e!hbA z%o@-o>2l|@q%OE*w-(EtvaF|7x=&)gpHE4r=Qof}pVaBJu$FXs&WG(ha>K_b1NcaX z>VGeBS>j~cE#NS1KXF@#BV%9ttamt}8z*i;_}c)|rTLT>(@he$C&uj~ZhwqBL|okY z@bv6=JS+Ygz7G(0k~mXaofi7%j+pi#p`Qhx#63>j6EW^75j^9)Na}c&xbwuB z>i9Hqrap#sdF7MiCZA<*NLm7 zbamps662c0U5;^Ai5v4>C!g+X$8W1kgw>USLOXYxesS(4?|Vn3H9+dS%A*2jxuejK zor&(*(Qn12(2@*TGC?^YaHie|ark|?WOTAv$;1ZXDPhVfu$=Ba3F+l`tdhx&2sLnf zpe)JIMoC$tq%2tLb;F-vWlg~ozxSSGW$h6D#SmE5ldLRU$HcWdP>9o>vi(xb+VBz6 z8ZS+s(bETv^LzC8fO)FEmK`upy%be`RK3}UtBGtgIXy(kM?t?5fv;6#POH{U31z(g zb!za#TlRHooH73wOms5g&Awvo$dEq5cJ&h5OZcCo+Aj>>B=ull&IbA&`|35&?}kKV zzpHiH?}kNJ`knD36$yVtMEASAi0*fzBD&v=MdQi)-4~PgyK%9m{jRni{Z2qTVYmbQ z(cha=@wjK@z<7{;oW$UsAWUwIXQ0YHL}eTD`xk^i@xZ+p3FoU`dH%^=eXifPrv%$Y z&z`bh;r9hPiQ;=_xTnDwBh>PsQz-W&jSJQ$;6`h`+R1p-b9SF2wDXCgo?e zomDw19ny8+hF?CeJQ`@2nvQow_f{@%3obKb7x51B#lT*TCMrGxo>A6@!&77YwyLvd zYbl?2gSzC6-5f)TZvGe=75W$&Q+R!GvyLIUD6oNJC=qX_?W@YUJI9bN|6|9{0`K!9 z8ckGuiZK)k59@gz>X~7V_EEZEg69O%1*?G=Hv}$E9ENRSZdCYZLi0r|x{wv-wu9S6 z9EMFWNA~Jd-Zz!&ZHXjjt99)rsmI8BU>)`*9l}3*Z5rB6rr!cIf{L2~-LtFB)dr5kkW*?+EV5r++cpee{d26(8ha&2BKPgAL<07TcAuf_ zByD>V(wu@c4?vpJB+Y}6=3#g~64JI0Mb!T>QjW~~IH^BvJFNc`;C~YQPsP;#6Na*} zZs#G*IY{%gp>EHK;Mr}oZqG*4?bDFHE(eDS=BaV?82h-Bj~PsUwC%4G&; zpNJlK`$hD?I}lm@lns7PXur6DcQBF$!One1q#AflE|3q4?h9mgoLp=j^yA?+cz~YM zs+wScHiWJ-9}}GlL^U8+JAl`D;2Mrfa0M=I=_Rm@T$4+#9H=hYO7ylCdqi-X!?bK) zgS@y5E9*cnPU;o@)}#X+4Yz@C`({{#e~p1#9dS$YxI{&RQqsb^k}{*x|Ci{e$GRkG zqrNV-QG#;6vKHlfZ9)j@08^xO?NzW|Mt^@wTJ?7k=e!7;usgt>(pm$n!)(+jae#p# zw%xP_M%Zcb8`SYDt|5NKu#xHjljPKApNOG-0Pd}Jles`g^_a_uFff;Me%&{KxikCB3%#KWoVl< z9L>vXQLcnXPy9(!|20Uv_shT^ijW|`X)8aR+;itszTHq(k+AaL^ zl)5>}_X~ZL9}tF7e$GdH14elhaVv;4XV&<=`;b`I(cN5+a`OZh_Q%5_T>6-BS&j5L zDE#wy(zM0VF480~Oq`~XkXpwgsAa-(HPkvn;QcZPPjflq`a+!Pw$`L>jJSz)b_L@0 zg}8&n9ZKnr5qIA@yCQK(+V+7EcZRr!Qo2Wp`@}lCv&5YXanBIyiK>M(XEMsCowC3x41EJGT*<{>#A8i z1}qzMP^U2wZQ1xF)ce!Iw^Dsd_*%w<@Rl~YB}gjEs))6Db|scim(k)puqdj}ze3l${FotnX3^2S3g5a=EDtGcPVzzFH0!83EC7 zRn;UTdnG^jOZLV+fE_g^MyzStEs(Zf=mz5h!e_Jik@e@Hto;x^0O4PNa9OTg0KO06 z^TMAnN~7vgQvMi}I|1d6L%EYA4I1U{fHd17%}$bLTKK0_$TS7OQ!XfcM)*d{GR;AF z9)h$70QbF+N5=iS2wv1J-;Y3jq`G+q>i7jnBlF5hd!6My4)6D~yeA;(N)Gvb?z~M^Dw}99N@?$tEZq& z=b%ncL!HhG|3r{ygs#e?I*(6DyJvw;Qmp^W5ZIpt*w3eg{VAa5lfwT9><-ZC(@@{X zh5td=i!AOz6893&^JSpLE5h$l{L@F17Q{LEah2I|t7)z?cQR=Slq+g?~A?Y~wEr{W5S&OWPJs@gtlDz?B-Goc=XQ`^vT`3*Gpi9#C-^=ys)@?mrfa zw@9h*B-F1Ab@(dOaSmv5t(`t>{J*8*T{>Ndd`|&9iO>I9nES>6snZRi_YGN=>>FQe z=l_m~=7-of{+B}ajjur+|5xq29|*l~`~!IYM*!pRi}1c7@uq#_KNj(QgYi1&|0AM# z9oGMUD}sGvKD3ga8JRb_(;E$L;py0=R=i&TKS5ygj;41^m&nYt&;~KPoQj7 z2k`~qhy>R^vH4Xfrh^hwWk%0Y)u&+5s17MqgMgCYn$>X)9C^V$f@$&iUOk>4d2xl7 zPxJLf=X)m#@2FnFK2CW&YJnn(F8v}|8)h5zE4=a8!hqj1mCns_gkeP?vX4c=v6-z` zIX3L84vr1Wkp!C|5#48oMRcDTiHXn0#E)w6&Sf4yCZZNt_<5h|z~iX;j)zdN=bd_k zC7lqR4Ts6-e8%Bmf-|L)=2=b}-2h)ySUFcs_{pDa(xew5>PGxQ9Ck$GaBa0XM6rix zBYwYW`|H&?B}5nicK>_ezIA4aAJg~nM%Ew^_e8_F#^9Vx+ClavB*1_qfD$PT!gLG; z3h5}YFN^~FqTyU+a84wpz}`Bb)cM+G4LLUpeu?L-Y#ME;B=9{G$P?tG!5V3xS#HjhG(K_ z7%1n#7(7U@bz(KaR31+k#lP4^+FU~ZA9(sF^z{EH9{;2kpLv(ZKc&Sxb3Fd64!hXUaRgMU0{UqhQrpI$F1Z?yDuky1YHNz4Ddh;H96=G|{R6g*T&@7tn@gyHK8~+&LX? zKpQHrqFf4RJ{27O^XVk@SP<#@$D)=#s0~#_IRBUNZ|m{@jK_ajMDxpPG(8fnPKhFdM7|HA*a2S;uB zO8=-x|5fD}sRQ13yEb2a=B^{2;{?wf!}CQ#+uL5a8XSmM&Wu+_rsZ+vt8YM5U+-r5 z4lq1f2K|;iqMu!*>ATg>hDC7UU6wH={O`~c;O6z!9bLpt!rLzN8wqv^{|jAu5B47A zKMDQ#?_xN9qdhJBuUhFn*87zI1myb~hVK>h3thFbDE$Xj`UB`s_;u{xrRmq|i%i_8 z4BHUJ^21be%unM72!3lBE;>jrNWBgV|J%<5?cudE{Z@3f?GZ)y|37=*0ng?4J^t8K zW)clak`~cY%28==?VSoKD>F@Pl}dv|QzcO$BSP1XQ1(tpwu+2O@%Q$A9-jBx+iHA1 zpa1vwc)gDEzW2G$x#y1Oj&sgEcLesoS?|&<@$}6WEj$Xxqp`MgzihZNkQSE1@mM;& z4Me;=)~$B4AgBM+$Bg(e#<@TguLA`<|IK(jvCdJz+U5pw8OM=$<748D9S05^@V_39 zi(_!CgyZqutH0JY+xUO5{^UETzX_x^CD-4?f1v&paXgt$cP5ed6s+wQ4I)okyfTib z;&>AM*>&Q*Y1r=_x0~J)ri%axI>Z@r@(nvQI`eTlw00&8p9y-z+FP|HOcxoF=V!p$ z=1HQC`&eNNf6R5vfwlYlrT+0dCB9D2bJOo(OJ@Jj~#A(vfZN&3j$s(WmE#bEP?*e*$Qca|?5_S4d zd9We!yRzwXg#PL$Bt%Ydexb-manr_gcrCarDbhJ$A+5Rt-=f6Dr@4$Y2_SYiS?cemT)>FQ;8Tk5pe(C;pyVAZ zx|+8cRChg@x6Lue`PQNvbZF_!evM97^Y+j`G;eoG2k#T?qE_@c_AY>%1RZ&RBcL}4 z2qz@uxs3*dr#JypoFEB=a1dsSr$OHiK#T<99RT@mPN@z+k_5xL03zLB5{&2oMq-R! z;E42tNHDqsAl@B|1Cd^y1Vo*UBY^@52H>DX0^%9jV;ezdG6^P;pdSuXNI*QBMgqWL zMh5^rgy|%hjl-M{V0I6IpgW%g3p;^DoxoxomXbiD3s^=1qOFPWDiW;e0#=iNphJY$ z<3KK>JQHl{1UBQa1&0kJ=&0Wk?`*|^NUuqP?OlKl33R&vJrd~SAcXz;ccD)LLlO|< znl=tbI2hny+zFWAV2Xno4w_6r9y`r(upq%+94v7d#{v$JfK@q(XNO2Y(6e9yt4@Gu z=OZ{ACP7EzA@R;p9EkMCNN~IhI7I^MF5nCa&f*}1{rXsUmIO8=Am%orZ=A!SCw=2Q z4i|7Z#suW~(-wz|B)EjbWgI560DBU!Dkt&GfdmA-i%cNY2EXH>vGqL;{5`HYn%lKZ z@c@sGj*+pUsUaar($Q&8g`h-}8M4z>1)LazzOGyTAGt{eXxtcAi%STCb_egX@|DSg zR+}LmH^xTXsWqRVkEc-~rMJWLI@d&hN1>^Gd0EiLT)Rn|NYmXsdB7J8K`jd~9Dxarhls^nAJ~4~Yhs8i~dR7WK9aPd9_9jmT-r z@iaVbY3RIl(f+r{Jdo0`C~F&^tjfidw)NNoS=Zgzbsp{BW9zhkam^ypU{xO?%=JOk zX9x7uyP(I}GkR<~r14=kiS`K6?~hD%e~__koCBDica$FIo%1QvbG%=+-R;&%`!Un< z>}K8suLCiCa z!xfN3Iw=-$ye#6RS;VzQ{i%AH-&Pl#(@1r+fK*2v_7l=r9MMr9p~VSdZ8J4tXVcid zygbQybx?+d9h6}a3%c#T4OV5~#yAXFjc;9)qqF=5cB34NS@lbDp4=uz&C3;IR!=nd z>Y9wz`-OK*Sv=utF>vnim2JO}4a+C3?{u&TSWRloB`hmMPt(%Sj9!G6A+x#b(-cZv z_>zMrq-yCNob)N8Wt!k5E>R>?qTdxpd(@l}eYEJt%7pxTH5udYakTxWw{9b28ro{1 zKfwxTYqjh&)-~8oxW?UXsBdIus`fCv$>Qk=F*HeJ+LcGVedOoOf2K^~kxXHhm<^}%1&`~7vArod z{hv0l>LlNhe|XDHC4= z)g`Ge$-+H&3-7pQinbn?X&yx<)8c4SEArZ=gEc?HKAkBY!^q3VH1f8-y|;NTm_(m4 zFbmR{-O`$D?LUN;hjO$Rgz2nua$#I8w*dUWI=;$+vB~H~Bz4)%th!`dnCmFMd}w(O zYsHLB^}9_@%_P)k+nqfA*_mWDX@tP4@oZ>c`{3@e(^H`3X-uEJ7)B-{z9U{}&sQ?; zXxpl?&vrvY;~mY5H)BnG6Z{>lH60h$#3&Wyr_mu_fqZ~Pf;N=qw|&|HJH4d;0JgUspK zJ9&=b{h1UlCr@9bX>K5^cp809Z|YbFJ-L%nWq*m^E)04YxDB#pL3;T}$zH@|;CkS9j1-66E;KO3l}4 zsj=_O{4u(*<{ChL-Vxt_x;$iln4Dm@a}L1d=(=5f^Z!Nuh+czXUAKb()>%7n4Pxe7 zu#0jF^{8B2r#h-%;V$wK=_DW8^S(Xeqr>`L_q+8gigjAQgPHke^1a=z++sbeU+a$Q zm&x~byL=e^X8-Z;YUA(Lt3;>uij^^!hsoD=yYe!+(q_NP{Jb+)+O+3`dsZ(U&2(4? z@q;?*H?1xS7aJOD%`nz6!C1Q1CUi$F`bpaJAwA-k>tuJ@1;fYfj-J!_hqA2-(iBtb{JGpvormiNBkWh1v2 z-DeAZJY~b@eP)aKuB^R#ULN{*Ov|5(p6>u6O!gTg+(r3_-6vt`PRpBcEl9R>=f!FF zW-2>BsqA0yyofp=cqHmVxRbizBmw!4SW*9aJ<-QK8gG0!7Q`|AefqcX1+Oa_?#omr z5qi3Hc>4|_;stQr565CyTOX!{`{P)g9={QPe=zY}6vy<_4fJ`979NQGUjH1#)JB7` zw(*1w9}oX_|Ln$j`$6RL{&PaSw)+1Y`3hj&JIO_x<7wX=`P2&G6H1aE9?`_qjzj4D zpTy($klT?sRYLp`ZOB1_4*q&Nj-J8uFC+Mp!0|A8{7NET8f)usU+{8wL&#c2PXDLd zr@Ql#45{vkZ|85!OSHVCF!K7KtCJ$F>z4cEdcrW~by_<*EZ1or(P6{aDYSU9U!IFh zW$$PV=k5Y}bpZS%5bOd(NI>u-N`k&5AfERpfdmP}NI;xSA;JU=@`(+Co+QTlGzhEw zhLS*v1VnyBn4lp;f{v`IyMo~)5XONB%aefMSAhhIBpBZbOu|8#1d~WGwLSdumBf8$ z78XW^Cc380saQ+k$u^*^-#%ecFM@SEpNe5^V|0J~osJ~(gx<0T$};AC@M`Atok?u5 zF81#PCeveQq}S>7>?|a<4A-TxV{G$?{2Ex~*HJxmB41g<#1G4Ibn^b%4)W>1X=;aZ z66s&t=dzZM(f(*_09Kc^kkZ^_^)4yz{)~CG)qOYXN}z)>b>t?T*)93nPRoWd{{6{* zK$^dDy4Nmj^O)M{To-lA+^(xTd!HG1Ri2L9u%kL<9p6cPt?MAaZaQyfE^kM2RC>Gg zwEWG#5?S-XwZ8LqWBi_Je?nN>xX;l%?su;mdi2@$w0Ip=H;ZoAGE8>e{PgPjReqZ3 zZs>r&e`*+HDaNM#{SOR-8~)-j*b#5czO&bKQAhui!z?rYjbXQgHe?;&N&Rj7!^3U| zYL;IecH4$ZF4}F;mMmZDatz+UOjHQ#Yz*$GPI@$E^t7zMF$Q<9%sauybR74`d8DP^hS5a>w0Kog8OV7_ zVvL4Qi!r*mokpK*#59CdS3~jmVOX13fX3%=JS>Z~IiQdwl|hb_zZ}jptuAI@bi*fo zoS%)Px}AyhE{yYJjI|kzk=sq4solonc`C4IHw99=DKfR265eh!F+vH}CN5ow$J30a z9{#2Fq=6g8Xb`7>*OLN^wpGUCRY~O}=Svf9sNii&Gv10|YTGGz-YP8GRu#v<)V69k zo`Ki%bezUaQvNfU+Lq?nqL9?KMEx_jZ7EaPX}@XTI?Wa0uZuOhx2>t?vvoIk0rS&fhld^XAxGqx$$Y@d67ovQEVe z?aj*msWTh$-4z zwzVIonJnAd4bx1PZGT;S!NPj29nV7;pHt!t7MNULQk|6ZPl+{{t?!ugVcyEKu+D2Y z4Rm3h_eaDXEUfaF`LyQBHi|Kzzt@sZ2&1pYcVUc18r*o8Fh(P!5@aJnl0jU{lq-O>AL7~YA)VPU zjU5NF(F{qN((0P+PZ)j>b;5_SHW(sy%s9WThADU*vj>sYJ~oV_#o8 z{ju6bv~^p?-9yK8ySF>(h%RAE`?sbMajeabkvxsihh+Mc{a>9%*zhTxusn=f^d%Kxe@NIH_GCyS+b8}5L)wOa zDc_BMGvDrxU9uR%*!4HhYuFIwmBBINlDNlX7ju8MVcnm5Ja%}I{>nsEs>Wb#@cR{ z7(0ou6bYn%5k`_g4hO;;`@bFl<^ZYN_Sd}hGPSI!I&(-w!?ba42*zj}Fa)PT z3-jUOURax_lOZ@=S*&f&62N+-$n4TSnn_{#yB;+vZFbe!C52oa~5-`N>}Nb=1K)-zI!$YcI=yn$?%8i3>U{o z;=Rt!aCOWF{r$Eud#B}Qj%UR$lOa;OdSNm|`Xw#@|4;S7Xav=@uMm8WZ9Q{MA8+XW zmsmFt>losF(%jd&E~53%-yHw{x=#^ey~fHgnZ|Qp0FOX$zl>~{Om7?Vn97cIz5b%f zOl7e)-pSuBi}CB1`U4xDu1#G^PxrUiZ;e>{wd}9bB2;`m?K6DXKeeW5JHOtGl*i~! z>Xy}hnui3uIR89M(3xD00D*3RaC_)!`!cLIxQ(@%pBi_#OJ}%or=weQrZRs;xbzvj z!lK1ltxQtX(4IPTAs zzbMvr*JNk;(^f|VN%<2XiPIU>2^}ent_`96md5Cs5`qrd?4JfQ7^B})G98+vb0q2A z(KsgjosR0{U2<9EF}gI6prhEsx>UmPc#O4A)Ny#eld<12J)hp$YVsdkJ!*Fl-;TeY zNn%)g?E3{tOoyw-9d<3c4?UlD;JJ@cl83e4!1c?k%x;@_cW%<& zHd6O8(la#OVyd}KPm75H*~pz1?#!tD5A@t=-RZf%+hc#Xr~bkHz%TX8zuhDMMo;9& zIvqJ}nZF1DjIrMdcEZ;m=y&EF#M`o}r_RbL)M+{Cs~K{7VWx75bW$GL^S+%tXHnjc z+J9|knFot^1;Ks#Os_lQ^4gI#C5 zcE+0|*0T?B2*%j4rFzCYxnt8dJ@kG})(Ul$Io6F)WhfJK(w#8(2S!8g?u>@iEd1M} zaHq2{#-fuB>sh{gI-8xlA;r@s!nA!%H!Q@alhglO7GjKAwZEtJ`%P`yf5rnclN9x@ zNyK`}7um}(8~zRc7x?03~q0f8jf4lH~qH?Qh$-hG7g|rl%3{ zy}MO!ZFGA7v+XtiU&?pc-^iB@UxyHqbs83=$D0%J0^Ms5TXOn;qh9`GzF+WsX}B-e zHa5}1BJ}W7A}mbeQwHnWV>BUmC0oSNB5BIUDSkL;@5bWQXqyA*R zG@8d`gN{{sTxbs1q1M0)mbd#`+YMOea9TMS!3qhqXUJ z#R|rffK^%KNigp3!mplsq1nd_ZR*f=23t(}y%7K1`i;@ntF-=1`>t4>c9F&Ffwqn& z`ZS?zA<_``6|}TGcv?PETEfPHA8T(XGZ-(gES~>#jNY$l_s!DwCba&hiZNOQ6ZFJ! zEP>+zO!R1JcVdip|I+wciN}%i8pI+m8a>(^MC(^XA03MQj@?K5dsx{R($jDLB0p!x zI+o+sV846PrZwMg-?&PQX{^iF{>D{Cdyjwr##JV}k#@d<{*4=MnfnKZb=?k3b|8QJ z21F)%kapJ2zjJeKcllth4>ElwD}i?HB+#?`JGx18DC>Com`cu1ut)iI?ptkaYTOH$sxF`-QtGA19uECr*c!_eQ*|D_d!|c~2qh zuSoitR!2@eq&oPo-3W=dW9umh^39n7ozxRiH~e_ra5B{m(Jz>9%fy{sm9eUee_D?A zEB&0GsXPMoFmbcw(0`(zcVq9onB1QK*4|l=R9D2Ujekc!*47@3>zAJ2B-jgQ0H@pX zyc2pX=b?`15qb|oAJWtM?8{J}e_tQ{E8mAGQ+xEo7`_k5|H${Tkl;@Q`*Yt1A#44e z{wIv{+SoarF8dvs4*4XI7^%GT@$#~F0A!gx^NAm3Am3owr-vYdwS5l+h>)Ov4?&Cs zgx-J%6CQa6lYs20XGnYatrJKBrpB7O20gg1ktTJ~C;c|Sf4BB*yah1}tb6x3lAja1 zX5w-Q+Dc_SQyMWS%O&@b+E zf9=E&v*f2MNe>GN%&%I|>UFRq#rp+)jKOfsPVz;=B%%4G^B1J_vYkmJSk)(!1gKrV zVv+#0>sOsga9Hrq(j9eo`jr6oNBUKVC&&alp~Eb}`2!NqU+qT*ct0{A_oF}9e{_1< zf5?(27LukvT|XV3NbH#etK;#dGYJ+8IvpLiVD{EO{}=mfr{}H6D}Q_i&2Z6Mo5X=$ zUP4ktNRUWYm93WGbZH8`9}?1(ZmjSbq%Z%R)|x(^{^@)Px;@=7_9C^(KjEKF?_hVaEBc%XYM8u7AN_YsXIM5IcA_d>kfrzP~3~`R|;JAWKvJiIWlC>F2`OKiSXQ zte*a4z62c$COTy42U+?~JZF}O@REQ?L$+2V`#<6T1?WoxLNY>@bP#d87;9e0zKRKN z90YX5Z)X-Z9f1_q-pYn-QA5Nr`$uNQ%Wx8m_%-+y%NLr=g=W>=ft-fpUEA;Wo?+DJ z?Z&2z{hEJ%LN>z5^e*9dhxWV|)@HFncoz}C7(KM`V{Lw%gz)#s@(NL$4_d!cz}mE1 zyYM(Mygbb5X?X&U#qqWrfafEDzfVZu29o$3jHi*r+Aepc;h`+bIt=eC1pmYFw6dhM zBk{B}8Oms!H=2B98&jRi;pu4JUdG~dpE93Pu73X~l)^_ozEzYM3Q+OIF%-k2`ab8sMHX!6mGjRH9 z80(8siSo_H%QvSP%1F;?bvL&eyqM~4E}q{!95c5U@h**i2>$LO{N084JBvy04#C>D zHLH3h_v>XiUXGV>DXHFRpn;cZMKcuR@2w0uiQ?NCno{SjmQMhG6) zHp4nneQ&_=MjUU#X|Kog+025*)@INo(b$gT9XQs)X>7x3Xp`zd zd~zvsdM5c~yQBO}@=wgZ8v1e5m{bIqV$xxzL_l6+TCv1xy2fD$=F|X&>jnVD&OhMLXDvY zTH1tg?m?!&B;k~1qQLAMVQE)w!X&ZmqD@$F#3a+~lp_{0&e_a43XyWH@1e2Ge064(Mtd%Xbv{=1@8o@~+g6g@a?!H5tFME}MNT_;z69g*cK(i;+|e-K zraHE>{5zYUnB3SfTU1YDPS50Ori;G8Hm$3A8Qf|9?dL~HhV*Q2So6#Il*pHUM-Qhb z5f;MQzC1Qd#>^NW?(Kr}A)Rx+hnaIJjPWZQ+G$NPehtktXgsUlNuwi;bvg>RtHEK6 z>6JPgHL*^famqIjTrBFSvpSQ(I{ylWW1W8mvRG$d0c$@i-T4YSyJWSwEgYEB%VC|5 zOe3*QXR8&|!_m8|FKJZgWw^%7dvpvF*y9+jcHqtKrmeVD8Z+Vw|y z#^(b&;G^BDQGxM!yH(@3PRc=hu895V{=?RRN5{_VsH68f`<9fjPRGS|tB>&v={W>h zrDx`^rS>Rj)@jw!>5;;Y-CYk)XQbQibcibBbKD0cL7$Gn7{*{Z-?C$8!@Di%rH6$~ce2 zSj>U7AA`xpKtz~)9Ey0CY?RXzV;~_Ch+ysA?vM?G2!$RI79#eX-{NFHIX|1c? zEn567I>v_EOqfV8*|D_qY{Jg?ygySG9o~z=!T282J)Ufvi1#|X$CJ6p9r4+jmW+8T z)9#rX#_-nxA58pX=r{TAu&KE94R`2jTe5yjfHgHXbtY=4t6Jze$L%z@oA7Y!i8&!s zDLCa#4UOo}byz*;l5Y|Fv@!Cx8RC)OW@u_^sNV@s-12nsQ+uYm!DL4*v$%uyztus! zba&#hc2+-a-xHHe8Ij{V=AUiBzsXO>e9`Y|W)c6fXLPuTaf`S)ruF8K)()bjr=_oD zV9J