From db16b424693d27e7c8f151195b812ffb91d17512 Mon Sep 17 00:00:00 2001
From: Riku-Laine <28960190+Riku-Laine@users.noreply.github.com>
Date: Thu, 20 Jun 2019 11:04:32 +0300
Subject: [PATCH] New framework and other additions
---
analysis_and_scripts/notes.tex | 153 ++++++++++++++++---
figures/sl_without_Z_15iter_random_model.png | Bin 0 -> 48731 bytes
2 files changed, 129 insertions(+), 24 deletions(-)
create mode 100644 figures/sl_without_Z_15iter_random_model.png
diff --git a/analysis_and_scripts/notes.tex b/analysis_and_scripts/notes.tex
index 229ad99..6102a24 100644
--- a/analysis_and_scripts/notes.tex
+++ b/analysis_and_scripts/notes.tex
@@ -9,8 +9,11 @@
\usepackage[hidelinks, colorlinks=true]{hyperref}
%\DeclareGraphicsRule{.tif}{png}{.png}{`convert #1 `dirname #1`/`basename #1 .tif`.png}
+\usepackage{wrapfig} % wrap figures
+
\usepackage{pgf}
\usepackage{tikz}
+\usepackage{tikz-cd}
\usetikzlibrary{arrows,automata, positioning}
\usepackage{algorithm}% http://ctan.org/pkg/algorithms
@@ -71,7 +74,7 @@
\graphicspath{ {../figures/} }
\title{Notes}
-\author{RL, 17 June 2019}
+\author{RL, 20 June 2019}
%\date{} % Activate to display a given date or no date
\begin{document}
@@ -83,7 +86,7 @@
\tableofcontents
\begin{abstract}
-This document presents the implementations of RL in pseudocode level. First, I present the nomenclature used in these notes. Then I proceed to give my personal views and comments on the motivation behind Selective labels paper. In chapter 2, I define the framework for this problem and give the required definitions. In the following sections, I present the data generating algorithms and algorithms for obtaining failure rates using different methods. Finally in the last section, I present results using multiple different settings.
+This document presents the implementations of RL in pseudocode level. First, I present most of the nomenclature used in these notes. Then I proceed to give my personal views and comments on the motivation behind Selective labels paper. In chapter 2, I define the framework for this problem and give the required definitions. In the following sections, I present the data generating algorithms and algorithms for obtaining failure rates using different methods. Finally in the last section, I present results using multiple different settings.
\end{abstract}
\section*{Terms and abbreviations}
@@ -182,7 +185,7 @@ Given the above framework, the goal is to create an evaluation algorithm that ca
\node[state] (EA) [below right=0.75cm and -4cm of MP] {Evaluation algorithm};
\path (DG) edge (LP)
- edge [out=180, in=180, dashed] node [left] {$\D_{unlabeled}$ (1)} (MP)
+ edge [out=180, in=180, dashed] node [left] {$\D_{test,~unlabeled}$ (1)} (MP)
(LP) edge [bend right=19] node [left] {$\D_{train}$} (MT)
edge [bend left=60] node [right] {$\D_{test}$} (MP)
edge [bend left=75, dashed] node [right] {$\D_{test}$ (2)} (EA)
@@ -193,6 +196,67 @@ Given the above framework, the goal is to create an evaluation algorithm that ca
\label{fig:framework_data_flow}
\end{figure}
+\section{Modular framework -- based on 19 June discussion}
+
+\begin{wrapfigure}{r}{0.25\textwidth} %this figure will be at the right
+ \centering
+ \begin{tikzcd}
+ \arrow[d] & \arrow[d] & \arrow[d] \\
+ X \arrow[rd] & Z \arrow[d] & W \arrow[ld] \\
+ & Y &
+ \end{tikzcd}
+ \caption{$\M$}
+\end{wrapfigure}
+
+\emph{Below is the framework as was written on the whiteboard, then RL presents his own remarks on how he understood this.}
+
+\begin{description}
+
+\item[Data generation:] ~ \\
+ ~ \\
+ \hskip 3em \textbf{Input:} [none] \\ ~ \\
+ \textbf{Output:} $X, Z, W, Y$ as specified by $\M$
+
+\item[Decider:] single vs. batch \\
+ ~ \\
+ \hskip 3em \textbf{Input:}
+ \begin{itemize}
+ \item one defendant
+ \item $\M$
+ \end{itemize}
+
+ \textbf{Output:}
+ \begin{itemize}
+ \item argmax likelihood $y$
+ \item $\pr(Y=0~|~input)$
+ \item order
+ \end{itemize}
+\item[Evaluator:] ~ \\
+ ~ \\
+ \hskip 3em \textbf{Input:}
+ \begin{itemize}
+ \item Data sample $(X, T, Y)$
+ \item something about $\M$ and something about Decider(r)
+ \end{itemize}
+ \textbf{Output:}
+ \begin{itemize}
+ \item $\mathbb{E}[FR~|~input]$
+ \item curve
+ \end{itemize}
+
+\end{description}
+
+The above framework is now separated into three different modules: data generation, decider and evaluator. In the first module, all the data points $\{x_i, z_i, w_i, y_i\}$ for all $i=1, \ldots, n$ are created. Outcome $y_i$ is available for all observations.
+
+The next module, namely the decider, assigns decisions for each observation with a given/defined way. This 'decision' can be either the most likely value for y (argmax likelihood y, usually binary 0 or 1), probability of an outcome or an ordering of the defendants.
+
+\textcolor{red}{RL: To do: Clarify the following.}
+
+The evaluator module takes as an input a data sample, some information about the data generation and some information about the decider. The data sample includes features $X, T$ and $Y$ where $Y \in \{0, 1, NA\}$ as specified before. The "something we know about $\M$" might be knowledge on the distribution of some of the variables or their interdependencies. In our example, we know that the $X$ is a standard Gaussian and independent from the other variables. From the decider it is known that its decisions are affected by leniency and private properties X. Next we try to simulate the decision-maker's process within the data sample. But to do this we need to learn the predictive model $\B$ with the restriction that Z can't be observed.
+
+\begin{quote}
+\emph{MM:} For example, consider an evaluation process that knows (i.e., is given as input) the decision process and what decisions it took for a few data points. The same evaluation process knows only some of the attributes of those data points -- and therefore it has only partial information about the data generation process. To make the example more specific, consider the case of decision process $\s$ mentioned above, which does not know W -- and consider an evaluation process that knows exactly how $\s$ works and what decisions it took for a few data points, but does not know either W or Z of those data points. This evaluation process outputs the expected value of FR according to the information that's given to it.
+\end{quote}
\section{Data generation}
@@ -200,7 +264,11 @@ Both of the data generating algorithms are presented in this chapter.
\subsection{Without unobservables (see also algorithm \ref{alg:data_without_Z})}
-In the setting without unobservables Z, we first sample an acceptance rate $r$ for all $M=100$ judges uniformly from a half-open interval $[0.1; 0.9)$. Then we assign 500 unique subjects for each of the judges randomly (50000 in total) and simulate their features X as i.i.d standard Gaussian random variables with zero mean and unit (1) variance. Then, probability for negative outcome is calculated as $$P(Y=0|X=x) = \dfrac{1}{1+\exp(-x)}=\sigma(x).$$ Because $P(Y=1|X=x) = 1-P(Y=0|X=x) = 1-\sigma(x)$ the outcome variable Y can be sampled from Bernoulli distribution with parameter $1-\sigma(x)$. The data is then sorted for each judge by the probabilities $P(Y=0|X=x)$ in descending order. If the subject is in the top $(1-r) \cdot 100 \%$ of observations assigned to a judge, the decision variable T is set to zero and otherwise to one.
+In the setting without unobservables Z, we first sample an acceptance rate $r$ for all $M=100$ judges uniformly from a half-open interval $[0.1; 0.9)$. Then we assign 500 unique subjects for each of the judges randomly (50000 in total) and simulate their features X as i.i.d standard Gaussian random variables with zero mean and unit (1) variance. Then, probability for negative outcome is calculated as
+\begin{equation}
+ P(Y=0|X=x) = \dfrac{1}{1+\exp(-x)}=\sigma(x).
+\end{equation}
+Because $P(Y=1|X=x) = 1-P(Y=0|X=x) = 1-\sigma(x)$ the outcome variable Y can be sampled from Bernoulli distribution with parameter $1-\sigma(x)$. The data is then sorted for each judge by the probabilities $P(Y=0|X=x)$ in descending order. If the subject is in the top $(1-r) \cdot 100 \%$ of observations assigned to a judge, the decision variable T is set to zero and otherwise to one.
\begin{algorithm}[] % enter the algorithm environment
\caption{Create data without unobservables} % give the algorithm a caption
@@ -223,7 +291,15 @@ In the setting without unobservables Z, we first sample an acceptance rate $r$ f
\subsection{With unobservables (see also algorithm \ref{alg:data_with_Z})}
-In the setting with unobservables Z, we first sample an acceptance rate r for all $M=100$ judges uniformly from a half-open interval $[0.1; 0.9)$. Then we assign 500 unique subjects (50000 in total) for each of the judges randomly and simulate their features X, Z and W as i.i.d standard Gaussian random variables with zero mean and unit (1) variance. Then, probability for negative outcome is calculated as $$P(Y=0|X=x, Z=z, W=w)=\sigma(\beta_Xx+\beta_Zz+\beta_Ww)$$ where $\beta_X=\beta_Z =1$ and $\beta_W=0.2$. Next, value for result Y is set to 0 if $P(Y = 0| X, Z, W) \geq 0.5$ and 1 otherwise. The conditional probability for the negative decision (T=0) is defined as $$P(T=0|X=x, Z=z)=\sigma(\beta_Xx+\beta_Zz)+\epsilon$$ where $\epsilon \sim N(0, 0.1)$. Next, the data is sorted for each judge by the probabilities $P(T=0|X, Z)$ in descending order. If the subject is in the top $(1-r) \cdot 100 \%$ of observations assigned to a judge, the decision variable T is set to zero and otherwise to one.
+In the setting with unobservables Z, we first sample an acceptance rate r for all $M=100$ judges uniformly from a half-open interval $[0.1; 0.9)$. Then we assign 500 unique subjects (50000 in total) for each of the judges randomly and simulate their features X, Z and W as i.i.d standard Gaussian random variables with zero mean and unit (1) variance. Then, probability for negative outcome is calculated as
+\begin{equation}
+ P(Y=0|X=x, Z=z, W=w)=\sigma(\beta_Xx+\beta_Zz+\beta_Ww)~,
+\end{equation}
+where $\beta_X=\beta_Z =1$ and $\beta_W=0.2$. Next, value for result Y is set to 0 if $P(Y = 0| X, Z, W) \geq 0.5$ and 1 otherwise. The conditional probability for the negative decision (T=0) is defined as
+\begin{equation}
+ P(T=0|X=x, Z=z)=\sigma(\beta_Xx+\beta_Zz)+\epsilon~,
+\end{equation}
+where $\epsilon \sim N(0, 0.1)$. Next, the data is sorted for each judge by the probabilities $P(T=0|X, Z)$ in descending order. If the subject is in the top $(1-r) \cdot 100 \%$ of observations assigned to a judge, the decision variable T is set to zero and otherwise to one.
\begin{algorithm}[] % enter the algorithm environment
\caption{Create data with unobservables} % give the algorithm a caption
@@ -262,11 +338,30 @@ The following quantities are computed from the data:
\item Labeled outcomes: The "traditional"/vanilla estimate of model performance. See algorithm \ref{alg:labeled_outcomes}.
\item Human evaluation: The failure rate of human decision-makers who have access to the latent variable Z. Decision-makers with similar values of leniency are binned and treated as one hypothetical decision-maker. See algorithm \ref{alg:human_eval}.
\item Contraction: See algorithm \ref{alg:contraction} from \cite{lakkaraju17}.
-\item Causal model: In essence, the empirical performance is calculated over the test set as $$\dfrac{1}{n}\sum_{(x, y)\in D}f(x)\delta(F(x) < r)$$ where $$f(x) = P(Y=0|T=1, X=x)$$ is a logistic regression model (see \ref{sec:model_fitting}, random forest used in section \ref{sec:random_forest}) trained on the labeled data predicting Y from X and $$ F(x_0) = \int_{x\in\mathcal{X}} P(x)\delta(f(x) < f(x_0)) ~ dx.$$ All observations, even ones with missing outcome labels, can be used since empirical performance doesn't depend on them. $P(x)$ is Gaussian pdf from scipy.stats package and it is integrated over interval [-15, 15] with 40000 steps using si.simps function from scipy.integrate which uses Simpson's rule in estimating the value of the integral. (docs: \url{https://docs.scipy.org/doc/scipy/reference/generated/scipy.integrate.simps.html}) See algorithm \ref{alg:causal_model}. \label{causal_cdf}
+\item Causal model: In essence, the empirical performance is calculated over the test set as
+\begin{equation}\label{eq:ep}
+ \dfrac{1}{n}\sum_{(x, y)\in D}f(x)\delta(F(x) < r)
+\end{equation}
+where
+\begin{equation}\label{eq:causal_prediction}
+ f(x) = P(Y=0|T=1, X=x)
+\end{equation}
+is a logistic regression model (see section \ref{sec:model_fitting}, random forest used in section \ref{sec:random_forest}) trained on the labeled data predicting Y from X and
+\begin{equation} \label{eq:causal_cdf}
+F(x_0) = \int_{x\in\mathcal{X}} P(x)\delta(f(x) < f(x_0)) ~ dx.
+\end{equation}
+All observations, even ones with missing outcome labels, can be used since empirical performance doesn't depend on them. $P(x)$ is Gaussian pdf from scipy.stats package and it is integrated over interval [-15, 15] with 40000 steps using si.simps function from scipy.integrate which uses Simpson's rule in estimating the value of the integral. (docs: \url{https://docs.scipy.org/doc/scipy/reference/generated/scipy.integrate.simps.html}) See also algorithm \ref{alg:causal_model} and motivation from section \ref{sec:motivation} . \label{causal_cdf}
\end{itemize}
The plotted curves are constructed using pseudo code presented in algorithm \ref{alg:perf_comp}.
+\subsection{Short motivation of the causal model presented above:} \label{sec:motivation}
+
+The causal model tries to predict the probability of adverse outcome $Y=0$ when an acceptance rate $r$ is imposed. Estimating such probability in the selective labels setting consists of two parts: predicting the probability for an individual to commit a crime if they are given bail and deciding whether to bail or jail them.
+
+In equations \ref{eq:ep} and \ref{eq:causal_prediction}, $f(x)$ gives the probability given private features $x$. In equation \ref{eq:ep} $\delta(F(x) < r)$ indicates the defendants bail decision. They will be let out if the proportion of people less dangerous than $x_0$ is under $r$. For example, if a defendant $x_0$ arrives in front of a judge with leniency 0.65 they will not be left out if the judge deems that $F(x_0) > 0.65$ that is if the judge thinks that more than 65\% of the defendants are more dangerous than them.
+
+Now the equation \ref{eq:ep} simply calculates the mean of the probabilities forcing the probbility of crime to zero if they will not be given bail.
\begin{algorithm}[] % enter the algorithm environment
\caption{Performance comparison} % give the algorithm a caption
@@ -374,7 +469,7 @@ The plotted curves are constructed using pseudo code presented in algorithm \ref
Results obtained from running algorithm \ref{alg:perf_comp} with $N_{iter}$ set to 3 are presented in table \ref{tab:results} and figure \ref{fig:results}. All parameters are in their default values and a logistic regression model is trained.
-\begin{table}[]
+\begin{table}[H]
\caption{Mean absolute error (MAE) w.r.t true evaluation}
\begin{center}
\begin{tabular}{l | c c}
@@ -389,7 +484,7 @@ Causal model, ep & 0.001074039 & 0.0414928\\
\end{table}%
-\begin{figure}[]
+\begin{figure}[H]
\centering
\begin{subfigure}[b]{0.5\textwidth}
\includegraphics[width=\textwidth]{sl_without_Z_3iter}
@@ -406,23 +501,22 @@ Causal model, ep & 0.001074039 & 0.0414928\\
\caption{Failure rate vs. acceptance rate with varying levels of leniency. Logistic regression was trained on labeled training data. $N_{iter}$ was set to 3.}\label{fig:results}
\end{figure}
-
\subsection{$\beta_Z=0$ and data generated with unobservables.}
-If we assign $\beta_Z=0$, almost all failure rates drop to zero in the interval 0.1, ..., 0.3 but the human evaluation failure rate. Results are presented in Figures \ref{fig:betaZ_1_5} and \ref{fig:betaZ_0}.
+If we assign $\beta_Z=0$, almost all failure rates drop to zero in the interval 0.1, ..., 0.3 but the human evaluation failure rate. Results are presented in figures \ref{fig:betaZ_1_5} and \ref{fig:betaZ_0}.
-The differences between figures \ref{fig:results_without_Z} and \ref{fig:betaZ_0} could be explained in the slight difference in the data generating process, namely the effect of $W$ or $\epsilon$. The effect of adding $\epsilon$ (noise to the decisions) is further explored in section \ref{sec:epsilon}.
+The disparities between figures \ref{fig:results_without_Z} and \ref{fig:betaZ_0} (result without unobservables and with $\beta_Z=0$) can be explained in the slight difference in the data generating process, namely the effect of $\epsilon$. The effect of adding $\epsilon$ (noise to the decisions) is further explored in section \ref{sec:epsilon}.
-\begin{figure}[]
+\begin{figure}[H]
\centering
- \begin{subfigure}[b]{0.5\textwidth}
+ \begin{subfigure}[b]{0.475\textwidth}
\includegraphics[width=\textwidth]{sl_with_Z_4iter_betaZ_1_5}
\caption{Results with unobservables, $\beta_Z$ set to 1.5 in algorithm \ref{alg:data_with_Z}.}
\label{fig:betaZ_1_5}
\end{subfigure}
- ~ %add desired spacing between images, e. g. ~, \quad, \qquad, \hfill etc.
+ \quad %add desired spacing between images, e. g. ~, \quad, \qquad, \hfill etc.
%(or a blank line to force the subfigure onto a new line)
- \begin{subfigure}[b]{0.5\textwidth}
+ \begin{subfigure}[b]{0.475\textwidth}
\includegraphics[width=\textwidth]{sl_with_Z_4iter_beta0}
\caption{Results with unobservables, $\beta_Z$ set to 0 in algorithm \ref{alg:data_with_Z}.}
\label{fig:betaZ_0}
@@ -435,9 +529,9 @@ The differences between figures \ref{fig:results_without_Z} and \ref{fig:betaZ_0
In this part, Gaussian noise with zero mean and 0.1 variance was added to the probabilities $P(Y=0|X=x)$ after sampling Y but before ordering the observations in line 5 of algorithm \ref{alg:data_without_Z}. Results are presented in Figure \ref{fig:sigma_figure}.
-\begin{figure}[]
+\begin{figure}[H]
\centering
- \includegraphics[width=0.75\textwidth]{sl_without_Z_3iter_sigma_sqrt_01}
+ \includegraphics[width=0.5\textwidth]{sl_without_Z_3iter_sigma_sqrt_01}
\caption{Failure rate with varying levels of leniency without unobservables. Noise has been added to the decision probabilities. Logistic regression was trained on labeled training data with $N_{iter}$ set to 3.}
\label{fig:sigma_figure}
\end{figure}
@@ -448,14 +542,14 @@ In this section the predictive model was switched to random forest classifier to
\begin{figure}[H]
\centering
- \begin{subfigure}[b]{0.5\textwidth}
+ \begin{subfigure}[b]{0.475\textwidth}
\includegraphics[width=\textwidth]{sl_withoutZ_4iter_randomforest}
- \caption{Results without unobservables, \\$N_{iter}=4$.}
+ \caption{Results without unobservables with \\$N_{iter}=4$.}
\label{fig:results_without_Z_rf}
\end{subfigure}
- ~ %add desired spacing between images, e. g. ~, \quad, \qquad, \hfill etc.
+ \quad %add desired spacing between images, e. g. ~, \quad, \qquad, \hfill etc.
%(or a blank line to force the subfigure onto a new line)
- \begin{subfigure}[b]{0.5\textwidth}
+ \begin{subfigure}[b]{0.475\textwidth}
\includegraphics[width=\textwidth]{sl_withZ_6iter_betaZ_1_0_randomforest}
\caption{Results with unobservables, $\beta_Z=1$ and \\$N_{iter}=6$.}
\label{fig:results_with_Z_rf}
@@ -470,7 +564,7 @@ Predictions were checked by drawing a graph of predicted Y versus X, results are
\begin{figure}[H]
\centering
- \includegraphics[width=0.75\textwidth]{sanity_check}
+ \includegraphics[width=0.5\textwidth]{sanity_check}
\caption{Predicted class label and probability of $Y=1$ versus X. Prediction was done with a logistic regression model. Colors of the points denote ground truth (yellow = 1, purple = 0). Data set was created with the unobservables.}
\label{fig:sanity_check}
\end{figure}
@@ -481,8 +575,19 @@ Given our framework defined in section \ref{sec:framework}, the results presente
\begin{figure}[H]
\centering
- \includegraphics[width=0.75\textwidth]{sl_with_Z_15iter_fully_random_model}
- \caption{Failure rate vs. acceptance rate with different levels of leniency. Data was generated with unobservables and $N_{iter}$ was set to 15. Machine predictions were done with completely random model, that is prediction $P(Y=0|X=x)=0.5$ for all $x$.}
+ \begin{subfigure}[b]{0.475\textwidth}
+ \includegraphics[width=\textwidth]{sl_without_Z_15iter_random_model}
+ \caption{Failure rate vs. acceptance rate. Data without unobservables and $N_{iter}=15$. Machine predictions with random model.}
+ \label{fig:random_predictions_without_Z}
+ \end{subfigure}
+ \quad %add desired spacing between images, e. g. ~, \quad, \qquad, \hfill etc.
+ %(or a blank line to force the subfigure onto a new line)
+ \begin{subfigure}[b]{0.475\textwidth}
+ \includegraphics[width=\textwidth]{sl_with_Z_15iter_fully_random_model}
+ \caption{Failure rate vs. acceptance rate. Data with unobservables and $N_{iter}=15$. Machine predictions with random model.}
+ \label{fig:random_predictions_with_Z}
+ \end{subfigure}
+ \caption{Failure rate vs. acceptance rate with varying levels of leniency. Machine predictions were done with completely random model, that is prediction $P(Y=0|X=x)=0.5$ for all $x$.}
\label{fig:random_predictions}
\end{figure}
diff --git a/figures/sl_without_Z_15iter_random_model.png b/figures/sl_without_Z_15iter_random_model.png
new file mode 100644
index 0000000000000000000000000000000000000000..bdc11440fb670283494c467b67999c20ed7d3bcc
GIT binary patch
literal 48731
zcmagGbyU<_`v*!0NQ+Vef{cW8gEYbrq7ozB2uOD~9Fz{}ZV>5|Zlt@rySwY|Iq&=4
z-(PoK*J3H0ab|z_e)jWx$`c?bBZ-Ycf`Nd5fc-%V3PV6Zaz{WwjChI)ei9Slhz5Rp
zVk`PV;VJm#_Eg^=e2;D=rD}_SfcYH$g*Y<raR+|LYbUN^_r>y?oujUe0fM=%ot2rT
zotg31*A50Yw#JqguQ`}mnK>9=8`;@e@v*S{-wT*6Z46nGG4nGK5MCpEfWB97O4>_y
zbRuY<^jduJlrsIhh?g9dom}3tpE5=;e1cr*Qt!YAd^yImA!rmcMxYcrJrbeFogAJh
zo?I;~n<s!+&5KGT;`$-3e|;e_aqSmc{H627hlBX(_;T>eFeeEw8z0rDMTeCb%lDC+
ze7ex<-#?@JP@b|$r2gm6ZJJ(eH(399)d5oQupxLc_|A$*Cs^pef3*%l@9gD&UNVP5
zoOy`z9@-J$|2&+=p==_?%*kQr8~8g1uOUGN-1&C&1H5k_lI&lK>#&s%BkMxOV!Zv&
z{;XW~n9Z}l&^t4D4S`MM;eU7)Tdv;uh~Ze?!T=Fgn{4E-v+tHV*l<n2^Y5a;3+T5H
zg!Lw$u@i*P#T1#AsT287GOZ0*pgd%fAraR7?96)8Z`=%7(&rmtom*BBq$)&28G9an
z8x5-~Q)TjL(Of|2KwO?K^l!1^`5N5Vx10u}r+V{3L0TLE?`8b2<<K=_JKTHyCHp1S
zrRY6{E*&eP>@nUY4ToJ_*15vC<#{tvXgyTl%fA3Gyc}Zl1^Q6AQLkWyw^(?M27wI+
zUgp;6qT7ChX8HTOK_Ht@ss-Z*9cnyPot#}`ym&gugN0VHxe64DB(7iTe`hM(BnjmU
z>UxgHO9Nqp*nEI`5QfJL7$Dx6-IJN}xRLMT7k|%QwZ^NVge*hK<e*6NL(0&#tn5s~
z=hmTWd>f@6u0dV1kTPXxQ9u^~9`9R7YJe2FV^24B3Mw;Vo0nd5BxEJJ&%S~%Tyetw
zuYVUh9y&p|iOexZK$jEVGz(-HVj~arco9AsA$9s#@Q!VQ*D_GbBFCvxmmhDN5>i)K
zF&Nl&3t7=4;%RrKhnz##-Y_)fLu@pm4*p$1%yt#Umed8CJOB7Y1G?1ka+9IXiiF|L
z5SyMnj!mfkAAf;0{6R6O7Jm3*p{x66s0Tr~WMnN7^GTWDDCFcH|2y62%^W2sW4xCP
zkXa~I>&Rw_e?2YaI=Ps@2y5+mCmnTL^2r_zWVf*PB@={SJgYeIW}<wAY?q3{s2lfP
zH!|b30DpX2k`SKilfKe_#P-i*IQ?Xy7=a!sZG|>Bnn^RLb=7~w*GTJZZa+dhrY3nA
zv?9J>B(*`axU70HEY=kU=>GJ6s9Y>+-8&@=M=RiS?RVJ_Z5OX&NZ1tDuTva+Q|LO+
zhBE&^Tbmq+H-CXc5UwF(Z4)5nx$9hKwr!&TZG0B)n!HkbZ#FgVR|aAKTPw>`xGVD;
z_k&=5s_OE2!D6mU&aX-=*t^Af?%{o#KE_Y7+86Tp^Xlc~Du_@dMQ?HuSG?TJsGgqm
zxq1=e8!;WaC3Fpx5|T^`1Qs1$E2L^7Cv+z{=ng~nTyn7r{uELYq9~%U1y9x?(&UiA
z39cNH6V{a~lw4VwscNMVr3azG+rAm~BYEo20Dpzjq8#E943%%mr%#5Ipd@dzLus?K
zVA26yM9s6!K1w9q&@4nb^0>^XTH}DGlC$}zcShYkNXg!^QiiyG{+wv^2owz27Ah=#
z;mw)N&xwu~pU@yFdmR?s9lpoKeMH?WUx_ZUE7Iu}MyU5DnK+j%{b-wZep?^CL7f&K
zEi0Iok2#)-pq8E^P=Gpv<l{X98BwAm2qV43cWU8tL@s&$*62Qk8bS~a^U++>B)%FS
zaYVfFutS;%lv2zdbFMQr)y+;Mx*~svaD|NRZCJA7<?oc%TgOpC6c^@UL5|e;B3wBN
z#-X_tf#2pOp|tRbc4yx0<BizNGQssW4bjEFBG;c&QMz5tfRki?SDl1dFV;bn7Xta>
z?cHpB*5HpggJ28I3WV*1;GfL}Smda^>-H&ZQWU2Wp^Ij<OMJ${7|+@Cez7yOxvH5W
z<}Txl0E-TGt(M1+#$rLQ$MTsTcZbPoGR@GMf@)9o+F*nkf2k7Kjs(W!PM}`13PbX0
zvuIt(7h|cxS0rh>owChlY3M_kZw>D)2PK<>wl18xYx`;u89+aQnfL~sEmn@G23s6b
zAa`WW{<FEa@>z5;dA?b|E>RP8KBhoKm|@$Q`x~PD(4a~y|F=Z@Lcu~ebN`1Sce1OG
zbbtDv=bR_+SiG#JiZ;(sUysXuo81@w9_{8whL?|se^6$gCMYj-BIAlDD_+wwC)P?c
zpZE09ZR>nzyXn=+WHqz)*|E&M=<I&<ui|f@B%v}RC?>V0lnf+pNKCI)uF&hyyk=gB
zvZZA3wy416>wckG_vPwvD^^X9*7?P2!m9!7<a~8!w`Is4GDAW9d7b~cYd}!-f=DCQ
zZ|&1U0TT5l2Gyp=LGiiQ6>L^do?2m?J_($@*V4^C^$oI<p1s!|rLWWW{>Q=QJ<kk=
zQ=iPjWJEjvxH@iycczz6Zgq~gmtT;}d3m<7F|D(xd*WB^#UU08zui9v^6#uhYX(%7
zv-ci6r};c<8-zqWbm|ljku)@Yz9gUzo>8Zj1s#3$3d3~vx7Wg@y{3819Z*kRHlT{?
zG~InVH6L_4Kja*(C?Vu<`G7QI<T!+WH5Gv}(@m(#e(M@7C+k8TZYX+N@#Lr?QrNIt
zxFT|0lvLkZ#$KlS{tnANXP^e7JYU%qcD+Yq(iDkV-m8un#;2(rM_p`4YiZE!@O5$m
zfqIz2l_&W?Of+TEOIl{I1!r%D`y)BJS^N3vX5xflTageMwKab<=KZG>>u^iM-TH8c
zvWmtp^GrU_^Qf_^0E=k7Z{+bQE%SC(vs)i`ocj9hE|4XY^i_4w6s@ggco@udp4T3r
zw$4}o`rg5g@iwZ*=HCA=OW+#G?hTb*a7M_i%IhXl=fhHxA8s;}HslSzB^Xt883s|}
z6{;{)T*z!AEdP*Hz+_H%V}vweQdJcMhq#-*Z{u&}X2s)D2`Z;#VuqwV4j~?&8TM!$
z4~#KfsK?}QJKm)?im&LuOK(~tIEqd#$e~W$mhias_AKQP*=r>)A8{TIv3s7fjh<V-
z@b!a0XZK5YVU_bN|Dci?C6~k6@XJJ{Ic}Lv2XRs&irMyi%y$QN9xlO+BQhrf6I}P-
zXiYziP-s(olQvU2?j+^2mMv_3jKZx7nSXF7`w>&%6LjTMSVGBs@=0OQTe6@;c28f~
zt0dTidvB)=^FtQTJO-{qgr{CNL$&HdcLUf<WMv6$lUZn04;u$AM|~nMqq4MM42;T@
zxs`OQ=3ToLnMGI+)qKQC0&M3e<~8!fVgls(3S<uREi>dtl&e_1WMrTrHV2M4?)1FR
zgH3tPlN*$u&(Zl}KZtIwY!!C}zQZQ1*1<a>uWh-Ri}>>%m%2W`fL4{xUcQ3VPbgy(
zi&+vk{s1Q3UXyMsZHIiKdNcJRuyG;B%W-e*yr4-q`=`avbH|X8wu0_sgd6JF4^WNq
zL*g^5K#s&D6zN=HDAn`d(V{8bP%0Le>T*)Km-DVVVfoIHeo?T2!4cJ6OG^HD^>u0p
z%J3J`P~<k~UtVeJ-Qs*3slfg2waHgb<mMU#t=8lYCK~}lULMQ!E25j<dD9QJ$iCfU
zOI((qW@?-1Zt4Ef`FQwzD`o3CcZiJ_5(UE<Y=rIn^1lmOb^jFiNZ|WII=4xDjaBY5
zv_2X({4z`Z+lq-nWByH|)%>pJUtHTjJ=qMyQJHU#X3U5;fn7Lumd6rHj`pr5A6-&I
zvQoUooAbBA%2*jvMMcnb@?kr_-Pbtbg^sM8Rr&c-DH#VTK!PtO>Lb7F=28r|S;=pq
zo~fQG8!Ooep*!pD9&*qnrQ%w3ZrR3%`QwWZF24;(@jLU}id7J}*=?Y!(w3nMPqU@n
zl?rsT$U3KXDO!%4<GyiG^GwDrk=f#PkhO9ppKF4!qr}DOy#5vb{lszLltG8y#oMo8
zcoS1YSz7$p&!Xb<%R_xVV~os@_3-Y$){4I#Y49!uYIBA)8^t`I?MV)%-Vxp3MM(~#
z*eMDZRsW=B#3@bvy>G#LrRv#>m>ChUDKjtqI@j>))0>nBpMzX}@O2_G$Y6Wvj8Ith
z(<9k2ZBr0My~)ShT{_oOoC<4;8A83^+6GJ4v#G8POFPWIcV{%Zzo!kjS{C7ASV-16
zt5NsF%n+U{uJfm`AnJ^V4*ohvB;Cb-q^@AQ<gA*Fm%_V#GXKKG=S+~4ITbN`jJsFv
zkP*V7-XTw9H}7bKIeIK(UxlR+KEpknY}2O@JRIweB|cbak`Z+?FC+qWsTFExdhTMX
zQ79l?ghXrblZqIdDGH`*tk@Icc7v4d#S;xXoJI4OqN^C*WJc!`&WuV(HNoDOlTOA9
z$KFC=c-WFyW77#MOBAqZZ}265a2ZvT8o{`a=0mIwO+kw@fo8Ga^b#!n3}P|zlAiec
z5HfMtE24@=7x&L{tj_Pc({Q|i6sIIvI(vrfE(iS2qFTYTm-MR8EXtXsTXMn9ZjWIZ
z*^1#<@;N5IC+?FB+5e2j8zKdypbs;&`ekg6PAExhR$wyx;J$&iDK168^33?opmYSZ
zkfsdfVnPw&SZ$RS-HAn=IrYEcFRA8IS3Z0vatw@}xM2UI?;sQ|lIDL^y*2I7_|5gX
zU?K!YY)(40qCR-b-^@QDCGjy$*kx-9d0&WLa8SaA;%{o^W~qaS=hZos)L|fKa^yzF
zJ|8RMwW4`OBAB|wu}ssFC`%MX(dASw7CLGloSdU=@EklzMm`P|zn>YiPW(=N&anR@
zgDJ_Ed4(1$zSacFp|&|Vj4i@b(Xwe0Hx7=EGJ8x?g#Y826yJ&L*19@opS^W~T_9TR
zB>q{?>e^JOg_XK0!(qg#qg|%jq~IU(YMU?eFYe1`^=M@dQfS7)6q$X~N%<vZkS(HP
zVLQ)j(b9e2)jW3T1g85H^URXE5q{9LB#YU^A<}Ki%;DGn`&5*;mi$ce(yjnmPWlz}
zk{(696Ma1e?Vq&+SCbEMMZ<l(%_KHO!CJ@hL%p#x7wo*NOuX#1nhrGGyQm5L?WLX3
z`SDMt$%%B|X3NlLZ=@eub1Kco1Pn^FKJAUn(7fkF_E3=Ky&TFuPfpljFL_gF{xxDf
zZkjlg<@Oy4{vQ`>gjx2p_^eE{377Xt2nj*S!hCpip%2fcnHJOh>U{&He%gXF^6x?p
zkp0Mo&WxL+ZR9iD#x_7z|5TGJBXKx?>LySMZBqpAIBIO4w+3IZU}BVSWw@>J!kCZ<
zPZ`&vRoS6JQ-cC8mmYGN6;*1aE$=?0T`BuE`-YEV>%&>BSCP%E=WJ(8%3+|C*O!DE
z@hoC5*bd{2Q2~1ONw=YPBZaE5$6C@a0v-{lM*rmY>rkOCfgxtR=3ytW<+TMAkP%{@
zIr{3Bsx8WvP4TR7+qB5iGsKyxcsgb`sz51w>rqbwcB%rZBGZ0@ib*}U=bIYPs_mkU
z8+R}Ag=EwI2AheR`s(GT5lG)nVu?tz2l=$inV-D)+~NAigT|bg^DPu5O^u3=tHVLU
zdzyu0BEpGbfl{r{gds4zt5n;S3AE02&aK+Eo7vulw?yI0)!clrkF8uGczBj|XUIdB
z+k`UaG3}!eST}PcPnF*{BI>S!^e={YL#Sc*tj@l|q|&)>jFe%ohAk2WtnP=CCmc-z
z4Sr%-R0ZTRc7AHZI{Ze#F)a349lFL~*KQnAG~|hNnf+$pc%oDh8j~KICj3tdr<xEK
zI$d^#cou_;%&;Zo<b{h&G_A{lgzXtwz29g$ixGv1Wtw}@A5qeT^WhhJo+6I>v*i?m
zxLG;=e^riGL64!!j`BYfxfCfV^EGtL?0xr4km?+FhWcaR+w3C$F8Z^i2B))~{BoJ$
z1H{UJ1zblS7izr^yGbg|3zAB2&gLSuhb{$pXVkVCl^ulElg{cWA=(+6L3St`blHHf
z%AL9W^?{sV&-{FAu+EF9iT3{XA`hpY1(G~`{uk5SY_geM0|*6BeFPWrd2PrE@(pq(
ztILSTmx{rB9~t@wikZn3<ro~?3|F3srK#UE&a%V4cg$xW%c*Xt)etp{mDAiEvR~$>
z2z!$YExd*-s^8$hza*xR7^Fthv<Mjz%x4nUe^o0K4J94?Gjz_~&%gCH`>Xhx-kAkM
zT|pUhu*dO_^bmT(iFE9uY{<!*8Kb$hsF{K&HA=`~W@_o130pVR`AjDk7f~0Ux!DbC
zG#6@T`7=&GOb1<E5y!g6d0md1%pESgW3f7~nyjdsss2-QmA1jbjM-q;{DEY#sz11C
zuHR^;+ruh*oL`rSL6tHNul2SJ#}pgX{8Fl$i073$5rdM565f^K-BZqxPrZexdwr*S
z{b<-d8VoZD@$W;tLG%plvaz_`X27Wb6~(u_GRA6G7rm5uAb(bV7Z0Ov>C_XlQ+@cs
z+{Cd}dqq6s8}~I<>zg2HYUbgh`c`dKt9TN?wQW*0SQ-&M+-zLyanAbdu;s^_7V}sp
zx*M<Z6=84gXBk5*ZG%H0R**@|>g}?2<~{ogj^RY3N%Ufy+2TZoSSCr>5ZAMNMO&<S
z5p8JQbsWx<8tX#3T0RlS@W+21A{v7&uy$)}9T+@b<QEjFFAo?f;qSlZST+?Jxq19@
zM%F$yH``gW@3Q%lXxU@SnU=KB)fxGIy#|ZlQ%&ZQNxTYoOG`o=<G^Dm--@7Q_`?E@
zVD9yE7wTziyW-{sAfBij*iVhzBh9?(Lm+bILT%h|uvB<AvwUF{Vb$*K@UwPyWZCj8
zXfhhNmg<o1na;Kc&V8R*)yU^)(-8dD1My;inIrQHo*6HrLwwnx)2D3(wIn{=aY~u4
zZ=P!=2GvK{t?K@8Rkh#okAA>A`)+Nie3R)~`YR)KQ)QFAbVcC1M-il8m)}&g&9vHT
zGa~pP-+Gx4y#dCy%!_Vktp$l?in(K}kjFRwK~akNvCLhX9&3N=l0Tn(QDk!0y80;A
zc-#+NN%e2q>>pZ<aRvqty0fQUC>6TlYl3!sEC~AuS7nM#o|FL>Ou6-d(d?`RIjlPo
z+`muxy<al;Vm{Nki#})7$Z)78X=i+PE3BeislHrE&}1KCUIjqjj!zJ0P`WDvL3edK
zB;;@`s;9X|&^FvM(EO`OFI#!&6AC>bBmPYHUsy238B(W<9$^iDD~{|hIdTT!iX^Jf
zGIKMr`+xOGN5qaTj(?R$ooY;0+y7G}Kgs@|P<cGMiR9_&saR<?{i%QT=6rXpH#T9P
z_fY85OX@JX)F?KCD1l#Yu#PDClem))DFyVYulOCHZ?hw2PDXs$ZZza7%~UHb7gKVb
zzyJ7wM@E*E>pZs|$1Y9&MTkzX=dJUn>#Sh9%r&<4NC8zvA;?Ooo;^ps>3;=Lp8rol
z^#Av9nAJz5^!2GW2U9mE3e}DI7AGfTk<oDUOOzg&z|sek<%CjFQt%m+g8xCm{8mK|
z*Iy?i*CkT@=a4_4%pf2YAfnP|%Ve`BQ#SKA7fa{T$mo(ct2bkcAXq&m&WyU@yH`%b
zZuES(_Jd_hFYoV%_-FBj<SQ4$(nP$2<u-+bz*BW^uV+R<`1lga4F{2k+u(~Xi)Vbb
zK!8V~!z4%Ggt3LKfHV={Kg*mio+ZGF^)3|}RVwn7^@S0x8A?o`lpiO9W#{$jCY@p)
zzL1cRG5ZykZyh3DOJ^tiL#-d&U}1Ui11<|JIN07ri$co-G8BS;7e%wKUGW?~&uBb~
z8;(NDHVHI~n3dQ`nU1UbdGU+zz~;JU-0cQ`^(qUn@Z<%3@UBRfmWQ0OKJbkf>j)pm
z<RwR}q#<@gkj`CgE^f{|pn46FMY8vz`~Wq^Ttg(}?IuFxf@W!$M9YYV1xs~E9Np>s
zvx#PClz=KG(|qPj3f^Ymb$&>rNwXHwiF>m<T{0uYW9!<>l0KW6CbAQI$v@I9^_-{O
z?Y}zsgQ@=ck%jJ31m?AP^I*v7BH9DeM$XSpgX}1qa7oeX|0w5TztW;IHue(8nyiu{
z&t+kMbEF*dpXi765*q4ZozXAo<Awhcp-2nG^cmnSRM@113?dIA<FjZbm6ViFt=`<+
z*qrZ-M=8j}R9C<A<6K!;K^RH4FtIiiH;899n5vspL^I!5(U8cL2>H4*mdm_~ih@$=
zus?$}xm?`z0D1e?e^BT)m|3${-{}Tn(f#t(=0I|$=i>v;Nvh{V{6I2)`L}+Ooyp?V
zErHFG?(V;T|5Dw4G9a8R((H|KJH-FLT92ELbv)fZmYpW5!IeGMpJwU(4NH@D$Vbs{
z7<NRT6g2m!K0vJ4)YD<Ml7M2Umd%MU?>bh%c2Jdy)BXJ;t>cCww`^2Y=+4oP&_(3r
zeq!L@$iTuueP_)o`j*#)Xw=vs>epa~GO1U-yE_=meeuU+9709)m*({3q=#_w^?*SX
ztt`3CdQbbW&xz05ndH($er?6Eq-rS~fPyeKC@~qQ2e)=0(Xw&-wdHsofqJD`y36^_
z+F+{iuru$xL6D&H>BjoQ?cv~wTU%?Z)PZW_EV#q$zk(w=qXx}A_)kgOCb13sOSz&L
z<swvZ`!huKQD!R4BIom_yG$}IDO?7UIW7@zzr4HD+I1QhpQ6@z#@gT0hWXoTwZ8Ft
zjT_y%t@3c%dt|>wEiJ99+gsWQ+V&7qo+Ge+tnrAyn)3&CZtm`f38q}YXyK0*ez^8o
zHuWn%DUQCnz2dlT=|02V39HI+4BbWP-}ynt;Ds8KDu@8hVjP;n_!vsa94ff=gObV1
zK&gJb`Ey#%=yiiv!-R>h7|GUfCK>^Ql75>nHmA`Lf~BRUL@3#ZPoJ<Lt#yZuyU!k%
zT73MU1`)A1{47>2_gz}jIoTYPeYm@J-b(jlH!}?Qfk^~`+?`Kq2ehhH7%l5SKPU4=
zqhgaRZEpIdrc!{1*vLuskY)=5KN~92tb?9-c_C=I9bfQt*Vt}uO%_u`(<`=uE6#4v
zGn6As(c0E#ZtZfm)uOEJ-aecuiH8UMo~M|PRN``O0{vd7UNy7Q+1Yuz;Cd*Z3!@Ii
zrxVSEWfQaNAWRl(SG9krrdaI=(K}h|eu9b$dK04BACvZUF+W)SzgAOLU%J28naYHZ
zTDfWI@)~MKC2?=5a@et2+vL$jU%bV9S#6Fb&XKTAjX##;U;`;;k_7G|(rbu!v-}w|
zi5>!=vI7ADT?VhK+~Z(7_5PnweEqu=b!dlP83iO-=#I2Dk0VdNYMkA_vbMIaq||J7
zuyC2H9K-qDU*Rx1;?~Fyg+hnw94!}Ic9{<94uxiqa_Zbp`+4m4XQE}2`HtI(w4V_X
zb$&;GVKh<j((~>ZjaoWNiY;VnY6{AZi;F8DAh0!8!@<&U_B@D?>1$scTjkMcj%>MZ
z8|t^dINiu(0p|!JEhq2hW)X{p`tGVlkB@Z@`&a9++D5|};(0~A1bBFO_;2OB0$$Lr
z93BqFa~L{3-d`Gn54*kCZyOzrv|jD7UT#IfAs6Tx$^M8KagNCAbfoi3Cb_5C8`*fF
zo>wxIOn0p-d~d!kmdkYFUA5%(*{E!;QXvV4;Q$pqePD|Z+Gy?<+KYqvAV0rphgs0C
zM_RjkX<s&YaAICCHuG2Uh$UhfSY@_cIoFkwbGdfIG67%LgsNrC7k$qlQoj3V!FNAF
zQvuN?{34}Gx0y=^p{h!E_g9o!ywW^&0)ukV(YUhqaHY9=A|nk=N1^TN>gr2y{^qll
z2GeB*+~B76(`h#ffPIl2Q{Vcep+N>lU%(n@X?&41pCQ_fZgx9k0pdY~Y%XU;iFjw%
zo2iiu%FMNnhZg*XGZnL_Hg0YLpwKqJ6ZfSE@{0vGl5m-@8uXZZu9lXTj^!(d^+Yqk
zd*4izrEGg30Ua#sDY&+DIP1QCey1&H6i2O?EUflB3PoBCJ$hk+66vBp;Ds*yc*lV&
zbd|>B(dc%qQu_7B{^|;t)@KLviSu=iUvgxU?X=-gWbSiPolEF3Y3FHosO-@*T&==K
zda(qYbd|=h(}fAM`qw*}reX5~bboohNf}(CVJ#oJR=unz!l|<={ra7sdx+v{PmSun
z{H@plwq{l5DBIWl_QHCj?<I{)($<i=a*59So$>rmUu;s?r#H$)8aQpq=ue(JnR+;!
zuj2uAwKA#gnf>8K1Jh}Sc+l2RI`YHAL#0#G{RJr{CFR~!sgLL5{UlAXV!l#B(csH$
zd<wTCMB9zN<*VaW9I+BVoL3p(^Kzvv!9o1(=oraYW<hfvOc8W`xZYH4@@ViGo&<;I
zx3C6w6~Ss}jP5_)Z=!~OZS!AkTK3b+__@HQ_VPce2GCQyC}tJDI>nD%)5~S;R%m*b
z8rd=5@0DFI_C5@z418fu2VLDF`vpFJl$q1+L?NAO8C80EI;Zi-_hf#@GRs9Fi7;~W
zAltnuHnl1X(3<S8wlc%sYSwas=Ce^)wXm``U7lN=mle-x{M@qPoY1oInx;_gPljB&
zC{$8%wBFga%2jf|#z~$<z>ys3^5Wu`PRl=)M^7+_95mjhmMqaQIfLQfHZQRjBK6BG
zzBP*JOSQu(NrTD7SA|n~g|`}%VDTIPd1hfKtD88xxI}<b19x-#b(7WdKW&E)@e13X
z?nv6v0@ZTp_F8`;kC@%r(a};jZL0pkTn)J4KA=*UH#a-maD~l+%E6=pO|%1?(EiY^
z<Ngen@&}rc(Q7qu&^2z?Wj<!dEA4h$LoK>pe1e!t&w~|x`=0zS3s8M@WJRw~J--E>
z?8VEM#thnU`#mf9x4L}aojc77&*ZN5tsuL`Dh@rh;N`{)#|K|6_{tN(e3j2r$c5r0
zDHp1(AJs{OkVG-6Mt}i)G-Fz9kSF7QbEZ3z^?^~lF^Q$=F2Z8ImbW@4ATSW=wd28T
zSo1=I>vW?#KWOyQ-20$6c|2Sj`nA6^ohVrD2qDEe&dA7k-Y=ZQW2u#KR$pK59~vri
zoB*l=3U)xX-iZwik;?9-s};;B7^mfeKszTKMJ*+emh!}v+v2S9gwo!rz1wG~G&Bn;
z9b;1E`MiyS*0$dJXL|pEWbjibvAZu;x?~vnr+D@spz%G&$NyVvziXB+|Dr3Jfe7@i
zEYry%sJQrWjjbVkpg1|bYkIh0h-OgU@O-@04s@<b7~S07{xDx#MbQVAC6}+nq~*HL
z1xD5Od|j2;9U6t(PcUTQm5z!<+zh%ZF_#IxK~I$Rt{M@uhSXxC`}TA>ExdJh1iuVW
zVy`lt3<aIQ>vF+0jL&}OXf;guWZrR+TKcf@_CRmB^+#Vae{2Gm>B{P=*dLRCc~#UW
zPmX#RidoHPI8QeEKY~7HmhT8q$_G&MU&iL)Y$x<@M`xMNB~phOaa%4!juH7(QnqI7
zNY<ywJtd|#kj9`CgRe@tSjzx_J^166>Pc=4rdp~?g^PgRqf%w@>ahM4XHf8xY<PG$
z>aF}gphM88r7Z;l_(;NS7AK{ZC4~opK*;nPTra;JEwx0kXfLE2Vl4q+WZ3c#O7Ke#
z-_Te5Fff<l0|U&D>Eu>JRWMoN+4N9He|^>i#rXsU<rvH(8nq>MhdmW|E#DqC4T2Lk
znl7WxS1QbqCg|LEy16_Yyx5<0zCWL=v02AZ%u`qfK!o`;1zb}^M8v*$4!`ZuoYlia
zS+KXxo5_wEI1S()`m}Hz)29?)c(ng6ktb_(igTgfJvQ!?HP&oA_pB6*UaKl?;77;6
zfX#x=&3*?O{G97y!*<WAjAlb(^xTKpU;{Kq*tPy&j!42jLiujJ9$JHc>640(G@TNe
zT8qE^o2%2i)UmfpKBY_@CsNUp>|;}-ZFRBJ4RvyX)}ulGdvem$O+9#{qp^Cd9V!3f
z_;^29U~1TX-_f78>5EHldV-?AvB^zY>VGVpy5^Tl7XC+3A4bji{sicn`oW<|&sB`>
z!5iD{FZsANg6xA>fdSrMa@?vNumS$6FoXU0U!>-acKY}d$G4btVCPLfI^aTR`@(73
z$sSJY7t+^joEi9i$(XZMKCqO%X=Itvo^tH}C2g)`OAqgaBz!~G5|J0>rs`N<sLQT-
zo<I79<;ecpxV>B)DyFzQLdK_mLM;s)sV=qrwGmJ9zx2+saLaIzB)Td^>!zoh%!9!7
z%W*e+9}=uq9?K09E$Nn(_M?+t7Zt}2ER{;EB8_g<AcK(q27i`AH$JVs>?oq>HLMeE
z8*Phh2|+zoTCIA!5_!!}aDU5Ee-^=gLc-~Vf|+JxaCNi$X^?N6IP*Ugxb0Q<_3mbT
zi8WGr<sd|>{q*ZR(aRaTx6~L>QrBx(I@dN%56;zYFBja2P;pUBGk=i<u-({Tv`j3P
zn2r5=7{2^oABSV(Dx9B>dM3(zGQ*M`cL@ydmwc}Bmi6OFliN#FEUzD!=?>;9<)|iO
zT@eMN|L<1M;b}BjM{hTMHXPabpiyJ}gM8eTmCuPYI!IxF{gubV%M^FQfn*=NH#%2V
zU6Hib^(CxNF=zYtwUYNh5BiS`VZ%9Gl@b5Po$v+s=hqPWe<bq+TBiv0a#icU#cW&c
zlhj|+^d+Du=C{T$$rWqG7^G6py<+~2-OcX#;0FEvMlv*uzi?@333|K#%Mak{gffHP
zZz?WeE>D-`e3g=V3OZT;i5oVMK4dn&&3*o6nfkdMm;U5Nw%FFQ`TlO=k|*X;B_$hm
zu_Ht-pRg|lySNF5-*Leil>)xSUC-g;^50mNRb=(mJHi81squKl?WR28b&TjYd%3J0
zJNGZnKaLNw3x552wi3M2e|~trt2;t<bTPnAz>fbvF@sX_S@FuX%j0A)7HM~jzt0$*
zydN&Vn^axkewka;w_As;Q8XMfFCy*lmy3hEv$@#>qoamaK>))IXMZG;j%5z1s^S5+
zUKhX*Fs<r>o7}E#@|251V`GT`?l9<%pbmhKZa+zK=O@C#!u1Ue#&gwdq&ya%<YRyT
zehUZK0HjH#2spo0E+Sj)45JkiTA*A2{l;jff(~HBY=vA{@GE}8>^a*g=iLb=Ut9_q
z03X39#x>}TF}ONfo~yO*C8T9!#KNOwWb9j>`@Vegf>s7$iX`uNwet<c|0&q@%B$VU
z;+)Tk+*da@5iHt5CgXYSb2YYKvU|axd4~6GFs1(|r5H&20ht8pY-c<OtSyv`S9G@0
zygOA`7|y&v;|wU8PQEQQpXHvZFcu4kh#wsvr@LLB)Sr!fw200HWlQw~7O3c12-%8!
zUCXrV>*rKbu*+U~IV*E3y<a}IROkFcB4kA967oj(+(a<!P=@EK`RvSq4B31BK=Ly5
z6xCH*2z?VZN6c@gxxy|f6$`Zc3_$b+?$1WyYyiMDwr_pJpm%qyv<JOfA?LOE4(M0(
z&!5i%2^cy6G=bAZfJ3su7sB3LO(fvaHh?&Q*-7T+<`yUD$_)VXHvss?3e_jG$pK8R
zFrSUNzrE!6`uC~H*zXpAb6@g1FcPt7=>u|x0Ovtv1|N}-kW@-^5CM3ue|)%2;<N85
z+;!i{%~2`SM;G(M$^H;YV>Vs(15Qc+f@{252~vojk6}_bnku1`6~5<WsXszO$o?3M
zT44=#-H?RGqOX&}Gq|Zq7z~}&weCo|9h<Eo${1$Nou89{RXxSQ`3ClTZ?P$rXVL8i
zU{qii#G5?si3OZk0is`tH%y@fx(7JM8h2q(EE3?La<!X;PB#a0zzHV3v;GG_w46Kv
z_^e5>@0?_Fo44XlUHIhObn@K+V?AI;_Gj;#-B%0y^ZBlq4*MgXIkJ<px%Y;0((!HE
zztyXXq(i~3&=TUvNpm?qR*$d6WcE$+!NMgn^hC-`XuIp3Pw~@mD1`Qu?yjz!>_K4?
zkdWj-q2f^JRBHl{Wf8(3gI>Z!Ms98;#ax&;m+{D78q?{r>^!wUd9~K7zGC%&Eayp5
zxGT_QE9mQIxq#hN(9ua_8fg82Nsu;CtbLd^T5Y`=m6MZmvzh7{1(U{p#isy(X<j6)
z?97Trox^89XrS9BshmLV%Rw6ZEuTrYrq>y&=rvPr7zUHh1&g8+0mnT+sU_-6F6=Q!
zgFIHE+eWt1-r1QC(4AtHg(d+JQ7p4&D1~ql!gz)_rk@i)-Y_o=Z0tYX%v$xo0cg(_
z6%{pOcRc8<G@m6q`Gn1!8qe5#y+BO%UtNV(aJ~HeOljQYJ%jkuys<Ud&gwozeqt#P
zhfIyaGt&IAzX8(g-I)rdU!~n|%{#IZ70=P8isniz`;IIm*s%9iru*OA_!&v6n(s=9
zlLUl*+F3{T4-Q@fm_I}G2Zocg^D=;RhYJnz($-f;x_;G{bGEqHB%CGp0{|3i);X+S
zE_#Bdk+QXA1a?gps9MKIN3XfLW2UB5qF@yk^9k6b+*G-pfJ&ibVP(l@N>rYH)zfo6
zYC)^9-z5ZUBZ}Rjnhz@W+b_R@_vgTU27s#*znF{iQLDj)x9f#RF(|Bd&bj5Gbg?8r
z+#B6*SC*G=<UbWM@6T2_UTOX@QAC)jG&haBzP^?(P+^lv6~bP+xc{w?`}_w6!O9OJ
zEgit>D^EeUr{vm0Uu=7$EhOqwgMGD{DjYZCKrg8!*m`Q8+Q|`X7_)18#mZ@pVcCF!
zD_r1uOYX|pJKpsijB~3X!xU%V>2kx<o&4gH(?Q{K(1(QXj>HhYe*H?eLdc>eFnrw}
zM6|ilpLhqLnnt}-IKYD)p|65K3AM4=ZhQg?VU&@F77|>226rxCAYWU%c%#;883_R>
zh96;JQPobz5Cm%JcNVKJo;_Q7mZJ6C?lbXD|5MGoyE`(E%~Ap1+6dY?9>+t!;$jv6
zX*2b^BjDXHitdXq*xyronK_*i9oDsXIO1d(on_=>57peHvgZ13y~kvvR>9ubs1ZDl
z3yhPrLge2D#)?$HXc-iqyxk;ZN?6a+fB8)&l~No$j!L5&FKF6W#H`IUkKj=?J08Im
z(C7gjNet{{Kxk-3Tboa;wnw~>`%U$yH@PXr5b;2{A`RYHOfiH+*{0{z(p0&U()i38
z#9sfBsoy%0etv$lb_iXG$~-V0%LTi4!Yy$h(Se76FpwgcXrVbG2H5{xBgG%n$&KCO
zCK^yc;B+~3Tb}?9(vc$Q`WsY3tvwqt)}x?|hEPk~xt>aiA>u+!!^;UfVjoiMUvdk3
ziHwdXSqqAK`)Wpu<Qsk7u<WwW$&7D$Lxk{Z(Bd#E5TB?ZV<|{e^S5b1k-T4Sy&wA8
zY_G$@Do=OrDPtX}mf_q4qbKNP!b0G&n0w94j3X=CX>kdp_PSutn=|7|i+K~$m5`ty
zz3q|gv0`ls&=^nFdnEy62Ze!azccn>q}f&+E+yUU7K6dnWWF^*0Lp*u{&I2iY<qOh
zc2E#Oqsb!~aHd~zlD55+qhOG{1f1bBApNPxLj|fF2uR51<YBIt2kSr~sc64l_M>>7
z9ZxHp0%b@BqRuBUeBn|OTqV#flCf2eJ^LAR8_w@&LBeU&JXNZvTKW|QU<OurtlVnp
zFyE8Iqs?h!FUiNxiFL^vv3`EdJaC;M#+;81<#cn`-8OopW5IUtJ@?h#AYs|q)(f}^
zVTJ63A=ms)C)s~YL0VIG<Lpl42`kw<Wc(*Lh{FC-hS_60GRK?96#U%}(wE|QeyR;F
zf6S*?5w*3o;iDN`<~RXocA&{MFD|A6ah$Q6JP$nRXoIUG08XDl4aIYr5_;aB6M<n;
z)(%v5jQ#|!OaN;HKzo5xDu4f{4hM4|fG7p1UQura5HjFB-w24#aDfU4*DK(nkqNp)
z0c>ytIN|c@YFpZSAGmDc1X>F^I(lj0!ns3R<~Q(Kr<DNOr|9UEpr`;RF9nkU+{@k<
zvI6h>6J{{o0-jSF0m<$K%_sPJzqSXK8Vy52$psVM_TL`Vg#jQnT%^el?%-P>n^1p@
z(FYGH@Nl(SVKyB>r0o_C<R6=*X2gM1;XwcZBFDj`W#-l#3F5z81)X3sQdVfWw39p&
z$OG`Rf64I;2<wP&rby`mm#I_ze?O(MNa*S_=LiT1Td%oI)pIIej$)2GQ-~>8n&Smf
z{p!JW-^_VI;H@&Vx?+#Zoa%nKJQ^#4KQjGHcMoD|v#GHZ3y&3v$~Kj9b9Kc|{IBu4
zTq<_<5&5AxzB2Fc<G+eR!K|jsKHC`rlRWqC`ZOObIuBf7Rd;|-0+fomEf+MnEap_0
zv>P?R1Akucj+9g^(okj6s(&Znczb!M474|8fNK<;PuAqKrK4fM(Nh3K6eeAx)sP<r
z16&vY82LQuXnK0t6oH<>L0J?Olw8p7b3u=vUfG{26$4`!>g?>C8$task4>-RclaC0
zU)3(kblE`Z$ptL>S5FjOrp@L+A^dmv_ZACw8mbQ9>YX38;WB{oQml8f=F;za#$`B=
z1kHEkF`vl-EIkvbA<qjt?g3#|u(B%YiDgNJf90u`|GL>LAB2sqgQ;Hm4n5y-voC%%
zZ4u0B{3SVNX*vR8;@>V8d+~X<)<;Y4#|zbYfTgLrzP_G)cYm)4z|lAWWM3F%%UyU3
zzy62hhPhvBgf<pP6urvDsQ)739DE_?-ngF`swI%%-ZNe5R0{in$q>y`3>*DL7yZ5M
zI~HwD;R3-c9@A&x+Fte2xg9bm=UY71r<_?eL?B0eUY-025kxWCMl%e7Wi!+0?i#TS
z?(Y&%?mvGV;Rf2Mv^Nrp1?E46O}g9C7sBEh?CkHXXrpB%1UQTV;tmEhhjH*yrX@<Y
z2RNuHB_*soTzQ>+i-zpSlK~omHPG1GT2>k-tV_cF8^eWo|EEsZ1}bGw$ej=f8NKKg
z%a|Vv@`NE}Vsr#iu=)b626b#f+`)KY&cWS1a!7<a6bY*TMDFF)koedG#6}F70o5<`
z=lonZqbfq&u>|*u+g`PWOp&_TUXarF{cfIZs{h>PN>X-uoy@pj%PC3h#-J>EzH_u+
z#p)qoHWT#+Ra=HCpU4IumAU4p_LDn}F3n@|quTRFt~5W`4pGj#Ag@F3CU21?Y@I^?
z3VG=t*-FO$`3EowQ&%dYm)}P`6+JEf&cCiUV_md+^p&{e?6DX}sJ1~t#U8P}Fq!yH
z;W=iZilu*jbnscH3}ymMMS-Ij7`^x#1*kMK;>0kT(~~U_#S@CR#Yl4{mSxCJpu*d9
zUUK#VO={A9L9LG*lp4KKL0m*$_j4j5NpKIE+WV5f1M{{fvS+g^{0+Fv)V-$-4*Mk6
zkC>$2zCHif?H)E?Q4}D3!+kOxF!+wZJ}?g@Z-gU_HsmAcz(|92BtVCv^_N<T5J<h#
zN%xntZCfPS!Gdzzr0%?%Q2_s{d>zYca}J-^Rt>%s*0@BV5v8+7ZEzu4a>YYOrb&0b
zh~V4I^DjV3`A9QB+f19bc}C5H>bcbHZ8w-*8@ZupKm?4oO&R)B&?(`x!_c|$n~)OR
zfq(f$!lh+XR?_q9)=c`IKvm>z;6i8BtETl~`mi#O`x`QOp~>RiS6o7Oa=6PyDu}oO
z&-}^vdWCG%?z{i><+M`azT9ZshJ#KT>G6{W$T(a}L-ipw6KaU3Y{JrQ>&A{^Qi%Bx
z#<bQ=A{){BW;2-GYynGF5JfVyx!L<@fPcx?v8XPhK&z`1$>=p?{lyaqjL^vJM>YoD
zF+u`Dz^hE&4u+=!Pwrx7{$&k1)w5GC8vXp5oZ>oCxi*S1J&$$$aCy&~#LkMw`8$c-
z*w|_vi;chdhy1YL)!9+dJEE3Ux;*Pgd_ZF)3SZtzsRF4pS8mCki33teZ#L;%wiw1I
zf}D2@&q7m1FUjO6oJJ1jT0{tRUR4p<U{$VM_KoHLn|oNs**N;q7K}mK9_!kaCAT2@
z$%xBad8bLVQQ*2voeRz3;Jc8!&Z5xPBrc7tFTZ}}$I6IoiB#B5t$zj7vW+m1+~yvo
z0ehCK?l;m}YbmV_tqXA#+ENhr-37wuXjs!sJO2~AG92yK&&}To<wV^GauhL$NfDpo
zV)n`~!TFLEkqylWmfmET!9h_C(5Jo{MW)veUxk>>Q%n+6S!^%QTCb+5uw5;qH8@&?
zQoW0tiDhX!NGiWIjX7+{KO?V8{^D<oWPUZ>tF+YdtORC4{qqDh|6g5YZXHmhqk~_L
z@z=MXK&`Tjy9Vff_LCRzJO=6~-tFb?Lc=TmG(*{|CFCR9$!T?mH2j)kt$qt8vM!td
zYD;1lHs>9Ra1novDYC|6m(ShgTo;J=;P|<Lh!53BO3CMOt0r&Nti<A@owbibzS^Xi
zs*AlNVrd-|7E9!wF2wtJj>(I~`x79mHxM#j8!CAfGc!hrsFak{1j7XdO?fXTXl#I1
z17Y{o8m&x{2sBRE)8p5=IV-UZGg~R!r)l;dCmcO5%7nh7F$i>}-aqOuWS-h&wK%Rl
z<+Q}HtWWo|zTB@v#%erl@+09|8lldRj`^-sSf>7l6drGR@Bz6K;A?F`CA7ixpDHuJ
z#1%Y~<2DBWfUe93beWC4II(P8`ePsneaV5osyv_VD6{CTF8-}vm-dZLx}`wXg*K$Y
zRd-qRYJ*fTjHT1&p6~1s_Z7crcZ6IOGq8tH!jkk%;+Hi+xTdwYHwf6e`c))=fdu(0
zA2Uu@eGlK2u(?{%C5f^SG0Ktqrx(?*%H)I14X40m5;=kU(fc@lpzjP=6vA|F9R!Du
zj&$dlcpVs>oJ;vZ%io*m4^7hZAUPlg^iad1<`f{pFJN-b9)3^%3$Q2$7?D6Ucq^BN
z(BN_o0e&Np#D;*K1b3x@6{lLF<1OX_wEk?z!-XgYWgK8=gNZ5o_;6oiw++j10|=)7
zzakKbm?-2a1b}y~vlj%u=-J_7Dm>h>zAiqVr+B><DQmXe^!UIJY3+#)m_$1V{}~R5
z03}BNtmsp33=^GN#anRHn?V1TwbLxJVy=7gkB`r*`kKzae_4%&UIW2O2Al$1{sAqL
zOu%V~kd}(-J8-p;m$tV9{HG3-VQqE5am!FDR0G}*5ugn&mk0g8n<86@Vp5L*)S%Sq
z`0GK#r4}5I%gH@U<TjT+2H7RCiHxWfzcyjH%g9=R<MS75{mHQef=VUV)BU^?OJ6X+
z2;3R})3!ccu6w>iW^vTs-g0$3R6+?AZ((U_c0KftRdz7lcn)N0c`UNHJcH>4Yipe*
zOOw``?dhUy2lF^?rv9AH_I5l{(nP5kEt*epB34$cn1qZvU@PmdHxiwkoJve5nL$LV
z6Dad=*gGh2iuL~edxiX8Jm%=QxH)#)BcBqvv4J`V7nb-B>w%*43RGl}7<qVjxB%E-
zEV6e$0aRiKo(>TFWr05gx2Yp)#U&(0s;%`hGBW|(O$Q_C?rJp*emp?mgWDi-O&;#R
zj}kR9dJDi>MY8MZ#^)GD)m8vV19Ea`5jsO(WrBmQzrUD;FB7=qHy8UK?d&*N8n3?A
z_MQOKNu$9fdb!PSb$9n82s==?pWy-SzXYBH7@<Q|mf8bJylp^InyxVZ5_o&+anem&
z4_sZl!-Zt@7c@se8sM<olJmIT=K*MbyTo!0yd|&Q)-!O-)RqhN!NhF;_)&Po0liS}
zx7>iuJ}`9)%>-(~(@p)#w#R7buN0R~JB4#3&&i#n2KG16as7z2hGuiP%rQ<k#ja22
zQeqm&gmD^N+lre+FArP<_#|0qxdHPD1?tW>=~@DMg_3iJKR{vw#>sOM5^1m#;65O|
z2HF+i4JBsN?1{sZU<+sdm>Pb}0a#U`KqVZAc4gqm{2DZhfiVp<B7Ke{U@~<8CG9Vq
z^cD>BxZl{xqzH6_e9s2}g%OPAYq?@sw7;Eg4a2t|Y{)1`FMwz51Vr!X{2XNoJmnGq
z7r?`wjadLfhWR5%XS^UI>jO0$$7QPIOYGj**ywh%ol|Lg0c1@;h2Q}txJ&rn2UWh-
z&RAMKwF6h+2%*N7_C5|&92-y!!7T#5U+HHpAb})soA(%|x_9IkH=0CBN=THsURvB0
z_%gkI?FAw_$G{dn0#}trdQpc}{}un>zW&sVZh=r)LIjF5`)&HPfa_zq)5b}k#dw6l
zG`rKWc(uc4epPbT=2BUz&UF0mzO9cBXpj;kJb~-12Vp<G;KpD`w4m8Ah_*E^xI5uk
z4czTlEzSi0cmncMBZ6oJ4UwCDPgcvogK)=#fD33ktAGJ>PXk>LZbYe6TTud1O9i3{
zgbqM^M9UlmJJM1OqDUE_y8w-oUUgmxj2>iM0TJMBk-6{!jj<PK`8^N!x4Ek2G~gh=
z{`JRW1gZ*<gpi(r;{`Y4w{}zGetnxCE{DyU6t~%wh?7&bIpq&!REKjG)w1tEx3&eJ
zR-{q$HIjnQ?(MkoD?U4aAX{>H+&Q5+1H<qs7MA|~%{h*QHpu#zqf&^3kZ>Wu7!5&$
z(wQvq-CtmBsnmMX{b;~E^+>#Qu36}Ha^2Xw?_$;(a~O@)7M1&{RbYMCFkPsKp;7;v
zX)OC=x++JZosN!!={h%<JySr)-~+{H_mda*@)wG5&cT6z^9dDQ70({%<yYVazXSsj
zd}sz-NaeNl1xoE4WlDa92&vk@aT2fRg8(o{OR)#}Y(W48s0GOISPC%n&-Q0ydt+JP
z*0)43v2KtnaJ>Oh^$82Z1ss|g+=bVGm-QuZ1;H7)s^FZ#dQUX`rskx$1oA96`G926
z((>|q(0yR_O3FaM!W+tWNk)JAwB!e^T(y<1-y(X4nc4|AG1m6#SoD`*9Mb+f!PDb`
z{eJ&%Z=4#Icds_i`z^*{`s4Ilvmotsmf%;l=&O%PZG&n$(fmcWrRfobrCOI-c<Vp1
zv+*|gWApw5NENA;mbt~tXBHJEaSZZNAdf-;Oa_Zcpe-`RbHHSQ(EnDcfEfJT@ugO_
zG{Nj6uvLF)H+jItKnl;h@I)TVk@C&`{iG_h=~i&S)792*J|}RNcF(gV0ja$gRNRYh
z7L95V@J?n;KwAF<iZNc;Q%JdRVFhqgYW9IS>K1=YTwGavLUMAMzo8TP|H}f%$;l;*
zfG}BnC-Bpp4OjX5W`r=kly{qYe*O}B8^c0-%P|a<Xf-X~Ca4+D-!&iGLp8UdVPj*p
z`+f0D^ACN22D<%@ewC6jWOHDLb4>-QuxJk~e?}kPF2LemCdk^^*=3Md1FOZrh<tzM
z&njrHa6K<AEe+~-|I7^ts@;h^ETGl!!V?>&6Meev<_Ze<UOodrI{tcnvQD>yM@-xW
zQZr&u==aYKpb0_6D&^(n12(pP`1;bQSw#*Z6TNzsY}N>zqI9R@6`a#+1i<(Ez+Qo<
zl5=1!DVA@+vUDGTM8T`~*W2sYOP9D?n&mux+Pr;Z4qpzR`dH5M{Z@Uk(LXBaI$O1f
zh`N~SDSCf1v*dbh(114g%H+s<dz&N68(7isNdY$s1YJ0RhPKpjxnQfZ9*<XHG9HM6
zNhp^ofem<77jP9Kmwngaf(?*U;Z_sSj55Ieg$pD;Xt-t8s}g8Fg&;#BaMJb0?P}RC
ziQh4l*LLGw*7$GvH(+dc^|Cbe0kx&Z<DL&h+1eHt7j+Mk8EPH&d;Ia<Y=Ia-al-`(
zfGeNkSP=nU@N<IiEy(KW4W+$@3pXH@!K&903}rAptdr$e2azy#AUM*3oDZ|>K0E^%
z$EN4B{DWv~Jf8@(8n}ywMamrv)_B~<F_@JnfDIzCFYFltaXlY@I(oXt1`QK!D;1LJ
z{nIJe-Xz^ayNQsIq0{v&jt!Y!@t@C$6@HOqpWOsLr4|)_5f&CCpHCNaFpahVscOn`
ze0l}{!=^`apeMrBPGE&Otw$?IXi_H<(tJ+%`zB1F6-?)?g$8~QOIU|L`H%p}IZoF*
zbAa}xv+_3pZX<#2AG6*YTUVa~f+VM4_O1em25N+?HIR_02Oe?*a++qXoo?4pO3*AO
zD@A#E;fDdd(RZ#FjA4Qo#D@)+<X?V$f~SRWj^UUEd@THg;2#aNp0A)|R1&nK3)~&G
z!F6Z1vtdb2v#Fn8z);6u0;O0M3|L@?(1r=!yoE=E%x5ZeK{I`6(FIZmmo)^ALDRE_
zU#um>7c@?3aHCwjUJV{D<uhx~v!O1=FwG=R1;0Gn$)Dj@4duBSNE+5k*LMAUcz>~(
z`iSm>JA^~lzw88@GiB;dFasj79sRnxq$J;;Zrrd=NhwI6q8((Nbo(39Fcqj*F{__6
zd|?II(`k8$QtMKa;S}+_yR(zPX9}w|A>+O;NXs;E9i{TzzJ85zxi!xg5Qw7xB9W`@
z(XG}n(icH3eSJ;0zM45koBEc3BDH9(EbjqfHk$Mz<mEy?oqlFMAzYBq7Y1pEhd&YU
zD8l8PmI*;|dmgRwaN}3k*t*y23c%J1(s;c4VwuBjfPv2#H?Nj0qEA(S66nwZYihh`
zFub<7XVLxd|J<W29`2Ia?aNiBOiO3CMyt#@vuWnr_Yc<@J^44JV@@jF-qHFUo%I{m
z)ot75Vw)JQL4M#tsd40j>gpIBG4FF<TtBKgr}^#wh5&rczRpgYW2y58gHXk%?SHc4
zN3C&yi%;Kqn6cdFja=(!Im_*|f)1!&$tia<M~ywEMA$@($#j`x7e?gS|GTib`^N*r
zzqNR7wo9aU?sxGgOZYdFyRr*&&gJTV6w;%}8(;C^pKSV3e@ru-+?(jZ#(DLRRzt;N
zoMGyCoyqiJ#5qO#e{FqzJX+?|*Hnz(hlEADH{-*GTQXlx>AXg1u(`>q`U^1kaOSo1
zy}i>K-qn@f)2Cg4{fN*NZkZN)SMj*ra|t{+fh%M1zvufz1L91pJi(i*uP9b2$+1jW
z{i_n|Lz==o!R34Xa8J~DgCX>&fxG@7?0Wf1+5TME>ds!ox0;|mIzd~~ICNxW<ZpP4
zv59I7Q(<2cFU&R#va%5K=5s%p@dkH(Doyq97(DhOBGMt)B#cA+x>fODoqb`j(+BOS
zCiT`i^|mI>a=?j#me$bY23_bLf8<UrGGDpBCqgkl&BOg_oA#~7)NrA+V#FCb<@aX}
zHVk`&nU%p`TPxA$!fFSJ20h<<9zNjrtBwtSE}v;|EisG#aQ1Nc_?nM|Ug7^?>#M`6
zirzKpkQ5a}N{|$!y9GoVX*hs%cXtU20)j|4DBX4FMnJj`-QC?FvyS)v?#weY|MKWQ
zYwx|{{oeSte{TAhBCW&G0=1dN-FUS^xpueB2!{H({3pEG^Y7R%sHF8?C^b}HGsS8L
z;Zxl=5IIDCouy@W?31z$Ud&l{dUIU3`aZugn{=MD5Iol;F4__8`f==a;q#|wTZjPC
z{t&M8)DWo8mu_$y;`H26%~QpM6$hTIlg+;lj>=NRpbRRD3faQQIZPQ`!U?Q%Jbd)6
zkF%>KLL~N)x})>hk8CdMSyO_h&sYN=Y!Bv4n>yZ805W~M$jQ1JckIZhTW)3Z5RdKl
zhnf0+-9E!po3Ew|%^uj;q<^+&bMw_23B$CRawN-NMA$+qBFMSEf|JO8C8#HC!Lx5o
z7I42Ef(-mGzTlzj=~Y%3??dvmC&9yfGO9Ajyk`xxuPXDAPk_~3-Pw@=Hc<fkgh8Bw
zmY;uynGG2k`S|?2)3AJM{icUl#8*%@j>%)IwzSkPDc@hO!PUmHJK_P0a`b62mn`0!
z?3MA*DhMpJQS0<%^SvLz)N59GI`4Rv<kvJT%@`(q0I=k!&i5zUuRmgfpbR9@q#mK7
zwO<`Ks1)|V_*@<}HpvI9Zwwe5ew4mikvh*(WH^se-r>m}C(9n!%^oMnR%K&M)6ijn
z>M*!SGCa$PsV|9nq$<NDx2EHkJx-rJo}J)S^E%sddh?kHHgI6%0*PX-(IfitE6Fm~
zeTeB`iZl>N%om%5zz$p?I|rm=Kqka&{AlL3nJIJGRY$Lq!if4<Bvcn?DWblbs<%iv
zLoo33lCXTWkF4?I&N=7ldn28;HdZfgPIH8Jp`tamIV#p?=Dco$&LKV*#A`C>yX@pY
zq0?f;$~2V^vvYMveJtPUB7bSe<z(#~l_>WjYogDgLl9pap(Dj8KZC+9f4F<PF7YoR
zr=U9`=^+)yDi1(m4iB4Yh@42CNOS=at$H%~@lcq>-Vj0eY?%QvkPdQzZ&xmbmu<la
zxInUjyfxPlKdv3kPxa~*FsaBE+!+B|yizEkyh!105Pm-0mN))A?A&NR-{onaT6m(o
z>4|i<;-8h63sbpw60LfEiJ7y{7+*5=pKYCQ8-1}$_koSPsR~Hu$<!4X6VfV-ll~h@
z<mOyBN?Ytw;Dey+Z=Ru=Vx)+yBzj3#`;l;ow2XbQTqeQzq!MM53{hRvY5uiZ!B!UF
z9$0(2iybZx=G_;)ypr+W0}DmwAkY$Uv9Kz4?m#uO4Gz0);I^;;hqW${@SDKj+&2Nf
zks$(pfWsN6>|~LairGMtT0HP^Dkie8xA6h2G0t|V)`(XO6?WN)K|9Nmj&=rCUl*Wx
zRw>IYHhQS&HF>Jl*vzVf^BMyG&qQw^ssPU_r;GUg&8$9HXwv4fg2`<TBx^J@H0%PX
zz6G3=mc>wD6SxvD*M@?_$Hrm4&gs3|-jov2YgMU+`uZLCKLcXAWmxCQQVPfu;A{bW
zunTC09U!w?C@L!6kG!YY1`t@w#qY_<#RY5u)0Y>xDvAJsP~-&62lMmuE<is!12Fq|
z*a-;NrGU;UoS&OJ14?Etz_4js=X(Q$EgXteFJHD&I*eh}iS&edMPBGm=O+IhRb34m
z6k%^k!{>Y({QiqTi9;klS@C(-H%N_DuFFCrBoGFMF`pR-OXjg6+e|Nd>U1?UoWupp
zphYfmOb|Sh92eCXquFRt%Xtd@;A*NC>3tE|B!st+-PLmRDp+oea&kaU>r{7L0*qgV
z!38};4*wN_X2Mb<s;FIdOfu}oSLUOHVZG~mTh16PBw{eM^3-p!TbPGNrRKE;w-YD%
zdq8#s1Oxz7jqP^1R`W33;Ks7B0B6J^(!r*l{S*|E-LtdF85u8MtKHaro&huY8c0#w
zGf+lkG;DoXkyaImu>l07oU1?vs;v&dbHQDlfPZS|JHYwM28cM2;&=x*1xfH*2L}f!
zLSBS!Zf-!`k_7RPI-vLi6q*UvJgKnvRom_qFaQG$JO^&OBH@XEGikum17ZSLt-vVi
z0YVV@z}En{0)&65n@O8;P4n@*$3Sy#9~z2$^z8K!kR0JcE8H>)h=$eFMc@>JJAQ!Z
z$LF|)0kCZQ^mHPi^R&;Ksn-E1F<zt<4bbf=5UJq?e}Fh8K-5VSOeq|aBw{?c{rFd;
z@K0bef;(4%k!uWUr2$O63;-#H&QeO4I-J0y(a>;>7ZyC$?E3=z$o%S655-(3_rTUJ
zh>n{;-gJGs+)rG&EHRdo^ZxLBvW>>aW;40!$bM^bXb<tSpQFr?rt#?0FYVZkY}Zi@
za?P|q$H@L_4d)EmDXdP9h1fb8Ojw0N9E?&%A;Gi+9na{Wok9^x^jZ`&HJ^7D?>)La
zJ?$76_y%amWI(<k!-+4M82UC)U#tv7VgWZ3ur7?}tDyUXoXfG_!=>NdNgoUN1Kk3?
zC79oTeylH3yk}uy0S>4ac2Z#7Re-W#x$P2&nn2SG1d{MED8r8~X6=xNf4%=F`U8@Y
zoSYnC{2NE+CD87>2L^D!f&x@#+S)GQBe}o~FoaJJ+xZv%V86CMTo3}LW^u5y9G~5l
zR`>$<{1wQykf*q<{dw$rcifHs)lmjG?c%_4EC->9*BT|*tsrS74ag6WidYQHvI9yF
zn3>EA_k|xakw7g6Kc^XWu$%wX8m($JIdCy6FrO8Sgni`ljN-BS_?~A!0E?)99O`i@
zw;2!}J#B4xq+f~uO5B#N!7jbW_;Nomg3Mq|D&(5yBlR0-baP{Cu|&ioN>Qo}AHfyv
z#xDL6a`E}1dRK_g+y;3~@Ecy<a8d_fk}+9_4d7`Q2JR?yObSuJ09uS@VFGJHxLZkZ
zaCcx3UMqMNV6r-0>}eIQg7gd12(<-Jbmc(e2{3)XlkMEWEhC7Di^B^TA+HPX??>6c
z!Gf@Nbo>cSg~t(Kx9Nsk`hk1u1g32##Jd0#TlY=1ybl3Di3S_fvqw4hPQ73hADTQ}
z{$92`2=Q+0?ZLU`<I~gCgZT#IzPN9I^8`t$boa|cU;$d%&P`WP@}^|3C&}b_Ugtsm
zt0toB=bNh18S-?9@YBI4?&`gIpIK{S0lTzo5pgT`ZJNrF-?xW>@$tX3&FKikJffPs
zl2Ko@Jm7-#+m?DZV?=^WlIHb6_x8<mJSK!kNa;wYAW?5wY!n|~K=(B>#uBRvb*$lO
zg2V+N0hs|+NAar%FDxP=!tLAER*}Cnno4wGA){8M={g96s5j*R8Ut7(n}Y`dU%^si
zJMU>P$j<%)OlJ}2kAStU1x$V5>g8c_)CEv{-GzLBLk9uah&D5=;13o|xe+j*!Aygv
z0m?O*;Q1Q?w*!2@F#=l>_=CHrC+&586-+`>QrKBx&J__5#-9Q8g#Y;QTM3EZfv1K~
zF0qK1!GDl5&)e2_b#se!ljm4(KPxiri`$RdmzJ}xOPSi$1r<6df`!gFVu0zS3I+;u
zT(F<P&+$Z#P4Uu!v9Z(4%^7XLDF0|txga1QSovPN{dq3^O6-@ImbU>L*3cVjPW~8A
z)SZ!^qF6F!i~}R7@A-O_f}wAy#L8{7g<(a<AXG^a6{YZ$fuJ`x{o^`xvgAWm-gV3B
zFy(0!M0wZj58$@ah=o%G8(bakV>Nh3E9s)h$t{|5sf+I;z!u^b_S<Xb<h%RfoOl&%
zldaGVQJo<hDPTm@derTSR1u0?8gC?>>%2uPr$1ndcd~YVf;yW1Bg=d=us(tpG_`Hb
z&~y-;r`x~{j)&U%`jt4jXM=EA6>v7eGk~m-2l;~ArP{DCESdb^!LZaCA}>MikHE9K
zLgO=CnOw&XUaZg73os*pfmeg~MG+AM*Upzb!4|xfpPSa+)AJP^veX`mvC6fzwSw>$
z5^ycP1ndU2b#==-J2y*>Q@`^8&ksUmt8Ytx1_!@ywCWn<07<J`qn2*W!$iQ~gDhY7
z^fccB)ONA?n{p`Frk0nN1HhQTT=Ah_INzOQP%qLf_*+!h&~SQpI+P1()CBjVyvmB{
z=ofyJXP<MB&^!E1ypGWGo9Pgzk$q~LpvDqk@WQJwRBmflA<ve3Q_rhMPyH|U2evlC
zSUfXSF*Ab1XVCM{II7Gad!Ij>Yrz}KS)R_pNwX-URxli$In=bA+!pXpbc+ALx3u)+
zqZ7Jm+Q+%)irq26($8NZ@kWH?g2BzzIs^hw7wyeezDh1Rlkd^R#RecOA}3{^g82#@
z$;THLz92%?qo$c~E|da@uUhJaC_!)(0a)hl{-M%Teq%5-c?}AnNKxoM$XRz98URh?
zIASQ72Twy|(s&bPMtScih?oot@)DAf4S-SsZkg(ddWAkXIH+0-i3O`dWXl9)3NY}2
zC|V)?%)JDi;kq33r@F-T_Y>bjKR&~}p$f@9WLd968sf{6hjgR~4Hw4e^DE~@Bvww=
zvfW+hCU)=`d?Fr{r}G$!Ztx<4)O;a)li_T^VR78Y!qJ9RlRmC+qmD2Ut4w2rM}KN2
z^DUv9T(<vbL{jTl6gHiZN>&MmSX6h4D-o}vIr&4lD;hYsu7OVW&rG{9U<9bGBF*wo
zz&dyYH;+j@g$qF7&;x;GcuIr`P}-9py}k_(e*lS^qp9+~SVr_#z+6pL+mHyj?nzzn
z*)<-c!JSq>RWbn-7@Q>rM4X`W0-QW^SVaR813O5iQ39O+IFp9K({Miq%-tWbF2GpO
z!3L_pX$q)tIOPuhztP>*F)%uR0okQvcpPVdXYsf8Z-ZhmaYaUcMu0!Vy=ZVT0^kiT
zfFa;%F(5RW0MZF>kc875A8u{|Qh_PBg*z}O8$r0|DU&9CE((dcY0r<(Kzmv#pQV!U
zG#()&JkXQ6?pe2}tibu~_|$Np&PwZc8&dGBc;h>FDj#ykm$#oE<q&k#I^+x<-Cdgo
zhx0l0Kbn((HJAWJEs=!P3>W*kelv^i45%C*it=|23-Ml(<mB2=<VeUq(Z|d2cf8=;
zeAYdFK!U%Sz{2`Cy3wQC`>#mrC$0X+JcCI6_6{T&<VPg<xQ+Vq0vy_@&6;47hQS?a
z%$j8xgFk4(2BYaB0S;+S<1hm|dxWq3M_}TaaJeCn(R{`H`n8PP*6>>ZoAns8$qUAS
zY;+b}@)#JEXQ5WPLPA0=7kkq#0BfBACYCD$JecQ)bAa~lEHu*qVy=+))e(9mP${<F
zlD#b|8iNNC_g7Ya0?F?F55Ess3|}wjz&2wcnhwHIZRKu@J{n#g9ztg~KN`#+$^1S*
zL>L8}>A0<}ZDi3Ha9s-O=ExE6@9Y$SX_6byr1cY=+C|AglRRv!EuAONQqY@;nLdc@
zG#qK6xYyN4gJ`XWF^l9kesne0YFu0n+VhCCof`&hL9w3C#P}$3v7X1fUFqnmX4e5*
z1DoEBOFAmCUEljESI5zMW9i}=>+=0`6)i~NSbTwWx&=|Ee`l-i%jcSCF;eylnEi&>
zi@e=Q`mU$7q0=L)*f|=dl2mLJhL(SIBcv~?5uwF4MyB#Nqe(%R^uHK%YRkxh<n=`4
zLHb|_Owy|4<m8LfQ?Hd;MX?{vRSX>S={UJOTS**Nmfj;YbTQ*jm;Dn;BRxbX&~u?j
zI1`@BV#UdvBQu?ycPw}P9FW^vE1#1rdGYK=$qVTn!O#-O#9Yjf@EAYx9mRLpk7m_F
zdKe@fHHpq?(1M_-dID0~+DoZl?DBkAXycWxlz5nz#A{yHz2{)Vh35!u+1lIN;Vdbf
zbp|<FxHbVKTVPav3xMDrp+-oR@np8{nf7F)NaTZb=HBOoQS(RXA`g5egC#v8(~>`b
zQNC<t`lH!6wHKhBhwpq~O=v9m;7yyJlfxyKj5gm8*wCN_9Tn{1d@GpQ&n_Su!kG<}
zW3Z8Sl(40wrh-%>J1~#00MQYSec-VKVAfO0`#{;yMc6)qt0yA&;Q#l|^@!)hbfE9g
zq~z#vc7%IvmcnDbVuMNM=&d)J2=UB`Vm*h=Rs>BmP|7Iz5cc#!jpZ;F{fXBqYB#y*
z8TK=I!3mG0N9iQc)^~Q`!8NjZY)a*c#96cFRlfL@m;GHGoq>o@wBh^pgEhw?K$R<|
zafY#M41Rv~z8_|0*K9%1)b(uuW=_>rphmmg;MC0|eK6i*%|Nv$`9m0xWZMg|-x;yz
zd1LYga({ks&WC*#=DPhZqP=o!F*^#iI@&GG(XiW9cPdhU!vVQIX!y>vzc&-d4b(}V
z#n$os59L%v+UV;~3O8u2<$pBOWkD5=$qU4c(fAcoMJcWKU9W5;ZOA*XY*y<iT+)Zq
zA5#88Z%$Q9a+STo*0y}H-@CVPe`Iud&~~y;J94|r!eLh`hWcEX6jB{z_|@j|OZ{$K
zz)bRbKJ`39@bi;0RY5Qi_;DtLNJK;g-oJH4ZXrIguBPUccxTmTseG@;mshs8ZsAjc
zpJtb9u9xcYo~wG{o1Da!Yxn2TYc{~J=b5!l30i)>lh2)aPm!p=c#JVJ!fJ?uO^Shv
zl|-UdQvyn&ZK=^qlt)kLTR{r}dJv%oD?XN7gpwc=N()Rf0I@B_8o2or<-eig|70PM
zyWG;2&Ujfeoag-NZudr|PT!>A_t=YS8xA2)ETSYA^g5@GV2UJ(!wC(O7mx~t)K7}S
z0Rzd)_m>8}PeB^%dJ{7_I_nT~tv~UJKj>#bKumlsF^PeP=T}BiD3vnzD=Rzrc6DXt
zQ+)hn;36?)nranIQk~c4O2b0@SdF4eq5B)c8cn%xhFp6Dhg~?lJ};Pcd*7b#n4(lT
z7;&&qmt}@aLEc$a{*ug9$e|sWtE1J9iy#1$<00lch%3H(h(bj}<2OdWIg>b08eRJv
z(l2T?Fxnw|thL#Zxf`6PocBnC^KZZn&F-v83=K6(S7fl#uYOqRbftCzduQUt0Hwwb
z6T;wg+l3(4ef<%tF*1I$h-o=Z-WX~hF$~I4(qb+vg_6v}MW1O83%Vh9;SRgQ`!0c@
zW}|px7AmA1#o)HUfsZu<Np07s_{1KOLn1>qr|!fvi+<vT>Kl@{BxpijGD_{7Vy0Ap
zBb8FAPMH-;opy2eIoQ>~YZ@%dshpbhr5Nbrs_|$yD^IBR91^yYr0k3P{6<@UTlr7i
zKu#um^O1dJ0RXjfG&?r48E$scSEKmb+fT$WqG<Q(8qw1cDAdVxKWlKaG+SMf=`|nG
z*lJme1SnWtohpJn*u`FeOsxL7rTZ#QqT_|c#Xbk4_QikKa^D{jSw#H1QPQ2MZ$@yk
zmF3otvc}8y3yiFHa-x|wxzMM6=F=xR^zobX4;5KjiMMU&c%|()*g>tW^P)BlRj<U$
zH1+3roy1Sq0v7#vw7-XNO8;A%xDJ00Jnmq(Y^RHfwI&X~>;{_8+_TX=DrRz1y4wPW
zOVjhqH8K#qAB4QGX=>|dXmq+RPx{8ES{93ZLHtH8l<7n3NHKx*cR0H&QBd!C_@aXc
z$81z2y38Dl?Bky3qQ<Q<tJm9x-vVBQ-aZBw)*Om<NBUbc&nnDnWNX7&sM_~pZ-+#T
zM_%N}2XfmO>j;wl_qA%h9+-`)!wsy5HN{?uUssLhVxN&qeVwEEHwT~dKc$stce<)~
z<#M_b6xt_XcDh={b9xm?{@ME)V=z5EMb8&hbeN1_XA8n_FKMUhNH`m!{$lf@YoIK-
zQTssm&hqxy`mfnO0gG*M(|3Foq2H4l7aj6ZtKFH}Jetr)y)kTsA75}AYv^b6yUV?&
zhQCV#qNtIm{yR5XUy{4p=iX71510JnJGhxL%dcI(#%mS}`T5c7Eif;$2QTSAXaCIv
zvv=Z1Je$<(llbE-^Y4LkKLeC`+h0^{4ARyfZ90$je2s*MLGMOc_gR{>P9!Ws(TBhG
zC)DEwCOdmgcNMzrCb*DflYu7BwrttqdMHDc&am()^ShmY4=G6AyjaV_nFda(K1|%J
zo)r8ph%0ojJWF<+ekX3ZKixB!n!{t)M+mGR(Qh>hH^2iS`A)_cgfEsgRj;z^?$Oo9
zsXH8Vt8}R~?2XvaJoD}s`?tLxN2sW8Q~7hKb?nQwV`iXOFSuiZ4!`O)pxMlVP3FoW
zfp2WiT>6R~``-jV{#~s%rO@JXuj%FZ=A9Fv$&L1yg<AeI=0RrHuXNChy_f6+z*HH4
zyu2V(sc+PZ@F%u!<joUqQ<u@)m;#I|1gL+bs7vuVb9RhK*;}i7KUd3N_+lZPrr?&&
z@V163I{J>zj*RDaSIsq`WCvPv`vi*cdt3BN&YTNG&wPtqxp%E~TU0<8VWtKy>4|zp
z!i&X(*c^IV_GB3M-;-jmjjmBxA=&DiYjWxt)WH`2j~8G$4!LE<*SUBhmeiJS4?Y)!
zzC*mYoo+b2PjQ&EkQ7C1>J^eXqsou52WQfnA17p$Vm{D#l)EBOOxNndH=pmaZSl_=
z4OqwA<mtjx2t6A|1MRH5WrZWX|Lg9lIiE^~`$xSRFzpckB6@kSWSdS7iK^C2K2t+X
zvMPSFvrJq!iO4d`Eelo<#QxP8+AE&OwTAu0#};4L`iPo;=f>fEc6wSZ+xXjZABPxe
zy?2^MXCssA?jm`69@OcOYR<2EaFm`X^$u3D(1d5~^V<kDKmdlODkGE~b0g2-@}9v(
zg8>T4lD$P+nfwBod<&WU4w?K9+2dN-<AT}aV_#FeAPm`0uFXwt1U}<aabPu5gg>D9
z$tW=O<h}W2{;`BFSb*Hr@mG(l^c53fg2<vA-R?MMng?37E!Tt?^|EZH&uxaSuK5Vk
z$;M|KlKz3XAeOQ6>Rh`gM^I%;trR2vjnhq0e=Z^=84rYAR%zqH1gk4#M49|h$gY>P
z;5~ANAnRjZgr|r%$O19Su~671LCOQsQu5N&W2+g4ApAr%cU4{$H#w87u1p=QCdo>(
zB(~Z?{WQj-6x|gXb*sZjk)`uV6Hl$s`9=%%WrakfI;Z3zJzvbhLGWDc@vWkp5fgqN
z>ghkG)w<PW;3CMwZ8TnYX-1JjGJ(Rn5o46S0*n$mDUG7t4={?)p|$967FvaElI;7^
z-7aMJCF3j&7zHGTCUm;NjKlo?<>0)~`=;DowyXpEf@e^@Z`Bzux8_M;rNKXF<TuA|
z50mpvsZf9Z8hBn8i8oh|x<4D3p1!^QBr!>?1p?Xi3z%uOkvmL8sz*8y1?)K`<0=>8
zC&MSr30CPy+{p6L79CKDh*ZQ91J8S~SOsars>|#j#CjE(6_aCa=wp82TZw4(%_vUJ
zyjd1-$6xTBwT^bZD$`2z+WMYoN{gxY$Yu^htuRRfAf2mAJ(d0s1j21bdw(*af8Sf}
z<E){`Ak;x&rIy8Rf|89Vt$!sVC=kfKU|*DXv@#IBYmQ#4_s_N5%H+RH90I%N!wBSw
z^Q(~8V+BWTD|q$}jt0}WeHfsQ1|0^bX*Q-SkFGaqiL`14nP9xWr2Nj`S9%#-{|b6v
zWlp<ab>Uo{YdA#%Z<KSN#EYL*u%7g+aZdd_*o`w)<TGUNDws`itpW<Gy4Acun}c6Q
z4eZ7t^qpB-Oq2&9hmJSnpwS}YgZ>xaCVvHgo};fY&x2PblZ572OY7@uCLqrW<6i$7
z6cqGbw;+^O`X9QWRy<Hzaa~-HZ@It6_=cNZ=4AkhH<zDBa&b_<NZfVVjKFf*Bs98q
z<e(obnno4A*<z}5d^bar_9pv0oiaBBBNQ~i=-eC-fA`+0+C+A++WIR^tV~(f`R%HC
zzKb7-rc|m}>2gg!fx>!x5eS#E-wUdvhE$*MKW;6Ph`>YqSyU&PvMGm36@aFmvKuaq
zG^FLU50of#FDziY@(Mpw_@7~V&kos5;%smZU;V_2WGHGF+$}hKKVPHWxTIf<jtAp6
zQz%YNsrtLeQ9P7VuFcyZRVgA)7zXQ$x6-Z1f@Pcw9Gd%_7#Hxo8dyl@q=Fvcm5FOd
zElwF;1QFqzuGiJDYS+@l&~304(r14$@LU(3RJVS}fkt#Oi82(#yf(HWr40ImaiKRq
zG>f&kkxijzxY?N+ULumTni@G)Wv`->Y9mte;<3TTcI<1^``cto%G-;!{zPS+X<Fb1
z;92m`k(XCmhgumN8&+0+!IySlvoBCCU^phyIMa}!(UFUK#$jFG*XxeXdU`|f8lALi
z14$;|SJZCFs`#9garKjyH_}6F(qpj@?ER~O)ho0ccE80t)oJzt!vIt)xu0$))b3u%
z^GyXB04nrv4w6+qa$Aw8L_7|{G;RA57WPY<cJyo3#&9gkPPcb4n~4?`L5dSaCC+S(
zdY3k0xVlX2I3ykA1MqTM>J!vaGh{b+ji2x4LjMl-CX}B(EFceB7PDyI3IfO%IXKVP
z4K(q@>q9x6RNZa0UqnroXH<6Tw?#`a5HDBL^47iSoaO<j)o{+Y=^Sc5-}g0gSsama
zH&rmue+cx>a}qTwK2Lkvl^mg^TCT&e9w<$Vj=bqlVW_Aq^9)@9!53oHl3+xEuo@r|
z;3AV>>nE@<0ig)zE2YT^c68}?Y4VRqs0)6hGUW?~Q(3am&|CJcBca0R5ur_{&&0)<
z50vGz<CL9xAFP;`Xu9kzH~~AWZHDVO?oK1~9V$J!RP$H1)W16S7EDw?c5|&C%USs-
zGN=(n%MG?+*ZswDo`3}Z=2J5u)cbvJw)!n$)$X_VZz#FSd;I>;d8AhQ{=2ph{ZtV=
zRWQHl6&LtX<S1~*{i`P2)O*;y{b#$mZax=sBl-tHH;0B!L3jL{i_hzQZGC<H%T{BO
zGNn`f>wP)9XP}_`K$##g8Dkt(q7sFq9xsLXL1rCA#vqiv`xkV<aDb}{W)`r~lj4;0
zu(8xhZiQ1-y>jzbRM%>T3RH?(QW6+XYUGW_<CNxbTLx~|7tKEGzl}L(vZ_fi8)^rU
zlhdK^LJj%QNoe6UWJ**bD6w*L6McC(hXiV;_s1aJBZ11P{PnhqOr60OQ|*(jPlFYS
z=>Y9~0=%g2)G@L+56=2dQ(>dp#v8OrV!7;3v3rE4H!jcl$IrGHObD8UeOkxHY$-WF
z5g){?yR@ZbKKHyU;`3-Wuuwz=XwHsp1_EmbS^}!~H?7|zFaF>;*Suc^JM<kLJO7Vq
zul~Qan}=z+2V*Y8%5E4FVKa6f)>0VZ6nMWR6HB$BNP2n<O@=Dnkw7sjbIg`L@qHj$
z@E!vIn)o|Z#by!t&VOp2soK0a*Q(e`zgn#l$jJa=04RNWIgmTI=IC8_A{)9!-#()V
zJGspDn|-q!4e`o<N&OppK8^(s5s}rAdT)^hGrf$&%IhW+5y+_{GWk+3wNtu%QB0BM
zt><~cg72?_74cRefr<pIe8qRAp+#5vsK5v!Ic*`<GCm`V3IOLI0Gyp+N87oq-nCj~
zltE-#$!GHjhPm5m=&zATjGBi68Ah1^u0{jb5~?Ev*Meif?Z)iu>}!@P`bE(HJQ+d{
zwd&ZPjh${1;5NO3c*Qcx77H~@L9v?uzTmdw&erg~9f&b}O!W*vAsgh^29}F&5YZU3
z1^wwWh{~pg2cC$9^qch6y@B&9tz!Xj%K)*w>5)aYV$aH%W+x$e7<{mkB>|Cma_P3}
zEq7P1ye|^xjH3#Q?lp|5@wJb7DCm)~JNhx_5S`0>#(<~h^y+f7<qGAh>}JZcvBHDx
z_}ka2;!(mHCR^7qV)6rtN>gR4p%P&6>wOAx(*4W3GZWsa0ySWxVW+eT<FG@o?)LUU
zH!q0keN<~zY&TsF%Fn1wvS!}|L_}mwXZyz<bEm9D3#|Mt<B8TK(?9%9p78sU;!=EJ
z)Nr)lCb+sy)}c<m7c*7Zg(VU&-ClNBl_oYh_Qg6m7fR5^Sp)Fuh{|nn(V+XpOUH@~
zEoi#j2K<0QYQ0P7+g)@vvqMrr9l<TVLXmmB0|i^R+j9gm5!-6(RF3Zc<-YU>;b7dl
zXqNPO@B_Ju;*sQK?rbsih_dlwV>#@6%Dg%i(&;dMF$OiX5QWbdV2OWmWw;{vOLCXk
zY{jq8`?X}Eezd{V?!>Imt|oy_Apyi^omx9Pi(X+o`ZPRjf3ogn9!PVXB_=LE9sBr|
z+$SD)RsfQXW1$gLHf62~i;yKQ{Zz79u#kyi--b!x6v<(YpJEox5ALCwHUMMaIT8Yv
zB$`gX0{9APr+4Rwo9_u95pou~iHq+O@;i;pobC9Wew)rc2*i^+xr~voE+%R>Et0?K
zg3E9Ut?Ww#8LMgdmANILKHxoAXz80OHX?3tVWNCY>8*2Y<ZX4U%&sU6_9u@~?A5`v
zd%cRu!uXhg$}peX{>CPMggng==UiC9I;ex9=m98{XG`}&6rym5*lS`8!e;|@W@42O
zK?e`srs<UYg2x3JC=!*LMLob3U3OZGho*`7UD$=ZzS4LF1mHifCCRg?xiB`bv2AM&
zVD(^M>opPfyQVQN*Rvp_RPLQgAFQU{ng(U#aoK;nuD*X0uYKzHl}DxtE~pngEitS~
z$W=@>KFm#*tC-9Tl)iXGRIUA|pOOECX+P_1?mR$_GXd!o6)z;S(81Thu16Mdz;SuT
z1Y`|R$q)mAF4r!7W5Tes;%NSJBjipEjMV4F7u?TUmEeNSP0|;kQ+7%E8`U0_#EcJ@
zdK^Il4pgcIYv2JGaBT+@iiJD{yGs}zv#ID30-R5b6;n%?F@KHoce2&?b~TO9$n(N<
zfSrWD%F4`j{JZgx0<#WlG}Y!%@XvQ3C*^)9#Lz}HN+x|1guj9bicueHE@4FBz2?Sb
z$LqfG+D^j--wm4vTCZ|Iuz;?XiN>YoN)17vN2${O6dMLEw}92zp&j_zS}0U17hYk9
zbU$_H|L(hcQ@%b&D6_%PGc@44(ea34WzRVM+}dm?$=GfKH1+bira9Yw)sv=M{1FFO
zq+v()?zcA2IndI4fhs^i_3mG>&S=w8mI}Yr#To>@X^u;(T;E*}IXJkM%QjY;8BDQO
z+P>tT-ksO(ZJC}(<j6tkJ6T^^ek;}U!&tB4H@Hwc?Y}V@?=@6J{zbgk@-^m5BToyr
zVEuBTHXA+BDJ<|NMZoMw6NORSFw_D(D(dOVU$>X?c4;jwJ!Qdo!Jv!A&Vep~3K+>A
zd%#2%`|J;v&IsiFbh9*A*W9EH4&ea-$XaE7h<x*o?4Y8lSKSrEts!;&>h|}}uO|(5
z)Ja4;&8Xhbyev@!waPSl4`|9Hp5St$0r`guZC7+O85;2@1I^SkhU^zef&N5>xJ@W~
z$};x|DCWgPQ<?DXX$L@?;6+an)BUv1+JrpK#m8@itj?(IV2$u!Y24OK5}v2_O82w<
z(X`UJPoVId-*3S>^<ua5L%~zfgtPJZ%P(!3pn1($@EZ&5)-8mfVi#W(^u{&EkpkR`
z)MQDJj=@xpf!s73sy{A^YsO)r4?{vzhVF`Pt-j|)jv{(kC$*Q)txScPTT~RQ0n*h?
zl3|~(?{FnuZG%9)iP%&uRLG~fbqKOM6r-#al76<2JTqNy0EP~+{XB07uz=hYcL3lh
zEOhJsK@g^RpDr~P9abT)v7AKX!^$DK3XPvZy6s<UVWm9cb?CUFt6e#*6-B+z5`&=9
zdVCpf{gaqAMh&R$Gez&fWo5*u#1im11C+ES9X9`x3A-O4qfL-YBzl_wYH6kO;YZV!
zx%7qd1VA)Ab_Y669U4XcH4Bc;g+Wqk3hHS@*n*ZuZ7Z!@R>-fj)4;<9?mzJDufiC}
z&ls7DA&YEP@YyZS>~Vu^)jZ{$B4x`o<(*8Ckr-tKnS5cH{4tq)`|NS@Y*i<#S^A(q
zQ|-Uni77bKems#`d9oDfhG}<EHQT{5;8!l&?soh$A4$1_S}qYO-P;g`EtgD-nLhvS
zUE)Gxg~`C25Cz3pA01};pP-`s?OT$q;ur|dTo_+Yta6Fy`z01D=ASVq2r$IU0QH2A
z;(jkh8Pt@ld|1Bb`XR)i(;}h!Q;GWN2P+p;z;1L>oY0Yl{C!&o6E>NZWw>CF((0UI
zv5;^$tkG&PV{`A&A6yht1QqYH-emgDyTVzTg=|^CYTq3NCQ8EUq1)4pTBBf=ifPw~
z1>=@^@c>)mq90zitU$e*7;iXu4SM<l_Rp`#qKnQ^UK@BKEJ9EC^^j-+09aiG@VTj#
zZHZCenqY`56B;r3D=%gSJ0-O4z(2tVl@)W*zW97AK?ONRIop5PA5fvj>?3jJgnN3e
zAcdS#>+%T<h@)ZJdPi7ylKGff_fym>gB{iAJ*Xw>{f%@*yk&_lhgy-@Am?n{-33YG
zX~bSEBfC)H7dI#@*T10k#mj2NM&UB7yd0>ook=6D=m+a|qRkm#L+hzp&ZIOhJ}2%U
zcAT^6(nHq?sKZo9$5NHD=GDZAVzVgL@-%0N%~4-j&8CQ+=MrGlE-~)g)CLrqZZr0L
zec~E$17(=V@;d(B%cI29dmR$No#;AU0buNKArUFRvE`w3J6+HOjdleh5foOe{Jd1K
zOfYH{-T#vO(oiPfEC8uB6NT*nx%)|q@5>D*KDomTe5$}vNae=bi~76S&mC>S4AG_7
zSV=5PO(%!TSw*S6)=ho!@<R(T3^`_p(Te@!<q!8m8=1d<?M)%>&n6E6OqM16p(F(?
zYKf$uRSLDwdD<{#>JSWC1t7c67#`)Q{=c<&KbQBvT72V*Gvh0)UjNgntihKkkqO9<
z^qQwlKs5qrVx>mNRAx8b2J*cXW<$1<&$m;7tk82WJ-0CHmFVD!=5@H&`L1rbSdQm~
zJ6Ed~->e{PAeos-^F#PE#z#+xw<YdxG_Ka7Ie<OH^i=cW`(xiP6?08}W1z8y`??I(
z$=b5S?S)RU#*nFy7x6Orbo5+$RyOP&SPcdhid`m*6~`fTw}?6@3soC)#wAY>v1Qk;
z@L%q*m2y1`L3u&VD@;}!^XEF7&FQg-^uY`;T8&%~;^XUI{%4?H?PjlS>f2J#Ug$Zu
zji8>p%{Gtu-QUOS8tOI|o%1Z3r*aj-6civemf7bczg#g*G`4Td==mYL#2i%@XX7)|
zTDv8>&}OIZ4TI5hAsJLhzaMdsOZ2JC;y<ZH`&!9(-jcmUZ|B<gW|KdpC3F}4jCt{8
zJJu>2`nLrIZxQ8$K3lc_TL{Uk3kOOGwDQZ|k2mONW}Q*b*7}knmCoOfEO+9%*Sh0B
ze{MdesZrXVr)0Wz^w$RnR~+GtSYj&E2lF|tC!lfOFEqI{>}QPpVA*APz-WEm2VBM~
zYqL3>Fh^xbRl;t}7)zWGI5t7^*w(jJFMh`C1T#FnhbkB+aMOMvITBeJ+6_ycGf*R?
zl$de7x-sxZ_R(_Jr>vWEgcEN6eLQhTx}3ibC%5_0bYy?E^(uNVwp<a%?3mu|pw%BS
z=L~qaS*-}YZy0!|ebaLAiKC#>i{lAH@BA&l3YH*R{b+vHYM!ksU3@N$kPW;iXb<~W
z$5Jx+rED(uHx|V$UcW8;48in5LhI!F%T}hBJkc%|0@@ZAt6)P!LioHkzjMi|-@x`{
zBQ-p~(^;jhrQao#=-!r^u?Uh1-wjiHne@#F#z=+2SnbHqwhY*hF813NDNd&T3K3Dg
zs^s3ToLK5~&UjMxuSh5S{}<_O()0Y;s&keFV2eOUtb%F8DzEy3&0AkiEAcX*o?3-a
z5+lkrwZ$uqlH|+LYKzE9QOY)?m~jzy7l$<cXi5+UwF2w!_E5fJvLajhBF&>~BU8c=
zwD1wH+nwx?WobKX5s!TxVH=CT{X-4F@|&PHpDc>Ow_A3QR{i<>+lyNF3!u1PoeGdr
zzUsX{`k#6|k8JbSR}^Cyl^%l>5n^7fF#|8wm4_z0G>Drmc%vT@!{2F}-->D9G|$9S
z@r6-8taUrz-}v=>G|!X1xVels9-<Tz(;t5MT!h?lgL5{i%HoxHomiIt1dX!wJ1ZYT
zDMt#O;&ThRPehlRRvFM(wx4Vrt$!6qB`Jee=dl}lw%+_oEVDscR$8tq7!O5UK&Fps
z2pTZbZNW9_O8C^9nF$*8&`x_@Xe9>|^o-2xWsDS)=+x>HG#PY&(i-soRP&Y9yr*mg
zOmk9FR&P|2KI1chj|S?}qR8u?!A4LP@iqcCh4nKoMSDGGe5vNYkvz}<XSsC6rv3W}
z(apg!QyjsEM(Mu=vu1k6l<coSp<8`yKN^phhdYPLJyoFuszy3a_893|IyS>o4BMdj
zjA@R6qr(07sHKm<#9c)TC&PL|mt=f+e<@!<N5ac1WloPxY^Xlq?w-KZ=;LKy?eo@#
zlAXh9^UNn@${}0zV}E<-qRlitrNoAV$d2%stAf#=E=rv`x6c5_w=MAp=0kX|S56{#
zH{{1EiIVrA<_15SPQE_-L2!}FRiufiTMxS8=G^uEi2|JERvcy&WwyBjc;h+Bmefn=
zOjYvHfLeZQ>`V>U+U=SsHV@J|8ftl3pjY3LIxT0`ZvIn7&r-v2G$rNHf){wg2Yksv
z1d1zb(i|}jo=}fh+}8dYdM*H@{28_`|NiPJy=mYs1|@rdxh!YP-UI5QsGh<c|63^4
zM8dQ)?Ri>BaK(ske)-RzPw#I#f-vz(t$#EjB2!O0<sI2SXKQB~w0B_BElm|jF8z$}
zdj4BFUAY&tvab_i>H7eLI_S@LrkVCpaam$~UjM*JT;-lZ0&w(YDQ!z~D_#JZRwZ}o
z?mC5(7XdgAel$I*PA8`)2WHy@l*8Ycm6o92;=dRT0o7QU>$xp89?P12dyE<;;tRN@
z5`ryeHLzn;xJ@yf?#J@b$B{O4wA;`!+Nk4{5@Tak0;bw(SE36-TibU=4QdjedYOt%
zbF)Fw$r6yZVeqVVeKTIT39s<B0bj%ODLA^>ehK@i0d^B74}k0+{J+B2JfKKHNyYnV
zoc7~xZqH1(=!kMqp;|ZA<GdiTl1N+iy6IBROj%Wx{@Lr=8O}I&?8lD*6g<-Z(2Kq8
zdWc(cl7_L20mPQ`ayM~M?RLjGX&-5Ainnefgl;hP;(ha5s_ac)Ua_9_&<2?kGR_>i
zfHGDN!>Q7At#4L@j#Hc|?xcPXHr0v(jwlNr-Og=<(c6n+yvhZKe*i&K>-up{8gA@x
zxWj|96wBW*;Vi`^(*gH4CjbfeK=3<<WBXTRhTY=TY&I&InEsOO%Hc5m-a$L+Sn}m=
z=~2r=qKCH~^&FoH7${&VW0c3pSNS1U_pvxLuY#n4;hZnPPE34NK_DW-z5;#h>|Oq9
zTBmZFbMX4qbQv@@!+(YfK<AI9(QKO+l7e6VNU1#Oo}6st3q*Tc`f<s2Hb@WSc<sl|
zwpX-q^~bP|nN#o}{*;q+AL$<UvE2Gn25aMbH~DN|IPy5wuUA>5RGJMqmnCwv83w?L
zGZiW5wMKIcen-!gwPCLxEaH88k+`O~KuS59@}AZQvimcC8lO_)#$n@v7|rf^zhZ-5
z`<pP5Mc7pwdFk29^!LKaU3@OR)LtU@#?23RPSJbMSZv2wSEK>z67p)FtJ{#*Y9>wM
zbO6f%Kk}s$C{NWJfL@zP63lM1$r+>U?3w?>$MT|$vB9Qh08libVovwto-M7m!CLgi
zQWP?QUkm{42@YZ5oT-Y)$eu1CLDG44N(q7=Nn;w%YYU(eq1XJ<kliqDn?!u7fm)z^
zzv7P62g-Lyg~V~^sowD6gs)(1U$xfM$S+o7jQ(UZZu70F5jiYI<=;655;7>L;LBP6
zfU<rlRUE@An6Z>a-&GYC`IbxoR8Cvl+@G#Kio<VT2I();{y4VqMeelWKOclkuGo;9
z(3ROEZ8ojQ_)-4;1zkjI>)ysQ5C1}c`r3Gmm5QcBFL<HJpbI<tGYd#JB`4o_ffJN6
zKdbbOJ=Nf3g{F9j92GTH5a!z_&zoF0YdY-4etKhpx8=e8P|5wR(yvAO!uNh69Q$Gm
zanTzY_0be(xD;ksCsX{vShuy)Xdw}3i!xmaFWy4BW5&fIh#i{$vbiL^3bM~C`@@>E
zm(X4&-%(5MnM!MtyreN3DTB%gCRUn|Y4n%QF!_Vwh8l+bjcp4s)NI!SPVu^jJEH<!
z#UzyVUOs?ZLg=WmfJQAL@&0JFgyMK{X(;=<>^O@S#Y6qyM%6W%<gRIGAGKFs+}pvF
zoQ^yzjX%kg@FzN7Zvy%f@IPy8?574ET@!kZ5a+=iP<F3lHsZyg5~$C{tn`azi7Um?
zZ|o&@WSj@`I&ETeg-xu*Z?r7ww$J7`RNL|NEEK4H7pTc!ygi=;i7EMxkeCWhePCP&
z^4g6{)Y;y}s&nK6caA|)1}B<h5vvV<2)Qj0ekZX<IzVW<u*|@3zW}>q`#l!%bKdVp
zJdwwS%S4G6FwUL$qR0vZyFpm-(k8m!UklIEyZcNPx*p2WEFrIc$L)dSOi3J$810_w
zjvFzIZpVs~Mx|Efmm(SdXrkO1WVFjx#_=#rN^2Qlo)MSL=Y$eaFkj<?ZGp+lk9c;*
z`d8b-iz%k%*{UAZQ}^{}Q{^2`HHCd<!k^2u$9K)u8Fft+s<Lr80X|2-TB@B+;)~J#
z(LtCf`b)tLFS_D$AN?fd*DS2$d1k>E;8%b}N007ryICrp0Ly-sTr)e9d9(jzZ%c4=
z#f4qVTCtv8wGIvjE2I7UZ=YQEjWf-X*7gU&0PiS)0va(a<WY(n_v?mW_s?Xp(AkSy
z*(k}(qpp;c&*EN9HG-#?j3G!CGQ>~8ojxlEOdSam&V&>icL_`i)mq)zTt(VuUmSR>
z{Xd}^>}g)@FivMs%7=!2%9dST^y?uc#b#+TrqxylibotK<zD}c-?<`k`7IL}ViK}n
zm-ixmt<9_{CZ+Ogc;F2ArV<c9k2dN1V<c6s-~^D1E41ev`i09%uA(MY=IHD{UH}h{
zr_9(-=){t_zjGDJi(zQjdl8uqg4VhHnqIJ?Ai_7H-Qd+3H!OSv;Mio}?gp|1O(^X~
z)T7VFrO&=Vm=8!V!pV#AlwmkeR^EwF-YHSG{Hk1#5IaOx5OX7vJ&uzdqb>vW{p}IJ
z7D;!vWeLP~sZv+Hg#qoB#?C&DkETQGLg6;7ad7#q#8M!c+8ffbGXc7ZAKmQQZ-ghF
z`mLZiA24C+B{^Lb<$K<bs|*(>UpHD4qo7y%43bJ;1^tl#RKPBdG3qdc<%7(8cZ}UJ
zf-gc2#OZN%40;)EEnZdo-pMi^3Vrr#%i=>Cr>}#pw))SGEZ8>>PzmCq_-y3CkRow)
z1%AU%2macSxq~`lh#P%sp$lz8p$+Z01rDP|Qd(20K&HC`XjWxD(QBPbJDNZy(9e5Y
z+LCmrUBxQB7{<b14ZIP-K#3WH1fMQa;a0(91F{wI4`nl`Kd4*zkcu)uiFm(A4|ma?
z-s!NNK9Y^{9wjP*;(C7KeiT)bQtR;v&Uf1_3jUvb_vIXSU2`W=UwpE$-K8x!K&)GC
zdgzu`45RVe)q&Q2*T`n1zbdJM*r3fbhN|6PfcOm5QWt-~4xmOfRRYB=H~6pO9z6pp
zoDCQTPMX9lMJt<f18q2s^+hjxzqcCh?g92q@6!_=kh{TB#GBnS0qWlmnw-3og(f&B
zcJ>4)+tgkbgH2xGZ?z8gceUGJ$arvtALgii#W3EE3i(+1$FmhWDar652DB1Ec8wJO
zcU_-$loWZ=sJw$~igAO_N_pVhE|#UCai|NoHpndq6R&4&9T}qEjPHqOl76;!yx%g{
ztolRkg^njSUBT+c-1s1D9QBQ4EKW9l>Gp8I)q=-P3~I^hOy784-KZrN5=>w0vVuo%
zf>>ZxY!oeQ_L3m40J?B4(H0Ym&}CWdMaO%9cLsxKH;79y^nX7xgY2dgc+BDPqLX&+
zk)ALd;STFr?LUNARin-zxqEo1!JA4!S$6+RVh@d{0TR16r4r?x=Ya-cs-8p;%_q<6
zo#S`;EJu!)GJkv2Ln73RwByt3oWh(gthqypTDo)7)7oZlUk?s)V@B<;?+ksYenRJN
zIYRa5?w;p0f6^|Om1D^}tb;g~2hH16)A3s)S;xp9w3l=i|CiSO^`F+xyC~32>b0yi
zBz=G!g7AgL(HjH*{P`rqXDyV`8@9B{rx?f;bZ)kwX-riUD)_F8kfbNw<sXc17!$8}
zzKwglUz@o4sm)}WiKi!EU3vS2IR9Kx#6lM&Xfj+rKZ3A8vRs_2<e3jY5bv1%F&qKk
zT`Jb|8fa7k(a`z2N?3QK$T1Da_<i}4y7$u2x6xJO^&+jfHt(2YX6_$q{2RZL4m~+d
z5V}X=M1HM*@tBZGJ}f-K7<Qm_ReZTBc7(b>e~t$HV#}Y!SP<;L$}%+gQs3QG65|MZ
z5WusGS4rHlugNWB+-r&TD0l|LYlxV!JA%b$t1G;|;IEWZ$Dch@l8Ft!p5YrT4-RFs
zDprI%K14A6s3*4WAE+)ULSxu1NI)?}Nc9bn`Yq&Fv4FJZ^F8G)6+C}*nmm-NehlHR
z^OXqF`p)K@2aJ8t*Fc*5Mj`_HlTijf%b|Z_tb%ep!xbGnlxQSVT04`|T8&TTB+a~i
z`zFe$kKCbztI}5;;TJsl5H0bqbk8Bo5b_BjUYIxA<P{TqT+HQ1`Va``+;2~8Uq8)j
z?Tur!P_2L5i#9!Vf15Mjd$eqSkdk;R;=#^oP2ck|$#LuZ_QdVA*=Y`!E%tna`2Ng^
zB-psA^Ro^)1b!#uIwQp~W_z<X*#@$OG5dxHNi(RC@*-YY6yL9qUZR~KqZ;9=smXjs
zFfaBC<J(k|L#jhj3Fjlq!FPVrZy2@qNE9n-hdXs%BE99dY<rIut|-jaY>rll$)>3G
zc58LCFLhHG{KlxYNBro~=?y&pYxovlV1O6e*#*+V!NG7nm-GIkA;HxxaPYE+{KKZx
z)wns(W+IgbJ5pA3j<p+}VYTj3^gLFfV9bYZ5)}9D>t>_=_j^tV%$<Tmct^jBco(p|
z^Pqn_T}vfHVrHE@@PJ=$_0p(C1sEe~Tlf-LK3=uc&Sro9l0?QK84=VMpQ-+4%tJLd
z0&Fsp5-3bM=B|TNO(c(=vZ|HMLIdxv$>#aN#umNvIpu$==|5^~yREbx?pH-S=&aN5
z^H+eQ0vqFb>wm1GlOzh8mZ`_~$5O{oaWT2+YNcOGJ8F5_M1)KTc86=}jLJO&$*eXn
zyf4Xg8Zp-UlFKjgL^zwgf$0%lxk)8N)p_{<6^_Nj{mhnj_GA{5e73cIZ5r7nsyA@Y
zHR~qMH#`Usl3H)vMG_-VmL@OHA|hv)%Fi~VMIm)ZCaZB&nkJS^-JBb2$e9m|*6V?)
zl1@4uKYvMb3rEtLyv=R4<rRpyYp6B11EmnPl1T1vPfx8=w{<V8Ytgo-<rMzSWpKP+
z_1lhjy%sphJjvjrDeRRhUBF`9N&M_JZ@r6Hh@1rQSnkmSWPAK}l^AOBwCytRx^|@-
zyFe%;_f#--#p?ZgMl>9DviCK1$?wvT(OSb;W-^wk)`8-CY&dp)GbHkzhvjS^3$*;B
zoV88_-ls@$_$8No$9;UKu93yLahHL8fb-4lQ-DQ5tg7y*-tCPGugjys^r5e%ESy%R
zo);h8DcAQ2x$QS264QAMkWz%OvSpKJc#9zc?|V0`A5s2}{sG*wB%#$39Bzj?AY;4M
zpUMIMNW?k?Q?BR84ny-rXIRdd`Hm_F8Iwkd6s*)OTbYK0QzvUKk%QHN3W(jbnx0<C
zpI>|<0Uv1qPRZqm*@nZbyxyXK)fVIcQBi)Bpby?kPai8i#(Mo&0O=z1(-PJWWP0M+
zhwoIOrSun{K9y~KVth>%RuYQx_AQe?(rEJ8M<i<MHID~go@4{|e3j^w!%dF##5$+L
z)J@?;IfcP7rY#OyloTg9E$+{BX2LzAWU_pPZCZ_E>txJA%cB`|>K|(qHm67g2qVZp
z&)No>^u2twil3Wq%g7-ye|J5jXIH@!fB&XNbg{X9{3~16Z+xtqaLmzgE{^9%`v@C$
zNkcs@3Z^IW(RF`gb^~!Q;#h(~s}U28zWB*HPYOR$1zt+8Gh3-pyU^7wr38Mpz$QMY
z3<wv_8M!CMj~_pE`HjLc|9lcQ3f<0?7a8%B^fN6=eX9(m?s6s*zZ`3LqQP9NSNJZ>
zS$8f{c$8t5+#8?V=jjV>Y^65&V=}?KvEs%vf})Q&^2s}^ADfnHSN~nHCmz+O>*)x%
zSC5lfx@+g@+;Qo=jyy(52fJCX{HH{7S@owIZAZ(jHc4H0_cz<Gb<TIz7kFq;m6ZvR
zdfqnWGGq-TL;V(J?Lf@T{-CZt<QO`+^OULUxeC%$O)T5zD%(STv|Mb{cDpBe%9)SE
ze*_O-&NY(~>D^&RI=yJcp}^%Z-6|p%4{k@s6CFs(WNcu_`HTg{bQ`6%8Erb?znWdN
zCXg_a`)d3)oegy=uKwF{W$-C%$WBS2cLZGNtRaKKg<$9DZOUNV+4hUQfmE#X-Hty|
z_CrD+pq3a|cA?CSSj~6J!7_fgZ&dS<F6Yrfi+r-z-`H}3Dl&!=1ovcTyZ=x~vRf|D
zkFt+#xk$30xjdKnW%(`CXRwIrYVFB>WsN=}r)`mJ?)i?1Y>DLjatFWVMkkSryK~1E
ze(QQgiaez%lap{ZT0*U1v+(o|;$RP(zgFQ_z9<cNhZ}J!b_QFW;(YlR(w)L)rnGV?
zd2I_7CibJ!{Nx$hJRs>3^qJ>fjZcF@+Fch(nGt%5!xuC^o<(aBQX{#1LvYt=;lGH(
zby?#w7jaBZ0|s=a?_|=pudvNHd4El9$|(jrta2JuxxT5gEuQbNZB)uNU{k7c&Nn?@
z@YD1Apj*G?Sxd$*`h9Hwxse<bxaAlRE#@#hBELTwF>i}znnF*b2hCP+p48>_yz3O+
zYzt)dthyMRZ6uz7j;Yo0yH&eW6{L(k=3r^exW9O<Hj0UVd+{0cwEQ<88L@8U64z5!
z!_@Rx^}7%VZL}qJf5W_eXE$|xRwg>rG(8rD_kZzN#*{=Qr+_XY9W?QpWkhP*+ewB#
z$M3H8d8Fim3_3v5=uCxU^ffR*X>h{~(HqtY;t$04iIoPpbmM{ncqI@1+S>Cnd<+XO
z@&(*1X+a8v>!@MpHE085z9{?~mFFEnja|BHmiIo_R0S#oI(oim+b|$G_}r+M6h6WV
z`(3fLs|(DW(iDZj0vz|#9iF<|@P_;k&F*9uAQ~6fUb8UTmOiP62R6(^voEzvIp%~c
zUl>a=?m>H}EU)Wze#ecIzBlWZ;S7~vn=onb)nz1(VGofGqf@F>mY39F|Aao|@jd7f
zW+(q*y~~l`1G-VyR!URNJlg%Ocwb!Rn;M2s9l^_b52J%IB6kQ>r1?sLS6Y=zRb&4i
z<X3($%=~JL2ih-#FM1PDjY1;B%vw>A-HYyx#`9DgJtqI{7N7Bihl7-c(vJRDW#1jo
z=KK9EiBVf@wWFmrHDl8#YPBd;yH?cR)Cy{^Pg*S{ifZXl)ZS8U?LBG}rDAWfis#nP
z_xpQ(e?E`DBCk95eck6=XT8t4&N*6;b@nt7HGk7xpr1(;a=02Q&!?r4#?@ohnCU=y
zlq^YF1~A&Ffo9{;cMP<VFzm|FtI!qZzO|5$^g1tm^aNPu8_h|X^@6lY;qkM>Zv|-@
z1;6@--J-mY%FZQjr^Rlyj2-WL)T(Dle-ShT_BPMu0-rs(_9A1c>HD6lQh($=WIFgH
zhiX$I(@7?qX)SbbX~oE8xbUlYyLwCbb_X9Y;3%HbgTi^mh<g8ghjzSz%s>hHI)dxT
zKC0fo;BxyUxG?X(GU8R1%BhJ6&IaVV;tUs1qI~@Db@P#n!RNro=>FdvZ{H%z<m^dD
zK9c;8ju3AU#Lj=)-f5qFs_2~RbWqY_UGb!|^W~^iTc*b44o7NLq~4sbA*V86>p2Nq
ztE*l4q1WUugqKz*>cIQZ6dI1%0$K0(HZ~2hGrqJYpRdNYX8$?ciE}z=OB*Y;-dj?5
zQm93%TXof5Pm<BosGKKW`&qmmGdc`bZi7&PqT6DW2yD>fJBWaUcm{OC?nF&Zya)r@
z4{9z3JB(e2>5ahgM~i8O@*BOw%g5%1tEL~i%>qMzByUP3a%?*~eqHSJ`e|POsI5?6
z;p^^X)pkvbIhN&~i;KdY!dqmV6lXv%JmI+AE2ogS^rc+{2fWv#L}Z~Mf*$1=>6b@_
zLls57f(=c=1ItjxN>@aP3qdn`!MU^8DQpE+4tpKT4ukd8$wfEQG2EXkK8^)Rk!}w!
z<;vIfb?P;{+#JjG2}7F=cJw?=VDOjT<c>fdmL1MUd(2dl3|<cIftxqYH2BN!$3qpA
zCZm)9yOW?Tw?!Z|AbKc|HwDFacw+D1xNtwfsim8kZhG|6B6L4Nn?r&l6Nd?Dol-JJ
zq>{xZbTstUCDU0ZC6pa!FAW3ss@dvPH06_=liS#r%boN6h45s%yfSB7N;LlUu$%ST
z&w7oq6giFK-Fd+Vavo)?)72C%>Ix@yOAgDB&(%eN>sNlmX6fvbCEq=K8{J%jZCppu
zV&OHWh-R{qd{kaPJmY@+p*-z-%EqmlFz^6FEm5X<x<>Z$!}2h4!@kgGg58C8<?_YE
zt#;@StD7>8YL8$!c$|6=o|ALEE?BD4wMe(ff78t#1bck!bEclmPSz0MV!AcqEb}QN
zlXYW=4baC-D|uTcxC)}F3?jLhZDk^6vLIU2%LfByFgN|FE-T~BN0)}~;TNG>&*OGG
z>e9R{|9ERE^f!l1%D9c(u;P=0n_ED|>%K=d#p@Wx8LEy#YXt?fMjMu{_O4p_Bsd-b
zq6UU>?vH{A#N*BES4JIJs~STt4ktDa>l5x-7}xCftUCwo70D$CbGO^{p9!CdEo*H&
zodllImy^{Qa&VvePf0z+>mcxC4XyE4eGg%?$9o&K-X{>3NfYO~%?aZ;Ep|ZiGLd={
zzrb#kj?LFkEhA?IwxSUyL}5hq3=BFtI^=d}A6<QY>eG;P>C>A8N$+onLbul%w@pbi
zM(dIrvgh<?68(?E&bMMz`b(^sfX*`4R4)Dm4GnjD&&MdP{B&0q$VtC*JC;A|PV|E_
zDxKS;{!aZ%M=rzf>uf>5xRrov&wk(>{KRneXQzXgqQe%MKOS@bp!k#<WzPA#e^nsj
zi>WG_=7c~eKWzFO%ZB~6-Mj3*GhyAmbzm=#zYQq>0dsTKo;`nl%LIcpGBUby7?REy
zGECTQr4`>E>KFIk?oi)EhLJ@pP(51kWKS~DTKl;o?P1@(+A$3{Zs%wDj@|W0x&EQ+
zODe$exZe&30qZD!eW2F{+PdleRGzv?I9Sq5OR{aX$ktTrxZ<SKrg+p4*mKtx_v`jX
z=^tEeJ;a7^&*hcA3*`QID5(zAbD=UKLSLa(lO1eMR6JA;oJa#%dj$OrLAHGCl>V=O
zcl&c-$+(f+O|O;bciR0!QREeTT0jjZHNmN8V7}j~9%<9mEmko<(8kJe3*f&#$?e1r
z^1eD5BG?ckl6IeRFr<eh06a4QKPjJybUJRP9X!o^q>%0%KZbAV^PN*|F99-Yg+FAd
zbhuhq4h+SIf|*5!Hto2>h+hbO{f~mbZ6{RQ`Q=GfDmH9OdAynPs$8tyw>&d(`ST-W
z24F1phybe3g1CrO7|;TP;ep>!s)L>BAUZTZ_Ob1IUyt|w1RA8GRRHaFHS8G$4wJn;
z;~d|i=67>RABh`-E|VxNV-)FUb)|!Ky7lb$#S4O+2vTl}wJE~cHk|{D`IwpQbwK3W
zx&1#>_N&G%qz)Hb&o#@&d5&|!!Nx_aSEfCi2Su#$eSkpx@#$U#1THxVIe^C~-hgpN
z-dk=A2KNwgdH20LSr<|Z_7P{0BVNYs{R!WC8Ky@#ZM+VnIbz>tEdvE%;fJ@ch_^Ei
zp$oKvJimr0DWAH?Zq9Z5&=+4?$W|>3TWF;%@HZf{rQ$C69AEYL>p=cC7IIh0WT9E#
z*7fJ2!$Gxs^xJEn|5!KW*?s$$HoTgl`z5{}5U*RH(_%hsyfxm^2Re&)4=s=5-@RFJ
zGx(zBknN`Hl{G@i!k@Q$Y!Mql3bJbta9=t*%K|;wfsWnc%>%Yuaom0{)6je}bq=n-
zV(;CKbvgiyZE5omw8qTlGgHLx7k{l7{f(=})tN?{$e;IAEO&WtIZ@mFC}F~k;zVYb
z_esrUTgnI478J-r4oz{z6}Df&8#f6jYgv{bO$2*b?!NqN^~t>Hi@nUJvG-6aE~>ye
z_nhmO{%%EH|I=jORKtw0uDn;#56bPyi$)bw1Z(v2U6V>;Xb8ZG2XY1+J{K5SC1g_k
zey5!?P%VbSt{GJ>#CJ<k+sXL=k!LIbC@TLKmx`(|9dI&;d*SUZX6M=JQe<Cj*T4@)
zJifkV=l34|Aa~DzqNed@ZUo@2<pkOudbroSPap9_{JHnzQL2}D4Gl2Fu`N0Bvjq@6
z<f&(k>fT8VSV@*_dzwiSRydCF4Xs#WC+*|tNXrgie3~U7aXz^-9X`1MSdF8C-yA*J
zWa3T4+!4k7r42M(!H5jqsT%Qb;O4vCJx^VR{fIXs&I_g#{G(LD_Eq|t>4C*omU{*n
zi_Cq#NR5;t+nqVt(*F4JyQ0->e*22@`@79J*>ON1cO9%)0?$L=TU!-b)?I{KWieI8
z3_cw=IRH$esD+<cQzQhSYOQyruJfg$23$eS-Gw%8HhEtjo6_Sf();Y&tM^NXPJ3p*
zxBp-XS_gEp?UM!S@kTZS1)Xd^plP8p{eOd}`P$@iUMYt2GtbZa;w>ZYX4d4M%^&!@
zMLz-ec1#Rg%hoV2a|A^jkxIJ%kY2quG9S}zgKMrnyjQ338=Pj<_@Ls+RtqiQG^A|-
z{fU{(mi2h}zstIHbgsJ3Qc+(;h8$m<mzKo20xz!zctwk@orLR3FLoKZFLoqlghv-&
zrS@;*;Iu|nkItVjIy$;MTlt;ge-r4Hny*D0w9E^D4|BKly&ard_g56ahe_g<(ThxG
zNTDiU=S$H{{sJ(+`q3CgY|qu!?{fmcZSKEb-DV8y-_Sj{GrFg|i3M5J&X1bejSA&V
zx_`g2JY!)tD8UPKMa>!vIkXzD)NwA|oU^HTGQoK*MY4?*AZ9p#BX*Z<#qOkOUht@h
zy>6dU?JK~m-T@5UF;|Tfbt56GjT@*b55}(uT^+M>-dl^Ybm9j_H=P298AeQiof!^q
z+@<hadBo)Lug^+F#D3`Ma1{70IPhyjdd9AjU*9dI*>i9DPy9>4$-ck{CWYTX|M3-$
zX1O^kT2(4wUSQd_BhYAB!T8{4r_5jY_rfpSPP<IXB+wJ<FVox12h7^s^=G=`;e0`B
z_?iwM^L=;)Ekuk%?b~DOJtcloYJP)}^CwBn>Kl{5;PUCf+`cVWjb)rVP)7u)`~IID
z^~M6VryfpbA2i~0#Zv;vfwxbmW8QypPmy+G@pvm2KYCp1v$zc1eC*3&XuW?;-&MMI
zFY~j7e|4JN+*4rC#eHEb^#Thy&i8g2-{Cn}uO8M+X&uEouY+Jpr^S7qzJGS=E$-a$
z>vjQHkNA>TnmmzRoxu_w6qykonh>o;GxXVC?vhu%t}^|5nKt2r#j|g9CUc5?o!whE
z#s~TD(R|=k^qZgmVsF<;udR*B_81~h7q7(%O|jay3NpwObL`jkuX)>5pMh8$t%+43
zpZT6#4=x8DB>x{v$g<Roovmz|9@*t56`(D2oWbrDIXFF5ZOX8Yg8TaIE}m{ilW{*?
z3T}@pd=KC$fCem{9r^GNaMS=c7tZd{#Psog(j{g1WgX}&@0t#*X(c}30@0&{W<|f^
zn?c(G4h~LqHt%phC}BUMnvFK^w((l$UMTd3U2rteY`p;#8ej=EvQON>n=Q_1ZQP~J
zAASGGx{?CDF`J&O__**ReijcV&(?Y?jNZvn0JQ(Z!Ossh8~gy^qRatVQ9~`V|C*cE
z5A+~!Oe*m~v(F+;TREgMi<ssoMTO|8_jT#*?)K^Qx@#APK2S<76@Yt2F0WjY&mE=v
zd%|Pg$!ffRTv&b`NQxUOvBTL_0sRY;xfj~L=HL1UQ><SPMCIck25=nhH-TKFiL!5@
z?3lH8Ordc@KujW2ft7}RGTfBEjAR@)^Yc{cobBT~v*@eHyD!bZ`(_AiVLt|-++(QQ
z%C{B-C2k12iYyB|-b{7co^3s3CCo?whE6+<pEEOb&p`|ss$LS76J6T!GjkzHTzYBw
zr@mP)l0hDdOzeUKrxt<8?AUO;#&EdkTkHz`bse2Sv3$uw5k)lvIz0D?JCQ6{>n!yH
zi8QH@`t4BCoxnC0aSKxrWf!ampgBxR`1xT(NN#>bY<a-e!#-C(u7wgdt#&6)XbXra
zj%EeeZ#J>2=cw-p)|<qebz0onbO6Y!a-<gBq?wdwbiH<-)V*2XpjtC7MRGpl#~1T@
zHT4;;7W-eQ+sM0tyaV4urpH4t#7D)ga5+b@&d$#H22utifYe_9K`?OU(dJ)VNt|Wn
zpN}atwcYP6?P6L#J1!FYp#XpwGm3k9NA0S98o!UmSDYXEI6AIqe6w%E7Uj00`5txx
zqj7=oT*mcw!_&6Rp+%l;fjUUtI-b3AhPFfj9%&hpO1*lu076%+Yxu!Y-G&^1Iffl`
zcTKvCMP{2bO%ERKtcd3N=qI_Rk`pk|!F!eDhYM*3$Npnq`O?%*39^*h+Ql}dIlOZr
zhTLPt<Y6bvCc-}rJdy=dMQjx#D|{#V1CKe4X^`=xK6r1p*b>RCcNL&u$1IYt>wuA7
z5iXLt;Yiowar}b0%0gl%Heot+wKx4ahX31NJ4H<hKhe>eC2V|xAhM+sevzazudGNo
zCORE7AKxxm(G`z@^5G_3tKeg-_Pai;y+b6A1FW{1mQ-W3@B4sjedT3ujS<UpoqqDz
zb>vQ?Z-uufS4pt~c5F|!ufROTDY+#;sU|l-(fhYgc(P*PDx;@oBM2B@pTs;_Ki<^P
zSLGbFKjnFKx#^;rFr|AZDiV_KY>1qco_r_O_8m;N&U4Z1H+7Z@l}5T!Rg#nl1V+@a
zj$_Ff7%5NgpeELwrM#BZoIdAj0zez<Ot5rdRb3LrZ{a_IxpNU?k!;+dd9DlAfHFTV
z>AUjCo;0d5uvAP|$thXP(CEt>r@hT)cCoteOO=v^#X4H2-M>fI>p|Lkiu{on+y+&u
zdw_egZ1-fbK7WtMa*~~@#~<{i%|FO0{a{n2Qfhfj(bT`H|4!arp{FY<;=&@^tL4me
ztYu^HR*|I(TTni=@dI;n3D=hfc}l_Mw}*;y3SWe^bKY$TsMJL!las}zBCTkr$u|g<
z(NprW@29TfYEO{c0SW^#(`PG>?aUPz{Gqfbh->7mB$lz~6rm8MUzn-8@%F-?Zz#oq
z6&@CFnj*e=Z_1PN<kQlY0)9F@PmBl5>}HMy5IH>y$rJ^0f7aj$?M=l~6?&m-L|!;l
zDjQJk@Sy1`&>5)&7|(v|^TT^$p*l4#&uO|?6-~{0-<wiz*6UV0an5n%vq=_Q0jmD2
z4FMUxII`{AB)^7TTO{_*=}`W49Gs?3M8p}YgHSh-A<wxop2^BGnD-Y~7+egZ-H-VA
zUtmF^Wg$`$JBqc(Ud4suRAD}^eLeE$nlC`4r|uS9(@amtE-_=F--7(&IeIB~WLRgt
z<!P%!{#p2y29YKWKDeKvMc9hl>x*`NJvdm)B*f_D^SlPwxs%fF&1bv1y}Q>1zTv(q
z)L3h1TvAoA_!dX@i#^(9u)eG%JN2T0#xQh(t$w~tJdAsk#jHjoN014Y1LXZNM3r8B
zpTTtKIn)cRGBzog^9HO4QUhti_-+MHIe~Oh`D$h@w*nOKOEG^aT5W%qyr6m|^EBf8
z-0-7maaUdv>EK~wGk~c<@I2QwEtrZ7MmO#MWBb9S@0}=IFYzjVJrt)R?r0(3MwK&t
zurc`3^zYyYT<-jVz~^utMRStBJ6^dRfqlyT^smTIY)g8eNCYXhUoQS*9V<CVzmiuf
zEFrFlg8)lqLT^R7b_gu*rS=l3?89Y^5&WPt#I;#?23Mg^+>iREHC9<y)&-W7qDRMx
zf11wbg4h4Fzt6dwT`BF}8g}exHIQ$>$tpo~zRQxLuP4bVPxPGMrQUB=DL!e~-*c<k
zLaXJwxty8(8-9jy#H@KeHIl#I3=EH)Hs@d;aismH%*NFo)o~$&p$KKRJ*CGLYH&3C
zgyS!STZlRLfKK6<&q?td8hB<KqU87&=b@MG(n#XHP{~uuE*&_R+ab>-!tiJmQSBC~
z&WHP4BYiyb@kMHD(^&xEek6W{12Z0bSI^dL>S<;Ki?9Ijj?V#*KY*ut6JGx`Lh28Y
z*o@zSdREdwuH=@NI=&4#TrAUc)%XSwY(7SZyi-NZsHwY}{0?*@tWBtpg@b>B6R;3$
z5zME^ck>MPLY}HG9Pd{B&SPP=`tT;M6(s%fXS_{#WV`jxYbH+oew)RgoGe4Xq;TG-
z23(|^I>2c;_Q`n<bUpPyoAT|iN|fPGFUJ>li|gWBqMmLkly(w(JgE|^YI~Y2kB1#P
z!L(2;7~iwAYFaI=`6afei5+8V_wm!g;Zx#}X0Vt~UvJ;S_%2o{uh~~jFfHsqFiC%w
zw2RTacyWBd+bvt0Rh6t!c{JR26#ylvkIxRTnv`HV@0+>*K3S6+ESsM%Yl_f8XJy?P
zh4tMK)k-sRomIN3mus*Qy!TJB2%+lnkCz5H<7iCJpi#P*sKHsf3;;U_-#ODRq-)`O
zedNAlGHIG{XG#d#GK1f#4(j5D2lRFIbiH~jQ26|1=Iw`%uC5$}asmHXD=III1y}Xo
zJueR5W%+PY2h7xt@b0S;MYd==pAfv<Zi`vygPnImH*Q~$QTYKB2f^{Fi5*&4U)T_U
zrCrCW6hsaZ0llI90dq^jSadX56IbO*Yn-p)2vpxX_1ccySe}qdT^(i5zc|TCPiLe!
z^L_9|L+Z&wyKAhtBos(FJt4dWYF3On<>8o^lqlG!A)BLrdW-!ko+wzZ!GlQldBchf
z8gu>Yp9;Se-V!GcZ+b1Jk55mHTH91^{co%VIj0mM8u>0po}d8L*<KqPqqoZUx{v#I
zF5Dn0?(?Bnsn{H1QDe#vH^)z|O;0fU-*_2K2DJRGES<%`A134|tmGvd<hXIM+pD#3
z*5or#oX*Zh5k74AtHc~Ep4T*hCmSbOYoAVstvo{wJ(rg&)Qk5UkYi#MgbB#we-g!O
z8|Ntxl+v}+sFQ^zAEu^2JV4D|o{J?lbNX0Pyk%b92%nH3`N#WH{XMTn-&mzTYlk20
zXaTxIKWfB8eg~10Qy&rtj0u~(F3ta^eBj*sJkm1H!rGn3`>t3CCe=eZcUs*l>{0W|
zD66sbCGet7c0ZIKX86Xu#txf{otjWg&}Oiekr;u?F#@c)i%=o9rbkmO^Q4@joYdrU
z?LBW5MZDU~k%Q)ZKNj(m%z2-!%57v6t{EbhtugWC7Jisr>Nb8%S9^e*GUOnNQmjm!
zt#Hv-O(=$Pu^Jr`WjD}c0g<*<WIIdzz>PDdjgOlUtg!#M(#gtT4W&K-&P!>&<*TQf
zBChQ8C#|Ql_}E_(b)IJ4x5cO&IlnB#C4<t$zUZs72_u*%g}+lyBXBD{O<imnf*<Do
zW0OvvTcBlk;<xch&R@vE0Sm`tmj}JNMe7k+l3&D1mqyj(F33NqwU9G(l8;Z)$^Y{y
z&ETfn#RcH3(Cr8cH^~He-muN}Cq7#B>Ek3lU!B=i2;dr!tToDd??nUjAR(c|&BxUt
z*G#GRS4SerxCWY^6p^mdKh}HdnBQA?u0HjWXlnPd^=D0kvH-DQGM}o>v#nCeHpb{)
zV~;m{QpWf`i!vU3wHNPt=Zb##e(k~yId7QX3Ybaw_Ls84o*Q@e9y_5HlExb!CrhG%
z)FXLIrS#F+AqKZOW^-_C=PA~D6ePbPUAuEzeu1i~%A@yi4{s5{$I!@!4WMA@;iU3v
zjM!UUd9e2ph`@m0`{~cwrmQb}y*MDtK<J0Ow+6ayM}};6l5ZQROca_)YT{90*fu4P
ziupJDNnqL_lzp~Y|JN+D6v=n50X!rK6D?|yO1%l7NL8-B1eXD|=z7-@;MZ9)73k_h
zTbgm5O*N0_bG{cnn7qh`e0et=O!|1v^N%q6X0FyUZRvJErX>nUJgkn^K^ixvfLqLR
zOjgItHu-`ZzkgqTG4Z{m0O{*$%~rpT>x)?G-No)8)yA$aFDet7WYh$h2tE3*T&;)!
ze8yfq7yc8?r$GDK)2jvv<&e)SZU3=yWzrzWC#qrCux7<LxeK9Hw!|{v&VR@gpSXPO
zYsx^O>4j|xbh5OMpHddcWmM#qis<T(qWa{p7Uc+tDlpGgI&W=OSpcheV8R(>*Y(;@
z+O}IfZ|ctGfw^#jZh(x2Rs|`=@~SO_2Z-wduTEGvV7+5Tsln$Rp%;=vxs@!1Lu>hf
z8<%?B@8QaS(H+ZGc_6YK&hGOGc)4ogP$i_y0|1;zRhCy;xs!t%Kdg2!?o-jvLf#$w
zY+-;JE6b-YyYZRuJR-n@uS@{zUDZK8J^4&A(>0zntc}(F4l&@imPhf+i<ub|-O@4W
zBi{u0_ps{CLpKH(jnzhe9%@q!WzR5({mkaY3k)fFy3ITX0ICiESYX8H6t(d4^L$}8
zu)bk~XyVgut6v<q%|f@WN=GktN(XqGqDjM>yYl%Su!BS+yUCSWN9!Gd)YB__bsG_(
z9^!Uq3<&5px=!5gO;|`aZ~mk~OtBPNv(dl6Q8fr(qK{m&vv0fotj|zEHt$o~nJIwp
zlXsjeA&l+#<yFpus8>sxWCgIUlOrc{Kce*f!1I5=wd?H-d-7WTKHxf8m9(&}HHiP9
zi+*Uafw{hskI@zJ4q)vrPU<$&7$1raIrpfp7il@Tqwfs!>lOKixtQ}n*{|LXj-ZAP
zRaAF)WAt?IC!+fu<Ub8=7YXiA<3gkB@7Oia+b2sdu`=&#-g<cT_syHQSa!Fl1#K^^
zHAP>W<6!}1rKzhDT-`znrv!HXO&|GjzJU?=MPV)Avd31rzRG#_sQVVlH_!*hlC){r
zO2eEMd7r2%QY_n|hiPxPC3UvzC1Rc6q@J6=Nh2Rm6Esfhzcdo*xH$9ui98nGm47x@
z9$jsGWxDtw_xqI%Phjh&g8(Vm?n0b)1V?gAL6it#n+zRQZ|?U!S6C!=ozVXjS?=Rs
zc3=+RF_}b{*xo`$3MFgpvC5n#?kkPAnsV-x4su|rVgsL|I!sutK55ioE$w8rPZop#
zSF1ZNq~1%xU^lMez^T-Wq1<D6*uv3~bQ3GKR^S>*bzs*^C)&zxw)C$(8cw_XN)ia8
zpTSDqEUo*KvQze`7b(ywQowf8>$Fw3sdO-eKbnvSE^9-eNWumsh{Qs7G_?B{{JYx#
z_)U`k`z56+Y<<UiuT|-nyI}wf%^n8(1&2X(=+T|9azg|msY3uZRQpf6DLopm=>Rx-
z%;-9p&pp&xSAuo~)La1O0xQ=>cw9LfWgMn1I}!ghzlF6@6Tpe^kY8qzfn)?TU<we|
zKu640vCU*&0$wl)W5h?sUySysSu9VUR2j)pT72cslqahMYY=rUMJ8G3HCT>@7aOpa
zWXbCWZ?gkKAOb|97~fAsh($Tu>t3|mkF3cOJ65n8YU23mklAW;EWPcGfzc^b+8{>O
zc?|msr;0Hm3&Mw5v8xrkeEZBy=9Q(H|H=0;A92+d(zQDgai4Hxlk0&?=+Hr)drWGY
zxUiOmIWhwVw3E73<Rvlm$Qduwb++E94PDgB82NzDzMFSZ_QK%aFoG_bDam-6Mq}FF
z3qLRbPk4xZO*9n<k%^X<Jv~@IZR&$xm~Z?!RE&+^2b~ZBBmbT~duF8O9rZjb>nh>!
z#!7MOuJeJh#HWHk=4XsOt2ZJ{RgaUZP7>PEMI#GYUutj)slopM^-_}@7z%6o5r
z>FVh4PH){C`kU_hPmW12bM)R{ug&$~axu!peU+ovyqNpPHbYZ;!<A2)3*(Q-g3nJc
zS>Kk<1~?2a^TIa&R1v>q{x$A2TPc3t&c$0qJ4esMb(BkVdN2O)^39S3{r_f@`k2rR
z<#LxkP1E5+^>$Mtk6pMdO)be$7d)fs{X_<mZoGHvND=j87HRjF<~Oeoh5g@LQ1fLj
zD6*y^kp929pc^P49*hsY(s&KVgsQ@ZsY!)`mGIL6P&Sf%7BmD~inJX^M-|2VyNDOC
ztNU;~O`^7zEu#d8(^3)=^c|kDTJeeKcV|w*Iyn{DYtqL=NhskQ-<Fe&L*joZl&D6?
zUP9y)@hY?%O#i+!6_mSa4Kqcw*gyj$VmLv!rxfrskcb11W6lUhYN}kZ0QE~8+5Qe%
z`*7t_jFKhq@zmiZWMSoy2i!|FA0X>gdm-t5e^Yw@AvcASf4b_Y+&`Kh!?Wb^tIU<p
zNy6^IV5*Av$=eMaU;gDPNrggJ?x^9{zYKrg^57G<>#WRQ;$F)Ae?KGvfv%Vux$WN@
z{+r4CpY);LXxhQQO9E8IFo5p^FeB-M0TvC|vum*Dnhv?~>nvzP#9MFPw6NQVW_q+0
z>=$COM1n<5DA7U-He`ypoz!s?3cm(pM=V01F3br&T6gFUh7kxLRd)dMd4LE=)J}}I
zWCFf7l4x-b$FFRimGqKp!3(Z;Z1HBaf7Y+<=G6Hb9*bCHgdTUCDl3Y>x-}Q)y<Wnf
zMICWn#q>Mh5T$oaOXrE`znat$w6!(P3pMA38767JD9f120@{}#=m#9<ha%VgV{%F9
zQ9t|YqDeR+h#p)i2Gv89b9@(C&6l((3sVse*Sxs4L1o0qD^$d*Aa{&!L~PCeJd48_
zCb7iiIoOT(Bd96zQzSdaH}`4PHk5vS>s5a+&^-!50|wa7^0eqXnkj0Fmn`|4v#4Lc
zXJ6edUU4tRr+KN9;JG&~uJ(~(*o4afEoojj*n#+1dE^Rmv8Ik<`}+0=-_4~mMupa>
z;sPNoI1enrI!=V4rUDfz;62D)=W~1mD8Eu$;SX;NOgmdh@h}{@@}`R8XXLb0)dDq7
zvmSj&Mb(3<WkBVq{K4jG<d`8BJ_&M8tdd5k`tgtxbfFH(X0$b^EEP6gZ9m-%R+t=v
ze)4}>ohqsh$$W343*?Bo@D0PKmR6h_u1PTpV6ns11)r7;zQzTHR`USYS?Zh;EFdEv
z*QxvJs<4Mx84}KQOgDx7#3RiwYaQ?!K5X>?mPUHb>?-jA+b$l{ji{hw;}y2sC$Z9k
z`6znv52%V<PcHc<JQg^nyLy8SNuGX`Y^|bk#<mJ_yLB-rBpUUhAaV9Ayk#R~Fhb~>
zNetEZsOAAU(*q75^555OeOj9e@V_$nFpa01NjR#68!9LO^M)DOAV|RipwhUq<UwW=
z&IE0JTgL4`{PZSWhA8vXUxDi?P@Th*t6{=0me`UYN#0C_u1(UrI*6qN?bvuX7PJ}S
zt&g@Yn=<d|YeH6tIH}6^Il1WNT08-a1_&&T5FY`_mkv34DeS=V;#YZ|xZh(?2-t0o
zdL-G){L%f6Rc*vCRLy|iTqZ7wnL*>8>P4Irp_(Xy%)`I0_lo|`BaJfueWfz+MvduJ
z5)0KWNCZfX<b<dy(o^#q+=L_<^mSx#XZ9=(Yld}P-kQDuUr^l#6^NuqbZ;USl6Sr`
z3$-yeq+)zsmQ`sdtVn2xq);wcNWQ3(hxdGV^iBSO_nBpdk9G6zjt;tFkCCcqwU*H9
z_sb6U=q1SCr_o$FDT<$7#$+=mw*H8JG>XpnvPy$lX*7_2XYxh$R$rZB*y}!KbpJqd
z+wh`8-JR%PKko&+Sq#HZ^K5x1#;23rj@Q=3PY?Z*E>W*>pghC*0Eamh?b!YP=2tlF
zeT@s3_i>~@*iivko=U^HFkTURG3Nc4J}&8#L$l}D3uw9~?*~3;FmtcQzE@uu)0^@i
zQF&|a@ERy*6QzVc!U%<4jzGUtm+;@0&0?Cp0h4KT_f`wuI<~`GKIyn~m6P|33y7iU
z&=HV8GK(@+dOgmK-OU!l)H%~6w1xWaP;~R%FvA*%0Vx;xy}Cy_68?U2lX8`Hz8+OI
z8;Z~gvSem-4aPS~!_-LiXCu3U?m$3zgs9-L4LG0Pk29fzK}{8Bho62*_Y&AE3sU=1
z3b4k?D}|{BpiRw|^a{_7NZ+P*yE9rvjn*ADda$$1osYJw$b`Tx8}13V*FzFK;wHKI
zT*#%5mC_-n6u|Ewgd^h6iKURkB4YFEQ1by61!LeA$1Oy3=)z_07a#5k85fg(1U52n
z<*OBP{PZU~A$#eVeW`1(-vj*c3-OF+BAE*H@>qB1WN5IY1bqKTO6}c<*pK<X%IB<a
zc7H_Kt}avhc~lFZRol$ZTySK;Zr>ls<%NE@GEJdJE4`;7NT<i?PVzW0{z7#DA9R?l
zVbX?uqdceYQkTzgw*(xn73UVHd68GFF0j!vTS|U1*U!_kphzV5h7-;?>AZ0p*M_vQ
z5E+4<kUWbCUwOS2=*kA@gUAy$sz&hiudtUM>e-m`fv+6*TAGn#_qg!7`F_2_EIrQ4
z_i&uJfXd?!ae=S-urYuNdhnjl^VE2j;5+~lKSO=P*8aV5+TrBE?A({Nf-vPTrdn#o
ze)TVzX1B(FVC`qCb9a{3sVyc7tb*UIkV=!P5W<7VP=ad4En!(tTBxGWs*8fAd}rQ-
zS$(;ywT$dW9YEe9PU|1#h1b9NMC_&|1zj7wke|{5l;OM7U3oK#cQ=+Pccn&1a-7`+
z6u_Ywhis2oHe3^5DC|~K)}QnWN%L6KQjrK<-XlM}aZF`TMU|-_O!UhGEu_?T%RuKI
zZ2<_Uo##~B?cQDXuNepZEb70s&4oZP_W${T!5g+)5-$%8mYj4Hq=2%Cz&H!6QjrYk
zLTf9`-R;@7%m;Qr!7Tq9#M_%*&krlJ<G*WGC~599-d&1T@f(^bA_7XX1=TN&He%9W
zO~zK?Zrb!(X4VZlG*AKa6#zb%_u1@WpUKL|YALOv-iPGnAQt_g$0k6bUx(dgb>G<;
zq0Ru&%z`$7PKd+yG~a##@}RIG8^p^5?fQ5(W;6@p?X!Lr{iedDmiUe&Sh*E~CZR*x
z6MC?IR-PgLq;+a8p~H5oszTGDJl<_za81hEbhxHb5Oz}wA;yHZX*x6)fsFwT`$}2R
z!9aU10Gb0F_&jcuA-q1m0}Rz6JO9;0K*jekp=V~{_=Jv8m3DrJR8x`7>4;`|ze<y)
z13La~LWcpAjSp5uA(SD>-+0HUA`z(mU_dVnYOZQ2;sfG47;#UaA676^3l%nsK*9KZ
zH5*$#FWA7Ob1qq=*?^3A2+;@ZE99_!_2Hn5V=g3|$N&@@d^(LYYdptl<OyU&hLRrL
zK^XY=ITGPPzJvaGbA;51<rz6bDra2Dw_qWOEAHowqg!F;WI=2}%vN}UC)NQ9pMckV
zyfPLczRkbQRZm#+#@gDxjIRvn<K&4D$Ul&yvbqsoQ`024EmlTz#y&`r$q&oH_}UP)
zfDa+15L%FGaJ%S1mrQFzq1iypB0Ja+@)@xhik}_?cR41LXCTENy2&+6d|sbMwGWbX
z^zXg4nOANZ^K97~_|d_zD{m+xU3+pZ_pp9IF)*-YW2SM#y%De9gx8JqQ8}$Tte)Gb
z=mi<Fc%?*<>IM>-tGVZn8KJ0fh+OvrQ~?$Rvy8ZJ%PsNKGsGBoWm_BMsJGMgvk-`+
zg2#55wqNxV97@H+sJG9xBGtpED%N%<krAXDPBrB~9gCH{-Zg(RQy?Yrq`w6uyD<D+
z7)Fl#XFC!IN4W5CZtBq6_&sa0H=u!bYbj5R9CA^3>mcb5As6rty25{9PPux^a?A!a
z6SQTZ!3GFZ$O_U{4n$34xIY#4&e!8z|EU{Ub-_eytMx@N5_cLoSLIIplBBj;4Z!)C
zNo!wLG!|~Ss}%-{W`IN~y!*Nb)COBDRl%o}=3@4GnGg0#B>7>8MYp``+nD%~?cfy#
z90&^Ugx7%ICy>&xLooan?prD?VXGjfL8rMteQ0+JR9JO>?{ku1TK8(Mr?Puo#FP2@
zc`yDDX7oK)R8E<ePS%1C=N5NPUMp3fPX209M?6EZA%E_v+d{4$QQyjLqkLvnazD6n
z(+1y`c?E#SSQ!MB8&BJtvRgK>&ym<?n&)4C8f-Q5mdU4u1%O<TQl327XRifF#|5C&
zzjha?De6dB*D)az&ZjQNoEd4YmKS1M=VG=xs4>dbezM8w)vY^u3SJeyn=R@^8DS#H
z>poTnIIOQ3${lzfP`JX%5CAiRat>_s4TxjK3Q&rg4tc(t|HSuzHtEp$eRbbJgbPqF
z?f`WDynQM>0jPV-@s*;u*Tb&q;1$478ed6vH9uOeoD(E;%E>~Tu9dAGHUku&_PhhO
zHxY6Z!)}8)?7S7=bvWk=ML!={a+>S0uLn#Xi%kK@)gC`R#w!9yCh=cIq~ToYvIV%g
zj2A4Li|&5c9^X{p>?pvB*p|dMqDVOe@*uQaetn$Vh2p{Y5h|oAYSDSJYFWOU_wkmA
zz<D68w&p86T99XjK-lhG3Wra5V;M_D{09fCgd-!4)+-omA0nC&Qs&t?zS)`%T-$Qc
zi6|%DkbAXG^NoytXQ!Hrk~n|p2g=Y$nswwAKO!TH?}bR)X?V0O{`_0W94VHr^3_-k
z?Gc4%i#4pAof@2>h&N5@@Wr~qhFTk9rFqCVwhn=KjK3{&Ii8?R-!ReZ@$P~&G&W$d
z-&OjtKaNhK&v}Xsi^p2~skAsE^Oi!UQ;>x1vcPlh$Uc?Q3<wt~bN<(Mm93qNd^WWX
zwx&}SoThD@*~<t>@1NOH<?W<Eu{hZkr`!cf>FsHnMpDa-jpysWp%fAV*n`r&CITRt
z0@B*ETN@P=tdx}^I6SnBW7btJWa~w&EazOJfo4UYXG_Sy2kV;<OMDJisnXM4XgkhK
zAmz=Od1dxY*RDQ>a^&+j;IIIxCwQq!%@aS}FenOyl26n0Cm#JbR{L+z21pAK@#(SI
rzEb%&P{smhaxPJdoBuB#a2FurSBDX_UC+mefIl5g14N02ZPfn((V8X%
literal 0
HcmV?d00001
--
GitLab