From 23616739e00ecee0acc1d2826efb2df56eb32aed Mon Sep 17 00:00:00 2001 From: Morten Hjorth-Jensen Date: Tue, 12 Oct 2021 09:12:19 +0200 Subject: [PATCH] setting up project 2 --- .../2021/Project2/html/._Project2-bs000.html | 466 ++++++++++++++++++ .../2021/Project2/html/Project2-bs.html | 466 ++++++++++++++++++ doc/Projects/2021/Project2/html/Project2.html | 406 +++++++++++++++ .../2021/Project2/ipynb/Project2.ipynb | 275 +++++++++++ .../Project2/ipynb/ipynb-Project2-src.tar.gz | Bin 0 -> 194 bytes doc/Projects/2021/Project2/pdf/Project2.p.tex | 433 ++++++++++++++++ doc/Projects/2021/Project2/pdf/Project2.pdf | Bin 0 -> 239136 bytes doc/Projects/2021/Project2/pdf/Project2.tex | 405 +++++++++++++++ .../Projects/2021/Project2/Project2.do.txt | 40 +- doc/src/Projects/2021/Project2/make.sh | 2 +- 10 files changed, 2470 insertions(+), 23 deletions(-) create mode 100644 doc/Projects/2021/Project2/html/._Project2-bs000.html create mode 100644 doc/Projects/2021/Project2/html/Project2-bs.html create mode 100644 doc/Projects/2021/Project2/html/Project2.html create mode 100644 doc/Projects/2021/Project2/ipynb/Project2.ipynb create mode 100644 doc/Projects/2021/Project2/ipynb/ipynb-Project2-src.tar.gz create mode 100644 doc/Projects/2021/Project2/pdf/Project2.p.tex create mode 100644 doc/Projects/2021/Project2/pdf/Project2.pdf create mode 100644 doc/Projects/2021/Project2/pdf/Project2.tex diff --git a/doc/Projects/2021/Project2/html/._Project2-bs000.html b/doc/Projects/2021/Project2/html/._Project2-bs000.html new file mode 100644 index 000000000..f9c800422 --- /dev/null +++ b/doc/Projects/2021/Project2/html/._Project2-bs000.html @@ -0,0 +1,466 @@ + + + + + + + + +Project 2 on Machine Learning, deadline November 15 (Midnight) + + + + + + + + + + + + + + + + + + + + + + + +
+
+ + Project 2 on Machine Learning, deadline November 15 (Midnight) +
+ +
+ +
+
+ + +
+ +

 

 

 

+ + + + + + +
+

Project 2 on Machine Learning, deadline November 15 (Midnight)

+ +

+ + +

+Data Analysis and Machine Learning FYS-STK3155/FYS4155 +
+ +

+ + +

Department of Physics, University of Oslo, Norway
+
+

+

Oct 12, 2021

+
+

+

+ +

Classification and Regression, from linear and logistic regression to neural networks

+ +

+The main aim of this project is to study both classification and +regression problems by developing our own feed-forward neural network (FFNN) code. We can reuse the regression algorithms studied +in project 1. We will also include logistic regression for classification +problems and write our own FFNN code for studying +both regression and classification problems. The codes developed in +project 1, including bootstrap and/or cross-validation as well as the +computation of the mean-squared error and/or the \( R2 \) or the accuracy score (classification problems) functions can +also be utilized in the present analysis. + +

+Important Note: This project as well as project 3 have to be written as a scientific report. The instructions on how to do this and how we grade are available at https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md. We will discuss this format during the various lab sessions. Please do spend some time to read our guidelines. + +

+The data sets that we propose here are (the default sets) + +

+ +You can find more information about this at the Scikit-Learn site or at the University of California at Irvine. + +

+However, if you would like to study other data sets, feel free to +propose other sets. What we list here are mere suggestions from our +side. If you opt for another data set, consider using a set which +has been studied in the scientific literature. This makes it easier +for you to compare and analyze your results. Comparing with existing results from the scientific literature is also an essential +element of the scientific discussion. The University of California at Irvine with its Machine Learning repository at https://archive.ics.uci.edu/ml/index.php is an excellent site to look up for examples and inspiration. Kaggle.com is an equally interesting site. Feel free to explore these sites. + +

+We will start with a regression problem and we will reuse our codes from project 1 starting with writing our own Stochastic Gradient Descent (SGD) code. + +

Part a): Write your own Stochastic Gradient Descent code, first step

+ +

+In order to get started, we will now replace in our standard ordinary +least squares (OLS) and Ridge regression codes (from project 1) the matrix inversion +algorithm with our own SGD code. You can choose whether you want to +add the momentum SGD optionality or other SGD variants such as RMSprop +or ADAgrad or ADAM. The lecture notes from week 40 contain more +details + +

+Perform an analysis of the results for OLS and Ridge regression as +function of the chosen learning rates, the number of mini-batches and +epochs as well as algorithm for scaling the learning rate. You can +also compare your own results with those that can be obtained using +for example Scikit-Learn's various SGD options. Discuss your +results. For Ridge regression you need now to study the results as functions of the hyper-parameter \( \lambda \) and +the learning rate \( \gamma \). Discuss your results. + +

+You will need your SGD code for the setup of the Neural Network and Logistic Regression codes. + +

Part b): Writing your own Neural Network code

+ +

+Your aim now, and this is the central part of this project, is to +write your own Feed Forward Neural Network code implementing the back +propagation algorithm discussed in the lecture slides from week 41. + +

+We will focus on a regression problem first and study either the +Franke function or terrain data (or both or other data sets) from +project 1. Discuss again your choice of cost function. + +

+Write an FFNN code for regression with a flexible number of hidden +layers and nodes using the Sigmoid function as activation function for +the hidden layers. Initialize the weights using a normal +distribution. How would you initialize the biases? And which +activation function would you select for the final output layer? + +

+Train your network and compare the results with those from your OLS and Ridge Regression codes from project 1. +You should test your results against a similar code using Scikit-Learn (see the examples in the above lecture notes from week 41) or tensorflow/keras. + +

+Comment your results and give a critical discussion of the results +obtained with the Linear Regression code and your own Neural Network +code. Compare the results with those from project 1. +Make an analysis of the regularization parameters and the learning rates employed to find the optimal MSE and \( R2 \) scores. + +

+A useful reference on the back progagation algorithm is Nielsen's +book. It is an excellent +read. + +

Part c): Testing different activation functions

+ +

+You should now also test different activation functions for the hidden layers. Try out the Sigmoid, the RELU and the Leaky RELU functions and discuss your results. You may also study the way you initialize your weights and biases. + +

Part d): Classification analysis using neural networks

+ +

+With a well-written code it should now be easy to change the +activation function for the output layer. + +

+Here we will change the cost function for our neural network code +developed in parts b) and c) in order to perform a classification analysis. + +

+We will here study the Wisconsin Breast Cancer data set. This is a typical binary classification problem with just one single output, either True or Fale, \( 0 \) or \( 1 \) etc. +You find more information about this at the Scikit-Learn +site or at the University of California +at Irvine. + +

+To measure the performance of our classification problem we use the +so-called accuracy score. The accuracy is as you would expect just +the number of correctly guessed targets \( t_i \) divided by the total +number of targets, that is + +$$ +\text{Accuracy} = \frac{\sum_{i=1}^n I(t_i = y_i)}{n} , +$$ + +

+where \( I \) is the indicator function, \( 1 \) if \( t_i = y_i \) and \( 0 \) +otherwise if we have a binary classification problem. Here \( t_i \) +represents the target and \( y_i \) the outputs of your FFNN code and \( n \) is simply the number of targets \( t_i \). + +

+Discuss your results and give a critical analysis of the various parameters, including hyper-parameters like the learning rates and the regularization parameter \( \lambda \) (as you did in Ridge Regression), various activation functions, number of hidden layers and nodes and activation functions. + +

+As stated in the introduction, it can also be useful to study other +datasets. + +

+Again, we strongly recommend that you compare your own neural Network +code for classification and pertinent results against a similar code using Scikit-Learn or tensorflow/keras or pytorch. + +

Part e): Write your Logistic Regression code, final step

+ +

+Finally, we want to compare the FFNN code we have developed with +Logistic regression, that is we wish to compare our neural network +classification results with the results we can obtain with another +method. + +

+Define your cost function and the design matrix before you start writing your code. +Write thereafter a Logistic regression code using your SGD algorithm. You can also use standard gradient descent in this case, with a learning rate as hyper-parameter. +Study the results as functions of the chosen learning rates. +Add also an \( l_2 \) regularization parameter \( \lambda \). Compare your results with those from your FFNN code as well as those obtained using Scikit-Learn's logistic regression functionality. + +

+The weblink here https://medium.com/ai-in-plain-english/comparison-between-logistic-regression-and-neural-networks-in-classifying-digits-dc5e85cd93c3compares logistic regression and FFNN using the so-called MNIST data set. You may find several useful hints and ideas from this article. + +

Part f) Critical evaluation of the various algorithms

+ +

+After all these glorious calculations, you should now summarize the +various algorithms and come with a critical evaluation of their pros +and cons. Which algorithm works best for the regression case and which +is best for the classification case. These codes can also be part of +your final project 3, but now applied to other data sets. + +

Background literature

+ +
    +
  1. The text of Michael Nielsen is highly recommended, see Nielsen's book. It is an excellent read.
    2. +
  2. The textbook of Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer, chapters 3 and 7 are the most relevant ones for the analysis here.
    3. +
  3. Mehta et al, arXiv 1803.08823, A high-bias, low-variance introduction to Machine Learning for physicists, ArXiv:1803.08823.
    4. +
+ +

Introduction to numerical projects

+ +

+Here follows a brief recipe and recommendation on how to write a report for each +project. + +

+ +

Format for electronic delivery of report and programs

+ +

+The preferred format for the report is a PDF file. You can also use DOC or postscript formats or as an ipython notebook file. As programming language we prefer that you choose between C/C++, Fortran2008 or Python. The following prescription should be followed when preparing the report: + +

+ +Finally, +we encourage you to collaborate. Optimal working groups consist of +2-3 students. You can then hand in a common report. + +

Software and needed installations

+ +

+If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, +we recommend that you install the following Python packages via pip as + +

    +
  1. pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow
    2. +
+ +For Python3, replace pip with pip3. + +

+See below for a discussion of tensorflow and scikit-learn. + +

+For OSX users we recommend also, after having installed Xcode, to install brew. Brew allows +for a seamless installation of additional software via for example + +

    +
  1. brew install python3
    2. +
+ +For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution +you can use pip as well and simply install Python as + +
    +
  1. sudo apt-get install python3 (or python for python2.7)
    2. +
+ +etc etc. + +

+If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely + +

    +
  1. Anaconda Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system conda
    2. +
  2. Enthought canopy is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license.
    3. +
+ +Popular software packages written in Python for ML are + + + +These are all freely available at their respective GitHub sites. They +encompass communities of developers in the thousands or more. And the number +of code developers and contributors keeps increasing. + +

+ +

+ +

+ + +
+ + + + + + + +
+ © 1999-2021, "Data Analysis and Machine Learning FYS-STK3155/FYS4155":"http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html". Released under CC Attribution-NonCommercial 4.0 license +
+ + + + + + diff --git a/doc/Projects/2021/Project2/html/Project2-bs.html b/doc/Projects/2021/Project2/html/Project2-bs.html new file mode 100644 index 000000000..f9c800422 --- /dev/null +++ b/doc/Projects/2021/Project2/html/Project2-bs.html @@ -0,0 +1,466 @@ + + + + + + + + +Project 2 on Machine Learning, deadline November 15 (Midnight) + + + + + + + + + + + + + + + + + + + + + + + +
+
+ + Project 2 on Machine Learning, deadline November 15 (Midnight) +
+ +
+ +
+
+ + +
+ +

 

 

 

+ + + + + + +
+

Project 2 on Machine Learning, deadline November 15 (Midnight)

+ +

+ + +

+Data Analysis and Machine Learning FYS-STK3155/FYS4155 +
+ +

+ + +

Department of Physics, University of Oslo, Norway
+
+

+

Oct 12, 2021

+
+

+

+ +

Classification and Regression, from linear and logistic regression to neural networks

+ +

+The main aim of this project is to study both classification and +regression problems by developing our own feed-forward neural network (FFNN) code. We can reuse the regression algorithms studied +in project 1. We will also include logistic regression for classification +problems and write our own FFNN code for studying +both regression and classification problems. The codes developed in +project 1, including bootstrap and/or cross-validation as well as the +computation of the mean-squared error and/or the \( R2 \) or the accuracy score (classification problems) functions can +also be utilized in the present analysis. + +

+Important Note: This project as well as project 3 have to be written as a scientific report. The instructions on how to do this and how we grade are available at https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md. We will discuss this format during the various lab sessions. Please do spend some time to read our guidelines. + +

+The data sets that we propose here are (the default sets) + +

+ +You can find more information about this at the Scikit-Learn site or at the University of California at Irvine. + +

+However, if you would like to study other data sets, feel free to +propose other sets. What we list here are mere suggestions from our +side. If you opt for another data set, consider using a set which +has been studied in the scientific literature. This makes it easier +for you to compare and analyze your results. Comparing with existing results from the scientific literature is also an essential +element of the scientific discussion. The University of California at Irvine with its Machine Learning repository at https://archive.ics.uci.edu/ml/index.php is an excellent site to look up for examples and inspiration. Kaggle.com is an equally interesting site. Feel free to explore these sites. + +

+We will start with a regression problem and we will reuse our codes from project 1 starting with writing our own Stochastic Gradient Descent (SGD) code. + +

Part a): Write your own Stochastic Gradient Descent code, first step

+ +

+In order to get started, we will now replace in our standard ordinary +least squares (OLS) and Ridge regression codes (from project 1) the matrix inversion +algorithm with our own SGD code. You can choose whether you want to +add the momentum SGD optionality or other SGD variants such as RMSprop +or ADAgrad or ADAM. The lecture notes from week 40 contain more +details + +

+Perform an analysis of the results for OLS and Ridge regression as +function of the chosen learning rates, the number of mini-batches and +epochs as well as algorithm for scaling the learning rate. You can +also compare your own results with those that can be obtained using +for example Scikit-Learn's various SGD options. Discuss your +results. For Ridge regression you need now to study the results as functions of the hyper-parameter \( \lambda \) and +the learning rate \( \gamma \). Discuss your results. + +

+You will need your SGD code for the setup of the Neural Network and Logistic Regression codes. + +

Part b): Writing your own Neural Network code

+ +

+Your aim now, and this is the central part of this project, is to +write your own Feed Forward Neural Network code implementing the back +propagation algorithm discussed in the lecture slides from week 41. + +

+We will focus on a regression problem first and study either the +Franke function or terrain data (or both or other data sets) from +project 1. Discuss again your choice of cost function. + +

+Write an FFNN code for regression with a flexible number of hidden +layers and nodes using the Sigmoid function as activation function for +the hidden layers. Initialize the weights using a normal +distribution. How would you initialize the biases? And which +activation function would you select for the final output layer? + +

+Train your network and compare the results with those from your OLS and Ridge Regression codes from project 1. +You should test your results against a similar code using Scikit-Learn (see the examples in the above lecture notes from week 41) or tensorflow/keras. + +

+Comment your results and give a critical discussion of the results +obtained with the Linear Regression code and your own Neural Network +code. Compare the results with those from project 1. +Make an analysis of the regularization parameters and the learning rates employed to find the optimal MSE and \( R2 \) scores. + +

+A useful reference on the back progagation algorithm is Nielsen's +book. It is an excellent +read. + +

Part c): Testing different activation functions

+ +

+You should now also test different activation functions for the hidden layers. Try out the Sigmoid, the RELU and the Leaky RELU functions and discuss your results. You may also study the way you initialize your weights and biases. + +

Part d): Classification analysis using neural networks

+ +

+With a well-written code it should now be easy to change the +activation function for the output layer. + +

+Here we will change the cost function for our neural network code +developed in parts b) and c) in order to perform a classification analysis. + +

+We will here study the Wisconsin Breast Cancer data set. This is a typical binary classification problem with just one single output, either True or Fale, \( 0 \) or \( 1 \) etc. +You find more information about this at the Scikit-Learn +site or at the University of California +at Irvine. + +

+To measure the performance of our classification problem we use the +so-called accuracy score. The accuracy is as you would expect just +the number of correctly guessed targets \( t_i \) divided by the total +number of targets, that is + +$$ +\text{Accuracy} = \frac{\sum_{i=1}^n I(t_i = y_i)}{n} , +$$ + +

+where \( I \) is the indicator function, \( 1 \) if \( t_i = y_i \) and \( 0 \) +otherwise if we have a binary classification problem. Here \( t_i \) +represents the target and \( y_i \) the outputs of your FFNN code and \( n \) is simply the number of targets \( t_i \). + +

+Discuss your results and give a critical analysis of the various parameters, including hyper-parameters like the learning rates and the regularization parameter \( \lambda \) (as you did in Ridge Regression), various activation functions, number of hidden layers and nodes and activation functions. + +

+As stated in the introduction, it can also be useful to study other +datasets. + +

+Again, we strongly recommend that you compare your own neural Network +code for classification and pertinent results against a similar code using Scikit-Learn or tensorflow/keras or pytorch. + +

Part e): Write your Logistic Regression code, final step

+ +

+Finally, we want to compare the FFNN code we have developed with +Logistic regression, that is we wish to compare our neural network +classification results with the results we can obtain with another +method. + +

+Define your cost function and the design matrix before you start writing your code. +Write thereafter a Logistic regression code using your SGD algorithm. You can also use standard gradient descent in this case, with a learning rate as hyper-parameter. +Study the results as functions of the chosen learning rates. +Add also an \( l_2 \) regularization parameter \( \lambda \). Compare your results with those from your FFNN code as well as those obtained using Scikit-Learn's logistic regression functionality. + +

+The weblink here https://medium.com/ai-in-plain-english/comparison-between-logistic-regression-and-neural-networks-in-classifying-digits-dc5e85cd93c3compares logistic regression and FFNN using the so-called MNIST data set. You may find several useful hints and ideas from this article. + +

Part f) Critical evaluation of the various algorithms

+ +

+After all these glorious calculations, you should now summarize the +various algorithms and come with a critical evaluation of their pros +and cons. Which algorithm works best for the regression case and which +is best for the classification case. These codes can also be part of +your final project 3, but now applied to other data sets. + +

Background literature

+ +
    +
  1. The text of Michael Nielsen is highly recommended, see Nielsen's book. It is an excellent read.
    2. +
  2. The textbook of Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer, chapters 3 and 7 are the most relevant ones for the analysis here.
    3. +
  3. Mehta et al, arXiv 1803.08823, A high-bias, low-variance introduction to Machine Learning for physicists, ArXiv:1803.08823.
    4. +
+ +

Introduction to numerical projects

+ +

+Here follows a brief recipe and recommendation on how to write a report for each +project. + +

+ +

Format for electronic delivery of report and programs

+ +

+The preferred format for the report is a PDF file. You can also use DOC or postscript formats or as an ipython notebook file. As programming language we prefer that you choose between C/C++, Fortran2008 or Python. The following prescription should be followed when preparing the report: + +

+ +Finally, +we encourage you to collaborate. Optimal working groups consist of +2-3 students. You can then hand in a common report. + +

Software and needed installations

+ +

+If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, +we recommend that you install the following Python packages via pip as + +

    +
  1. pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow
    2. +
+ +For Python3, replace pip with pip3. + +

+See below for a discussion of tensorflow and scikit-learn. + +

+For OSX users we recommend also, after having installed Xcode, to install brew. Brew allows +for a seamless installation of additional software via for example + +

    +
  1. brew install python3
    2. +
+ +For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution +you can use pip as well and simply install Python as + +
    +
  1. sudo apt-get install python3 (or python for python2.7)
    2. +
+ +etc etc. + +

+If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely + +

    +
  1. Anaconda Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system conda
    2. +
  2. Enthought canopy is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license.
    3. +
+ +Popular software packages written in Python for ML are + + + +These are all freely available at their respective GitHub sites. They +encompass communities of developers in the thousands or more. And the number +of code developers and contributors keeps increasing. + +

+ +

+ +

+ + +
+ + + + + + + +
+ © 1999-2021, "Data Analysis and Machine Learning FYS-STK3155/FYS4155":"http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html". Released under CC Attribution-NonCommercial 4.0 license +
+ + + + + + diff --git a/doc/Projects/2021/Project2/html/Project2.html b/doc/Projects/2021/Project2/html/Project2.html new file mode 100644 index 000000000..33aa54c60 --- /dev/null +++ b/doc/Projects/2021/Project2/html/Project2.html @@ -0,0 +1,406 @@ + + + + + + + + +Project 2 on Machine Learning, deadline November 15 (Midnight) + + + + + + + + + + + + + + + + + + + + + + + +

Project 2 on Machine Learning, deadline November 15 (Midnight)

+ +

+ + +

+Data Analysis and Machine Learning FYS-STK3155/FYS4155 +
+ +

+ + +

Department of Physics, University of Oslo, Norway
+
+

+

Oct 12, 2021

+
+ +

Classification and Regression, from linear and logistic regression to neural networks

+ +

+The main aim of this project is to study both classification and +regression problems by developing our own feed-forward neural network (FFNN) code. We can reuse the regression algorithms studied +in project 1. We will also include logistic regression for classification +problems and write our own FFNN code for studying +both regression and classification problems. The codes developed in +project 1, including bootstrap and/or cross-validation as well as the +computation of the mean-squared error and/or the \( R2 \) or the accuracy score (classification problems) functions can +also be utilized in the present analysis. + +

+Important Note: This project as well as project 3 have to be written as a scientific report. The instructions on how to do this and how we grade are available at https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md. We will discuss this format during the various lab sessions. Please do spend some time to read our guidelines. + +

+The data sets that we propose here are (the default sets) + +

+ +You can find more information about this at the Scikit-Learn site or at the University of California at Irvine. + +

+However, if you would like to study other data sets, feel free to +propose other sets. What we list here are mere suggestions from our +side. If you opt for another data set, consider using a set which +has been studied in the scientific literature. This makes it easier +for you to compare and analyze your results. Comparing with existing results from the scientific literature is also an essential +element of the scientific discussion. The University of California at Irvine with its Machine Learning repository at https://archive.ics.uci.edu/ml/index.php is an excellent site to look up for examples and inspiration. Kaggle.com is an equally interesting site. Feel free to explore these sites. + +

+We will start with a regression problem and we will reuse our codes from project 1 starting with writing our own Stochastic Gradient Descent (SGD) code. + +

Part a): Write your own Stochastic Gradient Descent code, first step

+ +

+In order to get started, we will now replace in our standard ordinary +least squares (OLS) and Ridge regression codes (from project 1) the matrix inversion +algorithm with our own SGD code. You can choose whether you want to +add the momentum SGD optionality or other SGD variants such as RMSprop +or ADAgrad or ADAM. The lecture notes from week 40 contain more +details + +

+Perform an analysis of the results for OLS and Ridge regression as +function of the chosen learning rates, the number of mini-batches and +epochs as well as algorithm for scaling the learning rate. You can +also compare your own results with those that can be obtained using +for example Scikit-Learn's various SGD options. Discuss your +results. For Ridge regression you need now to study the results as functions of the hyper-parameter \( \lambda \) and +the learning rate \( \gamma \). Discuss your results. + +

+You will need your SGD code for the setup of the Neural Network and Logistic Regression codes. + +

Part b): Writing your own Neural Network code

+ +

+Your aim now, and this is the central part of this project, is to +write your own Feed Forward Neural Network code implementing the back +propagation algorithm discussed in the lecture slides from week 41. + +

+We will focus on a regression problem first and study either the +Franke function or terrain data (or both or other data sets) from +project 1. Discuss again your choice of cost function. + +

+Write an FFNN code for regression with a flexible number of hidden +layers and nodes using the Sigmoid function as activation function for +the hidden layers. Initialize the weights using a normal +distribution. How would you initialize the biases? And which +activation function would you select for the final output layer? + +

+Train your network and compare the results with those from your OLS and Ridge Regression codes from project 1. +You should test your results against a similar code using Scikit-Learn (see the examples in the above lecture notes from week 41) or tensorflow/keras. + +

+Comment your results and give a critical discussion of the results +obtained with the Linear Regression code and your own Neural Network +code. Compare the results with those from project 1. +Make an analysis of the regularization parameters and the learning rates employed to find the optimal MSE and \( R2 \) scores. + +

+A useful reference on the back progagation algorithm is Nielsen's +book. It is an excellent +read. + +

Part c): Testing different activation functions

+ +

+You should now also test different activation functions for the hidden layers. Try out the Sigmoid, the RELU and the Leaky RELU functions and discuss your results. You may also study the way you initialize your weights and biases. + +

Part d): Classification analysis using neural networks

+ +

+With a well-written code it should now be easy to change the +activation function for the output layer. + +

+Here we will change the cost function for our neural network code +developed in parts b) and c) in order to perform a classification analysis. + +

+We will here study the Wisconsin Breast Cancer data set. This is a typical binary classification problem with just one single output, either True or Fale, \( 0 \) or \( 1 \) etc. +You find more information about this at the Scikit-Learn +site or at the University of California +at Irvine. + +

+To measure the performance of our classification problem we use the +so-called accuracy score. The accuracy is as you would expect just +the number of correctly guessed targets \( t_i \) divided by the total +number of targets, that is + +$$ +\text{Accuracy} = \frac{\sum_{i=1}^n I(t_i = y_i)}{n} , +$$ + +

+where \( I \) is the indicator function, \( 1 \) if \( t_i = y_i \) and \( 0 \) +otherwise if we have a binary classification problem. Here \( t_i \) +represents the target and \( y_i \) the outputs of your FFNN code and \( n \) is simply the number of targets \( t_i \). + +

+Discuss your results and give a critical analysis of the various parameters, including hyper-parameters like the learning rates and the regularization parameter \( \lambda \) (as you did in Ridge Regression), various activation functions, number of hidden layers and nodes and activation functions. + +

+As stated in the introduction, it can also be useful to study other +datasets. + +

+Again, we strongly recommend that you compare your own neural Network +code for classification and pertinent results against a similar code using Scikit-Learn or tensorflow/keras or pytorch. + +

Part e): Write your Logistic Regression code, final step

+ +

+Finally, we want to compare the FFNN code we have developed with +Logistic regression, that is we wish to compare our neural network +classification results with the results we can obtain with another +method. + +

+Define your cost function and the design matrix before you start writing your code. +Write thereafter a Logistic regression code using your SGD algorithm. You can also use standard gradient descent in this case, with a learning rate as hyper-parameter. +Study the results as functions of the chosen learning rates. +Add also an \( l_2 \) regularization parameter \( \lambda \). Compare your results with those from your FFNN code as well as those obtained using Scikit-Learn's logistic regression functionality. + +

+The weblink here https://medium.com/ai-in-plain-english/comparison-between-logistic-regression-and-neural-networks-in-classifying-digits-dc5e85cd93c3compares logistic regression and FFNN using the so-called MNIST data set. You may find several useful hints and ideas from this article. + +

Part f) Critical evaluation of the various algorithms

+ +

+After all these glorious calculations, you should now summarize the +various algorithms and come with a critical evaluation of their pros +and cons. Which algorithm works best for the regression case and which +is best for the classification case. These codes can also be part of +your final project 3, but now applied to other data sets. + +

Background literature

+ +
    +
  1. The text of Michael Nielsen is highly recommended, see Nielsen's book. It is an excellent read.
    2. +
  2. The textbook of Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer, chapters 3 and 7 are the most relevant ones for the analysis here.
    3. +
  3. Mehta et al, arXiv 1803.08823, A high-bias, low-variance introduction to Machine Learning for physicists, ArXiv:1803.08823.
    4. +
+ +

Introduction to numerical projects

+ +

+Here follows a brief recipe and recommendation on how to write a report for each +project. + +

+ +

Format for electronic delivery of report and programs

+ +

+The preferred format for the report is a PDF file. You can also use DOC or postscript formats or as an ipython notebook file. As programming language we prefer that you choose between C/C++, Fortran2008 or Python. The following prescription should be followed when preparing the report: + +

+ +Finally, +we encourage you to collaborate. Optimal working groups consist of +2-3 students. You can then hand in a common report. + +

Software and needed installations

+ +

+If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, +we recommend that you install the following Python packages via pip as + +

    +
  1. pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow
    2. +
+ +For Python3, replace pip with pip3. + +

+See below for a discussion of tensorflow and scikit-learn. + +

+For OSX users we recommend also, after having installed Xcode, to install brew. Brew allows +for a seamless installation of additional software via for example + +

    +
  1. brew install python3
    2. +
+ +For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution +you can use pip as well and simply install Python as + +
    +
  1. sudo apt-get install python3 (or python for python2.7)
    2. +
+ +etc etc. + +

+If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely + +

    +
  1. Anaconda Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system conda
    2. +
  2. Enthought canopy is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license.
    3. +
+ +Popular software packages written in Python for ML are + + + +These are all freely available at their respective GitHub sites. They +encompass communities of developers in the thousands or more. And the number +of code developers and contributors keeps increasing. + +

+ + + + +

+ © 1999-2021, "Data Analysis and Machine Learning FYS-STK3155/FYS4155":"http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html". Released under CC Attribution-NonCommercial 4.0 license +
+ + + + + + diff --git a/doc/Projects/2021/Project2/ipynb/Project2.ipynb b/doc/Projects/2021/Project2/ipynb/Project2.ipynb new file mode 100644 index 000000000..cd41515a6 --- /dev/null +++ b/doc/Projects/2021/Project2/ipynb/Project2.ipynb @@ -0,0 +1,275 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "\n", + "# Project 2 on Machine Learning, deadline November 15 (Midnight)\n", + "\n", + " \n", + "**[Data Analysis and Machine Learning FYS-STK3155/FYS4155](http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html)**, Department of Physics, University of Oslo, Norway\n", + "\n", + "Date: **Oct 12, 2021**\n", + "\n", + "Copyright 1999-2021, [Data Analysis and Machine Learning FYS-STK3155/FYS4155](http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html). Released under CC Attribution-NonCommercial 4.0 license\n", + "\n", + "\n", + "\n", + "\n", + "## Classification and Regression, from linear and logistic regression to neural networks\n", + "\n", + "The main aim of this project is to study both classification and\n", + "regression problems by developing our own feed-forward neural network (FFNN) code. We can reuse the regression algorithms studied\n", + "in project 1. We will also include logistic regression for classification\n", + "problems and write our own FFNN code for studying\n", + "both regression and classification problems. The codes developed in\n", + "project 1, including bootstrap **and/or** cross-validation as well as the\n", + "computation of the mean-squared error and/or the $R2$ or the accuracy score (classification problems) functions can\n", + "also be utilized in the present analysis. \n", + "\n", + "**Important Note**: This project as well as project 3 have to be written as a scientific report. The instructions on how to do this and how we grade are available at . We will discuss this format during the various lab sessions. Please do spend some time to read our guidelines.\n", + "\n", + "The data sets that we propose here are (the default sets)\n", + "\n", + "* Regression (fitting a continuous function). In this part you will need to bring back your results from project 1 and compare these with what you get from your Neural Network code to be developed here. The data sets could be\n", + "\n", + "a. Either the Franke function or the terrain data from project 1, or data sets your propose.\n", + "\n", + "\n", + "* Classification. Here you will also need to develop a Logistic regression code that you will use to compare with the Neural Network code. The data set we propose are the so-called [Wisconsin Breat Cancer Data](https://www.kaggle.com/uciml/breast-cancer-wisconsin-data) data set of images representing various features of tumors. These are discussed intensively in the lecture notes on neural networks, see for example the slides from [week 40](https://compphysics.github.io/MachineLearning/doc/pub/week41/html/week40.html). A longer explanation with links to the scientific literature can be found at the [Machine Learning repository of the University of California at Irvine](https://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+%28Diagnostic%29). Feel free to consult this site and the pertinent literature.\n", + "\n", + "You can find more information about this at the [Scikit-Learn site](https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_breast_cancer.html) or at the [University of California at Irvine](https://archive.ics.uci.edu/ml/datasets/breast+cancer+wisconsin+(original)). \n", + "\n", + "\n", + "However, if you would like to study other data sets, feel free to\n", + "propose other sets. What we list here are mere suggestions from our\n", + "side. If you opt for another data set, consider using a set which\n", + "has been studied in the scientific literature. This makes it easier\n", + "for you to compare and analyze your results. Comparing with existing results from the scientific literature is also an essential\n", + "element of the scientific discussion. The University of California at Irvine with its Machine Learning repository at is an excellent site to look up for examples and inspiration. [Kaggle.com](https://www.kaggle.com/) is an equally interesting site. Feel free to explore these sites.\n", + "\n", + "\n", + "We will start with a regression problem and we will reuse our codes from project 1 starting with writing our own Stochastic Gradient Descent (SGD) code. \n", + "\n", + "### Part a): Write your own Stochastic Gradient Descent code, first step\n", + "\n", + "In order to get started, we will now replace in our standard ordinary\n", + "least squares (OLS) and Ridge regression codes (from project 1) the matrix inversion\n", + "algorithm with our own SGD code. You can choose whether you want to\n", + "add the momentum SGD optionality or other SGD variants such as RMSprop\n", + "or ADAgrad or ADAM. The lecture notes from [week 40 contain more\n", + "details](https://compphysics.github.io/MachineLearning/doc/pub/week40/html/week40.html)\n", + "\n", + "Perform an analysis of the results for OLS and Ridge regression as\n", + "function of the chosen learning rates, the number of mini-batches and\n", + "epochs as well as algorithm for scaling the learning rate. You can\n", + "also compare your own results with those that can be obtained using\n", + "for example **Scikit-Learn**'s various SGD options. Discuss your\n", + "results. For Ridge regression you need now to study the results as functions of the hyper-parameter $\\lambda$ and \n", + "the learning rate $\\gamma$. Discuss your results.\n", + "\n", + "You will need your SGD code for the setup of the Neural Network and Logistic Regression codes.\n", + "\n", + "### Part b): Writing your own Neural Network code\n", + "\n", + "Your aim now, and this is the central part of this project, is to\n", + "write your own Feed Forward Neural Network code implementing the back\n", + "propagation algorithm discussed in the lecture slides from [week 41](https://compphysics.github.io/MachineLearning/doc/pub/week41/html/week41.html).\n", + "\n", + "We will focus on a regression problem first and study either the\n", + "Franke function or terrain data (or both or other data sets) from\n", + "project 1. Discuss again your choice of cost function.\n", + "\n", + "Write an FFNN code for regression with a flexible number of hidden\n", + "layers and nodes using the Sigmoid function as activation function for\n", + "the hidden layers. Initialize the weights using a normal\n", + "distribution. How would you initialize the biases? And which\n", + "activation function would you select for the final output layer?\n", + "\n", + "Train your network and compare the results with those from your OLS and Ridge Regression codes from project 1. \n", + "You should test your results against a similar code using **Scikit-Learn** (see the examples in the above lecture notes from week 41) or **tensorflow/keras**. \n", + "\n", + "Comment your results and give a critical discussion of the results\n", + "obtained with the Linear Regression code and your own Neural Network\n", + "code. Compare the results with those from project 1.\n", + "Make an analysis of the regularization parameters and the learning rates employed to find the optimal MSE and $R2$ scores.\n", + "\n", + "A useful reference on the back progagation algorithm is [Nielsen's\n", + "book](http://neuralnetworksanddeeplearning.com/). It is an excellent\n", + "read.\n", + "\n", + "\n", + "\n", + "### Part c): Testing different activation functions\n", + "\n", + "You should now also test different activation functions for the hidden layers. Try out the Sigmoid, the RELU and the Leaky RELU functions and discuss your results. You may also study the way you initialize your weights and biases.\n", + "\n", + "### Part d): Classification analysis using neural networks\n", + "\n", + "With a well-written code it should now be easy to change the\n", + "activation function for the output layer.\n", + "\n", + "Here we will change the cost function for our neural network code\n", + "developed in parts b) and c) in order to perform a classification analysis. \n", + "\n", + "We will here study the Wisconsin Breast Cancer data set. This is a typical binary classification problem with just one single output, either True or Fale, $0$ or $1$ etc.\n", + "You find more information about this at the [Scikit-Learn\n", + "site](https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_breast_cancer.html) or at the [University of California\n", + "at Irvine](https://archive.ics.uci.edu/ml/datasets/breast+cancer+wisconsin+(original)). \n", + "\n", + "\n", + "To measure the performance of our classification problem we use the\n", + "so-called *accuracy* score. The accuracy is as you would expect just\n", + "the number of correctly guessed targets $t_i$ divided by the total\n", + "number of targets, that is" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "$$\n", + "\\text{Accuracy} = \\frac{\\sum_{i=1}^n I(t_i = y_i)}{n} ,\n", + "$$" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "where $I$ is the indicator function, $1$ if $t_i = y_i$ and $0$\n", + "otherwise if we have a binary classification problem. Here $t_i$\n", + "represents the target and $y_i$ the outputs of your FFNN code and $n$ is simply the number of targets $t_i$.\n", + "\n", + "Discuss your results and give a critical analysis of the various parameters, including hyper-parameters like the learning rates and the regularization parameter $\\lambda$ (as you did in Ridge Regression), various activation functions, number of hidden layers and nodes and activation functions. \n", + "\n", + "\n", + "As stated in the introduction, it can also be useful to study other\n", + "datasets. \n", + "\n", + "Again, we strongly recommend that you compare your own neural Network\n", + "code for classification and pertinent results against a similar code using **Scikit-Learn** or **tensorflow/keras** or **pytorch**.\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "### Part e): Write your Logistic Regression code, final step\n", + "\n", + "Finally, we want to compare the FFNN code we have developed with\n", + "Logistic regression, that is we wish to compare our neural network\n", + "classification results with the results we can obtain with another\n", + "method.\n", + "\n", + "Define your cost function and the design matrix before you start writing your code.\n", + "Write thereafter a Logistic regression code using your SGD algorithm. You can also use standard gradient descent in this case, with a learning rate as hyper-parameter.\n", + "Study the results as functions of the chosen learning rates.\n", + "Add also an $l_2$ regularization parameter $\\lambda$. Compare your results with those from your FFNN code as well as those obtained using **Scikit-Learn**'s logistic regression functionality.\n", + "\n", + "The weblink here compares logistic regression and FFNN using the so-called MNIST data set. You may find several useful hints and ideas from this article. \n", + "\n", + "\n", + "### Part f) Critical evaluation of the various algorithms\n", + "\n", + "After all these glorious calculations, you should now summarize the\n", + "various algorithms and come with a critical evaluation of their pros\n", + "and cons. Which algorithm works best for the regression case and which\n", + "is best for the classification case. These codes can also be part of\n", + "your final project 3, but now applied to other data sets.\n", + "\n", + "\n", + "\n", + "\n", + "## Background literature\n", + "\n", + "1. The text of Michael Nielsen is highly recommended, see [Nielsen's book](http://neuralnetworksanddeeplearning.com/). It is an excellent read.\n", + "\n", + "2. The textbook of [Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer](https://www.springer.com/gp/book/9780387848570), chapters 3 and 7 are the most relevant ones for the analysis here. \n", + "\n", + "3. [Mehta et al, arXiv 1803.08823](https://arxiv.org/abs/1803.08823), *A high-bias, low-variance introduction to Machine Learning for physicists*, ArXiv:1803.08823.\n", + "\n", + "## Introduction to numerical projects\n", + "\n", + "Here follows a brief recipe and recommendation on how to write a report for each\n", + "project.\n", + "\n", + " * Give a short description of the nature of the problem and the eventual numerical methods you have used.\n", + "\n", + " * Describe the algorithm you have used and/or developed. Here you may find it convenient to use pseudocoding. In many cases you can describe the algorithm in the program itself.\n", + "\n", + " * Include the source code of your program. Comment your program properly.\n", + "\n", + " * If possible, try to find analytic solutions, or known limits in order to test your program when developing the code.\n", + "\n", + " * Include your results either in figure form or in a table. Remember to label your results. All tables and figures should have relevant captions and labels on the axes.\n", + "\n", + " * Try to evaluate the reliabilty and numerical stability/precision of your results. If possible, include a qualitative and/or quantitative discussion of the numerical stability, eventual loss of precision etc.\n", + "\n", + " * Try to give an interpretation of you results in your answers to the problems.\n", + "\n", + " * Critique: if possible include your comments and reflections about the exercise, whether you felt you learnt something, ideas for improvements and other thoughts you've made when solving the exercise. We wish to keep this course at the interactive level and your comments can help us improve it.\n", + "\n", + " * Try to establish a practice where you log your work at the computerlab. You may find such a logbook very handy at later stages in your work, especially when you don't properly remember what a previous test version of your program did. Here you could also record the time spent on solving the exercise, various algorithms you may have tested or other topics which you feel worthy of mentioning.\n", + "\n", + "## Format for electronic delivery of report and programs\n", + "\n", + "The preferred format for the report is a PDF file. You can also use DOC or postscript formats or as an ipython notebook file. As programming language we prefer that you choose between C/C++, Fortran2008 or Python. The following prescription should be followed when preparing the report:\n", + "\n", + " * Use Canvas to hand in your projects, log in at with your normal UiO username and password.\n", + "\n", + " * Upload **only** the report file or the link to your GitHub/GitLab or similar typo of repos! For the source code file(s) you have developed please provide us with your link to your GitHub/GitLab or similar domain. The report file should include all of your discussions and a list of the codes you have developed. Do not include library files which are available at the course homepage, unless you have made specific changes to them.\n", + "\n", + " * In your GitHub/GitLab or similar repository, please include a folder which contains selected results. These can be in the form of output from your code for a selected set of runs and input parameters.\n", + "\n", + "Finally, \n", + "we encourage you to collaborate. Optimal working groups consist of \n", + "2-3 students. You can then hand in a common report. \n", + "\n", + "\n", + "\n", + "## Software and needed installations\n", + "\n", + "If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, \n", + "we recommend that you install the following Python packages via **pip** as\n", + "1. pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow\n", + "\n", + "For Python3, replace **pip** with **pip3**.\n", + "\n", + "See below for a discussion of **tensorflow** and **scikit-learn**. \n", + "\n", + "For OSX users we recommend also, after having installed Xcode, to install **brew**. Brew allows \n", + "for a seamless installation of additional software via for example\n", + "1. brew install python3\n", + "\n", + "For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution\n", + "you can use **pip** as well and simply install Python as \n", + "1. sudo apt-get install python3 (or python for python2.7)\n", + "\n", + "etc etc. \n", + "\n", + "If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely\n", + "1. [Anaconda](https://docs.anaconda.com/) Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system **conda**\n", + "\n", + "2. [Enthought canopy](https://www.enthought.com/product/canopy/) is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license.\n", + "\n", + "Popular software packages written in Python for ML are\n", + "\n", + "* [Scikit-learn](http://scikit-learn.org/stable/), \n", + "\n", + "* [Tensorflow](https://www.tensorflow.org/),\n", + "\n", + "* [PyTorch](http://pytorch.org/) and \n", + "\n", + "* [Keras](https://keras.io/).\n", + "\n", + "These are all freely available at their respective GitHub sites. They \n", + "encompass communities of developers in the thousands or more. And the number\n", + "of code developers and contributors keeps increasing." + ] + } + ], + "metadata": {}, + "nbformat": 4, + "nbformat_minor": 4 +} diff --git a/doc/Projects/2021/Project2/ipynb/ipynb-Project2-src.tar.gz b/doc/Projects/2021/Project2/ipynb/ipynb-Project2-src.tar.gz new file mode 100644 index 0000000000000000000000000000000000000000..7daed6b8392902b06564a415f318e80e1a7579ac GIT binary patch literal 194 zcmV;z06qU7iwFQOHDzG{1MSaC3c@fD2H>uHia9|^(zIC%cHu%O;ssKh+Ekm=Bn5kW z`v6@jZi)!`Hb27*!^|PuZ1-8@?k-pjLMWvaretX{Cn8Hd!cx$8Z*Y9|q=Xqax01+e+*#HOt0AZR{&;S4c literal 0 HcmV?d00001 diff --git a/doc/Projects/2021/Project2/pdf/Project2.p.tex b/doc/Projects/2021/Project2/pdf/Project2.p.tex new file mode 100644 index 000000000..4668c209d --- /dev/null +++ b/doc/Projects/2021/Project2/pdf/Project2.p.tex @@ -0,0 +1,433 @@ +%% +%% Automatically generated file from DocOnce source +%% (https://github.com/hplgit/doconce/) +%% +%% +% #ifdef PTEX2TEX_EXPLANATION +%% +%% The file follows the ptex2tex extended LaTeX format, see +%% ptex2tex: https://code.google.com/p/ptex2tex/ +%% +%% Run +%% ptex2tex myfile +%% or +%% doconce ptex2tex myfile +%% +%% to turn myfile.p.tex into an ordinary LaTeX file myfile.tex. +%% (The ptex2tex program: https://code.google.com/p/ptex2tex) +%% Many preprocess options can be added to ptex2tex or doconce ptex2tex +%% +%% ptex2tex -DMINTED myfile +%% doconce ptex2tex myfile envir=minted +%% +%% ptex2tex will typeset code environments according to a global or local +%% .ptex2tex.cfg configure file. doconce ptex2tex will typeset code +%% according to options on the command line (just type doconce ptex2tex to +%% see examples). If doconce ptex2tex has envir=minted, it enables the +%% minted style without needing -DMINTED. +% #endif + +% #define PREAMBLE + +% #ifdef PREAMBLE +%-------------------- begin preamble ---------------------- + +\documentclass[% +oneside, % oneside: electronic viewing, twoside: printing +final, % draft: marks overfull hboxes, figures with paths +10pt]{article} + +\listfiles % print all files needed to compile this document + +\usepackage{relsize,makeidx,color,setspace,amsmath,amsfonts,amssymb} +\usepackage[table]{xcolor} +\usepackage{bm,ltablex,microtype} + +\usepackage[pdftex]{graphicx} + +\usepackage[T1]{fontenc} +%\usepackage[latin1]{inputenc} +\usepackage{ucs} +\usepackage[utf8x]{inputenc} + +\usepackage{lmodern} % Latin Modern fonts derived from Computer Modern + +% Hyperlinks in PDF: +\definecolor{linkcolor}{rgb}{0,0,0.4} +\usepackage{hyperref} +\hypersetup{ + breaklinks=true, + colorlinks=true, + linkcolor=linkcolor, + urlcolor=linkcolor, + citecolor=black, + filecolor=black, + %filecolor=blue, + pdfmenubar=true, + pdftoolbar=true, + bookmarksdepth=3 % Uncomment (and tweak) for PDF bookmarks with more levels than the TOC + } +%\hyperbaseurl{} % hyperlinks are relative to this root + +\setcounter{tocdepth}{2} % levels in table of contents + +% --- fancyhdr package for fancy headers --- +\usepackage{fancyhdr} +\fancyhf{} % sets both header and footer to nothing +\renewcommand{\headrulewidth}{0pt} +\fancyfoot[LE,RO]{\thepage} +% Ensure copyright on titlepage (article style) and chapter pages (book style) +\fancypagestyle{plain}{ + \fancyhf{} + \fancyfoot[C]{{\footnotesize \copyright\ 1999-2021, "Data Analysis and Machine Learning FYS-STK3155/FYS4155":"http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html". Released under CC Attribution-NonCommercial 4.0 license}} +% \renewcommand{\footrulewidth}{0mm} + \renewcommand{\headrulewidth}{0mm} +} +% Ensure copyright on titlepages with \thispagestyle{empty} +\fancypagestyle{empty}{ + \fancyhf{} + \fancyfoot[C]{{\footnotesize \copyright\ 1999-2021, "Data Analysis and Machine Learning FYS-STK3155/FYS4155":"http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html". Released under CC Attribution-NonCommercial 4.0 license}} + \renewcommand{\footrulewidth}{0mm} + \renewcommand{\headrulewidth}{0mm} +} + +\pagestyle{fancy} + + +% prevent orhpans and widows +\clubpenalty = 10000 +\widowpenalty = 10000 + +% --- end of standard preamble for documents --- + + +% insert custom LaTeX commands... + +\raggedbottom +\makeindex +\usepackage[totoc]{idxlayout} % for index in the toc +\usepackage[nottoc]{tocbibind} % for references/bibliography in the toc + +%-------------------- end preamble ---------------------- + +\begin{document} + +% matching end for #ifdef PREAMBLE +% #endif + +\newcommand{\exercisesection}[1]{\subsection*{#1}} + + +% ------------------- main content ---------------------- + + + +% ----------------- title ------------------------- + +\thispagestyle{empty} + +\begin{center} +{\LARGE\bf +\begin{spacing}{1.25} +Project 2 on Machine Learning, deadline November 15 (Midnight) +\end{spacing} +} +\end{center} + +% ----------------- author(s) ------------------------- + +\begin{center} +{\bf \href{{http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html}}{Data Analysis and Machine Learning FYS-STK3155/FYS4155}} +\end{center} + + \begin{center} +% List of all institutions: +\centerline{{\small Department of Physics, University of Oslo, Norway}} +\end{center} + +% ----------------- end author(s) ------------------------- + +% --- begin date --- +\begin{center} +Oct 12, 2021 +\end{center} +% --- end date --- + +\vspace{1cm} + + +\subsection{Classification and Regression, from linear and logistic regression to neural networks} + +The main aim of this project is to study both classification and +regression problems by developing our own feed-forward neural network (FFNN) code. We can reuse the regression algorithms studied +in project 1. We will also include logistic regression for classification +problems and write our own FFNN code for studying +both regression and classification problems. The codes developed in +project 1, including bootstrap \textbf{and/or} cross-validation as well as the +computation of the mean-squared error and/or the $R2$ or the accuracy score (classification problems) functions can +also be utilized in the present analysis. + +\textbf{Important Note}: This project as well as project 3 have to be written as a scientific report. The instructions on how to do this and how we grade are available at \href{{https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md}}{\nolinkurl{https://github.com/CompPhysics/MachineLearning/blob/master/doc/Projects/EvaluationGrading/EvaluationForm.md}}. We will discuss this format during the various lab sessions. Please do spend some time to read our guidelines. + +The data sets that we propose here are (the default sets) + +\begin{itemize} +\item Regression (fitting a continuous function). In this part you will need to bring back your results from project 1 and compare these with what you get from your Neural Network code to be developed here. The data sets could be +\begin{enumerate} + + \item Either the Franke function or the terrain data from project 1, or data sets your propose. + +\end{enumerate} + +\noindent +\item Classification. Here you will also need to develop a Logistic regression code that you will use to compare with the Neural Network code. The data set we propose are the so-called \href{{https://www.kaggle.com/uciml/breast-cancer-wisconsin-data}}{Wisconsin Breat Cancer Data} data set of images representing various features of tumors. These are discussed intensively in the lecture notes on neural networks, see for example the slides from \href{{https://compphysics.github.io/MachineLearning/doc/pub/week41/html/week40.html}}{week 40}. A longer explanation with links to the scientific literature can be found at the \href{{https://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+%28Diagnostic%29}}{Machine Learning repository of the University of California at Irvine}. Feel free to consult this site and the pertinent literature. +\end{itemize} + +\noindent +You can find more information about this at the \href{{https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_breast_cancer.html}}{Scikit-Learn site} or at the \href{{https://archive.ics.uci.edu/ml/datasets/breast+cancer+wisconsin+(original)}}{University of California at Irvine}. + + +However, if you would like to study other data sets, feel free to +propose other sets. What we list here are mere suggestions from our +side. If you opt for another data set, consider using a set which +has been studied in the scientific literature. This makes it easier +for you to compare and analyze your results. Comparing with existing results from the scientific literature is also an essential +element of the scientific discussion. The University of California at Irvine with its Machine Learning repository at \href{{https://archive.ics.uci.edu/ml/index.php}}{\nolinkurl{https://archive.ics.uci.edu/ml/index.php}} is an excellent site to look up for examples and inspiration. \href{{https://www.kaggle.com/}}{Kaggle.com} is an equally interesting site. Feel free to explore these sites. + + +We will start with a regression problem and we will reuse our codes from project 1 starting with writing our own Stochastic Gradient Descent (SGD) code. + +\paragraph{Part a): Write your own Stochastic Gradient Descent code, first step.} +In order to get started, we will now replace in our standard ordinary +least squares (OLS) and Ridge regression codes (from project 1) the matrix inversion +algorithm with our own SGD code. You can choose whether you want to +add the momentum SGD optionality or other SGD variants such as RMSprop +or ADAgrad or ADAM. The lecture notes from \href{{https://compphysics.github.io/MachineLearning/doc/pub/week40/html/week40.html}}{week 40 contain more +details} + +Perform an analysis of the results for OLS and Ridge regression as +function of the chosen learning rates, the number of mini-batches and +epochs as well as algorithm for scaling the learning rate. You can +also compare your own results with those that can be obtained using +for example \textbf{Scikit-Learn}'s various SGD options. Discuss your +results. For Ridge regression you need now to study the results as functions of the hyper-parameter $\lambda$ and +the learning rate $\gamma$. Discuss your results. + +You will need your SGD code for the setup of the Neural Network and Logistic Regression codes. + +\paragraph{Part b): Writing your own Neural Network code.} +Your aim now, and this is the central part of this project, is to +write your own Feed Forward Neural Network code implementing the back +propagation algorithm discussed in the lecture slides from \href{{https://compphysics.github.io/MachineLearning/doc/pub/week41/html/week41.html}}{week 41}. + +We will focus on a regression problem first and study either the +Franke function or terrain data (or both or other data sets) from +project 1. Discuss again your choice of cost function. + +Write an FFNN code for regression with a flexible number of hidden +layers and nodes using the Sigmoid function as activation function for +the hidden layers. Initialize the weights using a normal +distribution. How would you initialize the biases? And which +activation function would you select for the final output layer? + +Train your network and compare the results with those from your OLS and Ridge Regression codes from project 1. +You should test your results against a similar code using \textbf{Scikit-Learn} (see the examples in the above lecture notes from week 41) or \textbf{tensorflow/keras}. + +Comment your results and give a critical discussion of the results +obtained with the Linear Regression code and your own Neural Network +code. Compare the results with those from project 1. +Make an analysis of the regularization parameters and the learning rates employed to find the optimal MSE and $R2$ scores. + +A useful reference on the back progagation algorithm is \href{{http://neuralnetworksanddeeplearning.com/}}{Nielsen's +book}. It is an excellent +read. + + + +\paragraph{Part c): Testing different activation functions.} +You should now also test different activation functions for the hidden layers. Try out the Sigmoid, the RELU and the Leaky RELU functions and discuss your results. You may also study the way you initialize your weights and biases. + +\paragraph{Part d): Classification analysis using neural networks.} +With a well-written code it should now be easy to change the +activation function for the output layer. + +Here we will change the cost function for our neural network code +developed in parts b) and c) in order to perform a classification analysis. + +We will here study the Wisconsin Breast Cancer data set. This is a typical binary classification problem with just one single output, either True or Fale, $0$ or $1$ etc. +You find more information about this at the \href{{https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_breast_cancer.html}}{Scikit-Learn +site} or at the \href{{https://archive.ics.uci.edu/ml/datasets/breast+cancer+wisconsin+(original)}}{University of California +at Irvine}. + + +To measure the performance of our classification problem we use the +so-called \emph{accuracy} score. The accuracy is as you would expect just +the number of correctly guessed targets $t_i$ divided by the total +number of targets, that is + + +\[ +\text{Accuracy} = \frac{\sum_{i=1}^n I(t_i = y_i)}{n} , +\] + +where $I$ is the indicator function, $1$ if $t_i = y_i$ and $0$ +otherwise if we have a binary classification problem. Here $t_i$ +represents the target and $y_i$ the outputs of your FFNN code and $n$ is simply the number of targets $t_i$. + +Discuss your results and give a critical analysis of the various parameters, including hyper-parameters like the learning rates and the regularization parameter $\lambda$ (as you did in Ridge Regression), various activation functions, number of hidden layers and nodes and activation functions. + + +As stated in the introduction, it can also be useful to study other +datasets. + +Again, we strongly recommend that you compare your own neural Network +code for classification and pertinent results against a similar code using \textbf{Scikit-Learn} or \textbf{tensorflow/keras} or \textbf{pytorch}. + + + + + +\paragraph{Part e): Write your Logistic Regression code, final step.} +Finally, we want to compare the FFNN code we have developed with +Logistic regression, that is we wish to compare our neural network +classification results with the results we can obtain with another +method. + +Define your cost function and the design matrix before you start writing your code. +Write thereafter a Logistic regression code using your SGD algorithm. You can also use standard gradient descent in this case, with a learning rate as hyper-parameter. +Study the results as functions of the chosen learning rates. +Add also an $l_2$ regularization parameter $\lambda$. Compare your results with those from your FFNN code as well as those obtained using \textbf{Scikit-Learn}'s logistic regression functionality. + +The weblink here \href{{https://medium.com/ai-in-plain-english/comparison-between-logistic-regression-and-neural-networks-in-classifying-digits-dc5e85cd93c3}}{\nolinkurl{https://medium.com/ai-in-plain-english/comparison-between-logistic-regression-and-neural-networks-in-classifying-digits-dc5e85cd93c3}}compares logistic regression and FFNN using the so-called MNIST data set. You may find several useful hints and ideas from this article. + + +\paragraph{Part f) Critical evaluation of the various algorithms.} +After all these glorious calculations, you should now summarize the +various algorithms and come with a critical evaluation of their pros +and cons. Which algorithm works best for the regression case and which +is best for the classification case. These codes can also be part of +your final project 3, but now applied to other data sets. + + + + +\subsection{Background literature} + +\begin{enumerate} +\item The text of Michael Nielsen is highly recommended, see \href{{http://neuralnetworksanddeeplearning.com/}}{Nielsen's book}. It is an excellent read. + +\item The textbook of \href{{https://www.springer.com/gp/book/9780387848570}}{Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer}, chapters 3 and 7 are the most relevant ones for the analysis here. + +\item \href{{https://arxiv.org/abs/1803.08823}}{Mehta et al, arXiv 1803.08823}, \emph{A high-bias, low-variance introduction to Machine Learning for physicists}, ArXiv:1803.08823. +\end{enumerate} + +\noindent +\subsection{Introduction to numerical projects} + +Here follows a brief recipe and recommendation on how to write a report for each +project. + +\begin{itemize} + \item Give a short description of the nature of the problem and the eventual numerical methods you have used. + + \item Describe the algorithm you have used and/or developed. Here you may find it convenient to use pseudocoding. In many cases you can describe the algorithm in the program itself. + + \item Include the source code of your program. Comment your program properly. + + \item If possible, try to find analytic solutions, or known limits in order to test your program when developing the code. + + \item Include your results either in figure form or in a table. Remember to label your results. All tables and figures should have relevant captions and labels on the axes. + + \item Try to evaluate the reliabilty and numerical stability/precision of your results. If possible, include a qualitative and/or quantitative discussion of the numerical stability, eventual loss of precision etc. + + \item Try to give an interpretation of you results in your answers to the problems. + + \item Critique: if possible include your comments and reflections about the exercise, whether you felt you learnt something, ideas for improvements and other thoughts you've made when solving the exercise. We wish to keep this course at the interactive level and your comments can help us improve it. + + \item Try to establish a practice where you log your work at the computerlab. You may find such a logbook very handy at later stages in your work, especially when you don't properly remember what a previous test version of your program did. Here you could also record the time spent on solving the exercise, various algorithms you may have tested or other topics which you feel worthy of mentioning. +\end{itemize} + +\noindent +\subsection{Format for electronic delivery of report and programs} + +The preferred format for the report is a PDF file. You can also use DOC or postscript formats or as an ipython notebook file. As programming language we prefer that you choose between C/C++, Fortran2008 or Python. The following prescription should be followed when preparing the report: + +\begin{itemize} + \item Use Canvas to hand in your projects, log in at \href{{https://www.uio.no/english/services/it/education/canvas/}}{\nolinkurl{https://www.uio.no/english/services/it/education/canvas/}} with your normal UiO username and password. + + \item Upload \textbf{only} the report file or the link to your GitHub/GitLab or similar typo of repos! For the source code file(s) you have developed please provide us with your link to your GitHub/GitLab or similar domain. The report file should include all of your discussions and a list of the codes you have developed. Do not include library files which are available at the course homepage, unless you have made specific changes to them. + + \item In your GitHub/GitLab or similar repository, please include a folder which contains selected results. These can be in the form of output from your code for a selected set of runs and input parameters. +\end{itemize} + +\noindent +Finally, +we encourage you to collaborate. Optimal working groups consist of +2-3 students. You can then hand in a common report. + + + +\subsection{Software and needed installations} + +If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, +we recommend that you install the following Python packages via \textbf{pip} as +\begin{enumerate} +\item pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow +\end{enumerate} + +\noindent +For Python3, replace \textbf{pip} with \textbf{pip3}. + +See below for a discussion of \textbf{tensorflow} and \textbf{scikit-learn}. + +For OSX users we recommend also, after having installed Xcode, to install \textbf{brew}. Brew allows +for a seamless installation of additional software via for example +\begin{enumerate} +\item brew install python3 +\end{enumerate} + +\noindent +For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution +you can use \textbf{pip} as well and simply install Python as +\begin{enumerate} +\item sudo apt-get install python3 (or python for python2.7) +\end{enumerate} + +\noindent +etc etc. + +If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely +\begin{enumerate} +\item \href{{https://docs.anaconda.com/}}{Anaconda} Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system \textbf{conda} + +\item \href{{https://www.enthought.com/product/canopy/}}{Enthought canopy} is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license. +\end{enumerate} + +\noindent +Popular software packages written in Python for ML are + +\begin{itemize} +\item \href{{http://scikit-learn.org/stable/}}{Scikit-learn}, + +\item \href{{https://www.tensorflow.org/}}{Tensorflow}, + +\item \href{{http://pytorch.org/}}{PyTorch} and + +\item \href{{https://keras.io/}}{Keras}. +\end{itemize} + +\noindent +These are all freely available at their respective GitHub sites. They +encompass communities of developers in the thousands or more. And the number +of code developers and contributors keeps increasing. + + + + + +% ------------------- end of main content --------------- + +% #ifdef PREAMBLE +\end{document} +% #endif + diff --git a/doc/Projects/2021/Project2/pdf/Project2.pdf b/doc/Projects/2021/Project2/pdf/Project2.pdf new file mode 100644 index 0000000000000000000000000000000000000000..0cf55e304ce9f3503b266549b5ca73b7a3633d9f GIT binary patch literal 239136 zcma(3Q*VQ?YH^wr#Uw+qP|;_nfo;tM6}ZfAeb0*2d+$ zczPdw4sv-BF?uEjHdyl6#i3PLCIAtD=&zwAEDsMXql}5InUgsYGbuq-&b2y^k4386 z&$()UD=yJK-egWU?3!sI=qz06Iyzca+DK>rbbZmX?^P=&gb8I(DFsWg-aO{(Szz{D?h5-Eu`bI>Rr^wkt3FB%$<3 z{zhY~l^bl79HH#~9U2l6c8(8wfx;k^<&x6ioB_7Q2Rmw_IG%^?hK?Y_2YOvKT)KvJ zBXbVU>zxrn@S*8z;36uW7f|!dO=buZFJMkOokuPQYT7;cl4SUdGiPpnTq(S4)xB0r zF<%xC;V0a-Hg56r+M9w{!4k>7@s0Ub*}l$}VR@z%#_EsrL}i}O1-}Q)q3#xDWl~H& zgW{%g0I_E%pB;Ok(;wc#Z#c*a9=@(E6Q|*lEDuW^w<+P3L00G|iQQfZ^A0JErR{U2 zg9c2x7N+GjGw;EfjY!%#+$ld40#_9+I+N&IjbtyJap#zE8?Z5a!kg+ z%ZC()2n3Ql_<8Usb6%>UWRph~V{I8uiO}od@*V0_S_yu<)t|msmsnj7DF)H6rik)?t$$?4DlM@8)JI*AnRi4A^Y^8P zDWJzF0FTt1lnpueBV?!b*3n5)VKW3Rx>S9*cz_wpNNCeyLGZe2#=E~0y*2u!q~o9) zwXdfv5&l5uI^W5jO^lYV0;3g;8#psBH9Ty(jt%wZiV4-K=n~py!|#g(Esv%DseW#ubaxUAlY0fW;bdWh^8Vc>vg$9x~rR6s{(tXf`3$Qj^@mKqMz5Z=`U|Toc zzaX%<93yW!$&0iETr+6Ax?c{$&QF5$MKwZ!arcWkz z9qXonEOLut$RSK~b1h1=p+|8T#^?*BioUx!W~tr)k=FXt4sawR%os+7rO8zWdYmd; z``28Ny70eaJhS&PZU97r1}S$p)<4cy-I0MrSaS0e++cQ!WjM!6~ z-Sa}m?yzan!H){#(~@Lp>H*;v`!D3WW1?FuA{>Y^F`ol*p>}>B3wQXjbo;qH)K8cq zS!m&;6FC8pp64KZ%C5YXy6bIe8d!6r9;w`Pz>iO!{vdeO9(hqjHh) zgpy$qbvo-vsFV+c`-ro8GJ3Xn;uBH;>aHNxs?PY(&q|#Oc=U60;3yi>$~<}#-TwVF z(EFtay^>{534`D2q+32d?@7LtwxkgKm1sG=oDj5-;@%>3(Ej?C^rB#umJi=N)8d&v zAG8n{mOhXPN?K!hXh#jVj|2neA)hK8j}aJ3E0-f3O*tMgX0#R7oeD{K(D5O31lW4} zUCSiBZ2;EF!*VAkA}n?>b6hCc7T42-g6IG0L$nkruMh&F5MmC?c)8-d)Z+31)*1e7 zC+z#T@3FJ~E|<|q;5M`Ez2)5UM6m4V^y=l(mxudbd&13rPsnCt9;Y%lx30Qwx+o`@ z4M}QH1vrXaA=q`Md(dl@;!icgABP-8=;Y+Ie{u3DM;UzM@r}h4K1cjc1zaDq=^;qj zWwW7w0s!B$F#Tu3rG%@v!7qNDd3t!Dj+9{O^xEsm z3T$2G5AvJKsmW%3oWvg+B250_kZo=DWzL;%2c&CXn|EgK4k2C9VxeLPRUhztTjF@b<6+38o2GAIBg;o%GGYCbx2*i)L|@aC_v(#tcc%SLFzg0?-tRdTi4L)6)5O*EotJRFnd zCUCCtee5#CNh5oiLG5)B?CmvKuNsUPi?N9_NU73Vd;C6^-6mr1kGaKmG+^u!8vPxqiJe6Ccj)H7VeUUkp0QyGn}b28%>>%S-I{9%boKM zNFcTtEPxvKIQ^L)2kwe-Wu*ILzUtr50cM@#C{GiovzxvHao^zcRq$3 z2g>o|bLhDP{dmL(Y}d<^Z(Q!3pjQbrh*8&IPd4y|iGj$TEMg_5)O|vwG-#mX7Mf~K zkfJ2XpAANWOQ;=gLML7F2GO&?(_a>ZlHYAxnz4|_=>mI+6bK`BnE%u`FHWGoF@r;P zLIbmc{NDqhkO{?#^N8&C`5X#FY=K}ox{e|OP{G0`!*kM;{;z0e_q!G+WQubaHBUVH z9Q7@!Z+T4O)PHEFRc@n*#{(N3;cE=xeZr%=BvnYy)40Dg3uuUY6{*T)8vYUr&!R~c8ie|1Ta%KAI%p(0;BK(0eWcL}S3RuSUKZ>{vQ%+=tVM1t*i(J^i) zQE+*^i6X6;bXI|ldWwWIQ9G786T&{C0_2hQt?Kr`A&jV*B3KXaL`!=<^qT{qr>A@h z&tiPfU8B@pBa5&nvn23m@4W+mh5Kc##P(|8V7=JIv1{9PrDkW9@yxsg0f*Wlle zQy+FpQ>3xX?lzFxeAS^r!^ZUMBCVp+zuo6>hZqi#%VNqVYFe0D`tEtA2vc2q}N_0eaZs17KiQb5Y++tF6`y9?p1F z@nzSTIrLC#F+o1{Bj#Dcl}<kE`7f`s<%+zwFwy(2H#H!njO zLFnX}VwM60>G#drA+YSg=S3(FCC97umvTFixB3aadinEeYt?ph49DP1>tcl^Af!#z zxbIxWRXaB^9^d0yaIHJ4cUE- zFay3I4pFB6A80Cy(yhb9GniK&5>xXrn?Gq2K?$@tB^Wb^md&M=&j>Q7M|(`}rC{58 zO`(~}x-kOF6N{l#nQ}b~{6G@@Y zY9&Qadxk}VPQp+2Lzdfi|E}gQlk$EMBI?bfq2{7&JbN(JP#{{Y>OKDX=ii&Z)Rpn; z{uBl!fB4#`mP5!mA8=orOr?!FS1_&|2UOt>AR}Nm)QBs=WUNZ(yyL4RH5f`t@_k_n zPBlwq(c)BF@%zGD+^h*$yhPxEVc(!~0Nb43wy_$S%*N5qk%mzLJ6an3Ijkah6C;;Z zyPaN$7>?n%HjmU^@LeTa*1Csf;|r6}r-ple{gvGMuEJ$D-&CXwfdp5!zyRO_WSDfiMZAYGf=G1QYJ;=kVM16LJQGUXl~? z70L8$NPevu2*4nqyCETZ0#w8CF zA$+BQoKSJ!X90#D;A6s2VakIK7{QRUGg)Vz97DzD!293XUcp$D!6{~^|9Ii-C+0~8 zPuCRpuhem%=v|P}@r7{bqayQ^)m0#ov&wcvNL6BA9)FdpvDRysxo(sVbAy0a$sNay{UM%*bK z3ewcKVV$Uu$>PW(oyC`HL#JfbiOx8_hq4b7Dylxe`wH*_v_#Gq$#2KqMZ7k7&ckBK z{$eeaUQ7Zk8f7|REj0h$=Ujlum5@i#S)aKbAm4CFxo zA2Z66JvHR9XikRMZ(n}v>^%KZ)LcxB1Mo=(Vi%`>CK?HWqDv+2A*=aqy+~oamB{s} zl2w8cih#lFYk;;F7S*5W(AnqXSl`p5c3sIQ_UpVG+Ce+8W(gJOA+_C^-5G! ztUpz^#czW*%o4yhE_6j7ZDOR73X@)__q);+GsyB*pE?&V%L8^sRcat z3h3pE`2!nbtb`q-M9`qNQPx}Q(}KVd0sibf7NOEv%EDPi1`J!L3MqVy8eu&!Hkzr& z%*}zUkXi}@~Q^Vu7I-VGL$VnrN zLu|z=8D>v(ttSYFtZ@S7hi+b`p#1x4*JQu0mTg`F6WNG>G;?A2WKFs`baDb7Ci5XP zYm=+yY_Jex4edg@b^U)$gp5lCT@1NE&LZ5g>yQmtSI8m;%J*a`j#Sx>u@q=O2}6 zKfa!hPv_C&YVD(F$yK(tYKs40`G#oQ6Eq`P$5g{*gXdkG9ksvkyelH6o%L!t!v7Th zJ$LAyYY!)QUUI3PH<#No+)&G1Sr!x3V`Gcn$~nJkfAMR~T<~AvS4`#F(iFSiZQ*6K z*0g-|6pz+3v^n1w$Ua`&$$fvS<@f|~TxOO__ zKIc*X;`92;m1DobE^f&r{o3T)^WnzVzoYVRe)6WWXqI~JJD3g@I!sB{$+ zZLrHQk3W5Z@MJ*3)JF}!Qh3jGBK#0uqU5@~@*Lmx;XFEr`lhZtA$-)DYwPi#rK9qF zid(ITty)d4W+o|4Om5`|b)(Hu1|pIzW;1o)%~zoM%Dm5N67UOWh*HisknTMy;K+XQ zUTI=u5dq++5O0$Mn}`e}D9pMRH)z!ToIWa%CpRw+$dLXn(jSGX_>_g=eyX|7z>FGW zx2i4(RdaovX%=nm!`T%>y{J=$XU z!UrNXDSZ#n%&K}fKdB;hH=MVW3YI>UwP)LV75UyiDIvk0j`v?Sq69{3uPMV>ITUcM z79)bG&T*?>s-QqpE`#q3SHR@1{VEG@oh5q<&EUG(Du2TAx44;$E<)pAUPr5*Q8zAn zp_qJZ4{QlXXE!TBn`0M61d&dMnLY6og!@9MKzLHm&l5e)g?wlY=kTpeI33w%kN`i% zaDLUxTjqDcAlxh7MY3z^aqdL>ZU4j#5m0){(JsN-NXVdxr*pGzJ21QZZE)zf9EsS6 z)GB)8R`HA0IxchDwIiW;bykFsZ9LqV*kki$Dc(50kdWU;yztUkuUm4z-c`Rr)=(6W zQmKJ|7K}F4s9dt&XY#Vbatvb1y-+FMm~O*%!%taM{$O&tT23RLX8Q}b3+`RY)ssuU z)48?A`Ds}pRoTtX*!{75wIJ=L4C=h@*UXgQh97xBlPkjS{cj1?XT3B8GX6apY1ybA=2%7>nOIljT{O!xmRJBA0^ory48x!)M{=# zv2RGHK@9RZI!tZihQBS9V~X*`QCQ#2Vp~@Bh44ZYT+6)y^rLw4f{5cwPeuON;PbG` z(xgWA2(?7eNo2pk1L{Hix$22tQ3Sc7MN^8E0*Ju0BKknOLZIR6iYRu32qT179g77Q+UhkE}p08K3z&{Z5!RzIkva1VE6wMuR0r}~M zz=CcvYK9pGjuX<}4|{*vZ)hJ`v+HjPsWEtNjF$TFdqFwJsPc>)*+Z_dnHEU=z-`X9 zFbZQrS_=lX+Rz4X0$khVy&KpSzd)%*9@MK7Mf$$f{a4Z{y0`&&nx3f8O?{Gi!k$@f ztS}x(VR_eK4%psschk~f>$Q;{k-uw;Z8aVt{vLVUv@1I`)~;Rz|L_dK9o=mgUiqb+ zwBrCh^BeQ2*9+q`YFF?czgj}ZsO>AD08e;rQS$HjW>&V0i#RU3pwZwpc=w*yd zvwp-on-A=SHt9BOo`X#;6N)OFae?Ju%i5yib%5Nu%B%ynn3!zgy*;)V5=r2cCqG`B z6NZm{9bDFh4)I^#bGO31teQ1HSpqyWO2ra6D;f_xe98feNzT4!Omh&!UA=m5FF^!i z@|b>J*_PY4&=&%~{Vuc!&|_-@xy_lEF2z0bP_h%!LsUoYJpou9=(0x{C%FNYu&y!jCAh$MLwx0SMFqEkeOgft1euZHN*V>~hxKcXMc2_UJW)B{ys*Tb4(Gy$W>$AF*) zSUpE}&E5AIyl%a!7TD|=u2|V`Aof39e!Y3nnP0O8E=X@Yl3de$ZHG}s8)ghK{t_dN z(x2}1_zMLfOcQIc_^*Im?FU*<~QDkKklrG+DU!Fb#udnOv1wngS|7t!(##P(0&GQNHbai z0ZH^6Z;xsGggIhZv-RlAY75y~t~(tum6ls{KY6=ou&*=AGQ|);rowlmx8BmSVIenrv^V@p6oz^ltGH7KUdgJDFfp6CEv@GLN0wcO4n4z z4G^b5tZ;jt3PeJxS4Qs;g#X>12U`Z8?zOKCuX2yCDiGAn>*H zT0gtoT`R%bho2XSH=?0>Vri0R>oSCr7klxGv!rLIt(f^#Y(Jnui*d~(Syg{Fod<>_ z8010c9x`-L!L#Ady!4g|Pm2C|GN_OJT^>6%Ls*4ruDor8;IWh%fSJ+xJ{ovsA1(s( z>w9GXoCr3N9YG#_$a%L23{f`=b*p>50Ffy-{5bK!_q&liM!NKg&Ey3;-!?Sr6#;~UVBQ5Y(6#{wjezCDz*>;F^duyOv+2@5kj=l@>kXh{FvkU;G@ zQ9Erx*PpXy&uup$m^KE$V21mz4H+Y`uBEi)OZXREy7_+jEUC!GH(Vk-lc+v5cI4=_ z1}wQ?RBiXd9DXcZWnSQB*E;aa4H!Fn_A2`1p*}*;HY;j=SR_;@0H1{MYi(aDd%JpT zo9`ohwEsF~cpO_HSGHcr;+fyfI1_k!^Wt246!>q|n_b7+Y9SfvFu$x>d$PbguQ2Pm zyoTtS)TI63oR-n!n1;;ZeLnTt`u6s~=%KQ~I`aQnC@$k*TQ<*G);OK(FgWI|xid1S z&%mFa#x7Su9r8)LQ-+u6E-2n}nx}dij_EaNH?1_Y5%&x*$R0sqFy6tU}n6zx?`qp0ld^b}66AU=zbBK*as2Z%4e9 zd+w>uXgmjLXW=bDU@~-h8PC%SNjjFSUZ>=ucdV(zPu{q-;rCjNd2o?9RRQ+avKQ65 zEu|?il~=jvkA_pmlu|RFWHo6mou=*O^>6{E<}##86Mo8#`{j0(1xJ0vQ=!W4N%{uI z_@{DkCE4iSuYyLsC})9ndWqdpC;7^7zf0~YnJvN%#TeKcb6Sd3O0b4Yzv>H{iOZhf zAMf0;?TzzkA2oW{{gVI(FMlr)o~5f4wZ6Pf*!zI31D>bCpjiE_IOE%u%zuz(CE9i* z1QPruZG_b+%wBI2m=+XxMg@0 z;V5Hz!$wacbI9Q{e!zB~G@+j9xuC1-8H!}5VI7XNl(+O%m`L8&RB(zf%JEpoxgYL? zR+8kBd?juLddeNLG)leJKQW?>#_jThX#3{LukIBr8OW&3>s0&&3K(L+sg50q0C*ui zs5J+Zw!?br_ji7VF@YY*+S;&km3`V|2CNU9pt}oZbK^oD9woGNMwRUur2-i&P>^e5 zQNX?mH3Yu(A$hxM$4?<0ieawEd)F(iYxk)bx|6-H|eR>WAjsR`&r1{<$*N2?|RSSdq<>gm^@qh ziDAm7Znw?8!oD#e582^qhThJnh((kPl&y@co(KU>L{n(>iij(L*BHAe0%n? zFIM%-f0Iv*XR&S* z=-@Sl(MG^!(aUmB$TR>nVRBW3ei#jxvFmnud=2B`rQVgd!YcSVq)FU zR^^~$IS27;m<4yHDT|f}kj#idq;KeEifX14%s3Bb7OiyWPro%cYOi&7+`<(LrWRNg z8|jI(RRXRGdTGb-eYX(G*eatsq&^B@AxS8tD(ygX+!B1~(U)t83I<&=8G%pArS^{u zG>{niEvaO~`XZk1TE=}#3uXL@PINwwK2Nb{Yqg*mhm)L7aBd#u%gHrkC6_cTp}NN@ z6B{sD7?IC2a6b7XQx}i~j(E0jkS#yJFQr(_PQ%PPZdCvmX=Cmm1IuGFlx8qL=rv?G8S1PqEbhG9}$#!&-8-SdA+JE z)4H{8DtQgOeo>ZzZ@d)5-g88?Y8gDM8sILy4cz@~K_4P!)&oNL4Z$m6gQ(n=W?2j1 zS%V%c!lDZL zkPHq1qM z{tX#09HvlBdSL{y0JAhCVkAbvbt)$N@#L@TQaAx%P$u+w^X~YnYnK61f%$nk?4#Yj znTp32o=5QAorSD%Cr8T<<{N|gmDONm;Rkl01Do-I zeMreg3%L;q4t=dQJNVV)iXfWWAbwD>O2M6JODdPp{+Au6vy`ume| z^KQi~MgIU980z@_S;rrvNh;($ghTkSTsDI?xt0q6KL{gU@Ex7*EjBrs7s&c14jYpS zT3hT%D`$_V6)&AgHYeNYDDEUS=NB7yvdBvkk-bp+^t-LQgv}Y$a_gt0DNdEl8}co( z3{#f_BA}kKS;U4wCnmJ;j+1z+#H%&DVCk!ut3WA_CROUQzXNAX2@BJ|&L6^x{xmI* z-)C=c`Dk8mTK=%shs}qMl*E=s{-%qrgEiNYEh#9ynK@{9vQKB3o*XV3Qidh#Sfy&P z7*l`1d3xskT-;L_|0H?#vYRq;;k6BQAoS>o8T1LzzWF?z5pa(I=1Rzcx}=HX4e{4nMnOdJWLa=Ll(lhKtAI{dT`_ZbRA z@jM9A`OrS!eiNrhX;|>GmcA4oewFe>eMg0B8iH`sZKg?!TiW)*b3cVX`idt0!zzq$ zPmWjXn<6ijLB5W?1*85h{sfWmV@N_ zg^tGJ>(lFd5!KHzkB`qq1(9d}%~=dc#;a~W=~TN0LR01+S2sui+*&w}27N0rz-w8? zp4gJWsL}wIu9#kJXK3neZRgji$AVbT09@IHHI8oojZysl#m);W#~FHw-sUc2JHk-- z7dt~2Hk|CwVsV(z7DH;sSH5u;bo%|2-D2)c^VvToZQvck+6?#Vc%$f)%J~|^`jW)o z8qO05a7)oO&p|t5kb1%xU79Fn>q!9Ez)@EuwVMm5m*LF1`-Cj{QeH%fKE%B4Ku;_T z`S`F+uwr`Wnb%DVXqD*$4jWiCu6QzzKL_8aKWV*_{k>?IyL#MI@A>qZ zq`XPN?b_ypcy}-WJy>2(8Pb%2*x?)ifAe}_AJ5RXEa)9-Ri>Q zW32LJwlJToMo6JJ@K4GISG+(prm%0f_k*5gB1F{xq*@N<{|nV}urd8_b3@%48!`X6 z_y5JdspLv_@gugnZZaD)>(Ri~2h&86jUW2*H5JX=;@1wNPHvQ1OGG4%D8&rn0hX$I z>N+~k=gKUaS88+TowWs#JN$fJ&Fzd0D%dIM)2c|F$x`!mGV>}a$N4b}OM@;?vAdC~ zbMLuGzUe16*Us*I)KMe|_;de-i!_5Yd|CwH;6B}3dm;~jz?vH;l*j3{k>M_*HDH$^ zDa>#VgkAygtc9+!qrx5EW{S7wZ^_(qc4GNgUtvy%55h%{6P>NKmH3EG%Y$&MX&lz` zJ>P-YCI*$aBCo<4%`e>305;#D;19_Xf>0lRW%?Ok&fgkdRyp=*7Vm{u&-W>TdtzG) zC9M=)fW3ihSE*3Rn%!1t=k!lAm)L=ZHrCE*fwb_o&W`k!BJuWBNo@FGSp?j2_$UaQ z;y?t|dFsnjP)bULx{J`1qpvsMY5A*t)vn0klTN~Le{2aGea}_A%e9(mXR4r5KHm0v z+!G_aAXMHwAY$nV<}~0@w-Q8UBz4IB)@H#D$zr!nC;vxZ*acL{E0z0EiL^Tub|s|o zm&xZnLY7HJKP(l#hNhahLucfWJE`0C^?2+lT>&u4fitp4>lLEyc?NRPq4yla{d0#C z0{p;)u0#~|IM`H_U^y6(>Ucb9rf3pu(+Nj;N;z80192WCarJzT{Xp}?UxpM-zfgE@ zF{6K3Wsrnp-<-R^;y-u3e{or3o!9A^``dfS-K}gN^r(2a9T6_Oad^MdZC8?7Jz=Xa z0Y1yY}@IA>2H<6Etc(&a9!%k&Nzqf=MAL+2{nIFLF+&63b%S9*r!1(p=idiGa zJ9QDVlxe<9ztAarK{@B0mJlI$n}9&-+d3-%h)sDpM5?@qc~GM`K_n!xYj-~Vs|dUn ziOtlEso!t5o8zYAM#0Ocz0hn>WWN;M2wWwEs-l^B`YMocdCVp)rEshp4~>-;J1MXp zcl|$|UebG1um?N_!PwrYcDM@Y8Kiyh-@CHGsPjQAyqtflXjC39Fz{ z-4W#dLPO_cV~#IZv_IzRdTzgnD@vI30i1e^`=r^H){xB%kfC|z5~k>rW}Ux!8rOjP zsQ9F}v5ND49gd@zh=`t-&jGH2z4P-@sGxe@>^ z>k*!L|JZT$fevW>$(yiFCetWO)*skW@bt1l9oh{|kh$xB*dm+v&$PVcw38d+&npUI z|M4|s>g*?~#G6G)(nC`zDUX|hvJJ?8mI$beo{gyG4#bUn*XivXnmA=O1mWCc8MX%t zdDLZ_qNS9O3pblR4z=0w=9O9|@JQ#lEyBx#f9&c1?WkUZgVp{CO$zm3{V`VW2jq#? z-@(fYCXO@OUiXLGjXad_FhDH@C&lrIiG|mfE>eZsY5bKP@*q5Ao#14R$D~~(!?uV^ zm+Ti3VuRqCcqAqDY(jMI8L>BaO8xQNCg_5N4DxYTgYhjaLN>!G|m zO*%A=!0Uy*-;$-zHrST8?nneF9Qorpo*V{~QlTlG{%J5|C%i|!QSf#Un>@8b@}cXLaDyfLL2_50R%^*dy@ z4DK~WUQOa$#u7Uur|)wBSi7w#^H8)CZ~fr6Y9r!3&|oWCYHiR%=MR`eeQ1o!Vjq)n zpc?ME10#;M=!E|*ryW&vfWTOLX9?&>8`UXU6+~y*Eio25e2B%Z!NXqWUozlL{NrFb zf<7q5DHVrr!}&Ju6P7^H7y`0$VU%|uS8X~T;5BOHggLs%)J1M zuI{dNzv37Wa{tI-|KTR$GU{dYac9p07KpfJA!d01TceTa zwLfgha=2j|C*w9gLjmI>R#_zm6@s}!!RtF>T>|U^0IgXuw{w6guB~rNZX0MBU~{V| zoQpO;CXUxdg?5@a3%#q<;g32*LgJ5jHmoqClhm9q87d!PX3x4*ikqWp_zUZhgXs<^ ze>>#EgF8+ZSLC9@_NnkQosfDTbW7B;yOU8TCCU@aRNBqO{4O`Q4?nN%?n$lsNp|g? zwN<92TO{*oThP|q28__g0gN6T^}mV4o@*p+xoxBP9V8y9iVIZean{9<;D;p&{(Hz? z5sSm!-2FQit$Qq6N)La>>6$;-DJ^vTLMms*QMG(|agWN}9qYx@!bh|a$Gq_`_v-Xf z*i*x*hL`3FJfbVZDW4E%VR{YSXYL!V$tG>8AS;SQ|19J4y64sQ3g_ibQ28idX3gxX zk9_(VtBAEYHua^7usWRiTuw5f5mLGY_E6o>tlt7k_q(o20~74v2x|m6A3B>DO3_* z{Zjkyc-VF67CP#jkpTFI!*qXi5;1+Dc}}<{5-4S_g8oFS}aW zP~PJDMG25GNLD6h2=mN8Jhe~Dwg2(xhBh!8%B)W4S^5<|(|Hpx9p-A+tOVg)7&3_8j{Gl4h`DfV}qU zWL&+xkQIng<6}_q>yGjaHNn*HH#iPxtiJC8h@FW$79#tc?m66#%=-3~-bWfRyLJhl z#b$@vn~^H}tRXx-ITH-1&}dgjM(Y4)+xWkUApPasf0&WOaEpw)KyG)iN0A;O9F>{c zyvSs}w_TK{C*J(ibxm7`KP;khEPFqToW9+m5`_+WZy$w4j|%Ia0JF~LeE-|?+iU-Udc4X#zxCK(S9V_B==B(;E_3QH7-Y(bombuMsFa` zQA56rNr8P3=ngs#t^ejlJ;1FbQJ?L=E`qNV+!5MR4 zV-{(shq@&iPspF}hoe-sCC>X~**4pNG$v+U*GsrO1vrQJ`KC z>h&LvwJRExaH+vd?3(ut`_2U}*%~bkiY*9kjeE__t^U0>rWUW|C)wK|=FZ~e|v)WfGFELey9b!0zCG744 z^XVFUQ;=%E)4#S}L#p_}&R5~4h}T9HN%+>0p|Py1nY_<@>4{HrUSPe@8;w|d@duTA z-ifcPdzt$<(hiI!(@kX<7vHOFvaJaFkatNLk5x$&y`I9!kA2nzb#9B(yZf0W?jz6?-4wwWl3!Cvgiyo^P64rlm7iEq5yf&?BH ziRpUqE?j%UiG!7DF{%68iw4#QQ*rN; zx_S+(L$R}ajPD1o8ze!M$RXz)0QP4#xAk~aeWvjNM1&LsLDL5a#A8Fa87vDqDX^rf zSW?+-I?G@i|1T!^Du=xMZ8+_1g!PY1RF#3pSyGq(=AM-t0pfFe6-S33fby!H;wLOV zN#|TWAtdnyl3pRv<4T;N?UQc{g7{mtR3l0R5K#p5u=L$N-qi?fG&&Uc4et@?Yu9Wq z(Bj!HuZ&a}{edow`WSO}UKTgoUxtSB_gWu!zgKzj>7Fg8*f0cW7G8yuQR1y2&5@*|qAz0C&V;@OqN)Hh~f8z1r+IR$N^Xw4w5Piy!0b_6i zpUUFPo_z$Dm@KF(!rX@fAIhHl+5(PM1YD$xOuF@ZSOpn@4k%0+Bu&)+{VZB+k-PECO} zh7aO@KE_2+|1(pltx#_+mM^(S3sZFS^4!vog=XB}DLX*E;XC|Y6;`0C;5{eO9EmXz zS+?Uv670Ygvd2x+#tSaY#$u7E!UL>Ek;n>*2s;+^r%#ZJO%%wZE#$C3$h0pk$f(jv zLK3&$x&>>7`~$HOTBLX&ahw}M(@v2UN;5v*$ z;J?cPre&mXEJ(^;f`D%K{zp_o8W zm+y<5XwC0gxpH zu^p3>@nr1eIc4BMS^X`o1r@-I6LV1R1&L?2cW_55=&y4PA$3Ii#5~8M{q|)09p^of zwWt}=NpJ?q5t9E1U<;~hdm8xsyy3ChI} zDN@rar1kKu^>A&D$<^nMa`YkU$d!twAE^DDs;u_SG*`hd+eBD1^JwGu;1{Z1M#B6v z!}Sr0pCWnrs3C9+wz*HhK2y8AfqPmeD5;Z!CwQ}KfdCbEnmjZ~Yyb9}6s6Q2huuBq z=>r+QJkCG-X7hDr0^9= z{QvhZ{J*bVXlgT2&RSb$i|B+}f$iVg;%)21kZ*%<4c)y4cZCLryS~0g-PHO204+e$ zzd~?6(*2&3K0WT=J8DqHTsrq!Ywul8k-t||p%*r>GXhE2**eiPGcfT0WaX6YY=E}R zO!Ojl)+PWJ1|}vp1afjw2N2N7!p>F<=mg>ca5|X-6pWqT9322GOibJeHfRg;71iJGsvz4p{$_Ll zdDS=l?>ujYzuo1;RfScx6vdetf3E?+3~&KCI9mKp`!8-3Z_EIHsJ*$GI@sC#l>k6x z?&M_8!^r6B>dIi|?C8W`=U~QQZ~YfPRdWkRfUBK@72xgD0b~vOt1!;CCU50*G6($~ z;CD*^vKGc5TSw4uCkea1l{Rmsd~s5Cjl)c5wKeL+(E+hySMe4|Nf{x0UHy zdwK(1|J^a5t+S)YKWg*8%Qm*Nb+mAFa{N0Y2w-Yq4f@UA@psQGZ2ywU3Cn+!5LZ#9 zmwg*NTY5RWx9Zq3IJr6fW&Jyzu$U|lfQyL(z|74CV0s%;aa$8nI~$uf){Y3j%O_^> zRwpMr2Y1GQO|_M+ovW?qzon)YwkD>(t8U_K&!}c=@yQwVQSAS0-$V$1WM&{I022W8 z2>^04HfQ{u>aR)pEoS~Lek*~ur@fs$z!Yfh2=cZt1-*SBcsc@IKmaEPXOOq&KNbIr z5STduCKkp{Z$tjJQ4s!${?XRd4#54l_>IYbRR7n0s=qcV_1g|Lv9q;y2bh3N5g6s| zoZg0k>i?fT_ODzL&eqoQKpPO%{}%M$RzMpIYxn=P`iW{|{#iM+plzkcpy& zld<{VRsCE3(FyoA^uo4g)}XhJ{3TNRy^pNl2KMdHvG{#v0qB|8IR2%3n?++QTacq8 zfa7mB&|3}u#rLi7zu5v9WhIp*B-CmDH7csA|Dd-F zDt{0gfKl}idW)&?2XO-!wf>8^-*mt~=q;boAM|Ev{9nZWrZcv)e(TnMBy4QIB{nvH z{F#~FD4P5OzEuJA5BNse^bcfti*IW2Cjskk=<*MH*0*S;)_)9syW2TC{3GC-i`hTm z8zJ*Q^xn#B?rv`mvi*m}o6O=L@NM;0|A22xwEhQt<7e{+GQY9;M=0jEwb}hi_GW8m z`w#BSZ)ENNsNan3fp5q4-&_5^;+pR$CH{kd-$DQo$PHwSurzOH%o}8xQXf+4 zBc4B522IY*aPk#fKvOynYAJoPl>k>PXNwdhSm592mqAfrCo9FH=ab}bTck-Igk}Bm zqbads;>JWtXdCDw=R<%ZE(Hz%I3{V?666J2~ zxQ}8Pe?`AaCelYvV>nQii4dQclzC%?^3IKdQNd6rS_O(oe1KZr$y#{=CfxajXQkm^92;t3QKo<*5 zAqY4;6nO}18s^an#5BQbxER2lgo5Vwc!$Qs_{5Z(U+aUN(bMlvj*8=hD!^g+?oa$3$GS8$6Vp z7gb(QXKm&)7jaz#6?%S67>7*yVg&SwC9V6NaEo|ao=eFL!7B^l+ATWmMR(UOXfUiG zYsJkmFDDBTz<#j1eMok1l~R`waQ>clS{sFM;gY3pKHYsXc(%6gSGj-Mf@)v4V=ms~ zbLsuNE%ddC@g(F5!l?k5g+l15+l#iAo_-6ujQt$BqR*|#*mf^csB2vnck3*XBb(J| zlp9(aN5Z`qowpx__i|uFdEKVGRqUalbjL{Vn{5066FW!^Mb9xyqzN-z*&=z_3c!-t z;(%OXNGNb&P;e^yrjVbIC20TJ49$5_qYfD@P><1Agv{kOI=KL|9H`AL6V6%{7gLRe2vgt^L>Q4yt znUH+Bf?#&~#S>$9Zikz{GoLUk6B+AZVsTx@&k&OxPKVy`q=dnXW_n;*4qkgflB&0) zCA+2Ua?JYVn?APE#X&cS3JJlw8Wp#ys^LIWk>sx&(K~tGoRDm)KpNqrowg&!O9bQ~)h9{&8$rUUj#SKE)hZFFlw4b9r^du#lXz$Pf z-cst%vNzA`sv@E`EVF1& z-B~M#09{UW1Ms%j@5k0xm|`e$6IY2Fcv$(_Q*8@dU%>>g73&KZ8jceZKGX}DO6Y@? zDx1<@HEY?mJxPg~XLIFPF3KNLKOT200&G4+2MGH&{+v zs6}aNRrrT(RK!*A``UB0UasdJ&Yk7QhCW1Bd6iN~6Qq&-$Xcs+*qt!hz59|!O&yqe zr=eEqL)7~rtv&l&)Pd+=H)??p0QIsJ3iB!T;)0z^o8)8R&=LgVEfwL{Y1Q=b0~0AC z1(%jY0$i5u=Z<3Un%O|cOadu@(UF4a7CMR)U%AhIH(F27C6i00^E@8J>j>Mj%a1wr z+?72rlhKng?Zm^1(DeP#)xedyu$Q51U%_Y!EV|sBBt{IMv@{vjcUhfQsO=h$DnHjc~S;nLWK+% ztH9TpFFMO=FuLpQuK23RnN8e(lsw90;Rm+~<5fxx{JIRyWpy;lLdB|6?}p9^BR-cO z20*d3$@Qi*-rQuu{wT6+U@y0=m&#A_Bb--{G+j6khSVmHzzPKyx>n=!(iweVBvQh| zk(I-*SWoUUh@XXb{$nmXH7p!I#e(&@JMZ;A%F;8htMw9cm6ZWKdYrKQpdM)HLO$w z;EAOGU5q0ea@bba{xdJzM^S3~Xl zx>|HohO$b>4Us^Mo9T5)cS>%{Q+cH5u0_f@ptLm%z{qi*ipK63T3(8|KuGbh;cMkK z>E{IxDHP%+KeJC5R`jZDTSb_y{XAg%Fg-wZ8 z4H;zDWmOpX@xGx4LP+0-9@{-Xxs#QEM!Y?<^l zMAe%=j*o{dhS%U(WU;UXqAjeApps?Z+tR&5s|xAqJ}j#P0~Wc+t9hM#UeTZtsk_Ml zP`(3S;6|sEh?*XX7KSkYG6toN4qDy1;eB`FFY z+E$$`^fyau@8-&Pg3IgX5Zwf7z?Cj$R5m*q&y)QGi%93z<;=R*x(R#bj_RX*PG~|k zBkigC#`|PjQ1Os-G#gr?kjiqodNr&z5*G_8tnw7p1G?BrI1Qw|o792e>;mH&L=73B!A}{(JbO%`k2CuSE_utdj-`8xh`@ zhR2^r&UY0nC1KHIj>D1#kmOH~-4Q8R%9o2-n6b9Chj zvyM!XmgmFu4FRnmZs6lo!))sLw@}XOdd1==GFd`)y^cM=@Z#nd|+|x^t!oFziL%gM@m|lb@ zd*3>WjCA~Sg?qkD4MV$nem95kenuFG3GsZ+>a1go+3ur|OT^R7KhWE1KnE$AS{cKU~vi}%B9w2G% z62MGM5-P7a$~UbahxbgJH*XTqUaz4qx^}CI$kgV4Zf_qIuBbZ6WTa}v8{3dtsl6_o zmASk#LJ-WFWH;XQ?$CWKx$eGaf@nCQqv6}oMH=m4?!1NHHij%`5(X~i`eq|;GN?+2 z<>@0fin`T!R_|#bGFFPm*9w`~d*t&r?)ex@=uu8LOHzJA7XZX3A7MmOM@|;36@p{i zJA!8yA-6PG!Su67g(f8?Tnv@Nm$Nyww_6hXL3Qp+nSWM)VHrHT?7<>eQQRi_aLiWU z1ApnmAU)Ngl{pPH7Q!hcx?irlDF} z0+{bU&&L5P`D-9fUwqeAH|F{n5_}+?Bw4l}svbL?8`zPyt8Wug7gX4z{O6Cp)0?H4 zyP)XEXq;!UqKyo@x#@SAC;JvtFvWI zV>ux#m|n|-i@Yh`{UUjH86wBGo+gioXo6>iT8NTSLt->zVXr-Q-# zsVIY91n3`!5?y$FJ2jJQT%q1V%ND7IV|G&(uujrp(*C0RSSm}Wxn^*aM|8PRRT`RU z+%6^2prQ7`_-xtL|4R>pf29ST_C_CntzUVg;pevE?`)kPB5!s0XHtz)e<))AFs)sg z<4s&{GbLcA4|`7)mZQ>7lUu98OPmf}x{`rWxS$oA`o4vv%mhH77aPwKwCx8Ax8#g9 zVtt9xU6M&J=5HJ8gtb^yK~-C~p>cX*?5Hv@>v==cKIkX)sHJ|cVURt0WMc9x;7Q2A z{_K-$s1)7vkaziuX*fc8-8TtovrWVsMp3Y|QZ187bO9sC9}8|ik#iVKb5DxR>DFP? zJd!7Ul*q#=G@U&$DmTq1lQVL4iWlrfJ@Sk14E`V(JUZSFUCNJ0#C)=!1BV!n@6MIi z=h_TPD86o~=lr6ZMcYWqHfj3==f0Cfp`{42Aq-%HTGqoy~!p9e_xaMv(%fqyg1yfy>dF#(&E1Ld!0MAnr zUOg?nU&lO7ZMq+QxzUBM;1#0FWrq+kHc?}ycSU?B+Rn=QeKCI#h$HqH`n+grE;y%m-AOH>-ACccp-#|llqCLBWQB>XP!Z zn^!p}5fepTPEN}FF6F4NPR|n152HCxtTje+`f(LCOPS&|fLpDnvhytUr^@hF*Q-L7 z5|&lA(-}eSi}@=l$++-M8!2m6=6o()fb`yh?c&a|1pN-XR|>Ywt0-&H;G?iL;AsZ=lBxzkN8MH7wLq^@IBmApO$=WrmqS zTti;f8jE~GV#ATVZW-GjjxlFVL?lCk6z_8}QbV?8m7rR1Xj9PzD>A#|6`D;Ob1ur6 zjbfPNNDM@?G}yYsx;Ro0(-M%&jt}t|F?wLsntU@W0RGJ6S-EJ{wHeo;9;_{7(FULW zm4W0tNe6qqc9{*$-mEZL!p2y`auTz7;^G4eUQC|XiDdtXRR02r&Cgg4ugded-5Ta7 z7GygAGQH6twoesf)6-+|G!>H6?-$WyGV^l|?pZIgbLhaeI z=lgQgA^r=?pz1k&o}2fII=>dhQ~a4l`;qzamGqQHKC?8P8!VZcfpsPUd^FH36c@SS zJhx(?PjT|Hkv^5%cWXBM5O65HBSE{iXNlRiTH~ld`<$yNtyAw+)7Q$D!e^SKe{+7S zPdE8O=xGd&kAMrX)>zwIVgRG_<;g-q-enB`r=E;+D_zS(7&84#1D|M5CXT@VtI8 z#t%QwKz=Z6hi$0paErASjBsu0zM*>PkIY_JQFaC6woO|_(;Z1j0e0?&Nw9X=yrXZ7U;TKEGZ;o{# zZFdN^JA3L8AEF$@RT4$LcswONx_zC4kzVH7YAzh&KQ*Dw zZHgVY@ly?+$D50h3G8UCr@QlzqiSws)FnOoA+)chrZlCqnQ)bIsYL36>-WgY9C z8An*tI@i4O1rP^Vx=n!RKRUrZAKX<5WhDvJKaCTX z7J}KNV`)PH_4z?FszQERU(|Ig3j$-0D8i^A4ei2Uh3vagOf?`Ge>n%=c5PqP5^{Mra@p?weXnEvaG@5fAQe@RI(L7*r}C6dVa}RzV6ZajK<}-MsNdbb+4;@75MnRQyqprp_0l%m+85Ana9$ zVatoroZ#knS4iM`C_hl?Xaa-s?{oy8+>8Etr9dgMwDlK3$MHQZ-gs6>CTv?;lwsBae{-!`~>rrSh-#ucm{3ywb2EXsO#Y?;3-?L==y-iszU4a3f;;MjZ9p=vC z$T(QDo}`mT`_;v`?RzLD#5Db2qgt=vgV=G=cj9LC%d3KG zQ3keiuC;m$h%1L6xqG_qsF(`B3MCQ==}9xbafL=B_P zc~J7mOCFY-Iv(1hNmqRCzHjHn1fKH=L$t36S;e|1!C1I3!wfA}+aP=m#d=;c#^{4A z)!}gz;P6p=D5cR#z8j;4;(<}QR4MCSRJKuSV8FeZ@gr5$OsFHn9{G-SS#3)TeFtzc zCggHeXcAXopP}_!cf-u!%30XiVaxjf-FTI=pwYBdEgRWk$^(&BAi~Ep`tHz-jFYhp zHfpe}3S2@BWgjwEMukY3R+PW=gi%6nXH39lh}DYBjAL`6avMWOidtYSDQ&o-Ue$m8 zUJUNhUw>*&n&U`8LxB2DHQT@FmE!~A?zvh8Q{<9Y5zlyzkZ7j@L31f^t0z6PY#Ec? zAPm_-kE^GpAOZ1Q;mZ&8jvpCAMAfKecdIz(k0D9N;B~-E7IVy8?x^t(95c?F0~QGs z2u8C&ejsy|^92v!06j9-BOOcfa?u|XBkve0`Tzs97IIkyDY9N43f`|pF*VWB(x4&% zzoMXExTm1~JUseoP(hO^6J#hn7m4|78@SY4)6FAqWf&FGq6jg;W2p9qS~n47OPo_keZc^UD{P3kf0&$QB7S2*dymJ_*OHjNr*aF0b~`9aZ>Ee0OdOBo2Xqf%ycqXg(-(G+=u08fP=fyq7KvD zkDk)Gdyi#Qxi3d~OCV6+@}$F@T~Fu}^Nm~FVtd_+CA&Namx9xa9md1UEcj%j-=dSt z+!!M04i%&8DwLM&$o*mpp3?Go+hM|ySWZgP#>5~5LhwNOn0&}jvp$lSM$=5dG;WbY}Vn<45 zZgIe9} z>$aev*cf!W{P}gtJ+C99Xhaitfo?5!m;PbIBuOI@b(?{Lc-KW-IwYri_v$;{oLG^rO^(5# ztU(gAeASR2w1F9h)i-E?6sHg?5(ECxh)$EUp(*#rs&|sc?9=<_b+bY$hJ$cu{*b7+ zIJZCOv{(*x)zTC6{eIGXU$>GWo&uia$O{XzEee8d>}OxA#VO33#;9)zWSI{}oTu0N z!2f)AW{g}_@Z&U63UWTR7#Zs(k`Wk_2@40Af#2D&DfRm4>qqweOct?at@6u+;m6Fh8wFQ-T`i(Wa90Cvbk54Lf)XE;8v7H}%a}*R}Uqy#o$Dy`-_%qAzbvNG+%W%$&kb8m2 z8Wa~`RK&N8d-q+YK#G6(zI6PxP;{|6I^Uz{Xi;G-m0zdmQ|3kXqjzAqmB&K1p67mj zfw2%ca>u9~xVFws&f!aG&HDg(>&70Zjb5Ey!f!_i;zr6-Vy039tf1-M!9bbA6XBca zU4&bAr?3y^e&2RNnAUax6J!1Q^ANwNYYl|P#WOT$4{~2af~<0@YdW?q@3+xs2_x|Ek z`AC;9@~g6&8^WS%z{i!;E7CC_gweUvO-_E;eH*^er|v8AvlZ44=g%~NQMzyFtMh8@`P9?%@zek z(>A0aV;~X-jtcq~;KRp|LaEGL!%iDb_maCwH16N5_&uIxsd>y$b*@IE3JNHOq zj4POC#SfpDBU-#y{7~q^p>%dTL-RtzTW_f}uNNX+B%{Ml&!nc3E32g#PV1*A)C%jJblEFf!){~X z&M(dWqAF=QTE18MT+NP>a^(6^&cJXJhko#GEJ&ZkuVF%EajKRWgoCKW1iso(p_o!( zW}eJ5V?zz~hB%44b%TW0zqq8S?HLjCVrp*hM-QsLcX%aSHf*w9k84Pagr-rQ;aY+m{x?bW;1TEs*6=y3yu&i9M&` z0Ef{Y9}gg2U}zRQ>!SFDZS^7LIhOjK5Lp|_kR`73A2XsRG#6Wzv1FbSq>7*_2&YWM zF({EXEOQh|rOyZsP0Tokr<5^4Vx$X$LDd>_IfsGoG!1@56DL!QvvVYdZ)ssNN$^=s zip}4$u)aNwn%>pbb+T^wL=(+_{ixlpjL0qrq?d#25z$=ho$Po=E=%U}k$>FskbnJN3t}D5jp<>`=r^`byZ2>_f{}d~d$np?cN2*Qufso})_%I&) z&IeUi*8uR@{s=_HNU+-Jnou5R{~~xLw--)BK>lr@OOA?j)v<<3=*}C};`6&Egn@@( z%JBCh6O+q?z?vM&H-YD^Mfc#OV8Liu&YIDrur}-bQUl8BJaR z26w|RU!}=ZZ`wl;XeCJXaDprXAbT&4^ZFukULLAf_*TZUCZQ{X*BIiwqUQwhq`R-d zZ$t#Qh~y~9_ud``G=tna5=ls06sR(cU4r=ImcdBlDe@h(CaNlUyMa?L0Wn>QTZzjM z_gBEA0X*WbI`uwxPZZ*Yu)zc0>Ew?k9}MP&9J5H^=LXsxU8fVW$#Qvj1{B=5i4x&z%x?DEoef=D6av z8bi&}abx+GGVaYp7;GMHxjkHIA_W$r!m^gLn0(iV{0j7s8e6lZefEXRuPM%_uCcRE3>j?%`YR5I`<=+Rcs_k zcH*L5EA3~;GP>f!$12`BI? zpo1wI2_FS)F8IjOU;naC-^;gDygj!+odbvbr)`*_OEwFQMD zlA#5Ln1fMSl@i;_N4aaKfbGCPF15%Y<(~|5=ys$wiR;#bncu8FyuViS!hc;=s8Do; zYvf^i5>TRfXyqUQ1>kGT+wian%zbm59=hLhP%Z~4c zm8~MsLqFHVd5O5VvI*eVA=oqQenpz?gD5Spd! zsL)Jk;m|Ni1vI?&v-{-<1H-|k6}_hb@bV3tQgZCxj$O1%zka?+?wPderCvb9k5n;H zVeT4ItHM;+g5BCtut=pYcEd}_=!7WrY|N0VbT6E>JrEw6$ zwg-}5rG(S32} zve@pcQ};&(0zQzGIX37W|jpm-qoDpJTz2wBqc=R7WNBenf%ze0zS9RqB z#iQsUatZAk$|m{6p_++&6)GMPfZG?$-8p#Pr2ducigf4z#?x(Zv=_afU5>weR5zCd zMTFW#QML+&9FGw%7$&t_GA>oD-Wun6%M|8zw`#9;d3DL;eFr*+{aGb7mye7~iSy;Q zu1*(os6nhP^+zzzwcBc|cdo@^>6k2wza?&V*W>lZ@{)@AikBg2YhQkmR#uSED(9jK zYow+8=O?|3hH`?z)jiTx!SY2NR3`)tkMha=C6}9fi)3*tWd#rZE0q@XZ5#0$G>A>3`A4c}_k{RH@HyeYB*oy`FeJOhL z%@cbukZkC=yzjg8PH2uI@dDTDHb~5Nspa@gGtS9KPdV$*v=7v%OUraN9ezln&?7NK z1f<&8*@XJ4VW!!{=b_g*b{blTPY%65U`U2C`B8V2Ju@#U7@ zy@mZfMkr=(gWlYy>w*(4d!HO6-Ye;Ba@DwGxNP_Rk8*X0h%<)X4HBlz0W_D*=e6K_ zu?XGu`e;qZ*@_UA7)-sb1}+uv*+%77;EV@xhyw6{uf9!Vg9k2D2~jr4A=mz_2|bKd zh-ve%8YBJ+N=)#*cSk5*K;*POU-u`;_xsSUNt(3h~X**Vd=RN<{+^a{}KznQmvp{A?PWrSMoK%Uz)f_zqNrsKwc5yHJI0oclfrtsQ92Kj`aK z^o~hbVJzfiz`0j{A`^!uS4@<5tl}5%NpD#1Rw@p%+UrbuJ1_a`UxaahF|b8KOqPo# zBBpKJq<2PSlTh2*jU_)Juj7nqniw~KRr8wcC3&3v?D7u$5IBmyJ1-okx$(1s%Bt6U zz*VHA6Me`i;`EEhqKO9vgEUWDc1$|sL~`DXSe3PDUvX+lqgi+Qez(J8R;mWOwfy+{ zaHW$LBL6hv7?!TZ6YPlIPk3IhWnUb>A9paw-g&#_%xP`n`9dC&-uT`r6dx~l%W^cx zD}C1Z)fLAbRh?h$K9H$v;ohH;ejxO)=EsitGUvBYVVd{D`^&dxXrpo)x}puV;)Q7# zpIs~0x*?9on<&aa42!80<$!hFHbEkkOl0VFS+Qw87%$8R>GAKcwkcqUuw1lI;S~4X zQCvRE`8bI`(X5{`llfiJlf1&w$7!N^{r!go8FwX!&(fF_Nb6d?cJU24er#K%fd}L* ziPiGK-0mXHZ7@QCP?{hO8`1V;*!M3B=2G7=Jnt7n1TK3cF5Bz3Y~nE2;?Oktw>j+wO7W^fKg%P4=;CX*SLJ}Gr7f`>vsMF#cg5<&l236m?iS4` z+3L(tb^fb8!aSUVG{AO{ua|!GJ+jU!b8PSMz>SSl1WDtkArH3~tP#D0;u@h64yt!4 z(q-iXX{IQpB4@G?@r@FzwfmEhQ|jflbn|`hSz>z9_v7ehu8wNJ!Ie39@5jB9*)`s`9B^-Q zCytwg+@J|!azkM8l!W^NUP70yL(6-BoZaudHx`aNK71HZ?03SFC!(Zz>QFu1qigd^ zRhP;TPPs*9i(Q(WmB?t)XHo$)%olyx{QhtNIW4mmbjFE2xmUJ;R|2@ApjoGB-rTXR zM)89S;3kHDQF$mB%ZhZ3cbs)q8?Jx7kU^#W`k_0-LVU-<|6a(-Rso0&Z~Bd9f$Df` z_4d^>2=;-jgzf_sSL_Bsudh8)@<7wDiI3Q?LAYwjc@1`f@8;{_pANZWYwAMGh1bE$ z>fshT=UbSI=_&%aJ`lA!7D2Tp6N7xR48heD{|9Iz)v8ryaKNc%m-3F#oX59kUc z{w_i?RNPvmnx%oW?GEbFK!@%>CiNWGlNKLxNy0*?b$ZV(KXpBa_`3JgdR%MkY$*Vt z`<`;sy2AZt5K;ha*0|ck7}m!=D!L=W1;0>pKhpH%&V%z?2mSy)Xp~!ql%h;w^Svtg zk12B9JNE(3$2|elc%IKu8WT!Wx+K{@A4x` zn+f55UlVkq2&;X(6(rO`x=Zr)Q!x;EhR5~*b~>o5%naFlX!6kVtkQ1pc#4skB6g(w zrkBiLvN-d?D(zEHD^^p-9RL^e$0_CjGQ1qrFXt&?C1G zmNWHT1FMX<#|K(l{}&f${}wGO(vP*|Vx?B8$_WX1lXeKKD$~v5Z{KHZyM3#&Ld7^} zkk#t6LY=EMe~w(y(7wK0^gfuEg@62on1R8>7c+kcw6$^~OjN7$OPoJ(%Fm{XO%Yq8 zK?#;~=+^YKuAn~Alka04&C_)hcg-V!Mj z;@?S@s8|2Qagz<-GE>*8)-#94-tV(_Qz7uClfmtGHd~?BE2P&YqP!RfWv@O3o?dvp zPq`1dg`K?OATC>RcIsfRq{Fn3a%@pMJaX3)w4*vMX&gVaTw!RqCC93L^;EkIy`{)K zN~AZuB*EFPCimY=2-v{PpUZT1bVdzv)?ABog(dY^^cA-`gAg-3!Ha~yZYo~m^2p}J z`Y5M(DoCRp?I4bL2@TGlLNLvvpdUGyMyA}i?b!J-e}S|NdsWQr{o0%BN>{7xlrtOwfRvQ+x3~xmpF>H+2s3a0zj+Yi&TCeVy~YsTUb@3cOb8K zgN@sE;S+vh%z)T}g_bib;e+vw0KJaSmCL?! zDrw!9Bxf+ZwqIcy??tX82>rWABv2f`*hVf9y`dqwDA=gjUZK%>GJi3u zN^-4ghJq&=*MP?QPLqP@%hsn;CNug3Eyc{_m$oh5$Ac?YMkzod;bUBXaPcc&VLElWZhFrem2h|Z1%HXpS-6SVx z)E|&veZ*I-bPjZp*S=4Eif?!-oD7es?a8E~^}PPIOB97UaX*j9U9`3N)jC9H$qGISbQ1 zsnZzCmbz^0y9iVC`JDM<<%Su>ND1;X{}FN8MyJ`88l7wgx+&hd+njm(*0iJw1*s&1 zZP)$`Fir`A8N=zBya`eBpVP|KQ`O8!&R>I3#W;9K2vQ+prRjg=(8bevC+b(iSZSC| zs;+9fKezyxJrwnw!!fsBL(5?%ZB+mu4X|rxqRJl44jPF8eD)TNZhz9Ts5zxlzfu=Z z8wWd^6FzGtN1HZg3+A(**%kX_5$|y!kH!{n2fFah%8ySen}@Je-hmF$HgQX&9_?d! zVNq1?J@SYoSdX(OEJELlp%CB2*8?`h(Qkt$`yMO{x$C1W0^BY{KL*#9zPl_W$&SM+ zcML?DEwxxF2T-lDKUxlB$OD8Mg=mc=Qvz*IX}y{W$tWAm)8eaXF`7HS#kg5`}wY-RS=0{wVdM>M0Q{(cXk7LB*03%xBvd78v0)P#_8 z(K71sqkImwAK)FKC$UorRD@&O9r_6onEoAx8_%apOpW1NT?J(wg_CvwT%JLb&B&j2 z8R}yF;JgOxy}RMw@$O~(NZUPMlEPPrVTG0HP?($;1`YC9DdAz+U&nWCgxeFIN$Y@X zI_o17U~52pWa+%3#@C)4MV!)vC8F|fP=3}Ya0wxG?D|^l<5ZjII+Y8 zBtNijU~m7S@IeJ;ERt?(Jc%FyiE{HI=1uYboB(MyRTxJDtE0LfX=u*p8 zhjvAOQPI3G;H3|p4xW7e6?NWWx{Rv%bkB@bBAp0MzMiO1d6hxM{7Gz;(S3^$Ff+h_ zX+7!7v5e-l6V_rr1UF3(66AY*0plRRU?Er)y~+{fP$W83oqMXDeCwScy&0cLrWpe2 zbN%6_mlIpK2|_3gi)uXuik4>MBUk#Id*qED*1tZ)5eK3?+H4SXNqWF%qR>d)T4zdKt!nYqQLn%P057f-m6ucbl8>u5-_@E ztdD-#2y3j<_qNPrh*QH|OS{kgvoQ^-Ua#WQbxMUSGPZ;sQ77^IIfV-k#$yeDjVA5R zApgAud?zHTvlV$n(!oAI|K;{DdDJkEoLl0!NYSAcCNN!pDa(AxpeV`>M}B-*!OMgY znfiklmVZX)ohXBSt^)nNL@i9i0&(1(gU0cugFu#-m)$!oM>u@`-AMKH$1pJ%25J1l z1xc)g%7vX$p%)x76Qx$v@ZwgiH}pAiOQy^F0*+f1#aE8U1jOf9OMLhmJ^%^xL!;*H z4yY3Sv(`zCXP@k0iL&gsKGyki23)bl&~|I5)po1xzfoJ?wsX7rpEb&G>&wl}cqUk5 znv`i;D*Ml3GT&x$F!65Kf5!}WX*=v?rxE?W(ZLgs!#Ix~eslY9QPsqZpVx*aBEPJe z*{QA2lgkqHhUg!W7wxD&+eC77ZIfLk93oAF2s{we9c3a`3azqE%d4OJ3{b-+xjudzu;?=yQf)PzF ziwm3w$ytlY_?Vx5FOsPXaGG}QjuOgrQLE!%khyU=^X$)%@mJ77Fu=@Zu92uKWo>+J z5RqV=BB6TLZ?BW^lxqam=ic}I;LH9rBX)dv8b9b_MDaUCD>>>UB#-+<)a;`S+GzLF zINXPe!Nf!&V{<@yYCsA+D>(Ln$nzynSd?y@wM*62bjZbRq2wfD{z%a@ww?ck#|e4` zV15Z;g_#esOJfs9qT!x9MgzaV_9fi(fLTfT7v=8kIueVx8r3T}h`(V{5BWg;UuuHm z{q_4|VfL8(fb7|S9tm<@$=^(B&X*0k`Oso7oUBvLUYHaFukNidU{!*Q;#3XD48X7D z{-I;Tefl;I8f;6t9Ql1tz39YX)UG{jGm!8EgbeV`QBOTreP9pTn4=+LdDbjd#T{p*;^o!&!;45 zt2t1crHlfNyy0`D>UL$+h;LSuTM*bQa|!y<=mjtwI$keT-Q9I@(3bK`+9PhdFts}A zP*}X?w-uXsDX~~gcTo4jbhq4nY$Hu<}S9tW|V{LO5j;1EOoLZ;jC~wDbCkAhbVqNxF0fHiIy_Y*&pUP>*J^Vt5j=&rl-0 zNF2qLj(otS{n5n?a!osSsdtSWor5t5`K{;Gl=MB9o+Li8(tgi@`notv-D($pej#Ga zsPey84MyAQCMrs3k^<+y10AT!_P#6dSR|*-_6Yme(Dd(aIRh*G;Ij!*Rl@R=A~idI zWcYedKN%CC7kp_{f9qF7IKb9<%*0X@7Hxf5wTc6Y&Ju0xFl+5c_i-HL>-P43MsYP| zh9Meqp<&`y_^8YW3SxRD3@MCY>HvvY7rFNWW>h}$X3Mr~)$xjuKiRQ?G*+=mrwFA4 z0z0NvZp5<)wQ2Ya_LJWh#?uj_e@fr`r2L)H7XEp^xVEqE>9S=fs8jfTl3BfIkvjcK zYtbCU^24H7vAW{SY)#fM%r_h-o#xQL74^c6Opr>uIXmc(uQ(Fld7-3qol#APYN@?4 z@^$ZVEW=sq`tLg&Kv+RpA;l@m^AciDp|D_~Azj<=(=;KmYa0+}L5~!( zlk}%C_(B<>N?&%_-O;pCCH^{gy=GSltEn7q#L??xfSPm`KSl0p!dwU$;g&&k@riCw zwa5+YeUk;)GL*_|^?Al;3382bv{Yn`<`!%pzYsGR6i6Fy-7hj@x91M{_c$c+-w}fM zAtfc7uA8)63O}N!KTvH#F|cQ9VK0&=aK7q{UUBY8kzD%Y{lM!UmSVH+F~^D+tM3fP`$=gm~ErK zyibP35)pc#DR*Gd4$8v4Nh?el6`ThKdxMk8)`AlD9&A?<7$vMA37$@U(;a<)ytpUt z{N@8x3beIk+bgl8opp6{ffLCXDvS=f&0L{L;3@;Ps!0Da{Y}tVOLwck_1(ZoS;TpZ zp&$ogM1`8${||AveJ7AZ`#gp%h#pYNOs<^SyXYwbvppXKEgZ<$B)Yi+#iLz-DfAU^ zIP^?GPF;<$B}C}bb7AoMUz#_6n+o7JK_Hq^|4rm#R9y@H*{553@xsAP5~@n8#}Z9f z4Kn-Ey?{ao?W94yDNJuAUb?;m+W;S9hZ!H#-{|psqIA})13wx68XBPe(N6o|3*o1I zeoPKAA3|x;E)hcNa?$G&md*T{=9xJP!%L>umG{6`tl|8xP6NTesCx0tP3u*7!Ma(i znEQ#h+#;Magv;@d@p>+QeCG)c^Y_UJhU+iX(R;s9g|=?`sHuGO$`0Z#9B z#B{@?Gzyqys5f$~$_+jXbRW#A{XDg)P0IT)IxmT7J>`s-%1l6HUCElVGS;GX;+E3x zju@&M!KdpeW^$kq_{^OP#a)o+a5gZE(2d)1A1f!0*0bX0!795_rIF3!Ig6H8kEI(q z%iw1rO(t3Y{g@$a$OPn$8)AxKv}24BgL)aygXggLALf#r!1&$-Y3OPK^3OOSty?%)Wac?AB0=)r zaHD_@yQ5)>zDI5w*D&0bOv7XU?A z(0`=o^06w)7zB*Xg=O}DZj||skYH;1OC02}Z?aI_h;`pALLR!e2!6rKHzR+PUDYn8WoCb@@n zz*~~WlS6R-?lzK|sDn-+oEsYeDUMA#DX(EfbVL_9k{I}&>u{1#*?z*UitMDRVuvfp z4mLUHZll(wHkP1l#GLxeaw8G|Ik=GOH8~gt3rNPkw;xfnV?3OKY45sZ(&DUhsvKHv zr`OlYoxr3UB|&#!WYgN>GN?+Mrh;SS^fZ=>1`y9Gq@lD<4P8tISX#pih779FURt1g9%H$s%P+rDWfMKV;ej0QY0 z*v?-i=mJ#sF{-7|I=RE4DXZ)7UV83W_vX4F^#)3ndwrgax^YtHt#Qp-gR(}`f;Px_ zV!WkJwr8uW_)S?eXWf*#;2MD3?ZeG87j!cuuE0~fHjdGvwzx^vSJ0I}`4%0zZ_^cB zm!%+Q)A(!ebCX|9KG01J~PrQ9x7&1=mPiXi+ zvL<>l6emPQz`Fi@5Ng5f;!nm%Ki3?)Ft#hijv<@Yw54eT7*YRyopv2JOaPBXwkH%f zonNKDZK~%m5-vB0d8)JQx~_~4^;@J4x*S6<;)iyw9THm|=_j{acG9$582(ey;3=mf z*?nV+5j>+)Q1idsP=96R0gOLXn(mLhnrE!W#GMZb2W0`oZF%4CG3+>nj1=~YOewoER)GqwsTB4sNAA&XKk%1Y2dt@t$66x&eV(Ca0oe}C6;dOdq!Bcc=X$amrTrM>y_pj5u> zi!SdG>oz5C>*i0v^0vkhSld)#)2*kji8XS_D;iaia}ZO%ch!zbC1z}IGcZ&6rrjkn zL^-?hdB_9C`ic6_?0{_3vl>?z;0FKNJ><0B#Lx|JVtMBb9I|4*8?i+9muAS%;B%&@(~tmI zQoFzy|2%6Z^7VKTG5wP{TfcESDm1VO2PZvhX3p|6Yed+L!_g)j#PxLe&2=D4kD7cP z-VHN)Q4!&i=hWFKMk$2xM_#Kc%XD*2?}KC}k*q3j+BowscgH`gxh(b}y>z^O^iWqL z;*3;{LXyd=(7@M-NXtUqbrKp@)qN6lLtf^sgV^1tGg2&A9D>m*8{fPWI`_QkB71&e zl$Tq^<4^CJJp1L+fJ65&)vuiz#;&mws$)cU5ZDvMgU$aL0KRINRHr-M%fp=WA#mWh z$3Un0jB^F*cnnOAKHY$27Q}^8^X5Wr2>v2Ir(B!Lh|W%B$5(d+dJyLCO{oY>OE*<8 zD(L5zAFbe+^~;WZ?gx~m4>rrzrAZ=&)&5PVJQsR=f-x)LN#>oKXD5FM2Cex;7qe)4 zTdUdwl5kNU+3%0UK7+ehN5f|*vVb&(cU6|#E(SY6l34?nvl~=O=H=G9dZHK{AiQy} z-oSyPTVVnDuhY>Dt61+B&0}6rR=!m4%kb0G?NmleQ*>9J{z;;6{1`cL# z@Pa6gMw93Xv#NEFsG$4)=Hk&4_tg!n7r^*$g1sE9IGJ#LPov=*3*elNyhHRH{tI~ULnkreVY0Dyt zhuv4dh&3?nRhSK?3Ei`5TwDzu9`4GbIc@QX41~=r2Xa&0UMzepYcsTm?F;%KniE+) zVYG&ukJ+#_Hm9bZs?#=1I-dd}o6#h;_OJQ3it#5Ue}{JN>Y2e{E!cS$X{!ByVyYl^ zLo$cb@#(mBbB^F zcy0>~V#meNApXqk*0)ChGa6AdPwoi)3ryAL@>=MIP<8}CgW#si?( zFTc|89u-ae3bG_8dwG!FwJjc3!ML@%Y0J*-pwDluM9O;=?bjs|!?*wei3-&tnU0|D zeL})l7**rdw-sX0gckC^Vicfv@NWn3to$#C9+OdH1ickeEtZs3*l@D`c344oH3)y5 zK^bhMoe}AvAI+doF0f&lpg9P;ES9$ho!`RCeZ#<}P>p@`nD*v0Jzy^^!9nubX^dA5 zYU)1!Y+j)+mIkvl4p5E#UAW8u!mQc~z!uh8H|OidR}{a8SA)zt$StyEPd6}Y&%Y*A z%p~~~v&^7(=<7D}bC{Bz>@+$ym4rKdIYtVyVH0I1$i3}23N2G1j+)$>IFlJ5!tF{B zTbNbSYKV&imeAZ-LXpr3wlFnO0>(@xCCgwC#v2+X=?lu!zu7gjgj_En#OuVF1C9<# z)~XrL`RQ9MI9G(}!jUVF*OdbWFWXXC;Um0=r<@qiuI{FW-M92ba{+ zwKd>!0}pGuf!t{M$0sAeMj-@FYylf;n(#}uFiE-^fb`7~B^~nc6la5G$XDfc5XfL! z%LWtV*hk3ulb7k%R{)cePju@o62#ekQJ`-28YVSBT#5#g002=y)7l?f7{kSXzS)-E zw8t1LMXH*d2AFUA|ovy+4iiqIZusjRnm zi4rH>b!o}p;$5xs-I%u2x(d_eLbnfP9(gyAGsDY6)J?0cJVX`>XVR;!r?zifdmmE9 zx~mtpJ`&RuWigGqtac}229^Ee<+!RPjLS1t@&Zr6HJKW~@a&7^Q2ROnih8 z#&UQ1^p&J*`4mbBAE;caO>JhT6FU!DP`Qt-SAhrb`(FT%C-h_<+!Bt$*f^JOQT8x7 zPCdp6F@&BESiFtcRQ^RXKicmZI3FXR^YMiMdYs|rv1MvYTo@Y8sa+>L`Sh*GU38`j@lf+ws+DsJVH0bDg@e8nE zn0(7G)1x3A{kNcu`lo&we;&1zrYf-OV>k zX5f3VTJYntRuR=#%{ zW%Qja7<<$|BRtFC{$I0RUd9Ao+51%S)269ufuoML9SO4QkY;X%_7-esYh9lKTpOb? z;pzg(T%p~!gA(DI_#Z;9Cc>fsC$i97dHB7U{k)}K;$9EW*I|>9o_s>HXiyXj+fCx3 z5u#zSrA{#y_Ru$mL|&j*feTiJ=r?@9%oP>(;_pCQ4&&Pz!>N z&SVQ<#pHYRMV3-y`Z+Iz%3L2mXmY+|VZTRAN4QC9wOlZKfd-qx+2E~Vt^IZKf@QP@ zhn(xejBseQzU2DkYF2!b7=;MTo-5evF$`Lo)~j4H7+M?&osX4|;X4g^h6e151{|4a z5Y8Fr8w3*lzQHuo@L_4GXgb5SR7@7f6eWK-D-t;p$|tJ*b322AXqb&u;s1(!B?5Nd zE7kKkwcE%)zdI=H_x=Mq*%DCfA#W6jO)U{VaG@LM7XxaG@4w?&2OTJWnONefE z-fprmr8f%C?x}8(gRUcQZS!&vM1KeJ5^+z-37Mw@G+IV)+%>e8(vBzP?d@rC#x<1o z??ZXGX$aQ9`M+dHD3cvPF0YfiFzib+Aok{i#k7Q7l05KepDP&KgGPz>QmBC*_P(CDQv123D5k=fsY8Deaw9=)RzdONVw#07<~pTXwDCizRaYzeAf`u_$kk6K81gzPzG{C z&Vs=o`nrprX=Mj45?b% zTMe3}=R>FRb+{D<=b3F}-0EOU;zu^?x_z-RS9VjQb@6Z{H0sWaEk=OP3mau%3~i|d zD^Jd@+Fazggw=4h;gE22a`TfwnItKg((v{k&#a)f@rX52Op{R^e$1zy#u@wRxPg4e zz9w&s;6qLBsMB#}52e`aC$shRvoM!*BT~32P^La9NE&nR-1f-_H(VvN%j;?-2hx$R zxvOn>J(7#~*);?>uRJ3_rg1UPHtH!r*ndlv8DoCnhFh8t- zXFjstmDgKtsi~#PIo(tHMG_$Pb||%Q5I>SDI0#}hFSjP3L86Gk!fbB{>&UHE&ti%B zO43)vko9Meej$!Hi$9KJD-d%)$-d6Rv;_35{+mAt9y;SK-K)wR%|;2-L##lECSUXz z`a-Hc_G9yfJl{E1_cb)`dR>dT+ZGO*uE0WrDmamNDeKf%9?|rpr$7c4cT4Y*Sr~v# zuz%Naoe_?Dc~CM~^(#z=z>JkEA+Q0RpsB0MzR9+=>n*SqDId3U_yr3T9SX2_z^hPt zfqzP8W=|tp`O&4uR;8{wG>rS=hH&10`V-bhR8SIY=vV2!W?ua$xgYs=Eq(A-^twn) z(&WK9@2)dN_aPEiNn&cveL>tV_9zpLlsBHXL(R9%xfFWh%n6vBS5?Rp*YOcFUpn`I z{?$X!$h*<%o{_QpDM(B!9=*J}7M;))P&yK7WHvUiN3yVdgQN^BQ;9_C{*fueE624k z9{lZ$B)3%2QSlT%&vzk~EprC|NmwKGp5^w9b)Hcs=yK(43S^AGJowDSh!$el#{NSV zjfOVl=5@gcBAf*D5Xd(DHHW%6EiZ~ed$`howFo}krQ`F5yka#cll`De5yxqb5sI&H zckuefa@IAV1tI2d$sUD9bYhE(J;Ifox(&(oGic@tcGQ^UF0!rVLAXz23m{b3rj(G_u^?K(xd5q`d+TR5dca-UD1$XDALZ;P7V5cBH$r>XVsc8; zSg>fvp?@5);HK&kY%KoP-AD_gs4=%XTL3TomJu!{tc@>V8D9}=Fsa6UTq<8^_ahTE z_&=_p{~t1GJH1+PW-e|aEbd8xm#BduSFr_A2|AhcB7!D`sWDGfrc%YJr^Kq;ZU&2E zU|^X&gm@8~G-y8RVgz#u7TM0@pwYEie2T4ulon(jMS63cUp~Gf-EV;yh*@v#N6C{h$Isqs>X#xYnypN3Iy~v+U_yKj2 zC;xv?VU}%V_RetZp3$BODQ^5rCQ4jh2`J@nlDbWhNnOjNw^`I zlCDgL8btv}_9O2C=+7nL%HDz?hb+ObKro)*HJ`WBT0s7h537c^eK26NG50~;tDiu@ z4Ya+nSAn!YEGaY{qmvM>!k+}?R3*1*2!0K7EZQWIo0hQT2KVG_npnaof;Mu$^ zFYU$fZhrpR-V>iSm#K8^j!$!XH^|2#(+J&Zl%(a!>zhb}PXx6J24#2lB*u+drI67lgh{>q-(6br_54y#7T zA98wOR_Jp1lN;*h>2Op4?bb2u0h9z;ia%R zWwL#2EfuU%+{LA^Y%)(x992&{o??-L zko)yJ@<3Nl+KuMsIZ(a*`-6p>UCCKnDwZ}JP$d1Umsm)i5oW%9tw&-cLIpK~irQ-W zDtt_+IfKBguy79atG5e@#aw8#sW|QRCo}6YkX%yyuU)u2r%jgS!=y`H+pWHvCU+_p zB=&kGjaUzT-l-@<>(JG_$Fo43NB+wrK)WP$z{YZb;?P@CR~@T${vTto6A|G+_y$%tb&2AU8m?6(<3tMKg%Sep2J8r-)TvDZrO*r_o*gRm zM_AfL5$+#vo$m%qm3d4c0zlkWTsKK1g-UIyh2-yy>XHR}qS8Nm`?J2KTvaQU(5hJL zOsP34kxvVv#f!HKThb_nR>py)TqHaYnr9CRzh`AX!SZh=Lfv%e*+bstGjGfT_OrTB z^ZZ$VC$pC<8@zrnWC2P>Ee$8ScB~DmEcc1>Rs0bCbygG%SFN#@(&s_;5tmcS{-$d1 z^mvjZi4LdO!*Z)OC|d2B$#KNkW!nrNJgsR{%czU{FHVGbFz75jNGT#HPeItFH&NX4 z@9A2Ze!3MO%61@|>OkB)6)Oz0E=<`QxK-#CUi8FNG zLO$;N@fb??ir} zv3MU46kblx0Z5hoMP~pCm!f6yktU+Eb!|-NuF?M487yiUdURTs%AlP@#a^-4LJSOSS%Do7DJVafBxlJx{rdTA76inwsr5wiy5 zToQ#ZXPsQC1>``toH~9>Y3}s_!!T5I*vmQ_m_X;TJld;9#wdv8I63HBD5iaAevK4CxP3?sSM}cU`*?G_w^|OJ7Vt+&NySpwzp)P+#};qq{)9 znlc~=_bjMC3LyC{f8_#Ok0cD16rXgm2FFmwR@94<{4s1U6(hib%LMaIszgqHcflkr zl`>>;2werKff~PaL4kvn;LMrULtJ$XiVs| zCN74QQ2rEEhO6sES=HQ?o<0G|?XI=Ja$>rl_$_|Brm=**qPQI&TNGM27rWq5rlt-D z@WnnfwI~lSm?s7mgTu>HM0PDq+Ho@d#tI|pBP{fuT~c|;@|W$Sk+vnBNYu_vXD^I} zcF|B_^R5-zOY>G;{6dx|w7jln&bGfd(R>MqJPsUrWw9<&4sOJ-gTCE41xx>u*AuP*@z}`_)7v zA}f?rtP;Y(DQ4Q7gsXekra(GLK9sMFIJ==jAxes_&HDjQm(w9J%FzYM;#_E7wqY9{6fz*D9hmRhsmH#CnQvPHUy z*4+8_odB)JX-8d+nDk%{=G?FjP<{9CcaOhI&WmhrzbF@lYXDEJJ<;TST-L)(`I6=K z(BhB2&{M37bLzIvXh(%5Bd7TC`v~HC_(a`E5*Z9$!C0nR9*))NJa+B3@Z9_vPTYWL zt*Bs@VuIN{ghNBlq|uy;D8=|lucz4q_v+t#75f@`RL}vY)a#y){*P?1dTq|u>sglZ zsK@wASHo)=vO$*T7Bl?rI}17SFaZAp!GHAGBoYGB$*Q<&W&XYKIUzn|-4q8G6~i9Z z4jDQVP0&Npsb0=}imPeAQV??d1A872qr%e15Y|y(LCtRIW!vNaSDIwA_@INPxP;EQ zZau@j7U$`-9P2R_?UZ2`RNx zsGP+#`!?@r!F0L|&mg|O{UoUep>5bPO#Uj(1VwrAjC6PdODX?W!Xy`;wA7Ev28d&X z(7rZ!=Ud#dt`bLp=Oo|BVz)GjCv-H#YvHZ&JAF9=O}G!?AnZDfOgTV!*L!U9PA^_G zg|nWaDW$5cEu7qU03S5`8zpcVT>CBvS}Un%;fu4_xR^LKmy71VkGH~~W!IhoavrVi zt8|KV5M}Fi7nwuXdz^$8DhtA$v|l5sO}=iuFk9JlarjQ0^j^1l!dH~RkFVA^0u)xn zV5M%Fe8Pk==kC8*I4cX0OzKqtZ>?HxQNT~l5+536N)UqzHpS-Ewp!`u~v)? zOq5j&%#FSw;1;bt&H)G~P0t_}+B3qFysc{;hSw9FeV-nx;!c2@Pz!epzy?Xu27@&Q zvSr5xjp!mv0EQ)ae-4Cjv+rg_@-KwnAB43{KYgw-@QFGVhbzJC&lXC4*tj6aas2Su z1E#IzR2Qg*!V}4t$9th(tTf5AZBh)gO4W29u%VMTp{L?Ad-*aF(^ZqIkzrS}a2R4adVk{nHiPAQFI8#tafCiL@Qx1l@?kuvj9w@F{dPvQL9(^%i3eD4ut9Ee zugCj=$>v}cZCBrpBknsd={4I*k2(eA#-T0G^&VOT z^N%SgJ>(ww3-ra*ueFk0D1Cu4;B8Br0D}8KoSY`=YX$$J| zmPI2$C_=93yNs(mcK7Va8cY}^VyF!f9wH+VLi2!L*zcJK%)R?{39`=KqC5u8pG5%* zcl#+5f@ei3h=>uchL8kggztD`MLS9`WBAbZ}RSVD#gNNMVK!vpiq=9CIJcl^+C_=8}&NQ^DBl@mz1i4 z&6J)!E4;GH%ZB}5m@4v8yeNgglMgSVXIcQyAHgdZ-~uvn#f!f~Kss3{2p3(j2 z?e`R^p6+M|5J=W#FrfbH${d6v(%a_~uQGfMk3KcRBw>d8XVjT$ftftPy00?Q*ftit zud2NZ)}0s8aZoeE;w_ z;bx9748t99wlK0qjz9k0OR&sD%=0SUKBvvkEr!O#Vhb1BbiOKmcc?VJgeT5q&TPtPN9OQgcbgT#ss?7C^HzPmWW1G;;JmOUOdq!EP$1&oVRm`hixe%wjaRV7HZGsJ5vjP%etT?@qMS&j zk&l(!mCg~t_1G&dy9hNc-wNbqK2mnV?h-LS#4>R-CLW~mdXWb_6Uq0t7G`RfsxQj# ztKI-gQGn|$oI-f|gI zDnSB)=21KHc3Y66FV1Hk397oM{a;yJP6tMao55|OID~60osy}|;7-4A zuD?n<1^j~cd`*HhTT6$2R+&FiI^j8GL&lL5-~rP5g>EOS73RbLLHUAOVXmq}dKnJ7 z1n1~R($wQs+WL(DdvSV&c|jvLo~)TfZ6glae-VU6`iyBsjqQ!Za^Iw;%lQymjeg;q z8ZIT&%LYs!ntLOaK-;i0u3~$->cg!kwCIUcJHDYHuRceEXKmbANfFttVHYajJbGnd zW|e>S4F}52VW(#Z<|>xXd{rIdS4H?(cUv6dQ0x5E=Bjq;SS*^K654>s__YT|GpL1AX#XV7?y#Y1TrZ&^GOb9q`M?e4!ErZpE}`)PRoY+`gk5Ejz$RSa(^Z zF|i$jL7#A{Ln&dZ19L!3^C%CvJ6kTIMXCRWN-xky3Q&u5{SjuxD~(8s$g{MO)>-$5 zxfZ+lSNB0w3Lrf3hBp0wF)jatR1(=6Tfy-0{V%KXKc*!MEBF7-{$pCQvvB>tnU)+J z?419XqW%BDv}|>DDY0G^*lx7du;InuYP%^h-nw0PbLaWbI&xWVr{dm+iIM>hjN-=9kaGOd-JzRsNi za7BXtmbQHTJqce5-#!r>7Qi`pq%6DXn+BdWH~&o?T^#>lG|&`m;Q z2rrCIjIL(wzh~SSTboXKGyt0cBB_&?_jelq$)CIsv;@R=G2Iul=;_!SkiPx$Gn0C2 zSk(JJpZ5<%!coOW-c*)Eh7=)?M7UkZSdcuw0c>f#4JdrJo=Sw zjpd);+5Pe_Jym}#JN56fyS(hLFtBu~POgPY$1$ zj0-ZrTpKKuUe5#q-TFrgW_;&g?u{wDtdW-jItt5#ymb@;Dl@t1iuvYGOt2~dznvz( z$uiq3tLnSsvoV)_xSxIM5^m~kdB3)Af)qIB{uRnU{#-iT(K?fa=p zPAp%z-h|%J+KsWu|AAUM$J+=(yqa8S{%mLgd6~<<{x0}V7e&VQW;U}11RpV`Zb%{m z!tW%HejC7du0_Fjsx%m4t-^`O` z<^=xm&;f}91m}+jAPx?$58xS~xrV?0wZk|`90(Tq6Uze?kNYE0q6R2^+YiGSrhMZ1 zzE1#R1-_2~61qSVAHPI(@l;X%q!>WOk9+5&Wt09$Z<1)meG8IkM}5eg0Sm{15wSO* zzPc@awV9a$ix~h=7Er7i0I=X<)-Sx4KvAa74`g#QAkgl&aqUh)xi|AF~^hph44y!ETCC+k1=-DlwCZ-kCXSrD5UmU=M;`%6gSB*7rB zPQW8e=pFtc3V8VF^Y{8q4fwuG{Xn!gGAan?DTWWT21kKnGJWbEcOkH~G^N|C* z06whGe067khy2q7`=_Ueror4ik^eq&5A)PgUsvaAE{UPA*OA-1~mU~2k+iHJVJ)W$@RUse1Cl%xtk!i!RiWKF!Km z%)3$KMMr$(c1SPh9@NqE-|0s(U!rQ4WwO%24dJDSxt!lftdew4$Ptli*Y!x@-uM1} zxcCgjlp?;=2sPT=(wX(x4Y3bU(khoK115hnVHIb$_pB6MTKRoivNFkV%d3gyvp}b_ z_!drMy9?mu37YN3u3f#QMA_Z{n4w}BCkllkI4MMLe=>^=P0cdd(MIdpUsQQ1m7`&X z>`6b}tGerQW)jUSfWWjjO+P-zl{zz`47enN$&W@N)nG`>&b{kNa$AKo=s$Ee-n*w* z8EL5@w`H~=#4KS!o0+QW`2*o7??%?WQ7QRy0X-IcF|KcCCW>bwA1g}>x=sn5T(mu`0Q9R;h+aI-zi=H8;N%Mu3TXBI+I2G3PlA?XGhk*f_~uL~12Bc-5VXF>iVz zBt^5AdoT9e5Iy{?+THF!;SoC1Ka5hrTw^~^hidciHGfBZo>O_Dh`NiLe(szKQn^|14Q~uW ztDrs#oqU1ZVPp)tTohxW11`h3EXg3qc}LXSAoN9y;6Xb}V|Tp3yL(2jbhI=uUiAz- zv92)uX*%hLh-I#Et3s2mMQfvOO!HToN}`Vi5*m&VWv#>{kzG_KH7pH`ewk7tj622M z*oPCR4JRbK*e-H$!KduVjpmqPQD_EIKW#g99gpWW0pf11L>p$_wmgOirX-@Y*{?XA zK6JTsB_;9a!A#B+If^F6H!wg+=oXXlP$6zfqLKfSIc1xS(G&bl0;{g0XsrjNqGfc@ zU`~@IM63_5g^{YmDx)}F)9RG@sz;??lEm8}BdVtUa{GbK_4r$#RW`JoZ26Gkby@>y zR?*jKY0WR*XScOCyIzmFX48aAJ-`bH7T3{AwVqSSDHx-S4C?}mx;o2`ktgu94!l19 zX!d~IE$gbAn(haD)FRe!Xuu&^b#N+rY$C0Exo&T8kXsjTAT+vFrPFsue*{9GUULym zpY}E&csY+Go-dmk$a-Uo@((EI$3L6lC03y0;CuU?hvMIt^}kbrFX|rmFoiK0MzEAb z4(cOo_Art2SRMs*IjMF+UO9p|b^Nt$s@{}b!Sha_Yw#0PMtKgOu|UR((f2LwT3VG~ z!k`iP{Oiq<({e)*hDd3cYBsur111B><|3`GjmlEEOt-V0^0S`v6$SCsO;3Nm=g)XJ zRIaO=tKI8wh`_k8SU+qm(fl^CV(PqpNX#5)RE7A>lhR68mTJ-EK9W1fr{;8g%a1ib zg>Nv$r!9CUE^`#|7E1~X%x_H+8%qHi!l_;0yOL;b5Svbn&Edy%Tng8ulzThh>vHyz z2)T^Y%B-E?z+hVZZG6S6V$(Nzy-8P1;Va0;<`f2)27R(t+6NVSiu`J zqGKysM-`}luPzs0HJG&AiK9l-Eqj2Z+J+Ue3&3S|+OzhvSmz^kO!$e0JlPoCHs!g-$WeaVF#KJi-hZ zvoX(v%F-($?9J#XkpG<43*I;1#Pi#h+g~`nNi{ioY#LYW$iH8k>#VQ-mXvL^IkEe# z=|+4ckpUM~UBxJW5gjlrCPAfWae|Pq zx?Jf9dF^F)@x=wJTPe)_jeQ_bD6|(>U!dAjMMZr|lJcJ7b1cKacNgt|90haLYXb zx(Uuvp^l0~yI#fOT|{7at{&idiiu@&D{)E49P!lPmAhr%AhXXLYT5QAkbCpsC;F1x zbQHvmp-uBAd6*{s14DdIv^7DA4_xCyI_8i75}}*GR_+N`v=zF}bVR<%6AtOuUG1cq z2j)JrTJUf|s)Qad$eAQ-F92Y&orFXwOXjj^0>?5EVZ!Np5_l~2TICc*<^IVD+~`A; z#UxPus@gD8K$git@U<74!ah@Muu(%Xvw_;JwD30jpk#A`(*yd94ZovASI$?Suy#IG z-<34PB7?{AS6U=;0ph_x#M_;j$scNgoya{1ht+<`RADa@&~mhRGXmmX{gOww${xCg z)~^f9r8r=1)nAqc=0EkcA>*W`-iXtzKE)#$(lPwYBP)sxS|=)W>jKgnC(l6}Qvh*z zTdCN$yOK%^_FyZ2fKnbr=RuNxQ~fHQD+`3-Kv$HOLnm2^KCHyjq<# zx9+PiB8cYspNLudE=K`=ox$7U0FIRzYqcDPK9|SC1mrJMLt40gG}$l;gEQN^b&6do z3sQf^)oOS2Jz65TtY!sr>IG9+Qi{?^V(KJQ$A1<64c7qg=KldbK*GPb-XC(+r?UC) zAR>cdhWlp`o?nqS42qBhGu^5Jh{5%GvQtK@WaRJJON8Xo4S=4=hz??{OcefzM_Gq} zEtUi{n%^kcyWSxzowz^0P9`wYz6Tc$IA~M6wEM@2jWaHPWB3}!g$vRqE6?oWNO}be z>x5lXX`JL&X7;8BgSS zfq?>WUnUtQg}W!T6zra4#F};)m`}Xk!Z*Z*SP^qjW#5&*;9glyV{y2`VED-;z!YQu z=7bUv`<`TC0kq=?16sCMS z%A)UB#*u8GubEp-Vves&v~>y9R@y75A0C|cD3mSnEFE^iAfNP+pK|27OBV+-Ir!53 z-FFE5t3YlBUdzi@r;im;rZ;DbtfRf0R8ZjEE|okbDh4RWIvsOs>@BC@+7 zL+J(tp##WuOzR@DJrD-;Jbs}GYXoY^J1Rq&EWQFVNv~ua1>!+v2bAY^us&mt`@|OO zP2V~A4twuL6j`e-&tAzfnYOJ@l0YXIAnRm(ezT0_C7$GARv@}NQy#p;WQG`mCQ38g z8OUF=qmjV%c#P}p`x=Hf%ozgjOs%j;DPTnAm&FdJ zQ+>9qgZD|_ya@MO_1Jzs6WcZw-bpOveMi}Rmrn?D0Xav5DbhZ{1FgQ6cQ*O*UL&Wt zm@U*?ayvR?XA&#mHray=Pp#0}#sP&-|DaY^d?zW1j}AV|Uai;ElcrVS!b@(`7i>QZ z+DBOtC|;rz zsGw6yEA4C9HubW~E^k9th{@WAVMbL?tr_bL`4kKAUJ)Axys3H!x4ZTG3Ur^H*eU1q zdRdL-6Y0U;=uB!IqCGM_@qaxMV%y~TT6w4`MihXwR$nH(h}ugtPp)+@Lb7@uE4&I+ z&&$qVM*94EKKIN>N#AoZJ*3B$y|>w)(6~^czMKIKgH>JHS3fz5a0^a7KcS+ILE#JU zRpWs`OSmAjusY8u&1{J8@k%#7oFc)?P@`P?1Y4ouX|mCQ_50^^O9?Nj`%ljmB82@H zm;_j~IxfSXx;w7!km^TqG$1cTER>4#WjPTZmguNns5LVM_PK^Vo-osP5C+Ajb0f9; znz0Y=yct&2kY&*KKfY%a!i}*zSB-L}@ESId6&hC|x=8OW!q1g+rbq0D_+IOyl!a<0 zym3{FJ5M+|nd&@A(XE@rM&_+qr^_NHOP~z93YVG(jIw6?(`&N;)&pcw_0sOq^JE2B z*S;H^La!->rR+1us&|o)morqwQ}pmLOogN2m?eL@_(O5=F3rz@=;qdfo?Hkd-cQ8%?Ff#u=w% z5r^hHl>;L!q{A0%8OK#ezY37#QR`E#=c#n6>uw7vm(3c0;U`u)GP>mo^*f2YAGT09 zeUor?@N(pJpxZrkS33&dLcKx^$B!??(o!Fk&*g|I_4oYxE8(}BwWRg+Y3|lsL$M#l zfcR$1Ro9Q6vY^l<&ayZEw%^xI{qS9&>BCdtWOtz#_smvas84rE*zRb-b60Gc!*59{ zdGfUhg6MHx50^7CP+P-PJ}TlD3p>B?vxvgjNl>oZyTOdr4teuJ{QB5&lDb2;<>x~= z0<9JlApqM|N%V9ZqFtP2q5{^L7Zf2z^2A^Jzc(k z00@l&$vuV(fk{p*OoIekVUEN&31K1-A8l7(Srl@3_|W!kt83uotv*IoHS4C&moCqf zmc_b#94s60ksX|#Nh-{p^jnJ3`$C))2a`maz*R`!%4QsgX|q~fgO%+)LV|vP->;;_ ztdcLIXZ&B^zd?m)l85adRu#9NfM6-S%VD`!+?rUe+0+~VYNa_K>8F6yABipFQm@$x zGL>FM<}^lhE(aHZS6OK5X~{in-F}=J>eyH==IVvCi+5)tm4pcBWKp+IWz)zWxzPh!-3kJEjB2rJPywCQ4g zj(q8A%GWy1|4q%*>0rI>5jVhHU)>#YoJu?@^QdW!r&94r)v*lrX*xNwQd1w$RkgeE zR-JtiHnS}q9o|P5jYBUv7ccP2x!IkNQWSo6ETv$gjeF4`CXVZ$Wp!Y!(5kOma?y_` zq5}Opp%z*$HF?zSfFKdQ3)(d#lO0+tP}6yLHok_=DkF6<*+Xyw#USBvyg%W)7i}!x z+@K^{k5|KRg}@?*tZzjmeLaYp*W;(5li?f_@15yoxE}_~(!?3UrAdaXs_&4dE^e$2 zXA`-EhB#Vb$Fedbi3hZYd1RoVq~xL?1Lf#+qk2wrn`MWYO!y)~&scT;R!y3o+4>_AYA1Ys9uI2vx}pgOh{|TsBxOpvE++qyFbasn1Oy0Bym&C`nH9w%LT5 z_jd~M-$I^qh@JO5dSZtuBUFbyC=5D-y|BvIIu=y*{*`$aUA;bq&D?h3~(- zop3yUnc{E2gKMGApo%kgOYN(+n<5(Sm%VALTgAhcp$qFI4JDv0zf4c^E*SAuUr;P> zY3!@o3*-raz-7l=&Jb|TR%5ax+;cL5Q zMN{7D``JgEW(9o#kb&3Owb*0QD00;@&m&k*BjzLE`$@)~g~w;^c~sjqfetZ8QWhn1`^;sfj!?k2CJ77+_LDvKyy@BI^B#8@MGvjgpUh@# zD4(>9x4NMvB7;iKFB}liw~*Y9&g#xOU}-TIjz=ZRy?tx;L+YFQU22RBlg#rZx>%DbOPAy}&7Ak`1n_bvChcl^fy&g@6cN!WF z&Y7?=91tglwHD*DL#JH;@|9QmiQ+vd0fA-Q09@?Nx@NH*U7Z9w?lM?L0SuVSso}7%j>{r>FS3CLOI4F&yGN#Y<6jRF62!`RZlHj4e3ow0GlQ z9gj)_VR7jqH&6yLTF?Cyp4frAsvBlD7ndlrn{^WEL`7s6zE0>F#;eYpyj&vRQ&GZO1Yr7PswN;T#jPH3Ka8q_b#g48_N$j19DO{JqW7g^J?S5KP8i4Rcgf{(2;owHHwJ-o_~l{k!n*9v5i7F$YM7M5>S4r(4nK0Jmepn z%Foi<^Jxrit6&Nckj(p)WeP$sd(6i;fl2l8sry$ZbP&J74khv`zym)lG%@LCgsM}) zH8$*=1SVKcglLTV+Ruzh{aXN0zF)^ahq+(SF z>36q+Yms(-WTPp$si`Gj2WzI#AOx>2B)#W`08n$9djJ`|3kW&%l^u_t4=sNGQu$uw zK=B|g3~yNab%Z=YGF>^kc~tr{s1qD4WQ4FSJOQVOm`y$TJK}fb2&tg$lqZGf^RHOO5rI6WM(cDWg!>L*P-skv?3Qg>?q^g;6!t{heL67R1cf z4mn#bc0<3rHu2+4*EM;2saY!K(>lJh_R(*CG3$*b0_-y<1_-K<2som1{wIRb@W>cuUqnRraq>Lj1Ttj{{jt3>XD zMu)GDDlECVu{8{g*G z+Og)^&3;pI5ZPAq)wW+RVe27&;~Rl=|QCnn2v4z<~5p=E_YIyDiTGH5N(f%0?`SACHuUgO^(CEhX)s(2XSJiFG0!b#vCVx@Ie63~t^ zSD5Wml194pnmy0#igbj^86gy`j*e~E$NkD7HP3FSd1K4o`&wN#Xz9H$znf}OGUhzS z=118fHM}|-yULAELJU^lVW76n{uLCYoI*Y3s*iAMAp6@&GY56S8sg6$%}NY-Ii3Bn zm*HvQ$%~O$54qf^qMagKGn9o;(-+`iX358@^*C#cP7YX){Y{sUIUAABbcV#wFJnwQZ z@FwuF1BF#t0<_6_-|Bev^CvHKjh~&mytOQ#TTE$d+t)Zj&BXUr{M~Ja9^x8G>Y4Dr zHm4W^B{gHGb8xG@7afsJO+&ZD9m$)Id2`RzkwGAEDXwwsNe@qR#;n$KxyMq)iSGu# zzfY1U7)*~yYVBDLi|gA>Ch@%ZC~(Hh0QFuRsWNPuU&E=UBnpBThhpaz8OaYrs0U9x zOat8|qF3A3l%iE|VxAZ}x8e;zH_se5qw(~_Fzix-#af_0p!ia@$bJUf-Xv&))fdj*OUh%W> zGGcBKWPYz85g@*CBJMxcAivsWOfMdO+05>VWAU4&Y$2=_H`#&WFKf6hUZQx~@d*%e zQIgP2?0Xf2Bt{og+@^LtL8!BTEh^0kBbIi=5P0-6aHqHfQ4+*j#5mQk?J2QM?*uhq zXC_`m} zowlNQC>#4{!0${Ps;}n&r@9p1^a5I6L8sF>;hjMIQVQ8HXb$)mB*ajoc;{YzV&3h^ zRPH$XH*JtsXvvbwiIjCIGJ+w{n`WX&2-4jx?{H9=rba;hyJ5UTa)bT~WWA|k+1onN zy7zKoql!9&Hbp@=3|V3M`SMP7j;?WGdP|q&xBFZllIo4>2=hUtq!JTbRBqF1P~AS(?n6f|`+ui%CKP1C*#d4W z58Dz;97^fd_jS|b9Be3v3!^o>xD>2@t?X-9+==PEpR*$Z)o^l?lAV5MN5DV-u2P7S zsrm#*B#Si{qpz{)bCI8FKb3rC1^Ci3VI1lph~sZ{sJ56hN=z*DO&=jX?W`5|ezk~F zW~0}8xw+8tR(C;|ojf1pS@m>Tu@pMBIeKQHj32}6r1knX8^PI1+*uZ=>}X$zca^u9 zPg=C`x)s<8uk;f_Z5nCj#j>%clb8$Ml>8OE{Z{^UupcR5e+72 zfZlH+hhW(yXVUcp6s=(QmVyzta|caHE}Bp(M$fIl^GgZVofeIf-|ydyXCoB9HwW*M z_`TiVu@nl@0=D`PouWkumrUVnFzzP#rRl>LyEYlHJFKC@GmFvlk_>f<799>w7T5Q8 zW`yC$_^Ap)r&de-mZK$1ZWe>ds$YyKfHk*x;|KC5$u^>>HQZ6@NDtaZpT-n!;*HSW za`8RzdAm%tk!aa@x#@gez=gWWspusxdoe`QH^UjPcveaXu50U2h|ikP6c#VFwKFHV zH1k}0M+JM#;N42wTM>=no~p2?6HrWuD~>RKZ&+^2X%d)M3y?^ zm9U@Dg&WQ@OUi;OG?I4SpM(ulJYBrNO>PtE&v_^{Wi_y9{k3D)=E*d}wOwn8DV-Mi+MK z{jJnS%v+YTf@cJEB-VTO8cjAG>&9UA{J51|>}ura?%jgHCAuc?78vAe@8@UjOCaay z4Fbv9XBxyW_d}O5&_7qnwWm_*Yya2}Z^RTaPi=c+>}b=^zu({QvMvsW#UhsJ+pxsb z0wyPAxVjH{oMFX7N&$IeL@>Bf2uqTZzoV5D&+Oe`0Eo72@|yYk8830vJHg5&E_3d z9bY(PR@7QwOlV&V1piPu4Z%B*;pt(R_45^3HWj-T1R7OshbbOV@si9=~L1pYMHWHR*4o-=!-E%K1UT zghI+ES)ObN9Bf>tt#afoI*d<{UAqlowv(n0&OQc(`NtD};o7E{$=}}FUQ;ACnzO#p z-g>btX&BtWFY>-3QOWZcL4tOW+;@maC;s5q6b5o!V#e;JS#ah<;nX=d`Be_%d2zc)09RMZSm+T62{ZJPp1R2 zV!>)A`ka8g#G{Q-8&bsY6wxi1gs7U%u|#7pw{sJPufSIM)HQlU5kHg+MM^(EL#DE9 zJMFyzUt4mM~aH0r!Wnr>cg3&a=Cff z0GFTJhrxSSIZlJ?m1)iNkLM$Bjk)Q*^NSsDI?}B92f|?g1`-_7@jJ$`sz2Ah5YP(O zrz3chZ_>ZC1UVN0i;jm#UJ2opS>73+BB}%}+LQ+*%aIjP3LoH~sRSnhExFaugBTnqX z3&cuw4=#n9ry&gLszvOEoo9$o&RQnn1IogLBQ6mN=x20mU*M)g#_BMu^@N4k6x7yr?^r+Xpv>Vn%#aOMtmP;fR7KTk z9XsX{g@Q3JAYL8x1Tr3^~BBgP8}|?o%o!f?Sp%&WF4u zUJQlY>JoPWk?Ux9@bUD?qF&!#nXq+mOP zYAR+jnLyn0N$VWjYkSsy$V*OS2bV2(o!kDDk3Af{{hLY5yPs$FjRQ5e`iqbKz6mT8 zgffKFj}9SSl2E`q~>Pw=o6&mmbX!N2rZz zpkXf{!)@8{0iQq{TqRsZB5rxS z#%TK-c6Y%a@}unw0q(DB7r#;;*=%Zdyz5AwzsMV$D?O6IpY1K{^5s5{-7Zj!zkM}) zPn*Ki8vew4MDq>YX)8aAPHh!3Ai6kg8QEhmspJ8NTi$ZB@zWPEJtr<7d|`KkJC`E7 zxUj8$HF7?1flO11h#9CKw2K4S;zZhH^IHHcU|gccILAr23imPj%1ZuLspE;0XTCBH z1dNk!aOmDcdh>RO1DU~%)ZD9^DMn9n(!2}S+41ZIRCjvz3IC6bUa_jRy&w}%5;$xF z)Q2?;c^S`IW32Odj8~N)pAZP7@7p9hm%Iel!d{b}>+n>w@y z?EuiR8?biKJe)3}Xra5T9&AizZmnC1g}d4BS>EjnMpyNN@!bi6_w7%4ZJtt4(CMF% zsm>JE{4}j*HV+*|rPdi(8PC9r6dj{|Xz1XwTGBp0|9W;Q2rdb6&L%o=&;W`h75ws= z!YJ4>k7RhaFu#EsOY)VAk8n?Ek&RLv$rxllA9sakSTn6L4J@K$tqhi-R^Zs_LL*Tj z+_H4f0iH<9<56OmE2LlUz>9vN;GoceQQE9S5^6#p5YV1XphyEDuDP^Af^hjd^%a{h&Gg4rFC*1TFLo+9=9ob+##Ma|TsrBkoP$ z?p-^X_xDe}3d-leW6a(xw0S?+A^2cZ{0UsJig1dJj?XYFUgig?6h9tUPER%uJEQW{ zIu>s@hPmM!gXrnSrI8D2Wm%tIhN8+den&=(y~j=*-y|l2u$(H`_E@ zS-nMHfk~2G=6MT*1XM-NrW;>X?KCSP_Xi=_P%;ot{ zzX^Qvc_)O`)1J*xZwOzItcJnjZB!Bj#QSly94N6NWE2o^sK%?YZIKE&)9>SXnjDtM zB_a`BPjMY*CIh6?_wZ8G8N}!bdKA-0i6z&c=i*i;&)B1qvKbM|&qVDo)rD&H%nfPVkwp zO;OBmbg7!Q+6I>E!eH{)s3~wFA$08_^Iyz7j$j&fm(&s4Ikcj){tW*WmhlM}-=%va zMIJ2y)yX4)O{MlXr`^^>zW^R4HLp6q%UPg|$>f&24rCc(JRa2Njyh1yF{Q#kA%)H=BgF{=c;6HH zx+5*<6jBrNEPLj>#cd-lLTNcKPNspDIM$5Q+kh~mRV2h@v{q|t*9b4>tNr^HW>BUX zdHNF5wo;!|D~0Qq+EI(g{bh%13jcz-D}icC^I6^|a3}uhCfO5K=iHU>Q%j1-axAXN zpJyCPxVQtGcR|4T=*Nru<*?d$c0jUoR2Jz-&zU~{I(!n8ksd>C_?H%D5i*SZUd?3q z5nF!V>k%6v;C^>nhXJe@QNLY)YUm=)!k0mTPJ00-bJ#W&t{k_ogU8kwlZOi#Bx)m+ zQk+P1E0pojZ<~FgBrDbLBQiYqAXB&vzizl0Vj~+TaPKK|cKmT+wSgPUXqeKoRM&!= z=Fwy#Fqo&B=Ny9apA%?ry0?H}xwx`iIlieJ0lE+UtZduf%1nr_t?_&E{gT=JVNDDA z9gd@y7AnpUTr*bY&ovw`Yhw6{!!}(N{GhDX;@IyF9BCFuc8txJ%FL`Q7@Who`=5`- zYfLP|%bA$p$)hY9Vf^!@of8D=S<&ds5^^NoQbRf{tX7wvGT~q73$MDtG0q?O+>=0f z_C$(4(2ak+@#UoR!iww2|K8#(Tw1z-#BiII+6*P3NnC)T z?=!e(LH#z9a>Au-ZEdId+!)2r_C-8mM*?+nV2&|S4ZM=yZU72?ae5$CFj`xK&>0M^ zwfKh%0^ajZH2uMFb5u)W?gLsA=Wm!7?QYu<8j#C44F~pjwiJ|V)sXJ*8-l`>rz+8ljEL$@ z%p+XfZvl-Xg>F4A1-H0gs%}hhD&kSCIIoYm!9}KNRRp###o~=^XlaQ+L_}xJBL?AL zG$Do|Upm70PQ2Z|f*IbWo%#?$8o7;hS+1yy8f2exUni?G8sL>@4y@|mRT_c!AK20> zY`pT1OKeUPGoj0hKawH1%)N2UzMXnonu&pmrjK?sD}SjWAbpw(BMqyN@b2Z$>}`|4 z-pp7JIW2M9(Ev{RtW75%K!N2bSlW_dQ@)Oil^PYJ&`jN^g!v(Rs%Fc$hdI#pj^UTP zjdzwSj@N5@8i85-5Z)HO&HL)Sq6KkSpL!rwWfwBg7cSqJl?%gK{^RH?ibN9cMAY{A zeoJ;*EbdjZX>oLlN#xFGvlRtf3Z@^~s(RZjUT1 zo3+i^Is4RZIUX8+?^7hjxC)11YLC)MY8W-W zPW`>{G(5aG>_1Sa*qxyT1sD}Z0Ul_>|44v|A&B&$}`VL?g z6Ocxy_%&d<)=9v0dQ6E}V(3Mf-!-+HgD5 z9kN{+{U%&TvZn_0;eE!?@JN5`9C1qgI*m*jT){+JV1$DZqCKr^y|D=@ z7*WB8a}7Y=Kv%KFP)vl6&!FI7ke$+rHYbBJ_|ZzvO+V)Zuj!8MRQ}5+z{ITh8L7~X zC`X{k_f@(1r7=<|d|X=sy{8fi(I9cb)Vt=PT2+TiUYSpeafE~ptVPV7U+&_zix+d5 zO{P1810su%AX>ikT;KPIfvl_Knc(o0#T#c^BD&pshrkS~A?+!;B~J#WC+xZ1xQE}O zG*4RC-*T+Ed)7n8m_@z=737m;5Kq<7T~$iPn{Js~;A>Wu;42M4l#3m1O{osY1%u;2 z289b5N$TMgBD;R7Y&l@fb2;(Rw8mvCsy;>lL9d;_x-Wgp^sYCk?#*Afehp1W!G7VF z=?vNeLOF#*Jw(O0I<2x{;z5h?WS^Id%5c(TvwpBqC*_ ztkhZaNk6Af&5?C)@Ob!25#b0#5sUm{YciSCY2Vk8XWZ7?u{qhnVr3GvISCr0mz!L^ zZw>)O!hFTG-LkOxgTG{RI-ha%Q=>>~553t}2G~TsIMqWQJBBkCSDhgY2K(G+NwAn{ zp`CTyiYx%C<%Me(Lf&pt8FqEnK+UjX4)UQ=ME4?Jy&I;+&0>|d!^x~sw*o2TqwhMN=V!o>b<=tG?81v;wC>##y}4Vx6aY*XzPGMglCQ0o>f?smraga`S+DE z`3KRPwpHmGQ85{zY+#oV>k#ycq_r4|_Jo6{i*dbL@sa({w)q9Uw=8hJm4Pg_A8~>y zWDvJOR&i?u(TGzIX!}oE`RcGPO_B^IJsS3o7sFbGS)VS4*TBm`s#n!0{fPI!mBt1i zHgya=pf!@RB#pGU$eAg9!){&*2B8YyGKq;A{2?&qn;KRy{>1}J%i3Ob+zL@H$3|3H z$l?w$F{vE)owt_DzBxWwsh%VfY5c}PU==nx13{{tPdPAC*H41$E-LUKC9yg$PfEYU zgPn+4(>k(8h+)6tIoI;>HLG_c|G4P3A!og3P0WI%4H<3%&F9pFv=sHT(&?1K66_AD zs!GyS#oZd^{b1bLDeLStaag5CL;6IxK-kMAO|s^4n+Nh5QV7e;b@zSuSB z^N(B|V%I5aw>AlB@;$019!Ld?WH##z7ZQ;e1P_?o?Hr4T?hLZpm^!3E+a}=}ft7$j z3>q?NPfIWl7)_$HCsY1#eMY1euix`#9B7p~9?(v)40SD@4VSLVpiO37iVVqK3uy6K z$t4L+mL&5PtmEO_sPGrsxZWUc-o4I}=cEg-%*ycW0H;{t9VSq@YHO6=wM|pu+af&s zoVN|&8yeQ31k;~j?iR7!qC0ia)+Ft=?qiD-bp8AsECv~7#|z9b!jYq?CH;niR%G*4rKrZ!xZ+Zm_u)XFXe+z!>#^`Mu-;D zVtP*Pf=I*I`2KC0#{3Lwp!-LH*%^eQ50|xIMha<@`IlSibT(0fu^XybBPF*R+}Kt^ zv%J?B7R;g1Bq^=!?XAE&h_GI=p7nw(;hgbq@yOk;0mZgV@TPY$LKc+dR=ES58bOvy zk5SfDF=G!N)@S_4PHl*28Ff=4tlP-BZJ7@f3$=Fk40=q|WskinI&)z<(GAdbqt=#S z;-g?4#<@f1Y7?{dSmaFU8JS!Wlrc@_9^4iNQkg9p9EMD=b%t#}SOnNyLH~df zO3K+omR&N31xde3&9UXz#c|Fr)MTltn#ncxudskFO-MX!nisD77^`Forqu1kXidEM zaEdzigyOjk1ly%-f%FxPWKn8JY%vCgT1JKl!H3zZ5j33llkJZZTlW}y&)Noit_3uh-0sh#}d+J(8UU+3wg)t!4AYHnnRpp#%H z#7m@oRZ5-36A?#5jSB|KK1GC5`rcW_aXCK^=j#e+naf|2Okj&=aT`VPN_gd|7;-A> z*b@Xmm0u{i4h)ByNR`(Kfw0dLZdlo|<*`?oJT7Tz-N@-L|%VxmxQi?J4| z&~`4-#rdgCo$GMtMl+5Pmwb-b3p1+!EaRj*&ejm*GuKVFft)CtC0WQ^h}$O-c(%GV z%W)*I#VFbrB$V6Tq>Z=@-8zi% z#JA98)n21{npybds*v^G(w2n{7d)M(&HJ1tt+eAg=ugcUxft>>v6|K4VDN@IX{l@*b?4K<#OB}TX%NoSJrb04oD--!}pbsZhPPu)8o zvft*QxelcIQQ}TPpP*{R3*1lFqWUz);oz^!B+rg|n=QYUo#7c`o7cs?qNDM$VzEf2 zoun35ZP9Y?wAHhoQ2%QR*iios5BbH*@l$~FMo87Wn3nZ7ejSDk7DfsD8FNZ9S1LRL zGmzx{KH(#8U1c?xbdKC*6dA6VD<|p`Dnbh7X#-4Y#PApB?W4xH9*8EeS6 zG#hWBXY=eXIEd&4I&~TzW!v@1Qk)Kns`1GFUjUjWW!V&=zB802y%Y_O6_am-LpheT zwU$NF9UQMoZtr_I1`^oVze@@OWnXow#+tOcI_VxF()FYA=1mmtM=0b?VsR1@hHtzg zg*fKC0g^hY4=Xcb&KG$*pF6g45!v)BzTK=(b2Gx0P`;J6`rYXOK{W&#g)Z0(h}DYn z!(I4GqwB!6%7257&Q1csaGP1ILA_2TVGIp6gj%RQL`k7ilRLJH)8Z^i`B1Ouo2^2% z0ref=l06NUg!4LJke{4xDh+P*Nru_p0H3gd=B!)5+6WcUS;e^kV(g76gJFou#Cyen z!vR16dwt7@U2BA}d-q@-YXO^k`&f8a43j!c1iNVS!0WZVqKiOvwGUXbVPr3# z)&GGwIC&AftdqB~0Eq zv(5;h?nC-m*w4K8GApiVjs0`?5m9NgD!8`5TO<8WtE7@?$n{6u z*5|QD)jhD-+lcb8bVpMM99;Tk-#xfkym!GANGwF!d;v9~q8jtIXA%!~LIi1V8 zSc55@Pi08Yp(;s+d;}9Z4`3Y;n~UBp^@s!uK?&qk)8!|WqZ?n69C!<@V|-0@hm~28 zSu6ELPv~*aLjv(t4_3B<6GS&)Fsanmf7T3oRTwCNFlX-Av`>$d7V6DU$#K|kVc}yGYo)?YSClB-(Cy81D*xqgEM$` zB?PdI=^yhJb(f^+{e|JYkIh)l-CJ$uGe<3GoKvfd-#(HeMY^K5ez~E*Tq2ixx{sx& zR>Cb^8;QD1Je{pJ3}#PU1X`n>Dr(>+a5W$J8>h z?!c~j+g72-?fKEbm(HRjbVSu8!7(*7_)}p>(uJ0{#T6|enQ4qU9Q!S$9Y8$93|}kh zAv&ewxWLwBek3dFybWMY%CR8!?p@6y?NFbOa}6sr>?gu(!MvkzTL45#sHtO1$1>w) zRxt<%*)pKmuRny;#l{Us5!1o#AqYUR#6x$5wLWP+eH#g|)qzmD&2m5yXtOkA!kBckhxG7dcyIu&&MBF@|2GGAXkPIo(vVjCj6bOFAeAj=jDYsendX>Dc}o7`lI zt}f0ri`1sBpYCZ4T3V;Zo_0!%@KQubyX1I)`Kp0Mp@LPvx*6Y=HO&sfSE8acc^6B= zX^5g@GaK7VA#->I(813)>g};3750nG=o-2hi27;Wk6Z_J9gHM5pn8t-!j)Q&k3u6_d|?b1e^Vg z*uW7(9sr-oAchgAW7~1j@)m<{ z@|#A@Yq!2!u)TRa^RrtM7Y8utU5d#Z|F@la5b!y3XkFQV18dd|PBY6Wjjy+uJM%eI z3XOt*^H*%#)bbPz`yf#asD{scnE~rp)1I zdhLDxbF&oW0062co03J^e(QxDx-E%auahiDi}Hq^vDH_MG*On7>d-K;jVl|HjV!pz z_Op_`R=HZnvkDd6WD-LRfOb5KwDleTU;2q6dWF_V4Wt^K1_SF4>j{z{8@zaBp!Gfh zo`+8?GoEyV(D8s971?UtO2{6xYfvvt>rcxL(63B!^&J?^*h0fMX~oO`JsKzwzVh=& zF_$9RNIcI&!r@Qpn(s}u@d(f+)%ma{F&KDxboW_5EXdCQYVV5?2HE5;8`s}fWIktV z5?roA=`F)8fO2s)c4l@dF|eRQdKIOGqzzJKq|0#-5i=^TD7FWhYF_p1)ZXhAPpxos z7arJeU%fwG$i8JDu%m~BzN#p_RAIl?VwSIstx{=rCf+@mcv%V!83Dy4q%x$X&@<-??dDqw$Vx1YoMT1mbIQ8%OY<53UhiqLL^}JkIW~?5f znb(%h=<8RmSvz|lb*PNF;WgvG{hM)KZB2c4cMv?47wY1r*yi}{FpxaI*%SR^aG)L& zW@?1r>&2zq1i!J=oAti6Qa#qE8SA&Ts`hARHJUzmB=i82JnQs*n^2#7j?nW&aYYi2 zy7IR(m?)aT%kQXB5wL^ddZqFYPFIssAhW zvFn?dVN~a5^xboF(%_KJHED#zB``Aow05|)PhTXO)buom_8m78`p(3(&|#-?+YLvH z$`#Y(7SL=nh-#-FBaUn9d?rbSlXy6Iz^QErJ{Z~fjQBv+4NN0H+>^_B-A7<1!k?pe znKL3CjgK(~#!N^?gvmKesWHeu6EXQ8z7s=clhi%qoWBx;Euy4wbk$2(#A_RbZ7YdW zt6(mL6*f)a>@LNtAxA<$$wgm-%n{|*L6KPEZ;>k?WD#G?EDHpVVzmqQN}7fYJCcza zpNLkwff!FhR+&ps1=dvLkcNA_brxtiDXo_PvmjbSFcy~@@<{6?axKPVuwk4jG>RT^ zI4p!{p8AViZPTBiW}0%Nyd_{3ftuv$*;rG$Os+&$TGxP%`j#5c7+t~RBJMN> z!(R6*Df~gV$7F^A-D1Nm-$dPxLk-JzIN%Et5Nyw+57h#!yAz#TtlT<3TH*)XB)LI5gvoULQ4_)eC3uVoRI&bd|FXaJ@r_qw3)}Y6^Ur?RRb>v+e{_E z(cxY*m9S#DGwW6TJJHZ9ZlGyABfiBv0&g3r~wkf8sYf|=-pZGU&PplH~@BRVub zBL#$%iwQBRL=_Fy!*Ad&!_t zELE7zYK2(1#E0iZ88_1rLLjXNl0ubczx7o&^c}^K%Rr-F&`+WqEP9I+a}K$-oa<@> zVoXkjOI9`>V8mrCpB*wW;L3ggVEQ04F_o z)MKqIGt(lrdlzD$0}}!1`;4*jwU-z0p(X)r*N{9MLg5wZahbG=cORq)*c0;V)a#|M z8UWB3Z(G1cLASW!x`>c70oK^Pq3g$+A2iseYQUj-{_{%Ut?r>!Nr2%yH~zyt^Rhq3 zzwM*uMRE@AeuQKH_pqQpds#87BpC2%?NTUuUo>_@SCRudxg;iShcqx7r7TgE1`9{q zCy}c%jJ;E{FiO~Mw`|+CZQHhO+qP}nwr$(SyKE~vsX8a;;;TyTx_>}d&7S@A7*jp` z^`+4Ykw4lRQj#ysiC2oyp!oSqsQAH6G-XCu%G>;&K0uejU>?JfSTnLo zwJOd0hZdT|tg(u8TK4{<4} zW=*|Vr>af3jjc$m2l@@fWhYYCW^h2YI&i}AX#!FgXn?#runX9f($#6DFfEd8lKy4e zqSzg}S-AWHF$wxB=Dt9B^68ux{Vw14|ABW7+Vq$zZp#vec|xJ0$1>uErKa^h+K6Ep zi8-KZ2X{7l6GdinICpI`&`^`31M5pCvw^DFr(KeYfvJbG1)2R|;IUiOLE=`NNs9m3QMl~u9=KB1#gYQZaW17!u0O)=}@*XbytI-aZ0`Amxh`yJCA zgQDWyTM4NLr^s=0u3vH`>p`~!6i^3Cm2NLeB z2m-RzAxa@9;0Z#lv|XU+p`xZvvb)T4^I*Zc!+Oe3e>ehNB-1xtj#=H)OP!*{x~5n+ z1uqy}4GZQy(q3wQ*@nr zrX07ilvU4IPm0V;i*QYV(F$UmR`I5G%M@lg^eJK`lsi}YMyi$Ry!OZPQ&zCs=37-J zxv<~{n&y(lvxa?1aD9Std{D_`Lyxq^0 zJ9@f(HON`9x)&NXIlgTfsM*v&TVe9ft4ne;=~v<9Ru{QjX-Xxz2Fuf_O_omfvUW5@ zd`|PL{h;EBVJ$_ap#9LQ3!+{DeN30(y*Z7XglmvPx~sT1aY7n~aq?zv zmy`>&Wb)Fi;J7E+Lry)B58XlVJ9d+R<>UxOtdGY@`?A&p2E_d#`kmaELFhNvENVFL zOa-6AE=1=0c$h!AJWQ+ClAVq=ASP2tBu+XM$Oz3lp`C8W^dZRtwA2Uli0<{2)-AFB zQDY*S&2yW2dEqo2E0ru?Z7OG7nnQ84X2jdEIWJ5Jflnsh(*WeH5c6OZy z2TNG7%pH#WiP=z8>L3vZi3Du`yiG*PvSo=2HkUV0G|_YC1S|j_RNvS<>&;cdS)9e^ zG2e#F4crj*wX^kJ!E;Y|e5(%gfhkKw%k(F{bsX35xX?8{#mLnB8SRI4{BCX^9i+oO zitzTB@>Lmx1xVJf8lU{K?dg-(s{zIgB0ke>v+h@T$|rXwQwOj`ac-#m3&CJ30Usve zYAURxbm>fjz!}G}uC>V}(Ed*OMZwsgb3$X0rmg+~;i>(`3S#n}Cn|W-}=?C-m2vZ0YS`H-E_lvy#El zy0wbDM}iK~IBtNBbE+E!xYWjo7AOPnP~CO?IIZI&xdI#S^B6mj1`VnSuuL|yxZD{u zmo}{A-bxVl0@x?Aq7q?mKuV#*+>Zl;yTbIP#2&bXH$uEkF>eBZ5EurUDkm}3#7{`| z>2=Ub{Q;^T4P<#bxrRWG%k>Y*Qr_6hvO>y?31YYxUNLIon+s9ugMD$3A$NB4>F4M3)$rcOrr%Bkb;>68j)v~E8 ze+A5HlXMlk9(0Bx2I}TG`HjcR^uLE??+|zfDVnTYqR?DK?GDaQyVmXdWy|<>iqnBl zvVZml_IC&P>8D|j8Exq}uO?+vRigzSea)B*=2)6fJTyj{_^5ep(2@`;AM$Oq&@>D~ z?;JvW^Q%es3rm&8c z-$+{?ai4=gFl4*`lp^ofKM8M!!`k64#gU2Iu48@LNYac{e5rbbd!vl*j``aVu%D>* z1Ei7(2%oLaO|ILbaRd+ly5 zAH93$4NE|+!;A8wD?5aohVfbSP1#TrX}o=6j1F#Q^n_t)*k`K<(ssA6x&^vxk79Kxr_78FP_UO#+b|ldmI@lvX+$DHcP2HU*l#fE2c>n_d0X6#Z-H~A1NrGTk~wv#Tr7xCQ@u-Riwkf zB1TT0BVaJH!2RN)#=tNvE7qxf)veu>ya{0I`gyp6HP7aEo5}YT5)Cb!qwf$bGmLd0 z08jS7X)2oZC8CyJ9j@$<&LR^lvPgA|;Q9F2DowI0HBV1map{5f?0)2MqixX#k@02fh26W{f}$Xkx01|x#A=-2fv%qZSd$7kY9 z^W?Xmq%Fk8UUrV>j0oD0ZFy8q7F8cP+?e{epBTXVuPn4P@ne21G_2W&q`!r_eRqvg z#?PA-__bw9G-6d4b~A;)?3OZ1ev`!dYp3vtPrT=G(ke&elv80_^Mr{mCMvF+{{(xo zQpQK>Ym{8{mCh-{jB1Y1xs=owy}=q{8ur7uA~96Cc|*jgG?C&b9<;TK!H z`$VTzfP!q4QnD*GN={Dw;eL%*P26`2+k56g++*IyR12IF1?(J-m|2a!0DtX|YaLOh z$6!Iz-*dWpOX%7m4HvdxwdPw5bX=GMDjyXOoLa2vXOom=3l75VbB^uT zo|74OrsY6?NptYVS33ob)?OSi6dpTNoMzyS~s$Wb#!3r zFPPe-hIXgcM;^UOd#GA@DN`Q^3VQNAHAkxV`PYX;r(lAdvy~j{`<<^KZ}4}m7%x2X zsX!Z>Lf0v?CB&-{lwWoh`H~12D8-oT}5~(O2`U8Y37^yPcA9pC2?3LlP)liGh>& zY*T>{9tBy21C`NcZa5LB#-5btqBF_Q6xi!?=Okpe7@R$L z^4p?npsJzy?3yBsaFFirbk3!^C?-ZdRPxWAxx#)O4+^a(T{a%Vd^f*pIZDi1x`sn1 zh5HA>;%n+#_5%EGmQ9~$lUN-Pb#LYps}tc*U6BLfgtOH_H!^&gruFoSvx^$?8n6p# z&{(1J$|sv7d!)db8ek7cYN%pxEThOsnq!R7M4vCbD?e!Ux(X<){reFeio zDYeS+nC@|pX6~$O3#v7|7R2*vts$jRzGt^n#VD&B=2|*S6bms^_~^QYLDGHusp~sf zzR6i@AX{)hZ@^B0EYy0JhbrX$gac{SEXkj;nq(mS;61N~fPSJfqyi`QU320Dp>S_a z9kJE(pbcK^GzpJ9Q!B@e+(1wUP(VQV1UWe8ebQE?Z1*8nWJobIYV-u4XDgd$7141Y z8={`sRyD?b%VrzR3vPqX-am8*-$4LLcIH+d5QEX$RK;j$_VmiCG|Y5Va;rKDY&B#` z(TYqfk;zi}p={E2{wkQ**+h_)e^F99bGp*1D^dLVj&dz~8E&ML3*33(t#`$ra!D&` zyQn@QLk5Ksn5uYIuqp49Fg#EGz$^RI0S4K;c^8h7vT$DK`@);n$|HS4Y6{FTSYy+6tz=fOYss$4o0D_TdnbdTwK?X z%`x1*2%RX~;qYe*iQ{R=oZmMr$v;DpYGqKEBeVfDgIsLS5Jf3lBXA$9)*rehAZ8cU zcCuI=cTj0U-m`(zt$Fc|T)@BS(>8%zN#J>=fGr*PBex?Wi}m~2612J;fHXNf6+?7t zZJV&O{@(^Il!*`9janVq3K)zsF;X681B@I%+X9qVkhGC&tO$2AXI3ThtYa1rM%-YE z%A7gaKCq+7WcH9>Q1>XoR5uG`BEv+>c=NDsXc?A}Rl?u-O?H zl$qC^92BvW~nj*DPok?;4WKw276J4GY(m@5<;lqelj%%pYW_vrzP-vMh zFRb^1*&#t1y&+Xjsz?_C(pZff1>c5D^3e$}@V7Pi^I4qNpIku5S3YEZ%FJ=v!Qm+c z->k8yzDhP;^2vaId?TltU__@EZZjI$dX)tg97Iv$EPs%K$<|!%%sNgeSyW@nloa{= zAdl8URs(>{Ap!g8$YXW}xNMd6~UbG{B}U*__5#~ z+?9NGL=-MZ?G7_<&0~`t8cjJ$W)`KZvmWz=g-)fMVDuR1mF;Q)V&Ua}1&&AL_}V3~ z-ER>>u1y1S<_xP0;#Y@6>1ri z7#f0|Lkp^aisBS_PRr@42uOfK9;}~Cl`|p%C`ynr^h`k^0U$mCNhOHL8>{Q=`uQK) zr3ZK1w*H0-Nj}0%(nZ>s56A>)PUV3BZ@v(?RdEB)c-p*iplpw5BkZAPN(p37p}+rI z0N~U8!T!1s@8aF?sJ|zhMMe+H^3-Y`b85e!&|0y%kG$NR1u2;s0!>_%C&0dKTzLlD zONxHcj?U7be`2{^Wss&v>>|TgIIxMbL2ZFeKRB2`#l?fkIyde*(E;>Uix~N@FJC>g zi*-iyb$e`JJ6Jn4<~wfghE_~7Z44*VDt!IHxHtDppDt!U2I=*xK+EKb)F+g3>Q%<+ z9}qT&_SzvDs@`J|W0@t6mCITV$=yiu1Evbo=cDEe&=S+v2Q$JcwY#3?RmUuAj(t2V zf!N&4w=)%ecy#hVNM*Z(@EOSbne)Z3!tL`6C5o|@ybPHBtk(Fu2enj|U>*-dFSHSXTVr*w`NZyqCHmE~e~z7+`gFl!RdQXT6nB7d z84e1KG7mE^{4YjV;T%h<$S(bYu45}>chI6Km_;zcV4!I>Z+(?u&!5iml~S%*N?Sdu zWr`999A#R}d^T{^$^IgT#{f(PyPyQ?P_^pryJ9tYTACvNAyqooROVx_p>0q8(eEhUgI;vanJAW**<4}YJh0Dt$t9ERG9XL$(^=+wYl7MoB#VV zQE-9~(ha3fi<@AIJKow@o2(jC^KXA`B}zrP51J0He6`@qa);zKltU&szN7c_1B8OyK6m8_RtB(WP_zp!V2B|l$DL6pEj`-u$?jw0#2SuS~YmZ zZPU>Xjdjk-alXIM_R~O^ojf$`82*?l)E_b)4d9mNPP(H4ojfEFEBpv;jOYnvf`r$y zdioYdm{v>}dzuAsJ|`&~6T|ptg&#|P^yv9={nk>Fi@8F&QqRjo0|-EA(){J6L`n}Z zdf-NZeLlH`HGL0qWBjb>YdG_fE%jRgYjGEWC5Z?z<=%h-m-wFX??4iPkJe5*oCFX3 zz;@cT1CW=Kxi*l1?ce)2kRV_HWxYFX<2UVr1Yf(1ABe0ajp+g;_Kxp_wKt-l>c-rzD z3f!+$tr=@WrM9QPbh4^8<9Dk{N%YY5U)0s3a!*!-^KZIeez2W|s z1l+vae3KTTo;XN%AqoHav2KK388I<^wto4!PO{w)f(`j+4cCM>;d}`11zbL*yxAdm zaZa%moHF8++4W{xvWWuue(TJXylG5W9))x{1S8T269mcKomT)M^j8NRoMR?_+Li3P zOGf7XQb1Q4BSQ#On0fap2Wy}Ki`n!;C_0QA(xh!y|KWbbOAkme*e3ef=V6FIjXK&O zTzCWlpLnUunyWot)64KvY_V?sAm{DF0LUtP1G{6iPr{W`9o7JCPqGI0{(F4eJWCBL zPm?>HldIWN?)l1`V;_*ve;F%w@Ng72h-J9^C6T0sZ9k4@K${DL6kOUX3K z9>XLeL2K=PL3C7O5QkyRtar{w=I&%uQiA%tqESxECwh`y!oll(g!2l%Ouo`7K9QI| z8gGqmL4ir$p}IBhHtlNiy2c}vS2-vnIyU<08eJDFa<+2JwarAmgnm%&xx3qP6qrB} zwUP%(k%)8PhZN=3q>jdfP&8;O3_&Rb^tCRq&y?mWi0ia7Q>kG((YAgk3Ozq_J)gj? zZIOOYpojUU`)yq{YIS}4)Pa}R&0g(hYE^roMhEgy5K8Akjk$&)bjUxWcqi|;lAV5* zFRjvlXV=87vW~~8lSk{Gw$L4vw8)A6@K~DQ7+G_)DSPnMZ^J%hlUTIdem9E(tA(L? zZnCPpp_(mhZ+~<}7aNjm3pgC|V2Fa9-D_mdbD-+!&Q2J{@u-7>A>TLb_eL=FG!Bk` z5DpWRjYwd!kbs0^JJuC-`#G)oZ{Ft|L68iuA|u0$uA*FCd`mz@Lm)s9sq$j)hI?RX-*($Oj77% zFg}=43|9;=abS5kc=$9-0l_$d8pYTE03i|+HbEE|_+>Hyh+-sv81yh90Ohe-zzHlt z2u#@oqX6Qb;Ymnr35FlDI1w9(k^@jzQ3}FxqTn6{c6jan=xI~R#;pNB#88ShQ_R{i z0FerTGyzJ0CNnIC;~)`4A`DBW9IOP3;s6DKK7C|T&|}B_fvK|^05uI87A(mto>ZhR z0^S|GP_PjSG9YkwSOfI2`Kv&HO_xd$gtBakgow}^i~z`)V3g?z z+=dPU^Bn@Vw2D&6eh=WD1;!Wx0cv?^34~HC3o$_@YM_AmSlPk>t0XZ6t$-cWbwMU3 zQGt*P6ZV>NfDK~gKxzebf-E$VSB|#`iIkB0iJL(LWO!}mB?K62XufUHpoNoNj4T?^ z3x-^wo>3?g2s4IpiD1%37+xjz0uVwlBC%+Lw*+5jHID1tn!gegRSts{+KIiltdv0&^!f!s~#lm)^V zvPTu7SXiIDFxi~AHk3&Kct9mU88xdAT7V{hjpj%eB4Zg1uqA+%jJPmzUD&_T$)ZM1 z1cnsEtR_BJpT46ApFk*eCW%iAAC9xj#E>kIUtY72li;gf3XeP>7S%cn7gdtNX#lE< zGYDfzFv*a&zsw#z?@&zFE)5xYPyLfNcc8LySIj@&QgsU;lf@?^NvZe?uMY<<82NB2 z%sC2HUX`~EVoaD>afHO8;0jqYWgc!Pt`b`p{=;in@OL76M|%9ES|w?LEU-OZO9aGa zXwFNpXl#gLVxge&vDlxcoS6qCA;~HSB2bUd2tYAK4X|-w%aP@AaFfUdfC_^yO%VYK z0tv8EfZ#+%Kvqe|4Y6nuJpl>q7q6$<837jLiTMwlEOJhy`y8eSnyY2x>Me2Sly?qa# z3}Fk$ZsME<)^evku=H9J84&>^H$%9GH_Dw*|J7dGFUOZWX?m4pvc2J^(#_!-Qo)yO za_FOT`-Hum|7>xKFNP~;NtN5Z^)?(`kN+eP^#3SE{ybdO`_4z*+ve9e<|w$mtfl)~ zY}ZA^!xXv%FW7Cv)L^?eHD1Wibk`G3JvS~2%-X9ZNu#eXi~HDes@gJty*yktvua+- z=K5SZ-alNiySfiPH@Lq9+G5F`E~*p@y^p%Bv~Grdq)VeK?4uspVXwIESmm{fVl6$- zXn#we5|d4Tcwbt$O_im|v?@|(`RnsC&pjwS9hM^dYTL2LdRtxmPgxl32d^wlgUoEo zE?af#|c=@1jM_^7h*4^$C z<$iy)0vRfv=UEl5T|Zb{SB7FY|Kc7^VS94Fuh!(xP51}SNVLpWTf1zEGV@o~<-E~K z;+(jWl+rP7ZI1tV?gLjt#oiE?Qcc&C*8@=L(Y2&>x0A{48u7RvvYfx53^?TzV_6!mzu0hIU`+9lxzaidM4=N(Qre7$aQB_g7e4$!#RKk}M&-%B+GjRgi;DGlCi)wkBsWg6zliKskSamj z9_MQP^AZRpubglF%G#RrBloY<%&dQxRin0^MYhp0Sz7#X&K&sMG|GHkBDvB|_%yXH z4B8DT)~h12nbNh2;yc?+KP$(%({Efldd_oA)`OKNXZ4EQ@rbYOX*I9@LxB$uIu#!> z`Ea(Y*c{*A+S}SE^}*ppb>m2$W6+oG+1_+?TnAjyoNvrXO$Vyi%B{GzspcZ$@3ps$ zIBUMf8(!jkf$;}fYcQ;L&C0%44X-3tanjJeI_-4~oSr7b=@uTu`pq0X`);eJ+@^Ka zd!sF6sFU;)o{Pg%x1PIj?%Gjt=ruei?OtSwxGyfG?VMK(%8S$BjO56LqcM&ARDq zxjzHu+ZyGY_fM@M^4->=W=V6dE`|ENa_XuB@%Uw@eIJ!XpI-cfeCh@U*NBYEjgW%9 zVP0adx}U9~_-MWvQuLgro*Px6;lw_7XEnZ$?xiok{wmyL*y{CvXyQxKp3fSyr{&1K zulD=+m2wC`=X{&6s%=7;=&FxkX4jq}HqT?>?Q^RUO6hJTqCsfCy8NSaljkfSyI7p% zDfh32fKH7~)OK~U^ag$yXd8T`Zxx#y{b7n`K_9o@9pgZ+{M<0BQqE$EYCQf&Z{$su zhv~S`O~(5B-9n?QtbBMmr#KC*lRG`%=d0Vx(q;2ALCik(7EjT#XD@kY@x&>v`wPEx zU;FYur7-sYDTQ$`{_n8GM8M3<$@;&d*Z-NqSXi0Z{(q-1RZk9;o4>Tu5)u_Rxy>;q zj4!pOJL#Ag(lE|plbl5w456Z?G+Loo(nu4@2_#yo&7vZa!jg0d(#Vh8#cq}Lx1T5X zS=X}}^Zxv=&U|x{w{1>>mqro$g%AkE1PV}6+kvSqEa3;{asvnt{1l`myysFR1`>eM zf|`;MLP|>sAVw4(M5ZAH0b77lxd&jJ0RbqC1K1=04$){~BnD8z1d#E@0Cw00FiH#- zS_Bb5h?z1&aT4W(Jq#usFhYGEkf{OEun_?)$h#|X!)FCSBFtQjRWPUk(6RA+44Kz) zUjYx0=5r})R9|jN{fe}E91yBbUhBkKqnHyW&8ypjqgE6?30Rf^5 zdYUje3vp%70f~JuTbKa3$AHRW;H-yOFhY#}dLDz|u;Um`2cqJgEsWZDw^n8`N~Se{ zf&hX9kum16-!NGyF|coFSh~@~2*7M$fDrspq>9=K3>=)A18Gualn4_2P=O&bz83sE z7!YL1Ux072KoEyy462aOi<0+1ek=wz@P?tU00c}EluQJF5MTqUL!u4&4{!__kkG~C zdtzLI-~)Q&Z-$duEoyFVY65`Za0-HuK%WB7pB!^4b&eRmjO#mPc`$ET|%-v|Uu`LYCm7Z5<4JOHMJWK8d{a5mqy zPqs6%F*u77)EZt0OWBakKoknXlvx`weqUm0TLE2ZRDnTw;qozrwFC?VfMPOdhJVaM z!#s)+Bw0uP5IgupV8VufsH2C3G;xj48de2G#3B50ptG^RRe`!Vf1f^rPz4|X>`{kdl#uYyBw#=sFo7^($BqQwM-aruK>`6nML~y#{NR4A912(fXm`Sa zqVV`QakIB)2yqY=KMunZzi>aK;2}*wK6xew!qKEcAXovQjyO?lB2YANBY*RmU$$wG zsR9jyAOR77ve$Q;82~(iul$ex)6QU+9f)iqN&sQ^+A#8LgW&-h+&rV>ci@x!-~3t^ z!hIkJ0!1pQDJmVLFg)NAATK$9i{s+PPM|5OD}gbPAcc<=f>4WU%D_Sj4GxV56aKrJML{7X z0;K~4sz`AJ5GQe9Y}XV5QDiaeWT%xe6`%4+B2=ZrM1`wUMmF+ggCZTlcg$`Zg*;g@DTSM7$;9 zrju-G=ts}=COAyE?_F$pX>?{u%yQp<*JoeFviMvyImNq$z|TL$0aJ|%nHU^7!Bt8Nyjl`!!*iW;h2e^&(AwOXZ6}U zzGSX6;Ke#Sn%B*xU;*DTwp#?2D& zejxK9z{B0fx1e-g=3nvrp%^@g%@K|>m3Mg)-9@y%rmRwe-Nhf~>DB$rN=k5K3K%F@ zjEOqG6u2P~Cmcdp3gytmA|5?)goo(7wtQ0&hUq%2JhwxMBE>`NVbZbrx}jvl{9}#e z3~NW}be>&EDBu8O1RIl?!Ci8@JAlG|;bB`$H@tGar~cB7)eehALtbrZyds62WPEos zstb4_BkzP4(?a3rmFp6hDP8&k?Ctoyc#=n+We(Bt>Pd(@Qe_Fh@=9_N$yacG=437Q!hJ;0tK%1h-8%w_0(VAq}c zB4xV%{rrRV6`fXpQUCs`V|y9f=3?velNDuzrkZQP<&ns;8iY5Q_;DUc+YH!af)SPjP*8k=g+Q#}}F`ZoQ za}s~zp|dT;WTd}V((u~Z(!=n!rS8)_nqiDQ^N&FHtixD$6w_^=Zyw$TacLpZMO&=T zS6iLmrnRBgxa2Hzsr0nz0IFL0yF-3)m5{f&j`ZEe^!8`;8x%TrCsN5jLP3Rpb@e!l zsdcJpU!aO}nBB;#y_t`~R@wVtM@>a1tnF-2{re?if z9^KRKXM+7V%0?Mm49S(;U84&h=jpa%Z0$TW{qtv+jF z)X` z{o0&5gaf|wWiO*!ZA|Uvqa^}&N!z*IrOKUsnDSG4`audn%-uTd@_BV#eN{bjzxa6% zB&F`eVUp6l25zeT&R+wo+Rv+8rg&-js50-`^bp+~=~f$e(eh3Dd&A7z>MTV<&5FhR zCU*Vh(}wc_Q}#zeH{%X<32HnxK-Kl%qMo2uS4u;(upr9Xa=N$tFsN3u^PrwDa<3=4 z8Qa~`gl-#Ns}CuzMu_i>81(FVQ+@q$%<1C3|2<}NT>Z!xkKx}~yjovwa;IP7X~1;C6dSyF-Mmp{YI@Ld=h;5a zJzRBCmi#Veh)Lbv;I`>4-iAd_@zZS`GX0hJb&@;1WnQAk$Ia{JYS2yniV_RsqS6jKp)JVN{v`}i^wc&!I{IL(eoIPs(j*)p?6^A9z!X zTDH(KHj+5q{UnDiAJy}C<|CC5(nvgK{Kep~5RgbC-EoXQt!a^(oA}GBPK7|wdGW*S!#7%< z;cX8>g z=y&_6+P)BvJl9odxAs;?LR)(k_sGeBzU(UFn<+K)Gv^<|cp-ekGy3?`mUg(R9eV&>QYc-ANtAi1( zuAfWhE7N7d?YU|}35OD~dRuvgTDfPYw75nq;Fv(~T#=ma1kheaMH$^|zqp96(B?MF zK$2HsTb7ShiK2+&^tX`z?WPW9w^tO_o zwP>W=ogEhFHUGxh*iydzW2;Le=jB*i%cDB#=ZZDxOdy0pTtR;TV#Ol4Y*Wu3c zGqj7vYIBaB;IPy(IzW{qbA6_3zB*=XXt3DjixUsy&6m;1>S~Q2i*cOl!mvMkcL@Kz z<%#-Ii3p#^Dd`&u@LvY z8l8binoN@1mI0Cv$>@1=d~{={$*4xbcsk=t*^#tfva>W5^rIMQyEhPjITX%FdR%{- zVQy~nI+gT8!%W6kr!S$j!PLfTPVi%_hH=oe(+`kPsq5Sdsp#&iLH8QP%1^Z1=g<2V zE6yw3c9Z&&b~;IS-`Jhp>X0e~YLu%%b3L9xtVtH5yTYEm=9^?Y!_=SQctN=$wH@W= zn$#q0`t5q?yI1G!zmpa%L9M&An0dNwRw<1|;54m2aL$73`2VRT;{2alA~p_=|J4gI z5il__|L@)Z*%C1`ax(w_Zi!Spc~q?N`%^~>PSBhw^RA`MVrZvn-7!)!uBWBxn{y6L za+q^DF+@s4Tv_&bfvLFt3^veY1@*7HkIr{7mj#`Z+hGJzJ5Ra=FI0Z&)fa- z@1A~mwW@V^VJ>u1U^xpQ>?DdYGy%=g(b4dl0L>o^=dgs`DL8WGhX7)YY)nL!mYRkM z(BO(8fo=Lus_(Tdgc1QP;Z!gZR@c7ibs{A<_$w z;RDa0BH+O}hH)Seg^ozTUFI;w6eLLy-M;w;Ac!0sWq)-z5QG-gZG|+3Hg(L^)y)yc zf)H_pG$6tuMl?M-5I|wCrd;9xQT2` za4>w<@A&ZA#z8?G64K#NM+D-@fR9{+9yJFeZe~O0at;pe4KFWbF613q043cJr09Ui zk40NAX3nV%s0`&S$jkt4l)FK!!^J`qD?m#cd1f>SBF7mms_?JBQvit-G5QTp1}E0Z zdxz;b1gb&@vWE?4WB8RIlquo29YlbO0cb%LAd-Vn-SJRkH)58xA%lpFMJxasFn$PC zpqUAdfTt9jhzbCR3l(l;^mp~?C0bkqpqm+loJ2Mj&Z_MLJS-e=nv)^;IQkTZ=^8^Z z%6g7SJS~FYhjJjqi<}a16vFEI%a8mjOBd$IpEn5=?DC(!Y9Fj&$iRQuKim)b5HMQl zl0ktse&~M~6av=(anv=;L1q$v$nmX1!i78&9!M~gVq{`u&wKKSB#)Js$m&Qlco-FX zkL`Owv^otTN>G51kuldr7@TH5h#W=`23Il17gy)E5AIiAvZ>)666|C^f{B@(0fZjb zP$VH1!W%O)BM1|-$b*hFdjYOM0U3=%A%GY+2817#IASit%mk#OX`7|B6bp!&q3aJj zp~_p+{J{K)e2BNUak~jLwxBoxc8-+(rY8ikgF*YiM%*eybUXTZw0}~EZTU_4(QmS~ zRaA^2gvG$hM$9g%EevEe6Km8bbvmm{)?7rA@zabp#}r6_I7;B!R)~9pgkvjVNtUcaCsmC4H;WCoGQh5c-i97v7&C=^$LRKq=>amN}*4(4Q z|I+-hn^|fPUk#qJ!o%0c&~yNJ^UV}fPjH!_UZ;9tQJEV)FGDN6JKiO8x6+TFl3wC5 zEvdAi@b?p)lQrntYq6uduKqAAuNq9`U~xZ}hrGW$nQl_a4tAR45-*9@wXZ~1Rpq99Gaa#pGJARw} zTNX7o#m{WO0X^sMgHKgdcFDYDbAp3&@daWo%eZnsS}a`&ci1Jr7}K0}T`*vVawOV61WGge7gjuZ6? zFIwD=sm-xX_!fwr#`K_RbyDZlEpvqy-H2>Xc5_63g}STvP0jROFl3*=oSI;3^Xiq)dGd3L?MVt6`BjoNN(yZy@Cu4fLimZtk_@;LS|Hx^5n!vSgI zgm0fq^-?poo9No5OMD7l)9l8pCc<~;NshbE`|-iR{c3Xlm~M%LpRuyUOz~w^`*?^ z#haynzK+pa^b>q>J_irH$+uoKJS94XTbp{8f3a~d5ff$`H3GlJ_Hv=xRxfDj@}9)a zB&ru(WTW=%HYa`M-@Qkz*odQpv!zvFq2z5&6#O$2o};F-MqSJ?y3&)SBDPDL_*%5U zf0gk1dUjDRoAfFt@0aor1SV5uGnFwzH$E4RIoeOt1T z`|HEt+qtvk-?{u^jzmgSCwZ6S#KPI~-EB69iKohZdg#n?ycx>|$IMzGU49`Lsl2&0 z2PvvLHyV1Zxm5Bfm92of519+w`Rhg~3G1qRFsi6x zU{bmi!{o+r^^yv@HmRycMymEj{%C{;Pd+8tn&*`&v->5TQwLi7_^-_g(!R5#KwZ#! zY0{C6_w}rE(l`+q=V#1YxWL3~@?=7LbGNiXWg2y?h23qhTkI)Od(Ux8a}?MbxQ;ug z;MOF5#|A$pZle^V<^6X&Q&+)QFhXuN+4ugRn`_DMaFW;spI+!7o-r&gZf(j~WU6K8 zW>^iVH}un7Nvk6|**lH2Nj6){w^$(!1UawZm2k^N^&&%{`l9PJ|!=EJuprdBauy_{2LWEXH89v zQy#o%6rB(zXKVF)__WMbAWCUTa@Uu0`dU(Vl?ND_sOW4yZn;-lR(dJR%H+6zatV{{ z2i0g>caCmRWc2cQ2B!9S$E!JA8^#|>{)_a;P@aZM!>ggUNQJ-Rv^#AVU->NDGij+G zjmG|d)ydD1JzM&_-YKqpJ>JL?KerhN0w#}PmnJe?)S_}srr~K6mzQba)ON5KWN0$- zGIn|+Y;j+E!P%5HR$1@seW+36;K=G4lioWstZ^G^6WzHz?N&GP;w0UhfNJR(Px5eY zoxtPEwDV)eVx<$I(@s)M?wnRPYp#4Z_m$)YH|GyMNSM{tD}lArK&MxGbX}?Z_VA>x z(uy;f$~6B}Z9V^!ZY9YY*F+mag+oVC?Y_h7dE8vS7Ot10H2?+Sza2*RPwQ0(5vS+5 z-^7@!?@62S=;MqJtmQ67ILu%_7{yO|o_9*?o4F5qJ8Q46F=I|xPJM+ugjYR{HN;CL z+Zo4RTA{~KrLjr2rr71MRkvP$`He%g9(Zd!she4R(jQ7>1wjBkYTE~ zkp}c|F*O*!&8OK1QJt{;JfZj5SorYy$tu}R7Da4zY!%%deO2+%Rf7eIbv8PSGIl$A zz=IpJ{>1M{?fFNf{n*7zDU{G%n7D&(f0V}oBr)|hx1if(&_emo1(g!&wuAa z!3nm56V?Kp?`Hb_$UW7&v|(bE+_F+{H8tbUt8Kaj{}q~;{i-DYQnbtb#iE}8qbBivfOhfZ zn&}#d>y!V%oht8fPhR5wn}+564|cD?UhXT+3tZx^UfS-R#OzQ{#EnZ&*7VV`NN z{)Km;S`2pl5d)CS-eEbuK5lShv{Q%9fonG%wh$o$`MXb`C+J zg~1gb`;Beewrv~ljcwbuZQHhO+qONKWRX;=Qk8#qdwCc4{-;l$ubp99S6#R|Hhanw z)_pq8=SVl8m2(0I?Y9RKzDxLGznErI>b|Ur8?+SUZBL4<% zlUofPB)69BbCR<%H1ohX!98Nwku_l^OI0>MPhxDGLQ)xz{FdC9OwxVI%h+PwnM$u6 zelxkunk_^R%aBnYA`?`YooM6pdIN@vX<6T=>s;#tA-s)0-N$MgO@Xa(!4@@yIodrM zdfAkxjtU#<^M3<6FNoiolFnACy3bm+bi7z6>Zb1#_?O!9=hrWf6LE)ra@}S)xfP*8PdjWU-yoJ_#026SP&i@a=!^rS|2p&c@Hn#r}JWP!2 zjQ{=kzZ4HMGd<&fNuK|II7NMW(IgohQzO^Dtr1UL4^m| zvl#-BpYxCWsRp+D#HpsPBB?GaA)T6j_kaa(_T`!#S~_-F``H%-{s;Cd^as)WZ2$CD z4iF%V1^tq^cl>;BuTTHz7|Phru}&a* z^R)(~M-4RB&;S;s4af(4jQ!hU7LbsH7(yTXCjyd1@ik*7+zm|%26j_ zVIlDiDEk8e1Cr)HNf7;4sWyYE_Ye>35X)5HgYMT?F3gCSxWOF%;X1s(Vf>NO=|^B> z67s;~Q;PY~Wl)OEsCdy^|2%1&!V#!ut^T{>V*~<3ZFT9_v)1vg%Y-${$7*DU;)Y=+ zLt{)Le&CmyghO!dtKOTv>(l9Tf|R0&2yk0>2eA6q24ERWngpgH&dJHnuilaWnJhOB zmXEYLHx%*o!`h59ItP6C@Pl642&kd{y}aE&6{dn^cyI}pnDUP25(;?~GX(_!RtxAm z4Uk*X#&U7v|MNIu!}MXpBn-iaMFdqw1NTWL z`0oT%eKj?JpbB=4mi)&s1zdv@==H}anEq`i@5cqaUUb`@seC0@<@C5S@-7U37ERp=K5^`GoO6y zjGtRU<$Iy`k3SUtX9zw2+ol^p6-UPml80mXXM^?UcPsrQ4vi(hDgdlCu5VdHL>U?s zRv{XDYa1Z%W$bV>fcy6XHGnE=ED<^)kgYxN2Ou?$j=VP=IUJj>%4f@&UMP4klvRI$3!I<*mfYKaOWE5! z2z-MJk8u93>0elN4g%U!_%cvyIX}<7sgplZtpwl4(a*80Z+g&ZD>e}U138(Oma!BX zpP&TqKtG1Qr;lKo<$U`Zlr=CBt?h}{;zmeMN+wkb~&(=@k zzmve&+Rude50}=0j0rtR0s>VZfl=ZX(g7Wl^$@Im7Tv_R>>XF7-zw*PF@DpDieb-Sb$fn7S$8 zP`>9y7LslkKdHr3TnXEu!fd597I1`Z3h&v-ee;I|a)2Am!f}!gK_Wku0{DyC7xP-5 zXDX^FmXuOZ&-F8kzsJ9YYYK-(_&SZ|V5e1eVJa{4b9dBUknjxF$+`#WO%&0i z*{Xea7dO7#p*f3%9~siEskC;Vo0zw<9?t8ZdGiGdEyyStkRive{y-Gm@oa^A*Jf z)O_xuAE z7p9cNKb-Di6R%0gQo3@E)9n&4Uk_d|&XOxLxW$M5 zlW6w|c$|K)a7x7O_JpzI1zy5(@M7z#ZBW2g!v=!F zh=|J4%>={LA%Qz$A7yFS3u$7dTH{$9Sm#^lSEq7^nThRr4ZvFkuq^VgJ+Mcr2ApN6VfJ0JSamUgaCj* zt!!4St6X^=UwWHm*1Y_L`c66c5Qe&yU2)bwbSSR6lg!EYSb^kjl@q=nKHI5UJF8Tv2yJ`~pem?1o5_$Ff zY(e-LES4A?Jnh^w+8wEy^c--HtagPipId&snHuVxj(xR+G`CltSaKdM-*W@0bppwK1@pA(Ygh&_=K~z3QGi75QNiIv{P1qahuHmE@1YyF|Wui)syIm);^)YpDFAa2x^^=!nhU zKq}v6UXIRLnla)Dbn3XziKqYhg+UX$w!@j=`9i1q2QzbE-uSXcmh+f3Z4Tf1?%#hM zXT+vii_(OPuL|6RiF=QV@V)Dt>`Ti{^HNtlW-Q$a$Z$*(l%M9IR+KrAL}EjR4Z={J zHYn=-cNgv7&!BP0&c#{IrNnjj?G=BxLZPZg=5ve)sy?;EroFif9u(Q%F7n{Wv!v}z zp}!!-k69f=3E#y93Hv^9Y=}tIAgiA4)kUf`Bu{y}d~!DidhO{M2AD>*Y=k2C&6F!! zj!(=#DdX~0#WPdrvWa3d$buh0e@gf?_fCX8?J6H7-H)WPlc*#ko|Q@)GvP$)^NN`+ zIZ|lok}|obA_hp3L2`jEf_J{+Pjnzw6 zj01^xF)kt(@$M74ebBBT9v>zjA0=vi{=jyTdG0BElieG55sqhXPs@vTn&|OlAUl5+ zO#P9D6lH)}v2MP8D|WUC>e(Ba+o95#;c`=#Ff10$Y1qz4_+oehhT9QVS7E6?Jo9j8 zRdp=DTsv=#PX~A*AW_CzIM>j)-erFkM}hK!sp)iM={JTHrk84{=yuoO-#h_qDvwoK zF%Z*`#zq-~CYQRnodcz6Y8j<-WVgLH4r@3SSahQNd6P`TiJB25nE{*Obd!%z?3Yz7IYvQvVu@qbvlx=0ec#9g?g`%Hl`ZT1 zQcO9lM?+jsmYpWQ*U18M2k;*48{v1>UCChq|5uG#65(X`8t*t%bvpKPXx`g|?yoxq z{6c{INTtdKCg^PZDBqyat?h``vetZJn!q&43KEIaV3+|gs94TMb;%>(u!KSOl_fqu zP37|aYDBEVI19BoInF^Z2e%$zF6XcMHhN#9WI*8 zaHcNu7Om)m3y#({Q7i}baY^ChrVf&h%lSi$Fcd-ZjhdIfo-Mbb;=Erh0;)6ek;mXJ zn^>Bd#G;Dm;!5Pm=#&D0R9J%PZ@=prP%IYSCNexns%IluXMG_Hfz4+!eDS(ylsg}I zqMc9~aFe@3^+yK&tqdd1Mc$Y%M?R*pG`o`)9eS$U+0q-GqD4AIK+W-`ikV{V(7{n_N3sQ5a?cdmS8WwqdipazgR#*?@X zf%w$S21jqqV459x_@d2AdqpyulamosO=>b8ITnLEh*(BoIIzKF<#s0vXVqUpA31; zHmIg_O`E|GMZkc?OLP+S1z^QZ+HxM;8^_eoZ#j3}oK2#@95{C|gl*_6P+Gn} zm&v3TkF9z(=i{s8OQY#h!7Xw$nQuI|-|M=5SfKZx&n>D^pTMzhY?e9>s_0D*qp4)y zyC0d86n3}(vw`r%(N!o6%mll3t+|ViEQYR2yW4l94+n`#V$b{u=EBO{a!I_8Nkc*Q zlKFYKlrfyrpO0|0-^t%WSOdTF+L z>QwrzlQ2s+>zWtW@=}`dCkt&&WEuyUy-Q54d|olh6qUgrzmadavl+{t+>>_L^qb77 zP1u2lr~@A2KJ~$gb30^2 z$rE#u1uu&{Qu=?Ip3FyGBP4D-C&AD6lU&5+$t*lIx+H1-<#(_n>5&1vJ%+{3x$2i; zx!9TN;lg}yjBF}J;{$JxY&F6wn0;#krL1OSSwh8+3gID;=uKA@N*}cg+tg^U4O0OLgc^G+{Y}GaD#T+buA+82e4y@mU!Z@?;c>2cY2VE%mrQIT zrwY7hb#txQj-?(8E(fXpY}h-4X6b!0O`jz3IASFGrJ!-DW&C;)Oe_K@VnE(Etx7eg znu>XJH~yM7Z|-dv*TrW*=1s0shoDViZz!C408V??i6@_Qh&7&G1!!$(^9y}R?D?D0%=S-t= zdp$t)%p@^}4#tUVxxbZj-1@$l4Dvve!()QEQ*~!iScqz;>}HdO3C!)+p{T?-%fbx_ zLB>+vnBrz)k(rOL6GzN+53d(LZ)I&xtvk5DhkMUD@#bT-5{$kT(_LS%5)f;<>n z$9N>|Q=(EFjdu{hGBX(3YSmbzSfS5BW}jCsyvXl4q; zod>W2YP!E?bzSZM;Qm3<5GqME9jT3JZJOVMne~|&<@DbsSeSdWdzBJtxYggwDhqZ) zj+b0JF_s0=Rf~cga%}9qfoZXw;_)?J>0T*WgK^&PC@vAnaRZ5LAxi+HS@u15bcA#Eho$* zcSDhj7PbnZY4!+9-t{_3}K6ajG^e;xRo}^v)E{<(YqxU$rE|Mm{ z9)FO|6R&){?^Lf^eCTYiL{x7el3&dJiZJ5hs@4xV%RBeHhxDwnSg?#67VpfGZr}1I zZB?twKp_|S17QDUN+fQ!g-6ZTp3?E9{0K*M3;Td~bpue9n4j^ zFwR*vVC5vt)F08qcGL`2RuRoI$_!97B3~`@8E~X`OE|--h|Jq%5<&;rXYiYPe^V`I z&{GSjEhMGo6q(Y4d%8>fFoylQ?QR5T_O~Hx%bF$QkiD(@7f{$b{q`8eR-qfm(5!kn zj32ms3E0k}J)grx=L$N~66-FsVWE7e4vPjv$_y$9Ce83fHQtWAYLj2!1)!^Y+3Bi1 z-@6s)V)n%j(S(2TY=KJOH&OWKkc zyMGl}=%=c<%~7CS{|ybK2)v442ak%a9^ZUkrFNw--CGXBDpGsTwbPT&2CUj4Y$j?u zKcIvyl(Jx<9Z^rt>+MX5(nWs&NPNkm?sOG<2`;i!*DcMOBa6sY%xHK8(qe^L`4MsB zL!NQOSUoyhe#r&X2xM|teoN*U1NZrDAu)XVb3O|HmKbaq! z@?iS?iRx}h_s2)s?zD_^sO1>L?A(rFM@YXnC!F4o$;kPS_B5=y8Gx0kyayfQB*+8y z0Z}cQPfq zlc#1Bojl*2q>(dA+W!u64?~4XNuPX6MI<#&f!64r(FvRq(hwdTkIDOuV z|2!^_wsX}dwAfPhv`w_Fj;wRHXfgghUrD8Q`;!$Om1E#4^8?CbVQf%a0UE)FS(;f%Z=191TMRgt%nIYsVm* zU1@;qn}5@;#;WL6isLAHKXZfnd&OOem9Zcup@b$Y$O5l4VcULd$L`^e1GGsnX9|G% z*Nfl-;DpaD5RzmShOgB5J*0(w`??e)@4F~m1=W#+pN7a(7Oa#O! z?XV5&-}Ki&!wGS0JV()|N0cZF)B;=d5={#l6vBL&IEbu4L-kvj4AqObxPx^Kf1Wp3 zM!6=Hpil)I5cGzX&t%sC;s;ZS2UsTEcj-6y5JXj#RNv6OflWjAI08{&2b zaPGN1z~B+Q_M!??!%_8l34~tSecnEud+}i|Ax%f6dn|;#uUMBGu}kh4h)oz{+Hk;b z2SQx{iyo`yXG4u2gB_QgIXz*H@)3tmnzOp(%%l`$2c4EI#$?e4!fii0%NGsjO4K#k zW_z+(`0^Dua`y@Dh$SlnS8+yp&f9rdb4nnRY{!7+3iwn=D6R&+27;KL9Y|7Rl0Sd>Pvu#O9m-Lxf9wp!SK0gvU^tdsgi?Dj z&LNV>?kMzr@>|V!YhE9_rkn-eXql`n7BEqXW(mZwN4024EvPl^yT>Wq@MEB~yPX@w5rpfRqRMjOpSSz67mGW3h9P7h>tIl zQD>}X>_ke&kAtH#=4Cz}4`wZdv{>|wYlo!6=p_G|m|w4Q(hYS1om2*n`LTMzS17&P zg$;AYA*ir6_`|(Ag_i9)_(cAIsOp~tnngWzJ_!>TH12LrATFwaJpt|6X`8O8K~G>d zmIZbtRJY>*U)~~z;E_EH#L#~O13HapH9$U}$gH0c5MxZuS3aQ5x@(;=JHcHIF2t0E z72w^6vl{;`kay}HsS!!*tl{xuYp|Ul(ho#*zX7vdO;k7_lF^EP%E34>gts;{dNDjO zf*MV1d)4*Cv-i~DBw*KlLFf%VUd`7i=P5_&76Gc^0*TO;WB}em*JGXkv4i z2S;T}U?o^^NefTOs)=M2O%>{Om0ap~YFQ@ZEh&ZJnn;?f(cZJ*#kr-?C^So1VdEZh zVyoY;M7ABB^s1xsb@W*7Zi9ur%Rl`RGGS)6+gePuk^S!HDsbaP{z9X9EPvW-x*eM( zRVVE@s8wtu6<=_@Ei zH4Sc1YccZB1&z$+H<)J>8i;8+V2$&RarTu!x4fkYU`9gQ{VeEpoiV%JsC}s_GRZvM zTv;E@lF~>A3=SK!8J9RYn0V4}jFX2yCTj1FG-|6F*-xB zV|?L6J{Hq5eGI}u2;#vWNmMSgB?U}LdhijzlA~xNywRYF!f5~-xK3r7P_SG^8Af5z_Gqkou-(i(Ba`*>d9 zGgkSBeez$utb1#AFuX%=s$I1n<4A4pKf;x^&MHA13}Qaw3R@@S7X2)x-jA8&Wgg&tLVcBgDP_674a#4NVg zOO&%j;XNf6a;s;AR=Y`Z{^R)xqEczxufCZ#&&D0@g8PX^h2!d6Q(51xR?!nzc`Xj9 z5fMgfz~D}xP(caBDe3mpMHS3fUIdV(%Yly;vs6SzT%9?0X?vFT?hj}&HM0SX1F>+& z72R3_7xW=W9!y%jIJ^klvKcM9tIE`zxRVBFsLcvd)jjO75H{7i+H(t0(_@^{TZ#*C za{K*p5su#K+s|K^+NBUwNELtPJ>t{pd>NZWci)yD=Ld)Dd%rQ;qx0CtRwEb)BqIV3 zc*X(hW;^&V;4fJBIIc`Py^d@Sz9NZTvAOF;lN_%EqOOdm_JAZ`G61KCEeJ_OxOM4_ z0l&*ADce(l$H>G0Q{8ZpKZ|lpTF)l6CyFSr9kQMEUDO7k`km~YH)C~|Zmk>N)mz0+ zBR015Bp0A&f4rYA$4!YxiHtV-*vr5Xlo%c?T4{&eh0Ka+!4eo(^8k{xT;a>^bQ|`r z$&&IE&p+FdDOR%3W!=0f-_231V@VBnHdmH~Fk<>|&C-^mwf!4C2yBhi_^|Kp)YTKC1uFl z$+~#+{{#vAE%GLc0zManjOAl;58a8!84@hvGy7Ioo3Z#py)bM0cLSLF*A+I%f}}c+ z(sZljy_bd*#qUD|-(HR^T`P!e#v{9;L9_&$IpZE-@FuGLUWch+Y1K2ok!VT>ERT*6 ziNj0bWuhU~o~1?haubzd#~{m)1KLw&W}C#QU2>8Wkx%<9{)sJuS7S1U@@Syd{=9Va z<)JEXo<5`~okmqOF^uwYn+4TE5$W9H9dU6P!O>XBIX{k!|P z+xtug>T$*ovjGo%r8wM5<-+@9`9s3HUpb)_Kipqg=W>Q`NER1WEN5f_6>UcpibKIA z0+poTx6|Q1P^BcG6oXq|qLTy=B_a>3gK|K7U30(q<-IxuScboEzP$@IA9K2nI|8?BK>2pU(}l zq#}39eeb}fzBc0PC39Yg{BF}ypOSZ{YPmz0r85{fqq*vF$6+|F7ReWYOFHyxw zD{fk{gK@-66%JEk73`z?H}I%bEtwb!bCzj-!;1}2o{HP@!xcfxXb;s=a6>2o4{?B1 zi9eB&)8l)Y#CuBpM{2zL+U`+ezMyc04wwQ}(?A(V#yOLNb(XeB>WU&_a)|lD!v}W{ z>%k9b89fHl3T6D97Y1~a6*mh11bUi6tSuxb5cD{8;Bf?6S0&q#PX*Ei+T2Oj!Hp*j zl3{?0X!JwY%=?IDeC%)>Z*c4i&E}+u7Dn>s((wW|zIla}|D8%dLxC^wH<=wT^naxq z+TTt(LzRh?HLE_AP~)J?O1V(os+M zA}ExP>r`f4QJ#sWv3lRcEH1f}@y;@0_oF=c9PJgWy6<7+4SBJ7viydz=Tb92j|sM3 zP81Q`hmnkUg4G$klR4&FSWOX~PRQ+g#vNDg((nc~jjZ%4O1qiO>`m{o;}wPnJ5^hZ z(R*f7(bZP2SRqdYg}aQ_8AYy7PQ|U{wcgIPMg!Pql~E7fkLlXw0dRYB8>wd=YYn}X zaX09qT2iydRo8qj=Bdgr6NOSrnm+_W0YN%0-=|L#`Er;~rmzUAi=%4`5rF;0Dpg%f z0`XHiDglTKHi>A{+~-<{&s@$2voyxc3#3k_bAp5g*}b-NRAjoYFQ#6&nZ^-Av{vr? zpP}J(goH8&Os}bKUD;|DIKN9jaiE7j+*+jt)`&tGSp90fmWSV=7*ohap+cO0`wTq z1li?Ap1R8i&hrIrwnCDnowGMNGpffo(x`S!G~5XyZvZ2wG%UAKR5c{-(llas|5Up@PPQ?0v4 zq1oG><2V8FJ7j4&DZ^dSl5gRPa{mnFJOPe*YxE96x_E08jfqT18FN=2A04(Rs**T> zll%}}&KL;in7Sb(tOy{9ukl;_0(&+02CuzFNw5olMpQqDago&s3RCO}|7T5B)Ya&o zd56|LHRjE+@3DWwFZ;t?cAXcJ=lBj3?Yc{5?tU5-HW_BRIkazZwvj zK?bC#IF#1yFCtI!j7_U!1V?$Y68hq;Z@_Q<{Vr3E<$PKZ&}=L~wugk08lj@vJ9UYK zLg;COYK)}EC+%tDE~DMngAi%;C!WP=Uduxc&??ae>Z_#VF{MOfP4p{fjW-?Z&rX%1 zuQ4?1jdRTgP|<#RNybD~k*nyo{gPv|hT6W9f78wK*IXWJL8pd|Gdb`9iim%8yh{=E zGw|0l^VEvCo04>?LlJ2n&EsSXC~x98WM>9SY++WOC^x#XZA&$7qsKn3kq>Ll z;f+_J3;}Q#+y3g{_S&*be7}`j$WA@S;;!q$fhul$=nS>Vxkwj2uh+vH3#4vT!8tb0 zIu={cZrHRbQe;S3{5HtqjrwXZme;XPC&nE8OKAEKWx<933wIb#BiyV`1zx7vrX$&W z{bwNNtAcrXknM_&s)AryoO(WlI=V-8U-9uJX&WrlC3ZSz5F>ZHs&{SLKQ!avH<_s& zr_0+et3!~nEK$KP*WCgHZ-#R@mkhjfPo-E7U z&F)i#yJF^DES{z4C;UO`s^`VM*D-jh3--~hO-_4wI3E5Y=EEj2W@Di4IWk!f;;TwE zuGX3?Y^w`0v!&%I6Im)xe@%l`7wnP2*t8k`xVRRmC+p@=0l#Y-4@xr!tt-dFk9L^P z!8!7uX@`!doV^(_80T>qJh)i)@ajYrPsUNv6P4xM-Sl9Q3E4{Z`;UA|3+JP$=uqY* zg|{%(5|yLq10zm)1e`@myo-{6?dk_7jX0N1g+h)?CIV1$K;#t)y$oi1BM6R^f2pML zZ&2CI8oLd>1{Q%55DXp(a zGH0658EST83t}N+h5D|IYoR78>LGlM;m=SAbv+H)7oBR*R%LZ)kyIR!r1OYFsfU1U z%89|XGnnrBqCDzFZ z?7KXOJ$Ej%TGp#O&z-L#8@asVbb}i96hm!p6JT{4*X6ze`}4EuxoqQlXo%d}7A`w> zH{TGv={w!a8m_*{&TXhZ%{6rShvnvCV69iLCgwPNLPf)+aB|zww1zx{@P5#CkycW^;dlGy zxk)*Y66QtcBl{%n=0 z1|;o4dE5lBwLKL;b3Kzs9m(%RMzb94li!TVe`1rsS^=K>S3_R z1{5irJ$RT-Hl!7#5~k*U}f1 zBNxqdyp?Ej$2}KZqN~Sk$Wc6T;H(&}44b3L&6a&um}?1zA~kLD>zvUjrvLhiVdO;AMsE6f9zRoA6R_C{e6I1lPx3KJdCKLlCLqsiMVufkB z4Y0Y}p$Y4={usuto^r3JTP4&;9{xohcR05%_r5JA5Qf3;o!R$Yw=jHZ?aVRLu1?0f zo=VEQhKkgZJRZvbMqQ3wSk;S+`!@M5@hxT`2fa3*Xg^RK3ZS+j!S!eYzrdAr;+K6# zv;Rwn`qDdSBgQ_V=M)=hGhpR|c^jEG9IXMb&?PNJ+M!WKaCP^yc@$0)EzY%+E2x7{ zZdMQNo^zO&naJImUtkPfY>W?X7QhK>cR_C)#YinGuLo#TS7RZXY*9tk%!^9WKaH9B zyj7X33R{3|W(-CLnwHO7am_mT;6WRh&8yFUtE8Jikv!*~7D>Fa zI@zjQeVKL!!8lOk>PqUL)#J2r8WD~k%h+ewN5|wM;lXNkCIM3F7RN=s2sI|7nuHc6 zV~_61=%fNz4>GFV_StaFIj}lYJL#&wE-yrheCCL=kP;J9Wqj5|5+TQe%M$FtxgwK3 zrMf7;5cWpU@@s-2@c&XS$67j65AJ7~y5O1mx9aQsaI+g91Y^79Kmu>xS1TQdI_8-E z%w*OT`9_*A{;e&A!~a*Smrgs0LQb^|qCx~`Z-ljAi?b6dCUa4P%`zoDP~tEyfhoGQ zy+P5I+9EmYL^i1`_&WOSwDNgCp?ikZ^fq6pDvTYVNWG3Nam{PDpouy~Wrtxy)aGx7 zr-y{Dex`Yx&R*kXi4|(bWC*W5wPZ9frs-5{Tn*u|qrfT(YYrJg!*QkUY#FbqDu^u^Dkc9s!RE~g=hWkLncHXN}!9AG0JLKCV? z0}3`Z!z{eA9c_@v~*0lbWMMgYDz|} zWEK^5$;^j%gy+JB0xuM&Me1~ES)s{2w}LT$gN}=uRq8V+HMShx4*^$pb?5W7d2%2a zjAUM;G_GRlH}SN(;*joK0& z;vif{39c}*f7czkCxOfI+}!mt+|%<1{%jk^SzW-D`u-yi#jOMM2&zKKH+QI^QcEfyqOvUL5v@^qZ_X$$( zonNi$NR0ocl+-0vgk?wQQmF=9+&q7VvUECedit%+L9ikRsHcRZrv(=@7ewOgiFt4{ z>RD`|bVfZ!D<(xULFD8ggDm_NlIr7>tB+ND8uhFC@S2OWaWxxrm+_cb~Mv(5mW<_4vlN*?ssFc`{mkHD+RLZNnWo6*0io=KM5cI zO#=jtt5&YoZzZXs9Js5ukop2mi1PO;1h~hrX|`~n9QVg(3imqk!`nkX{^demqr-ET zJawQvvi2gaSsrF4K_QTtmiJXAeXZwe5r47u&w79By&}^3h=?B9-8$fn-%HwSR|Pq) z+Sj;g244^xFOmpYHf)q+YKSrRQ^M%>CUsx+n)O{~uD~BzN1^lyB$*LjUmH^UJbsDk zW2!>zx3@Cz7|k#UHnG5T`h%!a@VK(XQZmfD1p60qeZ_m7>!84dLumz9j7eoR7}vZ# z4^q7lkV;HqsaX&4Tw^iYN9Rv^c+r~m8V9t&YvOMFg*h~M34M348)BtyQBN_q3RacL zZk{HrlL%;jM>Ouzw8O3o?BGZlBrPif0@);?9}07Gg3~NNk=B}?L`vss($w{xHz7;` zK9`+^QED>VOb1s`DA>%7%?#EP$Q8>_!~DiM&VqP?rqccm&eQQ9f-gGkV6D(1)K>>bRtK}0K`nRs=PmD)jPoIh*bYZo5Rog5M=x5P& z0AM3WW#-EstY%ixPASBbxECvcb|LDnz$D9y*(@3;X~MYM)37y@g@B;(Xcy9F+AEx@ zh5}#7ZP63yAs&h~*AB}u&P69s!+{E^@rc3s3l0e>Z}pWDo6on~QKV$T>KE0#>Su91 zlfvN!22crf(#Lqe&mn>05(6uouDg}qcEAp5L~G)m*Os>SKt8dj;kl?y9HxFo3TB8s z$UV0;SGidKjs@uNUGnx+%9k57V)ATJON%BbK8jPa%5>$^b;m%il_w($DvGA<^eU-h z4PDYkF3kw7$15dC45XayifJn)!LoDL4HBT zr6PB!ggWl9iWad?UIO;vm<(Qm9@Z}V-aj-LF5Q}sj-aYzAleg~529XiLuvN6s;;7P zG-?EGC*(7lxm{18*GmFIRiQ3Q)&Q@E?ITH&1z0^R> zB#=3R|BI?rR@RUx^KH1!K1;faI;uQKflqI&U~n7)trq}AlrE_;aFl{Laksbd=rHdKuwbxokOH1<}UgKnk?_C&LEk)e#HnF0q?H(fDoQ6uu z=nuL>$iWU2%&gxDLeO(P3F09XkZ;1uK3kl~43QFBNNEwBGL@yaqAv{B>$2G_1ofgS zCeZ+>D-+NE%^R&R>-|k}tiIDRIPaX`W|9e0D%*YGMRz4A^2Q`rmE~|6H{_yBGGl*$ zp8LX#*vGvRZ_FwSpr}|$@?rXef1vMKB;iO9e~-o=kVa|H#w!x=K?a zZA~%N#mKZndvP{3$|vrcq+L3MT=fh!aftG(S#iDu6IHmkdLqQtKfbf1HeJv3{{9Ya zBphh~fI7T?rMSaE;gxx@#HD1fny*pTZ*)OAy2SAr^Le%`b<6=4(V4ldGpkV9;&hhb zWUcV*cSpUna$o)Ob>C&7;f=~-H^dYs86tOFvi=Aq!Ph)iihYr&ERG8^?W`;N11kC!(Qzlz-fLBL3}1ymEZ*y)L~1 zO%~tObZ)6bD&y~d3h;xq4?2A?NNS3;C1}3n@y+?QClEsa?q4IH_9{>Y;B{YOv`%}G zLjLKWQf_ce;WySFl0U44QM5&ye4(0v6p0SFh_e1FUMY^AmX(GC#-cr`$WPXf=MuvLp^C|LFR8qC{@OF7 zt1B0q_g9jZ*k-UEx*i~TY6U|jH{7D~Ry%xMP(#qFaocl*m5A89-SO;`wT%s{o;`-} zM#@IT_w`CWHInm>W7N^(FF~~UuNUWk?yz}NUfYO|0sKMCM%oC38sCNzU1K(SF$KOg zoDwHh#PqP9dq2y0T?A)cu51x11+lL%A0>q%157x;nux^YUO&fzc;Quambub!+jna4 z^Rh&JaDrFc&suJF*)mINHC(d}72cd2=)x`H(}8Bx=?;^08J(jWa(%>|4jhEx6Ea#< zw}K0a0a|4)rEXM_wGlY9VIH8pQPd3}(Aa9wBEKe^ibn9?zj3d=)k0(x&!f78icFWK zd;qfJ3aW@Hwud(+o|ZFh8&&Il*P3NIkNv-38EM0uJQ(-!$5nKW`|0#t%q?GuM{(Xm zb;|1%^v9hObgJ{3HMyIhlj_3oYl<~$Fg25!^~1y0CBqy|9qIAclt9htwutVR4mwo| za`W!|90;oy1Iqzcc@rmI+YfwKOi}c-I%178{IL{ZgyaN%vSrf}~l~zt;@`q_81Tzg@ zE4-Fs#@2nSfB6cq>}mVyYMXGtj6ibjSmGin3qGJCpWDb4&%ZN2tS_yNu1m9wYO>e; z!r5*zfV6Dv*d(ZA6fAqXa+Fz%QNm#$SH%RDLK&9Y|0r=>gUjf4WRPNjCERn@qmk8u z~U7d*buZRa!B68CRW9^6$Tm8d1dB_w!$v^mgT0S4XHz^O#4i3!d|M zofz8rF@CK72pN~ZrLt0=*-kIAd0EuSmc@^i$Adie&f6f$2RcU|Z@IxGcHP48haVx@ zaYYi-AAj?NnIi^|6mgzu5*X<*#KG7PtOimshQ1D@#wyF&SqkTj zeObBhygxW z>nn@<=Mu@I$zXz#kaYdZDg@n8i!?Qy{WP3{16WM;!;@t9TU`VjU2fqdy~;l~zl1-b zvk@PhY8A{V4V%Q()VX?&2*<##-=^%UFnI;MQCj+D>#ORmpwc1@^%(c`sl3W*H3kax zi#LKOws@jBF}D^7%hj(Fe?D?A@&i#@0n5f$$S|aqra`#< zGaEfxRot6*nMGRqz@bO@2Piv~?lCMaoI}3bLD^20=6 z*{)Q0gIwMM%bOw~^xQ+Ly(h#_w6-8*kayLUbhYdQ-)I&}R{%+k=>AI;4VhaoC`#>< z4Q036wBk|lTcYwWDPIDt4bLZSP|wjRFtT4K&+$XJ>1~*g0S};)Zd&6o}>e*^pvG! zqBJx2)m~ofD&Q%kWo0YvzRA}iscZ^y4S32K#E$BH(R9qv)P@=Tc6YG_(%i}&bXyc5 ztI%;!KTRk`uV8SLLSym)UhMjkeo{7)peG^QWQdCIneo=;M@IcmC5Hk`fQe|hj-s5K z;dQ{x-Zm&+mZx7ic6jx-waXGJk}$aQCC`zXyglW zJgcGC*he!n_4d5M`!9Wv1$B-h@dWNlAgix#VixR!Qu16ICp!~H)9bHQXA^)K`Ye;E zF6-&N;WM_Oj3n~72kDlAiiLXm;QdbzezH|~drwJ48xtoB*)X>DSO();iX~V!%v)Uc zw-a=3(k~Psb-SZP8{M!Y8FxXBk zi9p2`D@sk1c<$#1>+UJyyHaXMfv-U&r zs;~XRExXBzj;N+jF6yb{nVv^T%SqrW@lI_~T-=b?UEekSXVBrwv>Hk*2Bv4sD5Uwd zQf#nQ=7=&>{?>E+R8+e@ZC@j)e@tk&D#b<>y&Zb~LI@twaRaW~_8a*{VwRCAvflaH zpFif7JG(!Zd(-*L6XnV?jO!msB|a0}V%9=4hfXxqrV>#|*%AEFbuTQAGR$ZQ3{+#) zr(9xu_RpgFm$_rU9+}C%$;x`rymV*Ys33^u8kWeB2kGPm{pklw=HXNchfRT0k*E!j zx0YdCJI?{VPQpN|Gg$CD0`xyp3l!mhWJ_e+P|kt`JuvN{LPd!Mzz&hH>r$JfBC3uA$xm(`1$RnO1$B-0apwK*Hqhq~J8?tzwfx6vK(~Ck{K; zp^vSlX8ipA0dbez&=&hxwwqS$-s` zrV%>XQZ^7I?`ZeVhk_?b$@39S_MFOhj{+)4Kp}y^F%^(L`T{*~54p5y?v6RO2&Yv| zaV$8TuK=$&C?tOV?WBw@&07@5rM8@39`BF=y}*58{@#wm?+brt}6_T1ZPn6k#c z^z%3_4Bu%K57VK2@_)P{JF6EOVNN*IeL`_L5cg&KiM2M=vM^=qxr<)~ExAt&`ECj< zvUQ@x0rr2WF9MrSefXXFdZ}m?$uAR}81{@!1QzA*6jsR!<(|&hzB}6WPUmi0YsSV` z7Gn&J;I;k~dw++QC?nb9hLvoTvLOQFby2zSfej#gz{=IlR)%11f6EXs>6WecxF3Hg z6Ff4a+DkEpHTBtMVW(T+H@}P+^b24`&ZEAp&NsVOb5g1PTL6Sl`nUNG;ujd8GAe$- z-V>nwBEjPvf(_RRyOVWEm{_PR9rpXR33jgO!O>6jmLLTBO*m=tLoFU1W6t9Fl63~- z29-I%I(!$FR7u&I6)vHnk;q6DEhMi%_N~;G$)$P9+gM@$0Bk>#WAo z$CiMHiZWww$|U@NAR*769HZgPNxC?py)(79W1w#ZT#uJXWHEgrknMu?EK|gRk(jkf zsK+p13bo!pJl4*gb8ZJGix`e<2j?;DNp&m<(-v!vRKh55w6sUgM~_i-l2#lNzTy+X zIWFUQx*V&VgY@BzZYQOAlKGDS=-GlZOL7iLyQYl5BC<7Q4uNhDEqSN))HRiU zB56{OkQsE%G;DhpZp$t}`QcBA=|4bZ5ePH#c}VQrbBOq5zU?J>ARx+A7 z4&yc_)C@2o63iiKzG)(lQv|R`H9v6@MIUN@Ptm<_r)PJ~GdH%IYKxf{^5^u`6?9D2|E4bgez_R_O#wB56?`fv}g9M^qq1b3del9+EY;LR4PPe7*|vmSZ{T9N@?-rRn+6W0A4crUHE-FkV+52YAoU z`wOC)khVmW5Yds*qN`xPmW5VUvu&`#R@yPf}W~pju4%U#2MeI8n@Zkwe z+dVdTnCuC^?^H&ttOP2%GZcAN<`ljA94GT9?rH~aM5+R18kt>?Rf^P^Dz3}*WG-pP zQh3U1$qQifbZ(axY;sVTq|aiEse7SC%C#<&<=49A9uLp4Du;XsCR9lnR zyUh_3vulT-t6IE<*XcdVWVB~c2TK_e<0NIr9a?N9Y|zWQ@QmM{Re4%6K&ntzOs
!2rO)eYOKGXL`-U4ut9*ndHYp z5>H1jXEqDkz;ofg>($q3#I1?IXBxN1F=v=u2W-~_38Kh|O9za4cEFoxD?}$Zyd)sH zxyt#+hEF^*(u7}X9hWbU4eo_?;Ypq~a0>hRd}J$h@-Vap^dRI|R`c>(yP!o>ZQEk} z^w@kNwzk_?R@Sw~)AE{aitY=t&U#Su&LZ_!($swlj(6g?TC4Uff9awLhO09)Vsjxu zGU#WmL&9-%Gj9aRqylQPmie#wRnhLAy6sP5Q371b7Yw9Qd)d+3ZZQXm4p?p0>Iliw z37a1srJ@EgSrBh#SGRX7ot#4yS($8$_$SKMi6;_^&F_JUH z=JJ#?^A|UV^zx1@Foa}TH%%w4sleCK=u~H~ zdcWztEamgT3v{(7C@4D}#3`g_f;vCZqp05*0-o#`{jatQf;##?Ol9Ac)1#>8ZHCl% zAtytdSATE_$_`nadM|B@$6&$A9G-e(6kW^u3-aBt^5v?hnklu(-}o>rOHwHnYuj>D z-8e2#cnlAb^C&rPdDWZ7T0ZA3d)3R{ZzgCTmOB2^;kmH$KBJfvbxKfmeCiGoIgHx$ z2q(zzoJNqQk;d+5+g@Dt{psI)a1|sW0Rq#N%4y+t^n%C+s4mipg0^eM3H|6)cwdG^V~8y z#&>g8g%lGBVek0lz7e6@o9^r)%gLPXtD`a^&+<*IIh};=E7QTuTJCm&et{aG%2IT* zixFM2RB+`H>h9Z{c;bA^Ys8hgKBroA9(KTviQ$`Kn=f0JgihKlF~2N49b!XgUM|;# zjz*aQLm<40y>NhYTW+-|RI!nhR`>oZsxtu}C+oagL=4U5Y>7oh7g|c1xSrxHSlvYB zG&EkFYV9|2UiZ(`Tu0b9b=MSs2PjJmfRduk-gQ*_21X!3h2AHy8rWUrJXudX*lQ?l zCl8w!hLxb|#q*MqkjUx54eSOrU^d-*d}{Lp6g`CKb@0fFB0>W}OwTOO=j+#z{__jU zSOsZVv9fN1<4hL#bCH_qu3<>xkr&SwO5GmDkRb zD2GBjLWKq1cNS>fH3>!l zUDYq4zGCZ>_i=>k7{H{IjBIW7S59S?h`=~%Y0u0hFv!88z13>_dqB)a?vsFTJTE1<$vu4fFZr#FS!Zr`& zJVT+`Z34l^;qBPbF`&p!MV)ZrZKJtxqqn$RT`HvVQ9aFfB2&=2OnAa}k|(JfLb{iH zkODmN-*WyMOi1{Mb#H4`*hw`2Z1EGJ_LI@oCDUCvF>=7SuBA$MkRI=;G0|pdFDu=w zh3t+k(!nC3lRiz{-31C7h#uDQhh|RI6PT6Fl|#YRS>+9ACz3>puV&|Z7%Ng)dE*=kdJ*Um zAw-(lwWAyhlfj_r@3MlR#)d3B)O?0|Ehoz7=YrrkAV*-2*2etXlOTMq_Ye40G5@YA zwu15`iD^5D<}TY(gU9;jb_}!wd||LAlmS5$=oOOxh|SpHA^{B`+tI_0s)cEPo_cqk zs8TU^_Wl47!U|PQ&+saL7t8gNHJv_12zgPOsS3W|^^w+T_sVh@@0)kKxYLxKMVg8; zoLGY*S>6#C%oO;JyPh4#Twu5p?kYuv7rUw}7x~5f{J)4M7)L-5A4Xy^K>9v#g>S*hq=+}>MOKVL|gvBLhrLjK-P}JtH^NiPYt{(VK*{a z(*Dq83wEiH*$reiYy3B;g%Fn;yOVI=3wxLDeit0}Xh_JSpA^{O!y+g%Ma5sOPgr zeir*23?W~3NQv$S?k8bhAO*2K9tJ*-<@9>ICFbyJKcByX1{aO(K0kb7Q!B+i&d6CuD=G`CwPh0lN2FYdcEwUi2xZf|LfVtw*ULi!X$2((I>ah@ibcb5o zcO^i(S;KNw)jU#&ZszgH^t-cxftz*wP6V!vS;REp>{>l7m$J_re8;M|eR_ug@)v*e zjJU(GDGT~_Xb}BuQJe z@&aZ5Ti|NqV*3&Y8~WHgmozsANQpzA&xwj*cd^rIHc(=P%JAfXYQniRVMSg~7sZfk z`3RvxE*zjUVJ3on>auaB!$cwfHZCi5t@Ww~Q+BVhSJKY}^kON9tm?TYICB6C{f!2r zp-h((O~9frunhp2{6RHYYKdt1w^Qr`j=1NSr1MP7LzeF7RzT!~j2enPx&i7WuVc`!4l7+9pI5g>QwB1??J9 zGKIs^24&h!9QjW_~xm$78mj@ zQ6H0x)Hy|4Sr}7pbAe?8GUIB(#SJ3s(86X6f|vl_SZcuzsSKE;#CSwen&G@S7%h~$ zzzMXXbgi47L>UY!i;VbY(#6_$TO?=GBV}KF#K{KId$42W!t@Q@;SmhSJn0f^dbc(t z!^=%s7CvS~My@z=Pxz|^*a0C?QZt>hD@lSo3iRBDADs+5qf)9YE)g}A9Oo@ug2~el z5>(?7Y1rK}jIPm)L>GP{u}Ftn2;pt`V|OMuHH1}0Z93F-7rT=scg-~8%i}|J67yDh zaK*T(d+yUTKw4`v{K(Z(s^1=3_>s1or+jsB$l#Pl2l^ww3|vd(^tXT{{h2V> zs0bz>2OYP9x5oj3%Ycl?UPVNu=wqzBwW1gzx`iKzhovvj}-rN=b z>Yk+boyBa2dVb3TARst=xvp#Hi@$JpRP0DURX;=u7tj`Yq_h>p3$cf0S6^Gs7u+5_ z!`l$XDrwb$xuIv>JisZ~7r0_oq?$r$?sL=Ks8|`Wq8xg&(R0f0dH0_p!%v4NatkTG ztG~^7;J0Yd8izevpJS}bMmzp_K~Vn<(I>v0s9n8v#i!CUrChhqOdz%*tt7y!{jUp6 z;8LXBq;!6MX-LI1ID}?M3jR*-@Q86xe(K94^Ww`t65760hIVmFmKgckMbQn-JgveoxjZU z`jMn(ATGv*KoZclkFGc9kkIvqdDq`57{IJj1i1Me7W$noAlpc2nm?pYWro6F=M6gC z<|#Jth*DzcM}}G6u;JPtKISl-6bxI|*5ZrXDPG^@TJx5cE_+4HAd+^-oba9#ohx0z zn~wNE1`iDy`3FoGdDZn_)<&%obnIthGSSc0;oBbk0Yo=2OTB;rq_oyl7CPRLvn{rbT-fO0@g0BbMttS8?5MW-KnHJXz;PHviIQoZ zJbCUzfH2Rr+P`T*Q3D7T1YIIim~VQXH|`&HX(PD!W{oK!H+Dhrt_i7=@JKKD4GIGV zZTSQWaR}HuLy(3Z2<5EbeCU6LiO+F^Ca}o)4K7l((y;k}renI_zV{vp%`n!O?<<=? z>i!6Fr;dMV+f!TgiKRsBEyU6 zjnC=MWP)PQh39&|YoQJjw+lb^>86+2`?y%*#s!B3IpvIHSu zi#)k|N)NLc`eOgNt|M*w$6#pRj^F*vQp&U;j z0grL7b>xJ@W0N7RKpvEmNV|K=fvRr+gsj)~U)u{eA^$}2JIrfHv%Q;^=oHSf zI%2~V?;}F&sVvDXB#wgloC95ahVhnoK5Y2BX?;n+rK1X|TF4R{hxVNcAWydvC>q_g zddm#mez3P61sCZj7~+`{nwJiLfkH)w$*^C|@8sHrtW?*3|931|g*D~aRY}D*wJ^@( z{d`?8nUai?^6I`r0X%ZnLE`wbP=ZqH+FemOaRrRYASVg)_xPDwP|Di2Eow>QCAN}w zc7#KG8fINnA~bFc!`*U|baj}3FnqfHnQKucwYjcM2p%{J!JnnBD7V7BJm5Q3h+)wS zOC$>R579K>y8vbS=c`_cz)6%!NnZfls2Mp%t|VUj4}ye0z>G0tZbIcQ`R%ZA*Bh{T z)NXDeHv1bM&5z#yRjRBTnp+>FT~OdjLL?jTTyfahm~eG(mH?TPx}T}T3m`*Yh5M&F z#Qf{HFt)D3OOI+%g{WPJO%S_^*J&I5)3#Ki^g<;R3N}$N`Mir9uCi4A!fL|cRdPsy zANHD6boxuGyZ&AULD$)Zj=e3&np)=T{8wH5t#U`&cT-fSR>?3KO31D<eV2t(LfF{(@Y`blHwV4l(~Df_3`m$nZ;dsWLkHHgkw7;I;9nXWTYh4MJFf zf|PcvPvb~8OB#<^irypiGf0TX(nu#L^gj3GER(*+H!*S2u1=>F+jUW4V~t9bpY;Gf z3&>svrcu+iP9@!A-_eP`!p=TI&v?|!=`P>|2@Zi;rczLedPVB|C2Hw-fRRNa7X6*1KY?Gt} zT$y2Dx#f zINe)NG||-6HL0OZzuUHxXFGoZ>FSOUt3ZcU$J@LrJm~QzKLf3UaP_zQ3DwvE0mixY zWFwz;8I(%oJ|E|6t=X_2=pv&FrJ*6F6!20y6w0lD)@==8>&DCJCx)t`+~;#r@TT2r z{o@PB`IRGs#tGQ>t@7Go15}b-yf!nb zjpI5mE;$MRbe76J73z{6p6ik+6Axu22|m%18@Z14{i2rrj)q)kaY*;%H|?EYhHW5hVVgvPX%~?FgIB@Ni-m{@5g}dHPvM_X3$>q zKEC-0r7Vx>Vo`Jssw%iWBaz3{A-Qc9a1<)e*^Wc6C*VUA$Jd|b(s!#?8XLkeD7ef? z17D%EBSVGET@ukA%xv7gfUXkvI`E{2n#M#}jKsNGuRoyINFarq%HILI4 zao!p@4+3G1thR4Utunnbt{4{nXWr7v^r)G`;8{p_1x#=&VO3#YTyrJ0^%!o{j&x{v$wX^hM>lkso$yC zzTWp=&iRH86D=NmzE8A5TZQ;00?y*RxI%>5$k>yL`Po&mTQXgRyj8)t{N4jBSr5o%t^H)~?Y?-MRcjvCCf`DpTZT!l& zS2WrX=bEbpA!N9IkFyhd>mCs|w0dp}Ghf3EB+LQETUG1nY+;v6Yno|q6^&fTf47hR zPh3+f_oX%Ak{{6@C521HsXj3>l~-xs<+koe2MWSDLa!kgvQ4&=F2`LYlg8uGP{w-F zu|DF$jbs+@H7YeH|CpSL=X;F;uzM9^`YZJ0*_jrtg7szO0iO>vUB?qhHa9gUW-COu zlO3~0dB++;U&tFeJG7WFprJ0gJw`L}u(%oqC}SR|Tcr7_vzbYMUY5C~PJ;rPcfukL zmUN}|tw0gMFJ)7<3Gx)@1rj~=>3;v0wgoUU-<+|dAbenO6-SM?%BZOxnMCUfr4qq1x-inJymtpM5VJ5bxlA#c_v0OmzPO8w z8tJ3ugagTZ`d$4sCRxyWg7rHE?9O@{=|W6^TTGymgoul6Fi_||)KzE=r(q`)VNx*o zXm#(vSX5#|!H_yaDzV$k-XMEA%|X}U`(S|ov6$bJht?0mtRf^#2CUa3#wt|o zb^aVB$Y4yuc)jD+;q}iQ@%)Z}eQEL29nWcLW!E*D-;Gw9wfi3%Vc&?#S+f|`LYHrG zPf_&?tO7HojC))3Ue?7U6Tl9Qb%&Ot7xwe^1V>s=>T1oB=qiE#(wv?7H4;_r`3|+& zT@5%s#o-H`Y-zpmwQ+f~{yMF**$~sN6e=%aO_8lJ{iyT=9Sw&7Q-Psdt|a0L)C;-4 z=wwpmORN)7T99SpWVaiLoc6^X#3Cy}b!XRz^5;rA5NOz!5KHUjGEB_F&^OxcrJEM* ze%62SX)K$bu$cYa*~ycrSC8H15!qE#OK^+-wlGc0$HV)Dg$VJ=AX+~PcN5o0O>>k= zz^k4bCV|(Sp=wSmpMZ?^Ra4(hlE^ysH8S?W-@zf^QeM=cDMY4EYXtndO4|O7vk$!Q zgh!LzI`MftO>x@vi_pOoA;S@6xfD@YGO(>gPE2-GLwrV6-oo97bpQJ_ZamN^mgqV} z?;l0pQ$TS)9D#6I%Iq3!iAn4GSr6j?gd{XI9N}{43)vlR{g(nG1U6+NHr`u0#TPTK< z!AOWo=e<9li8-=$2bzFwjF)RDJ{4oG!N_L``$=@D2+R%3wU3RjDCq;G>gWSX4`v=~ z_04Ceq!+p`Sh~B>M%c4@NW>`GrHHe^@u{76Vx)Wp?RXp>Gk8*10$x`h3{PF?ISM9e z7nXIpsP$Cg`>3U0@>@XE%AG5hYQ(`-_S1_G3F zc-&u@82m$%872^dw=F^v|9Q^KO}TOc!hEObJ^uqT#@~maQ6BHN%(qRbF^{MU1UJs@ zB_>@~iKp6e{l)qB7!jD>lgX7n0)a!gpJa+zkxJ0$3cHBqS8GQxZNV4#rZRVyk}m3- zTyso3gmzNPhsCQcj#DG{4}7?^?+$zk`Pd)e7dgk2gW!wrG9%FZOL0!TeS(?<=IYfD@| zxsGWUQEX!6iM0`L-Ta;v2j2_Q`?zE`s5zS3}ZDp z$i@_9_y8`H?G^$T52|>(2>S9W5>B~v;@sZ85lnK;-=`7 z)@`fJMPAniC9e^2&_LvCgA#9xY1dG}Ehai>`tXIcR-mu~AeV;Mgc3X0XN&gk<3>>n zY@Sk5<5=nZCtJU+UYeqKBG^flE1h|sQ_lH*T;7N$=!31B_Oj$;v#63%imFfx!r?57 zUgwU~1$S}MR5}hFtD|&`jEG{HP$}#)LXFP+`q1pJpS;5Mih}3;^O~7`JcR=d>o3;A zF;;`b>wSJ;lB-58_@49~T{Zi~le6OtiZOthF>FXmu_yAPSOgz0T3`?>lKdyaQlVrG z1;26_k&weXGJjzPMHa>3fcXlI&GBM+ILmA~1WbI-*4jo6S0DoX zCKQsjmM}^m2(}`*=K&#@wQNy0)@20{w>i5pRzm5=v35vU{cPCL%{(J4B8J%L#WL?Q*=ECs)yj2ij=Ug3X|gikp&KRk94%B9fT&?r|( zx}HI zu@Xx*Dd(b(#)5*SiXkSkDa{n7_ExHYxD>V_b3Ujvw#;N=Wm^PDVMS&p$|JB7(WeJ6 z#DW}Xh(;Dvj;YPG-5+)vOUZttmK$f2Uwf1LTzxA_n^baZ$F~VCAN4w@_yM#W1T(!G z5!olw<_8M*wt33PCgeX<#&~{<>tQ|aj+^ITupk4 za;tg0P^@CS+eU1qVz^782^ZlqADUEZ%Rg2cyB@2RX$VZ^o3y)pSLg%g2kLB723ztg zQ2;Zl`q78646MjDs3x>FGQDwJf1FMJ)K8B{zZv3Q+x(W~X$U+$H#4DQ;~JuGF@k9j zK{G+NXcq0-t3q!tPzUIpaDh;@_|O#0Z~zyJ2x}cq%gJ>my*1wu-(GO0z1J#I^kg=> zLtifWU02(2x~-PZ=^|-RLY>WWMno1>O`M`7Y@V}~Gp*>6!-j_TXqL(g;EYoj$rKED z>FCHaHfPH^)4%)`4to_>-BY;2_7-U`atNqTEnQgC&aM-zR=*gCY!GeyDvh%g#5;BE6 z?NX@YEhq5qXC6bGTZ-)hgLR!?=Dm%nfKs&2K}OVgzj@?vq!x7;dE`OUu01hY;ol1J zQ0!x@3n+qeDVTA1(-Vixlo^pv~XkMUINOb+VcU@MBTH<{R+cMW+N9C?P@}pueXl z{5h^evM^ICso~4{>dnue?byJwNuXF}R%%poWu_20!kDKH?wIxAPw+luz1PnS*KH%ei=KDQS(}YX9y3Sybnl^5Gzd(>FF;JgW@8 zfLzLjIDJQ`Y&z9d-HUiA+StRhc(vMYqK8SYUe6h;fS-yyvL)JOHV+{!rpPXC$N05K z`wmwn%BSmr&KWqXONpXH`PdFIN-AuCuGTekLK_vKb1M-E+%HvmXH?UrgY;fD{ra3$ zAbGy=NOV&MHA9=_`68{XSCbf$Lpx&)ZD8V|nfE{5jQs^Af&cFPjz6=ydG$l4l{Uyy z=Z?+_P%TA>X&5Bmao?P|(fOt<`$|1V3rDJcXrMu>Xj+iSyz{2gD zr8IR({%EOoG{!O!*IEBcR(oL>FaTlbl<4DfhOVGCwabc9XUTJvN#RGx!H7XbIO@hV z-$G+4GvQYkiK{ze2N5z}*5+?qzKZgJZaH;9Tw65i=J9MS3zr&Y6iiW*yD%< z1n{>kpH#^wd15n*zBe*RS+gZ^Bav`SS z1P5(Hv3UTz!<5dDT-JZTbv-2HTVbm7z9Sa%HIbc7-;>$L6hA`qrS zV#OUHLf?L>$nL{VJ<9EpX|uq#%v}pgXu+j9mX`W(ob8J z3TvH#0XnKDqvM||3KC)n;d?Uuoe;AzR0+0uJ+Ta*)QI%?>&wC{bqR6_u3D@;V*?QT zmm-Prkx}F3C)M7O0*^512~K%Ek4JI!*#wh$GaWdudz3%+c0&czVr?4_VW2ZcLG;_o zRV$U!RWJII(LWD;`9O-;av`Dm9%u{}Uphu%7ugdgXqE`5;Y`GP3Mi9MgiR&SA5nB*)ri0IO>a|@@S|pB@{L=slP2Y->K(rt+E21?Y z5oPgyc`|bLG~JBu7A`B2z$D^@kSw+nRQMyG280!U&dK*>b zR&qXSX;%W(7kdZSwror#=j$LLRl3&;2KCFFW4k z=?RQ(F^xT)c5W{$bjxptQ}d)9n$?JK zoWIN3=!}R65sVqa{DIWxG;c4GDz6cuwTIWbA&2pb+P}F83E~w-*phgrG&Ajx>nxA@ ze%|C4Sm&E`nre&+;{ga9mza*EV3K-l;x{2sfQ}WE!va<>pZT)a6*(rkl`f=H8$P1- z@lSDrzU^1!Z{~cKZrv^o^vEnf<=n4~NX3ohV;@mE`xF~uL%76rF930!E=lGt;HyTP zJ$~qd&A2#s)W^oDE2FpMGT8z$kLmFTK~GxVW$frGpjMYG;`1+7-Jr)Uv0_U?bgfjs zk25OO1PnN2BW&I=*)ebd7Dx#XZ}O`Tm{ee=!XSU-VT}_ z>dR?kJvNw}vG$W2pohF34CSV+ZBrj?z{R`mcey?P5M}O@#~{qffzT|SeFk74(o~19 zi$GOwjf4{A#3Eh!(2!84*sSI|UHz4>=o9kRB}+B)hkX*T?@E^w#X!FtYN-yrUtFbI z6jddx-gf_AY7#a?+}{^E>WEqWcC};(*zi2UZOg2ql#<c%hcY2*IxXEmUI91Xc45!cmr1y9g`!Bl{nTcaO)7km#!`Y`+ zUXkD=q?fC5CTc|Hv)t$`J%8gs9U#ZxdY&l&x@%)3xq)CUu*@ ztZvi4X%1}g#Ob$6=QO>UOi?A|R4_-j=BvX?Cl3{#Szt*AXSt}$vqNp2<_F8quBVAg zxLhr8aiyFsSFixl=~62AEklzjQ8aWdMDd#3mykTeZzxi_H7GrKo3{QFDbm4|Rp7Tc zxNtp$?dn^_dVaI;i{T4+z44Z)pxpI!jT#yER8_Z^TW@7+;`6yzA8WK?KDx+cp5&2v z#OT$aii<1i82Z(WMxpGUBE2f-E~0d5Rz;IE7m$&@0O5v4NN@aZdx=?$@;XFo^3!%C1*$Tufv@u$ zpOj5jt5V>v?xQ<;$)gF7GR!^8hyoaD`pEHy z&!J}_6!E?ECOKWbX*#~oWt+FZaBCCXeR6v*I=&;a`1o-Ln`2H#W`S*hAANgt(%Q~5 z4+Uf#{2N#Ys7u7nOwXat1zofI2B=3cCh;7-?c!21??#&7oYiqB}o2_X!70Q zZ$Yf&uW5N%^lui0l?MApGQ9U};O#S78t!6A^mG@Q!FSb{Yug z%-(~v=zl(L`sjb?3Wr$42rxC@IUJmbItVQ@BZ*){;=6H!iN(dtGDpE6YtUTGdPDO(TBpH?u5U`V_ii#22d3o*jzh6;hCGSPG0-1uV^fqns;vXKJif_XUHJL4U#V44jEVb zU1>n-d)-9aT;WkuGHjw4vsoV%*UYgAj&Cd|gWbs>CdbhSYnjfbu34%LoJEs$^%8}; zmMH7L-@Cc`jf01b)ecXx#SRnK8#kEg8y|F`ynj- z>LJDd%LT0r(4+y?tAnFpleVEc8{vlfVhq;(6V8O0u6iPRUD}s4NfnMBdv?V|UBLgB z8!ZmqRNyv#pAZ>8WquWj{Z#$cG3b0{116|g8^HN>CX+okwE~)RtmKrmh#i>VAuN+3 zd}#+|W(sgrd)J0%cgi9jMeRturID$Y&W^*fbBI8z*2h+rHV9Qxh6m1`10C$J_ zhip@{ntDSGVWI>PSOL`I_1zV;ya5Q9!Q0jnxKEI`i|#p@}q2?X>}2;_=>8;y)278^P@* z=8vBYn~6ekG!2F(F{Xzuv{$V_>cXMoZHf{Et=o~rGQg-H=D$#o+JtWE(#_}?KZCO` zliC(3Wow=nZe@NLMATimwrKL)z0{DmLMix zx&A_WN(^N9k<0lFSgsT)^BU56FO)5$HLBMPl5a%4$*lbZ=FwLS z+#N|A%&92!XE1U!QH|TaKNntpukG%HWRn0Jk(y+ws-Tee@t>@%C5m?1+oMAnr0)qh zrW-SC>p5Ia_lYw1{4I#3g`GVViB6<8QRdR2-tzI5-0e0!-N;dbs~McdMk@A^Y%g;9 zIVuWv5|uY$7NuKp%*0=Q%uA7UGIgf^s#iK&xUA3x=sAo_ITZuiVT(=>OHC!$4^kb` zv=HpqADP&Npyv`C;yrb^YBAE{Bw0&YdG@ZpqfXwNS4`Agneb2ApROlMmWEb0WU&7-fG0v9?*1u#YJe$_VXLS#eOY5cC7=HISf>}9wz3LAD=>Ff7m|E8( zDW9)TC~Z`R!$MxC$&1$_d_5{A8n^_yq{|-_fD1M|fa^qt3wtZvgm2>$ilQ3`Arh1X z$Tym4oCE8%bBDmjMWx3Op@dV6b@#CIMl@j4`CI^Gs4cA zr53sF$Xxt}d~8xS8Ut{<%o=~!*>9`=2`mKD``AC`$}JWP%Isc$Bh8jsd|e{WY+L6U zo?8g7r9NeBb6^UPkiTVOO9gRD#j*HLaTWU|e`2YTMMG9|LC}$m$V{h%s{=_(_g-+_ zmW0%R4~Mz^uezcMGRzTKSZ6$T3J#isV}3It3k1HU`S~59xRRRr&ct;c%t{=o*%4Iw z48dsCYPE>wC;qfKCHa$W8R=xo^5;A$6z}Jl|lyJJ+yr#$1 z^L_z*I$xwWzW``@-M+rf&ia|?>i@?EKDIUd%)v%0+;tKZP)Pw;35+|MRIFXJkM^Ja zy*BV{aNOj`Oip+l3*{*_qrFkqw%RdQ8}0kA@MZ8ao|dYJW-w>tbgJuZHxV#iluuPn z>IrC;XXUw)S)(Gp3kMN3(-=yJ|2M@E&KB>jrZwSo~oHN2-+)G^QtQ)Jpi=uU8)8YLwaRU>05GQSA4;tp(XRr`S|z{AU!? zsj65Hy(hl<652<_PstByivkK9j*#a|1e;F5QuV86>sOr2uquGA*J2nC*W(FMVZX!+?Mu8_As0lWxRVHuAqLSQjkaiP!e>?L0oCFS!Ug%>ct}hR~i2%h+r5dTZG94;OI-Jy)!az90VFWJk*Y8)|RuguqBcPsy&_S5n{pkXSS&i%;qJcDG5omuVEPEW+m zd0hvvW#pScFM#hBdkv^z3ICkcI(HXO)NP|{>cHS0T6!?`0fWLP((dDqK=;1gmqd7m z{v=Lu5Tg?GYHGk88^D*#{|^!Cw4krJcVa@LHJU}Da`9sJPS#ms#el%F;^PDZbY*`; z_%)tKn^w^F3p{cmuWbH)94^6RMMIY!d-+TDo`SrgIQ_mD+71ghpWkh9TZ)9Z(K0XR_baG{3Z3=kWwD)6lWZ(KV49B*eif!BK*mf$m-AOv? z*y-3t#~s^A$F}W``J{j6uXFDA>pN;x?OGS-o@=ga>`g(Uq{<*-=3oMpbg*}2U}0qD z1<1*(IM^B6b1|p@E!=F4odK+j%*^a?6cl34Kx0=c2YYd2S0FEd%heK~XzKcz`WeB@ z%mYUOkOJBRoj;vs0242OJkZrx&C3zU0-!ejXQAZa;>uuR?DFXb+FMxJ18F`}#2g&G zoUJS@UH{f#V_^7O>2J0uBjAg%sf~k&i;We)*xn5Ag;Aanpy1%~X|w`RJJpsp&Rq6&~wQB+q_rD6Omtm@|I=-~WcSj1G-)TQVF;vx!a5&)nEJwQrb zRqgMu8qod|zXd%&LG9E3x6fzb-*kBiH4!y!B?%U$zxM!O0k{L5U9A4L{SR!EpU43J z3+*%2+}XkIp9TPGOIKG%UM40F4-ZBQHy2k%2WJaLN85k!saaaN06ZL=Z2+HFXP_%|7XLwFLgFz~4;)$XS^J?OlL>lO!Gf<+S^x@-ykv?)pE#K8bMsThsPm;Q$vP z@ZZ{48oT_HE2pF+2e31?vUdgA8{3R|6;<>Kn{uZlo`xs@&OFM5~1 zduC<-k4;`gK}J$SRgFRJQ+W0a@(!Qu*fY9%y8aXWx1NZ&94~;InG?Xm!wz8n6sd&0 znV5r}-6v`nxWDv?TYd86>fr3f^#ADE#@@lh-sk_dn_JnNng3&%ex{NdDXPUlO4H=b+Mj zo>DUhds{Dn8PFV#Nx{MOQxMeu|2b#>(M!_J)>gsT4oLlTR&v7MEz*S|&nM}#Ku zFIH*=2WLBD+yC%cxky@h0?m}HTum+iMeM)qGOos-+8436umyf@%0DLczsJe;Q@5Y9 z$LjCs0x+;Jv;K$gQ1?Hyb{Qvi-`u6_V>2WPmy zyUED{U;;Y18QcCf|6^h20Wdk(1OI2u!3^>FxS{qG2s{+`f(#K-)$^UsL2C)qZmO-|N3e7ytx%0!`tT7aUCaf~`}VLd*Om3Z}{- zDR>yqDzOB$zQjW;r%$!xVZ=rm_lhjj&Kl*1VvT))a_s3s zKge!965Mg{pHMn(X3xTN7r!|cdNNVQV%F!4_;q>Q^!x1V_0LKgj9>|;p1yfiG;)Jv z0#}Bz&3QtIKy%-%tFEdeg#BbCiP6SiEPgI0&#t>XXZ`#<`LMYV`yFnaHXzpVbM{2O zVF!=IjbGT-#eeDwJmRtOg9S5Wg8W9BA2{+6O7AvpT!h8mS_TuY5dSt>4esioQ_jb9 z4d!^HcV?(-OQ>gBCU8-74ttiSY$^Ai2N;6D?ED@Bx{#dAri_yTW9XipQZhCYD^*jMN>V6QIaWM}#B`<4QjytuI8my1{SV zo>8o*lkwm=Vq8DVZD3DZSC0MyK8cb@H~_~G5YE_QknZFO{p>{vbGYItq1xx^cE9A8k`JX9OY1v_F@e#^K z4#lRVVna&ERtEMX#C-+5Wn4rJ=?OMDflYP@B*!6)WG`v^oSrxf?{N~|@BWG|mygee zN|kynsgY-8s{I9IrboBE2$g}1p&5t70!LTsJG*)YSIT}g6O%poYA&U*WwrhNZGsF4 zeBZV8^hcdmSz-yqwu>Z;{Kl>A!^^<4meFsjmB6p2k;q65yKEvlk7dObWmo40fp>; z;Ajp9nXNyt>_<){sHVP3D=Yn4mzbRvg5AI1@p(~?* zX)fF53Bt#;v}7-CZL#0FqRIFEx#Ye*e~P+M5C2dwPTWi6M0-A0CjHohw0cn$8-I#^ z*k-s&pKvQqGPDDKxK$F{N=0w=P#~VTF7Vq#t-Hg`1m#aG=%Qw`ARvy-Z9@`QNefok zNn#NrR=D|@FbDCTvM;H)mEV7l1{v{tWc?n*czOJL;TcmKuZzrtPo?iN_*D7(DdJK+ z)L}`+3nw5U=PmqN2fx~V@Z#00hgnR)$Q(S@8+(qkD9UVz9eQf9|10)LpmhWBHw(+nrgL%m}EQN^@ z_C0>`O9Bx*MF>j&4{)U0*a+u-x*|Rd*&xy*hUbIDu~h^n%E;!f59-u;ObbDI?4@sv zNQ_6Es}wnBZuQtXT_x2Mt{2SJ^qw9VEgqWzzn(AJ$MS)L3MJdWJKA+1MLrC0<<5|K z<1({j`X*4xMcQZ$J$3>=C^Hq#wm8}xo|T-z@1;_cm=dTG5X$ES@TJel!u@Un&-?Pp zbM~O3Rk8v266;bt0~iuaTkSTSQp#$hzbSHkUEj-Tr~@cbfAMn}GzZQ*P$5Nm{4PTp zdR~w`!=BYRz?Rcf*RL|TXhDVBvz#8+@^+yIp8C2(_-P7~NgQ(2=lFo~1pVM<2!h3F zq}^s4=kZu)mrk;o(#jpS^�jbnAzbVSiDD=ctNTzXYFhtx#vk%jN`^;{6p_Lts`*hK>g z_zw;Aq6L9{DYWkhFBN-4Bo@WhYV&LO+ktTVRiToj{S#Yr*hg;5AO#tZ*GIzhR{D-> za;f28f~3`FoL&w1gJEasPUvVl@L-Nx3{w>ESrq)ffeML2fD30o2i`Xxfnsp$+o09KqA`jE)E{7egKkh*gEC>`y6!=o|b zucfy7;QK-`oli80(U} z^i+K~{mb4kM~&uuGGCEQw=07=L9|WYLrjZGwHv}449GcSIpOWp?&3W85L!sp?w`&H zD=lF4cbj`#&~p83)^L&DTV56_YaYpEHSx(?&PyaNA5V$Q`Coa@>xmuwz0gZ0f{i&g zLYvIwpxt%{qmyz%x5-ztwJd%zY};F@RkI0)iM5Si$T_jAf9SS@)|n;>2m&qudE^>w zFWyeDV9?O;+4+{q&7sacf=xK0mm1gX6%=b#bW;%rL}Wq-@9klKC{fARw0!k19j@Zf zQ_Jj%7=8pm9nVYQdox1I12c6YjSIdoVL|>rRMU3&W(5{CmVBkg3Q*hPRk>x)KsU!94eu{nnP%l9t6Q$1&i6v{!#>n`p-K=LZDPzVdEg*mml&#fGdAW@ zzU^E6kSps2fLLUhk3_Jo9m!GwH@F9S_$?viDzxUFk+#>iiKA- zYy|WE{B)u>jb#omEwV7So>IbaN26XPseOev-en2$=2)+x2%`nUG|!@&+j8}BgXupx zhBHtE?#%6vtC1<*Z})wIwhY$|`y(|At^%mdAA>LWO{G7OMY?8>9a6bI_`?lg__-O- z4*B=o`p1UP;M`wRi4t6uS1TsnY7KPe`JqlA3nWf62l0LOppn+2el-@ZqK2|IaIEWn^O#{eD(LUs6k@U=4Gb#sSNd62W3mAMj5b?WcF ze$;^9cepy7e9&IWSpEN)w3KA-G}Ly|3o6-|RU?8XIBc(iagr%V4S6r+%R8)svlDRqgOF9!SS zVAUCoa_86^dPi9%VOU*6%+{sQ;~O?O>s=Aw7MYa#11UUQzfb!KO|92xH>Pz|@6bB` z%>9$|EZUBe>LoNIntdb{{UP57wmQDLKv2}UgGBHOF;sJUt(q5kMMFPhP@Rr#8?XG- zfA+Nz95fB}rcn@@f%VI}yR>q6Sm4{S59M35InwwdtQU z2WrGJ{NOqnzY7$$kUerKShEh?zLh5oXsd5e0a=PJW^I|vF$e85=U(~IwGEyRyLMg< zHUi*54FXo_GtV9DecxInljrn9f~whd-crNG`yqS- z%?F zUjYVkZi_9V4kGC+9od2x2)B*YU$Asl=z;nw_0Kv@J_}K)R3Mh_YZebXaW|2r$WRfd zNhK}&{4*9-@{g_M9fJKZvb75tDN|n=jQy87#KJ0gu zDX0NE0x*=?qlr#FzCYJtf-(=?dhTZOv1Jah2(F13*LXhU+rhtXoZ(LX>V?5nHvT2) z`@$`OPp3nGL=tf)7yq*>m`#NvGxyNP1B*MNEUeCafvzYR@2+y~*8x1-TCvoDO(1d5 zRrK?AYIgc+8olPs+%~gMwz5nR5H%-@Nhp1DQpiL%j3XvcrhFbAn?k{kCQ2n$&DPSs zCq=_A52eJq{EQ%Gx}R3gPV352FnzZb+YzN9%KvsMYgF%dF35?0Wk)k->BC{qyE1=Z z$RFEi4d2KWCAlh2mlfoy*>f#pFs{)Zq#=vz;4zd6uFyd#UQ+ zkR}Y*O`f8o%Fh<${DLi4>?Wx!N?3g;$Iy9&{V>z&E#_?5XYUeiWxCotm}klVXt%Mk zI+4{$hX2yMtrACtX1Xec@&(XWB@cTf&OgN7Fm8}TaxCnwh~-3aPVnGg(=p!lN-q)e zVObxnEa>5cP2RaW-}rP-qe{~okV9WdL{sx{MZDiH9+VxKw{EH3T__LJ^g3jOS7K&I z~D6Lk@%RA?3bTjMTL7hW;dAVkKP2AR3L9csCtEe}r280KoB z+2=#@a65HLjnKJechm`kkYn%R^U^SxC?-K5Q58)mWxdFH9 zeNfvG_5GlV@3GmdkYZw1k$IW*=NjV>?$qE4DtCVJnOsrDBp7}cP0KR^8YwKI=NHzh<$eL1B zNBT~pd~U&-aeJRl-S<$~ve#zt*xr5aRLvtffJ{<937^1R@q-`<}&I*boOAA>w|bJrxokLtS(uC<|rN z&4^wb-SvT*K;zO>(&?h)6{57`xp+Uh8eJvR8{zZ zCX&ya`RHqw{cXg{Hqv>k;*GC)`W<^-vue&^w49}W_5^(mpLPl%HTr{wen1S*(VT#_fHdVsced(LToSgZQ z-ipJAWBK}JzLt6kFdMTHToVi_3RtoOJ#nsR6@>^b!68r9ihXb(&%0X`1XN=gy+TPT%=4py^|z(n5zVC zIT_-+Q71T73_bx=gqN#J-7Oc)zKUbzD?8!Y1Q&`eLu?gbK{I8Bs|l7TUlVq3u89v* za{`sh<16EShU$0^WP2R(zOyr5xDK`9OTy*<&>;Q0*Hm zKge;>z@afVgC9a(+XCVbglChB4A{iM3g)Rgm6e5MwS8@5q9`ibEYz=gdumaHtMUqF zxG&!i=1)pf>6CujA65eBxB3w@Wwy3Jh{5o>x-x#(i=?aZTS?=AcgI9w`x%(0Y!2O= zkSo-*PwpWvmWJ(q7PAIYDx&Y$p=P>fHr2SIN}iaUsCMMX&mH#TEiAgFf_2fUH5Rfj zK`xQiu+FLVmW-lg6i3H+me-#dYu(i3W^#=(LBCz2(PS01j5m3Tt*KnTC9*Q}=6Gs} zZ3YWU>^*Y2&YZSU$FWnTmEv9z=H`kCySsTXyl=dNZZO|oTWHSFy%Ls|TIWeq-ZXxM zFBLt-fGoj^2U+L>6^}+3nQ`D-k(+e(A7z?0I}5(pNr51HEF5ouZTcpiL79y0r2z+M8wrUjjilIF)Wk z>K~+S;9w!-E1}YWW^41rar&mWH2vh-_Zi+~5&FL2$Nty$Ek%CdNql&+h@_tcoT@Azrs75X5y++91J!(6UGJv-!7ROB=1eq>J&6DyVH1X z$aFchd4pe~vTY2k&{=aZq}C_7{cFveSmPGv5&oE5@kqw5S0`n;Il;o;q9>Cxvz*Po zZ-lEXgHuMT&p(`jqRzXviX=P1OAS<$coWe&azr%F1(Xn=$jWU}MD+OpX|p9$OZX3dv>)e~)-Zw0 zf{mjDomRD3XC1IJJYyPEIECJ1!aI0}<30XFwc18@q^a$=3@qf~l?5dcbQjLQdfdM( zFNbXus#-!*DUD8zRP@~@aES3lfMwaiOqBiLFu~ z{R2TOs4*<|<9E`ddfY9iseXs$`o{N^XL4{aojS}D0LQ}SdHFIgv@ptmiQTdkWY7^k zs?SVO_>XmSI8Km9*Y9pJ^vqLg^tmU9$vI8tz;BrzQiEnp&(^IVRD)-eJn%T-Ow6}^e; z51vE@V!o95)YW`&`m%+^QL#xDCv$F4Gxv>DDn=B@)3(SxdWd6k+HKXm2z>7L!2^WP zGwhndWWSK+z6@UOkW@7uQdAtJ{&A*{l?t(7nW<<0s;rR8<5nH4W&VS6!_TkQf*F!% z|1001SI5d3IZh&9=^3%nG;ySO>W~QNc}}4!z_Rft|tOl6N3Q$qpI)g9ztHw zHyMQh#2*VxjL_idU6N_)$0ugH65bNUp=AR-7xPQjxDX+F_zhbA!( zZfIVS1j!3~UZA7Bq<6_1j>}Q2pg*z2-zXI3QCUEp`MO33i$pCM&XUa;v$RmlvSzqf z2arJMcV{M#RLd$na1@Ei|Io#UZc{N*tsLn7FoCK|=1)}Gch{`<2&UwRp&J4}w|Go; z;CJ&#Tfhsz2JI8Z4thMIY%!KGHNheSAZX?7kIwj;@EPtpN1!H0Q@CFY zlpy-xMi6bMVe?b?@K0CEQ|dPZgpQ{j0pdS^)9W;^d1yO+_*d3-0W&L!U%m@^NiM%J zNt~L;%%t#&g2PQfHC#fmEs5aF|Ai4S4>t9U@}&xq;)Y(Ap(xE90$9h^Sa*=($* zy9&R+Uo^!d5435M@%e^VjI5CpJ!b1jRMvtW$%?&SM4dPS*}m~A8HULGJo3vXZsLn| z^r*eo%G!f0eI)6+mF}l+X#5#pzI9E|BdX}SXHt%vWrtD1aS1U8h8_47A_58L0Fg{&Ez1wIC*nUj+5raW7O!hTbi2( znIk`nN;mEedaqB{S?1Jblk{RdHFdV@JiCc%44 z4xvEzmih}JZgc5Feo&RU-*mY=mAUTir0u9Xv^12wM;DtW$;XR&;gVOIUMJN2P5Uu= zf*>eb@X<4-f7X#;WR=M(HnJ8f@zHV1@1BMmpOu+hg`XpcSr_ho8&<~Q4!DiA`mSTJ zu^N*E^&)w~=FomzS9YETA8qfQTDcHIj7_fleMwVGo8mprk~6X-DVWzWNy+AQ9eDfd zbvO!HGnxdYil}}eT9pf}AU_6AYpduHwpVU0MsbsqCSiq|=dvoOSCCedOrQUl7BII# ziD{sje;00{ASiDh8``|lN?EIc+7Ogolp`qpju!*doo<8)HSbu7qXxma6k;e~kUrU^ z0}N-KGnvhq3VMVHp-X`}+H!4jrNxoz;bcdcm)IAYk+i-mf?@N-`tWpsuM#rYWkXPFcb34+3ERy+?cTE)W8s@^Ku?T-JHmxbI`3#R#R=EqWwM{vs?pZ znCZCbsdu~``Q{J>!}gab-}s!k&sOikhIVa-Nsd$+xtp8x*E)+!A$v?`l!8L*FU+Wl z;B>h(O90P@@=aT8^03GcZ^z)8kevDMsYBdC+!ViR+7lAgD^t$Cq+sy%0XsqXMg44qYE z2HqT|&4)R*C%pE2s7vion_Ra5%smrJx}Ark`67j2dV^+%`lhKWZcw8L>EUJ~AL?h1 zP}xCk!+wbO!COU86$3|0p!z=CLl1sv?S6JHfB}$OI_u*Cu^g8@FY1^=<9{7TE#C>NpFh{Z^`RV60=nQ+OSAJAlN{b3naMxxX{_kte2m;e~a;9bR7?(3F zY&$POVx)MC5w#<93~knZNjXnrXe=>+rhp5@_xHH(mc0WMZ4{OjrUuCK1Ut5_oaylJS+QtHyY1=k!jegNf-VmC$%iL1d9 zpXozAh_JAMTC)lTG4?^fn{?PWpe8-9A`r8gBG4oYA@l5yV&(8%7zZ9Qs}y*e_jBZbRKCgK^Mah8vq`=`u?S?Y&;8aD1&3Gq^-ivO-%Nlp_ZPt|SKp44W(^ z|HG`I63j_?;3`#rQ?5vG8tAuZIJCL1k&}7@?on{{+%?tWw{6Wbt>15|wLI9eh-oM6 zRX-r*jnI<5&f3%sagr*7X$p|M!Got{SXETMA5Sh{N{Mjw!2s2ys<+gD*upXVmtKOyzHn~8D$t*m_0!R zyZSfna2A*^L6~_=DRgS=->=s`)UzKW;il~MHuGwk@({$A40fQ^$$br?@hMS75T`RB zz2WyAe%LeJP$iN^-fHv8UlVd(8>Nla&YcwTJ|}GU4>=OJcKcnoP#};4#&~2x$Cuh) zxpRs0hovZT&rv~{`9&tU@15NBU_vR=AZo`aO4dG75z^A)y1NErr%Bf!tr@ziVl;S5 z?9!L<@TRal`bHB4VC3x74D6w%lAFaNxK=a2uuo8z{A55q77K{u->4)LAkOMc}IXyVr>+fSuyVXj1 zb7ER&c&`4swOYWe&_0`|R+x81t+q+amkQXR~Jt4qm8UbumJY#fU|d_{E+4;O#19f72Ke!Hr6vv z+gCrt@kx`Z_<4#eE<}sR&OS^Ly0W~ORPoV)>u#xzQcz~4ht1=Dg8XgwX0-06a&~X| zZ2@RwJyO-~P>HKfA_`&IM|5?JG>u zE~YUX>H(@ghCGEcM#3@^jj~}3!bsl$3ol=UW58-%`qeru#|3#ATIu}G`Y*ZueulLi zk4DAxOut7mGSZq1vVd;WA_o4m0m92YJ(H%jR|$d3vaI`GcCcYbT!@nkCQQ`YjF`cTvN%ACltF;#dAb%@AjTJDe)vueDmbrqDnCDqg)li{GnE8l`= zTV0zh_%yCsXlpnwF&U0M2p%ST=W#db{;_&+GE3^>EWlnlT_1aR(e}DD1f~BeB2Y2L z8n=^PePj z7#H_=`qLcfQ^~7Uc(Ix0-;fiX;*Sx_O^{M6+8B#AlrTX)eIysPb@yJqQb19=ZL*Nb zlj9ah*s3k~eE0h#fGZ_GN>!&VJ;0$E+Dh=7yXi8mz#q6fc$2;DfmQP?KxaWiAG z*G#(R7L3f+4R0SBNHDFjNFk2Lvu>X@f6LqUng*I_5IB8|*4)=wUhv3GFL9Yia06v7 zBDv~F3+y^NwceZW<=q!0k09-#qtVW$n8D#gw}!B|!YKeapeB*8guVuVOpM*3EWUH$ zHxt}x^H<>fflofjZaU(mZm)?z{TlXuD-Se<541bcx`J}%FeZ4ue0+0vLk;OO!ct9@ zbXRJNb3$GqrM9TpJs$zKMXCAqQ$wAP*_ePhO5j&dKk&4W8xw=v#->$G8*FbYX+9E< zluVo)qA?8t8CV$n&iqa+#YN+bA6>fWTRLBEFbs2FHC$e%ueK7wAMcxt+4|L+E#e^R z_1(VU6k`_5@c#BR(+?N-`5KusKE9g6>X>?YDA1a%6yP{;eo;>4RkaU(*-4o1L7i|A ze-Yw-X#XD9s)|euL|M8{{|&Y!Bi8CabR82{Glba$zPB3GcgNNJv--Mkm`<9n5_S^($g0}{WZfP7T@dIpK#|D5lid4WsEsJ8WZf%*LU~-pl#t zETlvz=X1gNcnb`c78Bxq75OvTJKj(f(=}NFx)*AEE`SXtnI+-Z?_&`~^WE$wzQj2D zKSd&~*I5X%x9Wn*7fwsEL#%r%_ODCwDY__ks&oISw2UY$@ zo>9&((Okem=x^JFl<#Yh){m5&aAyzw^-4kXmC+P;r{Z+C zj#3{uHk8$bB7V zvWIkeyXbNY4i+5TZieW3Y*;ILuAq_;XA*Y(XXzZ6y|P&+FrJmp2K zkJynFH1>>3zz}aqaVTp?kgkX!X1jLw5uB)2d%hrFtr4bI4|4lJ*kYj|6}N)55pfaQ z7Yr&5#LyNIWiceg75>)`@yu(YMI*UplV42F{yrKuqvhvP@Cx)ssgG_pp(<0160oQ( zuC!whVOl@hX_PvS_iKBqpB9xgnuQ^M(=hBmcz1Ht&C4z<-PwPY3zUD*9x;KhgWolK z`e7BThXH-_wV`+n0=?^dPqHS$FRj+y0`WFFZan36vB3s0)3-^h-{u?%U1%8lmq<@z zA%dHTq`%vc(GCq=jYvE)a~=6Th8M;pXUiZ05hBPOZ&%@yH*$P?>*bOU(iO?j&KwD7 z)q8oJw#mT@#grAIO#|+ZlQ@HlutZy_M#LJ7paezl7_x>c+y6xM=ed+sspcUPe@>}( zLEDXwly9{6l`1*St%jEGa|{b$WoB~0J=8sNYIt51?2@WI(d&2q25KDp#Sh%!vXtN z(b4oH!F{I2Zt=)E5=>~q5bxt za39Oz$0J1LZ}zgP;X*`Q4NCZYKfn1B#6vJmx<@+TPzd0P#T`xLK_1iY3d!PfR%;0y z-TT%&QVL&Hein-bHY!tTk)etF9@KoY5HX-#-PZddoK)}ybjd=4RceV1 zxV^N~{DOGj_+qCgqLyr`yZP{=p<6NYG-6ZeqLvsoKGuA8zcz*XA+T1{Z1u}(Es&Pu zTg7hFzMhsmB@W0}iKNnZmyz8o4;Q7*g!f%BXOik@cDdF|V@QOmSKSxG$zj)c}yqPe- zCMEP=IkDOO&_m#aXVTH~cmt{H@UHotZP(-=>%FogHL4-m=Pb);S&5 zt9>X%M&@O79RzK9rja7-lg4-r%-b%))Gsr|hboN#FA6{bYXKuMeSxS0ch!z}JWsHg z)|I+i>UFQEWz@{A6IW3dQgcd(p_{P@Z^!Y?q|8Y{+Ysvt^-u-G;j ze;IZtWw;n&$CG=Ku^(_ORs*Qdj@)Ht(=FxEb;;y$?21TwWC6AS35~sDw8^&wZX>?} zJgtn9p93N>yxDyNc%m2e)pE9rhqXyqG$xjy%!{!p0*3jw!$y~yPZY#u4TchZrVzk1 z-M(>n9bC<`$jABVhd?Ey%`{}sg+#TtSv?O!(4)+)=|jpnqYV+k_hPGG@ss7VsI{W> z;+H+~AStQnk@1d81D35cswbzEL{_$Q=+gR6{8baPIE~Muc%AP}5|T||he_~NTTAhH zCZ1;IAY)a}teuI7Q6p#OiAd^fAG|KqR~D*hrzk=0ze)T&r5GfZV$puX;7yX7Q7h{1 z1}aTt+y0&-6C9dv-mGe5=AT;Pd366{o5iSuFZ$~YQxSVg9D zeTP^UGDCc+mAW=`DaVMuhP;0zW2lx$K>i(z!KqwdM06Te#>F`DiuW$~o@qB+d zt6hsVcq(9wPf@x;R0I>QqO(K3aYDY+9eG>4mBKjgLY%=d7#Gxz3LjoT#R>+$H^7Cq zs9vm7gjI?jACbWgHETmx4=MF-!QYx({t8F%m;eBtWvY8@oEvF!Vm=)OR zy8&ylt{F%N*%YT=m*UhFB{#+{>LR>&5WxHZjB_@iX@y9ZrJs4dv?k zMt|**y=O(WKqHAw6)Br)r)x9nv+BnYGKO>KC|xSu|2Es~%aO>-u?f?N)?2lUW$(4i zx*#eDU6Zi16C+Ehb8^SnEpnY=<%4`vxSqZ!d^DlT3rXFLUeFba)DL1};+ed0R?nzg zg_bu>w2O8e8qk9)D6Ji2iXvqZ=;0DHr2~^ge*rqof~>x~8JqI~p1nxK#S+FKRF($An^g65E{xqnuqaI!nBimFwr$(CZQI5@wr$(CZQHi3{FPKvS!9t-*Yv74y_uSCo_D$k z^kX>0$J@Mm=r>-8k2X5q5~)VCSXw^LVW8NQf~yl?I))s)J!i%u?T(lFK9a~xr8v?6EQ~U zSgPl7YexGm|+9b{5R zKwE;XqgXn&Lq?YhwEm=oklAf(5R$n|Y2Iy#1b6~c+A7qsc(H#a2{|oLq*sYgn}nGe zl9x+XTN%|WRF-EW@cX0aEWwmS)~Ur1c}Vn7q*}6 zk85%Fd9$HchyItrM%v^NF#5A;k{qWsF!x0xdLoWJt5)@|;uQfVo8&Y6#XW;NB|tsj zn4I{o%KF7iFRt|t6ezSZUpabhOtIz;EEk?5_fEOGzl>B?AXG#xdXx< za+Ok@Z?-eJkei8kre}xI9F@f=WJHVUrPh=+AqIbC9y3o`A}bTUAO-v0wMJIH8tysn z`?b}`Bx9y}@$s5egpJ}&HCUo|LpI%;peN!ALe4C!1M5=yUJofEfNZN+TJ2v zS)3TgT1BFyuzXxwnv{*aAd-= z(!-gke-!FsOuFWw{b$J_bBerAiq0vms@o3LmDkvo@dcthdzT+neR9x>kIUi--=h5D zn^8%&@+}WsPLI?Qrug6kb@&k31cVJ;^ ztx4DL$%%F^P1~Jm=^i=z{N|~MnPUrHFH&jQGF(%`#n%OVdF|-Wh>vhJo=5X3aB_3H z&HOZK;l|}>fwrvHG$f;@S<%35#8%I6C~&)WuL<22Ns|w(%T%+%+LagmehUyf&}3tK zKNny(<%G20>~50@wFeahzF-VR+bWVBKWUT`#{ZE(R4dN#TV8yu68?Iv;P0Dux0_aW zUCF;9kxAge87JLclZ{_*)d;#kJ)IGkFaY9O+B_c2`90fx1)G~2?V2UH1)nF~xfA~j zM2UY8UcHuJxE(IfeKcw7ua;#fptcs~<2kXiEE#CSG#;K>z870Ly@>w8A5;VPyvp~{L$Pd9AQAMqDornC+*ri zwFI?qRwM~w$#k$&w!aO1Ejbik`EhFMrtIj^x0j#9T|AXXk=%KCLX26Mn?WBWJai9B z{6}_a)te)hI~!Gz6XwH`Qrp`nXJwEhooZ33kRSkR8PRZ^Ul?D+<|o7RJh{=otfL|2 z-_Jq+Ag5{ORk+9Iw{gh|P2h!Z%>{5|&Q%cIu2#cC;usEX%@K@_^X{}v-HiU^2RTh$ z!b|sOO1Y{;^RK(c58p3)1jfy*VS1D+ss#k}S|*_ScJ-%!zk=ksZu*|k*oXN1H7cP! z#esn}KGW!##k_u4<(3dGB-o&>6myI~w>V3|Mww#lh6kv|bMJUoiCdePtg^!reGK!^ zh6==;#sP%*KWRP-Kl1W5g7F?6@A|$pjGR2=B@5Aw3G0tJF2%(IP#9^@ft@2y~q}=E*Bb zqKk8{sM~9!tTtRYs%{EC0Tgwv>(<;=1#JIF=J2Ty!6d6hMIdd`Nv8o^c}-}!aH~S- z6h^`@(sUI#B|m9lyxXefeoJ@a`s`p=m~;%EKoe8HWI%3iPp%RBqr@UtHDQ&<?TBT6+D-etjm>DiBl>&R@JmGHzYi5`+V)XL z$YG*8Rl%Hu5_e9?GB8z5r9?zmUybD#JhJWq4;X6h*q6%7%_k{S)IRB(wjq8MVZL~{ z>Eq{6ov?Az$GvOuBr9rWJ#@FXjY~DY9NcoheAI&#bu)FcI6vlaIk0nk*=ApRQB;-5 zvBq>&~m{uXIekecl91o|uVoYhL#^8aPY=0wZC z=%475A1`XBjg@5_q)Wt6k1WG@XwF=}7+4l-`cbouS?kNg?^-4IOI8Z>ROlW9LHHLq zo8iMLtl^LA7DB)M0q4Q!KTt^)%qEVE4ap4O0z`8%nCc{`^@tUZI=AlL1sc;o-Q_X_ zCxnW2-<5wPekKkd%Z|l{>0vk(i0F&?{#$`9D{L5YRT)wjNsdkr5_Duy% z!t9LCk2>Tg+L`^q;C~7K#MggGAwa6N!c~kL-$Jte;O{fg2qBNhTR^+zYSlir5{EF? z=%|%|YA!;X>tX*u2z+4m6pW>03SsegUYn&zl?LIm(I1|~V=VO3Xf`hsny=SSVyN;f z%7B3P^~vs=o>*us1h~*jl`q@YQ z&3*3e+i~tyiMS~%p$0D{{d|%K)_M=>5bg5oi$Qz3qhsmGSmHtIcqvUozOBaSI4%IZ zueHmvf|Dgnf|ZjLK>ohtw6U^G2x(aHOxhe5RKrl@IP z!qT%M*5cJtSKeTi0@T&KUxL%FFNzYxJZ`ysyw6z!D3C|3~l_TwEZA)u@Udy_@VW@)FsC2TGU1HED@d*+V4UL_++an z8gwy!@ru)=e-xiI@T>B=4i#QXCI6mv3&+P|kJQ^320qeq@M;`AVfvb5wOLIdG2NV) z6Y3JT`GkdOnm*e0!VjYj>P_HJ{g0&2++~6PJ+p{Gwnpm25K5D?X4EKHMzVBnH1dIw z`DWHjz^y+DkCjQH0id2ai?0&~DI1C+PRIi3{v?y$zNPHI%k2>!+ikk=oUC=1Y8__F z13Z2aJkFT=g|s;e0DZ2)v=jq-26?NZDV#@8s0dYd6y~6InW;GR zxfLqHj!Djx3b`NhJDKg&i@#8Cz)=B&cbpxO(buybykL{{O%5e!GSza_snThs2c=W9 z0ilfGdy~6u$GjDEu)t&B3xt8=tv45WMauSCUzkIbwy+ZxVvIB9rMv@&OAIg#Am>}8 zIf{!Ri5(3h5Pe{hv7sBc+vf*f#QY|W0z;mPe?%k#A)&9DO)i>*{zS^*3lQ(3_NFP_ zoETEIr_Aohql3SpS0!GZL<*Zf0QlNYQKiX&%q$5CT3Xk}#he8Z_Nm!4%((vWhT5Jo zCrM_Ov}aXoPy3?YChWSY-fheOwU2_pxU;d_Tb4&&1l7M?!iUB}<6(S48|4O(m&O)? z0oI6lF4;H8KSrq@$u7oyJi;U%s$YiGHgvjeM?D$nf=AD@$?;aThc?Z9Cgs(CHHo0a z;D|J5@foKdfa2nZ7{-oVr<&+n{qXEt4?B7nq`P^s(rqMWe87uO7KCT{nxyC752=30 zC7LFRD#|wpfB_aRSzheq?q&|i$zbZxC=$gh|2t-9c4*}_DsMgFGBAwkpwgq$;+ z8+?&-@+I%^-UBDpSz6I)V2z}t$d69pDV7PN7=m@84HG~d1i4o>?qNj?$Tf{Ej)7=k zFGVa5%7_8G{#iE3V%#EkIo&nF)L%cKWOdH^+ejU)AVM55*Z~cRJ~x}(>*G}c&4q2~3&LmIv{CaT9I^H{E5z)4*OTeg z_iY#=J^E+KZ2Jz-D%D!C8NtgRhs_>3Bj}v6N78K|KgbdcIxPqoD03J{1cdK{{iW>t z;|$1-(=@G(bPj2jiK5El%R@$A1TCqjrZ(Xna7#D&CkNrIJ+JwH4Dod zM~;RBo*;X_b)Xk76O5`3se}Ym%op}&yOfMO!U!%Btn!VZ}G4i zJVzmjz6qw41GnOkdH!4{=B%Ft?o~8o8O4w?d9Jy8iCqi9{6TGpCzAh{PyT=D zL#+P~eTa#fmGyrcLrerr9L$XWXZwHaLrk3PjLiSP`cSK>iq2LmYa}-aEK;$9``SK; zyE|nzRq;K*0B?Jjv;)K~-nMSi{%vZ%_-n`8wePI2zkZb%s;@GPhsOGPyFz&dX?9>K0ycxhA_@=+z`am4 z;N)ri+VUC%8=D`BRR9bEh>M@A7Qi$QV5C5JMQ9muJg$KGC%OV)mq7ZAjIQ2}Kzs@; zgv4Ja5un?%OKU@8=y%1y#PoOBVxEm7T0nM12k!dL$SjxvR3pH614Sc1a{q0e7&ZVd z1_%SNR@SCAP=6`_Y(AWTvaXaYEfS9uG(OerC!t>-_BxHW^!{y)78SA}#XJK+^dT4DV3inkWE5JISoq#!d7QgLZ z6!r`?bN}A5o-%FD0GvN7z`Uj~K`c;p_3iB~jan@o0$P0&YE;pW{x?gttq#DgfjM{o zRUPv@-g_EDEkTqzC*f+pKk9oEB>*V4Gr>6MJGms*H+|VqRdStVo!fX{L=P0;hw{?z z|I&x}H#M?4zf-F63kv}p7+L@Y0|=DkKdcFcQ4TMG7(S+6Kf#;xf6xU91O`_&KIGv? z{#-f#x@P{A(82p^r~7xe#+HBj+^12FFArbfZGW9NP7UItVxWgl^icx=q_MsC@n7lu zU#G#I(_|H8B-ON}bb~YxT<-;@Blp!^1!xf9$MqNdFC!$WAOPOiJpiu1bpTq|2vdMG zB?549^fbHjzZ2vaPgThw);E$~{-#@SA+F&bfBW2|vExHd@K0!Ua5Y&4+2m>kG$;Nd z{8$UVZJU6E0BruJJY;A$^I5i|zw|gZ{X6ze6%^M-ypCvKX>|wm+|U5h`tUuSX&Q2b>t`NuuGJTSWZ#831`Erk6ZMxBGS zzA>=;+vmeZx5$Pk0F8-Y5m3kUjeekh^EVG#PgQkc5mCkXo0suHEdyl&*IEy# z3CPjz2~avaTbfH@Qy0`fFavvU=AUB;oj<0Z1Y}?e7)V%02_UW>4&RS-9sOv|?%57l zZR9WWC)5p~7SV@@AE5fU7oG;&PxZs@HIwv*{Q#hW`YTqait_j0lKdO<&H+$ur@`c?amidS6*!MX-@LJdTR~Hg>6*n?)aOdNe z%TM`-H|L+Vp|@u706z|A0Qw8u>(ut$HAn3(josGwv+ciapnm30(jBVt1Kj&q|C_Z^ z^=~ydc2!`WpGeP^?JuBCy5SGphxF#2-Nc{t%bo76N+5r2;xp_oh4=l=t!`rf&!KP8 z^?U25GbmQT9Q%H6>ddF)>KMwe2W#e6@ZZnMXB=r`ACuch*CaJ0y0sN_g9&QKL4RV^83V6@$H%U z?W)-LcyKejq{BI;8wfF=!crSb;j>B^Y=3RbrtH{1OAKt=eZ3KAON4wIr#{* zuEQ|SR>D|3tzKzRiIba|chY`X6CSIBv+eY$^2H4_v3-Zq{}%E)lgDHMH2{SV0^^(X zEjYrpEa>q%hfN!ilC*|aS;#Nri|V^)C=n5JYSCZ;Wop~7$+Rd{+Dw6RU zNy|K)c^uY_nbGt<-v~CNh z2TwF48Z9vE{fq;nnL9tMB&gPIqr<*v4!P4$7|H?neIOElh$y;2btP2nB@G;n&1Yf_ zec-q9@?AeVUB+NCyc`ANF*|2GM7HiW|&N zjAKS#XSIq+q)C~$^j%j5i(YOsbiO;|bA|O~uL$-O*6}48{!#ZgOttTvZdS7jLg+jc zI~a5HKG7~EAcaZm!h*)hJ3$kMugCnDw|!5Gn3#Cs{CJnbReo2|fzbk8G1s1|)&c9N zbs|T|e$up^(#=fC+c#e+%htbXOdm6Pt9M;z$2QUa&n$)u@;lUYtX70VI(R11Yp=8^ zpDV&?)*6|#=v^pWwim_BANcCrNz79ck=vyONczD+_Z{o}Lyyql>>j5GvsMQ@9T$u+ zCX@L8{3-dF&^_a-C^*o!V4DaZUu!oyFk_xTUFUff055S6~1N*C>;2HV;Z z?MpRjz@^IH;PONY&89c)vZlcgWkq2|!jf=r$D544SeU{(Mq-T(WX9(K)0ZgtNcj)% zE{i)E*oEzs2hOnF0^cEu*~f&qDDnSfqU2$(6Z77YY1Uqw!4)wHNfF8P3iPchm5!_r zwfQAQp4$UYd?Cf`y}xbZ;g8t#zG96JS#JVE~))g?~!eUgvV|sz5~`OnLTd zX)^%A<_yWg2GB5CZSPV_lOTx5H1YJFQ_3i#4unwF?=m(u8VTRX&{emxK=@Lo+^LCEmv(g<9WJzJaVfHJh?4m- z^f-e&Hz&^ExMi>Ra!BhAGR$DU*%)_ZUi@~fIU-GG<{yk($d1UpOMqu-?})hrlnTv>RLj^Z%*;&2Gbg~7Zx>CYEGS=~B4 zXfAOG5wTEH*6$u4)-W4XhT)$CTl93V;@MuaAvn~dj-rEGq7*xt$7E*7@M9_>E~;!l z2~=0`WY2uxS<$fx=44d0WghYF+6Z9y677C{a}spYRMxN02r8(zxJAiO!&IN{XGEQ$ zJM(XV(H*QepQP$55%C>&Rd(TL{bPvdbW!jYz@A?V_*wpiw9ovTGEfTB<*F@OT9Z<> zgKV`2m7Kj@FI__ZAGeauYqQM4LSFj0LQ#U}>{#@rk!^aA2zQq+$ERc8H zyIe=~O-$!o*H97Fg}ZVJV)0hcRJD@2*tr8kf%^hO#pVpF=!u;DL#ajhF2;98tssW; z_e8glr58;qhzP+Gw}hy-rwkVn1SxYVEwB^a#aVf5nsg%b8c7|efO(DchwRgP;4!{G zoZX%NIJq<%@!rHY$nH5aB|o= zIxX;2`NG$BX}QzL#iEw67M1%Of+}CZ{rON*>+sj9S6PhQT@vL@Um*sII`tP}x?jWJ}gqgbFrqU!h4IOQw76S@m10UFrdtI z&urF8yV_9;C|r>5LnCkNLsdkaO;jIaAVo1$;Fr})qy>Y(OKiZ~6V5FQ&~+J!%|?RU zmH1)X=d99D*+C3o&fH!f12;fEs1${Ly6vEpqi698`fld|kH z&sVlrf)E|jjs)iZnjwKz%#BfMccRic&y%-16yFcdZ?6tdwjkO-8M4iQn-`lALh3OU z=sUJ;YH2Soa`7c$V|;-(yb_Pl=D8%nnvak%O7b4C0Z2zEEOW zMBqBmu8)s_n-+SW`?1_X_!dmCw+#esn)>(~4kN~&So9Rl#o0R@=vM;7{Xs1%Ss#+7 zhaAcW$MZVySEK44x(1h4jI8TOc)>6uYq(U!HJ_pV*t~Lm+e3`G-lYM;8V|<9?!Uxw zpEM_Kz^;iYiKNE+^RjDDms#0n=~+5*+` zPsDaFzm}&0f?W5N3`H6v+s#l{BhRyX1Uuq{5;35%mzn5hBK4Z2#is# z-F-pz^+hkhe^3iHf^hdXG4!F8@lBh{SPwT)ll5@VL_GHGxDcB*Am=1c_F2-<4J;50 zJv4tkZ%dx>t-w1idVD5#>nVpQ(52LI(-27!1lL1p(MDVGsp^0Csw`35C#{7`o*dz7 z{x3(#!^6LJBiP-$+}g?LXQyWIwnV{tRcL+U!m#u9Z|}Ahv;9FtcdsNQ1$U zl0rsMvPicIV>h(ha%9eAp5<6&da19}P5kvm!)-NbL&oHbu*K15%chnrA z6^b~sJDUWsG4ycS+fq{hQdKqJYV&uCu&;6 zAS+EFA`5Ym)wn)JR9#CAW7m2|TOTti(q{zgn;FA;Y+}IL! zsIoB!Qale<=c7(h+QU>pw;jn)70l;-mSuWV_NJUG6XMk|bXJb!b2(_*YSy>G!m6Gq zl}Z0@KtPT#>&Siv@`7|g-1lHtF6#dvS`dEhDPn#g{`nTA9O$zWXSqQ1&(vruo%XTLV%S2Vvx(Rw1hpEZqiwydlWax^rkzk%Wn4HJdgHE zk7kY?i|I-ghp9C7xMY)v)i1%{iaT_F16dkjFbSD;Sv@dQh)F=sy{QHM4nLunrU;7 zX1lIOOYwiSFC&wY90cHu+bjJre9ua+qnmxF>U4zs&g!hs-OC$myy49l1o)VQ2E)2? z8+{U6-nDYh+bC_@sWSwTq28}kT~>jx^TNOaefsqV-3&J1Un1S6$RwjHnk?B+oxg?g z|E2~0{jvGJHAL~RPPh2@TR!I-A^u||B3DNH_WjQU?G8a_BDePek4a^Ce_XkjEx7*9 z`gng&;8J@OYrRu`dUK!9?ipask!%9k+vCh>StM{>oeQua<~MfA`HZpozEjz>r9!Qb zxD0GPGKePXx(p z6*w^}FF$)RMNln*vF`+Ya}8RY^9{80NL{_fV|mOsJP$oUyQEJF-P2IXKm=3W{i!%i z)-dk4>}b0RpU+?pPQJV3QJrL#$u*$ zcP6e@N)ePfe|~=!KRo-^I>;KX(Z--}=eHsHHj)6)h=_jE=XEtiNZk#vW3F9V<9GDD5L}3D&CT4&^MeZkLG<<#W|o+jI#D>jEhYz zxFgt|?I@UMIE;O=j78E)G5n-hx+GDgontlKdML|m%XUrm9ng0g+5R={*i9(9_O+xX znX}Cro0v?q;`1~f*q67p9s6WmR~dbHG&+A|IJHyawJrdbi3;!ZX!_6fz1<{?bi5<| z(6<`NlgqROk4ny9<8Q2&J|dlft2w55US4D;af1IAX%GMTo+9AK(nYM(c_l1}q1#f}#p14tB{dC)| zJDn=WT0fSquUd|ne-D^<;wsYK8-_zBHolO~PmrFB()i296mweceZqdI=voUi2nI+^ z0k%LJHB4)qYmP$xh?Gd-(@;?)amHd!s!FMif+Y%JOdD>M`iFUla&cCh{G|Xq*O)UR z4wgv)m{k0&g9yJ)Bn?(OoW61^Wb+*~1@f_u)iI1!;@v%#*2dHr_E+!pKQ+)1C8j*N zk~>yuZ-cvN=KPQ6a zb^LjG1;ev>r2vvYyXh}zgTJ-r?dB18ua!_H+oOH z@dVdG12GZafCQR_gJa&33+o80Q$0UHmoT^6&YN51Tvf{~%2*pGhC!}SS{UoDoya*X zkvvavmk&*WJb5!Wm(QVh?B-gQzO3hwr8fK54|U;oz68e6)XRX%dolGX>AY+QzRo|S zNy=Vk*fz&hVw@-@{xeE+(r@Z3)eGOSlYx+dK>HEW1zb?k>xZo@ta13uqKcsMbn_J- z(+tZNp%zFTIm1gFI3ZadP3;g#pD996jcXfkO=O-i5r?8%ZaKV1hZSo%VH_zlYN z5pr7=bma<~@Qj7bWs@~96g7qfEF@pZz1STeGt>L-un~2w9GttIBlm4%cjbP%hO7-a>Z;R~K2OFkdbFa9pA)mMaeuLf$| z4*zZi`5;15ass`s+gwvcw}pxZ*iE;ee3&hi>DCLhtsM*^p;K-Zcg7u=ZXVhqKI!{_ zn5BFyZBr)}#dA1xZw6VVJBf1f76v+h^d+!-J@KG0%BR1d)MzUr+E?iE%x#9n-SN0k z#!~0Z^+ZR#kl>>RyVtQNOuXfMG7oHB#}_7%PDT`^&@OG?f9c#?ocSX_^cj*E&gkiz zZ{DDbi;IRMXLZMFgmS&Zh4!-^*ZRSH)8(_}Q%nG*&G$$>&Dk=+hQfBu=WL9tlm^g%OR`PFU)`|WJYtv+D3kOK(+ceZj zZ#XPZERBkmIDJ3bM6O+5wpE721-N&bzpLj5?iGciWRn7Pq|xL{=DOYrpZ|I{V~}>! zqd7iJ|5{;rUt%G0tC;3q0NtZR)DqEEs108%EdJl4BR@jaE1a7*4QP>YiTIWQZgpF2&rr&u^nidL&^U3Z&wZ!;^b4)J(xLM~#n zmXPQn=kSHo-aPD1=H&ppLY4Ca;R|&q?gC)*!5?|zMUO~$xhh`Ii~9Ty@_bcEP)Z{-`&c% zOl~qkny#_BYl6uw^9~pE+R|A!5Am78;g!+>@3IpVe<-J4Sr(29{IHI%ZpN-D1b?yo z{hV_ts6L`|pU;{yJ~194lMj{(sxasshTMqRoXQ%{H4kib_|h;QsU_WV0|%*u>vLHn z)f;#kzQTt13WKk8rbLa9b<93rgWYQDj7>)t2lPL;K%87?GJZc{H2Mc=l?x9~{aL=U z2ppRVRrFN=Nj$@UgXlJQ@QmUp{KJ}?0L-JFc72fs5&o@J%%BbgVXXuKh-tSPU&aOS zR>_twT;IM9ABLvR7rnmPSDOd<*;1!uglXu%U~6M8I=_fy1nm!%t69m7YjWF7 zb7L)DqPq?sAj7Bv`t(hOG1oy;W*i0}>MO>qqUEKq&S+p5b zLq$(g;25q^N@G}Oiar7!=i;y5r3HLCa83!%jm-_kHp)R!Ydp650aM}=VZM!=J{9r} zJLOs^t!j%j|Hw8h!o#cYMEi$OpPD;r-^z3Er1}3eT;aVdqgq*y;1qPGrXEsvFx?S& zFVp^VZ3TWAD1bvh@Vd9hSO|%|^z1r&I1`#V2X!g3)q@ zBdZ5ElwjME8;97KZCp|e&BfAxaWgX_O&@XWNSixqn`^^xAXg{`4?qaLNmI0y#XRus z^I$p4T~8HJB*v^1&riVjYO+_bpUeRhmufsctbt7GXPQD++Y~*(lD=aQVYmc~U#)g>1vEsNx*B_Fu zM-JHoiF}*XuKj)_X?I@Up!BmG>vi2msH)ssnB39n-FiZR`-Tak`RbAGg<$0=^dPb@ zoD7GP?hz0#VY?qS*SjLY{oI4WH2I9#&dDLMu%9dw+r~GS@cV8-e1V1#PP7W&RMqm8 z0Sw+ZWWVowSm8aej`hFxI5Mzwy!zcg2$nMTz|>)hHr%A{FXoJoKS7=-zv+%FHk=o7 z$Jrg&Wy1M#@OcgUXnakV2ek-?wI6y-!pW)|7?s7biuyE1{AL$f`^vm0GcOpiU_Ev* zADkI^Ca^~LvB=A1=-g177Ago0nxCv_)H zY8C~~vY?Pzj$-Qr)p`_RA&He+mLcSPzkf@hoHtzM@rGP~E&$a^#iq6Z@fThpHdCg> zpy#vesqCnO`dWOYvQ3lb$o^vLnXJ~~7Qxl}i2Z~xpK7!XTt_!?eGqLgKD@a@xQ$6X zpOrs*k%Rv3$rTGf{3m&oKk^s52dIxB1!D$Hef#xFI_Jl{e<{hlf-Tc~#OoQPYm%@a zCIPbtdTnNr^5m|}-lZ7l^X#CE@$9ir%tgu=Gn<_0jhv*)@%(_+lYLvOUfCMZx=_ zuJ9HQmTy7`x97syD<0F*vo<_!JR;)Wc*H8D0?KTjc47+MZ!JE0Dy8MjU{vbzHZ*Q( z<2T`K&B)@e3xxsow0Ojcc=%Jw50i}Ef&U!0ADA$)A`?wRJ$2}joGp!-x#b2%e76z;ifQa)Y~T z;t~&XH3Pfrfz#R$B(tR2Ps{992bIyq#&AB#?V@DHMQzeK7*ExY%l^o8iB#wuG*{9b z25yYAM&gNbC(#xA6fd7i)%IYA&rMm&dzBlo#WS`fVp<1k9CxoRU-*J4aJ_6uX20J2 z^z-$-HE*PQ&%~5f5(f}ccQp9kcVYCp;THk+ro7T!hVeat-qm!%Bku<3aV^no2x#8U zu7;D+`StItjJZLbG`u6#J3TAOLM%ocQvpfaA-Cg$cEm7p`r%+_wc3VcD*jGkCRy$a zx#PaP6?9F63_d3mqlR9(`@sbwy3XwUmAb884S4k<8feZY{8oOOTs( z^M~S%%|WLWU1T;nNdg;BhMYr2-bZu{MzX9WfD{bdYjSu`iY)4e(X|@O&-L1v9f+M< z!+wre^jw_t_`a8sO&6VA{40ubOQK`vy9FknT7tJ{kOppf%bbfL-mkyE0d?=UI?((n zLqoRiEThxJj)3L&a4qsuLLN*+T9R;2*d174U=zn1L>MvI%vP86WU&QGgmvK%-nOsC zqduJTGAR+=n!8A$XZ*0_LIH*>)&07d(iR5mrItf4hCKj}3pHKROaHWZjO6F+vDKPr zjsmqAmF-XzIRfLbmE2`cA}l3ZxtmJvgc<9&@fC=E$7u_QG0jUDT+qiQ0)E~G5rbJZ zs-r86p9dw+i&NV**vLnrc^P1FXMW3XRVq772=GF~z1nIc!!hFyOAuF`ot#r(q zeI7B}!#n5_tE#o-ld22@d|uTgzaFiGArQM@+{`@a-74It397V&gzHj9>%Td41|A1tR{Twbot^I2c!r&ro= zh1j1TVyh1i0kc(-*AW8M-w1?M2uJ|wl>3KvcJ{YC2zfXRUs*}emdtStiM^61E zL)5B0(Jj^L4N8G^tM*5r^PzCcMTDN0-@l3Ed^V*?1;nMKX1W3jtdC$A*|nrI1-HA_ z{9?$G8T9cC?zJxL0J5-oZk6FVu)Mw!tq2jM4^$Y%S{P4l6fo$o`f?yZ;bW`&ywI5{ zEjnC=%SEP7T8<1|>MBUlXr^^P;~Z)(zqa1=FRqD5Yo7|M0J=Gcj&1NW>-JGTEKjN{ zT9nJ1yzv*@%M%}5U_dx9_=`3wr0AU|OO*_h`Fj+MnQ{&DudX|CnmF z%Ttah%-d1GgT74l0@mX;Xx@ZlfZL;BWsiEVq+pE@g0n;0H?gHG16 zJ1j;s{2`FIQwDkZ_<<@fDb$S-(?Y#O@0L@D_u`}8#vt@GnlOQ!%@>t?0GR}DvMrDnCshM_?p_ZWT|kk z74Plrtdi9r`B=j6xlbTR-%HSNMoxM;aXzan;gC{LIyLe7i*WS-G(_EU@$EV8W0j#2 zKSG&1Ot05B9rNJwrkO#|MNg)rERoa-yL3qU6Fz9M(k7=Jb+IReKQoxx`e^@Z{gird z%U&nqxVLcTlI*_8cjfmig6!XI($yrk8Tt75U2_)vs!-6!_zr;g zsovFQSx^`4nZPWL_#}^s=c_7Z^ZRM1jP_gB5c?skeHklUws5I8cM=@p=An{UE@)o^%gdlv__K;z4z~K{RcZGYIIJiG!oRl zz&;vRX`zp07J!^%ETa+^WhDNg;O}N<#YX59r-5$l@Rb$eH~lNvJdaM2rOo%HTG!KM zCbf>z>!F%yoFJiAbxIDigx4d`V&rhI#XRjH`toS=S>FY!E`Xd=Y$Y>#p47uz*ob@l zduic58d6j8eBr$W!)(ib_lhabuzM!`H^R$VPp=YGv>APv@}uHwZ~v`+vB^JaD&SI&>OjS1k%WEkO4N(GR-M_ zHR(HKT@SS7;zH1A7({l@uif7D^oPWMQh+77aBfdigMPh=?BtrJB4T*od~x?tD>8+^ z&JaACFtbMHmXtu+I1;Ba^PDA$MwuehA~eyskGDV4`+l>_tvTwhYlKxwe9M_@=W)i& z_*f(E`y14BL?V3Q_@W@TRe1u!m=FwUxL@igk&sG}^9jxE^qB4DpVF9Xv=>JsnZ>H# z!_go`R?}O2p#?kovgIDy<-oRx2#wFCJF^gTew2>p?y0fNC+z-cu+^@GUkF}HZ%5x# zXS9>vUGZCb29&x9;oYfYp*Z^MNfF;5RpyT*LB|j?GT<}K-*T(;k~^A^+N5w{&1=cS zcjQQ~acACF0BNh(yD~KZfV)^`<^TiPi>8CCOO3cVx|689``w>mC-|1BZh4k^YISAY z`?vTk|`J?S=}>s95LJ!7PFm; z*|*jM!QZE0KYzTGgEDd!2@}2BtBuO^v2Dl&^r>55xaO@5&0yyvPqNa>Le3+-P}%cC z2zYJymxhB+O*(cT>PgPSW^mup@BmnAaG=u|+5gdZ5jd6OLJjMDk*4LPRBb^S9VGylQ4%MfTXNdMI_tzt)O z`Y|H;51wFlKuYHUo(?t9aii^!Zv1tYo=^{4c!NFfjnA)&E3rK+t#;%dQlC48kjMZ4 z1B|^ED26tbGVMD%Bw(sI!NsTAy6oftRD0y;@B&8b0ZF|Y3Wr<-Im3o}a95|tdUAUr z?~6S?E)uu{W8>QMu(OyUti0EOgND^F_a+?fJw>ku8TC-TP?9EI6;&m)z zs*y8nmI03Kh9X@a+c6oP?)iCf?8C)L$GbPe4Oo#B`6!2w6z3PHF@%m>7TnP7Rf$YX z5CfOkI~L4|W{ zF2GFvVm=hK+%{F3O1;G+tYZgbx0MNsvHRk`2&;_m;xAhRi(BRTE++6UrqYDDz~X#v z3H8_bXdSI&Vc*)MZrw5se;(!&?4*FOpT_wccre4KGFcAYvF7vHrc3%oxzNRZ&Wqv5 z%&}2IqW-tbZ`ji0xcz%{rFQ%7aPs>uA2Uhl|rgL1%(u@7!%czHq;1)ZnJr>M!xBUDSG3&**hanqdL& zNCeo6*`ne!-}n@GuNjb+8yHeeS@@_iWjc2*_f`=C(QVsK@>QNESk@$sUOOiBy7>)4 z#7PZ1v2)&snKPu(DvUyyQvf7p_(!P|UDpu>Ar(2adzV#r7;XZLL$$tM6qtb%I6!05 zW%N_USX0n{BLI&g{2Xvk|CL^;NnZ*EcqN8LJPx{8wJ&^H{tZg)-zUY{#Jm~Us4rp}GOw)xVejcJo7fh=B4jsDy;e~)6cnBJ2mh*OYdk>^0C zYbfZu3}6V!XGZ8^NtJcZMK*W!8sf`#ABaV7``~!0b+Ws4Bu1A}pg6noo#mRXWK+L? z1Tn?^Nu8);v9xuGSVdBZGQ5NS%3d2CW%N#VB|rKf075{$zeVuoYT@2({s)45GbwNq z6J|hvy5EpVo8HY8aak}5T^EFzeVb@GGHon&SB`ShgIBFx%z+leRf>o^GiTN_7*+Bc zNOWNpE}tR+XH?C-TU1%iuK$_jjNrMz`j#vFK3Q0*r1eBhHHRcll@@dH-cW~oPr|sO zMq{<_0P_>g%%}uZ;Aar;Jn#W5oX!=BXrcBFd>kl7Tj%j;k!49!+oA2txA8H~hjnpN z?=ENsnRiGu5Mp0es_FAfTYwtp*MpoTfjzz}w3N{7Gj{-@@DdnKBCuMwo*?gbvl9)b z>2ntRco3*VW-+qk3tNe4wy<*?^+uPMNV|*ZbZ$b$Jr;qT%@IMX;!Kh?D+n_KcBJy6 z2l*@NbQl{&$+hmJ%L8PBZyCo}*7Ms6XY^ibvvH8FI1)8pa1X}RlDGxoF3dISyfwjKiC3QCL6vqeSM zY&XL_dxxCk%3`IXAnJjCT_ATH>5C@~HoBKfLWh^2Fj7X;r@;`&R>2BEk z4FeNtTu&Qmb=u{P-hP{btbF|~>+-2M4$3&kvz@3C2C1TE0*>7>m`zzDUQY_vhahl} z^qf=Lf2?9)is~u>dA-UVQ|3K)qCJQW#sWMJ?GjPZQmNKHwJ<#867XT#oWI#U{x_Tf zT;6D-eTm1_~uf=4%@ z)C(n6_THP{xex{nXCFV?EoC*OsoVU9b@JH$(nfP zS?~9~`Zf%@&j@0`UZ4uocX`sumU8bL3?+MMyY3qMlAzRlx-0TKsaoq3=tztOFz25{=Q9-!8bm}1_}Ihg6T z`3GCr8Jr+cnm9;bH|(EY*fa%y5z)*^vwYfM)GV!mCHbk@*(`jYLOjN{JH2rBK0q?X z-vS}INm*=Y{s)tqaj7Ed0b9zNIN#u}-)5AaunbK3JHJZgP)=*QvH_`?*m4>}neMQR z!MLKqbS;k6%i1IU%DE?9?j=ZHWp=AIoR{4L(RSa6eq>fG_*}?wzE`|NK}OlsCv>09 zr1Ct!&uVf8nOUq`-%~hxDYY`E z%JraQ6mzem)cP^dZBe?XYCvv-Dg}m!5hq%dRa@D2cW|e|WL9>JU(Tzg#p3#H7??AM zquSYYI~&!b{wO#ifstlHi*>>+3+TS(sUlvqCfq8YR*M%So$%C$Bl5obNi)nxYktU= zSC{zSI!ajmk|8g>H($YtH?RdxSI066?6*Qt!iL%D{@3JVxTCuDBvz zfeeo)otg_3MiH8#X%`ym&COHKx`+${DsX6`3XQmhjg>e(@RX;%rI1PWSPF|$B-GNH zR!*C;ol*&Fk$q_WQLalegV9VY)9**H7L2P4#b(UsF9E)Coq87#ou1`FqVl+NMRB+x zx~HqDx*(y{O@J|la@MNH;GJ=-ulO{yl>YX>)T*M=2WDP(G++V)4%%K|2O)J(gbUtR zV^e4&^~V7wN1s!$6T~e!xWnG^eQZ@8nqZ>ilq$Reklu^*_3g1%h&TZmImTg5!IC^e z&+>DMiR#3D;u5mhS2oY;ACnuEBiw|J95Y42EZg5}(9}0m!WuN=QYho%(@CUW9spCr z3l086*D=5XJJ_@Wnyn&l8-3B-km%4*qQW&p9?@!k+0704m-PGxNH)J*qWn=OVBS+- zY}hbNz&pTz$>h_xH<}UHPe{VMSTD2q#m(IsRUwJ8VFvZmhx3w4T*`4EkQQ0wUDwF_ zuLLUgrKcA>-7Lm~z^n|eKJ^6IAC*zG1!@v`NBX%E#BmX!sdRXVNt28v(Q?og9s2m0 z2Jcx*%a$Z_GuQF6>kh3wRQe-=MBnh-65ll^p2=FyvnjuoM-`71KUYIaj|W zi|a0mtg8v|X%k-LnXVh3G<~@pd`WTT;nx%DRty90qhd&U#dTG2u{(AFX~5*#W{o9N zSSma9F%6H}b_l_{-k-!O0^h@*7y9qR`BUt*PChEY6Di0nfT{+d3K(OhP_0fRFIm;E z?Qm*i`e_04T2)Y3k2kXTXehB^A!P;Z%Ls3xZ}=ly&Zvk3H>v29En;f!%n+<>%(07- z$dl+RAZ5v&P^Dv}$Zx>pj3xYjIN=f!+HqV-ixIi5KPNN$UxYc0xOFvDEIK_67=8L3 zH@_F?P!*1d^kky)n!Lw#2uIhs-+E6LFt*tmYUBcBn_tE9U-x3Zspm&!B|Ma8`VcsS zjbm8-46Rz6AcMxR;xlNi%|f5dCO2=5_bbT5tA4Ky+~$p4#t5VF z8@^j6erR@vU8q~Nk!R9<-{&zfju)ZA6d3kFpvkCw<39KNC>Gx9DB~#YEM7beDbOvr zn>c~rJGBu%naj;|#<9mtt;0(FV`t3>UEWUWQQC>DRfAy~)zXCm0?a#6A*{7uL;S=8 zPrkqM^xAtHY7I*+&@pk|eo0otPkKm@E1xP7;6vFYbuTwd^nK2RZG`8wtpjQh^92DH zDr+{jY#J_dZtS(>J$VN$6-Wwe|zO``z7AS`Q^{P)iq;-gSIG;WWc9^IvE);)d2nWG~+|iXx3~q960% zae7*^ygxVI-4m~)<eREML@MT6vxp5j8gvTPa7@eR!2 zP-vP;ik_X{wr~A1yvwH3F~onG>$+Xkmsn^VDS})$Skd>5)2ZY4{A%h5SI#VP_iAo% zsBko96Ojb=p;LrTw9a#If~kpl$PjftMtB_3RNXb`>v4hYZ=nF%sQ7>J^GO6k6PprA=!JbMXllVNEUlJ5k zXI2#D^fWxmJs$GYR*xGF!pz>?n8kdH?P0?VWW6F|Pz^#M%aDnK#>P9spknvyjsSV= zQLCV8>LV6s8LI+t@=Xp5hCRm)SyrrCtKCO$XPUC>MAOelD30)TEhH6rkPWc2kjWj= zS@^QLgt>2-V(PCuPbBx>5cHG39xAG0$@JfH*hY`6!E3}zK-lqOF4Se+%g3gIo@sRw z*&{P;H_VE#xmuS`Qzuw@rulA&zJdYKM(jrc3yBmnN)3gj7uyR>d}_vL;XBSuxJ?_2 z2j%J1?biG}Y+HE`mt@#Nz98Vue2%;}`pHE+AMnudSs6L8pCT(D6#!64k- zI{9wq)J9I8B^K(2KTYRN7r?^qq~yN1xOL}DX}uj%_QLT@4IrO8%?qE%=g7=h^y zKIw}{qj z_0`PVa4-`EsSg?L2Ak3=jQm8OLoWw+&_fEeOawy2m)i2ado||&_M1qWMCsJDBOpC1 zH_aDaq_yE*XmuMzy!>o{E5ELtN`}i5B`H(%^~o~mt}e+`TcA|ikx|>o4J;?9$5@2N zeb4AAw#=)e<1BG&wUY_#kHdVe%{#nNl8B&gnPLI3Rdd19fh8hvy(oyS3|R#;#?@;(pPE(V=8{Pa6N3C>MH}I zS2dg>BZrj@uNcbX*HzzaMg`TM*l%EOQV1Jq4-k=K=jQiZ+bdUj`V=q34q@_d+hpzJ zn#HQH2;Z6LpdsLa&=Mz*m@?Asz3##&M4q6Rth6{}mUz3eOofw|Y~MBbx3G*rMNe>^ z$~*Pq$c4dlOvJGn)u_o9UGCFEraE8|#OC8rZN!zvXS&7%i6J{QRBQ(xKE#bYX45}s zBg2+mxzsWoC_im*og{h4+urNd1atJ?7LYB3SLxMDm* zq01p152rxWA6eJ4s^t8=f$nhr*L4c8Y8&et&1GY%h@l_X`EPQjvMenlf9Rt3R-T+< z2J>u<2zq%iU&F;1c9>MDsWLDn=k-i&MpCsT0RCSa=hT}E!!YaGT(xc6wr$(CS8eyM zZQHhOyIr--{Ukfz+4~dbWHMJWGwtMIzu3m*7QQ};5Eu1#b9|~ib%HAUWu%v68igUAN zber!2nnB~;?7dOcwQZacE2MOc4;redl=jh!7^KRD#{A@84CC9-2*15uBf5?Gmxv`R ztgA&%;IFMesn|}+?DeAmP~x4=yk2gEy@WD)*=>o?a=-Xr<=u_0Gr!17PT+)L-@AO9ei1G?rv@UtMxku~0)cnPFzWu9m&s2U_{8C*TYk4G~V|fy!_El9F{NO5v zISy|zTapb^t1T7PEkNP+SQAl~z)wpVOmDSZXX4086Kt!44RsBhAc zp&-zT6GtbvK~rWSv)hJ4?iYZu^sNnV_@G8&3l$V9)=hPr0zu?aiYJ)Nvr~ij!DL)FApk-Sx z;noR*n92lPsKy%&AVMj1&rl#cSDAJP;{C{N9b&v}VQM+UVvVh(Pe)Se8U&Xfs#nK6YpJT<6% zeo;!$shpv#P_br3d1zRc!%KgMOJ(*k!*}70)Pos{R(Y^#%QBnh9$C&Wr%RV@YNV-*&qF!&;i<69%~oN5mnMyHB8dvPUQHxnl-jT4LttX5~9Zn{XLZ2isD_J7m*i}#XpB+qG_06WVelK| z@g)t>1fr1`r3k8jq+8ulb5p)IF0Y%5NCcENAs~55G^|DqVvr&x0s%pzA|w0VgY&v` z^AVMnE>>}-#;MzML1xaYJD62Y{#R<~FZXG;_j1zZ9E3e4AGAEJtFVuk_2VE<4buTD!t%yhQAmKHJ z;*yD_ik0p9Hb6OKl8P@hj(9F4%`pk$MR>}<&yC4MDYZCd+K>f z1i(cGe*x18m>@TPJBG6u^~rVW5-E_@TyawG=|P>&;=6ZtMCD9PLzCtPrF*o=JKw4F36J<`rjPI)TjRI%QDgy)5J07x=tNom~Y+OK5{FmgB!73KkcS@r#GjFofG7$*ikGW=q#!w&cMVu$ zx55%Xsv~oer2gK_!y{vudd0>%$;ebA)R7NN*n0l_xWUe9`U2Q6+Q9$0aJcXdSDP!%1}}mA z;Q?6VpqpDa978NsdvP=N5ID4gdw@lpxShqdb@Uo9~UyPI!(m`pH^n2QP=xW^P23hJZW z;e@JZ(s^G?KnQ>09$=4!I%#itt`!EQ3gJB@@ukC-UU$ph233 zNEm%wwzA|bIKDFUmOM)XW>9Gr6{UN+#c@pZ^#@g}%C<36XuwX-jS2( zp`9V7!$Zv&=h9Rg$|L8xb_|W9hfK}8>Smd+-|indZIAXur}Xh zmign`;unF@tuDd>86$tl3N**pBjMyrVrNtfKe_>q?hz?Ptrs*?>q1z8CD9y-q(B1w zPVG`_kx&LR59Kdih2@0}8%h8v=8mC#p+>NEX6wgI2=UhqrAYkukvW!2(J@~N>W&NQ zaJ4gMr@WwypB({x!Eic|Om&XX^z9+CV;(6|>5!YROQ{i1(pp=*xYQqTFmwqZ(FkaT zScN7=PE>zJe}3AXEy5WR;CU7E#74Q5Z-9Mx6$3XxMUlayg74HFWz@hNJ`Xd+=|L6i zpZlzaJ-7GJOT2>pWQpJ-L}Gc{<%7D4Qv^S~2a$p=O4!e%2}5B|HQm5L1DkPCVr(Cr zk-Y2lQl0*s_c`wobTUHPT@}Hs*~xQY2;Pa&y!X~yaY>G+umG|Q9RdO6z?c_o6ga1&krWm9nwDDnj8NZfT^k( z;^@ZJFMZ8_g0DN5QpU;#)ZF=Sqc2+%(T&B*4^Fn8C~sAHTnKj=q}KZ`ndn&5q(tw6 zK{psc&`d>sh?Yyv7|g)qd@Z--$%+&h3lGTDFdR7uym4rSoKkI0djhYr=yrS0G^(~u zlO5O)zd?lLJ}8D7YUD0gq}z?zcg74}?W#0I*tZX=)tk#9K4ZSefbdYBn$G$zRH}F> zpcOtNj_|%ObcdyrAj7uH4f*wPS`D=RdG4@L0%0;H*cJvOygV$IV-yO>Gm9oNdqOWw zoj0A3(j(in4RS7X(`HdayYqDFwj>mr}#iar@MpWoaHA-x0lV zUhMxdqSZ}BtqYomwFH>ZqjrWhyukIuy;)8zi(xF)iRi={?e+9(QKK77kZrg1d*>~u@m3_; z`d+T+x7Z777^q&f=aQV)Tc^Z}uO7IKv_2bUOk-nHP>3YdX?WpmRE6#`vBWDe?Z2HP zxgq#+??%6{tO73+p$9y45pch)5u4D4CKt(13HIkut=8pgsLDI&GQ}0Hm4C{^^1idE zLax3bpN6!psqPn0AO_MOAw29}eXZ=*nQC3%W-H0Oe-@RiZb}k#;|{m{1xzHjVh%aK zy#bdc*)jaikpY%C*8PV6@tr=yK*5 zQK5-f9A=mVC(S`yBz8!Ll|S*9(T#?hc?S+i#Y=WW;^EuZufriXcXzD)0f+lAT`}!v zM}o*|aKGWU5P5moSnWFC2*H(6h(4d$Jf$9_O&1zZ(0H0D@?mUs%ul;^-ws08z*U&c zAGkS8rejm7Bd5IQRe;%~@pDX3xDk3DWqdd8gx!CZ2PWL_RO1vT9zP@A2KjqA&9~7Z zGq}!2*5Dq~BrN*wBxn-X_Vf{FS%P!|-8Ucs<4@{AUH_@vkL0FIdd}OxlVgj5)S;p5 z{W60*iNKRv5A?33ytDnzpH$;NX^f)0F;U6%;!1|w=JU_-W>;%>`vAe#MO-2|_!=OgWtW%vq)wO_hNc-fp@m*G%m%~z@yGqb>Tui&)s!+T@Hx|a za0P-O$Ev*|q}IsI13qd>wohcX_#7*=>$Tq-KI=Y&YizSfw1|ck%QC;1N4qms=Z_&| zfkRyamc<^*%uxd_dAaMdlODcH^#O_3)oPi$Y2A>;Q)srpyu6MU8a()H9Mm z#^0+=DeM8NoN&r^%Z@MYdXgr#i#4q*FWo_NXEBX*`7U*_1vf?QPwbJ|XeMCJWP$G3 zShvW{H`G{8c#pS#yK=Qa3o(GPfB#|`DD{6ByTx%n8DQm&R%xE4^H?JX*tf$i+62jN z4eBIdC>S)JCD*u%J5ZPMM?NJ9(L7PDLDR#P$=2Jc!>DC#`&`f?G7ah2hWsDxt^jr9 z=UoVhttlj>okF0Z%q>Sr@t)yMhgQz{RbZmb;qZpGm!u_ulrfX#U9VSa#O&*l+;$YK zkD_|Bv(^^xBdGb*X^?g1cgJkCt+%%&HP3e6aCg)V--<)*pMGikEr7G3yR*q-E4fq7 zA0J@rR#&1$!el{*%>cIZhUob}+uz4RU1#G6$;{B!ixPZeC}nQhe;W=#XI-PaMDdmK z1-pk7*ABUH^CL*3bDmp=$y#NTrBIj;KO-j3qKl)Td@ksI`=NG$C_qt(2ZdrEeAVNq zU#Q(^kipxqJE~E@xa21tVuXE5Mxb0DFp~ucJq-h~XUou{Fw;uLpd7%tdJwO3TJO

K?H&gvD7+OEn3hIGZ z(w;Wa`5?h8JKhY#kxC6ojg|DnjO#9#y>K zqY0JR`E8<^5jgW$A!yH6r48h@P=T6u1z5`OX#~x&&Zk~6bM=LpvsEq<@{H+IP9E)B zF%R{058LsM1DedHiWtwQEDNFZsWu7aPD`$f3@i}KnWhg|wC*uaFGTff5%7P&GlmeJ ztE)Q&VssIzY3V-Bt8~sI zot_Hn4_S@kk)BZTt$TLIHLD>8?#6ok9!H9g7uw;ai*PBP-UUcKP@%N>Y{IOZbWG9m-m@0=OR=K+Ding-Q$V2TGO9o7bY3~d328eq9AkUWY&n@+$q0Ak;70V5x%@zit zAAiUTsKcJ0z|0!O&kQz1e`wiMLU4-=wJB2tTBfV=nf3W+6%*Uk8iZ573EqZ@=Y6Ki z(+*YgFOUCtFxLVWHKuY^yuZzeUCyQq1y$%AeRY&m9vprsdJ0x)x_9-LqI1n+x zLIDOAS83SW6u8^_-71d2y7&(7R3oth9VnuZ()lx!ACdj(%FX;1O>V!41u@&v%4)5r zCl7b9rn?<@*aSs36L4kAYyxJjU^@64glL&yhrcP0%){9fQhj%Glv{4eKBXittA*M% zU_>JSy)jZX7xMX6#XPpW(N%9d6kLMU&yt9=I4?^O*zwYtu4eHFumZ7I2%6!ezaZxZ zxCPJ)eEkn|ey@WHbwLa+Z??j<39YjdIrB_v-ZJd-O#`kGcX7CEkmx$g88LpOfx_|1 zOVm-!m^AE3UdlQWa>QmF{$k!BVVgxqzC@R1Ht~<%jxskDq;38r=MD4G0+`5m(aNeF zYjdw^z;Nj4hpx@ezHm3=R~dV|0SSjXIQ()@z6N{C1K~$pi{AHQaaQ{FB4elcVhOYD zLI^ez;=;P%lo5vZ$A^d@#UmDi*j?#T31>KiBL6AI6gm`_Lt8k)SM%Ea|^gBS@-9&zjW6GEFW?#kZWx0|!iy?g-3s z?!0~1m6TE*W5kr?X~c!dGxhGiSSrH4GevKOUk__ zW8mvO!TZrMKyNKkJjx0N_Z!FIlqJK=4lZ@#xgOg6+_A+qaV?4%^Q49b;)|^-*!>R* zyD>5m*lh>0Yhp_<#LR&OkZozReQF_KBZhN zjI9uitlxUtA5Kl@gnRC!GYgX)4QnlhCwJ=Q>9q8FA`|PDt+Xo9JnhLb6#K&UGfEV2 ze0|0vowS0*^nZvo@DVkD5g)SIZEBl|->phL;gt0?3W$s2=yYR?x1T4zn3VPgq31(w z#y5>680(_<;1J)N7YXYO`wg@dB9S)E+NgH3LBa}`ns>hM)X0$Rb#B-XJduM(=~3$M zpnP+p&mJand7Px9rsA~PjGo88?Qqb)vTsTZyLTLsDV+qFx&~{FrFx@Wutm7KLOpJWgxm1pymai0 zyy@RGy&*?=b`Qk#o-~U%YGVzoH*m~Zt`-k=%gJlB&S>X^b)h%qPOJH=PKe-g1C&fI z&hbmC{yGTNzzxqS?I{7|oxN?D-``F4SF^l{zNYU!1hQU^_go@CZJvwNZrCn8Sv2ly zTTXf{XsQl*!L5a_xzUH$A6kD*5T{wE3mQz6k1)A*PCx`xMj0%xMK~{7^aJ12-&A{# zHCGJ|HN)NS_C{p3;f-y6%83fCL`y^c6x^2u$zDZ?ts+(-p`9KvwhY z0(k#{VyZaH;z}Bvo;{-1z}d3oL6KKY)&1;DAzzENmYnydVCUZWfNLi-Q(V`eJwI59 z4M}23^<*JBq923w7_lJ_ml*Sb-MLV&=RK;_#*lui^BAj~8TU)6^OqD27W_J`<;6`0 zIzO>GKye$L{UOdwL#PcHq#Vu9uU>o#E!SykEf3q|=8sqZj?*`fuS|??)q7JVoa<>i zstA{Ucp|y44^)^Dsv?BF7ECq9l+27f37^CcZ$R*Q+g{o3v98uw^8u48PeZ?VH?D3x z!kgCDLxRxt|3hWq)Zmi)McRW8a_PjU=)^Tg33&(PE}}r`-Rq^5_Y1Z5-~^A8TW1rQ zHxBK&wFpXfxwnn_QN7?~OGUqr&jFg{pV)146!ND-1fr(pOTkiNln(rk+T2}xPlKdN z6y0>SbuX9~83x-YsKBFIj*noI2g-2#Pcl>-Y|$Q!r z@2tluo^G;k*e0|sT?0%?w9<~-d|&aTc)r$w8{bra!zw`5#B=2#U#I zer~Wc)>=g7Fr`BtipAh&L4g5>=nW{Xd!5O57?t`$cL~OJE}T!@1j@P*@T&5VhkP$D zjS5lnRRK<)#5^8@SIQsQ(dz(tOE04)!*H-&{5fgMGnO~BAo|4oi7kdwHDK8B`>^n= zRWZ2f5oaU%5Ux^6fmnoRIa@_~#`nRo# zFnhRtq{rv5U@jZuLv#pe`Ltn{`ekSmyWK?#^`4kaIu3d0LZXF*8_sJTvH5Ys)zc;@ zZg2CkK-_*@@lLU!GS>N+j&g`;Y9f)KUJ`rUz18eVf;Z*KyJ?F&vG!fo;z8%tzT7^1 zG{uFBK6;{!173QBC9t4{jR0*rY~0iU(R$#_3u~rcy_%#0RL093mp50WBRRePM};&5 z(bitDpTQ)Mt$&Z6vTOf2@@^N4b*F5hzxIUcIBK(<-oF_*Ut7dHe27Hh*(ObH08n5W zhtelZ>wx5OmK_*#a$7ZQE3c{zQJ+4o4%_HwwSvPp zua0iMhaot<@cd-&oOIkf)B79uwf0x9yQTGcN%{Z~58h4a&*IQ5qpkJpYO_k>G#|1W zG${|zLRGcx1X6zY{y4wutjO65CQMr^znEY*b`?!;S*#oGLDl8@S+~zZ>C5t6a2Zdx z?H*$oDI2SQg(0R*M%q)_o&e-|yqxOEjWnW9=UUJ?cA_f{L7M$pis~GrNhu_*z2ANU zjO$|nT|=5lB0jqDVGc&iYKM{e4tRT)xNGl+w+dvb5{L)4ojvm060z<` z`&z9RSD{>zA99&D3~G{{Q-sIurDcGZS@vI89pGeIp9w3Mv#D2a{#F#vt25GTubfW2 z$Lvg^5I5&Dit7CDlTOl9T{Z|8M{$1`xv?XQVad}U{-qf9={oV9O95$*{R+M$eO#Z> zXgg=nKLviX@WDDjBdAdXL|j?~F#FDI$=Kv+sz$Y#whb_1KSos6PP(=rlM{zl5cdPs zVt>8VSm6O<-et8jBN~QgSqRKsEyk7egO+RgAb{TY5GK#Qp$V@#=(=ADnc6CRB>T33!B z;c@_x5LXv6wdTkB?6oOoQ&Q0h1++l|MMbCWjzkc@_EPQ}Z7H}hZaNf6bAZr= zV`?(c1QVO>`y?noRdo4p6jZC=tcH5t?~r&K?c&$_;()y-gC5RJVlaR#HA z?>&l*9)U&im)1kl$)EArgjO0NJFhLJ(n1q<@&Qj`J5gYH20|i^-Bx(Y;5|(FY57RQ z#ZW0St_w~WFzPS{kD2jt6_cMg@ZS~8X^A+eNe@<*Ozp!_nP0)W>>JW*2HWjjEfH+^ zEP7xzvdJrPd|N#hx8`Bv=_UnMY5v5hE!=X29!WZEe1*p5B&e66|JHcf)PP5a=$aMJ zn}r#PJcgn*AF8c0AQTV=j|ji`U3tWR%V~7n7|K@lzQeMiSfPmcS16B|-AC@pNEHF$ zXw-?eO+6>0ATjdF(vys!y}uqF7E^UMhg%UO<0?XjE^dg#dP(5;WCrhP@$QWjhdz7N z_7j{rjZK)&W*cs|u~TWVr$C`k6DFS$ZH7aIPB_iN4y}Fk2|XK*j~EVN@Vb{^2RYJX!x`M-cT?wjq3i_yY-`CFoj4ASGqkqH0C^HMAm`GlS z5;eZ*jbqsIhZLJ}%Po%>B?Y^CX|&3r@+Sax415WF=a2^5@dxY9j648KE5`7WAcf46 zs^UMs`Uq<&pbg=681`Ap-qD-YHcRWr==yoJy`0W_Jk$Jya@N|r5cGA%n@xOwNRAtL z9&DW=vlY-!RT^^06+0P8w$6635)`^?MBcS%i^phsKUzpg4fskue*aSNi+hrk3n7sY zNwSHmLno{K;|$&kPQ7rkuv$?eu`^xoKWXV*Z%C7TZ(J%%P)>3x<(*&A*0k6H+Bp&X zYLcnW-9|OM_9q|cbUyV%=3qh&`?U+MXrn>t#V9uWa+#9%8)#`7aKE~)oMS{ToyTp) zzK2DZ4@Xj@7!&y%Z#^2l_HF%atx`N$ju9h zM=K`9(V^+~bk1`j4sO3mzv(xk2FD6kYtpSo7C%U~nC*ODg=u81uFB$`jdLt0)p;Ee z^|c1xR%R1nIY~ST4wWvw@g8II5*Cf`FtJ84fH@3SRc_%9tT$3u_)CF00}=@i)PUSb zN?iXH!A1mAYqf&1rtrR4NNwh!gM=w8V*>gX%SppPKp|MRHR7BFwagsenaK?O@i^9q zS|+1AyO^{tjl-Sw2oemOfPs>K;vKua_L7gnS=7-?jd|hgaw@vr>Zq=RCQB*v9=c=V~xnULmVPyJ;l`kIN^YmCDT3v6824t4=H3~~K7 z$^X#$d;v+O6Z+~%v7IF2&P;*;+#A%zvX#HH7gD-$d$XBSwboTWwTR0c&M%C#)-sGG zLMp)wSLT*V3ZSo)n+t1*=-piVLaB6kX&EQ^@}@HL4>JrV;xBU0A_*3#ax&!}!ut#R zS#t=OG8-=k=Usf?6UyMgwGx5_bv1;g;I``>XVx0c-Gb%h(>KA?jI5gYBK##| zpMa7?)A?TDoFCaK^RrA@RMAaDKmMp)q%>c0kn+}D1j9@adEFSia`HKGtgvqpyn@~1 zX(r9{+N@v1`APCx)Vvml6S3-2Dc7FcXW9*jz|cAkG0$7!MA{IuLQuVxmzy|ajYEH8 z__A`v4jzoj;r!aKTqc-MOzO4`xh5mOrroHXC|9rXuh$m!HsGjmLrWKo&Gwr7+6+wG z_iQ2@?Tgt~fTw6b_1G`7Eo5PA8O!UhJ<^YR#n*FZ`J#=jjE&_{Y>YW0HBUYVXfn`3 zm?)k*PCB=Wg?S$8#Fi(6sJY*W>_^8P)_q(9k8!;NP4|jDLCx$96kN*HS`mv8%8t?S z2?4X2t3xRbI$~Hx1D`Z%rEuslF2c;Nohtsohu(3Fm`~9mnQKeKITd7^l0R$8M;?*f zYOFKreV&8*kcK@`DkvQ0k1$nn0rXmJ@oVys(&d5YC)6+&a#1OQEPSd-t z1v~_mn$Y+b${ol?Nu`~gwgE|!UUR4$eUXQfr}8L`e5E{}z6~~AUplI{5zf48M{5sE z<3_OudH;O$VNyU5E^GU`9=isUsRf^DD$mi?K|knL#)@6BIDb0JQ|4>0zQJ9$tLG>E z@CqR zVZ9?Oc)U8UTXRYZ^r@DA$cJF__HVda;!Kg*00*N zk5Y7aRbYMS4#ny&7}qL`iITgf;Ana5U3GDHO+7miyTORHzBAcreO9)#usajmir31JK??yX8Z>&VW6&5QJ9Y;o1; z{}$u(Jd`GcgBMoeyk!YP&pk(-?A=jq1uja2A`{nwxWzaJ=fVKWW zk#!?JL=&%k4s-VEp)+U9MO#L#~5v!SylU7jVC>f7~G#kIQ|ct2d~9>P0Bo#LpN>2bgR zXl8F=XZeIy!y_C^ralV4Y7d%FRse(MfzUWoEUh*vwyB(zyqnLE!*%bD$m|zt0kcdR zSM-Oud>9-loiC4>u+aM8+*aA;kkf!aF>QZT~}P*Mm380Wrq#Dpl5Q0S!9+9iP# z_yE3?8s~!E_~enn-2Vc*2`fGRA9$XL;s3((j127m(ejM=jLeKojQ=_O-*}#pg@uLf z|B2_D{s+&u@Mgs<*LM^~{L4JQR<4TqN9h0C(VKw(+5A)bSp3?4dgVR){rwpoRrDau z>sZsiQi1iHfG|ci4}r0X1tel^jbn~scFqB~z{uFb+5+>$9Ko+dMg{B%Kt#jC!$b`k z9>+7+H@~$ay4Hs{2C3)B3_OO(;nU`$KDX%DmjE#2$IAZ+U(C=q2a6Bv#4PX92+RmD z2Kxi_&;l3390CT1XOS0O-VPggFeoy%GBdriwKR1v2SbXL!59DI2E7183>!slt&KxH z4G6|M0wAG@sp(f$JsCQO!*62=Ob?QQsh$pCB?BnNEYr=7L&7}?>GuS{fl!8hM3KH)mwNooHW01tpWY=AQhBhgo~ZQaEO zv8g+;uX!*!c&>jPJVXGpB1EdVjT;s5Zw)-xXw|Ieyu@!MMQNDHf~4S-=DBz1OG_2DRg-0yGW>Z>l|M_|SFhH`t_#@8iK@--dcnF(iW;2we5|W(Tf`T zFG-3W^BaaYAhr_K-e|@biQ;=>W)<(q%i$TrucHH?8HW1$kJKweyl4t5KR7r5)u=Az zo*4R1EOfKwA|OAH<@02i4UonLr0&sFa5m!2rulh_m8KPZp<4sTEhJ{SHIjRD|f z;7gLmHel#cl?U zAg;Erc9<=&sZUiiOKFsAe6~b9n;2pH;?=~`;P@%K9_Y8mzcq@DH6X0-%~eGc^%o`` zb+>_5Fv4*B)*Wkj$`RAr-r5Zy`5u`dCg!G;e5Bm@(CnTXM8p|8UjG+S$fboP9zPWb|GA@(!*pqiu=v9^+#^3bxc z@_^7B%ELS;hYi$TAn$4meq9Oe^>Z_Oz0pF>J{+d0zzRXF>Xq!mqrNQOwbI*8HmcSB zy3))9^^*?5-Lq#i0NP9((I!ryYK%NFtisq*gYC39lzZ#fWq_4SpJR8LnG{7&znYvc zPG~nZS0HKc2a=hkD!Jld;>#?#mTtN+B1{i94xOnb_uoa?5hOpFfs}g-vVGT0yiL8G zANAq#6O1~ZD66XG&^RB8Gd!oC=)@N?hz>#Bj)Nvo1zqE~LXu=4{`d58Pe-+K{ z#1HJf0myTmhU? zTIp8kKv|M68hQgGs2h#5AoJMySf;MVnM6i9P~Jbv4en6Qei(nOsEON_f5!YS5Rh|} z3Np6o{4o6CcGS&!YibxZZJH4?;>NW&8($vdUc)Ns$&b90o}%5`^gdTnEY{BgU}8j# z%4Pp$)nxpdDvEf+iE5h7GywKn_pkBHoa^^wwbwC%Bk8x4mOoX#4J)2%J8;D1D&u2c z8tR-sUhVcvwwtmSrYdh)(ByA<>2K!I`P(L>B^Gs5cr&*;mCOv3Jn)u;(nn%AFC_wv zO`|3;dzR(9S-I58OXk`sMwV@{Xse`9G_}8?D1wHlYOpYo55ENh`|bqz+|bnoHbE|0 zk$y%5CKj!$(P&SOx>0eX;(59}ItYcm7q4!C=1kZ(gP@OtS36y{XUvbjmI)aW_ z1^g--nmP#4iAvykxv+wHaUZ-C?y+#<>8B)p@2Re&0r7Tn^R{ngJ{;~oyt3U=T|@uu zj-BZ=1nK<*IRZw0LGxuqYerLOUOaSI^zfz^*uHO{;$6+;pTn%}9zyTqEM&a|O3vix zDO9CtYPLDnxAz>S`g9kufL+jUBnpS(&My72O7<*mV_uz) zK%AINZ0HluL+dn+v1Lv#!mR^2^==`JL!&JG39e4Wnn>M1?Ax;{-V4hQ8oK=N%50Py zC=agu0qGXRZrHe1>}W=TQg9o5q9K|9DA6K`tmPz+8<_~CYCk|N|x|`nA=}mO$)_R?`U|7k~TJppq{Z zmlKPNpo=O+Za8ik)ATOc0^D2rnRsnCT6G4uM|7EULvocUP`9rLJ6APvXi36&A=BXc zVR?{dH=O>sP8lCzQaHFhz*rEI6Du^KGg0{^Oir2lac1NPDNcjdS)j99o~Ka)yxQifeW={GioAC+ttY zF{JCR5c5atXlD`>mq?{YNY*Wnp-fJEXd_{iBu-72)_Cblf$V1B>O0@|kr7Lu{34w} zAFY1ZQfwp^w@Zqfm6a@-3-k;&9?t&vqw1HY{S*#NR{B!5^<%-qUtyG8Npcxs=WV{o z4iZ9Kb9TqVnwkZlG;(DG7wkq=3q>4&?s(-{Nle6w1l{GXqJ*;O(y3ekSEjtR{>mAoHG(uP|>ZQ%}I${^P$w3{nyl zYeJKOM0cbWf|u!9wve14L22P@N|4;KR`AjO;78I|#S{t?<&zya{GbX}q)7MZeJe4p z(yD*`FB1{-Z}JvZJP%oYJ+@Fel0Lg#g^r4wU1JNa2Z~WHD(vrd%b`4ctkt=k*ku(1b~=b9_xbQ9cciW=(Pobb=d&kAjkj|6;JQ&*xeeP>4% z0h3U!a_5n^bPTE5;H)I&g~%n##Mzf#H$z`5&X>o9Rh|8miWDzEei^?=)~#;4r}s^4 z6_iZH2(9d4S3s;&-mTu_jvQkjm*^5bvk=_F_{yItV9~Azu0G)@qC{04h<63x@aQaJ zXz0~B67+>A3#o_mq%aSppEB{%Q0MgqY*F3jrUfl*^ zB^ir+B9w`ed`l(`y*ZN+YWW5xYh(1wsqI)E9rG9I7_ak!37)NZlM^G=O(*)P*UTE2bESkY5a;-=gz-vj&Bo zn|Rnyy@F4}+1qo5>UW%5ed0>+bz>?DN{jl(VHZ>|w7S!B#xb8rz4C$jL|}n7NVaJb z=)Rm@Jjbr$=ILXO(2Yr}r}0S?z8dN870X*q6(I;m+;$J+gVzzBBv$s3_aS6*NNfD{ zPq4-3?5fFApL=V~t*tUR4&;A2H0&^-*WgqYD`IeZe#*3LC31AvBnYn$BDIY!r()V0 za&=+K=s;L|49ndJXL$Pz9_<7hPwD!`IXlg|P<@kZG15fEPsgdvfHHU(PnE5n@*297k=j=NJ;JC;WyBsKiqTq$+eiKVgsUL4_&C_V^VupNQOr=00wW+kZ{L> zpy#UJ6K0yf!^;s0sx9~^>t<)d=qtx^d~DCwJu+cdOb()@p0(ZU19AM6$*r_JkI5q! zuYlpQ57bXqgHx6flOJ*9l|UR08{Z0HIgc}SYHY$EvF&r1J}XLHp;cd$ty>F^k7hKn z-dXTXS>?IOnB!(v2$KlMHDR@9(Yo7iPz-=4x34Nq`f)!6K85)toz0(k#b39`E`uhG z0UzIO$iFJ>#dExsi61KaimLw$=JJub>Apl3Zc;RnWH~2pF^<9qE2!479 z+YgyZqW!Rwiz8KGW z3lw$cR)%k{$riz081gBCid~jQ@haSk7$-McdU)7|0l0ifg0GtC>1=IuiF7H%(&`yJ zZffRm{`qsGsgJVGGyb5G{38yJs(z8`X)U%0*PF7Z3ZCHgemZL>`--?ZM%WU&Z*0RO z-3_MOB^+cRD3yHox=e+VO)Cup#+q#~JQ}0nP{ND?%Bi_80_0#rH>uaSTahVd^K~`` zTkXpA{wAu!jO;#;!Y-|&q-MO?ISZZRoCpo;e2ne`RVut<%Z!8>CLC!MvKW~*FAWNb z7BR8-ZK4(AAKM{IvE)hM@cf$+=_WBQIAp_UK9VRiBa8rDTKlg|jo9f5Gt6ChwCw%E zL7|qT@SzCimgnS+f21o%5y}JTG0V~#AEyYlO_C{q- zmtg-D@gU7Zpx(gTTZwOVbh8tWiX2dNx4UnSXRQ+LoAT8f7~Ht7QT6u8%ZczIH@q0O zsx_K{yc6wgUn(0i5U8kMNgL{_7e3Qd_XdhD?aU0^QCW}V&n14^-%vkmDPdjy&>)aR z{?*bG`)nC+DmIphn=sj3cxn-QXBeSZa1kF)CX{Dc9U+`!)cmb^9(MDwl1$Ny@qEC-H8FIjRCtRFJqkh5!7MgHiN9?hoY}2PT$|@oDVt` zuH*;|ha zMpH$m6a8exA3T*VrAo$>yyX+`l-l~NM%C4#%Dyt}>9?l7Ms@ZJeP*{_DYJji>XZii zMD8-BxMMcyO4(o=JYzK$6bY?Q)jWvPVm`?rq$LjX#z@lw$oslkr@j0%5H<<0kL{-- z`pAsTLlh7&M?aY0hp48B#YKnqm3WFXF0m!!RT2>|SD!zd;B*#l@TJ5SkJ>*)u`exE zP@X?h^#)ctW_$yUx^YMkc7G?K{Nn&fLW=zr--aT;r z0jw-%m78Yc=bPDq%;Lh5m26v!xMqpj7;n}H^s2+{S31~=DLN~~LnoS58%GrCVlYR= zSix8hoEYT(@}`-qG$;xe>p3ZmbO5*$u|=3fFLWs zf@-RSEQJA}{f0RO!M&?Qx{yYD3llgY^RI3zm1Pjc#xLI=-q?{z~ z%AD*e3mdn}BeCz*8#+06-cDg_Vl$EIb+B7kW>Ri;d_#0TS>AtC{By#C;ZACGXEhN2 zi+5$9={~gypC)sMm=ox)Ro3a?K97>QI{LhVVs>iPpHusP3suT--iOYI_OP7zT1KCs zGxzUv^yePQ?Vl-% zt-doXnk=6_`uM3uP>044A%@(<-x?K)R#x``Tc=Rvn#Fm;ZNYZ6m*|F)n`Oia{bV;S zBJ-9-!oaX)t>N;#Qc}(o&x2iAyUYgTm5kj96r)Z-nk-o>Ff^u+;6Ji z;xQELhH{W)PZY8XK!i~7QUwwoZfm?+&KctvGW-TY#z|3O#KcwXT06_@VaCTxcGa}$ zb-l?>v{na1`Hy9L(wuRA^L@_sr5ZC3YlWRW(28l*Wo`7^>p?qq0U`9I*zX1vXa3mi zsfs}$YRSG{(h|-=a=12sm|oC)F#b)eT8f7I*34SAu+wiABB}@OR(0J9Iza>qA!Nx? zN8gi+Y|*z?4__mD=Qh5P<_dxDyz~@Wt2?n9!x%tl3X3Z7H;$`U=qk4d`?vSbm)EY_ ze3b#rIPwdB@$y7*U^BDyYTUqF(^w+C$oLg*&sQ&dio~9UQ{1lugo7S#_DrTE!f&T- z9y!}y%rJidJn_9m%Dgi4Q_ZqJib&y4Qs={4G6jo#Qu1^ZKO1e!xpSsU<@m{?pCX{t z9-|bOsRcc3B@^w85-FQ7BZgY+H)d7?!`k+NE{PwP!2qiz$=&%I(Cv+?Vk^-<;$yAa zEJ9FSo;ZKr(TCydW6FD0$L`V?+|>WkztZJ@t8%sD>gRcUEQ7^etQ$-f$lAi&^7nEO zSziJ9`<+$v9qJMl*Yj!8Fn+fOP|7v~~^zXZwTVZ;6@*Ep2 zGF_o_YEJ>dsF1TH+)l}b*XBISEAMqRoe4vS4}DDZ3SoiypyDMjs9j~WwaDrLC=f*5 z5|^lcHHT0X5VF0NZaT9Fx+5yoRY8&QGf|Zi{k1Y#OV!eFSm`LhBedY_TLZ+(w^mH( zTkpb?O{Z!^`Nfhmgq3$Fsrjbv?C+@war(d>LJeu{yL2SjhYPLvbKhk(wv)Y4DaK4a zp&W6kHZ~$o8>|`hgTHMgFg;-Twk4%-wmZUDZ2ISABj=1TV&*g~(cuf1a(~Ka(WMmg zfV^Gv3a51wmo&2;&ss`GM$q23fXKGU*)5szFKfAyR%tSVXIyG5fc=i3V_6_8B*&4i z|2&I_A6+NNjO-0(q<@5(lE)Q)qXj*%gTr}x3nQM|BfL$R`>y2Ia#NIq)LERJp{#vJ zGzQRUU{{BkXyqSLncGW_!ljb|xuHWi%G*7~C}MJ|Fkj%bMoKBH*bY0wH6L7*6Wz0> zO_GL%D|aEwlx*$6+6LW~5A61l95d`WZxQ*BLw?#E=qTvp2>6!T7#B5f|3jX4CzJ>@ zrpZd{Wf%QQG4!uC{MW8M5kic|kXe%x{n<@`v};gsZh{3tQFNjyng__tW-LMM(H@KC zyPkZi-ZufbeIrGFiFEG>!{;2S)hF}RYUCpavw98^Qf^aA{snHud~%w_5A$A%5Po2) zMNYZJVqzVGt5Ul-Ee~nX=)&zq(h5dKRDJo<*yTma{JKXFa~1<0X(v1zx(2ugbVvW&L0yg%iGcA!7hVhoB0HR^()|*WlKr6V_=Y>6ASczH6HZ&QV6>oMnL+7c&BTj?I}|W z1k0*#iCpRy@`4ndE#c`c+_}sGF>u)%8aO8etDBX@m7jY{I623fXN1tP?;hj|yST|s z>Bk4zTqgA+`PUXcaa&iCeI*DL70L?b5z&^j_)hGN|6)b`Q7%{tWqP{kw0+=eKZ?+|bDWyg!pXABJC6}q2Lhwoqa|g6C`cGr zG#OYdRWU;=eM4%$+``2(>VlPApLk#)DMylyr$8h3OW<@N*D#wEvn`n)2^q=3jK(VU1Rb&q$jgw$3JV=~2voMR}}ArAKGAEC(S-96&jm*pYX zicC|zC)}rz$U=@Yq;C;VrsHvf7}c6tfySjWS0d~%w76AgBGv)Lp#Js zr>xlR1*$`ci4)544$Q6eR`0h12oim!SE)nmj61b6Su9+mMX)6QaumgnsFxCWd?{|t zd$+hshpD=sk0eoHPiSHO`X^Cra*nP;z^yQDgN-7<(W8)G55LV@bf(EHS7`?urZR(1*-Aq7I7r5E$0SGWywr zgxmmw_&0nXX{UYJ_bXlsOp_C{XuDIp>RK5PK**P@P0^jyfvyb^*TRs`OryE3nfiFK z7hVlp=2-}F7~K@~03lTKzW_Ht$iF|H8bF{`DI;b0O_@__s&G(A zs`u>%_Dk71Ak%Wi%-r1oOQ_peDjo zTZl&%D~GbJ<@O{qfKwz?Ok`^+d#H>wkFjQI@xbqJ&LU}<8*7Z;sAJ?JwgOYNfwYvh z?Hq+`E5fbnoA?v?MrMjHs3;jvZq!%7@OWcHvLi}O`yD}zW9NexCcl@<6vF|^%UHHj zxL38>!m+Q&g9f5OToN2Z1bpnHdDE)g%ku8Z?)Nbg;+Gt!I-?x z^S)}eelEva0-u0F-Bk_sCZUZiFk%Le9W!C>F8bBXKW*fr^P~8IQw`;J^5|)?^f*6i z_{=c2DSQxV*B0sZx52v+=x-IvlZ>8qNW8y#97cvS=udRWiz54;lWh?W3C8EE4DY^@ z8M2`?ksgjTKH6oBPAC``GoRK)=^O;$Dm2&yn{fq%#_cvC41PiJZV^z-E25x-h+TSg zCS);ZKVo$c|KT>(hk(0{KjUihIC8cM{R4q|RtJh!kqIW@0J zjC-`!mD19So3uqG)Ysb@w=LqxRfJJ%2d}^7PByFbtMU14`>au&8DIKFIQ&QxR?N6gC`(cJ`}M5KZIjZdKAcmhv@AUsD!z zS8EikZg**IKjCvh)rp|Or>v~{1IscZCmxO$(x7eb<8dq_E~AnR*;GxwetLStI~u#x zitw;`u?+q$&bMy^i^;^D=~yG$cpH6{z1F{1wG+6Y1md`J#Nf{DF`=2wFIYW$pADRO zYn_(BzI!wEbMa<`rMdD@)8BmOQWkXApcTW5q78x@FBa*!%s{;JW9eNmx#{f~iLne#xUjjd=>PRSN&hJ~ z_KYgqwc2w*d*zOpnEKoiJ}9#{R$oOol7VU&jg}ki+vOX0jby!O`nZW6ue}l^rdJqJ zs4-70cN0GP{1YvB>;2pAjY`byU}c{9FCu<@O7u${a5C3~4>1(I5Ox02sP)NQXNaMZ5qnG)cXTOLU> zLkv*$LJ<9C?FW_^w=Qgs@vMi``2KZ&G>s0q85EX=^%(3?p-}dB+|p>O$NV#MpzByV z80>~Awqpjby$LM`#~|c6aWLu^6?}O%sLfB2GJw|aHewr7v0<4;E6rZh0>Csel}Au_B~U$g zVPn=XNkiR$ZxM)sPpI%N%X&yCqU+xtL{uq}(|7eF`ivNEY8f9M8VLO*`FCm6r66p6 zZB{1j-tFQya$SaW*C!Z390r!G4T>sVvk+0f0ND*v=o4zNZ44Qxo4bw{#pY;&c%wUb z`{h(&5`F891IFb_tly&0eLI6Tg(HuIgK+}L!c+F!mi+kNr2+6+eLI+0ZdEg|G;gJK zWDd7-uSbNJn>1=O6{rGNj2&$l1kA5Il}+!W6t|fo3sI}4m}-l^8jiG3r2O(o;}uWi z;FdsS%OMU?`-iBYmDsUWv2m~Rxv9YP^T6~&L4&6?Q-@;enZ(gXN+krq;S^QwspgtW zxy_ohV@Z)HVz+_oek%}qx-hLCtpQg;{^rTa?*@Os4{QLFX|9`MG8?chxG?{{IC>r;qri_nkS#O* zpfUQ6QF4^f7yV0Oj2qF@&%B3ix-(pu(WNEq(!1kdV53k=PK)J1}XyAwFwi) z-BbwD-t7_rkgn#5x(;(43d&k2?l|j~U}oi8z9;k3;{2^=B5aHlugpZ805_wMPYp))TZ6mF{e)GS@7QPU<)g=|f-Q8yTZSM*C?TIVBBYD^PeM;JVsP zF6i<%yHsLH@h#e9&z{R5NbzQ6GLP(qAuFFr&&DeF^L_cJ5=vbn7>If^0>Oav+r_)J zbZFvHY8SP(Rpa}rj3rc3U8v%WuvgfZ~HZL1#dye2@Buy|=REJEWQ61lK!xJh5UssSH zhLf!cP}b8K00LiqorOV~wNyI_SJJN4HbG}`8j3(QW8@g&WZk@PC2~%kR~2OvM2-k; z0_G`jf=cHY?bUwG9x~sL{_4JJl{heJ;7vV_hteCc)(02VMdGO#|iQ>pJtt#4U zhY;TP^MYLp(Hv?Gl(AbusnT>^7#W(4BBvL6hIK&w7xfxNnJ40YWN!>)mke4Gu?OCe z8lDcizm8b?)xn998$Yya+(F3oN5y05H&DM+LwbJtFSJ`*eJ(M$(U`x5#7(%J z8}_kZD9hC$VD4J6t!Qz1eAA=^#2Bv1__sg`Ya?c?mek87Z#=NqlMCny%m?QPph(F` za$=>%b&Gr5IaDa|Ny+(?%Miut3VSDa>OFnsH2 z8sF!}$%Z!CL^&c@gsIOrWUKIWv*Kj7HlJC~A2r0ey>)Z!KwUNwR)h#9J{d98aG?&1 z%#7Um4L>#8N|F+x=3whpUapjHO4cgF&@k-7-nMU_S5_YyR@!I4`0I#FCaeK$o?Q>CuMc*e`4}>WEJa_98Wt94 zPfw6z0K`^(la*KjK<&!%6G1>u>g z^)hZFxcq7Ym}s`U0riZL3m9kYj!L2^$?nTZ5u_0Vo|OKz&l@*?Jk8MRBGIA{*aj z|DseA8;qG3MPrF~TjhQE1j9D;;to$XXenIpnR0-%Gi&H;bwH*V|LM9IVIlB${MAt4 zT@L>tV_X%Jq2~7DtI+uaFGuk--t5mYNHp#;AK%{#m&+Mq^IylB8GQs_ zg~*x+*&$mV5*JEv_1=g*+mMTw4LFMjzjXY>qOziY;+a|AtQ3tjvA#_nk_l64xUG2( zpXY{19sY{oN3pjJ(^D;S`K8Vy4<&T~k<($Wp{ z_NFc6$R)=qkie$RA8LCXY@C6y#P7l!fq_l-&_yw`VI41s{JbMOz8DjlM(0t0ZK@JTPT# zrfs#m&U1|5S3xZjan2CGs~;exQ){;%()meZU<@~m_+uN0Js6rzEHG#p3&UHC07*{; zPLg7UGJ+tJ9%`7i=dHMFl57+}Kw7Nw*x;AOpbj3!PG7G~3Dfj_Jj3)xu$AuH7ao8# zvfJT43St~C>B~xl0)VwNnB6Nn*(^JbKq6l-MNTC?bvLCj88y&hU)+0ft_}ZQI0ua1^qF6cC=eU1oMNwtHHRtJd2-<6>)yE zBjCJPccrGK<$Oe?rRadQgM~wDWgO4np->TnfxDHoekbvZROb+g;iC0D+dkxsl?6rm zV*hqnjfZ$12NC<^>(AS91XrWO^y2t>mhmY2xPvm1n`P8D2%|29x_awX8T@jiavKoa zH_ToxS<0%RX>y9*$8y@QKAVJ|)QgY3Tr)Vv^6N93xS6v?ir^Fu{NxqB4fw|J73B8J zf$!<)HoX@w*f@uuSc$vc6L;+G8oo9|1c$t2vkAK9p!@BPB7z@0z>ST7tl+=wy1CMAEFN?83An*g#v-=01 z7GEJSHfKzsmT19Ob(K`vHcMWfUbj^Tj;8rpN;WHA6m5hY{cL-(~i;`a2_-i@N&>{&MF4VFWv8;mqrpxkkBo`F6k|smqLRJYYsgvG9fPB9+Ot2 zCr^64cj3Ix9gJ`fUt(DyP2-cL(Roga#r%b-+Q@Hp0~9@ePL9u4Vg(^ApME7+nlEz( z=WTy=HE=r420)!fQUCDrd=I_&pp;_tWWEdL-*V&$5@mMYoTOh*(R`6B1v}XWuQNF& z8-xNAQoN*%#NuSMbpWkb$rjr&sv8t0=U6Xhf$V%}3)>lp?E^X&@}6}$5ZSV6fz+Br7fkf;)%?saHP;WI`` z1^!iTCL3i6rPZDbw>`yiTyf^i zdum{rq~a_1yVEoN9@#P>w20_lr05-HfNP5BggLSUcR)}&9YJTx9o$0Ad~cQSfgRrc z<(AW`f}vRFa(l6@E2^rZ{*K7iZs{$Bula^vP1q}bNdK_URdDGL_{Ndj4-{a1h6XO( z0%{MmC4?$hRwcpl0K>)7LX|5jx)&ns`k6yK^a2j(@Yi_wtf~EtETLv$jn1IA!m}LW z?q$RU`|gTgHIMK(8~1LTYvy?LF5Kjn*d_4SNl;)~Ntk(OM>7^GTE;YHXM8aW+3~6M zM~L}CrgiOUGfkp31d=5E2Zv@>;&YJbln?ID}2RQkV;`4zCX~pr9@b;?Uh8HRPHLeszBFC;6xvww+ zZCj>4uz!5@s+Bq9{8B{yu124E+wbRqc|JskG=sN?k)_l+-Eep}v2xVGFZPS^)FGoR z>7+29;o$&=ouxH7n6gEMs~No}Q5?aEJe8L-38=v_&NgHE2wpHXv zqSQBr!n|E(NO3$gDj8!TQBQ-(-&1XMFxt>3+KsoVS437vrTY)@pF#Tse%LM5r4Mt3 z@1;XeG%<}A++e4F{(=FR$#1Ih_6~SokX`LWkq83qqr(HTGD!K89;YHR$@3|TcU>xI z)7FMq&#)gFLg5iObxT-K z)eRL0(NPM!DHc&3{tlvf8kY%-mLiLzAx1oBK&U4qVEF=RQeIV7Am?Z19biN|FeOyW zO{VaD$vJH)tCqUdM+hxHg~n5Ho5rJbT3$KJ_0h&&^Lq_}2!vJx)jGk0L7Z>P@b-;U zw+gi1?|1sEUa19qQnrrwC2O~u*g1}N@)|k$hT~-uodvp5MLHw+lC;9*a8?8u4MW1# zg9Y!YWeOvHi{?bqTuID&>7RB?Oh;`sen64=_QqVv$OaTgC5M(0{t_@MC)m^_gjrHl`XL|NZAe7kNZ zk@>`th&&^fiyAa-2EVKC#af4k*{+6ySNl7zX>SNJapx1{9AYzu8{JXG%<>#f+eSsE z;GG*G-W_G+g1FQh_4Rnm2m1!3BMvPmi0H*T2O>*90Xbr23w|qijdHqwA=(f5-fM== zW7%)X`n|p1Tu~!Y{KO@ddeg^Mo%|fKd>TyMQEmtMfB}mfs4vTPQW%BqhY~N+ z1)|(>nV95KrF0yX*PHf9;LTmSE3fjDwTfie1(vp>Ynh{+b_i8H8<;WNp+?u1g4W=I z^6%$5mf)}`$$lm+N$>YABELc8G#W#2&RH3z1@2_ukQ!4wZS+{5lk>DYViG7yjlIHa z3Vm8U7wHv_+X+F5TC3xUx?hb7(W)))s;GiJGOz1%z4@1^Fz+GKzrJD?;^l`ws&EAG z25Txb0-S~yE54ub3LCg6eLXrmOd_bPBwB68BxaTpdLenI1zf^|tQmZH3}G9v`b&z6%&md3|^)F-o@@TJr-&jqiza)X=|&a;;OynAW2Y zsMr*FuEH!$-TMZeob2Ptc)SHdFdi~UvgxHxc1SSDOq^s%4E{c-dPqk2QD( zN2nfU^z0v6`Ng2U9J^|FLm9tj0&zcTAVZ~id-<**=>hxYHxcMstA5;*)Mjto zjUHq{F(G)U` zW&~52KNh!C(yuo=EMGVuhgCWuW#51u9o*5ne-=nFW25dlk{Sypvs}-O6V(q0)J3QG zE;B{L*k@+F{?@{N?*Zd+0CHb->h`TgMEG|s;oX#-ZlN-K2(^+v{cUvAa~i#;{P4JmNCH})52iE4& zhAIM*)F=_*Y+mp*zO5gG6EQuEyAgi615T$c!ixNwRus^ur@K4fJuRG4)h-_BC^nfr z@*_FiO}!X5&of5QgLh>xo3C{uFNXtK&$~;r`QKiGsTpOY94{9ceP)h}eJX}H!+E3k zo$}j>L?L4BG4F+?Ry{YU9}Ecj#2u+^?TrWIz$@Y;E}4pbUYgPL`RDrGJ61q0ZVWxD z^&hKU0=CrZehGT%m0n%TT=33~5!>+svzoF_0NiG&F|5<1+gL&0)Rb;Itc{hVr3`1|csSD;XXMMvvc?Jj`)Yugj zu!Yd7WD7>d&Le!YUQL0yXYO0tm{Ws37Rfuea&cxLtFhMr*@*vLj*f{->f$Xyy7v?+ zr`MNY4Wg4swE!Fr8XoiDql7GcfaI zT{+=%IW7jNK^6|3l;gH}fnp5O4Ns`@0)8^bitjt(zwM;gYl8c;hWB7q2Le^Vu+M=? zCt64nUjf`{BLls6yE%FgH+U%Amm_13?6%|a9If2PkNl*y%qWLOG8xHuwV=iG!&+#+ zeGrilOPO%9&L^TeNNgYy{f>2$nv_6Vuiy#%EjvKC^^pXwixJ3E7Ie(J_ExUkao{+u z8y+(<@MqKt)m=7>?dj&OClM)VV{6d_29Z0uHisc}ZY5{dgon6<_biZxpob-9n->ye z>r;rZoRY=ULld{zxbA0uF52Wt9UV&>E;nF!!WvS(lD~b$5BrR?#bM##&lL=TKmu;! zAN#Z`N{ygk1G8zNeJLRE^qkhUYQ z@%k*Y?qnb@wWrjiq6J*jT~Z4`;a|QNq(FF1Pg9uFu4-Jo!A<*PSZIXVR3@bHCGU6( z9NONs5#DYGGQ-;c^}j$r*Ty&HlnN3&L$q<4wN(?NDvj=Bv9F1(CGAOglpy@cWBT?o z(U#MQ%%{YhDt$O?j4k`0Xc?_D0O$5Q{InP&%_`JlbUY>%;{C{#=}r*U+C58a%8l}y^*ge2dw%AEh`&AIB3+mqD1WKA zWTDR5E$3jVh-j>|T+EEZHxJ@xg9z41(E#|4y?PRet8#?oUUVQ=>&QQs!-{z02;$*? zCeoB(33`BlLLFC==PQxHdv$g#9LBWx|8tyLeRa3Yv+^?+3*NWW%oQ~q;idJmZ^)Kz zV)?lz7YRCo33D+bq3vis?}V^uo4da>%<)kJzbUxnN!t{o9=dlP0v#x=z%|7OMf#M;38vrunTG&bEvw$r2PuH0Q)cvE&` z;IQ2mn7PrScY4a!SKo0)H~Mu0E6Aaf45w2V^MW~^D60ni@COcqTo`SVDS#ujp-Nsu z=r{c3kFuPZldFkG%-Hlegf3u!OknyOwfy3y7QBOpN`JB@V8GAcgFu#`j`6 zjRJyO>IQ?7>Fg9$gtD8%ZJp15IyC00oif^Ard>wu)Nxj4GqHOV#^8p>(qr4^PikfT zyk$>{iEBVmfma&;5!p=`NfYzTI~o|yfPsD@Xsfe1MWr@$!qp9Uo>nF^Vl)rBjHH{I z(h%ab*LFfYa%L)L()FhJZ~LXjCVDC$xXgP>DC=dxSC?pfb~2%f&qSalAdkCn6LXg8 z-MHnO&HgCMp`~~$pOhmfLFoq)-`V~%G zh%2>}M?Hd-^$#CriNgQvFd4^waP-pk#@F;lWWWn<*D&(yb39B}K?zn;51bZoimu?< zt~Buh#hu(&yDRw&Z(Lq`-aBF$sIuKcN&S0VFHx9D?#yAd|EOxe2xLa87}p;$UPscB z1R2ZmT-7n-Z5yF}X{{DVRY*_$p4|k6Y`OAeq-wh(oyAKlQUvyw0cbuSbHLfWHU=x?DVI(39C=Nk;&Ei4u z!c}G4tWTEmBX%I~sl+UM$%vuT4vRx0OU(n2-nw;r>P7l7aS2dTxHnuGaubje6aLJY;>Y7TX2``N6L{6(^#tUUsQ(U4 z>r~|@2^0C7dZ(Dun6BI~`NHqrcoKwbouI&flznw1(|yJ|KnU|N^LTmSbaP;WWyMow>!>E$zC6R%t-_?Wcc zD?C4JiFIUKgtBv1F0(4#P^gZ7s7oHKLOlE8tD8MVdm)I)rjaI)KPhfD9-)Yvx+vUf*y{*)?$QTV1czi za&V@e^VmSPZYoBcRKXV9_BJu%!{d5dN_SFL59xc3nV~BD%uqzxiM+1Q#;pHT3Wsu& z^Z#zy;8sBuB=M{)KDgkit@_pFToB zZXFbtN^eiOoQFnN=yxfnAxv|cle zy*X+ z8f)A1DiRB^%kh}(oMgFZn$@H+THy9_zQ8v@jBrk?sGwE&qT@a&G)j7&$nwn}$*BaY z_)$>acZk4eo!gQ;v>mT3Jw!Fvo&k}qDOKHz8a2<22{ZNuLu>9pORBAJA~Gvq2961q z@|P#*A>`wLSeGSwT(2I1+_ngwXL6ulb0bgErsG7GsAka3Bt%ipD~ZBVG46`oCgQ+O zjxedp68FT$JPhML7S*3%ux)h^IRPEHM?%l9T6x5?bTF@oqxexW+5jaVT24CIoXt1Zu2WMh_2o3nv&UZs5C|V4ELJaD1@&j1p07Sa`A-7|~REh&#p?Iid&Vm4~CmpGE z7dN~1a-N8m$jS0d-{PsgGt4di`+z~HbBTj8ycwsVt=|ho9a4j1iXBD~K`zgdty4D0 zS}UD~L!Z)R$qdPVBY3&CGN#|Ccunrq!bQ(^i>m|Na1#MolRt=89aWFAHDK$ZR!XlQx-xwAkU4-;0JR15uEF-nn|U+p^Caw2jEWD{;(gl{gMz+`I;& zY=BxL2O;GbP!fY@psI3|T$Gjy#e}O8THew97H$b|OC#a~gPb%Ci&EelZB;Wi2h9?E zp{6BA(mfgO(AGkf0KBjiWwN{kDHbc>Kx3aS>A?Dyet)xrIo{}8<{1#3#jwk$>ZUO% ztn;mzLL_~&72D!6i#l)?dy}UPf-Cv0|FjM?lGhy8+;{8 zSg~|9h4Rb9)jLq5D}Klz(A?=!Kpe9xDZ*c@p;!c$RVnb)Al28KYaO$fKA0(}`xafb zkkU|J=z0SJSOzo>ch4B~P_Y~A^o;KlrBUO|C$tS`lblNu_qdH{(sBp!WTfO%2H<39 zW6JO0-3RZlxxgZuSJcv>G2>o7a>H}W*H&%%b%|LFZwwr`Mm|lK<64fl+FE_kw=d|gAuAqh&7|J&9Z<&@Z8436oCUNv_2ZJkr2Nb&IAT=Z`*j$0on)}n zl&}@KJbk0UdXX|enIHqWneo4P)H@mjxE1aNJZ7%mI6@n!nARob=JB^Bc6_sn?7Q=qse%5RHG+}2q~ zH~MfxZcx#G(g;uLLx-{VYRmY@xqa|&be%V3SBEzRoXkc+YR}dzebxl(PjM6dZBo9%JR$51 z58fknah{efQ+hf>oMPh;AXEl~5@He**34P^Y~DBL`kiRE?h8j`OErWa?@fN!2KKNc zvo3(HYJ8YfijKg{6lVY5ZkpoIjJhV+Q}+VpCHayz!SYfL;YE?b3!tpQVXxy@^t zzXA73Y|u`C{*37SBo8wq>Xi%#n6$Acp{gj(-;{)tbHDN-;g4ZXV797oH}ziy zQuRJ7g%U zv@5=xA8ZgxR`YCGE~#D&fkwDsxa4|32Yt4rH07m`jd+_gGp#$oIV%MVl|a3xu60&l}*3 z4UG5aV0%XL-9lOj5HNhfIuKB=+h5WL05QA2?oeGoN_!Jr}g0U`z#^y1p!$Q7?H^zws~%NN+9HO%li1 zyHVZ-6%2fOf?Wu?pd4)RF79<^kx}{LLmii>2*vix_&l~mCK@J}r<)P|pzfEil_75DaSC&5Wii~1<0u9h8#}CLH=A6{V7LxqiA;4W|ZUK%-u>~p63XI{pSRg z;fP9r8ruZ)$4lVH%A7XmhlVK-#MWfj;rke0Wj`sL@P%?Tzs0DEQ^a^lVy*!<%bDqL zNERR6J$btpNWD4XRa@FHy}V57u~gr}z>kcO7kXNkzizzZ@)S=oG`RM@6HGJAbsdgR z#&fck{Ad33Hh^`UQfgvd9Ps&o&|QTy6^f4CeAzBtwLJHSXnwc2wufqilEVP8@2`i)x`8;WRTQBKJ4tnY6wv$|XeE}m-6t@M|o`kq=a8sv#tz^lTE&E{3? zU07{MJ*J;su2n8tn1h%rHUbpIAA2Wu+>Q64xDL)Umots!$zijqaM~i`PhDSq@a|{* zg-8K#WL4XB!{AElm{Ac`W*@TSt7`P2QmEBm5YdYoiqD1je;W)56STu1@~ePS_YE{d z2eg^8l%d4t%edIa8LtK-83fU7yk|-trrX(+v>vwrlMpx zIZJ}*ICtr7U-$RJPT1jKT`3iS7}*@C?_E-I2dfte;iLLY9AjX|;xoTj-lhF|4(wob z@wYQhBaNP7B@~GU-{2y`!tiV)DWh=kaHM=hyN@{~ciBk7r7JPO=BD(AnifdB=++;1KgJ1iSak} z*CFfYnqnXGa5Kt5F5p^4o6Q%IgnCyRw&H+d24O4p+Y7+NJ6%iTSB(&nlH^LsC5ift_n=?N>Uu!ByQ_*yjxe8Eb&M~K6g0i}Bx3Ssh)ZMp2ZTN-nY zVzx^{BrF*9yze`_3%uM?-Qa;=o!@S1e=&L8Ik|@fj*q@HY#R5EnA(Viw4UoU27CQ& zOB~mC$c_KbBTvl3<5pJ8pN+g`(NU{`Sw<@eu>d7Cjwm2gf+b%UuFFOpw)s4=g+FLu z`FI$!hQX4CI^TgO@%%O0oYPLcXh;~VC7=3VeH2Zcun3%y)}b^(9h8cq`3}~g9qMft zL94dC$`b4U)HqMYe6(nv!2gZQ2^uHoB(qZ&4MxLN4`7CA@I*tR=)n3rhNm z`wtcM5<9p`RCppu)PYd*4s_x;M1mA$V?}G1k*9rMb02JLjR zz9v(YOb3$aQftX(IgEhc00~xzwr%#+Wy< znJQw6KwY?@bV}AFqCHCz=q=m}#wf(iHo+?)VQxiFuej7OA|~W*SD)L)*RCGLQw)u5 z^{`}H#&|>{tDDdQieUHSW%q@p7?&D536tVClr;w%c3itCow7}l@fIJn0%*S%Y!`D= zX9hPk+O&3;_sawi=9H^QRbjFHwAS2CkLajb-?unzG7fGxPv%c zphQea55l+e@#4;rg2Z7- zqk&0G$*KMz;nMNCEi+^7`XBv!##K80K6}RWX3Q-~i%Wxd_Mx3$H?L5Z&T$_8zGqAr znE`Hx^sRsgMt!CTYT|q6KlUz;j$D?hKLjp3Lv!$2IMfRa>1`sxRJ)`*6KR|ZFJZ*d z3l|5d1tJ;)&|nWe86AsgSwnoQAO6)*F6UM1Zdc<I_$f1lxMI0o4A$eGhYW7iYFR;(Ve|{{Z)_ zsw2rj`(!n*#4W9!H@=SBAGyS_>NEiwG#EGDZTz)$Q>ojwRogJ+jb+~h|3AAD8mjx^UyGK>@#b?tqpM>OOoA|WF|M{;-XUJhU6 zLR_J?QOQj#;fd_BB{Ap_0!F*M*zz4L%%?D12nAO>zKrsW20mxDC?Ou;!Dh)`4sIOw zb@6R-d1o4pEMg=218u;lyw}w+Bj04f&@a}?w0ZUdr&@6B#S{022={>ss_o6^Dk4qQFOC8WLr<6V!-b&%^a9Y@5=8j0DF6Qc`sQX zF;Fl33t7d)t#sxj*?|0CIFP8p1JPau`-7E?8U9DXdk3MY;bq$^GK@O{?cwbjDza>z zctQ1nUpg_7yeitZn?&(oyZg2EoBCwW+=i?)A`sQE?N$4NIXbF?$d!qO)$mbEGZcgl z;Lk`NnOIa@wHumpYb>0O=bKRY)rEt3La?h>M&&e+mR9AQ$2?{mMMLoO^o-&=IAf84 zWaF9)VT;zen4ctr{}Bo-3y9lh7I>~>{lEs@WkjKx>M&T&q_mY@{uw zo*5MIv6i}D@5cH6YD*FoGMY1}&@?p?t)aJ$ah(uXHN^PhH#P#FM`a_nz{zAxy&dL^ zk%bK?AE3g}ywbD2vE+ftI+jI0B*&i)OL9}~n)F;lOEX5W~3ih*4^)A2%O zJ*7s{eA|r}U;Yh3bI=#aeT-1`;Zl4a@|BO^csh>jDg%%sOKgu0RSX7C>W*h-lF*%i z{H=(3R9qA?ZNSEFVutQ)Q*h`M@q-LF1XR}z3E-+Ev$`^c2&gRTe%s~FB}Il%uFs## z=T#^Iaa3+815B>S@<9bjeN+Iob(kg0NH*O8^>0*+Ij6X-;q^6j4KZx^)VIR75^rD^ zhH#m_4YQ{y;0A*de6v8ewPyePi^l+1QH9glKc~IPx^v5gM+tI9DdOGxKl{DwvGDRT z93@wV=1Go~WfsEvs;7>pA9`78I%bodeyV+y(lt)t> zL+Wk*0~+3pSSjd$n?T756b4CLQ}cK14%C{XAEWudxxQQi=VoX-ml1MghMRV^;1OS| zl*VBF6O9-knwieRkBmrqpfpLO7N!iK_q?^p^Ug!u+I9**irC+`>H_DSSRLVZN*Ywi zbX+c$DTM?1L5LE%i!4gx5Vp^2^Uh})GO*Ue4bWZAW4vR3-nhPEgLtk^Yf7s4ezYzm zf0foH7XvgYZZ2vLxD>l5q1d}Wo!8sxi@Oh9P3@uPPLB!fbCoE=oZ<4C7Gn-x-3aw& zio<;ISLTy%u&=}(IbmZJ#XLE@$R7;#66ne&Ecsm6?jgO6DCa&;L1(b`_ePE|xZ(w~ z2dm;aa7yNhmnZ&jitkxzb1XWC$)$iOfljxKt0jw?hJrZZao_7_7t)C;0bKrIYzsa+ z>^z%@1%St_muV3kMk_l58qwtZ6YHzdSPcW!l#N{lP+Y;54elBsK!6MoAUHF_0E4?* za1A=Z1Rn+u5;O#t;O_1k+}#5N2!RmXEg?V%JK4Xx^=toD{jKV%*K+Uc+g1JUdwpKt zKKFTMo4^!v!sk^gkEVvEl-&Nf_H+N{Yyb?P#X)AWE@=;R@a@K6f;sHNng|468Mex0R zjlRS96Ne>tN#vXTB>eesc4uvc>#{-%T1G?WWr?!5x_1PAefGgd*?>(H7|{4*|0S5B z_t*g&tUvf-;f^a6ahn@mQeeDa+NIf!$NAc22A~lsVXf8y)>O) zTmP`&JMU@-NfESIr((kq~9FvHoM$935rUh>Fdv2$;~v|T;dO(S@& z<}n`n`3pwfycq)?adVw{D0Wy^Z$?MP-M+Nea9XO-vq~(ys+U@^IQFgsyF8S?UY298 z?SF}YMmZLlXi6P6FibUz&SB@M302~agDcS>%RHB%AFNgq)%MZ8hn}v;Sy()lsO3Ae zy0slr_TI<)S2hvQKXHlZc)P#=yc(9TwULhayy^fTA0IxiJRISM1PFly{+mt&z%R%T z{Z9%Jz~5E}Bn$!Im3Oc}!en7q&ekw|adCV%Bm!pPi0`#>Y~tv=DqR2P#4L}OATyXL z|CjnJIpIJV-_X{JQM zh1(;QBgkTy4G3;!19o&Qwzer4GoeDpE2WJ}e^li#aU3aB&}IY|xk7A6UGc={QzbJ} z^Olqo$>@t`TceF&V^&%PU`oEJG}S>>!&rqUd!PhrO~w^8#U`$vxbeoJME9PBbX5ft z=Cutec^I}D)d04(&>=XT$z1DzqbJn0)JTwTPmAoD5llP1E8#)_jnd_5X!V@3@d}AB zC+;&gq)3xVxp;AAE4hWQQ}k5GZb(W)9cKRQm{wmDlgaQIOkNwSv~QOA&4>v=jT~6L zcLj*Yx`pExVSlx>L#wjjAhl*NCGPXFWVGXYRE$r(%ZI6a!C=a0ow~w#_U+s@cURN; z>n$IyEIMpHE^rjRILZL{?ua>cct&&MhH(?WXaHTG3t~V*u1tcjuS#`pw;d_fX#RGO zFEor)6E?Gvh@=S0!^gINCsjcO`NeXcA7c%n`i@`T%7xv`tg$vnVvUBup3j)Ugpm(< ztNZP7M3{pb4TvE%K}TGRh-6;N5G45N-P1BYX%Sj9dMoluzz}S%)EtNWkTZI_ZN@36 zga*G}!_J`3rw~GJQZVmBWCNU|q?!wkOTkNzS0@X99;U5{1Kc`0>*ejX-ZW%o;GfP= z{WAU)?lq^x8gn>!cQhR!e=RB6ovNGWWfW8KwslqY$pUiLl7orvU1GUo@}jE~dhr*n!!Eiv3fILEhqyx-W%f4Wh3B{|Nb8pxxX(D}?kEC+{hAj?L) zd`Hyt{FNw0?11SP$;#`2ahLec;P)4O3396@Qm~=@FFj1=qiuxP$m!$~f(dqprTN&+ zpS;~?j|YywL65A(^Vh)3e2roU2ieX)7&f@u*i%qj2Z=LL*D&${w(yr@E1#*$vChrh_+!xc-6yGZrJf{7HCzc_S$fG!iqD zNzI^_gNcbruA@zKn=Uh;9Fc&NouXe*^r&IW#t{q9*6;6i$aD9tN}p4r;|Ck%$-=hR zN@JwYFE&b&9`(vNncFS=aDc*2G-a@z`Yj#%{)1e6BN3^TTSp8o0d^~C93gYvL_F=9 zNuh^9K;;vK$MrQGX|ZHOyOj_h~@3S zWo>=$uGr%CJa#NGYNR@NG+1B4zm5E~1(g=#Hx}3ZU+?^A%}!42zJFuDr5$ooH^Opp z`9a2Ti+k+1kgHJDn5}KYy&)t?74Qfcnf_CaHKlP{dM>X~ROpPs{8Tb$13v!V?tS6( z25qhm(YE~e2ht>Nf4;G~Vh!mAG7$(2#sKxf{8jg-zaXrGi=>F&=DK6pB!1E^4ab_i zeTIXmt`4uAHEX|w<36b{XT?D)RcxO0QkT~9r>czV=5*mzA6@b*yPmM1Tq?Zo*kr%@ zM+8*m4UAvBN(8EsFA<N~p5okXeowgML5 zc_$m@HqtC=JI!;*?BHUXNDhP0YaINH!cc(LtlIb>@ij}!_h$=fmXh)qCM(W6UK>yN zV&)1d&$u@B5vSV=vClqyL;vi`AEe8(QZPHC*GsJMlH`X)*}I}gu{zIX(`amK@A=Y- zQCyP`1?C!XBaSm0(OhpgQaPw{-L9*gNJcO)m#R)6J7T{0=0%`jD6x2|xu5GVe?@z9 zjI3O#miUusxeKyC!8ANbo_MTs#f>oEvdg{S;-Bw|{tS>6i8_X2CVsZVENVIb6qa?4 z%s{$dJa)sIUkhC{f2Lr@bR^43^epCBx}sn*dO4b+S0scUPI(OTu#)nY4e&iy(&ZI4 z@nvel{-d&QcA8|*8RQUC{xgc=m1zt`*Joa>)`7&9aF(}$6nOM6o(G2>ND=;wn`^xn zZ5F99!wPCn4=86WFMWdEHWH_Sz?wsmb#fA4n_?vsD3szplV?!DSw4%ANOCtQO%b`f z5`f$OVi5Uuc$xi&W?zr^Q^ThUK|WvF*L*2NkW#T}^~W{hSt++$zT$??iodbF&!pH3 zggp3gAu~Z!(Q|~NTBl?lwHc0=2qPvS!v~z9v!*ocl=eL^j3Tuka{FcAzLLU6+&=kx z1JS7Em+pAPX9 zB@28u?b2*_a%;C=JfC6$$|}3{YjOIO5pErwKG*t1$Ns2U^)nMAq^Dl`PMpmeOL~TY zD30uqkPS0!XrW|MDfvLp(vXCs9olx4ZS&JfTZ7*;l5t9`kshoOo~szm)%hz(&z3U! zD%~iJ_$=A#uwt(&HEuqVpMfdtb4G-%+V^(0Qbnswbl7!>>?kaXU!!|aO2nrqBHv)) z$i~`C1v`S>Pr$RxbAE9NtBFGeggKgSUVWSX{pQhDF84byYBxlUyvWfcS9i9Qvp^LU zExQht3(o3H<3Jfh()hj~r^=-?$HE7CwVD*^+Jl3#>p0M4;@DM^ztXx}NcD*n#<5k_ zD1+$i^Ht3rsLyePO>XdHj26!8mM}=EqWo^DVBd(9!u;d}80iiWk!RFjT+{xzi(n@> z8oec&8fYT*dn^3dJe^zjd|=D=bw%|DM)3YDa!mcWdgNYN3BI3NtdV2q6%Y=pPN`eZ zz*KhUu^{E?32}va_J5$;d5{piCFR5RkmSL`CHG?IX=062bU#@AQQ+fNi0G7mgjv}H zG*Huyi=K6bn$T~^6HmuEj~-WfT4OX0^|*uF@B<^#UsVpMwvLH5y`8IV>X{=A{l*`t zpwbzmPO(p~c=F*W_w8F;mu*}h`u-rUo?k`zyuE5$-IW;K*%sOjAI&v$uW2#bx}FE% z1^W0Hf&!UsajD+`rZDazXX8!2kPBh|G<$;IC~FoMKuO4`#400_6NEv8qQJ#3^nCJ; z2`+G<94Eimsh;miDDGFGxGm2qk1+3{_Q8RWZ@kWqpK{29_mSM5Bz?_4TN8fWAIJk! z2rp#BCf-Q`Q}QdAcbLZshCP}zzn_UAr4}l-niQ>J(bkHKbU;h)wq|{<2x7rhfteGi zZ`5kbaYQ>abxB-B{=C$<}jqdRQ+j>Eg#Kj*L@@m#@YnfQAVIBcLcKQuak79QDb)f?Y8b-Ejq<>Y$=5kHd5cR%V%Z$V%=U)6@lPLkfTv#YK%YX9Z-A5gxFHT%f3k$ zh?94dpOP(oehMJVe1%Aa7KyXQa3TjM_U3ZFaDYw!8i07FFgyW!1viq`4lCVFTy56h%YiOFSVj% zS;XIw{Hi#=#uI#eo@Y2UXpkAv5V>p!p=|o~f_2kGRgO6w&SGOJKfLC)DLGN@y%sfK z(wLu>bev}JZg9Dyd_e0BdCuVAw9>O-h;=S+%ifzehy15LViLc9jLx@QFZt-L&o!`# zg)2(PT8Q0cX9q9Of!9v*cbQ&Q?261>0+C;q&XilksysN>w6-Ox^>zwxYxgQXT;Zwt zJ5?_hbVV>cT%4U`39`+6PeizSJP>5vED|jrD?whVkKHVA=!S8nbOHI3FNm6EOWAMn zxqS{~tNo&dP0cclel{)h@nAgDJ68KZuR9WeH~+q8#IezhTN zWjg_{%%zZYSCd`W*iZPOeezrTcCwEEl~L*U@J{oVucvIw>X?P`MbMLaD^TyxcncUd zZ+f2J)xW-}h+a>b%wKt=X6)Bj=y<~5tlg)lq zm+kMKF1)OlkwdQ6eolH?S;=p>rwi?^O*4J|-Cz({-&do`m$u#*Hn%8{IuSi-j}&fW z;#6c^p!uY~gJf=fI9^^$9Iny-7_&axWSI6?4!kE!S9$0qHnV=Hxq;Q#K8B~5XlvgD znbCWHZl>Vr&V1B;fAMQ{^@*65>clHsM4pU7^u!fu;vr-Aee!I4onIEMDX6eZyn4*@ zQ#ATeO-;_4#Lb6%3$SC?Q(i9VyyVY3hgTx^(b7jAK~62?$lLRq=z^g&o>y^N*-qYH zSxMjgvYGqbwtD32eVw90`$-@0<9@$3Z#`AI{I)+FUn=+&YE}K4<5m?^rOyc8o==Op zKXcXn^!5|`W+CNi;U2a6M8aXsM)UU!v@p3{$pJ=ocnHh}V!_r;Y?;=u9@%H8Ss@v9 z?C^fj@G5IlFu9dp-S%CQpSE?WPj}*hbY0fOiGcD!5{Zqn2OgsXH#>k=a8j(3<}+s1WghA zc@lYO7W}W+7m)v-(JxJev$eYw3e0-pVr`=XGvEgD@bmG2KwhPrgRpRMfuUl$b+)!Nuz&t>HAOs2$fPlF{JOckcLX-l^ zc2&3s4Di<~c7VN$1=7w9W&!Z{FGT?!Fb|mhFR}-7vUaw7jSmE)ScL(9bE2a7y#E_G z1fjTx|A89<1tI@~8w!Dd|9}nuAGl#>vYEyzA8F6}CHqOdljZqxjFk#*ga@q0sB|o# zSj^fHGd@>JEw0(Q+ieb6EyGE=ke%87B`d&lH!>KVIu|`k=nNM#a-XbFIu4sc41rxY zN6JV&g$4o4qcb5kwI0z*j(SOcrIy35RGQ#-Q7V4lp)@Fjj;}-#)`7*&sUcMp!s<&l zKUF9%<-bT1idH>~NkQ9MM;md~qgLDUB;-jzwl*%wr6EnoE^qgka-%X~pOaSMPofY7 zw*Fbuhf8yzshk9wi#4Y|LPQ)bHE>{>?KDgrM^@=3Rw-vPvi>@50=^5Ndrp7`#p;|Y z-jtQxbfeG82Vk=goVO4p171kEXpZBFpolV0IsYxXF>o8%?`ZwR7 zn;&+-cK}9t?50~7W7illJTX2gB?LVmbDr)S*%A+~pmXGCbmlm_pu&88oBEzOQ%5ZJ z$~4cJBgdXNqn$FF3GM&XZ40%BBzNZ=owRy-3BDc?Sg3U8RJ;^UB{SZ*RUF)7aL<4j z!6x4bSpPC}#?o{O^#CZsS_yG9@)8GWa!O5H zUs8D`U|(q@`QcT3i%0(8^QQ`n^n4TXM5FHMBGWq0RAQ@DucUAIC)J~s2#V&k^R=LP z4*3)D?6U?!^>Stu#foe!enS$vLUlD(TQB^IZN@3vsXlk7P#ALZ)~o6H3gfd!l@zEJ zYRD}suBU*s_JXSct<8rPWfAXv>&oJly@X#E&)Hu`if|T>A0I$_25~gBTbOn&7#^`8=eMq<#FA!IeY4u`wQFSW5UCpTXZ4EHuf~QQY;b-ZL9^&} z+ItaW59l&t0p$OefBV;m)IbYngU_qvWDWBI81n&u{QRc)yxMRd*xxKZuRg#S2mk|s zs7ftoXC$ifFWI=(Ukw{)ROzq7{9BR*7>k0WK=N_|GV%hld|)t84k8bhmIew5$n(hq zWo1xKO&sv=MNs{y!klbTUK7j@`G0!{`Z)!43@Oxdtehr)w0FxcSZxaWl#f~!1WjxG zhMk%F`5^iLL~@g~$1Fp`Sz7%)u>PR>`ZmD8? zXA0KxgHU@1YhB-RAYFpkBvwX()EM)!SeUQ?LZA)zeIBNA9^qIX{)G-g!m$NFA&sRY zP4r2co?og-S*}m-lMDubeNomR@u+9(BNA!SFsgTwy!;g*cMi%FE4lQ-xsN~QQhv

Hd>YJZU$5qvsc|h= zcopy0wla{CUl6b#ugQ-+glgM-%rBS$UpB6m1ZepQtU-X4lL#r0>OuE+6!CEdDeed1 zdiJQb;*1;34rZaw4sji}R99)380=h_fj@0vZ;(@1Z$&_5zWZ-{L?5 zoGxsnl`DsgXFw84>s|Rst+@xydcc;sjgxUEHK$>#M;}ciN zdv6ZqK3j)hO8@wYD&RA6Db!VbH;Pfrl5RG%m%@Eq^}N~aw#rFLO+(!YanxbT>CZ=Y!p;?Ygk6eEU1Li=RzunDBy%0*oM_SKmTC$Ca?~jGyK(9H98JOhEN_@MzVD;^!)%cX9}&oT zz+n=@>3JABbChPzcM9E0q~2ejf?;2E}&L&oZ?AHN`i|DLdCv==SDF3_#bEn3Q zGI3p0mrwnioav%ZG#u^~d%~&BRiOhE5B`wxMfbyV{BZ~(OQQyT(Yw?T$EQi#xEZS= zmb?=?M|In#%b;p^Xk8#VSJW5m-nHq`!enuyyqh264*dF4CNhQ|x5Jm_tH9YZ)YiIK?W%K(4fOwY8Eg8n)sKoQr=<#*6Zjoy5E}5L?SSruXY*y(F#Maq4 zX651BLywH3?CFwE)61~(i?~u!`bV0{RjSI4xXCw`@eL_SZvNzpv0l=-y5HF4Z5ATH zAINybW%N_AZ!`k@E7Cy93h{#0X4d&TDm<5$i@&|b`ln!?uLSsu)PACJU%)WR`n|BC z8NICg!KC<4Ys%-m$#-e4%L&DKsbm!V9k)iTP1x8l<7kez_Fy*tN0d$bFSVsQq*e*Ym{R-pxQl2{ z1}Zzzdl7uxp`d^L%x7)g?yImEcosP`y>UTKo!Xg{Oy=28)@%ukIT_KrYwr9(Warhj z|8fF(jXSJ;7fHRzvygC*B`zWK7|VmdQIiE^kf)L9{;_}oLust=v#&$pnJNZP#bi*0 zZv96x^s@?>e(wIep*qSid-;CVy~8?vaAZ${f=9QEb^Y@;D@x&-OAem=V|Ku#)3$n8 ze4~Y_KBLG7`G?H;zZ&xv6H!tL-2BXQG)asAfgg+W)g6x4fr*$qC z>cN2jd26Dy`dN1H#vu5hs~i^Qz5!+;GtVY-`lH`xn<4NK(|Er%5bocSIYu&|rAOh| z2tIU6sz#E8>a<)Giusy%O2X=~XvpM_J4Rm*TZ#+)4MQ)VZp?rsCf3=lu!WKg60l?E z7nVwY0o-AQK_TV3^?!NY?D^<(aMHq;49i@*nOhiPl)Um+Y1BmzHphT&6TZ-gvhI(= z@ou}!R0LP_(jL4|svAO$&n4=AMvJ-L^hUGOT~u<#O`xhiV_@wh`5)St%t-1*7K+^# zOLJXE#a(xDi)ntf?zc^ro=}?zSzsi_jsA1sCc8U(LCL9B`MlJb9{|y5xVUIkopJK0 z%C#+}SYSs+-1#`K1W#Po6G|&YXXBq+b_WKIsp2?l@@?DyKdI-?OXS|3qpPe5Mglr? zfn1%UF1>nwX8N~vTh{4hqa>tJY=1u>AB21y^k$zB#D4=D*;m` z^`YLauE7-$PZ(_v!JuP@GH>;=Uv!y1($lyP334mW0M4kz0F7-48lvhQMt<$^!VOiB15`3r?s}*$LiN{nKmjYZ z(>nrAk6kqWuXj^asLFB6usDdIJrwN+okTi09B}h}moooTxI|lC-}ae{5nKiP`JTOJ za#6iZr#}{>l{6-*`yEy$7L5+jX4~scPRq)>f}A)yL9Y~ah<=)$;D^g|<{`~k#cs;@ z+G1u`C|rmpxf`6OjSWdVj_$L}Bx ziQJn9dyjVB(a;~@2f>!A{3o@o<+rg7oFhzP*)pWs8+mW*0?H$^tL!YvNjtl2Et6U!adEW3-c@DF)ng@AXh6z%>d zJfgV^IlT_`jlLL2y55=lfzSq$n@7E<6m(kaRXXsx?d{_u#zYi43HkO#utWwF` zEQu~VR&z%b!&Yeo%ckS%+l_eQGsHM=%XWgI#}L_15>DO{I20|X2mdg-F#3P{>8|4U z%|zx@bZ{RHm8YuMe69H|q0#w?Ik$KMqTI}7r1_wOe8+My=t8cp2@Ec@xj0;w&ncCv zUR3}E=Ms^jafON*b_-7>3!i86kyX@WeQ%dLe(f8BJ<+!yN<`*q?1~+pvTgIc6O&4C?;(t zVsmx^-kH@u%_cb7nH7QJV9318y;Sa43e@}HhEZ!XFCX1jg~lhw`3Q1RHt|Uii26_+ z*zNwmw=CvFVUsyH%j#gBIvL_d`!6g%=}2zvOXCAZFKqru>>#I`X}Km%q(<&ryJmJj zxfVCajRn1sxje-3EQUKR78{L@#^X4{7unfBVCSl)WC14@T%g$ofaEehFa+zk-QxEz zwE11N{;*8>fTCm+j|O-ZCzaOpuW#}f-uck?v+FOMrylvopPEv_eBVNU{A^J#OwLr7 zHu0*f)x0dN)Sr95gdCHZTVqt0ZRRcmCYHGel0E|93DkGQ(x-}eI#{V&&eT;EDkO>VWJaeL)s$U%fw zxX{rJzIH;etyr#O(Lqu>3qQs%XO%Wuc4`z$uvwPa;)w{HUcrQgsV{~6jdMqE^luQx zyn)}w?7uJ;QbIn`l6dm4(=yvtq|D0a!|v>s8hniPIllxJttRGz)@Oi_q1Ig29^Ha= zbEiY~dIJ}HT#uM4uD8h2kSFLK3tE0_^5$3RlmJKt9d^+!f#kJ!3&%0mjB1!Ko7_E*UiM8cbUs8gL)B9y1dT`2Q8y$v_9294(!k&40uCEh96Y(Fo1)M2N-_O=#m$EEWjw?JoY7% zqw1?}1^U6rk`U}egI6}>X^@o-H>F^6$WeB}_1e7pXDGq|5w2)F3f0pkhI8zNwzXXaQ!aiXLjqEt# zrCYUoF3`(W9OO82g{_qn&I^2Ybd)R%j)sNf{6%Ae^_9!@g*n7IzCv(=A&5r?JDA^q z_P3@=mqu&UJR$q=so(0vnDEvPR<5urqu(Rv<)vEtH$omuEFROvbOPb#BevmrQ81n& zZa-i4t5q65`01_24&IN{zo1%7FPvANrp!ziu-6QnNP%$Fx5Eb%-uUG^zaiiggHd7@Q*F2>Q$b_+&nET+g6aML-O91cd)OI&&V)cia7Bk~Lpw{0( zj&QQ$ZQCAmt_}zE4_giSvH%@4iRIkzfu!+#8^~D-6>RBgU)D)VEcbzsYnBA8a)Hea zpTJtx@ECd{mLEJo z_9)%t@eNnrJ8F~{3fKf7 z^9XuAeqv3^d%Jm#$c;^{SC1g^w^N#g#_|m^39B;l`NArBuOb3y2=^lMjry=sw?aAJ z>)`PBS6Y3{P~>2Y6f{WQ(}*|Hxgy@E00cG7@d{x%vxuPRoj`!u(Zh8M=rT*8Z^Yqu z+8txHlk(q$T!$#ISEV+%&kO{sLjRLHT0!GbGYkx>dDs70YlZTQJj(I2Qv+20fuC%? z6yJ&MIk=XG7@&3TT_?SS?ckqzN+%0M7Lt)+#W@22UzyOGvX*=G&*lbrCSL&0DksHZ zJjROeDX{o%IcCAF2#;plF62Iz)Muqlv3^Lbx2fPztl&rv4o z)J2}4WPz5{VCsTc7Z|m6ghJA);wASLSg}*Q7_YZYa)-#r{B4kKyK*Tcc=-|bT1EfQ z0rPWC*+(<3|ITj@xiKf=?2Bg05x0=FejKmFLE*mpg2O;xb5l=O$P4nh+ zQbIBhVUGYC{S&gb1LcFx^?^c?t?j`9{4h~182Unp;dcOLeQur|j^{sme`Vm@v#&p? zO?UP;!)Gf(M3WdDp6lUSuoOK$cM=gpten#dXo|rIykqD#gt!>9bMIjKhieax_jW+k zkm<(BT~Kdp$B6*AF-Y5+fB4#2WxgU3N2+#Ae?3r+B7<3bGUj^ChrfF^Ixr#VB>f`Y zeD3k#>6yO2 zgcfB9S6xxM39gwIo2WBc{A`Ek6X`qGvCDyos)^y8r7@LJR@u!U4O8a9m! zHjD00ANJ2xzijzc@tZ5v2!{N8-P&K0>~|GK!y|rp0{b;PQm-)5u&PF zVu<2JpyMxB#v8%dWwPA(z095ZFBI6WyM!FVe#nu@c<+4PTCa~l9eIa^MESq%B%bnH z+q9%e^$_h0RKYm({h=h_+#xFDxmi1QFBCWmh7DNrdmgr@lTfT4($>UJ+h@bA zCFzmmTL($;UsWAAbDu7is3loKvP03&cbFb=7hVcqGl+gM^P}cXNj`5=YLmCYFyE7I zG6MQ5uL02C#0deq{X~xJY6f28Cic!ACy1OiyBQm(Ad@+N`)PZ2Y2-{5WEvA%q7gl zRR{iPK!f|kb%J5oryJ%Nx|dnZ(om2lg1vl!yBw$Cr2=loO}h&!RIwCxZd$hrK}JS` z3NSlz6A5grIBg@)U?=|=@oZx@)-AYUEbxk|>NM%D#kZ{#+K`KohA1^U&PqjK6ripL zt2kBbqvS#H41AJ_0|3V1JD)oR@Gb~!2F*AK$tGMZ@hC!R^a294J zZgAN&AXQKKs6V}}oOdfL4qdEQMF#?Wf#S8%u=&xV^|Mh5-^+_us~-e30BiR;lfkvl zo_|VJTy;46J?uJoZvWcc8-^RE45mv_)#mVJir$O>A z{TMlvE7bpFp9~fg(68v{FTEDJNNI@pQ+xt0Y{tUng_+dF{k39pmO)3}B6t^M^kajI zm@uUBWP0(LK6^%Bbsa1?`_#Z2S|a(Frwr*G8hLy7>fkeC9;ro*q4!Nf7v1VF@}MQ) z7Gdv)5fsR)Hsv=8ZwIzQMHY5eNYjs9?zVPgaM?Kz^ulRs?LWNCZ2bhv(bAw}m#2(D zwick|V%hi+YOy%rnbyk8WpY@?QoXj?a8UtN_XIG~8;!S8Dmo8DUc#z-qOg!wv8ZA& z0m+keY`(|*!)P3XW;=A418b~NfAtE*?0lR07ufC~HC4uD9}IT+#s4WL=SxRdc&R_^ zhZQkt>dH)k{Z%=_v&LpAqQUG0f639bfB`AYI=H0l9>pCFzupU&6HDT({gsSC+HkS% z57}wx1(`rFheqx6ek26d~{l@0dMC3`HVL|9^1hr7Yu`($$KD zJnRY8@&OLZ`%l=@Pi?9SoVrC@ECZQ^c$XAgfV0hMeky^Q*p1Hcmb$sa?&|>Q1Rp@Y zUm&P94Xg>KAArSAP$eqX3<6j;Y=)48bw#KcJRcjwr{}}k))5R)Mx6-)yj^HySbq-ikny<7~P#S*E)JD%g954U>4clnP*B=j(HPpxWt0xb{nz@kJ9`BMeTl|7G?qESj(nNO zYhcA@wAGBes<1!-mfW)IQ$2U~raLo3a^6`cdoPz3ZBsa~-_OC(70DstP94yx?gI3R;_uz ziRRy`nfF~k6Px8U7Ul0C&;k}J|1TZo-$3VZeq@;>Hr{T@7mF(c+v)c|uw&_k&4_it z4HanK$Sb7vc-B1zQ!nb+8lQp*`z}_jN(1ixa+E8!@x0FkQ?@8ec?#Z6 z-U1fShI~flAyBzy&GKmfia~PcdTD9s)R4YFpZqJ0=%XmEcEABn!UPM^)@XwCjSR5i znQI1F(RXaNk5>-9WGw!c^N zuLuVQ>BYcs#RnngO+JVI+aSZD+wea?B~dcRNH7_l(5}8M+;jNhiq*p@%Y`9<^wHQzbd1(U}ENn?#%s03sZRrAjUzt*kQo*1q;}GI? zm1P+sK`a=jzP!-Yf?5Q}SS+>+daD8Xf4+#xh?>jbd2rRu&0Ic#&_Q--^gIh;U!<_L zgD}gLoHUGfUWME`RMQJW>!NBDu`&mkZv&?RJxrhRkWu5URu4VsM8}_wO!9L z+>g|pHnOxGh?;))wPZ;M;0|J2te2s$Mm@R+4`_VxR}Q&N9CCjlYMXasH-?XiuFnC2 zGaO_x%?7^%v9;4I7Ud!ipbZvTHW%|clP;Q=Zx&>eP0M1vVOF-%YGw9mFg}2B<1Z&C23O8zmWV#h-Al!bA~DV+L^* z4Eva%OvO@z1ve>*i7l{JiiOEfzmIkuVWXp%m1Ce{s{~rtuC7TF9_GPGdJYAie=X3C zE2H7nDQLBSgRjdeA_j%$X(hh!36vfOZT zgh%IV#7R1#BpC3`rDC^UJ1g}0wTZV{_!fP28tKCl z4aQ*+{X0)KRc8yx70jJ+U(NBEmH|G$7iQ8FwB*?i^^cZr47zsv`X?VQOK$f3s1Yt+ z8@|1qn{Z~1DXeY&LH%_aN=}~axN6+~iwh$$)L9N`bUJ@;bEb*j$9R}G=9P<3_uaBu zD17v2Wu;+$KyXe*%oTuU&pxw`*!~1LrGU(5V&WqH*GSg`6+|($+=ITeoY6(vhOvyx z_B5pOo~Ybk{Vf(~w3kjE(w2X%u^#%>*6JUaG?VOC(g!(hJ7P^aGSGgW5>R;WU$0?{ zkF;o5!5!2oM>?L`Y)-T#%KYw9lVRpai;1agSN|g-B%v=s3>%yMJX$0bAKg<(@ER^T ztNmj`LFL#fpvGtU>~r|C;hNBU85>>6&tm)S9)RS5Spc>wMk;SH@n{z;M7jCz;*Wd# zm`EUu-9xM}UDPGuYumPM+qP}nwrv~lwQbwBZCm~QgHF1W9&`;>QpuTD>SXWr2ppji zoi1N!=}))B)6z5g_0`4UDKpJ_RFfiN4>lj^Ik_HN;R{+jzlDd@8r4c}x0`!D#wMiR z6>2RQ!B`%+DL96>SAsW1C=Gp2LVTd|H)q+F_&>5|m`xRuQ9Va(Y80$l!>or3*ruY8 zQ?Xq=u`DETz*dGs9czue=_-&6VvDW@Wr~ye!-5`#S(jz)zX4EI)K6H7S|;agvjt`$qm5Jkb?+leAiz##sZEpj!D*p`}!LX^gzxfKp9E^c;$Y zk^5S5)B@Pl!Vs_*uL#4Y3%ebi2@BB!L~9oFPYVvqTk4#kpSS}X`N3iP6^ux*1Hp<| z7e}*mWz~t2=isPz(r^q9_cDpI7mYA|53|pNlv*+m_2n>gE(Q<=;l>%|dOy``KB{Wq z*x8i4*66=Q4SoQrhue`om55+nUCOJ-zQvZ54@QR^mk@?Q7!IB&&EVJWuB3uDBqdP9 z+VvP`m^fuiyEziN;3iIsv=7H#7yNhs8U{qTeq05{;MeeyEu#MS3|%6etgkd|a-Pe- zXq_jgc<66q{Y|m#dQs9zrK?-AVz^dvq%dN^Lh)VWUD34?u_+9HP&W0!U4^1w{5@p4 zr2m&h%Krb5NEtcVnEnrml$n8@@qfSmUlJ(`6BEP#uSDu$s-dIZN(K30yWhVMCK+g$jGf> zrX|J&Mi&rMo0uJ)PfJ#Vv^X`gJ%=zTGBz*>Bqc?41Ht0bUf+nq-~#RdxSl-)@H9S) zUyg?W1C!HG3gDPWJAXf50mSeGFbSLuqr!U?5CaH3&=(k(oyc4rpT8J_ZDwt%2UYSb zf!Nyc=FrN>*OHx-^AS3zKP~RBc)02C|-vU8hRUWkjAOdotA_6GD ziWSga**KZ6j#_XN?#GN3kTV(n2R|pipYNinh$@IG3#usg$^Qmm0MZVGn@jUM|C`R% z2yYtTFa0IK>e}4&5e*<YwUK1k*n{Ke&o%VQpdkjd*TzetihS@ILkW0FssPCrp4qaD90G z9xnWt~H*cH#c)bRSq=<;=k%iQeh`1q}D`}@=ydt;MJlarJC+dzb1U}XsV zUA%iSm#4Y@YnqIjl8l0^s#ci#VZo!VrP&pKxxUER#r=o+W6mpxh)D+k4vZa;m<-x~ zzfe-tN;c@m#)e;+Yw(vMKYUm$H|rLM_xx8?S7URtZPWAHwzE9AI{YkuBCB(Kp=wKu zqZ5!s^cVBXFz~x&2Id0B01yBNP|(%D`Geot*Lf13`4In648y1Q>h=ntENo3)5Fbi; ze*|=OVsZ`z!ol4M@E<67`z{YMH4ewnz~J_4@f>d$_+#xVsjbWnaK=B4FZFx!X%4OM z&03y?m(Te{PsfOEZOti+O)%o0{@iy2V@p%b`%C=T zN3fvg!MTZ{^_M<3C^VnSDi81o6Js}g=TwL1i7EHdd=!c=-hxaw^oj(GCpSOE@6;kdKinBITGsT+ja7GnU^I3eiUaNsKwqkY+HDi z&WJyElo36$HR3Fak>sJ#s7uBaw~!@w4jlv@exz+DW=DVzbzBXZwiv&Sxe0><<%j~v zP&r*Xz>lP*!Q%}y>8D>* zlqa}bQkBYf%Ss^}^*M;-!j5y17uzXPKiXkNkpei|{L>OC1lOXV3>?x^R|ZY)&b_}V zRm4d~P3OqO+b~~4wcK%6Vk|}bC9trc{;lk1xY5UQC~->GTjVBFv2dv}KZoQ|`5VEF zGaV2YT5WPo*pPkWx_bmnNfQ)*^I79C=_SJI8KNUE;AyYm<9YJ}; zUf{KrRfq*&vmX6pxGjAU9^V0IinSqmc+Ar#u5j4Y9j|Z@0?^L@6W)&L3Cq zcyQnn55Th&@#j~mp?K^m-V+g|F!lVw=Oi2=(Sq~{{i!kQQytipZARxgR-H7qk4oHk z4m~jRlZ3NWrrTif%RY8))>2E++YrQdk)aVQI>|U=?_NPFwbl<$t55NDOMkhU4j_Du z*H8{1Zdsjx*h+30$!mB&>aeGsnI#46$J$*FN%ta2Lk!ruv)BT)+avDZp5E&4JHS7Y zf6%#IZaJ9GQv3Tz3U#!57Jd}3wpAkzUQ02vlvb~wn)ri^8>$K2-S)2jkCBP{Y!aS4 zc7bs|2JGr>5{-znSg&VM&16Qc$dh-(*6bf1@#25E#%VU88v&j!QVSyYj&Sm;vZP~; zlh`Y0ozghnF-AZ;>&E(qEB9yV*N;D!#2<7j6yDNo6wHXyvbSomc~NYxk(+NZs*Lc^-9@K`vj8A2TSevOmCxgV`d=kG{U z!Vgfeg_c)f!~zK1_8}Kr!@hW&mFDidvEhtbJIrnZy#|4W?W4-U2VrVg)e)imi(olT z5t0?WXI-Sz=r~Aj&Yn-|i5<_r#E6H<>#E5Zyb>{a#@rO*A?jz$gTE%{b~UwG1rM(* zu1_gUb`%k-Dp7{)w=iRdb@rVSSH|Ak6AOErRZ6OIG28o2q*_g&ByFoa#u7W4Fg440 zN9Y@)bniSbn~$rA?zNrCcAjymkTW=hclZenK{cVfS!MBCf7=gkF~W$5DoY*zu*$BNM=VC6HSHzx;@Z z&pUGyN)5eI=0FPKzg6u8U zX#jMsxKptY^1u}cyQBw#+0Ys^33Im04f5{m-~CdPX4iCDF{}tszAtgPGMD+EZ2+>V z(y23{CD;-ujA!~WQ>PUmx~7kwLHzSbJ@yU8W0w~My@RTb+$qaV$v4DBi0TjfNrfP1 z#nz>}YqG{H2Q7rtNRd~g1E8gb?gQUQ!uSAhf}BA|;&}mRSYv2C8TkhF@VNcjNmY+q z#+jGca>9q#zUsz`WGbAKdTJe*zuBGE4y3!xgar12H%Rm3~SpX zIaf5YVM@&*vF=*%%0WEEAIV-C+Vp*a%sY!i5$e$SEgs#R(ev2`%p5d&CmR0z1d#|x zfmH;<12zQmAt>hSINT@AV)Ccg2Edar#78QoCJ_76L?_@V0kqNt)_y1UyB1@jFBNgb8^0c>NH!B zg0v+NNeM5I+f{4T!jpCS($SA{ge`j@cA3k-wZR(ChmVTg|AQ&dYBQDpd5%CmN4MiQk6*^$xHjj>V^YO{)t@XUEa^< zJsvUr8dzC6Syu6i;dD%l0eb%40dQS(4x-HUl7}TB>S?;?Oq1c%LX)ugeLT=V<~d3 zSGeJx8CkWCDtJB_o*Zjm7(p9W!V3-(dgiF9jve~8F77@FMkxAM_||3{WF$42y3_13 zJmUddK_@1Qe%ODYr`8zhaCsY(QRasmJ@l(cYilu3KQ=QbN~;Dx5a)@^|@af?h_VY zEHSN8G6ClVouP{(#=I*pR;kg;GM}pCp7 z_f4po(Wa)X$ERBDCu!RAPG&Am?k&NoMoGWu8VZsVuAA|oAMCv~H1#1$ZXbr-6C!uz z-<;h*Q5bHc*Pf>;HmjD0;niTa7S zraXz+z&Nh+c%1Y>Vtg6sX{x0Wx?in01-jAap=>7+-mK>nw-E~)&5_o}VJ@%U4a*F7 z$9t6ATtcSg?l^>jx~xK~-^bJcn+q5^AIJZghZ4a=>c4$29a5kjBg;n3wuqin*>5;i zzvrXwyO%o~kMpbzTw@aekA2<0@TS6s4%IOiw7_@jNNxRGUm(9@t#71WEsR$caC3P! zE7TYx1-CrlhE^M&GHVY(@$1_6B!v2pxHhVSU)A@rlUXm-rSR2TUl%X5t* zPLLJ1K{rHrEA~6T8DIy2Wim>%s0)2N&b~wAJbIndgY@4G1e)S!rtxukhI08erF&W= zSyY*~^G{>w(Db#=Ncl*5oibXuPQ#+TBsE*A%Z0h8YaLkX5R8?w8RdBq)ClZR1y(C= zh5K^C)@#bHGBYC(F(bH*>fPqz8%QF)W)8J=>kOo|4QTz1MkYIgZfqizh;i#6lRfxz zLP8~WlYXUcQM1ZMeC2)AYAzEO!jYSF+Zb>JgiHOxSv@(1=ZwksNcZy6k`y>lJAY%g2O0U747URW&F z@jBz=*TJpCjAfncewqp2#m6!2X?^FgpMkHKeRMFcI`2Z|AU(fkf6Op8e+(o0B2)=O;Xk~h2Gwf}wx2vSZbG^e z)Z>W&sXa{Gbazdl(W?PZAaO)?n$8`)E;25mm2HbCgkRFvw|Z}*H9&PCSHe2Kv(0Q2 z`;E*u_7m`e8i8jo5_y7ApG5`i$&2=7o54=xHeZ@ut)We&%gG@5X(YxNR^RkDci2E_ zdbF2p0Jh59YM$PWOLqZwSu0wV;5$a0-yHyT0Ihv+iA(q1w%=X;|* zYu53n*tB64rmAd?p>1AqYaD#p@WRgc2cPeY>=#g(Y)*n}nr^8klde53#%Wt=4=6{5 znY8w_jpFDOLh*ma20KpgFtAe15r!GJ;?2DanW~KS4-Q&FM?*-LI{D#24%b!jhJ9Dg zNQ&{>0s&SycCS$@%ihesZ^v8#xb))|qaI3f@T|xPq%WPOvIuAJ2YWO0$FH}4-Bbo5 zi4~VtJab*_=YILKd(fzgmg?7lne3=dKdwUZ<1g5P=g6Mf(5hIv)dUm+3tl1JA}!R4 z^DC#3x{JK-lBpo?GArvu_($$!lLKRQ`aoKFpWFVS_X1g?EiBaZ#6{YgY0RTWmmrk! zzXSn<`Ym&V!{`*WHIMFee|C_bL5<-e1oWIfGc)0xgT?MCGCed6fWG2uo8>$BQBDrl zL0clYKm!;`wzl&ARNKDyM`a5gEDJ@^@!yJnA=L;J;XH^xj^A=E79#254K=K!uv|NY z{_XHP-~T3JkR%{^@mNy4f7{9-0vhDL0X`9l!axg3qd>Ys2_p~0-=mUk-Aja>ik3J8_^EYs%)C9P1v1Dc$eB0^eW;R3iZ)TBXuO`Z`{Kb+saF zV{&*{`N3iF_21TT9Aj|_m8>BeisZEe^P4bhBL8joz8H_emvO}$!cP0ib={A)oa10{ zS*tb_8Kd)vkipfrRTXx)*A{h{SB=rW2!EHzoC3&R)op2=W>p3lQz|Ir-cl6QX#_C9 zeKBC^33`d5vNhF6{0TVt7d9B5a;p4Lo79@8OyEWFnP zB9&=AZGk;yY#bD&mIBDm)B!q3_GT`_!j$7CAkJozMPSk!cLENTT>VfZ4;^1 zZ2gWH62snO{2b-`D&t&9XXFQ4KfiK{;}2eaZk}sCY9(cKX+Vic6V{U;INA2f|MxS9 z@S%-LM`VZ(JAQYl_x?sh3Qbl0xbUuufbCzTLcuz)Kg>3MpPS@wU~+=1zmtd@(e!l1 zf2w=j%23;WtAbg$vaw_RMvEYBn~dghXmIRWoF@&u`&N{E3U%8k=~o-OtQ=elQqnx;zVcLFyspXHD*Tx_WmJiSx&^nJBAL!q?^+cb0S+P&;9Du zOxS6}-n_L-F^>Dqo+8`ZUU;n^X3aF()0i zyKatP4&l=BSJ}Z_EVotpL93e?mT0b)@k+CLj0kWurE)z(fZMXJ^bOq44syvjT4xyX zm;h}-bd)qhFm6@dtaB`+32$u7v^?IDgmwmXR#CR|W6=3+VEcwA6mO1I&0U_v^( z8rB!qlnfhg*U7P}A-{vuf8(N)Yg;TMi3H4jhy}BC?Na=N`BA$u_>Ug1D^GKo=A`q` zige~HrR)j2(y?&8!EmdcS4__#N!kvLY7-yST5er z(SA^v68^s~lVX_0X~;IO>$smGniIT`W_*%hh||z)-$CqIiC&m34UG zC-oG}E2Y$qk_B~ioTaK3(8F%b=yj0S=PI%vp7@tX{G7YNhYr7nG);O(hNvaScY~e0 z#e?WiV*VSa<#L0~q%Puel|qdk^a{bOlo4kdpXC4OhZ{A4-91Gb5*2hZNuJ^buME`K zT=&16W7WCayKDXW2Kbo2PbxK&L7Wdz{}A(DD8}!+6}Z1%#LuUVAOYl`ABCdvS6~A3 z4Jud3i_bZfui7idko(B;|Pi}HWm`Th|-g27r( za7Kq5_5pPq{&^5VPxE~75x!CxjHZCmYFGM*76%`e)=OKSpsOH}FkF>4BrO_3>~bkQ zqH2#Q1QC-+!$jmepeq|8Nm1z(YwSH}Tp(%hyB3K0AhRNBe0VZvZvxyB6QOCGW+>+` zHT@{9P7^h7v3A@|a^J7u__)1FSik#zdK9y}R7(Y1_HaOZ6+k~QxsUR*2bLR$@jovl#{KSh*@Bei`TA@N&*!n z6+k1L8-BfD^37$}CqT7P?Hpb5-i^xr`ksAGlE$< z25yK-hx!~}?&ewBWKe5MtEksZc-z?0)TO!pmAsICJr4C5dxo0azR4QaT1`}E7RLbq z9=73MFxUw$l%I02UiT}o<#A#vzDZjah2k-5957Qte2|a{9_w1?Om>ylv*=7A;EjrBQEG-Gz!9k?!9@j&ngl1fX zz^ZUqnMcfN3-O}5R>tzFP>Q!iX;+sG1oEP%AAWp@Ts%c75b=Q$G*hKRDn4KWl0N#E zSa7|f#Zxf|@oBC8Hg~=c*grvknZOQ8*GDRpct0Z(~4{#|NpsgT3&59Q zFB&MY8e-vR-bb%rIbL2yGY)G%(QWqJXBQy(xq?Q#y*ixzr2WQ1wRc56*jR5SaS9$V z0|WrjokLNflv{Q;gnsL#ZL{a$mfTqAj@QCSQukEVpw&dK_F+ujltZpdPB^S&VJEaB z__FJ}3AL(Ic!@%IULlp}Kz38e@?d^kkGW-cSqk~Jk>;3S^4n;{%rOe-AsbE!S>W1v z+2uXL$0gdJJ<}9LoyTX(;Cmnt-5iMRE%-9pE2{l@A}rZYJsgnv<#P2e{rwDYr}Z;c zpJMhVVQvKN!mCabZbQ4qp%7~X&nt_{jm;l!1*vW)+PQhQs0CWrVAmjQi%70F&a2eC z!w~6YS>x-~szc^hXqpx5yA9M5zT#pHm69WnTaC<=k^EGs311|eS>||+)ZpkcW{Wcc z2A65>wq_z(@|hd%1p1z7idX06yVX;;2+yQ7E{<~OCP0YXMtyM>Z6T7^o(;HC*R!)LL)3)G* z89=jVOOx3)L;Tv%N-`^W79AJ~zOY7@if=u6xgZYucOpnV)?5=Zi+DO7z#Udj z*FYaACx(Cw{iu9MhOWl{G+Xo9c?)kkERCR^X+iUuxRCDo60}V;iQT`ER zaEFDlJrqn$4+<&&anH~sdJy}v2Y8@k_a@FHK!HI$*RxZ;+o;&xDK3VlXtmNTlnmY1B zkmx?F5Ppb8A7>aTbo^-GHUSVzP=(P~P8h!?*S^9fDuDolFeRZlsA_ymbK^lzia>r~ z(DcIr#~Aqp2Z(`?U8RxLW0%lc70Rm>sE~JdVh{3rh!Be0MLSv|TdxNgE^JD{>}&Ha zopyst9s$Hh|8y_8|8>rl>!&&2eYtmhaDEQ1noh%CE#r2%Vjg?Be>YMp+h0cKD8KojXf znnkg)Ev#SOd%@Vzkp8i9fv22$O;^(Vp)qFgbvnK*9|j)guU*i|rEp-Yl_kfei}Dpp zVCYZBBv3?W)@NI%n1rWJpo)6aJ17T zF=rfPcaJs|?Mw9UD+rEz(l9*Lc z1Z+WzmxDf?9TU7frqi$*sJhl|r8r65Uy!Qip>NOcf~=x^WRA&LjnxpMZs*WOg4%T% zszFb-D6b=emJS*jG8{iCIF)CJ3djnnJQD?Z;C} zJ(}ZPv_NrIpzPyjU=+TxE;dkAe#i7iAP;GKnp6v7Pn3?YkX}1I-x8-Pu4(EXy?+F{ z7o>aS>Vf$j#Z*&km4h|R&p)JPhg%_O~`rvqRUY1u%&<{oP| z>>4Es-4DDB3T8H2(+!oBT^d2+Shs@cVa?a`OdU>E6#k3^wwnbf`LqAVr;!tEaEfV~ z9xJmuq9O!S!1PxjIa+@}k&gCY9e5`XG!wAD1m+C1tMrdG!_1!LUJ1W};O7 z5dpitneyZKDO#>H)Cz^15YT>a&RlLUwnXv4&uW;HE$p&Uk5E&aJAe!05gp5(@;fkX z3InrYX6g4W+i%wWTI7OFcn7X&UY0=7c;+U@HYEs0x5W7Tjq_Tt_!9MIcQi?bHVIdF zgKNZ);>e5H_>oy?vsS!>Wq;TmdBBOL+Cqn6rIJIlqlre1UFkhy#=<=Or0I4Gm3*N* z6=$^S{^L%URjdzT0}K}rP6Jfh+id?pDZAe5khct5=SMIxxEo?;;p^Iz%p=dzbre|4O z2akA@tO+rdw8-$Lcaqv4iOC}&bag7m>MY+a9Xqw)Hl;1fu)WB){wRB%?%zTl;wrJ< z{<1^4&_KkmZa=Wzhn7Z+I9H%AG*@ih>amx&5up^Sok*Qzk9&=2fY;K>uWYcz8*|T; zbWaEbGBJnw?(-S6zUzg8(pnd?r0>X0m{Rcj6IL5gH25db4`nq34=zTwWrz$Cs+1i^ zJa$p_lZgms3EpWrYAFw@)uVuBl31i4Bxh=ij5Yn5ik_Q-MrOo&n5J)jtGbLZY3)x| z#!ESUqHiX@_8RgjlYhzle(x?H+RfAMU8J=cx~aQa{M6CC@!0}{r61nU16u5T@(Luv zwP~z+ej_ay#wt(>aJ2fo|Bz@+I8eeBYzM zify|-+1EyjOm6FLt;78D{ANdmJRSiGQj-=#ecRA#$M1c^t?WCMFmR&1(<=t=>~ z6?)K>(I#F1Wuz(I1z+v2|F@ZMThhVp1*pJ7X|*)&hA$D*`%hRBv-$bIi&J&)uUg(I zMBMu8#-)fMh~#u;k2#|f50I~y8jPuuC(hk6kwj_rD%hZ$X34H)*L1&^;zLrw}fX@z@yuO?x|`576IgLlhVUW(9I zq{2GxWN{@L5tL4l^KG^|$1(lHCm9Okl^L`_N7+sspTwyZU;kO_Ad6W+IGWvf;H3i} zoI@JM{EW3VqWh+Rpz6ZN4?}karG%Q-->mgoD!=vdMjw8N| z?Lz;L3Zb(=WVUcKNWjctn8HgZ!5WsX3AK@%@m?>H>G}<>&8j7VoV?+&K7zfV@Zf3T z(QP0iNboCu@JnGT?dh@M#BA|pexp$3K8o&S7)>4h7^&|Yvcmzmnn&y@ywX)-_q;%x z;R7T8ULv|dF=?``5Od4shyX;qu;iOIbHfB`NWz$@1Vi&X^`R+JPAS3EuxX1@ z*drZk9Ce^4XtCW%l-(bDNb5X!!+(#;BvJ;%M)eGY6STUW;8GIG!gXA6FBx_sPE!2qX7 z@5^;-iUE4Ru3wV*H$YQ9M!Fdxcp;MRuUp=Vt*53%fw0dvX+rJOC9j5SE7)SkE@#Sp*A&kg=*({eDyHq&r zsWC;!6$c$4QI1Q>ehor{&y&YI{NzG!Y#|XJrtZTT_TyXAd3z&Yl#Lsl#yrNx1$E^g z=-_5s;77&Wzs_Mp@VD$p`uw9!)K)uFhA#KLRW1)s2;q0 z*(+k!RUQG__d0=yjc_U#9$drzg?U`2)Ai8e92)XMw&8ONh@#PXIdV2gIHPDmMzyy5sA+@jxEw?9gE!kA=(jC$C{OL%ogyC`3WLWaZ+MtY zl+50J<5I%p;hsBtPP8KU9k0>X4i3apZ1`1s#5~H;7k{6W8)NkFJho*uM%td24_@Od z=@15jaL-SsWifv{ZI72$V(nM?Ch|OjThq9cK=7Qn_|9d>ZO2Qn3jE=+OQIwzK9=4x z>hka`HA{~)EEJ!f)@4Wt+O$suv}H3i^F-U;F@hkVlgfj0LU(-NQ80FaQ#cuKd7|kc zZPwuWMk+UuUZf;44liVr^4we}KQTi$3q>sQ!-FPAOmq)mrH2Fr= z377mTB^M@~NT@A*9|CM+GFzk0i$-84ZkM$OIs$)!G(gBiwA4PK%y>C1qnrqEZNe?_{xqz5mz{%70o;?a#kVo=z5-s+#g1z^aJb<6*;6pKc=GKpcXB=Z4rTS!r z7?nkYcI#*r>l8X}Bwd^hZm9N6Q13PY8zT;4zEkojQ6ph8!)v%YwpC95X=)gdHTy#+ z2kQCO8uD|ZbC1geEHk!46I%xWTSb*LF0+>r2)V*k>BSbNhEspS%UOAO-5nOHS7zn>k= z^RqQD!);b$yB{h6zgJlnuL*e|3PpoF9cGzHsHsfJGn3s;I|0AP;YcbQ2b?B0atfg@ zRbKezkKM>uOw|gc31R&(Oh61nBjivI^Zj@dd`x_Hn{X+ag=jy%Oa*qyQtPAh6Ze&2 z&)jrT`!{)>{^>JjIj;!m<^p zjic%^4So?YRv7@fz~9K#9wwHE3t+h^gC_iC&GDqk&c2oqC*V5 zq#?|ptlbY{AgQ^9(*gL+!i=r_&A!7#s-4+~#HAU?hZ=^DBSNzn`#-&xmtPiZjQY_E z6%#t;TF(8M@A>+hjn6dXs8GOQV*LF^c^PY}BzdqIzmsz#a7|Hru$H|cV4C9Pj8e8< zWEw2Vx7Ej9_c@or^f!L)P~U~P>IQ1_^XCm7vfgG)7&!n|Vg%zlHj;*pEh5BvFw-nq zLxMN5lW!UD*H!KM%%i602U@4PFxN{uL+6>C@LZ)E`ASS{^nzZ&VVs zdrtRfjO2fckX%o@jU%c(zD}jfV*JGQ6x!c^T#`P_$)Wan9Vn-3=fqgk_bq>gO@?hw9d(JhC{IHS;Af8FAm z9m!Q_vWDRd5C%zPtx#;% zY0~q-IIxVT_p_Dt7&z~uBZ{Um`TGWP=?k|NN}+=&XQe=_?ly5Hq`pY?i3}df;D}Q| zF^_0}i2l;qp4;TaEVpI*js_e=`lMe4V; zo};&jw75^_GuKE&lXBlq0k;`x3J=} zmVC6N(<>QXJf-<>et`hfkhX`lk0WBv0c(o$l}$ehO^HAV33O4C`B_f#UOOyWr%wfG zw4Y|g;Qg>L+XZx`6_n?!UEpTw*HSe>BYi@+g?gXvsc4+i&IG8I`rOItY86sHk%^xJ zs3{P*z8OLIJcxb$lMf56d-Ktq&70ZqEuD#o0_waf37&i~ss*W9ZD6cxyaU2Y#w7Vw z;Lh#dQ?T@3iT2LQ^{pN@atO^aJDKHd$Dnb;!SGTk+~L&V>Thpe-_Z>@OnVNxxAxpK z5wJ^o_x&eqp&5H0X_plk|$16w_ufWu*?|{+MDsr}8;RzhylbUZ|oIV+9qo(%DYD}y;=HEX)dP!?eIzizKhb?H?<(h8VzMMb@Qs57s$rfHuJgQiRO55N>6#8?)_?ru!8Y3!Jfygmh`ysN zCi%gQHhX+)>W+S$z3@3;J^k=AKrcQaar@YN8Zcs{a56!U%HyqAqe*DnBFp}ThUw>c zSDL$L>RO1Q+Juvsx^23Ecmm~;P@tW%3!%B&B6JZRq~_G-)jw+^UUQJI*BV0Ka{*Te zJ=J8NnNvNnYSNjY&B>1VLgmn3E!kRfBna3iZ`))Hd*=H}J(m}6PnOZ9s+ysb7WSgQ zi`|ULuPF(|OdjVGs?9b+c*`vua)l3Nrvsxdm@4*AWInUyNnj}*W9ZcFsuXp#&e6)+ zm2QR|V`lsXR5n)a%^NCBa5J6|g+)0s2}nwZZB!|=bUpbZvZ+#KI*J$LhN%K%rB-Ja zdtGe`{^kvpib$U*%Ym*}vYsT+0w6cmXMV2f=r^QDN}iN8e?`59gD^hdZEQM@A711Q^sE_2W$DHB4`U<;rTb}CLHMcefaFTgb1#~VUCvM3*q;E0{P%RJP<4bQc_`w_IiqRkjPOIC|j_Fo(`2X^CgGx_$x(O zqAQE5tN-a6g{V5afli>V_*G$(&+pQ2P&9#_rPoWO#jB#hp(*DATGXF8m)D#*GohAW zU(YTlD1lK!updr{uh)~@FFM&PfXCh+wCFyZ7`jWkuLoLxH(37X^Aj>k&9~d6swoby zqiU?N#ZA96XG&o3-){Rp{-dNFbB0C9NhUsINN?q%W)?PGXRj&fYljxX?bC-cyBSqI zy_9B44TdCw@Z=zL@8~0ng`-bT{tRCzu<#>57fUA=B`hbrK8*Z4K)h{i(b(YPf6aWZ z3`0PHD{rNEc8oA&#Y~nnwb&ATh%8xRwjt+T2sNmEiw9`!B5wfT@Zx{A> zP;;_|$XWzU^;!BI*ir-N}=qSF&4-QP@%n}Cw8ka6{}&q#MuKQ;#r z6XoJH(?W#T)P%>U34ay?AHVb(-$d*8-HTW1$;s{lk^EaJfo2qxvEMWsr2XB0XKLe{K6f>ZqdJ{%xn6dViZK zyt!7+l*=+x3l2kXx1;(-Jk?o;KpTJq_sneD$XmzHG9E62XHQYzF& zEx@ewd=vZ9X7H_WP>09%HJh5*SX6hp`^|HZs}eD`ga0dwOG1FJOluCO4GVQ1E=g(; zCE9x=&!;=7qmWh*N1>^6TG`geJF<_X7i=pV`%I3ebuL8~jlPu&%F=Yq^zIscj%4#? z#Wl+qm!?|h51Bt)VUlh1Q^cYk3kS6o7kQ&~IvlA9Sq#lbmbr}~-Z}A*&)8g|`Yzm+Q=I~B`$=>ogeU6O+|KpnGV|qKYfA%ekj-qoV8W24Et)kukj8A^ zmn+-H8-33#*KpxAARjh6sGaQ$!SkG!2v(s;oey~i5~DDz@E7nEVng5ji%Vu}QS(hpL8m}hV_qv>!(%flT-yvfJkI@Wl%Kl4=@_Sv6 z{hC_wA}61LGs($npci!qnU|;=P{U>ZL#ytISnk?WLdFyQtDgg@wbi5QtE4pXrShJL z4hRo|cFKAt;-!NONr&R`H4mYygoRCj9LQSR{F7#JM%7XtH943m#`qO!_jUT#SYix? zF{^u@(o4!=VTt@M;k-yWR4k-o@~W{^OSz4wUsx&U*sz~H^*hzE@7L?|*KC;+Da7F) z37sa-x-;wn!?_z=Yrg;^iH)gTu@if0XN#_hEh{n;teKcM!=;q@*@`>4lmUgU`4Npi`|A*{aWByXd8qZlJwqVPV2>{MZgkNIieP?v5s*A$r7 z^N+MMnJ5q%1?o{I!wiEAo@5i^cfFFh=h2&RN#j+%-Vhi{f1eE7ot!3+q=f7;Jsx_ zZTYg|z&sTvbDx?CMw9XtdE|2QXN?`r*=u@rJUKMh)c&>fuzhhX+kh0T!)h)b=cnz$ zZqAP~n;eUgV-SHcu@vQ`=|g!P@INy4xZs}>BRjldsV_CMS5~@uu5HksOfFM|o~?Bp zsF$jbiWfn9%C(kry|1|~>83}Byg_O$r7Tv~h#TMf4f1{BVt2&BAmbSs|KWA#?0SPb zMwU-fh54GaWp($~j@PWw0xzrCXOS3asc!YrWl-4P2vak1#Wqp(9qdW==+oAlBY7cA zM_fS5FctgV6&%k!N@BJ+dXb+Msqx4AqssLpT%HUHd{yJ!SOr_OvuHg5k+w$0*n>RX z&|ngVuvX?f*`pLwzk2%9H3?#N6GCbWBfkH|*gXXI7KV$Sj*}hRwry+2wr%X#wr$&X zvSZu)$F}X9GU#iU)4U&+_qnGkrGN8N{ukWPM*HJnR`W zuLUf*@6-HbIvXJ6yLx&BR{g<-Q_3gZjpn}00GXLQ=!>w#3QTj%4z>4#be>o%Ra(zvqgMn;!Y$#&>OB zFLR8;)(OWq41Rdm6>)-GHqbzPSs(*@S0_c-)QIb{H`~5Ki!fI_P*&zJ`H~ z)3rlI2tlzaHX%@$uv+W(VmT=SPPr;vcGLY0LyYmRpVSfdOs;q_)`o9x$OFX4rxxrN z=LN({2mXvlHNbxq6+}A+qTeXqS2gyByyCRO`ymdOycze}`CuKsO?SyGApVHSOQVB~ z1ou}N2Y2wUBW`Zwv|Hl>>2XZVc~Wuz(V@;q4n&JbePlif2rkpGhEDH%-zyL3&|+{z zPrI2oM?&*(Fyvn=*Rn{|79X(^L#)%eqwxkP8_u5I(-E0!rsf+6`|SB_3x;*zCVcnw zZ?9oFK?Euf>oIE-uk$y0T6bjrJ0#MQvNEe6`#Ahkm(`_C%n`m=qs+pO5T2D6luH8d zfvWp3dGI6}qL+`lDqLH;rVR&Y1&!?PYtQp{oThK(L>q6dAk8o1SCxTV!q@S_ z6;J?+e8W90y;%0=N7pBe?g+{K8Q{>M`*GDsJ&X;K%DQFjG>k7IxrEuX4{l^Z3{A-8 z$?QU&3dxuGPd~^EyNX7+=caxmKV7YAo*3f2?DQtdTQX$RGmam9oYiK9UxOAbZvRrE*;b9V92q<~N&j-9e1JdZ)|KNU zTpT-FKkyeu9@^kEnVG5xL`!(9Sf2`-gWH3@JG;WD8xfG>1jbipgb&hRRw?T~Pc9X3 zGDBRn4|epyYXZ+|>zguF-Wc{zg8Ri`M`rd`hF}6V#wp$h8fmM?f!adgtu^GyRfGZH*T+6$sTJL z?vwEW@sm%%Zo+gApcrgI`31=iAfcOQ6n{z(`3&?d_&&Pg80O)7flJRos_yXmd4MWi z$IHYn^lYbyKea)R`2efHXDCnpE1;LP_zFU7yt*H;lNc-YD zVe;s2+vY{?-_Rznaf#7nuQsM8dnHHdF3+m)Fvvc$C_ zT6glMJ&!UBW16$VP?J>|jmQrAir8I2meY|kd4eS_%wvQ@0Z`Qfgw^Km&p+VGl+Hd7 zAK$tD9tDv3v=I@4j(V&!^CTtf{H=+ZZ5RX%8$DQXLnJTzotRPp{2|{8DX?p}c;@4D zEx`7%EE+na2YljLW^x8kB;66YBjoo9mjCgKBcM-fhVRu-;^?((;}jGiyfsijrmIXG zB;2ndmG=Wls^s>h4l@n93t21*qPy5z}u-}*p zUSR$DJZ|eULIoaKLm;}}Xq2yhe-Bk^qj7Y4al){fC_3mVgLhn)(?t6wcO!z$Y&ww% z4ExqIB=07Cp4d(&67#bMqF#=GWADC?>aUxT8Lw^sO7IB=U&4&*G2O>WH+AGZ04M@j zr&`mWcLrL9^@EoRXXri^$10hrdaoLRzc)Nf1~U)QL;K=5Lt5ii?mCiuC|O}a?lGTf zK{>7PhJA`q?ufs=RLq`V;mk zCFFd+Lp;*H6~~w=oR8-^WXC~E5W)u=`BWi+%@DXzHDX|9Y+!sW=_C~uTgz-iCUU6N zB`jyY3_-o9Y~{DgNm3+Ee%L$t%PI^^neiySbULzHwWB0gsFFoAxD}q>_M$k+*#1dp{U8>oXn5{(jTx$^MZB!KR^2$OCDvIuI9cUv`lbr z#T&bd!u^eE1m>*u;}T_F8>*n6c974PzB(NbjJyp&dXpdaqir*mG_hIQ$R&dSv7X|l zZ_EkQ5t!_3Ipb9k?cMcr+#Zi=(x|2SG7Q0*)XrtZp~*);TtC_MF)`$fEKtH8QeU4V z@Sn%ENZ?9yo75ZS96zF;Ept@~L;BIy6WErG*$>G=W8WfCS3%)?u`58=eE5q6B(-xa z#EW(uGHLdg<9Gg2+ycq0Rv--pl^_I6y^kNkQaZadR|)o1x6qw+!msrY;Vh(8l}?ym zI|wZLZa3U4IN2-$8~Y|zNVBW%Ni^XJC;saB+`I%M0}7A3d0ChUvV-IuMP#Hs_lZ?| zzt!Y&*mzA0wH%0(ULM_^mlbK#Sm5eCajZn;bd`NlS#_CNkWV zScx;X6c}_e?jXUK$5TN%l9U%Wc;;m-7|d>e5E_z9EI#};D92iEA1n&l$Y zLOX4%(`YYul~kD-e%~H5zMpf!TZ6cOnU0$?MsR!9237f=jL@B6y@JRk+I^W6=|I`~ zI~>f-fuDCye$BDu8+#q_D9F5wRuGbX4Bz|eqW8wW9^YjH{%z9fpW)lck%DW!X~`K} zm4cJNNt?47g4!PAF+$M67P0^q4TXA!@jpGj3T~CS-flzI!M1!gB&RI&*?fYY?1RU@ zw(K^s1XU;;zjgmv1hB5*N`G2QPa)}Ts{{_vmjREh7(}>xMGjr!N`QZH3^<^+NR>9yv>t7-}6M_RAV17JxZKuJC z-aQ=ZNi4iBIB-1gqNk{+l;lSYH%=1f%bX_W9k(2Q;1Z--#whLUqc$=|;C50=e zQAbzX9MZ{%lGKUo1Us8eD(g_9qDCy&WgwP!n$~otd4=!q4fx6}{GO;J#+Z$`FnU#L zg?xM;jV|M?1XRu?%)A)*d~>jKjC^&yteU(VWW8l`3rgz`QK7f>sUi_y` zu0AWk0t)N>0R|Z)g+d?5NOebdnw{yNnA-YpER1>5-S?Mss2K|iEheVrH*xNKpE?BN zgd^=;i_G^FEcLCpe3_OraeTYN*4y%pqe)MTJ4Q(hhG~(8&9-84^=+V?&a$*f+(?6r za}L(oCUw?bS=K%l&tKC?lQ>3WD~MMNLElOU?MnW!%~6S#0*`U62X3lSKN07CdJL1O->Sn=FO&3D z8S~|*{4V_cz3qNmgX+?l7G#80W8^xse&EnRybvNdb1%9fCuld{3kXTD;3kx(EyCa~ zv_dcG`|2q(%MH1Ju>AF|oCl0e!LbxUljhS9q{{f@AcH3d#U7&`a_n3J`_Qj3(Uhu+ zP&)hXU$tsL0QSJO(KP3b23427Sxr?_H+`B*ZnBAn)Qb;nWE*IQn{p4`Q}OjHl9)OO z-fEa;V3T3K>e`%}{f5BJT)_g~mAT~_Lt%hY;+E7ngZEyqM`?Ijk3i$Cb(=r|)m?Rh zk^b@UrQLMcTFlyxpu%|njFW9aOKm!Odmq;ybVSBz*GQySzavA>;1rU9fv_{q_Ezek zWB}oNI&e~WlyVBe3!{7hdS)22|I7)Z!eyg<2}LTz*Uvwld0rWk8xlF8Bl*#a5GIy=CBDqf0)Mwx^Hh)l_X5$OF?}`7& z6HnTq;VD0MPpYcZIk|JQ)v(TP{q{mSJrw@D-le4utQ| z%<|vO-)%#jOfg-mex<3si=^G!TJe9!^ zH}T|UdEb+E8F#3>QB}?Sr|mV#jNf76%OCiw)Uh*3h%321qRvQLVAshC`)L%;X9>|yG;wHBCc zJTXN8B20xwb2VZKvA&BqI*@O}&1bvJUje884E!f=Q8V2sMqdXa6Wz+xAW+3P>@%Ko zRI!etenq}+llwSk;E1Kz(*k4+GXCdxP;)xtP*pw^vFs+))~g654#l4ff=6P63$|NV zwE?sZc&jvJsyu(C&EzoS9i@Dnc0{>BnE*oMkFZj%#Mo6PPHGw1`Jv z2LWNgIlXkZTO{XB8iUC{gu6_I4LWwSi?D3O4RGctt8*0? zy_f-MMQU~*i3Y)~XiL@nLrQfsCh~%xj7VH7I^Rj9sXful^o~r8f&|2XBV_?m6m?c3 zYCMafP(JH)y9W}cr1gdtyLTB$>#`?ujF*Ms-vJenm)CB3AGI%timt)MG&5+CTNiEg zcs8Yeopp%LC=pgBLIwSjkN>OQ{ooHeCnb(fRx>lc z{Mvkt+e}09U%OV=0~GdsI;xkTfH-@;T^-e>h-)RmFJ~{1eFHU%V^iFDtzyV4tIjAX z^5{o#>k`3eeL!H*+pW&&z10t=x_d$<_}a+Fc5#@rLqrnGWg0Kn_fDza2^g zsYepIMe13IK&big;I_>mGmWP%(iy>F>DkR3z6*ZL&l2neZhE*h+G~13xW(nbaar^% zv1(2vtmq_$;df?k5zol0Cq*sYM>o#)VkX0^d7q?vikBuN&6PwvX$~KG+GfuR2Lrq@ z8OO9X+9sgfNT<-hwiXc5v>g!%0S~_6$`M8}xC zT<-72)_kZ>HB&azEZW6D#aGS5dXv(dOkgF7$Ga` zP1aCA5eLUjVMtpGabcy6*9ptcTk!*^JF#eDt`v=#{jkn}27aLkn_Sr@XY(r^sz-~@IAd`8}6aCvbrJ?%; z)&(pQ$24*v^+32TTG)T4rM_6kwF163Mn-lBBJKsYJIW5>)E{q7tL~E2B6_0?MKzwH zz<>Xa^fYNLAz_Sr65zqT_Ru7E@Em&*+DRUp&DwbG`Q+f%vC$=v_%?m~g}Xx466
_b>X+P-&)ursP6;bXy@q%G5B89syY5S$v1@VI6CB z<_PHF@tmSg!_td|=OFU2bA9WnUhWv^zfw zro4`8fB@&LNl20$Gp65t(I{02kH-vk4}D_ZU-P6IcLHU4H2yE=5Rzu{P9dqe(Q)WM z_4g{7y&C}uH?hP}Bvq+_{Ct|{mvH^G72CtY)Jy`AWCLn6WmCXSDnzu5gB9=1_Ie{>iIrJ*XT|np36;VT(by7$ZN>7`DhD) zmoihWwE#_~4_5l?cCaV|5_Gz4bNBB>o@KlAs606Vc+M1R(%Y&dtpg`VoU*CXp%^_H zlRWG!6v)I^cqSpe`F-O1nG1Vvq&lI#)dn#wRyuNmJR(%w3A%?SLuCwBjxHmnRo<5U zig8$(i<*Ym&D=?-FT7Z9$Tr@iD6K%PhDBIwr@1*%5Oh#C5o`CMvn9vZ!*Ux18x$=K zYmv`*Od_BNFH+?(iygN_byiMaBIX4iz@3MDY;3K~N~^_l-f_f4#PzJc=9|M$=^K>K zdqyH&@FNxk)pxH( zbC-V&6Q^@?rUw)M1=ldXrGpGseD&899xlMOfN+v@S4}@{vcIb;($U_8yG1Lh9n@q7 zb4TpUDJgxRidlfIUH9WeF&;MJJzucyOkX-+dE>acBE23?Ng;GF{N=cIhl>aUX*}Wd z8zrBhoRzwCNXg*LhQ5mjMqA1SBekFRrL010#b&WhvmE@hM)U0MNkJit-5Tz8+BvVS z$p!4#PKxt8Zlgd(b`pk%M7O4u$wSiK+KzOc?q*RTlK4Drg5BS$_|xe@nyQL5;+LReD$r}c znE?sUnT2-~m*RjugBgkYt{RNdLnF&R?#2W!IlQ?wa}hCPq+EKd*y3aW0r}+s<)nt) z+y-IZwSOkPyDb-Lod+WILmfI7V$`)U`#rQ(cw9E_zH#+k$&BgI-ypo9%KjPY@9GGe zM^-e6<=O*Hw+~LBUZ5n=&^mcQ_p=sb<9!W3gr^mM)J}qZ>w_P8TuUwO8*=h99nJn9 zq6{V_O}mO0lj?>qF1e5J+?rSi7knw!Vyp!xua^!|ZdPkWipKtVH-&V1g8uR_vjiAF z)TS90-uT}qB;ptK^#Rz6`exRHfr`z`t|(-m_M+Xn!y%_3R`i;0M3(jyc#tP?5R-qZ z(64T?miChScu7ebAKrw=p!BCjP+rljs79G*-?8K~W_lxM813kGl*1%?$9jsuiPAe+U$e$uukd9xn=$m%w7T=>vW4R;BY9r zo7l_k?&K64!DOxUkJ_2O+KF0mutgVoH7Z*!HXh2o{zL!o^q_oi=W^8zqoy?=W+D^O zocJip@8)8j3e6tnEDZ*ak2xZFnATrse$wYpP|yx;Ij0>0>P@%_Vjv3yd+ZEF>MnQy zN+dRa3kSgbK%h3KKvAcSxK26*(@Y}*NkW~y}DoDnZ#VwGug)>!ITti+0p z?9aWGRLCNz9AJo`$_s^>VL{|Q^DeM4LJZdRaN7ebsq!0m6d3Hd3Un^~sH*y1SntbZ zJowjz{RFB!jV6sNG)^@k?#oYncg~I&bVy#`-SEw=#&c?l-h}eYMtxfMMR|DP;TQ0^ za>Kfv|8KP>1)e%gtSp4~t=0Rr!BMa5)DJI6Ya`C+XFhxyIL#4}%+eCK;e=9yTR+>; zCW>pX)JVq5Ck=69UVhM`YAhXwNY8TwtsE%*wca;8lnam=Bnsf)JP*)=op50a&pc?7 zKADFfNA)#e8%$@G^aMS)_%zZS&d5?L7{Ct+6vrNsFBG8VE}TRLWYO_Ug}t7BC2Lr1 zbM?Qu9%3Oc2L?}DwldiL20X$=x5eV63aBZv9G3-U^;Zzp?|=7M{DC>ia2ue=HOCp; z4@n9At!G^fhdBV_$Fk4#g2H5bWM^S6{946BoP48X`K|MOd#mSttI&XF08tqncAWP6 z)0Qr1!<<*)uU?M~RpNXzRn~Yl(Lz1~&Wm99l zs;;$IE{JV^A~sQCqP}}}2)(e2v;%L;a9m+?Bh1+$;a?`Pe)QaPa-l9`h4XVX3S|lm zdX-P3vsBV$^1RvOZl5PRU=LDh>BeI945ganM;Z6i5Vwqa#xIg2pBq0SZ7rkM>vDLr>9pXD>T~ie=}voSf_XP@)cR#l0Z@04kYMmczaq zxd2PeTo^?qwCf=0igp}~L5{00JHR5V4j>izs(n{bRV%y~CFwaxy*j zXT!>>Cf5L5%R}Jn?aOW6ShoDol@OLko~o<7IBbiF-;0TqRH>egHaXg0d;Kowy8?1M zn>+A3Lk^yDR|*;W9^LN?k#+UR3(+SoJY4)>{<#=4J=+rfL4mu*$)c%bwM}wXHL5Nb zdoRoCEg<1fEihNjsuw->#Cw+DFp5@jTrXYHeYfzW05 za%roFcA>?3$EYk#DvRovk=9QG2DQq^0N3bB{q> z!z)Hv#385xknHbE3CTa^J#=5U-X47*LIO!Rg=x6Oxj&y-Edn6@Tv-O;*>`dhKqsgI zKC166{g-k!oR;89C;H>!a3*p`hD$_n$3|O>$@1uaS)c@$T@RcPh2W@-fd@8cxNAq8 z`p!S>eT(W5PdsoD_-w)MUkEn_h#PIv4Zz5Xu#>A@I2tGG+XIW=bSwg@mS)}Cr@KUo zDiC0Aw(3F9kO^r4Y`@^XOE98R#77ZMU|?;3h{dcmeh-1?t8C2KzpJ_M_4ul$mYl9X zhT0{)L@xH47=k5hV5kq7e)zaN8OzhP#?Hzvgx9^roqi%Ysc1T<$w5h6mwP;7xt%}s z7bqhN0v_a7K!FQy(#!HZp+8rat>{yO%D4DdX_zKKQocvvCZ~RZ25QvKT3BAs6e`WQ z`VxxTG%^gSGyx?R_mzxEFDMQ7)olT50zqEDLqOdir$YsNw%?8_zoML0`e^1kazUPhC zEZ}7GX2))caw_a0=R)(MhN*Ha95vYzT=p{B-^jIfra9UfRtE@6lZ*#PL|Y6(OVV7# z8paq11rHPy1{W)%o@8zgFfgMwG>0_?Zs5iaIfldISIiBPj)}>+HwLsXsGWZhwg_f$ z2pkX2jZ^8q28soo9O?%Q&O>3YMIus$z%jQnHGnShlt64_ezRv`Xmh zd-n$WmP|uaMN*kpK{PG?ae@O9^QW1cAK#bX_}Lfx?uWg#vm&S{jdf-BV*^sihc^d)?V6offWNxCfoXyMAdd{M9E!-9JMHd(f73z> zjtpR0T!4Khk(#}l_R*i%>oARs&-K{sQkcY&+#%v1rh5rJa_Ebp$% zt4rY0&=omB0vP4pQJn0wqyiem-OE#u$Gi8k0|6vl5}v-Hfwec|IRSQ45a_=kjlW+6uwLhXaJ>C)dFDE8u(bg(>sDSbUhle=L3&?$7T}W4P}*h&h||H$GSzhXlDCyPI7^B*jJCg%Tx#eXqZ1!xGI|1f=ptO0-?nSaDCfjj|zL^B8` zuXaPg`5f;Y`(Sf8KQImT-uZDpcQ+Pd5D-A!;{t^Lq2eR_ZxtWX1Y$1h3)DB6<;U#~ zVc;15`B-qnPt<#j_8VBBUGoQr+4u2Vobf>1uG6nCZu|MipI{pI2Uvib=if)C2cHP9 zL=lq$yW;w*etigvo8NF5{O_Kh;6F||Xy4gxezvBeehi;Kn^4~a6Em@{YJx5>NpV@b z0{@qj|4eic_H}UAo9Fyf1Pthni}y2u`330kns2Ok1hUBB6o3fL2)c?xujcO3> z_*})hWh(fI`oRtVPba^v(c=Gf@+C(k-LxY9Ir&d$Mr1yJQ82Vg-NG=(^fBko^Mv#) zC`kHCF0yb9w5x^CqYb}W0jVl%F4l#V_132}7g14^)SRhBVrAIrM4gxV+p;n)q$x>t zngh&P^WQBwK>6IRv{L}U1Lh#>TdZS_aJBMwPcGz{8td#@;I&1}c~A6b2aKD?Uznx0 zFC;$t^?H5#5YL(aaPnoKCwl52|I^7Qi*#fN+CvP+#_P#7r1^#Gl+RlFb)|JbUN5Pw zb>+3r+d+3Oaz%#9-FseRDR>>QC2vE|0b~}_Kn4doDUZDri-K>IHy)eMDvj824YgiJ zo1G>g6ce>1F1w7Q>6cK13k{{@VDHDSE|8K+>W4oefQaC2Lgy}~6@)wJ8Jbw))$+YX6Ft+kO7i=;+U+l1WGsrS`xz(Z`vKy_w1M=oknqRRf7w9V9g zk=e3`=iMl;+nLYv)%8gD;K#GF$n>hNFP)#vj}^ThK;8vdih_9U(oTYLmXZGlfX`T4 zLBNC-d8|oINGh^RC*+Kp_^ca5AV=7`SpK@>f#+B*b!iC#iHrdhp9Kd4Zo;R5H^i@7 zii=-Nd(*0TtZSTkc996)7IWuYO!+f`_xTT37B86OyEui5t~}+XP6KVOy}0fgjMqUb zFfa6*?!PIx>|GXtCu%4f))I2_13KnHlmpmt5T-q9A%i6tUIw$r=Z3Hbm6;1lh{GF0 z7~PK{``hY-Otk?17YVkwFM>*3YZHDX!KNa+<#Yh)!f+o8irmEW>1ysgSw9i;e|cbJ8!z~rN`!8`X#3@~IH!ZeIH-!}VoM&B zmCBl9rbf#UFQ0;z>^iZuFw>H0}+rHaNi-{Behy?~kb!;j!7a=2*;GS=D*oUUm zYS+g?>t$Ojg|V<4L&;L|%p^u{^Ocm+8+i$hak)G?VI(dOW*+k1!erkP0th}rJTIGxqdm;21$|0S9?GCi>Do{9l!NrWz^YTmoWdCLFK^BYja*2No0j zws(KdW98gNj>nqU(@haHrub9Y!FS+aa?I+iPx{0~q=I}FejWE#_`VEAb@K9YG&yfW zaq`0Ep8X!#^|bTbdCVBn?%J3^k7gz)F0Z+-{_e>tqzt{EDf<^Gd~PLsfxTqzCYfWr zpC_TnEphMwRei<1*t=nrsw>S(iO6-=RYONWYL%Q++*9qeS0Nwk?a1op&aYE5AQmwk zVW^ko8Sm4xZaN$|!=pXUO~pcBK_%6Tuf2*d9_dmKufqyf_v3Uh^L z4gfAq8bU4E?d2ROE6Bbav`NIEt#&5{{g)q-m46UsGvyZheNkQ_D(_@Ry^)Qv8_89>Ug|$?wkJJg#x51%1&+@q zgBOIXNOjsFH+8hYQr|h4eoyI3RDkp#+oVMXH0N40VPR0s2KKr&f^RzOk5ENFMLTBF zD_Z1L`n~A|cyKHL0itnVMKaakV&8cZI&Ptkk<>t{C&VO-v{4dr#a8+i3Z{~}uEMk( zyzC|ky_UUBitS_F#>_f)h|(+3-zAB_V8>^2>A9@&*ER8#Rw*%hL*3YogQdp0GnqL! zo>mFYzJsDmvYzt|I*&slJfowf3H#vtkqgQz=R&cvm^7Zt-VkI87_(@`#CZr9HaP&8 zid1je5y&!P{IUsKA4!cYnQj&DgdnbUxoJ;vWL-f5Kh zup2gUh?DaD5G%?ubquK5_b&^PKl`78i*@f&%m{o_1^{j3k4qry=zt^2yt?qc(HN5p z4+uc0`Q6_(YmT$C4z2@pye=X4T~?W+D*%53 zs3n&QkZA|nxZ$DW(@WjZu0rPXfazPEx_-wgI-K#zWyWga*fD!xk4C})ax`!kMURaw z=V!YVLCL=1#gBXEiN!#V$A+~Et4Y)0B#Z4QAerDBM97|XZmK7o+vw>K%RV){#K@ba z{6Z+nt{=m?zpqAeI&dLbqd2n^`eJg(AyfV-S8$>_|5lwr5blx#602#LD#e5l$jIPh*ndJ0cwI#GANe zxemNYF>H>8H(cS#c89lAp59MO>s3%rO6*JO)r!NYT$#cg!Kxn7-~p(ZtTyZ?;5GME zgP?B~qpyC_^0gGJSiU)ST@Dq!Gw29>CXdJ<9>dhQ^ToOpNN~GZf{|TV5(!<^XH2#1 z3304tw3VJsGHyENdC<;I;%HRfi1JFfXekV<;B3%TemUZW+c({XHNUoqFcuX{d{fgh z_)6?{XmTNTIv^tCmE$0GfYQw(t^rq)Cro9XZgP&)^4dujW(1C)TVL{*O@ss(f)8~) z#f6YuO?&Bpumsp-=r7SbJqD4TLd63-zNQM;Py^4n`x)h~M25fhWy(Ay*K5(?tpj%V z&dTXB*LvJ6r+|{ENf7lO3D?xQVE1# zjr=pP%E_Pvgtv8h(!1*FcG`Jhfb!U8CL((v)K?-_1P9O08N(&L^BeQWq!DNZsd+eN?C>P#yVMdTzY*Nmg&f3`rw(?Qt zo?Sm|GP#Y}(x_+lwC*h`YwEOe9lRh9R>Qg_6M=}4>2nom9rv_<*l71HiB~4k%DmvZ z=H3c1dV&i(1TX>>c%JyBfsiX0J+BH z33@Wfp)c4!1X%4A?&3=j^uSYVeO%{Gjv8U@RtZpg2pDgc;~*5Ycs&g%ALJirs;Q)5 zOhBf4Nv#!-bLgb2s)NUgKa~ZVa7-g7GpL3K3p~0+;Nx)wBIXvG;$jPS$2P*BAb_lit z#eaC4U@j413wRwo=SzdY>l&EEF#oERdeM*kI19=&`0vj{-pBjE^ABe!L)>RwP5q98 zy;78%=ucsQkV0Sob$-qNZ;LgXeY*#YF=G;=g8PUNuxN=!N|iH~k1J+6<%aChHuG5v^Nr;bHWDN%hC z)>c4AVFy>`h@=eUGQn8<_+eKk^hbR)C^-Fvt((Zino6b&`e>iOo(7UUo#1Txpastx zn(Q_e_sRMNoI>lflv~Tv?YbiyIRFfV1;PQ(6J~h~7#6u+Tguyx{DJVQap;c?gt%s> z#G(1}-7%(=#_0%dRu^+$7a{AJ=EJfRm&!6e_485}{lI-Ib|&;M%89%jYObXL`{!zv z-jyMhF^$@BUzljvv&K-XJPE681G$B5K>bX$8p@PCqOkYLvv%0G0)85Vc)_E53}2hG(Prl8N8R3|-D- zmx!PhJnGwms|!m}8wjOPiY0o7RN0s%&^_7;qjm4-wU{g_^%o%380DJwx)nz~tF+eZ zq~;fO_w$Clnq@pc#u8%ZP*$}veHXztArktBfFvJs6}a(J45+QEyLvKG>(`O)?U}9W zUr5dE_Q#aS3CXe?^cvo3ucG#oL!w;9$Yy3S2L2q2fOJzPP?{Z*k9nHVF5npCL+=U8 zycwy>esZkjiWa!KJXz#lqB?hYsWsk_YlY2=ny)^JkGG__?y}Psy0Y(-rm?Z4)>Muh zdOJjT2u$HRim!yhM6WDMF@Z4jhaP{0?UX*zBxyiH1UsBxnQz5i^`kT$()g&f=<$CE zon}pL%F^jR(a~s*_17U&9&>x=-&C&xX}FkZuWRR$s-Pdxhs-N8%b1eLDKmxH zydOJOGdd*CY>`vt?8iQ~0fC}|pC<)q5OvI+(}_ zRd!b-`;EG2L^%JOB?4WnAw6OKp?XPK46me1SFAC?Hrq{VA_uhvY+x`r%ODjozr*B$ zTIpMRBdD1otRib$+tF4xvC7=o#A&DEk%NX5$c&j5f3sLv+9Z%YBZD6m?+FZ_ zClSsO(bqV!vQ2`ZmVlQcRE9gy5OwOIfdnPEKf+JFqlPEJ0NbP2{mS%_G~u_%ZImv1rym-?cx%0 z^y$Z-x~{JAK`U~<>F49yS;JRgY(DW=M$0xq?FC<1@&vL!N5qq`f#h^GFL75hSRfSh zcbnJ^sM-!e;D{x_><46sSg?Z;U(A*h6C$Avw)XPQrY~Fyp|L|(nwLq;vNgQPgr|OjyfJ!I#>T8h2b1f3$C!5zLD9t0#Etn&MGp88} z6=%`e5z;vf;`iHOw%sNz0X`cM954496Ba3A%KCgZK?!t>#midQI089RV^o z92tiJ);PhG1zB&gIN*pnt!I@$DkLK1(@`78RTz%x^Kwx!ANYlc>OzzIjyQ#SQw3d; zY%mEeOY>}p5nh8gWaKU7;PcxSOw#gN)Res|OH1|LUq5l}l$!8<4IH^Hgt)c3n^QK6 zZF1N5Tt$D9WVf(d8>r-l9!%lz%G_7s+^s@)=o1I=jSGi!22{2yaiTh_ZrY+q?EM{w zSW5QhL9sa;4st6zVCEdARLLF~dSxGCA`q`dC#bC&$3E5lX6y?^=Gg%z5WB+;HmG;F zrOwmL`2I;I&wATWnwxX~dUwpgjYoSXByAPUV6EI^dTLW_AZ+=yZakvop5mxzonEgX zmG6>oB>Ux@_OKscS*<1pA%mO8&PN0a$U)9bcgZvuNVA7rrv-RB&Ni#-RAZSV)W+*t z*vLSt^q5>d%?2#xeJ83E%rHcBazapA?Fp(5RV~L~{4BY7t1{)ZfW&4JSWR0NtPfun zA+pMs1j>%#7pVFaWKAI#uHLghI2zw z=dTb;n$>p|{lPVT{R6$cDpP|+F*^hwha1r}l#jg{-X-%suuwI*0Ih~8Fsd_@VeAvK zbvB&w#$r69uac0AG?=MxdED(rC)LXz7nYITlJ@bu_?2Z5&nRVZizr3p!=x*JNpT`@&2MdVbqR4>ws$JRc<_#ZACYE4&!iHYQ93NW4Hz< zv76b7J+fDtz-EPAFtkXR1T82*!T=Xm`m3f*I-Zm9E-4P+!~sJm=97_KIPq3f@rX*#KR0vdq@Vyw!80V2z{R$?NnZzIoEH$ zsf2C_lNjQO@Dbn6PnuzarI&8ZTCZeZk=FN3_b6kDvZ3N+6-BjSVT5KyWOc%B5uB)tRS?APb>pA7i%8E zPs16>JfxI`h{{_B*oCU$pO1zetV!BD-aXkMn}vzoNFp; znw4Rpx{?-+mSP(Ny}6cfW*>`eZ}KB{(W`EoIo+W#H>7q zyvYPr{s&H>3d)AQs0;Z(M~{M)!u4p98ab+Sl6Gx(gq6(+F8f9DYrD|%4I5h}bbBtW zv-Hz@_ze+5AV1hd!{ftEo0`hR+2?2e6q0JKU-jUL^BXpE9UlFg8cYOoh@>l4%WL1^ z>V`hJBnfpjy@ipKj@0#T^s5PQW>S;iLhv!1yvLdSsO5X96)=|QuYd{TraYdAp1Q)? zoou=H(30jax~+N0k5}zy8C^0+hWqJxlZRDD@$Q#3_%g^CPHL7Ybc8u!F8ST+kKyNd z>wH=H-i~%1xI5oM@|Gul+A7!@BYu=nDYynD2~_2raR38@p*L23D!7U$jQ+S#)fEuH z`Y*w(Tw!rW^0NkAm2L8NBlG`a>>gr73Aii(w{7dSZQHipuWj45ZQHhO+qP}bn@Rp; z7PFXTrBb!3ROP1bcMj3CYR6R?eUvN}wSdFL1aNlkd9lI^NCm$(#tabx0^y!ewe~Ai zZp1EI{PMgUo~+IvrefoEcUPCUhTsl;YUY7W;!!@Rx1YBnX7;aZ>Z#SRCi`WqVD+Rp z4gUs&p#HmRk2GX8HZv*>66Q*n8`V1X@|xHddN!ZXMk3(V#r*?w=ENG)x&j-O4B}oBDqcx+m1n$SIona$Gp#6SvEC-2&$^DL9kgBDcRLX@EF>A$23yJZkcJZd&y2+aHZHqli_caE3n#7gx7lux=SGu=B>lOIjf zRQAs$c#;F7)AP<|@VZyQB8shrs!f#vv)TTF(Byw>q(xK57WEuiQs~aS%{P4CI1fAK zpz{?(OSZJaYR=k)Nx@0^$Ln7W{u}sNkAXhIPHT{@Hnm;S9o%aD+u^y z5jIBdAaK-WQz7sq)Brt4WRakt#8;WkLMClBkquwW^55}onxE@Ngun0&@Sp$?4Kln!m{9uv5Yu}3-o?-q_64k0E^a@-4KB^ zvmdL$n-HLRsZw)v^ohigloJm(5PwuXr4yA@25nbxXUam!d^$y!KXUc_&~lSWJ$(H7 zc=z6P>WD-s{5ULqc^E3bVHN*`i>}%^b+*Zqv0jdhM)*+pwmQ1bj7`qQnA%C(o-5TY zWMTskDAE+1=N3FD!jeoIbfOQK%MpK2KIg8bK9TS3p& zbh!n#!2+rLD8pmMTN|$?xR+rnjgD~I{Gev!g+-QTNj%AiB!xWL^SG{sk$S^_|Bi`x z>A=!?8&M<8t+O~KEkn=_4n9TEfU&-BnJxwPnsM