From be1e16849e85c1090a6cfec3fd85805d3c4df6d9 Mon Sep 17 00:00:00 2001 From: Morten Hjorth-Jensen Date: Tue, 2 Sep 2025 06:28:08 +0200 Subject: [PATCH] removed files --- .../2025/Project1/html/._Project1-bs000.html | 238 ------- .../2025/Project1/html/._Project1-bs001.html | 503 --------------- .../2025/Project1/html/Project1-bs.html | 238 ------- doc/Projects/2025/Project1/html/Project1.html | 588 ------------------ doc/Projects/2025/Project1/pdf/Project1.p.tex | 504 --------------- doc/Projects/2025/Project1/pdf/Project1.pdf | Bin 287564 -> 0 bytes doc/Projects/2025/Project1/pdf/Project1.tex | 476 -------------- 7 files changed, 2547 deletions(-) delete mode 100644 doc/Projects/2025/Project1/html/._Project1-bs000.html delete mode 100644 doc/Projects/2025/Project1/html/._Project1-bs001.html delete mode 100644 doc/Projects/2025/Project1/html/Project1-bs.html delete mode 100644 doc/Projects/2025/Project1/html/Project1.html delete mode 100644 doc/Projects/2025/Project1/pdf/Project1.p.tex delete mode 100644 doc/Projects/2025/Project1/pdf/Project1.pdf delete mode 100644 doc/Projects/2025/Project1/pdf/Project1.tex diff --git a/doc/Projects/2025/Project1/html/._Project1-bs000.html b/doc/Projects/2025/Project1/html/._Project1-bs000.html deleted file mode 100644 index bfb4c4ae3..000000000 --- a/doc/Projects/2025/Project1/html/._Project1-bs000.html +++ /dev/null @@ -1,238 +0,0 @@ - - - - - - - -Project 1 on Machine Learning, deadline October 6 (midnight), 2025 - - - - - - - - - - - - - - - - - - -
-
- - Project 1 on Machine Learning, deadline October 6 (midnight), 2025 -
-
- -
-
- -
-

 

 

 

- - -
-
-

Project 1 on Machine Learning, deadline October 6 (midnight), 2025

-
- - -
-Data Analysis and Machine Learning FYS-STK3155/FYS4155 -
- -
-University of Oslo, Norway -
-
-
-

September 2

-
-
- - -
-

Preamble: Note on writing reports, using reference material, AI and other tools

- -

We want you to answer the three different projects by handing in -reports written like a standard scientific/technical report. The -links at -https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects -contain more information. There you can find examples of previous -reports, the projects themselves, how we rade reports etc. How to -write reports will also be discussed during the various lab -sessions. Please do ask us if you are in doubt. -

- -

When using codes and material from other sources, you should refer to -these in the bibliography of your report, indicating wherefrom you for -example got the code, whether this is from the lecture notes, -softwares like Scikit-Learn, TensorFlow, PyTorch or other sources. These sources -should always be cited correctly. How to cite some -of the libraries is often indicated from their corresponding GitHub -sites or websites, see for example how to cite Scikit-Learn at -https://scikit-learn.org/dev/about.html. -

- -

We enocurage you to use tools like -ChatGPT or similar in writing the report. If you use for example ChatGPT, -please do cite it properly and include (if possible) your questions and answers as an addition to the report. This can -be uploaded to for example your website, GitHub/GitLab or similar as supplemental material. -

- -

If you would like to study other data sets, feel free to propose other -sets. What we have proposed here are mere suggestions from our -side. If you opt for another data set, consider using a set which has -been studied in the scientific literature. This makes it easier for -you to compare and analyze your results. Comparing with existing -results from the scientific literature is also an essential element of -the scientific discussion. The University of California at Irvine -with its Machine Learning repository at -https://archive.ics.uci.edu/ml/index.php is an excellent site to -look up for examples and -inspiration. Kaggle.com is an equally -interesting site. Feel free to explore these sites. When selecting -other data sets, make sure these are sets used for regression problems -(not classification). -

- -

- -

- -
- - - - - - - diff --git a/doc/Projects/2025/Project1/html/._Project1-bs001.html b/doc/Projects/2025/Project1/html/._Project1-bs001.html deleted file mode 100644 index 6f71c9733..000000000 --- a/doc/Projects/2025/Project1/html/._Project1-bs001.html +++ /dev/null @@ -1,503 +0,0 @@ - - - - - - - -Project 1 on Machine Learning, deadline October 6 (midnight), 2025 - - - - - - - - - - - - - - - - - - -
-
- - Project 1 on Machine Learning, deadline October 6 (midnight), 2025 -
-
- -
-
- -
-

 

 

 

- - -

Regression analysis and resampling methods

- -

The main aim of this project is to study in more detail various -regression methods, including Ordinary Least Squares (OLS) reegression, Ridge regression and LASSO regression. -In addition to the scientific part, in this course we want also to -give you an experience in writing scientific reports. -

- -

We will study how to fit polynomials to specific -one-dimensional functions (feel free to replace the suggested function with more complicated ones). -

- -

We will use Runge's function (see https://en.wikipedia.org/wiki/Runge%27s_phenomenon for a discussion). The one-dimensional function we will study is

-$$ -f(x) = \frac{1}{1+25x^2}. -$$ - -

Our first step will be to perform an OLS regression analysis of this -function, trying out a polynomial fit with an \( x \) dependence of the -form \( [x,x^2,\dots] \). You can use a uniform distribution to set up the -arrays of values for \( x \in [-1,1] \), or alternatively use a fixed step size. -Thereafter we will repeat many of the same steps when using the Ridge and Lasso regression methods, -introducing thereby a dependence on the hyperparameter (penalty) \( \lambda \). -

- -

We will also include bootstrap as a resampling technique in order to -study the so-called bias-variance tradeoff. After that we will -include the so-called cross-validation technique. -

-

Part a : Ordinary Least Square (OLS) for the Runge function

- -

We will generate our own dataset for abovementioned function -\( \mathrm{Runge}(x) \) function with \( x\in [-1,1] \). You should explore also the addition -of an added stochastic noise to this function using the normal -distribution \( N(0,1) \). -

- -

Write your own code (using for example the pseudoinverse function pinv from Numpy ) and perform a standard ordinary least square regression -analysis using polynomials in \( x \) up to order \( 15 \) or higher. Explore the dependence on the number of data points and the polynomial degree. -

- -

Evaluate the mean Squared error (MSE)

- -$$ MSE(\boldsymbol{y},\tilde{\boldsymbol{y}}) = \frac{1}{n} -\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2, -$$ - -

and the \( R^2 \) score function. If \( \tilde{\boldsymbol{y}}_i \) is the predicted -value of the \( i-th \) sample and \( y_i \) is the corresponding true value, -then the score \( R^2 \) is defined as -

- -$$ -R^2(\boldsymbol{y}, \tilde{\boldsymbol{y}}) = 1 - \frac{\sum_{i=0}^{n - 1} (y_i - \tilde{y}_i)^2}{\sum_{i=0}^{n - 1} (y_i - \bar{y})^2}, -$$ - -

where we have defined the mean value of \( \boldsymbol{y} \) as

- -$$ -\bar{y} = \frac{1}{n} \sum_{i=0}^{n - 1} y_i. -$$ - -

Plot the resulting scores (MSE and R$^2$) as functions of the polynomial degree (here up to polymial degree 15). -Plot also the parameters \( \theta \) as you increase the order of the polynomial. Comment your results. -

- -

Your code has to include a scaling/centering of the data (for example by -subtracting the mean value), and -a split of the data in training and test data. For the scaling you can -either write your own code or use for example the function for -splitting training data provided by the library Scikit-Learn (make -sure you have installed it). This function is called -\( train\_test\_split \). You should present a critical discussion of why and how you have scaled or not scaled the data. -

- -

It is normal in essentially all Machine Learning studies to split the -data in a training set and a test set (eventually also an additional -validation set). There -is no explicit recipe for how much data should be included as training -data and say test data. An accepted rule of thumb is to use -approximately \( 2/3 \) to \( 4/5 \) of the data as training data. -

- -

You can easily reuse the solutions to your exercises from week 35. -See also the lecture slides from week 35 and week 36. -

- -

On scaling, we recommend reading the following section from the scikit-learn software description, see https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html#plot-all-scaling-standard-scaler-section.

-

Part b: Adding Ridge regression for the Runge function

- -

Write your own code for the Ridge method as done in the previous -exercise. The lecture notes from week 35 and 36 contain more information. Furthermore, the results from the exercise set from week 36 is something you can reuse here. -

- -

Perform the same analysis as you did in the previous exercise but now for different values of \( \lambda \). Compare and -analyze your results with those obtained in part a) with the OLS method. Study the -dependence on \( \lambda \). -

-

Part c: Writing your own gradient descent code

- -

Replace now the analytical expressions for the optimal parameters -\( \boldsymbol{\theta} \) with your own gradient descent code. In this exercise we -focus only on the simplest gradient descent approach with a fixed -learning rate (see the exercises from week 37 and the lecture notes -from week 36). -

- -

Study and compare your results from parts a) and b) with your gradient -descent approch. Discuss in particular the role of the learning rate. -

-

Part d: Including momentum and more advanced ways to update the learning the rate

- -

We keep our focus on OLS and Ridge regression and update our code for -the gradient descent method by including momentum, ADAgrad, -RMSprop and ADAM as methods fro iteratively updating your learning -rate. Discuss the results and compare the different methods applied to -the one-dimensional Runge function. The lecture notes from week 37 contain several examples on how to implement these methods. -

-

Part e: Writing our own code for Lasso regression

- -

LASSO regression (see lecture slides from week 36 and week 37) -represents our first encounter with a machine learning method which -cannot be solved through analytical expressions (as in OLS and Ridge regression). Use the gradient -descent methods you developed in parts c) and d) to solve the LASSO -optimization problem. You can compare your results with -the functionalities of Scikit-Learn. -

- -

Discuss (critically) your results for the Runge function from OLS, -Ridge and LASSO regression using the various gradient descent -approaches. -

-

Part f: Stochastic gradient descent

- -

Our last gradient step is to include stochastic gradient descent using -the same methods to update the learning rates as in parts c-e). -Compare and discuss your results with and without stochastic gradient -and give a critical assessment of the various methods. -

-

Part g: Bias-variance trade-off and resampling techniques

- -

Our aim here is to study the bias-variance trade-off by implementing -the bootstrap resampling technique. We will only use the simpler -ordinary least squares here. -

- -

With a code which does OLS and includes resampling techniques, -we will now discuss the bias-variance trade-off in the context of -continuous predictions such as regression. However, many of the -intuitions and ideas discussed here also carry over to classification -tasks and basically all Machine Learning algorithms. -

- -

Before you perform an analysis of the bias-variance trade-off on your -test data, make first a figure similar to Fig. 2.11 of Hastie, -Tibshirani, and Friedman. Figure 2.11 of this reference displays only -the test and training MSEs. The test MSE can be used to indicate -possible regions of low/high bias and variance. You will most likely -not get an equally smooth curve! You may also need to increase the -polynomial order and play around with the number of data points as -well (see also the exercise set from week 35). -

- -

With this result we move on to the bias-variance trade-off analysis.

- -

Consider a -dataset \( \mathcal{L} \) consisting of the data -\( \mathbf{X}_\mathcal{L}=\{(y_j, \boldsymbol{x}_j), j=0\ldots n-1\} \). -

- -

We assume that the true data is generated from a noisy model

- -$$ -\boldsymbol{y}=f(\boldsymbol{x}) + \boldsymbol{\epsilon}. -$$ - -

Here \( \epsilon \) is normally distributed with mean zero and standard -deviation \( \sigma^2 \). -

- -

In our derivation of the ordinary least squares method we defined then -an approximation to the function \( f \) in terms of the parameters -\( \boldsymbol{\theta} \) and the design matrix \( \boldsymbol{X} \) which embody our model, -that is \( \boldsymbol{\tilde{y}}=\boldsymbol{X}\boldsymbol{\theta} \). -

- -

The parameters \( \boldsymbol{\theta} \) are in turn found by optimizing the mean -squared error via the so-called cost function -

- -$$ -C(\boldsymbol{X},\boldsymbol{\theta}) =\frac{1}{n}\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2=\mathbb{E}\left[(\boldsymbol{y}-\boldsymbol{\tilde{y}})^2\right]. -$$ - -

Here the expected value \( \mathbb{E} \) is the sample value.

- -

Show that you can rewrite this in terms of a term which contains the -variance of the model itself (the so-called variance term), a term -which measures the deviation from the true data and the mean value of -the model (the bias term) and finally the variance of the noise. -

- -

That is, show that

-$$ -\mathbb{E}\left[(\boldsymbol{y}-\boldsymbol{\tilde{y}})^2\right]=\mathrm{Bias}[\tilde{y}]+\mathrm{var}[\tilde{y}]+\sigma^2, -$$ - -

with (we approximate \( f(\boldsymbol{x})\approx \boldsymbol{y} \))

-$$ -\mathrm{Bias}[\tilde{y}]=\mathbb{E}\left[\left(\boldsymbol{y}-\mathbb{E}\left[\boldsymbol{\tilde{y}}\right]\right)^2\right], -$$ - -

and

-$$ -\mathrm{var}[\tilde{y}]=\mathbb{E}\left[\left(\tilde{\boldsymbol{y}}-\mathbb{E}\left[\boldsymbol{\tilde{y}}\right]\right)^2\right]=\frac{1}{n}\sum_i(\tilde{y}_i-\mathbb{E}\left[\boldsymbol{\tilde{y}}\right])^2. -$$ - -

Important note: Since the function \( f(x) \) is unknown, in order to be able to evalute the bias, we replace \( f(\boldsymbol{x}) \) in the expression for the bias with \( \boldsymbol{y} \).

- -

The answer to this exercise should be included in the theory part of -the report. This exercise is also part of the weekly exercises of -week 38. Explain what the terms mean and discuss their -interpretations. -

- -

Perform then a bias-variance analysis of the Runge function by -studying the MSE value as function of the complexity of your model. -

- -

Discuss the bias and variance trade-off as function -of your model complexity (the degree of the polynomial) and the number -of data points, and possibly also your training and test data using the bootstrap resampling method. -You can follow the code example in the jupyter-book at https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/chapter3.html#the-bias-variance-tradeoff. -

-

Part h): Cross-validation as resampling techniques, adding more complexity

- -

The aim here is to implement another widely popular -resampling technique, the so-called cross-validation method. -

- -

Implement the \( k \)-fold cross-validation algorithm (feel free to use -the functionality of Scikit-Learn or write your own code) and -evaluate again the MSE function resulting from the test folds. -

- -

Compare the MSE you get from your cross-validation code with the one -you got from your bootstrap code from the previous exercise. Comment and interpret your results. -

- -

In addition to using the ordinary least squares method, you should -include both Ridge and Lasso regression in the final analysis. -

-

Background literature

- -
    -
  1. For a discussion and derivation of the variances and mean squared errors using linear regression, see the Lecture notes on ridge regression by Wessel N. van Wieringen
    2. -
  2. The textbook of Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer, chapters 3 and 7 are the most relevant ones for the analysis here.
    3. -
-

Introduction to numerical projects

- -

Here follows a brief recipe and recommendation on how to answer the various questions when preparing your answers.

- - -

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, Julia 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. -

- -

- -

- -
- - - - - - - diff --git a/doc/Projects/2025/Project1/html/Project1-bs.html b/doc/Projects/2025/Project1/html/Project1-bs.html deleted file mode 100644 index bfb4c4ae3..000000000 --- a/doc/Projects/2025/Project1/html/Project1-bs.html +++ /dev/null @@ -1,238 +0,0 @@ - - - - - - - -Project 1 on Machine Learning, deadline October 6 (midnight), 2025 - - - - - - - - - - - - - - - - - - -
-
- - Project 1 on Machine Learning, deadline October 6 (midnight), 2025 -
-
- -
-
- -
-

 

 

 

- - -
-
-

Project 1 on Machine Learning, deadline October 6 (midnight), 2025

-
- - -
-Data Analysis and Machine Learning FYS-STK3155/FYS4155 -
- -
-University of Oslo, Norway -
-
-
-

September 2

-
-
- - -
-

Preamble: Note on writing reports, using reference material, AI and other tools

- -

We want you to answer the three different projects by handing in -reports written like a standard scientific/technical report. The -links at -https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects -contain more information. There you can find examples of previous -reports, the projects themselves, how we rade reports etc. How to -write reports will also be discussed during the various lab -sessions. Please do ask us if you are in doubt. -

- -

When using codes and material from other sources, you should refer to -these in the bibliography of your report, indicating wherefrom you for -example got the code, whether this is from the lecture notes, -softwares like Scikit-Learn, TensorFlow, PyTorch or other sources. These sources -should always be cited correctly. How to cite some -of the libraries is often indicated from their corresponding GitHub -sites or websites, see for example how to cite Scikit-Learn at -https://scikit-learn.org/dev/about.html. -

- -

We enocurage you to use tools like -ChatGPT or similar in writing the report. If you use for example ChatGPT, -please do cite it properly and include (if possible) your questions and answers as an addition to the report. This can -be uploaded to for example your website, GitHub/GitLab or similar as supplemental material. -

- -

If you would like to study other data sets, feel free to propose other -sets. What we have proposed here are mere suggestions from our -side. If you opt for another data set, consider using a set which has -been studied in the scientific literature. This makes it easier for -you to compare and analyze your results. Comparing with existing -results from the scientific literature is also an essential element of -the scientific discussion. The University of California at Irvine -with its Machine Learning repository at -https://archive.ics.uci.edu/ml/index.php is an excellent site to -look up for examples and -inspiration. Kaggle.com is an equally -interesting site. Feel free to explore these sites. When selecting -other data sets, make sure these are sets used for regression problems -(not classification). -

- -

- -

- -
- - - - - - - diff --git a/doc/Projects/2025/Project1/html/Project1.html b/doc/Projects/2025/Project1/html/Project1.html deleted file mode 100644 index 30ea4cd6e..000000000 --- a/doc/Projects/2025/Project1/html/Project1.html +++ /dev/null @@ -1,588 +0,0 @@ - - - - - - - -Project 1 on Machine Learning, deadline October 6 (midnight), 2025 - - - - - - - - - - - - - - -
-

Project 1 on Machine Learning, deadline October 6 (midnight), 2025

-
- - -
-Data Analysis and Machine Learning FYS-STK3155/FYS4155 -
- -
-University of Oslo, Norway -
-
-
-

September 2

-
-
-

Preamble: Note on writing reports, using reference material, AI and other tools

- -

We want you to answer the three different projects by handing in -reports written like a standard scientific/technical report. The -links at -https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects -contain more information. There you can find examples of previous -reports, the projects themselves, how we rade reports etc. How to -write reports will also be discussed during the various lab -sessions. Please do ask us if you are in doubt. -

- -

When using codes and material from other sources, you should refer to -these in the bibliography of your report, indicating wherefrom you for -example got the code, whether this is from the lecture notes, -softwares like Scikit-Learn, TensorFlow, PyTorch or other sources. These sources -should always be cited correctly. How to cite some -of the libraries is often indicated from their corresponding GitHub -sites or websites, see for example how to cite Scikit-Learn at -https://scikit-learn.org/dev/about.html. -

- -

We enocurage you to use tools like -ChatGPT or similar in writing the report. If you use for example ChatGPT, -please do cite it properly and include (if possible) your questions and answers as an addition to the report. This can -be uploaded to for example your website, GitHub/GitLab or similar as supplemental material. -

- -

If you would like to study other data sets, feel free to propose other -sets. What we have proposed here are mere suggestions from our -side. If you opt for another data set, consider using a set which has -been studied in the scientific literature. This makes it easier for -you to compare and analyze your results. Comparing with existing -results from the scientific literature is also an essential element of -the scientific discussion. The University of California at Irvine -with its Machine Learning repository at -https://archive.ics.uci.edu/ml/index.php is an excellent site to -look up for examples and -inspiration. Kaggle.com is an equally -interesting site. Feel free to explore these sites. When selecting -other data sets, make sure these are sets used for regression problems -(not classification). -

- -









-

Regression analysis and resampling methods

- -

The main aim of this project is to study in more detail various -regression methods, including Ordinary Least Squares (OLS) reegression, Ridge regression and LASSO regression. -In addition to the scientific part, in this course we want also to -give you an experience in writing scientific reports. -

- -

We will study how to fit polynomials to specific -one-dimensional functions (feel free to replace the suggested function with more complicated ones). -

- -

We will use Runge's function (see https://en.wikipedia.org/wiki/Runge%27s_phenomenon for a discussion). The one-dimensional function we will study is

-$$ -f(x) = \frac{1}{1+25x^2}. -$$ - -

Our first step will be to perform an OLS regression analysis of this -function, trying out a polynomial fit with an \( x \) dependence of the -form \( [x,x^2,\dots] \). You can use a uniform distribution to set up the -arrays of values for \( x \in [-1,1] \), or alternatively use a fixed step size. -Thereafter we will repeat many of the same steps when using the Ridge and Lasso regression methods, -introducing thereby a dependence on the hyperparameter (penalty) \( \lambda \). -

- -

We will also include bootstrap as a resampling technique in order to -study the so-called bias-variance tradeoff. After that we will -include the so-called cross-validation technique. -

-

Part a : Ordinary Least Square (OLS) for the Runge function

- -

We will generate our own dataset for abovementioned function -\( \mathrm{Runge}(x) \) function with \( x\in [-1,1] \). You should explore also the addition -of an added stochastic noise to this function using the normal -distribution \( N(0,1) \). -

- -

Write your own code (using for example the pseudoinverse function pinv from Numpy ) and perform a standard ordinary least square regression -analysis using polynomials in \( x \) up to order \( 15 \) or higher. Explore the dependence on the number of data points and the polynomial degree. -

- -

Evaluate the mean Squared error (MSE)

- -$$ MSE(\boldsymbol{y},\tilde{\boldsymbol{y}}) = \frac{1}{n} -\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2, -$$ - -

and the \( R^2 \) score function. If \( \tilde{\boldsymbol{y}}_i \) is the predicted -value of the \( i-th \) sample and \( y_i \) is the corresponding true value, -then the score \( R^2 \) is defined as -

- -$$ -R^2(\boldsymbol{y}, \tilde{\boldsymbol{y}}) = 1 - \frac{\sum_{i=0}^{n - 1} (y_i - \tilde{y}_i)^2}{\sum_{i=0}^{n - 1} (y_i - \bar{y})^2}, -$$ - -

where we have defined the mean value of \( \boldsymbol{y} \) as

- -$$ -\bar{y} = \frac{1}{n} \sum_{i=0}^{n - 1} y_i. -$$ - -

Plot the resulting scores (MSE and R$^2$) as functions of the polynomial degree (here up to polymial degree 15). -Plot also the parameters \( \theta \) as you increase the order of the polynomial. Comment your results. -

- -

Your code has to include a scaling/centering of the data (for example by -subtracting the mean value), and -a split of the data in training and test data. For the scaling you can -either write your own code or use for example the function for -splitting training data provided by the library Scikit-Learn (make -sure you have installed it). This function is called -\( train\_test\_split \). You should present a critical discussion of why and how you have scaled or not scaled the data. -

- -

It is normal in essentially all Machine Learning studies to split the -data in a training set and a test set (eventually also an additional -validation set). There -is no explicit recipe for how much data should be included as training -data and say test data. An accepted rule of thumb is to use -approximately \( 2/3 \) to \( 4/5 \) of the data as training data. -

- -

You can easily reuse the solutions to your exercises from week 35. -See also the lecture slides from week 35 and week 36. -

- -

On scaling, we recommend reading the following section from the scikit-learn software description, see https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html#plot-all-scaling-standard-scaler-section.

-

Part b: Adding Ridge regression for the Runge function

- -

Write your own code for the Ridge method as done in the previous -exercise. The lecture notes from week 35 and 36 contain more information. Furthermore, the results from the exercise set from week 36 is something you can reuse here. -

- -

Perform the same analysis as you did in the previous exercise but now for different values of \( \lambda \). Compare and -analyze your results with those obtained in part a) with the OLS method. Study the -dependence on \( \lambda \). -

-

Part c: Writing your own gradient descent code

- -

Replace now the analytical expressions for the optimal parameters -\( \boldsymbol{\theta} \) with your own gradient descent code. In this exercise we -focus only on the simplest gradient descent approach with a fixed -learning rate (see the exercises from week 37 and the lecture notes -from week 36). -

- -

Study and compare your results from parts a) and b) with your gradient -descent approch. Discuss in particular the role of the learning rate. -

-

Part d: Including momentum and more advanced ways to update the learning the rate

- -

We keep our focus on OLS and Ridge regression and update our code for -the gradient descent method by including momentum, ADAgrad, -RMSprop and ADAM as methods fro iteratively updating your learning -rate. Discuss the results and compare the different methods applied to -the one-dimensional Runge function. The lecture notes from week 37 contain several examples on how to implement these methods. -

-

Part e: Writing our own code for Lasso regression

- -

LASSO regression (see lecture slides from week 36 and week 37) -represents our first encounter with a machine learning method which -cannot be solved through analytical expressions (as in OLS and Ridge regression). Use the gradient -descent methods you developed in parts c) and d) to solve the LASSO -optimization problem. You can compare your results with -the functionalities of Scikit-Learn. -

- -

Discuss (critically) your results for the Runge function from OLS, -Ridge and LASSO regression using the various gradient descent -approaches. -

-

Part f: Stochastic gradient descent

- -

Our last gradient step is to include stochastic gradient descent using -the same methods to update the learning rates as in parts c-e). -Compare and discuss your results with and without stochastic gradient -and give a critical assessment of the various methods. -

-

Part g: Bias-variance trade-off and resampling techniques

- -

Our aim here is to study the bias-variance trade-off by implementing -the bootstrap resampling technique. We will only use the simpler -ordinary least squares here. -

- -

With a code which does OLS and includes resampling techniques, -we will now discuss the bias-variance trade-off in the context of -continuous predictions such as regression. However, many of the -intuitions and ideas discussed here also carry over to classification -tasks and basically all Machine Learning algorithms. -

- -

Before you perform an analysis of the bias-variance trade-off on your -test data, make first a figure similar to Fig. 2.11 of Hastie, -Tibshirani, and Friedman. Figure 2.11 of this reference displays only -the test and training MSEs. The test MSE can be used to indicate -possible regions of low/high bias and variance. You will most likely -not get an equally smooth curve! You may also need to increase the -polynomial order and play around with the number of data points as -well (see also the exercise set from week 35). -

- -

With this result we move on to the bias-variance trade-off analysis.

- -

Consider a -dataset \( \mathcal{L} \) consisting of the data -\( \mathbf{X}_\mathcal{L}=\{(y_j, \boldsymbol{x}_j), j=0\ldots n-1\} \). -

- -

We assume that the true data is generated from a noisy model

- -$$ -\boldsymbol{y}=f(\boldsymbol{x}) + \boldsymbol{\epsilon}. -$$ - -

Here \( \epsilon \) is normally distributed with mean zero and standard -deviation \( \sigma^2 \). -

- -

In our derivation of the ordinary least squares method we defined then -an approximation to the function \( f \) in terms of the parameters -\( \boldsymbol{\theta} \) and the design matrix \( \boldsymbol{X} \) which embody our model, -that is \( \boldsymbol{\tilde{y}}=\boldsymbol{X}\boldsymbol{\theta} \). -

- -

The parameters \( \boldsymbol{\theta} \) are in turn found by optimizing the mean -squared error via the so-called cost function -

- -$$ -C(\boldsymbol{X},\boldsymbol{\theta}) =\frac{1}{n}\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2=\mathbb{E}\left[(\boldsymbol{y}-\boldsymbol{\tilde{y}})^2\right]. -$$ - -

Here the expected value \( \mathbb{E} \) is the sample value.

- -

Show that you can rewrite this in terms of a term which contains the -variance of the model itself (the so-called variance term), a term -which measures the deviation from the true data and the mean value of -the model (the bias term) and finally the variance of the noise. -

- -

That is, show that

-$$ -\mathbb{E}\left[(\boldsymbol{y}-\boldsymbol{\tilde{y}})^2\right]=\mathrm{Bias}[\tilde{y}]+\mathrm{var}[\tilde{y}]+\sigma^2, -$$ - -

with (we approximate \( f(\boldsymbol{x})\approx \boldsymbol{y} \))

-$$ -\mathrm{Bias}[\tilde{y}]=\mathbb{E}\left[\left(\boldsymbol{y}-\mathbb{E}\left[\boldsymbol{\tilde{y}}\right]\right)^2\right], -$$ - -

and

-$$ -\mathrm{var}[\tilde{y}]=\mathbb{E}\left[\left(\tilde{\boldsymbol{y}}-\mathbb{E}\left[\boldsymbol{\tilde{y}}\right]\right)^2\right]=\frac{1}{n}\sum_i(\tilde{y}_i-\mathbb{E}\left[\boldsymbol{\tilde{y}}\right])^2. -$$ - -

Important note: Since the function \( f(x) \) is unknown, in order to be able to evalute the bias, we replace \( f(\boldsymbol{x}) \) in the expression for the bias with \( \boldsymbol{y} \).

- -

The answer to this exercise should be included in the theory part of -the report. This exercise is also part of the weekly exercises of -week 38. Explain what the terms mean and discuss their -interpretations. -

- -

Perform then a bias-variance analysis of the Runge function by -studying the MSE value as function of the complexity of your model. -

- -

Discuss the bias and variance trade-off as function -of your model complexity (the degree of the polynomial) and the number -of data points, and possibly also your training and test data using the bootstrap resampling method. -You can follow the code example in the jupyter-book at https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/chapter3.html#the-bias-variance-tradeoff. -

-

Part h): Cross-validation as resampling techniques, adding more complexity

- -

The aim here is to implement another widely popular -resampling technique, the so-called cross-validation method. -

- -

Implement the \( k \)-fold cross-validation algorithm (feel free to use -the functionality of Scikit-Learn or write your own code) and -evaluate again the MSE function resulting from the test folds. -

- -

Compare the MSE you get from your cross-validation code with the one -you got from your bootstrap code from the previous exercise. Comment and interpret your results. -

- -

In addition to using the ordinary least squares method, you should -include both Ridge and Lasso regression in the final analysis. -

-

Background literature

- -
    -
  1. For a discussion and derivation of the variances and mean squared errors using linear regression, see the Lecture notes on ridge regression by Wessel N. van Wieringen
    2. -
  2. The textbook of Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer, chapters 3 and 7 are the most relevant ones for the analysis here.
    3. -
-

Introduction to numerical projects

- -

Here follows a brief recipe and recommendation on how to answer the various questions when preparing your answers.

- - -

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, Julia 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. -

- - - - - diff --git a/doc/Projects/2025/Project1/pdf/Project1.p.tex b/doc/Projects/2025/Project1/pdf/Project1.p.tex deleted file mode 100644 index 9f28fd21a..000000000 --- a/doc/Projects/2025/Project1/pdf/Project1.p.tex +++ /dev/null @@ -1,504 +0,0 @@ -%% -%% Automatically generated file from DocOnce source -%% (https://github.com/doconce/doconce/) -%% doconce format latex Project1.do.txt --print_latex_style=trac --latex_admon=paragraph -%% -% #ifdef PTEX2TEX_EXPLANATION -%% -%% The file follows the ptex2tex extended LaTeX format, see -%% ptex2tex: https://code.google.com/p/ptex2tex/ -%% -%% Run -%% ptex2tex myfile -%% or -%% doconce ptex2tex myfile -%% -%% to turn myfile.p.tex into an ordinary LaTeX file myfile.tex. -%% (The ptex2tex program: https://code.google.com/p/ptex2tex) -%% Many preprocess options can be added to ptex2tex or doconce ptex2tex -%% -%% ptex2tex -DMINTED myfile -%% doconce ptex2tex myfile envir=minted -%% -%% ptex2tex will typeset code environments according to a global or local -%% .ptex2tex.cfg configure file. doconce ptex2tex will typeset code -%% according to options on the command line (just type doconce ptex2tex to -%% see examples). If doconce ptex2tex has envir=minted, it enables the -%% minted style without needing -DMINTED. -% #endif - -% #define PREAMBLE - -% #ifdef PREAMBLE -%-------------------- begin preamble ---------------------- - -\documentclass[% -oneside, % oneside: electronic viewing, twoside: printing -final, % draft: marks overfull hboxes, figures with paths -10pt]{article} - -\listfiles % print all files needed to compile this document - -\usepackage{relsize,makeidx,color,setspace,amsmath,amsfonts,amssymb} -\usepackage[table]{xcolor} -\usepackage{bm,ltablex,microtype} - -\usepackage[pdftex]{graphicx} - -\usepackage[T1]{fontenc} -%\usepackage[latin1]{inputenc} -\usepackage{ucs} -\usepackage[utf8x]{inputenc} - -\usepackage{lmodern} % Latin Modern fonts derived from Computer Modern - -% Hyperlinks in PDF: -\definecolor{linkcolor}{rgb}{0,0,0.4} -\usepackage{hyperref} -\hypersetup{ - breaklinks=true, - colorlinks=true, - linkcolor=linkcolor, - urlcolor=linkcolor, - citecolor=black, - filecolor=black, - %filecolor=blue, - pdfmenubar=true, - pdftoolbar=true, - bookmarksdepth=3 % Uncomment (and tweak) for PDF bookmarks with more levels than the TOC - } -%\hyperbaseurl{} % hyperlinks are relative to this root - -\setcounter{tocdepth}{2} % levels in table of contents - -% prevent orhpans and widows -\clubpenalty = 10000 -\widowpenalty = 10000 - -% --- end of standard preamble for documents --- - - -% insert custom LaTeX commands... - -\raggedbottom -\makeindex -\usepackage[totoc]{idxlayout} % for index in the toc -\usepackage[nottoc]{tocbibind} % for references/bibliography in the toc - -%-------------------- end preamble ---------------------- - -\begin{document} - -% matching end for #ifdef PREAMBLE -% #endif - -\newcommand{\exercisesection}[1]{\subsection*{#1}} - - -% ------------------- main content ---------------------- - - - -% ----------------- title ------------------------- - -\thispagestyle{empty} - -\begin{center} -{\LARGE\bf -\begin{spacing}{1.25} -Project 1 on Machine Learning, deadline October 6 (midnight), 2025 -\end{spacing} -} -\end{center} - -% ----------------- author(s) ------------------------- - -\begin{center} -{\bf Data Analysis and Machine Learning FYS-STK3155/FYS4155} -\end{center} - - \begin{center} -% List of all institutions: -\centerline{{\small University of Oslo, Norway}} -\end{center} - -% ----------------- end author(s) ------------------------- - -% --- begin date --- -\begin{center} -September 2 -\end{center} -% --- end date --- - -\vspace{1cm} - - -\subsection{Preamble: Note on writing reports, using reference material, AI and other tools} - -We want you to answer the three different projects by handing in -reports written like a standard scientific/technical report. The -links at -\href{{https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects}}{\nolinkurl{https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects}} -contain more information. There you can find examples of previous -reports, the projects themselves, how we rade reports etc. How to -write reports will also be discussed during the various lab -sessions. Please do ask us if you are in doubt. - -When using codes and material from other sources, you should refer to -these in the bibliography of your report, indicating wherefrom you for -example got the code, whether this is from the lecture notes, -softwares like Scikit-Learn, TensorFlow, PyTorch or other sources. These sources -should always be cited correctly. How to cite some -of the libraries is often indicated from their corresponding GitHub -sites or websites, see for example how to cite Scikit-Learn at -\href{{https://scikit-learn.org/dev/about.html}}{\nolinkurl{https://scikit-learn.org/dev/about.html}}. - -We enocurage you to use tools like -\href{{https://openai.com/chatgpt/}}{ChatGPT} or similar in writing the report. If you use for example ChatGPT, -please do cite it properly and include (if possible) your questions and answers as an addition to the report. This can -be uploaded to for example your website, GitHub/GitLab or similar as supplemental material. - -If you would like to study other data sets, feel free to propose other -sets. What we have proposed here are mere suggestions from our -side. If you opt for another data set, consider using a set which has -been studied in the scientific literature. This makes it easier for -you to compare and analyze your results. Comparing with existing -results from the scientific literature is also an essential element of -the scientific discussion. The University of California at Irvine -with its Machine Learning repository at -\href{{https://archive.ics.uci.edu/ml/index.php}}{\nolinkurl{https://archive.ics.uci.edu/ml/index.php}} is an excellent site to -look up for examples and -inspiration. \href{{https://www.kaggle.com/}}{Kaggle.com} is an equally -interesting site. Feel free to explore these sites. When selecting -other data sets, make sure these are sets used for regression problems -(not classification). - -% !split -\subsection{Regression analysis and resampling methods} - -The main aim of this project is to study in more detail various -regression methods, including Ordinary Least Squares (OLS) reegression, Ridge regression and LASSO regression. -In addition to the scientific part, in this course we want also to -give you an experience in writing scientific reports. - -We will study how to fit polynomials to specific -one-dimensional functions (feel free to replace the suggested function with more complicated ones). - -We will use Runge's function (see \href{{https://en.wikipedia.org/wiki/Runge%27s_phenomenon}}{\nolinkurl{https://en.wikipedia.org/wiki/Runge\%27s_phenomenon}} for a discussion). The one-dimensional function we will study is -\[ -f(x) = \frac{1}{1+25x^2}. -\] - -Our first step will be to perform an OLS regression analysis of this -function, trying out a polynomial fit with an $x$ dependence of the -form $[x,x^2,\dots]$. You can use a uniform distribution to set up the -arrays of values for $x \in [-1,1]$, or alternatively use a fixed step size. -Thereafter we will repeat many of the same steps when using the Ridge and Lasso regression methods, -introducing thereby a dependence on the hyperparameter (penalty) $\lambda$. - -We will also include bootstrap as a resampling technique in order to -study the so-called \textbf{bias-variance tradeoff}. After that we will -include the so-called cross-validation technique. - -\paragraph{Part a : Ordinary Least Square (OLS) for the Runge function.} -We will generate our own dataset for abovementioned function -$\mathrm{Runge}(x)$ function with $x\in [-1,1]$. You should explore also the addition -of an added stochastic noise to this function using the normal -distribution $N(0,1)$. - -\emph{Write your own code} (using for example the pseudoinverse function \textbf{pinv} from \textbf{Numpy} ) and perform a standard \textbf{ordinary least square regression} -analysis using polynomials in $x$ up to order $15$ or higher. Explore the dependence on the number of data points and the polynomial degree. - -Evaluate the mean Squared error (MSE) - -\[ MSE(\bm{y},\tilde{\bm{y}}) = \frac{1}{n} -\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2, -\] - -and the $R^2$ score function. If $\tilde{\bm{y}}_i$ is the predicted -value of the $i-th$ sample and $y_i$ is the corresponding true value, -then the score $R^2$ is defined as - -\[ -R^2(\bm{y}, \tilde{\bm{y}}) = 1 - \frac{\sum_{i=0}^{n - 1} (y_i - \tilde{y}_i)^2}{\sum_{i=0}^{n - 1} (y_i - \bar{y})^2}, -\] - -where we have defined the mean value of $\bm{y}$ as - -\[ -\bar{y} = \frac{1}{n} \sum_{i=0}^{n - 1} y_i. -\] - -Plot the resulting scores (MSE and R$^2$) as functions of the polynomial degree (here up to polymial degree 15). -Plot also the parameters $\theta$ as you increase the order of the polynomial. Comment your results. - -Your code has to include a scaling/centering of the data (for example by -subtracting the mean value), and -a split of the data in training and test data. For the scaling you can -either write your own code or use for example the function for -splitting training data provided by the library \textbf{Scikit-Learn} (make -sure you have installed it). This function is called -$train\_test\_split$. \textbf{You should present a critical discussion of why and how you have scaled or not scaled the data}. - -It is normal in essentially all Machine Learning studies to split the -data in a training set and a test set (eventually also an additional -validation set). There -is no explicit recipe for how much data should be included as training -data and say test data. An accepted rule of thumb is to use -approximately $2/3$ to $4/5$ of the data as training data. - -You can easily reuse the solutions to your exercises from week 35. -See also the lecture slides from week 35 and week 36. - -On scaling, we recommend reading the following section from the scikit-learn software description, see \href{{https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html#plot-all-scaling-standard-scaler-section}}{\nolinkurl{https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html\#plot-all-scaling-standard-scaler-section}}. - -\paragraph{Part b: Adding Ridge regression for the Runge function.} -Write your own code for the Ridge method as done in the previous -exercise. The lecture notes from week 35 and 36 contain more information. Furthermore, the results from the exercise set from week 36 is something you can reuse here. - -Perform the same analysis as you did in the previous exercise but now for different values of $\lambda$. Compare and -analyze your results with those obtained in part a) with the OLS method. Study the -dependence on $\lambda$. - -\paragraph{Part c: Writing your own gradient descent code.} -Replace now the analytical expressions for the optimal parameters -$\bm{\theta}$ with your own gradient descent code. In this exercise we -focus only on the simplest gradient descent approach with a fixed -learning rate (see the exercises from week 37 and the lecture notes -from week 36). - -Study and compare your results from parts a) and b) with your gradient -descent approch. Discuss in particular the role of the learning rate. - -\paragraph{Part d: Including momentum and more advanced ways to update the learning the rate.} -We keep our focus on OLS and Ridge regression and update our code for -the gradient descent method by including \textbf{momentum}, \textbf{ADAgrad}, -\textbf{RMSprop} and \textbf{ADAM} as methods fro iteratively updating your learning -rate. Discuss the results and compare the different methods applied to -the one-dimensional Runge function. The lecture notes from week 37 contain several examples on how to implement these methods. - -\paragraph{Part e: Writing our own code for Lasso regression.} -LASSO regression (see lecture slides from week 36 and week 37) -represents our first encounter with a machine learning method which -cannot be solved through analytical expressions (as in OLS and Ridge regression). Use the gradient -descent methods you developed in parts c) and d) to solve the LASSO -optimization problem. You can compare your results with -the functionalities of \textbf{Scikit-Learn}. - -Discuss (critically) your results for the Runge function from OLS, -Ridge and LASSO regression using the various gradient descent -approaches. - -\paragraph{Part f: Stochastic gradient descent.} -Our last gradient step is to include stochastic gradient descent using -the same methods to update the learning rates as in parts c-e). -Compare and discuss your results with and without stochastic gradient -and give a critical assessment of the various methods. - -\paragraph{Part g: Bias-variance trade-off and resampling techniques.} -Our aim here is to study the bias-variance trade-off by implementing -the \textbf{bootstrap} resampling technique. \textbf{We will only use the simpler -ordinary least squares here}. - -With a code which does OLS and includes resampling techniques, -we will now discuss the bias-variance trade-off in the context of -continuous predictions such as regression. However, many of the -intuitions and ideas discussed here also carry over to classification -tasks and basically all Machine Learning algorithms. - -Before you perform an analysis of the bias-variance trade-off on your -test data, make first a figure similar to Fig.~2.11 of Hastie, -Tibshirani, and Friedman. Figure 2.11 of this reference displays only -the test and training MSEs. The test MSE can be used to indicate -possible regions of low/high bias and variance. You will most likely -not get an equally smooth curve! You may also need to increase the -polynomial order and play around with the number of data points as -well (see also the exercise set from week 35). - -With this result we move on to the bias-variance trade-off analysis. - -Consider a -dataset $\mathcal{L}$ consisting of the data -$\mathbf{X}_\mathcal{L}=\{(y_j, \boldsymbol{x}_j), j=0\ldots n-1\}$. - -We assume that the true data is generated from a noisy model - -\[ -\bm{y}=f(\boldsymbol{x}) + \bm{\epsilon}. -\] - -Here $\epsilon$ is normally distributed with mean zero and standard -deviation $\sigma^2$. - -In our derivation of the ordinary least squares method we defined then -an approximation to the function $f$ in terms of the parameters -$\bm{\theta}$ and the design matrix $\bm{X}$ which embody our model, -that is $\bm{\tilde{y}}=\bm{X}\bm{\theta}$. - -The parameters $\bm{\theta}$ are in turn found by optimizing the mean -squared error via the so-called cost function - -\[ -C(\bm{X},\bm{\theta}) =\frac{1}{n}\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2=\mathbb{E}\left[(\bm{y}-\bm{\tilde{y}})^2\right]. -\] -Here the expected value $\mathbb{E}$ is the sample value. - -Show that you can rewrite this in terms of a term which contains the -variance of the model itself (the so-called variance term), a term -which measures the deviation from the true data and the mean value of -the model (the bias term) and finally the variance of the noise. - -That is, show that -\[ -\mathbb{E}\left[(\bm{y}-\bm{\tilde{y}})^2\right]=\mathrm{Bias}[\tilde{y}]+\mathrm{var}[\tilde{y}]+\sigma^2, -\] -with (we approximate $f(\bm{x})\approx \bm{y}$) -\[ -\mathrm{Bias}[\tilde{y}]=\mathbb{E}\left[\left(\bm{y}-\mathbb{E}\left[\bm{\tilde{y}}\right]\right)^2\right], -\] -and -\[ -\mathrm{var}[\tilde{y}]=\mathbb{E}\left[\left(\tilde{\bm{y}}-\mathbb{E}\left[\bm{\tilde{y}}\right]\right)^2\right]=\frac{1}{n}\sum_i(\tilde{y}_i-\mathbb{E}\left[\bm{\tilde{y}}\right])^2. -\] - -\textbf{Important note}: Since the function $f(x)$ is unknown, in order to be able to evalute the bias, we replace $f(\bm{x})$ in the expression for the bias with $\bm{y}$. - -The answer to this exercise should be included in the theory part of -the report. This exercise is also part of the weekly exercises of -week 38. Explain what the terms mean and discuss their -interpretations. - -Perform then a bias-variance analysis of the Runge function by -studying the MSE value as function of the complexity of your model. - -Discuss the bias and variance trade-off as function -of your model complexity (the degree of the polynomial) and the number -of data points, and possibly also your training and test data using the \textbf{bootstrap} resampling method. -You can follow the code example in the jupyter-book at \href{{https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/chapter3.html#the-bias-variance-tradeoff}}{\nolinkurl{https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/chapter3.html\#the-bias-variance-tradeoff}}. - -\paragraph{Part h): Cross-validation as resampling techniques, adding more complexity.} -The aim here is to implement another widely popular -resampling technique, the so-called cross-validation method. - -Implement the $k$-fold cross-validation algorithm (feel free to use -the functionality of \textbf{Scikit-Learn} or write your own code) and -evaluate again the MSE function resulting from the test folds. - -Compare the MSE you get from your cross-validation code with the one -you got from your \textbf{bootstrap} code from the previous exercise. Comment and interpret your results. - -In addition to using the ordinary least squares method, you should -include both Ridge and Lasso regression in the final analysis. - -\subsection{Background literature} - -\begin{enumerate} -\item For a discussion and derivation of the variances and mean squared errors using linear regression, see the \href{{https://arxiv.org/abs/1509.09169}}{Lecture notes on ridge regression by Wessel N. van Wieringen} - -\item The textbook of \href{{https://www.springer.com/gp/book/9780387848570}}{Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer}, chapters 3 and 7 are the most relevant ones for the analysis here. -\end{enumerate} - -\noindent -\subsection{Introduction to numerical projects} - -Here follows a brief recipe and recommendation on how to answer the various questions when preparing your answers. - -\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. You should have the code at your GitHub/GitLab link. You can also place the code in an appendix of your report. - - \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, Julia or Python. The following prescription should be followed when preparing the report: - -\begin{itemize} - \item Use Canvas to hand in your projects, log in at \href{{https://www.uio.no/english/services/it/education/canvas/}}{\nolinkurl{https://www.uio.no/english/services/it/education/canvas/}} with your normal UiO username and password. - - \item Upload \textbf{only} the report file or the link to your GitHub/GitLab or similar typo of repos! For the source code file(s) you have developed please provide us with your link to your GitHub/GitLab or similar domain. The report file should include all of your discussions and a list of the codes you have developed. Do not include library files which are available at the course homepage, unless you have made specific changes to them. - - \item In your GitHub/GitLab or similar repository, please include a folder which contains selected results. These can be in the form of output from your code for a selected set of runs and input parameters. -\end{itemize} - -\noindent -Finally, -we encourage you to collaborate. Optimal working groups consist of -2-3 students. You can then hand in a common report. - -\subsection{Software and needed installations} - -If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, -we recommend that you install the following Python packages via \textbf{pip} as -\begin{enumerate} -\item pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow -\end{enumerate} - -\noindent -For Python3, replace \textbf{pip} with \textbf{pip3}. - -See below for a discussion of \textbf{tensorflow} and \textbf{scikit-learn}. - -For OSX users we recommend also, after having installed Xcode, to install \textbf{brew}. Brew allows -for a seamless installation of additional software via for example -\begin{enumerate} -\item brew install python3 -\end{enumerate} - -\noindent -For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution -you can use \textbf{pip} as well and simply install Python as -\begin{enumerate} -\item sudo apt-get install python3 (or python for python2.7) -\end{enumerate} - -\noindent -etc etc. - -If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely -\begin{enumerate} -\item \href{{https://docs.anaconda.com/}}{Anaconda} Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system \textbf{conda} - -\item \href{{https://www.enthought.com/product/canopy/}}{Enthought canopy} is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license. -\end{enumerate} - -\noindent -Popular software packages written in Python for ML are - -\begin{itemize} -\item \href{{http://scikit-learn.org/stable/}}{Scikit-learn}, - -\item \href{{https://www.tensorflow.org/}}{Tensorflow}, - -\item \href{{http://pytorch.org/}}{PyTorch} and - -\item \href{{https://keras.io/}}{Keras}. -\end{itemize} - -\noindent -These are all freely available at their respective GitHub sites. They -encompass communities of developers in the thousands or more. And the number -of code developers and contributors keeps increasing. - - -% ------------------- end of main content --------------- - -% #ifdef PREAMBLE -\end{document} -% #endif - diff --git a/doc/Projects/2025/Project1/pdf/Project1.pdf b/doc/Projects/2025/Project1/pdf/Project1.pdf deleted file mode 100644 index b7d7b5e2500c365e6363c8b555c4354554d4dc15..0000000000000000000000000000000000000000 GIT binary patch literal 0 HcmV?d00001 literal 287564 zcma(2LzFN~kOt_MZQHhO+qP}nwr$&X)mOG{+vfe}o|#qmqH~oeH@S=_--skt5D}wg zq+^F7onIbVhhk(VU?8wJvV!8_fuff&wKI3IAYfwR1~qO@;*7Tqh~-O zM4OkFU9}e!o=aMp-sT9K1&!z%5)|MC6XC zLxw4yt+!9E^C>pU;3GHj{!Ke~OC^v8>l)Jp5y`l%S9j6A$YzEasJR-dp(=oO3Z9fy zhv@chjyNL3wg^p)L+6Ty;L^K>v_bi*-7-Z7w%oK>{MM%23pp2Cv=hN%)uq{N{xxCw z!v$~}JVhu?zv#zvqpaDS*)0}0upvnYB=_VJLq(LDCqh}#u28xWGP1^E;pr}!aO6e$ z^k#pxV4`50r2Q?qR(5lo1`Y*XVv|*Z2dn@FOM|z*O1MFOuBQ%JDy79tK)Oyu-npP(Kx?e;>rFiIDT;S(`+NtLgR@aAaT0%& zGD$Y-f_}}V*MD?VNfUP^fH2t+hSeMr-{dDuP6M9^WDA4doF@F_oFAIhhKq})-8EE_ zO@Mk+aEI1f`8Am9be3wMYWg32yKr!XVwbUYjb8<|w^nlpLR=cfTK@N2?Lubw`yTu` zl6DmNVupz-$G0pzCZ!Ra3NT6v&$ZgkpLU-iF=_ZdA>U24ivYhgRr`>53b-&l5GP<# z{7^u*8medSCA*`6sQgJPztbC5@NB2;6taceK?kYMT?H(K@mGTrXW|c(`5ei0NXj7@ za&Qf>gi)#TRBR6g+=5z9iXccICXT?4N^K7swrMm#mXhz>QYsHjs3I4J9bNXXU&y9_ zq_dtpVXpog{TK9W?&eY>n>YSLVO5E~%DXq=|%_Kwh<6fCd_Idczy!uAOb0 zkYOZaUag-C&&B0*y0~?o*I#&VJPQ{rsS)L<2L+%$nAAO%n`#6rT)zR<0GDwI*hiNO zoIB%s%vuw3E=rjJW43wVlS=Denvttj#Q<{L=AW~KLQO;-YgQ*M=BK18x5WTUW2b_v z6E@UUR8ulWv%rRIc;ODU?>|mrOeKy?=;%OKd95}Q7HnQQM7>7^h#66^ZVL9t-@!NZ zK$5wJ^Do76>@YZJc z?wj;L!{B*^QGrfHJ`1J7@+5t)mnsv^xWJF|Tc#XCAVK+`>5jiP0+Jy(Hob;a=SDu$ zlU!T&nUvFgcw&#oQD(TUaBViA-24vEDRkx=hNevwwgbCsr^T3VI)LBqZ#skAYg^rF zN=L0aKewaWy{TDXU;`lw$NNrh4cFa>X~VJR8z*QXv3YEIQW!-kXCwfsG#*}}gOox| zB!$P>SR5~&(d2RCNK87gAJ==f@ZNRmx>-+=K2#Heuia8&A`#5Vq)#-}8 zqY6f14ejN>#ENftZdV%d!|!=?Hve|Lf?kqN#gk)#blj67Uo2K8mQIyGLT(faKQGLI zUbV?N&Jd}7@tHwg%IkM?R_=GYzt9iAXkTF;S9N5nd4jmYRmlV%2@V6U0#h<$pfLug z3X8Z*2DMGnulwh%SHV@SZT&Kdh-W4xH`ju6%jrZI+zGWvWdIBp_m$hwHCt@$g&vp4 zaxX63Y`(!gJRRnS#EE!Wd9i{eoqB{!1az0JxZb`hb^=DE&@}j z#FsD{5UJYHYjli@O*Rnc11jZrr7h!(_`^k_rGQO(+Ip=z_UJLBj0S7Ay?Ingu?QxmvML zY5#HwlH;c=i)hqofG(RwLD=&G{UL`-Q80BR-~s-6r~ip~&B_q=AZG(IeRI*B z)MHKiJY|mPv$o9qW?OY&dUCL3f>!Z zLDrvLt?89#y6c_lrV)x+?L3CinxR)^dvEV|>upJUeo&lyq^c(6eoWsf?Ukxc8*;g` zqDWnj^Ye)eE8Yil=x|Tf3%@$%d{qHrPjwN|fO z;Ia1_m!lIRp0(5RAupcm?<5h#`6vAiS!?u~)&y_oO~kfUM+ww7h7F4DHVJhwQHTgoF; zm8E!AqYuEG9>eU>BUN|RM%?$eJp*44MEiI7_4IW95k~S-iCD!wSK9PydN?IS^NZcH zz1aFttyiylYdrAfcQ!6sWPcHwbuV*Of?}cYy1p+zO{J;A~i*{l_K~yXpo?fAG?(Z=>$C zF3Tq@qvb&{gbc@7Q=N*3lM{WP(uz{YZ`L6cR9-;&{o1WEWH-W{}l-6s@-u0sMLcUa| z{=Unkl?5=35Z^-Y+!;F1t92#i;WNt zThR_1@~C6RBvHGobG7@9pOxej42v@U^R`A6gHD#9M=M|%a;Tn8Uqip6C)v#*?`2R3 zcqlYS(tz|aZSL>zMEBry3DShb72EB_?RsO`b}T9Ec-qzZPUSGB8|u)&iO{$-Txtbs z{w4NJ@qCuh3huDPnRgN^DSQ+nOiXg1J;{P8RynjpgyZtPEkd7%PS8dO`V3wQs{aDx zrOg5WRN0&@f=mG`i5Re5H38Ki6p#0LeFB5Am#2-edItq57YgohBSOGpMj%^Fn7+>1 zDl~2AGw(3Hbe}7uYxgStc`ve6}#2@Jt76 zJ9bv*{Kv{tq|vTI@){VV8okul_+~bUqz8a9z^RK?2t*kpIY6S6%x8pzg(EX4n;-U2 zH1~kKK#m7{vi(tTmHYJ0XArSK9y4leLz5Ud`kAK9^!cnR>1ODv)}|X=@##G`5ENB) zv4!Tt`g2es*dt=lAW`ugW13CYjA5rqP6!-C4@%eFx9=t*eD-2QZp{ui+kGfle5N(1 z$v&%Tpd+{+Msm@2ggzCO4xJ@#E|IWz;jmFhokew(vE?RiB}QirjVFZP@*OWKhYG#9 z>eHz7+ucx7&A4e3b&a-40)R`fmzEil%s5VwK*eRXbu{i5tYo7bXKp`c;RFRrNF_1? z%DPd2dTJZhW8YD0Z(GBh?WlU{zHM7 zE8Ot8Wnd5!UdIvpIOLg9PCI+hYk4mdCEX|z$OTUzL9V)j`)=;G|3K?+xy_P1cv@F* zRX6#@OCLH&<+dUlH`ZEgb$M8^$Yu*o)$-z%eB)-2>T+-4=RmEK*=ilW$zM~1u#^l# zIT&OQQnfWgp)!BOjwp<`goRb-kXaMmysVc6{NJgh$^e*=;X4m@Dxk-tFK5+L0$|2o|h)B z=%~^ea@Nxc;q56%kUf;V8DI^se2cDw;F(TA6r+-~YM}0qSd~DsyztMxhauZ`h;Y+( z=gtXTK=`+_86J=-7K2h{QZU9AE7_(-lS7_!_8nU-Bean)4X z+SMqmO;5Vmk_)UAFz~*DoT^2rnSHYug)+lnVu-M{Ir+CK<(Ne6KajufL-g1w^Z57~ zVfn0*L&o6|Qm$Az;6ICAG4ceW;xU6IGgF{@Lf|giZ7~}TMpIkMICq=E_Xr1Fg-`}` zsPMM)r>+)xa1<66TK5^>dqAuslML|n=Rw+-u&bVtvuLmW7Lm)p#XETFFNmgwFthog zpM*2qHm-y#DJ7fURkAi3(EPSwNzG;vN(Agx;t(v;Z(KnNH0>}28fBJ1!(||v-T5+O zxZu_UwCP}-lG0Tt!li);{B-w04~+ijmXioUnw>622n?Al@r(PD{SOg`oCEnL?62h( z%sr^tLtJd-7q;J{*m;Q07g~37la4On5flD?&rX!U2??n~Sa*b+4SUe>zlDLqS1BqnCo5@)=2LJ$? zc3nX$>;%+rN@#x-YN=a%_N$5vjT29viN2HWNCdZQP9(vmagf{Wmy5AUkneR{KDe8i zaa?O~I*cNqGRCPtrBi+Kp^ab+k*A7MYtJ&`#c27A6aT56%UFl0Ae0>7ghdULDiMH` z*R(bgVrE+7rd}sa^+C6SNJ%B*p&J4L18_AMM5qMOxlS5uoiqy9&Ue%Ve*3dJkz`+` zWYJzc>6I8g4<Z;j$9{lt3jw z29zZA@qq!17(PU^fo^8iH6H#E-rsM>to(Tg@|d_%0BD$p`InB#%$#n#H91b2->Z~Z zc)GQOAZcxYwi)H3IfeEmAx0u%YH2JPF?CdENQE-BbdO$AY_t^!vK`A;im9m2icTm~ z;vJPa@rgxE#4IhM<9jwjfm^t=dHIHC(f{|{Ujg7O7+jP4sNVCtfBD0|^Y8GqT~yt-XXb}i zyL+o9(INF~dm%2GTmiqc-Fdx=g#@nW*TCf1f_2qreG_sM?KUe*!2enIcmAqI@Km1M z-}&6T`}2I_ic*_kllIZ0l31+8&r01eb0n({C3sF+EfmUVYrTI%w|-#%7sThQg-7-KB`PA^LUw`ee#^f3&9a+eK7=;1lSsY#<#eZf0yh ztez9jisPb|`NoB)-ed1-`jwI#c2ud>$W7t3iW=6vdm_qDnVW)ZyGxV%ZiqYPQX4$t z`zO&vaPXCw=!?B~CFi;V9nNkk*842WZs}n3bL@l_)Td5c+_mfWd1!x=T$5Ed zcZF3(pSnb?a=n=Qrbm}hmUlBGl>3c-ubZVu7Yr)mHKqbVOF}T(x%)Q((36pNkjZvlfuXRuwQR!VJpv>BoPVY!oI*EEen6tGxI1+zR3It;!A+1+AtvBwXwo=$Qnc4 zl=ybE>dJk0XmjJ1Ui%8E!Ec-ayi)I1h?=u_PKIafz0DdW5|Yu{+)zoPMDGN2$8Moj z4a0gprtWR%!i%uxh8)DGUKal$sz!|RuyRgT+G+McxU)$U-_{i#MvI2cAQEvBC|4i= zG@~i0(y>4=UPPbdCJN2+9}=Jnp};zck7mrUYGn2Ncp@lEA5yZSk7juWg#@H|AU*^E z-@-&RVs$RM_F$uy5`);*ZIZ7z+%Mp-Wg{vVPk__KHVI0SB?ycoEs1+IzdBNw< z#ba{Oy|U}Hw>I^3)^U*v;Zx|8!sFWrW?rY#lY+uxS{)6N*Md7OlD@WxwIq$({9(eV z7q}R-f)DRrBINSp+7y0L6tUzW85>@z1!@nDj5(?8qJviAsIE4=abUE3?esFwkxG1- zH|l2hD+N9tq1lajp^S_k;#a2oR4A2W3(*N{s_ItJnagOkvOuI7u@*YceqwRj8i99) z#?3!IshH?FZm!Lu&|i}5lJQcdQ$@zqObPQNXc`p<)p|7-kc*5BC z96oooW=aS~_z+h#w3A_d=RhL=2S&ujD7ya(3#h_{D^DpJMr5Raug2acsf_34JBw5{N+npcS6$JOz-ih$OACc4* zG3n~tFeNfjV?oP6IX~wibSpkS#1y}*JN;_h*5}uGe9cFDa-_5uC$nWCmmJGZ5>>Ie zy=*4;6N;hkidyBYH)=L?Z;7LiSXUq&;+?Y1D6lOD?4&*drd{ph<88)y#fM+yZ_Woj zLf!OIAyU^|RnPj=GGT48*L@^r*wG{x8$?1LmN%}Psya{w1ov2G;_Odoz!}mfvT3(w zBFtqKlHh$J&I2M$5_^E)yzS=#nH`3(Wmr6yFopn@ckx+k?9?h1PkJmP6=`@??uu&& zi1S0CS?{6zAZ(G~-`;kvldA*VkFbB(e@Xj;>l6M9Qt27afdVncQ56hMk2hu^LR&3v_V;#0Ttba z>4LR^u0lLYXPl#kTa`ZE;Nc|jO;xpv1-!C|`B%vEp;3x~#4rNoVm-@oFFnNUhOIv{P7oiR(9=L&}^jbq@i5X5a{|mI~nzx&seuJ0dq7!`h^2;=vtI4 z!o-if!Zu8uEF7@}2ff8UmgRO-uo#LIp)($Q#i&rYU$7YOvw4>|W)SXT(3X`^g7Wum z%!jC(jiKELsck_{tubhwdZukUMNT@QXQai$z7c zrSR(4%flX>x(-sEGt9wJ&OHll=OD;;1X-*OK7NHm_<-Wi!;Iqk)|N=R*_8HmiK4~-e@CBTphEJozw3I z!ZLA~MmdI7Z4})F409MP>fY|Bm7bz+5&j~&I_TgV-(3Z9-jr?-ee`0)M?Xpjj=_5z zXmq8EIAOSIBMN7?Lq&D9t}b?#WX4yfQwtYf{ix|h^p247j02?>;)DTyYADEG>rSDb zelwVH&1V@AJ$5x{6=dmQZZTLCQQ+Rp)nxdlSdudz-u33m-apV zfVwiT(L;$cjyg@u6;M91crPcb`22hr)bTkgIX6;JIR(n^oCGo&yP3{zt-gv#bwb1L zNo3WtNeTqEBI%xlPS@V}zv1sJqX_;dK-vDcPCCbbr~k)Jda9PR^VT?G-<3K8dxOys z0GPb}jp5DuMoG&FiIP=s=`|pb7>N-=0b>B^t)aj7Cn!QenAVxm?DO-q%iHsj_4uw; z_oJGB@8j{?&CSXcykpNJFZ&adzw1|nRFLYSX&w4djm{2dA9lxr*U?v@({H0%V6yv< zSFzIGg8gdtxO7hbn&z>Ksy?f>o<+DE*!bW2eTl^ZyMxrsQn*@1opbXm?$Tq;(v}>y za_h|J+E#!6Z^2@x!>$#whg}L~+RbNpc6NNjoOCYSEQc-WqJPggGgdt?p|$R_aJlEm zGHOHm=KXV__1pAI2ON_4)f|s9WBr+|rLNpQn10YQJ!n^K_v)1YloD2jGDKegW zg>hi}b%xTh?9wYwUpPKO2zgm;f09{K>z?ZkKz4fv%J4~Cdo|;8G%5w_^*em_$NvzUEiSS`c7^OH$b?6m%UV0X2ZQwGVgf0qJ zI+c%WDhK$z{D$uMSH%P5Lr5ansD0+}Ww`}xPSQ6AMlW<2#ZY1W>x#YS+%M>`@Y?R0 zSN!K#-qREptcXf=;TU8qv3-*|&ros;6A8@@nKLF=L8&slmqc=fG5| zA(SLYzpC;wFsZc$3-1`IHUA{$$5~mh?^Fd?+5_tS^^S*L-jFB2y5hDWZm-Afc zvA`4($S(>A65wx-mrt`fx!T-rlzD%xQ}ahx9O`Z|d#K6}-t!VR3|~b=lw}R!ir;;- zFz-Qxp0f_iN;;t4Gj=cE^FC}9FT5KTKC1Zn-h1>NLQp^%zEr}|IPIsJ*rcAVkd;NB}(9KVJ#Gh?*w@R)65!A9-4!fmMb zzktN7o?&CvwV=hU&j?}&PG-I(Ij}Gn>;oNjHQ8NUPcTREO3>F#dSH-vL$1v4t*Plb z2yXzvkn_cNWeKi%9#EWrlMgK>>?2bSP^m&}%B}%|p`8Ir=?&ApP-oD_m^I9VgxpZH zU5PZo#bu(8#oIh9!aOGA)5r8}?3l(ez7!J8=6(1kc%)=nTad~NP1qOZd5{=AOF}`v zp;y46bc=xDsi>-ogc*h{=(A3Rc~qnat;-*ZF2Womk)__DOn@*`7g0x#evW6(teQD~ z6(T^>`#AF&%2K(|u(jd)5KJjFP8LZ1vgi0>qPvOSqU+oPokZ(B4EskydDRGt=)_q6 zR(UL1SMPY30hX>9+ic|xlyZ&$X^nr`jvYJa}ffQJP{3Zh~ZBFVtQ z*Klc#-4KJ=Sb7tPGT2ld48_+tQPPB4yg|sbLd@_)X`X0By|DI~D`X~1C>a9uh+S_V z)Hw(L{P3tDYI3!b5{%&e6WVaOs#OV4Rb1oL-=p@1;B(Jx<;UWre^ z9e`rjzABxgy>_y7S59O_TSW4p;I+!3vLtka-XPrTLHssmwVnb$cKFG9E8D_A{ww?11Cj|85 z4MQI0{#6AIh`I>V#JFkQ+q5XvmN(!MbEhiE?!0y0YYe#WDxN^4sDU&3k;UuqO z_=C#>bk-z?5bz4YQHn!j$oH7GiuPH1i;`RK2jbx7m;nJ=G9NtnaPs*1y_4%j^{5`2 zMT%1g!Xv=)cpr5;a-A|E5@q5ic2Ys9G1uMQXn{yx@~p#k6o)AUh_&f(GI7eBCXLvk ze34{CeN6ssx`o$CNj4;2aUZGM~ienyUIIuD)#x^3l|_rt6EUWu81A)Nz35_<$WT&(d+RH!baF5VAF z;PJ%trSv3yGu-jSV7_{}tm-rYunH*qz&6S8O{+_kH?G8vg4L$nVZLt!^!4R>WtG)V zJIaU#fm&zdgXG;&4g3yVmPM8%9fgKJ7B5WA>7O^H+xu9+%H{c;h8BZs2?1N7_ATe% z>|%pJWqc3cI@sIXb!57PDo8OwZsC(x_|_qBKDdO^g2uux0`PF2*iz*As_z`FCJ=Njx7x~m;6!a1h9qYp#IiQVF!3b=)ly1{VX|4>E zNuHBY1)T9WMz&WoIQq7_y!IcMlgS$(!4VXy>);$?r;L%J`Iq9$nDfHVTFK7WoNX+> z?wdRSvgPoh#j6uGS-E&}ROg-|P^#Oqt1CsX;=3#OmP?35hgUp zfx<2q35v+|jzJdFPvD6!ME6IIvP~4()#~RqI`igya|&Rhr1P=1t0W2hjG|_ zWOEp2*w`2v5JUor;SiJJ*L27IUIM#Fz?yr@bPwvuPVs^k&7avnCN+OSZH53$`k#0E zQGt1x`^S-VHwsk17zhC0kg^e?nNvd!kevm*9?~ep4gvhe_B?2cqm%n@dNa(5(-@mM)hn-ZcWW12NeMUQ}_v9K!cY1b$6q;&aM$d%tU=aX96y+5a zKtk*%_2$A@y|C8&a|;k32V@4PC>w%l1m=#mTjYmiFEa|V-my6z_NU-xE<)#yUQqZ$ zdH(wj^chf`;~t@hqvN?@|&Igf6ITF82_*Q zUeWyj_}xH--e4p&uz$0x*OU0BxLZ4{v|=`OB_%kZAOZsP4A4?C_513x!%F}N^<0&M zlLXDV^xwL>!1IXi4sCzG9cuowx1*!m1G^Y{+n%2K9+~k>+7mSP6W}8mUpmVv|wB+8UN6)|e4Ix_d zJrFx21M)i)%|lj~G;p!0nz+_VXIXV#bMApgh_JMN&`Wrp0wvqdb@wXf zj%LeIwwzzxQ%^uHzg7l%wuM}>dky~F(+++j>lqY^oW&_iHRUPv7~6>_n_5jZ?TmIPMXuA91bLkgVihlUyzUFt!b zxVj0PU_c-}KG`6~V=w=PVVeyAN!Uk_qF-E#t^-q;T2*=)(?(9Gi!aS_zzp z=}FoiroO8)+iH8{t7uYEue zOLO?A@?8;$(sN1EdqcLj)o4JEa|dI(eBkqf4xD{uBP06Mh+FzK%*Zm4S3|(bCX7mr z*inbWp(HKB9Qhx%NubJSxh!6;RsfcKn9;Engp;Phg;6>+OVIk9eCNudG~D&rV z{s{tKWzKy6V2=Qh*kNgX{28{U#I%5r20o=iZ96c0)BrJ|YXvpvWyR=Ri6$~|!t0o( z>~DK?Tvu_zM?07f@WyHU9nGSV)Ws8>lDg+)Z9CSV-E|NM2gPvym;{^eDA1xS!48Ce2O_5flNv@#0^evYbD1HqbhZQqe%p?Q2 zcV2)>t(<_&oa3Ibm9z3fO)tX_)-WL9V`fsOV}id0{De5!l7J4 z&gNsqUbZpoCOp@bP_%Pv^`r4Za+_?cIuLK0Y}0OSGTN&o)Vo+S`3r^J09drnW~PX< zi5DecLjxq+O&s(bCRDLpoo1~T2D=6K$dE!paGuBjL%)9C0Q&V ztAAW7p5)efRizEbz*Tpvg^9~LP>vmfz&({hx%2X>0EL~yjl_h6`G$KJtss~HlSC@t z#7rxM@fwHvvfvwz?C#vDl(2Ap)-V|OsP=}vWkvkxs~&*h>%WAGJ*(uc6OAarAb5T#vQ#CncAy}&3#qeyz z0@m{vCl#O^d9=PEun2ua( z6FueLb^a#yw>@N({@1gk-lCMAp6#ctKP#zV^N3plt;6ZA{RcTzPw_9=!NF%zj^!%{ zJCk@5WC&YE@6xTejcBTSomrXdHiz3osqw3g=RxM@|pHa2a6 zvD^_X+fh+cN3m@5twRW{zKtb-wzZL<6T!H1&Qh+WUCRzpb21yPioNMqfe?~+!h>9{ zZELMl7gro~Mv~9#KG?zgCrU45r(;T!>38&F|R&+7Mg*v6&; zpP87}$DaG=c;-F~dXW5b__rK&(!YH3dV;B(tD>&@PgEfIFB#`iS}`u$*zMGR&9}{0 z3VLTXRnU~_@05eufsDurwFE+fLM@I)Tb`_ihA49dHpVeT%dst2-ewzX`V21d-*%9@U3qSj)K{J*`kS9kAA0B2ZICdz;vQSXR>Baz`gi2s0P zNd;H%(Mmn=`2Y>yI^u3YhI1__jekjoxQE@+TG+~1b7R2a0odqBC1aUTMeD=JQ&ajF zVy55^J>TUDUlt(fs|##8Hj2jl5CkuTK?7v#gb#_=Q2+K>I9UH~Dr`SOpQA-Tl8STk zbq5hyH(~o#6lgZ<&5R0zOVxraj{3xAvXrC35}Iy}>0SFs;Hk(VckA@)VvIirLOYm8h%1m#)v$OYo=JArL+C7aK;hA)B zrvR$l6~-lzJ!U`GmbrpiQOE_}fXpk*^(DW2dyTsV7fwao<3Rt@=wBrFMQv=&n)@mjnUWMKHlvR{>+>LBcLhH`NzbA;4 z(de`Gb#`>Vv6V~tXELv)%O*^h)j@%{h^>ttEe)7Na+uSAlqK0G<(Gcx0>g*03l2IY=-h$h_ ze3MTt9nPn@pk&(P-UdeuwUH=py&78i92PE{!VJ$|WOKlolv`2Z!)NO(U9>6FP`N*{ zLCEzos67DkF6gXv#+`ckE%EL}+zKfVw z&oMOf1_=r9%6HhZYt+eNcfrcx$u{>+y*aarcFQW6lvI<*d7ijCqVQ$6EZ^L*Yhr-e zUyZd=kWWj7-aD4hzUtJpP}};Vp$ZCzcfr0nd0AMf@@w$kO_$3x+y^hj*2q)F&B$>?@`5G9Mr7LYTxp`;wP;?l-9ujxaj#CT#FUh2v z2-X&&=)&R=QfbB5i$QgSu+CcrkZv07Z zc(L92W8kqlc#?k^kS(*@3nuBN$dLp>mgX^^GBwp1E1nZe9DRf&v1s5(RhihO2+5Bp zpr_bVdjY>C#hnxF3F-&vZ(-WsL4S@5)`)0#ssoep2~9-OS>=mdbNR-#w+IQSWSHiC zaiu)HnifzLHFH0NxZ8q*_q6P^x&-6Nn`&(tROnH5>(s8_Dd)9@(Obg_*{2}qNMZov z*$=rVee-F_6ISW|#-S(&Ca}z88(frwlQ?ncoPLF44(28l(Ju zhN=O_%Oy3uueKxT7_Z5AsIX zbm=*2%fU9FzCN48{s+}mhCC|E+2djMTtiH);T{@D-_X8lUi7Wpj*&X#GC^-HCke3zuyR)&{UFJRQ(7BXs;2y${4N;yDV>vJ;X~8$mbMN+MuvKj>*t=BDe{{+^|?;>YsQr>bFKZOz3}cu zUAhH`V=)#k;)Ym!=1NeBk`;juH7Mm&y~Q6Mg%Z1(-nX}`TA0Mp&0AyLm@-DCmz(z9bd&#x&%(i3 zH{v}OrlXno4o*b8WGSu^0DuGmJdbgp*r1q0U&RNt25*2=fLLOqoek*BYEaQCth_Mp zkp39*!kz*an@flI1t5MVXQiqKILDRPn*x%_(|RWAwqcQwbXI~vRCuh1=D7gLHvJPh zhkhRnuRIP(7+hv^sC@Fa1JIT+PGE!%>gG=$;JO0~(wnYNL_`3{^E=$#sD%WqZR>Jn zzf1W&IRStf{F zO9L%T40n>4BOSzi^blf+$w0Q2at8)U;C(Jk!yh#@;{+3$0t_8)_P=rk{^AJ+asfC-g(?9dzfA^_7j3krF_s<1gsfAfbC`nz1- znRw&K!3vTIr5M0}1H(cW!o zA-0iADqj`bugFzG1U&#Y`rfoHCPiVtPXU+qFLTCpn{8E?`P8XAxGBy-&H1TSzcSr< zvgK5O7duV-a{F1W2$S{u#Ls5Gx|!GM65+5g&^l%1)=l|h##9PfT)(sUq{goZ%O}VY zPo_FM%VHm_lt8)S22$^hd=nua@9aH$h5awd4D=Yg<<>_wd96iM_*c3ap>Tdgb~y}G zM?O=-!HuTVOT>5(f>j+?Hv53Y|HQy>54z1{KSkLBe6N+7dui505;6OIk1j5Vex_Ki zS(Ac*pNz_sD^zrgFVhHk!RL=q^mIbzYG8=PsgID7T!tQ>_oow;G3gQvA34>FcSdtU zlOJkpir#@^`t<1tE@BXXgit(yJFU;MA9{9}mkg#L`@bH@- zYpgnhF%@^z@xZ5lplxDsy|9!N-#zeZk__qAums4Ku>q`I zWLsn^Nl^Lvey2=7k+gL?10r>%V8fU%pCmu-2xN>4oJeG*5#UCaA6xD!QOm%)QO0CW)Z8hRjewb?Xxr<;=30 zFDbOBNOxOZ%L6a23@1n(xE|Fu*t#C3*kohZTIqUf706bKN2A85B-b=MP5o+JeeZU? zxiwGq7FiLMK8rP^Rkdl)x9*Ln0o&|#?C1?6QJwXv2EM7ZQIsF6Wh)k+s4GU zZQFJ-v5hn5;`{4Vo$uoBtM2Pv)wO!>XRUrs{h4g3LEo%wo(N{cfzLYXXUyT0RA=3Y zu)J!tdWKuuY`NL$`{Ov`X^`LXN?42X?LaUPm~$EO)T^{7jPkev|JARLRVR&xH)bG> zl0=JK=%L**YD!B~NzPSG>Fw*WCUhWnVVgRf4%)}RQ*)@)CV4UKgf5OGXtwoX2QqK= zTgbYv22YrBV5OA_Rx)m&WtrRqYH5XxP$-=Xcqkm0j(bifm(?XKcA1@#r?Oj;dZwYA zXdd3e3ne>RkO>kQ!q;z5IOPq4BBJF$mnec0shOPMqid=U;LW#WY%~Hc8E;OfWV!^L z`kf!nm#0Ou12>Aum8qPcx?~B;cQUjOwfP6$f<*{;#yi& znzTr@xT`vK-YX2_CCQN2eB2ZeUs;i05J59b&H0gCY64_Jgj9(1Nz!E9gb(E!#XY8s z2R^d~>2g@2_6Zc6R>4nlGSp6nP39RzRjdCVpLE|^s5FNYJKT#5^Wb1XkuIvqY;xuh z;1s32G{S!dx;p+S^WCB0{mv3B*ct}6!uY&D=KPcj-VxrmSWD*w$~}R%Dc>-r$RAxG zQR1x)oi&bl?#n*~176TJmJZ$>F`!slLS*o8R|5~BtWg0W zbl)x`Ki~@!1@NqFv={6lqGO_Wzb5e;s6)syKT#Fgxi87xak1fGIa)|}auy@>BeqRmtFA{tsDzl~}6Q6$@QC3%6qAIeWr zb{=qM^tLa@f{D&1EH{=-2)dDA-glyMTCE{$RK{gYp*>i9ap14$2CSM$*%!*4+kArr z7KBZWQR`)$^Dinjv|Rm~N6RlGhqrpc`g**1Rodqe*u7NziyB0M1`N7p%1L*jFs@`N z0^+jJ14{swy!fjv!pj`~cS+ESG4hxko~g7HJixc18+q)uH&{$9Qm@xcpN6j1J4s05 zJ&?9^jWuRnK^vf|#Tog$F-aAxY1C}76snf};uq#Z(&Up1ea_;efuX=itvHpV8EHuB zLDJCkhA1G8l$pz4Nv6uE z?S;*HT?4Mk=C_^e7_d>Tb<}&SaW^&UKLyH6?3&hY_Jd+$t!ex^*rPwmMz^~N3-=5C zA%@o9xfrm*J9oWH!O}=bqmhDs$GtNx-fhl!?G`}!8q+QLbcydNrg*-re*qOEWrvS$v5DnD+C#M}|Exhk-bebz5T&HIe1U z!;(0WN!iEN-dLm#Wqs7=Jb0azainE>ulVJ_s|T`> z6JD?XH-DD$ExO2s_MEV@?JIU@*|0Yg;>H41hMKkNfLP_O2TT5*P1y6S>7tS!CF`Pk z)41z_Ss-np9s)}? zHt6m_Y|cOwJc;qn0((IqZ8Qdvp_*ZLqop0LD=jJ3Hrzwd10y5gB$iOC+Y6Ea1+51U zt-rqcwJNRu=@Rn#9hJqiyifE9K(D5yHj7aIXSx!aj`J}Pks>_3)lZzpIqtM#^H`8e9ss%_YwDPLr%-}x)VkN zx%oVUJAjDZ1Ghs330=jyd+H>Bji(&>`_HN_lZ9=iw3DD8CgRmq{ry{B^KY%@YVSVE z+hkj*m^N71SnIjT+lf5KOv}UcP|MfXF!O&07~fie28H>&IiX)=hO4dB!>eHum^+b2M44kdzn)*{lPDe)K zaTJvM^y-MvvMjeQ96O6SEddC-Sd{Oe?~n=1NLr7a(@0`n z47oA&8d<{`Goe;Efzx3ke>xSiENgyPF$C1?qeuMv8XD3VnO5c!t{^j3He3lizqzwO zp)s?qVoqO2A>7CSiUd{-A1?i-gR>U#5$Z#l+xJERkz>0O9 zE`&@nJed(((LNDl@D(jTI@E*)ADKBErr2^Ynb1;uM7`OlaAUNkZN1O=IHdYYq5+X% zJhOzcE?iM)rwEiDrL&5v{sC8*RoC&OntGm1uFsSDVz*_-4W=M{=Jn>uuMlK@`;Hs) zEY0Z~|G3el>r%s1(Zw&WZfuHeLX3xLrrj#x!0@E&w>pEZ+8O|RwXoMWo8D~6E}Z}e zqc;v&hiZK-1yS9=*vM1l^OH{xP)=H%#*F1t;*$iUc>doflcRfUr@j=W43}LxjnzRJ zic3S+pY^Z$4^X?h5zV7p>&31nfhO!ogbz#77RJ{edL_I#ze~YegRS260P~kYo9~v7 z_sUu`!1YDtv<>r(fy>X+XWZx4#g%!Z-jDa{_7q3!i~UCBFfY{c_WkIl;Ywqr>r3h1 zbWmK;6B&?c7g~{u>gm`UifJ%bt9^i63K6-U^>DVBA;M z_i?6)si*f(=FH?ExA(>!Lp5P3PaN;jfbJ z?XJ#Ft)rTk>cFIWf!0)om7$fQAK7SQ`oCqhq zZ3RxwNqKhxccaLC==vshDAbwsmbrxCB^-CaiGrrm{EH~8tL;s#`IUq3Zg=N}`NPG& zX=j&y?ku$rg&WTIQm}XkRFdiDVl(6_gf~Pbp@5I&u7bVYHAKR3cw_hM*QEas`aJYL zRln-Gxu3uL%Q>pnZF!4eAhL`dDab5l|A59nqm0H4QLzsQY(lww__t{lM9ryPy;ps< z0ep<=BVj4-Yo@2^)3cy}U5|f#h8;ixw7=*Bp}x%c^Bti*W*vfoOq*hvqcHEP)t17v z`6}LF-PT_OrNx#NIf-MfFs##ySqQgpvUQhzTXAr7+VPV+GIEAcQrz8K%>@&{ONA`L ze^0T>a<~1ya%O@%h`n!@-b%RFl&S}4quX(vy*cF^ys^>gHP}IT80P&v3G2ApN=p6! zx^lb{5V-H2dY>$JtC zZrzQQ*K)D_<8r&%PNb2?Fu!3G+q{zY(6n;F6Wl4~nZVl$tnx5V#2o&iM=zXV`w4RA z)}7JQj@$9~%lyvkmCLua3wLGPiJgwP%edRTr?E@d#z67r+Mu?!#QiLO?33}*7qDZ| z{DMt2U0YiD`!ANp>`E|`qo-FG9;k9ys4i@^$abjgH?JqzTr)AWmI$ zECmOL(z&8$2jhD+X5wZdexGmG*u#BJ<>>CZuy^YR!dKH?ACOEAy*S@pYkh{je$IlO zo!4h++G+F!)C%2?CvG0k<{bKh-h%qmB@dPR>*nm;1Q>!pKAvwEs=8&r>1dRDh{>J= zzJhDkja<3!lB?}KmCitCBaV*X(3^QnwZ6E#8Ie8S&9^h=M7_Y--@)n3H+cvQUwk)p zJ5?nhQqCql#m~LkgC4_DlFK{@!RRtC?%;AF4Nv+UY>dNU{HD*t|dxoqoS z!oiW0U%b`OxAhbzc@4XRQmpGg#NP%AK;XWV3=$ zIiVieafIgpv*t==H9>rDLefH3n~avySQQ@%9xjH@PdkSp_wSre zszN7E(jcYCUK0Wz=(pZ2WbNN3&FeclzC`4ndX`^i%b&u#R@ykiSnr!#2dz8IGyTup z-b%iy{a%pb@_<8HUY&4e18*qjoTtvWG3h5p?VgivaC0x6Uh=Ivj;CiQm#4E%j$`w# zHv2De^)2)v zMm+#1ZP7`5>*D(2@v)S-uH6;h6C@>-g70mK!L^K?q~vLi?jJB>En|&A+z$zZh(4#D z6C62-aDMOZ$^Px1n|C-D>m(0dnlrVARwn~U5>-%So>b>NQmSS49Xbr#lO-kYWJ_eOjc27SEEO=wtGd)iD~j=7 zpN2`VN`;Q(9ll)GP7(Z2{(8MVCf#j?DIx#}p2d4~DsMF*w9DeQ-HLM&go!zc)y9<2 zWE3gDXaFq9CWYHjc#OgpGd|i&`6YJ8v@9E8=$nzglX2fmz{Zov6K9Yh?LxU>ITxV} zc<}mUagTd~uUF8|yUp(cK!BJ^5qF`sIPYAim;+1Rf!pKvjH zMie}KQnspFHtis}f)A+dj#Wd^2hWMf+8J`aq_qE_Nd*c7AIKw(YQ#qEjyW<0HN8V* zE{F4T_PkHrxana{m=6sw$%1tN<~pvZMT`;6ceqpRa@3qZTCVmLNF?wmIWh%7?Oiap zONhZ#))}-cXhwQ_b%dyCy^DcQObk4kxtib_CMAGA5G;d|Eb)?|DSF?41t+Mmk@pRf zj0?_KmtIIMTn{?Z>Mn3g6;VED>6gWuHPJXWa!%S|3595yf@_FKJX3=LscxEBK|cVx z5adj~^t`R$^Ikcb&|1SG6r_zTFpNpW_7RrM&xM*;aVfW_SGe7Ck*B2L&K%{j>AJhruKwX=Nywtza(W&bxgpyLYwx~5&CU9YMNu4Ps z(#a>H^PuuDR5)U!;eu5>ksXaOJ7X$?s9H6mgC7;Z$tg4-jMuWzg%jeydgp3j*dk%1 zOQlw+N(-P4+iUx?Kw?K#No+#a$q-tzA*koUvYW&eh@q%lv1@M5a$vCDL={bZGGQKI zUo>nNSjsfm$JA$ERu3?#KzLHDdGjcKtwZWjEjtG3K82K*0Sxptj%Z9$@pv7KrTH-K zu%KeN2CjJOhvgHlFnKm8O=(@_W5ji7zx2<4yDx@SJ6>XXZQ(=+g~v29+wv}V;Y*Yy||$ao&%zD&p}2!zYQoyVgo`YXZx=3AeWCu_CP;vK=Du6RG72PE&7%7E%h_GB}E$k5VOPt7o1KC2@t4?J~0OdxcyG zb39}jp)v1XagvdlO6eSP%kzu!yHPny<~0j_i%zRGQp^~aSE7r2$HUDnw-6`Re+K>C z#)5{J@(x}M`Q(kWj;DXsEuf`)@_7SclTEisRu|u}n9&v~TQg`4 zy{8X`!e#l)@lBGpvllXlJuCzEu@y~lPV56vdMG7<5sP>#5S<$ zhE700i<+_Ob&|v?{Aki?%Pi>?!D_JDsi1&=pFV3)(rI z=~x?PE5Cpg@T?3~)Qh=r3AlG^^)+oQtXBsWIm__9eSyqMuCc@_h&Nf ze5%NlPMp1`<@vE$Kpogd!~juBqfd(JCOGhE-E<^Cc4|_#=TtIg`y>uRHq|S|p;?Bv6S?$ir=URO(SY=8k(E<9!3hcekY7g| z!HJ58pj~Dqq3w#NoVonc4o$)@)`cg(r`Akfnipd7bSTfpQOvQSrU`%P~ys7wp&!hT?mZ!hu|7u>= z7;VgS(AGTaD5}EYihDL!d{)PHtQGCyXYx=Lb4r^|Ix#azZIFBjzxY8flUm3M9-+A~ zBf-}!91@A5=To+ba~LFss|*v!tj+zpHYvQhA()yjK(3wFn7Wbp$@50%`%xtmxO0Z9kp&87p zh3u~}lm&0Im`6*jN3Os-??TZ6{kve42UVpd97pi37xHad7MLHt8HP7p4% zxlv0ts+dyeE*$&!+19Vb62YN@tH{uq{%Rb+w^UE+GhARQchC^iVkrNOIeErc{$J)b z|0RTwnTVN{i|aq#5f&m&E>@QRoc?)M|Fb3H;{4wON--|rYD-ruj6~bSAS$9VtS;`- zGQp2ba7;sRQ*hfbK}}F3R549#Qqs^;?(S*HP*NWSHy^nVInFmaFE!d1g0t7%d0!v= zHgX*hT`+DAy5TfF5XksWbG=!v2 zLjoBnC@9xbRfvLXyyktT06-Y5kPs9R5e+pF1vD54SfEgZKi)o+bO6g3sx{=?38-bC zF%p?T7RuH%B(ST4*!J?B!F%%UR7^r2iQY_MHXQWX)Y#$H; zKtxI4Cm~3513AFD(3{{8^8o#br~3Q}8vAOI8X7-}wzh!{#S9Dna+3!Ty5UX@#2m#v0KLHyO0CbJUpu_|Q=z_fqSQkbL=sX7kb*1k^<%GIqw5aaI^8oTU zL=Xy4+%x?99s8XF1@nOl2NJwrV@H&Zh;{;07uF#Jq^Q>Xs~l`di{O-6;GrH1r&N4 z0QUN2Kj<&>3IYPeON9kq7s@4M{Q5KFR|;7BA>X|s1b%`@2ff*&1Ood0{<<*p3}fQu zni@Xcz4q6bs;&A~lb=rhxJmf=rKT3*4f^RJE(Y8cL?RRr5KzV>ba{<*o9F2wJlFg2 zYFK~HU1Z>wA?7MgWoAT%f zaKEelohS5j6RXCrr}w?a{q6KBcuAyL7B#y|;6C;wv|TTR>o*F%=HHnq^j%m(vNe3& zYkN*LhmVP7guw9AAwi{rehy#17+TmD>)(IO*gbDw*e+y@q|b%&{$T}t3j-AR^Q+So zbM5W=O=!pXJ`UPRGWUnB90YXm%CAxG)^3mp4CE--zx9Y$8f=f!K~j=8{Dlw=^!3fi zVM!lq&#nz9fFCdL$NSI00g&I~ZiQfCKt5FfA-HRO@Tc;ZBzYYqh~sFhpjO-OL&7)2 zc48!ZC@g<-6ogb5*9tgW9!Aalr-N;*WNgz-CeO+73(nxzOGoMGq#^C+?{$;|IpXrt z92Yt_z3rcQ_OF@4Iau zxM5tom6G=JMw2ctNR%tnC0sA(N)I*XbeR|a%n z)Sf?3op}e(xbPL!y+K`PdQ-K{z;d3Ge8Sec$^k`Mp*amy< zP?)B-@dyE5$foEuhFH3RnO_$wsktB*Y&@28!|<@(!a>v>_%Ej%fR%30KNt)@Azaqd zf$7!4ngCYp$mNqJq4^6~n{Si3@JtMiME(tu$*P$o$UmSE+Q2mL{A20odV&u#hT1-! zf}T0o+@T+*1$y)taUH(r<=jcruh=I0fmwyFMW+O`w%PeLOxVv3FCrjIA*S80;l7&c zZ%H9IQI3aUu%3#iV+3mWc`!M>_Ks279FI9_9SY$4sPYnicd|Gqq+7@f1EVmFCeuky z4O7E^xvkX7x;gm`RP;!3kr_z9=)o!~#3Cs!(m;$M;N)6!u-b$4*u9(94naAZ<_{7} z?%;_>86<`g9nVFU8x@bHx+9qhYMz7d)}~u*?yV$XSTivvp%dKVh$}qD;|P`xVj-oTIMM7h8FaY)0PnvlAWl=?E?<7o(As zZn4g)QDINm%?(rP1Ze_%i|Xsjo0g(rd>Pv1yE>G18O}vOK{sT&9GY!--e(?S{9E>d zhgCJ^!%nt{C1?`p^j+sSv=yJZ!T~gI$+enI6BhpmfnqThH_r5Yg^x(WeMNcTcdN?%?q~T(a7fR(!u&Lod{vp*go*d z1pW=3ng!^pM3EWKC4|fC6PAI@_S`m^>sJR%ezh-+&OyhjGtzQHQ`D)ie+D7ei+|OR zVrRr~oZ;7V$qUEvSaXjfy~0;E|EX8Uw&qXzg+$8q58R7fU7rq%w}uu zz0k-bWnpLO4}Qrd1>e$DK}PLKpZCG7<96?;?I69`S{XGsz^$1s)+;nYZvQ$PGetWg z;=BL`_n7#lP{L^-s4pr7 z#N?>pb=2oW-x3y(vglo5m5$u<{?LUMWcNM-hI1I8G573y8Vvm8EV z@+))VcBYz;a70o&G>cP|N)N5KrUr$4bioi-^1=rZw5ruWg0Czs#Wo?(>6oVL+|o4P#&LQ^>y- z&@mM`TfHbwdEO+RjzMg3y!A(%adA(>B3G-2PLZNs)d{NHz4Th_zYCwzE~aKC?k#XK zRGh4Dp}kXW3rp}A@9S1-UrECJ49^jU$)FX6=WxT}N#Ho(ldga_wf^j!-lC&z`m<0g z%Z`?SPfyT9hApE|U+cS_=biM;<(q~ZZI|Bd%@*xcbI6$3*nICjs4q?}=()U24bQ5a zh&Pv|l6f8x_t}cCquMSQ{9CekM*jT(-h`bJu0bgX>N0!-2CA=uyy@IGbH3LTnUi_-@njfyl?0l@+e!dNM+rk409cHt*!_pm$Jf%Yslj4mdxgG0o+qiH2XCjW$ae&=BUYqp=zlj{oJX?t_lbk zqQ}(L{H}3nadAI~AKHGMjL{&3Qk_kY!FJ8Pi>BS=V!sh6c!g%_OjgqONJ5)(V9_tN&4M1(z$M7(jo z$hk+9IKw66LM>0ebf3>gWxv;`8~C)cAMp`Ula8CkdMcCa9yT}bd8#)b%*qT?c70^Z z`oU3ih9UCDVL6Fbwp{5o3+)X{AK&P=-A8Q95LkQp^k=k{s68A zC?jWetz1fIgTJ(+l!aCLqC?+D2BE-Ab#8t!rT{?cjdd;}k4~~1@SX+-41_=sMmrC~ z%xrz^pU=ci;*qf?YA- z+K|acVAntK%5Li}lBlUQaSL&uLhO|{*~c$|>5B2KP@dx$X~?f{bCS~|zT~}N^E^Ur z&5=6Ox!K*Jag1P3PK%Xp5wEgYVacKX7{p^OF^C~Y`H6GfJ#JXPqu=z*z)AM$z5@kC ztojHVWC%r33%CC%Xks}tn0+`tizI}2?&a*H)C1w-D&t-v#*cqW)gim?pGA!JKs=fC z>Ec(Lnnv23EjjB@ouuWEdnMW?Qop@D0bhMlNExbiaEB;Q4UZTh{+4f8gP7NU01=^I z7zb7H(8s-^6dgG9+!~-;PK!usV39A{Egg=tCe=P%WWD4lT+T_8<#Qo2Vr`CI7|hT?ju+p~?)#NcA{$7yevX1ijv+zb0z=a7U>q zkVpVV!&5UsSE*sa$PBBqqGQ<^owC9E>fA;8(_?imu4a1nI5Zb(Pw@E++ndzLW zzsCi2Y5{G#OIKF^=dSZFpS=aCbYWh?`&x)sw{%D@T?B#rRZF*QoA*y4f1fOGN@6tO zG~&&mlC%Rt{A>%8=XAul4ZVf|4~=`$+4_2Q?QGE`*J&;8iA>+te7QNs@Z8433OK3e ztK|gnb+i!!4i+hQ04tZR?c~c-16d54_VDcBV!P5ZtVEZb$A|K%^TNR6lPx zcD<=pO8emQrp8MeMWD>`I%bWcmEV|+@XG7u8Ur+EU_N35_@>^ZU>o|7yH7zqYiBG;=FF;Radl&#ooTkq@~* z#A*409#(H`k}#X;H}_cnt&;wH^n9zXh&dT1V88)l8;;-~vUHg)pQ7Ysir}E0Pu4JJ ztn#POa>3#?(J|ar47nH{5HhL0k&XJja?}5}r{NN^t%!?MU>My=k9AGgtF>6RU9Mzn zGh;7inmU%ZWi{|4)0U;(m(*ZZ#S3p|-d+$7oy+P+VdU%URPlwKCsH;!W9lT6k}bu` zTcY1q(4Cv3s8PPWcSQA}*p?MRaHpIS#p{#kfL2GVKw``4gI~Y?A%rYd0KnyY%~Llo z!`xl455ptLLuq=|99fdGji>8M`*2q+jFx9^p$yU4o9^AV8nOXD{n$C_vhwRNjg&`w zda;$E=1<{u5XL!UP-@a2+>?0-ooA;E(j#G!Z2Z>@urnT$O&k;T==4Kc5R*{cq|D_W zJsOoISjxE?&-O;DOVj*4&+)g&|9!;E{N;(Y}bF=m>i?0i@Ld& zdT}&!EoEAd;3a?BYd!{ZALesUn;njn9J3+;j&u3#Vo?{6)Y0MmVtmXX!PF29oOP7y zSGMe4w>}1^Z=2uAIf)7l@@PfJ z7cHwh1ZG01EbV1-c*UEDZ2DQL~tW#TfTifVj8Ibfw5$NmDYF-aDWny*X~w1FY=b7 z=|TA$@W1TW>p!jKR+z8=9zHp`NTQY}n_L@{(0n1#E}Hd0+rGK`F61oG_^vC63Bv)i zoeK$_2FkSQYMmbU@A_RkrE1r+##T$;3zwZ24gU&_cL#WfEknV;sJ$PxEs3WY;8NS2 zFh5&m*03;^9(+H;p*?2gD02wt6mH&qHRz`k=(Z%LB!~+wnfDxhep2d_w29(nhR_U2 z-Yu7ddjSD2h+RIfaRk&bIbZuleRuPau2Ka95%vrd0TTW~0P1C7g}E8hk_>u@lb~K^ zF69$0WX9SX-Bg5S{E%jh4ezW!s2?sjC6lG43!Bzlm0{d|)t@04bEw=7I=9l!Kydd`AEw#4S(1W%Us(xcw9S zc^I2FBv~&v?CEElpnbpxg^`4e)cf?7oqf!s9ZO%|7>p?s4Z^jXh|bq!uC|x~F$|;k zLud8Jm%fW^N27U?G{YgKBisg|a$^!7ge1aB+e#{XMql07g1$ks!D58-ZiTHK`w!R_ zo%aP;NO9ywWjlHN$JU2d*sG6b7xW~Kt=3GfnVw>-)3MC=d6G8$rb$D7P8oIQ@*wtC zoQkvc&C!t%j$t&sw1%B;Rm-Ws_xn-&Ax>MiwW<{bR(CfG-acU{`(uy=%0=OUFHh1xE5 zw~@`W)eXDgg%ihrEM$8S?wBSu9P8kA64!N|i0^~VsdLy&!DqVR3w2(c5_wSFfKGbU zRzx+84Q1R!R+@RoUE4}cu%+D#lHiMZ&qtb!j2{X6$U=ms&w}}p>=x_K+A~yVC*wa%jy<<&Mg) z-yqM{9<@TJG6iO96pwCfRq-~1!99s%&9=aWtZ31>2|>h+MBqwlT^oP+pqLKFa7qi8 zi;=9w)0+L;V}Gl<9GtrzJKJ{@IkgYpy9Qt&t>X6-GB^LA$2+k&x5<%q^%zwZm0_ur zHM<`369qIe&=u_0AzkXT>e_Q0;%1F%CVjNm$>27?dZuIOUn%R-_K1p+G+jQQ=w!>S z=bXA~vVOdQM6}0q?2p5sbb_*vD>ADsX;?^6HM_<;sLK|TiDy`rSLy8o>-F`G6K9Vt zcXFiN!S`H!v zRTIh*Tkkp?P5Ud^odig;Fu!=ko8PIf!n#F53+Qf|Z0;4be)~})hInr1r{rNhLsT#A zSXTFXoMx3|wG}X!*YGO{}cn$i)>L3lhVjIga00SO4}^Ei_VKimjSV z+Gxjzw#1b)6;L6ibn(4bxE*!M@8Ru;YhG#v05^QOqGsY>qWr`_3p202VyfqLR4z(t z*7~0HIIY*JHuuftvUojwgDC+{10983ai24Sjz+*Qx>3wZa^N}~_jff3FY7;L;`m5$ zEeMiz)T zyZ9w@$f3>@u1nnuAG2#?o*u4oL?RpKR^w(kz(QVb1CtWbTPCY@sk@F1LlVWhz#N-y zTqQxvwh#bGuAK2w#-e3#vI*HZnuy8BLPH$%yo;k;1X180%_=y~c`V zeybIc`PvuS+X~)T~!n&;#;l8o}ULzu`ZH zo7QQfD;`m|Mr!uujO{O7bAh30y^^y8mEDJozPStE0-aR>fS~-PK zoeH)cDu7vCb9QI=e!pz^m*i~RtTfh!wS3;!-w8^G=h9Z1E0X@V?K?FfedW`f_kOyE zQJzoPtpRbaw02F^qS;=UU!9GkugqBJR$L~bsv=ZG*9(g3@q5(Hq);-BtB6}!2O6Of zRc+imDo)I|qn9h4du)IRpjt*P;|igX-7PQpi04aY)zZat;+!tNbh*E%nn%-OI7D8A zoIT4lggkFVhnVwc_Di`er`p2sglq$uEkIeMF3D`ZhDb(!@x>UgZZfo8c>8^Ne5dSk z!M(vkH(@R#P=$|Lt!Dov40HGrK!cvb47~uEVq@cZ7RTj0Z*#*gUg0-ANjkyTQ%mCH z@&Jpu{Q$eDnEWzP<1Z(+yZAd8?6llndKoa=+H)Csr>H1F0N$&c2A* znj1zV_ZsN(n+QXc*v2o#%MD>rzN-3#>_t%i&~Wh&2&wu@4!Fs-=#fWySF$^M$(d{W zH5E6n(!`o$ZoF@De50LuykCFNh;#aKS`U?BG{sjmx>EW_=!?fY)aare;~T(vB%qne zivCxgm(Z6cpuGsJK-V%DA$qLS24O-y7)|!Rk48cMHQm}JfPSN-h{hXFr>v?TXzKN> zlq`DJC1FFEokmUnw%J+d_3tT~%T~S8t5*@BUHSku{~YJ)bt4z4c;ykB2J-pTn7}wI z7FrIGx}L1&T|+E0fa8Hawy?Xa#b|d%2A{P)snw?L_uO}v%B#pR%w3#pMBSI2#~i$@ z>cm=-fWTOkcwk>Mg)v6V7p;HG4r0}|dRd`x>Griis0H|bSY}s2$^FP6Uv=NfD6t~L zoinzhY{XRHbQm5Di03ji&$8f1!G3n8ah}e^*q_Ol!&#T-_8m2gsuAt|1#FOTyqz-^ zL5gy=DC}`0YdJAQk27ZSG%7L@a|=n6RElbs2mhLyB@Vj*aLm)~aEWo}jY@}vtUd=|%RT-lTj1#xXas!?@`6!yzMZcg)q3pnPIem(L@(q)l2lhQ4?g`B&y2%`gTS%LNt`zW@?wJ~%gl|r z|HLbEej4HdP?ns**?7A$)uMPFayg<2#oGF>a)`Hz26o*x{E_6gjgEvkoPsxb3;fnDAgiC0L8$Up^pm$+Y>!iM*zZW^$?3UEqRh ze;1}UZH-1{SagF|z>E36?39^Pr^M$3GJ*As`1{vJlsTw5aq%hmS|;hn*HqrAt-9;> zoMmHkpQ-9QMJ-I4nkr%%uf#wW(lFftDc?RJH*Hlq0R`VHL`U8}{+mJ*`>%aL4Cuf+ za}}6rdAROs6{I^~2-W)UZvJa-b$*s1N`|-cr@sutg>tQs*!|7(R>XdVL>CoTxRoha zcxXE}l1e1|q|})pH{MO=#GO9KL1#+;fU>u~82>kj<^MMw|AScASy=zWs<04ou(AHn z>HmdTIN8{k{N!AQxJA zmB^rf^g&Tb6A`%zp%>pvuJez7^^I=h%2d~f=d+(X-^VYDdDf#OccC>=btDHg!f-+= zXr(0$4p3lVp?oAkBr2(~89Ol1cHy5E)P^jf1Op~8y?a79aWb^%j(X}qWlk|$FueTz zD+nNHV4=c9B4%kKBqAsd5+8+-SFE?zX~;DCJ~Tc!!@b3bQNuuBLgF(D(R zIUzW)z+hfF9Pa?ULJ0v`o4|l`9Uy%~s!zs-6+DS8 zqAmVD2Y;+B=shUsU=SDerdMS@Dgz*_>)R+Ow+E4}*I?hie_8im@cZuPMgfss0zf{h z`+4tOL{jOH6H&XcC(o4mV99*$+Kxh;tkkKQ1JZtojIaM!9f2!C}E-iP`DtFHPIX+ zhk+l7US?7JAD-PQLg)ujEfD)7NDv@@-|vUncL`$zsJ5&h|G&MGbou2r`F~7@9!igX z0$^aE1c5RYm9Y9Ksfb~~0z$+?fK5z-0`mPK7eEKU$?g#?T-!)Ni0(9dIAh*4sLz^FxeGd#B^wb+73;#^!=n+!+iaAZ!SP!vVZvzUsm4Ba*Yq;ZT~7wU}8j1 zCtMOn1O##Y6sPxIj`X@s!$X9f6L%T)8ejquLP32JUYhq4)o?O~ug1{%i_T?zy}hJN zI6w{TuNFc_hJ*A;k*($3non_3A_L*LG$tS@_3ny5K-~@^m*Hae$$JU|NrENbuF44u z83zRLqrf?a?(hZ2Qh}#0G7i|w-5Q2}2_hjW9SN@8Apb1NpY6^Gehm!+gBa{wU-#yz zLZ+vM;jaDk65SFK!JLP^$9z`mgma>JHa_>FdoD)y&i3wNQ7s|D5NBtI2o;g&pObF- zFHP(|9qZyIe2?{*zlTI2@&{kueoIHD&8R2g_k{B$DEfb7+SrL_JZRaazM*t)FN9+* zzH9Yf@!bwQC-krB;O;+CP4|Tz*uDY$BtRepNZX(KWM_qvm^3djrWoL#eH85-Rk}qc z2zz`h{NTTi^L8Z_tN+%nhFhcPC&xXIV^ivcrYO{5uARy))bEJgF&{XqUVd!natXXi zCA{osY9Asa8qX;tDnW0UNHe|Juo(3$NX=DdeG_zp zLZ_NZ#{E{;o<#k_lw9e~5e5~*P)`8#=)0EKtFvD%@*WEUlFlJ_zXKN7TT6qk@wm-t z|8ebDT-Xv!K>+IZ=u7KN%9f=_9`T_o^g;g%xn5&=Zy3ZR*MWHQJF(2dBtqnin=SFB z+ZCn?T?3je|43|@R8!!zmYcd~$Rxtbs4S`C3}&xG$AQ{}-yPl?Pw6gMw`yo=bM!@< zP}=0L+#_P?hbnZ+2&T6hpNx9RJ^7bx6e_%2XLMz7k{7q1iumQq7@kx60l8e)#Sz~< zhs_9vk}Ap8-N*!JaO&X_%VN$gGazkEzo8$vv?tboMf< zj|zKTe9+)E{lTdxgg{yamgOV!UZ`*#YEEjmK)b&C9O{8V+hKt7OBUKwoBiuO=%TEC zBR5MB3r?KUgGIoYZl0A7EQBEaB>xD&tY=TDlwec-V_jVN?1!bQChvFujBaY@|l>6K0i>Zn@}(flAazkwb-p8{<-*B`#laO9+qK^jwQMOB>v-1FK3iJqXh zhGk&4y=6}N6L-#G+B6N-Wr>`tY<5pI!^UMn~ls!2|~cXM7v&sB`PxVgIqY??PO{8g^^ z>}#NU7W<1wVeDxrzjPBRc6p!16kim^*5_R0O~=F-vuCxxy~~Vh3^SojTx2ubm2Dzste(MNm_2epxYCrhiSo5fhAHkt9> z?x~1+$%;bB0$5|J<2#gxOOk~ZjI#wZ9%c2g^>;q87i068+~j7QH47AZi+}UVyl1>G zH2!3;9{I7Uj6K%jzLC)#^*>mYyF7a3Ze38iaYrt8Cw#RIX|PP4F5s4s&78E}Vr%DD zs)Gsf&NkcGTVE20VD7H)G|5f4(Y8eAW%@9-1omopi6n|`ejm)-J+O+kYSZwT`+I7X zLCSR0R&(ScS$a9zaxIoZ<*;2OjJAy8{FfMNCQ;_8MM8_?l&o;tLRw}SE7wkBfeyj& z9*jntt45I09df@vF@XD7XS9s|lfIC-9+5?jEw+ysAY}U^YG0S6?i+-}55_g2 zj5c}Kbp1n~|I|<)0xDJAyxka)=z|UCd=P?{I{rDb2LirBgD^LcF@MHwp{NiCMMX;c zs$dQL(?;L5{BQhUUb2b0 zuFFUkzR!OSizYqiSc`ufujs;c`hxTe3+_-Clajk>jYUNTf3^=|EF zkhVv2ugcG#&Ln={$^NF!CH9nb73Q`~tayuAexh#W8VJii zTQDQOUmevPm#ULXa$%H{oKSe9yqO=ag(X-%P$n3$~hu|v*Q%p{#oHz&SZ_-R2mHjZbE2=p$V=tmEui~AHsu_q^ zZ_+mjY5$%UQNW6BL%m2{ZL$X@-6w693*c3M$k&@bMj19%aaEHT0XpL6CF(nEqnhIC zLdba^*9}Pj`VIeRd=6XTyerR_o8vl*FJZr|IA4-*7C(A&mAha*U%!Mt7X+}Bnx5*rLAiw;-YaGqg={&lDCM5h zC*P}~rLxcob3#3@Ad)J7e*M(ykCq5_pMnvctqrlgsxwj)-`M&x_14vhw@jJ^1i;Oq zKf9pCH{WdrgkdDdlZAHivJ^Od>%Bqr-@JmTt(BV^kBszQIbw!TxcW?|*_+UMk_<^V z2bQHCvNHS(Wpa~Zb$;y8W!_mCcy|i(6$gCtN(UD`pj-fuHMta%WSGbR9UtkRZGkJSOkD=La2#<8B)r$a?_}c*u>dDb%yq9;cYw{eUgn zd*|}O0;F`b`{os)VVmoE!~hmoX(4>$a&shkUI;4eke)?iaeO7-#B#0eI%eu)s}gyDLEs%892u8z`}WqUG={^sNh{b`T-76;wNJ3@Ko$A zJL{`k-C}N zI%BTA+3fgvoMNV7SKe#}d-oZyVSc);NRAJJL{WTHxu;-|w=#P_11?hZcY-EEqm<9~ zPR;i72OZ#-)e1+%b!mJGWTmwmdxqJi)HVkrs~z2`G%8gYvN?}UWeXDWO)(n8V2#`# z&t$#Xx{W)ilHVuhO^h$ymj6uo>qN0Pm4k;8jyX2hruRiPM{zHhyZ-AqFD(*Ut>i>A zk94(tE8Wh>$HJ$@Auc$5LOK7hG*y{T`Sb&Vk?9nsMV6!VcM%ABELA_AASJp~l)0J~ zmQx1oZKC{mJW-eoPfn>tYzsd2O%*a__qcSFWLu|LCB`C4yedIFy{LQ%s-inNVs=$sl7kLU3zL-fep@?mvC4ss5Qe|Nh`gDuXu1;)Pkc9#L8Q1#4 z8|8!!@pESBM(BQ$B#J|p3~{*sy7rlu9q3;me=j`v|9 zlh*0TaU(Gja#V478HQb!97ady)|QZgrfFS{3s3IYgf-1sdflQ&(bpzf#YYEdu*&4w z!yxNjR?_i0v-l^Vf0QNM;*T0k*vuJV_vVT4vX((_g7Smj5Qd;TW|EUNjGPm$QRcS8 zgnD=&(h?N}K;OdDBhv-@YLoi<;!7OFdZE^UX4Spp+j(2p73Ax!8p0>rYZy`RgO`6T zfSC=S342&X!+up5EaYJ3ZNW9v6IxTS8z@`0hTe-m3%ZN>Xq91u46X1PnP<*MDdQt++Nh zGtkk1IJq5o$fPT$5`5RtBs-jABXG~nVx3kZ>)OeGow=EJ$E+coW`Cz*NI z-X;TPLRrPrfj6^e%nv`f&AB`rAwSBC~fR>)NX350le4UIRJQ>>@zr|`y;Q%#kRLgfC@!29n2^9gj zi3%Nwb1|B+caAgUF^pv;z~MUY+G$pBvy&0vAIIQKj;VdeX!RUJikTTPJG62UZ77bg zrSniFX+zy5=T5x z83lJuy}7=DN`X2$d3^4>!G*XrV?Kdftr!3MQNABU`N93=w&FQ8*32%Sk6=VN_L`mB zvAYy^k}U21(7Zkp4eF-{fh zBkjVtk*l^+ASv(kWFfT5%2LHs%vhjAAt(+6qfbrTPx@2*cP+vAo;I55*n!|e1H)PE znFA^GqDfTtC zklzhj2&d^=Fe$n5QwV}oPFc0#3LW9bxVe6U5tK2ff!ua(5ql>lrs^!eQ+w9WxyJb( z?(bVClT)m$rK%PHL~&vBkzsPOy)h2oECWX+!$%WMmoLu~oq9^iNF$NQM6EO{3D|tv zTNkgC2~rOY-4b0aB@xP7<0DE}zwAOuF3KZ){;}E#*VXB0 zUa1~*E(Q}?#q41bRJxNPTIk8k%ZyYZqFV362KBPHQv~~kE;_gg!%u(1`_r7MpLC6v ze_T2Nov}lV5D|aucG4sQDvIibd-=Y7hvbs2Vt>LDW|6xNN+7K^)2GQdsc7+)9|m%` zZ%!8JRatHDRTOLBF_7;kOb*#Piou8gK%z6FYAX^syP5C@;rxzvs_7*mN2b|(p>)dF zn{Fmr8q}%p?)%^`6PE2n7mNii=}10T>(Q(VF_cbt z6TC+yn2q;&zti**7(bdb3&QZj2E|uzWRnV+RfS3oIu-(5z@0X4=9<-M&uIwiKw}ll z`VhDpeji*Vnj@^#a2o_4FL%jd-bJ#kvgzRsDKiTr;mw$u-ohgKZV5DxX8TY|x;j(< zq8!xXdM9!h8NF**c32e;_65%2YGWlEI9#m5Up!IJ)zT9NQSTX_?YHWDNLx9vK_OSn z*%Qo9jf!ai0JNb*eOE`cjHUw%hR-GkN8_;C)AY1sD(6mkdlSzaxV3{RE&9I1DNnDD z6f;G5dc3D^w=!t9GpT}KIZg)CPfIuSXZYdtXJs8BMCxFn?;M43WG zP>CPQ40&9Z-#6z3J}aZVC>rmMI4zA~z1P=N2bZlFC^2Wx@8zZpH<31=2eVqzI39Zp+8kREQl*`s4UV$f+3+ zAn>OIDy8zR64)*op6i&pmJ!}=HB6le{?!SwoZLs=eJqUBZ+vY4@33c84h++v#hYLlbT}tl7xfq{w?8kw z1lWG8aHu=Az)prjv_IV;w1ptJ(2m0Gcc;f)wMAYerD9$IcNCl+Ebu$wC8C8ObB zsDFjpwV$3;=l1$^sOi15=`qi+YQUn8Q~wQ&D?qu%FmLFnBO^JbU{>up+IaP>`>${B zOtyP4%QlZ9u6xcwk{3Z3+Nc^W-7FYf6w6giEwmWI!&W;N(T5G74Wr3fhVIPK&m_Qx zwY`4to!b#i6 z1|E^^sQ0%3Mn!CY&gaRpMt{Zjlli88Z#YeF{?AOmFTe?gRCL`_vl{AX9kld}HhP?1 zh0J*jAAWSCI$O>x6UCvv^DC7lj-|@(x!l<=bL`0I+*L-R#U1?pGiD6JH!M)e>ExFpKmIm*I zHYx4udl6ZNq#2Rb|EA-Dd)sWUr`fbTfv0 z0s50xhD~jS?1S1hMCBZRu`yy(+T_=eyV;UB6h8b(c|BS^Of2L%Wh^NJXBaRN00q>& zoOUhhE;SXKJ`7(ZV_>CCDj7I~%VMU?0wsq3{?O6L45F!@cw>6E#L<0i@1;~bPrq$= zfrJWZvK9nI#+pG$F$$YpaKU#n?6~`5jdn90ml=NY7R<`Jj_Io+zE!v6GbIC|p?|UO zDjFs;c~%f-DRBy#5JB?6=!FnSt5~F5FW(ZSKh9%~=uAyRH>OoO*|2Js7KTBe z?%f0T1F))eq_~fs`VNjiT58Gb6kABJl;Bm1ZoNuFPow;vq*(%!hskaknk34l7bK4X zf!oXqIZl;N?@W9q&gbuFi25f4(2uzqm#9TE9KUBuA%xASfo$k48t*ix{O#EKc*0*>tYC}&#}J#WX# z;6jIB><%Z18Yajqor+0B>ygy&^9SGmfQBQg>_9=Jo;8&BS0Q}89&KnPsv6ANzGE8c z?!h0&FWGyKPC=C{T)LR4CpBB?&^)^BL!)!NKF(Qn_?X6gJ=-;woS4r?b<-JRVTW6y z)oanoQKEtE=`jtV54%5eOzZ0kIkDEMW zi*^kuIZ9tW@#_jX=v8yN9Wu~a*N}SBd~l5LeYGyH)3Lj98cJZoT3DU9W8@3Ph%~4J z{8`QdbgluYz;<^P^Gx>Ut9*$pkt1o)+3TK~6$l=YSBtUcF3;fp72!B;i=$gz90LUx z3KJXF<~%0nPb50c>}s}*4U}|bLd(SxZQM(w??5_Vl`=MwWaWJgf}Bdgu=B?RhZ*u^ zaFj0P{2S+RoYk+Z7uc#@rqg@qTFBzqBA2e$4h&VvRKv64xP zEh)Kogn(`CaE6GwG%<`a6PDDJ7!uU%8q_*A8m^Mw8f0n!}8Kt1jayp*b>uR`E|gIGo;00;a3=k#@S zE4X~tIvQ;hSN}Ga+kXcVNgK<&DPUxK8&{A(Jy45)ziv+;D0uG%Z_C5j^k!=7^H=3o zMQwg}qvOSGg?D*Q(bBw8QjN%VT!$iigF0cwtF5sjsT%fEiAX#2QSzkgpI3RS0usuJJ zY@HAwrV~5@YJVzh{_%Clvrs|mYnT^T@bxu8-H)$R#6JD;zk4VsM`rKDxCG~aoxwFS zxPXJYd(OH0W!+Y7NX8!MwRh@tzC!Ro=y zfw%o@@c60wZsUP1Ph$FdMU8~=!OzwLw|<<>xwM41r*#7S(Stz8kj{=?5)N*R!JWao zoFUAqD}a{z`AhkNRDNLefxn&|fwHW#{8B%uKivr97V*bVQc`AO^qt7<;u2x|u#SMh z$|lR}1iA=lfSkxHcLPq~Y-i}-Pw-9%Aed#@I2bvB za(v~)f4ilpuF_l@lU`q){dErf&-Zge4(bT5+0*4__tFdZA4t8vIR45vhIM3Y^6dpf z?Z)tFMIIN)C9iN&n}&`&!N-bC0M<82*f&et2MyE#e zJ4Ana3-thk;n+16i`;kr`S_Z6A30E!m5tN&3;BEaZyk2W+(NS3;K%ByH(OD0Mi)p= z)s_y38l4`(KQTxQsPFe-?|0XQc}l?V{!Oi>Ah8*+=T^V#t*5)>$L{q-3KH7O8jN_i zJG_*oR|E~l@Kxwc45te>naY>-JL~yNeEchS(p&kZm;BQkb+m@!cXn6d z9K{BGMF-GJJ@0ZGXVYX?``Y27)8qGTEiZ3jA0OZ8A<*}wKAOl>`|J9ZVfW3P9Km=< z0C!DY!PmLEioE;Gd4B=9p6{0M#S{4Jyx0x>9w>c7Kmx5T|A>NU1G!fFA+Y)DeALJ4&$s80~v2uyftvk9`LA(MR514ZYWco=umfxSSowpq9+O3a0hH=F5& zszY{oAKUS|zQ{I3jT&%~O-5DdTOvw*KKOsdx}E1Q#z80@sg*0^SIIL>gQox=T+&)CRsQQ+YR9p1osBxbXEgNFE&F&H}0%=Qjx0+xf4+;K=bMDd#2GEFtp>>+%((xC6P*y_RS zR|W)bv<@pQGB1!YT7Ih7pfTf@eof zXXYN2%C}{ z*>wjaJUMCRdVPjSX7LG57l5rm;}}ky86&EYy82kq*lzb-KG#u?0dB@v7Vv0?vQ_n2 z7BE*|YNy{8BCM7{DeJM1Qo+G&!jSryt7WxLoa+8GAlDPGrD>TjAAVW1lZ2qYlGqJ~p`k#aB*9Pt(oqG{`9IPeO5HGeiC zw(Nf71hffhB`(p%{?~^Ebl7TFB)w?f3w%H&wFs0L~B%N6!hy&*Nw+4!y`xSj?6O8*l zy&eK0rI^KGF}SD{cZ7!qYk`A@&?Q&m3fp691x~Hj zxX5WH1@i?(=V-MF!-p*=TtrdI?`DFy$o59sPa%8m>O7KMc?l|n^Dsv(vX%- zmmUD@K7*g3%V|L1g70KvnUe_Y zV`XRJ2+@k~2)|g*Ta>+{7C(|$@q3yf6_$YYK(Lg80UPHY#qi5e`oB2|MO(S=}Eh8^zfGDDE8J$k(L*p zpM3K_TuyCNZo&Fn77b>{fSl(o{s$7;(%-rG{XPb1wUKNcY+{_++eR?*8A0(!JG`8* zhRaFaC1|qfa5`js#ja){MI%~Y1u_hPB1g=99y;<`Fc;$r)S9g)8|dTuYCd|jSE0ob zdKE5teXNF1P}H$8%JW*Aw+Rg^DJlGmL6Cv37ly+hl zEC~zyM)pZFiW}PD>FWQxZDcphRzVNWeqYp#w-}$x5JwS3#m@DLh;O-8(_k^Jkn(Uz zw;HUA`v|2(}2>;Jg7APN6rQZgIFe^W@lBY`wf)+#h;5V&QqgM9erXH{|t5RX$J>K%ts}DH*{mQ;o^AKlObXsY>h^8>=hFMHYn3k7g65UN>wFabf6*gki4E&I(Drx zr9ugbjfcwfe0@)@LChI@yuAx(W?;L`YG`?#zhkCdFw4QvK79bDaeqNN&kS*|*=>qqe646%^?f zz8y7{FfZ;r9@KaluCgVU`5Fw@TYW4%R&x#X8eAQIMbf)A7+oa{*PGN6d_T=rGgr## ztVz(YXbBn)o1zG+4bL=KA5pk1^gIQ9u)c%7R+uOPieDM!ZH|^OYa zxmqcU^qV*FDAU=@qLx*~(?pTbmsaKj^S(M)6#<=ZuN6HbsN5VuA`NgAiZu{1!ztCx;3PrVH(C({lP)K-=`vkMJ0|=EjpsM<@rl z`50<{QzF~Ap0|Ax|@L zojVigWW%OuegXIX(&bwV?#%iqu3_cUOav5(A0wx>+SU9+$P*eSug1Dq0O1E7lg8UB zMgfD1okWw@XPgF5-6r`64+)40_i9cvM$%@v7vkxjF|hdRY^69K7DL9@U2oGT!p|s; zGngzii81iH#;`Z+Jfe)chEQW$x_hAjxEIeak$G^rRPMul*;bPWXa{i z-)LXp9HN3t?$&Pel2q+Y7vzmoIldwFdCi>rl)xTVU?%FI&7@IS%ouSUFlo5hkS9u- z{IPoqr(B#5qbe)WpJB+_G`*<`s z(80>7++WROZhBk127{y0=Op{JveQVgK6_-QUKeJ{hus5(*MX?HxO;WC37wXDN@-tB zlT$<|TtZ6{D9#V?QU?PJQnJ&02-XEjhN!i&lB@ep!n;AQf>CDH8r&&-fRMG%@xS575=5lt2$(V($ zk+#}&F{5$^w`-n_46_WcpynjbAH?i9X2NNDnLQ)U0k&TFgE_MgLp_DHVqSu=SizKjM`|&oWJ%Jl+kYLP zD#{s#d_B`y*z9t{qFx5W$lTjwTglgj&Do^Ho`>e@5-#3hmJw}tt1itk(Qf;Vn4qDQ zwVm%Ys`fCHj}!ej%p;?rZ6y`l)}5xio1{rXK`{~c|8#EbL-2KMMGQ_7)sYh;C52NZh%49H1vJd)P?kNuU znRG3sSN|aySK5&|G$-If`DK7SnQHq9u|&#sd4WQQpOhx%|4bKD_BhwjesR@GCt*vi z33%u37+d}q3Y|W3XNJ?c3X|3edZE9)_xvfDD*q-*hS3P3#=qW7VySXr74Q?ldrCr9 zhmCbk+hIMP9P|wV4M{F@@`z|qeZ*esMmev&bBM_DX5b|AbEM~*{A!x);v!g14fb#K zV5d09B!rVY>|sY{Z&K5+u;k=b`Q;y#?v03_JQf4jcS^}@*|AGQRzoV-=4HqfS3j%W zXFOK9ifDj2pI?j-!bk(-o)vVDBIJgs@S}u%cpNm2?{8jp>%&4RIfnsa(COFhK0Gjw zuNi8tT3l(pJ)QHK4i~-@rKK#17wyaDp9tPV&NprQqcn)N>!JyB@ydG&Y!=#V zo5)GU-v!5vF4@uzy~a}AiC*#eIl8t-B|#JK9(jg9Vu8MAGl|S6cUn5)0Sx9W2Se?w zPj&CCpJV1?M;{NlpHHBbRb}_GSq@I-&WH*%tDSmz4v;tQs|sV^=3K7wN0I5bf1mpX zdx?Z+>_#9jLPvQ6I$(&An@_Fpv~v{L@4*UFM*{_C?e-7Q0Rg-qX_k3u1_!C?@Q@EqjHsvmAzx zKha9GIl(xDc<@WamnwUTn&Ej}@qAjAoIB;7I8YzsIT}qBb?m7H7yN?nx)CojS|R^L zmZe&3twyZg=^=Hh?mU(gi)=_t+x`6)78!PJJ>KUR9wte~BHebEB9VD1?7T$HQY>5l zkHv51?$LNiky&fF>;gDF^E3Umv$?T>^7!1pybXes^sJ?ubR+g|wN!8o5n;ffgDGph z7Kspfs7$x){4{ln_)Ypth8*g!ge`*zJva-`_CkO*pK}ft`i0LOQ2n%h!-Rr#Ay|F& z-1ntLx1*Uz1bSqpTULNxWye#nhj%pU@NEUF8RJ9KSW0g>qJb&E(2L_@riwwO;>g7M zWNkQ^m#aN0yne~RdCq?9{ACjD*X+=%_ns#sC0V+0g{x;wppzA9?a(=|jWruZa~@u* z9qhObM^G*48>vk)*(!08cZx+B*!SpM$Z0HTU4@*C%b$C!AmmD4a~^HjiuLoYbFZnn zr?y1Hx1uS9D&1BWDaUEA{yxeJ9Hg$!V3#YQeDja^tyHSwD1Fx%_Ec(kxq8U?CX$ty zaXuRC@=on97?>!Tf|*RsiyLQ=8?^i$i^9@tGutNBwjbTMCcoyGKKie9{cte-1w=Tp zL4B6=ao^3M{hPPqr^S+taO*O1g`>X{)Aa`rII>6qj>4G++2v?XNT^m7Kfc+{Tz|rF z#Ggt@b?NoKrBA{RysbEm;2-#5wg4(x>jRwJ6m>uADXUD0LV=LGP9PV@hPdNl z{c#_&lZnj;xI8Y6e;3A;@K{~wv1(CjajWzo**>O)!IKABK_(vBN=9*DlllQO=)A?A z`K82z1o^;V-K-~oHtFNgmYFb2T=v;)Cbm@Z>t|!wTkmtA!}_(0=%}bkJv=M z=ee{Ebk5v*;Wf`@py@J|J_5A?R;N(Tmpl8|Ft*8Ni`T+M>Db8S&$6^q;h85{@cfmU zMoGyD|03QX=mCoy6}^FH(0JI~aUFv8ax&H_1rv#`2N~dAftF)A_N_)iydd=Lh2j}bRbeFI z#`r$t1651*%~t&e)um@ac!V$7<8(#@uT+rQ*k6ElbG9R;s-()!!z@_=bLuyM%5h>=uM}X>$hE`KUd1gANNtUdi9_FAYJMu@TN&ZQkUz zPZAkrnPA}5`yReAgTIeOeghQ#;*prKBZZ?JLM0?#J7k97=i^*&uD$xJIA;B`lzhnw zQGi1qS`Q7xg7fO!YOyL9J08@WKfxJ4G=`Ohq#3Z9J@PLY5?U78cI-YyS?aJlNH^0T z!B%hAAAgS7P76)H1`R8|rI3rgfdUh-={j*q zln1{EhJDyH1r*+1pXl^%Av)=tx2Ma z^Q1ug5J-}(J25W4g@GX0a9Dq^Ct@uDZ?|)t5FY5&ISNFPXH!!coa@#7r58`&fjE1y ze{KTPcE8i0251AbIL-&&uC=Vek3-WSix~1NR8j_3i3C;Mp3(Wr)9kiAejWAPv6wT( zJqPBmyD(EmLT>9;#ko|JeI55zt?6dCs%Rm3_eL-g^UCDE1A}`q9)rzsjpi{$0gSWxjnlKF)C)!aB%W^$RfgZauGY-`o*wT%RV&6-#K zNha?JFqhx)FLUKctWg7v^C}L2&KptUzj4mADAD588~WVG^;hPO68Lj3_xAl~4AXzj7Po(ih8Ec}ELJ*EPD5!nlRtUMn1glvZu>7cDd%wh;w=?zLQi~=izw}gQ*Lkl6TMyqzC5g`ABVjInC7Q&fn&j4 zirlWb1ycFOawdLbk3sMEvj0e}M8$iwTHv_e&y^r+XL1W<7RE%KsAf#Ii$){~?bpn{ zRYJUwK<+UKDh+0JL*ICN#hpBCKLjrs@5jB2HUy#UI-nSroK}V13Zj)(=d26yiqB~8 z$c+jX+D*_Hd*QC^3N-#RNF+Ba3`;5PmnV1QaZRD|1pVR4>?w9?%6o^!b%niJ9omVB z3s2-ie>wOvc$^XYhpo#*#(6nWJ>|G%OoA5$c1>1U_1yfyzQqnFHIsleJnQpD!oakm zLKi|oMb*^tA?Ud{YV6f9-OiUB!}NgHQFl%2OLB&n&s!5)m$G(M7c&XikVn5%8v-Hl%ADH_UkSo7Q4cv`|s_ zoxQN>@v5maYN+zV%Zs#JQC9>&av>b9w^T{k^7t=c7rKV(OMNVD0qH|N1s!@DI?vghs=`L1vwb1=hR?#ls5b;ROrx`j$UVsJ{ z84f=(&ZbMtl9?@rAmF%sz#`E*&lIzHvb9%3oM;$HNP>E&km>ZTMOTw0(Hhp@figuu zC#dG%2G*hyGL!}p-9Wton%vS2+cl!!VDfQZ!rD>uYGmJ)k{WQ(rAG1*Pb{ekqcsXO zN_q19kf~-9G9BNA=(JFX3rhl-JG;p*>Y&%R2ggq@oDwb|7 z#4R?)aCpS(SB1F2NwXZFa;1kh1M5Hi6r?H=3pF#wE6p_8DHCiW-&w} z&SIIM`%i*vg$P2`d-NKKznkIawS9|lu%5}~CgIxDhrb3-P08>QBT}9Oy?dxrTUyI` zDU3qQO~k=jaly;I;wg7vfle)1a{?3nj=@k=KGSzeZ|$nxr#UHPN3WW)Gg4)tTUQx@ z4)b4)+gJigX)+q0oAVKDLkvP_tBlmz&LgO#uMIUw}$m>=B5>F6dAEMN;A~Tz!rE zi0A=F)J6@~T}f2IQ;}g`Rs`JlazQF%Q?3gH3v7Q!{2yOGbQag`S!egitOf|!j7Rp# znh0?6n4UTnsUhlfD2`hJb{O>l5SbnwuJ(4+{`xiwQj|xZ5}ZVAq*_I1l8Lsc;Mnf_ zF^5GRXOj`uLy(7m=Tx_pkGXHV^i72@X{yIVMJ>)g%u^)Q(bO7iXf}$>T=X+7Bob)V zsCx05pK}od(4pf<&>FKyX*xMoncKcgrLK{k9!3ADkbKAL$VAYZa>4Kp2EX-fVYhdn zw0LL@0+Q4X8!JG9Ba7{b9Fv1Mr#1=TTGOj5%4+sueYoPR& zT_q>lvDIjY3<;7R+}ZpP)6P}@%DC^yl<{v4S)h##@una=n4|zPFKP4m5~`<*!+4qg zC%EV3Csem6estD{*}h%WY`%nHl3HGq=4>$|51na?lz!Onhqb6WBQ{|a%$;rQ75Si< zx)fGubM||m4!w$NaL+nIBT^i3vJ~S(aJ$+lxjGNW_4g-I_W3vS8a`#59UhvvQ$_Ct zo?>E3?@@X*Rw?!J!ysuia*X{QEhX}SHS7$R;1JADrZU=|H=3Vez(2`qE$+P0slFF} z@B|WCav-Yq6W3inGyz3n2kWAm&frMUjgu~GS%*Ma*~QATyLT$#Ev%rxPIuu%UE>K) zd;1cTAl@BO6O!`JwytOLf$evsIxCRA#YfSAArV_g-;&o-%Q1$ z>b-(7jf;-;wROnJgMKyaj;tB6iq0M?t?1WvQyYC7G5$$^N44Z9E4#;pg3wUTJ<}jO z)HxpCzhTlMu{QPoQ$T1tcczRPWsv>{dAJLDg6FYoTzc4I%W4pwxb%lNw^9TQ`-QYr zXTcbe58e}h{0sf$EBaJflSB|GQ=Bk-ekid*`Jgx}=HP^jm*m-eNdIuO{kU&O7e~y5 zLu%k%Ay+aUJS$8ge6q;dcIyzX({w>>ANTWLsGFii$fp&}jmOzYJKDQbneIwzJCyRo z*jc4!&jaaK6_Z_P$DwM@wo`16WdZ{WZxX2^d@(fI!1q|>vz<4XSxD} z5Gd}pkJLi|;;rbVObh}_fzVqgy*0x&jM*4JjMcm5P*8%g?N619uvGM5-+~M}k?hql zb2mLb+7cgQA#B?~F=D3gcyeWZ+lqM0+9!oc()oL!CH9cduksx8Z3lQ{I@He1A6d7V z;1D+A;(sF5_U)7u?ycA;3|9f}EpgYT`C*(=-i4^9?$%x9Cn=>E##i8er7GM9Oh@#od<(Ki=?EzTE0-0l=Ks3L-?Ui_*9Ke>eR`rc8o9?qAyWF~BYlH~YZ0j!5R zOap?g)_tlwO5CNmZb8qj2jx}dyyLscjV>DJ;YNA(Ii|hV&{dEYCI@FOt=FWoDUKpo zy%XKI?6~%c(-Yt@;wjcyhNSz~uuxfl9}r&!hLIi3IhqqJO;`NV{1;WzciL0EJh|it z&p?;k5+GR%c@Z#=+|&LoC{phcoyV#zj|$2?(0J=5Q$IZH52l}u?>4j!oqb@cE&rnF(~T?)sZ&t=FkbdlY2E5GkW>b!8rbRy zryb$8IL&WnLEfS^zBQa8Q2dbBZ?brOoQVr&- zfNhOPlx*a`HDHKr0B+Q7peCc-0Mk{oi!M2o>^>T=ZNeIa|AHy3fZ&?%;!XX37(0i+ z%z|!<#*ZW zQevfSkLz&o-PKSeXqax#6=@AW+h)B}n%+%{*Ksn*CovC=?cC9BB{RzY zO5487rO0H{Zcb%_ULk?nOrH59Vk*v}{H8sF+yDCFCGYoMM8SXpp z5xD5Sj_$QLiURKRvc&ADyM*4Yc!--EWgB?ZSqaY0FcoXYD$X-3hsuX{nP|n6WZnFT ztE03-aU&m}hKtqIi2IqOl6$g0MP;*Nt$G!}H4eJI@@_@1K$9;t6i&o@IZ!K#RqfrX z(-hZ1?O3bjC8~!{i4e!8fHbuNqhUYI_t(%T5fW5rf7qe}gK!f;fbOuvY$ZHa)WGG+ zYg~+&DSlEw?)d=g+X=;PK-OL6XAD5YiC~4Dx@DiosiSJ>d{CLJ>hVr(A@GdJ5BWt;T8oXZ z{i0zbsTTHX&IqmB67c+p_E^M@cSda~;RfqsE?d(a-6PIeG!#}q{`*(<`IOD}04Ps1Gg{C6A|)4^uhl!f&y*7Zg&oyd zE<(VV!w`%Tv&G$G_jFLh&a7Sgl><%gu-RCpu;`#mxH)fp$>CZsGc&6R0J`~5wdoR6 zpze^}*{Vt!bGmVPaNH@lujK=>hL@Cwp`gZ%Gu*oB47|#KzUBzbrvrQC7_CNI(WmAg zy!Tl53r!lC2LAavRa;M}nVw2`E7EUS_m-6A&X2%)CT+kJk=4)mZz9NjFK!h&&u%`^ zZAEjO4mztNCouJ$rK{}0Rlt1I#lXMHzbg*p2Q#JylNK&;rc=eAfkXGqpgR`s_Un5p zHQq>Po1AI+@Fs(Cm0AiQkwWRHPVf%@FeF(;+9=kI$o-SwN88NZP-ZSGF_H7L(JQ0j z`U;or(R0R*R&UcAp#;wo)K{{m38LJeW!d8DSqq{~BW95e=-PvUAkN(A|IW=t z)~4izqaVy$uY-5UlY27>&}*XMvT&9_5LQVwwqan4926bb;(rCJ;stj2kTD=_kexqX zo}~sN@h61lWhr+c+}{$r?tUCHOpQ3AWbLqG{373U@T!a6rMO`NOuISs@h$@({w}Td zJ)S?cM{mtNqVYcP+Krz#&~)_8Z?bc*v$6{wCO4|@vU!4D zOs0RGdQ}2FITzWImV){TLw#>SunSJHl+Wcn;jlhGi{56sO^|kO_l({kKtEYxx`_Aw znXMj*+yV49Pl(2N7n{R?%v~-4pjLDTv$W!Er^OLFzlSS6BSn;W!%Mo@SNqqZB#Lxd zo^(%if1f_=Z%k;xKsdYT5xHQ#u=yIh68(!$xiAT`Rlj{4()8A;1KDG=Ec}<%-~+j< z&@bG{zhW~s5oC|f3A+kpph7?6bF(1Q;kDS|UETBsIl{egC0&AtfmJHhy5Fd4iPnes(3}^i9%E6GN%g1zQac7xt z_FW*$aN|x3^8)^h?L2aV17{iR+qYCMU@ZLD`xI zAaCvt$ZAULyu8YL##;NkJchi)VNkgdT8yt=;B92|xP92}Z}y&Y81cnrIMnLduuQZc z1eMG&cDnNQuf^YwKEFJzD7OLpk~$@ExhcQLABznQ#DGdWJ{(qS{k_u}Q`N`#lh4jR1m zdlT3P0+*6fk^_Pw8$<+wMro+S?q!5p)nRAMK%T^h7a<<{!Q}75hZxfJ3YUe$C<*(g z=;B5Kszm~#B|+#0AbY$D{2A<7C8>zV4afhShPaBcP>2k;#0 z)Bs6q&;zMTUh>}DIQnM7PU2g`3WKoZVTq)|J_~73g5<{32MsXr{>b?&%p(Wbp<^Q5 z{xc_rwCBFbT_enEb^-x7hLLk21qU{H^qhTlKtUUG>1+8C?ob@6dtqNa*nt@MUU!bZjen@YM82hi z*VpqkHSVcc~)C{)mSWiIMJ`H`Motvs~XCG)~RweHFZVd(qC14!fjnr>7LiUz`iY)vBU2F zynnG??ul}J=m@U5wZdSL>{$~qC=K->RYtrL=77S1`uYmpz=pv9>skY)wp11L*4#tz zR@d$b+ydC0L45dB5@~t~xWFa}tq}*lkw5bCvKfGew?L0?UV1=(%#Z{6fVf)eAap?; zd`ZK8%A%bHE`L^6k2(grfDr{$ul;~?b$Wby)O62`LjI6=e*67w*ZBIymzCw?%;bdq z;3uSrxPgAYf&~G4`XeC$!9fH7z0?7|f7Jd$!u)JNoaSgM74pFN4*ghTddU2kuRQ=j zIey9mU~hAV8f8`2`++ilYB%w~fYleFXf1dcRKZR2Yk%{9voB&ji!`M(aJ7a^_=qSm%2uKr#v-V)J@^H^3H1AX z+BPz6;u@ELj;#Lv-kT)s*#lG&?~0}(z`*}J+4ULw@&BL?z)UVM*LJZPdoPvjCo5|nexLF~a`tF=E{0SM`J4i*b&KUGc1HbmsU-^V0q9{9@ zhE&L-$U(&@c959uo87b?xklNV<|`?Up#3qeneoKw&-9a>8&<}5zs zm>>~0nfNMi(vJsex0Xi2>(cEc9S-a!eaeaD>A+CQC#`3g-MAN5%1*o8g@nwlU zRoUeV*d7Si40y&k&C_bvXi;Dzv)Q0vClq)*g-(rCsqThI0JN01M;>XmmWv)~M#Kfk z;2uZOw>^UipC;yqKX|h}nBMVv$ik*{Psy+B2d&YFBu~G+&O?mfOjcG_7GD?}1^c-J zn{z8Nzi3vgMH`yN)XIC@u-&>q%6Q)17BgqfoBx)Xg;jfxc=J_eNQU^EqOR6N*jZTJ z#x{fB)jn?=huO%$fIcY?UzqcTQi5V6Z7v1 zk&Rd#kIwvSsPUFXnD4aUm(>lSS;0RE*w!yg1DYZo06WaHSsKy?){s=mjE{BR>nW<| zkq1Xza6WZP0Go1H?%l8JXf+wpes*Yk3Lm=a`N9_aGA+A?YQ!gyQ?Yu zlF-1|no9*YMh8*b4T4+I(m+s5uT5u;6OvP$IZrynJ-<|5PFV~Gvbr<~R-xjWaFRoe z?QiyFUa^0%lN*M}-#>MxjVMt?mV&jI%?b;`X(S0`U>fMnYoxm`ve!r@MESN|ZraP_ zcE&FdOS+$#8C=~P75foZ!Dttquym^@LVu=drVvD-Af+(D;@{Nf{ScjN-3Ix zMJH5NQ;JYp<{>v&&cFV4J%~EQTdGQHaM4-FTp9g}{+RVNZgm5vx3=E9&MHr&e?Yb7 zzyfA8)eccWlcqobOb=MpPkdHY^n`ehmDw2< zO{*JBi8Hv)4vCBrukGm;S?<$Z*V{9>+RuSVy1phVcM|)mj(2LFxgMEz&KHLOvEMD6L&83ZI_YK}!2-E7TEChs@`AgNSaa1# z5#{^u`t!W6zjo`;;WKjBxrk)ttV+)uNrE;ANc8x1S(`m_d3npw1opS#`e`6r~NLXxh$Xsl|-J9*3g}#GVbD30sx;X=f@PCsHI-twl>UGUEQF zL?W`|sfpXx%&GLI*3enr|CAKln=YblmoL#2ymnCa207$j4={Q=dib2ENf``73x)`R zdd9>KG)r;$Dum6mS|x2#m0+7ckJHdo$-+Wg1SQm*j2)bro=9#%fg3&fhXVIe5N(DT z%X1#hd69*>Vw7gH^PEM0&lNu(S7+nsa_+Y=(g{t_uH zhTi>ZHCBg@PREB>!ZwAC$wx}oqeA9}p5ozG)I0uldYk$1NJ^TGJR|;B@&y=k%X70D zS*pOM|U9s>;vNEHH0;pUjpY6S4P-^DXxL7Dt0Y-fbOrXwDRBzQ_X)yMhS9t#H7gSr= zHyrVGsgb6djfu9}rpR^`jRVM=9-+v{$3jV4V+q+&|ArZrBtq1$sQv1Q5D_BU%=I8- z_vyqF!|gunu(qM&jFdyD>@!)b_F@O@BgF+}&Cq<-+T?X6;yjs1iFdhYYL!C9dn^N)DB5 z!e9ar(g7U>6dlchOgi}MYC83mY0Yeg?47@S)qsYRm}*N3es70HFJ4k6(?vXUYWUDqK)+KeN3&OYS%;i<%Ui4um0V^{(6gJWfmIkgqgP~u z+QZ(&5H=fI=7#ckwYkx6d30AB59B1nIETIn9Zus8v#WSYS6;qKctA%b6nY9r?-@?) zoc9T=T8G1*YM%ZeZ~g@@1((4j^QU2%iG*p0!`Q!c%+g|h_)-~wF2iypg^vU5jgNnF z-IX-+b*J|z!7E>|ZK#1NjsZV?Yo5CzfPK0oq->5q1Ya@rG z-;4gOi9k`rDa@ajulk*hV3-FF8Ei^lsyl z`U&9);Gi?zDpg2v!=$&c8l|Z@Q`PC`E82gFnr_9F|LiX(FZk&Kl)dAAE8SSabPC`W z=L*~b4|S)_)tbmlY2c}}D{kzNSVEVy!dGU3g7X865d@+)ypr|t`nx6& z6$~2{Qhb6!ngj>P%%IwnkRMFhB$dP3I6jqK@3a{kW?QNT5+cc(p$K|4QH*ao)1vLD zZMxs*Qz}dL1yu3y)07spWGlD?45hqndY`$r9nyjSmaS82N^`!X0($Kxvy$=6-#7Sa zRU>wn_glTbM!vmDdvV?F0-uTlOi_L2p}D(DM4y3 z{C5AAis>eq^2mlWwkz5Zzl2^o%LNzqyOFkg~RosbZ< zrdMpxOchV&C{g^-1_BbibeGG&hNI+U)ox6im5u6zxVZk=*4asYn1TiV9mjDdbStvu zvsu#1z;@B~kW81#8z9)^qQ)iRFQjms5^#1sICJpX%Rd8GGL8}q?8>Xj$O;*+Sk`5s z^-?H7fme6)q61#4TI__|%CkqBlojV(S?BbU@>O&zp~5&9l$vlAKHtaYeZ7|bq*`+# ze3iMi6Gcl;2HPJ9|IkyRQ4wYM(7l+Ogs4=Ul$ zR>)S$`N&`rO+#iAZXE~{{73zn2$j-IQw^4L$r!Vu?mh8Ds3VK( zacKzMe2A5})P^y6zPbaf&wRMvrUJrlwVmns#39R0)`c66MEQ95 zAmH>irhnIgohAv){X&1cl0XTXi&&+=QG+4t#-uUHM-)_;(h=^cx-QjobKINR6`7TS z$i{7XI4lJz<(CKS7)&HKX~DMFY%@cTTqX?3r~HcP4%^W*SCf9oKr31gAIHc&)8$3= z#)m@oKZ-myilH6n(T8vAHuI2SxMWo8xfjWfVNi@OCTAVy#8}LM=Qw&Ig!grs7kwu7 zT{T|Cw7g)BA#=)=&F3zjo@mBsqIR|uzeXlp7?Y}mp~7~mtAJ4;>60;U-`tzh>p~(J zC`i+S0jiYvHOrj9L`=7im7dH#l}=PJjh)hI(Uzs&G*a94ky4KeQg`QwPzn!mqV!#M zfM6i_*2-l)_IR#29eBNw1J)fcV$+Xw7TM?-(DtvNu%6-xX)jFCJLU;iajxobpqT|l zwjbcDr9dh9-66xo@qq-b@VJks-2?KNqV<%dkzjbHt_@FMjEB{oAZ3PeNDp1KgO3nu7dN zTP5K0-oWE4P-kFP(KGDqC4^-Wf1N8@k6e6)ci`ZjZHmgYjq0eYzz)6mZ#Ga4ygfuD za8dilH9SQeL6M_lMrz8F^c(Ft!jd3ckCLiZd;NvLdco6DF)gr~APp+~b%@z7Itzk@ zM#i8ROi_pU9&WnB@uEvYG0L;NhkRX#By{cX!%c1$U2U#X{B4b98*y9; zkMqz;Nz{{V8jXsGxTT~qaagaa0-*w|LLXgie2;HSSO&g?o_G~`ROKpJ2?uW^@7%M} zii~*n<0{U~JE1v;VSVOmMxXcBPV>q}d$B72kE%Xn{4$kj*tEY;9d*b8zeK=*9mks#eo=V-C!oY%Y@?v0oVR!SGJ6E=Urof5z zAq=^eC;TUZj7>j|niWM5zjMeu0<$HX$DyV9_KCW3Fx*zCjR= zmsccIY6-v1XI_T*oA&%;LyHaVg?xBpF+hUOtJ$SlCwj=+ON!eg*2an= zZzxYP-<@c77!%($;oy$TsG@e@@}nOTbS7go-Ns9L$tTY8@K5TXO8(g0qJ}fg?Cs8k&Z9=i+o!vhV_L#AD?;?%fzFjBV>Ij=YKql5i;|`S2aOrzU>Sw1`_^ zw>(jg0~Bm|db-rQ6fbBdkJxpR0ZHJY4SF|h?C z6A5jEy*;0NJ|3o&ZuoG@P`0Td|7*wPrrqv>Y@2_7)fS`5enO69GBJV zpyN8SYbFrs$0pIU^T{CK3 z_6n?#5BdGyrXIaNhMVHkHyM3$a(T1sQ1Wb&&FHFK{OW$I(JWrAKjfLa)_)0c(b0{? zD8<*d)Xn0wU&^4gp6efHQdx}HIV-Y}3*pOmBX_@-th-pxB+@m3P4ZQR#~KeFk6k+bIaxH5iNAEx2nW{wHZ z72CGG_=Xb9ZMb?=BZ$Bx;7f09yUD8V?ESHOtZl>3o-IwH-_Fv$8RFz*&0&#iJ*#5i z>xoF`@AK%jqx0EwlX=7CF@4RvqRIR+$+L3FiHMouSC@+d zX9cMuQos$?_otu9H@vz!nMoB32?GNi1p|Zlc%MnYonzqVi0GIp$Y3+hG;z{LR*}Ir zbcD4S=C0NFJYTMtN5%w*7RHy{us5Emzuz~sw)T-vh~+PYY>+Flo^LghEn<{^2dYC^ zu9g?0s0U=sMZmWkm}{OVkgB@6{H_c~?jgtsQ63UnF6^DHZ--_MRM0>1yD25g_1=dL zm8%gxH@Bt{Kaq`%4RL5h2NI3}V1@$7m6Qh$h_sap?*KdmxF-YRk%<1*#~qdA7(@~; z_S(lSat;d71_}b<-;G6-67R?x{~*l2kFodZ&bKZZA@>GG(~Ap873#Cb z(MRM*I#B2fI+#Dt6jKFi95bvQ5=qpABe$lS^lHlCc1I|*f$RM@!9KXaZ^6rI zXdm+spmXtHpu_f|9$d!|{n%=P>p$0`Cp%L9-X(3_Cah?Z;8*(~B5tjyKh-kmVSa)& zfNjjj8P6kbAL<}jg^0T3o>lw|M4ulIH53;XaecFtCObY0Ka$vn z<$C63`5q@Z**`~0%6~tAy#00U_+^DOv4A3c$$`c$LVSOu4S@u{wMKmg&GV2<`Oj0RH+z&AS*{g9t7^gNGE?t4HtJwxC8s+w5t|Fmg8$)7o~KN<5rCT~6wUp^^? z*JqcXA-x~$2R}sonz7T}J_^-G7iV@|Fzu%oLHu7f<=CG!;L3lw2EJQlCP3}Gki;oJ zEaYJQhXMnKzWd{@&;MoalR#&{POlPdK6^+n7gu+$%&|LcpYD9$TzNP!;lW?pWXN{V zK>mIFo?$g1u2x~7-n$o30exQIqi8_Td9ws7#Jx^uz(62T#NRxOR8&BQ-?;Ave!@Vp z_?w{jHvD)%eOfgyuu&jCxSi2$1fxFlLV>`sGd`)hK!+bApEde<*;tcK>^_I~TVs6B z|8i-LO-Kat>`(vTZFq-D7z@vR+ z*N2sJJ~K-;(WpAhU6@Lco$J#cWIRX&nbj5|V`?=}ZVB@fKbp>}52)(Px7N+yWYNGI zO4h6x&YyV|@s0}6Z8vnVXFjWmC$?n0<&SG6MipAf9brv~uAzyu5))79GHWr4CX@g9 zy^H4RUeq(MA;Up8X3~~=ek3n0(UEmNS(IpfqclYgHdl%4mUt&OV}0tj>HBiOY@x7@ z)U8ztxFmMrIV}-jkDDBBm__Lqv}JNQF7AssQ$-buiZu+vl?s&55$rF??t@&ZSF&~( zvg(H|!9vriT#if$m+@c4{_23CTc~A9gH^x3JX@uRjs>>6qYABP(}_Lg2SCgcS35b@ zi=;+6-0wk$W^b)}U^%6DrPLybHG-IrqljD@)-C@^`c)H1k4w>S8PH9WeWV#CrfDsF zn-Lp8CiW8d_OS`WBi8w#L0(m^s|@dxZFY+-U7O69dfjwB?xma)Mgo1ewn8BgQ{-+; zG9t}|{t!idpRo8)Mi2{SXV8_nl%o1z>uhf}NG_ACcxJX~Ov1YA&N=UTvEn7{bON>_ zInX`@X$1Y0%Bftrk~dXj_hB}lh_d1USShtcBBVw9s?SJz#u+uvzVV9F?OV>L_heu| zTGdv)49t%Cxtz{N({&=z%4BAesNkUE%nKMA=}jf9U7a`rNpYm1OBu@^Z9R?s-oE?$ z87%W!DrTjLwg%qTZaFD`TcF9d94eV%Tjk7meEx|0&lSfMzkNeqB_v!28EQRjqChX` zCI~nTFgkp>L0&j-e57n;INhBI+k`zDreS!Bd)I92c#Cymu)s~h@CcXIa9SGrwMCL4 zL8;n9bP@rPeO;)&HJqr2&!!a*w)Q;KP3IVV{-{ye@q>oVT}U8SbUSBw{55K#wpMn} zbS;zp9?+k^Se)PKXV%~N0iYFY+wSB~mAqid8hvz{L+<_&U2V@sZ*41o;h^%KWb8xq zkrvsHZj4$Cq7$Z{md0z8{i%~zyGJHtG9(&An|>4zMZXY0b~`M;O}EOSii_f!UY(mD z;9|!(sU9;quY+$8Y?sG)Fdsh&iMx{6=#CN$R~+})V^ZJ}_c*ks{6GkMHhEf0$NP3&_mu`1lN`|Z7A z_T#Vc#wDxih9U`zmE>((vpEPO=<(4Qr0#FRZ?q|IB4V=B)95s&xZ@yylgD)2Y=yZs zJW~vK7mrctrl_KnVW?_~ty)IFf1qncV|Ui0Wu`NL0FXKU|7*O9#$1QKV1U3JJahr8Nm^15AQsY;3r&K%dESi3#xThvv+J8o$c z89*8AnGWiFc)L(Hl0V)dIE`|cGj=T_TDQa6v)9-=ZB}M22in8wv|PAi(t?zQo^1j! zz#iP^sWWv*Tm3x6s;MgU`yBy3!A$pr*le`7}EsdrlLqT)Vl<wrGGAc->1ehu~Onf=hFJxx=pP-LKD}jyILiA+& z0|IJuaIFU>b`~4BZ?}e+Cz}n>yiGO>?EzQs>S|Xga;vnbCprr&>nli}p;LYjSy*6h z>0CCdd8@iL=~R00XvW)m9c5fu&jJJr*MGtS1$|3x8YW~(9v5C!a1MgOLGhDb`-4wCaER=_Z!eJrkj$R^YVS1^5^33~ae&l>X@?-j-FBVKlQBHu0E z#k{>_+xytI7Dy_UkxPCh1h?b-ZO$1xzcqlj!ohb?P1pBbWT#$crkFR+J&jc7VNhDu zp#LR=t&lD%Rh>PkV<@bZd~h%^gUVj)bp^bs?B7IW7FCn$G69+4$*~dg`{{}$JxSXN zvi>t&;K~p69C597vfa}NuS>c!+cIH>y!?g)1X?7n^qF4CCkOVZ`|&01+auAPM24Dd z=jI&83n~Muy|cr&{#*FeN&*>y%fE|4PnNOL|NC>#Q&3_2iVC)@ndmSgmGE zK6!ykiB1`ah4qZ~%d^Ey0Ko~4}8KMKJrPa{9RYQm&rG|as+Ug*^k3D-?)-9Cr&kZrhiTd zYR7E*a8|3cnjM1-@q~dHBWOh*N>)4Ov^`Cqjhdjv9aaeya`QjDqO2<4Fl=E_wT6SB zcUlX(Qpr42iTrESi%~eNe(d;!^M`Q?Jh1f!e(6EBIz602Trd_ppoRCjI`t;jg$t zptIz1$Y#!pL%x|wOh{4$v3wXlePcZ@%*WE}clh@sErT*;#!-Uy_NLfH^3s(GJB=Hf zN7D2#s)jYj_bJ|`!b}bAPq3FB`xw0Vjj_y3`{NaNbp{5J5R$_?3HQCQKA3usPH7=H zXoYGstoaaJ&g*X`62;C6gAmWR1h8y>TXpmiwD3I5F&B$%w)MUE^+$*R4TJ^g(dYNY zvtneIiI6Hvx_-|59X5EwY#$5*wYg1^vw(hy81_}Jc7hY#pp@tld%n6fS?A?o2Bf29 z5|WOdk5??PMs9hIa>thF2`ro0k{BPsH*i#()y~B?Hg&URz>=QwqPS(9H8cKrW3_A_ zT}us)sM3T3jW6px#Ecnn3Cq2}pKV5NUR^^IT=15kCF|cNNJ6L@U0W+@!p3Ks1FS;j zMmkkZr?Yc*x-~6vBq`=Z)W;K+^jnjr4^kDw<^?FKR}b4~>nQewABwa5GCDW+5WBGj zCxzu+aU~94+ow}}K?wGWJ?_|B8K63uqXw~!cs|=6IX;84Dph=XGMeh%9$uo2CW;*z zwIOoPm^?X)^1m(Xa;?a4y z5|sdT2c34(sFa+~7-ZXIMG?erXmmQ{E#^!uE$iFX0P}F)J_?E!;Eh^=5uvCOF|EXWUMO2i$XFH>NURbGE zHQaDSV8^59{iT2iX}G&$2z{7CaSaX4q(NG#g?n>AwJavyQIQj@=8nd2K}D3bkF9OA zi-;q^1&(O@^yH`)`Hm zufnS25Amw<>DrVFD^jG_|=*vI#$ny z{5Mr6F}z3V)iA$t;5~e2d!^%%j!u^SKq-Y zlw8BO8I2163K0sUE)#f#-)C~?QD{7g$Zz|C1IsH1Cak=~yyj=8=wm>8=oDL5E-mum z;A9h6;~D;xW!mEEMBL7 ztKk+cy=3EKw-;yb7_5$iMcoD6Nw;YJuF!b*?K=EzMJ2j;QRmdJON+XX8w~Z&AH!4E z-@Up=G{q~Xud7+XHN=(yKXY$iR+nykHik3UO;rtfKCmKWV+P+A5jqo*RL_F2s>2)s zJKt|?N!t3tYPT|vHR&P46uf7$ZZcn59&9o#E_iO;aku+f>wa8gQib;cxmi!B=AUzc z<@zCOi|9>lSB84#oGE;}GfR48Uqxv^pCH;A`4HT-W3_^f0*wo50lE-WFbJAvI~=K< z2H7ys_mKJ4O+m2_e3-!h%t`9XUn|)%fW-{HLE~Z3`bwgtRS`8^&!-9HBheiD)StYo zu#g!Q+^eJ+{DARnajerdOOlirMUjS*a*A7|8f@Rb!z{~}l~AdK$VO88^Txk{J7qEckN<~A$k}aaD@uU`M=ptA>DeJv9s$aa3_AHw2kZNN6 ziz%1lWGnO#f7uO`^zdN&Z6)$wn#N_B9$6;losy!*wSU&aqgnFtTu-ki)##P(<=nFq zj`*E>;E@h#{ogDNcvMV>W8!U3(TkrP&KY9 z7p7Lj5yTNPRIp&-^GO z)$Eu4q{=a8!9b+A@jBFIG=Hg*Oo;?3N*Iz-c#5UAHcE#cl#rozQ!Opz%}QVwM-=6F zL)+`c33towN6^R7v{X%4ypUWb(Tt17E*)s-72zGz{`mMou&CQ2t6jvOz>yfXEDcI8XL)2 z9VD+9MoYvq*oMfVjn;cruYnE6?2IKH@?>3ex++0TlCE&>YOpwrT9FzVFHi6oRYqrD z|J1njNdVs@pM{g@!}k@#P%3|OaZEiyKJ`@*Pa=&ChTIGt?@BZ8Hn@8Za-Gm7Z7tof zUhAjbT9ulVwKC_R?Jw(1VaBC3+&^qlO6@t=*NyUzpK<36Z?Yu;(J>`f+@Lq4NBv_o zND1fP1#hsEd2L~o$?Rrfd5(({fr z{E22-I{S=UxHpv$PYDovJa|FZaB5bsMYcRlinJ?l-e;J{Ub%wW;OS+TBEtjx8+(&XK<|>%G!`HbCoB zJ(IG}Ewk&hvN&wMru#be=Ye60_OP%!^A8$`--iVxHljbfu|yieL>8u3ZsP4z$uvNk zALH1eA?>_}r~5r3^>iS!_qq_zx3UOQHm7-y$U(EPF@5Zuk|bzixQQa^g-;xwF&kA@ z;t?`3_?u+o7q#U`7x|l{r%(R~AnBR*xcX%;;is9am7s2_e>>{#+ z9*q+y+kEDBz0o4DO+@}c8d#3=isdRNpAm7R!r!E37J2W9cQEOJl9>wy8VLcMC$Yf~ zbS91pd7N&?+=Qp4vc%yvsF9r4Q!oF=iH!}CH zPG;O;atm_YW_)1krUBdo({2x)+QO-(98Dn%IXFJlzU%pU$^qpEn3%oht6Up*NOAEb z9bV3BsofD4w((};I{`I$jFR>M)IlgaLj}M0*h9^`8&F53d=bk1hJ^^MDpb)?hiYWycnf5o86$ zU{Co@qea9L@Gv&``S|ffRKJ_Z=jL& zN=V#*^!u}5uR+H_YA=HP=`t%vsDgU6U_L*e82x1NUkVkm&rb`w%*0~lkQbRr} zXZj?{#_RuL?46;6VcKZRw)wVg+qP}nwr$(CZQHhO+wPukg1KvgKe$0f6;)F8tXk*n z-E7#&nb>NR*dFTt-Q9ikF2y{-P=@NB-cApu&Z!jba9Wm?G*$+;g((4U$a3nC2Pl`mA}u1s+F>O7T+eBlT7Q?!K%|F0;$IO5_Kl0BJ(Ir;Xd|B7L_7n z0$pi})KIkDxstmN1yVOXD^XfEq}w>Nee=+qK-w#jP^w1y{rc^6rxf!RM`FW$UQW_` z+3*6^S8ikx@3Lj5W*7zA_0O-6-hgQ2qz=^ayY^y+dGbaP=_k~asEeJVPSMzea@_`F z55Q+UU6-r#Uo=vw34+2rw+w4F*GF$}eO)|^VU$Cq3*o&unkN_d537_B;j_Ia_3(oZ zBvo+SJk-?lm_mVF{(*T}9fVBUzF$8Xl7RytiKZ3k4)BIhZCovfr0 ztx`ROK0rD;o&0@WlO^)+5V(F^eJTl=N#Bg%3(ih@k0=Lh-Y!(@7AKDsnNOI2ZTyHn5-s%M;kP8RqgYeAv8ksiSDf`B@ZtWUbUY*%_#o&vpGIjXwFSG00 z3j`DzeY-pdsdV&0pxpt7ta&@}5(=ylt=={&W*!)!ia5*iv_l+RIy?lNKbZAhU6*@F z_ybdm8PdUf>rQm?g9b0XSu1tdjtq~^ENm@1jBE^UHNQ3^iQMeXiH1JL7S@S;Pk`f; zQH=w=go!3z$;?jOJx3q!epv$GThMa5Qn9DrXZqZW@*F0uG%-=tP`%fzG@}l=6qS1e zCe*L5JR{PdJ6(EdR*~%9k1b|vLI`)Lb$ikna+(Uvt}|1-sR7)|X1Bt;b&m9n*eame zWt;TOKZIBq_CLduTyeewX;?)Jhw2Q1WWtle#%w1wC26LDBV-A260}mMAMrF;RNEI9 zmm_d2YF%?p8=2104tFKg@HS8d3nFpN8slNz2g6>#VYXR2ic@uOlztiKphjRq9O&)Q z>tvsUAV8+{$+Y8qf1TTn;+vim123~h=~??_RVrR&f-EF4;2TZ`D>8y#f;@z_aJuJj z}3BnO6&U&n_k=6pD)}0}~fgd@>Hmj&GjKJiL0LRz(_+-qOkX&AFVpPyl=Ti)S9w75si~#E&@9r7D&fwu2 zf!H@s);TmGnCJnu zYw}Y8SmKe!7eI}{-i!48|IMR5tC>)nGXua%QI+*QU%H;ktK}0x+!q39Wd!2O(&5e` zFvSz?pV<3%Rs->G;DUTWSKp)gL*BdD0bEmG|G~bsz2OB^sdd+BY>cJUlu8@xuYcXH}3~;T^DS-w60&Zv4sj zDevD{*qOo8`%wTL+nB(4eh54{19kcX$cvf9x1IID{!tY)GyzXnOBVnj#fPAP+~l3p zS;l|pJx<-@)NBKMTGMSCf!F^$eB3RKp?PL(2=w3n_5WqlRh7opwns7_>7o3skj%wg%6D&ijev z@)3idr9u9Q4gDNi*q%JX2Y;QoJ8_wf`GvR#sDDcxgVk&Ps2#fpYyrvO=y|^B4MRU^ z{2_L{160$g-SIs}t^A(!27SwcYG31exNCC;Kb@`3BUo<@R4Zi`Xrui|tCvfw}`VIVBeTr1P9%XJy~%68RaC z%RT5hbpQKWUHdh4I0F7DIp_W;Iur0=aU8ExUkUZ911~afX{Z5f-cb0yO~?jrsH9a? z{$~oiU{fasPDjIs_3pWj%8>~Je)d+U(r9n1%~i5f*%^rn7F5=z+&1NLynq)unnA?e zhyH#dw`iPMEQ_c}D3BqcuxpUlX&|esE2Cc)$DK4K!c^#iCu)D{d0FG=-IMdEPPwpe zZ%I-}x=2JE(l?BYUG##y7n4Sm{y;shbDir_g*HthH>NMCZtOeMIF+DnTY?H#ar1t_ z$i6n+oACgPApX{hWybU>Wk*{D4)+XR1{?0&wxR7^@gP|>v&|V`oNL#8 zi9sJxVCAu|wRY;EcT>N-wm0LAlYM_xTG#4VG^53(S1c7gJecyc11|MbKLYw|ya+N( zc#fd`-|&z62D{Wf)%FG--j6RJ_s8_WW%&An&C1&I7LHRNL_D^Qp=+ksBV{FjY zOf6jFESKl1VD=<2<#s)j;tHDS-mR_oWXhCdm2$;8{BTyaWStwv40p5U;R*#F;_|qQ zaZ9I?G+4mS?&e zg}_NDXVv%UNyvmAEXx}e3jO!GfGu1*+&BR2COE>HI6VsnV{{V@em3wZ!+L9@?r@! z)iL8CuiRLvMx9r2FpwPV#-woI5U5qv!SR~!40Ov1xX_xu|IXzaOl5=}c zO2UJgB|Bv{9wP-%#)qH&f4 z4Z00adS+^dCzcoDtmgWkrse2hiee-e?c$3xPx5Bf_r{#8ioLrTJ0OnF^!t+BfHI?N zs~}x+vDrW<&sNgq=GI2?3GdX>Npye8B9C5jW@cohaRM))Q3CYtJFvHN(EyC_kKGK~ z-HtLy$9kDKU2r}T8zcQwIyw-y#EZ*r?BndUSN7W;3yD5LP4AJ~6u%VJ7-@IiHsnNU z9&bkC0(v!j_B-bHXUj~<^od|Wp0zX07gg0^apgB)o!1s~cDhYS^Sidfttx{~T8%ON z;z*KN2)*s*ekm}(S^Dk#TTWt<_&Fvot%{MhTCRygu6Omk%!mRZN zySEb9U_inZOOR5OvpGeL_`&E}!tCA~0&xuRTIu?Y=p_sj32;Q`arS}7N!117g0>AD z+vCm~95R&YkqMUF4IHmcqg8|N1q#}~pNFs8@IabU8MdX_qtwrH*XPc$w9=+#lWO=m z7aNb5qaq!JF{u3Ma|Pj5M}Q>{Hl^)Xb*KpczR(|*N$YM5NrO7BxfZ{xZ$^J*s3QD$ zdBDLM-&;woac_DBgErFN$RQ(6EVv62+O!e1E7fSW~X{UUYoZQQ`YB)k+D2mX3B-o^gr1w2# zo7Flns#;}WfIp=PuN2u4>q&9lX*q zT{#XI5eBPM-HTh?Bg^&jeJ?{ch5S30N_8fmBFj;+Zd~KD2l!f8F6_O+=r)-me z2Gi@dU&X}STr(J>95*Xbl#agMMwW~L!*B)!P%i|* z1gi>@6|3AH71=N|NZ%qBW;A{E2A=h!8e5k=k{%YrNg_LU3M<|#0E1e_wD8j?xdcY` z$@m%eI)hYgb$<0}@kb}-PsyhyuB{l4Tg05Ppd2k;W#RsqBLp+S7sl$;%S2+&4LUSL z0BI!?XAUA4ySI@Boe`p69 zc&Co`sW)>nv36=QEIT_4&~H8#Y5+D@sag3v`jHTK-BhWRt2!Lal#`ebT_Ky_cye$J z89S6Nvg^%nsn0$!Y+tUVVl*Kh$7oM565Jqd7H}gsBTc6j8n$VgFt$`(V5+%SCj}UN~$&Of>OsG`QTML z61a3GktA8a<17#{EmU3uy)zKCNfsm>I-$~MA_ZyGdLuI$@Ktns7C&#&OQs!s)CW~v zbDYVQ+`Nm^HCQdn{sTas7nHq zV{uidT$3%SaPonn>JRHAuZ(bvPHKA446zHnlcyVupaV&rbuUVWIhuFS{YurNpna=f zvp8%dt4-BLp5WeBi6p3aRYafxPor`rSO@mW@=O%^0x4T`p(y{7&zL*r@QPC8pu;Ki z6Yo08L4F-0S>-l*E2OLBlNb6rgi-Em=}YlcYc?+l@I!5|DhFw2WSUfEh8u@eEn6s8 z{3Myf(uf@_Pi6Kgze~^hp0D!S2yrUD0(i9jW{`TMWc&71WfMNRW!tPru09Q`9*QcX zD?@q&tWip5)H|oN3O0CA771f8Ia4iQ-pHDJps4u-h3U>|o`J_m9g1I8;|qUvz92>O zzu{xyoIQ8O_90bER3{LPx4Usb7Hvg%iE+al6DvqSAb@)-p&vO|cvJNDbRTS;6$231YiWpQ zd{CTdE;_?)YOlh}ogrO?Z@w4kuM-A2>T#m-N73>5bBiOTXZ9^8`gDK`{|1x>@W@gm z=@8fW8~?n&hfPSE6hkg<|AiKl#zcU_S>JCR`6jHy(Q5txO%RM)>x4XRPR`b?rC8os zczcP*mjhZdM3S-_cr#q~CTXZJaipX}zHaMgHB@}~tw&R{jf3W^4tN>%nYCBm?(i`Q zVzTdXLuAFL3{5SbBy-feFj8=g{5Qkn_E7KBL~6v(p->&?JcP@a?QA?*<2pPV|f@_Yui|e10XbvsKb{q>!1kt`rl=A4erq0qf{E17L@!N#pp;J&J-ri>= zQG*_EIW4KLfkrRxvKp*vJ@ul=?&Dwa?fWs@F&KH}xg?rw3{?CBfpEOHFgZ5Se8>Pi z{J)de7!6H~A~Mk~i3n6txf_V`o229hxCzgugW$99v%6OGxYn6PT5dj09o<|G!z6iD z_*>-s7`n25+Zlhl4sel%^mtpzkK(MgHx(!ooAE3mh$?9XF=*bR*3lfI03H1LyGQRO zk{?Ov*5*Ithu56|sgVlpU2L}Zc{22er5TFGT{(oKUs=)QyfS{eMl5fq+7WWR=;K!# z`N_1=W0X4uOlr=ULb=3i3#;DGzU`G2HA^?u5%j`uIoSmbxR(X-7QB#ZsUD&iSf7%% z&RPTPc9{UV6xJctvCq-}Ut0w1Qyh1Zrm2hx|IVa^0KO~0dL zQL`V;n?8dKt3C;y90f~y{^przp)Z6y4kPC4dFhPPj^ee2x z_j)@QJl!;y43uIi?wO}qqh|wm@x>!OjvkuzEcRrlmQ&s|?o=^+YtY_I+YskOdn+-G z<+{Z|n*1=h4Ru>ILr~{UDr2j?IrlvSW*nKUUj(IGtPaLqf)*fZ?C{9dK z{f>uwFfkgpXup&2SED0etpGYHAS2NZebVVG$nMciy&&@bOn9G}A-~`V@_wKf@>k2k zehB8ny_8mo=v~X50Q|VHx81kF@c?ezLNq=<%*CwQwWB;n8EeT{b@xzDXjT;Q5U)8f zi)|KZV>uI{{};4=eh-s#h}*#emSbvo@?;o%o(db?AL90&7w+kA52P#3P7KO+3r;&W}c-I*a`0Y#P?XA3N*Je03kbCp3JROMGrTXwQ~6qhN|v4u zLPx{QN&{ZHCRZRjePmI<;u)o-ex(WZ&(bT(t}0<1O$6bxcfQ?DyNr|+J-FRv8GO;y z=5dNRxA(S=JEV;N;7v~9cKY3^w^f4Uw_W_W)&OBnMK&F{$wS&o@fuO@#@STd^AF^SqjAQl zOYRVvIZ~{UtuBaq$*>HlmnJyKuwwsn7bC_n4Zgj+?>t%+i|m({(lp47Z~)(ftP}pE zFk6XlshuzA1TMIPT78s4Go84p!PSTm%Qq%aenvTCf%x)r{H#5!F0W?^%5N=sC^a-H zzkiZe-FqrynZHa(E#6L68?4jgOy4#7_!n!?1G0nJW7yMmei|HSAF;vUK)DMPZjg)= z^QeZTJhXs`(Nb4Zdk5nhj#mduX7nG?{$K$#QC_R(bp;?1Ea&&XXEBL;u2LtiD8jj- zGP{Qq$X3}!Z-RqvdN#7S;SR-*I8QdhjP>a&C;GI97AY_eR(jkLof3GX!H* zZgIwiiK8>E*nlnC&Mlzx94~ec&<`&&H%e)h zs-xhO$FEE@nGM#{wkU^OVcEROA}41kL=m|wwTzY7QQq+qbZXUuq?E+$qb%H6q)*;$ zLr(%&{;ZmfvCgCsSM0pO`$^5Q$=qAGi~A(}_c^-obVl8U>R)~yMN4&LwX;*`UP$w*z!-ALqU-cq>59hnm31raO=f0_ z`CmcG1=b zHh1-dq#Yray(5cMvpY^yaVR4fvlcnqu)#BQuHp;G2lV?>g+F(KV#8*K;JddDQQ-OV zN`t!0s^Zc-Jl7iQw5Rf#X{BMq)gDzzYRLwsq)piI7@wz zZB?`CrCZhovf6)bkTJ!LE4unCN)?Y93#N>SlLWnQEs~|n@@>@MdD2Hx9|+9o1h>Sc z%+5sIZ%DbIU@Fus@CIv%+4q)4C@ITWact4ZLSL^nDMKYC1ceb|s@j=s)#V9|bBwgr zYBTAw)tOPrBsjvM3WY6GHC---a`69pNx7tq%Jq}~B0fgF#M-(TCT^O}XAb;dTqf9g;abD2dksnDHP5)35WZ`7jexa6ciAN6~=b8Kya`9fKgnK3f9@P9%-oMGB)%@9uqk)shnf z?rq2KF9KbM3SaLXO>{5}*DK_Ght13~h{U%7kKo&8u+O9BOM;cXw)}D=C8NiGg8Qb! zIO9eD4;Myle9_UR*u0;g*diW@Bt>-WSFrjRChw8QVGGQXzN$SiuUv>pFaf@v4y6~p zE^>i?C6lybLL6GcD8@V6y)_@*xc+ z_RRcalI3C+itZGPBpFM!P7G@gYb+7_&eDy!Y2r$_t8dZjw-M_c0+ZuCb|j0(alNDR zW-gz~5a7SiNWm;M_1?+>PJ+we5_;X`$kyH6MMn}V0~#$o-f-YtbB1}GvFHgD8Di)u zJw?;fH~w8<$GT{;F2nj;HEmf>NZLUdm;BZQ&Vi@$?q2cce$af_xRGXOqBrIb=~a_(Cv=psji86 zh(qsM;worsA`L+Q={;yyVAjDTo>9RPO{ZYHT1f?iVH(kq0KN+T2OEUDx%wP2IeI}L z-8b=x#W6mgd?P*G2NZ&cZhC`wNSDm4{Y-0!U{~7{k<6DO2lZf-;OWdh zhQR^N=mcfrsMRgkC&AE0F&5{yX>D0m=qXn63sWW3`3LY9Suy8jWs^I0s^+Df#5)uV zT)L2T>SX+pq3@h3wE-WBX4UML+s&A9h%6Xd#4%u+NC`4KVIaRZW@J!7f^)+E#h6p9 zww(YKP=>8*!YilyqkB2~-yK;G4tld|7Vcng2LO3%!+Iwz{v)`Y8G=|Oh_!@UQ}pvB zSQ+9BSp~nF;B6&faKEo9=8MeE&T-<@_|1L>Y{wK+FwDVjyeo%9|8Qfqd3nqzLZVA{ zP--x5WW_di?NiBwv^#wm3i$8Wy#exe4N}l$uLY6*GZvOY?gX-nT%!7HC8QlVQ6JJccTly9?iq z?X4tEz!3tR&@+vz{LIwD8nCxo|0z~fyb?SGv8E#vQ{3pT)|f!iE}CidI{At?V$!dp zTbzRcmCg1!NfNQKXw*TC`Alg|KTQIBLOFrb1FC_O2rMj>$HCebF<_dDEtkBgJmF|m zA_i(B4n*4lx~%MZuJD|ODfDrGQpR%6zGa-bect1u&-mSnwRZZ2X@70JC}Ao}TE3IB zX+}l$q}~WYQ9ZTVJ%>H{I-ke)f^*XL(wO;}c}Bl4=E%Rk9{EdCZ#iFRkLD|@#+2IHKSYQq~x39~sJ)r4^EnOMtHcm4U(1t`AdsISJgnu!}^dxeU zqsEoS)6Mk)lzh7fOMV)Uxo(q)N}+(Fsk2m(Js7@1;r3%XO>Xz+UW5Qb%9gXirx?0$ z1#4V3VHg&Kz3pz@I-;jc*Jp5?gHmnf5md@dGCuCK0Nxye*vhz%0~1WS&X|WR7O~c9 z6Uml?ISP`q9sN@Mu553F-#cuIVT1|Pv=63bh;&iAy@Bl?q>Ufv)05Yz;ebj0Nk$``Hckv@ImdPK;o4d-|icF^#)FfTPujW;>5mt6FVsV z0oXYw`PGo38{JsBX0mx3t6AOD^3oekr`q|D1->}aAof0957{20=hlzP@7-UW4_pCqtHb{dhW4tuDIPnPWhDsOo+mUl3e8%;VwtZ&mp&FKvmFh`Z*9U- zNTrjU;c`led_-CFR7h>U6LHEc8-o?5U1eU+wK&3`|6I>g-A2rb8gu*xhM~JLx%%`g z>o)%z|B1eXm(mvi6W6wLeO$v0xLh;boON7a^hCO zoO8|WE|7yZEql#+8*RxH0CgWE6l4z??K9solO*%Eop3Lj+V43CkwG@*0C|zEu7wQy zQ$kuRpzx3PV?_pQbqz_*vW}jF+yRRKfA>>q@+zSg+eU6=_jYp|y&=DUK?m|J2fBgs z;stK~6;g3=UmmkUJH>6`WKe#Yo^KTdJun_Fhtn&9ZJsD(Qj?+z{GxL6Q5^h1f;3&g zw`BICNrz)ahpm7uo88K`AGiB{38dKfZ-ma@Zb~~$h5>=^Ww}By=Ov7o42*aj6TbNfeHSI+t4QPi7DVbQ8tr?=c)h?k1}ZnKEzskY_rlkA zq?HAOsw9Z=>1#L%KTvv$Phmpb%INi`8~5zl4iogLQz0Yv6=qYs^q&HLnaL~_r7PW4 zMA1@WX^U=UlaMeYq;{wO$zEQ{YOW@jp{PZ{{3lW0He;}fKtV>#o9>Je5esFsiL={9 z<0_e+WGT70Bu}KiZ!b>UGHQy>cMk>~ec|R;%R#lxY3AD#%mSKSvGxZU`0hNAZ2M|^ z(}+^mxS1BW2#*Atr}~;WFJL8r0jG0;o}%(##0B0m+2*o~F!kR4Aj$O>0xj{Fo;g+M zli39Vs;mKK2s-eiEiKRMN1GQ1PCDZPa*me!${vgtt*j9|yG?5P6N?bk%%1Cd+3-K$ zjbm22)}3^z>WpS$IUKy1x2qXZ&bop+E8mx{t7P@`=67Oj1f@3YZk6n~+y7`8upFg* z{7Y9&LYYjCfiF`z4j{M|E4%vZ-EkTyv_jx4;7hip?MbFVUlS#V)=S^)9+{BCu0*XYd?o> z&d@JZ=1VCNbA$U}Ld$CE5t_8z$>u45x7yq!Du#6c3TZ=BEu-VtJ7#XhCLt1M$f^3) z5e-6*?%d*0y=@VEAMj4OzO>oW7ST|V`G;aoKM+xBn*V(ZBQo=mr^N9USScEPs?>d5 z;pJ^mAA3j8W}m=%KXFC3MZ>wxyfv|T*5-qn{Jh%NR8anPZ<$YkM!)39?*#7>Y-!%1 zxYD7`$&MbXbdm8zE|Y~e(U!pfEU_|;QJZ< zYjoyiKUH150hy&A+50dVBCZphsh&%}bh$0=sip?-f+@H{o2}NHcr9X|%rheG?1p@e zA=eJh;!h2|E~@E1CR+Z~7HjlCz&sSGDP*J#@kYR<-+}S^IDzrvv%S3qSOPF$n8xDF zQK?>kYqacadoIarOK=T76kx^iNdf$H?>Ibo3I!p4pkI2|3Zknx1-xp4UCG9C0l*>x4V#sy5&gb|T2&zkVo zXFk$x;QmB-SC#BkQfZ09q%%=tZiP8#c0oHo5aW}g;CaJVh4Ic}g6A~1ob=F^I=T=u zO~-a5mnI~-N^;qE9PW3%y>G3Ild47JbVqO>L%IHSL!6wTzYZMZk+_wDY`q#IXjbKU z@BUYcx$nK`9=Gu(4IMAZ&=q{5vk~Au^Mbi(x}lK2Q4#ASqYad-l5pHt@+NO(|I8%N zn-`xM;wS3AzGG3bae{2G2!ttN6JdmA5TwZd5g)v6?~4!QYMbjpM{Jyixs=&%{PtY* z1M?lK(Yro2&Fvm6T*%DM1v7DB4NUAqC!&jD_64u6dmp#yUijKMEs^`(n~Pp*-G92p z8b%|M40vlPk?AeS-yO*{g;QYJH}B0&p5q_3_4V^&u#*88e!|C&=FzLeI^+)21}BwT zXjnVqt$d+}?nzZFGUsW*g6NUIvl566Jm2<=jAD;DD({F4;^fOmp@zG>!S`f64j)%a zKhDW>m3Go4Ts}@b3Z_YmY}}O06>r9fUSi~&_^s=mIB(gWjE`m8SQ#vFg?w%C-J_Rm zlgcRmJ-~7FwCD}~f@=w$bl1M!Rtl$v1d64MLkD_X+=&uf*s=LR9}Ez=U9%A1enNj< zjr9R@Nd6^Z4?A*LSfVeaZb7b5MBI|BXv6$rnK{XLA`;hn47YO9)hH&v#zV8iqW}1F zr)m`ic^8h`%Dc8_vE<9LAZuiCx)1viJ`w$8sTKCdk#(yd>biq4Hw?kgds~d|ACQXH6aOU!mtqlmb>)jin5KWePu8zTaCFeIwr}A7}fO@ z6h0s74BP+x$SP|3+u-A4GZlkJ1V-ZjV(zuExT$< zrCIWIG=nUER@MR8&5y7wYqJfbo-jE0v8a^B6w+^H&s}Y0CET2hFRu=; z*H|eGypw}@TTB_SV%Eg>7J(K)lUe#h=52K1j8y1nK2AvEassGUuZ~+jL{k++Pm2CY zAptH|k4fkwQG9#oS;@jWlvE3oP53`M^LG!-jz+rFtR4o>BTZc_(xJJZnnCu?O!od_ z$CcuH&HZe5Dl1kG1sS4SN76VimQKwbU~fiNEJO|B(G@&Z&9}_qq{WYozFI{lb8|ci zW8Wjxh1H#w*o`WzN-&|R=TL^>TC5{2omq}Er{`R`_d6#l7o-O}w27wg^9#rsPqk3M zM(I;#P)icJl0L38=i$BXT_qDI+K&4@fW{sXy@4i+n7NH}8J@#}cm*MU)O-egKz>$G zr<9Eb+gN700#@Kw8q3zR@J`AmhH&aWhUu7XDcj9g8C{+<7XV29N$ zVl>Iny$r7pERG(oo6M(ax2xmJ_yxY+Kemd1?t4Tran!^<6|`CwT>m*7v+qrm3p3ge zji`UGWNg8T4o2j(J`;*;yW#2VS6lv63ieJ~vQ7No%{9~|Mh>$Q>!u!e#~a&QfP;I9 z*Qi<=4!fM813DlDDuXBMr%p8&C1Q;cj^TA7C7(T=y9rpUEDMyrXVXdjr(=Z@TIi7n zu8vLE2d+M`l*nmTu8m>D5V8=9?mj} zz(OHi1yaxL6yU*bNRy?Xv7JcJA07y>8+HVMR<`fTqE5rDxLw$;RpLR)8vY_&3oKZx z*QmT{dc{2mwr=jL?0qPGBzN?j-;Kd7wv;h&z|CejVYD4fq+V(FsT)OmwhfNlSZ(In z515$+Mn4R~ALQRizG9r$I8C)tKojb;D$-wdo^1?FPPRsys5^%M*Ye-p+8?cf39d_c z9ZY0AgB^EKGTD|F%otOPAGYSrK%0|ueQ-(dFutPQYxCr^Dt)NPbSNt64=ac5_IG8e zN;-O0(R6AUD2$fdFGU)Uo^=z6*eo8cw25&@V*W=+8Jo?sjKyEgmn)!Q+}lvN2FZ!pT>IB z1E)Xk^7Z0e0us#=)KK;HF-YzoC1ok|dPj(z6ed!I+I)8OY$@xqX>pfoT#UYICu=&W zRA8op$^s3i&1M7iD6S5rgbTyABIP7Ugiw&okl@*0@5gm{^-idikmGWHN}@4KBm7ol zPg;XF^^<=>+!UZ;(m3!$-@p6=SS=zO^wzerQH`T*JR{PCQ(e)0tD_XNGM>_CNQ!L? zkqLA2l>NQ?w`d*)nb|m;5?e%`xhbZ|a~cpkH5m5u`i&+o$hD z|E&l);i2FA-+Hh|3@QfUe522hoDs{B;ALS3e8@zvNvj?V&@ltzH5kx2ft3iMAX7V#<4AzLbVoGczPn0X`I5Yg z5}vXf0illYetUT?u(r@*13qz}rAIjd1N&M0M>lC{Uv? zu2W|Y)!<7aj*K!OZM2{)}%f_Eqe*7CoOUz&;=Kbwl5*CYSsH-bEFEkjN``N zVtic-7bc&-T6h2T$8+czOm*nI**;8p{uVchE+ivFZ4ys!OF$)Ih#DV|8WR^`^zzF$d5v<^J!{>Q$W zzmIdTt8_-DOjAalxlF!zw+XMdPd1qHt5bqZ1C^l;tBn+{P}wytHlju@@Z)nNk;qqC zt61%}`PL48Z}H5?MNWJ(VRs(?ok{aQL1#!vw=|kz8CFPqqHSE@iM}P|_`4xJ;=m{b za(gL@t8OluZY1X2hTZ>B-fG zDHhzgUKuIp20fbx1CnGou{c~!;t660K5Hsw?hR#6;*i}3-!lTh*}QvA6--XXdqkz; zBgiy^coGkee5XJN+b`FxN?@u^nmi^o%0VZvPiWv=9vz@ZvXkRT!nD#%#`W2GM~fyZ z(RUqCYYHEqOe0U2M6-lXBQ+ux(pl0aId?QhudsYF3TNSazqg1?BZ}y|qyLo=Da(gv z1H0yLJ2P*T@}#Cpi>z0qI+yS$K!^XtG z{(p@Ibp};l)_P+lA;G7C825LE$05XzajtNNrKg|npPr{#)FI4S6sOc#T%_57fCyjw z!F%a_`@Q?}vGUo=Z1HuPeY)y=LZD?CAiNpF^IjjuaBYm=Q8P+3cxDm3Lq~n9emZs z#XkxN0z$Ule z2~c}LT+5#o0x+=4zZ#%#60|&v%ikAmG-wWHz6F%%x1W>Vsz1Yf|nhS&!`#NI8 zzJNbLALa263Yg2pqdb*HdJr%~r11XVq#PU!xJIGve++Fmt(yXRweqvP78Z0%lbaE0LP~0lOGiAwH?IgtZ&$&$G2DK zU=E;a1rq=6fmQrvyk4@wZBP*Y!5o3z+`p{%H#j^95cL3>1OVuGSoFw0IcH>;#@{9} z7hE`3VD!g=U~rK8pU^|t0L10NuRR6knE^kOfw#Ci#$`+}gs&v(%hcbb`YF8~vo9<9b-=&a5;!3W zR1p0yv3*f*LGRRZ`1?Pt(>|SFyUSmclRtv{zge*X44{vh#-GAJ_|BOPo`Q+e5(4UA-|?cai?XNaefB}|e2WUv(=>vNw zH*e)A1ug=AG6a@TRs6O+D6soSegpAjXaJR5wth^ovA-2L1AKq9Me+y1B2GkiLXT>u(Y6zB0Kt>Yk~;n@?Ik zu(I|--Ks|=Ybh4hujisfziEh*w+G#Jj4L|M?#@@izyeBGnpr1f|dYV$BLn3Rv_-RyAvO%3*5%iDgY5p84GeHm&5BA7M~Kzn;?ynWBb~_g!-r6%`O?)i z9F=}46zu}`7sG+~Ect~i!=<_4ZA7c~Sc!s3SdsWFNIy!*R+dvdh$MC_+c%SSlm|gL z^$Ja{eQ6m+e1?>-XQ=A1y+B-!#FAx@v8#f9oTtCwSPR z*_5p}t!^W{~>6!BE*fv1ZKxNR%9wKqyDjsM3=>Q#30-xR_e_aQ&qvI7K{EQ zP^Og3WSm-_#7b4Gmk9b*j)8(QJyabsC9dJuT_MB}?>a>N_&Z+ir_!Zyaecaz{LVqB z9Z6qaaRv#K1!CbfRi(6mih|VejE>l=4@OcaqrZQWrn27Z{@49NnP2BO3;(_n0-pi6 z5*Ub#0CVkZ+qEpH$o%lpV*MskLCBXm?ynqtn0JL*cpJeM8zlB=*1XF@%*BJo#Y&W) zhYC<()y z)@NA&N=U!lwC$XQ>h_kyMO@1E$C4m}_kq6b6(*+cu*E2}YuiV&!~l&=PK-}cN=33d zTPfdaX|*ghgkcdp2f4uA1!~LklGU2?+@Q)|zXH;=dIyXmg4BYgF<`r=>{6;El7=p} z=%oM6RX;dg9!PNJjfqnx&sleRVuzvm&-EXo&5}`joV3*)J8{)AQ$cDZ&-*r;Bm0L@}JLGCeP?wcEP>oBYJep=LKi@Q8k+?>r6DZ>!vSe;~{ zB6EHrGhd4hZcjqImFJRN5`*A0-c;!z#=PG-3)5z7RXVC(JxfR0YmE1Xu63&=rT;g= zaqg4mSBI2?zP8s21Kaa=kqMKf4K2o*4VM6EsxDkSJuEoA-L!5T4yYp z&G^MsXZbu{sanK$V2IvK$QVM1!+;(IMm-s`B48X?{}Q5f)Qad7V_59}!`M9p=h}6F zx{htzwr$(Coj2BsZ97@9ZQIF;ZQHiaSEp*9I*q@vo3lR~^Lefs_i!x5(kuuODf0e~ zOGv0rNLR{t$t_CV_xmLn?)^pp74!DkJ}xtCHwvlHC9+;jI~Og5T>A*y{V&p%3Rlaj z@ZVq2!C1ZcrtAZsoXl7GDit&+`5O)+nUmV05kMlPLh&oa1cWwcpZQ5{Zk2t_ir+#r zPHAW)vnwLC@6zt#d%J)ZhJlNCOZRPJR|f$jC(c`1>=oFNelMv)D6L=&1^F%JnzNm# zG?BFPnn0JhrPJUT7MIF?aaACoI?DcNHAsRQjVToNLiiMpY9i97tKG(0oT!~$2Vc#_<3r>d4C1Y<+e;)usmR9zF=0WveK8G#A$jsMkX?lV z;@=i}-sk9Eiz&N}iNrm}`FLP{*ZrnV2fo$6PCQUUp!6=>9Aq$`!YWM|xU&3KNZ$*% zb&JF(=iM{PfFwMfE(eyCDicDTr#ia0#;SWMab{P>)9T8raxm=H=WLvv+ThB#ePaj) zR0uty@2>_TI%)wQ(3sGkQ~fBfwNr*RmjoP^yzB3q%mf%w;&z+*Hq_di!06zdt0jrZ zP#wyOcKOkN2Hui)&<16_qnyrnb0@2B>j&P5lLTI|oxFdYVG*2No`0pfZ1Kw-4_>~4 zs2V$p-$jc_3va+hi|bK6o@`sZt3SE>|TaeowxN*zt$jiS-X=d3sL6zBb;3>8kav zv>V!pw+xc%6BwgWZ=s@JypdzBQxTe}a9eM;8lPf@#R$$kjZnF3z9n{O~MWpjF{7YrCWWZ;*O{X|@V%G#BR zpkIKX7#Q!dcTp&g5qhN%Ql<^B^2#VTt;Y__UV0X{jiTDoozrO9JB(6wbO*ayLAjN4 z?f07v@i4cD%dJdEyW%%}1PBC+$ZzeY{|h}uPX{7n;(wp(bHKJQ%2L``x*iswkn?2c zW-72%@<}FEzx`#omwCpf$A+POD(%Uim`INl{OYq^Z}L)(@nPYAUYCwSJDJTJHmK zL&Fsd=lFEvIQ85kwI2}$fhii~f1y=HnLzHvgFSbU+GHa0N~dd_BTVe-mb)4=%X1v( zJm=7*a#sqL=+enpbi0iy+u^%UZw0$R640C-B}JKlHFx2L(Q?hHivo8v6*xOFX`XvEG_vybH3cEz(N68^%80Z zk(lQxOQ-_8y|ig4nwM>px8V@>#v@Lw@i1nCmID?pY+-tIIr37?WFoEwc7|psV51W! z8)l8kSlC0_nhUehA>kAMzG@hzOmu9iIgdd5#@o0Xg+!0D6mxaR~nUee) z5q0b66{e;xZD3ltm5jK@>GNYyJD^%CYu1~nQ2iYerat1peCTR+_<($8g=LjaCIEBH z*ISe^#Jtdmi^g zN5#`TrGI_Sol>AZw!X6+s$A}@FazpeX_OnLekZ(pzJRo0OV22hRxud1(|W%FVNW9q z?45RBik(_b2O=C!JSaNEXdiap(`q0(>S}E!aw!Y?WgO?pZALZGM+d-#}&EMIHB z@ugP=;2bG)3$1Pb1K_ltSlS6R!9ZK4-592&?TKU>O{C0cvI{b^R6#3>aHt>eAL_>m}Q#b?(=nNIjw$VAF!D6Qd7dwzV7P z#vaTTscpDT6eh1j_k5}^s@4=?7!$-x(+;`rmtOfsO_Wu z>bUOs70oiLmI`1@<$I0F6tu3S58l)ZB2?x;GmE7wE#}UYGfy|hKdYsvzwiB?^%x|l zQMcd3$CXa>ZL{H^kF!roUc@T!Mvpqm4pk4GZ%Ea_D!XhH?~XCsJHhB5%Qkvi?+HD} zP(qQm*Au$mbcBDL=3)mA)&+i-1te70N+w2oCmvPl__2KxUjaG-L{XBT4>X`F-6Cfj zU*KTw718`+?5IlCaG~C))c@$vGz{P)|CU2e50aJF3HKz>>q9cxDmwpdd=LG)XLDP- z1R7N73qupCSfskeKhXKtE)HCAes+VFr}39p)y&L`Eg~c8X~qiQ(BYd6d8S>qFj~|< z-B@q_c@O+E^}Ub~uBS#VGw-!~8z_kcLa+x^%6q}4l{JQb_sHKgqw$eCF6F_3n6lL~ z@wIiv8i$UM!I3a|ihgDP57o!lf{-VR;g1wqnT~s~>X-J_u!5D(@xS+=_#pJOPI+b| zO{a3>M>9JXPS)n50jNp^ z`jb4c5iMKGSh-XdDQ~`#AF8Ssntx)};ah8wHsUFeshb~MwMds&r`%HQz*cAo3P5cs z5N}IY>M1ld_f7DYP;5YL-XM#DrwlojU%-y8`pC$$U|7m01v)wM;jsu?pqnrH7g2Ff z^&kHU;H(k?R@li)^o+dv&9Zc4u&JQsdY_1JRsL$|IbE)k?2SN~SAS#});j)y>r&PF z&WI-EX#Dy~PxfYwFE#uLkdJ1}Y#LshU{HF0P4nn+*(u`jTd~v7o4@^{zL>e90!n@* zn|>c$)+4~qoCos+GroKDbHh?~_BiG#BH%xunY)p7RMLSm2d0IT+H?>8G$5N5)qN0& zl+v|p@uK0%yAbjEATFp+l^)9`utw_cCEd`57)t3d5S!^?*4i1**YR3GnMQ zolS@?+*35PpCV0&O24S$i`hSo?ctpdCTV;|e?~w38RLpK#u=JHaAbF3hb@YqK9VtE z@kwn*wiY{L>h6#`9!3t#ZT6`Glb#~IXzDmYpOAiM#0xiBe5ycSCM8HMYR>d+N?Db9 zzm;D5NjXf-&i;tje*>g&e((&$di%vpRpx8SzjO0ndT|hWhE@zTDES&izn887a~!)E zT~W(jzR>G}@vQoJzp_rZ#2H{1pJPbfRrYK?#Yl+#sUCFCf&? z>UVpQZqMs?*>wnlIJTa9((%prpS|r-ENl@-N$Hsf;{%t193r9HZyLA4L|JcvO$u3A z?TREc>-Z;8?X?Lb?A~;glrmC^JOJp;E8NcJhm%<9-9YI$+W9bgQiQe`6U>sZhb{O~r$ zb3+(I_w)hB>@ko`i}P zdFP<1nPvIE9B|ntf_U>|t%nN+UB?HFT?Y%HU?afQw0Ga9y_kk9BlXO$W4kgo4ELRi z=3EI<854b+JIKXMb82gAyNkt*yJs2LV`E&cLh98aw&IVvQ{A?I&el~6#&H~$B|8q$Y%62lG3g z93ni%bw?z3LN?vcqBnj!wQPhdu&LLxeXN?%Lkmp@hhRBD2AQo@`?hxsiDW=h+sJZM zRm?EEeaM|bgT3#MdFYr@LO&%k{&(0aQV;9b-2^G+TBUl5Q|D+hhn+Mc;3a7@DPF_X z@||Or*g#28Z+U{Io6r_!{z1mHLP7vZBTKiiVb$sT=$Tfq7twy0%di&u5Jc;b)Dg;XjO0`lBBm8{&2Hs-~GeiL4fj zyr$`a@0z*JCsC)d80M(jV0*nX34YdBVV@uKtt(tY8;tV-9JgXIA+eR|OtKTDk?r8K zbfdo0-Ed(1ce!Zuo1bWyB2+VDR zyxJEsv3E_uwt8GH`UY1C>^?4O$9?^3etF`}(bvgdPyG}`+z7;8qvI#-1!u*_;c%O+Q&B7zxG2ex_dt-?}2l?kbP|43y9h|fvZ+v%Q+ zc7)r8T;5XcS?1aniT)e-(tRCrUa(DPE0)*{4nJ6bA>~$EstD?3x(Ifk*3FB=3rec_ zCl@hn^L8%sJSZ<%zgqnz>n{%-|=Q{MfYk!g>6#fKyz%xaPa>C`H&L-)YQw4Lz5Zu>JeCis_ zAzE*~YzOT4`BZ^IQ%^U-?;d6S=5Bj~nOv6aS}gs2{V|a6UhEIz&xk-&D0C+?GnyPI zeYQ1w7a1ZUpz#ky82~>++_3kCc53yS@~u2koa?7uN6GH|yZ>$aevjpR19s#U=T95Y z@dvPpICeEP9Sm|j!c&-L@iS>$*5^Ymx zr~8}ZkiZS@Ka3aX+8XfDw#mk+qAZf72OF2__eejEC=OUbEa@P<_)CzTyyA zg2lu+$KGND8g7k~05?i6KPW%pfH_C6Z<$CdV$G^S-oIh0N4l#G?Y?qC$72tt)6yz2 zkf37fL|TSmtM|(mG*+2W2y}np&MG~8gzKvhUz)ss1MDWvPHpcGWy&3*MAh;X&vNJ6 z! z%<_MR6CKv~eJ;wP4g0g{nJg7!z zV8Y$jsb@YtcT~a-!U|B7f)({(_^5N;vM#e>-cRsI9-$gtxs`-Fv9sX;K7CDhbAgNu zc^%?M=4GtY?XPVaFZCOhXVnF53Ur`XZNwc=i6^^hpP_j9bNzCqQrwjgcIAfN-%-C5 zBXbLOQ^oB6Vm|0NBldw2a&$+HZsehO8`v2(t4w$Tqw6mF@!l#GcO@DdDJ z%brNhcMH=ig{fUJzq2OnAftm&|ISfhg-$kWltsD?h4+>1SL}~<&{zuC2&R@@a^87; zs`SwA!Zfwc*bV9y4`8gz|5cctEaR1mad8^{hkyvvYVys11yf*r~hU=3#+yn-jO(I9CVT!#2g7*Ql2@{8_u6A5FrJ zSTA4Bb9x;E9&6WmG1azT!3Na6g`_Hs#H6}HFz-!>=nMrNs2cdfPfBLqxn5lu7?aXd z=GC?qk>Oa0yi{kfMdC^y=CBK_D1f0j-HTojoqS?0#3*`ya+Gz4M z>l+Q+l6<_h5Ui~J>7_zB`;H00ub%PY*myaNJ~a~7gpO1V`vrE#6OH*_8p`!QHI#$t z|M=&B8p^@W`hQOU|NJ2*2OG=(wT5~^swHnOaDaf2sQ*Efs)Lq-Jn!*9q9z7|3IUTo zTi{*rmP|l^kxGnHDdDW3ky7p?BEtVJ9CNwlfBjwk=uK@&^2qhFc$wOny)p|8R^OEp z8X&NSR2L_i10fR5M-!4B8wUXc0}3Pw63S;Hh74zhgAf0SMRm>)*;ZyE6S)w?lDSc0 z!U?7hp2sRgKmne63IG)%0ZPa~lhT8M0)qk#O!7xfyiox%ucw?M_^(1ZCq|M_C4xX+ zoE^_QH4h!DeR?Yi#2Eq!l$3&Y`iqaCem7{2i_&}+e{%f4$R~p zRPf`BMT|dZVB|ps1ln(mLv8VhHBfpEnI004Dq^ z^Gf-lNrH9Y!Gr}O+}Rn3EWi{(+7Ap74)Ubd0zSWE8Wel)NA%cPUT`rodW6QAD5MHAt*4Hn9VvD-d^!>-=K?T-y zD&S$(-^VOTg+m0Bf`*C;4xAqk zq~s8hJXZ%E;yAH-;MYw;bzN3h%Rc|ovbL2c!GF?hLQyVBO3_-LV_ZLK*GcXfBR*P7r=Tci~EgO7UCiTPV!HU zUN7sDefcT^z7Yf+1bwq)&?l4VPy%}G3;jW&f(bF&5&E_4_(gdCt@zSR`&B>t)laO1 z#QWKt{l?t?y${wmcFYrmi*e?k%b=bUHn9M{-&JEhGc2kCZ3*{k|7p_%g&&m*1{p*y zvI3st7vAgXbHtKCKY?#q4KE;8@UJ*w+&WtA9x#R~7eE33bQ>}%4Cw!TFb;z|2qB^Uz*9{@3VXN3`3*HD z(!?4Bea3y0tUyBa@PK=4D`oQivj6QR%U5!s3UM^XfKEhk@xxy6)2Qm-Bfs-PbMPu* z<(V&JltG*3eqEZC#ue1Hzep^s`F7dZhx>kyikMg*gX=b&9qlIUoq#ux(PR`dt$3=T zCV)ai$R+G?3SrqKJ;xs#VJ%!8e@%wVkg}(DqVG_dvZxyz=@REK@S*O#5wqt@*aB_` zkx?AiTEw1nf_cy3T~nmfTGyNt#Xy`mxNX<9_=9De5o2g77&wwg(>k`0^I5GuY6b}Zz@o{E?gM6@bi=#PCIPKfZdsTE&!%_#PbH5d;G}O_ z)c+o$g_^Fv5#wWM~YAv>66;8qu+Qsqofm^2tyj5Z!rP?FDxA*+Lp^ z?wk3aU!`t^IiGr68+zxv@-cd}r~NQ$G9+n~MXft@|}Qo=!pDLdKlM z6zu&N8X$ExGD!C7A8U%}_9p9yD*!)7N{&;rFECw)jN}CjpE{Xk}z#>IVy|hDSq`#4h{|Z2~akwVidqyto<`@mMLM zt~z&xbT{kECGpo{ICz2mlwi7VkNJ zfGA2{j$X!jBNMhoRWU)@jd=^1oGGE9b7`&lpthR7iXhw)#TbVxNRN@mbla!KhtF0X zBZ@+`s|7d}z1QxvJksddIa4s)u?tq&^z!<;X~nBx=p?gE)cUPdG)=`rUHy&lpqtqa7TNZGiCrnN?pj2t++ef`K*KOX_`T(*)(!77I{Tsdd4f{@I49QSJnpd){lO2e;$aqvkG=)O#2KDD&Qm??( z5MU*)zwI)gT(bVDOfF~(BE|3vboyH!ujbV#_BaCmsgJ=J$u&TVHz}^sy}cN%Y5JJG zh?r04p?wjQlZ(rP0Cv)zpZbu;L2OF-v zFGbM))wAzx7vMnJ0IwZy<=IKe-lrVg;cXVh98_h=86QsayAbzs-%Q!Z2E;DOQs4kr zJ3YN>!EY%z^hld=**jtK+WV90tNarAHdXm=yX0ywgpP;v<#sCbuaC)R8F)DY;1dqy z+93bsB=9i7HC6s(n4VUZl8ucbKVW?{3`e^)G5D&{0osR^{FA>Yt1v3^@NYdKcn{QX zW*pMiT-0BEZcLD@tV#aR#*h!sLGHa_+!A)ojUe&<28NBhadn>zp!+%`azqZlT^@#V zjo&FBawRzfckxh2>h>JzSFMB}7M1$sPTju6KfIsph|>8id`N7{z&HbtRMyU7h7C4h z##awPZ!-?`Ssn3D<5uPWs^!Q2gN&BGJ{DNXRNYr`4A%(>d(G`U%VO8*-cij-Usrpo zQ`~i~@ZSyYvuhvg;bH~Q^UVrF3~RJ^?zq9*4n^UUZU+gv!;35QZP&C*THbY}UU+F0 zC=!euri6$qrtyCI07s61PxVyL#3sCXFk(6;xqWz?gyiNiDc|Ow4ms5>6|5t}e#pS< z=5`u#t}P9JMap9PmGpj)9Bj+H*}@01@KI56kDx8fE3M;Vxi^Ug+0MNj(9 zjP>iuJ!+DZttr#eCr!xcA2WF(B>(2~0gqyE&NmSbBA8|UB5x@zo6RPvP= zdGgsS7+DW2u#+M^7)4oHcYS>%4yB0~I!4VDFGB`5JL=W%Fa&qbu=DWC-CvG53khQJ z+8bYb2SD)>_ZBI9fx#Zyzw3i#Ed~osu@PzNVNrNUBk7S4%IS4JB71sdg;KHLB0EXP9rnTp5F3C1fS?ND;|VKAOAX@B~LYjvs=msij1~ zEKLg^Mms$&BAVrkK;w3Qx_Mm4OGARkB5Js(Jiymw>`^yq>biOa&Z8{} z2FE1HO`|`cvcf6$m=t1r_ugMZ8^+rm5v!ht35wN~j#SjyQ4YVD>nkGyfPtS%D%89O zng<`KKGX*h+6SkMqK7!`vNRseC%e6veU;mwbcP_;Hs9rtJU#%fdFKDNO-hxHyJ&7j zlD4;O)V!+x^{vvI|GDY9l;4ae9!749o+xv~BNO$q#d4o6>Gc3?5n9cr9pLLPSUC3+ z#1>s7%AJ5w^2$Ckzbm%@lRDhX(b16czAV|?10b_)-Awy9?GGOGJssH!`VtGZO{IUYiQ(Sy|8{|4Bu33E$`(QL zr~k!WADdAhEIg74CO+WawTwu5U9`nvJ;9y;0dEUYo5aR~&p~%S|OC!s190L!ys<4cnI_YDxF%W zaUo`@JxHnH98fvUCs>!w&QV-QQbt~Bh+o!qaeGW{VPAWCMU$ert$ptK{w(@IUOAaS0CfXfIMh zQnVU^t>fHKDooY&Gv1eOWn?by5bc#rR~=#}{XJ~@JgIijU(DYPI~v6=bl`CCW0tCa7a(_c!ExQxiu|v!`_(LED8#l9KD@tgPCn(B_NX()3Qy*wFMQM>rDV=9^Qc#wl52?0;-=;)`jj& z9Htk`J@f$t#S^7rcuAq4^rBxa$zJ@ywTu)QI;gfuD0+vOzRm4Adl8d$oc1L zJ`_oi<9%>BcyllaaC0aNgLwrynA7uA7F{U~u}x#}=kjr-PVz|F9|hqo_sTN4Bd5h# z1(f`i8NWW0IGk6NI82eusZ2&`=Mc0&8&B>+;zfhFz4OQDuo?Mu`jDBNay>o-?;@zx z9vqe&t-y76&htGyDCzYxBQWy;b&5zc&WLp5Mk7h6LUJ!ToFF9d-~v>v%j#;=^E`G; z_Gp++Hk@fr2x*LN0Q&#Zf*py`d$-~7+EHoTtIxIzu4`)Vw23J)@LX6kk7Dzg_8Iw8 z%S!)JoBiikn+678gQ6DXD$;ohjr>Dg!6O4zhjlY+qIoUqNsKjSzjox0C;45vkoxoy z3-6smsef~;X<<;}Ohn4AE1Ys_Y53(F+ztnPXIz3_Nms+jv%4e80AXFnBxS|-t#Jyp z#-_el_eFk+Go@JtswKy>B%NCF8;DBJ2BRMYe$q?jFDv&OtlwXhJ|JGRv@r6yW=j=* zoZO3+wu?B?5qniOd5vcwQL5v>RqcxkP)09saO)fJ3vmo@Ne(xTy8qQ4MhEQ;eST@| zc|}|9#f^i6^egvA7SUndD0r1*4sEhO0t}lp&jJcsM$9l;f8o<*w&Mk&@K4xvtoZX?4I6?o_?d$m>eAG49Mhc*OLZ0T( zFehNiHUV5udh0hl2upt)PSUGbd&o#Q(~&d_T&&y}%q1x0rEeyZuoTj~7Jbw|EHX%B zA8O86T%y2)ztQ2XJ3urJ52yn6t~IGH^#7uwRy|`DOf)8+Fbe_Ai18DQP>`>&!lm&-&E)2DP%R&GJ;*sq#}|^%$Q1 zpopItx4~%xu2>0i_Ew!XD(;kTxHh)dj7zu`)-5cW6FdiQ=7Nbzn9DTpR3?kGc#_hSzKSg^at1J1|mO>mP zyZ}29w&*B1DS&o!rrbxiMMG@7VYU_LmN?LOvd8eyv`bJKA3MR(ECaD<@29lG`__rK zhj0E8MDNk(eNu`Wmye{l7d(LLyh%+!>&~UO-K;NqRSUTTuXMeiK!%S>lW-a&kDu0k zx>Y2?i+8Xt2S`o06{()0-p9y2jS_v6%?!%arjh4WJ`LYOc4g?dr=Y338X;z{I)?Us z{Zo_4k(F7dH&uCYR(S0R;aJ#t=eLOy3!!@uP+rl8_;5H+wqkf>^QMs&K)>$<-FD*( zqWjc>?h~jQ(ppX9E0F5k{N5dZ*gQF?khwbNKhM=NlHAiS&igXD7wJ^Sb1-kPsCc?K zjjZ}dY(3_k^DIrSiI~@P-0dW%95WG(f2lk7i{npMV`ZO^ubD{Vyfaeesaq_t1Uer! z^kbu4lxOy785m1I@DzT{E0QMnvO3vH`~#At1|iZzp67N^ApGw9=l}uHR%V0iz?6K= zQ*)>_lCpvCy4|^Bvsv4?j**thhbw=3Iz+xH&g!nYq7!$zB0g`8@l!sR%y=>;n|;N-#H?O|0}}Ovcjz+zMrTLiwM)F+cYq#XcTnASt_6m_ zuJ+v2GS|UU*nG;Yt38x{SQpt80uCCSrnb1jtN^}tkKs{%;dPQ&c)*}GDX4Pe40cdO zDL8yk_d?+96H_Vdb(>m#I$3U$SXU!Cvx~RwLesVVAY>QBfj61zZ@n-f)qgdR1BUXK zSzf1Bv!zV5@%WOk+g0#%Ie0%+W&X@@ZVW zpA^U?p4vx8$@ZF=KODEv37CuCl9D&8%n1vH)PKn%p<+Lm}CwC<1+Te=1 z%1yikA0Vl9g#4QHq5Fu-RJ)qFbt`!4K>IpesKIn?iAC-#=P<46ZqUaBVJ>%vRfXUM zLVAe+(Z`O;5NETgL`|USaMQC0)0iEx?*3<({KT!D!s16IvRYOAdd0Er*xjgKfllh#Cwe_gO?bHJW2L{+AnsHR-Zc3RUXes(H+pt-S{8GtKQunY z-9ynCpmp;05kom!e<&ine5-FE#(D*h?rz`lIe7eyV(##fv3<)Eq0{G|sMfSk`de~q z89nh2eBibZ?axvPB$3}aB#S_l`vJP6hQe*gN^tfAdgATNA{5vatkbTCXSIXi@CCuh z&tMw1=fC57w=-t+%kNi`^!_letlSsN6;?(SYcGpNurGsi^5&(BX#=Qa>@*L~gg<@T zTf|0@EN7%#FDkBorF883I6487v&Pv~A*XBjf8S)h-4+a zv`-bum58r__1a!%9vqwG4ZpFqGVHR_bkI9*?(dT30#~R}+=-upEo3d?p;$E zjQ-{dR!z%~;?|@!fyxI?(ox`{P8!2EL1rvHWfsB}acjoGTguBF>?Xi(l zH8E3CUhHOO9b;gns_D^JoT&`wtmeJ8nDC%Ma>B|P{}o8f9Tll-xy}MR?!I4pUH#-# zCD4hiEbrZSjZWlE#DzhRTQ4E<48hZg+%O2l;uUc?;)_6q<)YDC(JU76Q7<*s-%}+m zK9PnMWPV35buH;jd31|rL;f$Y@6N+_DpU(xP%z#1?a96q^sg>$5&E~>xmiK!h~ONg z5VvbkuV!32zsn#QAdz0<>%ucedpeX>-^?uLMmz}eY3;2=n`RPn5>d>Iz}e|DC=#x_ zqv%=R2d3T7``+XTK93$oGxU1-e7GbOd++Aiwibt($Bs_p1#STw}WS8URaFqWzr@&W;+$Pf>g^A^Yold-Bzj*}9swmws z*RsXCErN<`qFqC}7@LSw(i)skZT*M1Whs4V5|p_H~4J>ms9O3Hm^s+cjCsx~qpdE_#;KwzX3(8MiBcA(R%>3~v&;t;5TfB5O|j z6VllGey&w@FaI32ZgTwJBy$!c{I;=lSi`8r;e7#df_j_V+s79e4LJij0aIN`>I5mP zL|cVTM1gsTZy!d2N;N~#?QQ(Ku-2_Lk%)Yef*@^ieNSus1X++<{cw+I;5FCaA)^&^ zPCf*EKWta+e)t;pn~1NI249E0jDc&q>4oBf96o9Jne68+TG;aPf`{(EW8zAGlJzFH z;#>>seU&c%p6c=EICLHuTa3a-rflk1*jPXDbNn%DT+^4A+`(3P#9#^6dD3s=e<8K$ zTB1}cbJV$;a3a9%%5F4udY>-YSM%Fj-&!+{^hVv3)9SJGNd-a@7%`#yCq1mtph0YD zJtfGfe}pVF+{cz7j|23BWlGv3zC+xA4kdfM^Ir`Ra;0i1x`)+G5m%|SteY*C{LvY| zxTN#X^n)lfOqJdJJeF9S&_|-ancfwFeZi4E@UGte^5Hh8{>PWiSBqpJ=!tNX9bjNc zgs-&KAb7gH6y?s(r4;fDa>BmotWC4}toLV=Lri&4uY8i*B3L!SS)cLF3M|Y}OGpZ$ywbsh8XOG;1sw?m z1-QN`dytu-zK`C=n|2S6oSapdX8_6k^oK>q81mX>16biVg-XEde<#1sMe- z3JO~KH#$OW0%*BE*O5xe5U4y@tbY#3UA;IP5$0wuVgT@1CrN}Ef=fk5Pe1X>i(h^P zBRW8A1g1n7;}p_i;t>*{A3z}mg%@`JE(y)REeSU&FQGIyIVlTt6@?*Gh=TQ8Ul4YU?m8t3>#xCLe%e-Bb*Bt*9d65fwwWfuks0wW6C zkPRx|{2TxQU>YWD+#v)Zyt`co3Iz}O)xU9gsS|M=zBOY43vzV>CH2{Zg3RaN0*7Fm zn?vpQG<^#&0(-(lz6z2sc^t?!YKL7zkbT?c#wstc0|PUK{Ng3R2@2{mo&&LJpR7X$ z_fH+jZR7!|jB<4gC1$efulrfrqrw97z7h!X2drTKzn-kWHiWG7X(%ehzXdjpg8}jd=@MWZa3HdPf?~2Dl);O10`G+OrM|N?2LHlK^#wbsfV6iS<_fxD8Xf@N ze{H-*7(j?~1r2JVL{JaZ_ly5_3PVH(7cxkUm5)CK90&AO#)A>I{;O@uNC^H45foCt zfeHch{qglV0YS@1g|dI}NceI4iO!}n-&tnS|7m&PKc=RZup=bsr%X)L*WAvfq@^W; zM@3By5BFw?Ez0!efGhGlu`IDh8o1!!IYO}5$Mf<<3sl>8vIp_gw8BrO(ZWQu>MsVM zCu0B`@Pq#Km;Ltm_bb2OSNYW^_OlCb?b+V`h2ZiX_e*G8PrEg8$_PxWg9x_*^dVIZ zkMt|w67bDjBLD~U>iB(I&H)j_9>g*TOl#q1%nVH2AjpMR1N<5$48M;RaY#7vH8Aa` zJq;cgfK4)gXLyTYmzFT+KT`h_#AUimXjRSNA8w@e`Q>8h1_lQGeawpN1Wkk$3i1gQ zQtym1u+x^uw348d@#;Vh3J#7W1;+s%k;4pR6)y(tr(abJrlfR>`Q%LwhUOYd$~V^+ z=zj%zFH}(P_el?Fhx)7Y%fShE0v`-yQI>$T?pStlSDGwi4#D{25GJjznzBPm{9qQ7hJT{QrtL{x9QQj zdjix2E55)BCtEr~m*^nK-1*;-r5MnGj}5aMGVwY?r%tbm<9d)mOVSUpD|jytiK&8l zKcB{o^?M+u?Zngz_~XDr7F{j~xXdflG6!z&aCGPxVHMlRe9i}5-~p?^)aTN~tc-!A~!|qk- zm6H26yNmXac$gMY18O0*H}t?>dRVRaG=_&dBf}Rc9KaQ{Ot#jkpF!bQwkAGZk!Qel zRbJmrU*TPoc{tKGa3f;--VQPARaTvT|Ji9b`DFU2W6xs4`r)V0GhX zH;8za--53vw~wgub;G;B>_Yl=PVleoV62t*22@yh{4bvL&=d6S;3U_=0A%hSE&1fy zNIZ$prG-NAHH>ayyIMr!t+5Z`p}HV7AI%vC4)}*3{OqEQ2c8+G<-yv< zH3BvKA2F2u%uSu9y9sm1y7PpUPQZ#0;CbRcsrQK6C)I57Xr#IhF`o zoMgR)@%W*oVtP61-IBi&S}>nG5~B;U_X%knkLQBPIvH?!q9^tnKpIavOAog;UW2U3 z5?Z*RcB=JZ!U;x@s3jhVuHb==Q$V594KjVU+Ge|_T}_M{#2ysYRFxor<}6Ehf^~4o zr)h%^SbLfx+d__`r2xv=ML>bZ#nG@)L0Q61@~|o1;VM)Q~N-gJ7TA~!I#U)g({Zs4V1qj zJB=h;QER#&(b%z!DtQPTEfuvl(aKX(f<*P#(&Cc+JVJU|!KIXFc~(RCUIZkrd{bfL zAVmr@b?>yIXDP%_XrXAXET6DXP*#3Kvs)q3a8_%Y5bL<+9^93W-ybFrolnD3&KA!#MG<+=P=59);0% zJ(Opmca*7`6j8Kh#D2>8y!j3n zRp8MUZe6r4c>@gf9exKj98Vg$E`n2&epOf7wY$MVZ6QQ4(ed9%?#!YSAG{W(40JI+ z!Sm+F2A3NDfhY=uKxLUWWBa?0vB`f5ZV(M;8k=AXCoN&~Pix}uc|TW=8i!kmgw#Sj zCNN)%ZYxdxxKzoyW0|F&)yA=lXeDLSB9qjhPw5EYEgu}aYu`6qO08%m zx&md{F?7Juej~Gm&EtNXyN_Oq*W&kVgpw_!F0Xc4PI6mQP~WAWin>PV!~q2guYbk7 z&dH7Qi@kn_#WX8rqF$Mu&a?;vd$S^jsSe+rfqN9mSq^LD6o!#p)bO~YH&B$)7E;|D z1e5<=3Vlv%E)gU} z+}zrxe)0;@- z1tyoIC<4rF#TT3iuoJi4lUf0<&Ut=6KDl4-H)K~CU*8}Q3Wm}H$iB9AHB zoFKFBKL3(8Ffz|hk{YCyO*HSmw&Usg>32N*)Hvioz#i(6!cNQ1D7z^C+yWev4>kOH zV-jh5w{f!Xl){$Iglow`z>O7$y8c_L1s)UsvSlUKY<+WPa8l_w<7FM@wkXM8ElF-$ zR*k1b9uqLrCiC+Am!e~()#onas@(XvBJ8@Tc3W?AIo7(dzSNQftKtgXg0;2$`HX`} zC3WWqek&U}wq1=9EWLi?G>zAeICxcyTeUW>md>mT4l0zIql1B)aR z@0`_CN|D>;Zw7F0-;b?vKiwM#Z{%7D}Z+Cc{_)gKXWA-lz+SB?rf zfO*F7Z`=^hHsbatsYjwX06BXx{(&}_f~Z%!$qx-5LyS|Hg;6vXHdEF00U@OGev4?( zE5VW`=cM!PWJ-Iu)cl&3!8ivw&A}ki`YK#h6Y{Rm|Fh`84@F{|jzXv9L|9tOAzn}JcAH<9i9H^UDG;69cEy*Us08N5f16u~t-q#`Y?zoW|{egq_Nz zXr($mknmul-N@Zg-nw~7;za`3$s=3i!mJ`0+=TxW0ldm(Q`)J_yd@ZdZgYW+?yhPs zm*<>kd8H_C*kqerhXmTv;w%_hnm?C8beT3%4&$z?7((sZ?8+0(?DQ#KeB9xxtmWhGkuTqi{r`utb7~HSiP~&z+qP}nwr$(CZ96Bn zZ96%!Z9ADWQ}s>F#d|STebN07cGvUlwe%b=EB&-mo~?X)QJi?c7<&vhRg-&83J;%@ zq+N_u;}jk2cpycA0H0ULcY#8qf7pfhwQ3|YLB$C~C>EB)e4?}Pf;t+%1e*W!YlJD4 zu4^i%_A(#$liWgwf6pn3%dt|!iE;elaJ7S|>Y%us6YWiHmFV90x>w|EN~>!wHZ68? zmyEG}oc@QTN&mk4W$JPT4T|= zhi<#Rvi1))t~F&uEVCT;HMeZjeYWxhek(G_jpwLawrUNj4Yh3!F`tNI`nZJpUY7W? zcNrJ9Gx($hW_;TVA#7es%bM9QsVg)JE)a>+prNKL(Y`vHz zoK)l%bWi&GePiF;bxF44+HcO;Gxang6IjI%02+YaX+$CJZChwG{m=$}X)Pz0a3bTn;FYv!&d|8Uq#yUtSNFB(m}b5M0Mr6HnuKb)ITp;?U+xkj3F0j_fP$^WO9S} zN5;LEZvrlG~e-UNo`>&ILoji^5hG;$5bICIe9N3sarg~%ywrS z$`wDzt9ozcwpS$vgN>+1&x=GM^$PK;phfbqHhAx)G%W0gnVs(U)0Zj`J~Xk;W2w$# zRh>wQ>Fj{E-=S)$Ah_PpW|~$!NoO*Xs7FrT8a^KaV{hk~KZJcwGALahG-V5SjsxVq zOPt73pO*cMx91DR*TXoLEuk5<*^8jrhZ+#SYM2wkk)nyX=|wy}=j2mhNhgTpD}=0p zFxVx+AM-*2ywa!VdZn9@04;g#zUfVnfUnw@)g1F@_<`b`suC!M4AbHGD+#1NHaB4B z=fl$RVb(TbJ`z|VqE7w7<5f6K0Kx%lt4-fmG$-~)_H?rt6@aL`%X{7)vJ2wx_KMeG z?OoY~<2QEV5oXs;xOPY7NEdk$W8{*S_pVD19XhV{2bt)@+J4rdW(@PyZD6JsUq^LR zUy7A^n@Psg;}_OX z@Lr@L&X6p>rK5R9r=YwYv_bY_(s(^SfqN{nWB;R*9)OM*h`=B(SF&hfmDLl$-GG9ari zd@Bc-31B6y8&3$8qb{sGH$E80$Ob2_nG6Wd@i{oi?)x+op(?~*EH)~w-*JGz!xP@zkiyB=u&L*|?ghCQQRZZ17 z2Dkp=C#i1ap-|5NbJv(cGjJ@1bh#UUdGPs?T?_j4BA%KzqM;PXNp-1?U^l(Rg!J;3 zeB&q$3f2JN?|}h%;j3Oz+l@C4eL9S-cv(o7*Rf?6_MM;u>d&h?-3D_2-~N89cl=X> zpk4>Dl}CU7EQy`Xc-;_~I0=k8)dJp<(_=S!X-%$U4q zgFGk-X6D7-()uznKNk>93m|g~lad6m8%bhfGVDhIBHt6Fz%#;L)Zn-@*nI6Dh@DHuDTpi8Dz6O1_5C0Z$M>|mD}@_h(bvTaWV zgD+a5=@}a3wH!%ahqA{j0p>|52Blv`1&ivO9H8KU{i|j_Ky|wl$flVqRA}pNdFRDt z&$5o{XtFXR%qR34v<9^`Ol>Rdc5;1d)n|R_>SbX!-c0@uaTG~;2UJ!|(M1N_nB?el z*_!p_c2pOEZ|h0A3&WPPtJh~%zSXR+E)kOKC`@_f#|*t#_owya+q11AvT)>Jy0E!J z`I86JLh(C*u_vsV2ZH8#=b`%W{BQxg&Q4kpaT)x4{Q9FMfBOb)rl-roYl5g35sol> zUGYcG(toO`_7Q7io}9k2e#S1(hs0lJBrrRLQPN$4Kt;^kb$OJcRws@53fn7&G{2>B z%X93rWh)!q)p-dFo=@&!+PgP!)vng(FV+>M2K@U*6RPmo5E%_gq(8mD_kJZ*98&gh z%E-do<{i7fx|njh9fU{wb(MA8QQB%ggPIGsj4~T3a2v%QHr{FO#?a z>yENO{F~|s1c!D~fmC3RgUAsG6 zcE|M4OVDLlFb6*-Uk|1eRxRV7QgLJlK7$$=`HOom@#C1Y%R2FZZ%1r zGRcAK>u<(W$sEjz;P?%4HWQNWoIvDdpWdsDm0bq)Nd4pA_ufLUlzd&n#c^$xD_^C# z?&HLZ+~n?+o^0<2q_Wywu&t}nrKc7l=N@9J?^l<}94x%}y6gSclaxB7r0PH_Ica;Z zt3ETBW?X6mJyO{>8eu2-v%SfQC>!xRxm#Gm?lfYe=Gvap%^bTk%**huvY)Fqst|9Qv$C@&KCHPd1U( zQ4U@waum_JqCF;3~4Fbg#=ts1)s9<)USweR?u~YHnQq zCJ+NReB|ZZ_O!BlhBGzr#l5fEVBHc+#X+oB6HJo&$!oZE+jJu;MhUrIv57lMUxNvf zUX*VMg}5OBlCB&!yq!GV;;Zxhk^OL!LwXc<%C3}S1?^ysKiKxy_FxxRP|g>!B2|qP z9iwpD&1RzeO-Xt9@bJxz1=AZLxgl6?5l&OSecJ3 zF%2_^=A$kr*P3H&LP_mmn2cUlYWqV@$!%(_i4Ka*H(yfSW9o%xw#M0)*XMx!O4(npzMpm< zT8mc+Av%(s5_jUL*l_wXn_V(Aecx7iL|rv%Wo6d75`tMAk%+_f(&f` z69Y04Ffy|-{m<+FT{C53VqpIN4h5CLRgllo)I6!g9Nk3~1@GGkL?uWm5Ilj9xJD6F zvY1GanUbbUWFQ#B$m*fcbWpIznx=%&MHK}=po1zXCXLjIij0I2pl3#*?141y{!wPQ zIn5t_U;TE!?cZxhWaJ8Rs{o=y)-j?s7sge{_wdh7D`h> zm^5ky7?#huL%MLn90pr42A>6BK>4H?3m5@r1|r#ZBc(uHs8;Tw5t6wBM-*U)0DesN zf(z?|O!ng>SDZ+sLbwC##0^H;z*xk`-dn$2_NW4nA@GF)`9q=&An%Frk%9$c;``gZ zBp|5MAjV4NP@<$tiG-0rF_XnRG&X@ygQ5Y@)%qVWGytARg8?Lf;M2s&Gy(D;6k#}k z7zt~Gju;4Fj6&dxp@ND;KVgIA`RhfXih&LK00U?-Y~TtKE%k?B8=nIan)6}q z-!nv@3avn5frLb0f=O-oned`Ffgn)N0uwX>5q&7YgoYJJ0MF!cfo#@F0f?w@s^tgL ziIsE&)d-AyV^MrDie~UC`f-yb3{cb)p5AH2ju`UKy0Vx!wk%MT3$y6i34@?QKA1qH zC`lniO{D&M-xGu}B=)Ws8Q?Hf!eku+aRrzH#2rUde1LmcQEuIqoX?2BYQ>@?0}(2M zRf~o&+>0+#6kuTNIgwVc_}pT65ieK>1l&0yK#CJmoNK!b!a#%H41XO<&_I+TJro1} z`FVXtk)x0aC^ULz_xO#V)LBa~#p8Pflm5AIaUL=X9F3a*aeHwQDu?j~b0C4+D0JjN zV9FSC*RfZL2z&=VNAmqha+)pxs|G_K#)64O4M3a>4P0R_3|Gul;kz&V-gUYy|6)(Q zEdG)&{LMhrrbK)FtN{b~-47>;(BKWYQ;0Hrj}JJ$>mU3W0urHoU$~4F9)@c#Qaq@; z7DlllK|^yu$f3tdknF=HTjo!>418b=^pic40SA~kPRJ4=Z5AL0D=S3&BndG7G=-or zK^cOCN}zJ2MKE$o@*{zZ3pglbPYeq=rapEu;{G5P?|wht>5F2(fqsa|_v*1et~ou&$CvexoV%os)iV@| z)q8iiu5}?t-fmbT4o#G|S<}m|zgvN-BGg;qbC9$R^Z+kg|#a?5OK#gg>xYSh+s>#cBNc#drcv(w_KdB|A&$BI~X<>FkR zQc}cjpMPwrZa)~|%zbz%&yH*)Pwd>*kvZ7KDjyZFm|WO#PMCrRm1AAo-5=n>aH2N>Q#(wISF?7XuXcJZ)?E0jc4d@s7%Ele0Yhn4>eXz z6pZ|d)BVJ`*l;pQw{YZq_LaP%+^YT>rK@l_jFc|xyXD%M{nwgQGSd)>gU5;f7?@eSGKiA&1*?YQV>64U-`4# zhObYJq0Q<23Rbs&&8~BXM{M46iWTph;9Ta5bzXFLfgxuG!OuFBVz12OJ9O^0;t;+U zq>9tk-es*{X}dkkCT7k@3ynLlZQeb@*U1+(+4&To=Cmf#vQO)(71^WqVdH}0vw9La z&qL#JG3z7tR%G3h)F^8!pBs&m)qW~@Nj>@DV(*Tf?AwW{Ox|ELpWP)Of!_Tt7erBr4-ux35ouuVT^%UWy#Bxhj*>58_+u4qpZ4%i#(<48D z^r#fLJLive)4M(HWuokST*qq${v~5p(rl&6MxWJTEbcC&OZ|%Z-q^10Z6H2rqxpg# zj$X2;9F#?Fb6E}3X1SNryCM+SRvqOzl6RX$+_4E9`#R=7>&38C=$d76-rF&d>Ntz@ zz7?yz7Je6kX|G4mHNf1nrM+}d=XgBE38-lq7tfBdoDlDRgr}24a^OQ(a~H=bIsc#s zbDR5@&CgXvO3V9@!RJ&eEKrtF-9)(^gfy25ZCORXzRRav_6z#D)(N(`)Jpd9%y1jl zd6dk1Z`Gf?`N+3Alh1xVJXNq|qxONJcsBV^G_F|8ZW~PS^+tWilNn|`yUB4y=`$$@ zdZh6FbQC2n%aVbKRc>kJ(+>=*-N!fnZCj#dcHrRUia!pe^e4r`N#3j54#m~>cIxY+ zmO%09>ccYp&|OUfO*@e$cl9}~t#;A#XVRvf|1fH)Zauy(o>hY5l}eo<)2zdM&VD-e zvD)ZsSlecZ;zUY|qS1?pR25A-9@A0MU$nC;-Qd*KtpSYgAD&Ni++Yoed(D$0W5#(HyuXobL zz!u{Vrm+j|%q5L#_sag#J~O_lD&~Ecv~B#VX%4-N^9Osys#fKHjYKT}F%q#e{oi5~ z69E$k2g`qtM4ZfQZ2xB_f}$6*v~e+YBA^$uF?2B%F*UX~F@@segK~CpGBvb?^4PHH zbat?xS*69YLencAZr`9?!a__J)iY z$-UlP_1THVxX8H3{0OWf0@|91(WwCt<6P5Vd|V`FCPt?Q=SnnIXQmwx%d89lm027B znK>XcF%dEhhTvfDpX1Qf@EodurSPK-xxmia#LCqC7)HUp-qoRvsR87!v#Sfbv#Wi9 zyX$`GcS&+&8fMSR5{#j(krg=(9wXM1}68Q5!6#PAOj-!4@Cd? z4hjMFyS%%%I{J-&0-2R;$>rHRf;KbDyI5+^@zHgNzJ-GBvV#KtpC| zDlg1XpGR^dyE{Pr@0jhy$s2#6KS0MJ(#rJC^7NB`W9(J6P0x@2KO4S0t55&6m*AT&>zm6K+ zfm)tfS(IAsuY#SO$tS$(?~8TDx4GYs>KT8}FD2-kxt=WUm2JS_cBY)b??-UZ6+uc*4vDpd0d;>!Ra0W+)CV&o2jNdQl+VIT!`t;88Zu2j5 z+OOrWww#$+o*5Yfb`=~MPvI(N&3@{S(a59f9?N4>m=#NvOusdAk2}}6+E7{<*cdxU z{2vZbJcl=D!BL;jHE*SO+F>zBGlz2Jr3R5iuPTn)QW90PA6M>tsZe~_#E&+Ei@$DU z4+idv@U@!R*aK0^OVym7>8^`%$3EjoG7Vv;Vdl@ zk7C+z@2xFDl(?^3hCxB%)e32LYfmG4+D z>#0IQN1C73^HF=yvi`BxzRmpRbciN(ACtdqcIP`jX{ZIel>RvStC5%oVwC~FERQ1w z=NaZPAdF65#g=SH;dbZl!_F}uWKesiY5eRHM3_N^#^yJ0&T}WV6AURmO{MsqMU9p2 zdR!ZXd|bo5VRFr}cU{^>AGV)Jl_;N$+KyY~7Zvr{9JfMazn0`0DCsAGw7KrHa^_!K zZLa(7mublDZ!*&VeRPE({k!Ww_;Cj9^LIA|MDL(tVISeW^Z-{fPZm(cnoatABqAR5 z!Q{I$5;>H<`vu{3#;Bz|I|hLlN0JPhI!ajzX8-cCI-LV+u6T0^ABi^13v*xhKt~=g zHfJrGl8Bt;kon>9uRxV5*MlY3r=AJ*UPMEEQVm`VZuQsfAk3l3W6UxEzs^u-y6IE` ziaD>zRzc}JwIdQJ(NxcpLqFW|DlK3UM4+xHmc=x4K)uQEcmT~;r9-ib>K+94jLrT6 z>f@S(7^d6|SsG}>-7W$IL^4x)lm6@{y44T!<6_6lC_WzuX~ygDH6dood@yB)lOq}q zm5;L|(z5>(J((VXY(FsHYE7a3$pNP8iuOunVgf!zs*RrP&YQ2)s39gUwT@#(&oNXN z^=o@(1E({BDPPJg7)M0)1BB13R@hIqA;R9~-|sZS%%IrrkYl`SUFqak$+9X+!AHwi zXsfJc#pu#&bdD<8F*n|V8$0>Uv4h1daLeTRR;)>@rx!lZo8{v;r8-K9T~{-ALF`VRqJzVecRx?<33#S!LVN$$Byr1+3HCSf zCol=LN7T|8S#&mU&4oXnSe|mIA!u8Gm6zEfWR9A8}yYX1jJKO!9E zz_0PC-!^kTW@|Qb3AlBr`-Oa0X>W2t^i~Le4Cl{iu86j!Z9#)D-C8u~f`lzJiGQnX z+>*Tl94j>;<@md&v7$sBKuG(O^@SrW40M)oqv#Q{@JDr zhf(jn@=A2`y+{lS&c?thsLB&E8NEr&BjonI$!*KsvmaE|_jww*pw+mjDY%5>u3vBKH+OTRK^G zWxX~)p<2Jsm3741juRF>Ie5>8OgnAdkDSsmoORR&?jaN>_5PGEqj~qLNN#vewbl^i zPcqTGm!J3xP{(o7*M-0FRvbMyrLP;drECS5eSdb4?{*hnJS?eD5yYwgg0(hkR6vpN zs74%4T9!UT@Uv!TF z7?(+?T2>+r4KTY2?ub@k8ZE}P=5C^$@oGNj!~SNm(dKL5r8OnzQ5aOdZk?MbId-Fd zplFI0k*wfB|A}fm-_tkwcHC%xI_6VeMy>EPF2>2le_K{0Uh%D6$IF9zfjbVeSY|;n z$&5B!^DT~S1qAw(ep>2@>a;H~2tH&`5v0tKAEL%F5EC16Gb}NnScF|q1zAoIB|ke- z?15?(J9cj*2qgA8m1MtNUQoqG?rX>CaVe`u6gdO$9e2%9#HT23kwgYixi(^;6UO@% zHKhLWMD$tx+k*YAMR8da`kTUrRdyC`&uPLGiM8su4b~y-ek;n3g-|qehIfXCc8qXp zngQ__a&Qq*==E-x%`zGxlb>$c23kE!ey4?=vVZ2SY&kGw2N(uq&Gb8#kBNoor((tS z&+5@3SODB%cR_Lf++xU4E-`C-X=i*pgtHczF!L2iHd2r=_!Ns{aKkWP@x=FOY#_kA`NXv*+=Xyy8Nx%orL455eEjVwmk;Tq zU0sEtNro^QNB!PQ_S}tW+_}j-a`fYLMx#whPR}Z-5BYcd8GjNLJuLSsZv9Y6Hri24 zb%oS)9sY7ZeAt;#*YU6u_~UUsCH%l0m686l=D=|r;o{mVg!IzqGs`Qbq{TuxJd6rZ zUT47?foAmBAnxwsMFRW$-Ksz3>euq)aq}KHxzN=O4_6!&1m}-B3U2b1#M>3 z?X-W$7={RwQQeZnR)F)Bf@2u=iPt`ulm}DHyp-?s;f;K6{av2~i5G-?VnqYKAGiO4 zx;u8K?Dc}8bv*xZOYKQ>rJIPrHvJ}RKSLv-M^OA@ytOv!bk88%%s^?JuZQ|~G>Z)^ z4ugI8t51wue}PzhNa-?<6I7&KHPLg@DO6aiY<-5i7eN6uK15MjmB4*0MpJ!DPE4xK z9GN%k%e|#^GW{nIy(T(w)oiDgu>ZhS^$XH=rz{ZN&x~JrjCr#yUG7Zl4ShavSx_<~ zN*wLGCY1(|PbbsGZDa2s4Fa4k=Co+_=PyYj^sRVIcxqKoP@OR%61(+%ZR@NJ|Gc*{ zPIZ{q53aG(LYyCJaeg-sYNZ|WE4wS7G@jn+iZ`{l&G7e4Q(sKdne^u7$ex5Sdj29D zJfwv$*GE1t_HEJ*M_*_!rnh$)!FzckiLGu`%J)ZqjXNK@!KPJ_ME{?3$EMxC?M(Oz z1UkCfp&!E(lFiH)7X>h#s6{Ct188YLZB$`z|2}t8Y$*@dB`VZ6- z5>xct!1cqzn;8s+9>W9F73FSyKV%ry@-jcR)AjZfDyycss>_)OyjsNX)~Ppj2`u;o zcT6|iDnO$vGJo)}^SG}U$1^w~Ox!?r^VeA( zX_rt|@|27J`5hV>HDN73N$^O@;&PhSm{oYj!_lD2zuZfg26d2DIw*lcYT2pE=YQsq zPxFkTul8G4ee!~u*^LIt{O2HE_Oy0HeqLBKnr1BAN@Ry|g}SP%*Y-9Hg7;Kp>fuDw z?s?IW2$229M#CT*ZXw8aFihe=HNHZP5bTx`nLys(UrhLZ92L>st+r7LL&D;->?^${ zlbdyGE>68=?73%Dy~6aZ)(GJC0W!^)>Sf`v^a{4ahYlMyR>m37`zqCxPI%fg3LtUB zKSM&iTW#Am4|dv+fs%=j4i;(i1EqNpm7Z6fHnUO36~u7G9+Q>fQaShk(^WTIZDolssC;jEVB4h=7rpi+)yp5V!zw>oa&}XvE#uhV@SD)1~ zyS$`MRZXbsk-i-8Z|}$n&*7Nmv2r&~;!$wj*eK}Z>{Y)y`iLW# z?!e;m!rzZCiXgzSx$9hCZ?4}XC;P;{M1d!XBNsCXYH7KD@%A{Kd47T01Tf#>qxGwy z^tclr(*AkY)lijYZqg9rm`ekO7;RCVY}8})mxL=#3(w0z*zc9ObMz}948^Q7&EiLK z4%2LrQ=cZOGD(#Z2j1!zMQZzcR>EF&b^W+I86B5^Y&qQ?1R&HAvGJ6`2rlXmqMaC8 zj$j-ESpKZjn}4r{Q5!apB{&GlZz+RDb1or}oDItrjk!K}qJHpHLFKRo8p|@zc%eDi$_}ta?<`UY4wHq{SO*rRe4!WUbSi69NtnxR)2UI_0`+%|EJxH-`-ktaX;|F6V?j$W;w_i;cZ1w` zxU02K=rK7jUV$^e*qkAdL8gc*+?uj-i_(g9WhftO4uz$`3QklSkr`_%Ek9=&r2W#! z5B0n@%U(>;Zf5hc#zSjIl%1fDijKSBWBRzmF0OKk>=zCFE*(^TDT*2-w+p48L5TiP z%>OlwGc18VlS}Jzi7k!i3orv2iYzs{-NgED!j(Ul;h?#45vs)!_H@&GVfn;fuot- zr%w^o=%e$j4EmzRFI~%H^n7m=+dL2A4%7mJupX~a9^TF6Y*~Am@Lm-ntC~sx zMT2Mc8vf@mcwhCHBo^ixHF>_`E@On$bud{(<~ z#|gbd`!RM{DQFobDMKs(bkB$>w0AmQ6TyC4?8{$BJ035yrL|$F}oh))Z<<;ShhYrAA*e-aqxe)xf=bkD4 zYIn4Q(Y{!6PlK-glYhOjm&tIEWe|@-q@-amz`C8WWf-;uQwg|7-tF8>C z`Vg5akgEu2?BFnNq7w~c1_|Q*vQ$ZzXq}leN(u7GAv&)g$qmj*V_82d$7$0jSlShF zBJX{V>CPq&j~wJ&HC!5xlBIj-_dpHr!E}WleKIWafDPIGUPU~X8+>50OfwfVZvi9Y z5Bx-X^*Q}8`l*tHA5*jRHYCA0WDP;eU5xkVvwLd(R6^c=H=L@QCsUhzLLY~&&Mjzb z_}uw`=&p{MIvA$8hRo;@U~>$mX_3{qe^HFNCbo7Sk^o$OSaQ06Fct`&xOwbJ_Q{(v ztzGa5@Nzs0X9Pgy2#xy9ES{R0HVmx@YLodapF@TQXm}Qkh6|Rl4vrZVv+O2)>MnLoynVqkk=oz~11DByvbYmbA?d zE5kBSPsB4cizjhw`uAbSPjBd=M}=mb(n4S&p-*IMis`9tU(#rXr0we%x}7 zSsR?qL7ly&H*vG5#bFVJwoOP2()Y0!qeYw!!|ceOe*juX3-GA}`BUzs2F!+4UPfgG zFx(L|auz-h5lM}1A`Bh_nUI+YTefEAGLC;>qbLiuE=;6sh?VnRf1MP_hPAeJRgjb`lX@ z=d;1v_{b$r3^d3?g*Pi-tm8tHWC(J?(Pkn6TFY%`&@1Mp` zj@i5P%(=MBEE0bPhy@>izqls$h`LGOto#0He4*o_;}0+1%$=u5Nw0GyF6rh%DG^G> z8*Zof*z4jZ;XpR62NMRKP;(}pKCqcD(TmYSBS^wij&&9_8Pp3e2(`R;xy%_G0u)kV zmSXwcKC~{*Ak)vo-n%M{IY(8Hg93?@B_r+mFuA(1{B-0+2BqU7S&d!T3+)M9 zy8ACH9ya+L%Sn`X|KNtO$9ZF=IJiVdmiw8>SFm$C7mma)Q2uYW^AiK&Ic-9pBCX`X za?-hgT80Ea>?M{6q(@ViL)J-i%fSrjJd?)@CKu4I*lrfBv1s(m#`<}SJe|f>e1u7= zk`>I*GyC$X!#$4(oz(X`Xge;uQswK#k=A`Z6voCJ(T5BTp)2;>R@)$D<2MMca-s!F ze(HPb@{Ex=zJ1Q%`dp4cljRW!r=f*llrE6A2*4llEwQ-aHNXke=se5jwLg}UpQ>kw znulsUIer`$9(flOCh!(B17ab|h62wRHIj}X6MM5-IywSUgLNA7?n>6coW*7LBq=@P zW=&pxKwl9%i?PNQkS^y^H0ruJdRTAF6HE-4{3Vd_=@9~)ZY8djYN0Q=SbsP^+<~6_ z7Z3Y1mv(iWx*;^?1}JLQRGS4q37WkAcnFfl%RIT(9VMN%4Zi26`oPHv*L1G!U_{(~ zy98y_5C3RTBWj$BEg=hU#_f@q5kX(MX&mq~FyIMuA1&GV^?E|6vlcDkyjM3 z4AKXv;5-XU;WOtCOO4JpC0-Zrjky1C}FVmLv)G=Vdtx^~cV4VqC_6(JKzt4S9l;n4lIa za$ET6T}R7cDBNr|_hw?JBUC>{c7`ca`#Yw=-Vx8A6;1#oRzj&VqAv$+Zd5kFzQoiZ z`@K!9E$}U%g_7|oZ|`&j=rtzHo{Afg)$(D#H$&ydcv3}oOn1N#u7$(Y{O zih5rDa2q^;wLl~dMF_!j>Sd$335+K`aqA4uqFCn=Y!qb*sjsj3BISP(e)$CX2iLZhg+M-;EAH z-jstt*#fH1-e4Lf`1wJ2UD2s?HK!d?=)p<1_#Mew-S!o5xU!XUJdjj>hRuY=DXsQn z%*^1U8z7eK=!1yw5@&+Kq023-pYJ50P^@w*@c88QMv)a#I_sDrka1ERLIL*h-aW5= zq;uoOH z>jD|nV^TsZ%pxhtd6g9^SdFO(Eo@D((s=W?v7^EW(UQOahgTcyTv|d{Yv}=eAt=`6 z8;6_@jf-dQC@y^d={indX^ylF*u}MeP3H4B5WG%9EV8qUKaB#|bWwBH+(TB&i2LeGI=UF5SfLT+9tBL$NLs@+wDZnH%49c?yr?kUV9 zcRXKrJZ=5ANslVB8iCRV4mJlUy;vb`YBS88v%grm>FG9pHrS%ykrL*xAq6y%9GEb9 zzN|~4*d#-MKP#cnQ+*%rMGx;XxIENfOqm8xby&?)_PiyPf=KfBUioVfJg^lQddj4C zsABSBL&VTNP^nEbIdWB8x)_^2dddu_jLo0e?!FM+vJv_qKryC;C}qxJQ%qtRQN>d} zMAcB_hGANsM}P0qFlX6$VnH9ANGnpfuLWyT$Z}xtlkxQE1xB(S+%!pf5jDTM(^dAX z^XCeD^YIwM;%PfYry__L!Gfc<73#07*q>(mC;h!`rQfrT6Zz%Cd+f1MIhE>Dl6Lm%;f`4c|8VdQ_X!PzYwDlD`t$oC;-K$liRC- zs;4zMqg?#MH*~b2j0eUzUW`Md!>0P|w|`D!u7k7Hq?MQ5iG_$UpE(=I11*{D~$jZMx3-jJDT!;~KpW7rMH0{8_}0=oq80DN(#!gXpmk z*h4sJ3~j|g^i5qL#bm>$U=C8q6e8K<_vXb@sw9%Rw!91blvp?p5k3f25}1Tew`9C# zg}}vvI*g$g!V1&C4`wiF9Bu?nV{D}i4RbN~so}tJsLhz+#%kR`H@Nd>v^? zZ-zmiZ|u)e^c~vxELRDFYf<|o)_iLYd!+gbF5hK)*B@5uT08peb@9t3uV?I_*znN8 z4QbPdo3^P2APO~jjtDa+f9{Sd`2)h&=J);Kd+z1Y@DHO+@e8`k$((ENi|6MW^acl2 z-dnmldNi<>1I!G;PV}(s=SI%8eB6D)!lZNxF%zbm4dKEvRZ=r=ABci0db1qRH_Q$2 z5Ee1#(BCJA7uqK_>g)6|{B9F_eLzTkAVSxukW?=lE5u}JzHhi|KTq8L9${0T1S$-V z>H{u^xq@hEgTk%W6{soNd5Ol9@2pzyT&><0w^1XGWd>SH;E2=AP)5OP6PNW+Y)-22 zCB8kzO^}sji6#=j#7K5sxLu`hnW#qrgC11V$o^`c(h^64k08w;lmy(ohABruQA@p| zT$!UXIS%>QQC^r;+`8T<7WT9=kAhP%{4D@E$k=k*8KC1X0i6FCis#~wgtFD2?nEbL zToxGpMNWiB9kH>zOw2Fy^j$256y?LiN_G7_k>xmA;b#q`ib`sW7nNh(6Q4Z2 zX7a0{qDiu7LfQyP1UqS(+f%^u&29Cnxl6YqpsEF=VtmV`j^6%DdxiP0mp}bw(K<@7 zPb2*$3dqYHk+_0)@u~&i`lM`*ygcOM(Ts4cWFqV*vc?k z3yJ^BOj{SCEvux+T#KOeJfIfxy_s@OFv!}P*ZAS$Mh!2FH{*lE$_N*aNwy%r4Rd8iul#sM-@3+l%7vC{Q3((XsfX*W0s| zN}`9ta#WouM}IR@iX-8KZJYIxx1#%YaSauYN6}H;?PnH5yVnOChIa}wCO#cSoo1X> zd}JrTr3kFc-Y@D;CTJd&HnV=d6?SW8ud-?i_a~WLt^^V-Lo3V!)Uo99utg9(!ZBbZ zXL3aE)!o!PSyQPvCG^ZjqGsOf7%&|+)Oo>dV|3kth-8eSe{Jq({Z2_bB3V0z8M_*v zeQR>>t{?zm*ZI?+;{3Z?EeXp-)lg-5lSxh9ix=rSd=&4Uss914ApuqxN|?@CaohK* zbg#2@dAJ5+%*tSVei3!gUt5|&KRn-bLOYQ~253&Fm^ zFst{uM2R~2{3^SGfu`g?j7)Wz#On?0^~R!>hdAQGn`vhI@DDrqLQ{tC9F@*Vy50wo zmu0d~MbffPhKYVlByn{;dC5Y}G$#9(nMfN>z)NSf9QqT}Et0*E-eS{>^opoZE{^-EAyxY-bEI32l+ z7JSfBQxj{NeH`PuaXUo8A>Hdo&^X;W7}X)7vD_G8GA83zMa!2bV^-%3rd*C^)<{*^}Mc?cs#)=vMHFqYx~~c5M4#lX0QP|GW%FJQ4hQwQwov0 z*)uLx28)j1l_w~hdr&9xf3?ZiH@0U@KQ|?vgZY2Bc2!T5T%R5ESD&13(}ZkpAM`Ey zaLr^#ba0h(uwzBKE%-}kmE^lFXKpYjBHD^DRfvf%7Ln!$K~5U}niAW-IjFg0Xs8Ch z{kF@7eWnjmqh@6-5F)c=o+pF$$QDme$i+bB9-14w5fK8+r%VLSz$OfJ7wthh^t?gy znszn$HuJbA_(GWv?T6jSk{byy=uh?ibC>Mz(Q|p_pfT2L;%>C#hD9_XSom1+8UM zIG|-ARM412-D#gxk?;$v%Z$6Cd&l>E&T$c$Ud%5Z^~e=3W(R=PmOd^GT7mkukGTbm z_}V{ZvLAo)P1XmvraDr5OE3d#eR%bzty^4Bgb!%xI6D~-2NMQV)f!VfGbo^%B9Mqv zuU#<_@wsa~c*g6&Gp+&(q$^$*rjnH>NEHXXNT#5k=&h|1_Z7Q}_wRg>a3O^7F}r&~ zdR;zjBo+iqvDdm0!Pw4qtV9v6?IYO8t-9|-@n6UBL%KrLQfJ1ZBNM5pw!-kHd3t?V zP*ZUOG*E1;?gitoy%7@SE0W`wRA;nzjTTPA12%l#wmJYC7oMrn9pwfJ66&ST*M5%gfVf6zt#X>_K!K1;$Z|fba_6lrY7xv<%h`rk zRZuP*qO`8RTRP=;*^Bn^aRZ5SD$r>y0-i~o`6)8;;(4?0i&jTw@TI(ofqaq`)GPrJ zca``)_RmFD%SzuIZIeVcuU6$pW-3(SkI>7AY-WTk*n}bV5V66i6&YC9XCzWTT}fJ3 z6*gSM9HvlY36sPK$q;cBl`8Lr34>!YiD9ITfDGV>m^~cNy*r0!qiWSrcZ!-(HjhI57} z6f)DK5fFAhL}J?nyD%Tue*NT2CxEWZFvqe*fc&_gk(5M>yo+qOAnWmSw%?Q2Hb(+_ zSU@({M4v+|1v4?!FVl}! z;0hW@3YRfGZHmF6LfQCR<~S)blD^#Xq&^Mt`Bw&|{&cBIu6q)hng|D#((GBrwrt)W zGD~0Nj?kad?rltLWvHahn^N|f^dh;%)P1l6}EMEC^3p$#c7HJqJ zLJ@hF`TjWG7V=ePi$XJfmv|AEhgmP9B*)5G7&9TfjbI`pBdbT#{O+N~F)CU6+P0E+ZB4}j`v{cs(v>JFgERWQ--p$@ zgO*S@Q3g6!APWruNopc>prlobT@y}96f3i`k!maUBx9p88@mQ?k2vRx3U(&UU)Xb> zdEsVia93jJ_Xt%Z^oCWpJ+noKAbhzQjKO)~$o~aCK*7ID1MLmq!+_7b2ia>Mm=L_e zTcLPar6*9NUM3LnIH+zSc|wo=!7`Nu7m?)`ddmXcwxefgMTdMAMb6+5b%8ohPVfXX zD(3)F{nDKcL#*NZTWOy1E@w>&@-i=>={p%6br4Hk=@x@^Ohckz`doeOR>ENH;{J2 z)iY9=O!bZjq?d-JB2K#w>lkT@gChU+Ic-UC(X=hpncIDPf#^U|4fe3;KJdN0_$i$s zIyu%F^vqti{B}$Ph?Yvu}c$(P19B$HJ1fqMKji zCxxS}?2g@u876>?VG1^4-t~kdLfPOzc7+4ET$I?TiJG?>rdPkm;Be8wgK)7KhNI~(a6OAw35KQOOx!=k?@ncPZ|%2vE7BGatORX3Zd z{$3XnqW_}aQiSRfIdL6B1z!-^5q zxL-^dL`|w4I#$768UrPB-pS^0anUz1_)|l#!#cAV*}&$OK_ohzYIi;(J=4Gcmt@B# znoYhyc>FJOu{D*0@F-(6z7*9@*{WY-gU zX77V%hoqTm6H1z4jscarZMMQ}eaAW9*sz4mPqi4Lfa)ec%gng30O8VLhbOiOVE%IW4X^;A2i z0-6O7HDS=Z1N#lxs^bk^o~qtj4hV`+J%vBXgw)a2`@IxM_%UC59rm_5{2Ge91?lEh zC+QWy?u09Ti=D?zk@2(0-a5Qb{FPfdPN-vkRu_C-$Wlo$N@gD8!1asz@Rd`t5CP14!qsLDw~ zKoUa1B*)S%16xWrqPPKjC15s-OURu1zzS3Wz`V;zGSMI9?UGl_G8pe=F8reDeN2|- za;u3Tc0SS^ z=WYG^X1gWSi1+l3K=C8D$`3`)^I=vr%@8iRG7|7z)6u(Pw=REBl!aDkg1yfZb|&x(F@NmOefZbWH}vl*;HuXHHIOz$@$+Kl^@~XjERan8O2{J^C+8isLwqYCG04tSqkUY9<}Kdh;S4Plq={L)Q|!`xsms)X$l%w1BTxh z-OkTUnP8dh)YjwnJ{AZ4zMY=i48cp*{3qnCVn1KY5MU2?m;q_J(cHAIMX&r^N|xDg zkmxyZ-S99UILV9aGdo#+&4&%X90nyA5aF)WHIqLVUJeSDxBJqdJ4i$g>P55kvcRBt zzgK415j5c09C=I>>7&!V0_<6@xSyf*^3hG1U)Kq2UjrW0>=r0yD{!Dlh92}5WGdx0 zNa>*H*BU6gR9+!L?Sk;CWNvx}Pur^n0E!@bt6N0=WynC{8JjTt1!`$Pu@3qpf*FjY z1W>CP+2^!1!pOWihSc3e(6captQ#hQ9dZyS;y@XHm+Co4ls+OAQhv#JXp<#zg|7G2n`3 zc?b6~*APLeo!XG~HDy13eW$ z$IN+aj~jT3QL;mnvgYPGIT*3a}^3V2XL5W@y$KgTZNpQWx~%z>72e_Y6% zhc}T}m(cXO6_?4Td04Qt`R(h_D4v8XzOqvfpissybl$*d&YhtEGmO z4^iX~ODB7SjbwF3{fFLD>2QqiEGmp4EB-9JbQ=$k%flAfVW_XiYkTL~R%-ez*E6ZV z62cGGSBoNuOk_M6@gOj~6W1eUQ}xOuTDY9UaQIl;9;0$kXL5NVth1o1ZKf5wv)mOY zRL;(^hI95GatQT~-waX}8yd3OD-#npfyt1kO3xS?>ePMGQ93?>SV5bl6L{Fh#N(>S zXp*p-z*$BIR|wp$gXu2*!==Y4^5V@X<%2kFs12Vss*XQHC5#$(Yxxn101`seC>XDN zJp-p_@Gr4!ZqA$#?E@ZiruP#pMI)?rE;(NLq)%c979qRF;?HS0d|^yfRHgJV^0HR> z$^cEa_%cU78|dno>hw&AnqacOmvg>hB2bsR$V{r=khCc9Jki`dw$}R_@t|JZlAEWx zV0IQGWz6s&{k>$^{IX*D*tT?dIa(NzBHIg65k}x?qFIQU%5bd;Pu^$W2+B_e$i)&l z$gq$J5q;jsakQPjtyp`6Ld^k%Y$v>T12e(M6iwJJ=y;7wjYjb{)Fk_xQg5K~Bo+Ej zIq2mn$jF<|MCf-Jn%sM`l*CkJXKp}^ZhbO3nNkkEz626-k|*)-$iTLu@iN21e*%oX zLuqJ|Fi)vMa{e;dH!dcY8vrXQipW1^m>fbZduTDI%FFl>3*IkPyB)Ak6bG9(Fbl*& zgrSozm`_}nrS(}6HC{Ek)I@ z+o^ev02!B3sTL=dR2>_MPYi&aP|e--km@n>v$R_D76%iRO%o@5c1NW> z{qHMdGD~EDk7=C%nq=n!7hOwSBf8prctkJwNFGQ89kOwG3;jM>u!Wz8)QDkpD`fJi zitXlR^lCiz#n0aeqRRb@iAVN(GDWP{iR3-QQJD2j9g8i|N$X6(uJ`0NI+%V2BaJtg z?6#5^j4&#PGPAfnQ>2k2+kr6_lma|Bkta%Pd<%bDWM#2H$E(AN+Stpf=5<6e;u?`S zb{3r`BOsYUHokiH;oGfXd>zw&D-M*6E+1XGz$q#0L4LEJ4tdUloaH?T z28$O3>Y#{&Hz^&_AHJH-uHC3mX5zFT&xnUF}pn z<5Y^JN+GqTuO(Bfaq-bl5yM{iLH#+htC=V_e2sG{c9DsnSO;Cb?@jKWBN|gtGVsNB zittzwgce6$&tJR_RxhkYzSZ?Ax6gt*y)Ukd#A6q@lG&VfGHEFU+srG5GzkoO6F1tv zs_b=IOu6)Vk}j#J>f-Y$Zn$~&H{QOkyc_&k@s(E=*W1OIy@CxTDG<}RHVcs{r6&Fb z$XBIkR1<|U9pfF9Vuq=d17>;`?&0Ht4oAL*-}}u>glZhT@^Q>J<>iT*(!4J6V1xrN zT~xBi9Q^u(OWRojTGsS@n5*>geh$Sz?>w!*oF*nmpPVi20-?pmfMBdEG?>aP6K1*C zfK501x^~!7)Tx`jcxU_BaD7Uu#FsXxvm-M`E_vtxu7*4Y5 z4k-Os){`fdE)F8JCk&6iPdW7MtK_6Ja7UqCZ33Q#ar#k?vWQ&RE_o&%=k(lGbmdH$ zlFg7!NeoiQ-<*~XsM02Qr^9)(rGYpvPe=4FL0=(~J6Yb^!IPlL-l_tX&NzWu${{3q z%7;$wX2Kb*&C3Jf9y<9Xn!x7t<6EErmr?`oEXx0^N1(3``UgNB`?I%!@n}Rp$)dJ; z6}8?M;PCvU$16Ay1CR>phK82uaju%tTOtaZll@d&-}`K{%DW)uzuA9r9RzKeM=v_f z2A!>cl--L?_v;bTT7>fVU(kuB$mt)yHlI>`s6G(He`%9-N}ZxPmS$8>R!^CTZ9b?tH|9h>rr3z6aqq}{4H{SVN0TZJ4J()g$Ujq)MhkHCRb2wCw>BP1 zXIY)(geQ&0&{pR|$QAnRY4c@tG#a!zbAJ~veUTFmyX%)a`l9>NN2F8?4g~&K=w;a!{N#dUYcQx$ z-^Z2Dy&RG>ksz?KxURe>U6VF;7qqu?<+x^>(s#Yd)4LlP)c)ci&eZW_ORRWl@q#qg z?9C+Q3g*9{!~hlgbL0s*n(yRegD*$?;P@jOofop%2SyyDXmRfcu-1FWPtO1IN|q5Ct;Y;j!3Nj;FO^ zM+utCFc=joSFA`J^Tlvc-1be$OV4M3FJeL37g?tnh27L9dGpoC`EMT3H-tw%&r)Hq zxogLFV`238aDp=@@;O(qSd%+`9DzZEjN1KRmXw`akRBjyV zYol=R1nkyP)a*jKTj|7GOuyap?l*v%sfcZE#7qI65K|$muwpVNyqnit>32N;e;3rP&0;? z*nKO>6YbtdB9Gdv;wx}z`ynHxT=EY-*)VJOjIyEs_YeIb>dyfkV>YR(p5M#q>Z=)x zElTa{0V+25)?a>mQD^(025*ckcfY^O&7J0@hkJu%O>H=l;&|rhcB5Paxh5Mn`}p&b z*j)QMym_O85HmDE)Q9w#$aT{0eMN2oiArdMaMftNPk&2NSl%k zsN0ft-_1|LOk4LS0V~N)z*82zVf_SF-;=8`x^_$Xg?Ob`?2y+;mB%g!?BDH(Zq$^-pC$Pdz}ROIbPKkLNtiUfezVlEOUUq??D5f7Ky z(Q)eUZqzAY%-iUr=*#)`&3!r|%wO3lW{}LsI@aoVGc zp5P?~W*@y0bpL%^evR4NeSsLYD%QpsS6Dv!57kx+ev9P`RhXKVRaS8>G#^%fdd#d$ z?$skuLjc7XT!h>%Z$%k@m0zS8n!8*!yq10Wj2joaHGrVyCpw$?IOeMsvZX5B_TxA* zd(=g4{dB0RN>J3fdu;it(mluVMyV~eJAD~%ax#ulgNR1Uc`nEctwgk7w$%E5HywYp zQZJjPsbp4CwAsB-T0!&>#1j{r_qZp=vH02i#96l?BMwJz9ZMnc12DW*zff3u2)Xt| z^XHRB5XJgTWp3(GL~O4wBmVep64w zP*bND`g9TVakGh2-;jgw@AcLOo7p*;7ThsG6pWeZ6+4=ma%(&te~ozbT(m8+M#Kz<6$SVoX&)vhl% zPkHai5w`v_fX3~DJqM+!mz7in-yGsUcbJe@j2*Zps%Q?}Z#iD$rhG#i;Q>LfBKwLM zrE#oW~M#+BXRo&uOP zBDkx7kb+vF2q5D-N{_BZHff58Pjv>KpjWy)c9AmfmG5ebkLOuz{zlOi_vDb}5YL7V zMaIKDk&P{_rZ(a|uk z1nAo=7oBemZ}VPGC{o3bzrIC`lzF-oVl5Y!sF|=ly}@RHLnmRM67F|yU)T`t${dvd6mIfr)U5@NPBc0L0wjTqWvY3f`~ z?8SkB;A;KhG_fKr0~tKdP~qd*iESj@M+)E!4b`v2U5EusfLdH&%?mT)mvoh-WQ>qx zIiwV2W2SLO0V9zDa=a1~unKxUtkJgb$ug*RxhGeN zvr|zv@}f$(Kit&ai|-LOueH_^PQ(3-r9xh$3h5Al%BQJ_FToiMw#~rz3%->eN+2q| zN~uGai^u){K>_Hh$CDE&`%&c2LWJSnJ(amC^0FDG&NATa5=3W|3WK$4{O`FNc&Css zd=3VGpW0a{KuOk{n!FLm7)$!J`>|32fLW-Jm_@>%uv9d@oAw4D06wSyl{?tI!8YEa zPa72J@1?V?3!(9JIMREz@D>&r;;OY8mI;84SId@R2rQI)Tuo~|6K~DJuI?ET<}3ff zQ3L8nS%1tbdeYbVVJCmu-(0M+U|as$<&|@b2rE6P1`0#NGJXv6sKf3{M~K&I0u7b` z(|CB@>zL!YGy9VRre%v|+@=-#RofE5<3*lH!hN1gBe>gdPcsORpx|Qx;GP*dIJi?` zK+efzxrgg0qiM8h&)Dw>QhFa0m;!D_ulr#mggO?P`t*dsHY0NaAC_wN>ZuUze51vyfqk=50822+F6ibA^wm-FLm;1wy~H8Jff}*W1%79t+Tt?{ow;dWeictH-C(S z_YjbsMMcO&i&aQmg$c1m&KY#A5ODa4ysY{f6k`Mu6g?x=Ak7M2=T2mwUVj=~Yvq`v zNXaKIpe|EIMz!cJ3@nFv&WuJf5hM#0&9X z@klqGj|b7elb%X$EJ$ypsK_3Z`cF)2j}BmJVgoV7w(cvRo`V zELcNMGElfqHjgAL$yu=!Bf;Iw36v8D7Otrvy~kM}H*ox)K{8ga4JZa+LW z{n_(PUIJEcD4hzxM5}JOGHqYQjIGFx3)9X;(ChhX({r^((Q7GV1^0Eh;et(jTmWJ$ zHqt=F)4!Mz%ee{F1N&-Jg`7zQDgmzcD+s@I&#)mQ$ueb^;xx}LR64>9U{@QD<6(h(#f%t7oM{ih@t zkZll#3A4_>nvU!UvTzirPgyKMmPw@{Bjz(IW1Wj3El9L6oMSfwJN3Pt4OSxFdL7t) zE!#B%cXJ3RedJp3`kxoml}@l*X5`g>Q)S z*SEPyq0-Gb9UEzPczik(&Ms7{=$|3_6hn({PZT99Gduays7i`Ur9$Um@X*ZsCI2BR zGQ*S+9&&kN^hW8$gWhGw7lfEEq&rOu|F<-sVC>98mcigj02z7W8YKhEB`C!AUKij3 z1-b2o(U*v@7}KNmK2i7U4kc0b{TTuT5~&}sHHtTLD)<0d*wumKpt;AgSpd5hkzh~l zdrTd!#Gz$REs3B9B|p;ObL6E|Eci7= zzF@=9l};|+awVMb_hMLMpHH**LNCc5De^fGVmqJ&bMQ#+{j>c{LBtNkK8n101ZIan zAAgUNyE}lu7e*+J#)1MTO-=fQFyuO^#7v`Gj^JkLvx>1i9vT*1yAz^+<>S-60@kG& zuML)357az+Rxe#P=lOIz9-}0FLx!(YX_*TF^NdG50j?e@YYJsX@5@W0TXHWh=`j=g z)8oK8u2z(aOv?9^`64zRr}vJtDe*`6!0FAwLdVv_7tHINxTug0bh9w}G%ph#HIJ09OlME;msv z|FG5DGF)pYj^B-1=gCcDj z(mZo6p#PpXVw4>YMNw_psIXJ3;qzUGVS&wc-2T?XWn4ym^w6|qjaw~c@oIWs{B~Cy z!pv042&WdXS~Ou6^WH)f9G08N;_c{x1CaPHA<&G)g)6O%eP?Onl~uyGLN-|zU$6m&(S_EW`VpmJ?kt5R4P)pM zC7L&aqCUR9R>*g^m`{uCyrLiu>7nW_w#3Sl`xX^GJEjRpnoR;MtPi^3xW zAYf@hm8;cS@^2|tl1GgIR|jT!L*oAZoGI%EeFQEkhnTjWye=mDenA*}Zh_hb1ike#o;4VfCXTl3A+5vvdcMyPYP|gz zeKlKKTBg9}Yp49*i!W@2sb96R8PED9A=pF`Xx z1tHZ?qxVLrk3(U+H~NlbBIw0RFaMjVS7~q})*rEf8h&dXj2E z_7U<F|u0iMXWdM+7=^AQ_%s8;%+F3dDfO1mG+zgDmxeELF@f)RHd zyfHOR!~0>Fz^q38fQ=|7p6-jDGU^mk>9o~Uz?ANyhjac`$Z5jA`Dg?l1DGJ8K{HiH z!ktSf!xP!9smoGQCFr(^GXm$v-)!%cPVs=cryLM_6QTf;WCp&bYtQbe zo{q9#^p@XUr1s}!g)nra08CHMFCX(O1F@~=yh2hMt@2o>a19kL~ceq>r z6k?sliuiX|nfNC;*FkKfe`M+V?uwRRTuTW%Tm2!(u1cOULC$=>4I2+iU!-A@9Yu}7fM0Ugu68eCm=p)9Ve9uP07u}Lm~jcKEG}3!xphpOZHnLC z{DYu`E*b0i%|qZk zLMz2@cNBc=MFF5AK)Y3Bb}6-j+ZGYU*$AX>1Wo)MfC1x)^t0{sm_{|T^S}Ss3=!pTe%#+uUR7L~ zn5IqyJ$6$(Y1vdWtIOTk%SxDBg(Hg5b3=MDHxS>rN`lBgqS$>kohPGhvA@}x5rF4j zTk1^Zcg4%L?9vp>w|6H#SJRV{S)4kQ^$8!bFRCc3QZil4l-W|h2TsQ%gPWa4lp-kn znGpNE@t=zi=8vhu#s7AO>*|mZ*@We;2RFBx6#c`U%tU z9EiRub*E=H>A@L|rasywah2q777^kkJ|u5qn4?1A%U1c0M@fvFgf+k$+W1Xa`GZ!m z2P%z1Nz@Y|%^x1-w23q;wog<1B_&OhKC3Rooll3-84!M>1ptd1Ji~eI>R?3HDPYXo zQAZnsG(eA9WV@A^z^;lLCP*kP`vYRhed9r`W7MPDQ_QB<*U0ikD%ShuwkICMa}XV! zuX$PEoLSpBQXd-fwLXqs7Uh>Q+s-W>LJ`pbQar}#4H+&W&A(t@e(?Z$uey@n1t4uD z<21~Y`?#}n&}x^Hy<8@uMl_5Lg&$|{ z-MYe+#a{~i{1z-WO085a1Zzv=+n#$U&=>epKm{qBi-hopHlS`Tg0Q2dmhrE=zEqC& z9a82XF#uUZjvmci=d&BL*lV7myZpcrh-iJ z1NjD(Xm9dA!Mm3$Rau$#&_mUs>X{n#->76?syqh^#e}ok09DB7iW#2RAiFJIO&6Vv z@4s!MOmO1oJ@g-WUoSjk4FHMZCg9+Mk zehx^Y;K0?&R0Ev{RZze7aN;_v>7!M-C>Xn2no+xi)(xPY{yt7M3RUOIa}aKOD(e9# zIb=Dg{jPG+L)`{t;mbM=B*St3i?#57g4p0!QPrVf1dulj>LO=p6S_Z?Ek_gqSl0sb z@XRcZYm#!s2?eSw{Ke+9sI74KiS#60)Sko4c`wtxTvRF7QG-pmg8I`HABN%+b+5oa zIF!;_J6*E_1Nd?_ov+hj06&w;gHNBDg!6+wkHWL-WeI{Z!FFe;VH4_Gp9Kc*@zb0kQ&-mF7a`89Y^RNTmoH0>L33WKVG69G-7(yZP zulNe2aCa1pWbNJ3R^vuj8+1olT$HJNW!z?kFzb64{F%dCU48b}$+?@j5Ie*P18Vx- z2LpVzh=Vrr;Zj9q2;X!B;Kk>g`AXUD z18~&y@C4Fn06OnFeLw-2E@HAo9 zo*zLl>_|8yAG+G*QNzG54Di2yogUGyXk0bx!_*Bsa%OmGoIo_0-~WeM6HKhK$P>SY?kBtph#v&cKqL-X;uSBt654a0LsrX+i;$$`4vnndQyx_B#F_D4T_qK{I;Z!NmSk~zFI_~WsM}MOrtO8N zoH!iIC9m58ZN|^seiUrH5SsZ1V`Mtr)@r-L8(;Ml6uc>@{)@Gjp!%uKr(*H6Tj+&` zL1znaQ>&HW(>8%o+60H!O@{Rim>a!ccxyhb;1n;&TQffs1ogjlV&_JBD~GieF3oh; z4AhB7qW>1v;Mnh;3@}p;9Xv(~V{NTF=AsU*{&@G#2qn}?B+-Vm;~$ln0zlgK5wdPY z0marU-c}L#;;fccDD4klSyuzFl3zQDh)8Owhq~CTyvy^wh zax|12JLmPr9!BG!qh=GvAV6C0c3QEMpr>Yyt3ITO>Lg*}OGrO3N8AhePB_ znv=(?TGIfPUlQPfAIp-yn8s?Xe)Oz?+#SUW8;|vl2wS6)&X1t2+$&uHl;SQo1jn@! z>WCDc1+MjH*ML3)HL69sTod~p%M$$#0jQdI-V5Vt1-T3PM4A<8#R_;5F`y!wj09Q{ zE|nP;4ks}z!A^tx$nH0N5iEwR9^ohH+^PKIP*%yBk99c5%(9TFNd^oHwk_}0NMuP} zKMOB(NncKvda5G@iwaw@P_nkJK^QMoJ5H<{Lv&Ev_C6F0tfRPgpEBp$T&H=np7cUi zC4%_(J1C3IaUaO!tg-{)3M;N}7Z}5|KTQB5U15R^qR={&0VN(oq+wS&RNiw~l|S|C zUyUQiKIa-p{?7P4M@;hBx+40mPrdq?Ns8Tx^EjSJcdr(p%?;iotE$ZUGi~>NgW0Ju zU0ct4r&g$}0WC2ed*cjlNB4=g^({_CHouum;C{KF^#(}*Ry}%|sy9xEVKXimg&_cu zT9D<1_6DpWV6F~8K1O>iQOr_1BRq(xi5{W8J9@Hn3Ovej%b#G zckP?tTI+}W4NDpX;9LaA4=MTy|B64F@fO`QfZ9zpqw1dhic_5mBF`|STql3P!)gX2 zWV`UeFKzr=GL7~FJg4jN*+gcteaSvu{h`MLWt`WB2oeUDR~{HEamz*J5FWq;6cGGAcE6=vZRCdOd9Ewupfburn`pd&5?)9pA8w_Dx>_y_u0 z6uALC)m3oNqU$Ml_;7SGF8-N4_Dd0HWCQ%jyeSCK3SwEwIJymHp%Pf4SJ5?8r`N_XDA7kBhP`kvSvPCn zhWFuWRaD;^kNXbC4|rInkjUb3cSrRuFbnY?KbI62`R#eE$OjY=DG4y<&qLqH7+lpn zxiTgIBc#+FQdl@s$y^1%e5&u4!tJ0C>9FnnRHWb5q-(pOSNW8_ZpfS7y=ZOZi?Cr^ zURP5{pt(#XN0w7ElQ<8Z*Ez_~FA}j}MI8*2#1{ zyj3mitq~+ErKjd%=@)wnuRmC#aRZR$r#B$r22Lv?urgTYgorad3eXwYJo4tSPPZj1 z!==LH^#Og!?Ab@3_z&V=|71(hGKN>BZX@1yHkyGc#-bsYtQvaQ;4JQ(y<{fvtHU+4 zN)~$|6~cO}Zmm>s1f-RaJOSggqt0jlEf6&^2Tx$KY-}z9Tb{Vl!n5#n=fH3>2BpFaxn#7< zdZh#;Z!Z|l*k0Mj?hD;Nu7UrDt04~QGwtK9TMNI0FFGyl77VGFeDrG~hSjXW>tDz^ zV$qZQoWT@B%9;7UKf66ylP>nWIyA(SC!x{*YBUqq0mp|pn?MG7CHk#P=VHM`hD3Tr za@D6iqG)m?D#tFS2!T$CdIX>ro(bRE*+T5QV6CF)kdM4qnc@ch&UUD^xvKhaT!i1L zJvy*-Kc(fS%{R@qd>80__PyDm+TFX)9Z763a&J4)8Cs@usl3|r$?tl7uNQ+T3h(X& z#g)wpw2{nWY*97Wu5N<0l3DBQ9|%D*UzukVu{wXWWc$Yb0B3gUr#(p#?FGC_h)=k+ zzVt$X!fQLTXZG|2MqrLZZ%Ax?T5U8DZ|5N*bcTZ^Lk8uOKQc@aEKCfnnWwjuW*fV~ zjBJ1-(?+)yExp=Y@v_GJ_Cw;=H!AM*QC5Af8tjJ$TdqdO&~0CWho}B9>#P*D63eq- zK$wpIu8WyUzBu2#nFNnD$aio#J5#yrU#~d2sN{t(Cc8rmB`(J6X zILO0==3&C)?4OkFEQl}I<+yydu@*04=SCUYI}ZZORbpeg0m}8GlaYb(8vQtQ?uh#* zHbs`z{9iXz#}LScA6Y+RUEZOfew%U;<;?BZFod^hr-2Ecb9{LOFAj34Xl{qS^!p zH-*DDY%;h4=q7I3a&Ds)X$<`fsBwfxfbcr!4k^x>Zo=l)Z2 zuH=>^^FaPvlkIFK9R{w*r`i$x+z<2>+yS`sC{Bh$poQpApMMjn=;dkPV*ZQ2nKQg; zQ1vOE-$WbR69n1LlVz)d53vFkUUwdd92rEscj@u=dQ~tsbZS6;BXJ({(rGiM|2qxt zg)yIPSoSQZCV`rXnTbNS2No-L|CHq9osW4~WX9F#LxV=H^Dl5VOPI%gh%Kq0xXdDYT_M3eX=Z3{622M`Bbj^Y_xOpH} zLBZKH3>Nqeo%Rh~PWwFW-AfahGfTl*db>>OhX%!}Qw|k!kNR`z+J87%B~$D>xCYm- zdjVa)GUFC2G59dQf zks}mMbWMVp;L579U2-R(RZt+r8UaZ(CFN1Jf5gl9D0Rw2H#*w;!3}JJVjIDKhvmQ0 zev(m41&A#t^vXd#$@)u24!*ROijhRF(uMv#!(wL2nRl6It;Ruwog0cX z_t09?x$^msY41?M+Dk%hZ<t+2EtuOw^8oY8rtJUq#u zBh@!d&y*sx^4OC)q)|ogaumRP(1k_8*eOs$yEzQoO~s%bZF1)L%BAWdO<3mMJe`U! z6sqzI1v(i7y!~r8;RaR&Onfqqw$8NK2Jf4pEhVOw9n9c*cg++DrKX}U)HuT>y+0-< z%oi(^$Aeh@;W0MXuxjEnjC0(CB---2^Blfgflfi6u}yje7#%KyIqZ-?#?T~HDRu_F=vNe)YVo5| z-_`>>4`8C1HHE-)CDr0;L^X(&57<`zcM(o`4DRK6*Z8JWr^}r8Y`fZE_Kmg)Te( zv*<$k)Mhg%!fF+3_^|-P2sbBGLK1^;ya{Qvl3QD zCrt7N?l>Cr*&o19oORY@#1NCE*Mv|51+t}4yM@DM?XA^0MPXEZe>wJOnAPT-0}6Tl z<`8twiuv5|Qj;`B0#spgxC;NU8q`*D;i&-RZ>7LmZ!2wd-MVT2*wrosod3)eKPU43 z4QlE3*zw7z)?-YbXwBlzpkP}`Z4v?Ja*UlMPX8DW1wdPl7JCcC;dADzU^ZVuSd*qc zk{s_kUX?-i`4wr+W7q<{JbXS?N2Qh9aXrvZjc$%9)q|sW>auB2hfVw3L>y@nC=|p1 zCfJ=cF8CTn_glV(5MOQGHkxzOb6PnD~hcZ@h2N4nd-bG&daCFAAR?0eIv zS#KLY-_2c)tQYCZFLHm-6kJ1ra{)$kSfl27My1pj2gfa36iVEMmlnhCx!SEdho(W% zx0^R0^1q@|nx03~!4XgdxfvPI+qSbyB-gK%mIF!v>w=V_*gJKmsxZ9vc1;5)`Q913 z>>q4nh&_6VW7|hmXupV&XMzua$Gk zd8t2KAG{BwQS#1Cw&Q#4Riek{T3 zQZZ0Zaa-IBi!K>Rd72QGgWT|o>WSldS>tR*ehU5s8um0`kXCF*K;HSVVcH1E{>5(F z3>J7^59Mt`i@*yq9B-kxj`!;R$X*Tsy@195M}q#6n{bhtyy_=nt~$bX;JHvDY^5lP zSxNn<$Ht#8wYimcn-QPrgvD#6n_qIsL4kiZL zudg(V@H4a93aI?Zw*=A%#1fu*8wTj%<++OPqF1;YnWkK1mjL5Ta7`m*6YtU+Q;E6yHz4>FDr@(V)O*32;!d;OZ>Clro;>;nM&o+=EPg(OO< zugFjkKh(qVqRBof3Hdd0y?O~)7Zu0M;c*V^Hjaddqfy3Yyi{Mp7fam+kKTB*5!5GF zFNMd{lLGSK$;JF52Cz#DO%q8vwiYf0{z*O9nv*O|XYIa1Zdc@jE8(dy5+ADj`HbDu zvM5@VB+zZ!wr$(CZQHhO+s4_pZQHi3x>3&+(ft`~<;)x;Dq?)&Y!h&LI0w|^#m*oV zXTtAgd?O-(Ez&VD>JEFLCeY%=UGU%@9IUcj@NV?RBfQ*Pix-A|#awKn$8TyXNT1w> ze$q&|m8B}KY`%lTdi;eJ?1fb~nrIIXhk$K25O7Dbbl+O|wRjBSV@Rt+m+5UxM;KEB zyY~K0zEaE>A&N3sRtPbTTH^T%F4smq31~r=#NKbH!P1=fS@;Gm;;`>A?CD1H@rE*F zS@Zz6id-g#;Z+96CkFyjRgS=tgzY>OU!ObRmvZu<|1GcqVf?I&c%yE61YTWTe|&z) zlBwANKdmx{9X`=W$|t4G1?lIkH_xX|bKErdJp8h&~}Pcl@tlSd~0p(Jpb zIy!NHA*}YiFGYipZzn1slyrtohtgK&2qym{xnnYyxD5pU4rE%!=jwh`jfLE-Z=wVo zeuFeuxq2Syj6~(b9u5#0rr!7|4?#oJ+SpfO>_T#%WJMbIUcfIZncYhdogd^Mnez~2 z=L-{(UT`AjE;;?7LfYUbKDqDZw)4li6;FC1Pn|OoWi|amf83h3(vn-&GG1uqZr~Tx zyER2Ql;Vg2wig_9g@j;KuR&58tRLo{1OEH(w-1Hz!6Nszniijx78L(M|F8%)nM*=x z$mmvH;->@o*kzt{Mrj+xjps4m-@*hua_HJUej zVb9E13cIXn+*NXt!wEAp^XSY7ei1s=X9=-W66`h84YFlBj+jwj9wL@r{$bI=9Jwxa z#cn$|8->TElKVVn_QRvsf~X}El^E>6Lc;&cmbcX(p+)_>yx8>-4{5wI!U&!%Xd>s?Xm4wZFaL=-|<^poW%G3KTQx{xpx6tp_-6CGazX!TLV;8xTh< z|1_{k_jALm0OfL14`ctazh85tLY#pM_5h^3qgFUX1RzUP995G*2a( z+~jt%hkZ)s7edKYfU z{48B(Oc_|E+gDgu|LhH{L(~_7=hPX8bCc_^E=T$|P%udO(WY{;OkRbjT9KnBW{SZu>&aHNw#ly5w6?RjyBwUTJ-g9%tW~e14r)~;IsW9bZLY8| z8MsqlBB2AJ{}v)q7)s>C#?ZL4wEb;cVE1eBti{0>ND0o+(F?Sg3x;{b?YR-TZog`a zP=4<<+OtgtGSL~}2zAnt{(0F@=Yd7B2b#^K+$xRSraCw6{BT_o30;YGs0*;kDi+3A zu{?ZxQZ5WywE03X6IX`J8}c&xJ72(!T>V#+YA}q<{gW+yGfqP3fGZ0~HczJrUYiRC z-_-s{raYeSS-@Ng=X*^#tE>^LHfd@yL>m$#7z~zcu3rfz%SemS=fX``*|17#T(C#a zstj}pqR17zKvlGSLu%gL>u3I>hi!QfJMVAxIy$~?K2|)^1+4O;ZnXvgw79VcgF9)7 z>@V+o#J5S!4dkN_6Vv%KWELvZzt^*prIyzIe=BwtpQqq! zOBn~q8do=n)b$TaYvavm@i==D_?cG;xeu_HytajxdK4V*=UTV`ZJk3$FINhMfwM6K zM*CTTv}cMG7lz9ko&z^w-;kuywz6za9&f!{46=|Wuh7jv~sn*NNBh5i&2Zpd2AL{f6`9i#c ztyU)oR9YGw|MoThu6T+27Psxl!3!T)v1<>3B=7nVe|W3(H_>j6z}vpwe7HN6#-7im zeWZ10VYf`=gaGh-q9O?Xj+gi72ZB+wJWj*EMIs!IsFY1oPI?=Jx%6QnM78iVvlI@q zFzkmzRY*3ne*Y0i39;biyNfUH>S6M>`xWn11CZRrt_V~*@<{;@kdgzYDgmw7->hfW z!1{j6-rLaHSG}`WfY{>=(kZ#X22(w*WzRV*Meq7JGDv2*huGF2?rOQ!vox@7BljX6IUA} z5{FP2Gej5Zk%>XO6Q7HH0PU$0NDBqL%Tqa7tH5>8N({I1i+ynF8RK&ej?2XM)*9w+ zLd?`5vsjHqPz!s`(Ydq$09~NY!nSQl>?`D5_mxqDa76oXMn%B6We%11Ri3Ss4ZJ7t zD=p&{E+8K^&I~NO+v2asGr7l$23VH@Z5#>yR<|9&-4Pq&HSFvkQT1hINNeFonKm;L zVNjb>6slZh*lBcu@2V??sedv$Tj-_FvDcDNnqOwd3iXMKUp|QnL$*GOGth}te@Cbk+3Zi0xd42)#+ycW;;=hh)%uDXHt!>D2ZVV?t)u6MY41=zlmqvL_I zqf42i!)f7%d16r`WByB4|SNBn7ZYtZfYLb1i_=ESTD97yzZS zurao{F;qZgEpY!OKV1L(>gf2leh!Z1uWbM5$ojwqz7ZK*>nebyA*mmtCMIb>13_&P zfxzA<1Oe&moD(*=c%1+IXc(B?-(?)s{xkoS)SrLaz+cUuDem^K@k0On5^&5kY%T!U zXxS+nL-XIF?Dps|TK{8yLt@j8zw#e~(|0(2z&AcDC4kr{*uuZ7kq%Q0uDO ztlTIZfTcFEF*tTWKy50b!1~O@`sn!jd7p)^wKMhGpZ@osU;*s6;t#pW zErrponGK2g?Q{Gw*{9l{#LHcK)r0@*nSSVR(`UX?fdA{t`0D87&mYQnd8U7QF{+{} z;*vt~_a4SKH?57ag|&&DF_ePKr?{doc<~?TZ*JNW``;YQPx_?4Ih4Skzpte)Ik+=> z02%H~Us%Yk{+>VW&0jPULWAp5iJ5@`hywFs6TrFVdM3a%%=GRbeP`G#9NZ|I++W&H zzTWTEpLb47>`n}<;cA*KEN3u{Qnv7^ru@o!)K49y{XY7}gN3yo?xPh$X#o4;u0h>r3UXSNfcPHS|&knDRg~ zTMA$oWS_H?zCLuq7xP3)rn4Tcq5fnni4=aR`syxtTsv-C=e-57RH1BGPT>QrdAGb~ zD2Fwgj)QIgS7#d-QWknUw7YOnrL~&Et-U)&Rbty;#a^6S?*M{`B`nK-|6>)qyJ@Vd zgUW<3-^i`REnTgd(2=zWQj;NeZexv3cN%ZAnkOoyfS@9#~ry)GA=!?204RAViXx^M{mR~Ip? z0}1KVg|^V*25G@qky#8zef6Pn_cr)D6dr@8bWP$lJspml2}vjZg}-PT;MK;wmwN>E zQHwYV%~kXZpc! zkhlc1At%M#kIDgb{MY{9mVuAR7hQa=nWwtjbvJEb!|%F5hRQ3a;xlttzS&h-bId!X z>@{7^lp@%p+yai)EhVSqYpG?M% zT5dQVgBZB2j`sy1$ro-5Car=lCPuMV1^h3{M|PQjnunnCV$^Amm0ukVsaw7|?lUF0?U@P%x1umR_qNQ})&0ZF$@ zFt?SEq^*LsBkc3)Tb8g~Azz-{QR3K0-qyIF*AaYduRV=-M5VYpLyZYqd#)$#p!&kv zfPmE^aM_HvVy2ggtz&W#F}AR{q}(c1&jIlEw2?^xXFS~7MgbeQ(*X!#J!xCZ#c6zVwV->K9aL( z^VOVY-u<)98$^QqlF@Cs5^Rl;N@g(&4TZil6Yjkg1!Pqi%EqWoggQ8S`owFA`=HK9 zyvGTWeIskgdFJuFAnIoUohVYn_!us$!1(#P@)gbznor*q8wwZ38yN=u}=3%W+w z%cXg2uzgMA;)g{Snx5tR9~C(m4$A51TqPxBiD8~%vt#o>%S&aYeC&Tyv!fdyfSeem z*Bw^SX5s81-)Pa(xRhiBk)0OXrl$?F(3rd?-sOv z`sjEW0fPMzcHWuNY~&Rz>#kse^CB1()`={}x*C>v6-J<8n$8hTD(eZ|Ns5!xnnbT< zc+MVIWlH}n9y9vobMKGzM71zXg_d$VJdaqGF6tCBhGpX%X!5%lJFqD{QLbyWfeJ@S zx6Lr%zNoc?RZb9Ed>Vf}PP-{HWe;K-rbVxU!yY$KQjZeJ_gy-HnA53nvRbTFt6eE# zyHDLSNI!NK4%3!Z;GFv#aYbag!12WZHt-Vhh}*UoI64n$%?xx4ia6?X_bI4Pr7Npb zWN)r+`jR*<8Xg@$F+GW->*iAJ-~Z=`%xcck5OuGlAKp0&F_l<-705vItB!R#f%7vl0UTo^ojN?Y760b(1Hzz~hNV!k`Mw}serh7-V+6WU` z_g^#~6+LSsdT0ubfEgC6O{TMe;`TvOzeD-21EF#OrTR;=vwA7`Y&e#~Mb6L@bb(%} z9Vgs-#JKjCz=zTVI*0#v-EP-rNB%@QGP&Ie`UyQ^Q@vL>A7angHWx>_Kb4;s|(DieG5?UMzl<^xgk_f73 zHF9OucsJt#D-indJG&c8kFS&lHQpU74}-@&xSTjL-qaW5(4&NHoYE^XW;B~D>;57P5P$?j?vp!1Zkn~xSNa;S-IxgzQ+jJ?PYEc@U|k)dZXTwqr$6!`l+8bfpcqKC-( zdFG_DOU~^zT*M8`y+0}Wda$lzWJl@oucM-&`RN^1 zPH5i(2I@6EI-{E`p_;+07J+>$5N^}F9$uv7_L2~JledX}j;)4mMMs)RFl#@(T$V!* z&Y4WfXHvCyVHqL~5oeDy(d3jE4DXjB{@=~$gE$-Aah~s0MtP7zx{j-2_JWg10(Ix^ z1I(;u%JSRP^bbhJ)mP1Kk8TQpPA9cFEcdmeucRu=*z36s^h%dT=ZH=($2xL^+_5CP zzN~rr-z-w)g*kZ)zFM|{gO|=$kH9xQdmRn&tDfXF(P%WYv9^%g5t}O#nqR(=5>W$h zeEb{U+iBc>s8bVOC_xcT;$yZmk~1@1xZzC10bFgS$Od4nk_!JRPoBh zZTUNilzyEQGKHJ7wk-%E~4(>1dWI*hlB@I*O8f)4~BMn869VnYZK zT=avvHb8Mbavp6J@442F{O&T)C(R&i5}v`DL5c_aX0y`5T84cb+L7C@ztC^X=@b)R zU$&$$gSXu>)8t%eexb(9Q)PYB<){tg06Q<8ng3QT>dBdsSkB?5o`ON?Inmp8p*?;g zhL>LQ+RQS@{Q#LOvEcpkPhHRm^kn3|a1!l_IQz|r^u!{2X#YKpCM#@#4uMk$f}L`i zk32O0`vc_shNVHf_r*;0Qn$C2DyRET3e)!8m+eam0!anyv znr;p2+kWV6f6t6(dwRM|7em$&=xh&Gg}FKR6$!W~eIqv~u<%BrUeRX@bN;o){g5Ob zg#QNOO7Q_SA=_*QD7Li=gL+9fUzA{2zkHAwMaTs{`+X5C z#CqqE5GW?l&`fg9kL;^3TX(kLZVVvBLgi*%@cYG9?532yApDNz1Y1ecfH8lKMc%A2 z_t$zvYz>I80~pa5w~f(FZ;b54Q2H-FJcVsEML3Rt58Qg_zw5-Oqi>ma3N z{n4+oMm3lV#_Rp*hcNtJQyiH`*CKIr^ zo<2q^DAo%&ei2>g`b153&6p|EhR^#B8RXL4`UckGb%iQ$mu+|*ZvI8a+{NzzJ2J&Q zz1yf_rKlpaGbYvOFKz2~ilxMrX$h}79TgRUhM-5!E6z{?wqqqG&c~Q&1g98 z?6cr#5SS=MYGFJ*%G}qP_hzcNbPhf77bi1+D`_@qTa|MdcjhbJgJ^VRF)VW(<$j)3 zWu2?sW#axFv4LUb=kb=y!-(o#7Lqtfz%Gh$A!(zvaejMuwq}6F1Ec!@wVSpakV;Ox*ZaRD0BaPpl76{6t*w8P(uK zB>o8_kj~iSln|>xF|6YxG@K1b5Zd)gCG zv??DC=r>&{IIY58@naI%Th#$Y*RMzVzX0Ax(%=6KPN(?BUV9VMIn(*;jO6Z}K0lB6 z6ox-Y+Z`Z4jO|BcoO4K5<`NhZa-2QlKR-?~t*Ja>TBIn)k!*+d^B8o%?)y2)m`sk; z;T|P!ns%;vAj9_ggvf6J!n{r)a`t(%pxf{IRx5&?Q2{*O9*Y!QWmfa z^cRGr|BSC?un=YjI^GKnLc1vA-VwU^%}`d+MUIBM()m@>pdR};w6-KJ2wb_mshtFa zmB%H)74VD^irt5sYO$PF@^xCEtZ6%0zyJA`Y}vlko4zR8!>X1p{b-5SA7mX<@`sG6 zLKaWE%wr(ZR|{yfaph~;kdPm84ozB`KV+;C+^tbSAjW$H3NkTKUiG>}wQb~x^_%b&%#ps|%1}L*C+m+j((<8; zVDD9H1h?{+<00Va<$%*ev8c?RgK>GzBqk<3cXG>PDD)mmPi}4q9_vspkTT?NJJad{sjoteRJkWNW{cKwwBNS*B|fCL{7P#@`}t3_g+Xw#5F5t|@DMk*LHQ=`NjM8u zZKX#^!hJj$44rAJ^!}H5F_&)>P=)9FszlCVSL=FuLPrHdRN%gSXrA=_5TD7m>>K-_#{OjH3e0`?Se;3G@nd(%jr(4A=iEp}aqZ{3hvk#x`NrnAA#HJUa&7 zl(lsA=>J|Ch=L-+ra*y8wq*DqGM2u3OS9~RZQtx$CU#1mYW3#juf=)}9>BWvYnVTn znFl-MoMr#bCZ7dfsD`MFd2FQt`aMjlkf!YoN-LbD&CEx0$KvYUcU4*zl{JgtpgXtm z8$dgGF#2Q2AcI-T_e`^F!fo;yD)%k08;t?ItXvsi$?w0-=yPmV)bn0aIwtjW(rlf5 zMh#A@}PW86c9(+5+pbUt!(-&3VUiCy5Wm4G<64 zQ0*!N^H^91KMuh&uJvD_E1U3#r=L+eClDV64(n7s>KZ;GPQ;u;o%%MX1ZLFOAEWFw#4JIQ{nU6D{K(hRmBx7}ZS!~c$ldooh zpw`>;7Qe7nJ4}Ly61LVF`i^VtJi6z08~qm%^_1PPWZ->yp54i^9yd zg$aIirgh50*$^I=GBN3+JgUN}|4AWsqH6Omh9cX&nlZ(hA+vCMjsKvP6`tyULu)Ia zAhN2YJ}$ww2o?g?OQ2SY@cgHu^@>c~w6NRA-(%FOSeK zsERclJT&Di6-r2r(I&XeQ_v0Bg?Cu02B1Q^SbtA2ND6gK-4aV=YU2g}=W;r-1ifDf zPl0W`YTC`sW#?*t#h&_LKv|tQcF}2R-pb2;g?aT?S+Zb~l@;GudB{S%E7RKDo~Tp8 zB@|m&^ynIAD9wqOjqLJ9t5H$;O+@Nt4hw=6X7Y8&AeuEuZdR*sqOlNYd`_(k*}owy zUiI^D#=^AIhDTP3A1VS_@Y4GuN>YotbHyYP1CCY#@>c_3;IC?RF1!r?c(^Nyw;B0Q zsEt2GtQXJ}A_Kkgb$V9=A5mLmV_Um~XA7=T?Vp}1+VRZ6UR8mqLXcL`kJ!L!?PL#8 z-8NHim_$HHBsH7{=5$svoS`g&umU^CzalAJ+wG684zcTqbut{wi&aqOY}_%qd?XVi+3-#Fm@IMkx8n3}dP%e4y>W~l(Q^N<8 zb=+s$(Fehgxwak?{8*ire=f)O#1t3f#an84(p^H;gJ+Sf-=UU?;4kF+d3;oBOkYmn z5zvSzNd%8=SYgY&Q)5C0=)289u{lq!CaMO!8_UMjkLBL&D;8HxfSQLAcH2z1n7NRu zLPFNYRLUrZFCZY^&BV4?%v;LV76wER zikIw@_>&z9qQ@E{8Sx-ZgB{=nf@DJ}Hh#ho$(PyPRuQ2xNju2 zEYE>j&6rlV-6bnm<=_lU0i~+}5(hyvCQzRvEljd#_hwPt?`fpS(E(3-g-twUFlWdS za00-Z`s9<=E+M*eFZcUwJhKr-nS}b?Q~c#27a16~IiNk%0EyIJ& z>OVo!d8m9Vv*G9f!P*6ME;OZaqJ*x-D%g{5fByN=n)`?lR0319(5TjSX9d#*hGYCV z{bNeyjkUrx9^LSB${qlLNY%5@fjd-y(Qk-wTw&KMNH4`p7dh$Hp$y?2>Wj^NVk2J? zu#0ugan&&7ZWr9Ozn8FNJ&MR6{X1Tc`N4Lt#Dk*#i({%UqpVIDksFEdTe9BIU-(_OWviHV1azXCcXh( z_$cEUQVoj2RgcRE+)ZKK!<(Nwll67`ZvTEQd-?wc>2S6!F&cX=S(kg%5H{KQfSl@= zHjtoz{?#!hswfbwWO>7d>b?osOxA(619H$3PB8a0qVWTx$BVcYT0*o|<{yy7j3H+D zSeiMGd$l)gKt1V&3#VObVk%kB+tBxm{0m8P%tFBjE%p} zFv*%kBq_H;(x`qdz?n7*wrMQhs^jd%I6N^DuYn-G>A5uc zXY73|r&rJ%YxF#Iq{J|ER`eAcRZ(#M8nRpytmTAltq>=Bfi2AScfP(*)(Z(b7mhkk zyC&3t@lv2R8{r_tVW{$a_E0OIY5}rzKait6;?r7Yb%(HQ7Qiz-&KP^qEZ9Ih_)NOX zd7e?ri0=C--Ie3XZq_&dF&1Ko-8(q%Cb!Fdw2iU}-$;p$az0?=-K=E( z#{|y7Wqq-r}r77 zFe9-xVT`^;iM-(q$W7;k4GF|lsly8D#NbUJ!SWccy9KNdijP};e1O^zOU9b}^)Ob= zUEPhpx*zqPI@3Zm_xf&erXqKhl_yj^7{52nFQUtJBk%K$%`Zh z7G7D0hsP8HQd|VdknC8DI&do~o|z@)mpCWJCsG+smYU6$sQ8%l;d{R`duw>(Mw957 zCXmF~_h8T>8nqz3>B+9^1RI-^6sn%%cZWRRBa!QyUFPTJ4RcIkB_|(yB-(xkQhW^F zp1{jHKCC-$rFZ_tQ`|Q$1K54<45wM114kSI|aX?^!HP~6hqwhp2M%Z>4D`;@odCz&;~Ru+ z+vUvnTy}Y|S?x;4!_)x%eM9PD$%9zR^EYY`tOKBG>hNy}Nv(-jSj0BnE|!KK72XCM zv~Hyma?Lg)UH4<(ZPVw9q(vWiY>{yT%e{^o(Z?ari&*%YepJ_lqz&R45|txXm9A>u zrP^XfLL*JU^_#c>D0wqV9j>*RRUCEiDzrOf%B}t$pgIz{P2Tf2 zS!^*h=zP>c_xrf?tf5`s!>>yJ@0Aymt7YLzUraJ59-Hn|;7b>IJq_hcu0#OzBU3QFEmoVv5t zv#Xldc&S&;6BWkT2lV)_0{UlY&%X``7PatH(6UD2>2Y=$`cF>&aIV`q8Xk7JjCc|= z&U2W!gA2E#O$m@b>I-bKu^Je&R>YU*TAUx#j9+2*kbO}ThY#o&872;HL7+~&&F5&G zr?2DV20dM_-%+NX8*VJ`DVJiYKK6R^jecE<((t#0V^j>T^R816U_>oy4G{d@4M>te zf=q0%7?+5RwY9uQq?^O$G{kEVTWNOtRqO`{W$z85R$gM{dYzh^aaDJ%BOt?zA9s-m z3idsP;Hv!vvsY{gg6j@G5|KvT4)KnS39@xF+!PYg!7kjKClu;M>Y<72qvG%+#w9&v z+L{VtR3hj`un0>&!xNqPz|g*@%`vlB6)iPKECR0;L^_aL@U8+*|3bA|yf>{mnrQyOI22G8AW%#-4lkwxqlYvPiTe*yT_JdHF^M8iovEy*+02$acp} z@SHEaEw59K_#hS&L1Z&VU7EwWdiZf$>2U)zBF?qVz7d6c5$z~ML>{YyB7!Ix&zaYn zMBqqLvHSHWsRU^RKgIJs_(!Zdvj(M!5I|$mRbcB-NRe(Of+l6z%(8MzL!T7CK{l~t`O zdJs9t1*7VLTHjgYCq_AcqX7XL6OuUM^>H$|SI3ek^RePGbo+NH&%Gd-PtD11sy%P7O(gSB*FL#B54EAo`QUvCg6?TLv~%35 z@{T(cj?qofy+lneR9)(JRF)bH2h*052pjExqQ^GPis`#_qh!UIWP*7Z(>`E#&wz&D z(!{20aDB(iLQ3(^6{fA-R`Nk=HuP3FcKzqLV>P1FaHZEkoRK@Dt&*c|)Xf_A5-S^VJO2J{5;E;nhwwvvxVZf=egQnEdpWm#*orwmJkSpb z&hv8pG%5|zSdx6?@O;qL5VT0TVpl`xC}<~)2NrNuMhn7QOojQcmKD~wpznH=!SlHJu9P=JBi15tNvL%T&w zs;{+&wL8~#5jIrQVpsadlKRQE@&1&sqs(Ku*nw7AK{HP4LbazS8IhZb@&;|LT9cxb z6%m{y`~6b|=DxcoCp|xYF4Pny44DmO5Usl3D`-XiQY&Xgp5) z`3~v$Zn4VadXP_FgNX$3c+cyeig+rpXoniOv=oeT37WcHc@v|YHL@;nUPrLuloHp9 z)VHNS8N`FH&bR8w5cLaW(FpP&BhC<~s-22i0r=uw#i^HaMMJDd$7)1Mo0TZNJ=y4!s zr;=vqbW%SbX%(h`dpKE41#A@NgdQ=6-26>S86R7*NWe846%QQDx7j)(&AmPNAPSa! zpDPF!o=o~Tl6Ayg2-cGH7SfO3`{yzf8)?k7pll$3uf$MLAETpN!CC2A=Xcn}G-@;nW!rEIkS5Q+k=oGdXZj=hdy)0hlC4i>@fG(=;HCrk*f)AOc= zxo^PjwM6X$-V;z|s^uGRKHI#Xaz)$K(Q$V>rqJFLqC+mo1#W#(&<#Vx!Rss z@5Y)XZ4}1O&jgKxfdXOuE-g?~^gw+#_&qf*1cZJ%PaIS=JS!$n+dUkr_i+sT7{CW& zD?i4p(Ge+gqh)bXMRP*b-Eyg;wDhDi*f%qZ}4#kS>P1PYxQ*m&Ybk)kwZ9?A@WJ9IBi?)O;RBJSQg7KE3% z4rA1IlW82VCtccwy!{Gl5}HC|d3&6}v1mr#pWa`{EA$LtbY0SSUUc4WgdDHh2JAJI zezAJcpa!PKawhy~cg`-H{0>N6=s72RKf@!I-Eoz3D)T*kvZ4!ztoUA~wO6$YUrU|f z4Q6X%DQNs{ba;9VW=h)C@CHvC-5Oql``MPUY(ZYzoMDD;gy21 zU(5{jnz&wKFR#T1vkW4gf}&Ns<)MR*b0|jo3veN-#SNwnlD*5iw9_icDitA7NlkO9>~`;GUolSy~-q z`N=Ayd3sKUbMwK%5e#ya9fp=#fd{ z;gjoiVP>D1M#`XvD7;KHd)u>z^jGWM*uS!u6E+lopCZ0PAUR8$l}%Sa@yj_3VDPu` zG#@t}%VMselkT`@BG}gLLn88s%H5o9ZLlrXuC-mE+DkV>*o+$p)v3s`udtfr?6vRy z7jcbw@3j$9xJAHjnc{6vr_0qlImB(UCmgEq`S>Y@i z<9(O-f2T0z+L#)S`r=ZFpk5Bs=z8UP+R&;IGOUtVgo1iUS?m!@wC6KU^89!6?biW{8EjNPK^a{RDt&dCzRKt9crd**W8%3lSL09BJh-Ry(M54pDYMi0qY?{j%UI~EXMwy%~e&W z&coRarZ3Irj~q^_ClL8hjVJgS^QcTioe?v!9O&{ry1Oy}Ir1)ITKgWd@O`>)fs$&coP6lM?wgl$ zu(rpoi!gvbCI43;Lp`_^V7wGUa#dLn*!0U#_{NL=)B}$g8p{h9+Sn#hc#X9&0%p_4 z_xn8{mq|0#UqkBGltMIcz4R;2<89hx2!vGM>9Y=3UV0@st?o)aAG2;GobdFlP=Wmg z-$UJ6WT&h`aQp@R0}Hp@;e zcM2-ky z^812p?VjVI9k0mCnZozXb7O5F;N#v1%8nLMa%olihjeXD{H1;=ao$IuAqZj6B69-W zhK<|v9=l}3(ki5zuKNW*FdSUD(|K6V#7#~>=%ll@74A`48t7Y7|9xZ;n)Ve|P6QOe z-_yf{7-h(;*CrqqV>)j>{3U_vmvV7CJ?rN?s;TfiazEVf%+GWf(L!Q;k9r}RK!B7b z4}C|8$c*|)G_^R&Gh3{9k|I3uT8G#L(JfF*X?YTQAu|<}m@3Lz$8jJt-a39}?B!h* zAtnC7I1CGZ&DxOS&{`LWZZ|etV0zwNs#igu)E86@UWpOFqs%ZHP9I0Q*9meL?U!ab z4|>wfZTB&WI#bxZvFd{`@6Mg>svBG|AB7=LdsTbE`kaE*05Gc3`vQ;AC!`0WHVglO zXp1p4Z%=6@rEue?*bJpcVbm5L#!F)pkwr+{U}iiZZGvD3XFQpNK6w53S!F) z0bD=cS<0>qg3uYYj!bWH+$SJro@db{gDa46s?Dyn_tD`9Ib3s$O@d3?{`35-Y3v_x z{F2qE))?J@^HYLT&eDfDRoWyXdg5W+8U=CgY|qEc*$HsZ{Oa|I50~b)9&z{@dX%b; zotm4M09<5G4>c;(nxkSEtXv4BFk-G30wV#aPaz;tSf03Y!d3)*NPO`qSoRI@SO^%4 zLT4C7A*rA6WkBHPR>;Z#zSE&Etk>@KTluG2&8u=wm3V2lcv|C_#~K%@PLBdXSAt`o z{bfhdk0uu9)P^`?xtCcI1h{=%xNye2`fM#I%9uh8w48iR0$(X{_NqAH(N72)%nu4w$J z#)&%t1m}aws=haB*T&zFm2Y0E;6gqXvbCz=wo2nPxICT)crl?ci7hhXFGZZ92ohjK zY87vyt#vrKm2zr#{wpw|F&YF&ZbCVndK|-kE!i=bUF;9rR6@<~)0VzMyfBj2ZeUNY z9IQ3WDW88LYbe8=-aY==R4kWm1zz^H7PEGx7*s>>^rcCZIojEj26gWi z7TPLTH&v*Wpm1}ARyXVsO8*4i$@H6b@gG&FVF;PxSmW)EQDAx86bv)OMj%&?B&ZFq z^JqAs_Al;`(oyOb;OzY5Hj5+Xc;QoZKMHpc2|^ znOJ9Pm+(I9<$kD7D;*;F;drg|NQ{TZ?AK0Zu-*Qw^{FTm)(aL2JX4J$rmGvh1}b0e z_RUvDO(22W`iklEa@s zs(Iple3CWS&$z0a@ehQ(AHeBosJ38miPe{}<#L5e<32gfEQ*y3(4e0?V<4{q9|C-P zw$-B$2GPD`or%AqZP0~=~xl|wgtHZ)`$aK_-b$8GBv|c zI#BDWb~UtHLz_a?OqJXRM_WoxxmscD)^~r7i!4i~l@8;Xm8dSRmV+wOA{%zVXV9Fg zAxC(=)t?&`dUg}*xWnd#=}}m~5c>i%)eN||R5slQ&A<<#5W|UI6cZ!JkXt<2$n%#1 zcL5#QwLLFYTrg2k1i=_#QNpxHe9~^GJr@3lv2zI0L<`n!nO#4^yf#V3l)YBVoY+q(@7g z_rW|yBR$AsXeN5OV9$g+e!BB0WTJNvhb-8|xGZ7}2gpOOz3({u+4jt9UjC3Rb<+pL&FUokYY_%{7zr&mjm5gV*j*7UIc9F5v3? zAKnBk>w-3JGsZYTNUU=ZOz-(0Z|Fa(3@DP+ zQ*hx%{S-#tl=bAgmd{CPooFPdqZsjV_|ivVd9G)?)&PJ^*y6cr5DZ0DD=B_;j1$4TJDqNO~$K&w6>~|h5P{H zBMDH10-whAjz{2hns)O#0Zk6?!8XPY#b^vn=kh;c*{E>2KU=23XKzawqAv5Ls~ctH zG&`zFlSONSws^T?s)P+7Y34gB>BS(_5UJk3mnZsg*x)m!d%Uf{!LR*-%6#n3B z2Ti?yvEKxIzrt`#Gn&*OMJJIRoo9Hh&f+3(S&#a^Z#d~hi7`}=^S9#&hF#kD)46({ z!gCWb<(Ql;dVPXAdIV#K{C?VY}N2JzcwYbpeM_q=uE zT{?)!-w=;98PackA{z|qVS8*Zsb#;bw!TsZu+`xkYyxrYsF1GUwwBPF(YX9c`Oqt;67I|X@;15t9qQ!#54P3_qS63WXEWG6mh&dkYI z&HnmA)v4U%g#BcvC~=&FmAM0};H8DQs>TIdj8zelJNt-TNd=)ucH?|u7d%*srbCsB zUn_5YZyI?ISqNQ^!tKVEccqX&vHxL$;6_hAQ3fpJLQX%>#`WC$JURUjrkQ6x65cjf z=A?&fKL&qj{s^)f`Z21DD(T(*!JrLe7`7<_64T85#kDC12t}clF$|B6u>?ZN4>mWRcwXx^?aWY znjWjIA13D{zHiYWu+l+=7|kG}RqroU2^8o>?*gQ3Q%u$rUE&1CBG|X}t`uSyCDB#1 zHFCri3vw`?82z}WV9{Am82=wpp`C+fEST38c|fS~5hu?eKPcLa_9)Gm-2qvoz{Eb{ zQ#oi!vk6B|`GvHR1+V%^BPd|_=~?J8nI4YWJJ&9!_KZ=fESW*`COXqy0^_DFF27}Z z<`TG(RC&1$N#@fy7=80HhUt8(ac>;|TcVUqODq ziz;1&ckUFIjr8g?)H0&0`Wg!5vikfGPX?89Wmr**>->S>lCs6CJ%ziRS7)Y;TN+mA zCO=qzh^uosdDn*m2imOwOzk2$3X}Ck|9doX2t~QtGKtL)ERk?;fn z%|UiFw;}0LyCw_j2vCnmSG?Ivhd0Tft+^A{rdQqsm&#%PS$RZi7))EQ9tNT;F2#4z z-I`)gUeOIS(I3p|WH3nFVB`L|FV7(A%ylrc3=V%ub<1Glw9)EZ#X(8ny;V>vL+o-z zOPN(DAnE5y8x#pm&h>cUm7n18b%Gx+1;gM1=-YIjW|$={?je%ZpI!p9Z-iEbjzm%J zmE3?vx>O)E7Hogw90{JfIJ`J#?&H1h+oeg`E0(Z1K#KFbklNl|cVvq@ z9~7en?BHt}EQykJo%|Vbu+MFND9Wu(pzwj@D&IJR$*#`i^YF<_^Me%bD1}q2Cr?)H z#xk9)*I3a$w0NmIBJb95JaG&CB0s)W<9-k)k2Eb`8rt#aylD1iPlCrM^kY~o#Nv#J zEQ_xzbl5?7L6^}dYD=h)4u^42+k!#Em@-F%Lfl|P-(h9XZVGek%lM-yeer&P`e{4ACQ6s|5u<-v6Oz|STt`tC(d{mZ;##H~&6tDzIK8mvo#3MmQgIbV7hQIf)t zVKAgBWV;lG2BwZ9oUD)VJaC4*C@rp6ULi?bwE)3aMDOASg8c(-Yr%Zj6= z)V&K|x(Gj1k&KMVYzDB1ZBwWqvo}Xx@H3KR2PCrR@^vf}Ydvog81)-zO!Ei{{0JYi zO)7VAEgykUsW~l(?X10+=c;VKsE?u)nao~7E>8oN1OZB+KBPia)cwOOA9>s%;3CIF zDKfdEP=n5>?J*sHKu%BXU)Ps6HCiwiO9R2dvY8GO%&@Uz4*>LE_Aq9=E?=K6%fhu= zf&+*l)66EyIXw$0M2Ue1BiPLx|1El1G3 z<&*4YP)Ol-93Q{Xl+Y_LsiV?;n$1GT<$2*8zL17bj%f1IDb3Uv;Wd(t=|q>38HEJ(yemGkqrb!T zJ~8PK#87Ae3?gF}RpNv`ZH7&jBNi3Sb%@8AOUv`q!o%8b z#i_%n_htID2)1YQdRWZnAB8U`ah~eq4LJw3 z7aR|l2orciDX4fuB*E$@q_M!!LsID!POh@x8Nd4ss17P;q)#BUIh?jbPRN)Igvu zz8`l3H|JwRxjE8$+ZySldwkuIa1&zVkli7F#&K&R42=ZcsPu^W^Vsl-rdYZqE$!-q z>fWqV+?d)rIMcF7mRtE?+CqLJ(2xFHvbQk>f^n{2Yhf%B#ugn7PUqg!w`fdj4cNfX-Fnhk^mlKWnI~dl=Ws!T-4J^DO z`3-zYt+G8pcr!YLjcB{Pg|Yw=KB9-b!N(m#6ABJe&uRV^)=sg|!4ipld+&&lgNVFp zEdJN4$)fIHxX`b|OdiHOskhn~pAPKP+r&ULi$QJ@)Ga)&MxSE$qJN?pBWb(CLsM(Q z#?m`7_;1%htsBEb2rRMx$5w+&NG4~#hh`!&FssKbvwBY zmq9462ndV;uV5AHX~-7PE*qujvy|q_&lF^8)TC#6G&@Np?%^u&7kVdA4U(_WG2$? z;8daiuSWmN=3!@(sE)VR_Hi$8BqFbUlLS>bR2IHgC8E3i)+!FX=sF${nL=O9+rym} z=X|d?NV5}D)?ck`qq6 zj7^X1GlACS11~ zbFpgZyW>+%bzzDNgHKaKH<6i^B(G>c5KjC4BASf;h7}VAG0LR_ZMYnLs9dYWMl@M? zMHw%xw^yRO)|?``d$vnh>o8p+4LOEmKi!%&Zc(L6!a!Z~5h<)}H_N}lSLiVQ-Qz$C zIgYNZ0Zwg^skO+1A+N3vVMu^{`s@g3CLxyw z*be$bk|BWqhXT`KnxzkTiJ(+^YzP*Osmyn>mR2@k*-T#F3B^)YCs*tYE)nTKhe_ZlPwfw9FEtiBBQ z_Gos=R%eOZ8y>ntW#mEc;9v8I{er2rXWQRK(RXj`e;o!wtTyc*e17nd&e6J}8`|Pu zk)RH1_m@sXkHOHC%@kVW3e;**!oD16w(sxo@C44iz1GXK^GHy_EB>cLSifLtzn+yK zf-vQZ%uglE>R3WH0@v6gIOPLZPIwnf*ST9|@k6*dwF&&)Ah({l`@u!e!bHAQT0h8% zAmOMvo|h}*@M0592ND~Ptp0{WcRKx){$_g}#!S5V&ZwM|y!LX!zotG*Cq>@R+wnYiEAM1EjxTYgs zP19Q;#Q%tD2MMx+`Z{ETcs`^NDwf#iEc~a4v^92eb~G`t`Jb1ap(P9>E64vzl450K z`S<@NNwKmsGyWe*QZ^k<_I5VNv^bltzV#l4|1qTgD{}IdX_V9dR{YqzT+D7KE7c#C zzl-W}&lk_m%QZXAFd+!U7PmX+29{QPV0kR1Z!ReLMn;zUre%PgY4=(O@m-vFSe)l2szQly6{3FWO`OMX@ zvHuBaDFGGNbyo*fMwO%Ruj~vhjl&Zd+Zll%KHD=kGc&Tj({IeJEQ~zTd z2VwZu(_5Jv9$x^>n->4AB{cY2ebkeB{VfO%zT)XR2c&ZVIrjqAfTXRfrhsJ0RCj;F zYU(SKlFG9udrf{z(|$jG@yL~w5R{An%_d$(H;9!_D|Ik@^Sg~Hdn}EM5EU$w(tXrQ z-Rv74twMxFGuyT2NFm@UDl- zohjuTntJx7Uo9--O=$(AYGU5>URU@BU<2LuZO7`KX(OpXaN!4=VP5dzxEN>H0Yc{1 z0ixHuaPM)}YlS%V7px8MvBAnqR_Xe@9AyJ=|0Ew+SEPm7a2pVhPc4-Pnz`IY_$QU8 z6_HnG0;8E67S;2P4$R@Nb^EdF|3ij0wKN)XZB}g6Nq=ZYeKm#L0ZI{;7sOZp76LNI z`F36Ch}4BSJX9qQoKnNNvR}0mTPikO4KS+^zFD6(df0q@x0`VyC_}hD3LptA_(+cV z^|1DJN(HnfYy1y^2+1$A<_evICsQ$>a{znjbuO3d!s^2BLRic*i~ zuryRf;J^+R27#ZcG6?@%2(n8bUU`oK1N&4U4$$#OpC=^(<>RXOp;D}9q&J_jDd6Z%(R{tfGQH?{6HH*i^4_K=9FXtCe@EGb5{`O0eaoO} zhF}a0+*(yix!5Cw-Nc4np zjzfx_j}#BR6EQs}9KY9krK7Qz%oYh)F|+etPs_Kv*+b#m4)9N0c$L3gx=VCZ-HC$} z5`6E=(7pivPDPJUj-559pQ937PGALpq-`7xY3pv!%&Hp@8C{mgrfj zxCfNgxY~g_CmU2re`I2Pkfcg{L4bbIx#cvS&eVT&OZIkP9!)6dblS)!F{wxb&Q%j=h`O z)`|6Hx0Oi^QEvYhtxDROqAOJFwC$$*tvHC%9(a}#^+v;b7DccXhI!R~qk((@^x{|l z(&2KAHmD5Wd?-Qr24F7I3cRxHfz}hn>O`bTNpbkw44;qn<_(`$%LcOA{IvnG+3w4Ca73PY1k5uDxpxw;h~L5K|Mx=$!%iR^?B zPvYFlaKDH0y|@3vzR~MW?RC*9ORNdkR*BT}SUp=G_+U)!zc zLlMy)U+M7hxti#vmG(H;t!96-v%?6-M@yaG<8FEVf zCvdcc#=>-Nh&D>7qghHk?z30NMr+WO!^^;dbj$e(#l9sTaKg z_W)V63ShiI)~*IgTegv8Ab0=8Lr^1XJqqH*l{*-J?W+Nm{VMz`R%|j4-{$)^h!Nzu zXTNE?Z;`85SK7}F|62PW(G+d*P5M7RFJJRmebPOYLc@YOHp6FMUtt*ayG!REYbl2K zi8fRTh0sN96QR;8x9uZMn2WWVhshXnFh5Z8NzJZYBH~c!+Ms<#sXab@uHy@rQ3z|>$oV4ym=28poCcHWYKa%XG)Z4 zq}Q>`oDrCQ;ZD*BNBfjj2e3&2HV8<{FkiPBn^8-F4;%VEriBWIG9?+qPB0$Po^_@N ziVJuxc$3XZ@x2&j(UY41)?;gJ9*LG7C`sJKZsPvEHj7j)TImvyAEF^i5VS*O@@zkH zh+i45?@iaAOz92p_VD%YA8Di2fT&>M=ml zKz+g^O8}xy`xY(i>Byqy$@K7`w=LRmBwd*p;9}Ax5JN7`Z!w_p+2%M<|2}rR_i9GZ zUbn$N18>D>5?3YqI#S~M!+@{5BaO06FfzfTpn?6SwTCmd*sCnZZ{5hjM{|yoIruq< zs!lwzdm~f*sq5FC86%tB@5HlO#5)eTaP~oN?0p-RyukKwhB_PYS7<~uF30}+gZ8>n zk08XIl7iQ`@~^bThKu@=2Z@Fc^J0sLI1kqTq*ohFnsBkB80c1?22tA9u^@J$>>SZI zjElxe?&Ybpbh~n?QxI%$`Mm^bA+Y^Kl#DA-VXFdoBwLZ%`4k+K#0?t)!@fC=>*4~0p7&$yxjJeib9MU&X`5>gZ%QM`MI z0nqhfI1Nz*BQ`kpn7YKjcER*l31>VfC<1i0Uodz{#{>BRXeOzaL@fW~Qx*K>zIza; zyQ`uHO6G70;U~z^X#^Q3y>kO~^oaOR*Tx%|2H~}`F%Tp5J3B>$xC2^|z>XiM3%#wh z7`Y{^?o>}E9;Vev<&x9qFh0%jZtfp9BL3zO96m zxF8zb{QdpDfp}d6OH$lo1h-E^RaGQWKSq=ysejQNPGe~CQZ*^%%8C%9N~o6ktiwrs zPZ{0iasLSwty~OTl?lOJHfdAK1CyH1o&PY8$*m@`ecy3)q4+jk+6lgC0cPJ)7{}7y zFgeeE;S@H?P2lf=sp}c+!&*VcZ{j9LFeGh~XNVIiYS#dCP=tlb z5FDrd8Vy2~rdz`Y@V8w46K-&V#`y65;<)>(bWB&xWopuWSBrrFX-I{ix;~$W3)Dbk z65ARDL{_0i_rkEc+y9}G$dlc(O7L+0s*|hl@%!()=8tw|>VM=Sy123zQ-9$yyU8qN z0b&xHh1<(2L7r!{4F%hImkJV;DBCqt!Y3qExVhcd2S{ zKz8ARMBoJzgKBvx_#^G@)&>W+&QXY~!c*9dh56pSjb-SxKf$pX+y+-bj6Q?mbY>ae z1@I>kC#?B2|fQ z8rl7ZYq?%{==so3k^~ke907R_kd+GqR)fN75wCTb$Jd?HLz3mYT(7xw@cS06_?owG z?le4r>-51E9G2Ly7(F8O8+QQpTk*VvH*q5{ek4>p*2kujsWb@93)}hR!5QX z^Hz!n!(QfnLL2-f$jc>uKj!sBI?DoWO}+VsBWHF8ZoGp@?sxW?67e1e14;ajE#RAQ zv~h9mwA7E81tz$@(>5$cGG}G+% zm#ys)$pNMggKBs%h2d)`Z^goDDtZOs!k9LhE;7fBz`x zb{HqaYMQ)2U_%lbn8Hf6iI7MePB#m^B=Q+a<7wy110~eOWiD3=iNU&eNUjNvO^cp* zwzO-LhP)CfRgX4$s9E%L|8Qd3&W*~?Ol^=hT|aNjQy=MyxH zInjdJ;pC;IWMTIi2i21E8Yb2NtA9#}^k_#s+(gwY4+I zANc&S?*{A0F*4Os`dZyNZ*zSU9(ER7P^{mW+;U3)=#&aIh&XCL#gItgJtNEttm0XK zbtqQENZ+uxbJ3+!rU0w-ve}W?NPEl(6!f!xRasM&3zB zybk_%-w1&bYf(laoX=nCtY|_fffx2APp!0l*2g1tQzq3vRVHrgStac#wU!B(c%{Ic z9jURsiA85&{VHKReLH0hj|8}n8rn9bh8#DMl!@dLm_px}mu9~5fbywkarhy= z$=p%~zkw`#{9LgH)wOObcM#eNg0wX5d%3tN40Fep49aBDDxH| zW^Hm&sh+$MG%o)G{AxQ&eQ>U-OSklpXV9rfEgf8*%6y=l0Se67wi-na_%?Ul&8>GhRq=*2tf1<-2 z6dPB{(+g+T38YjVTn5bvX*$cbW>A}XhR`c&W~wk3L>47Tj>utr1GE46Wp)Rn!PF?$ zW>Gk-+s#3H^JCTpPKvgeKQw8y2b$FVvJTIWkY#+)fi4YpL< zUP+-ICpI>p#K)x?mG!|$q-3Exyl~DtTCTLEwwe)c@<<2EYCrB_CJO~JyE#b87Qm%C z3!z}q@M%!zQJW_|a#(mf{v2bF*ld{GkNpjY6}yChYH{c#AlBrLiN5Ad8;I$aec~6F zv2JUj9Q8&TVnT5y%1DN<(K1wAK@#~>L89HIbU zR87d!HZUn)cSHUC!jb0y^^y`YM3`RW_`c{uDA*H*dX?;5q>p)sZ7bDB`$24?iF|ob zw(-|XV(U6$qW%1GnBDmvBZ;yx#H&loX|Zj_F|dH`TxG_AxOeZ2qu)eX7gP&QT=7yU zJ1e$S5nh~Eq>E4nED%mxu*f^$;Z92UNinbcN-Il`>vqjUJE=q~-){44wO^Dh&4ZuP z^QQ-dLn>Ni+3}W+M-?H~g5MM&Y8}5i9f|-P$JX?uB7uQM+r6J;^Kl6*lLmm~bLMO= zEoyr!Q0@a+rwzysn&^evNjC>ADM&PgiL!G{x|P*soaR@#2(j`rXemN^ni!9{$USF0 z1O&Uo^7VUk^Gac$4M}~&l9?tLvk4tw3&w%EvJLMPm}kQFCi|xRmGO z$^W=dx%`2685)x^J$_-BA$yqRv`pFO3Jf$@!UTWuqQoxvgWY!4e;}7hBMxBer-pXC z6G(R0#{69>hp$F&Z0gNnP|JH1I4pe1201ov(&cu-aQSQ>IjVxxfzy)$UZv(ZP|8^6 zwXlvjw1XvQmmWDu#@1|TD?3&;Ka?!6UZ+@U!s7%EBsGCb`hgY4fTelE}aB8`{XDr4QI>Sk5t#a*Ym@>)^!S?DlYA&*v*J3G4U3q3%2+$&Qj?*I(01HXd~35Wx1SC zp`@70$}AC@e)Ntiy!{7SnOH?ZQKF{(-q?sH87l_j*dt1yG~S>=NnF$mM4hR}a z*Yhmuxq_ODV9$`ZtP9HE9U`3Wh}f7}by8x@Mq4*UXqxrUTWxNwWXyB!1NgN-?KZzg z=A&3D5n>|auqkcHT01jDRV#*|y61~!$vU`pSU>$7HjW=wT62DXy$0wZL;Nkq3(5-4 zFATgtnu{hJnu-`!W4Lys7wEtt3Ktyc_|!Fv@3+fIh4hg5hI#E6j5B8zCm$#kw5ec; zTHr=^?e~wZfJvr9tL|POCme$oX9|a_(?Fg1B1!$dcz^I(SRw zX^UH?N-BVlZeLgfx5hLTVTv-VFlFCxVWNR|_J%qrUfkkd2s>yi-~+xxnyM&yWc-^g z2KV@da=+kR_VK3G8tWVX=X-S>I;}Uv$r1G{T~0xOv5>^0NdP?T?~AkSlj+8}O2%RC z3Ya#r1QBTI?`tOhU8OkR&)TCqGpn1#+n`bUUUNBPJCvj_TbblpnqO0kY81%;8;?+Q zgJdlSm}uezSy@(!XRRFd3VgnPqaWcV((H7fy1k#xjckvyE^Inup;G06I0O&Iw?=5W zE~94Iu*n9PiEvW#YiIm)cfud*BeycOrG+Oj=nmSoIGwX_S##VieIXma zM|%RdV~t|GoBlppMw%MUnt+7Ny&=zzgWw9pe>*aCi}eG|AQ^+7*@w3R`0?h`L72e{ zh&ngQ%DL1Pyz$TNss_-4q;JFXv-?h(TuES((|4b@?d{u6;;CkrX{?t)mK1UX;x~3H z7U|VfL^0z2o`NT7o${8f%n;<9S_*ertTe^ksq7KE0hR^g#8_vi3 zG9AOy6F)ctc;9)(IpU%c366lj)=rw=CglZyM4rMZ??i1(xoMDrFJ8pmYvc9k7o;7e z!7*i;S9Q}zyJ~NsbfeYDgMFTLq^WcehXX-HCns|#$bkT z-ib=L@&c;YC7O~!rMGn=<+P08xH!M84uNKma5BfJk7ysQ3mR>P8!;~ML_w7SRo;4*5dn}f6po~-=hFSl+)S}avO5Y-SHgDP2?oI zI|Qd|T5@tH<6mW9@L^FYGnGJwtlA@ly}ZMvF5iI9He`(RZfdWR3%T&&{w{I`g@s5l zwznHnzK~8b9f|k;V|2GX2$q{8^YDfzlyQ1r6K+^%w0ZrdfMJUS387DrbIyGi?zBvz z@Vd>Em-T0e(7U`Goq`j~hxwG9izhJPWwK<(mll~Y>K)a~Q$*rI7X(>LnfA?`adYxB z0Xor2NC@W`V?34T*5?&>EHBpab96k*60Y!DkLhSP=D-whK0+VL(d`O@f!<<%&zz+R z&JBM&wH}$mVRZBb>oF zlyZt!xmN<}#goKB&0 z*VI8be3qK)eK#AifIaV;RZ|jX?7(c&Ola!AYG1xoY%GN`Z}vvRA5luv;0;^6Hq+pZ zR^yTs)b~n56MpKrS45>wTrL_9QY817bAe)<3XaB-x+@PVDkS*gWXK z?-ts|fAhkE9|cB9Td?bwR-j!{y?jI5%`z?=1*2FkdZQ<;lH^KqZV0j8k_aTXcFIb7 zp6CY%P1gk0lidNLho*O?@(Y4)b^WIPlcydfyJVFvj|*7y)ay zwXOH^uC9#4j2~DN4)+zLG@K&bt?D27A#x(UszBP6Zq^l<#-Ks^E7(_=_HcnM*k3p_ zVMs>(OEDI_N`Va}X4r4o6mzw)LJ>7Lle@xKed48PTHdt1K(Ah_Ur&Fq3B}TO$q9ml z<{ZmbPgv@_m}J3Q{u(@X6uH-%)uUpda}G2i(};JreE}BCcMZ48gh77z3L`fkuddKC zVX6P!*SyAAOYROS!G+xD3TvFyI2i8Iw|&GE3N|n0nu%A30%MU>!bRMmd@ZFdK`33t z`ix2Nv_ojM80(V@DfbsQj0@+UoD|_bx@(6HX{%{a_V9|^$i(h|HoLY628+O%N-%SK{KrdQ=#Z(V#;-fz9!v#7MCV?!=0ApmMcK|+`j>kd*nx(=> z9Oshu0oh#yUDok`!m=xt3@Wg=F-R~cE4X%U0CEstuuZ#?8ron5NSsVFod4qH{VKb0 zr8;s^&nvZI6Fzr#qK!s}Homjfa98Ak>5t*40#b`&u_gi9K`U2P5Ox#g8z+-^sKjwq zmXUijr&3Xacip6Xv3c~Z_r+BKN>BZ$-YT5dc}L_$9`SVefH5Rsbs1xy`|JJYM9^bj z)-R?!7q0$yP?L*VyRFLU~sIh#WaiW1ZjjP^NmD1leL zqIhv{M69tFP7Q@h>K;bUr8>$cy6xl-dkZ~~fx8?CNsT_?A#3p3fRU4UBtW;Cax`)A zh6d%b%!W-S`zBx56$bqguJO>Ja1NQh#oWs327N`oL3*9W@5pkpohT_J_|r-5Tli1v z(7ZYAP?_OO=h0vq*IXTlPm2-wv_^q9AIS?Et|8b>Y@!AkW}2(ewdufu&sS25t})I;Cf0xZZIq|Jw$F) z_6t-czrNyDjuJc#%KgX&1pcyhoNcFsA?k;Pg2A3(JA_{is!erddW7f?0q9_EY3QTp z{ZP@@13GC_@2z{=@p0L=rraTLFBd2eRbUVJZA=O5L|f5yKXc;yv@aiJ4Iq(Sq%-)f zSaHN2mX&*g|3y}2ZF$UAN~3&;KP*FLD9M3~3^BSAc=FoN1$q7|L7H;gACi>^L+`Y! zIXD~{cy)S1sor5*=_!YsSe%KCRnfz7|F}E2-79}!wTFIJ9%mm>9nGzyv!_!T%htE8 z{u4^kWPd1mq$)k_jyS<6&V(HjsW8x8SbeDi-i6#akFke`1Nz&19h+WiQqvdH02(_Tc&j--4?7DK~Mes(t9xc8Z zYI_wj-nc4~%H!zyweSKms-DeGn<6ZOJ4v2y+%vlyfo*($VMYaviK=~em1OG^28v|N zpv4RK5wya3+@Q@UjSZ!BDM=mtvhi?pfDZcOh@oswJTD>U6hmqt{I!ls_1ePz>a?&LJ*&BE*1ajuH{ zoEyfs+GL(~1dfZFCR&-OqI}+sw{M4hcoO@h0hkCO;hjMwm-0UhR9#BSK5 zca3-qfLnSYLI$Haf!u9zo3s2Pddz34h8_zhj5-yaorw>$t$W%It{`ho1zygK)u~N1 zl5)OQ+=K}7*>T9uTG9YVdNKxb$e~g(-)6p>G;twFzF*eW(D($ZFLtD{W_ll!L!rYF_@^FDcwihhbIjzDjzcJEBgA9dapi z>&;EhdXMmtelRobFfz#s!JhUF5}F?)XwG`4Gd5NE1dnB^bFfSrL&_GF8HY$OD;0Vs z3@b1w56cmQ$U5NT;_lY%l;(td?|H?6@CMObSzyFF@G9Av@gDd*FYya?cvj_kVrvQN z*n*$rL}^EM)~??MawNNZJa!x!UAugAGn0??eTGu~YdY=Ps@t2H38R}vgzMh~h^6Ez zD$Wge+4=ahuDIyvjFCfDrl@}HI7~ZO48rf%38neLw2AI~9LGPRzj%4snX-_cB;@S8 z$1;{vuB%^evlq(a;5LF zN@L1zECzqnt7Xa?Kp zeXfb1(oMMz*(y?5+#nb3NkjoJ|4^IB4|<$bH#K8 zLr1|o{u|Wq;#>-9x9e+j`11|&lmBoMJ+Ovjc%@yP4S8Pv;+GF*>FO@K{=%9)w#5yec)xk3XNx^u z%0DM4nxb;6pJm9VLEcmC{M)mKqrziJ-f!hb(9Naq?oQ-W8=?UAZz}qc&nC%oH#5a5 z_9AoEn^1HicuTdE-7r6yCLdb0{r3QWvs0fog4mygIhdqADG`d^6oIU9q|gCse=DdP zkreH59lOE+IgJ@;9i7G8ld5KP@?*-3hzQwu0Ch_1r9&|B1R61%(eQ^x+bIFykbbLxXv2~BD4m%g49Lj>K~t=v<)Wjt8#-oPXl8U zV#(HG|L&BpAFBP_cpsV1D2k!AQO?Ig%t*k+vpY=NF7rgZDaL^8tF$>UNPz?Zo!HCF z1J!(VNs|@e7G4<#o{P*jSO4&QA9XEsg#M?^1{(})Ui%Lkltgf-hnEn9>{%MxE=gi= zj1(pC=kW`ByNCXCD*{UXSEZJWh|S!XR%b$^4Ztq3QwhkFNZdf6D;DXO|;oK<#QzR$3U zdd)yVnaE&TrS*b*wa=-HF3h|0W#$nGt)+`yF8J}}-8WJ434EE8|HV$k{p_E*>}O$^ z&4tIsc2PPBP%O+lp3Ysw&lC-Afg&CQynrt7(?5ySHevUUqJ{Ns+0xh&+b_@`cDPt&B_h=totZnNfJpK8Brx_izDJE zM77+;NYtwE>`-~n>-zL>uMjh*YpGauUoQ8(dAA7t-Bn?pB zb2z9A%kj=U_A~P&sw_3vvT$wOg$WS9;dvz*IP_JiR03VW?999nGfjfsW6^X>2i52{ zs2bA%CuSO!qc3TsgFk1x*9m{E``7@&D1fIzXiPN6ygg(81nf$5I2nCrp1M#Ui+_Kg zgkw}pi3NYnjGmif#CV6vMuPm^g&1N@s8sgGCF9ZcKy{~FG3!DP*>WK{AvCp&x+S2H zP;%2j{Ma29)lCqk9(})@i*R7DF>t_LE3%qG)Fne0VJHG~#7;zXFAkXYfV&yf0i^}+ zqAWrq8}u_F6A?qH8l#O5xav+G=F|}-x?`|b9@ZgexC`DPJAY~9<1tzT+& z`+tm`Ly#b_qD9-bHEr9rZQHhO+jdXewr$(C?Kgk%c5m^kvPdfXR3*9R9@Fhw*Sl4g zFFDdtZX^usjx}7+;b?DK{#fdZnu~oequM0WT*k00o9J?Iqa=>{=8_J^N2b|{m{`z+ zEtU@g+J{o_LzZI4AP7EL*qj}S3SZBFl`Y@uzeEE?L!r$#+NB)Z1KT)%(_8)>qO)^? zkm!a0cpLxlL#`VFb=4VVHH3SCwO3XRb8P$s2n;=s=&>pt>IHcQEhx5s;j59N-GmF< zSvn+>pXluYVFV-#QgCpn{=JxQ6Yl4IK_{Q`0PV>qscRMdW(&ZaPNo6)4H>gV zT31Op$e9=Rog<{d+=P;A&%;rMe6<;%e!R?;Wv_xfF8s@GTM#S`h44w|9LsM-lg2EjfJwi{J`WS3`%Z54 zU0`#hxT0;_$V_)VQ(wE+xw>3C1jzMaZI!%*?@JhT%bsGOX=T~O%5+|U%i1c$6 zwOs9d847`HBtE8;bOTpblMb8;MzCrWynpr~nEcjWgA3#UjIxD`lY0 znumjf3LGjTE=G5snSvWjm4y3es7A4UuZ`|q2qXg$c<3gO;z5ef_P}uF&#(C!eFQ{Qi+A5 zf{h17uhrN0MwMYK+&&Sn(^DZUmsQ=@;lQ?Hs8S=H#s{n(Ce&@*@2F^=cugWgc-p4$ z-@uWm11C=VK|Mb#5eq1Im~M8wDS_W&v@60gvv7}X~M!3aqU z8AdlQr>B3?@%HCxkfqUUvmK6}q!7R{j>q;E6FrRS! zse(5oK?URDV*yoMoVk@OWrGonMmdm%UZaxMf&Fo83pc4;C-ZdoUZDlQPx#mPe5A0j z1zM%Bh|iuq*+8m)59OQO8;; zd!vh6eeR@VY4Giften|_Azos3&B#DY>5Jt_gd@Zg6N%aNtv)mXduxHxFKHXOR0mVm zqjy$a1C?snfw6+lF*04YXAaJlowY zsRl96M0Jge7YrPg)`~b8)$TWl5Dg!5vPElQu%lVlJX2OdO9`} zFY|s(cy_1KwyD!VYS>UzQQKJ#+L0~AEC7P?3J+_pODL0qWU zXb)u;^hc-Mt>Zw}c-?y9_lNvS+yc|l!H~`ZW&qSAk5gx0sw0}!5Zrr8{f@-YZXxq< zvQ@QK(EfDq&bfi26~Q#}h~?e@yXV967Jr)Kc8+z>4e#=tC{UPd`_c&TU-S!R3m&*5 zK7wG?vHi=IKSi12vjV&Y)zXbxwVT&HSS7){GJPp%_-5TLGz=(AanPdiN)m&kmyp%F zqE*yuN-}yxcA5&3YrbPr_1q5XYRep^nIZ1WBrzmLofeFBdx$Rkp8_)mEf+Z=X|M)V zJajJ#fltutmk$}RlCuQc7x;0dwCl|2+zonU53$k)8GFXM#{7IYD zZWLC;0E}r1=OtSgk4RZK>Gj%qRbJADVBg>=A>{ElGjCH!g#f@erK-4SG<0r>StT(6 z_>wWYgr9@mKPe3h5P?h;CC+}5X~1BT;+(hj-1Of_Zr9ZT%{+g=zyy3%MDW%Ur#(gJ zk;!IffK;@~JtfS=MoqPi#yL_+4M{<_w1HcZ?R`XE_qVG7j|9W_X3){&)Q8YxvQgOEbp8E8Zwzgu1Hw)4AQx+c^ z_H3pvwQoCYMBF>Ig(h8e_TfBC{ly;WeJ~@R&g+mUUfq@+VT0;Em41x9v{RdIdO`TZ zFh1I&P`3;(4!3nLKyjNn0CRVzeTh29y`FVyz~TqWfcDVSEM0MbUH_FB3dqTf`91`2 z>VrQvGslz3Yb@N)aoU)7QX_e(EZlR^bTn?BN0tW_uFGF304Y(OIQ#NhCV|wG_g_lo)UnyyCX;6Qd4g> z@4py#e$^4kPwWiYW9x#YlzrJ!6haGDQ;S%wbFGxQF>&IYy8xKP%=t9KQ@=K6N3gK)D5L zW1pTPAv(>w=zr|uHvuMdaAs@UsBVZN&iZXpI^f(qr40-nf}*`C?StwLZ=^ur6c<)B zhG+Ng!d`{bNm{`&&SM@=^4Qz5|(i&hTWJX(j74O>5uP zLtY_2I-*5Nob7bUTQ}Cmu2&BkNIN!9;|!jnP~#J)y7r2AfElR|xKUc>c%1kr-Zo7uKlX*_r17Vy>v3!L4Ec8HZ7#h{_OEh!LXj$u; znP{PMv(w}U!Kw@kGnF67V)&S5eC%YqN(vy++nCD!1K2&kJ4MLHAqoI41Kj4vkr$tb!kL_k`A15pG;WiQ+(ljc&a z$Nmeg*K+n9x67J#dg}Wj3j&c3MwxF{lo~f-^w2I%1Gg0!9s1|Rld=wdMNnJwQMiG|Lj1nQ2bM=tfzWw{V zy<_Rg)Ls_@sM2QQ5gsNhMe!Pl2COz4%h=GIME44MC`{;B;Z|hVNnZelIDDz@d@W3l zpuZ_orVIM^Y$umNcavg@$dOkOWyqzucbFX)=FAZAK{EX)9%=pPf)+K^iMXPSKA}Xs=BH!9F037}h|MeEYf$P2(5wEV-hEZmk2VIV%xAV5ojVfPh}w@Mm5WS_c85bjBzR8#36yqoQk>Pb*uW`Lif}udR!P@S zd#ZY<|Hz?49^acf;Iov`LY_HUvW~>7u`r zIH7n^SU0e_@4$OoJ1^dH;j6HMZ%u^%g#E|}G-qQ!6my0w!i~ZTUbcALf4lG+wrG8K zUW#6_m?u<^N}gVC+zhGe3=pyW_-girt9>Awv&Hje)%>TTVw6mP48M5Ec_+HQ&7(_nMk@PDxPzme5xz1iaM zd&9d;COZ`KoXF&4Je6Q{G8ie11ca!CO-ACzWdw!=rs*l`Ul<<949o)#?Vp9sriH9y`rU`DQyWy9J5)%U`YdQje>_lbcpI?>T(H2eaudOVK ziEn9bbd_%h8t=qf$H)dEo`H?E$&RJ;4^939p!&-J0AphlJ@q?0p762@VyzwhC;NxM z^g!PVHXcp&06if^0SE$mfdUff$tD^=(eU^0COe1YPY^4sbKQ@c%Mw88uZHT!ZzIf) z**D$A-V0t7T4(^EzOn5Y2t7S3OKU*EXPWH+D`^YB%5Pgt_Vo|;M_BwJ-xt8quemrN zJtONo__ehONs*lm3_f~ObBmKJqbq>;CKe{w8UTuQ^|Q@S)Iuw3tK;wQ@Hbv?Thb37 z=iO%dPZ7?ypY};C;fd{tsN;$3ueeR2-E*Aq?@J}t@2Sq;n3$=Y>fa>0uQ=S_uCj#c zn3T`o_J=&iA0L98#IiQ>UV!Cati4?W18gBJA}4Gm|H4H+%?mD!@7ep?y6m2JzmC&h zWz|5+_sHk64hOin)4)FbbX~f4=v`HytXW z<5&w(HjKwGv3Bi8K9!fn8%_H`SHZk%y{zyHJOVoH*@)FUOhJ!-T*C@3;Hi&naJZtb znOhBqBKn!f_OfsEyIIQ;B@D{LgY8i`d<{ZLNtecG#~|Pem23TgHfqHvbp5?cLMM+# z2I6OA8mZ(Mi8BuKcYbJu{VIrE2$88QWd_|$Jq}}*2YoAEK&~`Ts zM~tgBcC+k}Jc2g96uAUJJI*HM=TAwYQk)B3Qb@)%$``bpctOYNAvQ(&C^>U=CpKSJ zBh?MpZ`Un$lZk^G3yN+*1&TwnOIy7VTlAks2oiX(&%1)8$U|oIf6J;^SibBdJI@pF z%Xa*H_hPA09e?}BpiK(IYU78(@X2(x%-Q-nPa z(-<76^XgF<=!Q^I%!azu8fOe}6<^ri4Ikk6`!PI&eda)xiq#XVzLPOpqJFshnp$tQ zzO@4Bcy^kVZx+86XNRd1$(jw}9#$;18LObzbK~dGJ3IJ!po{}?9U4nJlU`D_^@aa) zMX)e6#Fd0j!=Q37v6vL ztIWbb3Qvd>y5pT!+jhDmcNqel5#@IE5l4H~++iZmRp)!D_dFTH)?IpgrAlT=7|-pZ zKYpG)v48IuhntJ%qnED)#rUThid~QZ#w#?42}21u*SNsyV4vF23v9g2UyOL>Ux3na zCjgyD2Yt_BC>1>Y>=CQ_b@DG`Vfa%PI>o#1VxiwiK)<2DiqK~YiT8auLEGxn&Xb_u zo=A2*TCt^^k1BhfLa z`dvHFC!>!OA&~6;M;&+N@2b4{Tz2%Yi=sI1fep-%^7N63hMqCT(gUrf6Wd zt_E$A*1P`eUlcd0kpk5}tdhns4<8DFo&A>~3a9!Sp~#YH+AK?}@3m~o6X`G@b2@Um zOneedLU!})r`PlsFcuLL(e2GKOS?$|jgfx_W+hd1_k8z_@VhS~1HH$IMR9np^nLkb<{1D-r|?I*n&5fct&e-Wpn-6s>zBMXMvR;jF3O zF3#Ol&=msR=lazWcdH+Zw!? zmz}5!nx(7|f2_pqO{Y%^1PA1)BJ5DUn$H7)kS|4ZN}~h>Ea90c9^oUs%a+!|+an5b z0~YNyGzI}A&qvs!$4NS~-%;M9q1(=< zGq}&}S&iE?9C{1mc?Y^H$;NHfh=yt`4p`qjW518>J}3z zTkG(XegrNI!xDgNvWk)Gq~a7WDC1p^HYh*Jw$V>gp9x;$Mb) zcv$0Gv~E@MOC#7Ir<^&SIn?xxWc*H6CewU$&@-mSM*ZwigkA{PNQ2O3#2|H7*#U>q z$)`plq!T_Hhs>Uar9EVoYq&R{oX!=;-WSMq`;`!@4VK2%Q1Jj3Qk;2nK041RH7xMgI{{Ij|PcP_~!lVa7Km?KrV_Lp_j1tg(89} zA42gB+P+EYt5En)Y0?#_igX5Ax$-A1<5>8X`_=+>sow%3rhmSi+hOiE_2I|LCaHwA zj+PAG(BMpeUZAyGwNL&NoWb{X-EAybj3%m=cam8PyA>@IUUT?dO64_f-GgMPz($*q zC3tBRCXSMM3h!ifs8fdd9p#_0UWqVT80N2AByXk~odUOT^Y=_0f!dNlM^~Nfz4jX& ze!}}@isX8SvT3J=!Jd@5m3554q{T#$xQ|5=hP9PCw{rdwwq^e;C27^ah{Lv$NV3OE zb|j6=Ez}QvrQTh76+tkXHrh8@T-u#+U#^l_=eLXeR4nf>iIL77!{kg=ebDZvSHABy ze!^1?zST7f-|Q1WXCG(@$G-Z~9|iV?aDv!G7LnfycgByJIqf`|PaCosdNFsx#&LFI z0|5YK;L3}@@Jbyj+qj#GmF|sbFvq4fNEWDIh!HPywrW=& zmv&NEsrkS3jow}kY~X8yyaRk)WsWl_#Tm{a?&x@+YGGZ4a6saI6);y=?CL-CtO~sX z4%V*JK-2HlHE5U%5RSrc&>Fo0`BeX;#kfD--T~bnLgCyJ8yM^23WYP^&*X2 z2iet^T+5OiHF?ugSv+zI%nW3ADANcj#Jl-ilEkT~x4wN6ZYqwUAMP>a4z z>yh*QqH{~S&iti#-12oRjGAehsMf%N`}wxOIK;XH-x1T9`hR?8roPqv|4gnROcXXD zeP5L<$!m}9C9;l5+#fu3zF0N=m}4yFIYVp$!^RvOX+@^r(w$7TJxDE3Z6+h=7dIy3 z&OonrV}3t$zcV4=dV0YqA02`=INgug!dV$@ztkfPddJ@K%7Gp}Em~=qvY7YL%Ya&` zyQ``)(F$o}%a#taDI^2@QveEXu*PCZ9HBk3YeWW3VMOzPLKE*7Of_&@XUI6@-K8Io z&rV?vX^Rkwvvdn`MLe~T(CM(c*>Zl*yLAQQq_vn*Da6x&|?c`UUrMcNfZQIm6h8IGz9&*<8kg@i?a~X{&8L$Nl6@G5f(tFA=P}_1BQfe2G zYl<^d<*lwp{q9-3{3Vz!1g7n+6sy*t!WYC?eYE?#ZTdYE-HkyaLC+jnj2mbXjdC-3 z>Xpa9*w=o-ATo6ERtzp#{>m2KdJ9u48xj%!tv4zV>Z)>U8azdkMZW$=t_tgBl6DdW zR(8U9S=DyQIzDxPZYv>sYksOdzbSLVA5bEJyTs8L{KYM+v0y(pCfaOzUWzA03XW5g%q1Ee^bb=uu6*wFavda4Uf6Ks267ahGnxeTIu z_C!XJAE@9HdB(9fy}t&3^8*~;9zdgd}lz|3f~`#V0fUr$3sH!`qNc*>@hO#0Zo zOWtGLhqt8%3^sVnCnp3mCgX+0sJl#EF&SuO&yXARk*D0^$&2h@UrR*B()N}um9+c% z%duD&GX89{PzK5Bc(i?2^~O)XbheKxBvALdZ<9&>y(10U=OT!=DIO0STcfs8DY?#?W~&Z+B!E3~V?K4uFS z+DGJUq!_}DM!W7gFJgshY-z-(s#A*smTrjBtGw6-c4cY_7hoS7GYjYrwZ}BZ?saHV!Q7e4=Vy;NfyIl>Ep)iDon2se( zuv^JFke>q;;JnTLjmm{JY#QYS!wBrbBRe_TTy85fw)KTf?{U^b)GjfAh|ZMEGW&@2 zE(0Yj1fWINHR_UYOTM#qj~$Qf^bef>H2y;asR{`8thYx4d&kKq<`rtaPOIk9?)UnD z|2wg1!`YvPKt6&vbe6mpQ-U2uu2&{atlxa`_!8OorFvNM>(~2JLnscx&Tsh^vnkXv9!PZSvT8~`Lp>IWg4-P<{yOhXzj)w7@$l_(& zZ%(JvZ{P(;MkS@}e3l^?&V}MuCUN<7Q~5H=TPLeYcoxm&X(9(a_t)A%_A0p9V( zxitc00mI`YQ|_nx3&bSE*pz90P(k1hBjcSWHhOE)rnj4c=!&Bt_Y%it!T3!c29tg0 zS=n;~q=Qe3>{O1Jvnn7>6p2?ttbG!(4!K2!RO2dgJQ8_L)a?Byc)g(8!!z}8JNzo# z$Egj^AaN~_D6|vr$<^3swCCWm@ucxagpK}35ZvB}_#L)(!v1WCEtWQ6E`aay>+ZKy zJQAL1Ot1(x=SHYK_`0xqvkJWj#!=A*g+qg=_QUHNW2dhq9^-NmHHfD(z$UI_IPbU8 z(PnhTt@pq*U$MMM>yM5Q-BpGSv9yEhp=Mwm5Py*o5Ypa5>^U%P^Dx+T4!k3F z3si(P7E4RQ)bWr(xFQ86O9A)G3sRM#Q zbEzU$Jf5)OU^2vc^Lgk2OTM~!6NM2PtcWnil1RQ`(8{Z z5LWj8DKsSZ^&5O^s^%OIED(_SGO|O*uk^1A(AvUj6kxvh(ki_Rt z>mbbBh_QCP4=Hw`%iHGEDeXm;IBq&s&rXMX#3?H6)vakP8s}ZJjY0LT1%ss5FLB2U zuQ}BaeDMLfswMZlb|s+mf&l_$D_bm4kH|tR)ed+T0M+93M`pSMDS?c3oXWymA=&P4 z+{^X^5Ahbga$NY>OZ_l$Zs9v^Fs$?IBYcNyG%roknO^svr?k)AIEtuo*U+9a!k#)rNq5@ z<>0|9MDsc;a}hCa7mWhnj8O0Gl^tg-M!I#{A!PrxguTQ zn%440dr8E(7>1^G$aNG_4_!|RyddmRa(&ubrG1CJJKsR45AU7-GmlkOn{sXbaZ~+K zgleEgF3`w3M|J?GBIb@AA%NjiO^Yim$du?qEzRK(h9Xk{uIebobeC(qo8H|U|A)!Lu ztkB};8n@zJ51pNs|?*b0}%L4HU(tKs>5$DVw(bgjfA*kULtfcM)KlqYq+ zzO?>q5VVK~7k__(qEw&W_W9&gJQmCo*l^%##X)Xu4xdJwQsIS*p1p1 zi>fgn&O(!uv(fA!vHSs6E3p(HGcW=+ukc6nz^Qcc(-PLeUUwOMAD~hpYmSr8!A48Ub)3EEN@{3zwiy>5&bS>6Tjt#Ott111I-v$n_s>6CU=iPQ= zpGexBvyWt>0{YS|cTZU5a)KNa(#@ikl@Y&HDl2SRzd4AhLyfi!=Us@VO1d`Sz~s?1 zIB6W1R?b;dH0owH`TYlM_F~~a=rNSx#lo8b z%jL(q_gxGMSyi>~Dn+Ej1AiYMl0O*zL99sqk>e`q@FkY1TmrTfLd4kQ6{lwYT$y}R z8Qh#SOvx_-9|;%!tGe1@>|yX%1l(bTF5Q2!ndm_>2+nhxM}sr}z-Et?MX)|?=K);J zlP^+YjGNduntbe0Rm<>r6SbtRfV11ZS~_NggGA{ojJoc;Q1se~bADVda7@5%oI!>s zaxNWS@uVX}G8iX-)qofqE+gV;8_IEIMzOd+!@!VFYr}Rdg!KAoJA zI95EN%&@9{mog;1Al@&W3-g^@%lQZ}??86dl;K+xnVP=&$VII(3TshTpf?(P#7coYMB}z^C)OBUYyNmWTLu)lCt2CuRmT?g z5=H(yG$b-r*=gX0zqpB_&ayXEIlV9SS(hBea$-Og3Mzcg$7ty__Cvq9Fg6l4BCSbx zgI#hp%P*geggdx7JGjJ4^`ZBv+W7YbjYH`W%*yJxlK|*Cr0WTj(~wRFPeIXoaM~u{ z4d&Gi$lH_3OV16#!FzYDpAW)Z_M}n(vuIS=$M4l9`?uRoo!R4$Xs&HZ4&Pjj?})CG_Rb`)7#CjB1<1+&^`D%oFO9!+IGvfbe})SIjf>?=Uxy- zUIg;_o3iIHxI<|6>j!h}!D|LLf@2dL#40!X?4sTNkSOc6Kvp~FoFtYO63tU%xE!Kj zx3Rr$tz>2g(1>g1_FQTYd|t>PlNVh272}^?|`~43=_9->I{6ODlS+#|3M`8Dsb!hIJqdTLoGll$dV+9NNpyK0dGKW@G>weF1<5ksK6E923VMbZfrmf2$P<_I0ZSLMzSC(wC!>lB?G(A=;crzYBw;=G@gbC$jeCx@N}Dzk_)6 za>;@H4k?v;?ka1mc*$Sg$J@$U4fVdF3!+$nX9CNwG7{znoNTFt1=PAq4;LauPL$mw z9`-^(o2nRr?p2&oL#xSoz3jQt4_o~$b$Ti7GOXyiHAM-PjU`MPsCsqV^w$gNxb>0hT%Ce*2a;P!HICao_v5aoAn1G zDp$WVG50IdwlK!zdI;a)&7>t_^z85CL4mnyUt-NwjF2-dlmNYK$Tk2IiL}`Sf>t}3 zv1TZw8k@q%mJts9TcnOpl7*lPc4i`?!_~x+@eSL^l_#adpYk!+^~*Q)xxI)!BE!(X zgvqVFE*5|hPSLRWXVy+G2y3p(4SUI|W= zc{v5o4uFQ%0>Be9iB8SfylW<)$5+Ffkkkw_R^)zH)RjhI>L~n_zzr-A~?SGgy(gm&xic zl;1P#4O^{zsi$mJM|g<__e%{Ed7L)XoG**{eO5IZ$Q-{a&Dd`S)blq?qrJjS6A8gT z2V(Qk3x16YRmTmcSrDXUqfd_lL!Q-UF9KRm&&vZ1G6!9Vh=YSG3`*vxXAWQh+2rA= z76nDjC2vDmgvig5N?WVeYtTwq;NG)!;~P-Zr8;4etp8YsnCkC2xopftsyjDOSPF!3 zxw5L`hQwDr->Y8s^a@PJ@h6R~AQG&u;e2k?w2IsL!b4*UzzZkDSp0Ox3f6p<+-$`e zPV6?PP%S?Z-3YieXZ}z)I(y5Y-o$UiGS|+p4ZjHC?ed*Plx_{U9Bwo`2U9~VudJWM zm5Tp$x)!|-;jjr_YBO?gC|%;l9+O=e^7y!uJp~%1$PNx%f*C%2BR$Q65kV`80iC1+ zkzv|S;_IzF<8n|$lKf#ILqS$J{z1rYpL;I1U+N{sgfJhcnpj$=_$+WO7PL!!ZorrD9 zMU-a^JUMIcuqs z32G37>#@z6Ty_KAEud^1@&l?Qaf1BrcDwQnLE^Tx0sDGh4Ug>3#dI`+XF5CJ-8Y`h z`x@Q3jKXt@+u!yYOrz=UwjRk^pEc?YZC10i(}>2T^u9yC3v6x2`$lu-jDe`_`;J(qWZ+;?yE zZ}KcXn7!J$Up%OdE!>;jXVx)W7N=wq46o~K-xHlaypsJe43;bP7spZWk<$8A=uV$q zwg$nu;x1kQPEy=tA73X`+Vt>9GQiknWSymmBJ10>l&DCkH0h(e68-JK*bbBuDLW0C zoeNvOm1#pxGH+;H3l+Ad+Cbc;81KfZ88`I8Ziv@jpE!DCL#xtb( zp&jI(0Kak`z?Ug7B&te_efHV|84**3W}u1{21GJZm^;|SJB@IRh5l{IhDkHW=t8sW zW_uew(FSkuT$UTNiUMX^NsEJH&0CBSH>&=S1u@umhofUR;k2Y42LB|iAA}6&D|j&D zu){3m=MJAz&BS=*p)@7;?xlSAn-o0fL>=^Lc8^`|DCWiU9`_Mo>TvUunnuoQ^Vt5J z`-k!i&p+?02n1SuR#Gz1=eE#f+vM*<+s;SyaWTf!G)y3pb^g<;#{fGU`8HmC1 z9Z&@yDxUM)&yvyhiQ~rAru|km z%Dh}insXCvT^fiK6*xfi2%wcSM7uA5rhG$)(9G)|@wZR1T?fK91x$IWhp5#nt9^mz zV#9J{f;aWLlxKXT~EI`lHJlc_Y&t zCsSB(Dor+cSkQ-USJ7katnEf|Mmb2)tHK`raYc17K=lk%98_cG4wxCvcV<$c(&!lY^KdaK$!pJz~0c;t(p&+r2(cNyIdzn#FZ!6v_rU zGn8mE{iLlVTl<~IlWZ-MQLrlp>kdw+?m1mne-#dZ-wG{MgJ92%F_G!D8vhi>k(Y<% z?(SG`q+jm*)zs~h1n@DFhaf|5j%ok_A7LQ24G2RfDGjV$B%uPjZ%MQ+-Z>vRVs9GG8*^22Z4s6YCGQb zb)>zeYcVa94d{$;X|>i+O2Pt@o?Xsw5kt3{Fk4Li3o=eONrQwdWw|;So!1CFyCzy6 z?N&=^P7}PQr73ruZLPV7?Q{-*jnq%x*V@nSQFhP>f2u`@WQ!;e<>ExW$H={SzXDI@ zSl|Zms!iICbq!0_-V;HySO6ImY`MEcd|2AZNDN}zlhA8&1k|p)1=KS_tW0h8!I;yF zzMvrxH!GCpg~vLgF{g2FnNQ8ZXfA?ECGWv;LS;ugrI)#Q%Bh-*`NFm{4J*ezLZu4& zRA6mwr)%qmKz?FJbjckQ6Wg%WpGh>D1uY&&Pa=ww{eD3Ensc=B09?*EbVAYBV689g z<-8gwihQ=hZEaigd@vD+mC6y{69{R(WM&~REBVdych?^+tuh`~Ea=A+k5Yr}WI%}*811%%1}~Fm5N3kW)V?Nc{q^}9XgXJ`m#Gq zhr<*v_k2!>4&qQ`ifLlgaw@f&pqs(!PUZKyb*RzkkRW0M49JDkm4Oh3F}fa;b`fF4 zFgrN}?!8i~hG9mZ(?4TZtb}g%!~u;q(tW~4M^5Th!rv2wTSaW_S=2&YdL<1T=fp8G zRIjib&@DWeR~hMaIo^K7!U2 z5_$-ze9k@1 z-c1xQit}r$=UKK@cArdZFojCppS4x`%;T^;+Lq`|5w(2VN7sfE0GXx2woyl(j7F7l zR4-bp)Z}X}Cc&E2hZ6m$33GH?jf~)M&1k`O*sN`}r<|=%6Bx{gU1780kzGIB2rWOs z4KV&V27{7ftMxF6)7~fa)1K7k(9Qieb_?lKr^T98t2&{TM;fC}_q0X;#6+WMX!=Nt z^DgF`=io`WT8_&tER)aCb}@|Nbk9H8>tipfpx=tns$k zvx0OGO{3n12Q_zFe+dq`x+$HG#)oM%kTDckrUA*UjeV}YC97UkuIDrq#6ug!3vX8z zrAYs-S+7MfSVrRwSWNg$U2od3$jM=8h34%L4;GWzQ^M)-k^~mj``3YNL;dUQpTUp) zHs7vj=tJq0$XSX17_ECCu`znkJw67P9yW>Kp2QTEfo%!s6<;!bW0vpo+($(585|oh zwL*n_C?T2VV*MWFpWz_Yh@$9V((YATG?i*BMX&J!sD~?eOEg5>n{T`_z|KlL||j zb{zCfT2-ReUk&>NKKrE$NNyyz1_AuIK2H1KL?{b!K|3M3@d``ogIn<5RUd>j%Ty&x zH2P+4tyejs2x@1hCJBJ@ZN$j`m49BW8jzsT2K|VLNMl_m3I22~%YI9Ug?n zl0Kx5(jI1-{Ls7Bn%J^zG;t7%CRO&p{a?M0nYM4PQT^@qpV8~&B68OeV+zutt>LtZ zm0982@Ck?w*aU;NFfMEM3A-ESqZ!h#&ajVT&K>}mAdd%MV`Q$5A|=bsviQ1w7ToKY zUa73vLR+5%a}EE+N-)?NZ*5E^f4BVGDYJlP_~6p=02CaV4ID9~9~>l~)vF~j$s&;a zq%f}2v)*ig3iBwqA`zR57u5sRDXh@5v%)k=g2g2eV3O=-<_X8dE!MUY;`71^fXEjH zgL0U6_CY6RO1HSZbJYcz?r|W0rBDybAFPZKFRF=2qh2erR<1g6fO)+s>Y#hqHdaDb zev34nt-p|$wB|*5dG+~}>G@;MUasG{s=0?Dq=*0={12zW+Y8KnuH8ERIZcrgzBP7R zBo)1hom;jyR`@&9P^PSBw9q87^7nr`381oh!WA#z08-%y4oVQKqpisB=$>=ueboZ! zM(W|qgSz!A{6x-#Lokql*PEbeG;TSOhIV+;jy_s0Y((s-yXl8Y#r8=M4LcT}TihRdE zE!t?O`5CeWKWxSo8}xucrXF5SNsMA9y4j*l3oUwj&5i4c$mFLQoC&fNg0*A0T3ey`Q4BPCV`Y)ha2nuc;+~#@DgX zLWtNgiqZ(3xK;{*e>b-`U7_p~gE7IX)A z*|u%lwr$(CZQHhO+qP|EFDvK2NL4OUl|03q)!k!!F>ZZ#kxxGST8u`C?^)Oq?h(0^Gwb;k@K_-draH}=9n^Rk(9=P^ zq$0(^25KLPSJj_znhd}AZ7qY~)(|7*5f1t-=$g$*`dSM88Xowm3At_D*M z9qVxTrRmdhep(EPz(ZueBwh# zfH!d8QIC`U+1W&YG3$j$`M{~EBV!i>_WMEIP@SU@o54guaK{$4-0xspXj?5zr7D26$X@+Sr`Ur=di6@eAwPj z?sImhnXbCN(n?`DFB5+&QWMt0UaRSqOR|(Fnb-paXL+jhQv7JEzBdU8MtN?)VxNV@1qevbrS^tyF;M_lEk(w=%eZfI1_K~W@UP{ZpXt}r!k=enut z)M_FVTB|Qyp^RdnuBqaDR@{Y)MB<&3|JtuIlI{7FxG;P8c&cVo0mQDpCT0`U7|0e1 zkJ!=;35?(xj8W%rRYxQgC$mvY@)fldPGmV?mnoMl1~GtZ4C!TCuxWogTi1iop}8t+ zRapp`WLXuZz`Nb^gH6toT3qp!9yr=W7nnt~)TP0gAZ8|H5pXWGo7 z`CQ~?DK(w|Nam#8xovk5e0vwOt2n&q;#*HzuyVmCyufgObD9WI&pPEDAqFP1kJRCN z^7WoX9^-a;ug=&-v>`+=ap&FBLyzG(2MYKmf>vhfl?N>+=X(F-2tc3q6*~B4N(@4N zD)1a57)ytf@_UcHQW;-*qJvVLhJlA&QgR}?Qp3Hv>L9vFS>i^u^#*9kbp2BjoNG6e zA&>fUlbT)35A5!Th7AnHMCBUmyrj4hbWqD3J6~)_sL;ad8n#qVnSeNH-tjLSm%Dz) za5mWJvnfQUJHrqOG?3_Seq>S(Nf}4lh@fPmzv4r}lj^n9O+oBw5ngiax+$vC8F2DEGp)?pu`lb6tI;Oc<$`hqkU#A6wt|g&k}6ys|`tnd{H7Ls20}}kC1^V zWJF;9MB;E@MQ#c7y^kO(D2StR$!Xhh=Oo) z42k3T7=p=X!8-xFNh{JRZ;em2hxRK^0f98)a&*xeuG`+jrQbrEPnIA$c}(98u%MxVT5qiP*>;Az0a zOB-*tQreUa=mV>hKp$exx&t!bP9ij@pl7SbxX)|R+bN4uV#l(t=GoX7)}h&=*X5?J zUGfOe3Z{tkOX5rTrHZL`^c;;RLe`q=N~9I+JI9?|K{LQ%s28x4Q=ysTG^!Wa^h|G0 zFfvZ!PuV{0mfbhL>z@F3sRto{mWoXzspJLAEX@ccAcEeMVK44H67;`|{ir+U+7+`4 zD@qiEpW}q&Tk1Cf!hW_@&-Uht94VoN8!u;KO`QHpu-8U1E3gK9@cYes0T?-Zee(95j1NlpFxL{() z(v59M?DWhalA|Tj@6r%~2I)%fS(|v;EeU|beaMqByok0ihyxpzU3Ed)CH;3!5dAER zL!W3RkT<}hmB?3XAfyKnz0nq}R%D9uGrPL4kBK`gs<=M3RCscALjRs1>|tXA9ioW` zm@`WLT}F<%tj?{xID5L(+=%!uS7n6l%xZrBWc?iU-F%Q`m4Ol=E|_uBuv1YyqzuNr z=#)#b&Jf|rk7`!X73MikU!^w6MC_a@*9_blGlKEmja*ZYy-4)`Fd{JYm3#&-;i>^qn_jCj3BTg518Gst!lZmNn!j+o7 z!_&A)K`C%p~#S#=q0;KPJ0S^H_RR>k0DpJ%O zE;(3jZy9OOl$esTiwX5Q&1Pv38@Jg=M(5f@@9)~pY^$m=^Gw?KY)~qgch!Z9gPv-M zPdfbV2^~|vTYySa(HqyZl%>?4x=Hlj#4IB>dI=XHCBD^His9z+|CafrG95`V?$+>}9xF2N z)ew8LPAOvFpK-5(r{>Tr7oZD)7Jo(sx)hex6|Rb)R2bWSGb`d#<%&8@&UX+C3~UxJe-e>_EB}M{sM225U0T)+MYEO%6-?4 z6R4F&E^Cg zq`2U$oIIEPFNM6GIjqGNs9<_ckq4p(=I>aWm&C98U>W8&-ij0F2Dq&s66*MoDPGn` zn~E4onQS3H?xBTnaS6RFiL>M&aG(@UoLM7-g5ZhItk4avYmpeSxn+B$JDN}mk?YI- zeXN*Ds}tI>1hDwr5kYmrY!7VXMzv z=~P`p&>|4&6StN7?P5zK?bicL3ExYR?=*zV`x*zIRL-QR%J`C~qwBy1=+WgJ$>R;N z#WHn3*)NN+Qw`>84% z*T?S|+a~jTn>IQgzN?>}-&2D|+v>u?$s$I8nyXkV`PaU3EQ7(HcK=aYtlWm-e5c=2DQbpW3^*F(3 zXpX1Sso5~bFOH(+8-%Y{TlweZ#*%q#B8+;s+YBiMDmMX=Ey8==@JDxQmv5D&03| z=#r741$b~M@9>r{b2+EeGL_vQgX%&t6F^bf6E3yg=(QS!6Dar@;$%8vlT)P0!~El( zxg$9`K4>rE0}4%#h@0q+P;<}VAHn|Q)=!4~>i*hvi(p1~*D0Ork^O7eRs^>{{$_{_ zgrn?TsGVTP#AXeTGYnY~PxD=m0_ z$lkQe?erUPS(gevHIpqfL|^+qgxZfoN9^&3{#z7MXEqS?hQNrp5xVEqjl?&=O`{yP z^SJ_fCp2x@JhV)^l+WAILNo6P}ytY)M-TZ(2?MP>b@cMv8`i+yE%hKKFY@ z)#5ytP$|f_6MzzeP5HhLXL`jJl-7{x*tF-c+z1p#+?=p%Fk91Dx(GnGqG!xsnFrTob0f|=Qp6w+HGI_%{3i6pW zLb?MuA{LT!x+y?)Nrc&5@-A7T@JjTgnNAZHk1D=;e3dFDX0ZH4NqW^%l$O>jy@NPl z0Orp7V>LQjzQI#0xZ{CVY^3{qsFkK+4|aW(uJfj1!cWM0{2I<08WmWNev}i1h)K{^ z>9n1mI{u5_d6zX8gtw}t!CfbTcyZgKWJo=nznHU=h>)ZKIXbiu>A~}m4Mcx@kCjYd z`{la`hYPQC`;KoD8r+V~_O6xImgq)eS}Z4A{HNoPGw+xcXRyviE9$ppY;d`4fXo;> zJEAxOC>hrH!cr1lBIj&8(PwpJYRniD#NjXgj~W8L?euu zkEaC6t?GK|7!1HRwqbM*?&w|_F7X@ceV1Xb7W=Z&*QI5_D!Oh@O%SqDkOWP$hD@}q zq17xw9q0YE`K{CpZ6SI`6k0l)&gN7E_QK3LRLmoZVDM&V6Gi4bH+P4_;v2Avs)`wl4vM$aoW?00BUb1GyU3#6rpZI9&1Mjz?W!Sw%xI z(gM4~I9|3<9vdWnD8a&|DcMj9Z617rvX~e>Pn%LOKeVTaA3}ygDJ!ZTUf+(tjGfxy z6vMh?Jk2JQe{dYf|E&xoQcmX=SKK`Sy($2VZl00n<}s$)OOQ@z|Jd`tj9m;W^tn45ymaTiw7rpE(f5C}+x0j5MXsqh zrBNz8cS~-g_KNQ^JV}yYErDW@RS~Gv3BiJW;&}_f!6eOwvo3)+|3G6_N6d0E>4!kK zr+Wd)AmW=k0AL>VN1~51PIQsXI0K)EfQxnsD>K=op#JB3N{jqDdu0X;CFR6o z*BprF$0B0050m=3Gc5}cDA|5BHK6ZMCmEdYwIJbEFp$7^BQ@?=ru`mn#sRDHzIjEe2w7Cmz6 zQbim!fpYXN{Fhj}qK-|L$SdSI-T2D{Fd60fsweZ?Qyk@;$Q_l1z=nXxq%j{g#iNss zu_F&K2)uxPI6w5E&}U(I03qF6}R1z797qBxl0icy>|MOBiLR@LyG!cG6D2}5$+nRbn`9Q1DB~-26T0gQ7q5)eTTKbUuT0Xak zAR)I?5Xn^7aT#C9nD^nNAzbT#sjY(|-Ac!<_Aa&1C$t>`r5ZBbF3v4df8Mz@-~!!E zu{#1kaggs^{q3n-7&~pE&NA2n;btguvFHOO$q4;YFYCE=i4|DUX@)Gfn1)w73?n zV`+gZS3aEio`WT=%L~JaUYinv@z*&Rn_jtj5VcX>^-|msll?HKw8hC!CyT^2NG%Oj zw{7hH=L5~Dg!ZjX7ZzvIHjAjMkdI`?(@(o;P0U^|2vX9nE-RKMbUhOO8o&9$qgw6O za3KC3t`U2H7X07PDOwekI(|g(h{w(svH`=u1v&l&iB7BNMgK9f#0n_fd*e=LX7+K+ z3K;F|q^Sciu0zEaa<@je3$O|FN&Od}(DQ_U;}|0SGUjj-ZbAk*ZjzlPlFrx!vB)I z)V&*eqy+YpZm_jalx|uf`6_q>t;nxMEVj6KmtulVsKiUlYWN1pv6n}+;*143HEu1V zdtDUnKK_cBJydc+^F9)s(_Dm=vhXJC_z$pWE0_=B5?X(&HbCW@Og_`SwA52%xjF8n z7nkP!RkB&&$BQ2jU*m_y=4UXxlfF55V7)TF*FWdf%&UgKnMph%#M8p{>4b^mV4HX! zde1@2NC}!9@l78CilMCI+kAsx+M%{0G*~Vdz3jSGNps38ltwO(YE_c$88fu49hu#c z9QhF1pdX##c!7~*#SiB7_f4X2W6gA7!olVkKW@f`tnB&a1E5om@kydQK7pfU9H~P$ zwwz$oIB^ovQkvdKqW1;+C~)subb2yNkUQ+49yd(3Ht{)dF64twe8UL8H~%xI+}Q0i zL7~+h7~%N7w@SlLk600|lTydcj~4=8Cpr*t+FECKp>ZTjeKqRL?<=Y*%=+i~0KlhU zHXnyt9lGIr#3o5VSDiGybx+}BgMOXpB z^u3YCfJub9m(5>~WV#qnLCd$GQ69se*mM?@6({`kZ|=^Z<@@gCM8#lD1-G#2vb9D z?9%9czR2$7_&2-N=8gmn9oINO^6GA?$`OAge>!{w<45PkR>t06?Cl443H7RnFUhZ4 z0U9u|02nwsd*auNPM*`Kcee`Y6jV;;??gbbOCf~ZFDx8n{P%ZE$Sy1%T&_=Fx@pc7 zBGi0+?|G>D>OlQc=lNpE+DkKQ-v@YdHX!I_&moygGf#jo=^)o68VQb7B1UG(cdO3+H8@(nv(!++G( zhAgA54na7ieqLEPd$%evB3Wie9}O>1g~`sUat(0cuP?$2&^by|cTc=_Es~%S#9S*u z3FPS|b8-P_xKYqQP0X$sblH-583WH4BFe@6>VCiBoag>`B%CBhIC)A&{ASr+K>!1%HQvR6TM6l8X08>RsyM5wlJ;ZZJx+?$Ho3G^rr3# zC-y@DRl5FT3!Q|W&B(p&kBjn&F~rlLW}F2?!rDf}aX3W&RiPYl$nUCuTU zReA-`?7^0TBfZ|y<9C4I?^G5c(=|T9!J1-3sDHa2txpccNxU@bn)g+bE)FYbCGQ?E zwM*K^&Wb#O}SRMP~eT) zFCHI7y@vIYhI6LnLG_yl$PqL&l{AQtZjF@7{3*soy@%D(?yv%-`HyswTnN>Ap~?(n zi?hGqELpo)yVEZfi@b3X{oxUQ`wpmF-ZhmsFsY>Pj9E|yIip#NXzAdgq4Y+OC#ORsg-_40DkO6qD=qsh9B0DJCR-Kcv32j_WftimF9DWn zU>FM;a9e%N5MPkTQ9eK#DHPkze!;%Ldjj0ioLab7%@s@5UYoIH&yu(`sVA|+0+$f( zD*LUT*4UYVLF5R!)J|&wX}Mi~EtkhuYi|Gu;F)L8Qxz*-Z)Kk!pK+aFK;)vR#*UFk>G)MH z#5N&i!URzT1qIwG+!mj$WJWgDKB!bV3NRJG>ta9ka6KSw2|TiOHlA~9|IAwd$Vhq4 z=FhgXQp$&Yo^5A`x#xQSATZ)ozEK-c@=xiSxcX zMhdjl>ZS9=aa&%(M*rqR9*$(uvqTEly-owTw3sYG9^jxOdAQ|}G%kZTO;+^XZPrRT zsEGstSRVy{FgWOUX+%83EN)BKm{F}OoA0pAn|?UyAkQK7)Zn7n4t1#V=@6& zX{%R2ivb^gHCI;2)DlT%mK;0jE$qI%2XipMm-1TEB?-`_nmBPL{_E+>Y!P90rJ^;L z?|%!`C4vzb;%}&Xtl+;~+UZZ_^q{8I2-r|NcZ%P~`Ik+^c5a@o8E0^Igbd)|SOURi zxL!=l<;||+{vA<#ak!Qc^!%o^TY>8pqaVv!n;z5ZOZO>v4ee_R_S=ZEFWCII4V- z^h3Y}Z#t2&a30LC{^nM-fx?8ar z@>AKbPxmo>{cl!=CA;q5KT^~v#GP09>Odf; z`o!iLWn)q8X-RAmt5-L?`l_>}QmIUAEioNSg)|R?`o2cG)~oGwRWhd!?zLk4wAS*T zIb^a;9H&NDI~4(89-4$>h7`IZ*yRGV4f%YoVh5aRn)vX|l#9T7{vk{+XapWfeMvv& zT=-7CQI4QVxf6hbr})(&SSbl3Ooi9ZFQ6ecm)6on52NYo1TN%A1CZJ69TR1JW8WVR zA$ds4B>*2)=?-;4?|SRhbTH|CKq{jgCT1sS3s>fLM<2KCwB3uhRLkdHu!<)w^#tB8 zrppX>8EyM!tYfh)3{QzNr_%W-epdJiDGx41Cc#55+o-1*GbL0rCy=pH5=P%=m{=;J zmkiELK1|MrthZJ_d(T zdH>UK-6B9+CyUjLJ!D1pl2xuY^KZl45mI9vzXe|&*<-J!+>!W0nNEV|X4zJ4AV!Nm zC{I=Rxr}AsX^kR(R8FW(>iblJ$?Jd)db9o<1LD8AkNl~fZZJ_GfVHafSi5fRt zROtl_$CPi{;bhyR^(3&iQ^FKHgugp)B3AnxX9~o`6UN zzp|Nd{WtUDMX+xr&B)!dEiBz%y3Q7I#e&^K2(sZ0j8lZwV#jdfuYW#89X{{OOECT7>qySlUmC?(l4%3l<7CN9O3nr@9NDQG&iATQ zS&|XWXHw&3n|wfF>F-%m4!zBxXa@s?*UgsF04}a&v}mBoKplq}n}o4fxEE*>cw*SB z?|}+8)3GIBr#Bi&O7sZDs2B_`O?#jijwBYYQ!!8MzB-F9I6QpqAPw8-b2tGGVm_?~ ziq4@RtvH9n#ED+IAUjrbM2<1xk3pe6xGO^Dr;USoxFw>|Pdcq;6zzo&kP;J2mdMMG#HIc1(hv^t zHlAO-EfkPOeLBjHP}hLWK4u%b7v$t;eIPNHXde4rkgNIK6C}u876_y^YtEkkf!Vs} zVN{4&s4=&ij?7c^D(J-0rrR4R$DBiZ7htMFeOCvl>k7a-NojhR$GfaZ?W$7Y1(Yz*{z_>*iHnG?YmKH;MB97`}1I|IQbqlSUKvRtc`ZF z10?RkRO7h#xXcM>_g`ggr{{bAXiCl`C4qpIvrA|`Tz{N;kfct5Phux&sjCIohurX{ zCo>F-2Q8u>D6`rw~#i5 z8|g+VTeH{Rs@~Kqo=FaHPj2Lg%@+xkhF)-CCZ@mLe0V%<>5MoM>HkrvbcYK&%27PJ zO+s-_h^r7|_Ee@{jKpd9Zpy28?S@_JAAE^OhNOM@t(*|Q0~|v&UTq4DN2qHR3x$Fo zq7YlVIHDt8_v4^7h}!Oo9lB-S+GE)}ssTPTEf9CizJ~VVZ^hiF!dSozI@*KGMU(+pQx) z%~9?wyH{S|@rb};Fj}yO&n4UKYa)mj^cuVVwD={U?FM$Iw8C^oaY5Xjxr?K06|67} zc}0j9B%Pn1=TPKE60Lj*KeUgIf8l|Y(V6iNSt*BouAtLv;`cn2Q7ABdRh)O2jP@+Q z*YC#zmWpHo%c*okXzwg8m%1lg5?jaW)+4@|h#S;N`xu&f%M`V|K4Nmqq`lSR#_F2Eql>Rm+dJd9SvH-0fbF zZUZTKt#(snhJY%-f*olmn(gANjNpmPw=%TI8H3111l}srL9mwhUw3&Z&&VGQ=`R*( zY+Im>0T)0!xMupN?X3VbIS+e{cLli2lETI$BB)`QcWY^)GdAd6(?-3kI=bu*QNaq zY1w%G5WidTmw+}j9@TOyaSC3fft$38f4j;lc@~P28>;#uAZ){dmBobCuq6jRj+?Ra zRB~|x#Zp_?V6bsmaI=?`&RtEa9WS(m@kMJ5u_ZOx-+ge*Z)z?(S9)bKgS10Ivcaaq zS!gI%`N2RFW;~F9azNIw-HziX9HN5!)u?OwZMRzZ32L%v4AsziJtE8@Qn4(fq+6vcvdR%E6KPy(Sb-47oq4~}o95(3UQ+Gc6&mW*2}oCtrl@}tkPdxLIIZT3%Y*W43BK@a?CyDGK_mH z>wkXXp0cQuD9lCySMiqGb{IWssIPNwg6iqBG8$@?Nv^>p?Pep=!12udiETTEAUYx#_^eT6K}X_^A5YFu|A68_3hFbOk9L zxlYO;uUGrL<*{Qj%Di^kk_nJ;42R0r>7mj5m1a#0$LFu$=wes0SMt{A4IE8Wex2Ls z9`N6HLrl9g?+Mc!k=Ux!m6u+nx2XQOqzp?R6M{G+^rZk*Jy^|l7t509COq!H1{p0R z1~L-E3jc*d>alA#Y4B>0lu%*0CmEsdaYFV$!*5WXZ=R=WgeThH@W@IJr{5~1hY%p~ zKcyXSYyyBo>75f?DqYE_cKcT1B-zt=(vD| zFRlydV2FmJZf{J0LH;<-exc7=OXoMz6ZSAx05>xAflx$-94#~6gZddS@*%3-f|q}( zoLmN><{0_}I((^(MomS7xyN7}5c)(dPoHc#nGmb+Y98wdR_EuCE+_#eM!!3YwUx^! zFSrMA1vqk(NJIluNFkxW2c2!;OHELw+@f*-$uf5IJmh@3_BA;$)+{mA2uX$+4IF=1 z7gkQ`@tF8N!z0I{QCc41b%4S-jLU|~A?MX$mXa0#gNB0sF#b^a4~FDD7N} zO{)e7urW~P*Ohhnw!Erb+~z%M5H zVLW!AKU@US9kSi+I3S*|`vhwvXlvd^8(`o`i6_1>@4Y7=BG+>aPKHyBX-0I{Lrb&b z3QI>{Mgl%T%}7c@lnOfk$PyYrd@(;o>G2)SHbt+yDW!qXvd2rcB)Bil?|6)1o? z1(Zvm{8F0D&nD=&52`syVkgc^-x#bc!qMmCiWFM&iut8u%_~wVhuX`hSeq=>#~I0* zQXvNgo#vAx5eik+Wb@t=gM1vpY#i{tlo9NJ&S=1Y)+^zQ>aEHxta@WnhLL@4#$j)c z@{hs*fk&Wg@r2-{i0bufRPtizXqSMigiPPY5w$k?euV=~nfi1ANxlX06?Xtz(u$%(%v)%0taFTy$aX>| zP=y&N7MuBxtS{|%5~sBFLqR2GkPwUn!A9qZN>A+skNuPTyb#(bWz`PVE!^XD&KNB8 z;?V>rS&H*N5h>}#_`W*>aq(=?blY7g3xA7Y`eA~VWZ%K4*(poi{D(h$vybTbHp!h zC5a%hEzFEm;#&80FtGJA#K7f`@l z!GBMFG{#D$JOWaTom6H5OcffG7xpQQ9uuMsIH$M8w&Edx+mihbTb2bZ0&|R9*fJz@ z&v6B>2}xu&|BhBHNl>%JBYVlU-AKf(0$lVm53(>Mk`V5VKkh{~)0N{*U1m}3g%OfIPu8JZ}?yEUJrmPP^)nLSJCn?&AD2k~J9 zxTlB?r!H}n4sXen@$x?ff^b_@O2Zy3A!}K5QGcYnUlr_fA%nUsfayS@4(M&zVaaI?oG>t-fh#Qnn6`zZBetm*+91-HJIW-+goja2^i=GQ>lQDrdceRRB~G&9wY|@Vy`}q~!(4ryyWofl@=|y`YmiCrlzQ*#s#t8RecQ#F!2q=( z%po}Vfm|MvOS?P?2h%&9_T{q4(qhKYoa@-|Fup_vbwd83fB$4dfe zo_ffRgU{%8@&*MNnO{B9UlV$CkCLW@+zfb2qH3f&x3; z?OQ@V3?xw1#>Huj4J!&i(IGc#OPjp^Hxh~I|3xA(GBR;6|8EEB|0IzZ89CS)|3672 z+yBO9rn_yfHpRy$A-LO;<88Z-l1@hdqas~xxQ(lgyOj_3-@EVk*6Xe6cU!INILB#P z&oqwJ^-@TZ-q_@>^4tW{z|h3Z^Z=+rqI#mKnVA7F6JrCBmR1liOkkRrQ=6T@oPb&Y z$pTghcmqIFAJB(@1|ktyncp8mFts~?DP%8vCxcjMU}px|2*d%va$9q&Cq#k zPOMIbE-s1&JlZEyAwW2^0RYVitO*2E6eYE_1d#-2DGA~dK(>H#VQB)V=)}|tk^yig zKrXC7N&qu7JA-Kac>^*uHnW2ML?ABXmlcC^{INH&ecGFR?_cz*iof}b$3W1~04OI!G7n%_ zz%_{)e5TtiwVAmA%=wqg>fZb*e@ahW((?nn=)+n9lnFHVTcBH8gA!vC1Zb_yjqQzM zUO*(YGlOz%0jb>B#Mt}=Gc^HfdVl*n{skwvv3*ttH`IRUpJx2!0&z+P>;#rc5uEoc zWn*gorC#LskUi|56T!29c5ZxkS_k+~ z9kr3Ul@Vw=2UI=l)5hY+{$;`bntiA9X8gT; zlYjolAt54iyVHZA*}(w-V^cE+g#O$|hv%-I|ED22I=X;tas3g$?`!@x|Bg!t6mTzq z7}{>mh0t#>)v~q>srOjaA!XUwjVa{XWrF&ccAaaXkhf zz8Q&s)`iYcz(u&QaH7g${rY15b%yjT?X~ySCjy0 z`DMh}LA{BX#pB1=EE;$m#B@sBisoX|7%yqv(dEV8N zgd!{2Q9oENgvp!_Ie?kX@lz#+506)ubs;yvN|x!D^BRBb7e(M2vBs|A*~3&(uv30$ z$Ix*U<{*n&r|S?KjW@G7z+98-m|20w^V#1j%Hj)qA_)l7vEQna^wb$Qog1`yb*oC1p8dv%3)$vur-*+Wg(w6Zo!ne}$RO}OTe121`E0r(=D#810`iNMYRutQlpWM8hq@V!G!%BGC^{0*W*@Sa5E zeU3w|h2{7u&M!h64oTX;oIz!^1_@4+0URm%&3d%3l&UUGievYvBcf} z_A1?TRpT7l_)jFJ zDI1fp?xAwK)2&PEYB)TouI^g$#DDNoX~4G8zJA?xBqgHb+*QeNG!^^a%l!0NDCcvl z8l{c6B^^^i$7`R=5>!rMZ!~ie%+Q&0Pggrr2Kzru(7NWD3KzZziI|^P(w26LLz)>h z_x7JOAJ=a!!RObbU0vVyU?%7yS8E`ugeQ7z@F z-!++~>&%g#8Sgqq&x|k5JxvF&j%HRg?UPkm-E85}Fcw0a9ZGrqWuyw&Rp;fq*TA6$ zGln((TWyFrd_Y-4B9S8P@YUKv8-TLL#-8)RmpgkNW$BJj!_PIz%bWJ{mMG0E^=K{V z(`deW@;>y32 zVaQdPxJ?2{vh1i9T6JZTheg-V6|I#V{ez4{t;W^%a}BZioF}l7Csugu)kS-fjpk`r zM-u9BigV}K8bXQ*eshZZ<+ zK>$6bV!mxr6jA<)GdENtoGQP2mS%cQx%yq*tY2*B(AnJgwWyoqM}E=EcE|1v=*>7l z7*gHTL>J(Z&V)0SoN!Oo5M;Lo5vcn+q2q2)A7ou-6(m<~$S``hYWZaWk?zWA-yqc5 z+dj(xhApRLv!kW5y?X;}lO@c(j8rYB3BlftJN_`PA(3ZNPD~9-Ci})^><^DSl$n`R zs<4oogncFR$VFQ6k`~n25&4**nVX?wCR?S#{9L}Yh2MM|5*anL-==RoZwmX6%u=Zg z)HheC{n2;H;_b$*^6veRBSZW?HXCB7pQ46>MOAsDz%AMpdHfdzRpq94Q-+Ou#60E& zr0l{QqJ1E#JI@X{2rZ`aPRwtYs2uJbyQbwp6u|!#;nWlbT%W$A8nk~)QD+bS+t!ce zCQ+lW9Oe{z(cX{vUxb}wk1$%0gvYjR+qP}nwr$(CZQHhO+n#y%W* z+8tfNT2$rbG*osvx0;&U?%jNbURyOz%rLsC01X;r*NtELs<`Jdq#j(9^z9kA$TI&JKPvOc`CVLpl{3X`AMoMNy z?EC4AmYTv>z;xl-U%BTV=1A$s=;k?AvFlU&#-}C6P6Pt6B_pb~Z=jt9e9Bu5&G=9p^-gt)qa0GRn3K0U(HfL4f6MyIg9ZFb;f&`YG!NZ|hT&ou?V#pe z{rbx3tA9W9Q1R`dwY?qJ0tnFy+FFy+nZJ>#A3ijs>7l}n9ysiH4am(_!v3c(6#1fE zo~ni?P(CaLO3dOEcg}AX7G3+y6a1gh+gy+@ASjD#pEP;BN4P#ou-Ps6G3H3gIZ-^g zbFe3#yotb-RU>Z8Eden@hhVl8Yq>+qqXx67&^Z((d^vYjU>pH|wXNCP^kTKDrApnq zW4-dt8RNSv0Lq*jFH|-w#uba`VmYmOo9F zc}EjrQni8}TG?tZbZ&*-TfZ|}#JyzH1j2wBdGoe`TQqC~p^5}@M`@6o-Q?RBK3s6A zRh$(37L3pC0?B&xK`7ALL!yac^tunQK;x8HA2TNT&Ra93l=KyEXdm!vIzx-@abD_q zpP$5=nOVeJ{$+rK64_ga5e0H-fUK8dnb)oj-xp&na6vQ*9hL^b3oJc%8g$O>DMVIF z5ksHxzAC^lgu-)mYTZ@J7b6Hyvh&d_-MKHB6f5MHU5dTBlQox^+6UM`Kc-RGJ!2OYjN{Hsw$tS-~y5i}<=e@g!uF z&oNH7k&xQUN}J1Z$mI65zr8f8&vw0rhZK(!q{MDWGFUI+-8^Yg04@CJ`{C?p5o<>p zJlTbAUXrpLesjY9ENoMFWD2)MnI(m8^!i1+|hT*-iOX@1P|zI&1W4lwcB>Atlt zR%P+GC3>`RmP0T7D!%F)%PBt2L>) zmdG#ap$9HhI2qjh`^HJioRD@9e#YKrN9b-thZ=eb*1|f+ltGMETx3Tj8pj_pb(=@0 zb`!nQ6#D#ZPdlaI6IEWxFLzE+_tIU?Fq;Gchl%?wNeP1zi$}ILakNzbLmErW97b zsXnLajfNexGC@Yq8@FmPUX8@C^MQtg^8~T7sQ>ksQg*!m!)Ujtu`G!x{l1SYpVW(a ziwQ@h!Wdw@xseEhO4FAtg#w3G?+)7E;q3?N${L2ce|c59m3MZ~x<1t*1QH}omLkx0 zJJULyb*6=bbBt*;_FrbWu)o!0<>-(ZRn1s_VL6F$;Rs1y$Zk^ICxClFiL7wDz1aF2avS*X#AH8sSgE-5= zJ<#yXt_(I-DC0|%Ay@F@5CFT^7 z;23Wlc#>T-SH2d#&jdQDXP24dwS|VTD?nQDiL<$fBtxq_MO@5j(GGjYc8Tq?YyjvH z=MTS{2aTQWAtu+aLAZaulH&RXrThTEsUzOD_!yY^BLWtS9-`A1bH69OT=S7Tl$pX} zr6Gk}2BY9nU|S$%pzBP79r$U6&HN#?8`wT(5KEzmCMhCYi5HDMTC&{nmd+kIR1A zHKP^i#KM(x;}(@MdeNWMFvQd`QR(NKg_fAkn{SMoLOY~*XX`|Jb|iu5r@@MpnOeZp zozt#AZh8!bT7Cfk34tQPXHl}sjQ4r^b?N>SQ>_%ep`%8XA^7gwEmSNQp#o z$j3m&A?R!rAd#Kjrc~bNX-<%NmCR>~$Of|iMkCsEGamiym!(jDg&2pfE6fUwNIO^M z=Z$je<=2O56@;zDG1fE}5{MpP#1N6ldeKGi*q;YUQpJFm^r^cO5G=Gx6~bS+1K=)= zq{h9WGF@gX8VTKX^m6APcUfSd2jG|x0$$OS6C&bkK`DQGXsz60i^CUG-km78H{bC=B&@!EgDY^Rz`)gf&m8BxA)wsNb)~wH6XkJy_42Tx(pK;^ zm{tk|Jic&Vk7b1N3>s6`+~cY~X;iy`75Ix@ab`u|-X=m>xv)+OT4Pljw;I!rFZLgx zqjw1f>`e?|KrSU&sxW*jY_`+eg*?&h6rY)T*=1&Sp*wjJgFVYTrK2>3Q6=ZEew$&5 zIltx6?XpEEELHTGj$=LE_%oV}A&V+jqS*=Z0*WDu$-My-c4^&8$3ex@EG} zO1K2#q4XwblOMlWV9J*iP1_Y}Cq4~cafSBN(XO7SyJcs8X{smm(N(6=Kga){gWgk@ z1Mgpz`WmANY7qP8#=;Up)GMS6-jO#+~3`9k%S}VBE zca?5>+6>zdqE;X5UvYc^^SJXSsEzs^%(`M@IJb{(bgRhN70Kxy&esa>w3+pKv4!Wv ziNli6=#gn0Lmk{59sq!VJiG|9E`=1ch08J9w!e;9jR?o>IS*dyr;v5{X47>zR4H5} zU;gFuIBViN^h;K|VA~U4SX4jesNq%ILSGLO?d{Sb{2#8gDp-#+%rwz5c=I%LQh^il zZrjXqR0-S;v5{CP{WwHr9=eZLy88P0@Zl03knFNmYk=|klE4xf_BOx*9M32aL0)>p z4IkHy2M)T_a-@8OcfU&HdK-9@gz&=}Vpsh=^Hu_JgJ` z8v)JtCmEwwb)Z^(NdK#|229IcvSe-gwwXf`{{8ug+=UWjgg>L&)}78yjnH;onHvG4 z?+u^R+vF=K}xp5xjAl z18oeO6o&^ZXPwj5V=tCpYr`(pFfT)2T3uQhu$^f0fvM5Wc&kiiJ;*8c;En*5 z26iyPa;D5U3f&>uLf10le`?#dGT>|QwoMpR$o5#@nFGJgQFPeXv#@jZ>&Ode&ZnnN zSt(B;8zrS$={B*mWiV*viRU**RgsgX+uS}O)Tk=8q>*4&Xp{o`gVSXzu9T1yVa69> zQFyP@%`z%ItIO$sJ#vk#PZP)7jDY@qGO)fqNQv);G{z3PZBi|{0Me8yMb{0g1k+?5 z&%?ZkJ|+xbSm^YU%*df(H$zvIC1&g}Esz0BBMCK>i#EUXH+GOOUo=bC(<>*LxX64M8C?%xkJa zk=i=<6r5l)ZSk1+9Y3*xEC}%(<7S_d*kK_a4YS|)c20622~%%*3EC}Q>Y#wp=HmDl zLrJ?pKCC8eu+f0XsEz&0`txL0VKt1I?sYy-4*Kt4lG}R)puT+_QoMn<=f;H=^;L&; zVB?a zv6!#KiiWQ~n_0-&bNCsx}htF6bt$!ijJUSnOthaR)6t=la)v@+~R`!GBk>A%(8TCX2Z zPbiaMHO5L5>y)ZWsJ`aIE|g{prkv>3_ZOlx7`1flr`4*he1fRpEQmXgDIEz0O}Z87 zybe#D-4xmf#u*ki$4@#EfECEWh}H&SvZE07{8sGA}irh12PLMsCw zIoDT4j>TA}t8kuze?6mTC^zSzyL0`hZ+Sb}4RG@G#_?S|v-e}gdcwl6N6-=4I zjA_qzS@wK>VQm0T1+vuG73c5HysVbVdx?^NtBzNOe*r`r4PWlBXq}X8w;(9(X*O<95p)_du|1V{JV1qtv%tfm*cdYv z+(@zy68)`bQf$~%d!0*) z8Lf_D+)|MMa5uo7-Zdu)Ev!dp7U7`_Mm=`XT8L9RS`1k_MigyRi%9WZr4%mP{`lae z^CjtsV>nNQ)Lby^UX&rnlLc=;&GN*H*hI?D?8CRjjp3J=ky#9AxTy~qap;K`t8b}* z-9)=R&7P*s8urp}*8K)ysIfmak+6ET%>PxpqWC&vuXOkZHbUry7p|Q=xQ#hOKldrvC~FO zak!SokU%bvX48V$raOCMHt7Gj_9}w&d%AE0piy9)c zA?(=CQiWdLm-rLss8b!v;}O|dL03q5qKAV7i5g@n5V)yVPPajYI5d!ZOrgbGPiP`I zPU7)H3hlXSbg`}V?NW(KP``B=SeHTr z&$Q>~Mg{J-td@ChXc$txCSmM3`0CaL4^Ifin|dD|!H(j7XbQdhZaUnEI3A1UDDPZU zsXi}GB(XA1j5)io`7|OIa$U}!q-h0puY>GC0DWIyEHpZZm{f=X&i#ELP1lIT%6fu{ zm2Tz6S>(pQEh>uh_hDqwtf!plx%x3@0ZOOmr6P0%DmF9l(?ZgYWomrcYhN{rIc49S zQ;xjFYmhU{#qYz|xsYyYbim*(yf6d==1rZC3^YRL?}bO|kB0TF+f&{SC@s78gdwV)Cw^P<=Z$G3)wPWi337qw zFGEuJQ)(*gMB!&OrF14+P28oJKM7$Df+x}m+lzA$SnU9-i!lg49Ol)VX+5~c)9Rjo zx1djZaqoGhWzDmzu58C&v7pq`R_T(eWVh-c6RF0yyYEzykn*`zJ=T3jWm9I7KmLV9#+JyCN)R;jW9wBhaj@uqkG zQ?LY3-u@yrkIX;i2X3Fx+j?V-YkM!W{$0=zr0m1a&+=Kl@7m;2Ymw>j+QAoA97XOg zG?>zF0%@_LBXaC|pN`-0AF2pAOwVaVXa`@0IPNyo!Sej{q7&U9LZwcSuJFkFm85pa zSvcm*Sw3WQKsdFlOv(9IcXh4-b#Cq&HfXs|E#z<}{r*NFh{6VG#wPmpegHvJG7Jg* zdD!^x`CXZ}eh&izSlYbDe@%NVuBboH#JPWJ+9rcAhiBETBh!`#j>hM1C(402##%>G@a(Cd_jT|rTw-JPiIe1C(zM&K(-dC4-5160f?fOX) ztFJgkqlF!T>pp9d?wD%0oT4K`vMbH?%Q{jZHM<>;@M>w-uqs3Dses%gd$C2FeUMGo zX0v$SRI_=D+i6|K9=0nby(rfjhc8=v72!M5;YRsvMYB2idQNY-p}2!|cB2J|jaP=3 z3glitxE+;bhVydX%2;nFXULEGG;Jwz)(4*P>l;N@x))J78WO!?aqoC6-u$%r{uvX7 zuF;(=n~?>yr(9+tnkcH;gWiO3kzXaKMde$G7Y_Z{6owH9)Hc)8QcF%u zv~E2OHN|_ozZG(|WZT4v#3sZDLG|=2+0rp`G^l3537>U?9YWV9Ng}C?XcS}9mV4Hm zC}hH{o;;eB(g7`~835zWRyWXiPCY5%j~j~tJQTF{K>jwV2K)?gZ2klzzLYWi7j2A( zbEvuitf);&DzjW#2~o^6J(YE@bKV1_YzmmyfW_D71@GIH#x$4L95f|fA3~V_>J4;`X~Q#U*y*`9>ujwY4Ay|bOc|)$J$MYH_;#tZ1{bZ> zHoZ#O?*dg{Wk!RL9-xX*5$!MjUY)md7Sbyy$}{hHJpRsKJ_8#AggV*o2wH3R2&qAHQW88255GiZvZ9w7=JJgzKM>#QwG&5Ks^TjLDT*ruSBz< zt4Gm(vlhlDAuRE|aE1>ET#88KD&m{6GZr!&p$4|vLesNQIx;ACX~FBw0f`e*H9KdY zmJv7RGKG}Lh7%`WBFpo4uA>_OxW<{CB2lp3X#PX5+a`AF+eJX4D|mLJ-gK$1Cf>y< zhb=2vz4LdWeiG?xE5dy6yZj(Kh}sU1aUFF>IvY?f?{r%mGQV|_kF!#F0KpC2+dkfM zZ^JZ&w+8-{zOkh7_q*k4QkEyabZ`Wlka&Ke}tG3yd(+H$!? z3ey`=9}tEgKp0rQL`ls^z;|IIugsQjKLh0ZD)5P?DUu~h4(G23LDNSCL27B2lhq2( zq=-#+Zozwe^b@T)m%0~?tt-u|Hu9KK%kjCBMqd?Dg|7jQ9OFcmXJ(w$?N6f=3mz;t z5!aE7B|T*45#Q&AGh6t;4VH&c!Q_QSi2eKqj?P%G@dnZPhjtesHz=;6Dmg zuz`Fj-Z=62o-@=^cx8PkK6&%zzV=nraKf}lh^_L-?VsPf>b%$*?v}6o74lqnTEF^! z9*ecMC78S9A?d2brRMc7Ojak*&V!lb7$1);NUcnueD$u(p1hiU8HfT%XT!Gww1;C5 zKZPHt>`^(Oxl`Vxy}*hDf5^;3fIAgQ?F1@P+l9ss+5)JzH!R^z^pgQ6)aZWQ*mFRl?k@Wg z8Zn9UHwQIx?`D^zwI(*hkajpNJ|xeie5a60hyj>wBlqp!|EqG6Dw?&p3XqRs9_ z721XT!S{WJLPGJ1`QQ|`@~#{R5b!ZHGIVeAc!-HK^AkM<<_`Y6nKog@oqS@x9}H~j zP+7uTM1c{e1PM#3&X^v;!HBPmF4FY_%civ}wUsp6*Xc~Qwm^7bEk?)7l4H+Ugr9yL z<_+ywYdtX>GYSKHTZ3jv1C5EE^GYGttYSOF0XK$7K3WBJ5$c~Sf^8M0NH%TER&+Kl zQ&nti=n?BU-NL$=;nY~pf)^{R@S4g_cBOr|CU4puqYBPCX@V8UR1wUV?RZghPj5ZN)sr5~0d<<1N_51Fo%T+I<>&%x=Hl`Y2Y87(H2Vg`+wJbqy9PaP{> zM0Ub^s#FzZtc)z`5or2D4a`s8I6Dm#8S9=C=Hyk`bnhL)8PLM-+mh(ELG6YsSsVqd zl=Vg2R{~a4AI3pLXr=-i1+1jJJ_!&~TQq?U5Q?*j!~uE)Npa{5W#$uqSb7r!QIf|yqDaqOi&AC6e=Wy0QWie zdt1GEwOqV^Be9Wv8_3+Sz_&4D$3@LoI`B!h{ZadM3Hidu&kqV02SuQ!e=sScqxB~B z)V{|b3O8>}JYRPZ&PUAo50cTu){c{dM*b*bf)bXRfr0CQF`3!Mi3e%@y3wz^f)Uu0 zFDb6mT^U}Ds)Lfhzfq>MYUft7>#g&UY6~rRS>v5|GHQ%3i2&;36iR(RmmGfO~k1IwYd0K2^l?nf)ZIE;>`S z|0}!AE!Ht3<)?tg;yo;3ytYXI--|e8N1P43+}}P1h@|M5zW*JOO2Jy>GEqv9ky~HO z(*iqKNJ%IA8-R0){V4o%ugs2TN_uT-OHW{_w-xuEc5b-1WJgZ6>daKAnE%Fz#6(tm zW9j!{>J&pTx1Sujoi7~Xd4|znv4H997;fvWauoWmz7c!182Z7GpR2RN{To3~(MhNu z+aWsZ87{+<%Mm4kC`uChuPiLe4COk#WkO*t6g0^?=L@~D$B}!tz1O@S&niB$^cjMS zQBr;jo_ejAcpr1*F+aTVMz!JLd#Imu;9)Kh+Gt{PA{cyllZ*Yp%&b{!KgalV8i9+6 ziez-olLd`e3&z7SH?wEy@td0!96`|ZYd~G=%#|+^9EF^{Kmb4G@JVYliV$Shb?XmF zN_l6digRbT(XoDPu=En%*NNr35-#b*3s#fuj>e_6YH)nLjm#V?`crp>Og60C6P=dy zo}4~81ROEmj=)nT8?VSH%6k@sM*L}2 zu|TGoq@w=uyn|P2`D`7mRVHfmaO*dAsr4>NQ4uCvG5vW8KG!ziqpj9X*Biwo_kQno zl+NhZ3yUu7J~N1YWIQs{_xFCMj^UX19pw{lEOe>r1q>;ZWGIh+;}QW~=qP``_p76Y zh#d^`<*ugTA`EwiUiLZneHB4mu+$XRJeg8IvU`SI$(q~jNQ7?36psWg7w%8yt*lRJ z77Y6W#-Zw;$|s~-jH4Zw2^VJduxb>Y(=ednh%-b)MK`jARJ<$ev5+!x-}%$OWY)-t z+86|dCN#ulDE%$SLBo$jbpzliV*!PJxq;AakLHnQjhyG*XNWqUdCk$9!jZH)?Y^c*I4dvweSMDv}P2z}td6Dci*pBe}h^zTT8{7dgJj4)3dS{<(FlR_Xa%-`VPPNKi3-oH4BHXijP_p~!(i0f^D@ z4MW^!Vx4M5A#}UNmY~RVLu=dXvoL+>+~WLq^&Zft{cQq`!0dQDX>Uyrg-h32iPW@z z@cVz`TK^Z;&hr0Z?d)s}oc|Zq&cVpZ_Plc@@hL<4JGpuwQ21B9GIkhcT89rM5d!@vR7&Ms*OH;6m*9UckC zV6b&S&&}Po?>YB5_iy>tTJ4>7dvu=G8s1gS=^?<8(Xr7v3XsK=6@~MoGc*9l=Dy*= zvh<*U*_?rQMN1~L1VjNe5D<$X8~~ji06RH3IvB6O)ZQR4SX+}D7=tFmKUg3a0yBXI zTnw;<#_)~qAle$hdtF`KQQcgf4MV!38~-d8L4ts0RtN*gumAw1Dl0E2DklIkPE^|f z90;hOOhEuEx-*0ZPyt9Az#y)`48XKE27r@%5CcPa(4c=yCnJCD02R~_XFuW@q%(VP z1vJH^gw<6=pb3db3rRp3f`I^}pZvE9=H!L^HVL3JclDPVfOG#u5t{!IJ^w5ESHg}!%sVaf4XS_+M58w3RJ=6&$u)*#smMseNz_ZNB-@n`b_^e$iY{?&gKjzm>W2T zL({)jyd3{rf8yeQ#*$K7-QJCkkIw)bdybF*9ea-tz#yI=zJK_gXmku15LZ9#hyG@t z?EkzI1qj>|K*yendP5x&WUZ7XGV5gqT{xMhUrH+3risvhmPFPw0qH)fSr<&+Q(ly! zS=Mh+M^pMcC(b6rhD?cLHEAwZ;~O6-t{93%m%fR%oT&G&{_6_#gm*O@<4<&vxS>%= zm>Z(cSyoR2UfGL#HZj9j^>?5_985iqVxqOZ79G=!|7LTCO*LjS8kt`3_XOQCs~pN~ zgPZ$jN$1wz8iSC9-SK%9lQ%NL)jzZUUlzCD0X7GkpvEYiLn?WX`#_apHh-V!yw!lL z+6{6qS@y1FxE&Mg8sh?;zZ8y@^%LyT3zdrTIh{M&*U{QN&YXIKtqp5fTU_H?+w2o}*u!n!WYUV+V<&1XRL5@5^fT z6$u5N{YoAe3JUYy6c!K9t7748?^s|hqRqP9sD4d7x1NZu#%7nktDhJGvXbPG{V->7 z#5qdNaCus@E2l1%CMqdMTYZ*MpKDCHss>lYDSQ%qf88ZvI}5L4v6s}BHT1xoh>?z5 z?czpH`+24ehf~6H%yS5~9YD$-n5hGVAu<1xs&ab*8deD$d=$=%O_MMmPZkgRK#uu~ zGhy$oI%ZycnyIg)Qhdi~x||HX!ETMZtL#L1I$#KlO;m+iizt~qMnU)HMP#2`Z1WF^ z0(%4W;U1op9KWANwJsgDzO+Qxby95SMH>feY*=R1_zNEztD;6DIRc_eS|1<@8@B%r zMF<}K%4jcgG?zz;4UX23HYM(3Rd<`vW_~hb{=Y7?L6^k?!P41C`m9|6S8-?RbVqkm z1fP9$kIz}of@So!{*T0gpT@(-c6D#duu7Rs=?BB@**+;pUM^-mq4(Rt_5d6oid_!w zkJ%pcj{8Z|-JF|R7fQUAtLlCYgn{aQhH4q_q2vZAR^sWMA5@VG%O9juX=E_sP4QxF zZs28&kuZmQU(Kb^Hm&Q{F_iq+pHR*Hves}Oohxua!cYLH6?tbb@N2-#2obde+0`Qv zvd!eD7M88tMR4kXBiBG!nU|ngZXc~Md_4siN3Mk>Si^Sg(>+JlNR0P$K?Zi=xTrV2 zI$U@=$*S4E1+eW$&v0SPYzBw3&p5YF*)^X@uaa_xX;s-+?cB2lX%04l-`Sk-?{(R^ zRj0{7t4j=8&?JDZZm~<2iR`*>P)rPE8S&oNS?)aI zbguynZ=}&&}KurAykrw+$l}d@4~v!LL|wd z?)BM-+Y==4qI9e4`sZ^GlxE8Ib4-G?zy;* zA5SX`J$3xb?b?9tT?zi+}_?-pMqg@I;2#aT^Qw1TGPjtAg}{M z413#^38@HHuLhet-_T<=j5KXsGjLAhD90EU##|_=%;@4@jOU%}2zJTh9}@7ZhXZy< z3gaXT{Vryvs#;j3HJl5RWY(A!DzvS?NUP!m{CNh?HLAy+Y-QOQhREb|OFuWKLk^)oPdKq;Ach7ZT6lFIHz7 zn-BU05OH>-Op)eFEqMzUNxjS*WpK;KH3Pn*`qHh=Vooj&q_bCDh5IF)$fCEm9}!D^ z-K!A!aSLOsttI-UX_w8`X4)yD;)RU;?(noKhuE{$56O$ry{{A5TBuw?B#s2HD{x+< zp6m0lJ?syZ=g(v@`tu2K@^MuuoxM_KT4 zL9G$l$oqevZ3m;wXF;>r$50CkBOHRDTisxwsr+&|diO`WQ zt>zj6)_Wo>QoCL)zR2sF;8T-M!dg0nMH%`sG#@O4QLLWN0Fl@aN%(~1((#wecR1=E z(&}r0>_v;660Vpg!t^KiThN<_X{Dx6Pobc$E?_nk1^6BB=12{RqKP|T{^`zuX9k9~2e^D(g8;YJnGlnm>t!cUq0XSxVS>^FRa zQ_JWZ&}cTZdZv)b)@ms_l=GUE*;=$FsF02|RU{DmZB*{O+#)F@s78a6jRE_zGP*FW z$px~S%TAp0kVfL24G;M?YG}&q#7u9U4&AmUo8xx3rs|)oVnr%+owch|y19!2 zbL}##y}!KH3VMQSNJ58139^%MkbTu-ek$UU;k$kkG}a1^YFPk)4Y`7JW&w{4R+7$% zfljB!-9H>U)yeg~lu;;tHpNL6?lwAUeRG3np~*|$llsRa*MdN>mWKGh&;2sBJQnN4 zW^{qb@ohymm=zl!Zru_s#eAk^m7f6VhS!sZkP41T>8iAL)D&vxF5)a-j#%Pz*itl~ z&>qoT{29kfseSi07mlHjrL);au=6q4K-&cqdImPg3>k5sKr_>5($?|qjgqX{{TKlz zV*TUY$`RpiQj!f5FnwKv|&?`)yqtofl?g)uF9Rqiw>g(w=hht_hbzFfAB^K5vH;%VVc z2DnM#idvzc?oms_(y(Zq#7Uui#P(pU#U@Tx;J5xR!>k@8qW1RK=pFdbNA(OK9{_Dw zD$lY-8qd(W7(1id5V%!FpUYcZ%Kp<$B$QRJ6(WA$)htPs?d=$0$L@00<(-N+2fR1l z_p#fF|NMZsHUx!A+Csg(H|Gv?o=xE9qj;euC!mZ@raT&bee%ngHG|~FKdLGyAQ4h- zAEc#E2o4SJ1K|z~JXA8nROV7GIt6SE8U|DzXI@dydW?vR?zMU29yzsb>PQb7i}-zj zNx>YPa&%YB+SfA8gN^74R+jyg5XPK%UqXZ*OSw2AnLd=}ePeioWdY0LPC36W zY>1)r7&8=N#$ErN^Pt=mH#Baclzrj9}nS@uJHCI|l+TVZ^ z{9f_JI}WA^R+6pi+3#N*c>377-{j|HIFnN{kms?O^p0@F&`+pgTg92wUWo`NtEhWu z$Rdm4?r(rSc>0mfFM-j%xcJx~w)M{J%dMo!;9ZbW`DIg;2!3>$nX8H@d)~fOBo$g4 z+^zAIsdnplV~RsOy-tNli1^svptNVp5{?>@`if#X$+57&{wL$noTPOomdsnG`?!w& z+*aT1(vt8e8phR^b`84=CIn>b#q#{<5c>lz*n4l>lOD*KA+6#L&_A)+?lP61-AWb~SL22+*MoO-y?(D%7-w@$-Q zgPq+?)x4U?4|xfAO7N;pCs+?f>EVXd>B1Eo}fpk@S?zx^6k>6iuVTnl*?yD}Dr;vD2=p$txCLg{q z^iu!Pj<8u(484*o3T~`~<`f(Zj4Lj`NNCD?%U1Qqj31o~44K>WJ@^52L1fY;sd}o( zJ*NW7A4X6%EgXIPQObXSRQ8IkOWXYVggGa$hwfiI;a)RIZS5r_%D*ny_xkALt$^o~ zoa6l63?P@#5I&n96P<%r9fUVDtHHKDXK3_(u}sT64kYVm=kU29=@D|exr5i%;B%KP z-KB>N#(r8w+`FF&-v_EsWBm#Qru~R#5xZlwr(X6V%dl znpA1qwtVv?EkDd@s($$K%K)xE94+{^81VZ4swic@04BOTI_p8~V9UQyz|)&OYB()h zR&(4#U1BtzqcA7Iy{OtXFwQf|0L;TNgwckPX);dn> z2!RFX|F13lB1(&j!0L`C5em8224wL8yIu*aO*Le!tj`MTv|G`QyvWbcl%|dvKR;TT z82sXKJ}rpk*z_&m{Ur|Z!1PKwk=G`7PV!$(G%DbwaQrfN<#fGHz|IR>E!W9qZxbP% z!F+m6C`nKiyv<8=lHySZ5sSUXm=r3`@Pf$v*72prJgMISa5@PCOo*p5 z7ep16Y$@09ljpOe9mL>dV;zbJ5yj)5#x&Z4fNlBV4n56cB!05V zQUy%*eIYlR%N=lNrZ~}F7{e3oThTMUIn0vZ{U+IaD3d7|eKd?07H`Gru995`)&dMc zBXTYv+*8hpuU+Z*uEG@PJ_cmt{yt`70o9#~1*QX6?1!|4>*Vu+N6V37g6B$j0SVEC zX+b)xuZ^{Sr&7|&)^NjsU+&z@zI%{ zV$mQJ^&2iGB6`778<5w1*W*kKc1((^VKS}b&Oj`K)YY3hUIr!eI#`T>oV4cgK!vj} zdGrmiM1;$haV*3P?2JiUl0p@CwR$H-daAO}$^nLbp}qRsIy^w`CNbZz;}Z zPaQ-RVO!T?GL3SmE98%x5~O0tphs5`*(wF;z%i*P%ccVr zK&$8}Q-JwmuCd?v@lbGpz>r8&*-=hV%%w(A2aTHAzw6O{`_U`+(ID=LRqg)lg$~#` zh-ilJL4oG-@Of&(ps5(5zPlFo>K!g((^g# zk$v)UA^0OMS&CteocG&WbRJ55v%STr^MUR?`;MmA-k5xKzpgmERXM`Iz?E}q2NF-s zN9#0(d{G(cUqESnPR|KaTR6o%V@ltK;xf_5Im=`0BFpG?s};Q|mQNxWLE9Q-?`i{u zLdYiBh2x4-Xg-+ge`{C;FdZ9U>H~s4)Y3Mn>DdAjukDZZl+piwP^=^rz6q z-8ckRB7K+drc|Z#+WOv6yUF)l`E~4X6|R%EM+>P}1?kDk#J3T*1I>dT>DC$gbjPYZ zKX<#*@1AuF0>s-RS&!L`aoG7a>Rs%3G^z_l9`8FE&S&k}p=!=pLal>u&<$F7A{U4vIH@_sU`e1e^H&CEfjYwZv^x3m}{ zk!tSPIgK8<`O7MPGwN7V(Dm0GCEkN$P*jtZIy|bOa9CR=U%au%AoZ(wCmt`tM9^Ve z2;+>;sie3~Q%aP_;EJtpkiqy>W}e zYwszaEkB*6z7rNZ$EF;r>eBOVNp>AXGHm%WW5<>(zLls$*rnfXA@g2*Br2lc`8S+8QX2NC zyKbyy$?C#&5b0-^JaE#hPL@>bHvZIf)awSEb~&0@%&{SC~keC@0Z;#?k&0H6G zP3bf`4S4=ZQYz>T<=8o6;>PBK8GS}(OmPux)y+`?pq+Z}Z=$Buym{$gI{K%6o$`!TG4lgDmSUkag? z%3A!qwCH`jDXkk>0%GdxnF|dC5XdK1J7v28C#i-?P@~H@Y}nl$VazB(yk(Pi9l-!F zU*Vz)DD@Mf9oyDH4Z0`+yFFdGE}M{|qOQOLaU1Z`OOHd2!S)JpiuXrrbX*z7ML$S| z5;Y=UXL`!Q1b@-hn2KjwlzzWmFuRV%N$@d0i3Di_=S8~=+EErE=6W%fDNmuzu|Weq zFdI|D$il`30EHg_#ssLg1evUkp+IG8SWb{IfsBB|;av(^^=?)e^d?7;I> zmXTHRRJMW1jepb+*F-a`CObGxGmSv!_@Ris+Q=M-zp+NNuB)Muw>s<%%D_oKf~+DD z%-)3&`iF?f;JR%?dD*8Oobc1FwpVC`>6;Kt2wYu8K0@GnMEls7f8;m5kFVrkM9!u@ zio&v?xmdkh*QFN$MZv+*(;&R|k z6&mZA?aeT;u*)vPk{{c>7d^*v(5bAEoEa|b$+9lE7_s8H~jFJ>-$xh3mc-k0>ts5^{x z<8f@MP5;L~;6taO!^hWWJzZ@+?-GYBmifT(Sb8U$x-lKG@asDv{qCd_w*i_fh#+|6 zi*H|IAj(i1x=yt>DaN`d^O9O&e}!jf?>Y6eXb5S}`Hy@EbKF%r6|GLp*Uf|v0 zJZU)xGbamHrP_99;mE978=2Q3c0stf3b1k3iL}++mw>;%eS(JPfK6T3beVYCD>WLt zC=1FL-PHgo=khgYKM8D}MK(V4$n0EBRZr%G=HRnJ|E-ttJ2QZb+@63@?NAriY}Lb2 z!lR{lTehC{D29lPUs3JGCVq2cr+{FIEYfR$muS-?>X^BMx}I|>f^c-Bz}4N{E+ps5 zvp6PcJXu043B5x>l$Bw`QV;o)ZcvB#K`wMyc%-n>O~!V3@cTI>9(?>U&6I4d^%l5&{RiKJzeO> zxCCvk|6A}YOm#GHv8!`fJ{ft0z>tI*9VsOw%EUxiFx2$k6}>;QbO!dUDbuWbG^1Oz z6s;`Og_z5hM>1)p1YYcTc*uDJkroO?+y2P=6^d*3PejHyJbvf$-G4(5SpO?}!1^E9 zfRTWKjfI~1Kg<7P)MKY-VflX=^{hIK;%ufHXtr9eeKuOGJ)4{VjsJXlou1AAruH7! zTkqK(dmAr?UlLK3zcCYG=_-y1h~b;-oRw~Cne_7v^GpuFNy92dX&CKm08vlW^v1?S zF=SvfXf7=UU^8S}0M5upehuL~e&LvXQ)6QwLSP8>jjwi1^h~Z`@>mMrn-KDitn@8) z^>!d->l&Nwni%Q8ZvU7A(0|Oq8LrOT`7dUXl_8KFD;r>%rg|oTFbWDw@ewIN64FBj z03@;&vKX+XUqNJkb1)+R%D~vr%6JfzzQrj--A^nKeM2Kl%bS|C?9(q?-^%jP{CjSn zH14$(fP*2W-~rSFB>)2=7%HNle>DTY>O<~+%Z%*SZ=Ot47R^sGE=}KiZ$fjwFX`Z~ z06&4z|ET$|E(o5_*#ABOSt(JfrQ>{_nfz{2Rc(v!vvJB`UPlH zXZQM1YZ9f6i6Lv5`8D~@RsIqCb!ACUYsg?a21;co_*14ACX+pprO7F{-mK8w`zO32nN{&;` zcwe#cr9kas3;OK=R2pnyNC5AQZgZVkQ-@bB$}yat=q^ikp*R#u(T`Z6yTt(z+X&aw z@@h!VTPWrqU;6N8;>ar($Y6s1)5U<|wy6gWISIcJ*hxCL!bDdEQO5_h=w)1(aE2IM z+~%z;%r#z+NVeB1YIEMbIyfmwf34$uLFbvRC|&Bj`QYp;kFwuej-1mY)#I_=>mlXw z&M`jaySUuV*S)3@{VCX5Jdo#==>Xo@$f$hHA6$K|~FMjjPhW zDBex=n`!DokLBBTY29V*XVG_$S?1y@H;18r>5Nn^qX-xF;6a;WZTF)9JQDaK&ELou z0$AC4vknG@-^D1jLhMlF8_a*+EPHJfU=WZtsy7Y1&v`h}%c6t+FbwcXz1c5i{N*(l zN0SnCB^~6O!<#t^Dc#?w7Woc(iZy4KrfMNkh26!8PzKF0>d?v4v=qP^HD3uWCD_p& zl=(>3u|HIY9VL3NPc}6FUK5M~dw@PW1nm@`2E&XJphWSfYfv@AzQ)Y)cu~!ZahELI z_WG5AlN8Z$$dCOFv>{{kO!ei&lSK_K1bqd9NZ1=Qv{>lqWBY1L+(k9wtU?JrPJKcc z1mNz+sBe#@0*JjvBZMEI{6HZHU>~&AamMaXxBC?sEA6*;@chA$N#=0H{5$b*;bp}OoTY~GvJN>lc!WLf^c@O8tgm%M9 z3Jv#n#r;OEYb4q>lPw9g;(fsCBpZZom+sIMpP5;{&1S#V)#BAV22Z*Pl43=vu1WZA zck+`eWBnyY`O8UYG!JtV>IT?(%GKY}AFa`>w7@rt^6xDwjqX5U823D3?5wY1Y0ey( zbSc}%v_0C-c8}j9{f?~IKQK#t%&;0t-BSOkbg?22FHsb1L$Wi7K%6|({;YZGNcP~a zY~$XL{pO3tBhc35GS}S3j1h|IhA@#ct#;4l(EqgrOb*d$!NtDBFwU-{@i=&=RMw+! zXP|#y;r0x%TtTN4wE*1NI{y&;s)rR)s=M8!&IgrrlP^ggmTd&J40jIFvB55U&dDC( z&Kk9(SaAxzvPsZ2B+&x}fxHyk1j0XO?cnZ4t}Q0BX4mBOcN`J(-I$*`=Rt-AGK+X9U`gN^DO8X|e&L)B$ML?+P55H&i!Z|%YBooEXP7*QWi56O6n z^3! zP_*1YJ88Fgj}V5NRUD$mLox*{rCS&DEn~hq%LVQAYmgJ;#+IBJTgoT^!r%4do48wf z@UfSR&vdg&k~MqO$ROe!!N}o#<>dGsbRuBa983$|QC(j*Fgn7S&O7gDE_pfzQi9&c zMdNLL#iV>Q#S?Wb4tW@Zmci1>xwGYpVI+InBKSo07N{Z7K3Q;iAP+nKh@zRSnNTlGdfrh<0HOA65Pvo zN=th<8jqDaJxeX+m^e9x zFH+bWY{v!Gec;KkjP>kBViZ71AyfpuBm?3(pq1Kv%Bb0!xt{ym#foEfP66ev?rIWO zm&sRdOH$M>Rd#m-*)1H}(*>(j^TOplevm*JVQ@e>%P=9vb$-VFcHcj=p`?jgM?l_;HqfQPp;st?BN#%{t#iH5 z2ltn6ibS}YE}W2_xGCZD$|o4XYpv3Jc;#BY*u484HR0b~tRWJDtT-ToeY@{?7oQTh zkoQn{HpzL8I@4=O<6F*Lak4}^V9f^aGF)6R`ct5%@PZt;VMh5Q0*;L?-wCGiwf=Za zQfa3f+!^$FO?igL_ccf8C_B65U>?}hG2fntk;g8A%Ef+;I#|~n9~mQuZBZD*c+2># ziw204`0~^NXU3`#z)90A9M+q2f8?{|xwquzYNfV8Exg$;3K9}`45VnW#>X}Q$$WAX z5B6MIL@eLV_2mS93!?@R+R>OcS{~vsZdW*fI%@v83HDd&Q#2LiQr;aE)#tZsnL}g~ zq_9`ZRg-+MYJyQa)QYa-B|ZGWsx=ASCq%%Np6wo5Q-F8~F?rIO{qTKQGVF(b#R8q280QVWha?!wMlxeci z-g;Xi-||2sT`1xi4D5X^)`@TR@TvIEM*S5kzH710=5>%Duh?9_2unTjU&?Ne8r1v| zk;D_Lt#B%ouXfzw*#T{0F3bQ8@;RfQ9IHR+c`%GC#BH}F?&bi$7Qy;|jpT*Zk#!!t z(Awt9!S1X*!sf!e#b;_&eVF@$yUUlPj`i@n&qdm!-p?ZH?%{5q>2bZ4$v?5j$Nj@D zc;O5v!PlvZi2iNSjzv^^A{}t-@I*@L9kLdG%s~pULf^id!N$Z0Y4&DDXyRrKq;PB2 zgLn_qkoj=k)jS^{Ny(x|=@wf#%EXF$Jm1Q^3*#*gnJ(%SgH(`$(?=h-s{{HQ);NUM z@r1pr#_`xz$CRAlG*hEKmcVAT>dkHXkoA;09EPF_(c zzEw-sJ;Q6k2>*`&P)fJ+wM*WJz=9On`S;h6=PYz!Odl6|Pb)e^G=q`?_l{0{$S1<{ znzzmODgb{>(OQ^6opgTM+=r7-H<(g$Uq1w)ug~DVd)lm^X?&^#&!+bJA4e1m1~{tS zX7PEX3{in2=W3AVFS`QaNgwj$Y^%z&qBtAi+<9?VfS zcwz^ad1t9+n-E(~d10rAY9!}nW10Lj7MO!SiU-Enz}}X##@YH;Ue8wF_21Z|*scWQ z<>9vFf(OrNCIv^N3qMyDqv{jidt%L?)*vF!0gF{G69(^hDr&Uh`z^dGoP{^^AEi;u znDX|^j15%b>)cXOAP3B}K<7={90<6K4i}iW*X0l9$iF;KJ2{mef3zI-a6|KGmv+46 zL0x{9H;dyRzq8$1=FjeBn8@m_b+P0e7x4epKIG0m{na{<^IwYWmXW&^_)!J2qWMkJ zmNwat7m$06_F?iGGtsoQzkq-_C0j9M9ZFl^GgpF4=*xx!x z#x;#aS2JqdxpCTIE8J~xH{+V98n_WksGr)z`lfWYsL1(;dv*vaE4;S2@7z--d~3@S za?59aDrkWX_n42L%|om-wNjS$N<|E7X(m}>i#%N>r4*OM|C$wy5VwP*v!vH3|^Kob4lH|4#-yTbAl#MXN9TRrQ15( z#Wa2;{EB`i2%Bk}0I!;t7udrb6TaR}gLJ=|1l(aC-KH^*x0Wzm=ziokf2iS^mo+Jw zZ{K<+7{^NV_r(>OJWi(;+`snM^up0Ul&8C0;`PKGmMg?1scNzj%j~6D8=%W{-Jh;v z-uroU1$(Yul$fwzBwHRLLmC^m=IUU1BN8SsUZw-X^#IHX7I-%&aIGrB_nAOD!$xrE zwA0fT$}?Y<7=~bCBw@3melCdJMiA^5kEV$9AA(Bl`y(yF%PcRDwmTl&_@$e|+w80J zE5KGypqg~1<1k1u%6EeJekKwEU>32%*vq{nY-oBLEFH6X0h)yTsTROL;qSR+Mfb_d z1pu}JY~>1H@o=}}(q>ka=(R7=eQkKicph|7;>q`ylxiHQpe?T7%<`&@2(3z{7aLbc zT3FGW&0x?c?#&C6`95VuxwVfhC;vuSwZyZQD5_nsKrzk7<4feIe7xw(ifWQ$Bg}^; zV@SBN&zo$J;-9LL_uG(O48RmOY9Dlb+tLCtyY(V8st1UIj22+`(5kAPfJc#AR`tw^ zgi=a%r(kPrRa?8z4|L>LXjrLj12V{zI8SW^~+BTb9*uP(#I%njv8t z&#)1_cdcIAg7ES%p)Fm;k6n7WQ~B#gNXDYNhRl(_*C)>3j;{m&i{lVG;0BB92o0Xp8Fj0pm@(n-%j1@a(pP08OEmS zu;Y=N4<#Ft^jW>cD!ZPAe$7{_9>Im5Yr_M~<4*X)_lBoPIxqIr&{(TeICS@PaIl z*kZ|tY(xKbh$~nGy`@g3LoJxz%<42(a}oEINb9iQFgR1C7m|0EJb+iCC5T|Ag7&5C zl*HbEnKjC$*{?h3ZiOb40dui!$YMnBmAV+NzLD)Q*0e?paKQ2UFwDEri6;;e#7A^I z0;fk>IH4gFyi8Jkwwb--WMV+h zdk7N&CaQfp%z#AS48wd^KP3k2NqxKa4C)bD#}GwB;_ZBjBh1#`vhHWvBuQqlh9GYx+$pzMl3rbTCVZ?HWbIqy`qQ-$J zEQ^WGSxu4;o8^!8jXy^Y8=Tx6JDgy*cxNnIC0wmr&xH=B=Xbq*05j+84*}!;!{m51 zL%MCf_9|=I<8(D2ngG%IBkQPpGou=@=asvTcAPws)aig~)WT58XYF!CR=!rv!t{Xn z0DL^x28DBsn24X1t*Gs{O3e&@KLC%T(jO6_=s-^4>7bw-X8&@YL{HH+KFO8(uqR;K zgLJtNV^zZBetGbrf_6j;V3xyk<%p8WD#iSF+q1xvN-V0QBZ_puQnl>EY#Tu@MJ#3i zU9r?*3rsPc8qo>h(q3Fd|52$4HRWlHuxV#)KN>cdGTy2b(S>zO!$h83+mRt7ltGtI z&PXiQb4Xuq*zVeR1f5s2oaC&C_HH0&vuT+0(*&RXpz!kgG>76T3_JN59%b%C-N^I? zjkDq*3qL)-NZ3*46)6Mu;-HR+S00AH#Pgqle0`if0ag?4D$wNY{U9qy*2WFnSg{lq z$&e&Zi)#kkgc@Nl@<-{e3&W3F`zN4jg!Io&P$BiSy`-2pOkW5?M^|gZI;ip$A~}b8 z6Hr0M$Kp3?Nl1YwQ)2|1GY=CJ`xH97!Y-^)ZW7f$^Y!z zNe_yMGAP95#4R&0*))fu@L%?nbutJs!(#OtHqGKW6s4a-$CB@kikKEEZ6mGB3eSn= zn;5_i4jpztfj(|`t)IvvWb32ft9B~1(lnsare%)c|F-DpZtQ1dvtm)w@szAkruLOS z4F=+%qUdYa_N}Y!ln`us!wzExdwV~NHsfDzLPHe%nnCWxt}p{Jj45IaNA3JUsgo~@ifVCZ{jSkty7sQ+Sj-hn$Gv#JXsvn{%!HWr zP`gR5CljB=>ge{HV!XSuU~<-8C9WWxsB-1Vsa9jZ2B@BGwmzc3y#mFrXj3 zrh^cvfC)k25oAwob!p_b5*cCC^i6^VqA!;$8_ z#kx_Z94mB+2`6MeSi+zcaGW;Cy6WiL6} z#KeiL$u(1(3}O`Ka*353gUczY9d3(YHp8QtH5I9ou=(>+EQPgq9D!~b%sep;lJorua{O%nBBrv;e0r72pkiM;Ug9coq~C0K{zTxpgH6b#M~L z>gP+yb;XE1Ofqtp5;kYf&sYO7n)}Yb1UpF#4d_@QBpXoeP@DBva4&;FWH&2aJeT#i z*wZTn;Uw-*jkvoj*1nTs;M%My&IDEn%K<}6 zWNA8(5;f#+6+Y(RXj{~PhVfAO14e~gCV2av9!n|_+mu{%ncVQFoI3FcXtY!$%NMQR zTgsq9bGFAjz*QIX1BCBp4W2TgivKehxNS94h93`ei<1HqZOIzNb)0j?%aBi~#BXk+ zJGlC+goDE1upL$o4;qJz7EEsnCn`tX`ocw2i)WHD50iqXVI*9;jJI@+zV$25mFWpV zPhPDmgMFq%sG?cfNIhh+vU~iKx_7o=7S0ua7a)CT5uThTyH<36n5GF#_!XNZd-v%lSpe~15g zh0R%9p-MD$3NkpeErWz!?yd9^-Jl6MT6Zas75Y4{{z%~gdyYM*AgS`;dSc+Op-ZP5 z4T_i)KfHGG>`i~O-n=(uH>4{Z@1ceJ&b`mRhazFn&doj1*Z9fGbV4ccxLr!onion-<3{0cNqSOQ?O0L`-UU0tU2gAHSxN-TXMAqNk(?7mH0!m8@95Jqcr zMw*EDeOvf@p)uVQSt^{*N$fQHiZtV0w0U1t-eza)wHptTRxHElpL^M~xRtizcy$NS z%hG$L&Fd_!KlTmEwUl#8?r`U&6;uD-BDydxs%qWDRMa4%zyj@Dxn~yL6A|tWHs*t; z5YCKVpDn<<|IxMsL46);$=U9F)pk)bjAN{{keABcdgdsq=uolB; zLtAhJwVe?S$~Xv?3|z0AN}b_6e4N+CsDwKK;yxQIRg3P^|h4`9PLOtHMfpL?SttLFy;HZ@XNrhyO8A>A_AvX;ld%vnQg=Pjy)iF^UT7r zX@1+H(QU37m>=O9BZCA=5Ezkn_Li7n1be6K)pv;v9^%s5Vv6WqoqC+_Ko=k0GvNqb z;hv6V3qf}2G2kE)e|4X-rs?-h(J`rRw5-TJ`Zjl}n4-*~ysGt^kK3FLlFG<< zCtMOoq$7PC>_i)}j>jegz+s;DY9F?eVqDSF`pQ&vuF`H9?)-H2B1n2EZ!7)h8hH4S z#ri3c-u+E@cV3RilRCZ2ss2rU3EYCXndb-Y@x=sYddpoT$4S zm#e$!9hZ(15X~HfXpW_ zOS7Fd55P^SkMN97z_XD^>%8+oT}|AkX1GJRUwC<(8Pg_Bj@v2DF$hwv`DbEkxx=#) zEgr!+$3Q!C?bpZ^xg7zU#%ZmnW~MYa@oNEWK}+r3;2Vr>(uLm8kpTrQXrSe061Z)>0<37VaAIgE5)U zG<_re7Ngl4ZbdvB3%KEl9O;iXj=_^0@_mUOqup={QI9~WCSl0~-Z(zH06*%~FURDb5vl5NPN7O=X6 z5DWQoYk)m>qY+ThWs3mkQ4e6CZ3qCwfRgmk{E6E^}9YIvbvUsd~d|&6{5K$6G zUwmWl*`?dt+D@hb#@$ z83^X?%~qlbP|L^%jscSLEa3u)EpbKO&<3Fav2iq8w0u!J)&M>FZJ2L=I}6v7N7``F zd-jdm4p_F@xH9Fh9X^Ru#Ie}&_SOP~You_QyGdlD;dUb9vXemK)2%%X2Y;1@)1N z_WQHEgE11rO-waYcKv6DIHoiEl(Ik(HuYr_C3pg2Hg`YyCi^9@G$Pc^J;( z&;MFkF*?+w#Sfl!#;*h~iBb);)V&O-XY3oBt&f=Z^C@)_m0=21LtR_6kT2Gr)~ z!?GW$But}JQ&^qqWaQ5Uou(eRgcNindTQl$aO)`k*~PrLnKXVge9$nt>Y!*rDeZi@ zkf1(wcK!UH6SsW5ia-3_fQ1MOG{?Ps{t>%~B<@arUf@|wY+cvxh3 z2}ppt)T24kKiqy;-oe%7Z-*1{Qr+(D^$m)SzGX^e<*u+AcLOzWqfU7_t@|TSZr9aE zqv!<{KTbIXz-By8jY_{8+X%XtlU$O;IvltBhyMv zDPDy)A~kaUbd~dpU7c2OGB2r@0!blLB0}q|Clmzj2vf zyJMLot|!lh6^RrP4E%ErV|GToH)%k(6}TKMxDI^G*f50_-t0vxVsfLxQaVcD8z;AQ z4FfhF@QA2QDUziTrj80Ci+F6}>uAmZt75HH!T*k2#4?DZ?GZ-Hp)iu_NJs?<{&x8~ z_BIkKtF1f*-iy4P4a*F-LI)VxtjJJehKj~NXCIUiM=77CUBJd)Tg%SZm&qpP4V9%uxNn;9&^d(oO~5zveC z%N~vt`C-`LGalY7C-NsGdrp)IFM=MCjf-HTf0`O!a5+LLX%2R#{*qwhC&jbOhCbpr zm;$lSb+yD5-k4T|y}Q;Fz*RuU)MuIAH$L(K%+xhoiS=^{vuwd#=SltJht&@A|3-Um zmr*=nC>FCBW#HDmzjQRFEaoT4ofuY9j(rp(te!elkSRGk}48mD)jDa98zyME?O9CDkaZ38D{!iz&WAVEKG9 z&)BGBEsC|1X~yVEtQ&*wf(zdpZcA?}0+d?4LP5|vsO7752eIsWzsM%^janS@P+zts zcfuMnB`ef0P-u;X@ir;`&R1_#S)uRJTu(z$kTG$*Zs@da8!Zp#ig8cpfB3uWuCS#{ zF%2a{+3V<A)A!qQy#9%C~A@?X$=Fmz6v{9=w?_>l6V|xrA8b zux!dPif`f@FYm=f9Vs4(zya+R)GTnn1%8!uc?U9MvgwQW>|wY@UFd)qeW90@O+Y*TGBF1CnIbXgxo=e-8K z_^@TV(NjN@s{tQY7->KPvQ|NBFx7eEIEnnfI`w`Qb`t0CdT1Xo2T>ogj8?Yk-$W~z zjA*!}^*#RZl|1IJHO{Nfs4?7ApzOqZ6N5fvOYGBm6`+DgSf4wKP$$r%-a<8F9br{D zSDl9$gZKyf>%;>LI0^n4;dTipMz5^>(;A%&C2pL z{jRZJp68GJo`l2owfagUw8Jt$SQX%qxrGdkl92`Stzbbd6#W&PX}qoB&L#?yCI9M`Me0LpxqDY`Vqp?-X0# zKbBBY!uSI_%N@-EEA+c?UNk%!-Ao#1p>*mo+&5!;gS1Ek)EhVmnRuUTirKGM=T=jx zEc3gVFO23(BW?UlIFm}keO>gv?iQaR0mO_Y>VJ$oNlDoyiVoHd?5XqaZU7B+;Zu@} z{7toRp`KXVgfPe`@LXg7T)Vwo?O_aU$$?>cLh>}tC+1Eo7;xg`|DHOiF0KnPxEXq%daWDc1~&P zvy7SCkE(Huz42a7-m(j!oNc5Q!>*$5y?Go?;d=Ci-r;7;Fir1y^_`=aE->t)n%;QZ zT2y%hrX0U!;!45!0686`!{I#mocory@Ml?8E0-^Y}O^RWN=`TjnAX%!xr z2T=B4jBa+Lgls#Dk`viNHS5R^tBZc2=X;SlPcBEPtQdXoa&KKwn(TQ&eJM_gw$r+s z7}A7%1KcVE8SoX5vHI_YlYL#cQ*S{#ptI2|;h>o*qh&OJBObs0C3toBvZOcqPCV6n z)zgmglJVS1y1BDT;=wi>k@0nvAu~F=C2qlDY7k4svP!!##$#aO$z4SOgQOPb`L}I=YHNsT&t51cr}MJCP7d*s>Z{Fp zfEkIIUd1Yj*pp)O=Wk&E$>v*9umC_+=LA{fm#G~GJ0kKe%}v6^R+lR@++UW-EUaU9 zv;=xy=Tqse<5u(DhDkDc#!RXi({O{of=n#)3qV^heLX*|vKs?c9A2+f=i6R|AeneI zE`BD})ygPB1_vARV06g$kJD~2;?JMYrtD)qEXzKRePb$Lmj+Jb#F-GL1yUmAV7u*N zHM)VIH-{lrLKy1Ca!q!~%fo(%%tV`E-YHlM?l%HhN5>4321}E1U5NAir~qn5x3!s2 zT<>h?=QE?1AXa!VUKlOfzr!B89Ip6^PeuYeV5zofs?@qGDK@{7i`^-G0#8{p56yp< zQKVl=yZ^=c)O>a+#D97|rU(K#uQm6*ba-EcK(yVgEJSh4_f=Y3#|vJ*P^-Siwm@vZ zS;R6Fxb8n>2z|r5>bkb{Z8XPbuU!6mZ^ySkEt-1ixiR=k=;2m|pH;=w!>-b@YQ~Y_ z;Is*^ECByCx8$D}gB~xr^l91xN;L>ARIHz+3YgH2m3H%ErGoa)8k$phv&!CL=}}PF zV7S5E%CZ0508&XTrV%wrrW31Z6MQ**GNt+>oiTR!h1A>?ixY_d!VmlgKV$Y+npTMM zxNP@LIH!oIwNB%WW2Vak*wB;+8Iu(qlkWV^>x=g=;DxmGUG)y0$W=y^y!b0En(<>< z*Qik#lB(A8X^f*CesjnxILF~n{h9+?1HWSi_ae4!er-)CsRUgM!R3d-E`iv#(6YzgADcjyOgH>qH~o;HK13qXD< zc3qXgt+sjJ$elk6;<$+;(ngN}(RfD>2e4jA?W$UJEBI|31Y>cEX*C8Mac(hEt#7?f z0gp`*|7OGn^Fc*w@Ug>Qko*Jb-(e0~*RA!$UN$!RMKk0c9VuJH24pWJN-KUh?wWN( zr?WOC8D-CLFXW6UuKPL&Q9%{E-)tMrg7*h4l}q_-X4E42Cc==u5jTzdte0i4e9>=n zLxFNTd&wmX{D+D%^%Wj<&57nMGG1 zviV9q1J^cuJ=FN%Tnt&i&GMp<{c=K;DGc^mHEB;qK#Zm~1Z=8g*HY)&5+G+j-Y=U9 z%+D2pRFUmC_HX};6p-?yDdX@U-;y)T;EJtC5#=9ly-&pJ?^x8D(i#d`aE2|BYA78t z>xo*U39SI*8=R&<+~Qk=r1|VTf|Qj6P1n8JV=^Yg*irukwm?OtU2(Y-zD31MWNeU1 zXd*g6)r}{w?P*O}Iv9_n0f$Y-!_yV}w=l@?^fyC2w4mp}XdHQG9Yi1Twrl#auxE~L z{xl#v+vv_K%E{m>DrQ#u-d;WkQdM0ylI?(pQnNYBj2o@P|bU4 zFiczZmxmH@rO>Nd6h_`dSVnfmW`sH@t*JCmSiFOs3DENnof2{BM+1!{atPG|squ8j z1bgZj=r!~OT(D9IOo4wolWs5=#)ve$UdwT&j3tW4z(VXv9gWf&+M$snnAm)s36xV+ zd#=1Ca!unn9qjWRmFG{jeP-t1J`H59sogVJ4vsP4amC&Y^D?CxC;B-l$Cd25b!xgJ z$98~~E$NThA9JB>@JVKzu!Uv%)JshBzuXa8FDNjT6bJ3$*gr*~D3Jf&sNJLR`!0jN zfpgnPMAUh?&nwn{`n`FGS7c%=(^JqO49^g<(1ad63z96v&GkjydvQnl*eSM%cwvpp zPD`}oCer-lyI^4zA5+7_3JMmNN1O9VN3?p%3MK%!K<89(CI-P0FEcGCIY=yCp6uul z4e!qg0l8`!`KS0N=4WHme->M6`nE{eZh9D5gU(djRX_+w6V!VW0t$2>_6-S8=5h|p zm9eP%`IXrBa#&lN>;tG7-NmgO$+3;W^S3yzDoI@36me|v$)%AFVUon2X?7cVizQY| ze;s`$y7dh|$nGm&X5}4x;8ePF-52umY~w-yVAz!%W(t;Y{RKP)@eX{64M+c@E;Gvu zdvr(~!5kOpppLYsyI0+WQ{ggbZ%WvCXpmkRLUy0|k~BH^0oVZz5~co=u&-Wh#OoS@ z2@hsn&Bs{hMmsAL7(efaR2S=Hf5i7KtvvvCaKz8uq%f^QVfy=dk$82?_3TU_K5N^q zoI|~ozQ!ZPTv1hY8hwj!(n1YUowt=aPjq@F+!@?g_Ofk)i0}DP?Z|#XN?g%j#s8tp7S|QDFv&BxK@PU zvdB{>Zh1B;Wg}s=mVX%OOT8wkNaw3UD6uXH!y#H^c_ukle6(9BMc%w6%)!U%3&pG20Zi1Cv}NEJF*Zoa9psE z$Au$c>En@o{dx)FVS2TaJ4Ol_YOG=vAMMWeY`A=XXv)&M)BGA=HS)ImIQNYWLcLTq z_4_G?72t%iMmxW?E89gwbJ_t={oKAZ%Jv}g4GaI!ObHaC&&OdQaf_3N2Rv}9B^oPx zS`THKk1f*5nd+Lp0V+j|OZTzx^AMQLQAd-b@Fqg; zi;WdrE5jz76$<(mGRnN)KESVvB3mPnNbt{E9Z4!KpCrmaKeU7_A-7^0fXeK?i3g`P@;}oaC;7 zz_+;O;R(IR74Mo_fSgRJBb!B)eD2M(!x?&TI}d~NtUI#Ys+f|mg?u?OX<%8Qfb)(Y zBVqxd4Eo@KQJP}O@w~2c9Yj*beGo2}()ya`h5yhU0||b{?v_-Gj?wVf#vj7#Oi}Q; z%=03_#r@f6hZJAdQ9#0rkpsq>-2&{mDbm63Akoh|6Z0sW05`hKb}Z13^u8!Mo%8k8 zn-G(YyPeevICTtz!73d_cIlDj|K@A$FR3mo7A#DySqEwGymp>^;G251I|bbdj!KT?MuOeL%vbY}JC38Sp#ipd$h2bvHbmHaL%{Z3 z9<#s3GV*w{T(5M!_i~c1vF_~{@GkfkeFJJ<$iy-@EtseG(ZVNqe zz3F22-q=R@W(2}H|#d? zi*}wg3oSbOa4lCi|HCJKRy<3?sTFcXtOgR#y_Ql+o+8;Q?4;Jhn_MV9{5g)|6do;+ zSP}p8V`m`O@%3)QC8UYF-NNH#s08wLt}l!VF7+FO4a`#=`G`IF52Hu*_%pG`LqnkB z9VY1sJ!PXz#3fH$s40hHYc+%&0phDprN$NM&SutfwhTy_Sp4DP02&-PuDLOJ9!>7m zvIlG=u7XX_o|5{HuZ2Ldu5geog*t_zuE*O9Bg~W+wAEUoB2v8TYn{O@^q#GPh^Fx^ zbnZ z=2{2b0!Jrv3p`2`^X&Lg=6mzR7s5wPy>Bjw)m~gw1!N|4-t!)~&6lzzxDCE>rFB>O zNP@vb{-9oT|1&ejafNWqwz!S~t#4NhBoKb}gEtArboOx0)`(CDkrNN!#w8jXeP->0 zhCQh&jJq{e0mIdooq@Isw0s#mA;w3@|1fqB%c3aJmO!^{+s4_pZQHhO+qP}nwr$(C z>Q>`LRO2DgOM7BEciJK`bz%)=WufPi;B>>fK^2!*txgQPFnV#rMuqV%F6@@_O-Q|<86h+T~D zkzDRwV%>NRECvj8zFydPaTV!gYy-8gGmdNF zuYQ-+t>9Lc8tLCCqj5RWSHqN>xS@K9O+1l`uKDQnNB>76M^6GY((%9>g!j}zJ|3f8 z!I<&qB!{jHX@Knm{r6;u)aer{x;$M;)swSFPbpO$|K&4 znN~T#sYd-CmTbB(GTevdf2Z|`AzE(Qohh>|SBhq4Vkok?8lPiCPxdMK>orW2>w7E7 za350^;gm%2=Vkuc4Fe*?Ojo&nioKg5Rfc?^9d)(}E#aX0MfX@EInnp%kpu)fXa=5V z|D_YngEmus5oanaT(2(jpo_PSa$g|M_eYm(%?6wJSmFUbaJD4fn0dA&H0b1LUq4s=-M)f6>Doiklto(Lj@B1WUtACD*9xW%OWX4%|YA<-Dax$F3C3;%*E7AvNnvQsWy*DUCu0fYb3Y9C4FZEgDH07rZ(5#$Zz?O z`&@*n<7Vl0x=Pm85M@DRt|Kcl)QAniV$yHET)+RjRC0-}=^bg1d#z}qXXlkQRun)0 zF2OTr(w<4Yf_f-U2Nm`}a2dG$?A0v(&b^Hh?z_lO+K`@3Wdl~@0K}Sz2^WYdo?lA?Q zIA`PkK(U9vb|%oOOFfR}Mb^j%^DI-%hP(Ke*klkp?s3-|58QROfZLB^OYd4HDqH^B z#mA)J&`mWzYWor^D@f_{xgQVSs)~-hbZh_a&)2r$RypP2Q!ZzH0Z+IEh7p5v&Okpa z%|PK!D|#4LAAaiE$K*=n-(G!`Y@oTTOTd-DrL#21sa}YfAOOmn`JG^~>qlKIi*U>Ql?UJGur+cNlXfX!x%*4_pLuRm|t{gi3_*@*ZMm|_V!AsR7 zVG)~p;?dWcMdFb2%ZMx>6n*A%F~-5w@ci`w%z`p}Mp05p zQKo2+yL)N~@mbP^;SPQ2q^;V4ckXWy@=0TI+|QD8hQJr-x@k9lC*Gvl{k|U0?g!>C z5~!KbhX=7!4FNt z$PH}e@KQuuLP;u zDG#r+R~iD6FLtuIRq_j0voK4gqIgQA zw2T~0WI)LP%F!B#QD&}WHiRXt)t7##M{Dl-O7iBJMNw_Oy`&CVVg8cqCeyE_^dQnQt>^Yt&|Ur zeQVl(aA0c8LJSIvgC`b-!E5LC+;EFTjLs$cQ6;$7EAk5PADN1X+u|8qFh@5U=^mhR zSX%9IlWDe+N~p8}Jv!f>z)A^Yc9a+BHq-sKO~%jAgdlLL1Wrf}CMiT;#15LQ5u`a; zCY(Pppe8-mw&NF^Rd@ueV4-~{sadi3WX#k>b;fr-Gs$LRe7K}5l`>sX8o zw@91os1fQ)q()XHTT4|-RGTJc6n`Pd!HZP@jOAg2;%_!;Nxf|AdX#ZgCR^Vop+BIH z4R5vS!x`|f1j*x%zwmu0{z#0`O4U!y))#IHaGL zCl$P&|Ekbgx{#1WjB6|5Sx}$U%(1dFQ8%s#BPEa&3%l4KhIz4AH(pgrU}8Kw0E}CM zs&vibmb5|>M|6)n-C*GQilG&2m|`SLNC1!Qh9HPAWYB}{?b>LAZ?h!t`RslXaE;r( zAut(U4(kxNq*9cFIqwv-5x5UjMg)x3xCFg?1Tnxg1dL!3ma`JM4iF&n=n|r`M-7K~ zF{&^scpiJ}(+tN{9Bwlip-`=_E^_8ihAz$&<*KFsH~xnGf8}pj{>QsvBw(UvW@G=K zH%`u_bBq77jSnMkxEuZ>Z*s{0^KV$mvfT~0a@SjD?%LjX@4Qxd zue9FyR#aB1>N-1Dcc>VN9K5*DRph`4ka3ZDh}rl>v{h0f5~5)H#pi&eM9B_8m>nCM z+mKlt0Ud#v{YwE*^MU^dz>)e7z;U~GZe(t=2b9m0|Ivb!Z)0F=X>M@_D&N{#@6yOr z2XNon-Uivp*15>h{v;3fwo_+oU(?0`qcKiTO&q(T_<(HLK5!V*b zMI|5`%%}Q~$zcK=m;7DYnH|aa!nezhh??pvECOx+t}`}z?n`|5&%d#9d3=u(hMyV+ zp>JV!0K@>Ep}o<^{J;K9P6p3^{*9%QTG$`M5AgU0zBj<6AA2!SI?&ATgKJAe0$U>} z5QWIK4X(D1zbpWUmr{Nb?!Ml37njCw(FeX7v6d9}+@Idr-x1-7;WIp@pzG=1{Mb8w zN=Z|D3HZG7!r}>juTHjhz@LH@z#kKvBfpZA+SZ>l`0vmJUz4S!o!QZcpXfJvvR~X7 z0TCHVDXHL7pTxagQ+iHBPen-%2miW71-M<$kwAQ!#QX^Ai(D{aXCct#{^>l!2n&{p?(6!-}2}tv^ zN4v$ph!elfKUp$@d0=y(u1;gSL1e>}T^!W|a#e0ENC74;n{v^nVUSMl6`z_QYYuY* zOMCkWr?N<;eYh|_Wb^4Zrv_S=x3F~zJP0>t223H8g0|o%MM*`ATeYT-Qss9|`~pAF z+E^PaB6we1kE_IvR;*%io{7vHH&x;r_0cez0Td(k-Cgj2=9K!zdwp_&60v~9k{?j; z?V^tuW-CNZXA_$H#x5u{U4(96HW_XxER__ zOwryG)elUzc0>|qw~0pRjD~}M&CdB>>izWG@R%wf4l`Bt4X6YJVRiOHuqjRNgi0e- zXsCT3<71i|2iP9QRY&8W1QU@5Q`;0OLZKOPOqA=LQ|K6Yay@p-Wxa?$3>!PZU zI`iMy*FM)H!Jt@tLrxZ~3kSf!98{dGlpnuSmMj*sm{{{>STpJ!!HTCkE513r7T~$= zU6#K2yNuX15|ZfC7m`+0FG$*?wL_767b4Wz=u}gDFM9O zq+gz26lLVB9~*lECkKr23Q$!Bv%W`pbYl|r=plC~fwAO~$?wGkKhf;88*aqf^mOKu z1nGhHd!(AqBuiJLRXGTU+Q=rwCQ-*zM)bFv_O&s}Jl9|6)^1f>GNw;;Tee@XxIJi9 zq7G;aDLA1qu0d(!8$&Yc?A;L1#)nAddK`z2+XKNkV1&F|Ip zVO&?CE^VPTBv`$?*MguX%LMa4lH80saY?XYT+vA78B}&4;laGRL5>H&vyo+|3$4G8 z=>aU#l}X*nnp0)?E3D(nS!o9e5>~xr*1xk`5}cdATRzU!9BNx~vkLji5GxE>PmBvc zdOq3tOI>W+X}e3{+u1p!``Q&{q^#%f+h3C`C`9NzuzM7Iv{3Q%G;vS`VtK}0j8OY! z?-VEeQEXFK2TNAB33#JB=%v&=T!K&w)!qflw0>i#isUV&CHw^V>=Do$F_lYbTKFdH zT=fDS-&4wivRg1k|F|AwbS${J=c8t|sa+WXYMs$$gAz`8Q+j-cK#TMnyXWlW%FfkH z!xR|+kM^aXEd^O$WVmHD_)MLD0dWhNDP-}42M($Y$3ijL>rZo?nx06E;Yd2H3 z(cv*L)!TuF%~7WkV=IYc5n>b{iDgn6?;*cc9}=^bXr+J zg|)DT`f+@?Y3-c$^95;I$d=|5$GU+jP!DaL?XzT0khv|EDtjCmHHl8{Jmf+~^-@HJUjlE4_nrfNunm|7?6iVp_q|=*al>5m?8v827U~7zLGE34lsi;F75d zDBE$pnCs0Tg+H}NxYj2k2k4ML!vrBQfN{8+pHhBLJK?iFgQw5{1r!|8<*#($U+7lu zO=)P=yAm}<_D+pt&1oJ~y32}OgUmD@nL4}`E!e!hJpI#;DdP^k08JaSR%1W7O-E1< z5WNHu`xx!$C7S*YsL_-9@P5)5c00Dc?WdFgUHTjnmpm1Ag_Ar=p~22h0C2ZP`S79` z(Xuf29Ar|7L$eus?7QI-Xhw%_tK!1au#Oo_Bv~-j27~}414mP;@=PHrru%K=RY=uP zaM%N;Yzw&)YTY3ORJKpr!g<&T=?FXfe2h0xfE%y~h-vKP+&(YiQ}Kw>8q{NP9B8^^ zhb#-<+ZNCxK5lOeV96u{hGjQ^f1UbCv+lU>?F)bm{DsN-Q3Y5g7bJ>r#7unB7L8(dk zfigc{HVP@%E$&Ih*p-GE5y`I7Eedin!2eFhkS~^Z=LQYzntdX07+>or8QrbHmaSI4 zSsu~Eno-8FNX`1NaV~OAa&CzU3Egxjt5_d}L+|u*7){y#L9b?}G@C-EaKF5;&$M!# zzB3-<(l|d;xtNNZ5u1reD4_#$*BO?Ok9EsK*RsztoY@IKH;J%}m`S@dqI-7uijGM& z$18ct@W{0VrS%su(; zUju5I^0Ors`l9XT@1G_N#=QPRt!l~q*@&qAgz@|f?UwcHVoNAG<*c8vINKr{TJSDt zu{jQsei(~9Ghty1(A~Q6*#Udm#X6_55m*^1U;97&74Smv1%aLIz zH-bz%I68&nV1BAN(1?6ZN4bc;NAoge2W@HvwkSe`@9kv?+`FJ8?3u~&x|>Z9ypR3M z2!Eyam)0*+^9lHUY!lz9sv!)`BMbPuf_3bjifl%v(Lhlh#fPg&oZ<6%FC+p$ZU zsfw+RP~*j-#-nk&wb#dS4Y=|infWP1iAUQOs^H$0R<6@m-oaEL^m5wFG)tVN~PPD!@`u?PDNi{ zJq~e&daLWubEaDWs=5gMOtEWQ0oxIknoh|pWG#||z>Bn@@Dq4qz(r&Rh4pEtrJgIm zUpjfbAip#0T&N55WC$*$HeUSUW+gYC$xREQxZa{woM6geOpIxq4UUP`1yWYN92QZL zylMO6sN75IuY&GtAdIP^$weSR@^dTZ@-^nJGpMb0GCKKPPZHor?TUr69qJd5;;p-g zeC1J6lIX~INBMKnmO9Vk6@F0haugv!b`>Lp!lTO{O)uC)F_uu$nd})7x--rH1(-<% zH<53qv--Jf=_n{4g`#rTKV5FV`m=-OZiZOxxGs^%%}|am2Gbl$qIYUQjc^jb9FVWW zR&qQyLj(S5!&NnMBVyY`zW*?+wZzC9z9f)>E~&6y;6&={tE?(0P7;TUsR5DKWg)d^`bPp*VBE5HLpRwG|uGm65hkh^q;_!Vat_%^8ysTIEmN%E$upT@-UV|4%{#$YnmtESl zj23{Ec-=*1C`w@y-q*T=TPh__g*CuUl9IDP+>#~74L?9L3p^D}%=DyW5>LCfP=e&I zbQo2~yK%3Qauu$8s7_Rn{nq4{gCGqFeoKE5WZ_tNY_7sdz|>L##UPaZW*w)kj5_Eu zUN?=yU)@s^Q|Xo;RIpU_ey$sQ7#qTZC8m;>TCR&gMz5L1BlI}A8Sn3AokqNP-d*TU zxNENelD>hcCRQ;mKG%dejdBN$f#bA-aUdhKy8J!)o9M2NsO+)Xg$3dB$(}t-{dd2_ zSj=}=^~v?b1vdXsgejxFHCc*{fn^xeMza2sbNv_`Sk*u@S zC#_Ga?XnHHdbmHFiw7QMpIP5$oG(hYo^uMF3Ek5_YvD2%P=_f)P3Fj@hT}C3K^?^F zV!h?SNIycSt+^wNi@bgJmUI2uvG`)I!L5>nWAm9}Q}JcqgBZ*2Rq_e=l6Fprxlhhl ze4w7*niTY?b~u+8Xa&E|Ort6Ao1VwS;EHQC$9DfdLVm#9QslG`>izgY@l-s)LJ*d< zr!7$iSV+U875YJzOhK5LV7(wd3q_?wcoDhrQbuHS<*E^R^0)zU6K&AsAeA#Zlybcb zD{BW4p^CVxZ1Sr(PwJ8ZaXbsufh_s5z)pdqHiyT2;5w|nZzAU0D_a{)wro{*J%{25 zX!CaM_f;#1zzYXu0@;kmr&z6VrQ^hPTK^eeaP^*I3(nO3^7at{RV9v^CD|pJSAw)W zyyJ^vaf)_|6cWtcl;mEQUj#aDq7@%NTbA&r8i%*-_xZ{-~)6_HHGqD@bnqq#I9lM0+)z}B$e>GWAbze zd%+6f)v&{{os%>_?;Ch^uBY{;ae}Uy?>XuPd;W@St%M;iEl4 zUiELpaPec#*_uYMWNNoPw3{nJ)OG_h4DF)gNWI~j&mpbox8Vv(zzK8(ne5>%?_hu0 zqnj;{6F1Qm98{wi%55-b_A41fu0pT2zjFD4;ROCTR1?QzHGvV|u;tf+YA8q+J&bk5 z{5loC;3EFD09jk;!p5pfWZ|q6H7-VP5d+zsZ6Ci|rwZ5KW5%KpvmEvnk_ZiiUkVyC z^Dbf{%qxe;lOz_yD)?uQL)gq%;K$6ym9`ez!Xi9ng{2+!(|t>c4d;t<#d75e%+K0d z5tggEWxO$jFj+E^jn~sE^CI@UykLOb?enqoDH(yULf=PLpJOP#o0UHTt z*;ikJBlqyTXVkw&U`t{)Vf2)0+BVl*m?XuLVK1r#H+rD;P#^^7L_{Ose)nX}=ib>0UQ4o!(z#4u&wIjQ6hNQK&?k#z0 zo2U+SSb0kR55<#K19t3F`wD8%|q%db?fh!ksC-Wuqx zN9y!}BLgtuMNAW$iEla=sk5*AQZfN|L}7fcgokd#w$s#}u7*CEHtXtsP0 zz`&lMK02Y67COG4TctT&$qTQWgd#UNNRS|WCXH$`HprMklT({bc=(q~96hL|l;(kn9EP^MT6Ui^OG4 zKUHV+bS(V!aqew9Q_vw=zDwuzx?iQVepQV1@HweYZI=SmXjZSC`F%h+)v-c6KUQp> zoM0oK=pK!&E9|LrR3frE{yMRy0j5kDyrjXU%&QT0u_(YM8vQ<4c6L5!6u|-CY~B+j z*ihPk`Ur*!XahUqQni3Q!rC>Mh@n27ZR=)?lBQb&y!}LKsW_{E5kZI&`4iTkVbjc>BAFdL0-@UN7YljmWa3^tB{<`ZnM$JGS}B6w5aG6X>I8Mo?HP?0JRYAAfv;d;+JhX#veRj z9WO+rMkBt9x~kZ;eM)cglj`>)9J-eM0z;SP*V+N}03T(=cD^OaLHqIdmE*i#(AFJ8 zVuq799r`0{s1)BKR=?ZWiag=6SB+0;>2%#<@y`TVv*a@o@q0V;_S%@iMMdH^_f1nb z?la+4)7wGHUc{!mM03jJ8xDnY%ftUNB3+a&&RGrP?Myntwtlw>3wrTsiuk#Zr=dPy zr4Li*ra`4O+*nRZXZlYKTt<^JEC)O02v-7nBq1Dx8j9Um>#b9A9`R3`6M;X}0|^Q4 z1!t_a{}9H_)OF(5=3$HXUUu>SF2W>|N~fdh>!#gnA8CuyN9QthwT8MjN{G)rhFt;0E8%5i)?c7CarzxHgdZz_pVyf1e`VImkB_fknrVn6# z!e;oYwpGs0x~<$LIZi-< z8?zwc4Jv%o@Ca9ZaJ4$tr*H4BtcgLjFo)6&jcg-1D63f9CPZqLs2n%*HN)aPll!%_4pYzDnqbcCvw900 zOW~Q%w5#1MrF^cmtjrJ*DrEztc9RP42H=vaxtM4@) z_mmP!wiMR=g^bT(TRY6hhRAAR)fs>Bu7^{0`yvj`0kiieQ~4JM_@$tKW~5crvm>R? zgGZbBp8@sGC)V7>Jz19_lFH6qoc;O;$C&pF*=2jk5};X`KOgyEL!z#MD!m{(&>)-V z?NYwP!tKYBHej7~_G$`~Z{kLfrPC)QpNr#J7)_!u3rS%qpJ6zh<{mWZs@XFzt~31X z<;~LvQF;uYy=CHWASndL2@-OPOxw;AP%3wpkC#tqdL{hd_)00}9R}N=Bze@k(#=`^ z>}6mfDz%zVa)!IIwlD}Y=9}b+`EX%APx7#KS=`ctjdj5f=jcsxVobVqU_t)^lcMlp zm`LDF3#N%foczDPy&qo)O5gy|=xoFV$V@4!8jjkkOtL75$vPK*-I0>ps4`EOP8j}E zAT2cG=zXPe9ovKds|j^8-(%B!a;Tm zno9~JHO4MW@p4s6(NL=`y0yl5;>r#J95MeYcPVN^R0QuuQ9PHN7QmT$i7@eFffEW1 zwiwnu?UVfuGu_TJKv_fnldY%QwlL31*jb~KQ3t#K-cr?dqV%lQHLY-|tcA~8(zLBJ z(*9LT%_IL+>oZiJQ5|P?sG+M7Wly%EY?5a|S8}I5xOtBWw}PDI_94@13MP%VnElJyjUHn)qem@ zImcyALO+s?{?IJp;4pvYe-aSK-#s&^sU8og6o5NlHu3x(O#tA`_^FTvlMFw!snS(^ zt2owXKQne>Q?=`*K*ZC8UHupBtGQ7U6REQ^6y0$HPvdfL$k)L{@`zDD4guYbWGVkO zIqe#yX^1-Vx9FUQwN@b%A_80c^A)28CSRPZ!%lvi7^qJaa3Z_l=3oB%FWV<42`E*8 z;R+-doU0NMW#AO$d5&XgzSD=v;75ZSjjlXgk?o!CHebLH92#H3f)A~yIUb{ z9qO#D6_Te$jl-&2gdBjfCG$PQCCS%bG&Ra_$jDg6TMm@MM{#FZD7|cD(7L@2YwEV+ zOF`e%CzTQ^7az*KU1jy+segI|+%nI;X)M+gAOLRi@Jc~k-#zpvtJmZ?)aR*G062Aw z)7;`TnoHEth)!anNXm7c9b(m!l?TxQeyD2~X;8r{5_(`Kf!x4~2Z_hk5W@=-*F*)* zgpC=@T&Dl_WSP#~KrcR(8Yryhc_tb%0I!xc=S@wxK;O2mD9C0WshIKF3or=k;;HmY%sj2M9lg=TVpN0l5Wnycl0r0jVOPLE$4b6$L`r77dK&|4PQ2VORfO zb%)K$fsG{^lp)kjej1igAivyxt~X(@;3w(}2iKLlQc8-RLV92dD3_L)vaig8C{TEl zo)YW_`zj9JC>EgvQfzfzYBhKXk^gt#aY3T}>B0mX6st zwlEM+Af9%G|0Q)+-U_JWY0(SCW}!nRDX^LN^}Ib4m|5NobgIIZe>Xf?$rax)<)bs5 z5lY8<$Tj%-Y;vsxDV;(9&6UjRvrQb=H8ODaf0ccw)racKa!S0p%2YTs>rSjsr&^cQ z*%pP~Si*|Ym5=w61H{BZkXHMe$9uz7X^#V96P;fp6Itu_q{YZR`?w zg>Z*OU{Em2Hgow(W%tDe!;i~rJ`#cem3}93YEQV0WQXGMr?Ug$OsH=t+9r`zF9`a@V*^fhBd8PSNYPW zakz#Q224P4$H(`AM-;*BqT`eOtS!_clX#D15&vshcX5O~REu4BZwz!veM{jE_R*-N z&l)S_UpCC2ZVAdA92q=(L~s@iObAs1mI z-;R3cGKKQ>8++=IZYRhI_}Behjd`oQ)PL3Z?rWEt(6vHK^CPfUMM2)lK>AQ-Cj#2P zB)BIa@W#zX*Pi_&RmYXZvybI4Pg%b+#>1U)c}SF6)7?-fvbDUX$qo}B=~ zQ2Jk0x4KPXuO9#FP^ywW^F70uCI>Jp)Adw5p;-O(h~pU|m5c*1Z$fe>DJe(!Hdh(z z^T%Quwb>tQ_V5vlvS1CNLd)kip&SPvC=H)`XY)PGm~^plzp&KlHG_8=qloOQig2;=?2 zSjnsnzA5&)lxXY58gQN(^VK+r6|d$_!bu^x5ouGkL{_pm&Eo-iBTk~p#*hulcb((z zyfUiG1enK1Frh9Oqw;fY)o?x{H-h!P?EmzR&85k>?OU9uPBs|^sgq@ zc1^xUPN#g_pvx4NVZb);b)->60DvGlWWQKmvo=kw`JYGCGT)m^omvY5O%mAb@+B&J z-A7^-4`zwVVpu`*j?~{xI=rlmhC@j0k)W!Ei}eX<@$Jbs{_vA*83_qi7U>SwDh$jy z_p^?5i1%+%u((ON=hR!HA4Ul63jr%3pYQfP8l>%4pwApahZuNaCJ%5;`*77a^|pR# zqi=grA>4k_0zx`v%UuRbmO%31M8<+Jx%Lg`3wbX2Qrn}8=KDKd{~lxV7@$;5aT{-M zWo|`aNji$J1@d-MmIQ@PutNj6tj#-hys&ZW9tbCK=yhNqprjrF+z zFu@>;LV#9rZ1)?Z?#o@wii2}h^-_vWk_#0Us6AAbu6-;#i>)^@vmJt>q$CpvOO@ja zkPRDx-7I-S+o!2O32mm?<7qCL5Z~!Wv4uomi4`lFt&BEryDUC=K=uea#J1)V$iJ72|FaK0pGz~pe9Uo&*5G8^J4w&e0J)!Ry~ zAM`8(9g{vD6i?cgA^S4Ixs047p5Ok^+!qINHy6vT3+F_@(h2kmq>puAH!sU4i1_u5 zq(hU=4l%xv69knK=wL*d0@l^py0)k!e`_!{7wd7uPF?1&FfnfEoN#pNcIhG^A^sem zbqF@jV(4Z7Nj1s{+MU_GQ2*-s;rX7BCw=R(R#&TPOXIyocdp-#F#t0uB7SBJ1~uD4 zOAHNU$*M|WXNekhk4jD>2#qb1(?hd)Cd8NVF}RFH@Y`)wrh6E}X{$w~*SHn9nq|xA z@^7j2f-~2yU}=b#C#JJ&+SC`LUOoTT$*mW^qbAuM2KoK&t|O5?@068c^G0d0nzeo| z_|Y|QWE>3G53bJELl4;z&z{Ca=`#+xMno*}hCd54 zXwPh4Sy`QV$ew|ENp;`b^%aQaflD3a>P__^C!MjOMwak6C;+zhoa<@atNLj}(_d(c z!ye!O+DhpCAl#O0BgZtWxhWJNo|i)x?RO7Ts)&c6w6&COL?C~VsOVqQE0n0g#e~Pd zyH9YZWSfS(bL-a>Y=0+d9G}-Z(x_y&F(F^ks7uc<~+jqzXwg`&nJ)11UR-O6*O zDJ+08bGL00@LQf;Dh~+I^HUqok6=2)i90&`mz)k3!x~N?ywYqrY2I%3fW>bNJ<#l1 z0N@;}CYiw8^Kq4wCE*7?={XiAe<2Qc%o=O?Q3bihaU2aoRpg$3XQ0fNp9GBZv6=ZB zGwK}pu}^i4zJnAbkOtj6>&B({dJeaC@%!Q_d@>Zl7AZ2&nHNmaM)xku6kQHZb5hnTgi%hghH>_p<jf(p4EA3&c+|r zxsy_>&VfTrd}jf#O^vW(v9!%YFgEkh9x@Q*yA5I7Pbo6a4usQciyHy0chR;$`+7e@ ziK3(v1~?1gOl(7zM~@@2A@L?ac@>0~T7+5>OnN)9OJ5(#K_I%EnZ4NkqBtq#oSw>a zvfE#%GU^5_7^9~><-(9A5FIYWPu|YinGx%!qb`lJ5eiGI2|*-mZ1^t$4?PZ-@)~tB zMlh3?5Z5lG@8`0f`zxIJ;{Y{xc&t@KV#sIQtb@^aUvN$UGl0|l^{6CZB2elbkU>5y zMGu^}aB3eCfkbuzQyULLLj$PjM#*GC`8WfKOHFlI611x+)aWKK2~EhIHh4vX_Fq+I z<;Jw00?jePwX&S*ve$k8IB?$j*4(^{22{_P&+& zFl8tSh)4TKf=$cPl-5+{!CLR$+LG~8q|oN{Ebj&3&vPxVxE5*Gd?;cGq~{)Svu;+j zD5k)vGgc0`Y6zaNvRUwV!+H~;l0B{04mhU#YSi=yY3d}rGL;o^^6;32nV}r-+hr~# zXS*&(bkEO=H<$BYfg0r`{M;{p&25dWD(Hxf{Pl6zw8?|#?XlSVcO_%?uz(vx6~z;B zU=fTyAJQQZM{;^$zbA3A<+JYbfO0Bu*Hb$r$2C`nj85l3Bp>dB*@V+X{02Q?uy(Y4 zDVE3H!%mq?SQh1g3ZI_b&b`hJ&* z2p9-fh`+=1K2KDMD@PwfLVSWdBxE^Hlv!j7bj^*|xqQP^L#`(Wbv<`ehQlhs7-Lef ztt>+aH*gorF4X=qE$i1T>26=TAjh*@Ug#<;rey59VNm3lXZZz4HwA$*=vl&ocf zBkHM9FclR<&Mv!7;NB?#rmZFk1iwRQ_Sm znRKpa0fZn+*@4~Ih|coiIp*lTzF>h>qDqH?<$vyXwPpop3+Ur5ctF60l{A2v5%CGz zo-(xx8m(cTt_`rwDwO>}ebY~^SH%TUL&}Y|P07#iN-A}ABzYOz&SaPsri`k7(I-16 zAe@PYd+ifW{hl8PNy{9fo#8vc{0f__oH6oL-RRRP9V~A1TU3%l(|?r3zBn035}ud^ z97q)5WA7#oUXAZnR5GWLI*;iu2_(tCI^zI#!sAJmdDU0{hoo^XLTy_<5p1}IQQ=N_ zxs|D}+jb09Nlqg7Qi9}jIJmDF7itf)HVO z0l&d;E+JQe>6Q~I;Z3YwmyhUDr6^O=e^js`EhKk`>$sT5Hd~Leq>)ag&9Tfv>~e2w z9|(K;oc~C{_H3xU1jCZf;{pSQxyR5ckmGm%J>E;oKG#+SG^Udiw64>8($&w~fYL0< zmG{(G5Wi_Z;oKCu=jEuQe)~Y8SDyO^YgK_Orza>?*9d@cO#R~$^K0roq=$7UE0o7S zB!-NA>K?N7B2IJ{>(vc7Ktf)S5`}AiJ4YqWZvTf}hiwwMP+JJWbnodP8>6{a9xUmu z)8*9bfXrJ-7;RvSXOo4Ze@ztiAVF-J=<{+3iZ_9S`2^PkKH$9&wKEbn@#cv0_(w(B zh6K*~FI8%#|3~1%1=bHGEUGkv1@{%pnHQmk_O;q~;)oE97 zdDe9Ofk2vSrJ}FpCF2%6LKhG4h<~sWueCOBjU)Xu$JeGzA$O9LiN1RMU(L1NiBXs^ zO!J<}+wE#RM_1Aas4U8HUEKYld6hrhi6siuM1Eu6*nfi@4VaPU;RxI;3Qo2ufVG}3 zWD3x$*u_o#QTFr??24Z~yg80z?76XFGoTDDa(1?0FOxJvhyjx7got-h63Q|(s+mNR zf_jBm4F$uvx_1rp1};Sk*Y>Z^0{)hm2uSlT*whn9eOK2tE_kmWvxb*v9aXZ2N0;~4 zO1YM>GX>RUW!^}U=9-H9;YX840pFF-wp_g3`%vn&jlEVB!pSm53v1vX!YGeN?-RXnDODV`m)nW;_m|9joy=+!;cX7%j={ zFzvfkU3G&|yk>5P{T!BmrvV?mX-E)iXjtJ2)RI5nEYMWX*!aUR2}yrr&cw%VLS)A< zvu639!QA#OH6DX%_Lg;i{;u7AV=&=d1O*M%SK@q`H2v3KTdO=6^w@bLZkQf!%h+$G zTMG#3b3t86?_!skdPtZid7GGXqpo$~d8I{`v^|$YH==?y#K%^xIRV1UT18-iQT^=2 z7R0&5NyEuna{a>?!uvuw^8RcMb(@Cb9#Rs?<1O{SoYE@SjXA7(vfCX(^HmbU7`5f; zc`o>ZmLNnbG6k+`w&_1d??QQZ&e<0==HyjsK>abd6f~ z{^m8T6@mWAi$$jZ=6P{@timM(C>;KA@{ zN{hUCC|BYbv4Hc9odl;mkzY!A^ez*(O=GhU@N4ry%}nXwsfv?62BIM^A1G-BG`RRG zcBlM!ehl!U?KbZBNx^-C8}4&af&qe9-Ae~#2W zXpx1lEmX0*lk<`m2W;y3g2>df6|wHVAPr^Ba)j*VkXKZhKn z6rtvP5U=CxjEZ{7gulH#%>cbR4QNb+Dw7bHk3%>9G%5CV_-8!=Pz1&1w`{p)He?Ik zXrXS2%wvq40cA3`cm?DLUE63!8PSE=o?dMx!|Mxi8w=C6&?&#vbC^jXb7a=q6~d`K zs~|_c)oHn*AD|uyClU~*2Z28s`lJIQ)vW->Q&%7q*fy3PIpt*_y%*CUSGvL7qwn!b zv=&pi$YWU$Y(LS*Ju+;&8l`;)lyhJ>GN?5LfhqWJ53p z1A65Xu*WP+Y@O(Vxi)(h&|)Zx7c`D>s{3qYA%8eR=8L+9-j_}@6T>z;MJLhF7aS0Z z${gJhfnzNlHD^u}7A?ODF$A{r%z3Ep#IY9K_OSNSpy?=<7h%80R6Mnt=w8Hb>Sz|nb#ahx|l(3|>xoIL?noIvi2>j@-LhtAER>9hh4Yuj;m zv?P@w_xcEB9uLX6*W{gRsEn_S$fl2l!?3n~{E-$DCVtj`>jW+pvuRT3a(d0)q&VQp z!&2Scl`C^CK=DEwvvT|GYY6;Lvhg6@WIEj4uX*Ny`7+$Xo=8Z2_VL6bX6Gc9gQb-R zf%&qzI35H{uII4$!abvK$JB=31R(XldlQV9+^MG|j3-BmH&Ixzn~AK3YhdBwfpH3b zEc_ZU&|`(?K-ZWmQ#=ea`)cAKa*aYRmeVj?`}&2Cd>_!+EUk8)tzJ&gjCkE}n>P5) zG4Zi!hVv)c{CU{jncL#S+-lCP;hpMI6UdkuQ@|dS^v{$F&DRYkQ&wRD+=ueR#&*r1 zpUw;Pz*G8t`Q|0vgVA9GOCDOeya+U&J&Y)?KAbs%?AYd#nwN!}BR{y>v7~Vj_VEzs^vNrbxlAHn}o((c)9`YQtP5VJbOkjUr62S(jy9VYedQ zu#6XN^258@1wIJ4E+Am2MOp(@fl&`oCUOF4`R=pK$W0)C&BXTA>Z%jciz~T}5p7bX zR@N;TvSwzBa7}37KKIo?U$KZ^1FcP4KeYGNu8hs+l$1MtTetXj#Y49F->4HC;d-2r zh0XmTQAgq(U&$;qp0AM&Ij2?v+`*TUfYFqg{3xZtLQ#(B5|QK62BK>o zKA5pfM`d@=peb6pfZI;N+B5ZVhAWBiDdjz@yHrAdi}m8Wv54QrD9ySRj3RR(Cd+if z-9G!uq=$~WJ>R9houZV=pV%SsxG`uGPa|!7xZEucJr*|#*CM71;xRV;0@IF{dG(N} z`Ww<|yqi$l0SA@eT}gGgIN3|p*v}k@cXLpG*8%1B2Y7`6X1KX`@{qc+sq=m@R_8%) z8XuIE4X;R@)F|p<|04HLdVLbBlSwm~we=~1H&+A)8Cuo2R|gSxbFC?^eEDFgGfZ(q zvs6a6%C*>M%;uwo*U9Rv=4Ek%2bVccTg^7Tn+m57izGi@D&FPG?4B5YK-(u{_i$v7 z(04CiE9+yM8)}hOJmFK_Tu%J$U}_jlXWHEH^$U-ue}X&Uh{;^rABiI>r?#nkXMr%B z5vMVS;@QLwa?%V}C;{{4-qclA%l;IX(Sl@;Z?i^_o-k)}F+h;DxUB-!w&$psVw)a3 zmqdCcBg+~TxwSTy#7#Uz2WY$Q&-j>E?;&j`nhGJ7i8z7fes-e$5=8`q``-_xeW^k( zse#B<$aM`qRJ8Go7g^r5-mii(mmeONp!E3hL{$2QZ9jp}pY1A_laDmVo%#bo-o^Fp z#K$a+mqAGJ`-wl`l#86%mf7HN!2AnO>$lcx<8hA^2Dez?cfQ=-kpN1guHYwMm;{>r(hnflus;C_hP8b#`DvupnbQ``v>#$Pc>nC`Sd;;1>z zgTHLqmtB!8!PooWYo&J}BlDt-qi`jP9cu(Jg0KYZBCSXShj5y=4jqIqnxF zQ^UJS6#X4V9wqtAhTN}ct$%v&2{PSh5Y&2kNL(2|a&Oa(N=`Itu>tF>m4!Q@RZo?R zEFeT&0v4#|YDRT59&#-ZR2P7=*;{QH_RMV7M<`LO{k*v{e-aEiVGbuZE8hAA*uAVuLum4EV1CN0FIl~jWn*0Vi1Z@DH zy>i$L_Na`}iIk%k%)sKD+>?SF+RTumwlK~qlWejAwy?!gF~~%7+sO;?mEnW(aV~lK zZW)Y@d%x)otd1Crb8sJk-;@SoO9q&NMg+f@Dco)E0!cp@WI<(7?cgpyhbN+%!X)t6 zNxvGAKYK(=?sE@ajnn(A zCknrSR@%>jB(w=K_uzll!tNSh2FkyRr}SEialk6#r=Ur{>aDWnK!u#i7}WjM#@w{l zgyS}^fJ2>fBWc7$FaP|(@}uZ?o=eLz@(P1XkLQEW$IO}tCJG0(-T1bY>KycQ?qq)& zNM4W~45AMhPy>rV6f0OBb)qB(Z1F`^<=B}7?>PMl(QTXmaBzic-163M_l*E-Mt^*y z%Z@EZte69jID@6}PHgj_aCoKV#f{=U=+-wMf=dU=<{)y+t{BPN&Y#cUkI*Lbe1fwn z6~*h7f$k5r<$GlzxC?%%rjF`cuWT8h26-E`512gs7PTL;x96>|l#3oNi_#LJjn*)a zaj%A&BZ-erxPZBYh%_Eh%GDh!M5kKs-by+1LzQ`>6qyt*gIPDNhys^eGR$}FuNIuf zcX<2imEo?Mf@F!x)C5>Gp|`s#JYMU~bENu8U@OK3j zpp2`yykppF`)Gc$NQeU>v@BfaXa@wOGXka?+_k5rVzX^DYFm3Q#oZTCjzB0&h>-il zxggF@BdY^$W^B%EM2T7>&JAT0$UPU9#z2B({x4{`R^#=%+wJ(PPx@gfBLCV}87jB; zQ6YF31kDWKVi7Jxk`_|FF387hg?p8Z-}Z_9sr@)cLBO8UT%X7zn+Hq5Ohnm8W?WnR zJTdCRJMl~O#W)lL@Q#m;g_5@`2$%1e!(XoO2Wo9I>wwH0 zmTbR#G$6j0RrymCMk{eV_$%bedY(MC)n&aCL;e+`3NDr47HXq+iHE3mGyHa{Gvjt@ z&bhcUVIXissvsgU2% za(yAP=|5C18PRu_)M52~wnjaHpg<&3LM8-!#lQw-Y8DTTdYmxi$hd-XThjlO=j3c^ zF^ERPKFnVpY9G)lcju1K^ritF0Lox%A3=q$b8ty!Sv1DCvqaL*-hE-v&b9gyNGzZJ zxhW64&|{0}h#%VXUfItmRu6~$EtX%;{$-1WE8>fMD)5a{>Hb^-*Pp13a2kR~dc750 z=mq9?C3v_=Aiim+(a01D>oVLrX5g0YhaHdjV9F8!rkHK>M1C~(C)NpqgxCO8T~?{G_~pNJFb;}cy+>VJ;gmkbzMh{r&=njL zoQyxrfk&cGf9SEvfo6nVCR*!=`|#VI%SUAa?wkvTvo7+|&MPY68#bpGXKKl({!1Pw zuO(G$zcHa$56UEv%oQ7J?P|%3-fj#rLX6aeDkwUgs=Xc&0`;1~6PloLjmk5gvIKb2 z6Y-y!X~u+!_K~>9RC34=W(?M72SsZZ&Q0@3axnMB14{J+nCOpEi4YXgwoWrkFMGe&WKy9_-aihRN;Gn_4R(fFvweh7gqdTqWQ=R;@XU}Q53@R|Lf`; zwnPDtBwDsz^~$zw+qP}nwr$(CZQHiZ=`*wH`Gwr%jf{ws!?@_aLc2AxmF|%DhoiqE zRCv{%t<3dDspCE|(uS_X6BVs;lv9zvr_dDJTAyLQej;i!+7qYW-YhbcJ1HVJ{1(7o(sK}?A={z7_pE!WGGYt0nryf|b zK~Iu}3@K9;NnBQ))vj_&VyJ^cCqH3h8~gpXzA{)j-B7D?p1EgUCPn+n`^cCqYozx| zK{SY}vV`?`D9o(z_I_ds$3r$OW-vyvtJ9_0zQdyf^bVqSAnWp9{>2i4cVy9xJbknO zqdU*=9bmXlMoL#u@3wt3XTZy+lzeipU-5$=A6a3#4&!UEmOg^LAYgLgdh$lg88XNJ z6Y8g{Y0+idVnNjIRS8m_eBvLqyUO-+`7mgHjQm*4)5!Q4Myn1P2*rFElo4T}woAZc ziRng03OzIME3Pej_nDZ~K4Z={wK3ycMR)AI{t@js_Q+j9lE|Z*u5yudeVx8W68S`f zw%6vM_#k#A3%Kb#BPifxk4423b9Pi<$@qv)X%#((eIL|~YT)T0u_!jiN~gNz9Vls4 zc^q)KX8{Nts=kdv5^u`8Jr2V9o)+~-j2y)Ne4vBULT+G$Ny?MeX*~GNv^^0o0bgTK zPmCqi&MO{8{dBCFd|bk`6kfxPdDkDmhL#y0rsSs5@9FFs?v(+dY2lxv74L2e@xYt7 zVoTvbjH;6-B5U``^Ek=ZUhDcvK3kD-#O-}7YDCm^qf?{Gbt z`K#v3cBXPP3;tR>eOJa9jT<(;6zSY@K*drd+AN@-f4W&3(PgF+UbXzPj_nbR0Z@26n~C zDQygHYtFr`%S}@Y&m!g=@)xNJk;ycMN}hkZoRt^scqO*VHXb6TtJ1!sDQS49=@>`d zQ#H2+#9FT0MeVcp&wdd?NIj6>Ui@iHLpP{yLh<|Dl$m{CS$~YaKD#AhzM7QV|-G z7zapKD)*?Cr7tX71~s`E&+#`UO>&R6i>U!lDAlhCdj#jolP1q~C&UsO(oKz4bHI^r zK;W8pPG%-ZQe#9)Irlb7F(CIn+ymE24}nJx^GJT}@dGECdohd~GHLVEC@#SpX; zT0yHZ86s&9JQznKQ{1@dlZhnQ|CV;%C?ng}S09r3(m`={K20lf?)DL{#Z@)=T}?O5 z^c#jfitXiYOI=GJ`Wu`-yucJLw{E-b4m~NKI&HEyDnrq}LSGBfQND%H0{8`k&3KOp zmvJiJ+}ul!sP}AP(~Wv_2`1KoYX-NW^ecXEWc%T(K%9eva?7YKl3f~Sn9=g){M=6_ z5J^-?bg~Ds3AdGObRT>_R3`X_^PGy$5e67dfAbJnO&PBofDtg2!d%%i-u-Nt5qi-| z!XIJ#vif@DI?9Jm+vmJkM5ufS6~&ni@%{UvO+4A{G~y{3lSV^zikEE`noWNJzZ%cw zdGvdewOu1N7+B+ZIU&Ky^ve|)I=6_-mi(Die$SX7KnRc*LvuYz?L{*^d?P_Jjg=9A^!B}Ko%uyg5z{^TJsoO?g3`rK?c&?`2FIY- zSg0=}OOJ8eW0NgO&6L%p%ws{U1gu8aAG?c*<8$~TZWZb1U&s9~a936k+EazXd)5^a z11DN-nh*bskkndimuIvkb!j`S3Rvf~IlipJ5!_M8d5)Y1{+6}tgQ%$Vo)tr=E;56q z-a3pLOzD@nenTiB*boOAFne_}`;5Yqm-Zapkh$=xuH_~nSB1pIP$ORx4xz*lQ!Xmv zH)mHZ-sz}Z30czrR9SDjIqc`E=|gwp>kj3Usc-~{Pk@x+*XS`Ry-dR%y(!8V2YMJGuD_%n z4(8zof|8FS2~zkwGz<+J!~sdxaKs74Eai`5Qg>*FLF4_JBu%HCwN6P^&i$Zt^0^p@ zQDN2AaKSPB$Emv3GFh2~Q%hsbS2)a7`#{L7bB_>KI1Ld5h(^|mtABYU1=Z1EL~h(1 zDBc^r_SDp^ae5;=9o9D*O2!wtN~59lOjq6?70l$JFR5aZWm?hXeg;-}CZ%<@Ds*-6 zX0{V|XCoWS@~8|(_{qi_E2vX7kuCbIo*$M%Y|R>!x#3kj{n*V#aNH}ZRf611U$zYg ziIwu-PkX?-`(512jaInBuVQ5`889bGw{~@AdvcuJ?9yO$&fKJBeZFQ~LZ%Q=zA!<$ zJ6AbnZr~Jz(In1+EpT-~^n%O=Z%y$xJ(6URV=7r z9%m1#bEa|x2g4q}2)oHuL!E~tPCseBk1Jwu5CkEkdYMDIv{=+O9sgfSYxtE_J~b%P zpE4HO*Y4gxt;ZqX#v(Nu_SYkx?xIrUj!2JnPXbbyr~1O?XSC7CGQ@}_VzZl3(wHJ< z5LwXAfaNq`^jpRT;f(#s{)?jE)sk$JIC{IjAZ^g-5uGK$R>2_1au4MIPJU(}aqbM- z^4CNz`a_PAr9WTbMLTxDh5+EZ!bHv6{0!b_NC2kpoulz75+&LS`uUh5z$wL~GD6a` zE~(@3@i@jHdeVptt@#AUumF6lK7isqimN5V;jl`*a$1_$zGX=iw*|ldN?pJX5`LEw zUh6^xCe#aUs;u(mNBQG&{k z;vc~`IPC_g_d;PXQvU<6%u>)MiiLjowU*c5>c-o$HQ01NkLMyB$I=JVzl&|i9A(;^ zF@+C{rBJ;Z9?4Whj$}opUI@*Ixf`8nQAzvG@2$C*@ElNqnEUoiM5|K`Q%rKW-$$kXR^*Oe5u|QRdXA<_GMt&# zr&IpwMfl#wZH<6^`c2wibZRh1Tvv{rm|3@`P%5@J98cY@o6zah#4OIb;#WLH8#9>MU1#JE|AO~STzLT-zf{`++lN?c8*gWD zJ>rZjMjz;5`L8~G$H~hIP&G-fm@J-{+Vh7IYbORmwrKU^8)OkLjFHK3kexh!XVMvZ zVVz`_8C-@d2XC)=3<49u^yo@f=eo2T#FZZZB#+o^c#KyS;2p;HJqwx=w777Mz;w}_ z#W^mI@s(}e2JGBGjA1T26#iCXC0JR+do6=9MJ$?#;1sp-n6*J+g*rB@oC>Xd2f*cm zW(1&_{do5Ux#yaZ8M?wKRRUPF*R4`zNEdF^&j@Gb0{ahdM0l*rqZP?j53?Y|Yybm> zOb69Ce6x=EW9w`2 zp|%{%Tl^{1K#d z#-|m*a=z?BJ*L{ARuT{D?Mk3cuuU8I**j+{@^p$Mqp8fV>QO4wa`%S?mQ+dZ z1cYpBy&r7DuO0X842O zWh+ERXr*HYVMzUVQY=#ki%`W6@jN#clNyFWoO%^eOSghe-;Z}sH)x! zO=LvP6*3;;&aGZS)zFCR1){n*CCo&ur|ssl30noi4FgLm;_J%oze>C9kw87bMn=|x z{;ZXE%SN+X_TW2xK!*f z%^*;>7_btu&Y-{WIXD42-Zx5(hd&U3J3696&qy2hsM>zw0_M^^rj0PBJi#KFDXwog zCpVLFI4U`dT5o#GKC#zoGP=g_6;SmX?oH1nhO{Lbt~~^d--NpPW>H~WSZUI@roi+i zX?x2-@1Fym+<3q1XZJUvbipOxiZ49Xm=K6zqy_tId@ec*Dt>@a6QNa7yZ3-Q?)%A6 zOT<x09z+ayevUG3-99cYp z?#9u0G~58`ax?7}W2d43p6m{nZ^-o$5H~=mC-1H>gquQ@osPQmav)obhx>YoGaGaA>o7m?h0Q8(>nG6w5mm&jy`G_>LB3OS54AEf zeA>{9m7^sf7r~bt&Qt46zPVrJf1iM7;GutKhv0#24Mm)cb@-bEAv-Fn@UVy&_Hh2SH)9E`zdfO+P=#AfOG*?I4e-oi-?e7h^or>v7f#*f(7Jm-Y>d7}*3z-z2VD z?j(!BF|_^d{4a+v;UGpn*PB~um4O8GL6@&t^Qku`L8@&CUxn;#%fT-gaKTrfF&aG< z{sU}@S>sGZE)(EvvqBA;XUS-``nG}NA{6DC(d1~y5AGj^lVXh!zT-=5n5TY8^Do6q zgvAW&WZx()@^VF!f!wotbE}kGsP7#doWKWCwKa|?x^E4De2(~N_6oypm)<(!j<5Ap zkzuQZcGoJ;r9n4hCCOIcPn~rt6hNxSHNBtY1AjIkHKRDA(x-IC2@hsml2)n_XU*rnfd21t_z8NAs?J zdNI*vCwnSghh&kalv|(oO;h+yfw?1*icvK^9W zjgT#kY)v52k7+TM*w=&4)k2jufwJtM?=uP)*4|vYfTV1d)fJ}6EA22Nl>NXcK?fd~ zDiez-j&&In`rWUBlOa%yz;%RtBtP2*cF`C5->9OBAI7-9{#gi_`V@ROJz)P@O!7t| zetzEr0foB23KphrO@rfTC3wrn(h|^Cx8Kf>!^sFDM%9Xvzcmbm`UC8JjNo%l{CWbB zS)94rqs<@sO0==-9v)IU4XVRElU~HY>LkbINclD&sKG1U9`=T0_@q(3bn^B;NRX9( zz`P+cqDP$c-1U7{KtlnA0Lc11)(7`>>Wd?xE#F9-SY6QDEQYFr3QY z%A?qSJ~$K&1aP5{0g#*Nh)W2E$64@@!_OiQ_u~_BBLIUFM4_r{o(_R!SiA#<$~DE`ac_n zf6KLSmMo!#@8bmB`Th$_`pqzu)E)U~tpl zZ*keD0y;^B`T)ya8Jd^NL_dexQcmhCN+rV;#a3ploY{zS^>`BAZ;$bBUyHrV9j&Zv zoAT<}gly|sjTQ5%3&l3$A(B37k`;kGl>$?)+yl~DMyw2UFEse{0+qy|7gA2d_IPpm z(ebGYxcf2rx=+W?v@gWpHF5Cc`v)6}^QxdsDMBVjS>YAtcX^!wwPS0|6aU<*G1CO> z$Z+cPP*6_lji-BGo$4v z$!pMo$4utG+5AVg4}QUsmF#*)NHA~_%mSdqdQR%BoB~4)Z0y-t*$ChRI&?()es-~f zAQ1OpaVv{p{(IM?ad6+pr3q>g8jb0?6sTvjAV=8}5>-{qVbSKLCq8pS8$|cx0QLZ) z?mk}L{{cGlQ{#R9ap?=U#gu|o6^S{}&!^{}7Xp{oy;w%Yt7BQ8Ge^&ubcJS#E^=%k zi3=vciSz}HYri>#L7B(9)DNPXMPCVR%K)J~1I6fo>aG`pPi8DnbM>zPw){}eCe3O7EivsH8Yu%X zxwO-CxNqMP?)j9s9+4ILFzgwLo_S101l7!@bI}34`ew{URf7k)aBY6G0?NI&sdWJO zmFej_JV#ysJJp%xmxk6pdty}uJow$&bRM#xYLw-2*2#&pYcj3^xFX^;W!qu8-I61A zuKA4*rHNterLVDqPaJ_QRjPZJ7&EATad$zZaWxVhu24eS2+T+?OY21D8F?51@gi4+ zF0j-sp;HU*Qo6)^pir4kU{kDvb4TiD(6Bn6?!ZPni271?t2kkL%6DK-KE*nVr=9S? zE=Vgik8iB+rszQXeZAtG|7NZ1DK?^>%)aXpr2v1N5iGRh`f2Lq!}shHbXswLKj5>- zG*zRVJai65>%`58Wv{N{Usb?%m?3cMA5Tp{+;ffA}A(*3E1z?H?Z|@?x@Co6>4srb4d$XWmF!ep0nV z@o35ISEK?_huWu_?w8Udn#5fRo1>!RK1!)Y%WZWI+Tb#8#)hPL*fq3{{M^rQ^hAdx z!KzI02Vt^=b>Ax_CR!o8!|{i7-s^KxUx?=)))9lWiz^JovUN^XMh5o}qvm@#7j@7N-hv2Z~(uG#=|0>*zwtII+! zpLEhhO4{UcaP8w9YbS)C=|);KH#~yxu6Ylmgi!f|r@Ka&IM9~y(>f)zb$}e_HY|{U z5y~#G3PxPD3{Gc}sA^5sNGkzdpUD4PsVOSxk4#WEFTkaHBiCBu|(lhgy{>XZCGl zh{5upCD`DpHwmkMPzON=x3bU?k!7*vr5B579G0K!-UUk5hCxN&@Ob6THRJbGbG0bg zG0O;rKMaz*`m#Y0hYtl-#ik>D$kHWJ3N`h{ixxm>+&OYrUEO3|8UAftJ;f9 zYgHxY#kU`+JxtXp*oeu$VI61JMZeG`(;Hd%PdV6>sJqInEAaFog5zv_O zn-w$>tAbU!8xz3l7Iuk9>hU%)Jqk>(N?Ls8FS|ypM;1b!;~o*;_ymJxFQ{uMYO>S+ z`o;k#6;#=Ha%T^p^KO;6$O=ypUHDzJcAT&fKZ+3bxGgt`CTh2LEwjY&<~;@SXz&3T ze3LgKXKoYU67rRjywDx)hDm9YKrPfZLTxfK?D)!_sug>vSE$eNEl3ppp?i<3`AMsg zcftoz*GAxTuSjg@mbvUZCwX#N{KG-sg7s^yMtwSMb;nK@9&Tm?*is+Ov&gOF7$GW) zclp!K-2LAJK&8L+;3%Aa#yXLG*(n+>0FC6naG6%R{QDR{>0xo=x$D^En!Rzdt+~|@ zLN}QLO9EHt3>wi?*G|Y(6UpoWt|88ZB3HcCj|6?AC#s8UDJCmas^m zYZ*}%3@PW2Exw^J%(ZZ}0`bQ4T=74t;ON>RYor}Vp=^=Ce&B{wC`m7eEaGn&K~HLko*<=u;uA z_Uw2KBvA2Z4g%pS334MOZvo9YSq~94WCo-t$`>ZV*X`?j+JQREEcQM`aFJSOkjr_0*pbTiHTPK#saqzRyqoAtMO^fS52+EH1pVN59+O% zttMA=9b$t93b59Wczg!`rS;+=_L^AvLE3R=yqZ~&11Tgqi_F@Em8=*sEda)HLt&ra zh@P&rFBrbmeK4|ceQP;Z6y5XgZ$c1Ds9{Jbu*@mX{$dLTuTyJ6zw6P|Xy>UZ@*Qn= zd-IcM_JkWl)89lwIMBrsxAVm-al&YuI~-PfnsJU6E*rcKyO8(F1kAoK{jL2SsT#gf zDz9AT#)Wk0a#Po`uXVN$Xl7AH({!_jFlaQZqmtP3_;6FSj+XiG%(GJdbJXfU<<`;A zB68jL-&ZqGCX8gSV+tShPcNdP6#t@`?bcwZc)YbhlGpjn^y8s6E6{`dx4Ok6l&S^i z9K{O840I5G0D94|Hs@5#*Y7{*!?>hj@pC!1q_%^eCTx2Sw(HWL)p9j=Y07j~(_w(7 zx^s68GI4}ZIwu!HE5pjq)#7-U={q=;YWfny+}@}gGZi0p@-5L|0)%V80vK09-@0k( zkMG1*3$k`4vyYLq>cfqeDpG;1a**|sY?u)~_iXTU z`a8_0X5w4g`q}h-o{es>4Sto{Y2oP~2n`Qr4j4+o&oM*|O9xyR&(2Ve=v!NWG<=AN zwFVdA-IIx@L;Pubbj~C>(DP=$X`c{&hE=4*hBRP&q|EOS=GpB{>gb%67OKA>t)h@BeX|*QKG7gw&*VjZC zt!}4VqEz+!p&L4`^RRi+YNz&{@^J$Gx^4l4-sAFg0H+PzBH{7zY|T6_s6E{HAIXZ| zl@c-iPqww*mO%8PIP?)aS4foCENKUbN&mHOYO%^g9Jc4(KV16UP_hnuG3ZZMe0>>| z9ojr3=76m&yaNapwwp6dVPV2pJ_#G1pRp@zv3Y;tOc(xnwQoxJd%MY4?nZHdP%*!Q+;Kb;Dg4{6qZP=8|!)siE*81eKG#PxgX75 zz9#Z_M^=-Kz;vz4Z*u1}fiZNFj&js5 zZp5phb`^#n{^30xQN-EP;Z}aokquwgA9g5d70IVyw4tOyANe9iks2uuaDua*w&#ug z3g2l(Vj1qkmsq4cBV{gJ?6)~*Uhj>o$l5{vQu~7L)R7{_@^I(@`D=9_E6cK{1X_>i zrSUnfC2ci!K8UHb|GZzm9sM|*v=V`H%@nj!3`^?7fr|$#gA{ zQ2t@7fddr7KkAzA6Lv%i)HqWv31RS?+QVU!iKj8zDFZAye8n@QlaZVSVJbqf_a!=^uvC|hD+m1CE>k2v#29Qk`F>jDRHyuD;)%9!o8P2-rqb!TJ zj|!{L?xHqhg1w#c3fPtFYmIu zxjbx_SXxXbUjYxzt8b^nCEFaOqu!JAP!O5&uGc3KO15(XL3>~D{b>=$?>VzHGde`nn?Ph;XJQcI(Sd1nj_OccK}+fvt@?#51ZAW-sMcvWI&O1N5Dkc$ADf{nEf@ zP9=iO`}SZQiVVCasu5feF+pEmv3R11OCl@PG2T}-PDAjaWh}3g0q?z4`2h>Nply1> zZT#ZcJZ6j1%Q>+cNMhSemvV`6rOOa;>y92A>1WqamO+sVSre$v_U5c0yQJDVbh*vF zQSTQ|0;M6%zlUWfJpteWjuuk++`Xa29?G}F*9L;!C(+C#DdzqSlmsajdm+Kkx(iP% zj%VP|MgyI&tNibNdZ3iPjC?c;1H_Cnq}x)?vKOi!8TZ28=$T~iOtcu~U?g&TP3(43 z9B{oIkhavyd{IGrH9)+Pl=WcGaVHLewBm+=G5(KUphafsc0igxE1iOrew9&Rmc7>Y z%oPP!2WZ_yr@`EYk1rOOYLK^3^#E0K{dRJiSW#gVHZf#AH17cbk5SYCKn3!DM=he3 z^C4mkP8j$sThu87(2M)9DvCMvvBj!Nb*r9~Ac=Ce2}J5nPN*lD$VA;4d8W$7ftcEg z_MoHO50`ufkdA+V-W4f~&xUoI601pTnw#C)qWQ=aPB4`g**n6z>BkCln?ZJ$eO`8+ zJ3bUI4vKQO($epa*!u-m*G2hTC^@CI+;9!N@T`_QIwVgAA$3F z_>iTM7h3Y5=25r&@eDvJ3CN1CXdh)KgP$2!gCtzbd;Ak*Z%7e3wYdQAEW9P`A(TV7 zL`u;AZpKnksd5r1&pi2B09S`fo%V6-nxOjqm?zv;PVzRhLA;Pd$bkhcNc##;{d2=@ zGUVgqfS^4RisrRkpxg~*v`=u8r8?$`hK$WaVRstBmJmu^jzPm)<3v=9_{uwhxoW+o z-Jya`ySE~M#Yj=buASiW8Tfg^L#F4I6RtQNs(e|X=p%helzp(*Z45jw*yArI>ROm4 z;Ee3EurOU2pW>9$T>1#ZQo&`}NUTZ6^;_#W#45-H4V;2iqFNhntP49Dh#smtD0^BZ zZAB*lN}l^Z@lr^H#WD*!RU?!Ij_%Mj89*2HtH)x8q{i`>U7R$H$;G%rFzFDb;9lfY7W;OXK5g2}VHU`3$1u zaO>HP#xg=_vJe`ko5r9~I)!eZfAkxxz|}0Umr6n1rdIFgAmxO@7srz!NcX`rp zM;G?v&W0S`nGunq6EML2nlP-=N5lFRo{_}l0j}D~7H5F@VJ1`7Jw~Ha=lbkl=#*cA z^tK9zE;ay>yrAIiLn*)EA*ji*ku7;~Pr3WTIRI<*KTZ*>@<)mDmZEQ_e{DN=(>YsL ziq)q*vmSNH?wVDrHzJtFaa+BQqQ~lTE_wbl@nJ|hW{%Hm6I%T%J`H}fUZL9g`+ink z(5mB2(%ZbrZ}M)Q2x9S8)Yd)1p+B*o^|psCQ&eTNXVRToSLbs~?3~Do@*-uKa0n4$ z06@OPO;b)T=7=xx>B0R%6dC@AkTM=zh1uiF?{EkBIK!pBwdc5S(ps<1E?DFKmT0A7O>iz+-w&t3VZPY9}AcZLX5KB2v33tac6RCuLwpN`d zHpG_Ia^q*CYO$B=TKATP%DJC7G5)Jb%YRHAv!bV$orQH4yQH4cYWoO{9#_W?(qDd~ zW9NGN1FBQ>A`L|Zgc3?Gl{HM1&Cam{ZUK9KGf>E*hZ`E zs)6}QxRKw#Lu)*i%NYo25b0~kvd=qDcSn0X#8G7#9Sx8_S#KZih{H9lU8C+jDgD6h zbU?}I0o2u0@1BEJ9#&oH6AYwGK_n4qR9PGcMQ7`roGv--gN1AYeH$pwhbdM>zZ$}K zO4Odl5Zm1*wF;e8Wc6YHayHUEVW`FUy$Paj?%B~^s~=*g!ToI<@&-2s7xY=0=`GTii0@H+- z{{h{5Fe&#*=dI{1q=`I}G-NW&;}903g3zV`;=$|g=-S2%QodH=bO;zjT#%Nb#1Am) ztyDWQ^$ML7iz+-y!p*vi`?lZNrLgPQ8b}>r3)LjKbyH|5a&58ZN6C_<9C;sY0@aN(t`uLQr}huKp?iTJfBc8%|SR0 z{aVHyGc^#DsSRE5r@>Z2)DYA3Z^TCtbp)Ew`FDYR0s4xhm~7Z0$s6t1}tfdf2hwkg5AL4D_#K->%9HJI`n7WzAF6qnUFglnRbT^Ze8eN!hr9} zvhGq5NV85%(+9MVDBP7__^UwrG^!S}{QlHisHGS*RsIGz;RVs;>OX*$w)&_lGtD+w zM$$N`d3sqfK3)QR%iN<->y^OQmfhl+3 zcw)P&zWNmyx=(yCV5ltD{tf^qpSrkmH~;Qar9w-SHClh` z^eryd#B%h>C3VjgzHedg1Y=a(>+(%lQ<~b;kSFib1W~h?choBb2DRQ0^A#a=(DYYw zYm)-|N;Vm*oxNNClIoa690{U?rOaD&<8T?0+?e||g*))L(;oX4{$^p+7(8W02&n_( zWP@ZQ?BCHcat@DL18$9Y?&IRHvNAnwQVOHevEHiEZirFn zK%cyVl=gpz!p$5nqT62coH@Gs4(ex)bkKW^C7V_Jg%NYqOFJvL2(K2fSd|Wg{tdyjecnQ@s`V$iz3;0- z_a1W5g|_b}FPjR_nk5)8SQc+=d(JEQ(SR#;uTZ_)QVl$-)W5aK!VyCRRvA>@e1Pr) z26jm$PN9co=5kET7(mueZeMEf=>v3+FZ6II{^#`i8Q6!mLK4u50q*vZOVxwvUEsX| zb%WEsmgyN`#luHE=)`%SZ=I+*+Rat0+`cEXME$LDSIn>A7PZ8zLzQwp$(brh*o#D} zH{+aZ>|R$5pi!J9$_zC9%}8x5@`wxB4(+E6UB+NcpbDo zka$1NZp?x*7q~ODc+@x-WvGbqesZM5+n#6(7}OU{E{^PPD87NF!Gg7TrxGk;lRMlG zEkrb{C78WO&-!5~5E`*wgZ5Td%IN;ZwLmi#fHPxk{nnMoKS^*%Bq6J!R!A+V`W)n} zE~Zo1TV?Rh7)Li-3iI&rqRO=G@CDgBH~dRGE^wb%27o-wH~)FF{RI6@}mh-a1DrfpHgaIUf44UeGS z*T`Bo`pj!D1-My#qNQ!ptC^sVM=CNM=Co8sj2mQyC+wXdf064jK1ErDIX%E=wyoFU zETB=>M4gS2rhS=-H3nF!&x**W%r=VZHB0?fdV>un!TPfSrfLPO@shYrcPU^ezExlc zDfpwksXIeYFAgz-pc;}nsBR_P@4wltt@%`Ke|Wb@FT%2RNDt6eQS<$$Z6LSsX1)|b zF-2vXWw;>ZS71FedK$1ZO*6U!hY7hP7pQR7)x;cD3AePcv4Q^@^o*;~FjK0hQYf|K zJ|8YPrc?1xp09od)i#=YC*3PWBy_bGl8UdOS}>9{31NwNTq#W$hN~8%i0wP?fcA`b z^IK?~A)zaVZoj z?!n!P6)Wyg+}+)^h2pNoU5dNAySux~2|t;$XYJ0NJ?qZ>W9~U?5fVZY2rDn%lF#>f z=qXprILStcBggi9(Gj67X0FsB_DaDqdrKk01eQTAKf;A>nb{mcxG#Z;lHW{_EL3#u zETbeMgV**YTpL}tM9eYsW|BIltLNVORH4jx3zOA0J7mM=O`S-w;h_{V5l$qVu3&FI zM)suNdy{=4Sj4OUHZLrpKZ`L0`isds&~XMOi&9@>02C<+zO2%@#3|Pqs^dLm>w6CN z;IKPvh>#7PGO86zXF|6R;d>GSC;}Q$yy!B7EWka}&UMWWUAT%XgVx*T;r%!5omy5* zZ`MCGzBm_FuIjfW6_ggT;_X|yC3|_4( zZ4#V0~*@Yum9;nyT6K ztCf>p3H|Dr(81<<1?u}TME#L(MXdH})wlp*LA(ljKISMUnU#rq{|0K1&vRF|^otR* z=a~$~A@#_SHOVo)c!x%S(ys0B%o;34NMhD+R0VclS7@G}&&N=%K6>P2#q+I~R7*nC zi&&~<$!SqnMl^@-TXe-t?47{e`ykA`#;}ozfh=&VOh`1#AJXDEO>WGuK%9Q0{xq)OGax z{R7eEChfLT@zpG%9&51s&0;%Z_nQxn(*1TOxw=xzYW>rjrn8xL%c3V#7zSfZMw40D7(;=ao;}aFUdg6K_8QeBAb`8%qLbMr$PvIQ|#nTpd@cVQVTi^;#jcK zw^kG1>MC+q5OkTaWqfLtEtE3OmA6ezLokSg{UAT}XhET>l8Qf7Ne3#vx=wc~zmRqF zbZ_J!Ji_CQTLu>h8{GK;m4+E@OZ5nCqKxJ%2m8`{DRUp?tvi(G@qy}!na#?tM%gGL z`9gwK#DQ^jP88iM`j=o6Y=-7sq-sk<*9@rlm-A9qr@+ zX)beCKc;fN3~G`ujn~hL@THMVI|g+QyA$Kf-qCjY=pG^a#}+0*S7CGVa~^3mnbcJR zMvXE1>0wpA*D2xN<#krm@xg$hLIir68aG^D=DiBs@NcP!h)1e3lX`E%D~Vq!RYQ4@1}LtA2cQFC2;Lt#UGD+5C~US2pmds{W}i8p^Rr5N?NU1#BinA|O8(p+w;dQ>-YG_~VMMS^0&M*c$mC9(O~`1y5? z5HZ;Z=9SPgW+c{!MYyXKvQt3Gg3B9zPCSw)TcjaQ5mkmVlV!CDI>THOsA7L z4<>-*8lfF_Kb+-CSO~(J9M$H%bymugpparT{Z%-}c_vB!Pb5#lCj<|)iTx9Kn~9K1 z#g1K3iF7>7hozs1qYaHk~lt{DkY?n*ErT>=BTi-jRy zn#pg-MBt!sqS;CYj8<@3!DIT%ZkyAK9w0oQhUTHG9jwUil2?0slBI-c zAK~n8;WGFm=5GR;lhsZ#^(TjAj_l`oRL^Aw&B}BXd9X^>NRy z9r>R$G3XV>E?e8ms?Oi~bU&&u3e-&QE=Ve1xz8CaCMI9rZ9!Uz`cPd5NPDSO6?m+& zDD&fmyQ%hIq*#b?y&hJ8VR5INmvSuO>U`r0+LD}sTiBs&$G?VbioFca75V0pG41s$ zbHko6M^86b&xDA3#HdtnsN1RJi4&JTLUVT8F;x|b;GNx8@P_(U!Upx0e*)=fu;y{f z7378Q2FJyBfUp*-RkYm(wu|)9$pVU4pRWqBRS5c)?}=b~{~8f%|C(0__g!z}_x9J0 z3plfEN?|LG8C$}2nhu*LLB99)xRK8y!Zt@#dU`3!WL5r0|1&_wr%- z=a&x~2iw2?@=<*Ir;F13_di`CG9NgSdDw3NbM+slHOk^r>}V37VS-*Bukp}}zwn?) zr>zb(ox|JR9y)N1E7jGZU-!)?^Ekz~RL30Mo?t95Z)>`km~(NftIwNq^YB&7wounO z>fFwguU225t@S0>a%+9$s`Mnsd6cX3@#Q^wS$;{41s=!BJFk7P`MC}q>B5Z%pJ2b& zZioK{(;)V$_hz@jEIq+_D}_sE`HSkTPPiRQ^C#U&xus0YqL;#;<4iiV?@A6F8moJ% zY$v?YP0!rOwe!_MbZT_-SHlEpK`${Zyuy?5EtMY2p5>2a6HDVawwwgC1`LEHUy3WC zmB)u)o6KmEUjUmCmoF}rQWK454Fjc>yOX8W&!x}hxiR$5#iOWCGZ@`W!= z+1%qJ=WhN0N7!Hm6$S8TBn9(CgYbu*QISKJgJi(=0YIiuenDe+cXZ1;6FYbdH)x=k zGJ%N42Z9u4j(zcI{{_Y%BrZ4(?Hg!BX)In4uQ(RBljhg>sQ_uzXSB=_Y9`@DzD06XwlHTL=29}mu4P6Dr z!OZIMXeu>Ygiplg?L2){+LIj;FYT#tTU@$|99LSC? zz|UU4!&)X4{DL7xAR;_yoe7qIInb_-SL@WASIg<%YF0aS-PDH#2}j=6etTlmuB&CD zgE;NpIbu#dBeVU3oRmZ9r5xRIRbpqvl~$$<=u4{l}epEa0b~q#moj z)~mD9M;lYUT>{cLc$INYk<4d5*o6dJnRo*iKVG)Tr?@roH{Afv3nFtKm`_gOJoFE8 z$-YLx6M@SHlR})uy23&tQeA(CTbNj~vsR{>UgtQwpO#rma^tBtqV?5QgKDjY^W0ZG zlk5e){2^NHai1nQ?{Y8K;=i@iSHXB1ZtT4CP`&!Fo*qGKzlgJ)d%Db-sU%_s2A^el zY>}MWfL^&LW@fF?;Os_~^*XiUd;2>MSboRBe?<@mEjRd4njWdYB7gf1u1h!<2tzVF zr?OElfsr+>)lOI$R~`}0v32TH2+fQ(06Imd?4;}@!1v_3Dq}x=d@RgkccVMHoJ)U9 z_SnoNf{g5pW#~0;7>fxDA6LUq=R9RVcbYD$v(Q4LeUj3&D z4>ea`ZjM>XKSCyDchKPNjh{Fl0EIeF0wKjG{w&=^C=Nm_a)%KEt^zU!VSo=}8oUG! z$G@&e#U0Jb|6Y?0h7|`SM^Xr6`m=FS+nNSw0trtZoja5+;~Zl{ADK@iMEDdl8Usti zo(Q&#J{@U?wJ2x;PzELav;L^I_f0q~PbUufq0tPNDl#aHn?<(^(^b;A|qP`jvi!Lx)E zywl$ly?UO3>zn3jebT$BKVv-e2MB`57Q}xZi;FkXAO-pjE3y7v#rg`@$czXO5wpFv<}=J?kI;h*EElE1;O&_?+`#!;l_ z+rx8G=lL5DT)wDuqoekvXJ<`W#bt|OGWV^c;gkkD?L>(iJ3EAvlZ#SLA}KCzJcVS} z6doR+Gb)<-3Uo#vwQGjYboz^G7t+m`D_bZ`F2}3Wd_Cr#8}E4vkeb(Gj(^P~OLMKS zwXk=t*tR5pt?jWi9iqKAd{CIIk(J`M9kIwXJ}l;H>!eE1vM=BNX+B%+;Dy>*8;u3i ze~^bu2OL8sIMUTQf4F2x=qy}z5J;-rPfq1Ve9l+6Ly^*cSh$jYY(KGzolm^JcV5E# zsl%;8r)uF?n|b}x#DUHBJhJxV;>CqUs-c2dNl7_*(ncZCbJ8<`c-y^yqEZ7YujDmZ z(zx4^T^?O>+s-}UqZ(8J0&N^4+>)Mpo59ZlhPr_AYCx86w*W4D82UCal$!YVlL?X1 zjR|E2scAS>G>Ag>yDKU_&tJVe=GM@`5!0t~L7*u>g5)ME_A&B@Uki3zYh%=%-L0%r zhfkFkpN)~~V&ug(@@#4Y+-@RlZVfCZ^+PfBsBKWN5E&9b#?{5 zMs9NV;j2Mv`gM?m`5(YJ%a{@Lx@`X193#Lud!NFN+TKd6ox2&6KV$>9A+B?=e)GfE z67L~5$85+4z2l5PUKUbDoJYu7kdqXVXZ!;gASM`q3Xl`<%q02P|IW212zg%i(8{a~F->3?J-N&Y z1%}a!zrtu4FpQf0XJPa$r{20u1OUURM;a3_jKZepOFeKK{zDk$a`m__v|C+-9_Oij z3!}5uI}e0Jjihg3w6C`nF|<-p78+V9b=CLa{0$rw5aYi+mLlDdl{-J%;120dpEM9l&tFZ;qy@fo?z_R{je1x- z_f6q1i zg&?K=`c~i^TmD;tgCLKiMx&2W-(Bn7|5-uT5?C7T0N)zx%>mZQ*c2+(zc{&8}90PEUD{$$vv(m_}E!bc~ zp<3ko+o&qDWQ%V?;nUW?xVZ2GozgvdN8DSzBAG8mLbcE9`D8UySXXOHopv?x zWmol&^u08BJTdw5J#MjOH(FebiF($rI6m-&FE9Rznk|ls5p*Z$N%7^~c%gdXI&|!H z`d~D=yMD|P&K$CpzkU_>omb->EQXh_FWWPE{$@E*E8Mlnr$q1TIrc|_9st%QJSJo7 zcQ>3K&xSM9Q5|JU+J{Uo&6d5o#Sc8?_vg>fp^ZW3hT1godN0(pgtC)dHcuKuUQ!+@ zf#*s~uM)JhJZZ>5buhSsSVcefHYE!ac3xYSMHpYLd(=lJA4%Y(S3_MK9(f6o7wKdc z&snh~sg(78%@>CPZoEoNdezgkOMf{G{E&VacampICvsMrK_s z6c4#;Q$TZ3P~OOft)w<$I)BEv?>$MhRoy zQu(JX*qzVs#m>K}Ovx4dnxN1Fpc~BkwY)pjY{EI!?82LqIARSll1c>c;UlS5?WFE$ zPa)Zxcbt6^MnCx^XSPh}l4JK|sN~bo23-pnw3wO;_+>-bm$NAS5*9%H_fv?btvO};d(C0a z&z?F_>*4t46>+v|kv4Re!;G4}h`nnHuRA#O1xTx(e>|T$+_bXKzqSm$+$!C)*3G}J z7`{GoKJxV)y=GLt`i;CGlRxss&cBivyatg!wlccJNB>Up5v#ysD<#hsQ3ueuzsRE)3Rx zolSCV+P>dTQqR|z&mk&#H(jXmF;&kH<|nN;8o!7)EvO0zlJ_kh2qFI(EiTAGL0p|P zh#3eT7(R$J;%C*O08oBJ8gL|#X;QsHkw5{&PF8N`*pJik1`0LkVE2xNG;TD=S?5S}A2Z*_3%wJm&je>$WvB#W9UCQw4}TUEiRexAz=RWMwL zWLCtV?+%Jh_#@VUIJ0(e0?97>raB2_$NG?G!g+L9FmvEKud!}d|!-Fk8?nc zc!Rr>xRk8-B}qz8g&GGXCZvcTs|1k8ne&4&c9i*(H8Xd1jyqdf?@{1sl)n$cq892% z3g%sWafpCS_dV{HpLZ{&L-OY}MeXw!E*xjok^0tEOS73R%~Jx>gC6}|agVhaWmK(M z?lXm5G$$L>*48;c*cr+ijUitFGoSqYsPARkBqKdU;#a4O!MCArut)k(iOi_-x3PCO zX7DE+maGwYudUL~zaN|-D$T!Y8NIs2Jf998d1)KIx+OoKYTmS#RlWKRy&!*lx|tYP|~KlNd7 ztX|^Y!w#Xew6NtDzJ^naQ>nu-z^y?cN@%rsSyeN>$aMO6ls&EKAZb?eP;_QDr=$6( z5gBeqZg%u!w~KXgRdn|}heJvxJ{ynRhsnKeLDa*<342i_$*^#rcvvBdc-MutV z*7L%ia7Fx7n!y8Gjs2+cox^h?zpRROWUn$b<0t5_0Mq4>+3S`iDsg!15u{p3^BE2QrDJwp8vvMQtU}9tDE9%e` zz)Jgkz-A4dNs-Ssj)-m80+By}11()=Tp)TJGSXiiE*mDsaq9ew*!4KdjJ2D65KQtTsf|pBnud9^bZALqcEa&v;-&rN3dY#K&T=0a+rcm4M^k z57#r6@*{j*?h9ky!&GZX1w~~Nqy0`XUrz(aVO0C5D&KKt%60EpOx26Ot*V#rirq{) zTKY9bmr*Je%Wg4qH~WaHk?^l#<}8gFlf7QObXwPn4QAKs4LbxHUUV_{9dsS)$_!Iu+cFq;MFu;M#bvEqigRrnC}Aa z5O@Oaq<#S(kG{)jBW$#5^v-~o#t7Aws>|z2KKOMmDBw<(^6^y6!Z*cJT^vw+_W75oonmVC-}0-rp8a2zqFD$zKH`*gG{?EzBqp6Yr&HVcez z$$H^-IhCnS&5TIJ(PExxY>p}+6u?TbeR8A86)ad2u+rvVxXf;ZGv3JI92DK;$(6dS zJAM41!>c?=!RNw>Y?X6Mp1_Mxf_JkV*6<>gK~YTKyIJIio$J>vSYx?+=&F6f4xOK; zZT~#t?#ac8IWbq>lV`QOUvIKhzOSXhFzcGREEMD%3B&f!=)Gx9=& zfnA)91X3!5Kb*hArw>{XJluPR@R5JVM#GNy(gz|u_TwQ$IDR;{SX+)xtV-?%E+OW)e5>fZzgo#gU^sWVzC>7{zKegBu`HZT1HAc&>RgKLv)lNjB zv=z~mI*1t!lc_SUY^%e^eO|g1*3Twn0M=elP#ww zvA?F(9@{F-24B2;WZD(G3ic_T>P{B~FbZj}pvV=ETQ~-TSBBfU{I{+b_m7Z@EN7uF ze14xM%` z7sm^tdjeV6JPu`gr8fhmZ#R5@&A_vk9g3_q+_Z9_+YC6e+s3CAnZ3d}#M1lz5~zfw z()k3T2Z4qp%!`7n15tf52h9|MX6@@DI$wi{@^>fcmXv;ePFM#*Aa7iga>;oo1e;7cV;=8e;5t}?yPPi zh?qZhLWWUSg^Fk|ewjCV4;Em12?MYRIJX-q@uBTqucmX>+=#`e_olwVWtHyw6^u9Y z8AT=}dmp@_43(QvgU=VL^Ew6sxo5jKp~e21F^_Yw6*yz`!VaYcbK0RQ?&jlwZH+#R z{hU*YwMq602UY4}oF}#b@+f9b`L0BqTfXfIeXO4C^|o z>h;NeCLJZ+A#W@N2<|GtC>3?PawOPRH9wOm%%gU%J9Kn!Jam+?{7wr09LG%c9OrQ& zbW~+Fbo6X8bhHaKeY_DfJ##g5bj+mt56!*%+HF)w&nwglyt)b{y~2`&*H8OTt-Zq7 zVBQ8kUNcI12t6-4-)4+Yd5XwhF3Rh;B*}|D5!MFZ^p;n*y!4AKTp{U;pTfBsqx(+) zp_sEZOFzDDT9GM7=ALkeT!iG#Go3MA=UN$TL&)AzZ%FIu(xainQB+@6kmaxQtnR%` zq!*Ag$X&|cLHnCQ5&+x3{=Om59KX#WHD6LU1${aui9lR03Vn|QBahTrl4M#li;+Sn z4~{DF_I3$7b9G};A@tbg-SFLT=bZ8?VJ-6_(x@fYvg!4nOsApwQt?v@7nHK{^q2-x zDc`F#&>XV>%`x)knE9JIPBxu%+QpMyxf(vqbwEF~lV8_rc=7Xfn7WR6WVIW1X5fr6 z8rF4ez+VP%B`1mwC^-E)X7_s(*sq)hmP|`Lbe8M3t@-uyb?7#x+`l)TJ~(0o9N@Tu zqZihi{hK#_eDlV6tH|xP1v^o1-gx!R8@CdvvpYWjsk7#P>+JNW8p~6l&Uzjp$qfJD zjj#Ul#=~PkZ#-LXGuv9PQ*ix>_RtlxzU2>Ze9@F%Z#Ftz;}yU;pxiz|JnxUe@Bxew z3>Nx5aDfE|Fg3R$@Wfu5VV@yrGzR&?Vsr*uh6{Yq7+*QPAE9Q>_%zfEA7u*KR20yY zz|0s-F-kacMaOZe9Qr|XrZB-YI`ymht`q6=T9Q{A9F|LYlXhTA zjlplI&}8dJD!0Mv!>8C&tG#QYv-uSHt_{;IB8hxMqYKfD5+=WE z{K{u;7A6h5lt{pD{XxPPdh6%U-=-)a4QPIDV4LVd{Ba8LPp z#4}3LvWU7F{Z4ExBNBb%!S_H+X|cZ6XLhlk)OCRuy1uLdtNz}Qs#mxJP;C9ut2`HkfF?PU!sr=xyYdC1dAFGF1Mi`1 z@F=H(G0+x=S^>cY^7`hFC&!|?dDXAJ_;^~uqh0UPL(?7d z`Ecp^jwRmaY8GNy*#2a*>(uql8wdUAjT?*^8_N9n%*=PSe~0XEMqd7hBAYo=wZp&# zl-ZQjVJXt>0R+3n+13ef5q98MO(mIoPB3w%GwsmAe~=|mtsI`9QB7aUX3cjYkU#W^dcWauZ`b4pB4z8XgAGOtGjh* zrss2n&fodN8oT~wjlUWLt+Cls{h_u-`IKvJoI^95#_Au|*kwg_sphXZTcl(1UyieJ zf5zG3w>ZoC+c=y4F!WAgb@+6d^6xkc`VVmy^1n0AKG45({T*jf|I;{&6+ZhP#aYPz z!{RL9;qN%>)cj_RJO8xCn$_6Lv+YD9A9hK9hwN{rbN*{_w%zg%mBm~U%Dn-)L6E$E zudnj?@_BMlMOh0^F-0j}+B4a&>4R~ipv$=jiE9tL;L_5A zOrBE@uZMMy-NwPAa~Ym+p~f2u2hEZLrB<$Gh>Joi?a|cMa$brbFNU*TpUWc~JTF?w zTL^B+qmFZyWg~a}h$txDDyW52EpT831@~74^>AMQ7qcYgM~$X)*%kUMzzWLsTjTQe zG9}zJ^OY$1m*rBP{@|5kUcj3*KHs`6^5hbJv&Mn_n7>w56w zYw-ZBF)dc%<>F`dy9@oQThj-T+0&u?IzMH`2t+gUpb;YE#@7X0gVV6%l^v6F?y)Oh~K59Kov-$-$&aJ!}`pDkY+fvk?4TS;33VS^tJUB zRF1S!wo$@CNFj*^D4CX`TC}2BHc1BELbZq0;-2v)r7d@a3 z)Db^uvSG_1M8T#BqIBMDiUoG}y`_qf1b`wk54o#_{H)9R83Gq!c9e-5TbW4zVdAeQJX~&d1e^X3 znVkz~JXlO&w4u30M*Py0;m{Ay>_=Z~fG}LEhY)2x{D(8{PAf;$-=k6S3Uz>=N%z$* z2dXUOp=@Tp&NB{BWf7X+74uQ>9*zQ4*3Ott#VFEBqkII|O&I~R?jI`4b+rNPnD)AM zG5%876rjrP1AC>zE~7i(`5!79GCFqJWwd7xR9PpW%G&*h zmHqUW$^tDlP-U&||7xJRfelnBuz||`qk*z{YoNZrHBeJcU*L#LBlBVZYM>58|Dl0e zxC8|LjqIuGD^O(v!QUFF6Weh;`&D5tkcr7%%il5kn?a)gc9|u48?jDK{C{-DF8|<+ zQ;6-n<{hkwF zz13We%=*kDWThiG?Kn(E>dFp%it@eAD!-qA?b#D1PLkE@DklFAV%wiQ&6nq{=?@;5 zebUN#h@#4n^kR{|P)|2BmmPZCS8xGXU*E#)XrO(t;axHLiQ2pdN-c+c;sNw*WdwyeY)UgG9qCW0OGb97vXzy?GhN<4) z{J!nNBw&Q26)9?TZtAG|$k}{?>`6)*!%*>)%QhaS0yf9u1(Ha_xuqy;xP;lylIiyw z6Hb?I?4crB_;6-I^(r9QHUx37*249^u&x65VTSfY=WqaBb038fX&>fDbiUms+&No( znM2>%4>JAdVD)-sa+AU-VfDI*kn6F4Bwmz9+;;TNYf@td4L*r89SzU$eMum;g;g(M zUh97~Q1vkXX#)k{WA-04P+ijVsOCUwYSrX8a4iuS#rl+rQad@ zn>nWcR%G8~R`5Rzvp0W**^y^M<3GYIq_VQoszzifo)-{V$0VNBf}l*RT1&;*?Yss1 zpmQLynwPF7%UusCiuLhAxBnYi{D?PXQ&QlkmU^jcct;NHSg8mrLhyRG#;Ph zNC*S4rf;1fzm>*+I%79u(YMmr6<8X3a7CTe@85WoYZTsRr*lVj;_WxX?%WC;1`6F~ zmGYtQ!7dK&UmXyO?(5XZmn=Ruti%&FW-DCMH+UNHHm9b*nV75SPu88dFR)t9IndLo zm^RA~HfTYXn!P47qvms;Qk*RORT?Xmv;3D!xdP zaRtafl*YdQou%>fJFkuZi|qfa(ir)zG?saW{i8Jg%Nb`u?uP#k+24#&0hpQp<C)c$qih{VuT_Z#@d)yt0%L6TOoAxGxz1w*QcW;kNbP> zD(Tkis=FeL)~8H9#G3T&Rkp{ylhVDz$q}dC(-rm>9A)IFjoq=?(HF{-%lwT3vnXVe zyMm%s*$AR+US!*slH)v-BhnM?r}X)g@?|QH#k7T6{l#H2U+slWH<$-|O^DA+zP{f* zAJ@ljma{)==f;9kHjBhAC-A7wAdU1Jc3kr|Gs*))vXVONv%v|QyX0W=j-Fw z-R`GJTzW=&A~$-LVa&tA#~z}a1_y!geX6plL!ftn-sgS1lmXW@DVI$fe|DUn^UheS6DbaHbH(0jo%j$ zt1pHrJX;uuMs`Lfxx;j!&fs$!B9d0Evx1fWPwaxJZ2h8>tnLUso3)y_C7MmX$( z8!?;mE*w=RP7M=`@$0OLJ?D;H1!T|B-z!O8WE(PwC;hiJ)lwO;*!B;s&`}Xkc-VQk=P1i)Uk!#5T`m!F( zZsoCW{rg&Kt^qmZ+3)IFDz{6@QW~y*OxCs7?#w9DHD3QHBG=xwy-*^Qb-mciim?RY zAOtSb@G>Z*nI1^ew0`B2k&hE)yLdmyZm@d#O@!~^E?k>y$Ww8g&jN28>vu%{W(JC# z>0cjg^Z*`O_`e%rpB^^;TqS-E)t@$zWm#zQlI7%J zPfp@;%OP`9Z-Jazrxn&6rjDP#9)q#!5|*D?D>T^>wn&f3*>f^=ON3?Wf5nC;x9CTG z)a0^6RHEWOdl&HVFt47uuU;`wSadvhO3ih&=yX_;yAF^j(O z7gX*r+Yai`MwTGN^3Y7N?(p&Ik7f9X!<9t6cU4vlTvcAvvw^Fl10G8svw_`1opeb{VU-y?b@ z-4I5#jW9r7&zDmkjc7>=fSvQw_>kfgGN--Aa74Ar*f|UTjD4f1XBDH-%)Ec?i*pg? z!S3z}YoF}P$jx-5>D#QwI}72k5q%rbl5kD7$Ct(FNJ{8y%T7Fl) z*2cV3imDWD9Ax*r*o(zoIC_drhdbw^{R%wpMC>T$OJ0@d*=4Y;-I-+1w&{!69(}u4 zM528n{a#!JZ{|pEd!#}=d^4ZqP zz(F6lz5%7R!B<5?6{f`gu^p}sxvk`Qq1 z1AE~2q;!FIA$BscHzqcAu{N}|HT+6NFKOuFWMyk$M+tn4KW=DZWho3?JAs%|m=pNJ zSs6GO7#W!u*%_Fq8JH*-7$|_ZleRMW&u*e)IRJ8R`-{{&hz-I^Z_DRBwwt7+M-w>6yYY0)?G{`0WxmHyphp@EH@+ z%j%gb*jvES%Mi1&zAYAEYiCc)!p;syFJ)+H1pK=&GqC=3@Xmj}lti7?o_q(;&7()!aiiJVX`Iy-5T3|+3Sk3PLQ9CbnT;N7ZhyBuFFbH zP6Diqd56EGzwO#0=|qo}i2DavJ8a#d+jT)5(%O1|iIaB7Co1ShjCCo&_(2#aa}J$S zVhtv98`#_705OOyx8^V3GUO5m^*RFHyAb1h{0F-)ZjT0Ae zsFKoZj|~Ao`w!e=8ml`H){StMAgR@uLwhLwKK^GR7%dRC!VSKtz2x$)NH%2Ocoo1$ zf)&4`pP)Hs%dUwKLX!X}(nvwW#JKx?yvfq?0iZbC+Z}e_@4~qlg75GdI#Gc8wD8Y$ z#d`-a?nT4x-Dl)Lb22y&+*%}jV$3qn7}eqJ)-*>ImXOU%^Pk6`?#`?Y3(efC?Og^{ z7(?1R8dkY0t+v_>9uuMY2c9T~C*W(a~FX*Obx@dFT>EiRm}Gk*K$$7kB}D4-?H? zz9rjv#QmdlU2#o)8n_khnDRZwad-H%R>~~sl;w8lqgHO+6!>CnBhK+X7(+6n>}7%{ zXaYn%*#+~gHni6hV+MPar1YLD+AakrUj90h?1axRuXi*VyIUJCPfsV?=i6tI^JA!x zRpEec^iJIi{uY%r8Z4*rF|@sgn^{nvfga8g0aE7+fJ)26x;au^3;t(!#2ZxvF8T{% zxL&bynhA)!){*NCQYi=?r)}cic8(nZ*Alx6-<5IhaPo%?erv1_06Bgima-q;>r&Ww z@???IsipV02?c2|Nf>oFOE2UnLa(f?IVDlYfT-$YB;H8nOj4B}Bb10euY24RlkV6f zg>~8aix^@^c(-%uC0Jp^$T7}nUk53j-A((hg|ueHZ^rjtE~m9SM>EE^MdC%Hv*3#s zqsg7rHIFJA0($7|FP!sbdaq)&CY-K9wI;nI(K94irw3`mzbA}SrA|h_NWW&HM)-Cc zqm3od?aSxuGG$T?8)J>=h+{#{r|ht#A980PUsaTTTNTSro_u_anH44V3c^cl)44hQ z-gg=cbKlF#`&rrb?B4n1{Bk&Vo%x7~(@s(p-)!_!sJSDUv4V6?%s4+Hu`YRd-NSfg zWf_?-1$?-O?mMyi{(S8N``5*;R>f3NJYO;B>GHB}K*L>~63V#Ao#v!T<)?jwqb8Uh z6Ewddii*5J;^_(;lhRO|LE|4db)&llGa>^y_fh_(drl7u`4XCKTKAWCPj^Spv~Y9l zH+q`L=A$pP4=$Aot%8J8pz}3}*8{F~6w3LxMBek)9!7@1-5>&d61(FcyAXqU>2oTw z?91KEngpz#1mm?(_aUVu-|03D9Jsp~mbV6$t)w=wyEjU~W@cEY_PM{CO=|TkT@t7p z?mPa`SsKaf#qIHYapCGb|4YpWW2Y@N30{SbTI*a!jIBw5NA7^;SZPy2&d^S6cbr`%w&UGFFuUn0bd%@G8`jlQxZ(oC*q@pN3jiH0#z_ zT^)WAFqzVS;+Ui}`tck$IBV3KuMDXlUY)~}4x-MA+Y+^84{>``wsK&ok(+GI$7TM4D43yCj=qr>h<=8+K?BvV8`C4a`*2$(9>4wlGyEV`0$Zy7o3x(y|h zja8Qmqby??i!nL)<)%v4cad-S?)lzfgkmG*A(eNy5nv%V44u*~$fhmILn7Kb` zH`xm>gLPju;VTCx!8wUlGvWYVk1q5DRm#>S!VT+o6??x0KG-K~sIu($M}kw_3g+)G zl7gKvqmRFU^s^Dq5PiM`DT5Vf1!b%wM|BhIsuS;eBsIm4X9eLXhU_aP5yZlacw&1$ zP9%5ORum#p&2UclEx}aC_Cd%LUxpR520!_du6IWUFM5&a^K*HANdESW4C|P|bcirs zq!Y-7wu*JIzPViPm zyVRCpFK%^8C9l}yVFz@>#flV}*cC#`XQQpF8=2aW%54jo;boQXmW~zZZDuqNH|BYM zs(E^v5o)!E#|-x`7gi_9NtE4|VYAf+z^_FGj87esDAmU|f6BzV6UC1HL_Ow_L^q^% zeLBHjl^oo~h@8pG4ynZC$#R2ze=Bew?mEFuZz<|LNNqU1;#?)McI(#Oz*F&SQBU8^+c>>=>?L7~h@8Knf+~Swu^2N9PVeoTVgq7g( z5IQcS>n#iPNY=}$EaATAd{r1)ToqX;H+BLopIEuI$2=INCq$yv&kXm)qb^xMT}nbC zb;9C!)%vGCT{B78QG3J2nJ`KEi8z8Jmw+*N;L&w7`CKQ))b0f_hNl$w+%ut)( zFjV@&_JsThM0fr;U8QJ*19pOY!uexe|4$b+h31LuzGUf=X}7kPC<|2DQ2S97=_lq3`#_s`9Z1VF(NUKWng}_ zgl(NUxotHMOI$;A&F)Zx*fPfm_ob2iSn0buhe?w6C5~c~Lj`H6sQ{l_9n<1#Lb~lGCN9Pi5zC@sc@OsyiagQ%-U8F{1 zz_@V=f2$|Vr5H>y>ysjMm|RW?0JKIYuLeTI`|Hd{>XU&`+lnXWxe*=P* zY1nYhx*~MDrrDfj;Rc7LFP_&yIzJafa1sV&RQQ!$HB&#M80=Q_&V92Wo${@9I15zz z#VNP_U8v$BIeso9lfI(A|IYkQ|pD14E-?Ed;9ANdVSs`p>{II&XsW=_{Xb~V||v)FoPm|cvB zpwu)jeb2m<)Cz@Y17L5(B$i7*II{}WaWUc24_AmbR4`UB1PTV_=a&ElK?_-fKswU$ zL46x2&qX2H#@xif(A?47#lXbW*%Zh&Gjw#ba5ONmv~+WFb}=)tQy{DaXb-e+V`@bF zK&R^Ev&G$kBF8@#&$HQ{`D>O-#?&m&txH4$gLD^hx@xTQ-K+97D8E|T_>2U%ae1+D731JiKs`Xf78Rxe%>+D|G{#4^l^WQgLo#amoTAVOqK%V=FGhFG^w5eM7uPMecmIA)YJ0ya zXR?;EZ}s(E^6i`(j`2G;YGXp>hbjOlBF%D zzOPvtk@|bP>hf7x?@L#NC3$;Z4Xa)s{Pn1?x8~LF`nshn>uSFSYY+v1eQvFT6B<+ND_$aa*TcJmPjZDD3f>=@)AAtPKnVCA3)E(wP@;b@*4A z@F#e|2djV_mSvM0KW#s7XS0K<{EB}ZI(vApXg40+-?-Ggto^L0^NZyTJ3smKrSGsg z{B)+!d5_9vCyjmQWqvZ2TA$Mv^xE_%zv|OiXUsHTT#{H+QUM%0;4%U(DzPx-QdM>J HcjE#8@`c)u diff --git a/doc/Projects/2025/Project1/pdf/Project1.tex b/doc/Projects/2025/Project1/pdf/Project1.tex deleted file mode 100644 index 2c142582b..000000000 --- a/doc/Projects/2025/Project1/pdf/Project1.tex +++ /dev/null @@ -1,476 +0,0 @@ -%% -%% Automatically generated file from DocOnce source -%% (https://github.com/doconce/doconce/) -%% doconce format latex Project1.do.txt --print_latex_style=trac --latex_admon=paragraph -%% - - -%-------------------- 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 - -% 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 1 on Machine Learning, deadline October 6 (midnight), 2025 -\end{spacing} -} -\end{center} - -% ----------------- author(s) ------------------------- - -\begin{center} -{\bf Data Analysis and Machine Learning FYS-STK3155/FYS4155} -\end{center} - - \begin{center} -% List of all institutions: -\centerline{{\small University of Oslo, Norway}} -\end{center} - -% ----------------- end author(s) ------------------------- - -% --- begin date --- -\begin{center} -September 2 -\end{center} -% --- end date --- - -\vspace{1cm} - - -\subsection*{Preamble: Note on writing reports, using reference material, AI and other tools} - -We want you to answer the three different projects by handing in -reports written like a standard scientific/technical report. The -links at -\href{{https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects}}{\nolinkurl{https://github.com/CompPhysics/MachineLearning/tree/master/doc/Projects}} -contain more information. There you can find examples of previous -reports, the projects themselves, how we rade reports etc. How to -write reports will also be discussed during the various lab -sessions. Please do ask us if you are in doubt. - -When using codes and material from other sources, you should refer to -these in the bibliography of your report, indicating wherefrom you for -example got the code, whether this is from the lecture notes, -softwares like Scikit-Learn, TensorFlow, PyTorch or other sources. These sources -should always be cited correctly. How to cite some -of the libraries is often indicated from their corresponding GitHub -sites or websites, see for example how to cite Scikit-Learn at -\href{{https://scikit-learn.org/dev/about.html}}{\nolinkurl{https://scikit-learn.org/dev/about.html}}. - -We enocurage you to use tools like -\href{{https://openai.com/chatgpt/}}{ChatGPT} or similar in writing the report. If you use for example ChatGPT, -please do cite it properly and include (if possible) your questions and answers as an addition to the report. This can -be uploaded to for example your website, GitHub/GitLab or similar as supplemental material. - -If you would like to study other data sets, feel free to propose other -sets. What we have proposed here are mere suggestions from our -side. If you opt for another data set, consider using a set which has -been studied in the scientific literature. This makes it easier for -you to compare and analyze your results. Comparing with existing -results from the scientific literature is also an essential element of -the scientific discussion. The University of California at Irvine -with its Machine Learning repository at -\href{{https://archive.ics.uci.edu/ml/index.php}}{\nolinkurl{https://archive.ics.uci.edu/ml/index.php}} is an excellent site to -look up for examples and -inspiration. \href{{https://www.kaggle.com/}}{Kaggle.com} is an equally -interesting site. Feel free to explore these sites. When selecting -other data sets, make sure these are sets used for regression problems -(not classification). - -% !split -\subsection*{Regression analysis and resampling methods} - -The main aim of this project is to study in more detail various -regression methods, including Ordinary Least Squares (OLS) reegression, Ridge regression and LASSO regression. -In addition to the scientific part, in this course we want also to -give you an experience in writing scientific reports. - -We will study how to fit polynomials to specific -one-dimensional functions (feel free to replace the suggested function with more complicated ones). - -We will use Runge's function (see \href{{https://en.wikipedia.org/wiki/Runge%27s_phenomenon}}{\nolinkurl{https://en.wikipedia.org/wiki/Runge\%27s_phenomenon}} for a discussion). The one-dimensional function we will study is -\[ -f(x) = \frac{1}{1+25x^2}. -\] - -Our first step will be to perform an OLS regression analysis of this -function, trying out a polynomial fit with an $x$ dependence of the -form $[x,x^2,\dots]$. You can use a uniform distribution to set up the -arrays of values for $x \in [-1,1]$, or alternatively use a fixed step size. -Thereafter we will repeat many of the same steps when using the Ridge and Lasso regression methods, -introducing thereby a dependence on the hyperparameter (penalty) $\lambda$. - -We will also include bootstrap as a resampling technique in order to -study the so-called \textbf{bias-variance tradeoff}. After that we will -include the so-called cross-validation technique. - -\paragraph{Part a : Ordinary Least Square (OLS) for the Runge function.} -We will generate our own dataset for abovementioned function -$\mathrm{Runge}(x)$ function with $x\in [-1,1]$. You should explore also the addition -of an added stochastic noise to this function using the normal -distribution $N(0,1)$. - -\emph{Write your own code} (using for example the pseudoinverse function \textbf{pinv} from \textbf{Numpy} ) and perform a standard \textbf{ordinary least square regression} -analysis using polynomials in $x$ up to order $15$ or higher. Explore the dependence on the number of data points and the polynomial degree. - -Evaluate the mean Squared error (MSE) - -\[ MSE(\bm{y},\tilde{\bm{y}}) = \frac{1}{n} -\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2, -\] - -and the $R^2$ score function. If $\tilde{\bm{y}}_i$ is the predicted -value of the $i-th$ sample and $y_i$ is the corresponding true value, -then the score $R^2$ is defined as - -\[ -R^2(\bm{y}, \tilde{\bm{y}}) = 1 - \frac{\sum_{i=0}^{n - 1} (y_i - \tilde{y}_i)^2}{\sum_{i=0}^{n - 1} (y_i - \bar{y})^2}, -\] - -where we have defined the mean value of $\bm{y}$ as - -\[ -\bar{y} = \frac{1}{n} \sum_{i=0}^{n - 1} y_i. -\] - -Plot the resulting scores (MSE and R$^2$) as functions of the polynomial degree (here up to polymial degree 15). -Plot also the parameters $\theta$ as you increase the order of the polynomial. Comment your results. - -Your code has to include a scaling/centering of the data (for example by -subtracting the mean value), and -a split of the data in training and test data. For the scaling you can -either write your own code or use for example the function for -splitting training data provided by the library \textbf{Scikit-Learn} (make -sure you have installed it). This function is called -$train\_test\_split$. \textbf{You should present a critical discussion of why and how you have scaled or not scaled the data}. - -It is normal in essentially all Machine Learning studies to split the -data in a training set and a test set (eventually also an additional -validation set). There -is no explicit recipe for how much data should be included as training -data and say test data. An accepted rule of thumb is to use -approximately $2/3$ to $4/5$ of the data as training data. - -You can easily reuse the solutions to your exercises from week 35. -See also the lecture slides from week 35 and week 36. - -On scaling, we recommend reading the following section from the scikit-learn software description, see \href{{https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html#plot-all-scaling-standard-scaler-section}}{\nolinkurl{https://scikit-learn.org/stable/auto_examples/preprocessing/plot_all_scaling.html\#plot-all-scaling-standard-scaler-section}}. - -\paragraph{Part b: Adding Ridge regression for the Runge function.} -Write your own code for the Ridge method as done in the previous -exercise. The lecture notes from week 35 and 36 contain more information. Furthermore, the results from the exercise set from week 36 is something you can reuse here. - -Perform the same analysis as you did in the previous exercise but now for different values of $\lambda$. Compare and -analyze your results with those obtained in part a) with the OLS method. Study the -dependence on $\lambda$. - -\paragraph{Part c: Writing your own gradient descent code.} -Replace now the analytical expressions for the optimal parameters -$\bm{\theta}$ with your own gradient descent code. In this exercise we -focus only on the simplest gradient descent approach with a fixed -learning rate (see the exercises from week 37 and the lecture notes -from week 36). - -Study and compare your results from parts a) and b) with your gradient -descent approch. Discuss in particular the role of the learning rate. - -\paragraph{Part d: Including momentum and more advanced ways to update the learning the rate.} -We keep our focus on OLS and Ridge regression and update our code for -the gradient descent method by including \textbf{momentum}, \textbf{ADAgrad}, -\textbf{RMSprop} and \textbf{ADAM} as methods fro iteratively updating your learning -rate. Discuss the results and compare the different methods applied to -the one-dimensional Runge function. The lecture notes from week 37 contain several examples on how to implement these methods. - -\paragraph{Part e: Writing our own code for Lasso regression.} -LASSO regression (see lecture slides from week 36 and week 37) -represents our first encounter with a machine learning method which -cannot be solved through analytical expressions (as in OLS and Ridge regression). Use the gradient -descent methods you developed in parts c) and d) to solve the LASSO -optimization problem. You can compare your results with -the functionalities of \textbf{Scikit-Learn}. - -Discuss (critically) your results for the Runge function from OLS, -Ridge and LASSO regression using the various gradient descent -approaches. - -\paragraph{Part f: Stochastic gradient descent.} -Our last gradient step is to include stochastic gradient descent using -the same methods to update the learning rates as in parts c-e). -Compare and discuss your results with and without stochastic gradient -and give a critical assessment of the various methods. - -\paragraph{Part g: Bias-variance trade-off and resampling techniques.} -Our aim here is to study the bias-variance trade-off by implementing -the \textbf{bootstrap} resampling technique. \textbf{We will only use the simpler -ordinary least squares here}. - -With a code which does OLS and includes resampling techniques, -we will now discuss the bias-variance trade-off in the context of -continuous predictions such as regression. However, many of the -intuitions and ideas discussed here also carry over to classification -tasks and basically all Machine Learning algorithms. - -Before you perform an analysis of the bias-variance trade-off on your -test data, make first a figure similar to Fig.~2.11 of Hastie, -Tibshirani, and Friedman. Figure 2.11 of this reference displays only -the test and training MSEs. The test MSE can be used to indicate -possible regions of low/high bias and variance. You will most likely -not get an equally smooth curve! You may also need to increase the -polynomial order and play around with the number of data points as -well (see also the exercise set from week 35). - -With this result we move on to the bias-variance trade-off analysis. - -Consider a -dataset $\mathcal{L}$ consisting of the data -$\mathbf{X}_\mathcal{L}=\{(y_j, \boldsymbol{x}_j), j=0\ldots n-1\}$. - -We assume that the true data is generated from a noisy model - -\[ -\bm{y}=f(\boldsymbol{x}) + \bm{\epsilon}. -\] - -Here $\epsilon$ is normally distributed with mean zero and standard -deviation $\sigma^2$. - -In our derivation of the ordinary least squares method we defined then -an approximation to the function $f$ in terms of the parameters -$\bm{\theta}$ and the design matrix $\bm{X}$ which embody our model, -that is $\bm{\tilde{y}}=\bm{X}\bm{\theta}$. - -The parameters $\bm{\theta}$ are in turn found by optimizing the mean -squared error via the so-called cost function - -\[ -C(\bm{X},\bm{\theta}) =\frac{1}{n}\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2=\mathbb{E}\left[(\bm{y}-\bm{\tilde{y}})^2\right]. -\] -Here the expected value $\mathbb{E}$ is the sample value. - -Show that you can rewrite this in terms of a term which contains the -variance of the model itself (the so-called variance term), a term -which measures the deviation from the true data and the mean value of -the model (the bias term) and finally the variance of the noise. - -That is, show that -\[ -\mathbb{E}\left[(\bm{y}-\bm{\tilde{y}})^2\right]=\mathrm{Bias}[\tilde{y}]+\mathrm{var}[\tilde{y}]+\sigma^2, -\] -with (we approximate $f(\bm{x})\approx \bm{y}$) -\[ -\mathrm{Bias}[\tilde{y}]=\mathbb{E}\left[\left(\bm{y}-\mathbb{E}\left[\bm{\tilde{y}}\right]\right)^2\right], -\] -and -\[ -\mathrm{var}[\tilde{y}]=\mathbb{E}\left[\left(\tilde{\bm{y}}-\mathbb{E}\left[\bm{\tilde{y}}\right]\right)^2\right]=\frac{1}{n}\sum_i(\tilde{y}_i-\mathbb{E}\left[\bm{\tilde{y}}\right])^2. -\] - -\textbf{Important note}: Since the function $f(x)$ is unknown, in order to be able to evalute the bias, we replace $f(\bm{x})$ in the expression for the bias with $\bm{y}$. - -The answer to this exercise should be included in the theory part of -the report. This exercise is also part of the weekly exercises of -week 38. Explain what the terms mean and discuss their -interpretations. - -Perform then a bias-variance analysis of the Runge function by -studying the MSE value as function of the complexity of your model. - -Discuss the bias and variance trade-off as function -of your model complexity (the degree of the polynomial) and the number -of data points, and possibly also your training and test data using the \textbf{bootstrap} resampling method. -You can follow the code example in the jupyter-book at \href{{https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/chapter3.html#the-bias-variance-tradeoff}}{\nolinkurl{https://compphysics.github.io/MachineLearning/doc/LectureNotes/_build/html/chapter3.html\#the-bias-variance-tradeoff}}. - -\paragraph{Part h): Cross-validation as resampling techniques, adding more complexity.} -The aim here is to implement another widely popular -resampling technique, the so-called cross-validation method. - -Implement the $k$-fold cross-validation algorithm (feel free to use -the functionality of \textbf{Scikit-Learn} or write your own code) and -evaluate again the MSE function resulting from the test folds. - -Compare the MSE you get from your cross-validation code with the one -you got from your \textbf{bootstrap} code from the previous exercise. Comment and interpret your results. - -In addition to using the ordinary least squares method, you should -include both Ridge and Lasso regression in the final analysis. - -\subsection*{Background literature} - -\begin{enumerate} -\item For a discussion and derivation of the variances and mean squared errors using linear regression, see the \href{{https://arxiv.org/abs/1509.09169}}{Lecture notes on ridge regression by Wessel N. van Wieringen} - -\item The textbook of \href{{https://www.springer.com/gp/book/9780387848570}}{Trevor Hastie, Robert Tibshirani, Jerome H. Friedman, The Elements of Statistical Learning, Springer}, chapters 3 and 7 are the most relevant ones for the analysis here. -\end{enumerate} - -\noindent -\subsection*{Introduction to numerical projects} - -Here follows a brief recipe and recommendation on how to answer the various questions when preparing your answers. - -\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. You should have the code at your GitHub/GitLab link. You can also place the code in an appendix of your report. - - \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, Julia or Python. The following prescription should be followed when preparing the report: - -\begin{itemize} - \item Use Canvas to hand in your projects, log in at \href{{https://www.uio.no/english/services/it/education/canvas/}}{\nolinkurl{https://www.uio.no/english/services/it/education/canvas/}} with your normal UiO username and password. - - \item Upload \textbf{only} the report file or the link to your GitHub/GitLab or similar typo of repos! For the source code file(s) you have developed please provide us with your link to your GitHub/GitLab or similar domain. The report file should include all of your discussions and a list of the codes you have developed. Do not include library files which are available at the course homepage, unless you have made specific changes to them. - - \item In your GitHub/GitLab or similar repository, please include a folder which contains selected results. These can be in the form of output from your code for a selected set of runs and input parameters. -\end{itemize} - -\noindent -Finally, -we encourage you to collaborate. Optimal working groups consist of -2-3 students. You can then hand in a common report. - -\subsection*{Software and needed installations} - -If you have Python installed (we recommend Python3) and you feel pretty familiar with installing different packages, -we recommend that you install the following Python packages via \textbf{pip} as -\begin{enumerate} -\item pip install numpy scipy matplotlib ipython scikit-learn tensorflow sympy pandas pillow -\end{enumerate} - -\noindent -For Python3, replace \textbf{pip} with \textbf{pip3}. - -See below for a discussion of \textbf{tensorflow} and \textbf{scikit-learn}. - -For OSX users we recommend also, after having installed Xcode, to install \textbf{brew}. Brew allows -for a seamless installation of additional software via for example -\begin{enumerate} -\item brew install python3 -\end{enumerate} - -\noindent -For Linux users, with its variety of distributions like for example the widely popular Ubuntu distribution -you can use \textbf{pip} as well and simply install Python as -\begin{enumerate} -\item sudo apt-get install python3 (or python for python2.7) -\end{enumerate} - -\noindent -etc etc. - -If you don't want to install various Python packages with their dependencies separately, we recommend two widely used distrubutions which set up all relevant dependencies for Python, namely -\begin{enumerate} -\item \href{{https://docs.anaconda.com/}}{Anaconda} Anaconda is an open source distribution of the Python and R programming languages for large-scale data processing, predictive analytics, and scientific computing, that aims to simplify package management and deployment. Package versions are managed by the package management system \textbf{conda} - -\item \href{{https://www.enthought.com/product/canopy/}}{Enthought canopy} is a Python distribution for scientific and analytic computing distribution and analysis environment, available for free and under a commercial license. -\end{enumerate} - -\noindent -Popular software packages written in Python for ML are - -\begin{itemize} -\item \href{{http://scikit-learn.org/stable/}}{Scikit-learn}, - -\item \href{{https://www.tensorflow.org/}}{Tensorflow}, - -\item \href{{http://pytorch.org/}}{PyTorch} and - -\item \href{{https://keras.io/}}{Keras}. -\end{itemize} - -\noindent -These are all freely available at their respective GitHub sites. They -encompass communities of developers in the thousands or more. And the number -of code developers and contributors keeps increasing. - - -% ------------------- end of main content --------------- - -\end{document} -