From d0b6368026c668c55cee7f4cf1b5ed3a71554d00 Mon Sep 17 00:00:00 2001 From: mhjensen Date: Tue, 12 Nov 2019 07:21:32 +0100 Subject: [PATCH] project 3 --- .../2019/Project3/html/._Project3-bs000.html | 464 ++++++++++++++++++ .../2019/Project3/html/Project3-bs.html | 464 ++++++++++++++++++ doc/Projects/2019/Project3/html/Project3.html | 398 +++++++++++++++ .../Project3/ipynb/ipynb-Project3-src.tar.gz | Bin 0 -> 193 bytes doc/Projects/2019/Project3/pdf/Project3.p.tex | 420 ++++++++++++++++ doc/Projects/2019/Project3/pdf/Project3.pdf | Bin 0 -> 231840 bytes doc/Projects/2019/Project3/pdf/Project3.tex | 392 +++++++++++++++ .../Projects/2019/Project3/Project3.do.txt | 257 ++++++++++ doc/src/Projects/2019/Project3/clean.sh | 3 + doc/src/Projects/2019/Project3/make.sh | 79 +++ 10 files changed, 2477 insertions(+) create mode 100644 doc/Projects/2019/Project3/html/._Project3-bs000.html create mode 100644 doc/Projects/2019/Project3/html/Project3-bs.html create mode 100644 doc/Projects/2019/Project3/html/Project3.html create mode 100644 doc/Projects/2019/Project3/ipynb/ipynb-Project3-src.tar.gz create mode 100644 doc/Projects/2019/Project3/pdf/Project3.p.tex create mode 100644 doc/Projects/2019/Project3/pdf/Project3.pdf create mode 100644 doc/Projects/2019/Project3/pdf/Project3.tex create mode 100644 doc/src/Projects/2019/Project3/Project3.do.txt create mode 100755 doc/src/Projects/2019/Project3/clean.sh create mode 100755 doc/src/Projects/2019/Project3/make.sh diff --git a/doc/Projects/2019/Project3/html/._Project3-bs000.html b/doc/Projects/2019/Project3/html/._Project3-bs000.html new file mode 100644 index 000000000..580a1ec54 --- /dev/null +++ b/doc/Projects/2019/Project3/html/._Project3-bs000.html @@ -0,0 +1,464 @@ + + + + + + + + +Project 3 on Machine Learning, deadline December 15 + + + + + + + + + + + + + + + + + + + + + + + +
+
+ + Project 3 on Machine Learning, deadline December 15 +
+ +
+ +
+
+ + +
+ +

 

 

 

+ + + + + + +
+

Project 3 on Machine Learning, deadline December 15

+ +

+ + +

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

+ + +

Department of Physics, University of Oslo, Norway
+
+

+

Nov 12, 2019

+
+

+

+ +

Paths for project 3

+ +

Defining the data sets to analyze yourself

+ +

+For project 3, you can propose own data sets that relate to your research interests or just use existing data sets from say + +

    +
  1. Kaggle
    2. +
  2. The University of California at Irvine (UCI) with its machine learning repository
    3. +
+ +The approach to the analysis of these new data sets should follow to a large extent what you did in projects 1 and 2. That is: + +
    +
  1. Whether you end up with a regression or a classification problem, you should employ at least two of the methods we have discussed among linear regression (including Ridge and Lasso), Logistic Regression, Neural Networks, Support Vector Machines and Decision Trees, Random Forests, Bagging and Boosting. You could for example explore all of the approaches from decision trees, via bagging and voting classifiers, to random forests, boosting and finally XGboost. If you wish to venture into convolutional neural networks or recurrent neural networks, or extensions of neural networkds, feel free to do so.
    2. +
  2. For project 3, you should feel free to use your own codes from projects 1 and 2, eventually write your own for SVMs and/or Decision trees/random forests/bagging/boosting' or use the available functionality of Scikit-Learn, Tensorflow, etc.
    3. +
  3. The estimates you used and tested in projects 1 and 2 should also be included, that is the \( R2 \)-score, MSE, confusion matrix, accuracy score, information gain, cross-validation and/or bootstrap if these are relevant.
    4. +
  4. If possible, you should link the data sets with exisiting research and analyses thereof. Scientific articles which have used Machine Learning algorithms to analyze the data are highly welcome. Perhaps you can improve previous analyses and even publish a new article?
    5. +
  5. A critical assessment of the methods with ditto perspectives and recommendations is also something you need to include.
    6. +
+ +All in all, the report should follow the same pattern as the two previous ones, with abstract, introduction, methods, code, results, conclusions etc.. + +

+We propose also an alternative to the above. This is a project on using machine learning methods (neural networks mainly) to the solution of ordinary differential equations and partial differential equations, with a final twist on how to diagonalize a symmetric matrix with neural networks.. + +

+This is a field with a large interest recently, spanning from studies of turbulence in fluid mechanics and meteorology to the solution of quantum mechanical systems. + +

The basic structure of your project

+ +

+Here follows a set up on how to structure your report and analyze the data you have opted for. + +

Part a)

+ +

+The first part deals with structuring and reading the data, much along the same lines as done in projects 1 and 2. Explain how the data are produced and place them in a proper context. + +

Part b)

+ +

+You need to include at least two central algorithms, or as an alternative explore methods from decisions tree to bagging, random forests and boosting. Explain the basics of the methods you have chosen to work with. This would be your theory part. + +

Part c)

+ +

+Then describe your algorithm and its implementation and tests you have performed. + +

Part d)

+ +

+Then presents your results and findings, link with existing literature and more. + +

Part e)

+ +

+Finally, here you should present a critical assessment of the methods you have studied and link your results with the existing literature. + +

Solving partial differential equations with neural networks

+ +

+For this variant of project 3, we will assume that you have some +background in the solution of partial differential equations using +finite difference schemes. We will study the solution of the diffusion +equation in one dimension using a standard explicit scheme and neural +networks to solve the same equations. + +

+For the explicit scheme, you can study for example chapter 10 of the lecture notes in Computational Physics or alternative sources. For the solution of ordinary and partial differential equations using neural networks, the lectures by Kristine Baluka Hein at this course are highly recommended. + +

+For the machine learning part you can use your own code from project 2 or the functionality of for example Tensorflow/Keras. + +

Part a), setting up the problem

+ +

+The physical problem can be that of the temperature gradient in a rod of length \( L=1 \) at \( x=0 \) and \( x=1 \). +We are looking at a one-dimensional +problem + +$$ +\begin{equation*} + \frac{\partial^2 u(x,t)}{\partial x^2} =\frac{\partial u(x,t)}{\partial t}, t> 0, x\in [0,L] +\end{equation*} +$$ + +or + +$$ +\begin{equation*} +u_{xx} = u_t, +\end{equation*} +$$ + +with initial conditions, i.e., the conditions at \( t=0 \), +$$ +\begin{equation*} +u(x,0)= \sin{(\pi x)} \hspace{0.5cm} 0 < x < L, +\end{equation*} +$$ + +with \( L=1 \) the length of the \( x \)-region of interest. The +boundary conditions are + +$$ +\begin{equation*} +u(0,t)= 0 \hspace{0.5cm} t \ge 0, +\end{equation*} +$$ + +and + +$$ +\begin{equation*} +u(L,t)= 0 \hspace{0.5cm} t \ge 0. +\end{equation*} +$$ + +The function \( u(x,t) \) can be the temperature gradient of a rod. +As time increases, the velocity approaches a linear variation with \( x \). + +

+We will limit ourselves to the so-called explicit forward Euler algorithm with discretized versions of time given by a forward formula and a centered difference in space resulting in +$$ +\begin{equation*} +u_t\approx \frac{u(x,t+\Delta t)-u(x,t)}{\Delta t}=\frac{u(x_i,t_j+\Delta t)-u(x_i,t_j)}{\Delta t} +\end{equation*} +$$ + +and + +$$ +\begin{equation*} +u_{xx}\approx \frac{u(x+\Delta x,t)-2u(x,t)+u(x-\Delta x,t)}{\Delta x^2}, +\end{equation*} +$$ + +or + +$$ +\begin{equation*} +u_{xx}\approx \frac{u(x_i+\Delta x,t_j)-2u(x_i,t_j)+u(x_i-\Delta x,t_j)}{\Delta x^2}. +\end{equation*} +$$ + +

+Write down the algorithm and the equations you need to implement. +Find also the analytical solution to the problem. + +

Part b)

+ +

+Implement the explicit scheme algorithm and perform tests of the solution +for \( \Delta x=1/10 \), \( \Delta x=1/100 \) using \( \Delta t \) as dictated by the stability limit of the explicit scheme. The stability criterion for the explicit scheme requires that \( \Delta t/\Delta x^2 \leq 1/2 \). + +

+Study the solutions at two time points \( t_1 \) and \( t_2 \) where \( u(x,t_1) \) is smooth but still significantly curved +and \( u(x,t_2) \) is almost linear, close to the stationary state. + +

Part c) Neural networks

+ +

+Study now the lecture notes on solving ODEs and PDEs with neural +network and use either your own code from project 2 or the +functionality of tensorflow/keras to solve the same equation as in +part b). Discuss your results and compare them with the standard +explicit scheme. Include also the analytical solution and compare with +that. + +

Part d) Solving eigenvalue problems

+ +

+Follow the discussion in the work of Yi et al. in the article from +Computers and Mathematics with Applications 47, 1155 (2004), and +use your differential equation solver with neural networks, set up a +simple square, real and symmetric \( 6\times 6 \) matrix and find the +eigenvalues. Compare with the solution from numerical diagonalization with standard eigenvalue solvers from linear algebra. + +

Part e)

+ +

+Finally, present a critical assessment of the methods you have studied and discuss the potential for the solving differential equations and eigenvalue problems with machine learning methods. + +

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. + +

+ +

+ +

+ + +
+ + + + + + + +
+ © 1999-2019, "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/2019/Project3/html/Project3-bs.html b/doc/Projects/2019/Project3/html/Project3-bs.html new file mode 100644 index 000000000..580a1ec54 --- /dev/null +++ b/doc/Projects/2019/Project3/html/Project3-bs.html @@ -0,0 +1,464 @@ + + + + + + + + +Project 3 on Machine Learning, deadline December 15 + + + + + + + + + + + + + + + + + + + + + + + +
+
+ + Project 3 on Machine Learning, deadline December 15 +
+ +
+ +
+
+ + +
+ +

 

 

 

+ + + + + + +
+

Project 3 on Machine Learning, deadline December 15

+ +

+ + +

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

+ + +

Department of Physics, University of Oslo, Norway
+
+

+

Nov 12, 2019

+
+

+

+ +

Paths for project 3

+ +

Defining the data sets to analyze yourself

+ +

+For project 3, you can propose own data sets that relate to your research interests or just use existing data sets from say + +

    +
  1. Kaggle
    2. +
  2. The University of California at Irvine (UCI) with its machine learning repository
    3. +
+ +The approach to the analysis of these new data sets should follow to a large extent what you did in projects 1 and 2. That is: + +
    +
  1. Whether you end up with a regression or a classification problem, you should employ at least two of the methods we have discussed among linear regression (including Ridge and Lasso), Logistic Regression, Neural Networks, Support Vector Machines and Decision Trees, Random Forests, Bagging and Boosting. You could for example explore all of the approaches from decision trees, via bagging and voting classifiers, to random forests, boosting and finally XGboost. If you wish to venture into convolutional neural networks or recurrent neural networks, or extensions of neural networkds, feel free to do so.
    2. +
  2. For project 3, you should feel free to use your own codes from projects 1 and 2, eventually write your own for SVMs and/or Decision trees/random forests/bagging/boosting' or use the available functionality of Scikit-Learn, Tensorflow, etc.
    3. +
  3. The estimates you used and tested in projects 1 and 2 should also be included, that is the \( R2 \)-score, MSE, confusion matrix, accuracy score, information gain, cross-validation and/or bootstrap if these are relevant.
    4. +
  4. If possible, you should link the data sets with exisiting research and analyses thereof. Scientific articles which have used Machine Learning algorithms to analyze the data are highly welcome. Perhaps you can improve previous analyses and even publish a new article?
    5. +
  5. A critical assessment of the methods with ditto perspectives and recommendations is also something you need to include.
    6. +
+ +All in all, the report should follow the same pattern as the two previous ones, with abstract, introduction, methods, code, results, conclusions etc.. + +

+We propose also an alternative to the above. This is a project on using machine learning methods (neural networks mainly) to the solution of ordinary differential equations and partial differential equations, with a final twist on how to diagonalize a symmetric matrix with neural networks.. + +

+This is a field with a large interest recently, spanning from studies of turbulence in fluid mechanics and meteorology to the solution of quantum mechanical systems. + +

The basic structure of your project

+ +

+Here follows a set up on how to structure your report and analyze the data you have opted for. + +

Part a)

+ +

+The first part deals with structuring and reading the data, much along the same lines as done in projects 1 and 2. Explain how the data are produced and place them in a proper context. + +

Part b)

+ +

+You need to include at least two central algorithms, or as an alternative explore methods from decisions tree to bagging, random forests and boosting. Explain the basics of the methods you have chosen to work with. This would be your theory part. + +

Part c)

+ +

+Then describe your algorithm and its implementation and tests you have performed. + +

Part d)

+ +

+Then presents your results and findings, link with existing literature and more. + +

Part e)

+ +

+Finally, here you should present a critical assessment of the methods you have studied and link your results with the existing literature. + +

Solving partial differential equations with neural networks

+ +

+For this variant of project 3, we will assume that you have some +background in the solution of partial differential equations using +finite difference schemes. We will study the solution of the diffusion +equation in one dimension using a standard explicit scheme and neural +networks to solve the same equations. + +

+For the explicit scheme, you can study for example chapter 10 of the lecture notes in Computational Physics or alternative sources. For the solution of ordinary and partial differential equations using neural networks, the lectures by Kristine Baluka Hein at this course are highly recommended. + +

+For the machine learning part you can use your own code from project 2 or the functionality of for example Tensorflow/Keras. + +

Part a), setting up the problem

+ +

+The physical problem can be that of the temperature gradient in a rod of length \( L=1 \) at \( x=0 \) and \( x=1 \). +We are looking at a one-dimensional +problem + +$$ +\begin{equation*} + \frac{\partial^2 u(x,t)}{\partial x^2} =\frac{\partial u(x,t)}{\partial t}, t> 0, x\in [0,L] +\end{equation*} +$$ + +or + +$$ +\begin{equation*} +u_{xx} = u_t, +\end{equation*} +$$ + +with initial conditions, i.e., the conditions at \( t=0 \), +$$ +\begin{equation*} +u(x,0)= \sin{(\pi x)} \hspace{0.5cm} 0 < x < L, +\end{equation*} +$$ + +with \( L=1 \) the length of the \( x \)-region of interest. The +boundary conditions are + +$$ +\begin{equation*} +u(0,t)= 0 \hspace{0.5cm} t \ge 0, +\end{equation*} +$$ + +and + +$$ +\begin{equation*} +u(L,t)= 0 \hspace{0.5cm} t \ge 0. +\end{equation*} +$$ + +The function \( u(x,t) \) can be the temperature gradient of a rod. +As time increases, the velocity approaches a linear variation with \( x \). + +

+We will limit ourselves to the so-called explicit forward Euler algorithm with discretized versions of time given by a forward formula and a centered difference in space resulting in +$$ +\begin{equation*} +u_t\approx \frac{u(x,t+\Delta t)-u(x,t)}{\Delta t}=\frac{u(x_i,t_j+\Delta t)-u(x_i,t_j)}{\Delta t} +\end{equation*} +$$ + +and + +$$ +\begin{equation*} +u_{xx}\approx \frac{u(x+\Delta x,t)-2u(x,t)+u(x-\Delta x,t)}{\Delta x^2}, +\end{equation*} +$$ + +or + +$$ +\begin{equation*} +u_{xx}\approx \frac{u(x_i+\Delta x,t_j)-2u(x_i,t_j)+u(x_i-\Delta x,t_j)}{\Delta x^2}. +\end{equation*} +$$ + +

+Write down the algorithm and the equations you need to implement. +Find also the analytical solution to the problem. + +

Part b)

+ +

+Implement the explicit scheme algorithm and perform tests of the solution +for \( \Delta x=1/10 \), \( \Delta x=1/100 \) using \( \Delta t \) as dictated by the stability limit of the explicit scheme. The stability criterion for the explicit scheme requires that \( \Delta t/\Delta x^2 \leq 1/2 \). + +

+Study the solutions at two time points \( t_1 \) and \( t_2 \) where \( u(x,t_1) \) is smooth but still significantly curved +and \( u(x,t_2) \) is almost linear, close to the stationary state. + +

Part c) Neural networks

+ +

+Study now the lecture notes on solving ODEs and PDEs with neural +network and use either your own code from project 2 or the +functionality of tensorflow/keras to solve the same equation as in +part b). Discuss your results and compare them with the standard +explicit scheme. Include also the analytical solution and compare with +that. + +

Part d) Solving eigenvalue problems

+ +

+Follow the discussion in the work of Yi et al. in the article from +Computers and Mathematics with Applications 47, 1155 (2004), and +use your differential equation solver with neural networks, set up a +simple square, real and symmetric \( 6\times 6 \) matrix and find the +eigenvalues. Compare with the solution from numerical diagonalization with standard eigenvalue solvers from linear algebra. + +

Part e)

+ +

+Finally, present a critical assessment of the methods you have studied and discuss the potential for the solving differential equations and eigenvalue problems with machine learning methods. + +

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. + +

+ +

+ +

+ + +
+ + + + + + + +
+ © 1999-2019, "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/2019/Project3/html/Project3.html b/doc/Projects/2019/Project3/html/Project3.html new file mode 100644 index 000000000..0d1b21df0 --- /dev/null +++ b/doc/Projects/2019/Project3/html/Project3.html @@ -0,0 +1,398 @@ + + + + + + + + +Project 3 on Machine Learning, deadline December 15 + + + + + + + + + + + + + + + + + + + + + + + +

Project 3 on Machine Learning, deadline December 15

+ +

+ + +

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

+ + +

Department of Physics, University of Oslo, Norway
+
+

+

Nov 12, 2019

+
+ +

Paths for project 3

+ +

Defining the data sets to analyze yourself

+ +

+For project 3, you can propose own data sets that relate to your research interests or just use existing data sets from say + +

    +
  1. Kaggle
    2. +
  2. The University of California at Irvine (UCI) with its machine learning repository
    3. +
+ +The approach to the analysis of these new data sets should follow to a large extent what you did in projects 1 and 2. That is: + +
    +
  1. Whether you end up with a regression or a classification problem, you should employ at least two of the methods we have discussed among linear regression (including Ridge and Lasso), Logistic Regression, Neural Networks, Support Vector Machines and Decision Trees, Random Forests, Bagging and Boosting. You could for example explore all of the approaches from decision trees, via bagging and voting classifiers, to random forests, boosting and finally XGboost. If you wish to venture into convolutional neural networks or recurrent neural networks, or extensions of neural networkds, feel free to do so.
    2. +
  2. For project 3, you should feel free to use your own codes from projects 1 and 2, eventually write your own for SVMs and/or Decision trees/random forests/bagging/boosting' or use the available functionality of Scikit-Learn, Tensorflow, etc.
    3. +
  3. The estimates you used and tested in projects 1 and 2 should also be included, that is the \( R2 \)-score, MSE, confusion matrix, accuracy score, information gain, cross-validation and/or bootstrap if these are relevant.
    4. +
  4. If possible, you should link the data sets with exisiting research and analyses thereof. Scientific articles which have used Machine Learning algorithms to analyze the data are highly welcome. Perhaps you can improve previous analyses and even publish a new article?
    5. +
  5. A critical assessment of the methods with ditto perspectives and recommendations is also something you need to include.
    6. +
+ +All in all, the report should follow the same pattern as the two previous ones, with abstract, introduction, methods, code, results, conclusions etc.. + +

+We propose also an alternative to the above. This is a project on using machine learning methods (neural networks mainly) to the solution of ordinary differential equations and partial differential equations, with a final twist on how to diagonalize a symmetric matrix with neural networks.. + +

+This is a field with a large interest recently, spanning from studies of turbulence in fluid mechanics and meteorology to the solution of quantum mechanical systems. + +

The basic structure of your project

+ +

+Here follows a set up on how to structure your report and analyze the data you have opted for. + +

Part a)

+ +

+The first part deals with structuring and reading the data, much along the same lines as done in projects 1 and 2. Explain how the data are produced and place them in a proper context. + +

Part b)

+ +

+You need to include at least two central algorithms, or as an alternative explore methods from decisions tree to bagging, random forests and boosting. Explain the basics of the methods you have chosen to work with. This would be your theory part. + +

Part c)

+ +

+Then describe your algorithm and its implementation and tests you have performed. + +

Part d)

+ +

+Then presents your results and findings, link with existing literature and more. + +

Part e)

+ +

+Finally, here you should present a critical assessment of the methods you have studied and link your results with the existing literature. + +

Solving partial differential equations with neural networks

+ +

+For this variant of project 3, we will assume that you have some +background in the solution of partial differential equations using +finite difference schemes. We will study the solution of the diffusion +equation in one dimension using a standard explicit scheme and neural +networks to solve the same equations. + +

+For the explicit scheme, you can study for example chapter 10 of the lecture notes in Computational Physics or alternative sources. For the solution of ordinary and partial differential equations using neural networks, the lectures by Kristine Baluka Hein at this course are highly recommended. + +

+For the machine learning part you can use your own code from project 2 or the functionality of for example Tensorflow/Keras. + +

Part a), setting up the problem

+ +

+The physical problem can be that of the temperature gradient in a rod of length \( L=1 \) at \( x=0 \) and \( x=1 \). +We are looking at a one-dimensional +problem + +$$ +\begin{equation*} + \frac{\partial^2 u(x,t)}{\partial x^2} =\frac{\partial u(x,t)}{\partial t}, t> 0, x\in [0,L] +\end{equation*} +$$ + +or + +$$ +\begin{equation*} +u_{xx} = u_t, +\end{equation*} +$$ + +with initial conditions, i.e., the conditions at \( t=0 \), +$$ +\begin{equation*} +u(x,0)= \sin{(\pi x)} \hspace{0.5cm} 0 < x < L, +\end{equation*} +$$ + +with \( L=1 \) the length of the \( x \)-region of interest. The +boundary conditions are + +$$ +\begin{equation*} +u(0,t)= 0 \hspace{0.5cm} t \ge 0, +\end{equation*} +$$ + +and + +$$ +\begin{equation*} +u(L,t)= 0 \hspace{0.5cm} t \ge 0. +\end{equation*} +$$ + +The function \( u(x,t) \) can be the temperature gradient of a rod. +As time increases, the velocity approaches a linear variation with \( x \). + +

+We will limit ourselves to the so-called explicit forward Euler algorithm with discretized versions of time given by a forward formula and a centered difference in space resulting in +$$ +\begin{equation*} +u_t\approx \frac{u(x,t+\Delta t)-u(x,t)}{\Delta t}=\frac{u(x_i,t_j+\Delta t)-u(x_i,t_j)}{\Delta t} +\end{equation*} +$$ + +and + +$$ +\begin{equation*} +u_{xx}\approx \frac{u(x+\Delta x,t)-2u(x,t)+u(x-\Delta x,t)}{\Delta x^2}, +\end{equation*} +$$ + +or + +$$ +\begin{equation*} +u_{xx}\approx \frac{u(x_i+\Delta x,t_j)-2u(x_i,t_j)+u(x_i-\Delta x,t_j)}{\Delta x^2}. +\end{equation*} +$$ + +

+Write down the algorithm and the equations you need to implement. +Find also the analytical solution to the problem. + +

Part b)

+ +

+Implement the explicit scheme algorithm and perform tests of the solution +for \( \Delta x=1/10 \), \( \Delta x=1/100 \) using \( \Delta t \) as dictated by the stability limit of the explicit scheme. The stability criterion for the explicit scheme requires that \( \Delta t/\Delta x^2 \leq 1/2 \). + +

+Study the solutions at two time points \( t_1 \) and \( t_2 \) where \( u(x,t_1) \) is smooth but still significantly curved +and \( u(x,t_2) \) is almost linear, close to the stationary state. + +

Part c) Neural networks

+ +

+Study now the lecture notes on solving ODEs and PDEs with neural +network and use either your own code from project 2 or the +functionality of tensorflow/keras to solve the same equation as in +part b). Discuss your results and compare them with the standard +explicit scheme. Include also the analytical solution and compare with +that. + +

Part d) Solving eigenvalue problems

+ +

+Follow the discussion in the work of Yi et al. in the article from +Computers and Mathematics with Applications 47, 1155 (2004), and +use your differential equation solver with neural networks, set up a +simple square, real and symmetric \( 6\times 6 \) matrix and find the +eigenvalues. Compare with the solution from numerical diagonalization with standard eigenvalue solvers from linear algebra. + +

Part e)

+ +

+Finally, present a critical assessment of the methods you have studied and discuss the potential for the solving differential equations and eigenvalue problems with machine learning methods. + +

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. + +

+ + + + +

+ © 1999-2019, "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/2019/Project3/ipynb/ipynb-Project3-src.tar.gz b/doc/Projects/2019/Project3/ipynb/ipynb-Project3-src.tar.gz new file mode 100644 index 0000000000000000000000000000000000000000..731636864bd749b729ed5bdf932b693b59cb6c76 GIT binary patch literal 193 zcmV;y06za8iwFQRPs&{Y1MSbv3c@f92k@Qu6nTQttzCB(^x#1l;tO=HbDi4EwgdO} z?gR9sco`z}cli?%LbBhi*1JvQ?k-piA%rsqV`e;^k|c9IiQ^1nn(~CPjPV3WOes$T zAj_Tf(pe|Wa7x`sZKZxYH;h%)hduKv@XSAPsFj6n-n$w#XmqkbJVS1 0, x\in [0,L] +\end{equation*} +or + +\begin{equation*} +u_{xx} = u_t, +\end{equation*} +with initial conditions, i.e., the conditions at $t=0$, +\begin{equation*} +u(x,0)= \sin{(\pi x)} \hspace{0.5cm} 0 < x < L, +\end{equation*} +with $L=1$ the length of the $x$-region of interest. The +boundary conditions are + +\begin{equation*} +u(0,t)= 0 \hspace{0.5cm} t \ge 0, +\end{equation*} +and + +\begin{equation*} +u(L,t)= 0 \hspace{0.5cm} t \ge 0. +\end{equation*} +The function $u(x,t)$ can be the temperature gradient of a rod. +As time increases, the velocity approaches a linear variation with $x$. + +We will limit ourselves to the so-called explicit forward Euler algorithm with discretized versions of time given by a forward formula and a centered difference in space resulting in +\begin{equation*} +u_t\approx \frac{u(x,t+\Delta t)-u(x,t)}{\Delta t}=\frac{u(x_i,t_j+\Delta t)-u(x_i,t_j)}{\Delta t} +\end{equation*} +and + +\begin{equation*} +u_{xx}\approx \frac{u(x+\Delta x,t)-2u(x,t)+u(x-\Delta x,t)}{\Delta x^2}, +\end{equation*} +or + +\begin{equation*} +u_{xx}\approx \frac{u(x_i+\Delta x,t_j)-2u(x_i,t_j)+u(x_i-\Delta x,t_j)}{\Delta x^2}. +\end{equation*} + +Write down the algorithm and the equations you need to implement. +Find also the analytical solution to the problem. + +\paragraph{Part b).} +Implement the explicit scheme algorithm and perform tests of the solution +for $\Delta x=1/10$, $\Delta x=1/100$ using $\Delta t$ as dictated by the stability limit of the explicit scheme. The stability criterion for the explicit scheme requires that $\Delta t/\Delta x^2 \leq 1/2$. + +Study the solutions at two time points $t_1$ and $t_2$ where $u(x,t_1)$ is smooth but still significantly curved +and $u(x,t_2)$ is almost linear, close to the stationary state. + + +\paragraph{Part c) Neural networks.} +Study now the lecture notes on solving ODEs and PDEs with neural +network and use either your own code from project 2 or the +functionality of tensorflow/keras to solve the same equation as in +part b). Discuss your results and compare them with the standard +explicit scheme. Include also the analytical solution and compare with +that. + + +\paragraph{Part d) Solving eigenvalue problems.} +Follow the discussion in the work of Yi \emph{et al.} in the article from +\href{{https://www.sciencedirect.com/science/article/pii/S0898122104901101}}{Computers and Mathematics with Applications 47, 1155 (2004)}, and +use your differential equation solver with neural networks, set up a +simple square, real and symmetric $6\times 6$ matrix and find the +eigenvalues. Compare with the solution from numerical diagonalization with standard eigenvalue solvers from linear algebra. + +\paragraph{Part e).} +Finally, present a critical assessment of the methods you have studied and discuss the potential for the solving differential equations and eigenvalue problems with machine learning methods. + + +\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 Devilry to hand in your projects, log in at \href{{http://devilry.ifi.uio.no}}{\nolinkurl{http://devilry.ifi.uio.no}} with your normal UiO username and password and choose either 'fysstk3155' or 'fysstk4155'. There you can load up the files within the deadline. + + \item Upload \textbf{only} the report file! For the source code file(s) you have developed please provide us with your link to your github 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 git 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. + + \item In this and all later projects, you should include tests (for example unit tests) of your code(s). + + \item Comments from us on your projects, approval or not, corrections to be made etc can be found under your Devilry domain and are only visible to you and the teachers of the course. +\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/2019/Project3/pdf/Project3.pdf b/doc/Projects/2019/Project3/pdf/Project3.pdf new file mode 100644 index 0000000000000000000000000000000000000000..6f43fbca5d7219feca479403a04ace82abd853c8 GIT binary patch literal 231840 zcma(2Q;;q^*fwb2ZQHhO+qP}nHdfoVZQFMDT5a35=J}`onVt8onlD>PrBc~RrH;#Rh$4yI<%{QS@^uFhsgcF>+1shV2xo8rj+ul0yuO{k34jSTqy(Yb4PV=v%& zQtlz11e~fiBN@aHq-^6qAKz8gRDZH)f&h(LSH(9URfp?;Y4iL%(faRveeSK_^+#DU zp~-49^=g<@g-@DCw2f5RyESKjBK^-d-i4>n6g1#5OW2pbEOGq(>8H5*9ID2c_LFJ4 z%B<_=PJVmePG?4l=aSZiONCF3=2L0ZpZ$O0>s)m<4weXxxXEQ?E(weX1dPn_dLQmC zA(BMroD(UiiGueN54CLBxzTP z)(FW`dZvd>hiqjOC@X0a0+Z+zZM58M zv})II_e@gPx~m$srQL4@d%8%R?7gzvu9*-sWAVnW7Jtj8+8y;G6cdM(LDu2Uc6Zr)$aiLSp+HUu0 zm(Sq2ysiUT;_I9M$1%-2vJN zcK{NL2?F7bqXKYHXy!wV?)4s+#`LIBQ3BQ=3IgY zZ?f_s#;M*GJ84F|4Bj7P?bqe_)7^sNQ-r3C*A#-!(pxu)T{1b>^Uv7m4mf#^*~CRL zXM1k0%Llh#y6m}*?(uaAo;J|@FyU?Wn?O){rVZ(RP$6|D=E60fza|wB3^>5-99=}U z)ciUxeK0&J+vvu}p-3_Hz*D774gHBh2PTH1yB8XfknvmhfJa_Lq{;X)QI4?*#?@5* zb)#8`CiJOS(-1VUcfeWZ7d-xh_#C!&IfCH^(@i*s{VS3|`!=-B(sag(K|Wgs!{4fr zQaC*@7j$Sb&u&F*T>SH~Ud%#C+roWzAHC@&2FH+ueHgT`LigRe{rdC%m`c=}Z~MCS zJ~c)RB$-RgxvQt5yATB8a=*!|Kn%(n2GmPn)Pupflu3-s!51j|5#umAg^@goJg^zX z7e3cd=+AxXEj|vDq;t;urZe>Of;Rq!EnVS~hq8ew^D5&1x?c%sCEE#P4HJX*)(R6G zhxQC;+M}Dg#D4&{v@L3o=u;k&M>$?nc&WixB~RM$%W)RLPO?goK0)aEJ6PLV+B-Y_%W4Qp)jZ@ zin%5!8Ohys--d9$V2EH|{$@#BCo@)D) zD0N5|Up={5XXQ%SV#6aipz+}Kg{xmBlwEIJ8nP$2a(&nmEz3QNX4ii zIK7}{g+iX##V}9zwwrWxr~rd2TwQk2;*0lTV9SMN2&$(p7b8%d7Qi+`C*aYWZ6Ej& z4h2hy<4*;WbO&_Haj-WPkIg}AHULpozt9K5T0Uf^e#Fg_0(8a@-$5kqGZz_ZbV?6O zbk?rN#)%)dmyE^6#FWKS{5A!JxZ)!FcI`6H;2IZ`cE~|ZHb|Di=xS=V${@1gN$gD) zD3%9i0k36-<-Go2r2h_~n={}$u265y66}9?-!X28j#yq5J?cxJj836%uV!F;bsA01DWEXK{r31ef z(XCB~oSX|vWmnJ z5p&mTFZV^6w>(5#D*N&^1=HdO%u0@|cL_ys0alO+cseufGG9!OjIwA{hS#lGI&Ydk z73ftAK!wfDXY?Gq5FFIGr95tMcC=p#*q(Vr5)DJlZk-+5P0VsMcLG$t5FJ?_$bJ9GoI;Q zdpgJm9~|kU-sDE{mXhNEiM7K)H|Bm0JUA&gQ3rJ(Jfd4hS>bnmBFgO=_hIm_BX$ZM zJI5$yp1ud)P)a%56s_)GUEclrH^bJ2FF9#?JX{k_0LdQg&XgB%CNpKwcBakti7C? zGJ%w3BQ~EGzRg~&eRpC|)*wSZdg{9_kJk8waA!lR)8b{;vCcC;b7-k1Q|uYVXSnRIKuNI$kCE z&0btWQ^&4L^KG~S?X1M~P+um_xsAp6k`D)gR+9nu>g=}T-ic|hU736OYNM-VxF2HH z&p}Obh^SVQ`RnI+r_1o>_(3w>Zqj!1KdUd7&GjI5K9Bwq>Gp1m(xr~fle%lBe1~2F zo*M1Nsuf=Lm0uPsS{rnWSK9AdtMkJ~oTFuf#^XVP!TC*sF{o3XdwBw)TGb+FA#Z#p z+Ua-C-`1_Wy~KgN_uJ%(xb6Kl15JO50^8jD{vrH?mpHZ=)r+26a|asA`M&7f9LVdf`^DUvW27MCCdS!E5ndP-ij9>X z6I+(+@GbNp9sS0GUp!WCA2_!O@^6V3JglW*F=WH%mkh^O656|eh*^l*Nbcav)7*8W3i zZ;-X$6~XnX)~Vs<99;@m#6k(EU#$$dBqO@O$L(>Y?S;9_@=!A4vQRRUys#OLoQ_-z%d zQqXS`{7kP~cqHHmww{;He!*;(!75Q1qg(dJU6^exv&`!?Cy!rhTxA*V8*H29gw;jI z7zWDzis~bQPHW$R^;_Z7-m+|3gPt=7Xs2c=YQCYop#YHy+&SE3oE}8|d%j73iWi|> zJ7`?xEM*s#{a&pf-rqPM5m)L4kI1}Gku`M3_hW8$>uFpl|8SyWYkB9t1D8CNA%_Zt z^s48A$z*f-AMv^z3DQxn7~N_a`+~m0jxSTO)I-kdVTqr}`BOZ;{<)=Ml4&w;rcjSY zp!I|-5v+P!6sj}U8GM>^YhsNH0xN0$0|6$(tpvm@Sh#+TSbNxqBYI80*xaZ|}@Q(BXkEe%soJEDRh~mbdcYz#pI7GEgv||dU5@g#Vb%@G$k9y#7 z9s;wceeXU3YT8^dZb)1ag_aUkyL5cE1PQO-IC6l;wh(A!=EUjL6~(llL=%k%7)tQJ z4;WyPi(u5RKZ3CY^&vu%>k8$X-~Sn!CGLC-5e>OhM)QIMsyUa)V;J#oK7mv$47__q zm(C{=Y$bqrKVu;7vn5G)oic^J-5lW(>*tLQ_;S9%!METI;1pV?r)Q6da=wyyQ75Z0 z12xy*QiMmE=`kPBtVSD;K3A~0*!LoWY7q6CT3WRY=G#es|C}&Z{rU(<6r&*jo2cB{ zpZ$aSTfiIIBoRB3-z<^D**rn!)3Rkr$|q>)(fqVYZc_qzvkrkMU`ZDeNy&M_ISE1W zD{}VMlji!&ch@Pkgs6RkOr)w+ze#^&^Tw#44(=qapnvBG zm?zUzXt2bL;<4SP1({shjl`#>NoEy|*LdN63qx^I-TnBe*}7wb$z%h)4Z)k%O(|K8 z_kcoWU0o5Vu7{7jch=x2_y71-Oe^qnZ@=wsKR;AE_WXOSYoI;3x zAPT$1PzX2EUKfzZAmUy7NLc$rJ<~o>D>(n6b5uraS{gN(44i7LpGi7j{TB^B2{Ccm zg%a$RqZ|AWoZ79u%V}+$Z|l}(GGh;vlh5^^OG4OA`R|4JBeTpSwV&Dnq$z2}8`Ia_ zV;EH5!5@yj;Kwd;_z-KV)xdsghkfFC=v$bPw~Dw_ofcWju(xDFMHlwyL{!=i*{Wlv z$dZIxtlu3LnK5xb$4VT69)g3g%X*+dUhYUO`HqBLD0tpchoEG;4%Je#l=2y>{u>EY zWInUpuz7AbbONrBAXh(KA(Ps0qI@8*UNfe+8j|1o%@3+Pxmx6-D!s@*hmHBSXnt6C#!dTOSN+{nFceA_1h;Zj6~ zW$sCv&I$s*3Q|gF=^$S{kn_hnLEelmb3Y~Ol7Ey@-5Xhbl*9^^oR-@EdOi4hbDcqC zHRaiBmS%VpJJi$(9OSf8ezYP>HRwkK@XDML37^jlsTEW`#~zI&BUZ&Bv#UVq%8dG$ zBKzn(s^6+CLaxr>1{=c(kb2H?K{TBN#Blx{KOxco+ZK49{BcXAz$U_Yr%z(q@%LYT z9#2Q;2cAIuM*g2BVrBbZOvKL4_J3-kZEY!+!!fkJON~9q)Cj)b;?0*5TFuHBw3tQ_|zr(@QYIKr7{GO8Iq+Ab@_q|99Fg9q?=&bC#cn_~CPQbn~i1 z<{WKf*5%a!vq`9PPZ8F11|ldzD>Wnrb@!U-?PEYLPM*do3{mt~n|}Rpy?wjEc0_*n z!2tvJvCH2f`}^Ot>dE%rp%s3`WKZ8O6Q&(!Q}H&a~BRE}=yJ~LL`-fPEL z;tXFsI*-pbXzcutXbo-)QPh3dS$7Q~C1q?21zJv*mXwK_-l>oL-?tlDpIfV+uI;TW z4}BT=%bUgcRr$KlI(>I1KgaEkA-qw&X|^N#24r!{9Hgyb`)-GHw5MLm8n=%lkX`}V zg7fc{=5AbD76G(Y(I(3*yY8y&Cby4J$3dOD!!lcsRu2sO{kpl{LfJHpSRRqzDO_Hn zEpCNA=;%F%0tX-bo#U&)bKvxJF(Ct;@uCT0hKFEqy|-53P%|giTrZjr41Qg#WyI*5?=YkuQ`fX zv-kBuXJImv7QV<($h(Yn#x|gM?F1y+@7~VCZCA@!Fs@QVST&lPR1;oZEHmknHM_AA{g33-s=qD%` zeQi#IcZH9k$Se(QmQPx6l)^ZiS1k00Y;5XWnG|2{xzIBBc$(i~PA^VII68xE6-cA4 zUlbPwWF-v=H_IAnipbMYHVrA!m(aHg9p;P)T?jVwlkmr71w0972o7U`%9wN>bdj@y z;sDA#X@XcMCh`TUms(kI5d;PhOpXcl7II~E(zN;@)5ooR=*|kbnXvE<*$!@^+{(ZF9+El2i>6Mq$VlV1jol`;WevZ zJCQHJTtsD>+P9053s!BC>gXoaV+^!$ZSN)(=T-r#fj5;}b(>>)J3MULB-2wDLLn)t zSKctZ!->VrwLEMS*M64N;&7ww@YCCe$!-{sUcNCG&nRf@#hT@cf}+SLm+R+0%wH- zmTbFGF1R3sG`RCAQ@%%#n`tl`ELoSd7_!yEY2ojxP)@hwJNLj~cS*WqRL6rp^HRZU z(g-;&JR9z>9`jKuE6XxD42TFU4kQFZwApS2Y^om?w*;Kr+F*glA39W1JVsv;4A z3!7V7q6s-4lGDUV5O>g*Htox@mf>ZC<3L#7g09mK&6CRxwmN5Se;j-Qv%qZ zbGKWi(g0Z`0Al0&!FBV@+Une!f|W%!u3KC@g8fr;dVsb;c#|q5>on<9V>gE`O4eDxf%|S z{vnS(@4aV;4c5Q$!ilNF*oogIv9@P&=Zb7Iom9QQz0&T&!PKf;k5p0_wPRzJT$w2; z9F{%@q!kz9e#n4!`>LjEE=n-b>Q5vVe7tFpgaXCU;>C1dW|kNd+PFHIR}xy+AnJIO z6;5spFw(71t-5@EH#GrKb1w{t{>@3i8XPT#{oP=KocjKxC#mh$z}Y`I;KPc|qurZT zqY|B?;}7f2gN1DU7{{NLo@AVqgzpAxHIIh%x*vtLBr%s3e{2G=W2Sr9vaw_uj<{C) zpL2)9z@=Y$5W9qbg4%alVH%!oPAQ*mSXS8`c0C68=2a8+_j&>jwOM9eqvDo6)$b^O z7s0x+ow4*_y`1~wnrS*xv$e|>gn>+3UJe<{W?ihesid6^pCTvHWhFL**{xv=s-SuE zP*EF=1>YDDJ##{+F!&6)KG5#TZm!xqCR2`ru#?SO(@vuj$%*cb$)3Im>fMZ^2zsm!SmxH zxB)9t?2=0Bk(SoEP!bv_7f|E*$=4aCwlnt#S5y6&8-GKr(<@A7GfcRBb%llq_eh2! zi0Eu~L2;uA3dssBDe!Rwe9cd^*6CqsV?(HoVFlD$|GTOvZ@J=%MvIw)v*2}3lFR+q zity*4eyi7V8L0qdCtQ_#=OE+bU`sVL1YpPk6@NSv?JxeEK0tpGm(t}Z@KI*@!b3)y zyzmr`D%JULTJm!^L8lj#Cj zvCoj~!fOo6zag(K6ZczTf9~d6hAE=$dLQ#;(cUla>=16tseQSzumYCY5NG<P|7j2y{i4@c+#v)S#BO{0s)bYY+%doBb3?ci>L|4+EGvHmY`<@i7U%q!KB zaXA`;>$|PbASsu=Fza74?R0kRiMOpf=&;pO)+eaIk~U5gk$=;yeYgZ+UGSYz%u013 zURPkf1A1GC;dssVe7u*NneX4}@%j{+lp}8vCCkq4j$P62Al)YY2UU7~V&4DG_hDA8 zmeT(HqHSU2`pMO%Z2Q~s{<(?3oH>uy7;JVJyEUO2J+onz2}1jL?pPBgqU4Yo?MRm^ z_AMgBFxqaSb|Ykt~>6sh@X# zu_TVYQR)|aks=iqwY*w67aX=#&32lS@UKtaWe%z`8D&;aRnI(3{@M87LhAE(*mg_O z5alcOvsEfod7AnKLu5Hyg#ttSB#(3ae_7aL-(8($(I%be0m^fO#%0Cj<7N|ZJ-EY3 zUlnhP|EgFAWFMqLEvIh6<3>Z4-UuX5G?KB89WvH-sFOwah9$&evE_Pm?d#R8PIlks zfq=+5I^LuoHERa2s_2LB#3$PHAY?W<<8;Kb#Z?OQHQK@WlZ~THnA^III@Au3aMa8o z2)JgVsd=#&F;^FjzSz9_U`4KNokcPK_*6|Ksfp3CXA@fx;Z|v}piTJ@5HqYIO`t9; zqQzQJ#U3N)7#671*qp<6gTyigP|P3f5Fd$aulaAw3kgw&VYpHbo*pA_CnXi-u?zht zIvC-OQpE45m(~WYu1@gN*(+g-c{AnOTt<0%jWN)-V8#U3+Nqxy#*1J^DR3h}1`U5- z%8dgurzX@5c~_PiBdd*x5%C;(pbez{hS)zUi4`cgzMn=ubSUlK-cUC*&1^cnG^x^} z)W?~c7_ycvNMku`#ip3<5#fse1VE#b^29N-fxwZoJb0c&UAp>esQu|~FCScLX6Yd+ zfVV3VKgGqnfxH$qK-=S&8DGrKC@sK5Bi_S@y~(k|R4Y-TK;z$YV9Jy$!NO?Uy-G`0 zz$#f4$tTFml_#HlIpjh-Pp*^?ca$qSFBV*DzN%`XdwPrF_5Qkoxy*;>BZxI9$1zbj z`sOZT`{^8WA)r4HP=7z)9em&qqe;0e%oth3cNxdRVQ z13tl^W8;4>5V+GgEMH}{U45G}M!-Q`@mpnw@aAbdYWxjPUbcL`3Su?aIg!hASqp^o z>Mj&>nl>rwy&Dibt%tw)3@bd@~0EnIJ z>?x0f}eY!{fL&F@+z z_zQwE{;qAMhgwA=0g!g$OA$xmdPEEVpcL55k#kmwpJp>1zsPv74edRNkFH@lPNbdn zkO`d+C+2bSbs;;1yx|Qg>*yxJZw9lc?#C?rx^klW!OV3^&mUA$$`d-1$p9s#KA6kC z5;jV&I0K`q|93ZnpPc{on{(uzYt#hYI3^8L((&ivqG#^Bn**fbA`Jix+W!UX=*B00 zsD5&zHJUiWiH#e5lrve^j%dS{Azm$v!NxG{4`TLTI|@Pu!cA~v(m(j*cl^Lu4w243 zmbH%G^`hq_%E(NjMkAz$55whnP(iZrzi$K`KWIQn4y_SX`J~*o%^HXmeEmIidCioB z9VaXfZ8GfPM2Ai*v(z5C53(-*=VKQ0nuwf?&x6;7w6h-zYe*cSu|c~#5RKY0y6psMI>YxO;C&sA>mAW5GPrvp8ITAL z5nN6jCsDYf*G-Q=v5thxlKQs`eOt%=Eb26~f!9h@Qv))PK8n}cGFnF(g0>>@QJM!% zOv!0N7!bFG_JJ?1jzI9kdgnWG$*fOGJE(c54|46xgeUW{hCN7;bk1HCoG6@RX586y9Eu9T)8|!iJZRN#F(?9G zOMr32#{?Oev}eUj!1J|izS;lw6F=o~!vq_MUYdBpZVAAsb;bahi(P2^qIfI7qDr?v zy_Zi6Y=}o9%q<3b1|HlHptQOAR#n&)UM~I9(}se$<+T1bSN-TNTCAj^nE5*sb)4>DZgoza zUM<*zyU(qE;hfrCmKNZ&Pwh?XZZ0s-s$#eb4oc$)=YC5{yJ_~L~Jhq%%9WUxHoF5*e8I>F++1c#WWu)USgd7jq}4+rzK^@ zAWn&5b$(zqEcGxyBl^ZAZtxvA|7MN-X8}&t3{kVL!_+Q+Em(J0-q%TeZ0R@M@px7R z$6UtXjlMkZJNdk#x@LHRFBLTFmFZ{v0_z4AR~ERs0jj*#3$F=~KA3H+?|0H~VvG;O zmGyxPt7rS;fH?)4_wkiPsDo?jpI&IuE_o*pJ>7$Lb50H$2yvbmpw98=?x@Zb@DXx^BsKpGowvjo~v0 zN!;>l?Ze~)zjJs9CE-%v;5EyH?`Njf52R>TEYGsE-aL>z_KY65BC`6lWA~RZ4TbH- z?C+R$6~3#-gxXa_wyqF|XE+YXF1|yac1|05T|J~BaTwCAG_sIYO2+n*R#fm zu&|n@_iD82e$QvN&D*Bg{a*>We!s4-!?U&OILTSU=-J&7u?uS2(4ABd%&FA7Tdmmv zWCkos`rn^74%1!PI!ZMdrJ11KZAdtY8dgdtf{P#K>cFI`E?Jx zmEE0rwDDIROdE55U+c2#Of&DAx*RGONt|{z7unn-NCoz7A(pF;j!zgj#=JP5SUGtY zUcpx39wRZg!tBClxwEsFQ~Ovf(m%-rn$h=2_M`bal3_ zRNkwQ_s_5~swmt*VMUqv`LM*SghTPc%hj8=&T%wboD?F=kaQp9Xp}0GX?2mZmnJow zN;6AS{792VJ1VB7%I_>El)C&$%tWs!3At3FnSeVfQJJ00KIa^?3|9hgX zTFLtHT8YRj>EXRxx#<+UGY0l*gNR;T{YMA1cBL5WDudn=>*(CTgNuE= zDtp`}6y#0(_qFAhcTpErQgIb#OP!mJMrv!r;2M(LS_>uL16@6?#C-1og&p)r&mFckLUIo*y)Z>Bw&gh_t4Sin9*$y&b6 z=-{EGDrq$8EkS{3XVAVCtn;$+Ao2=2GiK+>3k@bTo^eDFndwyd$dkv{GwG06sx}Gi z4~x47qK_Bvn5PZL+tR*L7jh>X8c)3@_s3m@11RMTo^e6z!mWo9QbdjBZ(;9h6$+)B zg2WZA^==S%b3ddNS4J2T?d`J(6Z=rL{IADKC#ghgX_)G1oJ`XC;c8L$Ij-oD|QTXZk#phEJx}=J&bBCW2 zBjp!L=eUVmnIO`7rZEJq7DV@)>A@+uggdM9GDL>yX4!ANC^;XdS2~_`aHJ3Mv=P_Aj-?{x0a(o4O}?sCzWBw7$GSZEo_8SgxR*(|W{M$DXh9Z99*J8+XwOWJx&k;s zgEBnpY-pOC%8I%&%0|GtOLY!~enoUU&RfWuuuDSQ3@>lEVF$wD6E4 zS;tp$qZg9|r28|v9~!|S**q{D3e3C^O6P?QCTS@Ui%p=MCHq?p-FkF@gjtQSDFllL z+e1s+6O_TYLR+bQ=pAGqrwM0AxQG8vCpgBf{(cm=$tu}ik-4rGQ(-a$E8TR^k_#r* z;103LIHcvg5Hcq}qDn(=whmN8>i1m_L()v<8lpSD1W;Kc;+rd7`4lb*g^wd z$0CxuO~tNo&!=lCXePSy&>~qMby1n5GzCkXo*6;0J0gTK_&54A*klER;Kl+CF1b?Bu6n1p$zvSj^{-vjwNTz=f2}x zmx!m?-Ja}O#3V8-?9*=|Iz7f>r}WR+A4~wz+GSpE2Vmi{on=7?nRs1$BeiIZY!}fY zJ*b5AFIhh%AL#nh2x7_VM;1lN(M*9wxA`4(rNV-djAaG|fdl`~8;Q;--Pre?Es3Nc zDf2XS5rXf*ybwCkcBLj4>V?fJ7{eFt;{IPW>eQ1(ts%5CGEODH*REKJa^fOp>S!*= z^ffZXgYen~z!13vW*f{^5+tP1|7})Nr(E175^SnP`6Sx}NbSf<$*OAy1GO`J?*oN1 zg#K+HM;VFqhEjfPt-CxiwRk$D6MXX>L9;JHV^K{kallC`jy?r5v|rU>?3$eYH||RP zILuX84LDCY|Dh|pY&ar2JB?FrJES;rUt4;hd@9fs*!{nkHb=?#oq%cbIVEl}a6$#H zs~|S_9ah2*=qXl{537VcU4O^?3Lu)`(|g`wYM+$X`4T4?xanbLsv5Mymff;nfMEKWfalJ4UP% zqnFo}Hi(r?<=0_*Le%Vi5w0FC3FM?lDW8_qB-YtvLz46k5EN97)Q44L6t-aoQJUsu z@985_wY~O!UZtJq{X${=jVxt6q&)#hM@zvExN$g^tn{T?NtrC=fJA|m6qUq~XKj|DKi>0tzG=Ttb|H$1` zGt+_Uh=E$d0T)yOG@uF)kl5K6(Fbp#7ZmkXD~F~K|AIJ5&HMa)3L>@Q-X`hEi(>#+ z>Uy3&dE(A8C_geiob^cw^yvI^GK6&@h!h2f#2R^}GOM71W`lt1tZ1wB-kC-O0=bsBhqCEAhiYH-(%Dh=HCD(KIu z5sTQ=Wsc_P1F-tX+ez$jG%Ji{+F4c`Q9)ecX8X=8sR0W@kngEh)Eb;+=?bf7vsldP^Ed=_$*bFK00p9nky?9o9Slo3^{uqlGbFfZQ8ABpiG z2@ODEeT?aPC^j1$fkU9EL-}!NG&Tj0S7#63>(5MYS6;_putRJS#4oH@&o4e1MrNg| z`KXeZB%j_4>3!p7hT0f6n;I6SCNEqqqbna3M&0m&TCaYXY#xG;MdJ$eb^Y_t!n$Rw zeHW#oktnsxzpDl`o@_EVRf@p+ID2a+xm}T-q7=l-#^Dc7?NS<7ResA^S*U%V!qpj! zkJgMqsJ{_sPiO78`tvjZ*I`~b3<8l+GW2*S);&aD6XUzFdFG;lz2s^Hq??5WLMFS4 zS+_CgToY77Iu!^~k-?dM6j)6KTBl!_BEBojKT5!hHjmXIRo{Xqhs zd*;t+CXr}9xG|x@_N!sg0)kVT5m~SnL-|DGhYdT5X*@CUIklOlJn)Y}A_ReYw-vhT z_hGWc^?yfgl_5dEN7qa82V^U6g!_lBYDg)9IES=sbQv`Ou?+lPKtyKOeemS`Q7yL4^aQvU*cT`iS{*WEsKi?p+ z-7OGhHSK8A1jcn^``S2qj!Lm$86sCyD)b&M%|-t4RwgRtkkg)r5`8mW)I;fLJia7( z)+*Vy&nx}zS7ZLR&yyiZE^ACJ?a%C^Q?fJ`hF9lgGch{zCNBZIfCNF;VMnUfQ@DUP z<7O^xlVm%Db;}qoY`fj1xZbLFi{52hKkx5l^s7~0MPqL;2ZRqtU1j*bprpt|4_2Txw3 z8kmutPf=#bcZQ@D`0~L%3EYc!e6PZX=@v@<0K?GBm6+#|=f55YWC9*v6-^V#o6?!e zD4NMG0e81;yWS3cUMdPcnM%p3q>`d_jq1q};@oUa^TwGwa)i8{7aPtBfIf50B0e<= zJx90dA5Aha)+$aK%y28;e(0EllzkyBtkVX@>b-}Y*g%bt0svCOQ3?E6`&)v{agSrCN8a+f_I+Pp+U-`mdewHGRQq%^d= z3CcUL$o%?+J!ANG6N>R~39=ZqS3jNE3Y?3{B*g_>S^e`)4Q7S1Q=)I2pQ`5J{ikOF zB7m5UN3t%5bcs`gm^SL1t}>~y=8h@PJHtGNXi@0&&Q@hNg%T(SWKdD%*KqsiAcxE+ z5IN395!H-#3d-E&QBufTPcpkwb5e>!5|*Z1y^*#I1IW>yQj-huPP6Lsl*DHO0HQV~ zM$MiX$`>32V~%F3E_f6m@Fj}x=Vb1Pf|pY5swmp!j+Y$H9~X`lN#%Nwi)Cf}TJ8tI z%}^AwUu(iq3zFkAucb5+EXS-gfGhHD#^?}$3>W}0ZCpcl@bFR>H`u?)l@845#Xw!m z9`+X&yhC|9jj^{r#C<-wkND>yq3_SBiTjuBcDNw?w8+?3aagH@b$7t#A0~^8O9TGq zoaDTf>F>SPmaXwnhVF4YO8XDft<3*fYQWONLg2yrvj`-%igye1;W)f-^!C7|{?7 z-NpN;-V-Iy4}t-8;X4__lb;_o?wn{i@+h!{oo#mF>%9!q$s`q0%K{Xwg_+bY7MFCQ zvAQ5RGvG5NS&c8ToX+V&63cV{t=yS605{t9L3WAOOtG+OrwD1Sz}$%^i6}_E3exa= zsYx)$Vcv@%N#0%qp>QBoNysrB-UuAaW7OF~>+c>#!Xq-T

G}ZPz0uz%vdY@y=kM zxiOKt5{?H34~(kge=&&29b_Zd-9EygYRUc_^)V)xi1;XEgJ}&q+xjiVY!&x-#b)QN#7+ ziwp1;kV4iuPkd?Ce=p>MZ?p0XNB6{7Z|BhiBuRO&@1e^o_f6gH>l3Y<_6j6oYk}i~ z#L=#ah_yvmCPCzlfm|9gn&V#B3>hSL7oXi)d>|2N&D7{dQxZ_jY7G3iK zUs$09Z-j)d&;NYrW(&$}j@@gHQxFnu9Hx4+V3Y)bi}GZSZRPO5yk{sBq$vwCS5@Ii zQhD{T>uQc_jsdaq8OZ)c!-{-j;*7IbiG_hf4u*Rdmip?Q{5-aT>0DLz&m?y$1Lo7+ zw#-Up1lvjfAYT3)C#k|9vMy)dlY6MzQ$29Wm8K;{RYS%I8t5ZO%j=##pBYo^9H89X zzR2oq-ly$KpQ4fpxlwHxPAIRPOmqx&=GJ_Pq{c1%z(p`=;Xe`;x+;p99#BO=AEdYL zLgOX$bgcNQ1ft-Xhgu+6=-C*2v>re-)&_z+Hvuh^ZTrW-zp3;EUP&l~7t@*BXC9C! za-rVj&LCJKo1?q&X32O}d$OSG(;v1mIzp35n7u*rnZ`bu`t{;LylptX=&0I$ZQ_{Y z`7N;P1je2Lh6ps>`~+R10$n0&qNYl^Jx)r;4-y3P7n4}ve$$#=BzZICf0WYY4`&?V$_bW>1Nq*O5Uw&ZJiy`c_wg!hYc6); zH$|l!OBE+N1(9By1N)FlONIR@r_O#Wzu(Q&uHb=_K$pfsj#NZ-yNFw+GFU&1D7&Mg z0>wPig8)DDavmeK?iw6fGCoLBV2SHQP18k6+{09xEZ_|9gziv35RQrtnCnA^_ z|7VoUM99p-!uJ1^^)eGOb1<VD48rE}BZRlp86ImJS+@2k4+90xmCjc#RF!8{m zkn!vVY{6>)a0YO#>=i8l9v$FTz%D2vC#(iXSXxz5SR+3CcA_~uySO;`E}%QIvcgt! z26R9ib%6{hAm#928rx06FR1uw`{9|(K5%dWgK|T1!!v_IFWcn;Xaa5o zEif*3jebXCA7Zig5L!7@U@Ra$`+)%!wYqwEUYQxZyu6rn+q$|~csX=fc-KD%<@RRR zz;}+0ZD2jUJRsMB0cc}zra$WP8Z>*m@cs@^f;$_qkj_9q>`AU~K8ruKDGuTe&a1!3 zJ~W}14*Bi@)gZco1HVuvW(N=HwLMKuyaPKU*hkQynSnDqsCsg`+q!_UfXSUaL6@q2 zC}1FC`?k+j4)j;T9Lg1p5hPF^Q1D+R%JQvZfZLP<)3n3A4%C+~E*>rrI=MT& zpnwg`-X=~@eDoh4#J!6v@TbT3)1hBtPjvdg&eqIJo#cKv=r4mEGeLBCKwkf-gY1vP zxxl8v_iDmYk6cA3$iy6sA$&`qg|e%opAGoMZ-DugK_lt*)fKATRnW?>&O(4q=9PemD=s?=*KnI?Objib4-1T0z ztIa}_@6Q=v#)bx$ADio1j9FVDK@1GQ>8|-eJN8%ftGg>-j8w; zAd@RYCxjNL1~&(QVqt;8;Ca%HZ+N^1cx&isZQ@#ZtS1LzWD3UB-IWEz+1L34U?42s zc5x;rfeew~={|6rfHOpY3GD&NZv-PChDmP>2gcxxQa?ft`bR$n91ITuh!}xV`v~5T zm2UUMBng5Di zIq=W!j6KJ*!#v^;Jq8G(15obz6Aoh&!&*NHfM7PWyTY~!!rDP;akp}Q_xEN;9tiH} zivH%mIjA*1IJW=?{}kSnH~&ER%8b4IH-pn1FwpLL^(&ah!T#$w`XIiuuh77u2eBZ~ z(cF{h_SO7lerfsOzOw;Dr~50Z^+WRQ6Z}i#Pt#Abu)eWRP-^TYb4~E4x#xKQVSJAc z4ub~;qdI`H=l$(4@sKW|I+f)v|7*u~>j!98+2i-WO~(%rO#PZ)KVW`tA*p`*cz(FC z|I4aY0_gwtH)K^ehtMfbZ+uHZEgL3O>ACXzh71}KiI#1BLWQQ0fY-Emf-zOXNF@K z*2#<5`sWNvyn0}AgTtQcA#@dV*`TdSeA{AyG(Se^sFbS6EBF~AS_gy-vo?GA=%Ccg zWssvyKHFE4wYeWgb;Mt@BL;){04-tD9>HWK5>5#tu7;|Jl>ZAcK+eDBShk=@r<{E@ z=Tp)rVt~f@i5O2;?_(V6=$^}{_}v7{)7{{%zaB9Y zVoe4_R%qvpi^8$TWnmoqdr@`iGK$oVn(G#i%wb~d`w0o`)u=&;8lFM*cUU>gkhLGmSH5FU=!-&lmX}bBASf0)kg$%*O7-9%@d_C@Nrdjq^>=Wz!cE*9LycD47{w`;Rh!lf(L=EZEU>9F~x z=KJcPtR?l&11^Pl&o7ma7`td2Q2nT0D3nEXrDHq1 zN~3J_*FJ2q#*XiNM5Wr+(K-<4w=gqG$CP33@Y-iKW1*%B_hH8o1W8@B>+6Le^p~CNj*@rz zh>2nPD&z(9q{6Ld1cuCr0o0x4VUnS2GlONAU{@~4s zf)mU0#ht4_ zWJe9s6dw)WVzsKuamBbtk&NNkY84JdqPgsr#@_Dp{uuRpqYfs<2mQH})-?9T-B+U3 z?$FYAa@wpAJvN^jJlp2kT0*Sm6t!^|aGpfcO;>ca@E4l#k`i-#Al*r>-yu-`fUhDu z13$^Sofu>wsn|iqKn3_oYre?;c%cw$J8*po+vVSWAw~I^$PRsRa(Qfc?ID<63gey5 zSEjMi&c|0BDc1_aNiEewixW+3&;bO8-x3$<9$Yay=tjq8RzNjX6*Jt1y7d!zqM80B z^c^w-BEC#rRLq`r9`)xy{@O7>p9{?hy!%(nsogcEII_ajb?P=AR&l{h_tI_+n9z-K zYw1$kX-W!NtFVQnAy~CidT$kdbZpeTVqIa0Uh7#R%DzU?pe)s)^G#8BQYFi3HQSz4 z13zx4+tJO4$#~*6>v<=Q7;Tf9z?i+7gc|JM)D=B3OvdA~{ zH(H(dr_A>sa*JqaLNXt;H0u0`hRCvd3hLvI#6|{CN`wI@S6xt;KQk{cIkJ5vd8S--vXmisl%hcM+4NCQkyl*D$?kfr&n{SOCF2Scxz z-SgZQ@gRPUv#+{uE@&1Wy5|rRY?#Xs zBBZHOhOT$Fg&1a>S|le(oCHZ?i;GbDs}n;vHdK85(5xz)8|!YUB;R+fuHf%60>*|Z z`fgp-f&625uUo=XMXTo)j*Sdq79F$65;1)xKVGDz%mRAE<8;lpNXaLrcBZt_}8UIzPX0cr}CQ;#xICwZ&_;Pz2`Dw!cQtI*uG zCljpHYE1jMj7y6IB*<1>G)T*lMuZ zq-o4km(f8NYLxm3kjRTSodDozr5p-fSC;Gg`4QTu@7?6B)2u_$Z zJg?%7#2-TNjf#(G@XS5~c}qW?>e_*Ctqq^8dI_oh7z?ae(9OFi2Iq)4TO45c zr6#EF&;NM{A*AIa)M0>!^#D~Qn{3i33B;tOLBDLD)Hg+H?+pEoSOsU~?zUkV1)g)U z#64r1D{*%SXoqOeh=Qrr|s{dd%yCNp232j#V6(~-my$h7=idJ64bdMduE zQ0@0>fj`bK@8Q5SUN*1N1}R{!W0*{&sjo3gel*$pwsO>R0YUN|yOnJzs--py{llt; zm3>smzXK~7Uj`wE>#a>)LxhA8mULt}6F@Xus}po2jAV*)*@*NmNow8MyFGo6b( zvf&-aB6v{={`Zo*=7Vi@T8n6F-ZS zDi$ot?V;_6RaS;_U)L8*Xs;0LGDd^$v&YNE`S=Pq0h<>+QN z=R?-P^n<lNJK>2WlO);%{OxPNKFhO42y&+hUlbMnSNyRX253FPutBImd39c)=1 zna(eT!wr4k2c{!i69PbxIH@?CgRYr;^qscPM`M|@cy--5aB|BtjJM|EJG@v{Mi0uH zy*ow&8R;`olq&5sFRFF4P16)X6=B_Jkw~cJCTmH0 z!mBtw9s1MYs?myjxzQF!UPxD9O)Y3@!t(}=tTaEg+@*N)R)qL?^QjxDc}>2X23F09 zw6|++QJBt-UX28A7Kzg6s7w?ltt$yWPhiIhpThvq-G3QEr37I6>Ct?c30K&ddkBJp zxiAJJh^k$4p#1fmpxqrJHLaJZJi8BF1sA~dLsC0QY+i7eKe|sj5WBzye2QPL_cA*X z|FzFC&yEi%HgwL`o;ACxwm$ciW;@%9kqu8?D%S6So|YCk$%Y)L;;cDv)=HTtGbEoe z94VNlr5ob{ujNd9yy#i3Q(@%MC3iJF^U7*!4GgbjM12;Fb5VP)FNVMPWXNHva`!62 zfkgb%4zf0dB`LEU_|Tx+Bz3d#L~z59ActtvzeUr?6sZ$A2m>9z#Qy$qiF<9SgPtiH zBFaLsjM0T&Zm#70(G{q{$~drRdt&3+qHaL!#RaORmHl0bZa$#2)SHEZV1nTXi7g-pgL#;S=E zZp6|9q9RR>At;uUwgS}pcAZCuUu-9eaHB3+)XAV863l~1w<5GqkTnotluqI;YoF{$zOmTT<++A~s-k;uhX(C32AFAPb#lML{A30J zJ7ktW=wf6K41AHcr!;1d!tS-=@C?zcoNotPi2cdRl?PxfIC5MN92Dy(4ChnR;&^64 zKibL#U0!01;EJdGv|Dt!!HS6vBPMu{$Xl(eB0*H-x3n7g`l-7bwn~pbbUBR&&~#P1 z;92;n_v583^@2IIE3NL}s)sCw`oU{BKvo~ddPZZBz3Rq4Tr-MNTBTY1IU{7KsueLG zS2DpaI#%JACJ>l_h!dbex1Q6QBXv>?X3>|xnCKHi3jTv;LznSo@YvExxP8j7e#pHV z5r9J!>sra^M0U&tSm&?Nm7Yf z>p75St7z<<3EVX|G5KH+uKs2|8NTNL57C!1FO@AqZ3YSE zw($X1=Uhx$m>FTnFHZGx^nvBU;LN(`q@e*}I>r-elro|Y-IC93b0ppH($M(PC7Xi1 zgLk`Uo~?X1gAzflO|j>Nu47;XEbm1=*?sQF4MUXm3Bk2i+MI7X?@>A)+H{Sr=d?)K6__F|?BFL-lXN7k;8hCCtt;Ecilp&l0WNc>`Vh$N-yT@n5(e9*lSHUQe~uTUX_OyqeCD5T>4sPl|Ed% zOCxS<4`qp&wunYn8KKA3 z*x`lM8yFL05kBx(^K%7)Sx^)#Ov`-~JX;xvbV-T#W0C7kp1T_B9*R%WL}v-E809cf z?z6b#Et#>CmKevECDt7y;}BbzlRq;~C~l zN=r8x3QH6Ex4NIQv;=yDyRTk6r%4;7@|8>9AlpZsQ}e%=D&84hyh8lkzJ{V5bXJuc zJf>Dn+k&xc{Xxh`@6S9=@h&8pKO!bscD%31(>pkwrwP?TtUWByntAEj`0$DK+tSo~ zaHgtiLu#&XSS{*4GtcIoWB1(EB6^T=ngN`?MY|Q%GMpzqlyaMKX8`0iUV2Z9Q4h;l zEy9U{UtL3=-4ko-!Ns(YjT*_${p5sj?{4n5Xs{x zj1Ao*eTF<$4znD-k)2%7RWn3*JM~q$FyGsKMBy_a=u7f0IYfkn^XTWD3ka5QQYxCF z*7E7-0wN8&2yj!CVLj$e+vDC}T__!$D;<0@A#nmW8)T6$U&+HgA~&o zBa9DgtXsqQ(x)mz;XlSN&GZmaVU@(=CMY>^ljW0&s++lA+uH*33z$?zd#3uW!2 zuw4J_z#sH##U zq@P$9LO0&)*9RM&z>apInAa}5gbhan(z2DZrRA5ebPRFvrz8g9-`nh$Q5z+o1)eEI zB5}>YwTuo1=KK;8Ti{L;8Nhg8E?pr@Wve2}CoqBI($z27p-^f)KQFsd7?&lYv*Ncs zMzp5FajRRM(FhUrMt)HKq(oj&AznBaY?vVxh8;iseu9;z1ipqh-T2;{$_S$XN0177 zsze@%Ax|%BqY+uqrfU z00|qtc&n6B!PcA$-4#x6#Ft*biiuWln?B0Gc%O4_wXN6uC}28Xvm*kAuEm+FvkBUU z^Vpv!+k0}qM(^iO}o#2Vvl@$jXWLGsK5s12v}z*Gt^-{>Wj2Z`ugwXVLCmxJXv}^8Lu~Ro5GfZ?Nur zHhgw}#L!^19vcVR6+Fld)%=~TYCre$+-B)hZ{w zrG{lhGqWcnHMma))fJs%_2LVoWzcvT3a(Z` zE4d#vqaj8YgFgV%&*mo`+y!jobp2RjF4%ZYmsrc4SSb&Ff&GLEiRu^-mAhy@lbZus zTC_(wV|puevLfgAa*OIV<}Oy%Vo`Ruoo3+GEj3Oqz61o)v2tSCB6qBFYldi?(CbB* zP#x%DI{Df&*Lt57?_M_j0CcSxJOH{m;?+XHr20A1;Yp3wl!eIfrV(XzLLZ(=5o1Pc z48Mc3kgXQC?#RWwtghdR}`S>hgWRcGFs5s!% zjaeyMKRCQ5?M2z>*JKKTF`8aQ`^(;NjU*^4n)TtM=j-#?$Mpom*BbRXm94RW`l}Pq z>BSdxS0USh4AcBj;$S?JUjZFoM~>X7Q{wEA!f%4u(hBG4SC~Mk5v4AkUHTp08nFQs>tk(4qOmt=EI;8F29#Jz{V7!(=h)imi(9z(z%`WXa3MJA5QJTYX=i0|W285u-$UsSO45@&QQ| zn8}(&h!M(sM_qq-FI(>9l7vwa!;m=V7l8&e^_66>|J?n_omR0**D>FZAu3cVEjKir zzxlz~p5$E z<{pu_$VR!q$r|)xQ)YCuF=dr*q1y$t(ji#!S`>YeTC()5)eCM`kC{5Hfa9LtaalFt zfgOOpEuPgL_RT!UnRQtkHr{k@K@p6GDq9j-z81+X_vQ3Ow)4Vb{2a^akSOHjo64(y zG?M32FPL&DdEM^mp!qraL?-6itDR?X+OcZSTKZg?r30&dm&E*?#D{7iK~4>?h}Hyc zY8~WiyE6*1)g(Hl9@FrbS+5D$hVZaBs;V}bPF?&;@y`k!Dp1ATKl?jEPczfc*L;4e zrt-%YHD1wmpUoVBMy)<85h9utk4Y1P1S8JBuaM_hl~=LFC3g3ILabI+D?Oq$pJAwk za(7aq?Cy7g=kwC)=$4j>u_0Q^%M;w@fvu;1VU&T*BvGkRv?zXf&(KMO-h+ZtopjhQ z=t7GW4u)&vi~PES2*(WTRfyS!*!)hw?l^Ifl8r>T@3mTA3<2Mm23L5qAh2MHgXOhT zdS}&@AecCTk-H;2!tqBOU@@_+XTA%RFxJe4W}?>69%t1Ei8)YhkAVjL`ur z8Z${V_ItSo?8uND#7rpCRg>7>K3?xK%%|Dh`NDC36Hcj0p}JvB4vt6=H4UvOsd>w8 zZy4qXqSv}pO$VDCtJT1BU&`$-i77|y{7cOf=vt$Tz$H#{1-D7yHKK?RQHJJ6*c>-n z4AmHO$_Zx}aJM%9D~v)K*~qam@}L1;_O)F=Vux?lEjrP{%W%sv6-Rsd(jX03w>^f? zJ{bG5{l&5x23I&1$+WGw3t446RL~k-Acxr@DxnK zWpK9`7iLHeC0oe@b|D0J-3BFBSc*aJL)xD4cGUc~xY$}46QfA#Plm|8BkN_NQ~M>A zaq@`jm1Y#5ErR8vk8`J9T7%J}Zr4ZRx26guDW0!gZ~VL77XY%zrZkeFA*Q?HyzfgL z020`8zIdVYV&j=!Ih2KiZJz~h_L0xSR_BDO!^%aaK@9`sUb%pXVSCG<;XDdrU}4jj9;sbZ|Y zM}g>2_71N`JJZ}6c;sSMI7KsG&-rpcGQQAo_BV^rS>XBn8n3c|IZAFck3vS#qC+WV z>@P5PsMW_yy3g>C|?j z_uZ~+3a6>7BeNE-9kMBE%syI7rEBfE#0>94y&cJ;8Us|2%dZ|Zz6HR^pVOh31VhMt zEssNr^lH9+_43mFS;yYd&oqS|NZjK1 z(sn-VwTYidN(M=Ds9H>YndXQya9$1jOO?OpGi^5f;L!eZvGs$T(4;03UGwPXUfe~{ zVGwHGq4>z*P7AB&I7_&jPhH!Uy!0aVXOkg6JU(paNPRlx$1NC=7xRT2@q51S13j14zAA%9JV6#VnLAvV|rv{!-E8*y)%BApmi30qe*P+CAdP6zIBe}GN zxdSNfFQUkPW}#U0{UUC2Q1}9lQijLgc7$Qb$vZAA^Y%a}voKVttQh3QhCFaNkL?-gUz8 z>{bzV-blf|w8Q!jv-L}bCQ3Y9LC|zaxrzIonr-t3AW7ZL9@}oiK`tW_X_Dy6bRjUvgG$zp}nqfq%>r zd!L|_B=(-PV?Ln-+5Ea9Ad!ih6YCR~rNGnU_Uu`sx;AV8nzsW)91~rKkt=mon}3(d zwtDAhOK6l)36`}Vfldx2hcB%!uRlb{kEeS+hs0>J;d#2JxlUV5sh=CNa`gG9liPjS ze=vkHI#gbj%@G!wO_>n0ihbncUc?X+=q7zBMfxGehY)>kW7e^Ep7KKs{-)$xcUfAE z73t6q#-1fHvnzhV8;n)ZPmQ^H)K;l?v1Oq4_>=hjBFWliJnK#$klQQb7?I)QT~4>X zK0>=Q5h^tHcu4;0slqEb5uViLe1wvz`u3@}tR7gk2BHKeMiEyiL()j(02Jix27O0^ zJqu|$mzWy#Jj3yqR?NicWKnR#Ez*95oW99kJ-7~{Rfhf}eN}wt;B;g^Q0v8qv3wFJ z_&h%WmrcPNG%j&EsI-7!L24ZOk8s7;Gd{W>g_PQur`frb0DkEYb}p;7e2@e%uw~rG z&>f`>iNu5N&U=59eT^M>vKSkb!?vz&Cgh8CeEvLd`YdgMS>6JnvTSmIcJ(fCon_%l zuc!CxrEqSY!gb{F{W4UGn=~GqRJ8{*0soPjFpUJqd_Rhmir^5Vh399eNLVo&Yd3S< zI2fyLVamlnkS&FS|TUL9E#Zv+}dVh+A! zIGZ)d$bEwR=o>T9Q)pvcdsB%POigRbW1N?k!%H1qe0wlc9Uy4|S#o9(X&VoTbxUTu z7o!$nzlo@(N1AF7;V&6&_*otlJ6Y(^=cj#F%h4p!bIZ86LQbag#em7IfsWMg+&X*X zpa4F^kxVAM{!;0ZpQw4}dgTuBdT-sJmWC|@AykKlN>k&eAeS>vMDYv|EH_6mrnr!A zPQW}=*NFMj{!FKl_6fmOBpqqRb*d~N^$c(DjJ3*15eR=Bn(XkFfWp;~CX z+YQPpaXwp@Bcw1R!`{2nt}|(+lH!n%q@_>r#dCU8!5qIH%ikTXhKP}D^$Jj6O4x%2 z<;(1#dOzexz#NT=Gq`CT)AU}li^tcIFl_G!o)5ce>Ij8*f585ZMFP>wR|QwS*WYP+ zUgQoI#1gKMe=5xW#6zXPIBMR+By=$0r}YZOQoLzK@ijS53EiCjl+R4&*Eu=9ddN$K zT^?qU2WP&n_x(6`&gVo#G3c#Cqg?P=;3WIb&76Bg6(4Qj>#oi^3MsE`3ZJ^UTi36o z&*(qU;xIoB(q}LOZnu3RU1pE_!a5tMKz_>+%_Xx00a(XAt>iGdxuTIv!_J->=G%Bz z{G}bjVM@?V9~AkxoB-}r8J2i;WGoqnRP*CxT5aFW3Cs?Q@;Dhj>Uus*dCCh?QjdflP zW%(q6v@hLCWefSK_Hrc+9?!`9g%$40ypLp?V?vmAv8=C*bn0F2IqZk3(u9FbHmrxV zC`@I$=gq9|N&Ct1Ap`o2B*au6@&ib))LwRqM{K?WntV@fg~NG0Oni4jFfmoo%rT2} z<@m7!!3TQ}z?y`uxLcb*7%)(7{Z9{YAvyP`y;ba?8gAdAKk&RrK7AfX-#h9>6bvMAvQ~yhZBOgsMia;G*S%izjH~_z zc}nX!u!8}VQx zp4bxIoasZYg;x(khDM5OEUenS(vH$Z2_>M64~|7<;sONGFc0y*{4Qz4yAbA z>ez8V(`N9~u8UZ{SGt8mGh=#-Ad@-t30q9yMMdE-cyc%7dGAkL!gDgAxf2k*OKOD% zm2cZ?hrtov1w-!wktvR3%@`%Z+Z?|_v3dp`3^s> zKy@(6UW8>ZaG1^_Q_!H-C&^`(q=&gcNDM^^T=~U5l>FEFN(GoQfsEB4mxv-U9xXG2j7;9 zG;y^uM|nVw-H6U)MeL-W&v>s2KGmXSU5IRYvSucj1$ZEn)(6wQ0W@q9;`FB*XjCtU zWJc@TlaStE=rZ5dbdsUGPZlo2?}n}fLj)t#^hcITmn&v9a>HxHsE&iXv278C8-u zzEV-jbSt*CIfbP-41yT)&g;&cOO36?F(XE1)R+}HI0c=UzhHgu)0;I=p%%arhzQ%j zUt}$YjiBa1{(t-&_NC0xgn1C`mP*_BPD9`FaZD~eQlT*$Gvot`BOO+3!^ZY*NhORHr?z(ynU6{ z(MR%K@t7pIy>5pDdLT#Aj_gwXR18FK*~gvE9%*NKXme@daNXUq#q+3BYV5n)tzfhX z-re)}?q%+b>CwVgVjWLXW6{s7P*UJiv6zi4)Fw_`8=1S{HZGr#&p63t!gp;dcro`y7*sX3_th$~xnR(2IG*RB=p$7&dDyR9XF`y+ZoFtMK%MeS1_D(p0VC z&pf+A(r19jzWV5hERJ!xz{DCEf0(TAky?82ha6ukz*Ij?;J~TSJqsV{Wjt^E4-*j4 znXButH=bf+%jGZVryO}m`mzhtX*rlsvI5E^Eq?eAW_;d2cR<`@fP3vLy416?L$6*y z2|siK_sBd9V%BW2>)Ez3457MY>(=@~B)odT*gaEFeM_0WAa`%k^DTD!Es_Rmb=C4HjyQ6L~yxbo4*T-YnmKXJM@o7SBD^oOu8F_F_ zcC<*y<++zelv(x?!;bEVi%aU$^UEUoyyg#((k6ND^1iFvI4OJ>H=kgznS6ppni&=D zx6w++UV*PS5DmN1q{i&s z#+FzrZjnUEX@*JC`#ReK4B+EUL=Sn2>|v@9EPD5D>P4f6v4H3KCq==8&=;0mPY2bY z7oskTaUE}=j5u>|YmU5c}SA?do!a8P#4opZW5sFeBw0(K&4 zf=;{Z`;1(JbhyUBnzl@{;%yg{_!laPufIF;O&%Y*8Wrs~`do(hfZq~rCkr5@hyJSC zU?-D?g$%qs2~t&xh^V`+_Vko)foSLemSYXjNU7ak1NF&=-f5MT)R=0^_j|V)No@XX zGa?N-4^e8ecHTL)|0&m_X4G7QH2Ux%F#qKTST2*#Vk0PbOhCqsA*g3VRG5p~-4PsF!C9NlQy> z>6dl8=Y76GQ{4`%S*<=~;xO?1%uVi==F=^miRU}NYc)?VuiiZG_0KKVf@r2B0;4cH z=RI5pxiojgXwqLXDKl%?bfLhqYoNImN32y%L)Gzg9LP|nne<)W(^nF%WoGBdk6EXz zdJ8$OxxRy&V&69*ajUTl=JxW%kRW~>-6hC1zLTRm$VrUe9q}AD(CK|pjda?b-s0N* zpx!Kv6bsfB4_T$Zv{Jz~B8`)Io$(-ack`NWcfe)&6Z3wG6|0RJ&SvD*6r+Xvn!_5c3lWVZ@L<4M(R(o@~mwQOgQ> zKDgZ{yR2cViR|Qh&9U;}rq*k3nCSD%U#EV{2(*{A5Q5_gF&5#crYv)A>JH=TqD3&S zw-iVo!=A!MyfG;nRl0sX__a#oD{GACQOwcCj>hcQrk`V9;_bNOxr@!Y+WNvi`X)Ml z&fmsA4cQ3?WOO;g1Ofb0_Vfk?S4bJ%WGGITq;f_);b&vg*U@-C$xBcluQoe;MWe2# zNS(iDG-}%(+z2jPfuu*Ea(eX6HnvC3L8bRDEWX+<_rWJ2(N^TdQzAeB5pOey2VR&`9OL;#{WE+IGu?ifX+1OhoqItzP~avnk4Pj~B>ISzLT|#LjZPJk(K`C)Oo}HV&l2$D zCQ7Jb<#R69TFRboyCutBfHg^eG@D6zS!=L)fxo0O9qF{tMr-uKds8%iCv2_HN$+^1 zpWKJr-xEk{n|JrkAZ#g^CvJ~Qte7r{`;qAP4Y1c&r!A5j5U`^d-%>?oEDr=weXQyl zT~1SVd$bXAYC`8l`*tgCTk;-WTl7Xv2vk)vt$p`F9t=w`xK(HUG8uXaWa#4%~xv-Wp ziMQGI1xCE)YxDUoyM*ttQi4*8i$%p43xCqjpoV{;NhQ$(wQY1p# ztoA)We9B*n$j~5C<;sGZPc1(pojPBjHeTXhg~x2+>%Y<6k^YVB*pAx;pV%v>Cx zJ4lv|4_3HnpPvkNT0~AG)@U2)et9o3T=J^Thra&3gAfID6g?D9Mqg&aNH(j)!OV2* zk{ys71S`yNlWV%lSQj()K-9Z2A3FtgE9&4R-I-7hSC1i1h32A~;#?7!FTJ;(?4ieS z_285&ZSTSlT(WUJw1I9C+rAbA~c{3Y|3`j04yN_N(Z$plEtmEfLAM$f@!9Ee~dUDtLvPoZ2TvT-sx$@V>J#jlo4ipz`!=a%+ai}+9`Fk#a} zQpzjK6qEfcblrcK=A(5+HGM!o2Yc{SwsCQKx3GMz4{jlFXq4Mia^0di0rYdOW^@D^6 zHRyWOW86Pqe|>?N-zuT*k*QE-uAGBZX#ZU;GtM3LOVf;--GY;K>XFA}#|`so>E#|N6t0(5{G zJSP{=?;x?^#duXW0aAXF&Nl8F=oK2n_NX<*mF~V9=d?>A^EZh`g@C%SkvHpMK!VZ` z(M&0C{@BdL7Pb>8Q5$jsSDo+K_*&13#>`}!T|q858VoMWkCu@YXHC@fz6({u6{e<( zGIyBc<*{x4Q_X`|5)vbn_Y+=qpKRNj>Q3QO#GFj&M7*3(V*!JHEv&rE7K2QmzrMtt z-6A||C!XE=sjzzLjQxYJm(W@X_Pad-4G?_iazTYh!<7QgfvM*2C#zU>L^5=Cu50-b8I%Gh2hzR6J+%N1>!9o(s~bmG{BV(B?S?`d549U zv+4HIPb@^nTw)K%0lAv|o=rvi!! z-NZkm2WuvD<)IBTolD^Zh(#(FnsFXCN_q+c}Yjt)sfiR8D#NN075}8N}lRHeV zp_XfAq(ZoS!#M0)W4{@TZ4pH5Xke_!^6w+dVt=DTluiY5`|u#;5v6cU@Um7`yN zhAH)dRryD}Uo6DtQDTyjX%t`UoT=KYn(DY_XXb2@B+t_$e< z8foOV;oAZ6+OcK@fPmC5kV%_CTG*=nnXCzb2rco+g;2cQR44FLJnnmMBb_bi^3}WX z!omtd#jO%Lm{*c}Q>d2RMj96|Aied}4LTpvElp==wfs7haxa0^k#KMCaS$0f>Lb}( zjI=f*bzQsu32Hc&P%fMD5itv$2=5-!Fvs5ARd|xgH@hQpJrqQBIrnPJ0L> z-@fu1IMG(-O5`~a{?0^bKDzlTqo0Lf&7+4!*}BU7 zj}@iN4gJzF?acnrJ*#F<6f%4-lyM^*k7mRQ&ne4r0PViZdAEG8f1i~&gNI4p22$3& zY2%h*-!k&ly%ytWC0u(S5J&WQgEgK{`{RXr^p&Ubw}g0Dk>AN0=e;94y8KgSACwpN`2p1ab? z2rSk3w@pNYZeYdg-{O zVFStkWaRnl$ zZ0*ib0{SSMHYq{oEb2t-<~!J~H4_QK6OBkT7V+OP(Ql8UmCv46gjszse9VKgPC~16DGEUv zrZ*>TBR_wWSQA*zdl5I0ch%*qQj5+x=<3m}(CQcR>X)G?`iluSA@c+_sX!PGtBIT& zkJ_=X$UzKR8(oBCpeJ}2ylO^)O}twq8Z8oq8Ax@LZ25*#!`Pbxug2ME&~J}ja4Y&B zJrbY&v4UCf&8omWF#5+TJ*l(FV@7BmuPkk=dUtuZm^;K6F9hlYK=Xv$t%7FP^KLAp zdFYEQF3nIevb${t-Oz%V#wnTpDl5d1R&ClosP}J$$mN&~mej4$dGS1qtM%v+B>r?r zlh9baLMCBVM!rbUTfRM|IXqg9u+f?2mp(WM0J44|ST<2|15+*L`DJ}pfHnGUkd?#U z7l<6Jm@XA6_`mC&)NQd_hrhbnV@EzHnQ#4}h$=cRM54u7?7BVI#3Z^WnWPuuzk?|* z_%*hC5Ecm#ICbQ_*lwG%?Y*RAO!e8Is45O%fYX=du750^sp&ggjSn`6B=DH>5xsth zr2uW5pN{Pr*dM8SWi%|>pmw6i7sh+R`F=0XzDS44ngww(T}^ChjwH64ubEO^IvTk2 z)AHag-rkP_Ca1D{c*YcQ%NiLX$S_mUY9|jc2}LN(+iHs1a(Rlp=;LE9ScT9_+?FR^ zmSvG4qCB4IqfGy6o$1!}e&FB`*0+4Cck|gRI3J6r3!R--`*5$@jlY&L4o(Z}6Y2o; zYl)yqFd?fZBYXP%VZUL+JjKj-#okk;gOd*dx5imm6}4Fi{Qzkp-sj{Yr2NmG0VrHm z6vXp^pyS7ex>P`etAlZX0WYNi1t=`0`6lRGxk7?uU(r4XoP{WC)){@mhqOk1GMKk9d{h1d!g$DAL4#w=)4TUqZ90 zbDT0Km@M0IxXULGIA^%HB2q-QN z0vh+>Fq#mO(XZ{sK@TtiZJ)<#@L-trPZ?E_cZ5fEsPEp}Vw`R$Z<8eUgJgdJJH=Rz zA(FV}HIC0>Q83=gnC{00c3$aCfa>u$|o$xt<2h7nwAEQD#gD95>JyP8EMik(pR% zsT5NAtXASwILmoBpX5quzP=uGvi{*RB3eZ~*6pH6^LFprs|FSL{loaGQk`(E{?t!!@9Hc|mImiXiFCGu`thMuHlf_&xXqnhv-sgkv=@uyFAM-x{5KB~Yf! zfmmP-6Hc<+qgGfxYI@cg-Q)a@7%%vTeCS*7DccODbNzZ>#V$fB$v{3<0VSoU#OzIZ zXlelqa6)6|8wLzF3X$v?9_!IBY~yoC2xPC(@f@Cbb8Fgwe^b9VpWX7!B+?mn5CRv? zCKbMVW@1^PP)yJ~$~G5{7IG@dRnjDq9`&Dgn?*LjpbzX#TQ)I{omyO6SK^)Yy3Hp` zX19CW-lGvpsBHU43+y-a-~Y4FGey^Qr_QlE#cvYuV92wjZ3M-~Btj5tPmQcn-GUTB zkT#fQIUrXUUkbn#(rCfvBUUcK&q+_T$Gv=)JW0e<@ZN3B{feLB3uO6`FiL-b1=4c> z%RCA{ZRfA*Ce7fj^db7f7i}HfU`^zf3cLc-Z4C16+uTLw*Uie5eKGP#;)hFK^!!;G ze_R1gvd1Y9o+u;|6!=qd(xsYDEdt}fMs>6yu1dywPxdM}Qwp@e?k0R)tI9#9d5h>N z#WzMum+a@S0>{hUC_XBVP(Z+UiHRU~X4IyElnAP3z(%_2=p zO^g!hjWruOi&K$+1aBk&>F;(pa( zshkA4d4{cZ5B|3&NKIu5saVC=p+W{SJHZjy9>alY@h8%+F za1<*ox(W*HwEwdJ`phNNmY)!G0c=Y)W8*>DpUy83`JJJgXqd$(!e(4p%560noce|N zpTpj-3Z?Gi@3&g%FmkjC~2>n(S|C&faN`ru={%f=+C1nDg$N^wc6XYdXl4rizUO#v7b1~duX2Eh{}2x*$KLEe}T=zE-OcluE>*^aXV zmM|eOUU;PeC`xfMi@3N9YK?n-NUk}OJ(_z3Kgz(iMb#Nw;P-ChkXqeR{R6^~{b@Q* zaGCB*@{-yV;Bg6<7Kh|8n8`0BRyfz#;-7*nF+p`R0U~tzfvp=F0p@2@b>2Fj&u#Th zZRX>A+$pq|pC&rzHcZM_sNx$pn1Lqr{9PM#GB`aYyv5#$*ndxYNvb%frj-1UmkvtOYqt8uclvt9Wb6VJIh6G zF`&xK@j~Dp)^%YMq@%n)m7_lvb=U9TmLN?q7D+EG-+WYoiQHjHi-ONSAWzwD;~Z@? zy!dHb)W`siVgf-uYfOu%5h=YlXG5iZ0^$%Rcy`S*Ng({bTq7Nn)pxEo9hOYFJM9y^t>GLa(liJYB zDPvh6&g(BiA;Yl0mg<@vjiCR~0P+;r4F6doc&9OMU}Q|z1XP9(^*#-1v6+2&{$4JX z`p)}L%t40;7iI6_225QN@EhKlOWT$!gdI4jGVAJFDLcQH)F*N0f*jO^{~~{NKvCSv zkk*0gR|9z4bb7-fdU-(hkdQ`4+zxbGeJ5>~WHp1ZexTxR(EyT{JI)o^Z!T$VkDwiz z@MMF$;AvFG6T;S>*NasDKp*u?02RN)dhz-4qCXfD&$FkRe~OSo8;h*419hT2?x0!2 za8rlE>_8~%qQLxg*i}VYFLr71=l%DFh9}C|Ml^RxMEg_eN^BAs0NDZ|w2oU2IacU# z@P$fTJic@i1_Gmmbo#DK{fuE-rN}-x)f{Kej6e2yMCQn>Z}?AN5inGm$sCtk=O3fD zIh&uz&W1B)?DSoUlXTH z*g?BB+J727wj=PIzapd;-mfY|89dBAp0%?oo_s}t(!G5cy1^D*Jy0?cF&9KG>(Adz z=M^lklzx6xNtpzZ$e+w=6mX7bH)SE(+-@+6-hA9$$~tVx(24b1Uuu46vc6JG9p71r z?1PoNj&;ybug^fR@n^oR6|19#__?fC^}&aZ+PVZY2!I^aAMG!>^Meu=*wGR{s#-}W z+c_`{L9da5t-=#uj_TmsOl_Bz7ou--BqMZkg>C+WieJQ#O~+mqgN z2MsUv0~O9pKakL$wVNZuj*3@!GBjuEFCa4mXVZg5^Mz;$Kw40vT+p#DRYx-;IEdL4S$q1MUdPN#+m-rH9(SqXx1UJvH{?u@_!w#p> zA$_XO>gF5D^!4WOe6qwO8{($Y<@L{a$V1k0IY=1*!3oD&_hbTPFKM2@+V)5=1as!P z47TyOaD0ePC+oJ&;TtV~+(ycpWf|Rqk9_VGoH>9Rp?pOWdxb7b!WU!bLV#W%UO*=i z;Phszc4xgHLz{~OcE5_*brxQ%Du_snaYGfS^K7OTN%#RvZ96s1jJ&g}TpyX=Tj1;` z0C?`VOPewri$nQ1I(rGV*Zs!$Ki0AUn zX{O|`o>zdh7NA?{U?eX|v1u>nszRY>(FVQjB@=D_@N?Mo3P-P}GX1W?Obe;$`4(J; z#HGV5Xotdbm(#kQt}7Of{}zBjia2-drfU@fXT*b6K4@=GToXBpY0PoD()hz(ey+~v ztyXAMZ=jvXbrP!X-q>|RKMDWii^U`sup8p}mmG(jS*Hb_-TtOJ_9L9in$&l?Nj#ba zHUpe|xSxg?R!$1=@)tLONXIYfi+VgvL*S?peTau)pwsZ6%HiYI6hsazyK<59)G_k@ zc~hfw?SUIk!sOIV2YXrE2TJcdP4L>LRJR7_kA7=|Vr;y7alQU8A8RxZDXu$C9>^M+lT^8{8`kR zbb9Ujj`1&xEzf|f&|7T!_NY`}zFL+;vWvH+j)Ek9;>mHx?b5~py#pOjo%!M0GS}7yp2sF1y{|V411zXIF&@Y!EMNnig`3(#GKaS= z#*mtq!Za-Hh5!~X6rGoRltvR*#J~$Ag z%$A~U3ReIuA>d5&kX4Y*o9&aiU;Y;RwDRxZ9O}sf1u^^}BZ<8u%ZtkPBe`Fd+}t2O zFXJTpL-#&w5z)Nd8-c|E7uA`91NycJ;xy75z35X{==#xfggQr7TVZPVOPSUYl-YyxJfV@qk$bs*gk3{$lZm(dCrp{w#F1lVmK z)6N0@64lDFj?@kfL+;pk-A8yBLM#gyZ;X6D6ZCH!5>rw!P5}C)v0n7&L=jbF^9{2@ z7^H2X_Cr{%>iE%7z|`wm`}s^Qx6Ngx34J$@p3LK(X2g#GW6PydU#vAgpfp<`WYH!s zpe)?w^k4Y>JX?-MRuBb;1zz1b{SYPzwwfR7@y?}pdnnmKc9}=qIT0iD+1`u z^iT~dasM=a(MwQv2ijDE1&vyHV0gQfP?Fblm@wZyegw~Uldakh>beImfP)B1)ofsB z8wbU0MhM4}z!|UWz;pu!1D|r*^g_iaDt9ghtL{AP$%N2wH>8khDpI1?Z-&^&lr_*K z4de+A57EF`g&b0x*o=>{nltI!(TD8;%>&;LX7O+%18cC_Unw`#ZEtpaEbvijn!rhX zhBJz0qMCTg@O@IOY?hPQhe@yoX>2W{ObHet=3Ql5K*zjFGDMphj!;m?L?@qZ(y|m5 zMm1_R`sd8wz50WZC9VEzX1%88K>Mc~230(_2MjG}wZ22o*}`6TRXe(m7Ta`hrY48< z0>-sjiY^2JITw)ZO7-V(@*i%{OV(?hX`wjAF`?N;T1@0gxgFET89y-yFjBZg`|g&J zH`-^+>kr!1wPfs?gVNM{;TbLFaTPt#!&F5$g%x3IO(RU;1MXXo2qv~UpGNDlEPD_0 zm^h6z%tt{#b&0DXe;gwQlKaaTt@^? z4lKLQR&|8vL|>?@%*JoG4c=0yRw&}9!gXsic+EBZBo_Q|&Luy8P+ z;;dpCW8C$XiiLSF7~da)0OYr{WCoJBioO6$6Hx2rvQdIUaSookfGm(p8-|y*c7I|! z0n%TjeA`FoUp%dfCg0x(%M@wuexsjJL-GiJK-WfE7mv8q3}2%n7x@Cw18+&Psvo69 z7;za$7-VQUbk(CFyHtsSP^#Kzao8Ay&2yubRr5ylS246p$f!!>{e960UI<>!MSAHCs#rfJHk+R6<71c__R0m(P3=-TJGLX25tmXJ-U; zWfN!E5-BT!P7uHTQW(Y&y9mHryx<%U4P;Qpvq(kIyq@B^AVx_N-(vbcLHB^6x9oHf z)w$k$Eevb>%YHKj_Y`Ynz8i!Z{)8M-s}z2GPk!~rev&DL(gV&x&>c$sMD1$7p&f0r zWC_NP4W*HcOyK(H>FVYxC#^;*mWu}Dz7ozd7+iBoVqwIBy1StxRowW{b{lND z>ZBFVH93k~o4gt`t7B*>YJCJUi4j(!+k7$x4kKy$4QRd!*l{SBCTjf@E3D)7ZzU;S zxA4g)sb-JaxDn)W5Tw6>vM&^s-LkpRXJp%UA9ryg%;je!QJ*mG8kP(Gc(QNZY{Wvd z#0JxnvTH?Yy0fhO@%ov~8dH_o?KH`LYS_&Ed71`FOm(k54OO5x4n2ziIdcy81sV)1 z-2F()sRRVE!(8Q-Mb+-)WZVE`{@IJ1%6c7AC8+L0bQwZAx5c=ZpiyyDi#aw+E%kZn z=2I;~#c(?{^3@>Qtmkp{NcOVKtz*%_B|sWu&0+@OfR%y{4)VPr2#W5Cd<8aG*d;&V z>)+GR0*=?VZ-N9&nbfR2v4;nLv(VD+TFY- z&ESadU)z-+ZJTa&u;_ynE~)y-xIe9yZ-HfX(_QMtz0El(208t@X_oU4^{uwXZtGcA zzb~d9DSm?Esfd!M+k>`sB)pQ_dCGTp>%s=I3M`B9ujjQZz*u}jIZ8ffGZlw(VLiot zuAFW2_Fm~OTF{;!lU^4pQCjzt;=)Dw>~}4l$pla%=5J=f_Ap7Pb@_(b_9>7eG7Z=g zlu@#4UPrQrFF~>=3b4hr?~^>I3SR+cn60p0VLxTXyHx^4V57I%)EFy)E1)AvY#o*2 zMeyfT2W|wOXnGxV=q&156SJ*i z9-cr&_P%3J!ka##MJs|KEOc>jmS;XWK|xkfWOUX(=7h|k>Tb^5TH*I2S~`Z~>OtVB z(xw2Dk*s#(6XISzHy7A7JvBht9luVoG+ixf1cG`h2M*s@T6W`H={vj;$J)E)*A?)% zoY_mxKa*qmV1gYA<&1QXXeXP&EetYvA0SQ`vF3cEEY)2TQ6qDdtXn2MOoE@5@cDHT z!fcS<9Ai4qxax8)^cP_$w2kK&6lsYY_cyY#x%oao*|Y1TT6~pQK2mw>F|l$#5ML?> zG*(qDq})Lzj{59RjFNb{(HkU4N;6tK&*K2zkKUaGLOa{fx3RxI;Xr^l?=D=QflcA6 z7jCc$d=M0kv+O(sHy)hoO_-`7X-9$yng*GZe>lXYU;F|PaMmAj&{NC)jJMUv2s!_F z(D^!?a2x5BQ+9((c44_F*T5<^&bPt|GHS8A##LT4Ug|0Qt*9M~-unU!b{^)0_WXPa zGeZoOD^EXH%NI1uqi_~D_AkSCVEB_dIE!WmULbqf=1WXAq0Vr3jHP_~PVu$;yGs}E z@i9NF+B@oyn~a(GPb*c2tVv9(1P?_T?(>&b8ktvI6)A%LJZfK`)JkSxp$_)-jX^gJ zhC{)XtC5Q->f0q--NrST)AVm=#!uVJnr4~x`LBdCBFDoJ=FY18pK!nc)e|Z_YzFHu zt`Z4Y*~op_Tu&G=HRX@tq4J+vcxyEOMTmFea0?XJc?BU|?S?X%!M9R@QHf1bF~&~Y z;6;eO9fpEOQa)yBtzfKpYUinL)m0O=o!7>LKtmaa_WL4j-;4c$!~AZYi#PeR{4XtN zA7k!a@|jopw$7syBD@><749eo&|5ojOl+vlJ|%RDHybA{;|p%k(W_U!-q=#Sh4-B< z5jA>9-6wi`)MT+?paoa|?m?xwgx;cm4Uq>F_UAG#=mEpe0Y`+|@6EAhPs}O-Dd@oT zmJ-@G{Gsm%^in_GV!PBr9Ja9Uyp6GVa7iqd#2!qmB_E~jVC$QDslg}? znj`nyFI%f>1Jin704jGpErGQU!GQ>3TU9_pkamH}PvZSk>UN2EOy$6|GlLz(I?D$r z(AV%YPgqPbH&qX(ZfBJe=kj4yky@_ZZrrqs+SR!cJYq|h<^qFIW7d0JFx)TRcPIf` zgxZ>6v$W$Qj}rJ$PnMQ)IoXK3b0-58XMy>>O*V1TL0pTn+FY-Y@D5aMQp-(>aC^Z6 zp;bBM-wT!aLD()8RCXNU-&4{U5ImRoU{LKI3`;?4OdW)jxN5k1oopRZlK0~Yu~As6 zdM0vmO!zl7FXZo3m^tj=OWDIX$$5G}$6Mq;ID22>@u~+6%+k*Jwi~UIv*Ifo6;h)+ zka9Kr_2ZiYP>3uL90N!#tV~`N!conj9&)cabb~BsgWbQs!sND!kwydaPe#e=Vg%qi zvVi$Q$Mmi86uurQzy!L0U#!Rd#n|)sEVdoxwM3%U0t|Cr*#K;ys>1@$k48E908S;T zA+NVsCcUJkSm3C`sxY*cVLv_tagC;C4yC1zzP7yU_I*{sVUmiD;hW22_e5{#WMhl0 zjxAGSAT>lYU{GY2qyYXk@&=1d(^vf;ewF%fNMPY>?zPL+^^Q7M#-zd-SI%!}-4>0IL z5?VXS<@=ZQ=&gj_j@Qj_y6WKtIp(Z)cA#m+5YA-+dvh5W5vZd*4m%9U>IslmBBQj4 z?P|dNAo|F;G;DdeJ{YP-BlTB+NNr*tiQqO2L9~OgcP<=u0t|S|4ck*$P2J>*2vE%T z)}qZ6-=&9JcxRcCz}CQcePZ2V{m0NIt7;e}Bl?hl5)v z>m7-{w|xSDDI2#KS+7c7D@hdbP{ydZO~EoA%q!~qi0tzn4JwQd5e*x?Aox5vC(~-thf)?Pp8N>`J;g|M|My+(x4PQ!;fo?ZS(5^BeDv%8L zqYy+Psj0hwA$iF%0wPvKT~+kCY2olX8QbJ?P%=WbC{)%oDo}U!Rmi^;ZmP<_DLLJU zSGu$qX1zxru1uN0-~w4IP-ryV0>b3M1!#WCn)nLJI;gJQLx9?yK@_2DHv-L8shmY0~j6K?H5?8KGB z{bWXWaHMbtS?p!i$dtfsfp!dr+{{A41TVZ>z| zG)n&i>ucvU_iA#@Kc1A9R+O$pR2h|AHM1jTr(7RDn@CU^>#`{$T8~Ey%!eCZs#QP| z&9hq3smfFn&0OO_BmqO)ql!R?INP*>v~pD8p)WJBue>30+X~_N)!+CVoA)$!1{&r% zcH<8lv~}F{!M6VhteXrR-8owO{V-h_P#s76w$(dX47Ez+1aMt*3m4=!(q{RMS8zs2 zysbHZ#!5*qQx~fF4MX2q6b-OYzm?0>!%-kmy&%rEPWtH~-@@K4xZqLMzv1%-c;YZUcq}9)|_Y&t?mwKcTB$wvgVKUZM>8R0TtH2V;&Ij zl$*eukAdTu!|Rg%y{4yjn}DCn^^*8N-pco4W7niX!6{uPN4ZY`Enqjq+H(;1wN&dk zU`p%|1*;f|A&<#*&}5+e;(@ZGZb$Tyd1(2vZ|6pg;3gI=5N5*22xstjK8RhgKmu%w zdo^h+H-sPn=oO`JgXkQKVlk2G3=&viUk#2-5d>r2!rGe+P+^Bgx-?{ z=8QD}b1sfp%oV0N8uE9)Fg9aT({r8eSPYzjxqC+*VCN|kyT4z0mPr!FM# zH>b~X4&lH!r-3v=QyMdWg}duoV8Ne^+M0XM(YTqP*Jz(Degq!rHj27!1I2%a;XXQrUaDO~W{HutOBE}X<2 z;Kf}El=G)GA?uc9aDTz=%FS}c9Y9u8>T}qHjd17JEjeUO{EotL<2XQy2S;8)f|~6X zaSF`{B{pDW8iup>QzONdSxQ1rrKYRl>teTD;5^l+9<|>_NWSn!o>}vkOeqG1hbs?q z+ScYS;~=c>{^Ty-LFCg6^MDd8LScJ4kZL}j1R8MbUC$u_3%RX`0=TZto(0@1{2r9> z_oWwK;a3WLAu&rC>r^mfD2L@7{tM_va5-#5CGxA1@}rji=aq1vQR^C93qy->g*Mcv zowiw(7kqcjv_$>bj@|K{ZG2N<7NyYc1>Q9OYfeH<;`qUmMI`8oqFOegsPNSLzEV$| zH2FE*`)*ZqXJvPKq2{=Tu(0sJ1n!6X*lN+vU?(36GiSy!G>~^)CAwHxZK=`1g+r_E zZ3SDs-N>w!;TW>+bKFo+yK7=YxQI{idqp%5ooFe|1QwMmWphtHHh+}JXdA0I)zJ*j z3-aR2@moe9XNb6wv;y6G30g1tW>*i<3*+^%YRsul^AIC`5!w zs;Kk#M1YwcRETUcIwu5tim91>@T!<3&}Zi3S;vM#etf^^Dy%0TbDuK5J;H=?>?aTG z$NLKE(&O!grzgS=TXMESqK`- zHb^QIuCOt}ugpAL*U;Ni@%%^CS`1KpeV2FU9sVUU)ro5>=*^L65bi%_u*R^jzapvN zJmcA!iwfllUWI~@AY!(d8CFRPKO1}548D<@`>$8N=jI48cZ}dJzjndVx17nKbh`+b zhKiBSBP|3M%p1%#FvJbQ6bEgGYv@hysx7&0MT4sJ~{&SibPN)jNzub%| zE+QB9ocDHyeWlSa8!`O?0@F;h(L~>yWUyOa#6jna^+31LO({U2q zd>@qBCYLBCE2If^Dpm4szE5{?Ot)Gr2Z*}jBG(m_cG7Cb&^Nhw#K3)S;WPKVDPF-1 zHl_AuCKlC=@Mz^}lLX`8!k)vh%18;QeC%r`y}c?9 z9b*-zF+24Y_x%A5H>-uJL?WQo{cQr}r>OEEXntCAUl6oz^7b03pmY4cMHqYB zWw(wi-f7M4Ho)NlMc%h+w(@hV@DTOgE494PA(ARYRFk3>;8>8o8P^^0+!%%0UYW(5 zJuyQOejO9869x&He9$$Hib3~6C+g)Q_vyyAqKI`fb<&hBbpnmpEkWG?#+vE- zEid#zI8tb8qRAvY53KpmOXqHKxWB`sw|ntK;_k=U0cGKdUn1UNRTf2VOR{da6KFhAOj<4dWIIRCJ$_y%c`AXlVhpbCeLAIQ7VK2oi z&G})McULz-$`AUU2l6(zdK1F5j|WC9^biTU$x)ja{^<|$=a3N(XUP@5le@Im8s1;X zf*&bT!wn;qNjz(4!4=8HnHaOvh!&CE!^}DPEsP7^aWRTW*&?8s^%pontkFLYi^d&% z>-Rs~r7Hfd0>S*N-YP|ISyg#ANlnQ&0saw=f<8zV;p7VMA|cOG78Q`|1}$1DeYqXN z`*>-$GQ5fg4W5j%g3v;k0P9fu*O9AiD0H#3;L49@;Ng-L$l7uQBa2sAm$%Rx&ZkYK z#Y%~wA8e6{znPjIvucZoEwZFAkEG%y;|hrh3<)JX{N<9Ili41H(!T_GwoV`#(i1oz z3{d{PhdZ(yt&Mubt_!i*_7)U$Z=|W|1HTg`j*3n<8!&i=5qvt2aSV+6{%C9R+*6`~ z_?fkUkR;v$f(UAO(!}v9Vwl=Iu`KtK8=95L_nADY{|)=c{y*Ts|FC~d99;k5{QblJ zv2y(%*gtk=cCP;~_OHWCQBP-`Es6)seNAw7x3+&PsI$@09Tt5Pq<`zbM^I;Hr-9S^ z6@OW0Pj4%wW{d>D=QzVz?P}e0Xb062JjMntpoo1rt|{i1&%}GYb{Q$ErcUO z0`vH#lac6FFL|bzQc@*5xofUi(i`$QM9W(bEKE zJTU1KJe`d$F6|`H`?5HY6gL)kKy~ay_#B*|r3pMEP}aE)HmFSuu(TjsKtoYP5V4@F zko;RVG1``dkr;)J4$jBV0daA8F|9Z#5_*Cf3NVnG1(5N0l(hdhHE@l^kr(h1BA)fqYfGk{_w>D&CLCr3q4;zm@hk4el-w*tZ6bif?$YQ%SFHGKA(? zwiXXa%~K7$00`a_UM?<>m;V~SU!XYZ4>1%N6xZ6u&l%B!efpRG`e$}W!4zP;iI%VD zqVw}6?o5Y=tL?{x{ioYT1jibZ1yUm1UkwZ-W{Nh@f0gHRjfLf+*M=FS0i}pIO1f#D zV4$%ywf-$5jxG03)q%mB|2Z_2j59DN<|~$z}z61K!Q9# zvrDrn-wIW|DffWMSHLGz&<%bIdkc_>uMFg_JB!IK0Caj#Xa@_*p~)S@@8z}ZpcfgJ zgr{$Ga`sN-kN^_=`tLc)@(4T7w11{R#^=V9|AxZb3Q@^BXQ4Rza+)0wNdXgXcXjMR zI40%K50}Z$g3RLByt4WZjO5!6=+_RVFAa-K>rdQAK9brlL1oHMyeSK$5{pK`90!Yp z$b!GzlYtww#24ZmtY)?*h=AK6L-KvTv*{brYP<~KawZqNz|7SAi(n@l9iFZ_5CH** zy`@*TFCOa?aVNh2#|X9nnzYO z8k)Y~JCSj5!IfT7kb<(t7ROIX;8kueKS<1t)xobdNp=LF?puNdgNR&UL56n@V~AOq z&)8L<$9aDeH=yoGe zY0v-G(c*mr3utnGcOS-!$CWa_m6?7zd(aIbG@0;x^K@awPtKqbz(H*q*GnZCC1Ug62eDq=` zeED8|k?1JSjw}VI9f_ni$^N4%iZ3II`N2hhefS??e_!C;uXH~-_I=0x=88#8$}pdspTI!A+zreAEFCJX0Tlh93< zT%ny<=l&HxjyHO1Sw_`icQhfL>8-gTI|vOYC1y-26fV0=CvMzT*_0P{rOZsM(eGo< zoyXmhg;K=lLfN;u@`y3S_LpdnqL`z$(;o|R8T*)84Y0G1JnsmI*$8W8T!T1&{f_u% z(e2eSO!N|G;6ZX0MHz3Y0@cdh&0@_;v8cZ!E`0eA`(ysWna9vxr@bOm-f(&3aH;e`mBiake;Sc>)5mbmLrBUi z;AbD=7j~Zvd#rqF;KXmBnyXjIyId*U|B>il*-$4$WgPu7?@<0&%Lck}15xUI_xzRw z=liwbxeHxD9Jcu|V6YP@a{9Y7MPhT@=ZC^WReK0$J0mKhqrHtR?C79cbe`iN$O$ot zKPTDt4ne(=2v2(;k$^Sdz^>%$YCz#BcHBo30A(_uJ1Vg%&aJr|rt7Cdg%^^LE6S8# z$nhdDy)*{lg728;jP>i9b$(rxvKShXdIh%jDkZbehrTJW=(7j!af0V|U&HwG#C^$D zpdap$(hlM8A-Tp($)<^2xiu60Ahr*!%xyT)NLN^g{>&2{CBcU?xTv}r1fQrhlxybb zq&p2O;|Mw?3gQ{Fxf;hmZE=hNXl}B8Ig;{9Gyo)gr_dC_=fmy9)+Kd)qDu0^13ac3 zdDK|zh|~PHGdovznLCo}Bid4hvzE9V%cKV5KXVncAg+7-*?T+u|?RH@+>*8 z8o8WRl~`=)AFd(j5H4G!{4eE(v+w6@hT)0th9gOtF&3mxL@a%kl9TX%+KE{q<*-J% zWdvs^8pnfKXWX6hKsmR)7IU`4_G^?emn%ju@e(m1sa&rJC_7i#Ss ztzPS2#C&U~^>OqiZ#r-o)|UX>GVIl@mtPu^;Kige9{fFxPTTk-ltbJ*zTo{SqI&ZO zlZAw_Qb(bcJl|A9BRH=JlwM2+rwhg)qx`M|3$+m7Z*h=nIG>9>(|X48R$ii8eKE7T z!C8j%zy3v~j88?ETr%*J&|g}cUVgYbac2l`5Een^tssvPiy_G4>!tZ7AnnGOF1!`K zLJ!tyQ?ogv3hWp9S}~d|%1!0jO_jsu$As!1+q+(xkXU@M>h~D&FGOe*gYPDcJbth_ z;KRZ*Ja*mFsQ>G=it6TlsazoWhf4c8jR@BKcgA8?Z|h7JKF1;^%l*ox6cd8Cz%3mC zOb`}IT6IKRU)t-|2ak~kt;Bh}^As3vv@Qqh15KQ|?p_CdcAo2mk6NxBZL@2+n6Tuw^AU{-?w`P;e7=-nYpMUQCLNIc-Jf&{-aa5 zWcN^wod59Z{=q_;F>#fa_pS7m-1mx=C>w7z#O=j|Bc24X8kMer=c!~7JnA}&d~ssR zaR-1FkyoQz%r}Z^B}%OOG+jGN57pGv`dXKTK>hg~`u-YQJV~Ra9LT`{m>a?(X{gEg~ihQ zb-?AM`ztDU0hhz7Az9tTdX^wN|FT}%76YAPEaXA-Iw(XbWnX`%&pO0POIMU4pc9zN&&%Ugt+Jfpb_kNTG)KFyEz; zuLK>q`SQYy-!g4>4N#6lVQJOyYZTUGSbQlwxj6Yy`ig9tGiKJ(ZMuv}rm!pB%D&)i z2Xz|gjZ7xX26*MCohPr9Y0S;|{M?UT^pvH)?0${XG=*wh)d!Awwp=WF0uNdEzA^yo z@nB{gfUq`?+8RQ{=qy$wyNf@l<7(?^B2BB|o3v|-Ls|J`$N=d`gR!Z#c?=Q^W94-+ z$-_kb9y}hWI+09Be-#5=MBPKm0yBN4V6;BXF9SzT71G;0v|3)@SO6!CHb10yl;;1z zoHC`FZ!1r`?-#C>(&eR1I)RdWjul2eto#sja9A|&seX>xx}sd$KP1tnoo1<}`WImO$9%Fy$8^rzQo* zj`_v}SJg6o1A&`PVb`T`i&##ZdI@()ja>Zeo(kI#GY9+SB|Ju=g#|?x$bZhJyLYm+x0Vg z-MmETZ+RDoidASpTS^mk*avQ#mS*PCMA97;+z4Ry)17pm_6Rr$r_lAIV#5U#d|LSv zUn;S0l&V)`gmTzVWKO?MPT?3Xcu7=hiHD5Vn~g<%{SI+TeW6C!z*3a1dKr>rOrhp@ z(;holk{Abki~&7?z-P7CR=>m43#^#ao<-6QuB0P=pI~lf7y`;Wa+Z=)B)<^w4kABv zW;-MWSd_6VA3FyWHfa{5XxS+=;TSS4R&_5$+$@L^=P^ZKJBOD>^lB9%Eld*$6D|;{ zh1e?>7Ip0O?dz1_>4cbz)%0>(;x+C^Wr$3t>T_cg;f(jbYf(}BA6PS5#L>+eZ_bIc?Lpd;?F!*v zh}&NX8{s+Ua7b8#h~h)@J3YC}r)b>oM-{^MdbapY^tN8eXztS-;ajewZ6|&D;9-Bs z-Gpd}+MOO#;3^4)Co4NoNFSr2!a)JtU??#UF1B>VH-=Fnw8Y6GlzxA8GN+@tITcIai(|HlRBa~9Y%OMI!nqBf#?y|&6CsIuq0 z!3}0M%=vGC={yY;PY)&InSSU_mgeRvNgi_hda`@V4kWKzkn?0Jdf8u7%$Tf|5lO=$ z*F-jzwjEg#j|e~;=267;H0nD&6Bf~jQ3OgrXAas@4Gbc5#kIC)Ka!94){K zhJ6ufp;RgdD_+z|F40{l#cZLe4i3v6`uSZO0xbA+5*9>U*!bnO2o77I1cqNa7S~(+=hDx#-Isz~ z2lm`D^X_xq>=t%Rc|GW;I)P5=Sqg-tU*va9i#0YRGom>utyYo#E^Ov=^=?o4$k(@% zF^z66Pa1xR6qa=^GD^{)bvBVp1x65#BA3)L{tS}r9s7!7`QA8u7mX&E$c&`JqIUG; z4OO&Z)uv?TsI>|KsaNse?4s+~8nq3K3cO*x3|urDVZRD9vZW$%?f$!(9XKlW146~# z&k8Eb8unx#^nN}My>tqFoIAI&-wJVI8+ob5;+@fJg3lOquQ;(rskRx`4LINDBt$J@ zBkn93NA~?%1#)QVa|{E^Niilgc+QHGb^xcpK4`rq{sv(lwfy;=WW!|F+}L=t#xgrP zm$8-ll*)SYf#3sFbe9!5xZpfL5ds2*#V`_k@aM|ML88$;uQtYRc`5n2I7?@F77mO; zLFvV^^&lV`Md!cUwxe7*1=cbEBInFEEutZ9FCaMZ;p$U`9Dkdy4nazsP^IACg@;7% zuH1#R>#C@@D|xJrRMT8L^viHAi_wbhYUd^mm-4iSf?*Ysw7^w?C%-%_mDxWTt3=+A zym908uTE@&ok%6dwzxtvg`bCZ`F?IzxGBMbt>)};(TWiMxIK$!DWI@ghanST(Huw4 zOV%4xg31TIjqr<1R%|z6%LcF!SgKxyct65f@*k%vZ>eOUz4f*6jX5^T?=p>O5-n)@kw}dZ|XF4cZIUZ6L&8&*ydJ%(^Ev7huf6X}S>!vxBk0YT z1jX*X<^l;nW>wyhRqWh!fw2PICP3}7_zcHVQ(OmO9e?Gw14k?y;C={Q*;{@iB8?uR*F`C;G$+gz8akn~uiQ7g=0B&2PPXxXv zab!|a^*C1A7<%H}iB=;{z?~hPZIEATKfQSDmG}~6$Ex2P?~eWoPFsi3<;g*yzwtuY z6ikL7E9lb$+ty~z>MY^g5#8@5^ugpjoWy&J{wIS-2W=`I|Log=iwwM*OYp4XUyTYxsL3ECtzyD<{5|7u!=V_khPBYN^LBO%Sg4{MlcM^h$Q^qDmB7;T$3 zERfuVUl2LK3>(>JI!a8G<$*BMZsE~{pgmS-I;+g*onisH*Gs0e+rd)HGZU!wpqUVeTOj@ zKkD}mKff6 zzgiq8|2`Wj!Nwq#z<1@}y|fX@X4Ie@S!-r_j~K1v1p zL3wV|2`HA_y9|~~$*rN;ibajVY7_N7Ne?-c&R|3Vhf3w9pONG;GP?niLm`9^g5qp* zXJh(V)7Ilef9v{UE=L!v$0;nxlSy&C^M1Y9yl zvj}r)q2M5gXw69aBso&uB2$C8U9lnKU3Qnfeym^L1L`k+&yX!#O?)?x^nOz|vfp28 ze#G^0(NS@M8le;oO+Kn$;VS#~q-0kB_cR$BQl_Q@XfNMcf2&B{51iVP3=8U}3B#ck zpNZmoW#8zCj>`!;JC=|$)WE%{6%obtPBGgzmuu2d%Dd>n6Hi#8yccxDqt<$}BpoDLZP5OID z2yG3hvil=Q|H(j_p~v>*BK)}Sq9k#WV1A6=veG-0k)t!S?b&ELp&pKU;IXvSU?Fr= zuU6ADVhEqxH7SJ96dKyxaTDXO`CJogW(k%>_x3h2B?{59r!AG>@Mz~x1D;Rx{l9^a z_wls9;yZ=gS&fWRkJ!^l+ZR^y(P*9fxLvcudRZ!@V@ka1aGUW@JLwpc*^jl(m#>X} zO#=&R65y?AOL&qE;Un|>Kf7**^b?A9=ywNa+MgFYyPi6HjXM9jNAAg7M?7Jpk|}1P ziEvkcQ%B#|L}IdKDU z^;$9H@%sEl-FuNr2Ex^vuJBn}I_bP)9ocEc$wQ{{(E)&974#-( zmd?(PrznHp$!Ey9kwk7vk`4gguO*%St?S*{V3(RY>PXv|m1tic+|(Y`svtl`Xb(&v zQOh;5XmE>)K0{aVu(mjbz!exO|w^5x$wbSALCBa<}3ee!+4HaiMZk~>=e19hwJrr=8L2A3_B5R5+ zYydF#sJGL-s*u9PnvsZ4e~zYi+3iM~N3Oo+tgiimnEk%$VUOdwYJ|Uz0HKaLg(^zl zA+e{_VuYl(SMa8xVipBgfG(&R(;tnt_%b=hGpo;AaY4Sgsk*0R&6fu0$tK(A=q@0q z{4BTDkZh(CQtymsl{-I!n_;ncDf+u;#q|c~sbOeuK+Vw_?HB~H%%NWzhb@P;(04Bc z;~lJ;#ToG&!$+#D@prj^Bsy}i^?QVpv-AsRs%V#>jodx8qg3Q+j8n6KFw&}}0J9GY zWP!8%CAOcOA5%RbO8iLFdQL zkS;mYU(*Sclu6~-Y(j8aa$K<7j9;Bc?@iApL=hh7!P!`kHM!)z-(zo_U?)OS^tOuF zkuu2}TPH8dw*~-;>%}lkSWk9YGbX1S&O2SCzPYOB&ocs9P(aiTHalR%Lj3a1FKiLf zH$OQaomHQ;!hOM%bZ zV=f(WAmGc#U*n;9a%vjRF;7mouw7fbf6jo7VT&|6pgtFw8ZhYy zkS)8?`6b$k;^SMu3BbkPsBRG1*3yi&;w*rp=fi-#T>cEV(N1+bD{_Vp zF+x*^f>=Xd7~$@u!X}9}Ols1bMJicXk*L9q$91b$yl`Kj<{s^P7w}%{YTjPr6H75O z^x1??z$ZcY3AIF=&pxV4+$@CY^--BMC!XV?^+JN|YRu5xd*6Dq0?viY*nyeR5@5F+w8+n4`jOOaM z6()RV!X2(*F4(M%n|&*17Bxj7*!D;1-03-V8AAMm@1Mz~Oa+L#4#L7p4Ktw}G{B2# z0-1{jml|A0z^(X7^kxol#o$+a&5L*%8&jOavQMMeoR!a4&5o9*}g^4{4K$me~JNvaqsEdfk@gB9DXC&{yq)jo`r1F#2 z!8-A8QM{p{$w*8OrfTEd2+U9qj3jpUb|lL$w6dAAh5JvG#Prb%d(UA?sp_P5Oy*^S z)*>Y8j9p+?-hY4bEa`-8Vq}vUBf={S*KPaMoi!Y zqtxpnG+~72s4^@HJoU#5c@*L>7cn42O)MH@j;{tO?{$evrS^Q9kfmMEGW^jo9OzJ8NGjGM3@uq#C*#0qp`z2Mat1c zd`n*bKKnPsI-2-=r~p(Jx>w*;&LDE5pG7wUWAkw|P(wE0CsH00ps$i7G|X=Rz4g}2 zV2~YHi5|l+j14~9`tSzd!a(s>f{#SC`7GKW6eG5+*u!tc0#iNK6Xm;Y=HIWVA^NS3 z3SYlKxaf69E)j)|2wulBTqYp9sgNtTK8jhK5FY-<0@i093cHZFc}DaBDhO0#cWH(7 z>n+XbxJJ@xH`4`vdSPV0HBg34mentC*M*1KD9U%jPRCF*ysdDWk0kS#U%pZ#r)EYm zd=h87ZT7yBG%xBrn}glaWovr}_+>cFkK|hE<>^_ng&1A$tM&mJp)Y5 z^o?eOAh=qoDpO9D1{3$qftPvJPHA$k?A*)U3J%$-sXi0@xXq8~Hp10$v$lN>kJilGUp_{QaWDRR<2NwD=awtbuF zr!=Esfx(1}zb_kM6ss9&2Cu!bpvlI68HSmHilRIU8gwzHyFkqbz4-&Qo_bLMWRv%J z@2VJUI>IbQ0`phG*Lu|kZ~i(9aY7q56+wG8jCU!zW}PV7RGk?pECFwRu2oIyuLH-Y zI6-rg!jj@_gy>kuDZ4<{NDWJ?P+s9x9Kh#$7E}8Ju{g*yIMj=6wPP<#E`h!M<{(V5 z?O#80DEFutI;sar&lkRGA>4Fl69Yi6&k5fc=*^@Ra~Cu>4M1xlph1%<>yx7JaBV!( z%mR>MF4VX9gNnz4HTPbFZ@HGtkw76W9^VJtA0ZRtcA}LaZR4SDYTi}?a8iRNER5Nzn^L z;n!3r#~*aH96uj$)8}3gQE=NnGIH=a8NJ6tsZV+?t&CcW_y;m0 z($rmLgE4N>b_?DVD=;1pk6rIfbt@U?2lS|H&2Gz`sh|7Qc=}Abz2Kl+&YqKmSaG5; zZ>|B3U1d2sXBzoe2?{-b*`-hqp`|g|8OavZyh)TOwPr8*0~8&zgRk zT)TNwXtF5ZIR+^pQtuF=>b3_|)}6gY$b;}`zTH;44L|ybmbTk#Usuo2jLI#^X|aYqJ{0UF6Zsdf<`GhPEP(qV^eM9Ifmbkf5T5O@ zTh1}ZG-=d$H63FRHb9(M)uD45+izv0>Rv>>eoI8aGdh(37qa8M9@<~X2g5u~cdJNb z5_%;N64^%^hab+t0PQi_3RF@BVOa>xcy&B0gkS+^BJc3!1{V#k!Pcg7_r4q|)0E)A$NPe3(MUJk@*T%EthOE#X3VD#|3qY(KJ zGQMm++_0H6r2Ay(a)^dw)Cxeu!e2`gHoCPVyUm+}{tP0#dzj#DBAt~ulJLBQY8{eQ zN>@8lHx6&jG+M3VHl6Wm{DxrJ=p`7!OdaTDCBk}dsKe~V=Jko!>Wno|$@#k?KM}{z zEtoJ&V6rPh8QH1gWy|-=ejwVjdLT8qTLb==+_yOe>gK2;{<*6+0?H|}>U^ivSk>25 z{P_Oe0~hI;O)4RDMe2t5z8D)qWP#rNOE=w(AA-+vc@NCLYU^Z$O3J(IvFe^0yy zSd}Xfg^>2rdH;?g4Ju1e2ts1>Sgv8^9=_`1!LxzTscV_F;Ew-85i#Vv)CpNfuN(>b zccen0SuawZ)-{&j5;~Zg8+8!3PW_7S^Ab*uf%>_fSuh3u71Spf3mGo_;udPEMc$Ig zUP*flxsEcU$49+0gAIwj#nwc%oKv1+0+*rGtY0gaF_|k!a31hnMqV+yH5F~Y%c%p< z&fHm9lTQc|Ed-trV;lR%IRom$*oV!-8!vUEg}1FmjK$}bR3OwFMvc#Ral=j|9Vt*R z;4$;vB24ra-*%M{hD%Wt6n>kI-aA~+-uQ{bnL#5MoW7VBb}*QLxCEgW&KEl;DWsmX zl-~r-z>xSzYBE-_V57n3>t9hn+dY6;P4LS@kz{*y6n)O|Q&G8#Rg3toE{?`#RefL` z39zO6(hj!qy6B!A@~CI%ZbVs_^8&AYSbt1pp zBw6QEbN$fych}z|nhoQf7rsFrj4sA&*|6Z4(H1Jap`s*WNu3SIYF1x)*D3P?WmIgm z6v=eH588QU0kmBEaK12oMaPhteWpN~iwD7Id@d*SWoZ@Zb9e?b4((f0<^5)lxJ_4y zv71a#J*Xut*Qts7)PvPZFl8@$OBqG!}#Na?2^^pMP~8xaB7rI2q7#%A)IKUH2Ul zUcXFY{hJ!rgzm`k_t_xAC45{*zd#&AiN2`H^pQE|>Fr}>>l6)0=wYQoxwupUiRd`n ztKADLdvtI}K7iUmu-fT?sVPHA;UGK+`dNSuA68pjyh z{*>EsGEKpfz7~_TTSYM^`czq{Pdki1xa7|D%J~k-7~b;;#X7#b!7G8R;@xmgvw1WM zanrR`?iAUe^*{S@jZ*_B_!&IDy_p`92=ov4eJt8== zX=2ZN>#oy*tCwpRh!pNi6ooh`e(e};Jf&3gm=_3+ZkP?i>0|hAGj28OCJuZyJJSxO zKy`;j+ApxzXXU&_g>t+z*cPv=HoOWoe=3JD>AkRK3QS*DVt=kED=0lgf>u zCArM#(sJLJ#yT#5SXgBbf8GFOzOHmiL^!HTZkJXbrKWYbrA~JtTco7@x;byx-Pt;6 zj6R{op^$pMw)Y2N+n(CTk3k_)#6yfkzu=0k0GbHkeOEtcduP9qB)!z?lUE()DDJNJ zozDkV9LFi)*$Hb*4QQ&YL4{`ja@Ay?=7$@%iRSlnkh{2FaGyw0(wd~7kcmU&1`53= zMT8A-oCYw4=e-v}cs(=1&kz}Xmq*W*z<_8A(*U55%e>9!An=?Ci8D<{6<5>``@nbw zhUK5%`M97F_8e-M@p%`49>lw&-(7Y$4i4)!%gQDd`hv%3L>D2~mzu);G__~r(h0e| z>o+wsr$-G$r7?`=-N`)ME?$TH5vhU7ChQ(7z%Mt}F(amL(Vk+nq^)reb&&&Cc`oED zKQI&TH%_{h>~$&bt#nM9H4fJ};vBgSv3!e?YZyFPxu|;&rtg?45s{nt($37vZp7NE zImf^eVEDG|Z#tZ_IUUnrX7B$vZ;Z^pz41xJg|{2RXK=2?$;syRrgYM52YDz5 z=%>oE8X0zrF(=&aGT@eqc;hGi%VJS#o311iRKB?2~3|A6Wv}g%&1+8 zUp`soRif>W&Gvq@))8keQnHLo_NRH4^fhnw{^?k&>0)d@ES6oN7$2Bv&%VNrY+H!ysG|MW1PIR6%_b0Ca{ysOw-GEy<=>S-A%eE&4S?z_GAO^l&@oQjft5GnN>NX=+G}Nd`0Zp;&V;7 zHkI0=VEOLvvrWW43mFw(WJj0Y%OzP@J|n48>zMq+d!#@j~0}=*fu+rX)oLl zCT)JcT-7+Z8!+v-eyuGmS84n@Z;a+~;qP0pMMe>}^5p_N(s+94q_l-Qhqq|m(mJmN zn3(y`Pc3@ChxeZniSgKKRI9LDCfG8I`j6Sd$jzv-{(8*tVf%cVxAbYA;82NT4=;t% zlo7-S%NQJAHJtSa@l8LFw|H# z+0?yER4oVAS9sgVT4cCtC=@8>RRP7`8J8RTlS%Xt!#FFI3{YiljyWbmIXBwTrOarD z$SwOhrh9t62^XbOb-t*vbHr?hD#!&i($S>zPx#1gKWi`jV}?^ zZneXu?*cUjsc>vwxGYU<_cg{TRbp3+PwuPZKjqvdQnRG+d3jTI&Ao)!M-}Xtr?`Mm zbjgmw%va>ME8a!a7#krrOfpD%BHLGtE6!9PC^Isk0t;=0d8Rt7FRp2tC!LZB6I)Mk0pGihz8MN>>4`I=;gpfGrArfMWHu zRDn&;acsF7SE;~)iibeJH4)gga^q?{yq2-+i@YQPdR)+j0f*!^Jh8N4k7PP= zu#tS)AuZYzCH&h`A{;HTL2?KC#UM4{ltQ!OJYtbG5_iiw#Nx7G_WC!!C1jnE<*yPd z@8_YN*OJv_KAti=BPk3}6#XVmkDP6OxyDCfcR8|K0_uLDax_}Z1>d^Bi73)n-i!Qc zxyF|?l_E@~nwbpB+Ff(=sxZ zFd}qEBcdv+WQjqxQL6{T567K1H{XGKFuHBD6eky=#8wu`QeweW2g0X~ePiF%dJ-E4 zSBdU|g7Cl^*z=mUUPJqmir0OXIOw{n^!SU7G zx;#z>DK#(Ob#4f%B;M)TE&*12%sx12(>wDKq9FhO@g=f;3sgaFHlqFUqhGl-pH1Q| zrn=CwbB?HpXa$~;^BEz^bWONT*6`ZF@ zLQO7J>6AJicr@&Ubaf$iD~vv~QONYCU(F%BQ#z>d>ITsG5UJ$#JshmgV}D|>y0mu7 zo?cD_by4A^wlbPZQvKoegwN>!Y3i&4Y-h7m_>}svRcoq6HDA(~G_r*))H3R>YC~o$ zRkwyLTQe`NXXZhsW$$l;VcYL~-2~Vt-dI*E_hDHW&~}P6_77-->_5%BD}t0;L;`6uol;oKq=M-o&!{d;d| zdTy=gGnswQ;s6X+baU{;AO*68Mr!}sBHG;I)MU|2wr|=={>{qe7YZbDBOZZzXj#6XQXJ-9?;fAx zG=Fe48#UL|0~dO8>~F(wE2$)bkJ(AwU9#ahj-#sArbd=Ey3EY1r&7LZvo^!XQOz{# z0iV)wwbWF86Z~vo*pN+k0p8RLaK^>RT*SpJ;77=_%+xHa2Ge#&sWU|vHzL}Y#zlk?#67>S276-NJyl3f16)cV=)Z;O!ChF)X^nq2P z5r*jrNdo6*Z{14kyu|(TWa<#0pkRCQa5j^kMzD(wc2={Hq8!0aqLJ427E^TN%=X|@ zusq5sZJ8)8xFlTowdmqTXI{o*j2nxO9Y)yG>U~>JIoc=ko6gVtg7sPdx5H~n5A;>M~N3&WMzmEk`ANsmc>f}$Pk0HRN7 zLv^-qv{;cg?~Ij7qz@O@-+aF`yqJy}QJlc);+nDa;?I{N=kGiddolUg4%)`ST>^$1 zNJ*fRQPTFaS3Cj(d-PM2ra}1+used*(0BX&{io_XUiw;eqrmbR!zg;%uSo`J&kspv z9!=V2p6p^SmS8TBv#vT+^@+>4LT*~lHJM`Si8rU0@AjwpmTCjT%^qB7V;QXzew^$k zg_oV^NTNg{uN)+C&V+jV6P6mudLq*tcF!0h4q8)aKyI28(N;cZIb`Ae((v{~WZn0ns@e z)e%iJuBXRgtHw?`V^a88`6-u)HmsC+T44BA{V53YqT|Vspoa68Dq2uaMTnjL#eHD? zLwPeqX9koJ5vKq*)`pHMNi^lpo2f^fN2`7-CRonOwlwb9X%JjwG8v}#hu>7{2Lv=N zJoD;p*|?2~_N~ioRD>ARkh~U&2byhnf%5NfP+G19-5W#{M{H>W@3!G6%kNCG;IJM& zM>LL6nBcuYEZ!|J5~I??3;v?J&zG$w{)h35`^hC!DQjfPg-mAN7v>wFH`;PC=Zorl zW&;DrDXg{S^$qk4qCUW>?2^ke7AEP7adzC;j0O`T-Vt;n#kjK6>cGMJPTS>YhvY9p z9ocL=1#$)CrZ3I}xL8g7MmgWQ9_s_?hk~_^La_7G_gs@Y@@vb4rE8z2_9IhG?XGy| z*9ry}Ra6ioh>VLr-+Df2jra$|&?~rd3~gNor(w3RCbVO9z5i~=gSnHtxSEa>fK=@H+-F9F2Vt|>4HZh!4za#33-L)3VtYwn6e&;=G zXErnHR4Y*r*K3pH1zspMa(|ZnbFq`di4nR-XrV9 z*4;z<_;6hof`09^W6dQA>k5C5MvoCCqgfHN&sddf?}uUmNkJ9-*g&FjY&3G_<&=2W=5hByu3ZVSBOr@*-db^)ue0b_{P zIYxXI?uO_{Gvx~JW0}<@iQqt2XDgH3*N1&!e@^k&Gwisr&^uT^RbEMCOC=jee14&u zw%>)t#CO;BU^OPZEYb`1Z-r2#Pad5HbtW6rDEDvva&Q$g7OEuWS1R21I&@3bvYK~9 zqNefCDJE4cq_Re7$Uyo$qN7zB?+H!Pw4HwZgEL|epnSScMAkD0~q z{ufv&KVkN7NKzck0uR67781dPCvO@F_3guV{RI0ZGK}5~_u(xF!CGssNCQgn{dcU&o%z#GvUYbG-rm;$6|~@a8$c|8ZH&)?~$m z>`@~+?ofUE7??;VS+PC0Yc=Ix1gg_db+rYd)1oU2wpIdx6ZI)rLxZmx5RA6x6UR~f$q?h1NRM=Th% z(pRvN$~zgIBZ`|hC8A*r8tjOv)@)vR!^e>@`DT&jwvE=6yq@=s#~Of-FQ7yU$R9eQ zFjbsd(6YPz#sHgi)#h|{$rz0Vvi-$<-WyZoe8MAd!_(NF;vAh8qB}UT;KXr#rzuN( zaalq3I;84Vwo#NawxqXE7@0(}1gx2AqN70J(1Ix$d1tyr45sZ&1hY+oo3b9TLF;vf z=|%w^bMqoBYWGui{A`iF*rQRzQRnxp$uu{=+aRX#&e7y&g)>@CxFWi4XktX*=C5kD z+QaIYz`;UbX4hBp+7L(tdWqT}pam9_pM1YcfWzN|U~}f7o1lg2f2r7kKyLDj8H=rOmLdUgr!-MZf z_hek41DYSC_L|W5E3CR|nGSZ#KMu3_LbwEL00eyhAx+9UgeoQt?imEeh8Yq|nN4e( zT$B~p8ToKFl~UU)b=d4vA)kACH6sp71dU%lzj3XY&@Z?I6`ag5T=mCt>))kTlyfaGB0nNNKK$1M|!8_jh(P9-d_tPGYUd-m|`;ercfD1C1z%MwADWtSt zrEp zEry^ab)|L=o_KiT;Zj0s0(Y!lV6z>se&EIV+P;`TY3F@vLgc}$I8&kWD%Y=Z67~DP zs-E+>-p^xSS-y1E^`#UgoV=jalcSy;rNgv{6mf~^G~w~OpasJkSR&)tw+3cNTvl&! z1&vH0-m*ufKNhfV%>8*j=`Kik;G!L`Q6)Q42A68el2x#Kbc z^NhMjKi!_)0FvC7D0Eb2J0Avvsz+toY}tP|sM2K*Ekd!Sr&5WHQ-r8zLd}v+Cl}(Pdl)f1INCHo4NT9bW2TeNOP;y7>LSMVSNWvsfjD#R{_Vjac zO^O^np2tCD${DOr!-tc6i>oUmJ9eic3?_eq;5Bi7DbtbFT17skpSvaf6Qmv!q=~HR zygZbxzjT;t5}aC9c!RKW-DqNJnLCk4D+-50k)}`0Eco&lxcVtrAkFWB0-d7IdRbz)BPc!^Fzm_Q>(jeRASHZHJ_wHCkM$Y<4gT{i4 zoFncvk;&h0vKV#g3xjsHs8!S29lF#FPr_yT+>c-i?K-rf9|ANP;2(nm`!0#tF`F6z z?*?nIK4cq2!5n!oZQFG_urKiWdfRl{ee!dsyFG0?wXC4x^u-4Sou1-5yi3b(sKvPo z9;Qg`o%jrQ{lTq!f(unQb>6X39)}5bN{$-Z<(FgcFe(pK#4u|6D7X2T?hX*lh0$xe z(x|mk{e&|pF02I$n8Gy0VMpT+U(3yGrYquR2M!({a{AAJ8KYtn`+ITN^xlps6Y5L;b#0vgU)$!iTnT_c!PvHK+qP|cV%xT@ z8{4*ROl;f9#M*t?xBZDzeX6=~-WQxft^-T$*iYBJ?fa%?2T%~8)fYrRBv*_bh~;%J zC#M_gwj0n6U=aDy4X%pm#6f+rp^6rf*Cpy{M ze}^@fi|fjkkc`p0WtL@4)#c_Ve1`v`gwGd1Fgk!=K-#$hIcUT;yP#C0 zR$X2Y;@WsT^g*bdnIme;FLf7bh5t*Kt2_EK{~cR~qaF6TQ8*l@qj+x|1-!c=%$5L~ zjFecqJzrwj4E&dq>DKLGPRs?v7Hp4AF`p;#x4$^%Gq7P~EV$mC6a+{Y3~1WWwlp!=8k$VIPUhssYJnJ`5d1?8N1_!-a8-*a`3;@W9DdZdnw?-H$Vh zoCTs+`$kslAz~r)ag;E1@ih@sJZ`u>v}u(FY5PA->xYCv_zsoo zY79N+Uy7}bpkA6?R5OSG=xpT>R-*QOcecKiil^KDmIlsdAWB?7DYN3I;t|9e`cE z)yX|94x)9OXSj|evU2Ec5lLzCckYI4gx&lhktW(JRHInQT4JQQhan(|ToCe=+s!uA zrlB?wP7$sP5o0glwojemOWml=bo3ye5DxFd8f<=e_Q?y>7~umfai0+N#KKMhPYL=C z_sIv+3DH%S63sbRxmGU}Dz8aOMN7KNTJSY-Uk$lAj8JMN6$3(PX59p{DAHQ7SPS^k z%lr4>A&Vgd@vN?yxRx3c$<J)tf1A;$miHKHr;tu)cIHS*0Oy075yP3FlM zr!X-GEnd639VDMJo|Bmus@IBcjs+wGwv=SXI-u0x)cJqio(jOyoTPx7A$s+P*l+s3 zZv`ZFv=qToY*dFQ@HWoua%|^4Nf4=ybwV3{*+;_$W3b0?m8S%3dCnLu>@0By_?c9fc;6TGKr?fzGEc>6ovbZNMUJ~OPi4E7_2Y4hz)-T^{7#QcL6 zWjmNh-)w>0T*<%1`+@VOTE+7;g*to4hVpIvaVD>`@wFABp}O5ipYf}>&GL1v(wA9= zIS>0KGDm_BmUYqa%3jg0V;^@}{C8Xgdk|9WM>Go4Wvs&mQf~aXBOhk=t6>Qz@rr(8 zy$$X!RrIAsll%SF`hZsRl9g;+`qG^OeuD>7=@?G(d5S6dJPi_KH5$Wt%q4^Y+`;95 zZAPw{M|BYb&Y|gvQ}FH|*~X-hXaq@J?fxj;tplc~ZzjVlt}zIGJ5>Y0&_NfTtB2W! zE);y;khd?dn&bS#zm6FVN!xH~0sJ!Hn|r&|9YUN!$o>w4TF&egwZ6%_yX=TZ=RFyF z^K$uns@A=~CDa{66*RZmE^FJbj@^btm(8Brz=k|y%J4HImOlO?ZRp&RWm<-K9}lOm zSFbQWp9`5;3iHGcKhtt{bFo0v1lJ{%9)^Vr4)_JDE!gS=tnp<+^roFdf-*XkWaEh@ z@&gHp=N~k-sRN^Szg%a2Q?oon%Tp<+k#_>jr&)&ycGQ3;>g$!|_}c(z9)mXgkb3v~ zD;mOmsRK+Rt5ck_XYC7}ORGick0{S9Q0x)aB)kC7pJj7yI|58*ShPPIwCHBjhnXMc ztLUJ@G-Jhux1W+>UhnVZdGgytncw>9e=$LcA~k{(ab{!!d^FYi0_K75^53Ioim-2V zEh*9q7>aOm9mWUe-2<90IJeT~Q-nq#J#V%QV2ESAOw=;#cS=-%A6`3O*=m$!c!>}e z(b2HLF~$Bn{j;WrqSmjzABYHEhcyrvcaPs1l)09kq$@m$PHBNW`}8>03FP|m)FkdW zvOIhz39M4nOlvlq74i%ze5sQF)H6kb@P3cbPVID&h&D)=_W1=z-m#e2P_qhPcg#7* z1#@Z+u&blTV$tRMK%lYLyK;k28&iVJk{-Tpe&yT2wagwG5wnW+kdx9Vxj)FBKIigEJt$36`mh zXFH7y^rF8~6!H?=hd>Bfb*B7G``|znW#|!bW1nf;NAQ1WmqyyI63eR20;ka%Eq!x1 z1Z6$C9h9wTGrkC5D#}(!UUT1qdqaDQ6Y$;F^AlNT(!efX%qcQ+fCwwSo)%RzMkk%E zKjVE`Fq|{DHY2S3!0cukC@nVn)^)jmrs{x(^BoE$5ca!u8}a`gtiJin+r&d&BZJ^w zEv5uq2Oj={Ju|QbP^X{!5e%?j!7B_t6o=9gAV6VKQhmZCCAx3L3KAbBiXqRevy^c7 zR}MsoS7TuvW3PCA^D0WAvQhyHu}?{?mTkg5R_vguav7ix{j%W>qLkb4jeZ zCobFcCitY4cg@{D1H8jmv44PsFqesDKSgJe1l9^Wht9($5+WI>@8jp5k8o_oiWy24 z=TOkFiKvUJWLzLD2fDf^(V#s&4?BTyCGOIKNHW)xxoEG zDqR7F)EevSbTfv5;J3SlXafswl7ytFC5#7CPe$VDwoDd#l=@dwlOxA?vvid2L}+@T z6Z~(vM!mn4$)oN9oYJqpqX%7dh9JJ9w}l<6hi>T;_M4AD3zCRWHR|1Hh=PWKANgr; z*mZEgj(8T^N6yDBHE!RnnYf)N4L4Ba!`Vfg?Y?&OatFmchtttj*F!!65Ffp}nN}

p8qziW*@X3Ol*V2((Go#^II0_=CXoo{}j6*qQt`eER_-=RQ3xckjVIqM$ljMJ{T*=m@#`aukgz1~#aQTp+)+9}JB>1h z0QI!16@EGW%b>Q=w{`kY5Oi0PLAlW9klWNTG ze2uyv26VN+ej^GPY?DUQpu_-UlYwx1(?PfTj#LzpYD*Q-|Agd*3Tbp2n}+stU2s=t zUZA1>ge1m&07y$zfkmn9E^#D)Z7uynhraG$s59koH=-(dQGNZ-{i~i#&zGquD2D&^ z1^arggZHe&N_pX~l?c5#h!DJ?=d|chHuGx}jFHRT##2)8HIMh3Q2@5$mQa1~&9V?5 z7d^Ebs?ZBR+K4_}^jvUGJlihIb`a4ilQi|QRGQFo51&!Vjx+~oQf>+T#izK+ddDy!A3;P;Rg|C*jaJm1xl9WD!pF3>L zM`8?$e9EoY1sP;#>d{rir0$<-!0N1EsGg74p1svy+fVS2@i9zjA^~5xPUM<+BvWfxlSYi{8_nba7z#LIBpja1TBTcS`E7H|vT z3K)y>`%{R;o2UhOF0UNA=OK}qHS27qJ|k>zP~L{!mejmF^7}!jM@B;n_NNm1+XH-2 zRJP<0@6)vVSd>!xH$skGK1&SE^y`3T+KBSDe#NT=W|ntIJxW3%HOP?6>KXg>4X<~F1ga( zBt2UMi<2=dd%dQUj2}v7c#v7bEE6t31KkP?z^dAq5v8`Paod=yI-$x-V@YnhmVK+G z?h#y!_(J)V}A*C^jl9j+L-ZWG}5?k2?Q9FcL6x8|RI$&az*5vvh z)7Ze{$}@XiiX}ssF7A%)g}RAzI51^<3d%8K1EIbs>~7aB+2iHJuw2L9cKpaCe6zF^ zLi}fh5U5vo=+8nIBkk&hWRi5pE54VMm`9#~TXAc|^;O`!q<4VHU%~;4$m$dhAv!~2 z7XJ^OVpb`jOKa~@@DjF6R#S)8p(j(G_dZ^o?ZodbjIphVDl?u2}u$w}pm_3#d zs;$=HUyD3D5I6J{0$44f1E*ZoUEOVi8B=8hW8a_|<2QwH zY)eei{+WXTu#y{J9TUzl&o|Hepg8mbeOTdk?J%@TS4IW?_v z-CZZCd#pC2#7ZR~Z`8eQRkPhvh_MDuYzbCe%*9~HVIOh9%rKsm1kf^a%g$#J2gF^# zfcsfjG5iWU2w`(zredSa*~{f z+qgU$;04@C+CJ6#`h{un^l8H>3CO}cnAN>(Rut1#H#}kY+$A96Ejo#-QN{-t-BbO% z)?p>{D=55^IOSZXRm7((GO^$!F8}+VI1E9sC-Q*=IsxCwKVP$WorIq>f}c~G-u+;7 z(kVjAcOIWP)8U`(8`~@k21nhw7i9_rf+?O*-5NgN4re6IXt>VuO$}o|8G~gq1>1%A zlgRRRkec9uW}Gz>7>Q8C`)Yzb8*Tui$xvwk7bMYP)N$-B1nJjS!gf-c)M>m5u+x2O zl6eZ266L%(U6+Woe|URZKxTSgJp4pU;Bk(o6R=$paQ30*oNn}qbbr==>7wPMK^%M_ zdS3|CoSdch(B|ob_!7zAcSs^Rd%H%D<^hP0e>WVk9e7`8C>urTRjdqRPo7Fg<=ZU> zwI6876bv{Xaof`ly#(xC%MF$#C~>Cx}@N)Aj~iYNrd*K12|5K2wu-=wW^ zSm81N_~Fgo1)r>vV}&qsdI)zqws+=|tgq+zY()@NcW_JONpoe+a1$k_R^`+OV%Vpp zHZ73%ZvaS)quLS4r=*HeZ#5%T0#)+y-4utg4e%r}3N1N4jA|LUH}Xjd3*9K2XEuzzQ*fqH*RB1; zw*4gOxMSP4ZQC|FNyqHiwr$&X$F{w@zFqH+s<&$I`)u8Fovo8K#+++hjK(Z3!n#Jj zJtTp}t5vIH%jIz;0f(k}FzcgrzVXrm)-GBWJz<|gBJeT-DbcZVM;_>w)~)e zLg+y$I>z65zXgnRbncK+Yr^+AF=fJW#zj7j;$Ktm6!5E>O0MX#W`cyJjCw)LF+wix zATJnxLSkv}k8I6Dz-r!Pl#I_V=!(2Ei;Pq{Zn26vT5BJZBZq^UUOkCas;$T3WVLXf}H%m{sMZvGf&Ap|&H1=ZCkMOJt%OH9C zDqr{gro3vpBMM=M@zdLrc4A0`5A}Yp$kf#P@qC8!)o`;MB1>6{_#;PD{|?{0j|ESc z#b4j!JblB|go=%u3wU@^ohALQX)m*=d!5XtXCW%;Vj4*oGf*E3eVYfIfs9E9hNl;f z+4l$;Q-@tqOWvtUXJ4_KAGa^*!475suA%)v4Mma!wRJEXg7r(Vd4PsywM&5{&3tefh9_l{p}J zE*ZArAfq&xm-C6iX4~0mh%T8i-yfFI1)gm()j&SfbRGM_jo}KkY&M7lEq1rfKdOs=hkVFLa zjNwVWoL9NDgfvuJX@qD7RrlXi&t)*g0+yCHQ<8G6Bj6|jH<*&d`R*k z#B(30l!%bSBmR&28FE53;SyUKXaVQdDed#~UCwc7l&eUC?}e6Q%kPEW&@_p?Y^(vS z%hN%&b)M$yv!IkYIgNPC2*R z9|AG7=yUi*u?l3<(_U+W&Ld^3Yp2*Sr;GblosZNgzbOXG3Bly5gdBQqfP zcrn`1ak>vM1;k?gSJcsFmLQaqts+uCS~9K}r~rp%B~MX2svO+-JRP8cofs%o{%c0w zd0%s)Y5D1D^5@OXUoMzcMy5uZ%=LX~G4-p}s>ZwJILLW$*y0!=t1E^y@;o~db#j-W z^ethGI}*||IqEXV!brLaEeKDB5Xq#fLthX_GEq22s#)qWZ*BS=y(RgFI`-90R?==f ztEW#jE>p(smCZShLR528%5FBv1Q@H|3(_T&Y21Ek)hkR7Ew1pJT|a|;%3*s6sC%w= zS*2i-D5B&%B8C58a{W~r|7#Y1d34#xnKPM@)p>}is2p}sD_k&eAHnGDbigc(p%hKW#-cCj+#*k_i-j^cm=);l33#eT47F zZE*E*KNl+h+lt@Q4-O^W4zN{>9*SUY6Qk~19D-^kOXScE2leJeE|_%lwVeHAVT^BU z6ZL`>9444a_^DsG!Z0S)@F+2ogwkBQ!esX7CU+aIC)d-+a>|k&oe`k(gHbgaD=!I| zFrRE$H+ZH3ua904i#uOZXgctq%Q9rbhKeF2K1^YpqZV60$z~xJ?Tl85?f8SbuME$GDLs*v3cT0>D`L#oz|ioQ*G~tm5^T0 z`d%H_pI76oq&Ue5Wu?zPf+c*=&f&1R}@+vOfq<%+H7hO``o2#o&{x`PrLxdMw9 z1F%UyVs`iD9XGaNa3V*J&DdP{zMXn}L^${H`6GCKu|#cy&_^F&?TrSUO2#jP3y4DR zxVYeYambauBEo`Q-|p+x<#xJgmmL@0N`q2i)M6?s!11L5i!rXx$Pdu^S~1?I9##a( zO<+U}1rjg3SBiZ=JtOxA$dVni683!;pPouT%Q-aUy9(4WH3`%umW4dm9JP{(cy3xl z$2y*;XzVI(osx|z7Q*C8()!#%Y;9|OTDr+m*By7Y?l0W3xrv_?W(s5d`9@Dt%2+CO zx+=?S6@WPosqnyuCKMNhje=dk7IN0kWb8>_{R$m z+2)fR)RWVmR=sE>!IP$^{tB_wSeA)<00_1YYC{~UlC9ao?TJCe3iMrUY2I*T8W+MJ zYtm}i_5}H_+}lVa?KOK{8q1ZRJgM%>JCUE2f)+u?uC~JUc;{UYCNk>$#aq!g zl$2h{f!K-R-!-K{ZE)DYcM(>p`yG?&W96xC3d0KVM`S-6>};K{?A}z19_ZY-5?j*h z%0N2*4^<(I93CpEbZ-QSlhM+MU|8oO#^SmIC8PDE)W2}j>1V;swq^?IpK%V>u;IKR6xZr_Rc11z=4g31?dlKGl*^@2!ncvrK+<|?G z%ThC#_SQi>>(4%gB1U_-`W4QvxMXOuI#qF@dOuEsoNEZX5YqYbfn2Cl)O%}#{<+6U zSfW7{Q^^`{{{R*%1wop~-?!=CN2Al{FD|f5f|}DbLhzVJm)pAF0iyy={{AlO%t+^= z<9LQ^m-Xib`w6{0-%pF{05U>H3WHdr4ogyn1Y}E@9#VEIT5Cx878N^I7(Vl22AT+52XVpT zDMP+lC-r^0lMN-f{F!g zZ<}UN5U;>5t>o&<8jj-CvpX- z_lbxW#_wt`TzWM`I#yk4k0>aYt-bmR{?=$hj%I<_7Q{0RQ2`}X3o!G!?!C?s5E}IW zZ-mjbEYMg#M6aJqMKhr?TP{Plz%r43FBkOs@s))Dm0QpmV^OHB3;<1Y0- zjiigufsRm7NKFP4=*N&&tk_uo4Lg!G)9GLqBNDE)AZ{w=to2N`=)M>i6_07{i%K%N znY2QoX%M7qw+i1ro7`$;787DTj$CyD#4Ux8Pu7xV7lj z8vZ@=88d4V_SA^{)?9x;qtYwI1v@aUmwkt|3=0MQC1IqbiRZsp37&5k@|D>9n6-(Stb7 z^hx7pK7&!o(pEvHnf%|UTtIi8wU$_q`O=Hw@(eDdbEj5{prcnh-QGcR4FyEIWuSxW3FPD!)jZKmd~Dy zm~+^c0p+#X8I9?2;$wE9^Yd{E2+5!bG!*i{EUc_RVfyQu{ zr)78aaAQTju#N=>dZnM3n~tb+f3NxweCYvzv}}qWYiul`H=+=d$rY}fktxYAMwaf| z{zuVC{zTd`Xueq5dkqA=VIRQEc)c~X45L|APqs(1568IvC7)Wp@GudKhlf(S#(KN; zva<)hYOxC`(1v(+P#(1}+voZ4W>?v6(MfeSGJ}+y+HE%L?tSbz)u6IKI;I1((-qW2 zBCOw>q=30|?Nx4x+K(uqRKWJsHrU~#szdGc_7NAA+h^mpMC&G~N~&Euzd}<$^pqvS zlAvYq)GxOMMu@`1v8eaKLj~wm;2sCW9)AOVvN$aNH!2Rtf1%G%cBcP_ zk;4pR_)h`I!okV$KTI4TEWMbejf<%hkY3Ej(8W~5)Y#s{6qb(<*4f3$)X)~zW4%e$ zlSk$KUA)~=Nl^&RmBc!2s{6-A9IN%RMJhAAaf}&bFtp-RxCFXpA|)9KS&?YI%^H>T z$$Y+z=D&OMrJH@t_2>O_`brw}%)2K4^|2=(Y178Mm|zNYTW~0Ha$bO_Aq#-n*%@?T zB2_Rqj<*&Y!g@AQl6V&I6E>_5SXlUA2@DhrErkS!ECn%8#nuZzI2i%>ZVBQ7GD=Ac z0i zAW#a_5p5)hRAe3>Am&yWJaVZ>FSuz8=uaRfd;*AOIBaXM9ylyafJ6Z831D^YZvN0WBl(=>fZ70liGoB-nMVeAIfAN;Qi6a%SufC|^BBU6YU zrLQ5iAqBSihc31ZcraSx-2i@W@ZX+!50sONZnHms%mU~GWlRH9K$pPSOpviY%kZ z18K@U$)!lIZ~dwZDEM#iqYU!!;g|tmu^`WRLO-o${VgR;sN)AuLE}%unhO$Q83m1l zdwdL_Z{hC2aR88jSPE(yZ6ekCUHd&XO?a?KYYj5z1CzAlV$x2W zhJZ>t29H!j^y9nNKV-hhOG!Wgq#~rBviz@~#RbQPSl|e5clv1NM0A1rt1Og!Wf&4h z;Q|QRWI1d|+G~|pD?#QuNMrYJjO|UQYk&BKa>nd45R@?$c|b`eD6$2M5C9ZR9$HQl z2fDvMSRQ<>f)qSgQ6(It1K=kk#vl-60WFuIstRbtfznD|bO5Zs*-J{o2zd7n$iLtP z%11mMY;N8heMK7rSbjNY=Kv7-nAc+kjN@R?Nzz2#Fk*v)S(rpVOe2Fy-{Sq#Nuyczrz|?Ln3JTwnC}I6oD<078i3=LRT;b_6IYfvF@j_I=+yyg0&C z)8THu=P(^Zf(pbTcmdn>g$dJ9wk`({4+$XsaR9;F+!J8mq;Ne%^LM1HPK+TyMIcSt{n#$!wjhCTN>b*6p% zOPegOAB|zLWJcH(mT=APqTVsmqnePl+gka5?+74lwM4xuJqKR%SbCTGR{EWhu>FN= z=>et`wmuvk6B=XsPR+cCT@h2;R89K|ivKREDXoKK#e=`}4!r{T3Pj`WL%a)Ciu?Vz zkA_8N)HFo+Oh%9N13WyX>z(|q_^dcPT|MAqI~3vK3Sw`aSDrwx>L=ksYi)etmcLy&uQcm2O&q&YXyWIDT^e7$SKzx;L`LFzg$~sBD?ODJ zjvcq9{t#P_=-qlF)Z4?Z2GwK5q$xSFc^f_65=m=>x!_Y=Se}?`M6xs*>z2{(7W2+9 zIi#O0iw9h%R>$t`X9zpwYbFdK)ipWGsPwvWN=aoFAIJwsS2y+4YHVVJtJ`J_c_;s( zb6i35WGzQ@Aj0mM7xP%@h*3(mGIdmunAA6)aETBLheb)M+RKqV{ePE#@e$1^i@iAo zecO6-2jJ)5y@|rhT2{2GJBT7}Bv)@rv2wk5=oRTb2I~Dfy?G1(IbwxN*S!xumF=d| zslu6Qx1P<}8Kk(eQR{rpre_~!tt(T#ME|Ao?ifaH_*Sc_yb`JoMIaGH-BqZUz9KD7 zZ`D#Tx#xps7yGGKgM1aB%S9kDhM=j<=jNFsF)0_B0oiRRJ6>l=v$!Aj)N(?m`8(o&TNX%V9lk;!+O6vt{`s{dfOMqc-rXdDW9fic$*b$7zsYY9Jm+Kxu)VK;gGLD zLg9A)6iS**zMsYMaXwU|hh{yaCUkw?;UmlXkreCimUiWbEorf|o7^g_MVZ~yQ8C-p z*G%t?)zuvrtmOCBAm`GdW7+QQu$DCHP#v_!_-N@(ZFAcs6?xRA_^fRzsgL&>XKshP2lAXKOISPE?V=7_ zKW_z1YIE!ynZuhGe(9N?)%*8$Rl8hH6quaa%RS5Pgr*P?&^L}R7&qMtY&i4r&pTor zG{-c+h`w&FOJ?$thV*b@#ewUsz1hD9K_; z*6afl!*h#7l4<)zGgysDWQ8%=%#n_6q8Fk3;*#64m_AFtZUPltHm}BYKTatkSHs23 zVj^EF>lW=&-UO$hctI;cjzHQ@Wjn>WqTHd=|3hHA0DtwK40EsT8=x$@e;Zp#kPHq1~k%&ys2{zojpAu@K8J`UsaM~ zBtoXMW@HnsO&chS^al|hg|(Xgy4v>h{5x8fkrEu4rWc;Eh6a<6p({V}XY?c^N^)7+LXzemU}ZEDd9 z;N3Q~861CjGtE{x=Js~a%gDCbAn>_8Jyc)pKoF8*^HK(JH8JNB^;0 zRi;LtJ-2GiPIrpX}8@%*zRV8a?#2sW6ZK0)q2Fl6-Atf;a9Udl&i1P z(An}P9t>RCDQ(}Y++aA5KfkU|)$!`7p2?~LL7SE2U!zvH3ch=>)kp;MRl(+d<&r;| z=BiP{{G0vu^$G4aL=q1z9CTx|DVKWNGSE@r56m{SBvH z_B1i>vqg5t>AbM5Z<$8fImlAwFt!g%Mq@ji*2F~E74D=mU!oOLAm;DIXLjAN!X2ZZ zZ+w_~p5oH%CT^DIdgCjx6+Mac#K|UZ5$CQIBIPkBlv`_juXV3iWkF%@p)wG3|1kA7 z&AZIk^ph(`ZAeQpJvJE!{I4imoyFAC(ZQ9dkQE}v0|jdq1eswPlL zajoSq9kpZT!-VV$&{NORt%wb=j&gGK+@()(7b<_~nfYG^9xaaStSgt6eu?#XhmPQ; zOz3+Zw@KlZPkzPoy|H~c^w$@knA~^U?)h#}c?bQ3Qn5hQ==Bw#M&ul;+}iUSz5%=P zbK-1wSzv46*_^Yc8=BJ@V+^k?7LLcLUYXm=m1yc0aUS{ylPEQRKj$z~vKg;4psig7Fr%@qpXgK^OTuJr#B;P_ zI!6bQnqu}J%%XyYeG{v$nWx}+5$H@*DUo+zq6?kRSq@f8>=$Td7P+jIZ1*$58^L}&d_R+<~~Eei8Q+o=0>h;g+vH#Em9KrcSy7IEFMu3kD}HqFJ}s#5VQPH z$MFig1-f@n|M`m*5l*FMWzWCQ-o3Dpo0Y7#kg30i)AGp zBJMug5fIA8B~_Gfb$D+hDjr1`$wS$ld23|vf7bgnlqd?JzkP`0(axVi|0c7@ULgNX zE;h*h;Qvu8f|aa?*RK(d?8CB)r2$e!yv=24@W$|N6TR+%&}7G?gXuK>q4F=BuiLTv zU(gBZ*?MGFUjm7V_&(2md_=ZY7pc}&*zoHD1#_zs5Bt-PBbV(17$*{Lb61LW$tJM5 z0nSd_uOEjn2l%87~ozFo{c#zI#g`6D?qhpS^^%s82QgsJul2?gL3(-$WsQj z!TE0(C5dTqnK!NX#dBeCEHU#tWH{-GL`2Qj4s|65>G!deoDL$XC!AQVGkN50VU0L? z{H$v^^-YJFW^WVa!wj`O2Ot;apeZCqYcks;KHJj)Dg&CJn*?1uEU={o;aZD6rVQ$q zVq+MF3JI;N!;5Aep7ryUS7}FZ0Po%XGA=1O6;6TYMn>O7BLn?LRH9Cbq)3K<{6BUd%$OlV@lXMDB4{X zP6&8KwgRJ>){ir#uXJgG$ZE*Ws~kd0wmNLZ9N}X%uNGli^Rm-@MK{W^_kIHDrLP@0 zDPA+Yg;*ltW=-%|9BvvFqT63)w>(N&?JkfB{+!NS>|J8{%7SY{F1L8OQ^4OMrO$ZN zPs#HmXyo?iz$uB?xW)LCP1_Z_PRWinou1gVo=%%jU5m7Ci#A`mN)2UC*tCM|9*awp z(JOmu4x?`$g9kv%sZikQ<(*oKMScZYN-xzj?b1f|FHQBvS0g!oeXjhOu72@v=h@vc z1ehbvQ=B242F=8v$zizWUFioK6-$wL)lvBKsm={Bd1WkSwYlZQV*Tp)B`Z}vHp2MY z-kiF5oYjlfMj{(RyA>1}(24W#9O8xv{riC@o2P{BCL+qS#!kN$>tBJ?-qjQ|j_Pj~ z_Q!n7YE45dLOzzWThSv6#)mPJJeqi9!x%To=^khIX#^Tj&ARkE%^&QpoGUUbctq8Q z(o!nXk0PgV@9dctr4036T4Y9nQC+SE|IRx)7x59_Kj73_y#uko@HjZ_OWqv5KEDRv z7uQ%N%92t2<;T8t2w+;*x8GmzWH(3RszDsxIT94F@J@DbX(_=z$2!3;RW7w5@>M0% zi~8bWP)a$`MqKb9N08Glq<_r3o6l>3tl+VI`Td;!=2Sd+lFm0`Ntq-n zMxR7XEX|@=e?)JETu76dd8Qg3d9vI`{Tp@5k{H87sY-BYL`^PIQL6u=yBkMVWCK(3 z!YlmA^_cxU^UO@dTz&3MyYt@Ovr<(*HU}YH*oQs~8;blBq{I|_b!{UNrJYvDjHwh- z!~9>{=tlzxsGPPmTtU9kG~^gGSm@7DXbW)U2NbTqppeACNYHu^mOo>ZQ87nc z0T{3g!H^4FOmI{H>s|gXv@`P0Uz37{$am%fxm@W+h)}=kaK2nZ^o5}!57qa~Gj;)} zX}Hi~4xl&yoP3ZUnY@S&0m2FtU<$#+0ig3^GW>uiFdaep-l?JaHHccb?aQN+{Fr}m zGKb(GgC};nBf4t4Wo;0+{NM^QK)XFe<(7`l>rOQ=%C;V;cA^Rr29S_LQ+kV?d%5E7xH2HhV!A>&UzjX*GB zy|&Hc-~++Upg>RP@l1jugLj2kTr!~vbf|lK@OxtfA&3!2-APDTPO{64sI9fNP|koN&tjZH6qg4!AQeZGL879f8? z0Pscgw;#ii?YkW?KI~j^qv1df=o?aIPJL4jXUA##nSM5^!s@P_^er>Qm06m{rj--X~ZZcoK1i#Z}Z_PAJWO; zar|iDwcxs$e`30yQ1^y+Sjed@A2f|_h}h6q-Y|rhq!2;`5TR!0bVCGO=z8MB>4W77 zdbPPTJW+XBzI7!d>;52P6f*+-1+?XffDLO03{jZWNVdGZz@PxDY{*c$GpOEuAhw`# z0FVw9=B3XD{yPRn%L|a*_w3932;$R=lo*gcilnR!*BeHc&X4AHaDA80(d3u;cZ0OW zq628df&5@FJ4^9JM0`@V8a3OcRay^SO7!XTJ@Z}yNzMOfk7&)z%)tWu55S=5ah4Pf z9y%Z}TkSyp_hQkAnd1COX3xflER+BPdZ7`QmSMT0jeyqeE>HGh(eTG1b!h7|gc3K8 z{@M zg_QK({rwe3ORCKK)5!-c*<(xLy{4YJi@wRWbax4$m6wQ-{bvWiw@inh;o~4xk?-rX zo~{m|Q(Kxseaw^&+IqK@ahr{|aT}nXp@WVp{v&QR?{=R3ymrw!!9n5hvnlxa3gP-B z*z+-lC|}$Q!zUQUp_Pu(cIV(xP zqChS)(8XoQ9Ul{-N=n9l(k{ANS2&hrxun=Dm+m#K{F*{I)<4rz4H zHlXB_9LQf%=^sx|#dgt7K`@5u7*D8dwd!(MznntXC-QHQhsS`a(m9X7OBCK(gIo@o=8@s+=^M8S{9Z_$%pMyr**0_0(=#(Q{;CCBoD_(2D9*FIE~K ze_frDTIV@lcg^nAb(sf=qB{MG^F`(>!VY6o+V?g;0loT-6`~r^T+M?QCeiv`+r4}L zI**x9b+-~P4{syG?sNuijr5u%X=?A6NtEn$33|FOIBb7gf~!%R9f@JFlP>}2m_518 zCZgmPKx&7u^=FZ?9($J9>x3{D=e1Q%>SE=<(MVX$-Z!!oVP`!%sVZx^VRehxYQkcd z?j!>KJQYjV+22{qEyw!wJa&ZzH3feus@$!H=mM`;(-aNa1g|V6hRk|BqsLTcMdRH# z1&3sdi^?Xsvo}+OMH2pxqnAG*V(a?y+GU86-C5dJ>GzWAvRuVQ=%qyy{HEbXcAHEH zNspP^M&5=soSzP+qUPg6ocJBw)VTI;9ay9`dU*Bkwy)Eb+TKYHq<2QJA+xTG!PD_H z!i-Qelhp5XCX*`g=}52lB8TT$huN5i2+5(~xRqQdYS+fDuNclJa~_wFaJJGhH5XF8 zY2lprDn)`v=2{mGbzT11=@?rlY{UnC^NpcuG&G$`sy~X2(yLN&iBcPfxJBD4a!4zc z+c#IG6{4QYnWLXz%6V{@9Vv2`%EXj&X}(+{jQog}CfIC?#BbvpWJwkn32HWdWt>wb z-_V@@#4y7SE(eyLkk&(jK^5Z4-Vtnjo!efM%b^@kFCbysPWca4z_1gPdp2ZmI;eH~ zzcz@&(du#6=aBI^RC1^(k+6_|axG=I8I~%Vx}wtFYB^U62zF17b!6~)?KD#It~6-Y zo=^NS?FdU`tsQG=F&&;pQ)b^3(%M$~@!)}#1qmVV>ocDtr0L8Y8`u6hd~(pfcXYDP zrrf{I?ADIKc^KHIhdc6NBxPr!?3MXP)$@da_v|?+SKs#A;J8HnrH&odJIcX(phS{O z$UTgB?Hnoj*UfnpLKmTl3(W*%4H`Ud$KOk)j$@81rWg+kC;8XZmb+g&=p}`=%(8Ln zP!clFA%ipaHgZkJyMbemms_5fik%Jah}Kf85#@%q;@t-j&s{fu3Y&@A;(g{3VO@Dw zno9~E)ibx{2mji^ZlFAlWZQJ+Zue>}2wZc~J0>i%Nm{;^^ST%BxtEJ&?EAAUEkPOBvhKMz%T>CqlGP%g&yXabBnX@ABcEYt;!)Jm zTV|#?E3hharC?5CCK9e_;-fY`d}Pz?vaU!mPMaRBr6^I4+X0>SpmLaxcUG zEIcl08a{`Zvo&F)n3XIPH2Wha@?5S{4v$+Zagv?T6~rBJ#pCCHyzurGSJKr>rCpvf zjw#brGV53q_I{ovCmys{`HY?&fS87;US+)GW&BDWrJ>`t?GmQa6 zH>%di+iR81(qGP;@v(#@d+It)ExSCBjY;{Lw|enitF_Y=wP}cyeOqO)upbG*9dodx zC!AEPkaaIuzFtIwbr9|&y*>RzrsPpFzXbL0&U+o~qwEis4r4^^P=)HzTk*cweQcKP zQ*P9Ae=yBYvHX@*6Rej0<8%R;xhD}ebmEb@(Cr=MtcFnD`Yuugp0h*)?yc=(vw#R~ zn(^e)pYUths9bz+B7~F8(dxOS`et-`xiZJu77|-*a9c08FNxz%Q&*?-nVewRO>%@<8p`u#q%zd1I3dQc6YxBb zx_qH^UgCkm+#k=z^Ou~U77l{G@;i8O4Z9xpmvQ^{-$&yuFDZ|CEA8EE@e*kBqw%A} z_W473Hnv1`J8ka{yx)@@4$pXp$y#yn!*HQu)onQBg(|CGva{>&J~nOn0O4%&DnQbL&->^iUn8lka5|iEddC)n9rH ztnXKFIIMxIRw)dP$r(3wjB`IeeH}Mm!7dph^|k-@PP?Qw-GgD+pU+DxTaX`Mwa%On zv#t-BcHdVUQg`|S`_Yfc_1`om(|^&JOdL$i|524pKxSqp7MA}k|9>Krg@cX#e^loG zdHkdC%~VBaW0^IIq>r~13+4*3Z)hf+{b0pm^$TGdEf3{bOsDxLfPu<|)->2?HJ8MRh08V9R z?%J@(EQDTk6OcG#H4{k6Kwh929)K1boB>QLV_h4x_YVLe7g0c2S4oxtSXEqEmM=E~ zPgQw!bOP=6)uT7K#L7~H1A>|_g_{7##UvHbQr6vnR!0Cr&-Y-E3P^(OeNTbGd}LKq zmJyfdRgjDg-+3bd*aEm`*M^SRFTWqe5KiH5N?0YzigNCfEx~~v`!-Jeif}M{TzP7aR zSaY)TC_;w*0_d#V1T?#60@Oz1q=J49_8}t1RqMw2Pf!?w&5dDPTnycPBQ_I2Ux3`b ze(%;cf@o@dEpPWv#fzeu>>ofTCBO7Igu}eXPQjkQ8bAQD00Q#P)_-ze59_Y6_&FI7=`M;<~VUK`nky=7YQaiUPm<_Lp1#c93ja+n`C~ z0$6?%$$qoXF7^#?pYP;-qva!fY|&PsuCDbhe);>b(J8TI3&3IEsXL$d6n$>com;`C z`76dH^CH*N`lqq@6H$4OUI<+gW$dhA1^kXqeE0W>t7B;a1JTt1WcrtY5DR_Ck03w# zKf@W!C84RV{4svQ>=pG^k z-IRR_BJ(}+hokl8-~Owze*dd36M-A9d<6q9xcrerbf0@p2!6~zAO(q}|JzG8nMs_8 z;L-%8)&5oI{IT-&jPqIprnLUi6Aq^|iEo3HmiMi}!RfD!kN<^SUnitye~Sd22mVzC z&_ut{hu`e^l(Q>-`fhR%NDq_=?y`UYyZwO+z((ed@cF0 za_v%^anBDcq*VC@LVRBP)j&Ners2)01t$0%9%%^86J) z<^ZJSmAomP1U0l1X`5054wW`NH>VD4n90QkrDsWNl27S zdDH40=#wub)wX}o6}D{u^gxSUBl|O>&sbFCVt#O)IM+5UZm@4g#kj`~uYm7(K-umS z67(AEv%f%iI^j3=4~gbueqdqe3O`W9@|u52klGV0>$~{v`LOiv9b&hJg6KSFcW(pz4;0IHEZ2{-|sQ*3+g=@FFhu`@`gg9<-cae$(G1@^>JwZz>0I z!}Q=6qSKg@&>zG!M8Y2b!}RY@)cX}uG^eU>sBgbu03a}1!1-W_td`Ti)K*N|qo~JZ zOa=pgiTfxK$wP?CNwP+*!JO+i%nw-iZNM+-U@b7`Qt3WeCZEtROTkB;;)FW!IOmieF&||^ghlAnn2Yp`xC@N*a+7h=jp>nil#&mmXX-O9oMJL zkvjP<1Kpw?#!{;t%Ak8nZG-oaeVgxfXNbhoNDtFuxb??FLJH#LMtPTHzwJT}e0v?l z?OSK0CqfsSw3{02^)CGGoYjQJXF2_479^C10-bgRiChUVO3CudLJ$@z$Waf%wa$R_}BuYO(Kk18s|f77!pb0zakn z;Q8zsmG}$vy!6^eH18&m29JD8uMddE?Oi?jW9=M;^^KaF0@p`pC0#I;&rko&vV#@% z;#GI;w%sBIz+?6)VjYMGNGw2 zE}w>_Z4S!sbBMdulWy(Nr?+7^*g{afP@u`?V|%hRi#^G!kw3c3!Zn2tkD~Hpd(Ha# z?lAB&;;u+Nx7~)r3YTf_S_?-lTRqgg9%Qht9sk3Dm#uc|AQ{pb{k~m@PzVw5fHSqK z#3HMBGm-M4IP0R_I_#F36WYUQImTUgK*5=(*E*Do{VW&Jdq3`P6bixHM33K_Zt&~T*5u>OgU?cH*q(z@;t$+N-~L-|gfL#{(-tt{Cr?sn1}ytn zSJ5k)|1x1y4u3tV&h3>GSCkaW;+#rM_JX&KcMT7vMfIoE(JGeU=G=9h$ z4v*jgq5*Bf4|Ni4oZuvkiARl+3NgkoBJBg{t^EEM9Kjqsb+wjmD%F}3)v%(+2uZ{*zh6BDa%M27@L>~?89%#Qe}TL0SQ6&ToPt3mA}UW#{JR^^ zF6*IAj!mPQ8m_G&QF7=)kivrxX_IP9QOD$MR}5)aucKQZG5GGuz6;ATAJ0x-6jn}* zQVbtmO}b}X<3EX-2j>#}p=*389SFyCZ&COoBlxSx6p zUBRnJJMdR3RX=@UZab~z^BsX+gD|X^6J+ms$jOpV_8ZRv!`CU==Hi|eKDPBM{>qW0`Dchy;S_u;~Ci;E-|km6FZ zm1wh9?rHr64e6*bhsD+RF3hEc5GoA_Y4R*;ReYM9A?x8I8;rkMaXj(Dq^g9@FNXc0 zM9Y@anG%03U%Cj7hm4(Iy0`F4pLF}~c)62|utcGI8wPyZ1UF#q{#cYd=W{yE%D2+G zC{8`7mvBWCz*XE+^>P|LsEXFZza0#w$_Nvp-#^g2Q7G9RA013I&$$VPVVN~n%O?F1 zqpxSQUXwp@9fLc6w2JeYrQBT!NSUi?M_={iqx?B7>my&T>lWxMxlR z_4k~QsGeS)RY*uJ6TW6ue8Jz>dH0yOn{D@Ce&_Y;lQS&H`MLAGL06MEA;&pCEONLK z8tKV^Q2$wO5^PN^&RKXT55l!5CG4Sp`pcaz+hS5Y&>RnJ$0oKlR2C9pKchK`<48C) zlMMLi!3cYyZB+od8~-4M6<8e7 zQtq)q`LP>Tir}kDz1Gda#D!&)&pYx;dhCt3@vJKgD<5 zvD2MIM)eflD4son4J+?rBH_8@EpdV)1!>}5+0kBxu}5PapPDnjFQ7gF(YHq^Ej(VE zDz9JVeANDYcydeFXjUn!bx7n=m|?)AQZtpt`Tl^hC)My#^=#@E*9BwEcB=;H;`Tk$ z8X2DB4byamew@4mGo}Wrcs^ly_{7J*Xi|@cVf|b8>?Rwasme;JD31>7PtqtXa#C5= z2-F`PQ=y*d*Lp9NDLmrK@BQ9kxk!o-p+g@5&lli^ta~bj<2_?qY|EPo zN~qQvR4yJ)nL6?y>OSp@EzQMte9DXLeOewdBythwhX=u8A&D?P1&=RUcoeV$qCE=K zQo>bN@AU?u*;qa4rQ$9Llc`mc4#Q1CcQ5@QKQr`+CO%GJIbRnB$@%KL5bCv2ZWxhR zUnA4bpV{P$_@ZHzWp$u*?x;`6xkB=)Smo!H-EqavtI-{*l=IIpHRkV z=vlKxbj&1u2K+vLjMKRo9)`7SraOU5@l35TAu_1si!?ZHNQtUNhCn-asnGt_rsQa; zhHyxYSED$2)d&aoGR6hxqw+6~WioWUt}tC?Mm}q4vSGJ(R6dNE*B=cns*6G0xrVJ|S zxlUc7jCOw|{wd5dN?5Cr#2Ee>edi*q-tL)ii>gqA1ueju_~PLm{1q0CPm$DetXT_7 zcS)gS$n}6^?Amld(a=lu9XpyE3|q6>W8&qT`aWKK)HC;(vX-u;Syl8$yK_pyQR26E ziK&Jk#*)txO&|Izwp%^yZJx6L{%y9*_HQzd!?pJFgI8KtR{Ib7W<$RL?Vla!iL8fX zy}gQII`m%QI;StQ-aoVOr{Mh(^)S4-iX0|Z^_dK&?a8mxMkLo>zh74JWyQB3Fi#gK zsW!wOhl9ej%U_%Yp-4WBTe*#dt5*emjUv5CY^!OhRdHfM6yXF)#T8k!G6@LxIFR_H zf1+m3f5p9qLFpWGY3w<2P(5~BZydoTraX+W@eNJ2vy_pBpILkb1`FFfa*BSP*d?(Q zYVeC>iE4opcSW!hN7e`%uW}J}HnuU4k2;hwlih<6_c-{#it{CxI?3m9_Teu1e!4Wa zj2yq|gd4_utFeo_@H{RyJk@6m?7Us&10NB;5DBY=d*9IZSwDg9ZcVwk5p3%3WcjQi;h zzLkvCw(v(n{%D9?Ty74JLymgR>Em8c*0rx3XL+6$4MxW#2Lg`M2khL$sem{HzF6r6 z&u9F|I`@RbXFgZXePP;OQGb-V9^CpQz;kW&7pu67IZT#UKgN@nSrU+%Pq zp}w3CCnNPrrwy`=um`tS{iH#M#R=OM8Xy)WQRg!l+G*DM$StGAhi3Zv?Px?_9&t|O zkZ#ZF1poaYv52P_5;|vAYP|UNNN+Qby)idp)LX|d2z>Z`HnK3YFHQXnDIXpBm0t#< zHys{1;jS%^nmh!ujwp+XM2olbghPxU>@41&=gGm(tTh;Ya6Rz(v=^(A@pR(vo{CNR z{XiHfHoTAzk3pQ$Kia>cpby6)EJmoqB$=8M%jbpgp6cVLAalbdm77=XIQ@t7iz)7z z+WZeg^1Z0Nw~9Yu6lm{==+6h9`d#FB_cCqp``sFTmOs{Q2^J;3PLD&ROdTJw+`|NZ z>48!`W0XIBdx|}4aHpp+#XZ3A=7ammcWaPR+;R4!5xDQ9%I2u?+}7jLuH(#)nAMtw z3n>?0xll+%0aXon*W3RPqWqRUJ3XG)e0_^|5$|D|!3i#QGQ)l3K$>@7(5|yGHiJvO z0&8Rw5h{lkS9VE287X=C?H&FC+DzwQwKpSL$4~#^C2UOhcB;eJxI|U%6eQ)jG(lhb z%A(`XYwCu*#+OFl$u-0)+&_!Z7k!cZl@WT@4%h1oi}&S`CVP+xG|J{=ubWhAf3;Tu zm)apeugd=!qY&)bj$k5%goVuRcCSM$Fvlp!md-GtO}nHYZkA0j%3UpT^;OAszN(ba zowE6x({hLP6&aQ}YyujGh?IeBS2f2@C+}HA#QE^U&(c}6pRq3ST^FKS@o2M3u3F23 zB&}M!f<)C9fE-6m#YLC<@T>Ab9iGKjZ-3zw20P8Is&1w!+wjR)9V)Zj{$(lhYYS!iC+1n9G{kQw1v%I64J$1;7khKIITz;o83AedyIrI+Xf?^ zH%?MBvIfm}tc*T)QJvY??SN|KRV4A|c=w|6>=46t#$-JD^=+7mPq0Jc%8JW$HnC&z zcnbGKz3J!2%=Mg5jTjL>-F2NWDyh9`_vAZhBAgg^CvApiz96=gl76cby8^TZyh-8{ zmo8O%t4k=!pxLa?8NT3Dv@*&8m$bNiyXzY+H51|{-Z?l|5LtVu+9%2MrYhZ!z5m)f zb)1Yz>_<|AV#wv!G;K>e#cpVXSTB*c3x-WHiKgEJb?|+ff2sPiJdxRutaIN5DtWFQ z{kr;L-;&Onv1utXR(|btE91ImctcBciCV9T)2fjtncsfar*(;D+3C5i^E&B_dsrkh z^g?ZRU2M}{Yc9hG{qowl*eK`8+MFVt?OHUhBVTa+)R5XHe;_9`jwRd9^V%Cxt%W9+L0rUTiAHXSEEjTWw!yw!J1L_NsSXj*=`0$H-_mYwtX;zkcl)`OLTR z`G5lVTT?sLtgs=W{-bM#DorL{XTaQPaz{DP`g_yR{8oy`@}#r16;qesmt;IyqX6p6 zoICb#%-ZERUGN75O{vR5{zSW*Vf5$Kt(F%HFkjsbnt!=XA5v?-tQn%lfDwE#Ox|c8*6Wj^}1j zU)e8x#j~*|EIb@)A8!e3L4@6aZ;D0qX&77@m3bVugVi{ey7ey>9fm2ntfTI{z3SI8 zLbmi$S$jwON;=6ys122jqzCPs+Rv28ncnx>#`;}zZehAdEG<0>UMQNcIna5toUl){ z`FRX}7#2+P+2)HEvccm02*#?ZBC1SNh!rg=CU@BrBPcHrEi)DfPM>A2Sphhc%P#7q z^1f0suJ|f!Rni3pvFy$Wyk?2u(C!1{)09zNI*R2_7KJ0=Sezl8rG4I*R^-DHVD^r? zsLmypfaQ1L{uPeKYTHdtaAW~Zi=~*}O}l+R1!r`5x5#TAa0U#NTfGTboGI8>y%bB+ zk1-%L*?+`(n!{HjPqz6Q9a{ALBMQg@0cFuOG&ZdB)Ah3QR(o{`GTY%C|1TSw*%1md zK31sjL>D6&<~F3S15H0Weq326!Ug3~ngL!{;zuAi`RKYC42ji6ZqD0u0%ccbsM5cr z3c!1-ZSTm&)2Qe=yYsw%SMXtrK2I1cKQ~|f>>ROLV+Sd1g=a! zxgA^~cfurtZ@4%-Vef|IOe&e`;?|VCkTBap42u}@AD(^Jl%Fj&e&~|p^$d~~uIB_8 zuhF-jT~+ju&T@-R1NhLAmLnLw;uz6!l6gkY(c3~^8|Zv4IKdB|;%3qEH<7yjJRuj6 z@Kw3=Y9`{_PpjgNp?jz{b2%Za`X8pF3?dKqYVO^sQZU^S2Iz78IjssORORiWuoCRe zyC$$<9OH;Ix@v4y&hgycR8l9#B83bZP$9<3=V|xC{E6QR>q7V#2RbTQ-N!oJ@&p~} z!`Mh%k4yCW*qu7GDX3kHoQ0Z>U~;2@`xWQ9)fQRNlC4w{8Fu7)uLX!=mFYAV?Gh;4 z@)j>!c=4G!vHiBIHJ^*ClpA9g9cl@HP4*O9`*Y(w8e>5u)K+Z~bmxtM6^5sXEcPu& zZS4FMi-B?Gwi}+A;+5nSLmtf3_~y>rMI<_lzG>YR4)!K2Gz+sXj&lTpX{q<_Fy%R6 zKW$IU)YsnWX?|J7@)`TZGg^z{=#tns-rMKY_VLxF7M##~0%t0=r6G%-6Yfv~@6Z)2 z9fGtTN^V)#u{5rY#5s0Hi#o<$H{9EElZW#)HckZC{n`-Q9rNSozL1x_Tyc!oC-?AJ z$dQ-P4Dx-GjG(CTHdr}1`c=%({8B_J<-9WELFy7gemOjmIpPr718KoJf8u$q#L1JhRVtm$eTC-Bcwz0c z`^HF8dS63?fU`kXM`HkJJMFSU=m8jlobqY(72riao_O5N3x9dxWAAFo@jFWz zTPW0WJL?q8-~yxHG$J9Or=neClRDUTCA82|2){b4j+$^tQuQzW__RysIldZ|&4#ch zWBRM!LO}{o8;GVd&NVF4q`#*3i2svf#_9#dq&%Mb#PMTJWY8(wj??OOC43zFhjnT2 ztNoXw@lk&RpIlL8SM6g<`8Q^n&a#6h?-L^Cz0mZeX6Pq+j-X{|g=(11$d^&6@Y}DK>AB=)+-8&A&#~e^UfA=>IP0X?f6qJ z=oy>XZ54I9tEQtberX^=)iTs~#*W+**{sSlgPJRmZmjxMi`fcu`JYTiF0d2{7uums z)xG7JHNmy*KDWxNh(HXD)9H%<@(pRe?gSilOxrH*#w1pz+#m9$(7c>fb|GK349gtL zz)&HWR+2?Tw)I>~IRnFSsu%CKlh$fx^RkfmgCyi*@)qB-EI~K*QBXXwU*yPyX5^V- z2d~;~zgF^JRuAVo^o2t?S`XiV1qOxoZAJg|QpAWrOq zb}i*SaXNa93AZl4kr5rUw*t~_#!orlsyj*`(QC@??Sp#cE#EZ*m=c*N`)7$1EMYW^ zz|gF=a6B2P+fozf4wad>yx1c17N_4>yuR?9iv@O)Ryaj}QgVN>kaZ2jjEt$4zCkz$ zp&-H^Y?3vBmT|=NensOO<479RgQWU2)W-|-_9LdPuCVDtaiOoKic$xcGzle#qReGF z(wL4|PryDL6pqknwMhVuzS`@rw41guA?1p^)R{827LQG#SVm5E!C9I6&kSYR_@{KH zM+zta@EUZw&9lFz;_^7qPF)KrReNIg`axSGS1sC|QB8wpwQm^aP!@4?~bYmrm3f z%lgMv@4{Y@V78!e_4tiXU*3ncTgTb{jJvFxCdcs7;k3IMXl?bl3qv);RuPV3rk{D` z^0}w1@|Dx41*iV5R2`ENZDnSOZ+4eQfx zB(Bl^Qy9t{W5UEi_oI}~p`!)*v@a3P%Tz3NA{yi515D{F)?-vBO{?MeEVz z{93lTN*weQVN-RWBCtpGS%s;8TTVCSLxYe9Cmx&b?kH3#KQ!(SyLGq2hdYVwh+Lkc z@7k|QneQHN4Wlu@tGnp@uTa0uvB9LRedc|BEo(Vo&fwdokoSW|H}QCX8D_M^C>shK zzkaNUQYX!d64-w)F)cuhv*o}Xq(as@VxLXi<$@{sl}DV}0QFoFqu`fpe1ROttF{9W zaouNkQ9(K9meKS9B`22PRZUoZ1*j#jMjX*WPTLfh7L3K`)K1RY6hCO78RHyq=HrMwV(1RQ{=k(`#Xg?_$QoN}nd<0(ZC5w)s@fTo2k6Y*SvY`vTjbE8E zg}aW|NWO_-MBCz{$%s7k!(473&f zaf#_c+e4OZ5&M!M<}#Z=^c*6}fbb1_z$=EKaXk+&0{Y7+l`?>P*`XLw%D5zz8ksYFpx?;2 z_VioHS4utC_W;HGf(qw2`LXvH%`{CSyfbz_|%Ny2+m>k}0_`r#$DT{cuC-9PXVD?Yd>N^Cl1pDBq-Oa{z$F)5CPSd5H=Q zl@`H+?dz=$Qr`whGYUY-ud8>rk|aO+ZXc6kTk+&?3~M3W@7sW}>xi~8W|!h-5w1#m zFV-a)u)U;PXJ@*Gpo71Hr)bw*`9~R;``vjk<~1J>hm|gsNSd;>AGqKxmS8YF&&YbZ zy7kWT?-Q=r70S=fzBfHYwp*H^)F%K7V}98162xa?;lA5z`pI>q3*I;=TcMH@I$H6# zj?0#;baC_fY(EJ13YLv(t~{EYA(4)VtGVc?EH)n7z_6R}WH}6zYYemf-F3-ah;yWy zhc)zffVRH!WCyKe$-y)Z_s!zev!|2x&k65Ag!?mUsQRI4T(EC; zF<2~IrKOEVmM7H6O&8Ao?h;Ta=;b`xxH0&Q6DZvMY(?|FQz8qj)xX&P9_QN#o zAJm8}4c1>A8&g!%2H)}FUtLGuI~hmQc_(XXu9q?NWX)0E2Ipn_GDH8OMm?;e2}R`2 z|B{lm-$1h@^-y17$}~nYvgXZq|5TLQ1JgSA?h$Yiti6B{g)zKq6r3v$*pz*$b^G<@ zMHq4Cy)ewH@ue{g(*mG7S+?vTjge&oc50hylm1pH7h~#82;Am<{ zQjz9p9~a3_0j26qVk06;R*Hu5(h~gqLM-x4o;q1B0f7K#b7xrc(6iPO2Z5o){UYf& zF?zd~noAM}o^9eI5pOmpjZZlxc;ZpPVBb=pDUvT4aGeJ&(GU}~c8-q4IAT)DWZVo# zcVxd4dODf5v4Wg{8Kx-f6Z$H)KTTd@UPsLbuo}PxtA0U#SlQc_jT)T8R+68-F5DuC z_UK`>*#K4d)i3?0y~~;%KP)jkVma?Q^>yZnaIQRW%2y>SJdKc+xuh*-tl%}dltid% zgcvtUbd~VofiPSqkGiuAuQq%-kxr~g(Y>SS+7H14-%srvf}r>M0E+trCn(HHjGg$4 zC2eHan4|B4n-_vgE||oADRH2!f!Ki6#YFFBZt9 z6MYFSZr&Xf&72cKt{n(4$aqXnI#u?is{0ADBt) zx34olw0AXsIe)qN`q~q*h-tG}g~U=*lt}UFD=C@gR;edgIOLn4;v>Vb@0%_9u;NAP4^6w?{NQmrmL9_d}rQ&e+t z$Nu)0;>-I&fyqFk!(mLnNDsVwNptr=m)3!|396sJ3DD$zZ}7EA0H{?CiN4}`5yVoq zI+_%>L6td$EtUN$r5VW3PpyJ^75+8cghoE*t(_&a!*lJgdBV?16G1DMBFkRM8uJzV zjPx1j^(zjmJuG}56?YWz>Y()`x+uaUqT6S7P?!cZ-C>@=aM31`>F%<6=n4`qR_x7f zdXk}j#9zZ$PBY>w*LF}rH#q1t65KMJQ`ln#C+$p+D+_@}fgiJbF&x7?n?BleO*6}h zHQ)>n*0nRE8SNVM;dgRaW{bWM0^nSan^e^X$|AoC^KVPCyhYpK+32unJGx)3wmzPOTV?xz@&$|63R z(VRikj>kE&L-sY=0@_VXWfbm`=6CsiO4M-fZ-z82G2{E~)}{6xl}ZowN3l|OO7C6O zVpHt4G5fu@?7ll{k*u;F)VlAg)A0qy*kt*BnKiJc6!x#hr2cHXRW01MM%=5S9*{1+ zAa4`o!5k)4zMTIHiZ^X?)+@7w9>W5sXK=6D2xW5dJCL)H1nM<99-Th^bnJSg-mUzC z+Uzp}Dn0?b>$@yl&TbS8{fDg8hy!WYyCyUA$60Ts{(verJg3e%Vmj(>1t7}O%xqoh zL`QD8*kf;)6bZQj;tds+3l!0aw{|_$$&r4M@rCJ6m$hWTi$HnU#Lgj1afu{j^5-r$ z7TY05!M5cj|5VJ@4yoRV8&7eAz|n7fO(MnNIa?!$x^4V29^t zR^5*eYSriKHan8-G|o{)+u|^}R(DpHR)Ln>UN2_}KFk!={Ckmn&v=oFXfr?8L zyFlR(6H(eeE&XWs!3E~7nOcTfAz6EJP>W5`AcKiRGU(LtPInY7Ya0p;qDe$fR@cH$o7=hE>gdSeoWYR`g){jJz z6|s&w%x+VFZc)^-7TEl8Mg#A=xqcES{P%SRW5c^H3gI?Nb!*jUvKMMCwc^P z@9SPX(NaHrN>Y8>Aim06htvAOOMsErci^c?Idm|<;Z?Bf1U;*`rbJYJ&am8b;P(O4 zx$@ZxBU|b^3Mm7;745TwK={s2p>BKaNHaW=%jPxeL}@M)?@m96M#ub^?CxCfFGu>I ze;Oe+a5oO$#x}jJkT3U&jx10RhneU5hjQCSFP=DY<|eDCo<;XyJDijW{l$(eo3aG%k!4DRWhux{P*M^-*Y$T-wO|A6(NqboP zESVNO#~Xa8vgLvqNru{u3T>;|ATb<8kM_n>WC>{K25XGq=qv*7Dq)Aq>rhE> znwfhIvLW+Z8^6J?KhsZQ3Go)0*)|_|GI=?WzL>u)qEgk~-6#3sIdO15_2g%ZI2rcC z7~Vo_z_I_0nd=g&OP%0(vMgGw{l7La=B` zFcx#sZ*gbwSyH{G%vJB;j68Z0(qKyPPopy2CE#86lXO7;+>ESu!quX$Hd^Mfksa3V zTlSl|Jd)_VDEaMKdi>M7V04?s3Bka|5DC2PT>QHg31h_?ig(JFclZwzU(c$Mc`o4I z1lQtw?)oW66koG?d}AxrOZ1URIDWG$#@v;+N%ENRipcN6!dKL-@=?&AoWpPy=dsz9 zlKX?IAQ7w1+Z%_pDGKzWqdK<7o;4AN3WLH=eIV=mz3`;dx9ezporBArMoP~fu|0E1i{#y+wz1B!ZM4w+crv@PK$|>va+&XQ}*3^mTbS?S^%!ZhPS|daAQ8E zugCiO^WH-mt&H6B58no8l`{H|^D`Yy)Gd+O*KkfQ?EJR*h6cH&lnFOd_pTSD6)pDP zhIVAc#cM4T@4LL4f`nAv#HtU~W_1${Wuq7zb>f`O6%^NS9oqu0b#36g(cZqtPH#uX z;8x7Ne_267c|U*I;NY52#wN+=v!=x8EXH@{++}cM9=?>J@3;#EGgXhH%%q*-R{u^u zW|dnk^ZArNw}JVIj{Eb+bfG|$j?^$>5umS_>Q0dBO$DLu@U2^zzf`o*paLj%stDZc zr*l`s^-;1dVHaoo9iBX}&>n2Ua$Q``oMy%AP=g55K0a1sd>*y$%!u>p%kp8>fj-7c zZ_6g*;e$&yvxRuqW^`+h4Ezv|*z@>+(IY=Ec4X2vLXb&{9KtLgf3EHK1o+{5saH@; zw#OrTKeP18aPtFStPhT|cYYe39iy_JMYD+Pq8ywMWrvm&hJN{x-3NewQ^a4vPGGJb zk2~C)Z?_6K23x#OJg5*+f{|=3vJTr-LJ6fdTen}$`_JDP9uSOt5A+L%xb#Xi_P!(- zZt%B@PBjTh9cD0hSdNiXN-Lq)5b|lxdRTr7yjqZS3!z+~mgjmv{#Y5b(WW$g#Vt@8 zlP_~8q7?~Ry~Ho+O)8W(;%!Di8lxyiSXn~i07?nTPqPX>rIRee1xPG15*v?ATdot{ zvV}i_iOuV5DkN3)d6zdu&)q$AdGs?0K;ux_d)VwEL;U0tHrDz&9?GbRHh7(wZ3`)+ zK^qy2W$!_t+!Q*Z-fzt`T#Fkna=Ral&MbA7laI}?$w(jHKepZq9ezOHlog3sKfOL2 zu!^@)+oj0WF=d$miz~9jv|G+!7zfc-W>2s>ldN*|$x-*yYoKetL0&#pK|&vQ+t~79 zMB{a2?o8S?#jTh6d4s_*$Zff6)PT4%c7lz)yr6^i`Ql9{U-n~2DTz3LcZ&&r>VDG* z$@LFjIXXucl;VWO%WEQa=0g-8LEjJRBXoi#bSV=}M9JOYE?e#7c*9RkF>`2d7WMPa z4s?t6>zxU0J#t<4-BKeY!uY2(rQb)O2biX?nRYNlialuK)ZOpj64uo54-vqq+YO;54_FU8QFE>+^#TA2{>(YGD5+-af6@>)1FF-CFO zh3luCCz@G9ii5Gu>^3O!ZHMeXdBz$u+*z6j)I8p?a}B6Sl4zoDrHD~Clhl;Y=f13$ z#9|G_8nY+}%WNED;Ayw(7hpJBQ;Nl}Ch* z8h0LvK;e~ntnUqAvAi!hu;=aNF=FAdVHxrOh<3Qp_uBK4`vj$G{n#ap$T92+xPFp5 z9EPJQGvs?bI$xv@xfo71#O%BX+OyYO{az(ND9;?VQ-s=9XzZ$Y(#|iQ$8X(MGPTTR94fX1J<1~}zq5CjM`tAq zE&hPG7PiK*&Q%A0jCNn?Bd%3BGK&HaeHaY^sH)eZib!9;aRZ%VV_4Igb5q#zd>C6< zga6(&Ty$Y(9%_Wv+DY!J$qzx)fnK$h?-*BoyY!uG+(fQK=Fxm}O?4~lP#AXE*Q$@j z(VQk6+2?i26FNulEZDk~)4GfuN{4=-k{X#+m<;=V?GMe@x$>-w$TQjHwS==!V<)MC z5^{X#n>e11#fG!!p~uPck?v(Gwcc(J(U4jVYx}N;8RNy1i8i31mZba&y)3)qebwdD z1od{@^vANwuM%wLg@HZ>OsRp|6nTUQo z<6#%7u7~MdeIsexc^T3lQ)@}QDEN$VpyXYWqluV@-5BK*`gKO8{81~Ao&WVpK`cA8q$xnO@T$kCb*$4gbDw{ zo}EmBpA?8K3ri9`utN(Dn(9SeL{F5ioOE^0$7|vK)Y`!zT2yv6t$8#?FZHQJU~n}(ou(`)k#>Vh*~jo7CllxgyLf3O z-0e1Xmy^iaNFQQ<3AJaVN;XANH3=*zvUOfSe^GeDJue{PKO;Am*UlN}{&Nee{NY$M z=Jpp7u*Zi+I(TFW(I)!Uo#uV8HPg&SOX0QD?Ph(KPPRtWJGS|kSM-GJ#J=}Nm*#ID z%B({`(+d^Y%czl8jMD4-aNZ#iwy?JT6f?|bMLj{XL@a%ZY4K-Xc&&G2fqlx=cZRY~ z6MLq=@+@C`SJlISVuq86$bD?4$Zh`0UeKuO6gj@ePclqr+rF)p)k1k`T(ga(MC7Xh z19l(B(-O>I=>p+R-#69RKliY@-L&^&`#cx$6)a>pmV4SSeqz#DEIbUs)-nDOPnJ$S z#l@W(y`%fuuybm723M09p=xEsz%+e>7Qy~zplCpepr*C?=TpeY1!*;(oipb*Y=>+1 zIuluC3V3sx4h`hA7_~6ioifFn(j0I|e;}3f~-Rv5|v% zoh=6!sU@Ent3!wZxE#SgfI%*AiF40>AIWoHaNh?OBs)K-{y5iJJgHwYY2G8sY1uk= zw;fGd+<5m>6-KNsNyyk2me6E*4qS^!r%}v6*8eWKo&lkht)P)pcbLzZ6?oWiLav8{ z-O$E@YIxo1Q`?!AAs__U(L9iU-$oSN_3~uG6}flTb@z>jM48x;a{VWp)o8;P)cg_H zO|8Hc0>9m$jpyKBEfx=HkNW=l-BuQ|&Tnwec-%fJ2*qcIFHS1l(OYrub!`O zgMgw26QO5-wWsmXWn#F>52#|NGVacn=$#W=a=<*rRvRQt!S@XX)LJlLfHoc{$?d`+ zl59Bs>#pi%%ap^*rNCK(QEeZwZ|ULY?f7#UZC$FfIIhieJ*?W3uCf{ie|@i5pMoDo zPKAx}5m&VrU&-<$8kl1N~wE2r(EIQ8MJ zW-!ni^{qiAD5JPAr~aVgkZJtMBi#;FmE}~g*$afnB85~;gCn!aC8XPJ1JpAo0S>Tok8~-^R>v`R zxuy*t*&3ocvKL7?HX{c2tf5#WD;IWOgpW4?$sMFjT_j9O|#|Slj zaM&~Tb8y*$k`OoXQ+2d(7eX61yWafZ(}vx$lgI{O(W-{mHAmM3QS4!=0AUdHh#)7O&8PeQkyBXYB|F+@tE#T1wW zMWGugVD?rYS=6t6!f_*FvIzTu@}jjGd(*`o;T}8&@(#=M;77ak3Ja%yxe;pMAvJv9~pXp zl}~qmvRbo6!cm1K^Qq1r3}u?9OAli^xQQI4UK*IQK!!50>fzf1Oaeyy3fR%$p4G%z zHgDu6ot*TqG|9If!j_W_nYT+X?i#8eJ^G1Y*M%F`PZ&1NwKCglJN)vAm3;$IVzK}( zop8(XPGqafiOO3!RdEieq}p!J5kIPi;#N`Sf>QY3hpq*Bv4YsflDkMdEy=9#xpR2> z-2;F<_YGFe6cdiBjgGPjEiu-uDB*Xi$~*>Trde3xg3o@_~EUt-@dzZ(2ZJuJ?EB55xnc!H;UTK zInkk6JCDgAh{A$cntyY;*x8{}px9$?lRu(4EwzE4T&Dar*ejFfg zeTtdfUDDU1YeW0qW_1Pa1vnq*sXb|&*487^qM_a`8w=axG;R$zEPoF+n0Gt`PohVs z>j!IxIEJ+%C-6bW2tf$PFTk^VE;Y zq=<`B=Pm1`7APlB9*sRLy|?>rGnHHp0j;P19{@r?y}uTCx5`XJ8h@p9pkA4lU0F$O ziy@MQD%>Dv<_R52ECN9YrA&D!Zz8eV>w3+7q*vG-?u) zJVedbeOi=*u7NC(r$Fn|hSajv72U+Jz@ik_j$eg8IlxC?iTQqU0`T70rjekZD+2Sg zMV8M_e)%-)GU(j4fP*+9KH{+4{hPD}ouGq;;)f$f#PW?0Gb*JL`xb|ty>40nt^_Ng z2QGkxYB0X6%H@&PN;zcefx}63v)9L;=nd0~Z~(DqJ1Tjot$j-PC70L~zN?{~5nGx`OG|{Maz^)`|cX8VMdE^4_0e zC$j=v^1~!fY#^7&WYJ$6Nh3hc-qM+^98vj7C+`^iqV31Wm~Yu)b2c+7#Z7CxQHN-U z+u6Q!^K7JMUXffFR%O~}BSySMZFMjv%jUC>f5 zj|5pp>;QWD%j_J_V0lB@wEl8!y+dS*eSwPa`iwFJAIxlYCJqW%`Bfj(GM!JCD563yM+6y;F67~h7zenInBQ@Iz9o#`P5we1Ca9>uSvYm^iY(acaBSa4I$pBV)m z1`?wlsz|7H<7kYNNhjpL-P{D3^jouxrE_=TjZ#;-_z8cw(yfRlz2C+iGs#UG?it0+ zg`&8?iI>+ndPr@VD3JYa_!CjsBoF*ad*+V`X?tDj`13oKd+z%7a&&UTJelfgrrC?E zp*zjJ5&8~C+NNaWWNVKJW}V!*PkPO=m?q5*og2;Cplnnmd(n$}5>EmTi^Y^-r)58g zS?{x+mT#1?^Z$qHeML(^VFGgYd>De+?5~`S>C%2cGclhg zxE4U<7ZvN$%5=-Rp5AG(B}?Svu?%fe(d8}JQA!pf!KEjw`ymxxn+F4fH)|=mPV#DaHxDl+iD2~uv*hU^c+>oJXj@y$ z5dk+RVu+p%H@P?aZMSh4m_uO35UR|jZtcyH1F4*cK{|ChR<Jm7L} zBp5>muwQ2%v9Ku8pM#20soa`?{5OtqoLeWv+rBMXpzDSBNX=jcyA3j!{u$)^;stR- zcA|o@dX0-s&IIl)Br*8L`cVw3X~Ww^tw^ej`4#5)hdM{m3E+xeXecM^zhdd+EaA@k z2qo|ULTE^54Z%3SJWtz2PB+0Qx6vLs;f6AS+W{v7uC5loP+CZ5y#fxCx6lMm5h_~( zb9b=1x@IAD#V%Ee6efbcj7~36e{axV@AUHR4t-_Na-`yHiDMv zXiN7}^{h?g{NOPHdi!v2m9Iyf7x#O1lLbborC*wMVKmxsVF$ioDmLgEr9bMFMf)D% z11gx*n~7bJta-~x=arBV$@Z!y?V|#-8652Ot}o*$vH^qJm1VD-$2y232xC-nQ`fF? z17YbO_{xf1&B;{(j2ESf4{wpMWBYL%-fPo(hV7Z_4!bs%NXdIFKCtR=YD8vMP0{W% zq3lShIOC%Jo74K9INvvR9@RH|z26?_4Bof>22r1noh%C~?j`o-gD4{i1;}wEin=EOAECvtUK$a8W zy`w+O%ZVCGub%tUhcWLSjIfdRH-eXB^^iJxCq<(L^U(MbBlgw z8a{>z{=BgKKZuCjYyhTq;NPa>S@ru@Ci%ayvg`(K*#^&-^;vtNH#f~)JKA>@ne2y(;5wKYg*IW|KZG&p7*qufnMEPnQ zpB8QvZ{u7+Bk#^Ro*6W`kN321Q-7angHuje}%gtg52~eQi}- z^%W$^C<$A3=gY&qcp2;Tz5-tIh>g}FrU9i_?pBk8Emik4zw$R4H(Bqqnlr1oGL+cU z+aU-+%e}E>?gTNK=0Qe-@MrS1U}e-uq=;J(Q0-=C($o&Gs>{w5pL~!~2TS0Uml23f zRF|Aw+NrwA0VVq5iJ`Ev!fMPWc5W4$Xrk-Y_xj!shwTyNop1w4rTf5$}T0mBHrxAF!0y!SpJMRo=P^fFrBppc4y(+M!G%#ApRBcTX~7t@r57#tW0{+$Izd|5-z%Z8B zyz=8DrTS+movQc@4_>F)=6Na4K|;+`zY+h`4SJM=_|>3b(%4Yj3Ka93w5+|efvsY#;qW#QI7zj-?Czh22V6QS-T?!) z+SrtDnMmkWs&<#zWes&fwrR+*P-e^YUb0%1an~sWMT-xw++^Ur1MZ8;XU(K=`J5e>6=(xq z(aPB&lp!RX=Lts+XD?V#haH0te5XzEoqpZMwv=!JPD*QjJ+-ZA(U>DUHr5)%A86{C^>x_<$bXlAU3_wP>&jd|TP&|l_6C;T(|6N#>Lp0)i1A}W{R6<%X_ zz*Y4AirU$7vHTGCwiLK&F=q*qer>OMCZ#*>v%NvZr^6tEz*gpk=Jk$=QDoX-TFflT zbFt$%^k1lBcn-{@=cd3iSJjY#?;{u*+Z!0UcdU4GKBql-0J zAC1WWjKQ66mt6306Is=kd9kwD(#_z=Jf3q8MHv<3g$et1eklN2vPX$Ul9owtVj9mT z)iCpIT(;8H2U3gzFEv~JC0@m_8~_0J{r$@K;|qiDQ)_|O(;}m!02M@PJZ%7xF>jl5 zMX$qL8{HE}K`#<`i%*ZGx*~1>c5nD8iC-vR5h5V?FaeMV+e};_H967sw=V$?lgGrZ zTL?WJXz95EXRD1z-anM$@Sf<;tf&;+;a^@W7_9I-rW(LLLWGck1 z8f|xTV&v`lGk;ic>0U#gdq1Xy0Xpv)beRXz-U*Y zulezlzV#1gn}m3YCWQW5Z)CkP6&o7(=p-m$pWBp{=cC~>%NG+055yr@T^9JR{sksu z2C!vS;l~bDKjF>ZB?QcQTpJ9s*no5c_sHyxCCJe$V@AHhPvw|^Fg;!Nf@<@%3v?>p zf@I%U{doczM3AJxJPf9Ko>A)3*`&$Fw4TeQP0_@5uCtB5DYnsyM1`P+A;^C>i!Vrd^1#1w8D2r);Y$etFX?Lw8rVIoTaO$=8Vcm@*~yAr zTDUBAn0&s0uSJN}-k&o>%LLs_5*d6E)*fC&5cBcflTrDPUgYUI=nvjCOL=MFqb{fi zy_NkV?A-kwpF{A$rFCQH0v!;a??W$&pyJ#%?;Sy~smu`0l2{QBt$)h<3aWkpdO(H0 zfcUQ0X9bumim>~LkMrk?mLkCLI$)Z|u8D##J;U(=SAwB_jK5+ZZI45zCQs#tgVe~$ zRu+Hxm_9%a;wn^^&8Rk+|2|7y=Xk$&A(22-&S(WLDHtE*TC9IsI~76p&2KA*!9(+z z_w5BJ$*zZ3KQG1NvTK6vOMdj|_^o#1dKRI5_n9Oh zCMs*#$jE8$LtPxAzbT7&gW%dFRm5*0{N^O(vJ_kVt``Lw3X7X|)llcG#S_F=#^57R7g~0;1)!mS$Mo){ zu{o&o=nAS1{mYZt@oZL8ya+OgdxZ=#2mi(Uo_XI-`|+RGUTL=_A=iQvm`)BCew=e) z={#BqrI&I;nx3@QnGs5r4O46!M090&m40Paf^_un+e!Vdb^PA{KZtF8gas3SE~a~^ z&$5?slqKm`{J1-xp~?GTHVs!5KIO~IpDvy6##fZMr8bMU$ffRy8Ot|Y+ z`Y#g9jxDamwRa@t%4i!$I-|fWUmldOmh!xjuk~f?(lPpP93K*N%HggUKe- zY;SNuinyl$vz~=3E2EH|DV1Um;dS|IG;(OAMPAY7g)u-N{;<_82UNQaVh27IWosXw z?UL2=$u*aDH~HwKG1@`R7{&9r2i7;zY6(nq3yEh{qTFJxEr?ic$=WaBvHya{GviNw zz^XqmIBMJ2t=$xsGhgg5_37bi$VSLR)dt(UuS~DP3s-acPq;6PrRhJwE>g^BoiwDc z`6r1`7ZzIh_O~$a&Z1mT+~Z?$p7$D8;Vfaf?k&;O3YdJpBo3R3=r_0B=$SYK zr6>9r<;Qf;3cY#`yC11xEz_KKdpTao>U^-DXge8)#ULMESk?GG?Gwm%gszQj5`uGZ z8GS-XIxZPs-#hVUVsDadI0dDEL;)syfYRliF_}AHavaMHXLnXX+Uup!o!ilQ`*9WY2P^>2ND?HjhVivjGwE_pNIoZ6OJjY$R# z^sdQZ?mNxxHvYpNs=h?D=32!-Q^4nOh{1$4qUZm*SzinVPL|9>^3`E%ADJp0j?my< zL>s>nIZuA;i#Q!<$YbJbZ@s6$g&>BsSNPM$9-ad9b zAKRYV)|e~>F}7MB=JDWeDC($ruQJ&9Za;sa$Z3gl2>=^M3y?Y@faf;@@vmsNn5h8T z58n0<8nUHL((+Z4Yk7^iJhS_VBAlUu5;q6#xz!77)v?ASAobmzkPSUs-j`h5#Y76W zV)MN!A10W%;0(`+d)ze&zfG_E#pK+DiQz;oxqbv&ds5{!_6B0~yCPP=AES7X5w|i` zQ&?szxumK34bu6R8DPHKl*~p<7^{FkpjAMG|CZS zz&~j+`W3Z2aJ>e5!H^GQ<>*f_hiFNva=mjmhkt(1i*d~c3?*Kb_+#YM&}~@QZfI3i zres}X7vQ=UGLGP~7Hl7D!pb*e`J5@onAVn2C4flQI5+r*q#mHr&m~ur=&J(bgcJHY z6uuALZE#M%5&;}aJ;CMU7-Wi#^fe8|0!Od(u|fxB!VZ`s1NH+h4M?r_FZ@pYgf6QG zQNdKmBL=QBPjAh_VDnEJJgEuDWDHetavf{glV*vT!>>M%?mjwLn9Fv4MU3Z=bWhRy zdhRG7)p>>Gb}!(+e#t>^0+;6@g_fWxu6*2WzA$w|)f#ngn@0e>HT_tdTmj~s6+JZy z3%_%!inetp6qYjiDKEqbDY)7v`i?GLYSrD_V5>$;j#mRO_jQHzp{8MFPG;kDs_Shx z5irZ-#h}Y{zZ+WV+e{m-E3bSK??>1H+qlfQRCo!nngvt8E*q>ExppSkz7$gV?BTHR zKDr9_D2$y(ZSHJ!(-qp1!lx1R%4EJTEX%n;qV+_yJMb`)4ehoZT}gssIMr}c*gwyz z3PWHZLthi$F4t9-%h1TSA>7)v`#2+qtGxMyU3n+!b|qgDyTf^%NS8^YFNGkGWwsig zd{*xMbsrg%j`9Nj-LEWawc^`X7dZ(kM{o3Bp`p)-j73mjd++;3sUYAATlaw%Yref; zzsa($U8GN%`?I1rx7v^l--0SfVPu#9n+12xUNS?MWXW8bD1?&*-&S@vyo?hw_oyVz znUK%g2mPGaO*oQDL3&P%OBJ>TcE&CC`!HAbN{A4By_9v@@YSQb!NzgJWhAyC;H3JU zVK57(dP=yt!F|s8$A2XtB@SOVwY#k61eADXOp2`0c~Kg6YckTg`zW!wP-1LcS>0UL z2)^6xnz34BKYkvl?%2=s{w3pVQf_rg%84y~EvefdD@-28=w`ZGo6vYp`WlFS|=d7`vS zW}`5>218a7W{E(#B^;`_%g5jHU6`mRYM`3n#}9GpL9pGzFeRdGp^6li#?|HR?R5>DmZ>cB6x-+BKYi~#3U8!flVLH0Hqu73p!*`So~FjyL2n%_o>6oW#jX;5rGxp52rp_j@P4G(Tuj zNxWLn2tD_F_@&(^l3KoE3M4ShXIkG8!E@{5cb`deH}psY(KtD8HkmQ)3g29s`AL-0 zt5(wVvz!A|O$~ggEL>qy&ik8L7Nr`;(j+00E{7M1jkqant+|1SkjpwdPOZRg5ks1T@}t&q2A6 zUJIS|G0g#b#F-2M`zy>iCZ9Xcu66GqU-DB1Z>}H_Z*yu2x8ugLZ}-4;sD#}e&M3di zWQCcy8y;VkXB7G2Z~)3d6Mvz>%JE7M0VoCfq~0!*egkwzGUNT*2G@_vM|L?ct*ghU z)$>IDnvG_zu1BVRrCR^RO22H7qjw)LIIz8+?(SU08l>dyxpka?kR14|=bRDPb74dOnDLf$p`SZj7xIV(}z|}L`+DffH1FRcCO*bxZICO z%JX)-r4T0Qlm>YfP6~%FB9+S|Tgm;tya~XHg2Rq&7@4$FTb^+Sv}Oe6|d{87)J&N#(Jt_{04&f=E@I-4h?4^xP_b(ATsgXOz%aLP;g&GsVcoLivE zRHBmNO!9m-Cu*93z0EtfESTm7l!4wSst}>3!fH>#$7JP9=aRP`|ErAGB6N6gsk*ZD z4>IB7@7T_h6t((VTlvg*4S&BOJ{eswQg_SZ2hK}gd|_4Ep#61hQ4cHCFGcS3n@CJS zNDkHY1F3{gdAlXXF}*DtvApc+m$X&xLUgRRu7Il1UD8+^7ox9rbU~;!&EM?9raG5x z=BB}NQ}ift*XSR@&U>eKwUUU9IPK<-`#%1IfU32x&t$zTwpXi=hwFUXGiEdj|P@n6`PTyF`QDY_k!NW zr~_iYJWB6?qDbS6Jy)8^7*`GCzbK+4avYo5YMD0O8d!1}#2RX?X3 z5TZHEy;l)ojNaNFGs~MMb|X*C6Pb6@UoAWIXq0e;>ezst>y8AX>IPxCZW=*yxzt51 zDdzIPp}5^7L5Gi32eZ@`SRP9qQ)>i4bcD%9C9ZqVd274R0YY@Aiz?joM*C}4#YSH5 zr1t#|X(?UHN?r7ll z9@?{B7yNfvHX;y~D>MME4BuBIE~uy4kr2B#Ol=slGm<0R1sc}yK+P%_e0wMVj&Z|_ zwX52&4p$=LE6+G9_>#wGiAt(P7CVAOL`mCrqap3*P%?5jdNoluz>nPWhsG~bHz$Kg zvgW2TiF=%WEwRbpSt}E2Q4AJhPa3Fq9~jkIx3? zN()cYtVKirOf&bCCqxZ@HcUV;WQ)U>gZ+BpiBS!;ciiV)HTe*yS{l9+fG)gV_mGWY z+;k$#(y2}sLa-rkKfjc9Ql4G8K!4s12rz>uJ@(J*RWuzx%^OqUQaGtUQ_|{x+(U`C zuc(#G#$8utK_6Wz26HvpEE(h~7T)w&DGCvU-V+^@6C`{Ez*`%D#@T#>(?J^JpY)Tn z&?r=kw&gxM2;mUV&;8GmJ#md0CJum8>2eTuz4i5W6Us*EpXw{NKyFimp(pxl5^$e* z3j#&vx|aD(>`CtFDbiHbyCcwtSX(ENXZFpQ%}PK{2}2RC=qhxu(}SW2hBxTDrOf}- zx1|X7Pw)cBGc2*?P+oRLqQZ$z%_+sp;=9wFh2G0cJ;`ytQ%&&JXW0|_;(zcGK!BqDx)EuIh zga~t%9osmVo3qi!d-?3?Z)gG8lA2tmz_;KNHQYi}^hkRp!(Xtbh;4r{z zf}$sVcd3v$Jf+JzTs*lcF8&2<1qUHMI?!+#4{LJHMEBa+JVCTDRchePi}$u2MT@+Q zwNS%k^Mo{nJ#qS||1+0ib_dQ*rhi5mt5w^TK=2V1AwUt~16y;Kzj!Zd>N_ORq$OXK2v z?BgVnd|H|Gqar`7KOymPvyCNG*UJB~wlKE7?$p>V^bk9A*vEN4ADR+{D)KI-u0`~P9ansapHGc}EO zWsh~VbZ-ZUvPhRCfJyQdyAKeD=H)E3_epd9-O2d)cs4p-iShQvXlZIM5{42iW05`N zq(2;7oPr0IT=#?T%_wY2X|mLX61r(ml2+G)eUF-8VEfCsDUQqAUSe4TjbFwRN{CSS zg%uwIgY}i#Od*E+6`Ns4bl!-AoldPm=rb5BI1OGotpuW~^P;HTTvZ=t)Ayg}KQYn? zgVGO!RdGYEmz51~w$5NwdesoZ9;^WzSz7TI`L#+@RJ~AYv#c%w6c=b* zM(g}G4k8HT;TE-hOgdn7x)3`gDqY?&JA8r{9Kj1 z(t%@kK#D!u&6^ETzmc#|ta1Fbfvh)PO92Y`W`-CI8fdMMia;_`ngk;^wN@-{OWs@J z9I3}DCyWGJ9n}Qm1XH!oOBgOuzt=VpJ%(o}4UuKMT`Qy7>~>ZFRgoI3^?b4%T?MCF z#$bT=MIw%@n-0Rm_D~;7pvWw24@@D**ZR_}v4OUFQAt%ps>z>8oCSQaV5$0#5(}FO z%p=3TI8bVI1Z~X@ru5&d3}qAtrGWzYCD~I?EEj0rQ*ytA-~<4qiA$e|IJu?G(FZabBH~@0LPe{S=Wr_zWClI2^7KabM6+fkE%1T!4|1%O_NOx75a$HF3m!_fSGvJ!8+GCink>NA2NsTTuqiB zjexFwd2nRb`O#w}Zr)V^n_-XgY(EPgoUr+1phtH13*DS-t2ThmY5`IPs4ogSA;YRt z^OIQ$qT!VTPAK%+GdISVn$Mq@s9UayVdV~{IAFZJN5hBXmc60fs17|8T~_1j3$@Mx z0s#P8&@p#_ux|D`1A3K$%GYajZ!-=bbi~H-~?5+{0_lbJ1gZ z4-nBJ;%_2o^KNTcUVL&V$8UMWCcxGco-w1_4yZ;2 zUQ_wbJxbYSZ}He{@&?NpmaGEdTqTgfv&veIxAi8ahH`%jkLb9de30c9#+{| zU=G36y%Xi4Ic?y%vqtVf@df}B1rvNh`&9F0CqdC7hTs3wPAh#{o=^35?`Dc*P;m)> zZPLqU1PtD=%peC?Mcfe6Stivt{O>6d&Mg8!b~wI`8PP%5%Y`6`Yh$*VmmCly!15$XG*zFF5j4Jc^rzC|fl9 zXlx(g>ij_aTpPF{#aDlNzLS1912#y>HpNHH40&r23{Tw3KO2GkP<`%VAGudKB-mj)rc|myB-MAQK|B*o z{dpFl&FAfMOuECsg!4DsHqt%MT1H-<057MS)KcQi=-FFpQ-h3D*bcPoBQg798*%oX zV-N&vA5)03nPOVgfoI2T}G)p>~dvn?|ksNFa#dgSKTd@u;-^`(sdzRZhg*dFYF z9ij4a>pw|PDWSBX^7ZR{uzPN*NinZVQ&0OxT&BZ^UU5;z#<(ynse29q^wTx9a#G~yNy74xT(CHt7=b0mbu{2XPoe5dmx z@u7eFO`zu}#;uM2mTpe#t8U_jV{C}crgb^|jA|uUbN1$sj{S5ICRRDh4g-9BBvQ^g zj^4q}TWg_hRMRur1h@aYic^|ecyP&58foW}2|W@WGGTd$2s8q}Y_v-zus*J}yl-b= zmJqWnf|1Z$#vYUfHV}gOt2Uc)-L*(}MeT9iIao`_rgoQLkGwr6nIlU~pwOKk({rGq zX^->dm;l=focv6>EasN1O>qnCqhTwE4r~VhsMVb9$~*>O-xfWEPWHp5l)uEq`ld`S zkyR?I+Eva~5mYaMY(UvLsGK~DBNgTu7%xuE?Q$ODq5#SrD!|Jk&wyY;EC36`ruAJV z>6eRG8A0b+r^W!os4M-nM~+N-&5@&8;_xe2N_X8#e)8Xl!g-={icA6ux-Ijc|_tW3#r!s(ny)0i7OLt$F9LW=#9$e zog9uV;qqTeOoM7(p%SgFOQS&yG(s#W z4At!X;{x=}nNGlr^70|ViDtV^o%5ZTfI1@`iKPj`BWNIG>4#3x75b5a%UYT25B2y- z8*>hCH3t||=K7OE!9=Gf084tmrB+lvtI$Cv(B>l0_FsgU{aXBaxA@PWNOeHOJ!Y5Y zB*H>8o2BH@ln0E!ei|4QGn|yAL$_v$b8|zi;4;;9y2m&!6r?$Dva$#n$bR-}VZ`5) za_N-x@ocDvI~2=s-3|XGe?si+p)p69KnOT%XBH^ge;($hF#f83A_#`FrF*4G`@Js> zLCMvbNaxn&W}FVeKc$#PHcZ&yEj1!=#C1r`zkNNPooiLv15vC=sO7n=bwP`y_JAX49|L_@)T`q=E^Vk#@%FgwYX9TSc$`j;spmUvmqv;NLp01R5~~0v`(s*>T;f zHX(v3J97(|fU}L>%k_hm&bAgoL-H*-tko~;OsdC|B0nQI&HUl>%MN&B=bERIvo*_& za-Fz~KP@1b*+Q7^y*`uZNd400qlSp-DQUh&&;7nWIi)B;i$dvHemZ|H#9E$VFtyqmL&f=l7Y zua^SK6SHF_{n-H!@wlKo>P+(0iRW>8Ydry?`q-~~#J6ZOICk2e4^s0t4TKMdoPQ@C-sVl7;WLSmc+>@1RWfw zBN&Z~-@V&+hHb8@Q_pEwBob-VHbqM6r;{nlW3 zmPauf7RLdE>`vEpnE<9`t+FC2W_4;JNM4TZF7|Lq@7u+rmqtAVd&4Db5Z1ZFplu%U zjERnqH^74?s*_yjT>=A<7oXXc4;T zQR}#3fb9^$RnY<~S@iIcIL%0A2tuaZa)O> z^xdxTNV+=d9wO59qyKv{+IbK2zaybMD;sIm65N|7}PQ+v2|6hdxAy?>y?7no4 zxvg4#murl7J#OS*Q1diW<5k``ZYpx}!PmL|E0_(YyNYZHZWn*ibs4~4)11gi(v$D` zji-6MPZpq?P0qz-Rs_yIk@_$GNlvNznFC1e(-E2%r{bitpk7VfnV&9fHsZD-7*bgrhp9fBo;v z|DdXid#A)8H&r63B7aV+SdH7HB_>*vziJ903`!D>$ z3~1jy&pF0Wp!2);``g0STD$lDeAAc_4c~!E*1JPQQ>I4=Jb1q@I|;otVJ8vmZmpj4 zR4xY4gq7F^ds{I(&c42V!lYGF*6NjU;rldwsXtQdBFb05W?3fkgfyZ$D7&2V%b2Qj zjpJOyX?Jjh;M-OG>Zy{4mCB;dD<|pmXsa)ji0XS2M3z55ydiR39W4f_lu4PnLJ?|* z`kEQJ(LZVVGP{r(PE9AH9+Lz{1EK@lDc#bYQ?EYf8Z5AgsPrSt(VjszW%)r980rh6a+BrK|t&n4xRdw8w+u`a}ayA_@npQfCYe%=#WB1#Jf8` z`a2Z--KOk#$qW2;7zy>`JwZQ%0&!kA`4CoD>6*|k;6<}R$=~od5>`y9>{;LG7uyyx zyE`Y|JW7?>9flqBvPUp@gkaNJ62-2Z&y@vgOKjf$MwB8V$J_eCqp2)XKdyM${ zD&mqvT!(@hP(+Hxn(aMJu1&Fszl00W&W&lq(pdFDdkD({0P|dHL&wM)k^L{KjaQ|3 zLpnKeI7QZw4h;%K`>p;L%o;7Mv5J8R9(d_r&HiF$RA1#cY?vBe!it$>TRffbV!rf| zV^fj16-2O=VRy9FWy6RC!z0o%?Rz+4Pm8O^%8%*0H-PvdB<)*l>)BhzhgG6bKiM#WfR3A7W*@oM(^XIDHcPoZId(#}L&+ zA)yJ}Ey;mj_O)OZPAp`8pAwvM$_4lIVxF z-g}?^D6u>i(e6A9Wif%P;%hd;_Mno6eigFn)bt`Cabea&WH`#AT~JxFWsiz7-&!{f z6k+1G@J;qARy*B?4rG< z*;u-Rk7`_EPr zfFX)&u4QR~hGTiO(5}vUe%{Q$I$AF2=b;M72_2s2aE!oG#3UzMTDh#?i?4)b%rK`3 zM7Ss6fPEN&%%AQu1IO8X9_}mlc+m4StcKMo*m@v47&}XRl4U$Da}*A%MDK2!G)=Z9T)K*&5eAQ%C-(#}RAI0@5|6`Uh?eQ4yk}10sxa^XvS+7j4N(Wak}5>D<_^ znf;nt2hmy++aIW{;(zQ&?aCv=qfJ&)Avuz%>92?xYYV@ldlp>Fr!y$t!GZ7kL1ECR z^nx>)0Fkq$(Ba;%q18JFH1w_>LaCsa5A%$ZV0_#wMlsKn(hUxKfkpgKG5;hyFNYp0HImaJ6nLNcKRo0ym)Ao_0`BycyP5;gC29h^pqoiHbR_)J&8IOm`;uGev%? z>(XctVJejRyQ0LV96ZLFrdyAz2cBr2c-_fX)d#CSUfmocqG?ec7L_!Jj44jU5G-^u zgd|DZ#Lyf4J4+5&Gw(-K`916=-irxSs+tIKxnI%(%$hV&wMuPfQwkAsn_W7`NOM<2 zKO*FVV^#g9L(i`qHeQ8`{k-j8dmym<_Rfjf9_Q-4K`svVAcLpAh}eP51Co+7Wgmra z5|Y|XmNBL-Vp2@ndbXEBkHlsyab^K2cb z7ju zU^$p$}5q32U>R5`5nSEpLCJ_Wt^$`&>n{o5dqM24x~@ zuBhDt1jpeH0Q>35HvE@{=jJ1cr$`buPK^4Ze)6xB2tvG?=+@Ejy;WL3CwuxZeuAI@ zF{wdXv4dAP6)7^l$kZKRh`}8uF;!q~G>gscDa9loaXM23i^W1ReB(}2n4~7YJTu2# zO8x(#mfyqfC)|wDisy0~&J?t&bGXV_6_LtUQY$hj%C^n3;-yFPn18>(Xq)6`4inFX zh%WB^R^W}lMUP3L^vncWY|9D4GsZn50`MWEl(WIpW zVzT!tQ24|{yENVzd;r_nIhsRb7V(}55Ls_WF62UUyH)Axd0jt9aP=of;@Q$4dM5O4 z>|>YXXW3R;)umflx*#7Bz6X7zz{W258+Uq~RyL#vT6~Ga!;RCp*%$gjmSWjjS2ik2 zTv-6JLc=H1J+*7^(9`-roBECYl0VlFVT)|r(@MpY?{7#Qjvg}$Ri{@3)=$Lk)6wMd zCihKEDqI*)!hx{|NwTx{MSt?Mh4c zAWKWJ|5q`N5{a>&C-p~Xt<>+|VuS*NiF&U0y6e;1PW4*VLyt@M^<$5?Ppk~eMKXSw zIMDIWYM?!>@TNL<6c^8gK=jLiGYQMV37=yU0=aaU13^DCoT2bNa#kejMecK>(-`Mv z21M&~ZMSaiW- zc_k$9IJ_qpF;PQNGKkIP(AK0t-i(9c>==I;yRv@NCo)?}nDJIiemw3%hhf4Yso(w8 zGtITjiva}PEATmVh9a!6j9hd_i)a8Ck4`X5MiBb6&7)6pbjDryTw!tvo$$1-WNiKm zFqh7&SA0(QE8g92VI@Q@|C!|O%^$L7io#sub}!(+e#t>^0xcR_ffse??{{aNd}!C| zaXI^Ty?1Z>%r}0}-W>%GNnQm}E=AaW1g#GMs{=4Ja zs|P3Ti=t=!Q6yC)$r;DSmfm&eUpb8G7uGklPpxp?To&M#H#NosJ;)72fft|$zjAna zKoGGR={X%(?xu(e2%MjUYPMO61k0_+c3MG~VM+Eah2Yoc;05dWp*)S7j6*!uCu+e1={nWcCTDTJ`>WThVLv^lP@sc+KA z#Lhmbl>rj^XG27D38bMx)Mqs`tKTpUkY9NLDf-Yp7^cmOC*WgWmnF8~KzPbZ68PBKV$mfKDo#|1L=Dnm4-p?=Yj)qxjOn#nR zj}1vSX(3WZ(wy-i_uhp+a>Q$JJI{1D^lfIe$P?po z$RKYE&j&H*at6RE;= zS_uZb`ANT|`k+MOrv892DIe|i{HN_SZ8ib_4T(b|?3r#L^fbuwvR+hW%8V#| zY&(>Dp#J$HgNK*ylDX+SS~(c%M|gpthZk|B2|93(5n*1Cu(ij)8#WWV%?eb2%SG%v zSPovXm*oJCSV~N#u&`d*l>W4hPEK+(fE|3Qj;?a>Aa{=~oa6eat#Vxa$Vvt?2#7?C zfAy?I#zi+^jdF;$2Y+#AUM05H0gSWdaWZvTf(6YOs(?VC>hT=B2=*yY;t+g>>zZmlC; zK4@=yCnH}klpbgD-yEnR##vpcLgnE;9^8dA?{oG%SZ`$^I5C-5X$<$xRy?(S0(>)^ zwPz`JfGIpR_};fT3tPNJYsL@sKw$Zg!-O#3)uL`?iEKFmgA^?xD83m+AhJ&=L~g>H zx@8O~duufG9^Z_Z3m%XXNVQwS_mZ-oj9c&#`^BWs;OuhQ>F+fwJ( z$*m%MQLcGFQ*_|XUj~zA@u8#=GaR>hg?8$d7n3YHHBXg~IbHh3N(PmU-;N=>$-m<_ zYfEX7u*JB`Sj)(%^eP}b0EQF95_4^YW7nG({3oSQWR_|8N5RXxy_V0ElsBNf^A}Rp zX@yl7iHPjz`Y`fv9nqvImz2@)^s{RL@v&L|Zv6}rwb@8(T#syPtv)7$3%A>`!KGqr zISCz5-6!Oo0&goN6YR`8uzsnE><;nwGTfY=>dtg$BQuR8D1YlA3JpgoQ|Z4e%K?xr zG<4(Z%J`hJH#cFS%_^3%ZGku6Zdy(!zADOeLOsGUATO)gVwxV!Sy`6)I8e|qt2&(M z!(@`_JwtPgo|Y2~ApUaoZRSf`M*uAd$W%kqPjLy;QCtI#|7mERK~%v|j-0jKhMljn zgt)dq!kXtg&k?T);=&-#YT820{DeOeKqjWiGiWhUW*a9A`razNNs4MOBQeM#9XvfK zMKiyQp?R2-nPf~)ggUZnArgCi}RH4bR<~R7U&F zeA%ApWQUaW35o@g1JBVv;2BrTOPQxoa;o%pI!R3aH;qmbHK<$D22VTrkcRv(45ppM z`iTp_jYC6GDVFo8xF-nJGW7M6IwBJ-6Ec4T#)>{8(g=gkr)L7SlWRMY_3W~U(uI1P zv45MSeSqgI-`M@|_u+U+;ox#wuylMDSZfXp=I5Cqf&Q?!UkO2Z*lFs_dvR)Jjdf$? zfX=7t?H6_XDu)!ThN=}0s%O^gXl)8FS-}nf;y~dj<=eI7`Fg4$@`Z)f6V#;0P`^W8 z$n)qWX7_yh4-R?atdP&lCV!?LVU5jm!XtiQkvQEm43L<2!Q>*>j{Qp8kv7|}^eT|? zj#Vy=Us_qil8Lvrel6D`xeNdf)!%_$qp}5pzWt)O^U9BXaTV+RPCWN;I|%A4i#w;l ziOf>V*XVN{%A52{Y{xWf5(GiT{M4~&bVI41X8j+BSU?frCT@W_%6SGH$J`JK8vj6g z%Ze-GRL0?lFfYy$+|6DAyNY#>LQ%eC6{(E#qS7~I)2Pe~&toU5CUrDZQW==oR%AJg z+JdWM8t?X0NY-_^$gM8OwD(z+J>*qWf)AvRBN%%SFjA4I75vAN6qXQ$S2b1cD~jm+ z5;)KReLr-$p8cL+Ox$n`DI|ex$lYoRYl3?+;qo!Jn4_vVs#7|Kpp~ujT6;nMDb(zf z@#95$!ta}sPXFV)6(C3OF}{eQPy;^UFn41*iN|#oz#J@I?%rg|l=V@&`ycM2D0`v_ zprzPl<31)&C6kwF0h477{Eag}xfTso(S{X*Q)=#RS`Cn)uypa}1-Qb5Iqb-k9aaEt z&O5UkkvD}1I_#d}&#h5J(?1@v>aTod@&g>?imhK{TPH5VmW!Jx>vA9|o`n@OK#zcC zRb9$8ApZ=9>x1c-c{_M3@H5s=r5nF56^FW-U!f;>>iydhp@fR|zdC&QjF^4T) z;H5%#94tQ(`@+nzta`3zGTO$Dr47y+B8orj`M(RQjHw2jz9-`+Z;RQG3GPP*!WXR! zH!jjJ-05?hu6*^JN}l>=)g#0|_T9K5G^(JX^aB|>U9(I`mn#B?(HDY9xM38UnQ3N6 z!Ci7}T6e{=>4}{QDTa$@jH@5=b``ZM!%tRkznNh~Q(i*y3T19&b98cLVQmU!Ze(v_ zY6>wiG$1e_Z(?c+JUj|7Ol59obZ9XkF*Y?b3NK7$ZfA68G9WQGGBGp?FHB`_XLM*Y zATSCqOl59obZ8(kI5RRgARr(hAPO%=X>4?5av(28Y+-a|L}g=dWMv9IJ_>Vma%Ev{ z3V7PIw{v)8-?}Xv+jc6pZL4G3sn~WW>8N9;V;dcJY$qMtwmarc|MofeoW1Y&@AuSG zwQ9Z`Ys@jRXyImS>^& z767&JKMN%X7gq)oW0y}i(B8tz9!T?H4<@8w11NN`JFO83A96O>G=J zTx_fW#`b1_FO2ex00jq+Pooup+QA-R0<<)?H3vAD1Jr<80CiOf6;*(gilVxbDh=aj zVO2LrM+fKs!Xl=srY=Pf5EoHUlK=oU=mApds%n3K)qwV&_$}xG3TmJJzkNOf|E9}J zsEMd)D@m|0{k;bO3&0)d>|*t|?SEjS{6q%$A84Pc=FSdw|1oV}YyDl`D8GHQajv3p# zxp@CCZvOqYO&#oAtXy1O{#6kOFt@S={zdQdch9Ws|FOx7D9A`ksH!o@eG1Q>LEhn$ z9eYMsPuG8<|JD-`m*WL+Gjjr1c-R5VpCXm8HxqNPv-?Et0{53bajQ?BTpgUfnEoGK z+t@pJ*!%pyc5^FxGxNX9o4GkMsoPsQxdCOw|10{_1ouBS3!p22831$w06k4Dnf|u> zM=gKNEPu_PH2C>AIyeH%jcr|kepcqd&kLN7i?KTp;OguK^z-@Oj(=alv2X#*tV~@$ zh5mV1;QpyDV{h&N;Q5#N6Ul$O{!0SX{~T1B&r@pVU~lUMFaw&yF)28>ehPy6|DSXA zAH5{qY;6^c?SRz(N$G#a8QWRedi`7Ee?({k|6-+9aB#LWw*3#Em5Zd6C(umE%GK2J zU&Q{)F5_zaseKW93tQmlru<`4|9hNlKXv;#d#wJBE&u}yGwXl&J|$^tV-Iw30kEpvtQ;Hw zAC^x|n*lxlQ3`;G(cZ!JGX>!2=IRG9cW{RLyPNDh044`};9uiEYYt`rldFfrKkI+T zqxAPY{v$Q!{{i@4gX*8Ps;jeu4N%j{>~l!}uMl}-S7$3vUFJ`nuzcD-KmYwt|KA!= z{Ex){SGK6AgQpJzJ2xAEft809z{0`#`NP7>?)PuGrvI85|7i4+$^TyeJ+J^E&=Y71 zx4htB$`@>%+7w#mFHta64oShocvguesP!ctVmW=P9Un(LcZU=xR2bMCkU?4KAScbM z@1GoKU#vwDjA{G6(UMd(d26ODyld<)?~jipAu^Vu!KfaTA-@z*-bqSx{3Rz}yM%ov zqcy&j2%tW8BF59+^%TuI@qv1iLaJBZx|Rp&Q8dT$3&YtK&2tF_uLFN+ssq&Z;|U`+ z%D7i#nReDFKNM^13zTC|7y3bV>yhA&ga3rmaWi`sp1b(XvCxx=G8VHwZ^W<5j-Rt9@(nw9ByRk|wl4ltSKtwkjUOzSArs^`()_@Yk5GEIY2zX+ z_SQ0(aE17{(Q0s42c2?0rfV?ABfT?2U0Xsu(=vgJqI1}@G-XS<_dLK51ZL;=7|?~} zWHx1-3>ZWA?3AimL@L$81KPn3HcDyZ9va=DiJ&LUY!0$kjZ2$9;$SSblj6%&in~7% z*C+6joR39Lz$c_Pow_QPcj{@Hbp6ve7v5$Rd=m-@RMZ06IM#x&qc^eSvSy_20G$9u zb~+*?Sr}LPA!&Uf%GC{i>-LOdMV*XC_k!z$cPOYBHS%Ur9*>QERYMN`~ zI{peVcH|s@Mge_1MjDFvS9BG;B8)C!xlQx{wy&a>?+68&oUhcoRH*<|<1~~kms(}b z!~0xuS1VkohZ9%aIS93M=UpfHCdvqNBT1Mp`T?M3HoNm|FU1e<_I0t-lW41>Gi1l= z5m&=&gUu!0pyT+0%I4WBL1XGv@8pz7b8ix|w{q9zTbz%?_ofHv98to|4d!=iU$53p zRviB8mWTz>u}#YkQ;UyKHgYI7B^4V|LbfunA0h55=q=+SYDiD8$q8(-Lm)X0VI+G= z+voJeS$L0=@P7AKbh&(dHdLzAV@ZuXD^u++ATvF>?M0{zWDLzXBo;WjQs3FtGq_Ur zqnViO!B=xBjV-I~?{5=iK;ZkXt)~DmK+nHF>a@xdOCYvgBw^$?Zfzf42A;Kyep9Uk zel?9mMrzn)6VZ7rE3Pmb!oX##v#-^k28n=0DJ49m<*CQBL#M#^H;^Z*;DA-Ig|Df( z5pLq=KDb_nb>8yVdzEP=XC)3OWd8$4b2$2enMF==lg$r%;k^&l@R;9ry!Ydh!rW(- z&Z~IDXUlc8Q}n|&!&Ul(TXB-19r(kolGs)%daH*5@x*n3-zI9^ z9d0Hle_}xwHJb$iacpiIlDJA*u)Bi+VEIQP>P@nOgYksdKT zA1scoA}~=#Hg|ncr_N(q2+CtGePcvoJmOrX$T@SX$Ij_0sh)7XV6LY3^uTEG*bMmf ze9=CZ4;)k|+5X+pt^+CZVSp=lhRhq6nH|$Nfl4mYMr-J?6Zk=ysc^Q%(dO{1*aNdCmDazw_8Pd@6g5(+Yti}PhoSwRVmBB>|D%_ss^thI{3qA1E z*CoPFQ;?y#*7$>yP3Ka>pniz+-v zRlNEo_>^mfIvcNLu`=93N7!)(vVUT+llI-ArZG*{na_f5K)7KhhbvpUF$F&>dzS+K z_fFu34?@Gr?`q=y$AtW=gn~U?DEsU~$icu=g3MyWjWYkPSI~x<-N5IP+;859(qIl1u z;P(wwNE8BGIQu#9zVQeYgInK5W#>2bA+~Jpju!)||9VixI(h^(TBQNK5woyo3gH2; z;=Is@M9$@BT2O=3mD5M*Xzv>yjR}7(wbcjT7mDe8qLG#n2IAQuOo9UcLg@CWj5zjF zucS@9a;|if^A~i_Ql?T0)GnMbZf_cn#B1|D-VJs=>lR#+7C}<0Z`2hjyw)a!cfjtL zfrbtKsy5Ye>reqUOpq#>sZxCpv(tK1^ZZ2Udv<|Qo#TMCH^hew^izGy2Hi~;Ybu~@ z?>Vd(9W-P7!c)Xpllvj%dSF zsn%n`FH?7rkZ)lDlRA1LVda4Xt>FomOl6PW)~UjV5bK=DaIpf6VqkHFuu{!m z5Y5KE3r)374O(pgK7x2r_D&NiR$!o~>ci<@_J%oXH0P7~ie$Q78O#ZyZSo#sT2!jt z5Z+)w&Kb)IZ>M$_=gEiALaKKEbWT`l0js~;+}nbd>u0lui}c>|vQSy`NG_|1Pu_A~ zB60b6N@ULe%6nc>`UBLM1nUJBov5n3LYsS9ab@P!Er^7o;d zw!=3ou&A-*D>YVt;+BYH8#>4@%AFIq_)rtjt+0eEHZsyk6GmptmBgnIq;zb(ZY`D<4_ias_q17m^?L zq0S3cg2-qSV{XX<2LZdpP|cgMF_-df-|B~4SuX&@BEx(nf^F?cmI}DRJTbGK>5b8hCvh#}V!?0qy5^QPX2dJ?jF1-uXrLh%77QcfxfLp}tE1VuT zuRdF_zAEy3Jnu^VJZgjy3k_B*ysBX%nD^(W6TN9Hb9iZyg|YRN5{5e(^(sm2E4=Y8 zOOQ9mdJRPwEfA)87Tw&ItB)H@|H(0&fg*5cZhu^jO!0oZ?-R6TxNg`VsabFpKyCgQ ze8F!j{edjfHGAxk%Jsn?ZUDp2&46~uzvtFJHhc!>{+ddZ;Htb@G3i!opfk@Oz2pS< zyWZ0}KBrz|oHBWguAPyq_Ss<-k^HN0cw2@Ow8ROdMId1=bLU+@%39tuu?JE+qZw*S zA+j7#3RX=?p4zgef7AI(4yG6C0sJdRJyxXem}=fE0LA?FJsM2d$)jQc2CX>;@IVl% z!|#Kyl?kt#JB-SMoYkz%m2j$4fA{sH2K>Im)#2oW_DaU;|6|fplDX4R+et5|WMfv1 z2%6xqy$Z%jrW`foy_7HSunLyLy9izoc-sT5T*O9s94CU$Y?9a>-;nKPtLPwJ4&jT(2QvIkyP}Dd?VQE`04^dQR5C0 z!7Id2&FQsjUgQ-G{ft3%I<{@R@>BoW*G6#AG}N0$L1+fnFYE5o%Hd&wZ^u5A&uJKI z3*qVt&(U9*x?TJgtZgM2*45Oef6^SN5zFv{>ty^cP}oBD$faP-I&k||o-m-TzC8tG zDY}@oWiH1Yw9}k>%%X4iR34Hxf+ zDC#B<`q{|5nr_Xr3f#9`=Z#p0ghixXG#AO)j^gqck6En=;-Ks>Z$O5`y;9iuW78s$ z=^#?yaG`6eNsMy{{S`OPWWXvt4ylYM>ci(|%({gTd%%k~HhVM7%+n^yW}$ptLRV87 zQoouuqaV2GoK_6g&)NH>nI-FP|6f!|x7V2a!LlGpTWEUHJ{h5k=r zHqY(Rj@vkACBpcsUjD|Hb#C#Ip)fY39MaA~A8L3>QhMPm7GN zsEx7Jo@81p$4&{J_0q^89oWt`9B4RoG$lakW5x|5<&yZe?9H65!3IW%MSkPEFvBr| zE7lkSL;QQj_H!Woc{G@_N6Qm&)Z=~y7|6LTwum~2q_cEn3tk}HHd24V(p8}c>Z{a0 z>ooZ+M5R)JSh}xSJnY2XM3y2$MVuy;wCwZGSXjwFww8AY_QS~5E@Y%kePuBAU+RoI z!ZaP~5T^=QV$VCOkoeeTS7rLJ-&Lld2J8sHP->4RI{EniT!#tDJap^1o5{zPIlv;g zCSqLU`H*i1|GII8JNc^@228okL#vvkGzbI1}7i)dwUJ9t%1#$ucct z>K@iE{i5h2{p%8@_Rda?hH|p59*^lFuL<5i=8s;{`52bbFb@s1?O~?Hr-H5()g}qC z)4TNzlxD=+r8q{;h1~MA`qEB%>;eTgKR7Cu8PTgA4NID6--M~KN4!ewM)V()ON})Q z?yoONw>J`cZ+LqVi^roK4(JXZiMk}-=1?(ebcF}I6L`FkZQunKo+lw07zvQ-C7sLt zDun4Je|^d6>KL1xl=7x$E@AAYs)IwCFkClzijFEjTafb$wp_8Bq_QYs^`RU?=N0zD zOslt;vt^&XOSF~gYV%;8CI6${#>VPIRwo($OY^o$92J`Bsuao>Kwp(S?2$PC5PQS8 zK@Q2Wu)89b6U8~fgMUrOc-JevM97C_eXz2ihZ8n==jwdp(>;wUO>aOBeI*f1&BGP( ze!qB7c4XeVrFM6rJWSK;kP%*qnH`bW|AiJrUD9Im=VKxEjO0z!MVwNhRrGI-yGUJl z#Z-e38Dkn`=I(Z=<$|_6I5A_GtA%Et56Q#r)Fm}S=a$`3Ck#eLg5J-(=&ZP@;#wlo#bvUZC(~Ez1&ljVAE~?=y7sBrBfU;T7_o8Z$cdt=*Achy^s=x}|FfzGCjlm zEat+7IM4_Y_Y>`@koXwtq8mY3D4T9Z^xEjI57Y!2m!^_V7bUL{r5(@3`^nYlG6~2T zcvM`LqrE{?lb3>sXl>wt3htn)!Ur^weBR7QU$g9QBVM+V&NJ|C3Wb8gR+^-hib!Yq zU=uPy8lvcL&k&GtjvWY(=mni(*fd^hUkT|%>wh5cTZ&sW)8mF7OS1_Q!s~AIIy29t z7^zj@4;1ttbf$2#Mg&&D2}>WosYXd;;1ky!uo#Q<&k7~)*O_8v(FVWT07>V zfsMorl(z{}^q#e;629(B-z4Vb%#ZX|96lV&*WX-v(Hd7a^B|df`18W`$ISV@hAtv( zqp5T$h3}x88}5hZA!|PSKA4!3yyz`daf(lh{P>MNNt}`nxDf~aAj9q1X23ey%y%1? z9au zNkXYy-10g^sJ=4i(z3q8(NJH+OrS#SH{o7G9NG^?y0Dlu5N7bT)JuTbn3dp~U`SEG zk{#%Yqvhqt<(M$<2l3yq0=9~-!UdV_re`T$GDf)?Mg<3Ji_SrN1=^Po7(h;BKcuuh zfqm!y#*>AR*G&ALu-2lP)_6T=_|1IZ!4W!bp! zZtA2Ga-OUsyLRaa9TuG)IY+618j*>1=5@d)zWj|~Pe}aEteRl(LV!mDnG|-Yr)6b5 zpiwktCC2GO!G0{nL0AFO!zP}T`}?I*a*0&oyUNmSa>H&OB-zw#9vJuYlar|s@^M*b z9(=^N4;&;sFE-~QW!mYTB*DX6C2-5h5Z{eD!LefS37{gpTwUsJxnTBH94lYh3C||D zP;41us{jj{DKlJ6usr#iuzPb&e3+UOs8k+b8TT_($9o_Xyial)C_(FL=KxBj2@8qR zL?(EdspA*!sf5uSUH@Rhb{)3~V+mV?f`BWv{&U)87>_pDYf~->1rLGoW8lXItZ-%V z8>pte?jtbzrCp^NIw%ILjnS6w+^!>hF;Nb-hXUQLLX&8jTqMwP?x83YqdJ`XvZH9R zk)l_S-V;|`ii$%po}i4wMsIiC z69Tc~!`v(7z;@0D;?vF^t670+-(dMcj*A8kjj9hS$}V z@w;9mU5(#L8V|fXCJNinz&vGh=;nl6p{9Lu4|%aPZ1=O6HIPyfea8+p(>1fH#uZia z#N(Jd9Mi%zYvkbMbqiL8coPOZ0O6eXiLI>xiS{>)hGrY1L&Ym^E4 z?HY|HtEgqX$y01i>$$E&7mc!?s48y+Bo0< z~fw=j?J z$K;AfGIqT>Da*|X7XB7JnVgyBZ1#O4TxA)YGE#m1;S3aY-nCUE;YpB*9p!-CXgkM& z@QoS|;Y-`rEX4C^Nh1UJbmnRC%Ffoq9&(SSHd?gpXb9&|MhBaVy|Zq$hm~?y!kESG z?+u(#7{3sS@)UexN{c#JcHX}51Gf|2!$j3)vt%MNCu><5qW zqH%?toTm205ulk2dAmWFM)a!YhlTDR2wFjnVX+^-lOEOMZaGc$J1o~XzNb8sgL~=J zVV(du7B&QjNPlLfWineWa{{)R&7RPmoYT5ef|AFcNFIE<%#N)P{qkzMaf)gc) z>3C=X6PcXnavZdd5ja_{YFe)7OZ=_N& zqClRuMefl<9Fx;-tL8=EbGHv3Abg%-*9<26g*5kN@N$Qws_~Gb;wbfxGkvU7hy}|` zJ^NQ>g;XB5>R>JNADkP0ezg|NkVGr5C^Kc>8uKS0A%9a4D2ed!tSYTp=4p-PJLW|g0nlClW+<1;g{&65U z5kv!WxtnBL%2o5unA((^$yo;H`!!ueu6-OBpbeEWrKM*bxtNNC_e(o=MUv-dR^6G# zfT*_ZSWUZ*twta*6JF(9Gst|@D=SwPfHCLDabB=btcREiq9DVGlFygFdbm>BTXuUz z-ZPuir#J=?K`?R(9wkWgiFP|QiFt5C^NJ)$UfA;j9qlE(OWtr?j#>r%i7oy{p)ili z0_x1yH9}Y|pD=dN;}K=MqtVZjdWxf*_I(5U z`MC`)wQ?Q5PP{6`vE@}pwC~IjMvg}8cVWq!%=?3(7TxPbZVif2M`IT49v@(t_Hben z6pgcwe3pfUhN9n7cS=|aZ*ev6%p2m+1%MX8EMb=8*+YHoZ!g}Yv^PjS-_12Sz@vIg z_00r9D{p^v#@~d`aMw8kH94BX{bHa5(FZqzXgdv?pTdWKx>}x6zZoEOJnaY&{{fs{ zr+Ljo+wsG{vaSo5SxNlzUC>K%`He~9)I4S;g;x|DZUU;|5{hj}1ZOT!8g@g;QySj< z3Vh4D9E?JsxR1aYD5v|9yAQtH0^s&xnV?Etf_yzu=DIR&CO`D9*H@sqGjhyH)TSua@7VJn? z?ENC@#1Y8$jaSJqMCRv_Up8?QU#z1??X_0c9%ShwN$*7w3oPe78I@8C$k|-Z?X7Xs zJy4E&r-Fp@Hg;TxCpo2rPOvOiXGL1a>utoA9!5X$T9# z!u)6^2jRuZn{#rUG%p^bMyK7<+&suMVX1ms$?~xyB-R@kb574kU1k8Bg^Y;L7dbU} z)0R|b>4^88#_LnOt;p69*uCd=I7m-sJaABL&^KQC#rfpO5esB4Qk8z>MwGFcS$E}B zXc2ECnZl~51xpzLRfE9?cvi2566xYE-2nsN zq2xWf*fdE#UepVhyxR0Sq2_PekI@qZLD7Pbo-zHijszpCOjfaxwNQzVj$3~BG~D>C z%;YNk96`*waPQl&G7fjZZLHOI9fOV4m?Wqd$rCn*_T##;^ECKqd+*fBg&1ONa^3Gs znp)Zv?{Su#ktIpNypBmqHm~cz+gGo{QOKInBq&ux^$XFeTxbRPF?d>AMUSw(a&s|? zo18QWE7UxfRYARiw3=l4{KvF_xeZE81I_%qa0>-NdGpxN=8aa$S`E~OpzNX?LFsqA z7?|#KBTT4y$4VSE2*#xlLji;I$u1pWIP09rY|d2BBSZ*Y3e?e-Ym+N2j#LjPJHoug zzR--M^<5DRn8TcZf&xfY6}(Z_vx7B8Ysg|$4yVY+W06k2~_MpXo-%cWTYcs`VG+G3N3MSgfY2G@k-%y&;6;uhkj z_*K)Mkf2_fa`q(^`!_i18>AKfB>VPQ{A}@+X!FH!*PC*%?`9j+ia!bHXxa8rzHx@; z@$~ZAC5wfO*2h%sIpa%phhJjotRge;<}htO%&|S;wdX@!YKPk7x&>hFnOM^8JRHpz zDFo9SG&|HcO;vG&8bwGCHxv0#KYN7A4r&|rL%a{(DuSvQI9dYL_u(FT@I!0&vvUCq zfZWnq9~X$_xa@h6=gZ)fo^5SD&VEQ|ix2R?7o;QZJEMU)k|oJcKc_)w*gL)Qqta4Z zREUDRHtXK;j&d5%@ct;Zjvbo0L<<*`0ZfwKvy}2+RCpw zWb(Ke(%wz6Jy4kdB79d_d=W!|f3IjC(AcbyhDrAt=u0+8-i!50(~>J+tj!;6L;OpP zt?tY9{REn$Xu2O4e@Pc+i|e>tNDHlQNJ1y15#?~7t}249^`C2QVpu)!CmdCJk%#8I zA|-%%E4wdv=bsuay<63sHS>4_QdCE0fQe3*l)9){g!~I|u}a}zPp@}Dbg$>l2RVawD5BlAt!@dDE>3J1_n9US{ zCRqrXXMYqchws8T@Q_)hz|$m$%sK$!1(%>I;DsM~6`(1T@FZ6uD1rG_R)@M*WSGd| z^v(!W-y|G+;2#qnGML$sszSJI->+HQ9ltBA%VKaF>OL8agZ47q*hEX0SyE{4^+JW? zYn_1cV1HO-gnq&9tnP}6|zdo(c?dul&0Mnnc$x7I@# z7UO$@6oZ!F>+GD7;7Faya)7COn2mcbz1)kB=3wN<@}yKJOCKkSo@L-S$Ah-ry%Oh* z9U4*jhcpx0q1WbRN4?4@!^pzy2_o3lziEfFzf3i7J9PodM|$zwhwFp6P}vku>sFn^*ptkn`FoZLD_gq=@%9 zVY7e8k-)Xv@4AHofgCW#BNIBl)c(qyOPoI}MUi`s3d+ncGQoZCxJ3dje z_K}K^mKN9DH5fZhx&~>@&{Y+q!CPXNzKn-Ah2`0|tS+_yCLl?2{#F3fu+wO$Ebppt zZ$xEYzpbmQ;ICs^vd|_^?Ce9Tj2nd4S4XbsCaX(kNh$ki^?e1JKjwjzIOO_ZI-R#* zWEe_647Y_GxarR6!P#DaA9LERR??dj(>lX*_1CTaddSl19Z2gXlAlKlheW zDK%j$dJ+1Q{rNykbb`w>&Z3wmTrKUPE1dL07w|RW8#CvGTG~Qo_CglsBi1fRrsPi_ zYX;c1U&z|tr6Pj7Kq2f20XEYpD1$2vD7=>m^ zuxAIHy&L6+G*4mDXIHA=23@tWo^jf~`XP=_noPydQ(SQ&T0D04VT#a|<;A3mj}BaS zOLdfjGAli79`_UEZ@V|6bvKo>d&_SNKpX3ks&nSWh4{?WGz1H4%#n-=B)opglL>;5dXV$|Lm$P-Hl-lj-k2n$(Sh&rs9 zG(`|4m=!-X8pTvms;E)y8@D!}LHniNGL*n$jM3RHZe;*2Z_KK<7P!?>Qgc$A43;XU zSu5*EdUo~~kHjtq+MmNh-|E0lhF6i2Vs=xr7~*6a3Yy+mY5bQb0YBg|VvRrsfDZ;j z+>+v}K$g3jnaUHSX4`-#3lC9GvJ}I(xX06<=0KlHUai85%{2dpoahvPj96}hlv>fo zShS&p3G(S9xu~tX_v)1birQ_Hg-o6tw?M*HZNcZe-zNcFDfv;VI&J9z4$aV3g5TUt zmuUt5z}>-{>}?O+QgM&MHc^k88I!$c(lxhWWVUX2`_MpwX^lk+aXg-N`?UF6-nQ2? z&`g8C>0`9!zRvQ3M{atF%RGV`D030XRYzK2*U_o<-h40bz9@MFX%8KZb~eQf4j;NT zgvAw30l)z@iF_sWH2`E{><(q|oeRI2;7*&r0_P8W@U_+`1jJDS zzk2$Cr-j^@800oKt!mm}ds|8Kk$9wJ;^YvGX$Z)`!r*u2cVa0n8ejbA(na6W`Er9{ znER^X@-lt3l?eWL-(<|zuik7C2T`x@_64UHvtWkzx2KtYxVX>P$ei)<)f`sG)XPJG z)@-E!$AR;Uaw@N?eelap!h8?vgoF5t5cfm-_qbM7WMUx7(slZ8uq_#}R{x>vn7EoD z%qH-?)u6sRuI``J*L}lu(tMS$i|DaG3fUI^^f!oZVeIY~9Wkn`s)BR^EJ|3O3Uk5dzQUh1f4_~oeS#OeZKTJHG zt*^`^V4O7G9793;l@iK)v}vgIDc}v5SP{hUUt>!%?T!&O(|nS-z09HAt!HBW`o>D( zrQm$-Pr-BK^?R12yuyz|t#zqn#X3e9W{gN&0wLSa1(PNw>JIi*rtbwCz=`f!zsH&K z;N0b-jwfN1$sT6YNMO$ae!qzQMiE@hYw0}-8IBESq`9%Z!ghjYy0!0pD7kTJ$d1nm zKh#1orN(X;O*h(M8=GS`4o&u6&Oc`%B|x7gWA*T9O@N-CMDLU6N1HMY&U* z`$wf^M5)1h69|9DO1Gnf=$gibUcvN?a(;>C0uDld+y3oI?huoL<1nxhQ2}m|%iVSL zcTPDCPgNxhlGMvwyoqgz(@s<7{%A${Y>5X1u^T+=PW1}wv2SC+olT}*@s(%jjtif> zoO_`+I>#v9Y;S`yC*hFzpN&5dTy2m*+z&DO6D1a$tNc4Zw&U-Ew`sO~_N^@I0*Uj? z3607}Ef5$>Fao!%DC_}P?4%1%<0VMdJ+s~>AZ^%*6#dtPi%yrd=gQfnDy689Qg0rr z8vbiI*W?MAYkQTm}YrhGWav zjDm*kgIN%7w)JRZZUNc9J4HmE7#iJr+638m`>`4EdXeVgJFq zlcR24c46tx{;OP|{Dbz0349&=uG!NMt5`h@=$o$%#bXfYUEh0>H4%PkweA*(x6yIq zDX)tSHi((NOdKwE6+(ab(-G+>IXy|H0;*pu_$nP<{FeW)$1`&u5 zLFRb73ZJ}@jF4qhmxtPpJ)aCe-<8B~NN+DbJd)?frB zD0;_`HB{ODC#pZsrL0Od50UtDO0^5xZhWMCqrI&6gk^f@#t{ z(gBA;09P#TXd(~tn08l47MHVHOW^30Ii3MYhd?t_9 z2uTw#^cm^4-VfoVf-j&;78vwVGzDUsh{@v>e|mcBA(7wB#vqK)y;OmA<=->|S}eD0L>h?}{Ne zR|LcbX=2KLSXqPb*$ke=e>KcPUsfSFkvUOBWjvcO$n0-qOod5}n~Hru@xu@p_vB=1 zYEMcDuSbSg6q@}xGMf9v%U&F+UD0w+9^j+Vz8NL`0_&F^fxCI=a> zydlBNy`#GGv{S(9s;EyW@6f^c?2Pk300}-@S=PX|<%Ndd*p#rKo8eJPxmzs;$53%Z z2@~_SR1blI{CMc>Ak*=dRi3lX>9}6)Ln$&cFQe-qXwx%|6k(q<#%o~Sb`hq2nJGS0 zX#{vt018+O7>Vf%L>;)RcD&C>f7toUutO=s#Rxl|+>?y`fMc;5Kz(-PE;E~MDUYs8 zCXZuRMA9P*umwnH>>Z;`z9n!Q`4!-4Wt99J5Q*W@sa zj7Ii)1BvYkVh)_>x! znwZ6Dd=|y)d~cGFYyvw>g0I?IipMkYG&2Vot9oYbOhk+tIWtd0QfK?%b)mkpP(?dM z33C5U;^!&FAh8sS_8SIolH811QFk{`X(HS9_Z*qv(0ub|RU(U`+3GM(!?#IleX;!Ca6wV_KnM)Wo0{VN$mwL~J+H(Gdwb8wC4`^#DFTCBlS0b_iM(iNg2m~a)H9rBG6@}2I;+v2Sh z#&H+o435FLpmtRF@B%7UF!;RzF0@7UVx1zaO3a8^1W(Kpi03q+Erewu1(QB^5Q3Fb zX+2qOvsn!9doz{I6me@g1^L6Qz)s%{Sc`SdKsv~#IQ_a5r>-cuF?LZG;l+ah<_BP$ zvjI&jM6xXX%Rx5Q@Df(a<4V|Yme+bE2;$=No=Y}*;G4Sn^B)tKaP+woI6M9 zQtAG;*=ApkL|%?fm`1eTs$DF5uU*y!Q9l7;=TdLcu27_Y5EB#6M%^m3ylJ9cwByi#9$Z0b?I2SWDT_c4m!K&fm>l{G z&|wz*0R%GVefOIBEKSbS@2H^Q4B1E|a(f3$%1&{5_qIB$o}6K~^XkdioDcBqMItVi zFb1KrG#K8bs;3*bZXMo{JV<1i!D>4R>PNxWwH4XovO$#~h`R)&h_ytimD5`N`9xccxF67%gV;%N~l=5j|)HG+Q%Y2!+ zb0e)BXwqRw_UKc`l@9tBdMZ|OB{pD6N}Z>@U)r(D1f2d9%#tXFA`SF@k_v`gwN<4jpnkT=^8h6#xS|$lVPa~p>T+14)0+xu-l&xQG$T9 z-Iwr9{Akn9Nm6L`#ULem;q@=TZ8$+1a+x2&Fg&?=_1#tn&>2ug#QtzH9i3?~0*&cUvu@R4$O&cwLeM)B-1`={vquZ?k*MR<$G*B*g zowN}OkE)Si3#Ub6ZYxO@#rOm_=@>N)5pTn-I&fVt{uj$*HfVeoS@pzpPE1PZcRv={ zu7ug^(T@I@cnvg`pwbZzZOqXNg@DWN_UXNr>U3b{l^bJd(LW`FlA80LNLsJXJ(qIW zuEik)kS{pf?sg!BYnSY9>&^nf_;j;5bb-IYA)J#kDcMUjTTn2a5e4o+~i<7+{8qbS?Py*Z6w=;Uifo1N$4EiD<4Hq zQDeOoznZ$Xq5c3`%=4v=uj1>pASFMA47p$QgPf)593PlT?53>)B~{W&_V|}?n5!QU zUr$~6V*pIqXaTw+0cd_avR8E?*f~+>WpK~CRqi9fJT%kbKWGwT;&tl)MT4rRIUCQ5 zBh5R>qwi5oVDiG-vBcv(m4;#LoKS33q8bg1-T3*#I9ci%ztANrl|N3HH8DD0$EI@( z-U5S6;sSBAU{^g1m=3+Q4%j5j0(Os+@*Cp;NmUHWt`TELnSOU zoW(bD=7g|Pe~COY<18;2RI#^f8i9oL&XTskx>+l^{x}iwRxiVyYjKg!`SpWV8JTem zwNJNfvB|8nL_ER<>g#cPIUgDCz@j$>Md9%5XE#A+-);?}QNMn-^fA*ty+b29b6Z1BUv?D~He7_qJ45WaJAF^ggf^jEC^^_;O1!ir630+wm z^5H30ODhe&%@`wIriM4FHcw~V{}8r1$jf0e`EybjzRYx#>ceoGinPTK6p#6QG?f*< zzNAG5ck@m+sHv4k?DISXG&<1}@pvHcu;fqm3ddq|tQ_;J}xoUZK61_e)@2DeZCjFH5nOViDwR_51MSEM~AXi)A*} zS*#@}U}pkkb+_&Z22UlSe`SUb8RHv4w_l9Jt%QTAfz%YEr#XWqrg9H)C?o{CHO)8% z=_1{PVw*xaZs!qb&HE;ImfG1GF;UG;yEx7eXHy}#V}T?WMWvIH5CzA-veH=mz{3y z!`X52_gN86eQkbi;^;qf=xliY<}};?UQ|(~be!-sT#uaGu{*~(yxTDPPIlD!lo_27 zDp4e>{6o7`Nyxf+0QstZ$zhdZe}aE{doDuVf0=)wluvhM9-CrWBB>0KWp?yF^S$iOVjWz_6eA;|u*VLHO`2PfA z2A%o$fJL2|GOtK0UL#TCy4&ygK(d(ncTSl>$wH&v_*4Bz{mJ6;r{qS&FAO;_l2B(5Ac~5AqnKjwFSIRu2%WmlEWF-E-O+3t?7oeunqqNAbbs8P-99=q75J2 z_OwbRQo;_L7xD8)jc8w&EZXXkwCvdBjVhq2rvd}uUncOg^E51IKspwtQkikf%LW!; zFz)$-?=vKqQ4!gLjV>7yO*TNvIvln!ucaCzXMFH)xE%0RB0DLn3c~71eteX{*S_Je z5pJNcqy@dajE*lRG>`|3Xsgp^68gj17H1poYxTN<1~^)xyR z;LvWhoYzudB~@d{jaf^U7bn1s!~|Qf%PTAlhnSm**N|40)%OijfLB&~^@Nw~>7$WG zcRnqy{dDumVQhZOTYd>0 z8ximKKMP)ZS4haOwO2(RX%O>JxbKA^K257Av0WJc9;BB`{ZaE+AFH3c;!yRbN&oa+ zqR7XN<5#yZ!TFb+9<6bC48OFFTQy_}8Fq3rk*|{6K0`w+v-!8*h~kW3uw)PBf8$H% zcTEF=^fX2Zwu)ShAxkGUnXOQ1iZr^}vOWeTeVVXH4%qTT$EQga02SwRN7x~QQ(;9B z$xQ-X{gfr{_m$lbPTwKqZrXG_<=4EXTH>5f0P_1F^37*{g|~4)0{YerxI=jy{Gco( zMo?&I-=wSFgL_pLQ0R#$R6?q`qMa4HW>j#$cA`-Y$4SkmLU-ana%SDINA)N;4k`eM zx##01FR(V@^&2L?%EOUaGOjqURW4eXgP1Ee0u;p`dnb0>jrXCsG!GabK!*n&*tDa1 zicP+2>1PP0wKpN67|1c7mG}dQNWn6}6QF39aYsfHM&rT|BKa96HWy>KZ=iwcjsBCy z0}19w`H7MO5*OBH%F<+p{fbT>KqKB2zL!$%Jjpu_QzgBFq&2{}fhHd6ElS!JTxSOn7%zYK^k=}My@7z!SxW)tw zcQ!X~X{d_ngRuAN4~C8l9vJ!yVxEI1rjDWp0j)-P({Hg)@{3jDP129~@(hyVSL%hX zvATBLi}PSz2aM1*PCcu*hO^K5lb^8rvqTh)97MF6K4X~S1sxxSVUES@mSkUQ{XF1n z#BqC}n09(KF4>Zq`430Rp@^G(WSF4v!&UiCl4X)bQBSmh1_7eeY3bdb-kQNqPtcZ? zyW7{9PuQnE)^r}5Z48Q5ivE^P5(uyWh3GKoVEZRIK9t=(@B@*rX=*N58pV~BPsYn0 zQKv%`VFtPu#2^UagLkU3-x`U*PP50-#{^g#(}~o=OhyBB|7o*L8156ha__PYuZk%D z6cFA0|IC}O+#^|9Str_Jueu@3M!PZxXKR*eU6CKTo4U?_Y)a$R5eVZjZUVza^t5E~ zUmmLh&xhGxK@XYjw<|vo<2AsvP>tREuqIrvzu93D@_(C5+N`htA@2 z5M1UriZ0m`@l7DYa?wCwQ#fEqfQaA*|CQi>9KlVDA~HYk#EgUp41=M85d_zz}gwUvF|Z_UoJD2R-b!r zrXlE3JHm&${f%V*i_)pnWHRKPGXta^q^}BH;04*p(T2@kuB|%dEfL4*5J$#&ZLrl6 zGBl7Xp;v*!@#9{Kq*<74#WsA!y)J-AxGm3y1R4dpuEygv5>99#M(d(5CSTI~!1p-NUjdOcWj9W81cE+qP}nwrv~t*tTuk zwype0RZ4k*>3^#~Oax3E4DA1h2l~G~ASO;0 z4yONK52)Q#MQ5v(HL9yi8N~g+0|6%p1*y;u28&0~-X-h+af`OSv%}i{l<0fg;hpci zx3^|de=huW?L6I4iAtIo#Z_FMnG2~f33og_J~=!HuArdO)Wiz*z)%TDt78*`3uq%l zLxTg6l5!X)U~Ep!oe( z&CO13O&!2E4nr_jG7h59=%GrNFw|LX#>3GjzB3TpC7SJw2uKp-0^fq_XB2q&Ol z$)v`w^rctb$+r^suBSgFZ}h>B|7aNhP~e<^{yz(2%j;+L+MuW?p#If$sACuq%wU;4 z;M~|;-5h}Y-^QLlpqcVNQ3MDCcSo1+CE#>kwvR*G>(@PQGZ-hw z$M0~qzYZH^?u6vjj9HMV~K>)hY#Ow@Dw?4t6d>XedBu;Rh5{Xh)Vo81*a zM$RtoKh&SHyz&qr{DAOK=>EwmKm$*FCCP-$_S~8qZfo)o{zplTo~vVki)b!O|JY8;(_Mx;OFA3LG-MPhBv(`qwwtfTyEuasu|$&ih}$2hqt5)Cl|6=RnWDkMd*wL?CK*0hy^4PhSOn z_-G-0l~X~}Hvrv#OnuBBZCCzc0jG7rQeS>jGuHxacmT`=oIp)UxldmL^WT5Y|BWru zQ=(&IlNx9fPWwoI?l_FKsj=h#BEIV&X!r4JS)|}NM$IR>>Dj6hmGhO|kgT*FC7tl~#+yD+V^+izc=mgYT z-#gcgIsa293CO@4c%8GS4Dd=fy9?mVEgr((H8BEdfc9hjiGB#i5d9bH0Wgj97u~K2 zFs$EUzuft#r@oA)e*rZCr~Cc;-IGrJ(S67* z?M^=1@}J_jW=9_IA@2y${sQh`Y5(m0=u%6=dq@$U<0IU&SNQk2&9nXc-#|BhVKw(A zesiZjYC+C@Lw*nZ&F$aZyWUK?{v7(2TK=|vdR1Qq#HqXb(&v8V)&@pyf8ba?z#&)N zulUo(KWEo3?x^laa%^I9Y5oxJ_0~V^)9!L|XmJDo6mEP`XYJhnBs}U49^v0-w{Gmm zU-52#;otS&i!}G{0&{r#ks8u(o#3R%p?6n zD6qJ>wK}dEy7kSg?_S^X{r}AG@nI?cdL#c5Y01paenS(Feju0{9~uEQICsxpbUozm z{>E8<)~asms-A!P|FHi&>i_}-+zTiN4|8%U{rAsGKKGIEC%Pl?=To|6$W1wfHk}w1 zqAQSVLOhT*wDAK5KYnBVELA?-pqPBX_#l4^gls1TY_iu{{=GjwSTn08+8X1F;tQcj zE*w@rX3W_7v)q`OW_=bSDAd zs4)y}2Ao&~AM^%@&5N}&EP&m9zXqV98gA2eO0fS>vSXNQz>W4+0Xz3i&m#$Q962Ik zzo06US7+;`>^A!%I8SH@5kq4^m*L|(CC6^%_8Z`l2K?Z0LW`zRlk{ffV^Secr5>y5 zkQe0Bg~7TL;FWlM7n)>x(&_YEkcUv$K8^C7Hsm}S=u6<85Xk9xR!wjvq^;K%(gNtU zhCAcNkr8KMVh0O`LY>6uc?ox4ABec4NIHACNtCT*vZmf=3xWyM75cy+P7lE$dvddR z?C>8D!5@xL{xluR0CnEMZce)Z+TJTsN9}HwKgYo_gH{k-Y&--xkOz-$C9C1R<3-__ zs>)iHBv$*i${j$SZpjI>Fqpdz3h7bEFfT(41H>e2ba4>mFGy({XUd9s@jaf`hkxxl z8Hc8XX%yyp*KVqrs}D_7s|vWb7Pn73hkPl2gT%NKU6oDJ<8;;yDvo5DmBhK7u@Ql3 zXFYk#`lJO1k@)GS5magfL2_+NrZQON(mB(9EvWJR<1Tegw`?khSH>M@@TMrsL_%(t z6&vk=gxYnj4S+JfIBb0Nk6>>MaWo_pUQg=}=22Sgs9g2I&(3Y;cZi!nn^w-KmlIv0-VPVDy*Xu$eX2S-?hDL?N^$>ULzK&EQ5{z(>`A z>B3*i`t_;C_l$jw&1A(BovSTz`xShkboYmkoL>i+ki|^z$w#n6C}kMZ<|9KJPz6yy zxyj6q3r@3j*$5~FO-6}FU{|H?%&&D~XU$k!Q4?9Uh|3=fmvZXD9I(W}k3!V)R}NBXL@|g1cmG1xPYcz^pYD^%1h zfcBw2U;%Ql!uihvPL(OiyE$BEzGH^%2z(Bt)hlTxT`aIcK{M4}XoJ2q{Fu;HrA(rt zp_&Tt%8PuToS2c!d7i@nZ7c=kfM^MS@nJW>D&PWj42-h7;dmj;_UqcL{JL%TB5%@( zgPP&jmaoK5^uIOb`)xF_VUO{g*fcp#Q%0^AJdbPi9^(^|FMMxEGA{F*s;mi5X35Z%xICgV&@?Ww;~J)?&oy|hF}OITJ4U(`*E;U83*4-Pt>>k zpgbs3t+XU*%iDTQPUjgj`4s+KL8M^JzqZ(yhQ#?lQ+`JHUo3H;OB^@pZNNV}8f|N0 zLzX1gc473!76zghOSUp$chHgSM9u^1{zasB#u&~(Vpirw$%9iMSJ6|cuCA4>`vILd zb%}dC?45-P?)Xyl{&3(2W<#MuPq>|Jx>seA6XwVBLg1?;+z71uTsV#Cqi~KtSpcF` z6+wl5T&-CA_2$5Vlu=s>@e;W^#iu4$Y=m9l(QT=j`od$A`o-gF5L{J%MxLhzzQfg~ z`BRMr_^QKwx!_BgVfF?h37mxl1r_y|uRc-^Kt~4HI)*Gm=K<^wNyM|=J z%jwPt7c1?-ENfpDtg~;1QZ+>z+d67Sn2mm*(ldGYt8Y*rwo}2EwW-!n`Meyn_f4w$ z_w1DtMubM;o;9pCX)gy|-l=CCNTD8NZpZ7G(|ZHkFm)97V`H+__}EynT3!*embEo1{JRJQBCzA3$AP(3^f8^0VryftAug&g?Jq_jRu z_mpJku%^#?6*ZGOj{)+T5{%YlLV?9E-N07m`;p<61Qv&~moPaBN!&tpNL7nDsW2gh z_kyJ2oTeXfZTRFW;e)iw8{eZ9H{~o6bQgz={{#OQr$#6q&A~Ps8oKmIjnF+cWV(7O zSIZ<>94zEU3ZLo-VzXJ5$+702RKQAM;_0_QP8a&!TBTyo?v;sI{H!UCHbYgPf(w;F zM{q%0CSPPSbN=G08vMdOmNVLJqZNOmycH;i!E%5xrd*iPuZ+(9URhjgR9Jvlj)SJX zcRb2dA@cmP(hyR}VADB}EBMOk0u)eWxac%$W8dok8wy#B7QmwH5WrGEm`P9{UmZay zP~cJ2MWF-vjv! z(TB==A(6kw43+`$4p*2;NK{u{?D9e_1{U*D0fP@aABy!wb!J1*W;la32|W);=)|w9 zUsAAaLyHM`D|U|%SiT0~!{_x*gTIZJj2JAh|yBOcUeMLUHx5 zKsfMADuO8fW+P!-U$;DN#a*1H4VXOe1_&kGYZ|>eHMe~Oo@DgCm)`{I|6=ouQ@Fw3 znJbfKc-~bZ(=(hn`=F~;-_LYVO+2^{wkh&u0QsFK8HOP5#u={~JjIm-=T_z>+7h2A zH9AI*OmDL7&zAOH^Bm!T%`rmbUefJbe*vhtkCE!N+eCA$?=CZve@@0J9e_#_w zlF1oG%EI@UM9POqQgt&&d}>HCf-WyE)_q%bndoeZRZM16m`}(}qVZpRoFG!jNfp~h-E9z4T_D)*+(kN8d--rH zbp$+NNkM7Iyf(ReJIg870so3l39kHw%qLlm>!bFpWAjPd3IJW2^hap1G2rd+klB#x zBLMlTKut@{Pw+s1t8y$}h{hrfyKTL63UeT3$ara8VC$M{@JOaxtzr^iZwuqLYB7cf z^Ll~Dyp2kY=<3Mj3cXE}-Op-x0*h+`5si8%km_Lq2g|e9)JOuZ7>k>yqGW22#{SC0 zWScGe^*uYH7SXp8y?IcZPc8AW&O@-i-4(=CVurk;(Ug=Q36ZpF+0eL@Zc$MZ--d?6QzwSuW#dEm zpC|n1w^u4yw3IP}yAOq-{Mi!S%RR)ugMW(?hNP+O(~j0(v4>*B$N@BAeFGA}Yetx{ zv77l%GNO^}ithQ>>}ms!r84BwctI+SYMDBC(zrPmvKAIr+b@uV6&#W>v$h3| zgDS72QY(dttqxoZm9!yaEq(bku}YOq8K?*pv22gou(eCm%bJDRb7rL&*$US~Ht!=@ z;fGKlh>k_`<>Pa@;$AT9`dR19jEJeThos&s?Q$a%C9J*;J;PuURdt>dkoGIzUPTNY zR{{ovnB7yad_-X}mI$p`2nf)$$0!<0TTx^eV962PigAbP#x(iOEHB&Z)W4s^u)1Q5 zL$k!gT+fstDbxm<_u-NLY}(n@unWhczfB$+Q^MIe+s zy6@4sA9mHeTbx73cO;|UC)_9941w^)JVmZG;X?W zR+vKGt`(6R6=-4J{ijrUS3l!RX1(hy(5`HU(x;Ll)6 z6J28-?Vae7+0$?oQDDl)7e~Ekb8T6Htpnfl{#Xe`il|0%lvVLko14YZShwm|(uspq z>cSC^vw4GeBy%%sP$tefA8_DMHsJkG;QV zKHt_42(}S=x|4+r?z34bex1{1LE0G19EFE^6(=|O&K`WtcGD#1r@Y_Hc4WN<4h^oX7UY7- z=gUufM>&|%QlZ+eZJH*1{7YMC@FPTxBKZlLkfeDCiRIK0W_OSbrq}S%#&DvNJ!kbp ze~)E?V5Jujo7rCRW)K#TK}P-?7M=1IfS4{Fgpl*nYN5cCV|Au!YlUyKk z+%yVxPzB#8zo`!Kc7Je3Cl4bLVl?%4b1rE4e(C@qMzyGf^rL5D@{1PQg(nq%>z(4F z4R}QmH_9v-jyruTwqNsdl|r}sopbn=C*6yF&~}a)dTUH3=0iIOYWiqQ$}fu1^9%WO zI4F9f=!@6ui#rcoc)^n9{)2^V>$&4oh<{ye3-#N-2ggwBQ*%j$ZL?#xhYc;EXIq;Q zoSl89>X>SzV4&FgW=>kUsk_VnG%Tk>YdPoAWGh!yf$u+U*%rxKnIRZz z+{Ww)8jmnty&Bi6*R-xf2Au>)kscL7i%6X?q^lJd?wiH3VDulE{V>2|W76inUAb z(&3&zL)mhd99!mNqXt!XSO9pX)37M)0H5%CGW4QkaCJx%6xyp1W(jkfASL~AdW{ZY z?;%Yt(LO~SI_c!@a3)HQXp`Ww>cXddNeWdl1cW+R!}e^0G6MnRO=9ON%nS2}m2`!l z_+;Y$>Yb42|9ybgEIh&Ci2+2uX{n| z-la)9gtgli#9x!+ma{Isp%Gsq*01UT%G)MgpC_I@EQ&(O4WZd*RAj&dlQ%L(4imI3 z1*CsaYDNDqS5hg{#>!*X9*wlGTiAt30RrZg}q)Tf?*Tpf#` zquRPT<0%LS)N12Esu*5cN8;+2j?5tlDiIZR(H6@S5wE&y&DI%gt|U-0x{{+=yc<*E zMQuQE`&QQXzYoZUAm>On*tYSw&JAqTB@1w>x!C5aFwH4?X~{993bk;PDnP1aiJImYlx8mO7T6X~>5V(gYLF|qCbPgctIrjG|j{0mz>w8zdI ziDg;Gt4_A;L)3dJiAIT1ffMg9V(Jsbrs*0t2@t}(Stg6wQ(4G`%YTI$V#MeVkI195P9 z^bRJz(|9M1yiATQS8P-=-TMx^!>v={EwG#6H6W{-1Y5A>_cA<|QA*+(I{Gn8CAyoG zPZFNAoKq0n!i%dP^`V!lf!=A`OKI1?D`_(V9gH4AmZR^y?M%%wo**Ti**6Hc%~py* zP}KTh0|p)W+SaH1T`*jR_u~Q(XTlB8Ncnbv*1vtRbl}IqXDnWXtWD+5_`M$scMMJ@ zhAh5gk2ZRkkQAG3pW5`9*-6`}azMQ_!`A&*g3Qr?gbbM1-rx4I@mNem4Bu(Ahz%%p zg2OtxUHr)2k{uaYLuN=s6URXJdU_JLt9HQM;T<~%b!Gi8B^W7YRVEBc&iY?pFWZ)N z-$GT;tCQWusqh)JO`jd-=sNkHSCkr~ou{Y4PcH=^r+hf3dR#~ddSLL^;IZ?19K6&+ zlv3`dDK92ccjoRr7**t=;ak@p;_@bTDDj(lja7#DK^wMTW8h4E+hl}FiLa9mpyzpP z2{F3O2=J#1tWE*y|bORtmxHm1ZL`#9d)C7A|T z$FY3q_)5wASQC8+Xb#VA#Mec3;!r88bU}DN+hU8Yv9-BW{hevjt*Gioi>14diMvjo z7>?Vsktu~YkI#Uz9#?UlaTEN3VJV>7j^o*TL$Uvonhms~DyJOi2(X?VLN|Nw=c(4w zx@f2`%`eIE*%yH(v9gfDxS6|c}2q&20z z3bQPDQc_j9>O#+i@fBmSlxTmkMG3C!4jpsAgbe@49|hjP5VJpj zG3Mgc*l$2q)dU39dh0@e_cR#|;;H~ht!2OP$az=oCs-Fc0dz&xiNL_NWWC4gOk{It)8yrSk^8vl6MbY&8;zASil1^ut z$a~Bukbwid)gS7R9CXn{80m;0JU38j6G6aqhmpHwA964KqF^Bg^6_ow$Ot#S^CLPC z+V{o?^FlMAs_+wm^?c@)P5v5*I`3L{;!l3oJ;)lnHR%)G>Y&?hUQ9g<>@U|o*+@KY zuclM(&FC5_8xqEf$aOhfRhUUnUyqgn-Bw|#B!i`_Q-?@=CYbSXI00v`}d^x_z{D+5p9MDY=YxClJR+ux-`Y z62W*UMdiQM@3~pl@n(3>MEr7$LU4XG3mU7oNi3yN{QmInejZ`kG37WWleP96C)$z; z@<)_(?jaroZZ*KIMD0WbcbZ8{o$`FqmPv?w(X2#^UQlYZhJMaJ3WJ>@Q}I}bD1%^yx{iXw+a2AKCOJQ2VRS{fowl?cWq@6c988 zJO93(dfTfs8CPi*8FhF({(I(smra3wUkhnA5Cil$t_|we)^LlyNSx@_(+bne^$r9L z^51JGe%P?)O-tl|T{6IP%X?K5_uZc!{Bj2}ws7(-isc%PHRbNmUE}-V_eQ*rGQ&0B zm*za=eWr9sw`oNuVloHe$-ZGR5Q#N9)a*O_%reR8lZsaivDbG4Kh63}VhZSevDXKO z{>i=q3a;vIJL*&@y3|kD=ToP)FE?A@qdNwVSz-YH?FzEdDX9mUzKRXwkqZkG-PD_@ zVm8SUjLQ$*pi}9V1nl+EL(3U@l)+!uZ%rM1?5m{k&9ja-P71@gvT%d*EK)tMaVpd( zNxqN}HJ^Uq&)hY5>byFzP=7oip=@D{V_?$pyxS}e)j+TTJM6S6%#l4%Utpd1M+G+j z6CS{FQ*aQ;;U+xF&#(n51_H;_(H$q!!b#16jRL_LW1V07w%55>Rk6XyT;-H1T96mZu%{-l$Y?>3kMCa3Ms+ z66~47m4@2PM2JUoxoA4N^$5dssQbg#^VA_lS&TPxU40;(b$`V+bZ3{@)6a|la$k<$ z)ak5DnM?XB#=LiI?@q( zfO3`yIz0XG!L}z8D)pU@_7f)vlp|HfburWKCIN~xLVv1yMO|Jqz;_^)hSdJn)v^`8 zWjchr06BFCvGr$G^u zpeJg=Hd1Z{5wEpca_O%#vznfr-YxYqQx;|R5GGhA;Mkxh6U%*X_~}9*9?4Q_C;m1F z$4oT)_TAG6A>CYgEOJZq%Q93z)CYZUxAR=8x?AJsq17dYQZ#izHz=ys1|0OVab=j+ z#@^sA56TIecsRA6GJ|++|488R6wa#Lb%U{MaTAY{u5jX)lfc5@SEO5h|A0c^MHu>v zT&+taX1iKAP{ut|zDr>Ni>#|)MRvrBlI+9;7tixr0n9oDKBFJ~arnCvn+H<&sp0t; zz$>|Qjz8WKUB%wWghDzU7jMcR3zU^ma6RS#Q%%M1u4Q~%hv1eIhE~f4kseet`&Ov6 zn<_BgWT2rl82G8ZPAtu<2EV=nly6K#a$8S{Z&ldRcROb&N@^=8@rbP{L70}wDyK%s z&ce;#YgKx^`JHt)McUIH_5npJhtSSnyve}4K82>480emnay*Jt9);O!APTWvoQUa! zuu{OCA?u^Ewoa{Rvxr42sr&9*0}^HGHwheG_g?xqlNnHUEZW}1{mOc^`#8Xd#WpnJ zA(kD@HWU=pstSKFHuUmJ(-%r5!DIoT?{(o28Q1KQVx7?z0F+H1lq}3-g+fTfKi8XQ z8otU@V@FBFkO^^hECF0Bv{24z9`kofhIEeu7O8>@7J4374lLZ#jGf z!lGbm2uu8j?t2=(?vHLlZm%+sTHqcoSIFYbiq8z}V|loEJv14!W7-Vg)KbCJopQjB z^(`sO#u9Mc>r4($M%$ZX`RJ50X?NN#?5}eUGW{0giG{%Xe;v5aSlF!C_aHc)jENxzbkKYfM5UDnh2=H9;sZ%zh4y(|g)tkx zNF0GGu!UnG^P^Kx7x``u5EG(lfhW^3KBl>6ED~HpdN?c_%kWPdRb&!S4R%Za^wIWx zTx_n+D*0LDN>Il*)XwfF67D3O!Y^#j$cO7`U&q7J?Spj>5}q*3L@SSQVW^OA2D(Un zSlR{(EGs)?N(NLe&^UlOmuD58nL|~J2S-2B>U>#T4@PL{wrYppnkz9T)B42rQk*>P3`gh^t2%jw zevmha;>u0_%eTM4R?TAy>{ARbIJSp;ElrqcD=0JOnB-o7X#H2lF+>j!{Z^R) z;lyi4nOdsB9raMl1TL-`y0`Wp)bfa#?9Mu9uQzRpcXmn>G58QS4G|KBPcM@&mq1iU zV|T8U@V#Y8ACFy(&D1%bVhuHl%|+!F^PuMH@y33Ek{%6mf~LX-mT{Ujc@}0DdI?J? zl$}SH91gf%r=c zY1ifq9~!cHutY2fRKhd-KOf2ntLsOskE(&8C)PlUFy&-NPL8Bq_K)>WTkAs-=i{-? zEIZht=T^wCTlP9EK09ULHt*q;&01e^(@nF<5dAb?3zUW2@TK}j7Jqa#G%g;xLno_te zo-4#e$`c4VH2QI5d)2tkiJtS2fq{o<*(jmR!Pe%M2cU&DoW)0BxAl|d*000C&pSvF zSy`zKr%zB%Joc6p(`U7Ovz4@RXht{*u98mF#a`csz`_Zq4aS>>o%M~E*t?OI?G07` zcZl~+RHP4N+;ys;9rQ9k6u$n5;`?Yj%;#}`li(~Lg7&G~%v+^P{^fIWk6=oRU4k>- zg~N>K@DmALJLb|vs`PJo>#U*(cf|7xwj66wv)1AoYfw+Z06kf! z4Zk>qQ!TuYRx8Oy29Z@aO zlpMlYeKxr>B|&`KCF653f3G|n{`TM!PhsQEH&jvBw4tR0Vqo2w&ZfbumwDJywNt5x z`>8KD(4nH{eFQ%^JIPp1sLua(5H2T3Ik47H{{SR1GGop+to!%reGoZ$MScU-ft>hR zj99eW4%t<5mt~j0madHsj7DRPNZ;Ywh&vp!!n3K~5S=+92}EPgD)AmE7_NCt_kL}D zwT)07=&dTsPINZ6tQ~PA4vaw-^X_1V6+L56QO81Sc4XAR#wFVb@{maZP|pV zQ#O^Q==q8?7C-mrwjIB}ndWI8N_=jX3g{_S6}0jj*a*Ti*?Gj|!cQ0rk=9|-d8DED z@zjR?K~RZI&ER*RUBU#)5m7@N^3tnV#gAI;!CBNmV#!vIi$Xi55`PIL?Vn724CBz^ z^Lg`g9sbnF%x(z&ZM1B} z_9YQw&NAqHj9zEHmc-pkaQz@xB!k9Axt3NSg^YfDV=KTQrv0y+=1i#^@|X zZaJLcZjWT3N#o3D<^qn(Y;$#EWpLS_Ik}+o5hAvFo(0L=J|_&1w1*_~&c$P|DBOWa zYic)HcphX_O3OSZ7mO5DV5U2p3oDmRTNM)I!L+~u)UJ36(mgVTV{YYvsh{@LBq%nd z8}bLP!qX1UWwxr!tMDZ^?`4YEm1$b8b!tQg9g(pugat`@bXR0dY0@By^gfwXi`7;B zdeTv2aN>7U`CpGM+%I1fVzh=QxYNv+eQ5c;g!RUhi zcplr$Y9yYZ3nt)l;E2 zji)=v;&brWZfJ#&;>8h2gfz8Jo~G69QWnJX#E9sdKD`StIO(QDKoL?>sy@0O?ZD?~ zT!bl5FXX2$FYD$2)ed~!*Gp8_k%=UnV5EGf&tY(g4E?~^M?8oA0Dp;65a#3(VWp)i zHCRCUUD>xyudo$b=qa{R+52}rjv6vsMa^(T7bH5arvD-4tCGclkzIG$sb81MuUPT| zOll*GB+Uv<7~t46h?gD=3pm)6H)ixpgR}O*5gRQvOH#vU0|tzW3$C#^-F3eCg%$*O*LD z!e(Yi8!?7EH3Ee%VkR%D_8{1E4n5;xlNrX|C5f#@P)cVDlmws>Sc5Ol#F2(TnR$g%xvKu$}Bs%q+uZ$S=`_C$@gB}JA+F4@wZ zHa8+Onn{I!%xlN6;Q0)Ct0pCzrsRbcq9Y(&3i59fOqN9`>sQkXqy4$))WtF@(d<;G96>kN>C}Fx_#dY|I!};>s?U3eOe#pVjgGlm^4Um>eFr; zaX!LF(PGJuj`6J^ZBHDrZ_S#8DI|<$bc38>`)5tr#j=x^(7GgNtzv(MruTUTSHX!L zed=v7uxY~>>OR1+Dv?(~n|L8sEwq_@TsV%&UyDX0V`;(}DG4t?gAr%rqL-&n)b|@@ zZY1R$Kyuv;H$&*hQ9=^5>Y-A$h1aHb9xFR6$V$sm`1+*YAdy3f5nTvF?j-sxB+EDI z=@-6#iac8VW~ybY?Hmg(a`v7@5lm^TpzMHZgq$4)FhgWMsuaS6fmTK4y=eVDB59CG zT&7x(H3f=X%|iV!kh>OjDrlC?U_;|AmPca!d2i62%B%ayI%ZN{IUb>azxJG7LlWD0 zGfiBQB*Ge|b#h8j1ueTJu6T=18@*y{_`^*{Ag#`YwcSdY#)Ov;@@B%79I&KPy3VZp zVqwMTf4??2eE>8vp|=**$>P1qwMolkQ(n^Bgs@sVi3TMpE_hSIW z?sV3P=JIPG!rQu6e_9CWt#20B2ubRz(l@VwccERV3h0e}-nP$KPQxQ~H`VAP&YlLA z#hzkOzF6Njo!rJQUQH(*E^wJ8M#^$G`6RQ_v=Lb+_o~L>yLsmC|VT+FH zYf$ZcVTE|Oe*0f}^+84N5gD%cG2MZhZE8>p@SYaQ^h=G5Ed`EcPgr)z2{trLQ5RGO zJGMZHJONfII@oe;a1Rsi{`D~qJC~?8t;0Oc7q63jZSn&ysUq3?(_)z`r;56P$F`5| z4#}bs{4>Zh=g|2LBcGWx=(;onRY)SlqxUHf zOlurWNqBDw$e+i87bY1jnHii`6)8#>fwJ=T z16Fe?55q|#qJS`RLH_+1~ya(S1K@0alW8&q;(S1 zfq7zg)H8`yK2cj6r<8~PZ~`A|h1Yn-#>1&sk(`(2pS=c&PUx6;&sdfH7H9DuQe9HU zBdZB-4?(U`e*{W(tg%(cWXj7Vv`R?v#QPq`mFpxqymAP|!K7)?1rerr*Vcr;lZ7;O zWzjIdz2A*7XcSapoR4I$?LDXvk|%Zeqd|1%PqW`mo*Li4O13+WB7zpjWjjfDs|4iQ zXxg_IRq%UDjhnqk^Vy8cgd-@ik9N#nf$sZvG@NUkN{EGxsELFWq@bCrS!+Se-nZ3; zbqO*EoY<;Tl<;;hI4!;+j9?@O=^Z-ZQcj-33|J^jh{{!_i`7sG-ySO$00L%&XUuo` zNw&V5V3>=*RC9?9;1e(Z+h=)iI+0NDxPc9_4+w?-1AYBeoNx9X2Q1z2eFo}^$u-?&P|V+l#L@cU z>7DGO;MaQ>PHp_M@yZ%?pwM7%0+3ADH?Y}kxa$ZCfH)dAb*!OkLSB=qlNfd#d?3`L zybCblu}b{tV!U{Syh~57j1i6nuhTH0s5LEKrj`#DxbDN$VZTn4pg*8Qc5(A6IsJn8 zP^7Ru4JQ3LJ8UxcuY>pYZ5|K!qF!C(wG~|FWvC)JD5vO#s5%b4?jkZpi>X9oJr#!% zs_zxl`w%NfEe9lT$zRzku zr&J0I0ql}3($V?SjhzvKi7scqc>G0$Rk`}Fch2*E`WTctsrf6)c9 zS+Vf7)6S;}(`8IygdXc=t){#H+35o^419B09r}7b#|f><9gpHrha-idade(DsJSY+ zU%5Mc!{~wJ4-dveRzWUa`Kg*R)QxEvIRwN4D_Uw&kdEjnlXlQ*uc>nO>Bc~54}t|v z?BfBE1l6xbYT!1nWKms=>#zGzSO8xE{{o@{HSm{XK~rU{!4AT<uSsMlUW9R%P7CNUSbn)ixgsPmE4Cl0xvr4x0^`|#Rt1pLN$-)1VcV6((Z>InkKX6}M zPxvv!Dx&moWnt!IPY3VmpYPr8sArz5h?D`!;fJizNKFEzNt6p8V+M959MF5roZOOMXWOuy zPf@UmQ&j~WV&^aZ?unNQN&B>xorg!c|KT9XPxO13ks*vGH=yiq|CV%%6FHnRfRat$ zZlOg>>2M%&ckoY?A$5G|t}%JJ8M+9LM{AlA;FX0Dwiv>b;e4hgbmr>uF}05sUZ6l; zZ3Hc}?8mJwdcC{t$4vxwad!E(zA328&tvxYvhWv|6JstLzorze)P3<0I{95>QOM&d zS5^%z+VHr**&+V7G7YJ1<5!M!cJfX;8#A?T`M5TR=+W2!%zD^Lyeh(HAXskiy#0qJrx-MkEgyUA4ybHY`n6 z{Yu_Pz^byq7ZsacGr^dHP5{x?&Xl?Vx~E%k^R3iq{>-Xd{t2ojZ~E9|f+OY2pIt=} zyw5bm8YXY8{Cnnqat#2yMv`-?lXytMwYKfFw;FR@9iJ9>J9jqfB_C1e{~q)-#?NZ) zJy#Iw3x?h*0Pf;;!fMF5S6*W|_!E4bqWAiF#_R;%p%|6@D8G2f=B+@CXm#(LPg6dP zboT0a&2vRN5&t%AC9<{EekSb-Gs6ulTVi({T{b*+D74?t8!|Gel5{$MBw&|{E(Ju= z<2nToobMz;?EnF*zu*z&EuKP*qMV}#bVW-DEX2r67Ru)!^`{5<($>EjZUjBHCxCke zL3kdP`+%O!AAn;Ze%c)Pprh^SHx7L57%HzrU&O*o3>hsDA+~&@Wtz4~PWaMP9qSZo zCHj2xVu=XkfbhR17!wzvp~L4*yax8dDTSMOlBk)hosD?T8|(_zr>96|l~5(J=&i-* zjAUdKL5{T=&j#Z+6Gh|Nul`t<(FI`IfGlQ_bmD-?6wHoyg2>dWPclXCZUstucgUZ9aWd*WS!)oH?F!CWU#W_JYXWRENaZR${km$|A%$2| zsjp^Hp%OKGb(MC)yVO~Uv8ahr$k#-|ILj-2I~Z&d!zZzw4*DJNUL(iGRtZy1onNA4 zW1pfIpw=@1x6jtuL_5=nj6y7RdpiBJwVNR^RnK>*3$NoNt?CdHCmfk2@L)|f-<+*` zR7t3zci=M$meRwNr3io@*}z zV$4}1=W$ihw=ar|75HQFaOB~!Ek|fRu}9gtpNI>ne&lGOAVl^C*`piTH~lYb(plhX z$1w|#BqbQrTnarZI`s);Hp>7D^ooktTHhEJwl4je&C~2`lnpjgfqujw&6nU%v<65p ztadg~*8OdqK7uyU_|C$lCaR>c7X3%*Hr{yv*_-BV)M(moh4hW6cw;V2?-D+aDqc~$ z#n62ra%H2?^jeCLQ-$%;Gtp3dIk>82L*Gr@GvLfK+orpQRMEn)%ro`!HD!ji*|A-M zOhvRETVe2h5KyF`99!dC7l17tLPs`iB;%$Q{V(?9p~DohT5B%#hZx|!W5fFQn=dKw zoaSOhSylmQSKTy9b6h`2Z4I%NyB5RFtc7tbi@AwOd8_Kp{GZe)^M=yP1 zhfK1UPdA9&gm~5YjYcZa;aTg#|7(2qX^r{A{*w`PlDX`xp2 z$?^`H?Ik8U;oiNh-FK{>B^Y-2Hk*;L>gIB32;6cs(6+7|^OQ&NxYI$BSE(d0%Ox1W z#&D|+e|YIV=HkGG22P3$Cb2PIrm{1$I-Kvtp7APA2ElwRI9_%nkprU#$OSuXHUvek z&_ct)(6nrOQq=$rMNvYb&c~WDZiwQZE3IkIZS2{+h|s07s&vU@!TEE{1{$jMbr3em zBc8oeE_X22+ybDec0I9B>yS@rHN@@6we)|&L{ILTW!>1!_vg?KvYDyxpoHCNRTz1BuLUJD z;-3@0WNYEkeKbe9{%v2oxZ+FHk5o=}Nw3m(u~!myg4D~ck*`*Hqfsronr@(aJ|ix1 z#sHIa@{hY!{vV9pLy&0Ek}lw~ZQC|a*|u%lwr$rb+qP}nw#_~_I$p%8@^ZbHr{!f)AML&zTdE*$& z=IPR3c-0Vrn(!LF5o_%s&agSytD{QM9)UIKO{H(vqQ6utmqpsfN@4!G0!8sVYb?{QAHMH0n+-m9xqhh9LCp3|4Ob{(6(^sKcyI#O;1- zRotwK2M*2sc9$Sn84i8FeGhC`b1mI7z(z8wqR3``*!XjA9PF1 zo>@Gz6bbjTwk4OJG8XUvf{7Hwp0mQY_C(+Z-K)Xh4*s*v&7&+-qqGP8tG9WwrUea5 zu)ncVZnN34t(Dc<^6#|i>~6N%bbXb1StR?!C(Fo~V%AWvDh;o1C$^JhZK9ZaTE$%0_8T;v*^Fq@tc2_RAYef&iG;@B7xofoEQ1k zC=6ywv(Jv>#Q0ct>WYJCj<|7RwBAfO{F@gP1$DioQAb=eLIrugSM{YraP;Z++*>Mw zmg46#;V$S%=^uEZX{)evk^BwVF-rcizr=22Q8xpMRkctEwcz4OC`~j<2)JP?l(UVc zuu~uGcVZ_zuArmO{!=4%g#*UCX;I6I+?#M^-R1~f)B4Wt2v-V`wMfh0-Rqc0wqCoo z=f0dC3jr243!ANS{p_E(ApaF}fdh^?H!j$d_E2eFX-X~80mi2THs?H*G%wdHKkvLl zxBDBW*XM`hDO%3ci~h#v)p)cau7@$eN$Bh)gl^M%)OD5Dg|p-}H4F%DqFX_?SLjg% za4QGimQg3=HJGLTSl$k*gTNw47&6SD_EO3`Q+Yl9bu495rVMm8;Hl%Is8r(C14qI+ zf79_nw&_hjxq<;jakv?PXETF$up3_twPC5QWDmf+(j8m8CkFJ?w)?JoD5#Mzw}MG% z!(m2+T=IR})mQlltIVZ*AA=ix)A&k@;_ajz(X(h)VO$wChxCI^_fxfPNbb)LJ1)q9Y3 zh#SQ<7jwEkwv-rO@u>z`&jK5jBd(KFev^3JL!{tsw@z2ODAtdk!ld_CS0(zdIbVPh zPEUdfC?xNrq=!0z^T=ty%f{(p-h0j4(+S&yWj;!95V>&TbSBti65)L(UccW7Msan1 zomW{_@v04`N8hdHvD%p;ro5pQ+8J&^Q6km^5+Qy|mU@YW0NMq4p-A3yd;nU&w93;` ziw=JPt2%n3<{$Z{o|LKW5rX3uSK|Z;hf&YR@HKk~u~^v zw4RVc%{wRu&jA$Y{y%7pE-W~EhIh6W&hPpeQ)z^M54Sqi72K8%)Vc)bn3o-Z*_%OL zx)siy%OdId^6N9V{s_%a7mApePt*v{eG_p*o(x8hVCa_bA^pp{mII#n+Ys}>RA}y% zxS-&3`DGL3uUXlTN0_+3k5MH-aXu5>=}v&J325mt3_}Jn%02pGu&S_#Fs;2v9biDJ z@WVFOFX{%_Z~3sZ1M;_MP)Cu*B3i{I>L*73SKIlxTRW75N+E8RCB42BvurRtz8FTQ z*7;%EAilRV@9t1i!6*` z4j;+Vw!UuEd-|2GuO}Iz%nZItiry@vim@+YXNw;CQ8vTV2Hh^9wNn|}&YK-vrMe@EM1vgHa>P6ipSI`7y}XK5H8Q_D>BXe4?u&& zYljJgE%b%yS`==S<^f>4+?NTW$_~|at4uehj-6k&-A*oGC*I4YT%&pze~d2nWB>J- zKT%C7>D;^5#S~~N?meHF67cfcg=lx;EdKs5)n>V7dm3Z1UyyhY6O$dIb{ImG2HOsO z|NTw6QKrRly>9&;I+(PPb~^H~!^!c$VKMp}2=DFXN4yLHhH|r<$Om{@q2lYm`D`LE z4dt^!0z)v;-9o}_8j1ytHC4POfjOO~DFs?&6uGJ$6WtB3Y13+LHM8jrnGXs~h1zg? z4h}|f!M7ovSa~#!s^O6zWyn3{+F4S?6nFYhX3`%(rfP#s&{Hv=jRe<)tu+k|w}Wwb z8Mcy=$Wf07xG>YNZjOkQy~{|#(+02Vkg!%&aED-M%2PAbeg;H4L10)a*4qJdjCNNHuIt__T#4NnJ#5x?(AAYPpW z-eLw0z)K>5wp;9;z%#a-CcQC|l2UV`5Z=YpxMJ{PEG3K73CFEAyJKi2x`S=e8T=FhVxpFYkK zsIe9v?O$kGx%e*|1Z&q+fbu(x7QylvO}m#6nfuz|v8gi|V$$Oha6A%#BYu^yh37-b zRLUs-^0a$7g@**4W~$!qtYJ2lh$%dOtYwY89Psz3{v1ZO7B?p353j|N@o@02EqR(- zhD|kS<8Js0dV{iC1D~X($s$;l(1zp0eGH{}gus52^=Sta&|^#^_xIrw8|}6pC?7Y> zmtTaUy)GNl+>0>Bc}{o5Y;iI=qm%Bf)MkeK($AYeka=RjBlrASp%t4K-(Vq2*PGQk z7aG-X&aC?=R1sqhj;wAz(S>d!-sg_@`pgzeQFxINZ6IrGWAjEo0mi>j8?T;Z>=)(9 zTv@3nxCsR@60(8N+fzmjtFif71Bpo-uHf04@t)epT62$Ade%0QrT?t>fH@|o?de(j z7#j~qYDKa8wS^~A`gi~Ie_}`ZIx*-84Wv&@2bJy)FN8e#OK@cXwhFu)_57yR<9wwN zgi?$#NGpz)P;0r>RSr?CyGw{^_vjR@*+Tz*Qc@TqtB!D}oj_095zSODk1ZW*{5;9` ziH7oG4>~CU#DGJ^zSii02i(_0xX2`_y{SUMqpahP3$HSZ*tnvpl%tfo&f_;Kg*1e0aXFlxeGB__E$ zBz|$>S@>LG*>qt-^T7hEqP&un2xc||CQU(f)+__R@kTyAzTO_FRjOifFEDTvT`8zn z2NV~aM=iRDaqcl3w%&Z!8!9?>RfnU4WU5uf`)gr%xqppV9@= z0W6B`Vl#b_YK}OtW`+Jz*N4`}x7u3t0ykJ_0k>th*Uc!{na!*WXOQrHbcHiz%IP@U zhWvG4q1LRP38ZfDthp&&Wg4+Wj2}v`J}t*|eaFN8Jn5L&mY=HAxl45Ermk8n#ouRN z?*=dkG2UTobpsRCr?%$P7U`dI2ky0uk<$g;+4#{Z696Khb;H%qg5uoE57b#$vTl}_ zTZ7`4R>8)HifL0=$@aaj?^@fsx7_p<3l_PsCqV}kQrEp$J4I`?^6tC{RRWt)B3?bT zu7J?Ap+vh?YFS!CfXDG8NcyO6qe0A4bq7u;_Ur!N0tSquR=2v zT$jya;3X0=d9i$+T)B)_CPNYx=z|U=_FxJXa5#A?!ogHtbq9{V_3mfpV$x-_e)Xxu zHW?p|fe5tz{K$6Va+Mxe+@1E+B(;Zmf@HQwj|vF0_pi01@mQehsnd0%baFhJ*k68+ zzrKa4xE+wn^wbLv3{1e>I*!_07`2gJow%F&xK3Oz`&FY}P>1%yF3uwPUf(--TqLYs zeoicx7Zw3%==rH!RMafCMydq;QgV7%6m|AT7^>1U?|6y>5YeZc?%8=YFIY^KHt_RM zbIdAOr*^XLLG5CFgSR3h_0RLYFs{k7Q(3F$zNFsbsueb8UcGh(ILE*l10QaqlV93? z8@|M|D)U4G6Sxc`2}Ush&fDV;rs7SXJZ#XD!W!rxJ1NwIS*U3 zk}8-lI#rm2&K;PC&A~vsf)eCTt`5|u<9SHD@m4ucg0-Vq>%g0hWGLTx_AqXg|(;Bwx5rNsi8m7&=i#^R`inM;&26r;+n$$TFF43?HRf>zs5^M+qri zS|4|@T}*x0BiaMGO&O+J^|+tztmBmoUq{dHs(OsuUjek#b36Ga7$gosW$mg%m^bDa zJ}S%_R-RNPFuEaDI95C&`#AVaW-wziQ@v0{!*^;I#@*aHg=d+qNV7JPws;*%T}m*T zk7_@v*g>7unuZ%}Ax#1w_&E1}Am!jzC#p1?Z}KIQfcu-J2(*9L5EO ztnlxSNUN4*yHTVyMn0T<>+m_D@8W|K6YNtQ)EYNT#|<~}cB6ET)?1%pc|Yj>+po6j zjeL+Y_S3l4G8WU9$v2b_#He{^q!&gvEuh!aZf2f&I5Sy7tQD6q6Le zqpxl6%;}v+cN7R#flQVM6+0VDYW;-6n%?GD0HK>9NzDdTka`nmk0!n*$sJ0&Ux>!k zYcG)OQV!mJRDsB=E?_?r-CgqC71$tjyWMt(EN#bJc=my)UE0*SC>MZPh*W>B=Y!j( zc2~vm%H%39Ftf@-nW<4VRn%gAtCha1)iH8I(F95B(}lnVOHyH>qYyvsT+0c0JPig! z!oZ0n%%nP-s;aiYh}Z_nb;V2|_llhk2T=UMErHQtN2*zBMv7R+VW%-(NHT6wj5Qda z@0~a=-Ak4DYwt`WaUHl{E|SJ6Q#ZC>YpB5oKhj#(#L9Lm|v9P^z2;I|H@AJ z7$T72^jmxY1azJg`uN6qtMW>ylI(L-E3c)i4pXEEs=B8+OgnEQ$Y}I+BJ0Z$b zOqlUWI;0d+YA!;qc1(wo9D11erL)h+7Q;$3!GU2q9pF*XaSx)Mm4=QQn?Vm|@tRO& z#hdT{0C_El%ERbFn8k!^Ecqj{V6|jXXJ9^Fp;K0juh;YuAmm=lv1@8qVkEgBi7q&i z+<%@hPqR{1p+5sS1lAtSTNCeLAXh{+x3H^4#$D6N1X>{_pEu8L%RvFR7)}_KkZKQ$ zpkK~r7hOf#%r~o3RrSu`p_!ZnE~R@ZFsq82ria$0Q*XKBd^JRoRipM*~7mN zf6uesr70Zr&YYTNo#JQM47~8B)EwamP0wv~DZ6->kfV8~HJ33xCP45fuIj9~9b5`$ ztxixoC30##3#lKv^6EBp1jWFyJ1|z?+d6L(s3=o^OJG5?xNw4L?k@wqg5(jlncU+`qXHZ@otS6>v~9ho0Zh6r(Y72T_#-Bw*O>{d=>Xqs2u87$7M zJAnnh8tW$VfDwJ^PU_xs(GUU;N8+ZFv$G+EPW7_W@uSA^7d1V{MEnzqDfg#NKakw%RGyw=~*OXXBy zU^}43WA31a$5pxC;cN&X2HC;nh!_9O3bph~)6SXcGJrD<1`J`iFKQd!7p(CRt9@o$ zT&4bWY*}^=-jNQ2X>diri`x=k&s^2<>x#+H^mLswB&qIuP9PM^h0WJje-~ERrA3IW z_n{)og(32jJF0t+6B!}nV7W*f@@>iiIxB)$HHXcQpxH2Z5F?;V+_59uXWV%PE3m8A z+s`o(ZCJU}jb8;zQv$^?KnVlVs)h?n~ELZW`-{v7j0OUx%)YWP7Eu9O( zWwiDb73?5hrK1Pt;?i;GG`d=HvDe(gczzkpqXs$FG05M^MoIapZtuT5AO|A(NPg4BR@0;~!ET9tTg z`U^xuZHiRN7m}FBO(+kHB`Rx>FUW8IqW{zY|gQPz<4`YVzD9?g5I@)Aw!FgupszjcV1Yh6qC$(#F<1oXw-c##bwxy5q$)6UWb(l^^ z-Wr~$48@&iAn%;cO|L@cEhy=uw!CN|dEFO=q;z&ovUO?QD*wQw_*6tg9Nox!xtE$( z=Pl9{VAJz+-Z95K>Bvynh;`YnvHc9c==><_;(JC9u4r7z3IEtf%^#NJWHAtFqnlyf z5X%hmj4K2(sn>RohzD-6Cp(~WfRj&Hi0B?xh`y8dtjqPz5v)CGaSBJl3`};*p_IwR z5YqG`fX9_2ExJBrrpr98h@9P311?b;V+Kh5-8iOa1I#Ok2CMa^;8(sk9|7!`FF*M$ zW1r=Op~ccOvJ$wkK~;+MOH-RG7gmFYT_jQoT=FoEnFHiq5}n4zNE;ql<`y$tp|)63 zZ$|@n3JDR}WiOvm?PeqkAv;ZS4;*v_+|%(~TWbvXrgT7lGv-NDg!qVT4^_OqSIs90 z3Iz@H%2Z05R0P{}G6h`tR>tBWrrNN8x;E1_48*ic>gk>a@ncO(^6IHW5RjyV5KTh? zKsrzX#ks|GT}yYRDdGgY0693I!xK@0qbJrZnIrim4)Y~gRfa3B*6D`Y;dEr6lMq^= z>+p~zXc~0u)oX9GdVTuF-nSietvoKs(A2b(hAq!&k$iaf4;f8TGYmuRA^|~b{wM~f zo;Z=N$x{9hUe>c=|E)UZWtQ=a07+kwm8BOV#eoMsA2qHZsb^y3C#~w!*yIcC!5%&d zgVPo1zvyY8A5+Q#{FgcQf-;(9VUlN!>h%OIQf!;rc6-!3DM8d?*eQI}7R4Je>&mNo z-Ba|KV6L7%w!^a=v-4sIOz?u*d$%}|>?ZjH0Xsy|G#Q^cs9Vn2j$ZaX zk9TzxNFSU-bqFq~jVTW*<}SuoHpmym;y+O&viFU;H0>O6sHt79tei%>+H2`01mx^o^xGwK{4GKtK4IJL%kT%YNC~Z0EksovBHpqR9F|vHe3F zGCFZ$n4XPGc0T-zeg$ju+~;@GHM&sDkiyqE$|+D5cm-?>Q}-h-B~89b*7&qLWxLqU z%=9Y!Aft7I;%+YDBQR@x@JyXP60%% zm(4aAWBf2pfk6r+fbaf;2+1eUL0*~oCfGY0P*1c`6Z>zd6g|kHS3yfh-L3aL8G`G= zZ2)tXKd&y{{go(UV!?s{GiMjtlGhZ91ppan;~6Fx$<0gF2KJ~N_2#&HJ_KLr^jW3R z+)N}}i&24_`s+;?UB>4JyJDy@Ha1r?&q3WZjr5zPo9ZVd!L6yEibslo=HTvNCa$Iu zRUq$^z}ZC!Zl(ZSx^Ja<0be@8O^#*{nVF?G)`}RY301?l*F)|bsKvNOu;FVt%0htz znpLNp-(4_AKYWsQsP~=6gVPIl?uL>i;+mj}u}b0TSfHCsmGL8+I?-v6@DXav=1izoWxnjqvKk%TuDs1sNOf1hnEpzbFjtAlT=Ed}?nf)g1^t==B# zm;OpdDxd$T^YXdZ0oj?P_K!PsaOz&I`3mSBXV?@C*RW``3D0A9^5B{Q38wDDW@{s$ zpmrlw%F!5Ex!!-Gp76c@gg5&xP{P>3Ewy6JKvqK}S3!DA6s5odz4_A-dZhrK&ew{* zBw5QAubcq}5moU@l-RB#^i*3oyQFYxZ&M3rT?tmf#6N-d5kc{j5{M1xbO+-zylNfe zX3XI=NwqS~o48e)_R!?{x%8c?-vf` z4s3XBK9pM606*|dRN_Y`vk2>nf+F2=`UypqcRze934AckGs%unqHBbzS`J&8LMyEg zbm*SlksSe8rI^e!x|rf~j#23Lha1Fy&7vdQD5abnktpvKu4pJ!w)YcF^ojf8wm&R^ zJ)b<&Pm(_H-CQ1h-&iRte6e~F`Rvd%(^iYTtQ2bq13q*DEAtB-=!w{|R^3Y7c9)2X%l>gBz0Ed$QDdX~W z&l@9e-`i{9CS3v!D0Vk}(os~5z{3>|nQkFf^d9-fn+7^f)QI^JvN~wW)oD+DGkH!) z7RNV_S-gXM#Jq|M>MkpOeof#o9PA5=?cc*$rq(8pNl63kt};_g5pDobYr)(kYK( zcmv%Pp=+9gghGGyNE}+^Y12eMty7tj^x%rl?7t4jL z5}ou}6laYl&ZIxzq8AP$*|vgB_M87mMg9E|J#cb_oQ;`T!-T9Dueyc_Ot#N!kbOP$ z+wg+B@H&OtO(igk8Gf$=DKR!u^}qQ^GfC62C@*^wmIt{6GP7ms_aTeC-N$2b7+3Hq zX89XXd3mNi^9+{Q4?)NdnEk^kw0{I-eeyV+RKAh6mX;Y?BlhVt5gJ${(sS)MWeF|x z3_Lx0qhtLX^c!<&_a*!iEC&dLB-V1KqZZ5JACBSR8(9+76BpT86lTD|TPH50+nuZ2RwmNhr6sQ!`Y+uJ6Ap zyV9~R@b*3gp0`ssE_K8A{u8F3lRj=UP2p0Kl)KdT?A;?tOJwAL$Ho2?-*_#>jQh_gp{g#Zbg zFZi~$%dH63lgh^>fyudV}r`AQ6>9gMne$hDL|$E zDG_UTJR1%vyxu1Aj)n?@MVCIZLheCco4!$DMd?G7Hap}VcYyzqJ>jSpb|A}BvxNe_ z=ScIkI1oH0QqJgs9JEg-vvBQbaYAX1wI$LJMG=K@8;zSMqiDIK;IY}jc)XNILgp3c#O zKj=`7ZLLZ4uud)SHQP|{3YmSy8-bTH?nQyXl3AlMMuSg29>B~Mx(*)6^W2HK$hi=a z2V9M%@b4e(1+NaI0O!SWqx~tHk3yH#nW;Cg;}msts|rDj*QvU%<^w3dP4e4W*X!F> z#n{Dji_&Ep4`3H$7wG7)mpt&<_Z4=l=qV=|Fkk<2klk+PFs8xOyj`tRw$pIEKN<`I|&2HS}w%;ksH zN24NQYz!Lpo$~VHy}M8bd-SBwmWWZOs?rii-(3w=3f)v+1Po|ML zx#tD|z?TzYBAFY_oKd1^7J>nyORi-{Y9|OoxY1PSMu(Tj)coPJ1#CBztx~%>Jjqc+ zCnAPHT8?f-#OyRe&-qsp@ox7C^Z{6$$)LH013+vwhe6TTTc&2qKAdvNyz^AZekhW4 zUz&e`WJL+TQgOvEf|yVb zlM+QI7r3?d34Cl@bjr|@`!lo7nJxzoQEDNBg&yGA>^*P2?f(>M<_A4g!As7b!Q9(L zBOkKX0lliF?`9cGgCO8jqfLQR8X}9-(L!r$2tpPxpTcj;?vU4iv{}aTvW1>N?Ng3% zF0fUbtj|q>YeeJsU_XO+^_6uz)a|UtHY zF$;1KIdfrv0}k9I5vPnukJ0eK<|Zl*TPQ9%E}%&jebJvxB`!C={%`~09vnx=N7WM8 zRhqdIkKbV+MT0xTqKiEoo>TS=Sri#GnyP%Y23S`oky||TW-_`hvDZu zamONQP%ILkS;}zz2j-WW@4|049pc+0lMEfeYM{AV^`2H72!6vpS(}Oy2tt*`ub?PQ zt`$u(Pq{0_dA+a6Gd2qDgE2ze$nJ$3>2qsGV&GKLFi!qOPP6J=@x}zVp@`O$VJkeVI-9DXfW#O=a z+!ZfxdRl0Ok%UsHuULp7n)9Kc+OG;pnOTUNElMi28jfCOCg`x?lSNYWC_FSyrM$R2 zt-VrGzBUOA^6d^N7b9gyL7j^HyEIU8^=DN3*N*q}hT1FCZJf@1YRv!IX?f}tXgqg~ zIJCy+lcIPNeNiOwB;3lyLc$H_bI@$4U7x7f37HCJ{Imt#nW*&{^9i8=*tMl!So4&W zz6hV9uLnS_+&$Sv1yRh~06mBq5L^RX#OKXgT(F165Ch4I>dl(E^kW*pv5aG}d>H3F zFucFokU3A;6X)0yP(vKjt0r##ipe+m5$zcED*r#_Hj zt5Q#@G02j#W0ws~!2Q#5REY>1|U_fUZ}R zqZoR*W=mxFBZs8MdjtXA)N)!kDj$9oCTAC};yAJZ%&-O04gb`n)!~gAt!0=zq%4Fpy!(I&<#wJ)g_E)Oj^Bt3`kx z{+ZLgT!$119|FFX7RtJ0X6@<^=Ucd87#J{H7~1a|%p1JK;DmR}SYL%Dl3ajv*f`8M z$kk;pm7XfxSJ6>4kEcyzM?Wh4f(WWZ2UKwI+|N7B%QezNqi`9?#6~JFPiY-{bl(ii ztlc#B5-QKtzz_}N?lOK|U@`+OUq(dOc?M{EQ83POvLb+Omp=H;KiQ&8Yl>NxhBr)U zs+EV+R8Xp~CYclrHOWSy=xcL1$WN2Vk(m4A&Lrhfnu?^2WG?j8Pkf#d`mYCV4Yz|A zHnLCo&8cc~!k)B8Q)`s7Y2!!z^^PX!zXul2r(2*fUfZsCO+FQqzJ5VW*6CTm zZMQgp_C6@zdkS=W<8V=1vtLF~0FOIpX|4Y^BO?=N>%el70bRf3vmODxG2nJ4HAGAg zU5uX>$e@S`Y`A=)TbpjXl?4?NgTZThUOht=ASOq9wMw}p*F~jo6=Zzzo|nKU=`?NI zfkE#y7eP0G(N(deEZ27q#!AYDkeZ@|KBc~VoO4I7psLB>*Qtu@lSH;dXnCYbPkRwP zoiZl&5CH@qChygCp>v0J&l#?lQvKSEV) zBsAX~V@j6k80nOX=pKJUL59D)8}}2-{T3qef5pbO(Pb`<(2amDpYV=By+<0Uw>;Lc z)pYksGqma1sGRE=GuH9ocEYwry7T2Ytr&HIc4C&EW<5K=%B*BhZVfWJ=GF;N(}ucA z<9c;SJt;l)csa+-V5GUA)&z4}e4J}-di6yRQrB=CPzRRVrwQ_If9gysG(I6NOuoST zCGoU3 z6k)3mTD?@i z{VSQLGQIkPlDPeckAw}u0RE~i_f%LJy--a2{E*##E;#B9yrIN$6N-o&C8{y`BfpJW z2{KOWCCiyf5Zx!2l+Yq*+d2=|XqeR1rl#yUx`j+0D)qP8DX#epGLb-+x^FF=vPS2w z_5Qi|lMgas47v$mlPxL&UECahKR!ATVaD~YFcTZN6kiv4db-oA2$t*lRn&J`IpX0wVk06dY&yy&2(TyHj*tVGvefnUXME?J{;rcE~wu z8we#}6}leO7=Bu82qu^YI&j<4YwW)zH^dEl@*b}*Fb>gpGC4hZKzK+Z7O)`^7!PmP z-Fy*1Ub9gsID9aIC7Z^w#k_XS2P=tJ9wU!4t%n--S8V1MI>V|1AmhrU+O`w=qOs`?1g~96RGyl-**H zk!`6}ObRYu7M~|e9c(ZWC_aw(uXPOxG_5~4ncWMbuQ=fs<&d{({vnkjk7X0)U1sba zn&}<;#NLAvaC^9=B=*~uKLY7dQ?tNbh&ps%vu6c(z>5QV+}xD8d|-po&?c~Ue04g} zD10jW*@=SNwW9n7ez!SrU(LYo+{u8)*UIXfV3H1Yg(dF8y9w*=B6HR6n+K_dEb7*z zUSB_upgo!|Vbhx_Qs<8Rk#dRi5GT`7uU|SFzoL{2qd_}O!s4M{E4A8#T&O61PtH0D31x>om;JNCcxj*CL?z7NEB+~ zjz<_EySWvhy|4(?$x_VS(eDv>6_rI=r0$1H;0Ef9fInQUOrRjz2m`Jjv@6x(^8>c{ z4R3kQc6{l_M6TMSeJfPsp$TZrp|8nnBrQv=Un3!!GO~X_*gxAzN?X9)(Ibt<`!K?H zT)Ky6UyKd~9znjYO3{vQC4o`?bW#R|vvVE(-OKuw)&#( z4C>T_3O_xYfzN-?H%r4}fg2<^4>bB6v2*qzquam0j}4agX+dx^L=uu^#xLy5z+eMz zV2=T4N2J9r83gK)<+SB^k?Y_KMT;6-P@Ns_V54gI2_f~HC>?yGTnZ}+6mfu+r7HC7 zt|l_eUHrb`ZL&eGHcR>adCObBJ9SI6;}{{Iv!C6L!l4SOQr^sg%4sTM%hzV|3ty6XEGATulp1 z?LLg7BZK%-4iOX@O}s}?X}|%p>C0?_(p)p#&YXv&xj9p~YmYt$K48-u=bSTqA1bAe zYY=&pFI4pH0>j_SEAG1L|t2}g%-q( z9EA2C#<{(nGw`nva{Ti;;cq}-Z*Gvbfc`jke7>Cbyxx9ZRBo`AR58A2JJld53WAzw z3LtY6=cVUVGt&~|0;BVbs*cZ$%%vu&LYki#+MGe?7Z~a52auAYx`JSFX|1hCVsHVs z16<9V0C*Uk#x2D`fPu-XEBbTHp`E?$vjAdv02l|%gi_(X@QVV3?CbIO&x~iRjLn@7 z!Zxus)`2Se7C~&TdvR!{XK?>&1ZAYY%u4z-gq#5=hK{JZ+RCb;0SIX!`=jL0;PTJR zTo0N9=G8TVVgOFVR?7gomH?pUf$^(oC`(`x)R5Fxeivts$TRWwZeRYT3Y4)bNk^3R?K`-Pts+xzRRp@1rYDg&w@`oZ@Kpbydp zgquV2Gxw9m)&Otf?Sw>{o#I$*A#D5I zFW3iV;C(i8b-X>Mh1^@Vt5eRg#K!tgfvau1Rj z|0_g*KyY<%_7*1eqW4ee%zsaFChUykUaWt4V08Yx#bs`C`6qO0+59}V#9UkF&}3)l z{L~X6=vx}VeiiPV&*o~Z{g@=7CMTgFtElFuyqo{g($wgTyI5QJZ=%zTM*tC%4gee& z+dm-*wC`@dsGx;xz?F>+zamHfH(74zpip+kH5Tvbx3ad{`g+T{`=@n#X<%jON$yxi z`|4c9hW4M)nMm{-{ln1zvuO(E494)^MrS?eKgP1R<2WwkKJL8`hFAB+^#wpl$cj8K zE`;*#5a{yQ_zVh!gS!L3r|akDO%7yo4343m!S%=dDb66^+sZ{kONkrcly3@O^5^)& z3|jt+wJiDH;+LVB)iodkVCKNWb3r`|&t8=5Mf zAL55zqNM=8CdILJ&J7IB-?Z5QiKzix^8)>SYwMr5fuBLTbL+SH&%hSbmA*Kc_sOCU zv+OFqkC(%%`Zp(MKy%Ct%-_k^`NR>`7JhPa0OnCWx;-(>AE`UB>)$cZxk*ZB>Tz@V z-`d=FYU!nwwV9!T6_~!E2`~m{XO<^_lQ(k^GF4T7^g}VZEP(Cb3lspJoa&rMF{r+s z;dO9*%TtI?Gxl{3fF~k9X+Odjc#>DUA;`S6U;LqdfIPJ~c0@RVKU+gdahBxf5c=K=ew|h2WAcVZ-4FiUzpZj_{}K>G;AoyuYclA- z2~0@1sYZ;K5~cw3WHW9@1d%3o3V|Ha!~9zyAfawbO4wU?mVv9IT_J!Ssq@`5m#WC@ zU{yd_V|_|_9u`1B#hFzoRDzvK)_I`6B`WxnIx)UVV~{azEoVLoUKaZo!pYT@Z;75QS(=ZB+j9e>#Xa zO+EggqCCRgkSbTSSyTw(sLeto<+q;+Kif=_`qBg>L~tnxOvfQTabeKl z?%4f}R6(3z)NqPOxC!+!P|X>0A;wa$TLcUJ?%T|If*W}#gA$`;y+Lj?5e<_p^|enL zk-HXHKh*|trqv?XfDPU=s=Y(NlrTo|Gn+9Am0BdMnsyuwH{@wU5QNbst6HVYA2i_= z&=!zW=mB0`UV)hRG40kng4@&s`Qy_MO|d#C2akEu$Q1^gvh5iLLIC>VZ_Lv^HEyx; zZG<>80S^vb6sq=j;3*M@NF*=qUvTZH)rmH2@)o1hEUR`Z z+j|A>8;33!`f>be3e!yx_(d-}H*2wl$W1U}oABVUC7nd9kynoZm1@hohvkRZnuVY2 zbUP5f`b!805VwrBe@q3pw8Ry>FLmhS_Vl7W_Cw8%yM$YTgaHO@?P*M&>dhhdPj^pM z*e&3%a0zryr)xImljPnWl6)=guDLJ8i%sRQz2{=|45j7EhX(!tN%FkPu4J5^m|$ot@a3O+SqKQaz;tZGRQT(VV|mS;5wY{7!a{d>7t zSuN2@{UCZkxr}fJx38zQB9|O-o9jNIW>Qr~{kAEu%aG72OFS0NDu!SO-XEhRaP9}I z;<;PWP zeg&|cCJ0ILUNg>8sdOA9*QZa%b;J&*pQ6M=U_k0o1-p(Jc7-A5DJ z8! zy=6HrPH0O#1teYWH$8U%>*myq%)f)zgTO1KIJV&LZN$eZ-NQ z0Y80-h)+S?OB*Jv2Ap*v(GdrL5zm=aqUE{F1ePhA%n}{aS`FbS@`;TrZ!Sl2F3r~% zEY5mz1U3s2#0vUB1W*zrMro-5zXr>JWt?Ln0BI;B(=8}UM1YJIIS7ML+7QsEps>Tb zHAw(uO3n0LwSYCXbi-y*@qjBOxs6%(K})nE9!LjR#QStaSjDkzlDYvwl`{L2=_s*C zSP9|gANxU93p*5gA@^N?u#37Om<_Bz6ESB>T_JBjOYRmMH9DtK3t@$U{`wG?DRG(o z+V~@zD4sYGT7WHr!g!<|F?Co1qHB2T>c>3|*I{2{Jal?O(A%qM%bu`Y7kxpT2djLu zAD0VomTz9Txg@F2aL__H4i|Vf*aKRq>)i7W$B*^%#LMcpC!FPhhBk!Mk&&-c4~^Na z9#?j|rk{F>F2%ol?y0OVOQgU#s-@I|`I+8oZbQ0BkBei^dx1o)fcgPpmO`;=n|HFYM=jC6?IJM}EW5t17n85R!0n4gDXf@>hdYkQeqYVuJnv(wN zt4y&4%1c=QkreR&xn8zZ&OchEEgpU=H4I2NQNbtc=CTBhP@HQ@#hJ}$Q^~EXS=}yr zbdzZjqK<4W1f9Ji{Z2FEY~b#`WV}F$>tcpafN~&mSipd1NmeAbCoS$7sTuS;`Xz9g zb$UIWb-PFRsbghoXIjQ7gwZiE`s@06`NMV6*^4mONgPaHvHgqfCRH&xRwna4^L2&U zl3d0s^6bo8M3I&SZcnlZZhM}jG|PoUvQP{561B)rDQ=Z289!Kh)$l2BBzGJCHv)BS z{*>^i<4a2I#k2?H1IFXyzW_Zz!oMETz_A*#+WX*?@PUFQo<%K9VfhA0znZ9S^}?5@ z!CAvX$zH6DD4guoI}Y?QHCp4TJo+&6ajx|-1+mqKpy80c&nqhAaKGE>@A@FcD8SV| z*Uqs_Nm(+ymTs7Nj{&v?$;zVrH~kCC*D)>~J+hM&(8RKMU#lgqt{4^h+L_6vtHNaE zTQqR(15Yp=pu6rLc4gxuuPNd!a>zZo<$iSTRk%uZinMdkkhd60E0}Z8xSEI%Rb!>+ zys@XD#69C=5RW)9j;d$o6c!baWFdK!7vmhHVsThpQiGR;3V)H5Da;+xF|v-LQV#bs zznJFTj$?a%;aqFWfH-|*#7)+~=5=km30wS93!+S14?CYZFD`)aT*UqZZ>fQhDgUsy zjr%IUcj{r_A*1QjGS(?F1Dt~A7e^6{_*2rWRxeF6=vEWzx{fvWPe8E)Wj{BozSfL@ z0+*%q{l{l=Wccb@JAnW-ZtU9cY76R<4mBM+oS$BT)3MdTG#kww8D1~;AAF5)&=92q zWp)@rI-%Dro@`HTDc(rYdo$wU1mrnu)Vo|ykD1+78p$E6Czh+r_lpRpYX-t^h2Y=1 zMpk!&P@P!q$a}oM$yR19m6X_zm1^LUS#LecnbXSO(+*XNOZ8o0K_?;DW;_?+-}kYz zzY$5?;)d`c6T6@7=j{q8VcZw4=#^8lr^E$;09L^w?EGW*R z(6aT`$kQd_k?Lu9dN{2ja9L1}*xOK!Jc;E%DPh|Qgn#-znZTe|t0Y-;eMIRz{T&Qa z=4JF!b}TD_tO7h#C!+3i44bOL%$C#auL zqwP~1!95bLxE=1u=SUf;Up0Qu^LqS>es-HiQWJc;`VrSt${DdS$XxQ|lKCWK>VbJ>T3ChPtWnDa*}r&GZb*l;UVM((jSY)9 zmcKY>PglGRGY;=PpKIU;NYxJ<7Pu*)@ zVpps~9;<-u>gUcvvW!X&)WLQZu8)+?ZwM&>NiRnvU5n!y)cQ{?OWhZU`D=#v;;P3+~7_-7Is)?3BF{c5Y4|DjD#h0&B7KdE1(r-yCHkEYgXL)t629?7MYExz@ z=%Nb64dbaAwjHT=)R5cul-^S_F#;no2W(aE+S2y~ksn!{D{HpR32STxu#xQ!Qy{f-QZ6JEhaDd7vkRPn!f&`J5IYp z`U#UCu^7@0*0*)8lAW_!l5|fKo2;`MInv28^lAEm3udycV7FU!7C*kH4mEFcuQyG; zHnrg<8N^Ghr+@yKVV})%pH|6R<0gKqTdIlgWKFn zCt9i3tEW>Eihuun`krYq+AR{7czlY2C>EFfF*ni*q%fhT9-^^T z>=-uvo)$YHE~FLY$$(1V86B5*SpvrE!Fd8l5jQNl-CnyUIzrXlXo?X2CHD8Qdt%oC zS3-3muYTv+n-p)^nfEv42cWG$^m-&uAjW*sP#enWxYTUIx)Zm3EZ$nf+EVJuf=|p6 zF^n~Rm-lxKflDtI^_u`Ts-3JnzS-(eEmII8!%`{t6IAiu%bfleMNTPFt$5%q8tQ`t zGAUVNGPrdPL_q^oZq@G2ZT5T4@(~jox7q?$LuP4%nDk=J9Z8^K5HN|TT;z5vK^5_f zryixgCT4T+V!vxO$n%Ys#8fh?n>fSVJx6OCA5Dj$x{vXE{OOzZ1x%YcAnYu>O0#6U z-#RhOZA*9xaWKu6xR%+YivIxOkK90jY8-C%iq+J6Gky2uI)EK7$Y{oRQce#w6A&f!x@Aoe=MUlE z%`fEk?*F_hU__Cr>S{dCyEo{5pU&S46;Wwd>;aiK6|?J$Ff!1o%wV*dia}cP7fg$FE-kTAupce}(Wsn#Q#>D=#CW zTUnMf#$xIODUbdjfFWPC&%(qPC@rn>cP{^K33-CDheRMRIeg8{hq}U}ca%)MEU^K7 zN7ve@-5-i_aBG6wL>&Ub7*cI*Pk$=heemO_(YZ9yqKl9GQU8Rj5GaQ{5&d!ePPJ%6 zmq)R!H6;x?-64Ot58eEx$&3<62|XS(lzaZSs0e_uPW%A+5h8{KprwigyM-Z%;Yaw2 zk|J5?Io7$>VY@+FsuV1Zv*fT5NBf+}3@y4G3ev5o&@yqa@cJv7%!aJVYiAu(zwBxp%Ijxo4}#E%ihEq)g@jP3y1QTIQOnfMY38 zP@V25ps!efFb?}L8eTzqMHM!)tw{esaDSn(W9B(j{VQayXOzmH&=5EJ7!uu%=K?t* zyen)MM%~f)umKXMS@YWl^O)J;qLiG|Xaim0*8`}q8W_ZJm$iMPekzi6g39&hC8Z6_ z#=L7NV#C}bRju3p#bF~2@EHB$pM6st>r0rQgKPBmIY%GD>-yU0S@K$vn7ddCL`xyC zBoK0KzEARbKoI#`6)qxS9}URw;$HaoEF}vntn$(LRS+A0qEsj~!QjL;$nc$#e_&+g z9bes%6Nt0(>p$jQk`$4Iu7>Fsxk2L!W8C7b;;9m zwazjd)l!UGoQLwpq54MkE!<@ePR;kH=X7aUYZhsuS7*;k@)tapfx*;_w2S%jFsPChMA{ zo7emdF$cEdy6{k=u<`4Z#e^wqVmB)KcH-HunhiUJtQm8^tcLjcMQNxaGCju^3^HAu zwT$IWp-(&Za#m?Ag5N8Ng1%^=8Xth8L@LRk-EV=~mpwUQR+3g@Y(-Sd507O2Zk%tm7n z`aZkL=x-EecuSWtOGKxAk+Tq@C4)k#<@ClTSr!D{sR`M;!fI~u`rLNVLKZbeExS_W zbjF$xsYJpk=?3lFg=?uC*%u#79Kh{#T%B#BOi2<0JK`D)+w63s^BMWB*}|qT z9=o1d)GXz?;%X($`;@2i8>(@k*s)>TxIIQ*dYIQ;XJG!taTupXLdu~Bxq!_+#^reC z9EeRjS3^#ZesOQYLn#lF)Gwn9vdjs$dhCwoq`XfK>a@43`+LF5?s8pd?y%(uLnv}(JEPRt#G;!@#K;GX`EF@GH zCP_S_(R!E_jh%P@oa9x{?(ghhu(1z}|N1IcWWpTz1%HHRz6vq?p4ATjz7hKK*@8e5 z{rym)$o2yuXXNUt&y6$)y5~$^crCFZbD(~*EznPKE4({fE9?%tY5f`;=Z87@o)(z> zuj_@^4hDI>9zq%mpM+=Qh5clma``tT$aQ_U;Y4!@UCz|c1=Z*`!SxGSE``+2-sz~O zJD-1uy+MOo^A2L-IB$TjIDei9g7eTmkBEIrm@#Dmi&pDj5oqE0X=~Hlc?H%4i3~cb z?Ig8WLN`v7JVjRW6d;Jo62u}W-38Uzgp!JvDAzZ=7CHo$_}*v{Uj)uXRz4m~=l4K6 zNXdnkIc5~-^{v1ZwsOg>4vlU*vrhc>2Ok~2QX2l?eqJcf-YV0eI&TgN6IcBuL`kjq zlzKYVmm(m%V`p?hdRGg|-OW(ANl1~2@A7!{UNkJj1{e(jOsJP*yB zj&Xi#NXs5G^ZTZ4in z*m&&2MImN7Ap=zo8kyrWmeG$zU20=bl?qYrB9`mwu^>-H=J|d;5js4gln9T3NG!{x z!&H0*KuLUG>KYEdMWf4%5RaC&e{Ao4@f-bu|Cs}D(pogKr|;HWDV4 zT6V9M_=Dw6R&eZk7FJyo3*Z0|-<0M@vc2PM8uj=_u8*+lm9*OGv`SL41YSM7r^(Ks zdc0$(${&ZxH1${%4Om8p=X>M!`Z{foD6JLl;HpPfNsJ-%Eyz$L(ALay(k zd>b0~WR6gH48b4(3%bLKg(=%^?IC{HrMJ%LhiN++7jf)E5=!tYtOcwhb#I0;yeY$V z({da&CWhrL#UD-B-jJ?U%N~(J9(shT7Xog}6Hf+yI`W;T@0w7bu8C(E22b1;BQuOp zmxS4JkeUZuo|~w>L*%4e1-;BDqPpbg)Cb=L2)sBExV0ZlTh&#-o;~)v zHqsJx?Q_;EJ;Vr?CYs3DtX0EjY8F|kH{P);NPR~}v8qYM3EQ$WDTzNS6_EQR+L~t` zvMUD{OlH!}fMDvDcWq`wnopV8cOYLyuQ*E>8G^9f>xVHHUxK9Cg>8+dgF-w`XyaVh z-(C!yw}g1JB!?v3jBE!7K)QeaCbCh;LfSIoi$~Hgmn~eS)^4Sppp&m@7&O5SzXM&~ zr-jILgIIM90q0PXP4x_|uQKs%n!aOP{0$pa?{@$#aI+TUm64z9J+4<44o2$iE?&h| z15=U?7F%dJ48ay}HkM4>W)b_@T9Rg<^hLqNgXwD*N{_bX_30ymrsY8^$F!}4&JpIw zfIEn4TZ{$OrhUx%*vaSVLX;U1cQHcS^7zk_Ribg8vR@B~KQXJi$29_DIx>hpPhAb{ zsw%;G!WiwtQ@syOCk$%tXwFhlr$n!po*RP!Q{I0um83Nvf#sHERh0S^Xe(2XVGbg9 zH;wPF!M-q+b$0!Drli_$M`}V#5zx%a%YAOWrHvPslGwdX=iSub{5raWW8wR46NAt| zdh1IOLEK4+>;m}|+~Lt2cb~pUDzBl7+}B^mD$kEH z!LgD{ihqbE4&tGWy(luX@lcd~A;SwKUI^e#;T{FY@MLo&0vBZsE2RL=rFqlLwmDlD zL85Naa?bGk8JQS%Ve|-w1?O&GA7PLbaV8DiAbLpa7iK;ygONc*YIMxQdEM{{e44O) zH7AE=SzJ%U2^V}d2!0V_`R5pvEiSjS-d=E&4fh6jd!TcEl+P~mM_ znoc@h6#IoBVSg?%K%y>YziqEkk{(weRTb^N!bXIR?&a&YKM3;5nOR;CfheE6Qpy|N znfZ`~M6X+dI2QhR94XBMp#vJoBOL4~iZ%8%yyiitbsj(ElDHtG1_k{ypJ9Ok#C8J3 zb$Q$MJ~=yYjRY_J*-GqHZr~{ILLX#+es|qzU9ekO{3uVyN*Sl}ucGQ9d{<-?@;uw1 z*V$5xIh-Tu&trma@D|I9eIlm&f(Ww_o9_Lrvt98`g8>W2M!Tr_^HulZ1m8`k*$LaT zaQrbzBQ-%p2GLqQI4_qs83(A0x?!vWtgd#|qmoyDK`O2Bzu$i8CaI{!&oX0HHDL&^ z-QsLWD|OTgv0F~od&iR5M39>Ny7LJ&)T-oM?7B8xz`XD2SI|@ZHQwWuqKH0wwk_!( z#fV&+BuB?v?47whEJm4qI<2Kf9E)*7)J@17Y!{CSb`Bm6`)%QbyaZK7o@EYER>=cR@eg~%t49=Tbb`>QzOAtupU8y4R*L`{CR}M`@59cHeyfhq>`oQgc zvN;CA%QDO_8kxL=yu%U{<9-P%I5>f?VW}6Vl)+Xut+JQX!`JU%Z0hLa-)NZ`s$Lxh zuS4ch3XuyY`f0UZf8&|Dvs)EVb^J^dtcF3V>7#7Tl|N+*;f6P#9X5#7l+V2M*7gx_ zVX;WXHMUZq*3vHYQ#Axq+{s!IqX5B>J{zY z#*$Pmk{v_YI>liq<9ah=@-sr(u1BOa;KX+m!E!9DXk0MWsyJD>vn*COQuv6?XlIAz zv+mm!C()>t9LB5g`R`pdM!yi*1|1#_SOrUaX5isco4|gVxq=p`6ZPBH)~RI}bY0vN zI6BpHFq_gq^zrDz(Kc=a|Bv6ahy9wBz4g*yGaQ_DY+qE}C7IkEPW-PnhSfz7yE@qM zqQOgSz0_M=9wXW&vk{dgqG9d6$}4{oGI)@M*2|1)=bv_$8*nWsO&#vMmM})&i6Ri&xNp!$qx!3Za8EepBDCj$1$1wYNRJG)`))r)KI; zFy_j%6091To+;KkFw^O%QunH{_zGce76MUmRBK+2eUnQx*NTS=W4tj+3U9!|Dc-XT z(s6%Jzw$uUZMa`1``MICownYtaM1M7+PSGu9v~Dlt7Fr@>TethnYdJlBzTOwU;QsInvdxXz2q`W^!6zt5=9GXg^~MQ>c&;tcc>B6xIVcWGQ1BmmvqMqP%}fO0 z_-TzjD20s@p|3lWM^Y?A6uJcH+%~Uq$6t~2Oeka3%@)FOO}ZR>k;~P6f6i-yXwE?n zEZ&a;)8Tv^;w&+rnAg`L@Uwt|tf7hdVRukc(V*?BSP7pa9$dW^*wGj4zxc zj0hz=Mq3Uu$tu(vIR>keezIsH=u-Q7=tJ6T7p2Ww$ zAm4(AgUmz5+yWvA2kSov>O(49%*Vu$n@7|?Vue$`D7=#lSzKR?O8o66;=vBB@*B$@ zrPd>No`Gz{;bMQjA}+#4OD5ZdXKA#EfDs;slkK->VT^f%sIM54;raR_I0@^M@CEx# ziZC|wu7nXs9oZG;4Jr?*(XJ;b0ws~pCkb5J!|K4Tcph}P1*ztVA&kqAVQ24G!^murMLXEnT_?N=Q@0>MfTh zIL3$h-`HsDwA>%b$5}OQ7mB-T1s<;99RdwnZq!OW0jXgJ25R_&Y6hhTC9bN0X9EY`PW>%{2f(=<7xF~VuA5Q)%? ztj+1f_$=&+OkbwJD>L9LOrUJrPCF3akV%X}A=8mqae5Na;*_Gmyk|6zdsU%xp@kvR z1>LKxZzdrkdG-6Qm?Vsb_Zo7})8l^gnuHCzRg%&+Mwnf=AVB!#1^8`}t_IcnM|a-a zo8P(^5KM$^xCI}-@Nk|l$ZEbO9i&YcD?sGQj!UeP-C;DVTOMgqwgXWJ3AZ;kgZQ)R z^5B#h2Pl`nop!T~0{HB((@dtpEU3pN+K55vM3;YEsJ0tkR$3?!;6BP^D}1|8Rvm3Y zwQe$XCSc73;z?*SvBAB#v>>CTn&%ll_PH90DgkwoS-OEU$A3A1$$zSe6$~WjpV)k2!&JKVm|c)xLI^%O6>3 z>_q0M*%#$wuPJIm`G6^5ZG(qj%60@b3_CMj00;y}xC!|>IuWfS***s&$6>AV{ji9% zbQbQ@p~QJu%ON_B3xG(d9VIvH1S}sso-@SrbT4CQNRNhBd^v{weXY9g?N6ec9fOuL z$2K|&>wkf8c5R@3Dl@ZjMRsQ2kQwgJp12_5LTGJslO0%G6JiKdS*MIXYE*#G&wgYs ztZTwogVgH2L}t6Fhy}m)$q^e6=1zx$tvCA_g!LxbeC`1hSi0)P=g(x< zG3Jw?nplRVrJ1Q8dc{_+s5K?49MX;&BiiYmg3c}89^~MZdS{fsru5A-H+?dr+aNd_ zp)if0M0b#PcsF}~Q&$^22d2m|MWC3hTTjMiu}#QMFoW@XHs1yOZ4-q&uJ`XzJ`NNx zc*;G8hD?e~-|({2kjmoUcQ>4iQ3vlmV*A}32-A&+>$r%}6yy6pze;v7i^R|yX{;EQ z-+Fu=vi2pz7!X6f{h4W_Ke#QudRmdL`>5Iz&=7XB$Z|k~&qqhy>V)pO^@CIo;?#7J zNi`oDUep!T#Lg=;y-N)WK3-btgoGB%e25Eav>19@pF*8kAdTZ8=!KB zCOxQ;Wx_3*!r0lBI|5#bNfR7B3A0nr&$>S&FuXJLYj_tBKafiRgvpCa;2})NqAxLL z^}r85{B@ni$5ey*-72i`8bX2WbhhK+$FgDy)7vBU+Y=wsA7PTU?k@gR21$ERrNmv0 zubl~!m&_%8FD)3TxIEt64LSAlS(r6ha&Pf6t^K6E4>_{bcu@#5!u>!) zxGvKay^_U1W_N!3_#(H90KPCWzlbrv)f!|f#{KHm5e#bzCe$4H0^OoTNDZGedkvMK zn}3BUHs#Qx<(+^^y67iBMs9%R^W5KW6efarlhl$e8}v|bd)rR{enrCZqc@q`{cw4X zF|R3C%Be;wXoy?3#;Y~U7afT%j%*!;TkV3r+#ohF91)*#lg1*lA(JrnhgTbFoWSOl zh5~Ch@wzw_=-by5^pSJDbjbA|*hXI=RoS`4LYebIJK;bgx7XJ@s#9Ip(GG`5LY2?`mkB3oAq%{(21hlHblrGSlZWyMO=q#tP*BdX1f zR#txf-C5}8*9KwSHDhl5DnNerO*DGRPXZ`WV4hqvOv(z&nMvnl-nwo;e)7bLDr|5Z zvo%JnB6(byJEKs@GJnf_;!dOp<90u+67RTVDQ4$*vWBFm>7=71 zniH$AYxM)W)b=b$1CL&lA(m3jtgA|)**i}`_>lmX#LvpH6CBlMW41DY0#-HdG+Lff zl<^^_)e70;tfyE%LDUqDjqx4d`O32gop0%$|e=&Zb}vzb>|}& z-m1b5lX4;g3RMLW%6C_F>Nc2L+>(zkef?j$+Ljc8%L!YtAK6Rb5-At&TVwMf&~HS! zpsyn07+#XY8G@SL`6B{K&(Y?=589!a+n?IM;*l!d&he4c%?a_XF!|#lW{hwDUiJ0< zG;0{J#;H_fE>o^K;LZHdu6Oe05u~JW3M(OvANqK z5wHejnoVIK_cwpBHsD0mAy>$83?S_PAj6kng6WkxpMPt#Gj8h;eD+ zvvb+1BDi_Xz?j9K`a~q`%y!5TR`T}CmrWS`j=iC8@cHT`@y|{bzk1h* zxmZ0wH*@SHWaVkIPF~?9Vy=Q~DOJEZ^Kn9r(c^nXE=#mn_ZERU%iQs3aZxpf9Kj)i zNSf6s+OJtJJ`8X)$d|x2r@dfv-WL%?mNI{RfjVD9x1mZG1XG%n2-e=SI*^t1O20(H z;!GSzC<-&hTlq!)E^X*+oQ%}Bn|;QB;fZ{g>Pu>x8yzSncZ@XU-KPQj@>s`Z4MVh! zjmcHP#Om|}NzX(|sI39o+<1j+XgfF)F1d&UhNooAdQDI8zzv?(q6(;xA$Q2P`fr8T zI8S_44YKGLwMVs)bqpBwJ`nDZ0sWm6ATa(MV)p`scmya?_d`um^Aw9U)lw9pcjs*+ z(T6p{$;JlDy1JF|Fwa*rW9JX+Rw)>GIlkN)F}tE> z*a1n{qE)P7$pN8akb++j$$a%umt^lbn2a9%G&631yOyA+oY(0cnP09dvkMaX2zG^g=HFCU<}J;DRi%9IOEI`=(EZg|GK~UCy4(*?pzWh>8Nb>!gFq@nWt7tJbkGud?p~ z8j~3${!lxeyx&l1f7Oe7nw@>D#A0xToMz?DIk*@sb{rUbRH5BDSUSMCwAkHUgyNR; zg79m3-()~;lu1Z_7;5bRi4~fNeW-YB#y=jQkw>W+TClkcURt7aZ`C@0cC1(0yv}v7 zWP7?CDMZ4t1dX?M>fT$X^LDquq#TGHYKJzN0v=PJVHqdwl#L=yBe~!;xrU`=Ze1(5 z4JRW$8Vu5q(>4v~?ti#aXPlLvZ=;vWg=|(<{xxGGRmXq8$j0g|XC=ufa}FCpX}ff~ zS712|UbQ9N0n(`PjkghiIBF)lQt$@4&R`bFJ_Mo@Hqnb^=&E&>0%pME^LRGSyMx0~ z*O3?9Q$|VoI~LC5?qzou?3>a0!E4O%{lPCs$w=<GDFBLp{QLR}aZM97| z`xY5~aqp#_-(_{77*(?zk(JzK(GleebdaE1l->wg-nByMhk{mew|c)1H9HM^H)8;oXS`vRxOtb09$OCasYkct{9#mWICE*oO1LeuQZ^+eiKt7XO>jCL7R z0%}&TW{vEuWe4|r3Y8Hrkxe)kUZmKPK(zo)$G>OjS#f^DQ6hfqWMr%n27KaIxPylH zjjCh9JU6ly{I-3>lBJAh>5GjPIECq2GSGDMVr-{Bu1*bU^Hf1?2dR1g#&$!2b^PM= zdD9Stl$7bTm7vZ8HzqgnIjcOs?wBV4u6Xhu5cS5K@MUWwgm)R=G;9s)PMFFd40 zA94OlDWqLZ9bI4geTorP&e<0rS5f<^Y@a`Lz+j@u7n)waB3eCESU6cubOoZona`fI zna#;opRlkuPC-Z*tsvlsAs=47lfXsFzCa!~z=G0u#F2&GCE#8Ku!Z2F1o&D1@^gG5bpE& zQ#WQRurEqkwXGOQ5JSqs3%t~qK zG@U36N}F?eD=I8lgRahtvmp*q)$-w|KIKwzZ`;`U{5)g}@{)eWocl~Y&``XEhYZeg zuLWF@s3dGO@0OSb=^1JyH-R~(HBIPqT?Z7TNz8*BG>ey!Nq6^TW3vL1>k}Pb8|Nk7 zmHFDi;*m~|ZN{EhkG!JM^i5o^U(6Po)-hM!DGW!tsR74xU)j_==Q!`jSfb{Nt8*?a z;I*_$hO%9yqJJrg$d-gd*H%Nw%pv{Kf%5xaviC)>;CG|duqP((ff9eVDFV${F6&vLh zX_9d?X*&EAy*5_dn}84YV@wd=;kH$8faewSo#gG*x;*yCJWtZ>X@ye%4EAQ7#{zbY zwm0f@`?{(DSD4z=`t8iEuzk5 z;Y%5~;7+ID9>(~Ys91D8U?;?8;jYfQgdTa!NP|=;mC%Ir0!AodseeFtRt`_Ht9t9R zc%PEsuoj<$EXvupaR|o>Q$OZ6`)uCTb6qlv>>jO?ee9?+DQ9xjUvQYk9x-MH%WwNi zDg4-8H|#4>>71W1Ig)Z}!M+vU6Fnll0a$d<`C6~@BTnpjNKAPb`@lF7wzhby`=q6j z^{Vh9E&?71Ez4`oi1oQ95?m7(GFwHPb9(Ai($J2&$`Echa*EHxREOHJSV)Grv}37_(FBCyCUw`mJs(=k}MZk0%Qn%i!!w#BZaGHo-~_aMqn3 zIr?qU-kkkYHcm8Rae^RYBT6aCW$~!51EwZ7j)VQAVs0M9O8ZtO_BAfBI@|@lnL1?< zz1!Fv3iYe`sr5l`Dc7~=Ui;76OSAF_p290yQkpe2i0yr_*(cy5qjwR)f@950f5g~x zo3LQ7F*N!mP|&TG)K_=5x%QgIpn7V~^hAscO1xLb)CCRxh-GFcQMW}`-5bg`i{`fN z<4-~vaUB6^hE#9f)f{=>DI+$HFGT&-O33o?!+nTMo3JcE&NkV)xlm@3ulg`i17pX(XK-|wyzAPBXuTgWJ39ZxBUYGbA8 z!qM>M6tOaz9&1lI$RvJQiwp_gDL0{GzBxMG-*Jqnno-Qu4 zhFh0!)>)zv{sk5RK`7rIgqz_e^MSUTOUUC_sq;Dq{HV#GW?wZd&Jf+BlwW>Q9U9JN zd(s(B{^+2TEo=Z&@vtsXG+)E9=qUAF+qA%%flSWsSZ;NGo^) zl%%zsxf61Fy!6SZIZVx+X&h33S89Y2Z6cKJawSh)aUuumyKqpTMbM+>v-Ca<$>44H z?44tW+91i>VE;VYQ5%DG+Z6(SGY5jd*ULjSRwA#Q+*28-`Z~3t59b|0xh*^s{lx)e zm|KQJAOjjF+y)ad8Md$CjyNZQ=sBr%;Fy46g<~D=^;Zw!WbHf`*v0SfCV@sK%+hev z(LY9}!fwfnu@3vAp^S8b7VlB=tFpm`_#3zme-Oo``D4FteHrBW&Ad#~5dTI`OBMpg zgu$z0gtXvf)$Vy^fKpxgr`_*3Yfe4;i!SRS}$@I zm+#8z5fyoHVLsTXT4<76N5*PK5U*VED}32gVbAFDEg@xDmY`rlec^rHK!>>;k^a8H z-q;y)27*(?UNTsu+CH*+Yq^@`3Un+aG%{2YeQ^m_ODZ?qS)B(EQkd(1CPGI4(ha;hM#jbQCg~S47nW|% zVAth8ToGDp6C^I2zH`{2(O|0$Lyc%R-}FH3d;dL~DdYSx*jf&n$lzHVE7h`q@CvRx zYtL$v5KBNk{IruMT`0qxgm!2-@UR9L8b++VUjEv{9cMiBB%&DS*8 zr>G*W2%{9y-&saP(NU-m7S;XEW|%LpyP5r>f>3)1`B-xPzOz zL;L(0$t`|K0%geDPA$a7s7+3N)jF04yy1yY;Z6%2U0hRy=;FO{((>U)s`&$%B=KeG z??m>magQd{^NO3cUkc{wH~(K7*R)Ux!)P;iwr$&-YO-xNnL8VkZQHgz*=}cRs>#MV zH{ad=4(nohTLmkcnpF=W)E(Cu{7_lL-or(^yoV~X4-$_*$=1$7`)t; zg>_rw-(Ll-g6XcXvzk=-FXHEf!AA_n&f(U?3xV6oC;Gy&EUd`@`gSzoeoy<~)2#wW zG`+9O44Hnrt^PxQ_$=Wk;tt&rs)E!@g2cC0IGJ$ zr$*~uZOkJ4IfXX3HgKyPfaV`x=(9Ig_F~PZt4ZTXaoidPtCsyetPeyOk76JAVfdY; zu%IfQ4FgbXoO`}Kytc1AY-Y!z!ZLcQ&~|D9YU&B z(}zuqC_#_ub`$O`#8Z?e9iXXBA7)hR(l||BWwR&$D?n_adz?m*5}!J6h2)-jQ4~h9 zJG87JV{3d73&RZq93m^5p%n~ex54dzyiMFb>d~XL{jLE|89Bbw;OdXyOwFTd9BLuj zb9^l7i(!W`&}tV}O;pE;oP+&|;MNjOTuNm13?ibRFbxygDX!TH84$ z`Z>DgfhB#nAs_?CdD=2NlayS#t&PvI!vlxMhR$9A=X$4P=?)=Fxr*OGfBXcX`w2IJz*^2d_+d!p_@~?G1KC2|!l5NyQzC;x z!B|S4{|6*eclB2b%>&>DYXiiEfid0oxAbWF#3Ld!uBa(Qy_-pVXul3D`L_?^vX%v* zYxwrbZFR;e1&Xa7MB=hypjYzp2&QaU<>>3@FWizu!+J~|=C7TzEcs6FLmDz=%9;TT z`M$(YcS7otxGh7C|KS2xs1F-(c;Xa*U74DYTX6Uv`~(iAG8}o$z&I_;<#fW?quFbg zYcA@$f@z1r1*SroUG=1nQf5_r*J1_wvGz%XGQ=(lej9WLYaOcKawGdvY8eONGU-|f zbJjeE{6i?;NBQ@u&+6?P&Pwt>dL!8zjIXYE_m_mOk@pj&pd}ZBjSJ3f3w>!VNl{U@r)`DGohnyI zY0mv&1A8fLPu-{^mPs3lXL$ z$`hDuEk6-;A?3TkyoisJ-W96VVrY?2hX8$o{6wyljh8C}iYs!Vnxe8iAD)jB4Yt9` zvnzx;2KqUG!daNF)|7S!#(!Q+SsuanTxZT^riI@Jte*9#faxn#FU;t7Y%zfU-9p zWX-D1y-HNHF_9F+^>3#cNPoPWf=*fbL@P!fiPsM*Mw_{0UY}DJHB?ba(%sMWz zoq}9H%e%GdJXA!#;D2)7$&0gCt||Rtg&E6Ro-i9Lpbrq*=iq&0VK|DRQX0fp+Rz_> z(Hv-zIhS|KUdK4esc2tKE|65)^-c60-g_{*N-6L026FB#8p%z*5x)}gU-{wI) z73PJ@X1X^;pwy4jBhOlbD<>&H1`Du#e;?GCHcL6B8qB>x@oSL@+WSJA5!Y0^3%_p$ z87{qd9GeYIn;&d&vsDLX?Y+wtLdtVKtkIsIkz!ycdGXOr3yCG;l;2Sj5#e~s*KmK= zl&B1lSr=BF2F}EJp53x)>7ZV+D=qG(+kKGT&{5g7K**4$?hW})79%6Ii9p;E4a`QR zbQ($o3X~kfAIN$2p_ma$dpZT0?5+e0;K7Qgap!{LB z|G|5$oeGG%o*)d;C9yaUDc7%+Wp2E=^~Vs2Jx(+JYI zEVgsUzdEZ*r)D0&;W8e7bfNjNM?Yp@>CPXnzGKKB1H-i?fiqZ8urJAOzJ5@5svLZG#A}3YJ*^|?GcKGxKzcUA<_XLrnq4ontSV9ayk4eaDv~HKdxJ(V9ICfLC8k7u(ka4FjPTjLFFk&p8l7T(->)dZPq=wcg3o?ZBS(!5*ptoR`N;bKQ)2gk&1CKT zsRV(duz=ZMnIob>`w|hagIpG&>+?9ddM(onTzWvK6ucs=}FAITub zxV+Oc@+UK}wj4d0Y0i&5;6iKOPqN38_0!*ClhR?Aiea>EMozG|1G{uhED<^p!!Xcs zYn)`SYrLCleQ5Um&6CR;V#wAb;*+Ct7Ec{?15xr^H?>gQYJ+tA)prFFTjW_t;F_3& z@PT()fg{PDU1#wQ!u7%l0@b`;$l=VUnQmm?yVHRlwa|;oi~&l{e@>b4Z(eK9)WGd} zOtJawo8BnU3T(>>oB_KbGVJ^X6>L~%b?cteO0}HQ58WTDhSL($p*$t)5pPN*9<-9B zxTrB2(iMmMr6QzGp4KCuq9Djr5F`ct0)8Yef@i7S_Vv4d4G^>IucP6#%QSB2(ur3~}ISjw}ORS%}m zKzcT6u@3=x9Bf%9(3G=mj0awjg0eNjxDCvBhT&9xf1^mua|=uV zd44S1aoy9K-iav}`uvYgwH21L<@x%_U&C?|7r0oGE?4m-7MVCfObZ5YW1M_axhO9J zJ@w;rQ1nW3g&l_|^8ofQ$Z-6fgDD&}Ecb>Hq8?wM=+5@m#7@LR=O`t=kG_gAb((%v z4c(tNwe)=z`LgIXAA2N&Yh;AhAZG5TJj=0#u4?24u<-d09)}D$t799Z3@52e=*1Dv z)^b2$hY-&qaF)258RAM2qY zd?xrMz{wP(70e`l&)MN6g7I#ZK5|T%c`;{ps%j(dmRVmfJ7Z__Q_VF!h4Q|T`=3XX z-zz()=6I8+O)nUn)2FDeER`@mpF7L3UDOVh1VMD@&Z~Ba>54}+oF~>TlN!25s&+DhO8>E z;#m9P!EKRCuR+KYrPd_-p-Le_n;=sO84=qjKjWS2qIic2^Bw|#6|9r({p{J@X@hVW zP9CIb+6+)vLqi<~DoxJ|_BfZ>C!_h6Cl-+@x9jzuls-gFT@Z%G$A1m<*BdakR{=qdL3gNL#vwb9H zNZhDzw*8Qnuq=HhLw{Wu7V$3!zq)bN{j7OKP;dz*K z-p3V6k(TL!kedpB5v2%2hat1yF%rAk>eA%ki zY8J@81l2`CQbhej!?G;{4G(e5|8g+K;u^mF1st+~|Lx=nbfB+leqw?%uUQ0sZP6JiNfz}aW?9T1rMDhX z_0lGHwxl_4f$c+?0-`NM3JWA|yikcebXl53FJo?k>5TU2U_I zQ0dZ{HlxV`V0-N@4%oXc&k0TjOH{4p7s%NyuvDAmaUdHysoxg4p2R(o%sk~hm@mZa zWN^o7kdewm2k{*LsO0>14$m1c`gaLt0w*m@3iT~5KFs`T;G{HamPjQ#0!DZe-QXu9 zr;uyp&4av#=Cd1nXEB3r)~rv$E!k58oa$OUjwG9hEOo1Ag^do*h?H$w3w3L~%uu_) zucl@_xN#>u0t^mx(}S%&_p$wuSOLpg)i;a5OQmd=<}z2NJ&=YzX}QeLm8I!Ok8-AT zrb)Pqj*_Q}f%!JMHwoWh;He6ZU<&qe(o-+0Dj0v;U+9xAgdki-#*4H*pF9?pliYy1 z2JF&83%3)7mAm|Rz4p|+k+DKJdiK*Y9SZQ3Jb+z8`__>>Xl?5Y=E+S>WN$X(9sM(! zNO1~R!_=}a(@bh!=JHz_^PJyx9`gZ=Db9mHaKqcO4S;w7yMP$)rki_;qqp~+^x zDX=GqsgIEaS5LqB}rawW?;n!^ag?k(ge^yc($pl-`Wx1DI%NCc!j< zr(4(wBscBNC?_OsXNVL-eq&MrEt<7}0YyrvdJi8A(JY?*&#<;7D`Pm+@OXA%r$Nc( z!pjgJwiDayEaxLjWuCfiQG31@;p9GIZg>h4j1DIeh>{N%aq7LVIYVyy%{m|ZtZ{In zEF*^RL*Xc8FqiwxpB~zT+zzwEYBzi(T2$U|r(ohHvQ7cXxzTYbo4N<(jNVQE_}dsF z2;$0=0A3!Ai!0cE>hhf=mHwKmhd9C@{BX7}nZ|!&guuCxt12KesoBi{AZy|PTd?98 zn0Qx`q*X_>aRfRU_<;b{+d7^v;We7m@Ru@^8JA3aA+l;xEgtFuz@>~7OU?C0lSd1^ z4O4sgN%iJR}$?IpDO+6W(#duel8(O&Ul?84A_mBk$FBVwd~u!ke84_4X>%;zC%eR%a_=`0NIA=p!$fJE>}f$n z9iZw)cQoKZ3U7WIg2VYZ=3tHzZYpVKjdu5yh1yzMuy?4%HG>+AAZ`f%DB>lLl+gK7}W0qs>y@n%)v(h04?$)N5Oq9e{2++IGR-Ia2oj1vl7wW z#k(~uj2{#C*LJ09b+!r%5XI)H<83=C<9e z)Vd@5d8|d_K#06CVts(J2#d|aIWVfaFPaSaQ$Ql z=mkg=4ylz}?|#vwZ+NKY1@o}rjogj5YkBk~i*2rvk84f-RwxHE`C;V$yjthPo3pQy}iFSRPxwN0u$4x@q z@c1q`2B9}Cg#3nTK{?7e`++H!KGPd9LvKs7s}w5UJMNQ2+i~KxO;#ickcPAhw0*A6 zCu63(w$??d=z~5VwBLr-j%!M^jVavsHZx=G1Bjn4zAhGWts4APFdRCpEsCerj(?QI zu0rW}X~aFaVUvzV<{pS1)$*6A9T&t_kGKe3~shWHgj>V?+C|dr*3?b-wO~R@EF3HI@!_xBzq%Q*ibK(R+;>LdnOei{k2fL2(3IE58 z99fMz{3h`TL?e|51WB)SewVpe`$V=|>;f5Yfjx%|iW^3rn~Y#XWwcz|mE>o)JQl^H z>sZeY!Q*2%(62$90IMO;%<~742`?<3uY6s$;kB2pMu4KZorLAHEBw<3MOIj#Ro8*f z%?a_azh(Z`=5VoKl;-McvtX#Yo1wyCYeuY)fmyoMNqwn(u>vz9qCe+OLOv6}Vt_83 zGB*Tbh6#cD+$-P85Ftp%-E|+NxYBpviEps|I>4#mv$9g8pw5TEXt2YX^%SBkl`54Z zBvvIp_S;u%Z_btoct}pq&EVa&+GA>p)|lekN^M%_RcUzP5fbo1seZ%OPeip*o~srE zGZVgTd-Y*$aMUv^<BWlyTIgxo3Vn!y-s5O>bhm;KwQNL=4GoFNaySKDxF1TDyK>3da=gd zy-c>tiRc7j;VtQQBE{!s=!596#oekX4);ztR33e^(fnn|zbwts>ydRrLm>A7j!7_F z&j(^8~Y5uV6@C4*hB1SYK-5Krbw&-T+Ed zuw|nkSDQO<7{Yg%GGPZ>$Je$C7H$5Qsp@8}cLF=QfAl&;Myjj!Vo*`p=;Dv__XzN2 z=W2&Lh9xg#C79`ntMWQG7R!7>WHEt)GR338Dq9O=zu);{k56`YYZq>J$lgQtT0ULR zz58P!qP7lcA?nnbi-Q-$F9&_5XGgq0FksIpNjQbHrcu_STE+Qt|8-fd87S+e z%0O2K>T(Y_4(A3Xx{OS;TLw3t|3!Io+jN;ZA@$Oq8ePVtgob_WV zKIzx2hvxgv%e@asfG-idAQh)5=l3hKnGdL+BU3*t>t0qI;1q?=Tjis<|4P=1{V%BE zsouC4tg-B|!4e_tiQzVVk{p_ECNTb0*CRVv0VqmCz@gPS&e}1%p3^UD-=bRhGZ$<) z9!rqhH~g(W!e*;fJs^?-%;ahpw)*MD&cNai4HKWrU(@cLvpqruWiXI;YqdZq@c2}J z7D(8iV)V#lu~GO_Xc+5XB2i0qks;7L<;^*}4^?NL9v_vI;i@2XFAl5x^a$n%vi^cx8enJv zKRU3ecqNhlh=;DJ;ZMS6&;m~-M5YVNw;0PnicfkJyyHwR6%`+-#}Bx~!j7uO7u3fm zF7STE&neR zpMimeh2#Gdir?&FqK>k$!AeLF=5#u@O}M?C0E7e#GYvy8#IEA1;2_gO{TwKzhMG0H!#Wxg8{cdJyx#9RQ~$W0q!s zhl~qOEKTVvx#(tX99+Vg-vOfP;%eeqsJM9L#S~CKCk?;{PR)(pUyHy#xwmR{K#L8z z`(4KzB;VA^2@5HUOUp3_`!8N#Kj=Luhmeggyia(mC%G|zU;h_ni*p$JHzt5SbJ#!* zD0{oM_V#+Mc7B5^1T!iZu}{9YrfTav0JgA3+`aQfOk?;Dp(yCJgK*h2ma|&0UOLb` zLnAmKA%Wh=BLZJg+?+D6KOPT*Kj3S~h;_R>*T0hB`~rEs(e+@0ze^S7=4AjleysW! z5HKK_e53<@^yn0QPqDY>=*Inb===!$8^i1GuQ1ELp;^CXQ@#080y(lTD?5|mXD_qZ zzyUgWIlsfxyRp`AfI`-J^k;gw0svCjo^!TObGxplz(0pd$;wJeiOUP7O72K_X)>q} zLvU%hCvcyM-<`QdQ4t9Ieu?;fq+`%~UP$HCYl2t@dv`%t5igm%n3?1``Uqz5F~7x@ zT=YY*SI@oQDXjX?le<$|ZR`$be%4vq{1J@56Fv_4|0tKB0f6j)1mpl?+glljuWfu5 z@7E_E*FFXPqmzgS5Op9MLIS+3DS4&v%yblUs33kEoqk?jKU{aN(9xHueFGynFC-KrOET>VDMo>`Mf7@Qwfgs zt!luWei*+J=X{TTaI7zQzO!Fq5qM{8>Y^-Rjlgq1_zc)(SwEQ;DBpwZtEQGZ|%PkXJ5KhpXA@X-2$w=xplktKfXgbMSpF;!36RFbZ7Vo=a6xK z0e+modkuiukCK>`buH_@=&rvc5Pq^yNqa#L&~=`;ZQzUOrVp^#IxZdUbrMX=VS#5uEi(0|JD8|E9^6?%9EG zZ1Pk8B=$ca1%5O9#^@ftqWvOLochK*)}CI+R)ac*toc7#QRn2m{7Nc-4rE#Xb9dz~ z4L=|Ge!Rc>iuz*nzvd3V)ml(ehIOK2)8zvAMQh!KgW>s?nj9aTJicpHzw6OWb5v#+sMGc;|fwb(TPh0Df$dy@LEVHXFPMW!cO7YQ@EXD$)k z7xyFyR@)_n9EB02KYmAFIP~{w33^=xo+UxZZ z{Uz5)vkDF6(LEd%0rn(8PdAf>7K9W{io@;XvTxllj2yl#z-ZmI_X#u_@l=>>Pl*zr z(?VN(uq$x`wxU;Fxt7Z<20#3&;1C!=UTv>~8QToBLV#=sgs7Od)?jyMQ@7KE7b)`| z(F$C!-e2M*T0?RJm1EtTj(iTGdA@jeIjirL5NCvhNL0V{R#PqN{noA{$9J1nbBk<` zVit75XF(KG{N8Dg6g%K|op#swjY5$!FW#k!+HRaAR(US7YyeHyL2%xi6rh*+erPFk(=y>1ewtO( zjyr{q#l?p=;DxNd;ot5k6i7q;*aLVx4&HZ%*d_4>_Ngjm%5I-XF1yy=`A@3c_=Iq5 z_jbO6qblM(NpP-&nt6NnviPC+$b{g1LS5u>d9e2w?kzi{{)}2y`n}NEeJHTHgmt83 zady%8#nqd(eX`SAV9cos%KYnTz7 zN(GW++aC;9H!9y5=g?Lkx$92|0^N@@YF$uDf10YzVDGlZa6`Z=vebZ@nf=wHiTn3b%QbM5B^74?|7ZFUOu0vaLg|xa$!`=&DQ!?3| zz1jD5zE)RQlfvFO<^jiu%dO_L1AtmHI?-Q&LU48*9_G-k+l!KAZ)RV)6c;DdO7Kpr zy>Avh&%`LSQoO2)rB^f%UPk0TQNfF)UKxeTKivkAPOJk*@00(k^~M!4f7B*1;RcdI7dtAGywM1sryW{-KWVC0(&eM1VpaWrM3<9O3@;i!YP#)u=3f z8(zkn1kgDW&G-Py_P+7Z{4usgu2~Ij|2N`E zh?Tr4&k9uG_tIx)q6EdQCmqw-8t;a#nuK-v3LQ*B-}2U$+QdE(;s#}Ma}uzW*CpGj zbGdSLjh|eZ_Ph;daD7^<%w$hO`Lq{kV%)87^V*zuT_I+Pmd@Ci73egFxA@ z6`s|Se25vUGvTtwZU+l^{7jhz8cS~`ZK3C`fC_}YWDaTaamU2+(lEmvp;Rna*Bg1o z3)`cU=Y-r@RE>5TR!geD$whI1)(F$mV?A{O%%^Tz)#1{*K{J$4-WLVSxv!6)bNTyG z+y)0*3FpWRi{k>K4D@+%CPb9xlvVTmC|M3Nu*gEydj4HGr+(`W1R?;fAQbi_qu748 z(9zV`y=+s~-(=^ILeuXBuf99I#s%TBECO<(Pm|89n`kay&rM7K|K&Bmg#!0dk%SyL z<_@IFb8>%7o<|M7BYIifwYmIV#i|fFdv9oKLR!q0o;Y_?3VrM2uzrEnAe!BFYgLPL zE8jF099tZ?*sluzWu@#;1$ERx^MgCX%zObjnRH<_Gryj7~xjH)pm(#v0lgz@Ja~C>8aoc zk4fZJ5PC&aR8YvN0Dj!8xmIHUXhs0_yfTN;=qTGfT6F@(@zHK>G=iU4?fzFx#J4Iuw*D}KcVD*^Ufv5aVtvV(}>V>9{dMt>|x=Zvr7hM;=4QCZq;=@MZx_mA6vnGX8?A?o&EeX?2k zFu8*Zw}*GY#f|PI3u?X7wjcG0!42}N5jD3Zqr{mLJKIy7M$fB2ak|au2-*8(Annhc zw$Tly2!0&5)1Cmt?MGi3RP$ZxoQ36!-o;va0~AQ(6kGloXM;DN%)|N;AJ2?W5xQf~ zD{;|wVB<+4^Vn{i8&lpIfM2HQoaz#aSZ@fKVRN^GOu{ag8U5>_23VL05c=~ z=F!$b|FHQ~S9Tm0VB3NH@jwg(V5Z4*AOP>V$I5o($XrYe}Cb|y^QZM790vJc| zKZG*INNS@-!Cy|d4`lb_Vngc{87T2ipN?|q@wgzGKnJ+nXC1*pI3nxkBNQ-}Bh*61 zMOkp_vUh_cgU1=nmG#ivnsK#d%+X|YTLE=CfuSRDVW27GQzmQ65M^K&(C{zZN}Go3 z?~>Zx#%`bNf*@J!`32XvY?9A7A(y+j^ek;;#KB*en#shq%n3~@8;Ov zLou%JzvFjK#`MxuU9N7+b;4!Fi1;*}j>c5dF?t^hD$XGiS-Uy82j6JdE@98zZ%*vh zCT7q!Xq7_oNKtuHcrB_r3nC^PwOw*&(`+^{!(Z{Llor1{K6?1&$RtP6DP1Z%l_1E> zlK;qUr}1R=*nU+d^(1jX9D)Ox`1_11JjR@ninUrT(4=fgInzK2z8xyegtZzr@UfH$G_e&Z{x-S>*4NrfpTf z32K9}7nYXjF6VLgQ2dt~XleggeE?+L($& z0UI#duC*%gX(h;l9;iua78owK-riC=JdIn9srG2$zWEKFkDMgW5LYDYib5Aag0szc z+v5Pk$-1RfUwj}`z~)a#mSL3WTT_|wb#cSdJMo}N8M+t?6~o%$-Rt-kMqITG&ZP_t z{_{}86uAZZ5pB8Z2|v{D=oR7XUx(72PBhlCf7ZS4%_1N%W5WlmwZsA-;Bx?(a<|2bKxn6;=aI2eQi5x?**)fuE8D zCCh@qgrjKfDia2(RQVyVAyLlIj7nUmu^H@1>i6R2_ca-Dw&E(Hh@;q#^I?kI>>X6K31e!ii_H3 zxl~g&ZERg1s05s>WKQx&S`oSg67=iO-X8sR7%*3O({8&jP{nU2r{VJPH6$Jgam#17 zd0Gl7#_*{Nei!OLtIV}_i=rpbf*Mtsd5bB(1Bwz( zdK*2I6e7M18yX0bWj;atz$#34jX%bbxKWZZgIkh>agyH&C!&YW+vw*P_>t!6>APw# zOv2g{u>jGzb!YmS4o5gDd1N+>ZAAfs!zfqnb;FP`A4ng-4AxWwf9U_QEC5>R{^CVg zmM3hYbnv|E3{+^iqAmoL!-9k7M}J&cZ7YO+_}p$p*ITJYp;6}_07!E`-FxH+6cl-O zGUh&3T84w`xBEjicuG2rX4-#r1yj6B(+qzcIN4|+NOUDonCwCfJ}|@B>oB~$xvQ&F zY(IJ=MT^RIddERV?-$)@KOoo!eMo8q%-}7w1w@v+aMqiZgj&0})#sU@ znROI@Z;a9ia(KAc;Q03+7%eDCHaejuBJc*-j9*!noM&02zYPpXFOlJv2TmHSRKN<- zB2MBPcNmdSWWO)yOe)Sz_|4W^Q!{X2j-kVH`FI`{BLtnYD$XeC1(50X&@_+$w(sVN z@oh3<3Ac8x5)YBk4Nt?#Sl8ul3z=Pgaw|z{(mNui-+%1H#BcCWzO5eE=QZy{7ic^| zIcoEmoIVOf`z(na3N@cS%u&y>jj7+;9!a6J9&hJO*W~xk6u6zqL1I-8d`ayR(>N~oREj-cJDg<_z(kVkB{#ruAXd3Cs?nA?X@dE{CBglb3I8Rl--E*Cg& z(!H>)=Q#`u14qj|taJe8hw#M<+nfuV^I>Bp$tz8W9wgrPjK#L_*t!84>B2UtFLu82 z6QqRGU7-0Kep^r_nR=7%#)wymhV2HH{dRNd0Jl)eF{!v(36rO^D1#=0b|eNWrGUp{V_pq%U5uB2832TOiWPuV-#R zrB%0+k2{7^HUZ*8zxju55H`ClyKPZ4hvT!Odn9N+$B2?Y;-jM|P*50757jh!EB(7&7NDn+!aEeg5_CZ&1RlBn;#8DD05>@$;e z!)QIvsTj{?%lsq7ag*-Jbjijs9*7F0Tw9MBb1x>@_xq3a_scn_I$3*|EfJ&^@ya;( zN#q-IS|mt9ylVv-{BN94!-B~)GH!FwIxu-cB#j$&tKbb1W99I8xAPqaRrPkn>(}iEY4cG7t;~+}xmhyH7e&Ov+RJst2a#?A zBZlBAftx4~x+uBwNfL4}IMS;MfaXXmEY$53a!Mq1envo-XB%%WEm%jcLL+B&vMtzy ziBYBger?RP#uWMzOI5cDQzrgQ&J)~HoD>fYw(t8&vtR+AdHF2G;pQyl@FP#wEjPXz z+$TRYX3sPR$L)2WQ#E$-XwLxza56|6SHmQq*mN`9i-mJ0v8&Fl=jcK^=axnDVp*H3 zSM|(y>zq(@h+*>u{in;=d^U*b+EauJZ^z6YQ&NIG96S_g$+*Hoe+=ikM#$7I6oi+A zOFM)a3UStbTg=*|0ocd`=Q2mD$krlN2iI90g8$!kg(I zbD1U%sZb_wX+FQ5$BjO`r*{}p#;b~Ma>K)HETyo{!B26dG?mjzWT)Fu z-nYlC;gvv`BW7-)voi;(jWq5;6(T_!8TqyV98Mo|Ynp0Dw5z$Lbw)PrIMqZgj`Q{M z2T!!or46Cy40KLz5JBrhPsest#q)c=g*nb0Gu!sPo zk$GVScY+RqC_?Sl(Pm%_W!52o-3v=5^~I@_csUR6Nv!(bq~-$8N}!E%lNK|6NJouF z`XI0v+uu>Hj0AqnjRHE{U7Ja3bcvK17&O z51cS^SQp%|x#49?I_}8LP-csit-Eo29$4pMruqJ9F*vNB^~dH((mJ%AP3Ee=Ot;?a zqW$JAV(naLK^X_rdFm;5qtaodEL5qDW!230k#``gF4c8R<%`zWJ?uT~{diysBBgv! z71?GB)SZxIThlHEAGJL`C zLLlF52*aqmt|IKaMdnLIZI&GlQgtQl(CCz`sqD^v3H0#rn#%fso_(Qj3GE+ms-U+KhO^E--O z&+Lb=AuR1~mH@|;r?e1Xo%${ide3O}_ixUiy+r@EqG628ywvlxVAOA z%x<6Z5?3*-ydVrBVi$=V3VDqybSs03B8i?Q5T)#BZlrGbo^W8ZVo^o?Jp(k9iBc(W zUWcGAMG~mOH~}dumr2t1fwH}lyT zgBH;%`+LKJC*ZFDb^l1*F!*#-aRd{vRTdb1C9w3WpHu1BP9 z$^3hOxtGu*&1X)F^68e=;Lm={@*#`&s1~Hk80eqmLr@;97_7 zIgX`dsa++eBPLsDkv%<#Z#^1E=>7#N!Ki)8#Ovbmf@2eG9SArg04%)50vj;jbbfSH1 z;eqQPYt10W+M!s(GD;mHi{mxIc780#UWgS3CM1{d<|5St;-(Jfbjn#=%A;#iNqf;v zh~0qE;fx7gv|d;g2pWdeqS8hY!ZhXiK0mY2d$xP#Zfz*jvSYQ-b0A@2h=XZeYrbU_ zGwLMVd|>02O#smG{1G9RP|w{wb6UH8pYh+4sv^3i=&?E6yg_D8F~70Q-CB;>=`G+0 za7WBjb3H=nZ+2LRS2uO2>m&JAjEeRKctGMndV-L2FYCV1J1?G{U3bszhB<5VLu5g$ zSD`_fb3!d#>#@{Vm47hAL=G&>!!F zT@RthEUMR_xj1htzsXTn(RcJiN^sM$m>A8N8ASV$=8kZXbiTeoY!tQ4342xjuUOi zjZXh?#nMZ#_h1@TdqpqJ`2HRV=Bh9Uv`p^;bWB_z_3rqIWjf%^OBb;b5B%xWEH%4= zmZG3&P7u9*W!kE1o6~9@Z6%g}cemNyhetiLwx)UrA=Jbrs}`e7gfevAX$H_OZbzYS z-nblp0$ZzIzCIm=q-K9cKP~nSsl*ZQaSN?6{9vX+-zFexY*bszdPQ4Fk>o49tK74- z!_R~c(23f1F9p7SMX*D3UFdG&>f_4qyPjVRMqJV)6J(;`kU)2vs|Z#jANPkDhvT*= z_*1xX1LUttKgoGoxX#Zvdyp6IU_2X&pkUKfgPPUH19>(yi3%*d-kZ9lyH7wk+j&I&sJwsQ-N zpdJnRMQ=plr2`U%E+S2od#9DGnj*SY)fnI|%w%&{!zX+@QgsoUIH1rhfe~Vml@-=b zDYQ3obZI8T1Q2)H6>vR;FlL)f;Bt7m4Q*mE#!JK_dSaRWc%b@wmN5!2XNS(nJIo3yMTkJcv$ z@OnvRS!EPiIct&y=xCSr`Mg@aHgI@qt@D%l8+d^BfD#TGx)MAhb7ie0w4VKF=QojJ zEP0@sjQg6~ifIXI#Om#wYTsCBe#tUSLx2`QG}qb~_Cpa)^!uG#m|k$ESp4jdx_9eE z`V42TID82#Axj{vVFh2r&mXs7Kyqk!>jf}_pd&fLuF>y?d^aJ89as_G^2LGqIx%({ z*rqM@xpOSK=vUj|O$_Itjq#y|qI59s$a5}ZD7LzPuL4u#S~kJCx)afwcZLRBfCGyvx#%P&W-fV&V90Jz>UFTTLY}Pr z9pWM^)<*r@ytIn4LtFoeUQ#t&^oJqmqz;9HMr80yLMpAwME{a)+tEEKWpUWvi{i-M zocZY>p~-%!_PsHSyEG&i;{P;h`Ik`qFD0o$qSx zmJiLfR(@odB{};7ADzbVmLTsbemfq@&tdqTd#VQ*(JhpyMn|P+$?)=<5z#dFY+TII zH8Zb2xmW&Q5_Xpec(p0G+Q^Ik4>^@h8i08PBq`5s$o2#|tcEWIwOuBpn;{Qs%}CXK z{=+~OBrXYE1=k16%TVEr@3j@t{q0s`03Ec4{<1pSoX?}p0%jyv-ssK>43{Bv1RG(Q z)5o>{NW>f(k6Ac!9chhOMZNNJPQe}N^Z0ceTlEzCE6ZZ?WVYZZ5S1oda)V~O-Dr27 zEoT%eE3$d`tjEKYxe}FuM>q+gA7iv-5^^Da$~0SYGK;4;a|hRGb)iVwB^qNW3PgTR zUI@Wz1rPL9UFhVyEy})ll#96Y>mMzoC`^|Bz5^${XIb`{#N4VSxvAY-%@)=9`_A%D z*`D^V)7(b)f^v8n;}!4xGnr3HrY#gUGkdE2?}DkilJ}k>Rhi;@1c{_vPO}v}sDc!K zU=_!Z#`|IGc|{I$hb5N2t4PzKtSVw+EH zXxW}YShB1R68{93oBLXe_*&02Q+dc^|CmROO!;6rhq5PvndR#A63^!&?VMUDQcwyq zRnD(qa#SUatLp%BGqU+zO0MFzJ8_(5dZ$N6JyH=^-nlkzBKF+CUc&yc`W$FN%Lk|Y zLgUYG=_pZo~-; z7p41itI+*q^X34GkJsY0#jZZWdAe5YUI)!H&CthB*~XzLxIiE~XVC>{af`^w-XUc} zkyAdhLDZUt7zQq*@~ta*BCfE7UFP2Hxc2~Zk#hw8>PC#g3DYi!sw=bAHu*)of#O{_ ze$?>TR*P!?bM~iLPx(XTe=mgMt7ZuJa@vqGJLw&bMdoEmPFSHI%#@2ZZFeHKzO}@= zpUt16gVn1{e)8AU0%l<>D*LtX4Wa*uG6Wtf9nX4ruhyao!2kt_)u&NjL1)1A>&Ump zHgM}cp#+7k85A~AO7wLvE$R=8o&>6)ftvT3OcYl3U{(ECO`>WY%W`G~7PbpT?gO_` z+;-?rPH%g>*nmRm)DpusJ4EOaheJPZ3%+;wlRE)$Q##pHXdd;Q+;+*vs)N#$Kz-BR$ScBVuK6=&Q2=ac(hEM_bV2nl1cP| zC8N7;=yhL*!-e9pJ@OMT+#CGEf_W4#ulX&OUu$-P#mCISvG}ebqR_CER}lKjFlnL5 zgyKiJgt)MAtu8p*@B(r-g?3GRh6J^RHEp()C@$>eg;A5sO9eNX?>C2M$ zl*mLswR=F;ftb&E;X6{;T+HUn5*@Y{@6{o59Rz(|gyb)@(bT_2MTnptYe)_Z3+Fr% zIcE5=k%R5M9=}&R(r%`tIJY!8f`+==MJkGGtuEUIDxdG#vPN_smeD0j*)8RquZ5=} z2#bzcR^Aq0K$TOkkTLKFW)+?BQ$(9WJ1nD3++ix! zUCDV$XFdWl2^3LX9+a3n=@S&4E-T;e2eg8Y=4un--9w3IA7JRC6j1Y?lE2u&4#O}{ zb}VV_U?aSd0gxBwtLFg3-Qm8)P;T2WW-m58Ytd((_Z16gD%xh%c*+2u{zx`C8NxX0 z*6_Ok^=s<68+;Yw0>hN~*Q9JWq#qHiZUV^(AvnC9Z>W<2F`pw{ZdV|+>*An#x5o)M zw0tC16%d+1WG!eSS3KS}i6m{7xv6F7E&{RSfieQBVtZq%fI>mo#jCWq^;MC(ugg{? zM@zs(ir1*xU2_x`g$1AG?}J!uK!4Mlm`PNVcANdJKuQm;)iohdWa1yegE5l0>=E1% zcyNsGSC-VA)kK;Pcjyx3C^P8l>&w|>V#|2M8orJe-S)AjFO$Ev-IPnNpZ@Y4wmc7V z9o3E!oe|Xdy=DeywU`ggJ}tnymWsidTZSObe@@i6`*^Ppidi5LvcJ7_M6t(MHm1iP zLu9be^#e8KQ=IG#^yLkWP<>IRlp~!3jwoVJl?@H8OE!4v?-r!skgew`mAVBxW4R8- z_to?36lUx1#)dSlHi(nkQTQY4tpA-FKp|jJMC-xIZ*l#d;V8qRqO|+2C(S9tW-8s9 zI2xv?Dd;|0t)c`yh6p3)#A-yzIz|{YFsVc+7aCdU~PXag{7Ke0l_QJ=St08@*ZmtkMC&@ zrai~JnF^spwb2>L(kg-QjS{OtOv3S;3-icltaN9}U!b#G5E;SqZVBjIo{@&Zf+;Kr z&QluqMR8X4Z`6#oMUbJ_gMIFh9(4Sz>wDuqC3!L#c73N{8usc4!I2HFNCUa&YS0nG zow-zqP)l*uPraI^2`FLuUQd57PsFNJV>(^QKHVzjB$BNiGqsUFzMAdvl726lO8Iq6 zauW|F>iV%W2rNj71gLDW-oEfy*e~qd_A69BnjoL}!*HmRoR*i|n9apK!he9XtA53b z^x~Ke5I$$seJZg4lMt0^D&x^no7U_Z3;ZnkH$Z7EdZ=0sjW9n!rnY=yuBuYEkb97K zkuKUU6olgW@|Mr}@nyx(+wTZ2HXDCD_ljhz%knN`k!}c5KpRLK;Z|`h2h{j? zEpX{xJAt~&Z_Q*yTf;O`y!_|*^d(CTYzCEvVBj9=^!XrJ;n3*=A8&lMAF&}PFi1GX z`8eG5$xb z{eR`zOsw>5|4*Gw07WPI*UHJnfq+ib%D~A)*u==z*aV7~7s}De!NkBC%59@r#hqK( z6WiA$5}}}>$OwA5%B7~tNbe-0US%nY1H<$rDxGQlCXC|GAI0!Y8qyAlorRFGgbBr# z16Gr9kyb^bh=-j07neu(+uU0-UbC63(@(3@&rjajGYcdM3N0*QP+UT^sDb=?G{4SJ za8NknVfu1}z{tUZu6}vEzyUxzb8a-M>PpL5KuUUCLk@ZUIiSAuJ_7(q<3WN%CSE2Llv;f+PC~sFG%YVB-)x!*B=lLqqxp za6%}O=o9HT;9%Pn1Bmk|Y(Q}H>dgJ9Hh^K>1Qo*K;m(Ap{h&e=FbDxqfZ_yv`$C29 z$N*(`ba{mn0I?*b59-$uPHT0qB#iM(tNjA2$?BIdcQ*ii6_UH#(0qpwl;s%SMCuuLOkry!Ux>-n=gMb;(p#eS&=%-r#VQvcw z=JL5|qe9s8%~_4=`5Z8^p@jvqAOPAO`N7)6VPZvu&VmN}ia`br0k_Wk2_m7P0&((5 zR2w*QB*a|q6?joNk8v2EM7)3I&aHs08_ zZQD*K9ox2T=XBMnbLt;_gMZKVY}X#$_w%fET@k{w zG)nxogZX6$1mm9ogo%e6iiEDe7&{w{l8L1B>Sz-V7p#D(6EBYdoRH;$o7M;cW&n?H z`1cF%^-dDo9D4S2fRWhS#HARcsjFU?u-|7a+T|YZT^jM-gOynkd~|i55S;`(KY(Z+ zQyx<6u;Aw>qtmnn*Aq6BA_VIE2efHoi%N!@ZcZ!QEzE~vm~e2a@i_q$i5HoC4W%XHyYUocFS+zXD{}R#4cXiy43Eq?)j_+y`HTaD9z3d6t zGBF)lIpJucca@p*YsD)4<49SMz_@mJG`N)e`ma=Mz8Pk1=0h*HOe6lCSQ5y8t z^L2c?`+d61#pfe%@KCZQ$FwaMEmAz^VnkPcpAoevZ@g-`oX=N<1cbpyi91uQ!+N8w z!`gi=pansfZ1u_g(A`JPorVXC$*cL{;Jqs>CQOP!@47rKv7wgp_-dQ%=+^D+yxR}lwiu7{p8hdtpj+6 zx226!`c_{BRREfg07FmKNhp}GI*VJ?a>s`bR((c_{ zm={-i*UarIYTcPccjIcer$LZ2L8H#LwxkS;ta?&99lpJ->33z9V{@EDn^C$Dpy@6} zrgZH*tu#@|#9ZvKsi22o0gdA{dV}p0Ix+O7 z4(4XZajBp2mhsB=-7^3mv`5Qz-5?y0{f=H45*

UIl)AV{TQXMvGe;^Rc9}$S6bl z$F?Z=Va(ig*!1*s(;*-l$xD!f8i<~{xU5cft<-WacY$$@MRXm!&+{^kjYra-JMT#% z_wr5U{s^e_CbI@dgr4VqwyFks^OGO#L5>fPP;V>-i-)#TjEK2axbyFM+2*Psm&NTE zH;*iZ5NRZ|Vh?=YM_bYXJTgDKvUxB1uj4xTIr>Ru3*7iA6aAzXVxHCZ+Gb_-8>F~Y z&vV^F`-Sh1uYqMtp$KUey-V>IWAFm-U>C5fCoK9zp1gprT-R)+hdbm7)L727lAC+`$Q zD>Gkp)d&mO$3v0UKXwt_CuQ(6kcDiSH6^p^H3|=SV@y5VpBfDuR2Xs|7baO{3v5t6 z5z|GMftpi=DCrT){BJN3+HFNkA7qMLpT`Jr6itdGA?vLa^g-Un*fpbSa8?J}p{t^@ zneD}V94H30sHmB9`$d;|#D*26FFv$ZL;agwN^vxsBe&#Lc{?Sd+IILzeU9JlceAxN z0uwV+wKSuY6UA3chW0(hDKcQw2Loc_MEXXMs15yR>i|3YMZ=v;=ub;F3^!$&h_$Ny{~6r3N~&0kIj(J({Tfk z`uiQvzrDfk1;?F#wVdu}L~T9iSM}x(`)+_M%0)y8GxEr}llc-0XT#OInk*x`KhhWb zB4^uoyy}+=8&vl_-6~Sb;BDIjDI4OlIJY+YegfO4X?3S$E>q-3y;%ZtByRDm~c&Hp#=mO8Zk2Vjf1ubE~Xs?0|SUb$1{C&SC zUpe6}Q!BUAOtK;<4^@}zjKs!zb)l-{<`+(3T4A#`^xH2~!in4SOKmbDv{+i%^hMrp zejm$>=1HpTeZ1gQvYQ>TR$A2U>%!9`V0ZPk_-oRoqH1}E?Wf)~ZtE#Ci(ARr4#OhqvB}BW1HZ_(gleIGv&D=~8 z3TRbQ`U^!|QYk%>_o8{H`MZ*HPA_@c$zrHD)?gz%a5F1#NmWJ?&Kpz*2cNl_XK@<~ zcJ*R%5pKk&`-Q}f z?4Cg1ibb1p3u{s4vaZ5v8^H3VpB@dqrR%FLiUuR>mkE+H?--HswPWIwjL>S<3~GLx zEHgj{MaP5bRrVnq;a-m(5TR`s!Q%kq|Pu`Ek@b#Zl)&8 zdhiA}PZP^G2Toq90m_C}-%lQk^C{ zlcPSA(^oCfo0#XRt?)GEOUHIQ_N;t?S1kHO1o*mkxG;|*0Gfbp{c}4+Zr-ckaf4P& zn^8v0n7pOf75^F^0{Bq9xQQZ+l?aZrlLUml<}19|_7jG+MC~iJ*H&9}g$IGUC1A4J z&F1`KnI@I%h25|X1HD+sk}t)ndE@libDJej9pO3q&a!C_NDBtOt@v%5_{ls}^knE0 zU%XEMLXT39F(s?~x`8Q=f3$lw?c+ph4bPsN+4pG@nxT}k!o4S45o+87^)`?M)CYyp^KO&sPv3y-^C>dLI%fswe;vu1O)|}^^<8T-Vd9~)3GSy|`djUZQ z8slOnNAZTt>cUu1ekqp5qGFol*|eGIX+F4!!*mHn0JwWWj}0A;$y6<$et369=*7;_DJasxk3^yHog1duvgW1nkGP z+1k;?TXSV8ud~Z}ig4h1C7=zp&6%m0>5s4>9fZoZLg?ZIPIM)03FJ+9Wo=q_tHW=& zGaWRjO}%f(m$ov%ZkDyas$I`pxh5cKt~gawmgr`z^5}d#c6Uw|iy1==^j&pgZ5JNR z7zZ3tJk8ad>juuLB6KIy+#pnat}j7pn47_AJ03P`dGC~2hgY#6^K)^hcdyY8a!9Fu zyzVLo=J|4@K2C(DaRpDvUy>>@UtI7$>+|wEl=khM{d1Cd&stn<`42j{ml%YZx2>I7 z-?CIgLEPz`f;bP8qt?cK`Ya}@z1<-!<`{ea+ZJav>EkY2zS?B5tl5yqNV27ev`GYT zy_hM}JGM*#fS|re5&o-=vi?^cW#?l0e+Bg4x-c`x|9|_Rb(EQf@&7Ih|8MH3il++7 z3ci?5eL`V=plD4S;buZi7-Kdm zW^FtLBp4Gi)JXBD%)5z_xcuI9D0UHik$?ON9ioGdjxbd?6Q^XbAuKGIumJh5(fojb zcM}O0Pa-nQ(5&dWQ8*|#IYIYiWGQSJ)LKcy zRqUh=omS*5yg|^pbS3H1IW1g+VCdfuq`ZjIr9>p~n-Uh~ie z!D=h7;J{lzel$xCSUB57t6+mZV#+bb_d}9Fn=p<@1d%qu@hPxXU=V^8h(#KCOKf+Q z_pNt_;R_e$kexF();E7sgd&M~&53V5ZnY+oT4bA@Zbu@PJK~wlEnH^o1+S4K4U2_D znL63Q&g1|4gmApTzr`%a7OlCIwB(4d@9)4wUK%-(P^y#_DT(;=D|xru#oZ z{ViY_VkVzS>MZP}9%VtOQW<=Iv z$tI?l5}*pk+}LJ&R^fWtu-MWe{sqyDGzZ4wJ80sk2rBJ~LWSZ%Lrm&%#EYbmr)(so zH@Y{6yT?-^4~Ma#YNE*8C<;PsVFc3561a9hc^MZgG@IoM!$w}G&o6(T<3=9aryn`) zzV?p=(c83XCC`|21qkHA4$c(FhOtbg7{+>@DZxcph=^IaA=`xsat#{u-0Q>#(dnysk3K+ z?29#`3mH+w{ZaR?E?}60C0?o1?NSkUA0QKjsQL{|8vQeiIWL)DzI=f7Bb24^R2ol4 z>+R(9d6nFBrp{4(fL^chS?Ww-MXmO5bwcd!;)D?`pyl%@1m}^6dER?|GIz=XCLopo z$#}@ZrREM4FhCqYSE7{J`!JthppEY))ZPrJO%_OTH!GK{ppXj%i3^W36|h>Z1`~5= z!)yCD3~8`#P`g_nbADbvkMBaV8}xaYJmi;MY80Ht?slGX2^V(E^4$}S zbDbQ0>NcBzvxxJFU3OaSM80J)TJAQI*fI?z(vDnZr}?ojxyM*tXTWWTl1fMKy>mXN zI|u9T@24oXZH%4u?~XO9hT=xglig;3J{mrb=-K6bU7SNTw)dCQ643Egl%n-RJ-1?# zxX64h+Ve?a7^YQ~_~Zo^N0Uj@^Yc2*%*bNRvn}1W`>oOy#WIW3%2clJ`kkOi>}Ogt z@Gp9K42iH73g4$IwJu<~rpQEO0avN7R?Y3+-k4i^fbT{nb#bX#p+@6{reNFOR?e9} zciT_$=qnbsP0Ka)XWYZBvH(XQl#L9E;Z*-x`K_mxg0YDe@PSu3yFql`GudVMF(UgL7PpF;5RX z^|QutPkz$@tr30zQnpW2dqDYPy+yG2N~g8DHfKd^tschbXv&Z+H|~C~6CU5ZR914D z?aZ24<-JN<>-CmijXP#LuhnC37hfK&&IWslf!SbDT3mlB8iW*`j=b+x3f#Rbh^=JW z4?#)XR+If@RWJ%mSI5vol)MaXyOi&o-SLQo)!3YUGn=>_W8n^Jp(pd_aav8u3V0tk z=4;XS;a#0{h_Z2i@X-N{%W#fjZ-z2n+*JLxsdWzj-Ys*H}qPl`ytAsn8qgO}cuUOHM&1~ZgGIe@3J13Xe6B!hQoK99z{u)EYwbTJLU@j&% z^(wE^g8l0({7&1ju3mXx;))Gq#O7{e^W0Y(AtlvCjw|gMK<2t<6X`_#jQQ2}^+03&A3l1E{FJ4GE?N7os)-;U<|4XE1dL63W!hRtr4ODemd(sgf%WJz|ckf0dSMjOxc(F@bm2sfq%(H+g(>s>Oa}ClOD_3Pg5CA zQ*AEwc%GrD^AE>ge=&8#0_~aD@LEfK`Ru%(Vcs9vtVQ)l*^Y^v(ityK!J<^RkYfRqiH_pu;o@uQv*w=y)C*oMA=H zLxmW};f<~9|E$ps1kufH$@gf{O3@nI$G*}Zc@|qQP{$3osP+w#q`o;KUqeN$^yR`f z{CtI#-iqY$ka4}y;->J)EH>o_dax1|4Z5{V=@yjUSKohzSk)Tu?#9s5>oR!f#_TW- z(-*Cs>n~hSrYhOmt!&X< zVD6NgTmPP%!L2Q#c5Mg9fh{5p2lqhj_}%OGgoQTyJpPxj4>)jC=QE$>TTpD<}n4*Dr~$a08?LP!b3UkgI@c0y!4{2n<2|Q>*IZBS;f)1+1T6$+$W?XR^_{ z{F{Q_Ff`tTPvVi~K}BgzDIp~ndDzvDj{xegF?2(qQOvA$PLK9g4dCdjDIjLpdsv`f z5OD}aX#V2^_y!=&tgUPioavy8{#rn;s*Y%?{>71nor%V2Ky0T#iRYD80KZD$D}h(-Szy#f{!hTW57@V48iqQG+JYLgx!LCfERdN$ z&+OXpx%>veq3G2B{1u@kPQ}?AqN@)ZsGL?;7ta$Di;tHVlXg=_2Mafw4ine*XTQwM z@(ASC!J!q{r<)tpHUI!OGPQkaoL!sR4?uf$rUc@mQO@syxs^{09_q_>BPjZmc(3pJ zW@v~LYX7dV{HSyAhe3SRx228I{T@$IR!Rh2#nN8u0=bI0k`(~ajnUou2e|>P@%ILp zCI1jZf`xZ-@%)?-{n)2H{ItGtxFUtcQ7kk)-LX8sot#^Ab98fSaY;Ba<+^)252o&5a!$pkJyAPyn! z-_cM5GdnYm`+5g*a$s}`57y4z1^ngnef??=_IDDBv5DFB3-SGLWB-?`?)vs7aQ_#> z14I&_>suT|1wcpr^fxt=Yjwj5c+&saAsQOKmLM?y{l|Rfd+r~odRc8?6<*?LCgGc9 zdUbky`J#{b2dMz+b0|e>U|?-}^_hU51_O#lK0hNdrK;jcU->glbA9PmAnZs3+(` z9J+sQY6-#MdKdQPgmJ9{D7qBBm4H@Y5}=u$kq#7X1_+w`$2i&{gz=yIw2z1Dw|L_i zg*Rw^$N4O``yd$oRD83KNf)5z`GoURar@LC9k~nepH_nh6b*jPO}+^rWdn&}Xd9>X zwU~cPIDO%KhC6)Vd>R~l6O7)^ZC^0F?gXj7JKub^D9QYAMf`Nc{|r{vPQ$Yc%F;B& zaTN81jN){0YkQtE`(A=Rgnb=c?DU=esDs}BRNwD3ivGss^*vo%%>aI^CWspt8~-$a z#P$10v3jqox~o(DRG93$%kzc;0`mfq-Ae^r z25U@bU-u88PyC32pwdEH!A%xYI>jgbGf2zBH8Ly{BMjI3ZdgcHWOgzsqpvePr@j0Y zLB_zJSTImPm`>D(roAC9;7=Z-S)@12nzE9&p#Uq7!;5focM%wAl<6(fDM>g@X{$pM z^zz9vw-a)26@Aiq58jAqW8H!}fBz20CotZm>6T$JSHXTvm*e z7rQVWhTQWxEQQC>VW-(?!!FQCVd=If&j2gwoa@kM64Zq*s!R%Fja*#GD6^Tlx?~^g%MLg@x|MLslk&iw^Z}UTo;dX zzeq?mrzK_*eSs5w)0lN&v5C8=0B!siSk}<#5f8}~B~2E1jC1DIfl2-2JJFCh0MiG! zR!5K_rF`RU7XQM_7OPK=-#?-O*-Rv61lNAUR}tnx%7dr=iWQKvePH>P^l+JOW%mfx;-{g3Tac(HaxzG(8&hg z&}7LkkkBJ_>K2NDkU+tF|3uc1waClY-aA;Y#zKJxcVVb#>1S5dGaWh0G`oWR;f!MNO z(l<;B$s`Ni#MgC^!8w)&p*z|>lj6FpQ#}W+S)EXaXOiny72G;-P_QWNJ9BPOcT$+D zf1l$L#Efq%m{Cfa4DO3w3yi4j0H&Fm(oc-3rY*MYowx;mr`9R-c9%I!4n3w=!Q=Fq z4bwYw42YP+R#Ztkl zaJ%)6{2z#8SBn^@x?!DhYBa}+Oh%xlkDHwyA(Y% zVY%S*kV0QgoM3;^`N}$XApNt#W>|+le9tD>p#c|r>^>N>Z=lyUI*NZLD=t7nyL!$+ z5CqUMpXEnSyhim4Nh-zmu!qGXB4(`E;RXRXl zuG3wIMk|@;Foqqe1i9OE$o|e4V^Fl0Jo=u3#hMHE=@XE2ECY}~XUvC<9C3|(v8k5O zqWfFG0|ZiQ~x0Ke#1 z$v0-}xP{0}&E>{1bk#cWr19xJbmbTh`7&H8>lvr6!Y4Li=tQ1eV*c2O+(+r~B>fpK z1wIX9s{?_XN|j98}^t1vO$hPa)(R*5Jd;HSTi!IZ!c0+Rq%Z|RWvIpCs zkUg!CRhuWpY-(V`fK!|KKK&I9_s})KSdC0 zf7`9={vrIx9U7ljhaaUGlKvBDB+`P>TLm-|^J@7l(E?JrlU~EVQI#Sll9g~Da__|d z_x9L(rWMLLU!J?8__0|Sf*E=d4{l+T?_C{ObhqawRsl^O84x|=fgxawjmOpbOEIlJ zOlIZFtF_f?7zG2^DVN?faRivsJcPn(+_X912Uj_0+)0PXNOgpd6F4aow%IgB6Ot`S zEWe;PHCuvYV=e;8{vVa4&7a!QT!b*w#acJwMn>)zUM>xO>0E4e-m0OY7KRO6mTvA zbgja0iDc$`*LJ36H-ir$p|aS<`AZE9QHHjyFv_~Ba)?I*vab{xB-t5@4iF<#!L)Y> zh9$t=D<7k3fe^=S8$+=lEfdN9(RL$4Wfl}{Q6?~-pE~VIz6f}~Dur(XTuox`_OnoM zp?Ftk!(!-3uqw^^z9hot+D!j+2n0pg5gbrt@1v{6tm6Laar|ki=wg;bQW@t{w~?id zOq(6~X_62?hJ3!xfmIEd6uX21vP=n^Uhw%dL@b5^;JxW7TJdwMSk?OS^?CJhBJ{Ln z^QAgGYQiK!$T->B8r7m`SCq-W6){I*F-Cie=#s^aa30zin{8bjwB{#?8>%v^eL_zr zI3S_bFBAQZbWS_I(qzhtW9nX*N9^8SiHjBRIl~Hs>`fQyo_iGPh>xH1>xVkYDB%M> z#gft!CClFEC7~hL&wcHV+0IbuGHqdumR+l2FY@EJZ>F8_7-%pa`o{d)ahMw4LRWj) zzCyC^+GGA7uQD2MN-ef|N;MpClv5ohIkT|kmA<=HQtCO^7pU;h1dJl)1#>=BJM`SJ zfr_Q6CcXEi2Ip&P{9E5i8$JCMF25xzie8whkro^AC^a7II3BO@Q2 zNc=6+-R9IDSY*7UT{PRAg`uAKY#!)q*V=)d)9(u3cyWPil`&tG+{L>~noko(-_tPq zBbc{ZGp7lhtMD*eZ$;BUg=@>T3iLw*8D3De&V?X_6@A7lmqlx-t10dV*at?*7vdh* z;Vqx`UkLBbr{x4ZkcQO4{PE?0+%_$$v|FVRoIQlEvxkvjpNZW2Qj!avbeBb@tk-#! zNyUEhcCp16kfCc{NG#-pj2kN6OJ&LM!Yxj)0>CUJFD47+)O%ZNWa5T#s8ow;)b*t& zmn9RtE1?imWL^i*hTa%B-%K{vh$d$9h>0WeAg;!$TZp;@bhcVddb%=d*TkH?B5Ys% zu)MZ!Xs1*PYBpXhrH#b=0&ctvhx_sR0H4~4xrF#2_t)A$jk|N0B?k(lXGf!mW;CdH zpQp)4yBF;Wu~2*)KT z>7#%)^KTj+`c>DR1~(ffVm4e3rW>OBQ)~(PSG1a-A$=Nqfp9 z{6NE&4kA9AZ)5SNbJZZzUjz%A(5L`WVRhGUy~^;LB4uLF_KDpT_iWTh0p@%S(W)>$ zl93gKt6f0(*o>tWMJqIs15juus!I7+_4~MQvUGUC0KIoD`Yx^j6hm_n8sS5jGxac7 zqF(l!bkN(j61!=*>=4!&S*akmfWS;@$H;3LB5|uaf6+p%DZZC++*x7fIPS_>G2loo zT!xIU6`%PBzdPOZ>%dkp5{zZ67JP;J!h%t#1jdeE;*T}_?D~v({i9}cLTkUnDvA6z z!fs@z!Q+ge-#=w%lo%SV%+mXriLljpl{wCm#JQND2*5uiBj?6$nc2VfKt@ojI2=`c z#AD7ke}RRO@>$M``Khtqtd>r`{^AJDmJgr)FAuEc?_(BBI|& z)?|h0kX&uTcGm}W0f;+#JZa{cPBuQ*wkjkS(BoLUGy|H|KFT~U!duZ)JZfpU*CNP6 z`rJ?XGZ@v`OzF4`?tPe{;rlaOvX41X6t4ak582u<)5vB!zO)Qs>feaUlV`CqXGeA~ zsLa)yTSijSA7^5v-{$qvovYaiA10f8m8lqKoU{7{1uky&^HRHJr+dTU!G2z1|Lhv( zCokO&UbrhWlJ}}bbY+x3xo6qtF~xe_C*;`Nj}_6w;|&$f$(YtAvdC^v-X4aNZP$W^ z_5Fo0H#^kl8aRNn=4e0}8`hz%%bdXwyv7$9W^r7|0>ZyJ@}uq00Vy~`6~9DZscpbNZ{GA=Mf8yv zqRy6#`1S?et4biM5*@*zO0%@-jp-*|)w?Z@JzEeb>?!Eq;AkgcWc0=i+-PGa|24Pz z1oJi7`HpNzGb5 z%6&4fGY0gA25{l$yQVNt6kt7Un`%zJ3Ho(%MyL!io^HbxZUr{fn$1r)Bl%-3xj!Ej z^Gq*8CZ}DHZz$+RPazYEE}==$D&7Sj3ppYdfm`DDT(b4xMGXFs=KuO*ObhpXO*}>| z?U@P;w*j^AA*Uz}O&<({PqPPK5D5dH9yp&ASc5Zk^MRY1@gRA)_#svao9~tBF^#Xr4 zxwgzr*2U>2(wa&(&(I1IuJrD(&ggWFRi7xEkL{qDT4)@JoblXUsbsXn-5KS<<6n7H zt)BBB5qwV7H5qHIhjw2ibO9{btCzxorQ7pIK3d;L72y@!&jCRQ1y zAa-$S=P)^km~_9%q#;04xEf+;cE6qw2cI zE(uWT^D2gZ+Njo0AJi)K>D}T6AF88f^C~efVoyraAJ>KW$A+(uza(A#J)3?dVL}=R zG%8MAP^Z1xZ-^I@hibKncRSM?kENmqC5vewl5&ih)U0ziwzezzEH#f!hOnTA#%ydC zy@Z^pYduqmkk>QIi~C!Yx(wY(?qji-cjnz9cl&U=m)-5vIxlhZwrAT<6XWQ&Q@+;H zj#zVqW(#MzjXOl$X0UxK0W@XA5AR6`81dW-IDhjcGDPoYbHeYzWpyN3f$!m2#$BaK zzruRB0xDsd(-9ltB(0>%d$21_LpTBYU=p9KY8W}bWazZe6bnPD3QGxhEmfLHR#;f8 zBOh_W>A+8$qUs{BM&rAcgU-4Ma4|l~;ZlA5s4vU6MpatyqysZ-Zl&(#GrHPI&+3u8 zmpR)?ysUs{;-+FB`OQeRWCzMqmC@9{hwfgAHithgkJnht$ENsZtABGRl8J3G2=tH4^#SN zuZkfidNUF<^*ZO(H5a!2SrzZ^WnMal{yGD5loQB01DW3`eT4f;&5vq3Af~0NRW_%c@-OM`hp9>S5X<^ z@cYM5`pK$AwWcqNWT#6oXZaGj$dztUDz#q4LXLNczNW+j{z>?Zam;{An@V&;p(z)& z6kH|a=Yd45w}H{3M(0EdJm{!-iVNp(ZEjdoT5@^lC}1`)rS~9H$17N-KKZ?RVPOOM zR8?3D730xNf~5KGa}m~0H=z|`f>#wq?}w8RmJ35&UfMX9us6g~+UoC2?tui59Y7Pd zbhLjKoDGU$c3iECT(^I<{C5~?>XDrWMc6=LWJFNh#yDf4xo&ql{rdAsvo|LjPkiz- zhdUfd-%U1O(gNNC$9#md+Bq5s^Egh@)w*D>M!POc#)6Xfz_nhRx?*PB3?mCfWWa=0 zL8O(GSkQ66>ndv;7qjB0_{N=u-E+CJ&G-I9=GyutYxyWfbeq#_j70W)){3-m#Eu?^ zWWxOEVdOOeBfUHy>_^}nJgP#qAcAdm z*zfRdsy<-^=zuPABdpV>dq%#AmTe3?*FYKgh{Tz#hs{<AV%NwBwsczrLOeR<`i z9e%bm;zW984`Qanxr6nJJ`L`)dVI_uZaFz}TXg2MS#>w?z9boWLzzMuVamLdze(Ig zF}GdNJFP{rh*NaVoP(NyR9-KaewY=@NT@O*R3b}SNl$DW1IYi><9=H87u*@WtmF|U zC-BjXn>=C3D{BWvQR=P@G7xsvexWPDiu>BPXH*-PTZ@y?M2lxSlv_nbk^*(QjoQ~{ zhnDd_WO16Hl~If>%%fYBKslPi#eaLfi!7Z!ou?NI@VZj8o^HdjiuOwSS5o*-kR8=Z znT+_g7NMIdW~~`6+I-w0EjEQ+z--s4H|CSRcT1QS$mkJ((%#Jjuvt^d5}>!H5j;X0%A4UxGugoC0xiT_pB zqCA6`!w=AA)j&|bpXU0uu(<3ghr2Z`_rHE0%|jH5{fVpkv1Ky5aSV#*;Fk4fVqA%3 zLVT+%7V!p~`N+5M&ca#V->?VXSMzX7(rhS)!N? ze7Isq8bYkVw6O%~H4|6E;KvtI`CGZ6eNYtD)469fR`=})9Gp1hq9ro?ArcJXHf=xQ z@9HH+qCQkf4UCHJkQNZ%b8x=ulLK(Fu*;LxA1A)UA`Yte9@|KjMNnl*O}Ycl8cB=g zs-EO0PR`nBhVAu}P@@;78z;l+>>#WRWlteIrwczuTcAA4L7fX@qP;3Ru0F4iVo{*Wy)>&vy z;xhZFg9(h2G7aJH$sYR3*#`?Rf(BL&5}=A2AQx{`Acy;8z=YnWi=_iyFXyxsWE{N_4yY0vr1#dZ2Hd{r6;D zpvj$4YZ1$v*9sHsPU20*d$sIEyUjlvaMxw7eMXixV>$J(?v&VMD%}@XthP@O^y309 z`L_iQ?W#wu?LA%)ur}&RI%eS1iyPTfa3YIJ!h4-t%f?xr9A#f#ZPH5jHwbK{B84uM z6?~MaQ|?)=(Na5nAP63BaPfJF$f=X!^;2&1cI!qugWEu9y_6i=BhDICNo!<+u$7$8 z+{&S{U+C7ZuTlo0k~9X@U{A)}dS{wKLxyk3Au!HrwGab8YZwjL$x8elCo8covB$_- zG`=$#w*2ad$75hKo&9B~LeF8ORh^19dawz?YpXJh_HTs9_Dt2scwpBB_C``$Y-52(CfWevZGv5zh4PhJ|?SC9uj#jon_iva4A8!pO@6 zfdMeXOCsxIk7aJW#|vOI(A(2-#~o;&3cB6!Bu9Z!Jr70m1#2ZSJ8WvtZMLcL3_--I zkn2sjim$$we(n^5!YR-mA@rfRGl$W;morju*^pEVM+~p#UK%aZx+UPFWENbe3w48w zwXQcoHcQ1Fa`Wz53;wDS1e?9egOqJ87**@qI(MX#uy$?M%x_ul-rT=;MCdcJ-0VrI z>!mJElL?DsZQ4vti+Ugj`T z4aeuB*i+8ME8zjVNJ&JZNoG!E&nSzNVleS5n9_i?SuD53wHx26bep5fhR5rLip${| zN!h1zO(v_((yWLV#5=kFP67lRO}ztEPr$pT$Rvtw@M6npuQmttwpW8ejJ&tu)Npxz z`g7HD5!)mg@_p>BPq&ffY&Fkus)k3e>7W5u8=^sqP*u@h zcr-bGSWNz_uxP{7_3LAd#3Iqd;|~^;MYDxdH;?E?f2mps17j z`|)9y;c634pdm8U63bd0hYSL)S6=9|eNOzw+A9-9+5iYjKEM62a znf>l(pdPf8EY70*qeQcka3iWb8^Yb%q#EM-@Y$KIWS;SLvUImmbE!smPYDi)bqEKH zmJXrU9T(^H-#U}Y(UY)B;|5XUs4aGPCp)rUH~pkBvi4i_*QgE^YtiL$eVR$W;myu9 znyaJ8Zh?+U;=x_}_lP}Ddj_|;`Vft}pf|h9^qWE=ZIDD?v}-nokAN5?hQt$bw&NgVLg2&1qCW!7~tlOOL|P+>;pHY&KuK^Y0}?JU`=FG?twFu%?(rQn>eS&<-{|s_EPLn7$IE zji^*%Xr&uLH^)1gsl?%hENgk4UVaat;LvzokrMp5(O7&IE$Y!3u|C?*N1X`#{^f`ybK{=G)?qaxoyNy?T z^|C2-tPin}+A@O(@kRu3ovr3;9ojq84+4ATTrtlP6Ct;};0%3i%cXn^DXsDLVf~as z9HSV07`T8;x+=hH|NC8PHb#hh%@SHgch-h=HL@5xs#o&2WPImYoX z@9Nw8(ruRZaVmGyQbuiU5_*ql)cdY#*yePBW+=nXK%Fv9uX~?Gq2e#@$b^DRY)Lso zEcGE?vC)qS;ywV&OhYnM0DFLJ%BIbmu=xHEJDe7JPN?ntBE6_WeHS&3ELtQ&;b4M+ zF^TT8HGd#zJV%{^QPeQb?eWya1yer1zVY8RASS|Acw;>EjSO_BkqxB)egs?*B~~Kz zjbToujmS<8%M!I}n7RNNa5BUw?{p?8&*~&g9=gu0a>$*=y6!&Ry(>`0tM7dl!Fz;n zqAojysO0V4SyfFlq7%Te;!Rm8)!jSz^`m5S4sR0T7Q=0yX7FET9B*5xLCOW9 z(g{iI|G@Hi$0M&WwHO=nb+9rg8=~V1*_HvtSBNIAxS+Qdf#zl5- z+gM);)O}eZv+Y<=N&90 zxM~}=hG#J+BDtTmIK*M+6MEpTLsNHvCfFq`Q4qO43N5QjCqKzOJ9DMo^oCTk;j z7S@q@)55Nim(aR+jDJ*v5e`3*k;HWMD}lC%@d5f#n$j1fhh;fi_KZJ{n37#r=IqFx zdKq8^9<>!6(wgKWCrC$w-murobJ-c;Q7`usu?oj|g{Oq`V|3<7%126eRJHTz76VnE zM!+hhA_OarF1f+3ow-*|LTZ2SHQ16C6YnA#T5QEVd?I-`aS>bH;_W&k8nmwb%1_%i zx7xtL_)x;y!aH)eGDzs;F7q=2)ItC-K@ zYcIVqJS;yTPsX=I(DeE<$loQOvXxdL{5p_Z(}_De>hmY{-vl`MIyUMTEL>1M0gYfg)GGoxs&YLk_$enRY%Do)hGiAE$nCS zohfgFKF6xZ8x%OZbH%v zc{!rhYHVoft~OzPhU_G29j)zG%%}&HGpzBvyM}ZN*XRPcN{4M;ZS)*;fj2&KXhkjT z%Jl_LvIDrV9HYX=!+{Zf8E&I%Hv&ibA&{bhti8?avJAQ8BS~uXy<`3ZSN|CJEJa}f z>gWUd5x4cc$h2|{T)0g#my33fw6=6sgeizN7$$IPrn=eo_ft~&&=yspLVslGTk%nJ zTDQ7J>T>6HLPCL{r-qs+%{3Qo`+l`6N{+X2Q4|lt z3i!ijIKC+!o^a{uqiZT`h4&JLsr&6my}e-OezGa~G&8-C!MS~jUU7q#zcY-H+Z3UJ zVg)5m(KSp-Pn09hSlobRQ&=T-<1~12!#la8e59G(TlhW4H;Ez)u0#+JN2TeDLosBz zOw;YV(-ev-?BX$uzEICI{ObDc#yOvZkjR{&O~?E=fszLga+Pe5v@l&zv%YN~5o}mA zLHc=(>$Ved04z=MK$r@qq6bT$M_QB~QiXvBX)xFRG)nUX(5Y=cKK7r{UD9@((bI7^ zXi6zhX}vma!OhU!z({4}u|8dI8pSP5yW_X2v~O@199T?Pk_}g=HDn0jj%aR;K(qs6 z3omiNKgpf+i(U0K^A=l_m68zGRg)MKZhEBu6=!^+o=WVrZ&qoyKG=*V5@N(gRQOMzbu*a&# z?#+KTUq+vup#4>uelH&yt!hN!7(psM*V`xT_fB%!Lt!bF`4v@n?@HopsR{quBVsZS z&Io$yP1V8XGLu0QM*YiBKC8)zkz+Wo6GOba7Xc1U zBt-JVSTLaMLv}35uGQEbTH6)mqIYLRk$p@Ces#XkNKeYAgh=pA!e?}D-Z>j+NKxLq=5CC*DVd1O zAI}4Pb*Myl8G_hOF}fGCi}iZbl=!wE!ord0C;!%^Ho+4Z9vdfYb~nAGYV$(l^)!*Y zk-U6!D0C{007AA9Kp$RZ(TxE}T4W{%bVps%uzO4Wb}CWp(P>BF(;fG5xLFt8nc1>o z@+RNuW4jD-$ z@cQ-l+^A!gHstC~-Z`m~P>nJ_2AnVxxXnhl3r)xqas}sR;d16h;FWI~jCiT+#*au8 zKV_@q-v|@4%m&GKiu3E8psl#i`E zTyW}ZWlXA4{0mu&lLhzPS7a-eiERS8n9n;G9%wF6G_xrel^{;a-+VD2Npf2YtE=Y1 zT9Zj?qcgG2YN9qjXy?EtGM?Pgd&}}P3;+t|e9F8}cUi_?(V*7#k0XB=t;l^6qVwnF zVdJEz+LjI%o}6!y-vixsLjDnFhl?*HY+ar0$HOqiyhWcx#G;~6sTxUxom0C#hHKQn zO-IT|18&s+hKjc(-E6O^P6Rt%YFVOSUeBc2yThv zZQ=#-GqLBt@S|1zGJj_M=q3%g)+U;E!?!w1&lg5DXOT3g^2FIVmG?lHZKDzEA;QmK zA}K~vf3B98yV1-=Q(= z0{->)aUq^kJlAKWbWl_Pp|FTn?!G|RwSV$-P-`+=d^ONz*K9M^U|>2Wvsk^nbR%78G9 zcZ1yHGpqR z@;G@=Kj0kN8ohWHPlKFUjoAi4G?917>}w&LF1hk?j{Zp_4RB~0x=-KdwMW`%+FdBsq_!p z_u<{Xi!69Po6y;r!?PG;zNwq=BLgqsVZ@r;Mydqbhot~bL4O1rS){avb4TVT2E=-< z2y%kQ+_9ud6utOMc(5l3Z!W;K9=uCiV68XoeUXwZb~r3< zRJRhg=QqtQ(147hr>X$oH9x63DikdfSq=cEs%d5q(|i!(r#(h_4~Coe`#E)+>J|!i zrq>RH`cq~6qxIMIpAkfkrCH}$PQ{9y-YDtC`sPLcsW!A_%#fX{d*x)gxRcl42s5-+ z@5vV*?@<~tLLDi|Vns97oIRMf9DIzLwV|`O*q%wnEupEf_9WYR&`5ewPbpGxg9w<7 zhZrv+;hAQ*4mBcQX+pY8E&Mxv`L#cdY}?=wlDC#S{~#3z#Te`R++CO5NFfOwr)@%F z&kz~8%5HB{UawxnW{sMv-fZqWbxQJ?U%4{%r(AzmN~)b_i+`WjL{0XJZ9OQ>Q8h;j z7IRw5FMj&3pw*3r9B%DT39-Uo!p56)`HP>M*GpA^Z^kqW{AH^X*+|BbHF%Y3*^^-A z?Tl_xs>0_F?=;({rDjsDJ zOh`6nqp1nF>A|)ah!B{X(j-~jcUTE`kZxRBHc{hB$X?9l4_A zE#*5JhUGZ_x{Z}ilyN+LHmS$wY#z)Gk<3cBk7CvwnJAe0%?xmc_9%6`wlFwWLir^H*ijfz%dw?*S|*GRIrMV5 z1OOs{=Ex?lO?b??L_F3k1EEmH0oJT^#tz14^&^5CXF4SzvAn`Zm(8=0O4$@^S;FJN z+p1V?USX=|sm0_FrpeG%HI~soK|Y<4w``lz3tVVV_%xG?{xzsEFyUD*Y3II6Iv8&P z_1sb?1V@d4DEKY^1TFVtLK(XWyg)T?6_p3{;u@GUGlXt+WLcskzcPoF+AOXQ5D_8e^Y37nO^Axzf+!0Gc+`q@MycBT@uaRPr#@81oEF;J=h1H8m0L7LWzpB$eNq+ zAb-yoI6jBtA-=op6CL2XpW(^MOKv?>Pd3y9ojd9wE$|)*qJ9o%u}iD!L^g?BmFugI zBtHN*pf+XsVwk#Ims(Ug$tM~_aW7bUb)OnKM8V0|=5M;a>El2uoP&>WBpbY$GQHe% z)&ie~TIC~rh*_h*t%2HV(<=J%()= zC!_jrEzCC&&F*8LjZ)tB5dh_OfYX5%kpb_%2(`Ng^-|2$Y z`tATAO=TYu2#R0BNF%tI<#;dE@XKV(sYGbkcf6B|G_d#VbE6B}Ga}DNwdzy$F??z_{yeKm$;ikJ>oMgf|xQrNp%Hw?e@0(lp$q51)J`Us>(MtJjlk) z@vWB||JZMa+W^6*fE=UQA;{e?F=O|5Hck_2L zd-lFs16hEg?nla_(qey}&ai_OIY`+4Q@=-)Qw)}{S(O8w?eoMQFZo5c)LhWvAFDMO6SyXn;J9CM@N;Hgey<9niO6mF*%B|he&>z&Ao$THL) zbe+F0%@NV*t}{awQPlTV)5jp=2(kC2as=LekN&iD?_owaY32r(A8Hav%}7M?$0G(z zoh&Uqqy5`w4tQp{`@X>H+s9eBw%ZBEmrAzn+8>oO(0mkB`c{F0kkmIA&><78CF0a6_vIvec+x`>}_VLWyP6(V~(+d>K|K8WyGhHx?K zbhHYtd92$y!zl_XamRe)H?d@-%(^XN+-aR2Nswy8;65A3E7c!34`Q<#Gvshnz!^nO z2+kzwpl*i{=UQ!}G5GLYyfV@wpFnGR!8=IbRK#aU8#h`|^ZkjR`*GcSGzWf3q=bu9 zjTf^YAC9!AYNr6nZEHZxFj@r9_@3)F9)w%9mlVDlUTw0W=7z6(zsN$$dM#m9cS<|% z`ZUi~O^y^f&HyvZOVFv=a)bW|=adyqlS89fkt3Jq(2%UvG<`!g1-sjjQ~dDhKk8hI zbyh9qTRt%p##9LAM$Ptf!>ZNTBp@QIS5>KxM{QESdBQ8HAT)bS*9*-mw8TJ>2B_I- zk*!&E@0QOS52_e+*pKATU+$Yqbm)qHOZ)e%N4v7aMqT!yK~qj&8k_*4v$TjRqsLd7 zoiH1;t*L}pGUprgc?y&(_qRP_lV{PX4LP^-&9R!Jgw-1rr>vX8K|lNUyd7ID;&^D4 zwU1JUHQ#8e%q8uDlh^AVqmq#cvQt@CS!goOGkU+!_T|hfQeor?J?Ie2>M};Lz>i(B zAI{+bAHrkf9A~8x4|37k{w<_`fR$ssH|MJ~)5F^3{-jAe>D*hZ<-ad2n6|Q8ou{X- z!lpql;&}HtYCkh|wD*S=JV;0jGD^KM%x@}}`(a&*_J;>S|JZkZN&yA!%$Gl$Ju|y* zId3Q0hkCU5y-AhLjX}@+9u|cvivTH~<92l8vc~AYUv-UIH+n9@gFq3Ul_)6)$OTbf zb#u^Ww{S8Z$_`{kJe*Dj%D$l~D`NBRi11efSrX?`=SKN|(b`^7%F;Bj9;09>bett} zc|$RNzhZ9%xALHSq@9V$Uz3eD1k2k6NI6uaqpP^XIkjVoF+%~-DKdvq{>B*B-OCvw6T7pgH z$79KW1VFvV%7?AphAa#sxMBVA6Lo(Iu35688SuM#bUC$w;U;Ded}AetIcU5hpf2Ww z=$C8R%73)(;O92}nK){j>Ltdjv8*=knp@Y+b5tyT1~YZ9xkQKZ zE1%Dz<8d~8H%2U#nt>wLfatJz0aP|mndZ0srY@OM_Jaon-)@+IT4T^79< zjGOiWTeB-0jak2&LugAOk!ugmL2o*)j3jtzoq*zQS{{$i6BR!uq8%$P?auPy*+CX( z)vPynX-??70$fEQ83y)n;|&7yDCh?bYQXqde_?8G@NA`N&>r!m$^|n_jqSyp>6aaG z)x)k%o8O|UX6xVNBoHl%4h8;ytU%9SM@23V7Si*+T>E#Rh$9d89oeIs!O_bIZJ?)l za4aqXrbmj6@6*4r<&(bBom&IE0&-iLQ4=>Jj%pI-AfjJVE{W4otHQS!Bb#V(a^45{EM;6O@MyrpuB z_BM_mT##WP-AHXHfAdV}iPs5Oc?UWuWofMV2FIN=BL_ir_CRux`MP&Wn%-;w`9<8w zdtf7Xlbc~#PilMT&FG8}c-wrDB3q3TwidC=jp)mm4pM@(;4cUC4Rn2_7m_@ZFr0wJ zLE;&8v?a8gX^p&%PNhAKSC5Vy%ayX(PU5_@II3h+)m zEG?Mq*aEV7OqHO^FRdQvYU|;p^hi_)KhwQkM^&suBb}{29$l*-v$c-YtPHh|bjjkr z$ue422y|vZLwxS>xW&rbwC?<%P!5^!e&9hPldEIdqtZq)D8kqZyW8~z*e=_al!woP z+~4EyPEpnJ&%cLJysYo6(r-H{sj3 ztt#(j-n2f%6e{gOLHtzAddVI5P-cqE*7_umADERz(d&K6n8{JQRMrFnSASrAWX4|= zKVZA3r~%*8Hid|(S}b6Cc;mj#WJsK%2QAGYHa1>}B_hI1i|1QTU^?gp<^6!j%ip5l zadiV5LuD7d+_gohf*m!w1X5D#n5{CwmPqov{Jl6w?H;G!>Hw&7QLBISHMQ=1;)X}o z%r*0EWKG35U<3$Yj-M59o4C}%zRd6=P(G(to5Kw<}-6MpwHDrWz zyas(P*HTh|9xNdEFkUE=cFb&IKbrN&k2LS_>OH&Q^=qsrCEE?@CPvkIzOAAPZhL=x z3;;@uwiUO4(U5vkQRgw*j>&*}349P53*1&u=&8h;Ur*G!TS}u9vuXWS-9Y@EYe{!M zr}fAMiWZ+_Et!-_u-U<=Il*s|^R2LLYUA3DGzo)#5`^yf20`~x(gQ1E(t7f3*=#`%UMt8`bS~TfBG%SKXOl>L|@(^_Kj zp2NXMX}F{^%6{3&$sX{AXcZVAm#hq_2#`ZzLB^yn3uh1%O6EHJvUyl**+p$lSdCnN zMB%V&K(s!2^Rn5yXB){XrQcKFF6C+7o$rzK`4Ugn{P9`hu zWl{`0`N%@MlnYq6iM{5+Ebo;Rz59oqx?SJJ;K~DXpTShk?d&%e*J1zWTIS-*d(z+ zwO?AL2m2w%*FHkJ8_&QihUb1~S!b{Tx7U;15*0l-i>yE&+~c$zUZ*q4XnYj*M6YMc zr!iX%ZQqtUIZle*RijOVtWFZu3TaJ!dGKZFKN}nRLW&4Uo}I72f7%2(o-m zFsyk11Z}0{D?fwqp7QpZA^$v$cR`iF*Fu31~~b*`SUk3tp# zW4fHBbfL4K8tlyJL&*?(BsV1aTQK9iwNyJ?-|9eUq&-RANe_3Y4W9dmuK?DU`M$$7 z+W<-txhky12w+~$I8H$_mEdl_o_KZ#)~fP*^m*46jW;A<2PtSdQy>FXQ8Y zN%mQ&$8iN-F7-XVf`zCh#}R=Zdx0N%Xs2SrTT*tyrEyk7KTFA*pQ4bVo3CE8<1*Q4ZjDB%t!4nmeKY1=AY|C*^MCaNVbKFW9(iz7lJ3T_QJ2QLie9hd;1 zzc%B6A4?Y}$UO@hxsM9A%0AK^H4wKnCrY>5ylvRwYOxD)44R%}CL5~f-M{b`r;Gf4 z0g{niw^LFTEAZ@5QNqZvm3(Bcm?FAu^|&~qs)O7p7(05*efMECEUGp5`u{q)1oEKYL~kiMhETd3H^*f){I{C z0)*Xs|04P%F>zt)$F%4fQ(#*OO!ALF%zK_wqtf|lY^VgSGK6vzD@mS~`LasX9J_;pVCqxr!qQ)15$uC#o zHUDj_n`t+?MfHa2m*(aIEh3R6hCx`6|2(SsL00_+yRk*Eo{L2ZH~xOzXL4s~Ua-XP z!lq7cQJJ=(CDsL@W!`^YPSuszeK7%j!fIaD8u&Vm2eBDx8p$t+1p^41IC3bd1s*B| zno#G&7~;1N6_xDo_7W$U)hm;owgJ9R_zn)Ji;eIW z4Lp5l)iuNi<0$313ZPGbN8{ayn`0`=cM zy3E3*hf^%c&cA*-#1fQY+qh{CT?j>WxOTX6_oE^ngVw8bICqx@*_>B6E*}=Ut#@pp zzuTkBE?jpHItnD$DO)MbU}4RxY69WM@tI}hy7BpnbfwLe|EdbhZ!DoWc2YNvsyFu` zyb24=MilxB^iDmf!VMMY zO5iUhknS(^QVQ@%{v0tae`Sx(6&{^1SPIJZfaP0-GtvtIS$=(5xr`iW@2Ab@zyTG6 z!MuEw)RPmQo`_B)gP**x9I5RkHux<5YMlIzIIR0K<94!!h>0@#Wm3T#?4Y# z=gy^ZP$nFUEteX~ajid5cLT^x-2C)XW9kToY-|d1*^b$Ts-h4P8@oIi+ANR$GK_cU zjo|}Ef1AWb6k~#cQta{8je&KTJ1k3;B1`~0k}4gJtB9^MWD}BkuCCgpcfzF~7vM~j zF?{kXeceLJ(!j1XdfV36RcD_T-*rYgci(62%Eawh51twdXejva9`%b?GIqVf+3`NUEK&wD*GNhAo0pgGC?B$%dhA(}WA zd9@+~3}tN7AZ%H;qST9OG+ChpB@OvTwj?|BLuoOhs$iFsu$XTIl`D11&nkQcYPbb$ zDm@5b=1Xa9u}rKmfbe>xg|>=`+ti*fCsW2QEjipt4=8QMN*Ng(2HXl%%&7?6pmDAw zoyi+kHIt>n-hgp$O84BJD78<1@*G66CM#Er6Yy>Vl#11_5&Yf|Mq;|~KtEMEY$D#Y zl>w;ZpK2{P;N4!{K>t|!0)l!=`V+R^WDJB|1YytEd#qz(Wl3qUsr34C$~2kg5Oi?! zeC#S^uR!ve(H9^kx)(TY&GhPiGEN^MD0NZt>}4U*dczRz zNW*obh`Gb;pn3jCDQgq`m(o_2&YG@SHXTte zRr8UNgshmc#2`wEajdZlib@F@3sIPru^QEKxydW3mEE zQ{C1)SdDRc<48r8h9PL6-wXbb0ebSWbK+SUX_|Qy%Y)S)2M`6IyhskJj_gPeTyE$4 z!{WnywHW0`1ipfAUO1?b+4Ow78qE4GeBuj;8@eit4@r{)nqf8XkEr!G$di4kSZcK?U#~w2% z9R9kae1$_qdP(ubW*C+@$*0c%DH#ic0=0sDGt&0yfq z2~aGnrYt(!OtyB`*QH|^3y1)EP8kRr972-|g-au%Mn3?h7#ZAE3-_WOJ|1{|ZKnyJ zF2RGx2=Qvk5C|MffB|P^;>S1;W8cGNurt=mD8H^g9I*9N3aE(o=44D8>MR$R@#4}4 zylK7-dq5-3mr(0UZAA#5KL|`;Zq)1qyrA-z#{dPum(U`GQcT{yxMG$VNh7kO-&i9m zi;_rIK9jh0MhYJLNy_M0f~E0E%DFNH)?-$<8yePWAp^sPL^eZG2Q&Jb%G5Ce>jdXt zkxM*c{Dbm!(^v(I?~p$YwmXMf?6i?kUDD$jEOghy$4G5}ogC+AGWS#-I19|2HGp|- zeQRA4Z=z>n^Tw5v^TOo3q!zt&uslOZb4RUl1}8^&pCdt~QeJwWPpgEA?0lcYDT_9p z)^Fxc4ckbDgx?ip+WDovZmhso4z#&sm2UTRd1cPiQ|71DW@oUn#i(l&|2}s(Wb6v` zWgOEZycAexEt(uC3fBg*JEn^3$4GB-BoG3ARPWzjVb`q;Gp?oTtu8C}0unY2l zzAbzH>Z)<1yQzAJE}_j}RWl{TvDEC97G7$nn*k%Rh~KI-*)9b3`zt{{cny0jVFSo3 z;f?iIyX0W9%a?gp9k?p=2Q5;UV(0ZCoY7ZeG51)^ul0mA3*aOp`pu)HieWk9^0T*Z z2jH!7%)rf1`y+m4{a?kq|iymD?Mw=Sy!%>77 zv0%ev$r0LzXI+s7f^AsyKiYKua(wf5MC3DIoxG2MdtB6F(MG^@=FDJStQy->`{yb1 zTt@CJJpnK(`XXQm!FuS>=fze@Qe$m$;WB`mL5d*=60rqr=bxeI$nH*{jSs(Dp+X2a znCFd7Ynb2dZ&g^rB%l1nv%u($ZTLi3qH_@4Oy?GDt#rCNQb2;8(u#srv)wIe6~Djm})F4U8Bd@LeEt zV(!jCQiQ)?3ggK)t}T&(aP7?aT5y)Wdh*yCPGlQkU;WwPrlZ?S818wjVzW*iVio61 zl<7r>+kS<`yQ_)Ds*cA$y=inaPH6#*=ef?;oi%l^*30pxk3$kpKAXyCxB1?nyywsLPt=QIT)sbWNR_K9V$epCI7_Qvbl-mFc-nLDfWgr?VM>}t`zdKRBltiGUo)Zv`Tw~(^6 z$?}jIV&rdQ#pi!EEk9eaWcwYB+ibUnm&8_zoW&qAaUojSq!^?Gv65^e5TNQ3IB zy7d%RNZh7>{l@e`RC;0g_~Y^yaJ)DA=1m%*(Y`^F7*4d&)z2aJ#ee&jPwOC`zyHbi zVdUeVOnm#5U9%UjZiXY)9w%vp11x7!;>bm29;=K0*Yy&f!dX}+BO|cxK_3)i>TY@y z_!Q~*$B)1-0UNr?=2TYkLONO^dC|sQW_vn$>`}v&pR!^5A~Rk_jeBF?Klyn4-P-cp zizU-sMC#!Q+dRrHu6}usuJg)nS;o#aYMFlwUB_l(%R$D|N$NE`2&iB+8T^kMLxb1O zoH(Tfe}W9=3H@+Vz>ZT-t8Ppb(i9cnUVjNE@Ud7rLHPGK0-q;S9`n1L=)N`py63B#tkQ0 z?Xx!D(3W_aolG}_6_n+hf*udFbEpxlairC*g@(QL68hEm%MJo9;drDETOYJ@{-(;q z+=!IRSlw3k`o_+uKd8>dGx@-)v(35qt2aVD2PVAiN#$JDG~8p#@$!a37V6Qxx#Jl` zAjc@XAnmOLuL&~JuuAt&I;BNnsdt?ZAc5^Dxa-3SXE|4Kn;yz&?zY7S)z z{4UhDB5tD?!f4w;Cul32{(NA7^`ecdx1qV3#F=<(rJG#b38U1@;Hx_t(&E!&ZSoCz zI^RR#Bh>HV^>@1?l&JgldvZWDRUw!CA2QTv$bCnB<{kC7@$IlUW3J^JS7fyUwTJ*B z-b}>9BzXUmAV02bK=s*u%u^ypaW5r}QDGiNh!Y=gcJ?n?C4xW6Vs5VrLJnAy&zD4@ zBvVd+jQPi{-Gc&On!HKOyOIwvg7MWmjaPAkNz_;MFBSg;P%nyzhv30O!FW@0fn@{5 z*uye&uH@r8VSQaW#>%MbLkq67bY~LWby*s&RF$_HGhILJgN@jgv_)%m4pOdU7cN4? zBTQ#vOw6XrB$-z+|5S~D*(^`35Fc-;w&%CSpQ7cnvE=04U+Oq3ZC0EK;iNrexSB%+ zW&4&Ktg^q}E!$;z*go(&Kcn215tmwbbv$(&FOYACSzHC&Cz8GM9 zp6;(1_4|Sb^${^JwW!t(o!oly5Y0JgXDmmdx!{Hc=bLsyPjQs$o?P z93<+8s!Q&**zz&GN0)u&kDALN|0%)Ov^=raXdE@C?6SL|T$_rHDeMw{H54Q}c1f#N zA&W`P7Fm*=zqD#cP-$?H$sa!2~gcx zmAQtB&hF9AR^1x+M?ORc(4RS;wxKikfU zqes*AwbNex5}%>}R-4Y6o?h`-zcBFka&mqt+`PEPYK#ki(q5!(ePBe`;Py6T}o z7v7NDzY2X8=>57pnf)i<{(`@s?GV-!1;5Ux%NY=paZtAMP`kb@ z$DTeNGi~L{{-ZgWqgunCWW8=t3x0UtUA&z5mF+*^!*9;`Ru_dDj6vFPULwHH>Tr5O z=C~K6uzEEp>-^C6BE9$;6g{G$83X)X9^QWwn3|h=U~AhVFN%H>iJXgpwQVj*UN~7` z#K2QO=70=2P%9lD2 zFy`$DRJKL|Vm5xyTSOs8<;Mr>Bdd_nYG2Q0lbet5`On78w&ITmc(%lGp@*_3j}5=L zUB~-1u@=ps@Hb$R--13$Ul3beE@l&!guNzB%@YMXU<2sQSsq-F_EG|qL$WKFUYafj zf2(m5ck7VLRIx(4FVCikE^Uq{Nfy8hY5iLLg>!Bq&Irr3t>A{B)$Z7U%1#@I`b6h( zUNpGr5AlnM=2wPBClQ%Dl{7LanBm=~e@y!Na+p3B&yB|rb;5FUJ~7iJ4$0deVj>Yl zQ^)peAael*)fwF&qqdFkdZ!_%(5B|=!K&Qqy^K(0(k7}|LtbWg9dm0)+7u&8)~ zjk=e`Z^o9`jaF}4nVlhXZstteIh8K6&G_cA9BO2bG$%KFL#&}L$+>EWwOk96jB0nf{*Pr)fja1Dvi9D=aEr~3D~sZ!P(S>iW=s>j%4LvFZ#ne zC!!MNW@Z9#0RRAICJtsM78Y7203{O>;Y(*2N5lQBhS) zlR@0g*4Eg_-u?%uPNiyTb3Si@6W@n>kW?=u{{rrjl z`E0UQ?q)rXTvZ0(XKg3HuG$cpJ z_Q4|(huz(>#%fR+Wrmx#yqug*^QXX63Id~S#qp>IsIpinxljdY6;Xg(29f_BDl!uN z-YM#tl{l3*m71p`qQDuCClXC@uyETF2qV-vEfqq6S0JVo*62Sm;T+i@N2prz` z7wJz**J-4Gsl~Zhipw$3co;q<_DAt!UtHN^Jw?#3HC(PxXUsvSseD1M`ZY(ojNzN= zDDKVelHX^4)H;u$Yp@mKh@0A}s;O?dn%$SdKhbNCV0pObOwSCFY@qRF2T$Byltyag zQ8hbQ*GYgF;p`zxF7_KLh687@CwCT^rlPI*)@VdjG~N>!Q6vPGhNsO<%~eXYj~|Qt z_5?o3kXxKt>gpF3W0C$<&3A255F<07kqw<5!0 ztqo1mn^^#<*V?Y|!7M&AG2(UTLCrbiR&QBuGzO`yod?5H+WOPg8$xP5D?TE$@c%vZ zkixF{psUc?Xs)*MT8uV{xT=A6)EoGka)9EMTK-qg&KNQa3C zz{;ct!>D5AZT6og45J2-4uFV-2=McwvV(&w5rBo`=Mc*O*~8r7XX`&>^gmmoL^`}; zTx_gdYyb{XF%}jPA!czlQ6UZyVNqcrQBENi02eDi(f@zQ&v|6c>@9wBAPXz=|H3vA z{a=h_{Dn4YF(st1{Y;l`^B(VMuH)<`eU}-oGR*oGBzBrLOX9URF#f>QwGB;ZV_NhM9_K)6<{&QTfOC+-%~R8K zy5_;R=RkYrAU~OabeNz->fwj#K}qXji)?k_KNTc1oW z1Y5smmWVB{D9FfEDsZHwEyHfbR zpU0DM!JBqs9aF-Qu*y!QfRZ}P6ZlGhw3hw?($!+l`#&!q7gr-^SI?hm35J;kz|6`F MLryNHAP)2Y075f!N&o-= literal 0 HcmV?d00001 diff --git a/doc/Projects/2019/Project3/pdf/Project3.tex b/doc/Projects/2019/Project3/pdf/Project3.tex new file mode 100644 index 000000000..7debfcd01 --- /dev/null +++ b/doc/Projects/2019/Project3/pdf/Project3.tex @@ -0,0 +1,392 @@ +%% +%% Automatically generated file from DocOnce source +%% (https://github.com/hplgit/doconce/) +%% +%% + + +%-------------------- 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\ 1999-2019, "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\ 1999-2019, "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 + +\newcommand{\exercisesection}[1]{\subsection*{#1}} + + +% ------------------- main content ---------------------- + + + +% ----------------- title ------------------------- + +\thispagestyle{empty} + +\begin{center} +{\LARGE\bf +\begin{spacing}{1.25} +Project 3 on Machine Learning, deadline December 15 +\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} +Nov 12, 2019 +\end{center} +% --- end date --- + +\vspace{1cm} + + +\section*{Paths for project 3} + +\subsection*{Defining the data sets to analyze yourself} + +For project 3, you can propose own data sets that relate to your research interests or just use existing data sets from say +\begin{enumerate} +\item \href{{https://www.kaggle.com/datasets}}{Kaggle} + +\item The \href{{http://archive.ics.uci.edu/ml/datasets.html}}{University of California at Irvine (UCI) with its machine learning repository} +\end{enumerate} + +\noindent +The approach to the analysis of these new data sets should follow to a large extent what you did in projects 1 and 2. That is: +\begin{enumerate} +\item Whether you end up with a regression or a classification problem, you should employ at least two of the methods we have discussed among \textbf{linear regression (including Ridge and Lasso)}, \textbf{Logistic Regression}, \textbf{Neural Networks}, \textbf{Support Vector Machines} and \textbf{Decision Trees, Random Forests}, \textbf{Bagging and Boosting}. You could for example explore all of the approaches from decision trees, via bagging and voting classifiers, to random forests, boosting and finally XGboost. If you wish to venture into \textbf{convolutional neural networks} or \textbf{recurrent neural networks}, or extensions of neural networkds, feel free to do so. + +\item For project 3, you should feel free to use your own codes from projects 1 and 2, eventually write your own for SVMs and/or Decision trees/random forests/bagging/boosting' or use the available functionality of \textbf{Scikit-Learn}, \textbf{Tensorflow}, etc. + +\item The estimates you used and tested in projects 1 and 2 should also be included, that is the $R2$-score, \textbf{MSE}, confusion matrix, accuracy score, information gain, cross-validation and/or bootstrap if these are relevant. + +\item If possible, you should link the data sets with exisiting research and analyses thereof. Scientific articles which have used Machine Learning algorithms to analyze the data are highly welcome. Perhaps you can improve previous analyses and even publish a new article? + +\item A critical assessment of the methods with ditto perspectives and recommendations is also something you need to include. +\end{enumerate} + +\noindent +All in all, the report should follow the same pattern as the two previous ones, with abstract, introduction, methods, code, results, conclusions etc.. + +We propose also an alternative to the above. This is a project on using machine learning methods (neural networks mainly) to the solution of ordinary differential equations and partial differential equations, with a final twist on how to diagonalize a symmetric matrix with neural networks.. + +This is a field with a large interest recently, spanning from studies of turbulence in fluid mechanics and meteorology to the solution of quantum mechanical systems. + +\subsection*{The basic structure of your project} + +Here follows a set up on how to structure your report and analyze the data you have opted for. + +\paragraph{Part a).} +The first part deals with structuring and reading the data, much along the same lines as done in projects 1 and 2. Explain how the data are produced and place them in a proper context. + +\paragraph{Part b).} +You need to include at least two central algorithms, or as an alternative explore methods from decisions tree to bagging, random forests and boosting. Explain the basics of the methods you have chosen to work with. This would be your theory part. + + +\paragraph{Part c).} +Then describe your algorithm and its implementation and tests you have performed. + +\paragraph{Part d).} +Then presents your results and findings, link with existing literature and more. + +\paragraph{Part e).} +Finally, here you should present a critical assessment of the methods you have studied and link your results with the existing literature. + +\subsection*{Solving partial differential equations with neural networks} + +For this variant of project 3, we will assume that you have some +background in the solution of partial differential equations using +finite difference schemes. We will study the solution of the diffusion +equation in one dimension using a standard explicit scheme and neural +networks to solve the same equations. + +For the explicit scheme, you can study for example chapter 10 of the lecture notes in \href{{https://github.com/CompPhysics/ComputationalPhysics/blob/master/doc/Lectures/lectures2015.pdf}}{Computational Physics} or alternative sources. For the solution of ordinary and partial differential equations using neural networks, the lectures by \href{{https://compphysics.github.io/MachineLearning/doc/pub/odenn/html/odenn-bs.html}}{Kristine Baluka Hein} at this course are highly recommended. + +For the machine learning part you can use your own code from project 2 or the functionality of for example \textbf{Tensorflow/Keras}. + +\paragraph{Part a), setting up the problem.} +The physical problem can be that of the temperature gradient in a rod of length $L=1$ at $x=0$ and $x=1$. +We are looking at a one-dimensional +problem + +\begin{equation*} + \frac{\partial^2 u(x,t)}{\partial x^2} =\frac{\partial u(x,t)}{\partial t}, t> 0, x\in [0,L] +\end{equation*} +or + +\begin{equation*} +u_{xx} = u_t, +\end{equation*} +with initial conditions, i.e., the conditions at $t=0$, +\begin{equation*} +u(x,0)= \sin{(\pi x)} \hspace{0.5cm} 0 < x < L, +\end{equation*} +with $L=1$ the length of the $x$-region of interest. The +boundary conditions are + +\begin{equation*} +u(0,t)= 0 \hspace{0.5cm} t \ge 0, +\end{equation*} +and + +\begin{equation*} +u(L,t)= 0 \hspace{0.5cm} t \ge 0. +\end{equation*} +The function $u(x,t)$ can be the temperature gradient of a rod. +As time increases, the velocity approaches a linear variation with $x$. + +We will limit ourselves to the so-called explicit forward Euler algorithm with discretized versions of time given by a forward formula and a centered difference in space resulting in +\begin{equation*} +u_t\approx \frac{u(x,t+\Delta t)-u(x,t)}{\Delta t}=\frac{u(x_i,t_j+\Delta t)-u(x_i,t_j)}{\Delta t} +\end{equation*} +and + +\begin{equation*} +u_{xx}\approx \frac{u(x+\Delta x,t)-2u(x,t)+u(x-\Delta x,t)}{\Delta x^2}, +\end{equation*} +or + +\begin{equation*} +u_{xx}\approx \frac{u(x_i+\Delta x,t_j)-2u(x_i,t_j)+u(x_i-\Delta x,t_j)}{\Delta x^2}. +\end{equation*} + +Write down the algorithm and the equations you need to implement. +Find also the analytical solution to the problem. + +\paragraph{Part b).} +Implement the explicit scheme algorithm and perform tests of the solution +for $\Delta x=1/10$, $\Delta x=1/100$ using $\Delta t$ as dictated by the stability limit of the explicit scheme. The stability criterion for the explicit scheme requires that $\Delta t/\Delta x^2 \leq 1/2$. + +Study the solutions at two time points $t_1$ and $t_2$ where $u(x,t_1)$ is smooth but still significantly curved +and $u(x,t_2)$ is almost linear, close to the stationary state. + + +\paragraph{Part c) Neural networks.} +Study now the lecture notes on solving ODEs and PDEs with neural +network and use either your own code from project 2 or the +functionality of tensorflow/keras to solve the same equation as in +part b). Discuss your results and compare them with the standard +explicit scheme. Include also the analytical solution and compare with +that. + + +\paragraph{Part d) Solving eigenvalue problems.} +Follow the discussion in the work of Yi \emph{et al.} in the article from +\href{{https://www.sciencedirect.com/science/article/pii/S0898122104901101}}{Computers and Mathematics with Applications 47, 1155 (2004)}, and +use your differential equation solver with neural networks, set up a +simple square, real and symmetric $6\times 6$ matrix and find the +eigenvalues. Compare with the solution from numerical diagonalization with standard eigenvalue solvers from linear algebra. + +\paragraph{Part e).} +Finally, present a critical assessment of the methods you have studied and discuss the potential for the solving differential equations and eigenvalue problems with machine learning methods. + + +\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 Devilry to hand in your projects, log in at \href{{http://devilry.ifi.uio.no}}{\nolinkurl{http://devilry.ifi.uio.no}} with your normal UiO username and password and choose either 'fysstk3155' or 'fysstk4155'. There you can load up the files within the deadline. + + \item Upload \textbf{only} the report file! For the source code file(s) you have developed please provide us with your link to your github 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 git 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. + + \item In this and all later projects, you should include tests (for example unit tests) of your code(s). + + \item Comments from us on your projects, approval or not, corrections to be made etc can be found under your Devilry domain and are only visible to you and the teachers of the course. +\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 --------------- + +\end{document} + diff --git a/doc/src/Projects/2019/Project3/Project3.do.txt b/doc/src/Projects/2019/Project3/Project3.do.txt new file mode 100644 index 000000000..6b2cc503e --- /dev/null +++ b/doc/src/Projects/2019/Project3/Project3.do.txt @@ -0,0 +1,257 @@ +TITLE: Project 3 on Machine Learning, deadline December 15 +AUTHOR: "Data Analysis and Machine Learning FYS-STK3155/FYS4155":"http://www.uio.no/studier/emner/matnat/fys/FYS3155/index-eng.html" {copyright, 1999-present|CC BY-NC} at Department of Physics, University of Oslo, Norway +DATE: today + + +======= Paths for project 3 ======= + +===== Defining the data sets to analyze yourself ===== + +For project 3, you can propose own data sets that relate to your research interests or just use existing data sets from say +o "Kaggle":"https://www.kaggle.com/datasets" +o The "University of California at Irvine (UCI) with its machine learning repository":"http://archive.ics.uci.edu/ml/datasets.html" + +The approach to the analysis of these new data sets should follow to a large extent what you did in projects 1 and 2. That is: +o Whether you end up with a regression or a classification problem, you should employ at least two of the methods we have discussed among _linear regression (including Ridge and Lasso)_, _Logistic Regression_, _Neural Networks_, _Support Vector Machines_ and _Decision Trees, Random Forests_, _Bagging and Boosting_. You could for example explore all of the approaches from decision trees, via bagging and voting classifiers, to random forests, boosting and finally XGboost. If you wish to venture into _convolutional neural networks_ or _recurrent neural networks_, or extensions of neural networkds, feel free to do so. + +o For project 3, you should feel free to use your own codes from projects 1 and 2, eventually write your own for SVMs and/or Decision trees/random forests/bagging/boosting' or use the available functionality of _Scikit-Learn_, _Tensorflow_, etc. + +o The estimates you used and tested in projects 1 and 2 should also be included, that is the $R2$-score, _MSE_, confusion matrix, accuracy score, information gain, cross-validation and/or bootstrap if these are relevant. + +o If possible, you should link the data sets with exisiting research and analyses thereof. Scientific articles which have used Machine Learning algorithms to analyze the data are highly welcome. Perhaps you can improve previous analyses and even publish a new article? + +o A critical assessment of the methods with ditto perspectives and recommendations is also something you need to include. + +All in all, the report should follow the same pattern as the two previous ones, with abstract, introduction, methods, code, results, conclusions etc.. + +We propose also an alternative to the above. This is a project on using machine learning methods (neural networks mainly) to the solution of ordinary differential equations and partial differential equations, with a final twist on how to diagonalize a symmetric matrix with neural networks.. + +This is a field with a large interest recently, spanning from studies of turbulence in fluid mechanics and meteorology to the solution of quantum mechanical systems. + +===== The basic structure of your project ===== + +Here follows a set up on how to structure your report and analyze the data you have opted for. + +=== Part a) === + +The first part deals with structuring and reading the data, much along the same lines as done in projects 1 and 2. Explain how the data are produced and place them in a proper context. + +=== Part b) === + +You need to include at least two central algorithms, or as an alternative explore methods from decisions tree to bagging, random forests and boosting. Explain the basics of the methods you have chosen to work with. This would be your theory part. + + +=== Part c) === + +Then describe your algorithm and its implementation and tests you have performed. + +=== Part d) === + +Then presents your results and findings, link with existing literature and more. + +=== Part e) === + +Finally, here you should present a critical assessment of the methods you have studied and link your results with the existing literature. + +===== Solving partial differential equations with neural networks ===== + +For this variant of project 3, we will assume that you have some +background in the solution of partial differential equations using +finite difference schemes. We will study the solution of the diffusion +equation in one dimension using a standard explicit scheme and neural +networks to solve the same equations. + +For the explicit scheme, you can study for example chapter 10 of the lecture notes in "Computational Physics":"https://github.com/CompPhysics/ComputationalPhysics/blob/master/doc/Lectures/lectures2015.pdf" or alternative sources. For the solution of ordinary and partial differential equations using neural networks, the lectures by "Kristine Baluka Hein":"https://compphysics.github.io/MachineLearning/doc/pub/odenn/html/odenn-bs.html" at this course are highly recommended. + +For the machine learning part you can use your own code from project 2 or the functionality of for example _Tensorflow/Keras_. + +=== Part a), setting up the problem === + +The physical problem can be that of the temperature gradient in a rod of length $L=1$ at $x=0$ and $x=1$. +We are looking at a one-dimensional +problem + +!bt +\begin{equation*} + \frac{\partial^2 u(x,t)}{\partial x^2} =\frac{\partial u(x,t)}{\partial t}, t> 0, x\in [0,L] +\end{equation*} +!et +or + +!bt +\begin{equation*} +u_{xx} = u_t, +\end{equation*} +!et +with initial conditions, i.e., the conditions at $t=0$, +!bt +\begin{equation*} +u(x,0)= \sin{(\pi x)} \hspace{0.5cm} 0 < x < L, +\end{equation*} +!et +with $L=1$ the length of the $x$-region of interest. The +boundary conditions are + +!bt +\begin{equation*} +u(0,t)= 0 \hspace{0.5cm} t \ge 0, +\end{equation*} +!et +and + +!bt +\begin{equation*} +u(L,t)= 0 \hspace{0.5cm} t \ge 0. +\end{equation*} +!et +The function $u(x,t)$ can be the temperature gradient of a rod. +As time increases, the velocity approaches a linear variation with $x$. + +We will limit ourselves to the so-called explicit forward Euler algorithm with discretized versions of time given by a forward formula and a centered difference in space resulting in +!bt +\begin{equation*} +u_t\approx \frac{u(x,t+\Delta t)-u(x,t)}{\Delta t}=\frac{u(x_i,t_j+\Delta t)-u(x_i,t_j)}{\Delta t} +\end{equation*} +!et +and + +!bt +\begin{equation*} +u_{xx}\approx \frac{u(x+\Delta x,t)-2u(x,t)+u(x-\Delta x,t)}{\Delta x^2}, +\end{equation*} +!et +or + +!bt +\begin{equation*} +u_{xx}\approx \frac{u(x_i+\Delta x,t_j)-2u(x_i,t_j)+u(x_i-\Delta x,t_j)}{\Delta x^2}. +\end{equation*} +!et + +Write down the algorithm and the equations you need to implement. +Find also the analytical solution to the problem. + +=== Part b) === + +Implement the explicit scheme algorithm and perform tests of the solution +for $\Delta x=1/10$, $\Delta x=1/100$ using $\Delta t$ as dictated by the stability limit of the explicit scheme. The stability criterion for the explicit scheme requires that $\Delta t/\Delta x^2 \leq 1/2$. + +Study the solutions at two time points $t_1$ and $t_2$ where $u(x,t_1)$ is smooth but still significantly curved +and $u(x,t_2)$ is almost linear, close to the stationary state. + + +=== Part c) Neural networks === + +Study now the lecture notes on solving ODEs and PDEs with neural +network and use either your own code from project 2 or the +functionality of tensorflow/keras to solve the same equation as in +part b). Discuss your results and compare them with the standard +explicit scheme. Include also the analytical solution and compare with +that. + + +=== Part d) Solving eigenvalue problems === + +Follow the discussion in the work of Yi *et al.* in the article from +"Computers and Mathematics with Applications 47, 1155 (2004)":"https://www.sciencedirect.com/science/article/pii/S0898122104901101", and +use your differential equation solver with neural networks, set up a +simple square, real and symmetric $6\times 6$ matrix and find the +eigenvalues. Compare with the solution from numerical diagonalization with standard eigenvalue solvers from linear algebra. + +=== Part e) === + +Finally, present a critical assessment of the methods you have studied and discuss the potential for the solving differential equations and eigenvalue problems with machine learning methods. + + +===== Introduction to numerical projects ===== + +Here follows a brief recipe and recommendation on how to write a report for each +project. + + * Give a short description of the nature of the problem and the eventual numerical methods you have used. + + * 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. + + * Include the source code of your program. Comment your program properly. + + * If possible, try to find analytic solutions, or known limits in order to test your program when developing the code. + + * 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. + + * 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. + + * Try to give an interpretation of you results in your answers to the problems. + + * 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. + + * 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. + + + + + +===== 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: + + * Use Devilry to hand in your projects, log in at URL:"http://devilry.ifi.uio.no" with your normal UiO username and password and choose either 'fysstk3155' or 'fysstk4155'. There you can load up the files within the deadline. + + * Upload _only_ the report file! For the source code file(s) you have developed please provide us with your link to your github 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. + + * In your git 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. + + * In this and all later projects, you should include tests (for example unit tests) of your code(s). + + * Comments from us on your projects, approval or not, corrections to be made etc can be found under your Devilry domain and are only visible to you and the teachers of the course. + + + +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 +o pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow +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 +o brew install python3 + +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 +o sudo apt-get install python3 (or python for python2.7) +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 +o "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_ +o "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. + +Popular software packages written in Python for ML are + +* "Scikit-learn":"http://scikit-learn.org/stable/", +* "Tensorflow":"https://www.tensorflow.org/", +* "PyTorch":"http://pytorch.org/" and +* "Keras":"https://keras.io/". +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. + + + + + + + + + + + + diff --git a/doc/src/Projects/2019/Project3/clean.sh b/doc/src/Projects/2019/Project3/clean.sh new file mode 100755 index 000000000..2e5da2c72 --- /dev/null +++ b/doc/src/Projects/2019/Project3/clean.sh @@ -0,0 +1,3 @@ +#!/bin/sh +doconce clean +rm -rf *.pdf *.tex ipynb*.tar.gz *.html ._*.html *~ reveal.js Trash README.txt diff --git a/doc/src/Projects/2019/Project3/make.sh b/doc/src/Projects/2019/Project3/make.sh new file mode 100755 index 000000000..e1fc0d6fa --- /dev/null +++ b/doc/src/Projects/2019/Project3/make.sh @@ -0,0 +1,79 @@ +#!/bin/sh +set -x + +function system { + "$@" + if [ $? -ne 0 ]; then + echo "make.sh: unsuccessful command $@" + echo "abort!" + exit 1 + fi +} + +if [ $# -eq 0 ]; then +echo 'bash make.sh slides1|slides2' +exit 1 +fi + +name=$1 +rm -f *.tar.gz + +opt="--encoding=utf-8" +opt= + +rm -f *.aux + + + +# Plain HTML documents +html=${name} +system doconce format html $name --pygments_html_style=default --html_style=bloodish --html_links_in_new_window --html_output=$html $opt +system doconce split_html $html.html --method=space10 + +# Bootstrap style +html=${name}-bs +system doconce format html $name --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=$html $opt +system doconce split_html $html.html --method=split --pagination --nav_button=bottom + + +# Ordinary plain LaTeX document +system doconce format pdflatex $name --print_latex_style=trac --latex_admon=paragraph $opt +system doconce ptex2tex $name envir=print +# Add special packages +doconce subst "% Add user's preamble" "\g<1>\n\\usepackage{simplewick}" $name.tex +doconce replace 'section{' 'section*{' $name.tex +pdflatex -shell-escape $name +pdflatex -shell-escape $name +mv -f $name.pdf ${name}.pdf +cp $name.tex ${name}.tex + +# Publish +dest=../../../../Projects/2019 +if [ ! -d $dest/$name ]; then +mkdir $dest/$name +mkdir $dest/$name/pdf +mkdir $dest/$name/html +mkdir $dest/$name/ipynb +fi +cp ${name}*.tex $dest/$name/pdf +cp ${name}*.pdf $dest/$name/pdf +cp -r ${name}*.html ._${name}*.html $dest/$name/html + +# Figures: cannot just copy link, need to physically copy the files +if [ -d fig-${name} ]; then +if [ ! -d $dest/$name/html/fig-$name ]; then +mkdir $dest/$name/html/fig-$name +fi +cp -r fig-${name}/* $dest/$name/html/fig-$name +fi + +cp ${name}.ipynb $dest/$name/ipynb +ipynb_tarfile=ipynb-${name}-src.tar.gz +if [ ! -f ${ipynb_tarfile} ]; then +cat > README.txt <