From 7d9b9fa9d86d8f48a37fabeaec4a067d9c955f8b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E7=BD=97=E7=A5=A5?= <1366971433@qq.com>
Date: Fri, 24 May 2019 10:14:58 +0800
Subject: [PATCH] =?UTF-8?q?streaming=20=E6=95=B4=E5=90=88=20flume?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 code/spark/spark-streaming-flume/pom.xml   |   1 +
 notes/Spark_Streaming整合Flume.md        | 360 +++++++++++++++++++++
 notes/大数据应用常用打包方式.md |  52 ++-
 pictures/spark-streaming-flume-jar.png     | Bin 0 -> 21168 bytes
 4 files changed, 404 insertions(+), 9 deletions(-)
 create mode 100644 notes/Spark_Streaming整合Flume.md
 create mode 100644 pictures/spark-streaming-flume-jar.png

diff --git a/code/spark/spark-streaming-flume/pom.xml b/code/spark/spark-streaming-flume/pom.xml
index c8fe1e8..d0e8441 100644
--- a/code/spark/spark-streaming-flume/pom.xml
+++ b/code/spark/spark-streaming-flume/pom.xml
@@ -98,6 +98,7 @@
                     </execution>
                 </executions>
             </plugin>
+            <!--打包.scala文件需要配置此插件-->
             <plugin>
                 <groupId>org.scala-tools</groupId>
                 <artifactId>maven-scala-plugin</artifactId>
diff --git a/notes/Spark_Streaming整合Flume.md b/notes/Spark_Streaming整合Flume.md
new file mode 100644
index 0000000..eacbfd0
--- /dev/null
+++ b/notes/Spark_Streaming整合Flume.md
@@ -0,0 +1,360 @@
+# Spark Straming 整合 Flume
+
+<nav>
+<a href="#一简介">一、简介</a><br/>
+<a href="#二推送式方法">二、推送式方法</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#21-配置日志收集Flume">2.1 配置日志收集Flume</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#22-项目依赖">2.2 项目依赖</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#23-Spark-Streaming接收日志数据">2.3 Spark Streaming接收日志数据</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#24-项目打包">2.4 项目打包</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#25-启动服务和提交作业">2.5 启动服务和提交作业</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#26-测试">2.6 测试</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#27-注意事项">2.7 注意事项</a><br/>
+<a href="#三拉取式方法">三、拉取式方法</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#31--配置日志收集Flume">3.1  配置日志收集Flume</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#22-新增依赖">2.2 新增依赖</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#23-Spark-Streaming接收日志数据">2.3 Spark Streaming接收日志数据</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#24-启动测试">2.4 启动测试</a><br/>
+</nav>
+
+
+## 一、简介
+
+Apache Flume是一个分布式，高可用的数据收集系统，可以从不同的数据源收集数据，经过聚合后发送到分布式计算框架或者存储系统中。Spark Straming提供了以下两种方式用于Flume的整合。
+
+## 二、推送式方法
+
+在推送式方法(Flume-style Push-based Approach)中，Spark Streaming程序需要对某台服务器的某个端口进行监听，Flume通过`avro Sink`将数据源源不断推送到该端口。
+
+这里以日志文件内容为例，将不断新增的日志文件内容推送到Streaming程序中，具体整合方式如下：
+
+### 2.1 配置日志收集Flume
+
+新建配置`netcat-memory-avro.properties`，使用`tail`命令监听文件内容变化，然后将新的文件内容通过`avro sink`发送到hadoop001这台服务器的8888端口：
+
+```properties
+#指定agent的sources,sinks,channels
+a1.sources = s1
+a1.sinks = k1
+a1.channels = c1
+
+#配置sources属性
+a1.sources.s1.type = exec
+a1.sources.s1.command = tail -F /tmp/log.txt
+a1.sources.s1.shell = /bin/bash -c
+a1.sources.s1.channels = c1
+
+#配置sink
+a1.sinks.k1.type = avro
+a1.sinks.k1.hostname = hadoop001
+a1.sinks.k1.port = 8888
+a1.sinks.k1.batch-size = 1
+a1.sinks.k1.channel = c1
+
+#配置channel类型
+a1.channels.c1.type = memory
+a1.channels.c1.capacity = 1000
+a1.channels.c1.transactionCapacity = 100
+```
+
+### 2.2 项目依赖
+
+项目采用Maven工程进行构建，主要依赖为`spark-streaming`和`spark-streaming-flume`。
+
+```xml
+<properties>
+    <scala.version>2.11</scala.version>
+    <spark.version>2.4.0</spark.version>
+</properties>
+
+<dependencies>
+    <!-- Spark Streaming-->
+    <dependency>
+        <groupId>org.apache.spark</groupId>
+        <artifactId>spark-streaming_${scala.version}</artifactId>
+        <version>${spark.version}</version>
+    </dependency>
+    <!-- Spark Streaming整合Flume依赖-->
+    <dependency>
+        <groupId>org.apache.spark</groupId>
+        <artifactId>spark-streaming-flume_${scala.version}</artifactId>
+        <version>2.4.3</version>
+    </dependency>
+</dependencies>
+
+```
+
+### 2.3 Spark Streaming接收日志数据
+
+调用 FlumeUtils工具类的`createStream`方法，对hadoop001的8888端口进行监听，获取到流数据并进行打印：
+
+```scala
+import org.apache.spark.SparkConf
+import org.apache.spark.streaming.{Seconds, StreamingContext}
+import org.apache.spark.streaming.flume.FlumeUtils
+
+object PushBasedWordCount {
+    
+  def main(args: Array[String]): Unit = {
+    val sparkConf = new SparkConf()
+    val ssc = new StreamingContext(sparkConf, Seconds(5))
+    // 1.获取输入流
+    val flumeStream = FlumeUtils.createStream(ssc, "hadoop001", 8888)
+    // 2.打印输入流的数据
+    flumeStream.map(line => new String(line.event.getBody.array()).trim).print()
+
+    ssc.start()
+    ssc.awaitTermination()
+  }
+}
+```
+
+### 2.4 项目打包
+
+因为Spark安装目录下是不含有`spark-streaming-flume`依赖包的，所以在提交到集群运行时候必须提供该依赖包，你可以在提交命令中使用`--jar`指定上传到服务器的该依赖包，或者使用`--packages org.apache.spark:spark-streaming-flume_2.12:2.4.3`指定依赖包的完整名称，这样程序在启动时会先去中央仓库进行下载，这要求你的生产环境必须网络畅通。
+
+这里我采用的是第三种方式：使用`maven-shade-plugin`插件进行`ALL IN ONE`打包，把所有依赖的Jar一并打入最终包中。需要注意的是`spark-streaming`包在Spark安装目录的`jars`目录中已经提供，所以不需要打入。插件配置如下：
+
+> 关于大数据应用常用打包方式单独整理至：[大数据应用常用打包方式](https://github.com/heibaiying/BigData-Notes/blob/master/notes/大数据应用常用打包方式.md)
+>
+> 本项目完整源码见：[spark-streaming-flume](https://github.com/heibaiying/BigData-Notes/tree/master/code/spark/spark-streaming-flume)
+
+```xml
+<build>
+    <plugins>
+        <plugin>
+            <groupId>org.apache.maven.plugins</groupId>
+            <artifactId>maven-compiler-plugin</artifactId>
+            <configuration>
+                <source>8</source>
+                <target>8</target>
+            </configuration>
+        </plugin>
+        <!--使用shade进行打包-->
+        <plugin>
+            <groupId>org.apache.maven.plugins</groupId>
+            <artifactId>maven-shade-plugin</artifactId>
+            <configuration>
+                <createDependencyReducedPom>true</createDependencyReducedPom>
+                <filters>
+                    <filter>
+                        <artifact>*:*</artifact>
+                        <excludes>
+                            <exclude>META-INF/*.SF</exclude>
+                            <exclude>META-INF/*.sf</exclude>
+                            <exclude>META-INF/*.DSA</exclude>
+                            <exclude>META-INF/*.dsa</exclude>
+                            <exclude>META-INF/*.RSA</exclude>
+                            <exclude>META-INF/*.rsa</exclude>
+                            <exclude>META-INF/*.EC</exclude>
+                            <exclude>META-INF/*.ec</exclude>
+                            <exclude>META-INF/MSFTSIG.SF</exclude>
+                            <exclude>META-INF/MSFTSIG.RSA</exclude>
+                        </excludes>
+                    </filter>
+                </filters>
+                <artifactSet>
+                    <excludes>
+                        <exclude>org.apache.spark:spark-streaming_${scala.version}</exclude>
+                        <exclude>org.scala-lang:scala-library</exclude>
+                        <exclude>org.apache.commons:commons-lang3</exclude>
+                    </excludes>
+                </artifactSet>
+            </configuration>
+            <executions>
+                <execution>
+                    <phase>package</phase>
+                    <goals>
+                        <goal>shade</goal>
+                    </goals>
+                    <configuration>
+                        <transformers>
+                            <transformer                       implementation="org.apache.maven.plugins.shade.resource.ServicesResourceTransformer"/>
+                            <transformer                   implementation="org.apache.maven.plugins.shade.resource.ManifestResourceTransformer">
+                            </transformer>
+                        </transformers>
+                    </configuration>
+                </execution>
+            </executions>
+        </plugin>
+        <!--打包.scala文件需要配置此插件-->
+        <plugin>
+            <groupId>org.scala-tools</groupId>
+            <artifactId>maven-scala-plugin</artifactId>
+            <version>2.15.1</version>
+            <executions>
+                <execution>
+                    <id>scala-compile</id>
+                    <goals>
+                        <goal>compile</goal>
+                    </goals>
+                    <configuration>
+                        <includes>
+                            <include>**/*.scala</include>
+                        </includes>
+                    </configuration>
+                </execution>
+                <execution>
+                    <id>scala-test-compile</id>
+                    <goals>
+                        <goal>testCompile</goal>
+                    </goals>
+                </execution>
+            </executions>
+        </plugin>
+    </plugins>
+</build>
+```
+
+使用`mvn clean package`命令打包后会生产以下两个Jar包，提交`非original`开头的Jar即可。
+
+<div align="center"> <img src="https://github.com/heibaiying/BigData-Notes/blob/master/pictures/spark-streaming-flume-jar.png"/> </div>
+
+### 2.5 启动服务和提交作业
+
+ 启动Flume服务：
+
+```shell
+flume-ng agent \
+--conf conf \
+--conf-file /usr/app/apache-flume-1.6.0-cdh5.15.2-bin/examples/netcat-memory-avro.properties \
+--name a1 -Dflume.root.logger=INFO,console
+```
+
+提交Spark Streaming作业：
+
+```shell
+spark-submit \
+--class com.heibaiying.flume.PushBasedWordCount \
+--master local[4] \
+/usr/appjar/spark-streaming-flume-1.0.jar
+```
+
+### 2.6 测试
+
+这里使用`echo`命令模拟日志产生的场景，往日志文件中追加数据，然后查看程序的输出：
+
+<div align="center"> <img src="https://github.com/heibaiying/BigData-Notes/blob/master/pictures/spark-flume-input.png"/> </div>
+
+Spark Streaming程序成功接收到数据并打印输出：
+
+<div align="center"> <img src="https://github.com/heibaiying/BigData-Notes/blob/master/pictures/spark-flume-console.png"/> </div>
+
+### 2.7 注意事项
+
+#### 1. 启动顺序
+
+这里需要注意的，不论你先启动Spark程序还是Flume程序，由于两者的启动都需要一定的时间，此时先启动的程序会短暂地抛出端口拒绝连接的异常，此时不需要进行任何操作，等待两个程序都启动完成即可。
+
+<div align="center"> <img src="https://github.com/heibaiying/BigData-Notes/blob/master/pictures/flume-retry.png"/> </div>
+
+#### 2. 版本一致
+
+最好保证用于本地开发和编译的Scala版本和Spark的Scala版本一致，至少保证大版本一致，如都是`2.11`。
+
+<br/>
+
+## 三、拉取式方法
+
+拉取式方法(Pull-based Approach using a Custom Sink)是将数据推送到SparkSink接收器中，此时数据会保持缓冲状态，Spark Streaming定时从接收器中拉取数据。这种方式是基于事务的，即只有在Spark Streaming接收和复制数据完成后，才会删除缓冲的数据。与第一种方式相比，具有更强的可靠性和容错保证。整合步骤如下：
+
+### 3.1  配置日志收集Flume
+
+新建Flume配置文件`netcat-memory-sparkSink.properties`，配置和上面基本一致，只是把`a1.sinks.k1.type`的属性修改为`org.apache.spark.streaming.flume.sink.SparkSink`，即采用Spark接收器。
+
+```properties
+#指定agent的sources,sinks,channels
+a1.sources = s1
+a1.sinks = k1
+a1.channels = c1
+
+#配置sources属性
+a1.sources.s1.type = exec
+a1.sources.s1.command = tail -F /tmp/log.txt
+a1.sources.s1.shell = /bin/bash -c
+a1.sources.s1.channels = c1
+
+#配置sink
+a1.sinks.k1.type = org.apache.spark.streaming.flume.sink.SparkSink
+a1.sinks.k1.hostname = hadoop001
+a1.sinks.k1.port = 8888
+a1.sinks.k1.batch-size = 1
+a1.sinks.k1.channel = c1
+
+#配置channel类型
+a1.channels.c1.type = memory
+a1.channels.c1.capacity = 1000
+a1.channels.c1.transactionCapacity = 100
+```
+
+### 2.2 新增依赖
+
+使用拉取式方法需要额外添加以下两个依赖：
+
+```xml
+<dependency>
+    <groupId>org.scala-lang</groupId>
+    <artifactId>scala-library</artifactId>
+    <version>2.12.8</version>
+</dependency>
+<dependency>
+    <groupId>org.apache.commons</groupId>
+    <artifactId>commons-lang3</artifactId>
+    <version>3.5</version>
+</dependency>
+```
+
+注意：添加这两个依赖只是为了本地开发测试，Spark的安装目录下已经提供了这两个依赖，所以在最终打包时需要进行排除。
+
+### 2.3 Spark Streaming接收日志数据
+
+这里和上面推送式方法的代码基本相同，只是将调用方法改为`createPollingStream`。
+
+```scala
+import org.apache.spark.SparkConf
+import org.apache.spark.streaming.{Seconds, StreamingContext}
+import org.apache.spark.streaming.flume.FlumeUtils
+
+object PullBasedWordCount {
+
+  def main(args: Array[String]): Unit = {
+
+    val sparkConf = new SparkConf()
+    val ssc = new StreamingContext(sparkConf, Seconds(5))
+    // 1.获取输入流
+    val flumeStream = FlumeUtils.createPollingStream(ssc, "hadoop001", 8888)
+    // 2.打印输入流中的数据
+    flumeStream.map(line => new String(line.event.getBody.array()).trim).print()
+    ssc.start()
+    ssc.awaitTermination()
+  }
+}
+```
+
+### 2.4 启动测试
+
+启动和提交作业流程与上面相同，这里给出执行脚本，过程不再赘述。
+
+启动Flume进行日志收集：
+
+```shell
+flume-ng agent \
+--conf conf \
+--conf-file /usr/app/apache-flume-1.6.0-cdh5.15.2-bin/examples/netcat-memory-sparkSink.properties \
+--name a1 -Dflume.root.logger=INFO,console
+```
+
+提交Spark Streaming作业：
+
+```shel
+spark-submit \
+--class com.heibaiying.flume.PullBasedWordCount \
+--master local[4] \
+/usr/appjar/spark-streaming-flume-1.0.jar
+```
+
+
+
+## 参考资料
+
+1. [streaming-flume-integration](https://spark.apache.org/docs/latest/streaming-flume-integration.html)
\ No newline at end of file
diff --git a/notes/大数据应用常用打包方式.md b/notes/大数据应用常用打包方式.md
index c9133a0..f05e267 100644
--- a/notes/大数据应用常用打包方式.md
+++ b/notes/大数据应用常用打包方式.md
@@ -5,12 +5,15 @@
 <a href="#二mvn-package">二、mvn package</a><br/>
 <a href="#三maven-assembly-plugin插件">三、maven-assembly-plugin插件</a><br/>
 <a href="#四maven-shade-plugin插件">四、maven-shade-plugin插件</a><br/>
-<a href="#五使用非Maven仓库中的Jar">五、使用非Maven仓库中的Jar</a><br/>
-<a href="#六排除集群中已经存在的Jar">六、排除集群中已经存在的Jar</a><br/>
-<a href="#七使用建议">七、使用建议</a><br/>
+<a href="#五其他打包需求">五、其他打包需求</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#1-使用非Maven仓库中的Jar">1. 使用非Maven仓库中的Jar</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#2-排除集群中已经存在的Jar">2. 排除集群中已经存在的Jar</a><br/>
+&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<a href="#3-打包scala文件">3. 打包.scala文件</a><br/>
 </nav>
 
 
+
+
 ## 一、简介
 
 在提交大数据作业到集群中运行时，通常都需要先将项目打成Jar包。Java项目通常都采用Maven进行构建，Maven提供的常用打包方式及插件有以下四种：
@@ -118,6 +121,8 @@ assembly.xml文件内容如下：
 
 `maven-shade-plugin`比`maven-assembly-plugin`功能更为强大，比如你的工程依赖很多的JAR包，而被依赖的JAR又会依赖其他的JAR包，这样,当工程中依赖到不同的版本的 JAR时，并且JAR中具有相同名称的资源文件时，shade插件会尝试将所有资源文件打包在一起时，而不是和assembly一样执行覆盖操作。
 
+**通常使用`maven-shade-plugin`就能够完成大多数的打包需求，其配置简单且适用性最广，因此建议优先使用此方式。**
+
 ### 4.1  基本配置
 
 采用`maven-shade-plugin`进行打包时候，配置示例如下：
@@ -192,7 +197,9 @@ assembly.xml文件内容如下：
 
 
 
-## 五、使用非Maven仓库中的Jar
+## 五、其他打包需求
+
+### 1. 使用非Maven仓库中的Jar
 
 通常上面两种打包能够满足大多数的使用场景。但是如果你想把某些没有被Maven管理Jar包打入到最终的Jar中，比如你在`resources/lib`下引入的其他非Maven仓库中的Jar，此时可以使用`maven-jar-plugin`和`maven-dependency-plugin`插件将其打入最终的Jar中。
 
@@ -238,9 +245,9 @@ assembly.xml文件内容如下：
 </build>
 ```
 
-## 六、排除集群中已经存在的Jar
+### 2. 排除集群中已经存在的Jar
 
-为了避免冲突通常官方文档通常都会建议你排除集群中已经提供的Jar包，如下：
+通常为了避免冲突，官方文档都会建议你排除集群中已经提供的Jar包，如下：
 
 Spark 官方文档 Submitting Applications 章节: 
 
@@ -250,14 +257,41 @@ Strom官方文档 Running Topologies on a Production Cluster 章节：
 
 >Then run mvn assembly:assembly to get an appropriately packaged jar. Make sure you exclude the Storm jars since the cluster already has Storm on the classpath.
 
-
+排除Jar包的方式主要有以下两种：
 
 + 对需要排除的Jar包依赖添加`<scope>provided</scope>`标签，此时该Jar包会被排除，但是不建议使用这种方式，因为此时你在本地运行也无法使用该Jar包；
 + 建议直接在`maven-assembly-plugin`或`maven-shade-plugin`的配置文件中使用`<exclude>`进行排除。
 
-## 七、使用建议
+### 3. 打包.scala文件
 
-通常使用`maven-shade-plugin`就能够完成大多数的打包需求，其配置简单且适用性最广，因此建议使用此方式。
+如果你使用到scala语言进行编程，此时需要特别注意 ：默认情况下Maven是不会把`.scala`文件打入最终的Jar中，需要额外添加`maven-scala-plugin`插件，常用配置如下：
+
+```xml
+<plugin>
+    <groupId>org.scala-tools</groupId>
+    <artifactId>maven-scala-plugin</artifactId>
+    <version>2.15.1</version>
+    <executions>
+        <execution>
+            <id>scala-compile</id>
+            <goals>
+                <goal>compile</goal>
+            </goals>
+            <configuration>
+                <includes>
+                    <include>**/*.scala</include>
+                </includes>
+            </configuration>
+        </execution>
+        <execution>
+            <id>scala-test-compile</id>
+            <goals>
+                <goal>testCompile</goal>
+            </goals>
+        </execution>
+    </executions>
+</plugin>
+```
 
 
 
diff --git a/pictures/spark-streaming-flume-jar.png b/pictures/spark-streaming-flume-jar.png
new file mode 100644
index 0000000000000000000000000000000000000000..a4e8bbe5df6b8e1637096a58d3de7f97dc0486ad
GIT binary patch
literal 21168
zcmdSBWl&sA*sYBQcMnd2yE}y7?(XjH?rtGy&|txJaCd^cyX)YtXOic6t?JZU=g0Tw
zt3sxxW_s`L-Mjl*Yuz)E3UcB|@E_s9z`&3sB}A0Kz#ue0k8@x@fc}zW7CnOg!8u51
zI)i~B_x<?=Po_o22R(`ABC6q{Y;We`Zs=qRrfg^FVCd{(s*5we1_nj~CMhDM;*ojU
z?(c%Xm-YVq9T@T%)sIck1ph0wD}-!u5Iy#tHL(VjsC6qDC5+c51aRuIvc}%KgV215
z5r=XAu2v^ADsiT=!mn`+3c$S_diPDZxjC|?{TDQL2Xj)iO*uMbjU*)0P}C3vb0_*r
z`tx8+K?2(+>Dq0st!p0HvC-?%YoBa9%l(Rp1J(&+WmQ#C)PSKONr*tO1$}>Ct))On
z7%+`SyEQ&-@=U9*xu|=4dqvBLTc1f6?nlKc&sWn%9%-qo>3b|kul7b{`>`Ub(V>J;
zLv*Nk=03m%k$`?OEhD=AJO?F2#pDc&7IzUHw&8p^f!U$YBRzMCWAPYn&K(g9<F<~>
zyS(VG%6qS}x6w)hix3Kz{d8o%H5lpLP3-&{haw;;eaP$5qF$KVe@?7sai46>bi})U
z2Gwcd5p;kgzIK7p3}f2AnYj8g0!#F&_dZA{u)cxWI!`?j*iG5|nA&~AF~8bZ#VP(U
zWxZi$KSKj*-nrb9%5f|tw0FynxOF%=v!j*syy5yJyz(F~(Q?sJA{)Z`vm@Xsoa2SM
zFm=5pRPLSz+5P!p>YU7iv&#)Q!uAG1&&b;Zuw*bp2&y)~-I$Q(YI2K@FXEtn*|QcP
z^276JrYE(SUoBV*P$uqr2Qu7oR2;-ti8i?<Mm9;6q>e+$zuq)I%x~ExcEC;d$oj02
z^u0CZ$NGArJkW~G0bicfhy|bK5MK6#7x3Zh4!T+!*ZXb5TD;x;GW6MABGa=PoyY$4
zdAYIPn)6||ErS8J=zAu#nhe|8MWWl?FPhaqLyWar_730vyP|r*whdhFVz#PIKM#fT
z^ZRgQB#RKL93yfkc1YQ|fh^VU4t{|b#z8-owQ=?*;k2=*s+H1}rPgaVU>B&JB}k^8
zj?_<l?@q(<?9z&+W1@DUFki8W$~Q#enU<?4mfLln&r!C$7n~=?@03jvNa%c=&^5wa
zeP0e$VLI16`H4?2E>)UI`O@|&GB-<d+&2C78RHR3P22?zJ1Is)b>MoA24P$Ma00dK
zE2256Pi?2~-t%%}!)taPHN6Mef_m3F57x9BtJU8Flbt=LLJs_YTo~wjow(n;!&AIg
zw{Dfbth|?A3=-O=i!&vwFeB=Co#^U3A7fm4zg=HC6YlP(ZiT(?b48}WKN@)YT1vwg
z3D;umIyA=Ht|n|_PGvuy+onu<$7DjKT&j<X%5IqI3*&b_Nb<5*)n78X*W8otRE??n
zyyv=gOi1+Tj>5{jR&9So#5pW|ec8W|U8N-z7UBYOt-gA%>l(+~cQy_VjxmaZ@n+|&
z^&`DsK)q>v3@i~phOKJ*xJB~JecF*7owoVf>&CF2qy*ffY+AYz0^8gSOya+MDeb1T
zgu^?Qj9{Xyw6x-c-KX@ppQm5NJDwyOHd~lk8^1SBhzY;>R!u7M@M3)}_@O(!?^c%S
z86^|V?R9)Si&r^|tnA?}-ac=SID0pqm%I~sKWVokd&RE)I$ysbP_v0?U^B-fn`DCX
zQ1bg(($)0Ikq?zWEtC+Kt4}~MZx!+tzAx{7H_s|?NBsT5n@`Z%oDqY7LATJeVGMoV
zz8ztr%zqnllZ?Ux1*WmktF3mlJIe5^8{KKIB&Xqsh?+kBj%d@NF$>=>&jCf=$+;IJ
za=6B)g@2-FPJ0?7yUcDO2Lt2Ur2QK&Mw>rwwt0wm>1gkD`rsq-88a<idp7U8Ha^Pv
zgWLGlD{Dd<lv=6wlm{zhKCNR=kzO$VJXf*j;mYar5dy*d#^*;@?g(i`R5|-fNBxN{
zJ7WRUo|zi=qPx72vCdC#A3lDQ@F4aNOyMs(a;Z9C>NMBBhr+q?UT=Sj<L-<#IqUw8
zYFKHZ7-1p++ug4YUKr-nf*UMvHHA3iqt!d2Xl=K>K-q6(D2FHz{+`$B5wLMR`|xx#
z`SyTd<2~2ij-pv~qch`z!^y_i&1mu~U=>PDMnkt{V7##L*MPn&B4-_F-tHA%ojU=y
z1i??54@e(Gg+pN6D#On<@<Y+iw{egJvWLxDV1!VyxxBWo9}w!;_z)WH4A12}j*A4c
zpK89pR5P?^%klF$COo(LwC4sB8xCFrUapyU1mBc4$Pl^FdywJK%%ppQlY=ag=dpR#
z8Y#tDOYII8NQpafbtd>#m(6=;uSw~^#2)$?KY94M@h&)oY@Fh4JyYuy3!cF_0ppW3
z{lJc<D%Kbs<mYZFr`AK|<sfLWtLCnX%npLTW$|XHUA~q5^eoE`$g!AVCx~0nn&20F
z#Nd2ChVFQuylqbV1lt3#olX&IZN%8<fYG_WRJpgw2E`t)M9pN!)n)UR$&UQyge+KB
z=bePjp>u7_+xZv{>pt)c!JvM0Ch?OcQ8{Rx_H+-C-e|NwnA^p@?mJ2<unb<r!iIQS
z=@5);|K`MhpYq(~2_u;+0OS?m9q<8Gv3MRB5WeWbaK4YX(WP8F$j$(}pq5cc=#SQt
zN>?hJn?Un9c2eIC&*ysv{&HPg<}B0n^TbttnE4KhW{dACLB>-kWpi_PDuGf6!S_{+
z#B4of_=T5UNxSt+i0oc8KEaN)QgA~KqwCa)=16@(7hatH4Rl$5T0VA1o`Xo*&_~hE
zh7yc(y<IgA195y#&H{a7Qr;_Ujg)IO@;$yZgEWhau$<qsh(QiwXwQBbaI^smA;AL#
z<zXEBF}STbtK;ig7<{)g9CjVQzb$5)et#x*KjhA0c2sla_4fRpa>aW{NA7uL?CI>O
zwoTf@-XhMThmLUHAA(&y(2WG;xlUJLakVUHZD({YypTiC8(RLb6URX$km+TMp))f(
zzAn3bf3sC^z9}(144b-cb)^&>Agii+tAbAo+tWji39qQ9=T|z0Du?dmJ7Kf?G#=ZW
zDX4j>3EHkvp>n(aYwISf{=0{x!pFeU{n^Y>Lm)muNdGS{8x1D^3jhiRegFAL!~pIu
z6JX+^Y5aWv3oQnf@`n;&zXqZO{>K~sFE9J`c>|I%I72}dtM1*>-&olS?@e1Z*JFr0
z^B(BW5YYTjYX~`Q&?%**nciDKiz}K^YU;e@hV+;F!&Ywu<BGqB>9PFsM%t6MVukHk
zn2gh<^iNl2LO}2P^`jO|eTwf^m2s@W%<Gz60n|AvCbXw^__s~Exgfu`*MvIX;c}R-
z1slBH&p{3!zqv=CCYSdmXMa<L)rJ<*h=6*f(gpX;CvexMW$<o8?VJ|vs-sgRa%-Hj
zqKmlMu|20S@zg1oOY_Qs@$O8$J5$ITU5c34sV+bNy=+DqD}X{xUziPc&0LU$f89Wa
z-B1LK$9a|@Tz)YeoLi32znr#rlXVFaHY~NO@A<hg;$cf$=~<_QiVfjB+#N(URUGv(
zE4K;P*-=#GCYdC&2I}%YX=?x45Jpy^?wR`5(*ZSiqYMruXT(lvUe5h@<c#ICnK!3t
ztbi7wR|fC%(&BN?17AEMN>D2}3@Tzc6(<JQuUm6~XM|J^3JU5Ct;-|rOn>^kQ5f?f
z@Z)n$?MRPXYZ+FKV)7wGjT2G7YNvX)$c|+ltqXqR$XZ=}OSSmZ5uEze&n%s=9@W|9
zzz9ivNr>5qugB}kv~<v<?x4FwT%a>4A)(6g)VU&ZPg?%sRy9$l?rH`zm;6`?bzFzQ
zLkmfsUz}Xd3pYL>*zkj}+_ARn)@E_P?)tdGWN<HNT#kF8p<cmz1}K7sp@!G$`qN!!
zx6hV_9Qu-XO^;8)N*h9`sA<PO8^_!N=Q&n@O<Me6WRy2C7~O_s-&tY;VPE+TSCO(~
zwrO}*ZqaTm^PC0P#cX^M{1|-;D~=#24+nYmGCm`J>J1JVK6bt25h1}O`Po;X0Vb+0
zmtK6cX<bHc)HpbytfikIq0~&0$*`Loe7m1v?{qkG?qp;JH<CiPe_9yL&FstZ>AY%J
zK3wfv%8NkR@8iNHv)*6tWk&?kRV?$8vAJ^2RW-#F9UEr9S8A7YN(4`gp@h<sH&MZd
zsp=sUm`KeFg-{o%>4Ptte$X0od<}^WhYp8LQm>gpMm6L^3}k>7RaC$fw4~;fyk%X?
zRKxt)W@UVaa{rqtSVc&Gl!81V5%XxU%wviJP)$g3SVVjJ>kiY?;S|3N7Nt-dn*d%p
z*+nV%re9aD<+siC*7f5BJ5T*Zk^G8WM6ID44SZCiVE+(SUlkP<?{mcNbpl$5vdlP9
zxdYeeRxc?wmZ)ZJQ`&bU*?iel5$Ycdq%TuW3|!Bw3|fwe5qV-o3aSc4g8gbvagx#P
zY%9l}`No*_g+0~&vlko2E_RTerApwFD=`%vKgN#Tjt6)yZJL9KT!~!=%qs3?qya#~
z`S{XeEUy)IYfHVNTHlH}<{*+$sRCkqb+Do{)5w_cz~R-Enp8!i<6n%2(`~rj0xiM@
zF+I@X5L_q5xs4%|lB~Vbt&RE!>ObZi8@M~<C*GX~dXk#a;V|H#smkuRuKSi(CEo{T
zbqbbj7eb;!U#~-bKdEkI-G~#wF&ZgRF*THhSI<atTaW?Dv*(v$8d@#b`@Ujm4F`&&
z%)$2@NsL3BwBI8Gyy~;iS6HN^QWTYJAiwkv$Y9FJ$X>rqF*aND4+acV?(|}&(rqO+
zz9n2&EvBp+oZGpnxHF}Hu2d@dY3d!lue;3h^W1tmdfpLQySTAkXz);AdrSj(zBY#Z
z!Js0RK8&aoB}OQ#AJl&n5$WF_+%p*8z#39h63{tEkXxhY2biQ%v%{6o27q)hJHh;$
z1?$^G*6wpVr)QEOnDoaPO9bco{+lqb@KW)-q8BET&D}dD?uzH9zs$LUMw02MSeB8~
z72vo_1PdR<-E@#tJ$9NMgkFHbD)sdg2mOko!t|%7!o2Z-;{2cb-E7b!pJ8aTci2-6
zVSD(#1G>^P`36+gRTXt~Y_$AN7b)A&;d4tBZHhA=u0E-im+T)tL;ku27?mfc*K7YY
z##F7CY!tug;b|zNS;5N(420SK)WQ+Reg-e)#a&zY+TvGQZ9(fl1HC0fd8mGvhn6!T
z0v;0kD6Pd+Ud|FNUeFsnf#b~sb4hEea-5N{CG>UtsoXhX^w_+PdniUMQz~k6=!5r<
z{eEh&?hZSL#xd)&WNJ5@id#QS(eC3hx$qI_1V7tewm|-b;6gZxwO$;`hz(vS)npHy
zBA>n>CQ7IAZW<bz0W*MRig>=FO+%k{NNK5(j+2@ZUY5!+#|&F_sO_amksgn-bG4W_
zZt+=H6>V0Hxf@!lwQ-iq9fZnct$kA5F}Ro*>#6(+idV+14s99eu4WRyM`llNU*9Y4
zm*w=WnawgFUfazU$GmYzb6Yru_|$V@0Bb&7=8{lAkoB{ZmRgXu!jB^3)2jzkB9rKo
zce!&Ti{XX@*4ioCuI??=aJ^4q5&@m{3V0A=f1I{L!EWD=t&6_viczO(AItS(62HRm
z+7Bq%)4ApQqv#=_f9_i63I6cNNR-HNf}*D19~L9TZY{;<@C2Su5yo|&pgHi<e=8yR
z%a#~FTfnee+gj9CL=w63N^)zs-cMg%Xw5Iz{nT?+#+a+Ajl!O{1sIksG-h}yoPd?z
z72Uze^3$xaHh^|=YS5wu;X*x>4G&EA;;sC^x$Xunx+0y=n1L#9^8{&2YY86=TAi9K
zC8orBd68z|aD^q!ZdsR4^KwU{^IZDx%q*R0mVFcr3$qpxT*k%tF@KQB(AmO7#9;i2
zvB~tKO!T0dSi-%h`v98GH~_i$1m9=5kG>V@-x|Cxw!{@%OuRm9)%E|<YDpQ56d{8O
zwpYj_GCryWP&3pT0=jQ*5S3Xk`cqkbyvIm~#|1wW#}L2BOEJFE0?+H;9lwWx_((nQ
zr+#h`-7}0=<!z{m(Y1joI`<`<BGR>i=d+ZhEI-MiIG$9tG9AS#OMTAgGZ9O;A~bFU
zx`3J*cP_4$oc?GWMU?_ky!<IVIhhh4-#w_*<j_djIi|O9=_J2vE#SH9AFBF|!yl4l
z_~Ej74hIAr1fFno0gw}%4!a?V6Xt5lCc)&Q;u_6|t_&J49SuClhyJQIi`B`QvF%Ss
zJz9)^5vv+(Nc&&1g5BR`PXyT~e|X3s_J2h+Zh5wk#w<;e8Ch!&0|s>Z>lYiP@cq&Z
zTW$95&fue1e|b;iVVAyIMRji#!Jd+11FQ56hrL5a?(}@h{T_Qj>tA{bco7a^Zk(gr
zmZD`0DEtS-ir_pE=zMRx>H1VvZd5;lmVGM;-I<R8HF8$v!=R?+sQ-`^pt6_XW(OEw
z030y`Kal8jzq^CpVk0?6&25OdKHZj*|4kE85kx{K66BXux!5zb*OZv@3vOGz1vV{(
ze7Ls+;e+_oWrI6LEgyp#rA+%oFn<G)9wNZ?1o4(obLF&P>VWhb-UHiMf|qmOR+P?0
zeubD(AUT@YI$Glyi}4jNg#;csw4`(nKG=_Or@o?Te@aS719p@e1|d%*pT^21Y};tG
zA4_pcy;)(ZN1%RBMVk~>6TNRkSuA5BoRCdu>a1w?pj#-!uZn`thoMzclXs1jQm#lc
z1GtvN{$$~LXnS^QZd``2bIr%X2YdE`;0qH$-m1)I+I-H~J}Hf_0N)`3(laeRHr_bf
zhjv|KK{YqcMLYNUZPrqBT@tnN{H7@SI~%FiuZ@l?)r-#ZO;I0!yDXTMdGp93GRqR^
zKgBSW?z6QFsbQNF8@qU}9t&e)d}57rtYzu+y@IN$(s%QcbAG|nE2;-M>~D9?Sa}?C
z(>PA|dO#5%vf?@^ZImLwh|`~4xrGH05g_wcsu44g9HO9daNrS;DV+VHpD)kHcAt<7
z`&9XvQCXi;rJh_wQqz(G&$!&av0=#$gkQ{<+_~nNu*et*Nrs3bhv#rRb{8g?>Pt(N
z@jvA;vaU;*jdr4MypL#YWpTKK1lKQ@KRe>$5q(xstBPJAg~$a1C#UF#plWfN7Ab_$
z*QQ5n<@d{;eu@(LJ`BC(LRv*o&bi@}D4gS$ioq(a(O|-C&Zs*Yc?7`o^e6Y{sSB#x
zy^kT2o)A9<77Up6ngOO(5pFn)XBw?z;3SPBy)Ii~rZf|E?8^Zkv}GxcF9mEUfJeUW
zS~sR)mhW)*>E5>;W<Y!oim594`AH2*_CMrLl$cBy#DoGB4yzzmLKqPO$4!eA)Fo$$
z7cw}eg9kt(v*DxT68YH{73&rNyJHCJLQNGFl=mPhy(L+i)4gAp2ezk*@9f_s9)?!{
z|765L9hq4mA*k0M-S8t62qq+`Ysgt&^gGyiLqpVqYOQutA+59^@nZqw+10eDSq4hZ
zM0nLyNvNRjvfI}rB!IG_Lfb2~TqTL0sL;}q_~)GX7?MmQCv1%^hqT^sVLF<qS`$a+
zkrYh8O}SAD1GUl7ck~`&`h$yb^TGb)k6B+;g|gn*gv%gpJaJ!2qn2m-;=J%1)-n`T
zv0UpQ8*GFVi9rlDbJBmpF(VC4kHW@WeS)Ve6LUFQ37$@5ksG9!Zi@o&QxOy++-uOa
zM(VK$p#L<`!H0)FTJa=bo$tq)0cuIV{M>8(5O?lbZD~QtaM}JgE?oI3Ya~qxh#KNa
zG9BMY8vsxoVU3CjH|Oue^v2uEwR<v}xHiz&N~7Vz?9wDgz_ZFOSPX6*Bm70XXPBk4
zvAq8vpo@v5`ot1ONFNyj7uo`+e(YA?oa;?l;<>T3SP6%P@O*&qELPgXKXiAJ`z604
z>gFERjfWaW)+A+4d*70L*=_3hddpgtFb6UoG^lgI!mH{%mJ;KnJ@@OpqsMb@QZ^%0
z)S6<FHArAs_^gUt%<N;&;+?yBSj3PtIWZ{8I<+D4>rdk2okkg=WpgeMuA~Z$KPc7>
zUfW%@6#IE94RBxvepcoMr2y&EDa}U${A?2MitJcrZ&9vCH4y0edMg7dh>9V-+yTMh
z8rH_qHG`1G3PtyX=O^B)YVFchxsmyV%+F%?Q{PM87u6=*j*al!TbC5qx-FuA^wk_M
zd|5Y^su4z&W6}D`(X`@o2~*l$hj=xm0m7JA!z2cNl%mzx70o!(FcJAK$@anH3$M&I
z{>X26qQ0ydJFh20@<N?K)A<M3dpvPY!Gl9*dFxBE6An<Rv{P^84Z##DU*(SFE#DFd
z)rGOvJs5OQ{hvuxZ7YU69aW|j5O|*_$MgHyqtI?>(|#HJ*fr>PZrf);6iO2ZO}@m$
z!KUL<nyBb#*^zj{=MeR;q$=!DNXZ!betl8wa|FKrU3GqPF|JAgHw2fMyi%;Ey9hmv
z!Tn!eM1E#5=QRjCFK~gNjuw|tnzIR;yV|I$Hkhxy50T_BOMX#E(+~0sW%zjcD1-TB
zF|D&8w0D>W;32}p_ux}g<c!*>Bv#mg2U<jS%O`knhRXfKG7?8E`br95vu&zVE%h=X
zP)|=jxUaPB>-{itpVB%PSHvt|KXqx+K|NaFkil5Q;MPbtxZ09NHxznVCoL}R+u_;!
zgdM`2;2cq1T#TDBW~NCLD^|0FZjV-4D&;t&8ENXW-*X5HIY;>58b7(n+R8NC(BjQ&
zu)xs(3|i?Zl8$c~V?(kr%pmodU&$nxr&W2kEVx~hHjl&5$+}^N>(QBP7mhWna_6>k
zOhiBabkCRfvopH7RSgFhEeuAOD^ag7`t#l))T^%Kqo(C8>k`4ypmkaZ_4I~uR%<dG
zXp$eA6_K8oQYuvL(H8C+Xs#fmnozl|29c3djw6Os5<RhvbW5SYeYm9MPh$7DNJ2Jc
zH%}p<GjX~z$nrkpmXYz1aM|bT7v<;@V=I;CEgipDGS_!Fn`W$DniNT~df-(*J}cbW
zsxPm%46Hf+MNt~NhqbZ8(NMh?MKs%FT@(k*C!8&>M@tNq<)_Xr`Ffk<_eE+$_NI7c
z=#*P}%OqYBb^q&UnFuglsoJ8i-i7ONW!<rjx<XTT(}yDzeHa;UUsj~jdq`MDN$fr<
z)lFT<^}0X|D16s^jp1mupt{bZ(yeLNT+WQ*&{xs0q`ZU30w|)0{({AjUI8cD38DUk
zU$imwnP}LVV}L*U&#dk*pD{Ms=tvaPFQX^`68StRCNccT7UHt4DdBEOAxnVm-o#HT
z9@J<#wTX(hS+rr|=kJ|G^ts$a+|YB5sJ)H{r`dsdPn?#1mwNUF`ezmx!UkQal7N&3
z^~hUbwzV(mk3~L>A0l!-UsZbfG}b$mw3Ja`pA^HsA7DnqHOc@fBY73IdMri8ArK5~
zP6?6Y&de{jMKQ4~6$dO_!Z4I+^!2&BO-&7rl-XlCuh9!P+SH!Cv;Gj-l_(0SX_}Ca
z-7T~bD!jNJ1j1(5H4*{$*6#I>Im{&k$tg572Cpmx?9Qr+ILIe`EPM_da&zM<zn1EG
zguiq}_XXsg{KHv6+@52*41al#MB{&T<YdM%Vh7R??WWDlIXjqfM^ZUo8cV-kewXGt
zUwPL7a|WFZh|91AK-{LP%6;Fn5;^7`TKX;;%b1Aq?0vJ=9$KUJ*Y5)X+B>S@s4aVR
z)ATTcLCL?y+ShdmX?<TN1V=30&8N~-vkbZ<I@lm4-v9q_-jwFe9Bvl7HJ;}eQ*Ygl
zdCWW{Ayg;Ih|Z!IjvN$Ti=ezU`I3p?VAxkZS&s)Dx5Qev7Gb$exkQeYR%+{;spH9&
z)uo96HJpJ1PZJN+RV1ze11^!~A6?89f*E7NDo;KhlQ<vm)V1hXGA$9kS_k~fPiw2!
z&aC>)RmfYulxK=NVric)CAo##?KBwH9ArCSAuSJ?Qt60R5JsfcC4Bx8OI2{jU*Dqy
zIxG7?*WGJrJiNz`g8~cb8XY%K`uglcqO^jtXj<Ft=pn2)FxhIo$`gu_i9`dRHVx9c
zza>7Y>2h&dIi@zUV%}^zEIQUdIe-iK2UOP#ZK4J`aoX0H5!N?(yRi^Bl6(GmYRN67
zCzTqOVjaRY_HC^na|7{VeC}^J{v8pAw1@&kW0J4_^0@B7l=)4=<8HoCzUhaV^08?6
z`UM-(`j*x$U2spAaSIfYy}vI=a%;VG+mO*<p<$_*sj<IIB_~qQ@INeDMJ@s_uaoB0
zKYv#NC6@@Edj{=~`xyjJH?5gznmx7T(cIogMshMaW51tiA7^s03dlRnQ8mXJ>TU1>
z2yZ!T*t255we*sZGTy3wIn(L+mjnKIXx6_{^`bdOtL+vh!P&!j#@Yw+W;<BVCu|9V
z96q2DkbS``bR$8h*ez#+h2SkTMJ3{NHXRj<sJ^TQjl&fU$1*rsOgQOe;ZG{~sBKF6
zP|6a&4;eF25J~E`qpw>;cxK;d4x^U}K0AFUd-cF#TsFgQfP{K=30<4bCPRtyHOAV?
zbs70%)P9$<`qt4jB8~IIR=`JqxI?T6`|ZYZ1x}G;R8Car@~()hImmw=IqE-Z0UofR
zP0fiDd<0m69HNGng0YapxqZ0$^75#pv2m1@$}Q`TzEhFUkcS9rjYFIJ4~Wvnfh*1)
zVwS=m-xMuMjZ0eFxAF`>6<p|nOzdz?!O}dnFLW;pZE~j1%W{0$l|RMT!kN)EQs=|)
z{vV9E(dPjxf1(}KkePG|!SzZoxwR2=JQT>}=0;=2e7;#({mDX4<{)2}@Gjg?1Y~H3
z`N6ODaPr^H!cyJrJFzQSD(r2YeR8tj_hLe+oypNDFyR%{&U+6c>zfnbVoRDO04x6u
z;&_kZxzP3c3KR@%)KLtJ5}hVR%ryBc0C=|UQ&th76m?0E@kFIqAu|b?(DUBsgl4x^
zyyOk0NlA411&bd7qGEqgWu~R?zvk;ZO-03nC(93nQ!{AxP};wKI?%7Ny+i&_3Nv)S
z4P-9X9>lWAUBCN4nExy>EjbiBX;Wq%?pC1Y?!w6na-ix3V2l5UCr9_xu&n>h5w_<j
zFy$N48HfH0xD6xyCvv+K!vBe5$OACmAR6mWTsSKoTv|(nrPi9|1aPl+zICBIVI>i{
zgN&)~{}abF6-?>~*>cgHl>DWx<4>DDmR<}vZv{Q9z?4{7%Z^9qM}=uEE`_@hnJbEe
zHa|=F6iH+?zCmk-UJzY>HV<Uy_KYd3EX`|bTC96ntdaL1$NZ|Azo@`{xhGz(p*%Kq
z4<mDup|sY0*nvImDh|D(nhXT#BI8~T1zS>u;~TFJV#+6JO!Jas%k-1XQKf2!E1#%s
zf8E*EbdrEBf2jOq)uiYrft*MPSp4hMItL9+Q|x$&aF`s%I~U|__2J@K&OZ)$=B?%d
z*^jvi3+L~6v0TZR-HAwdEMw8iU_riXn|2ustF+3e3|c3zAh|S`k%gR!K=!&EW(na<
z1|y9j(-O`ohuY~5TS{};B=;BrZ78>!8LX<J5;JB7po<sFU$lvAgjH27Y}zgFW8ncP
zHvae(NN>aXnf#XHgj?!=I&QUSa=(Cg(?8OK*?D@txWf499>wqjqMPXIj#sSShdzdr
z(0z6}SIA#F(T&qub(z7DUW2=vl0EYj5%R||bgy7wB11$8bD(IK!jT-Ff_rNHeg&95
zHm~KLj$dGcQ!eNn<Z-9PxQkVp8j}-VSC#*%Gomgv)Lq*^Nb+0Zv<X^NN-h%*S)OlW
zm=0znbw3t<TiuJw`eKusaM)^`PDtbT?rAsA{i<_X+QYZQQAZG<1jJgX<hZG1Z2=}a
zHdhPv4FI{3baNLAN$?%++U(lLwpy_E`UorHyv3T<pD1?Hy<F^7L=jFJma69l5fn~F
zS0vm&%oWzHh{U&KeO^#oYj@#mOR49eFIHch6>(jaDeI^JxpS#wTu@tVpU^2p0vO1b
zlylpa><8yI2UaS7T6kODm||>4%KHDbC6_fs#WO?YM511{@F+6m0XL@l%h<YXNuG)z
z8)lu(R!%g7k*2&3!#i+%sW_ZWIV7?W2$ta3lavrUb8do;vd4|TlimMQU+Bna{)oBO
zenT?9N)=<;4MpIEN-9-mf+VwHakgAn`eTAbh`|Mt)T}X`t{VpkPi;8Tva=1XK6E@d
zh1pkei!413f+b%Dze^q!`WL>o-wDc1GeHSG(17Bon<UfsFBA10-(+lv%4NtX1xQN`
ziRkS0sVhogKuMhNkaJXji1x3$#J(&l`Ykpx^FU}Ojyq%+@q2*BC0SHd9hET~Pj4i3
za#dC4W^FA~P07cB4bRAFd%d#PcDX;X<YQj=HxRH2G(Le?EA$g%Ahg5g<A=CRhpWd?
zt0<f5f<stZoc3|(=#2%nLGahAC?x!b?|>S0L0yfLe`$&YL{opc3aUG^&CChp@cluk
z^8YpKq&SEaGk)3U>?AEMEuNw(Ek%zlEw*CoU@R@A)2$6@Y2i_;)X|j4`1cTOo(Sc|
zrA`Ng?tP7O5AT+z%eSA^^*qVfCisG%y*Drt{>G|3>S1ldAy#bgT{3LLqcTZ9e&Csz
znWfFYgO}?xe=mFAW?XD_w>K>R$M)kw`mJx!|7&8J5<`|6dUN9qxN!&s?kyJO>`w_Z
zj&b-p?uL&3u}eNGEyQAjdekv2a(GWO+G6H98Q&=Qy+IzkRp50<P^6ggf&fR|>dI@p
zan*^IMnOqu^!??)KRY-Q<my@cf9UFEclO8coPhnp51iPM&^%#okZA0`jmlbh2$~;?
z_BuhrA5XaC`8>d%A^Xv5#k3?RxOpY*^>G3F<&IYGiIr{v6BL1|mSKH()f=$Zb~K%N
zQ4d4hU^J%h>`^G(TJAd|4V$o<KA><lyiNQ90$tVoWvk#RDl3)3;-*D~G`fa7D>i=w
z90FsP`Et`;@MNIt7g2=g*%alOMv_dYpa<_^i|fZaeAUVL>HoA0-HdOVPbo}+F5#J6
zSroPn?#lCr`E$9A&48R(YR*5@)P(N<*6k2+rD@j{`2<+v6Dgbshj}$TtE$s5;hMp{
zaf^M`STZ$iU7L%Vcqo{IHrnH>bU(d#xbXbq{ToS<yKFJ!$_bYsZc_?RzMvDvS=WR@
zqb{5#R=@{&a=bq0XK!MbY=Yqj>LNZSk;xqi+KDl-VBBa>Qq74WV`en<rS+@3QetFi
zjdzbKn+37RpBUFlOEnGnoScy49T9iwjeSE~_be@c46zNGgx^J@J@)nlS-833x|R;;
zjK@T+sc0ElBWTw~u#)FIpteT)0tY={oK?09!Y^UDaLt%@3p!N&(LtK{|3;+W_JtzY
zPm2530Et@>v_m7a3rMQAkwrTH5B}pV0yhBo1p8D6Gr)IT)<#uUO18QtK52?608-oC
zK(}fBG|^9><!aCF<2KS~{Zvg9Vxo}J562@$IgJoY7Jnl#O4&?u0fc{p;5_Dkg5ZLG
zgWz1j{}u#$171D2Sch(2heNt|lD9^T7-=cP#jeTz^)3;{^}A<h*`!k$%$I8SWU>Mv
zVFo;$cj?<xr1)g5=l77kEkVop(Yw_sdinrQ3G=Mw89Vk|-~9@rr?%QN>#e!3%xTkq
z<K0+SC5);CX+TBT_<BgiSo;@yK>wmsO>e<@%mF!$_P}9z&D3g!jlA>KCbXejkT4Wg
z{w%UcPe^8R7))B$Gz)C}!Sywt$fAdaL&O(PEPNms+CVSrp^^e7wCH7GO8a*e!Qe4V
zQSjk1j(sAr2t-y1tebutC3H`Y1i!W@99vcdP}bzv(qm_Z)6xP#<*Im&IB7ge3lG{r
zvGAn<HobaMn#CC%ZPVY_a#iZ5;_{~@e-;U@k*dJr(^uP<lkp|lc{otXs7}K$N$9Wc
z6(U@55DVm>=;tdVcs|tutFoJhQ$rT`uDPR`_;7nXzpr(3Bm5gQ|4ajn^JEl`j2jM3
zJ8QJu*rd>DitYY}+EU=67Q>A-gGalYwkbP^jZ;)n6;)Net5+$Q#mjfbf4Ft3FN!3*
zZ7Ie(=>zP+dl{L6f@T(AQL5QWZ#i)+kob?a$<A!(DH8+ABlBHUgq7o?TB<@-qzwb;
zQBE4}*es1K?#(LNr;rzG$4CZ1!(D!kctB36&E__HZU!*b4q^h$l{Q`Q%-D2w5|-?P
z31N`QIS_7L%}9TPP5{4I9v_NWuisJ=u9(3qlgjy+oBm%G=l-KZ^#3$Cn@2$gC%+H<
zss2wcnyiqPXP26^rMHkK6~xFjWWHFr_R+5{T4)$jZfxf!7T*jpWOJfp9m!sApoW!0
zl;A;N66*cEs-6g*jR^bbob=R12|1mFv8-`)2EF#v<eVI*-A6h(lj2t0_j1cyxjU4U
z|3d1*!m@|u1%^#)oXo)m2Cn#I+RlOFxiV@-+E18IB-Ao#Q^9ks8H6AsY(^wG0bZ-2
z@ylqLa%ZnXT^#MAkneGZ56I^D<99Zc$9h_^RZZf<43x(YMOyK^v|N;~$Y9L5tZJKA
zA0Z*fU=7ECh7d>FCEW`#sP;^cr_{Y#8#Cn_D3@iF`r|}K`h4`Z)bxv8(ybdHu7sD8
z$Fno%Hm&hCE6B+Xm|?*gIxCg;EB=?R89ZphF_YN36Uq0&@SD~|+gbZI9YpLm_>Q>8
z+$M=SNKY;qBnax8MqE)8s~LBCN0-u2(<C36kmbL+P()UZe)c9LR2dleUgroqv|A92
z2BqGn5*iEDJ66R;4(6xM7=0T85~{*d$fT-lv&S1~pt$|f%XtfNVCgWwraZ$f`rM|u
zmHtw*3}q&}%wr-U)47_0GTw1cRTUFIaT$q#*DWShA3492hJ#IfJ~MH3kf%hm6`Ds(
zaSq<sTZFOPxV5xa7sYw$L+KIjnz(dmzhA&HvY2eUJQlX?*qD#>zLogEH5u%+@V_iR
z1W9Hysz2RhXqv3~BIE9H@o2=&DfA+Rm;7XA;_>YpK;T5@GvEo&R*oeft!O&&`+G7p
zh{jnJ|Dkd1i)QtqMaMbbqcgl;opfJ3RkhdGIq8@5N6IeM$L~Jqf6E06u@ILWrjZ|O
z9PSq@<3witC~MtLEUHKs@BY_G9pmV*{rD*%n4Q?ZEsL6q&$g)@2%xN{+|$b@&;}Wg
zSXfxbfF-(-d_}AbS#1Hz{l?+6UA=#~eHYpFSBotuH)rx_9>`wm|59#Ko;a=-0Iziy
zJFUpy_1hxs)$&!y-kEIGVot65pk;V4PK6~am2c6|K|R*hYmsHL&&*R>@ps4>p%q(e
z!av69NH0C{KRG&zIAAprHbKJ4*Vs=@>s5vW6cm2sBTslEI$5odS>#j_<>W>3RKJ{5
zl0r_Y2mXW$)BLK0fF({ouE4nlPI=or(4kZr9M&CeDe4db1`3uxRHap+QG&}%=k1#r
z&ll{Xr005{Y`-gkVs^4hr!N^qy?BSir?;cuyfvx*%XR%7i7}JuI!~rjFOh>GEvj1d
zcnX-{Hh<)nB%@1b9%}#BSIs4y0+Bl*6IE1;_mPJOy|t}2Sh1Z|9*>&ZR8Na7ht2Ji
zn=*<OVZhwm>|0&tFW`@5DRea~pYsN*hS`<HKSsLX2nOd#6QKH2{kbpRxHJ)44Ze1s
zve&fJBot|=L>el1OVY0X#&X#`sIbGX4G16AkSeCnffBW=zMi=LGYXM4;u6RbyOkwH
zCO`6Ez46CX{~y5LIuxE9LfOA`IHP|%G3_?F{wx1a43B^UQ-9_Ju{9jC-uNZeDD5NZ
zK2Wk9IqW($a8m&+GnV?5yV&LHf&!|1dKnonzB8%7#&-boStt@n3RhMguLKl9z!2$w
zqSB=@1oA?AA*-zdevA>gJNNz?e~Z#b|3@Emyf~#N)2b_}s059f8GAiJ&CDqKT=tvm
zxIOZyRp}vh<%yDe5POEa11f+8h+n9nSUOW2yuR_(m~NdQeD(Y6hiQ@?!6x8Y{#PJa
zdcA=-jKzfgJ>8+@8vtg+uJJn7-e*|bo^W$yCr8<2COx8bMkb4^yB|ltFBFHuoLIz^
z2!!$bF+>d3BWiSeTKcG0<jm&E16n1CgUh;Ua_4&&OwwLIK{bI3BKcTY8wH)6IS33I
z-QUTjtCws*uJeHp7qgJTA^U8xl!~h29M#kzojn>nsEj8RQeI9zOXmb+>|(}c3WxmT
z>V7r5N*~hw>$VD2|5u6R4|W*Zn<Jkz(!YMM!!fmB5fd;1=0D0XcVDo#wl2`ti3NLh
zcg?v88+{&^%Bgk7L`}<_fl7Z{+o}WUpxZMTLk=AehEy)b!s6en;2f;1iFIm*Y#;X~
zVbJdUQL+E2#O%S72SsHl@qQvlQP*52<p7N~569_F(P(UDR@UAdm?~u*9fA%2c};W0
zmTf)(zNe`n+~9xA;2_XZT#c-xw0~==MvCQM`m?gKf-V>if&Zg^n%}-Z6FWN=WAvOL
zGrKBo8H!u2r6*E65uJmsv{$peC+oymVPp`~KtAVk)A^FJt#}~S@A{z-w+;p!yPXbm
zoMCo463W*X(1!)YLcnu+vR~Eqgzi?)54wEPlV~ws`P@2F1L3&nPqFBRxJ5i(`Y=HU
z7s;P0c-8m0gr?M6ile{4B^P}AloJmOj`}DiRmT8QLR@}%UqJP&wVRVwe0)v0;V5qk
zFgH#cB^*?8jQ8|a2V>%5OV>M5F|I@~Kht5t?;Y<8ugwkRPVaYo$r^O*xffn_Y3Is}
zLP2M;(H~){ndj3pH?W~+uQ6?wmoxZQk^wieW8~VS(>m81h115jOe~e8yCiX9i#+ml
zQroaJM6b<XKH>pEH}K8PZghpm%Fuh+FYDI*aj}jxUkLhhPe4@`&S30wEjOh()$^;z
zyTZU)R5X^e-RSOWN=x%|95<)W{x=um-!Dso)zVuPNMTotHBa|PT@0Z``kVv`#|)FM
zbrdLpu4HF55cHi^US8>pknzU&hbH^t;AVHP-}4rTDVNbDO5I=jB6~D^EfKsawBhyd
zX?t<w*PBvm)RM*RctXuGcdeZ-D!?;StN1N}2XSue#6I<r#B8H#@)!ohFHI2Hw);J?
zV?NGKyc^7F+Ne1|67M(AnQHqk$69p~p9?-}HqZ6)kRd*>ZB2Q#uN9PoI+a!4IbAU3
ze{43VEAZwPyuNS$+)TG;=Hc>A()3XShJXv*DEU7ube)d5L5jMJyO~y+nf{Mo2AIEj
z;0`|hDMh`*T4<d1YBvmLKxNF~bWAqz6V;_z_I$e-P`2zH>OEb31I{n0vA@t5I1zjF
zOzde!J|`Kw#%ur}TQP~&RIM)87!bL`VdIGf$45sIa}^>M^bjt#qu-+g1JymQf3p#9
z-13Qbwn40Xp^SciE*l*k{rt-7_ww*(%4brjINTGf2&Ao!>zkLlUEtkyEcd0Pxu5*<
z_PjVId>ITvMB-gE!4$x!(Uea^p$v#LTqkgMvX$YoDc)<D2D1mkR4Z3JSgajyUCmP5
zqcAIh5;!;GRfQ2(#H+iI8l`W(uNM_aaEi57s=q|DVnG$2F<DmRj*$9|-EccQ1njn~
zy|@uf7vW3`YYA(NyQ2-(1Rfn3s_wp`{Ng?8nC9CsqSN<m7;y1U6ha+FuAYH5f9FC_
zZ#c^JIB$4bdb;7!^d6nA7xK)=yu-z2D?!U{IOfJR$IQ%(jGP=u(GAVcCNWf(^J%oN
zHtLKdPrczen;LnFdu8GCa=TO~%1XM%3@%(8R;AZ_{X8h?JHB($FW0H+JW@A?V?}nW
z9WoGSXqF52Q#Cg7j{=$jweBog&SvF|2Wi^57w1NNM;5L&%Pt=1YuqZHq#AfzoeOhz
z%rW-v`J?wkC2c8FtrL$vW?W<%?S%7^I&3T_O(<J+V!N3bzuO#8>96CFPF62%R+MgL
z;BSsiB{j$nCLEZTD4Y@-g|MhqLd1pqXsVqifjup!3yqAFAVLa%xR8(%EL7ELfG6n*
z-(k863A*=#uF@QrQhG(?C~`3H4^WMZZ4O7H2=h#&@enXe>@j9M+>slu8R~Hhc=gLB
z;*1!F?wZ1Ble{BZEZkhfnr#d^ZQd|BRWk*xKV}5pKD@uPpX-^He`CPw>334NQ9gN&
zPKtSrt2z<?aIP&s9fgqQHEK5SdS-3~*v-(KTgBaFt8o=OhI&$S47>`*=xF9-(|i-@
zvCbaKKtfwb4%6t|+GY5V;~a63*-u04?-SD>omyitMP|cc%M+G&&6xMFBmAPi6ierI
z?yFXzHN@zDPewbgJ50FI9#v98-N9$L-z~M9?}Pf>|7OFA!(hL^fkc`xXumYpDPPJ5
z;ZV&A^vG6w+@y#IHT>CY1O>_WFc0a&)U4}GVtaF80TejvaETG;?cywPk{k`wPJcX^
zt&YKi;Z;0q4f4-fHP_-Cx+`!qKW9DCKtx$I&uG3#;q-9OFpXEWiwo7=_(4!eMlHf+
zHmwd?O#RcM43kk9=8pNicC^m4%dURQ0+#r@g4b6TLCp^V#&kTQV+ZlQ5D$FtCtep%
z+q85ubgGzr28*uF7Z02GQVwQE{59q|YCh3Evu5R&^Gfeb9^!ZH6T}bS0$kA34i9Q5
zA<8>P26;9QMR1Ep+q33O%Rio*ZTO|$A-MPxGsg9L2`587LzZ_kV|&f;NzQpKfvzRr
z@rnF!l|<z~p~ZcT0D0dxcJUsW7#I!L{P_75Bil0*17Yh`x!OA2>oCnlLHQWI+(T~5
zKIc9)osD67aZEk0+!WWcCl9aHv9%~jS*V-pt8<^aV+fC~h4E}v0S;RU4;-yTr?YY|
zw`xmkeA>jXA!r}6!^Z|t?WGPab|Xpcrh41va|>nM%$(Jn>Utq|nsxWAMlL)v%G!%4
zF<yZl&-L$md!=I<?z@q0^o_$`b?{d5X2;ejAKUimsu@}~ZPcn7{<xbq4Wd3Cr_NC}
z9N0OkA+}3iDJ3&)%ufu1D00<%4fqew?~iOW;2i4y(TrnXLnhAlFUxor&1LS|kMxMz
zyy)LAPu)851&jSI>Ov<Y35orGsyet^W1x6dh$)N(YaiFWx7~QJ`rKCbId))=?Kp1)
zbW{!PdH{BJTz#VHUmPCse+~VXY@m#_y%UffY;XL5;QpN~)2Wd*xu7F^ZEOdzOAH_7
z<1v@`pn+Ozndik6Hc-fa#>}q^U!`y2k*A}dVVyf-=(etj1E9C_l`_QK@}2j`_>{#u
zhi>g1ya%m%#`Oyw7A!p}e&sW-{tM22l&Y!2=WEAb<xZE4?xB|hM0|yIm94b|M#_h>
zH`miAa+U34FBez~3xJN!@J33Vbb9RDF`Kyis)H^9&<*@5oIO-Mz;r-^v!J&$CjI<L
zMgP2Zv_=v@h@LyI{D4|N+^yxV{mFi9BQDhmbjWxHwlG{$sP}@4yxg<WND00d>szo<
zWZt>PW;PV97jiK3sP+%c^|>i0QY)U;o#Czn%qF#oHm4&~$OzNp70HX4fc7FhjMr1S
zCqJw8QJEm8S6u;yjjD@cl$moIrOF*Y!S&}?%G=ObkTCZg!QOJPkw}-A{Y{zU(-V$?
z*3o>Y6i3KAu>IZ0hDkbNGRBIg<42jR^QHl128UgsxY(mX?bi(=<`Rv;PnLQCivqq{
z0qOvspJr?3Eb-_5kdziEj~>L(b+iL>gs0EJ3~cZQ_d~rlr`I8K+<8WJ96-S?(&HPq
zK@W35pA$2X^5ZUC*!@EimIbb5k*&TynfY8d_*vM<+V!N-;oTolDzy2Z#J=(0r#dz>
ze2Y6HHrXD2Dz}gNQ_gE+ux1-u-YEzwH}5@WaVp7<gq&~>E(6{Sx7T_JkFcnuLbimX
zspD5@121>q+kRil*5+dvQdP_z<4nid{0QgVz=p3m2W|{>N+2_IqU)uLDpa;s#Gd&4
zk(&0!Cj9(9T;M10x2CeB@%6B_i#8l1(?>t3lxJq>G*w3DDYJL3s#cm`d6SDvt+IFh
zK$Gm^7#q|0Y#6ZO@olAv8lUth@w>HgN2I$r1LU-ac}N$?+HjFy&h5~JGUNU{-YrMu
zi(U75UlDX)1k|{r_s5^h?n)3oC5OZ4+iq9c4|h%w*PR(pT-ctjaY<({(z%&_2DbG$
zM`fS0raC91gtiPwwsrG9y(oIF<Qx&~MLEiqWt{B^Ap!+jV_}Xq#W@Z6@8^8<6LDMI
z53vFGOJQ1-`5C@7;m2<RdY?gcLXfR3X%2J85WRHnWIa=nyH5iz<#kU#Y_dVioCgsa
zfp6~LFsh8>ioRvZ_IIn-(XA8xkN`Q%$g*1ii9M($Jiub%S8S{9#kTz|Up`WDIWeb>
z`(_sde-Tv$-S01a@l*+2J}&P+U1u*O(jt{OGhF3EG}jVd2dtYCL3DOxuWwd6g9!xG
zS=IAf^TJDw-(|0tFDmqxE_mOZ!&z*E+UUX7C|9u6zTGki%l(AIb7Djg1yz|x%2RMV
znLw!48v;s-t<a{5;fsrp*>ZgExDV*{2O=>jEuOn$Go8n|F|n|i+1X8*WrYdjW$qj{
zuIhDgbBc>2{d3r^;nNfcKlKvYEL-e+1uONuNT7Joje~3jEmJp@g=UY~V3|m=)Z^=j
zI?Fwryf&bXqhoC{{{Y_dC16dbsZDRgN@_P{tY(5g@uaT6)Hp+PNwr(HQSEw@o`e48
zF8ut5%Qr8*Eg1O|q94tRSvb%w1I!Gz-S@v4tZpXt-tNq~#GUhkT8Y^kGWzcg*Q+AO
z6PPfdJf7Z6xQjyhPgORy_G|o=Ci9x*?3U8Us?K*G1HZ36Vf3~|W{gpzj5PX}ecs1l
zc^l>|%b7C6$x!#cT;C`>ADiD-smpi?zo&E(nBOQ2=TXY;NojzTiD4B|!19&^14F0(
z^SuCgsX2Y&UpPMR85~`QRru&lT~*ov06`xr5h&U|WW0@gtlc^ze6jk$QH0&Jc=~Zy
zk^iOsSgNQS(&B(coSRV~*J4YwFy-V8RK<7iF07^7E*Pd%nX8FQm6$cucc9Qx%bdX5
zVCcpRFMW5Hjn=0GjYv9mT7=*T+L}JCZiKq)Jj)45A<y%T9+-i?7hp_<^v3eaZKYT5
ztr0(yPOC1d1n<E~%hby213oKeu90#PY#0UiQ!~f=9Hih2gQ%z|2pq5SbNC001>Za6
zmkqZ0Rz^eDZ?SVT9!&yAIbA7g4Jt3GMNlB^hW15cSVKS@uBXMng8%h#;ZGg>A7u+l
zNuOLk3J41Hv)CjhB{`gZ7yW`ZMI|FETVs`+)!D{#xvyZ8H}!7s5#pSZql;UXh`Td|
z-*`DD*hj?DzUn@LoEabE=StjWPT&+f#x>g*-8I)(qd@7w$mi8`dR~hq-8}!d<{o4)
zXV|yJy}NJF`aB*nzirprD?HUZQ_}?f79WIddnGz2dR)9|lJ4{)<=pUC%S4hZ0?i8D
ziC=phe#PgGW;#J4T(94(GyDFq%LuX)LSYUZ*CLy=($nhk{}lG-G@<Zl`<K|c|CH!|
zgZs}9LYmDJln-uXqOkrG5z6tu59(w7%F=|Me>4wN*6Tt0^ZXVfr9#&&z}YYQqLU)d
z;imVK&*S4CrEjvdRIbsSUR9yR?``@wV+@%|TQzrz4{5&Ic5dG7k*20$k>5)H@k3Ip
zjOk;n(_t`dRVU7WQ~Hu;IX*9Fg|y>4<$!Uc%k6&2cWipCK9CriA#Rg%baeE2*EG)(
z1z7<PyF2l61GxX3CpIQ%@Gz8>i%X+QOF=68p%auWc06n&#C5(f$_QQyg2Y8sYX5Og
zS~?G&@$c`<ekTnz>{ZDgI3(dbj~UY+{k#uNTz_U&l}Xv!_*SM?6po$pYU*ae%RN0}
zKpDP8qWY+8^*2@w2e&Pr9kf5PQLD$@0gBe7FQ_uH@8DB*3CQ^TjG{mX&6~2Aeq-!-
z2*Rkc8N~L8`A+<vFfn*1B#J~%u7v~g00;yfX`oZSA)KFITwL4VUTC=!PL9kg+bU%0
zHKfe5-c3@yFPnfqJ)w4B30C^}tbXS>NPca$sUF|>`j8VvpQ$6a_r4zWsYA>08_r7W
z0{|rQUv^Gb7FCQrGWIP5z26Em;_2d}=ir=8Gs3U$JeOX-QYkj%qD`2Eo4!T^Kq4Q>
zmWW~eKSlo2=R~070eHuyFY$tI%2XNz2I+$ZudsN`KPun;`La`EV82y2lk;FeW*lh6
znPLCKxfl8q5Kw~gPGENVI9d@)aE3<R)~LNT*@*Z$xIEV>e-2dh@ax<Pm^*{e6}P@Z
z)u7xaqMf`d`&Cr*$a2Be5GZHnhiPE8!AR}(>X=6#Sl~4%pGj4ITNDDiGI}F9=Zekm
zDwWmVk_hiS>rT_@{`M|lgUj`l6|bvP<IfZN*4rt><A3CS6}xJe^O(+U>X;5+N7hH<
z`b3|%4jw4a%YGJ)Mo`o_x8nEmzJ2h*fr10-S1~B;oL!V1UdH3u-o?k4S$;{i8-MSJ
zmrXsg6<C{CP__%pX=8_7;B;WVSQm*vFlz><Jf1N7KZRUnToYOw=W;<1q+6t<yJIqj
zq=XWZcce5*NC*;R2pA|b1(6;X$xDpxl5i6QMhpjx(W65~*LdfB-|z3IbI$X7o-h9=
ze!ufS=Xv&-)9X*08MlisFLnk(Yp#6=x*@`ktoi+Kt{%S(1QJ$LtONsATqb!&hxbv~
zToYb1ey`OqtL#k-cWO_BYUT4Jb+gNR1ET3Vjd|!C+#4v?`!vQ=q{xYH>3FLHPuCWn
z4vuc3*CP{MODGsQNc1Gw7aOpY?I(*@Zm#njvp}UGRJfM;sC?B|wUPeCE0Pph$@nT+
zKs!bWDHttUI-eYnhDA2*+hvyVMZWST0~F{!Te&;)W6q@^{t9TN+e^<#gXMXN_DqY8
zR}c`OPG$^Qdy^{d-m_2(ma?PKI8ck_L*YR{fA~QI`yBrwx*YG9gy)toFG|sw)p}gP
z?r}ohyi81x06;y9Pd$=;-~P!4i8OuR=7M|rH1aYUGs&$Nw4L0=#LrtXLJqhcjDY?>
zVRStg`?IF_oGVB{?7-!ICdLrC8(F+-C&tQObwBQ+-G%^Y$YP41tzf$aip~h11@vJs
zH-5iTu8)OeI5)nEr<o&+h-w#}z?(ARaIfIt6!Ku0;^N}a7oI!ak<fb#c4^ZG&>F`)
z`(bN&hb2!*fHgUzdB<x%x#M%&^QL{bCZnQcQCR_Y@(!k0;y!kMUDJIVe3<5UJh#RD
zqi5%L*8Re*vJxEf^!9rS8$(J{yAt!HFPr(q4#)gPH#2kDav{glKBwWdq|nNRNN#`i
zE#*34-g9rwd#Axq6Xu!JdE0F`94}<Kc};zBp|>JpJ||EGGVOT}Xh5ldJ<5QoS>Fsm
zh7G$=3qUi_`pQBel0s45giSPr`fdVDd>(Bsd4l7(@B&Ex+cet~6?m=L^%6PLOMCA%
zK%b+Xz0UX%8U&d!2B>BFg;(#LDO917qoUqu2bP*MrNM85rdA)+9+RVpgx7&ce}A_v
z=H4WwIgRFDnq0!^IQwCHe4sHV-B&(j1^{+chx$JZ!0^kaOQ7~6X1&vB(D<~1Frf|3
zcD*V>EKp7VdALZQdgrrXu6Q(=(j0b6gK2}1i{F0O*ZZ)jWvlV&jh=E@UhD~x%C`(T
zJB~r+ww252T~Qx7R^<rEyUyP*7<L!9!k#PN#`4wKI!9>ot|{Zf8W6G?a0g|xL$j>|
z@m7YalG7M@vaQ{#om^gFEUgUS?EUko#S788&p;w^q(M9S-2Uban@EBY-xo43WEi`2
z_{l;t0^7e3r$xcJbFJ9g23-PachKvYLi_5Q|AqFF)D!rVl`>6#;AJnrGw&QCF}7<4
z*;%|Tw)!|W4+e5AU0rBGKxTpxKg0S0t)0)!dN|jrxiv6;R_lF8_b2WVW>?TCZvi%g
z7~~#$C$YmS%7|dF&4+4yUaEq_wfZNiwKs^avDDp5sw{qmxjn7&M@V0v(fG38@ng4o
z6Te`aUAb&ibRTnoJMGP^d53y2j&6px?6=?a@9{ehPHE~iK?k>#+6D)QW$0}#ASAbN
zftdZ=V{a%r9Ik~D9rEGedv&nsFWgI?-4l;c`|3&T8ss;nv@yI1aKI(`hcK+iLYtYM
zW(bH!@EFG*&TL40>U<HFQ{u_&_x8i=k(VE!do26lf%@}*4qUX}9DBHDblnCzJA|xh
zO|iMXlThULYKDz=dvUH}EH>zpPbE9k&S|>}?_oiaqwiP09E3k~w!Gn1Y##3aa?4nF
znZ?wx3wDuD2d?7$2#<!->wRkuoBo&*dyR7!k#B7l3lYaVnjwV5p9F{Hs|<#eG<awN
zCMIMUrCAu~+bCNGSOfKa%i0=WMz;)>JAo(0eLc!EF!?Wnabg6UI^QDO<<LX3*MCH(
zPoRaCP$dK0*M^@dhhker1JewHi!dgS6%jnM0mW<;>M5KlSy{-7)9dy!*KfeIkdJE=
zt{iM-=xyr+6?7a{A)PJieQ?=RYrZPsyUK`X<tN}u{fs@e%&q)j-|RA;mgNuwCCln`
zsZ}~T)yv#N>P$Z^Ysn#{%e|~y7KqC(o#E}CQy!MY)GSG`i=D`$ALOAGC{_C>3)Q!g
zvlG;o&#DSGo)?7{$Un8z<!zh`wD2oeJ~;Y2#kI0Piv3uw>nS`Fvj`mE$}bd6t=vry
zIe9bNx>;2Yx^_4>jx)?vd=ja=Pc!0c8<K=d5lry#phoRvji^udEqf}k_qv+WP1UzB
zng2qjf|G=!03tu~U@BvgQ4G62%cF^#>TYe5UoSuE8ox7a&srHFdNSRwYNA9FxQD6-
z#!xA`v;ByY!ySsF!<Sk7*$Ue#!IJm$RI-&d)iF=&f%fYNrO258pbV;}&Du-_ytZO`
zQX0o;;5YV*LNG0%J|Grp)Vm|~$!kP+WnGmelB3_+mL=I*U5l@%P#UXa?){ksBKFKj
z!4vmbG<*D~CPN&7q>w8HGjCA*lDjEZ<Bt$3=%Cy+wnk4WRJp&Vf#H-3nEW!#s#&>2
zT${XKtFqcG9-s5@-<&^}?nF&aa__Hp#8HQuB=Ano3AjS5Wxq1St7G8qTlQ@<jG@5=
z+Ey_T@iJ2Dk=m9Zep&Aa?c+D`2*KfkXiItVn?WO5A*Ge)GWiYfzvsY<kFI{jG7y`c
z4jT1Vu2qUNMlc@tzcZ+6ilhHz39eMagD^0Gw&?uSfk*G`c$#@U(jJ1Hq`H0uO?4#Q
z0O%0sI{JN%412i@8lF21(TzEIPR_jb7?SA!VWh-@Ftj)>9tqy|@X*`zo3SW)_md5r
zRdk2#yK>U9jTJbTQ|SRJp<lfJq8IXE@x$WMKfMxlzm~IacGwjjTEq3ce&a)d57FWr
zbDGdFYc1fNth+VoGrvGvh_;x**VE9;*voJZnTx|g?$K^>c5KcIy-<}JVx>yVU-N>@
zw58cIQOQBA*G1Su#BGM3gwADOH1Et5-rML-e8@N_A>KsM|6~a`KvSD}s<XrstzWWy
zSson>p}%tEEfTk7cWPyCmT0MYjb>kR!AqY*+%SB~R`sDlXWtWj8K(^cVFz}DXNQd+
z|N5j=Nq6+{rQ1xg-mMi09R0`+L=|`LAt!`pB$PXSuYLQ}0$n-(pkWCptH1AZzX%)k
z_H~2-@#SAbnu@Vd5uo!rI*YZcB%<(R-E4oyZrAkR+xRi}ey0Av;ihzRHk;dolzvc8
zd{L2B5CLY?iWfxvakAFJH#Df-W^`VB(~iF$(+?_;p=s3l7DE%+QCs3O;0=lv5*RJi
znT8&Ft&Ltu-Z)mw0^dgf+44DYc*M->yfLLx1$Ua!@qgZi%7_*f^8946yTAFK3ltfL
zLQsaFO3t`HCgg3_q`ZQ7^EBd0*x`3MLfg~#dM?ZfCjyEZ8v25+-EP;JY3(D9WGJX;
zsl-k;%26UQQGb}jf%#`xZTHbH2KJS<NEB)Ur{ZCqLQ<z6_9t#`Gt}A{eW8*cF5)HF
z5pL0+xsj23=0UwHi~lng5(Rj*{g@{$dhDvv18ke@Yh)*;jjvDc#mPKeuOvC8QTw@G
z9TsfaKk4RigrNyMnd@uGiAJ)UrFSeCz1?rg6{IjR<}7Hn)N{4P<oc_~DosT?0s;W2
zvk9+$_Q5kyrXCpw6J@TMp+|wHo-may(Qi4IrIbv9hvrJeZOwLVi^+P@ciBh@G4(~S
zh%SU%ZQci>LnpW$v7#Gn#kd%hmoleX24G#keKn%Bep>DS*;JB6RPe!eH!<;x$ida)
zbmo=rlqe3_&WA#$2uEKLBJZtmF>i+8FzMsA3f|c7=t4Ny!YE+&K|8Tk%(KdJkfM0+
z=LVr3V{1hCmTRCP*S6QX`N*WA9SFXCln>@HGWXWV=Qo-1Or5Q%Si7_ME~q2OvoP<q
zjIgj0%cv$OO>6Z4IiiUzd6}EDq4YGbM_DA-ig}?s>t#&9=evT;BlH>z@v3hRwMf|b
z3n?hW_3KrNB=$}?tpaF|o!-2L^(!kk>X;yG$ZvXCMXJ7BCA*AYE3$8wL{x-=eu>AA
z!Bb<bLOvEdX}~IeN}xMLzI%S1D-*@Wt$ub6q^^`i%J+CJQB&&NTKhHci`@`7jvfb6
z5Q0+tBd`6qqT+`t4^zqvW~;-nE)OUwfKk_$mX-<&3u}(VQ_UHvdEb-HfATpyG%JA0
zsKwS6L-&NnHscff0HB#L%JUb)>FO=Ia(~`wYR|!!y4Wu(^#&C5r%XM6ahI8iTrnnX
z(+rt%mdHht7!q81iG2nqg(nGkPuZX4Vr#&UTtkyqlAU*$K-QOC>T@S@rvPP8reub$
zm;<cJ3-yq`yOR)x&&$kQQrYfl*`|7B@;Ywyh9<!B{`&YkE?TNR!Jx~|OnIkHiVQgx
zIx5N+WHbj27XP}Pv>>Vb`#?*U@!#%d{k`55e53pdXg^eU5_0cfO<F4XJC<7Z|JzfB
rykX7%uZd0$%k|%mAzSI>oC>Ut`yAoKYs<*fCMk>!%=8huu5teX5yhi7

literal 0
HcmV?d00001