From 1c0911a3d37e81db6f4931090d9caab1f5af471a Mon Sep 17 00:00:00 2001 From: mhjensen Date: Thu, 26 Dec 2019 19:20:42 +0100 Subject: [PATCH] corrected labels in dimred --- doc/pub/DimRed/html/._DimRed-bs018.html | 31 +++++++++---------- doc/pub/DimRed/html/DimRed-reveal.html | 31 +++++++++---------- doc/pub/DimRed/html/DimRed-solarized.html | 31 +++++++++---------- doc/pub/DimRed/html/DimRed.html | 31 +++++++++---------- doc/pub/DimRed/ipynb/DimRed.ipynb | 30 +++++++++--------- doc/pub/DimRed/ipynb/ipynb-DimRed-src.tar.gz | Bin 190 -> 190 bytes doc/pub/DimRed/pdf/DimRed-minted.pdf | Bin 259876 -> 259793 bytes doc/src/DimRed/DimRed.do.txt | 31 +++++++++---------- 8 files changed, 84 insertions(+), 101 deletions(-) diff --git a/doc/pub/DimRed/html/._DimRed-bs018.html b/doc/pub/DimRed/html/._DimRed-bs018.html index 0455b1a78..10609edc8 100644 --- a/doc/pub/DimRed/html/._DimRed-bs018.html +++ b/doc/pub/DimRed/html/._DimRed-bs018.html @@ -210,8 +210,8 @@ $$ \end{bmatrix} $$ -

-We will generate \( N = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from +Note that the mean refers to each column of data. +We will generate \( n = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from this distribution, and store them in the \( 1000 \times 2 \) matrix \( \boldsymbol{X} \).

@@ -220,13 +220,13 @@ The following Python code aids in setting up the data

-

N = 1000
+
n = 1000
 mean = (-1, 2)
 cov = [[4, 2], [2, 2]]
-X = np.random.multivariate_normal(mean, cov, N)
+X = np.random.multivariate_normal(mean, cov, n)
 

 

-Make a small Python code which plots the data.
+Make thereafter a small Python code which plots the data. Note that the function multivariate returns also the covariance discussed above and that it is defined by dividing by \( n-1 \) instead of \( n \).
 
 

 Now we are going to implement the PCA algorithm. We will break it down into sub-steps and across multiple cells.
@@ -234,17 +234,17 @@ Now we are going to implement the PCA algorithm. We will break it down into sub-
 
Compute the sample mean and center the data 

 
 

-The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall the sample mean is
+The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall that the sample mean is
 $$
-\mu_N = \frac{1}{N} \sum_{i=1}^N x_i
+\mu_n = \frac{1}{n} \sum_{i=1}^n x_i
 $$
 
-and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_N \} \) takes the form
+and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_n \} \) takes the form
 $$
-\bar{x}_i = x_i - \mu_N
+\bar{x}_i = x_i - \mu_n.
 $$
 
-When you are done with these steps, print out \( \mu_N \) to verify it is
+When you are done with these steps, print out \( \mu_n \) to verify it is
 close to \( \mu \) and plot your mean centered data to verify it is
 centered at the origin! Compare your code with the functionality from Scikit-Learn discussed above.
 
@@ -254,7 +254,7 @@ centered at the origin! Compare your code with the functionality from Scikit-
 Now we are going to use the mean centered data to compute the sample covariance of the data. Recall it is given by:
 $$
 \begin{equation*}
-\Sigma_N = \frac{1}{N-1} \sum_{i=1}^N \bar{x}_i^T \bar{x}_i = \frac{1}{N-1} \sum_{i=1}^N (x_i - \mu_N)^T (x_i - \mu_N)
+\Sigma_n = \frac{1}{n-1} \sum_{i=1}^n \bar{x}_i^T \bar{x}_i = \frac{1}{n-1} \sum_{i=1}^n (x_i - \mu_n)^T (x_i - \mu_n)
 \end{equation*}
 $$
 
@@ -265,9 +265,9 @@ Compare the computed covariance with the answer given above.
 
 

 Now we are ready to solve for the principal components! To do so we
-diagonalize the sample covariance matrix \( \Sigma_N \). We can use the
+diagonalize the sample covariance matrix \( \Sigma_n \). We can use the
 function np.linalg.eig to do so. It will return the eigenvalues and
-eigenvectors of \( \Sigma_N \). Once you have these, carry out the
+eigenvectors of \( \Sigma_n \). Once you have these, carry out the
 following tasks:
 
 
    
@@ -279,7 +279,7 @@ following tasks:
 
 $$
 \begin{equation*}
-x_i \approx \tilde{x}_i := \mu_N + \langle x_i, v_0 \rangle v_0
+x_i \approx \tilde{x}_i := \mu_n + \langle x_i, v_0 \rangle v_0
 \end{equation*}
 $$
 
@@ -293,9 +293,6 @@ Have the input be the data and have the output be the principal components and t
 
    
 Finally, try out your own PCA function with other data sets.
 
-
    
-After this we ask ourselves how do we prove the link between the maximum variance and the feature reduction.
-
 
    
 
    
 
diff --git a/doc/pub/DimRed/html/DimRed-reveal.html b/doc/pub/DimRed/html/DimRed-reveal.html
index d75134efb..5b1cbcdce 100644
--- a/doc/pub/DimRed/html/DimRed-reveal.html
+++ b/doc/pub/DimRed/html/DimRed-reveal.html
@@ -999,8 +999,8 @@ $$
 $$
 
     
    
 
-
    
-We will generate \( N = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from
+Note that the mean refers to each column of data. 
+We will generate \( n = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from
 this distribution, and store them in the \( 1000 \times 2 \) matrix \( \boldsymbol{X} \).
 
 
    
@@ -1009,13 +1009,13 @@ The following Python code aids in setting up the data
 
    
 
 
-
    N = 1000
+
    n = 1000
 mean = (-1, 2)
 cov = [[4, 2], [2, 2]]
-X = np.random.multivariate_normal(mean, cov, N)
+X = np.random.multivariate_normal(mean, cov, n)
 
    
 
    
-Make a small Python code which plots the data.
+Make thereafter a small Python code which plots the data. Note that the function multivariate returns also the covariance discussed above and that it is defined by dividing by \( n-1 \) instead of \( n \).
 
 
    
 Now we are going to implement the PCA algorithm. We will break it down into sub-steps and across multiple cells.
@@ -1023,21 +1023,21 @@ Now we are going to implement the PCA algorithm. We will break it down into sub-
 
    Compute the sample mean and center the data 
    
 
 
    
-The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall the sample mean is
+The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall that the sample mean is
 
     
    
 $$
-\mu_N = \frac{1}{N} \sum_{i=1}^N x_i
+\mu_n = \frac{1}{n} \sum_{i=1}^n x_i
 $$
 
     
    
 
-and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_N \} \) takes the form
+and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_n \} \) takes the form
 
     
    
 $$
-\bar{x}_i = x_i - \mu_N
+\bar{x}_i = x_i - \mu_n.
 $$
 
     
    
 
-When you are done with these steps, print out \( \mu_N \) to verify it is
+When you are done with these steps, print out \( \mu_n \) to verify it is
 close to \( \mu \) and plot your mean centered data to verify it is
 centered at the origin! Compare your code with the functionality from Scikit-Learn discussed above.
 
@@ -1048,7 +1048,7 @@ Now we are going to use the mean centered data to compute the sample covariance
 
     
    
 $$
 \begin{equation*}
-\Sigma_N = \frac{1}{N-1} \sum_{i=1}^N \bar{x}_i^T \bar{x}_i = \frac{1}{N-1} \sum_{i=1}^N (x_i - \mu_N)^T (x_i - \mu_N)
+\Sigma_n = \frac{1}{n-1} \sum_{i=1}^n \bar{x}_i^T \bar{x}_i = \frac{1}{n-1} \sum_{i=1}^n (x_i - \mu_n)^T (x_i - \mu_n)
 \end{equation*}
 $$
 
     
    
@@ -1060,9 +1060,9 @@ Compare the computed covariance with the answer given above.
 
 
    
 Now we are ready to solve for the principal components! To do so we
-diagonalize the sample covariance matrix \( \Sigma_N \). We can use the
+diagonalize the sample covariance matrix \( \Sigma_n \). We can use the
 function np.linalg.eig to do so. It will return the eigenvalues and
-eigenvectors of \( \Sigma_N \). Once you have these, carry out the
+eigenvectors of \( \Sigma_n \). Once you have these, carry out the
 following tasks:
 
 
      
@@ -1074,7 +1074,7 @@ following tasks:
 
       
      
 $$
 \begin{equation*}
-x_i \approx \tilde{x}_i := \mu_N + \langle x_i, v_0 \rangle v_0
+x_i \approx \tilde{x}_i := \mu_n + \langle x_i, v_0 \rangle v_0
 \end{equation*}
 $$
 
       
      
@@ -1088,9 +1088,6 @@ Have the input be the data and have the output be the principal components and t
 
 
      
 Finally, try out your own PCA function with other data sets.
-
-
      
-After this we ask ourselves how do we prove the link between the maximum variance and the feature reduction.
 
 
 
diff --git a/doc/pub/DimRed/html/DimRed-solarized.html b/doc/pub/DimRed/html/DimRed-solarized.html
index 686a4c32d..44ddd27fa 100644
--- a/doc/pub/DimRed/html/DimRed-solarized.html
+++ b/doc/pub/DimRed/html/DimRed-solarized.html
@@ -974,8 +974,8 @@ $$
 \end{bmatrix}
 $$
 
-
      
-We will generate \( N = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from
+Note that the mean refers to each column of data. 
+We will generate \( n = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from
 this distribution, and store them in the \( 1000 \times 2 \) matrix \( \boldsymbol{X} \).
 
 
      
@@ -984,13 +984,13 @@ The following Python code aids in setting up the data
 
      
 
 
-
      N = 1000
+
      n = 1000
 mean = (-1, 2)
 cov = [[4, 2], [2, 2]]
-X = np.random.multivariate_normal(mean, cov, N)
+X = np.random.multivariate_normal(mean, cov, n)
 
      
 
      
-Make a small Python code which plots the data.
+Make thereafter a small Python code which plots the data. Note that the function multivariate returns also the covariance discussed above and that it is defined by dividing by \( n-1 \) instead of \( n \).
 
 
      
 Now we are going to implement the PCA algorithm. We will break it down into sub-steps and across multiple cells.
@@ -998,17 +998,17 @@ Now we are going to implement the PCA algorithm. We will break it down into sub-
 
      Compute the sample mean and center the data 
      
 
 
      
-The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall the sample mean is
+The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall that the sample mean is
 $$
-\mu_N = \frac{1}{N} \sum_{i=1}^N x_i
+\mu_n = \frac{1}{n} \sum_{i=1}^n x_i
 $$
 
-and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_N \} \) takes the form
+and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_n \} \) takes the form
 $$
-\bar{x}_i = x_i - \mu_N
+\bar{x}_i = x_i - \mu_n.
 $$
 
-When you are done with these steps, print out \( \mu_N \) to verify it is
+When you are done with these steps, print out \( \mu_n \) to verify it is
 close to \( \mu \) and plot your mean centered data to verify it is
 centered at the origin! Compare your code with the functionality from Scikit-Learn discussed above.
 
@@ -1018,7 +1018,7 @@ centered at the origin! Compare your code with the functionality from Scikit-
 Now we are going to use the mean centered data to compute the sample covariance of the data. Recall it is given by:
 $$
 \begin{equation*}
-\Sigma_N = \frac{1}{N-1} \sum_{i=1}^N \bar{x}_i^T \bar{x}_i = \frac{1}{N-1} \sum_{i=1}^N (x_i - \mu_N)^T (x_i - \mu_N)
+\Sigma_n = \frac{1}{n-1} \sum_{i=1}^n \bar{x}_i^T \bar{x}_i = \frac{1}{n-1} \sum_{i=1}^n (x_i - \mu_n)^T (x_i - \mu_n)
 \end{equation*}
 $$
 
@@ -1029,9 +1029,9 @@ Compare the computed covariance with the answer given above.
 
 
      
 Now we are ready to solve for the principal components! To do so we
-diagonalize the sample covariance matrix \( \Sigma_N \). We can use the
+diagonalize the sample covariance matrix \( \Sigma_n \). We can use the
 function np.linalg.eig to do so. It will return the eigenvalues and
-eigenvectors of \( \Sigma_N \). Once you have these, carry out the
+eigenvectors of \( \Sigma_n \). Once you have these, carry out the
 following tasks:
 
 
        
@@ -1043,7 +1043,7 @@ following tasks:
 
 $$
 \begin{equation*}
-x_i \approx \tilde{x}_i := \mu_N + \langle x_i, v_0 \rangle v_0
+x_i \approx \tilde{x}_i := \mu_n + \langle x_i, v_0 \rangle v_0
 \end{equation*}
 $$
 
@@ -1057,9 +1057,6 @@ Have the input be the data and have the output be the principal components and t
 
        
 Finally, try out your own PCA function with other data sets.
 
-
        
-After this we ask ourselves how do we prove the link between the maximum variance and the feature reduction.
-
 
        
 









        
 
diff --git a/doc/pub/DimRed/html/DimRed.html b/doc/pub/DimRed/html/DimRed.html
index 7b1ba18c7..5be5508b6 100644
--- a/doc/pub/DimRed/html/DimRed.html
+++ b/doc/pub/DimRed/html/DimRed.html
@@ -979,8 +979,8 @@ $$
 \end{bmatrix}
 $$
 
-
        
-We will generate \( N = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from
+Note that the mean refers to each column of data. 
+We will generate \( n = 1000 \) points \( X = \{ x_1, \ldots, x_N \} \) from
 this distribution, and store them in the \( 1000 \times 2 \) matrix \( \boldsymbol{X} \).
 
 
        
@@ -989,13 +989,13 @@ The following Python code aids in setting up the data
 
        
 
 
-
        N = 1000
+
        n = 1000
 mean = (-1, 2)
 cov = [[4, 2], [2, 2]]
-X = np.random.multivariate_normal(mean, cov, N)
+X = np.random.multivariate_normal(mean, cov, n)
 
        
 
        
-Make a small Python code which plots the data.
+Make thereafter a small Python code which plots the data. Note that the function multivariate returns also the covariance discussed above and that it is defined by dividing by \( n-1 \) instead of \( n \).
 
 
        
 Now we are going to implement the PCA algorithm. We will break it down into sub-steps and across multiple cells.
@@ -1003,17 +1003,17 @@ Now we are going to implement the PCA algorithm. We will break it down into sub-
 
        Compute the sample mean and center the data 
        
 
 
        
-The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall the sample mean is
+The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall that the sample mean is
 $$
-\mu_N = \frac{1}{N} \sum_{i=1}^N x_i
+\mu_n = \frac{1}{n} \sum_{i=1}^n x_i
 $$
 
-and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_N \} \) takes the form
+and the mean-centered data \( \bar{X} = \{ \bar{x}_1, \ldots, \bar{x}_n \} \) takes the form
 $$
-\bar{x}_i = x_i - \mu_N
+\bar{x}_i = x_i - \mu_n.
 $$
 
-When you are done with these steps, print out \( \mu_N \) to verify it is
+When you are done with these steps, print out \( \mu_n \) to verify it is
 close to \( \mu \) and plot your mean centered data to verify it is
 centered at the origin! Compare your code with the functionality from Scikit-Learn discussed above.
 
@@ -1023,7 +1023,7 @@ centered at the origin! Compare your code with the functionality from Scikit-
 Now we are going to use the mean centered data to compute the sample covariance of the data. Recall it is given by:
 $$
 \begin{equation*}
-\Sigma_N = \frac{1}{N-1} \sum_{i=1}^N \bar{x}_i^T \bar{x}_i = \frac{1}{N-1} \sum_{i=1}^N (x_i - \mu_N)^T (x_i - \mu_N)
+\Sigma_n = \frac{1}{n-1} \sum_{i=1}^n \bar{x}_i^T \bar{x}_i = \frac{1}{n-1} \sum_{i=1}^n (x_i - \mu_n)^T (x_i - \mu_n)
 \end{equation*}
 $$
 
@@ -1034,9 +1034,9 @@ Compare the computed covariance with the answer given above.
 
 
        
 Now we are ready to solve for the principal components! To do so we
-diagonalize the sample covariance matrix \( \Sigma_N \). We can use the
+diagonalize the sample covariance matrix \( \Sigma_n \). We can use the
 function np.linalg.eig to do so. It will return the eigenvalues and
-eigenvectors of \( \Sigma_N \). Once you have these, carry out the
+eigenvectors of \( \Sigma_n \). Once you have these, carry out the
 following tasks:
 
 
          
@@ -1048,7 +1048,7 @@ following tasks:
 
 $$
 \begin{equation*}
-x_i \approx \tilde{x}_i := \mu_N + \langle x_i, v_0 \rangle v_0
+x_i \approx \tilde{x}_i := \mu_n + \langle x_i, v_0 \rangle v_0
 \end{equation*}
 $$
 
@@ -1062,9 +1062,6 @@ Have the input be the data and have the output be the principal components and t
 
          
 Finally, try out your own PCA function with other data sets.
 
-
          
-After this we ask ourselves how do we prove the link between the maximum variance and the feature reduction.
-
 
          
 









          
 
diff --git a/doc/pub/DimRed/ipynb/DimRed.ipynb b/doc/pub/DimRed/ipynb/DimRed.ipynb
index 82b60b04e..0c2ba7990 100644
--- a/doc/pub/DimRed/ipynb/DimRed.ipynb
+++ b/doc/pub/DimRed/ipynb/DimRed.ipynb
@@ -1070,7 +1070,8 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "We will generate $N = 1000$ points $X = \\{ x_1, \\ldots, x_N \\}$ from\n",
+    "Note that the mean refers to each column of data. \n",
+    "We will generate $n = 1000$ points $X = \\{ x_1, \\ldots, x_N \\}$ from\n",
     "this distribution, and store them in the $1000 \\times 2$ matrix $\\boldsymbol{X}$.\n",
     "\n",
     "The following Python code aids in setting up the data"
@@ -1084,23 +1085,23 @@
    },
    "outputs": [],
    "source": [
-    "N = 1000\n",
+    "n = 1000\n",
     "mean = (-1, 2)\n",
     "cov = [[4, 2], [2, 2]]\n",
-    "X = np.random.multivariate_normal(mean, cov, N)"
+    "X = np.random.multivariate_normal(mean, cov, n)"
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Make a small Python code which plots the data.\n",
+    "Make thereafter a small Python code which plots the data. Note that the function **multivariate** returns also the covariance discussed above and that it is defined by dividing by $n-1$ instead of $n$.\n",
     "\n",
     "Now we are going to implement the PCA algorithm. We will break it down into sub-steps and across multiple cells.\n",
     "\n",
     "### Compute the sample mean and center the data\n",
     "\n",
-    "The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall the sample mean is"
+    "The first step of PCA is to compute the sample mean of the data and use it to center the data. Recall that the sample mean is"
    ]
   },
   {
@@ -1108,7 +1109,7 @@
    "metadata": {},
    "source": [
     "$$\n",
-    "\\mu_N = \\frac{1}{N} \\sum_{i=1}^N x_i\n",
+    "\\mu_n = \\frac{1}{n} \\sum_{i=1}^n x_i\n",
     "$$"
    ]
   },
@@ -1116,7 +1117,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "and the mean-centered data $\\bar{X} = \\{ \\bar{x}_1, \\ldots, \\bar{x}_N \\}$ takes the form"
+    "and the mean-centered data $\\bar{X} = \\{ \\bar{x}_1, \\ldots, \\bar{x}_n \\}$ takes the form"
    ]
   },
   {
@@ -1124,7 +1125,7 @@
    "metadata": {},
    "source": [
     "$$\n",
-    "\\bar{x}_i = x_i - \\mu_N\n",
+    "\\bar{x}_i = x_i - \\mu_n.\n",
     "$$"
    ]
   },
@@ -1132,7 +1133,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "When you are done with these steps, print out $\\mu_N$ to verify it is\n",
+    "When you are done with these steps, print out $\\mu_n$ to verify it is\n",
     "close to $\\mu$ and plot your mean centered data to verify it is\n",
     "centered at the origin! Compare your code with the functionality from **Scikit-Learn** discussed above.\n",
     "\n",
@@ -1147,7 +1148,7 @@
    "metadata": {},
    "source": [
     "$$\n",
-    "\\Sigma_N = \\frac{1}{N-1} \\sum_{i=1}^N \\bar{x}_i^T \\bar{x}_i = \\frac{1}{N-1} \\sum_{i=1}^N (x_i - \\mu_N)^T (x_i - \\mu_N)\n",
+    "\\Sigma_n = \\frac{1}{n-1} \\sum_{i=1}^n \\bar{x}_i^T \\bar{x}_i = \\frac{1}{n-1} \\sum_{i=1}^n (x_i - \\mu_n)^T (x_i - \\mu_n)\n",
     "$$"
    ]
   },
@@ -1162,9 +1163,9 @@
     "### Diagonalize the sample covariance matrix to obtain the principal components\n",
     "\n",
     "Now we are ready to solve for the principal components! To do so we\n",
-    "diagonalize the sample covariance matrix $\\Sigma_N$. We can use the\n",
+    "diagonalize the sample covariance matrix $\\Sigma_n$. We can use the\n",
     "function **np.linalg.eig** to do so. It will return the eigenvalues and\n",
-    "eigenvectors of $\\Sigma_N$. Once you have these, carry out the\n",
+    "eigenvectors of $\\Sigma_n$. Once you have these, carry out the\n",
     "following tasks:\n",
     "\n",
     "* Compute the percentage of the total variance captured by the first principal component\n",
@@ -1181,7 +1182,7 @@
    "metadata": {},
    "source": [
     "$$\n",
-    "x_i \\approx \\tilde{x}_i := \\mu_N + \\langle x_i, v_0 \\rangle v_0\n",
+    "x_i \\approx \\tilde{x}_i := \\mu_n + \\langle x_i, v_0 \\rangle v_0\n",
     "$$"
    ]
   },
@@ -1198,9 +1199,6 @@
     "Finally, try out your own PCA function with other data sets.\n",
     "\n",
     "\n",
-    "\n",
-    "After this we ask ourselves how do we prove the link between the maximum variance and the feature reduction.\n",
-    "\n",
     "## Classical PCA Theorem\n",
     "\n",
     "We assume now that we have a design matrix $\\boldsymbol{X}$ which has been centered as discussed above. For the sake of simplicity we skip the overline symbol. The matrix is defined in terms of the various column vectors $[\\boldsymbol{x}_0,\\boldsymbol{x}_1,\\dots, \\boldsymbol{x}_{p-1}]$\n",
diff --git a/doc/pub/DimRed/ipynb/ipynb-DimRed-src.tar.gz b/doc/pub/DimRed/ipynb/ipynb-DimRed-src.tar.gz
index 55bb01bf7c59b9a7601fc78711ecc4c1db61e57c..ba174a3e865f9de050c6824fa6fb88627c0cf86e 100644
GIT binary patch
literal 190
zcmV;v073sBiwFSZ`2=171MSbz3W6{c24Js!it_}rHK*5sUUU(J^a7L3T;*olgnav^
zLR}TzNCf+Ce#ZG3=Tzmk*q4dByJ$5Cp)5^d$~>cUqRsN2VN3y&%x9kFzKAd(ltzHo
zTj^!6jv2p7dlt$H^|t40P4nTOx;8n_mA
s8i5hFK^g%}jlejUPHM%jac}g=u+Ap&ua60WAPBzp05xVzQ~(G70LboI-v9sr

literal 190
zcmV;v073sBiwFP+h6G*!1MSaC3c@fD2H>uHia9}h*^@rlzEoUiSbg;2uXp^Y)qs`5RgzB=KiWZ;oecOj^~3sI#_i_E+rN$bKL3*2$)=1IIp(t!e_wH(
zGh_96{}KED)$OZpvlT_L9v&(hRP}w;Y(%mCd85s`Je)Rs9bUc5bXGR|^oAT8u&de)
zZP(MAa<^-{ecjw2p0SnFWZlrSJ)hm)(Q{wzx(dgY{nkkZt{1#3%|uSDgX(50rJNqG
zwd1$y*+M5XMQ+_tuC^TdH0I;L1#}o?edn{EU9@>u461iKQ~QVp<2&>`HHR>@vz}LD;o(qW+p9f4F5iFXnU}
zX~WEVmML3cOE6oOrHE!sKrdm@1$9OVgc3~V^MmiIayWHfgOH%Fn_Wf2C#T^su0Fux
z5Df|p>yG8&KtnP@SC>QGHi3-u`=AgWY3$3POnZ7Jn8_ru^H6ma3vyx!XNA_nyRuNM
z$n0wC+(-$M_q$D@)?Gc2e+8^wx2G-)Hj=NO8v60|pZ`dMU5h72>|NIr#~GN^nlo+%
zG4|ai7haUYuAgczd|}q7zM?@%Z~FT92$2tAVfCRrJ|4o7cbQaOKhTe-8oKdo2B+sI
z=Q!l`)r={pfpASMf*
zh$%UXJFCXH;|U(&e~I}>-#W=^U&DrLrGgvD
zl#=u9-zPYp9%PKmn%jQ}3(U&E3sF_PTbe?__n0d5wdhf%
zSWOmv0sl01fBnWi3_Waod_j}(v|@6aKpB;CcHCFVDyp@e9(RA;WI~}jqCC&l(fLH{
z6JYI}SK62+k%WDw{ETMD)Yk(WRy2nw2)pa%oJOx&Q7Qb$7M4HzThfrq2xw%%IWmrT
zA&~{>
zdP~Je#^RHY#Ye5bejXmETuEQ%UCs!r%X;6_s16JIY8XaPeR?DcL!^F2DicT$S`q~a
zrGt+CS^ng*mfz)jgsv*@1`j}0<`Ds=BT~Z5$Rfr&SGy-%qK-b)V-m!I$Aeoj^z@5(
z!yx8Ce`KGKZ~a@>NDSwtf%)u3zLpuFR%nRsPR-7jc=8UBvb2?qqm0b<9XC=05MfjV
z5y^Tno-d?j8PqnT(^UYT9x(!vm51Kd0#FWS;2t~JfXvU>fxEcdo%-G*dAIXDrPI)r
z1$*!qh$2=aOx432fJUvLJu-o*I(W*ya&WDoe?#Q@y(9b_Ue!O>`_X&eyK|pYIUH`S
ztd4HoBmy^Qyil)z5kFdhk(r)zCXpw?g1YI4stioG$t14f$9BQhjG!hEISEFWEb?U_
zWCqz@$I?$A7RTyy;~<@PR5W?tj*;q%KBj61jgQpH4YLr8h}{u5+_wPu4@c+p%DUpO
zf7W49DE?UW9+CG`pc+W+iKQlKaajOVs@Pm_A`-#2-lg;RTWImI4;(Z>o#n1;``{iU
zkS28Ra!z$P^ehoe0AoS}bq1vN=l1w`@`0PnATI`e>=Q5!P$`c`=Z<*ef{t>_bRVUI
zWIfAm5E_&aifScI!K4a(R!F6}?@C;2f8iSw{C1L4)$-5_GtL$lnhU8dK#`qEoRVQMlC~_x>X#9hGlgE
zcnW`^5<9>Y=!Ciy4g#96u`Nnrf6A8#woXJBYzwHQd|-=Q(ZVrLjE-~k_j}FO`zKJK
z``J-NIbVk?TMt-Dp!)>MKnt>To~cAoUGAE|$;wn`0EzQ+eMt5J87P>z;x5^48m`ZE
zD$NcmyOh6qckq%~0W@pHuj2N)C!{JkG6xQSrMYkwdSqyDpb_=4UpUKMf7=}s(rgVv
zz;K^VGQnRLO8qjS>cp8yOGsR9k(znU^5SvgHLo0|W``{_sO9vm3-_FKk)3^PI#u%o
zl_yOW7QMLuUd{YAP#7d35dhD34^=~$^4VFlJq66!MU2`uew2+@XWW^(53fDE_>9Jn
zU44c?sD!8C%7uV*ZJ`!(e`bEEIxb2oHQ8)NosbcC*cxuz)pyS$n5Jhy@Vi6!1VF;J
z*DhkTU>Sw5M|%)>YwUi`;V&ZQjTL36X!-@
z)|AO*QwGOLLF28Eoaw^6j*!I+A)_+{8XntEbcXDL=M*ClGEgWy$6Y7z{6*K2(G7eR
zz8G~Oc*m|Kpjx~me-tVs%9L3w^YjDO-(GS@UQiaJunPWBEMj0k?Y-7Lds=^AlbO=}63=tr&3n$N<9(CD37X~pbEAX4
z1
z)3@_;{0uy2e;4q1h9aRPS2ZW_`F}!I5;%{?HsUsZv_Onna}Adiuq^PQ{3INFJPgvm
z)cXqLMwsW5cweOTsk~2qBJY#Gq4#AhUpk*op+w1Y`|AMk3rxzVmcIf@#m9rvD?R>;
zKEi)9d`Zm-f2d!?&I`vC2&3&})$d1L^8c-_7Z5;AYMFBPmx#DngrD4<{p{LV#A*x>
zL;rd3-|6i2?{o-ICXJmF|Mfi)lrSrM@x$$0ndgU8c7v{L`tX&x4|B7F*KJ!n*k_)ty)==xyVzM
zE^hA@QfuMTY>`?gl(vi8?c)38pQ_@%sf&H}v0SZHu9p3d8(ndKIPSThyi>B27p#*cv^8__q)~F=xlkr%76nJwQXqU
zuJOzQU!)>;I`h+pIhn`|NL)*iX-RW`>iJ_D_OV8&G>!KWtF_DX
z
zD~RQzv<{9E2)A&if9vVPL$%-2r?%`*ZB4zTH_!>2qJo4Y+MkfKjs_L=mKr=LTapfL
z`ljU}&E55XBQ&WNnTk1n<^+wCLgH!2Tx8C^=md>Rh0S8g2-4}P=#v~2n^*m(S4XY&
z@)5afPWWP$yMlQo(}d3;uC9DVUpFA3qHQ1PtvU6>5Ng~t`+b9RAFBGE2K7bvzWWsE
z8Qk9|Z$LRmycm%ivseTDxy;Df{E$dDnG^A!h&&5_+}TW|AS;_WVR9D0Yw!W59vxh2
z+0%i4M7NMH%XWh&ngeO~JOYj~9@#g&_lM92%8>%o-ss7>Lg%J9f*?hP(+4jVolT(LSz(PL>MuRmHTkV`?`PMbGdcFWRDNnog0&)*DUtB?zQMya
zNjThpe0eC*2Vhr_U+E=)J=>x$XcQR^bb)1iW5T|Dmp5}VD9Qzjy2&5}paxZ`QtfKp1joZPRx~Q&B^60#7NbFMo^mdBfb%zbVBo(qbgdHsfjW%%MHEGU
zJ6%xp=@&D*Ne#fwgx!8f)LiI-r4k)&SkG+efbMj>;qPbU6rJC0h9K>O;;-U9UPjA^
zn9o*gnIa6JvFT{VTqx_J6_XjtjWBNU?5>#`ggLjB0kn|MZ51i`T7k!rKtGTyjArG0
z*nvBl*;D}{OtKg5YM$QJ3BX>XF;zEzF(~h%zyWn^{Oq|AFYFmRkVA6ppxv^ccnr-R
z!(8_mQS1lA>!ey3lZb?Y%^O6N`^Qygm*~wBbcOppdE^Ai*;B*g8ftcZc|?X1#hwrP
z9@?ts&rD8BZ-j4ZdinbGm(&;#Fl`c6L8c2jnQ4lFh|IwGBsmk_@4Qsg9MDsLzk`+w
z;S##S`jB(m&oGCgVN7*;MbVe*O2dXrD_I(i{?!Jf4nt
z8_?2Ui!(qbN{DVx;L)Q*=zHf+e>C7Jp@3pbozo8Fs7+!>rY6xIqPk%>?4dj)@rcxK
zhn??yl$wqpy#%%Nff9Qz^|GUXIlQ64$yLjJT}RPxQxzEfrurCINQ{u+Iz0=c3<8o4
zNHapFaOG_bCX*@gs+?E|6ge~D`cyHQm*$PUtUKc#*Y#<7GtLpbW_>P_hJ$!~;
zFgW!?ebQ?n1AULZC2!qxZxb|i!j8_BP>N5a1XE8pRRE^!1_<4lGfaSgtgZnQ+8x-6
zPdx%c#tQ``}#B(DH#qPmE;e{*|C>vQQbhC-!
zXdt;g!x$Rel|5opu5l>2Ac~atz(c~~M3%7we!U0Ps5ai&Im^33<91hqR1eQ2w;*Te
z1QJA^GnNd=x3(6w?sVXPc0H6(EG8S?bwX<>ra(J35WfIDK)0&FN_ir
zNQ|iZ_PJ_;1-K^TIS;#PgS6}dbv}PK!9YtU2;nOmAVXu%`BPp(DG`oUYk&rYZ
z-d^N#NRyhMZg&4vPm97Pb%Y_!JEM4E80(kHhM0B!r4}sMgn~Y-|O!7_|
zFBhsqwJ&&FxX<9Fhk_c8`{`6r%=58hSeoO_941pKV6v2P_*8Kaz?6)1ILzLyF!BM9
zR?>*s!iPk4aoQ)Be28Do)(=dk&xdG}84Zk*^|y0>dx85I^yyg-{@n~2nE+%uB12+)
z&6Zz~z##R7h?5F8_U<=-zPIR2}1vivxABbI@`mUARMC0&(R0hwvM
z!z;Y_b>s5bEOiQU9jhis^KsTL(E@%6Rge+c@ev%jdS;|g%M9cJ7)?S}hwo<`PQwT`
z;TMKYIKG~V(C46^0D(H*^m@cjpN2u>yG98Lr6+R2Bm>E$8jE~EYB|KP#DRa~ORw*L
z=0FKvhyJ!e?>E6DFGY)7-z@8JBLZofT=LB*S0?lb0vayaM70$ziUK}3ZRqZu3N4`q
zJ7DQ^Z7|c}3V39)j8s7$037)2Zao&sLm@NwWl>+S@Go6o^!^ghfAEkM|5;MifNMVD
zb(61B(JX3ADm#=dL;nSgf<6K|PvrN1=rFV#E=1e5>_VAtixe4_3_QEC?~#%#_N}c*
zm5CR9d)>uh^up_yXwEekhlni6(|E!+&$tJys_F@0SmrV=T$Sn={KNZD4uJL;cpB7&
z;g2anNarHMv`UFiT-%Vu9gbDY1uVScZO#=-`tJ%t3x5uKpK~yfG~tg%sO;8%vO>WE
z&9Sczw3=MXf3fy4LXG0~`EuY~msTM6=o*j;ITzA0+)$Np1-m~(jnp*c2
zxzm~kyyI#Pek$kz>O&1$S6$XY=1Urg;lTMsq&*cgo(&lzp0#q2vc>CC2=GKRv4^JX
z=RoyPa526-7Ok>>X5%dPcT~}ToUt$@s;0`^fLfl|+yH9Jbw1sE4Ip{>`R4@EkJo9z
zf3-+Ek;1DXlbfoCaRN4t6S*zrQ18$@9>-aoPGXZ%d332gq6Mi?x>}7K_(#$QREg
zi$x0-(S7Gxc*ygeD@Ws%R#7>l
ziZa$Hqm_wHnXIgi$`oZiK3dsg6{ejFDk+yR9aYGBVY7?E+AQXf6b=2u>&&fGC$Opuq29;v!>x
z!Qrw81RCohvWC!r#@ieKneq*Rh#to=*gXV@YP3fu(FnjS;*bV55F(@oqmXD28tk&q
z&>h$mD$N*l$e18%AYzO|2Y49c0XNQWpg*t+SPtN=(F+3uiw-h!Izj*%19V`)2M|+<
z_=iqq3G5&CUwOU@S?o=k-Vtl>4yZ(Q8d$GKO|H!~UmhUdlpIv`arPj36
z3~>gjjc$N_&LZ`0fCo=V74ZCUeS30qezAm2$eLH@tJU)M3YH>2|N3e5@?>?sTEZTJ
zIItLBKn03wd%?hFn!X^KvFkpA&AJao`_@MsfRk60K7V4;99r+qFpsO(X|T!6qD6`s
zbeTtL{E&2hOiSrCJh2p{i?Z%_)`SS-!b3FgcH)DJ$7$ZB*_xk{HqF!T%RE}7jlucy
z&4BL!_Q;RBV1ovy*ry!CcJKGNG%|n>J8plZ()^Cn{E
z+uu6e9#Y^$AsbJ<7oIx>PCR>~UBnuN+j)oZW7LzVB|;^+}PH}70Lb2w6&^Dmw`&M%bm-ixOn9=GT{RL@MDsMpii
z-ah&${7?rhJJGg$RkR(SUR_;-&nX`s@*MZJlj+6y?E3cI@>V$Mv+-}^kK+&c0ie#f
z^#zV>N*QeC15!6cC7bo{}?=t94K7YPlf#>-5%d4Nqr%#^94&&2{)#dfocryO;
z^^g4g>8I7|=I$@!_{-($=5~2^H=r{t-+dh4USI!wH~w;d@zdqi^1Bbq^V_S-tM}iT
zKjXk#8!7XXM)`^wz3i>KGivEPkJe}0lH>1br<0g7qnz8uv?FcP=$-6Pj?4~xQhz;H
zt@mh~UUeG0E%S(dK;;m^JYp9L0#KSqYzn-?e!HhsU#u67R%N-bNO&s#fLpYR|E~a_JHl+cP@VB_N9ZW|DZe$O4bF*t3=4jh6
z^K9c03-0~3+n2!DHcE{*@Be#q`;&>65r2=+X++JL006R!N-&1pT9@-sR6wk#zbjoq
zQ4?@)vV&cZ_bWMmwm8kjVy$y!a7y|Z^M>nNo49_r<;BRgwjW1L%<&M0sDkX&+LE|R
z+T0jDBkV%`D7@+z|BBW$g%Zff1~Lkq
zjWB#{&+8jzIg)LrA!5CYmej!ywtpRV+EnqHko@2KD?a_&wBfMr1*c&Y)xA^VT&p7g$5bgtQI?_O!ENDYfq-UD7mis|-AU7YP(_1JmB}+|axMj-GfteC4ivN9
zIA0N$A|ipWV}PEsv4xnv-J((hMZg-qYEY@K
ziA-#6C}EO+J*ny~QWF3@?vtA71Go~jKu+>xd7vCAY{vG04^fk^e?*BSq>MF?=;FjP
zXpJ#>VxPorWu$M73A#2uIe%jA7?g<%%L&I#bRimnvmuJT0>L8x!$-d80ZN}pAAuMw
z*=O;7#}g*wMAb>41!Hd|@@BHEh6M0vJce&)(Ci#xi~+O)CHJ5cyfCv>YgHr4UhTKJK#eB!4%TEVxH;i%9pC
zkXGYiGNQ3U#pfDVrLFb`M~4P!CdP`Iv-n>sOtY~_q7zTeo-s2@(j$W?@sUTM*Opwj
zk~j-XL;AJ;i)pWiOrECfY1ioYQAA9MhSdYI({E|L1GGgX$tN123>V~VtHy^Cm3
z9-dUInAp>2qsi849e;9h{#~7aH^cqGoqsDpTNv~o+4neO8Fri7~f3erQ`%BQxe
z6mh9->oaSuHj=c|lv3n-7NRL5HOpL-axUvy1(1%fpa!|Da5XOkR!Vle=2O@8I|d0V
zSl75$I%#n=lxui56oq*YL+e*%3e5y0F!r@rksfPVKB3yqHh*|I=2?o?-+6PaEoGhe
z8VOF*;a!o6cQaE3sC{XnysqlNukIlRd_3IiN2&aCC{n&NP-H`hit`djvnWDtQn`*-
zGmG*lftP?nC1oejRT3?gSfsU4hun~d&QaV*OAVd8IB(uu0{B9(@aL^0%6)~?kUd(t
zwQO5shV1EL&0GJx985sjKj0{X~8}kX{0vmg{xuMYi;d0B^#8g8NpsK8I
zq6UDP2?8`sjZH2-O`u2|e@8)pz6nl$kq2jZ;jX75ctX(__(j@(
zapT4@1N^3jbA_Q%uD=ohc%86VHyI$%%gamH(F21OMxh;r-G30j_?bAtF#s2M3>>WX8H
zA-!myDje4*EDG%d{2v4Ef<$>C{r;C021h!;E>6M0!+#BEiiEp+K($o=cSN>JET<+dBa-QvaHq3o-aYjBAO%pBu^z0D~YfP=7cKiu)qhJfb
z;-82VED3Oc+hcJPI-HNh{I+l5}@Upm~x
z+Pfg37z{uR^h*WBb>?5}amR7N9RSocR5#Nw6@U02!}QBo4QY>ZfFm6NBH|JN2pSFX
zAqC+kL_}O1;0MMHz5~?z*9ZZC!blVr=K^r^!1@DVC^YHCF-wX8fZ7+LzmPNlXz(|Z
zk^%rN{zlT003hTy!dcq?jl^(K>~YuRw;l`v0e}vFKrjFZ{Tt#s26Vi50^l~lZxe9<
z5P$v$#AQeP4Mo5g5aMd*@H>YDE(PMElfQj%tX%)_xqu$O)nJ?h@(+mPg!%*GNV)xX
z#+@9{4T8qPAqWRJ4EDR$3po@GNBvP97+3!vm4I=T{AMMFYa9jv!8rY4fQy0s1LA7)
z{0&8LG`;_TIGR3xKpaiqU+|w(bNgbc{(ssLpxY*tT%hJ~2i&vv4-?#I
zq2bw{7TQUK^VwrFm}O&ld`ayh^6PO>}4k
zjrSZCb34o2SUW!U{Ddwc7Sf|U&woF8uP~fpQ2Vl5cPH&Ser$>2&lukI9MC2~tdn)Z
zd5;{0K&`t_w$3IQQflC9vDQ1GVJ$gQT0H!{8&8B+J8s7wc7M8(31bQ}kGYlp1MG8%
z?@5ivaf6%kwfzVEg^~U%>+PhCCWdK;z{u-aXHs5EvIFYoZ&up@p1xae{C_R@m$*>c
zMy~=Nvd1V~f3=Wux)|?s{DaSv~iEGS3mHuhULSmNCNfKI~i})ciTDb7xLX
zBx25`kEdIy^jY^ePHAc9-GBa!Y4zS?PHOfCY>{U<1!4k(+FM0(>*Q{;;_!r2I0DxBnMD?IV5U1fM>eKu0mHVaB)vsDjaoV(
z=Ph}D91n|dnhQ7u;QO#CO*3TQUms1{YIYM<>o*xxLk&WXT-t=D>qk?w6T%D|D2w@{x%$dH9k+Rxu@669)haCpT63flkp;F8APr>nZ_9s
zu~_%LI$bV`0)L>ElXQ<}>y(utXH4h_xfq(Q@T
zhm~<%8Q#0LgRO+K{x8#PLKzaWd$)zT`d@jii$uWjh@a9I5NYIPkG*x9;r946PSOK%
z){gtBVT=F1h^9R^Q-3w(%PrEw;Rzat{_<7^Bklpr|m>h
z1y*!x7&p~zuk5WLgm%ofxYO$xoR*CX88g3BrO6@(KDa)sV
zc`mrCDe=)9a*3EWrA{5?`B)+@Aq^bbN>{3)_KJU&1lu6y^mw;G{eD}AY_MT&ar9NA
zThhfHTP-(Q7Nz4E_2smGPTX7H?d@%tL-gVnA
z-xd1O-t*Fpw$05~vQd=Q7at^kw1SqVU0t&|qYy4#Jwtp_ifMvx$R{v6oc_b_pN8BL{*sBF7Q
z!++kNCsoxs0<&*hV4AX=rG+$g$};u|lK?j0&{08uBTVtoIP^8~^`txRG#)*6dZzZQ
z=vx>0+V&UU*O()ASXN?o*S&?866RS~Ury+{+de5FFzJ4$XFMctt3RswNPfZic(26}
zY);hQhAEEx40|sQh8lat5cyKreifjKEq}XTw!^afP*v}l2QBi)_iO>nfCG*zO+_50
z)Y+6T%ZFh$uhCv~(VIz5-LE>W`xy_vo9r%Li%}|gIX-M{Y5_P=98j
zR(%f=w9xg0<5hhs0Li+bLvDv{8cQaPQq+hH>%Fa6{c_oOMH&vQS+422|7Eif&*lCh
z<|A!mOi~MxuV19N0Df4a>bs*^J6DMu7WhWXTZJV2YnN@#C@oC`v6<1<-{%`zPQ0c<{;0D9
zsHW{7PCGX0P-QjbGZ){>KJcEV{276EE=AXPrZ{2x)>vaVRD9n>6BLn^>=Y8k;aAUoWouKJ?oOxdy9fyp5jJ*}^Y20*-sG`AaxtwdvH%Eg-&1Aks?>Qsp`lt3O?CkVdxv5#$1j1_T4mKBEmUEcW&O6DV$^~gTCnp
zpaudmV~ejH3~RQr=IO*M5i~#R%z{Inz@|&BYBrCwv_(sn(0{5rcLYCa@|80W4Cwh~
z?@NV^9p<3tuBoGQE1?W9KwV{xsr$>JE5lzUvZOaYdgIH=g@<-I4hKOwdq|V!b
zVU@@+IyNPS8j%xwk^?quwxJ#wx2I_M{6X@Gjx4oNdwQYUIG1ynsN=o2#QjDgIX`;b
z8q8w%sy}aRgnzF+6FH>**1qia?IT63Nuc288N?n3fw;oK`%4RbV;m&AL}5Ya&T-Ny
z4+}l$x@Vi$@xC?Ph`8=}w45Exe)M{HnMsw2n*Yt(R~RNhJFn$~&`CH?u?T0*;fO+F
z*OA;hMLk(j(z{2~^e*kTn)k%#pWK>fe-}ZyL%-+AHGi>79Qa6p!b@LgB;c`5V||)T
zv~NQWYzC}IZ<2U3iJ!L%_|B?9W)I5scC(^csTwvZ#AQLcnuR8JnB+{w)-ua_$4ebr
zb$uDI5crBA>p7&n+fIEgpOH4^o0`x431Z{@^n&xN{>ere&HntN@Jtf!qINQN&rpoD
z^hobb*?%dCS1&OAi_ls@`5G<3z&#R`@D%QxhxJzJGx{@26qyB
zdyz`HcT2BI??Y3oQ@fu$q+IYD+}%*=(9KRFbAP$N<)-Lo+DHnlmTcQ)~=C9g8$MQY5IMi7?|eXp*+N41jHnr0d*?l>k{
zZGTBwE}Jlv)ZD&MmV=Dw_`uRu3!yhEuvWI=oDUE6>GSorsgFL*=ssaRp{>{%19-qy
z@u5FdZBeJpZ}NG*?s!Jz->Pv};mCxssvx*&_U9z*o_6&l-uUdZ^US61<@Mz_h2@);
zdThPbT%a58vO>2G>^b|Ig3dGVDN>nfcYpR-oxcujvuC~>!M@m(dnk|C8Zr{mZ=Py!
zQ!wLyC<6(Yx;w<&U0L$Uq($DOY9OwM{W5{N>Gc5
z8MMorXVz1AY!DY7zEdLro_T0J{kWA5iks1*1q7soNGz7$OaO7JvN6
zmhy9S#p|Y={tu&pG4%BtVJ>U2xsDEZ)~S82H=c1n7pdr{Bkrv*DU
zR+S>>ToCE@{lj|sen*DtZz#c>?pgM$-+Zkf9f%;bPU<+s==z*9JbQGmP*#(>K9SKq
z6w7)PZ0TM0lP=2JQ#bfWcu4JQuYZPH55XK|p!
zyqnNrwUTr1@FruKE!s%y)hX&et1xkG7=Cr;8msxJ0Kuv>)Q}X6KYnv$9=iE+MU=fRo
z-OeIIQ_FtmVR&zO^0}?0X_;!eBy*Ix|J)Emn;u9OBZZf8X9eWUHO*SUdYvG}jW{q}{YJHTs~vk`W3)IO9hmnqd#cOvF%`!$2ooB*(8au;
z>nY5)D<;0~*m|n+Hh({n-Se0_fG{h6Z)8+W_B^rM^rlAY^@RSHgeSQ|6|cY88@5K6
zI2afW$mTFDL6)+Vlt>1w2vUDck+`+Eb$?54rs`?KgOr`xX1eFgl58m2j`paGWoax7
zbV+Aah-}jxo1NL^zdcR
zxKS2n2(jlY6^b2L0{u-23yfMo(^~hMI5ihJITPCb;N8oW=Sw^r3^^ubiiYck(Bvl}
zJa-P%1Tr$MOemFjPw3CCS7(lGqldMkO0(_P^FrE5oa(Hr6et@ni?3`wR(_Vj%27Y6
z@HxP!GuG#n`hOMOeez;6H+SN+O0zc-uXb;@r61H(3<6kt9ZyuFKGQIVCVQ22d!Lpo
z>Pzx?GJE#!&_1{NS>feTU|Vnaf$t~~5(K==;c<1OfT+y#3qjN|H%PfjGi;b;(npQ)
z(cF2tbSqy9_fq*T&0b6rUEWyDoS>*-HLo(4-8+v;Sbr0!EKgWDdcRj7yoZxXqcW+D
z`^OGK2#e{w4PY-f;e2jFDiPi~DNfC@MH#ae#8|V&SOY9UYW`rm?)4Sl4=Q@HU}{+4
z&(M;5o@?yNPLygwur;&zm{k;C%gP>D%T=p%Ej6^RC}#MhJ_a8lD)Qp8%wA{L$40W9
zqlc5CZ-2|l0+w$t9X3nckfL`A>%4~FyhGxqe#arEqI~E1_t|^^&3kK3;|8}=_D3P~
zuXFsoKZxWA_tAc)i3CYVKd@uMT$%LjP)2Q&i|2Nf<$`pT?&bK!>yH#AMGj7}4VHLQ
zaRt)_Y^yZQ?cK|fn^Fw2Gp5!_jB7u=dHnpdEq^&U=)oeVYIg$L?39cxQLm+}+5?@u
z@hs0{4XDZuLk6JlwDd+4zRys*)grd2c1`bCsEFn^=5%cS75=ZacWw%zuEqfsDIQVX*4w+YOyUpn}_!;;DuFha=I4
z#Z9`Kl03}cMlw%z9x^QNCQoUq>J51vlRe?v9V&lJu7?}~^R;6MbTw0^Y`c-qfoH2z
zXv)$Z_Ad1*Yx2Ng&&goPl4nqDJM*_USOgNB{24^|LqxO;-hG`~^dG?Vy`-TboPU~o
z|5H&B$Yxc&lycP~FkT?KR+9;&Gd@GoO0RI7`B6Z&sUEVN_I>AukMwNF`=<9I@0*)X
z0rD-fPJ%qCeU6BjD)E#^yFi!DSz2Yf7BWqSNjkUy@3+_|RIDgd40=mQ8nYAeMxbKYsPGDdo`S9?weFL)Z)^C)Ny}ulZR(Y#J>{il
zTlx{E>b5C&b98QksRKfI#;=aR1AMukpIrt!w#{pT7~<|;mfL>)c8gZ2nSYlBRbV;E
zZ(RHodq+yaGeJy$CelvDQtW5koRE;fCx0i4)A*-!B(u+tX_D5HeHMGZ@ekN6GrcIC
zay_Gj=euU>&(*bL^oY7Mcoxryr6>34U6Npnt^%%FDS75LeN|5L)tWCPq9T9fQAH(p
zO|i;FcJ#FhAj|&TmBoSnV}HvUt*aDz6c}jP6~~(6y0`Rbkkc4`4fU~Q?~`>IjZ4>|
zVeciT2Tk8r8}~xC8+^Y_NO?Ia4k{jgRP;fut2<-H7-bxobUaCftk<3@uOehg@?;(T4aUTpvuY74(@2Tsz-0Xe1N7Ed7%7aj16AhT0pMQ%|>y
znOH)OC+Duo4T6!>#+xzW5ufCS{pViwX|Y7mM$<$*K&n(0_DxiV2yId+XC!G&WIr4I
z;D0>+vUbR4&!9G*Vt={)2l2?wFN9%Es_9|iU8DV+cQz`tXPIi;xxtC$PaS98B&nmO
zta{yBKT?_r6T2B~E>Ie}mURu7b>wgao)L_s~K9SY2H4GEVKD$dK%M6f`NcEWAe1K&uNdZp*uTFqKo+MZkAFY5CXo4d85)-kpClR@
z5gj-s8#S7f_{5qgTS1;5Lu-!rr5==#l%yh1BVUEbyO^3tqbUbad`>w%PZlR)zDIi~rLhhwjEUstY}4z&&`3Ll@i0+2CiF-PT1
z70s`JHXX)>6n{vwa4)tm(?HI
zoCHh<97O9FmwuFNe+ac^GkScJbU0yUjEXf3nmPHH>Co;DcACFfZm^|(9-&^!7@1Vy
zV>nPga+kVRCO!4~Y*;QK`^zQzsJ&Or`DosOm3430_=DOqqMd?$
zlEvr=L(9$3y@)nGy|cG*Hj}G_DK!nSxVnluskeqB~nDrRBV(K|i7J^HxN^U8beGs+m^hJ~bA!nULVShw*=lJAxpiesWPM`>vb+^v$KF
z_KQFYy(@!$l_3(hx-=yY!kW9Y_*P#+rQY*{k8b!c$KI}bt3b@*8#r~RM6EUMb5Yk=
zo8D7C8@IcPcH?DaWAr86?5?jsvle!W)5@ku3COk+I
zVw)y{joS8hT+SeIn1lVJJ*W#4;lwx@!
z74gW~C7$kBVE#m>r1$#`0=uS=&|Q?PoOv*27j)%Jfi=Jd_Otp*Rn@s0Sz+-(oHi#XV
z>2$_+;yO<3q~78rlLuO&ZQg2&L?w2A{XGYuB#;y(EAiYmp#TuecfPaW01T4}NeGji
zNG0J{5@#*^YQi`(P=JaO_<&y43Mg>GxrG|^=gMJ=Cfsrup)3(XyAIO{5z0Xy&55ua
zhHFVA6MxvEiIj#a7U~%fD+yt0g&LmwIDe#mH>%{5q}5cIt&(a97Q$)3Jwq%=pqzg2HZ|1
z*aCetQVSD0>su^w&SD>co-hr(AjgRSoXi3hiQ=-uEYuQ%15gzSAc7L;Epa$Mj)x(r
zDI@|Kp#(kPKi&BvrP%}d{8IrG>*Ss!g1J2fN)mgbEaSW9{w+uQXP7U3Dk=8ZyhS^F
zGAN5JBH7bGAKP=pCrviW?;O&
zi(6m;^VqYI%w@^r2|ea5uAEdc28)$tZlEj_V#Cp)wdBLM>co_Fr!25rj$8D__tK!;
zQu(LX%<^4e2}}6bul=?2MROPrf1NOU^QZEHgQ?&wz8trQvh@;TUHZC=dMw|GMJc@b
zm1jAlf0j0|usS6eFyiYH88Xg)fi7(53!?=%tMxiUPEe89+o60AAQU)+wF!*SM(v@)
zXy46uuu`I3Xp2+%vK0=A1=5V=Uj5Sd!-W}!s89?~ZOsC;TCx$WLb~NUsZmcbsRC2;
zzs&m(*t{tjz}Aw5(CUP@xTRM7xJBRaUMgHg`Vn&3%fAZ@g>v~!lCPbAKmO>@%cpTh
z=+S6cyvnBCNj@${la8;Sr2~9(`TFGg^Y4eJr*95e0&n{1{glYi<=|lSNWSbsaNFe=
z={bf=1vAtAX*U}d;C|5aFdZLf`TawIA+|$t9elUT@a3B#?dRS7;e9`Y7HGIAvcWYF
zEIRb293~~s@{mq0vV#17Mi1#5`j%eMkMxFqrc?TZKGG-p?@qvUl*3F+3fP~{JAeN9
zEFnK!zCSrV`u*MM`Dl<1S+aXD>i1U5Dj`Phatx8^UI0_X(z=SYSZ$X!S;{Kc$oh)D
zrpF-h34KRT=^6b%&q4M}n$la^r9GO_J9g692(}9yed3E
zYP$__?~g9euPelv^}Ml?HUk&jkf3@t35
zbZ3LRJ~&@JcKYRMHt3=KWZ!0KpJtC;Fg}c`_$R0S&vyv3eR?;VO;+z_>n&yuS8!sLvSw@Sz--G9m
z5k&9Vu$N7y-O(gl26FmuHW^jTW2N~6~}`XM;@iFfYvMS8PHY|z~*&KA;|*vI1ZFnB|QExpwS
z@agFI!-w-ZTW$kea&8;`a*1*K%WS!1EnBamPq`fLZP?XF5X{V
zTpztH_r;5+!R@F&-I&`svg=bjyfQK-vLx7lCx32hlQO=rOog
zI+=`qgy!3H;Ed`vRI=tkx~x`DvbDo#^wC|5}
zLFwgzk7yYm?1+){1Ogjsj95u85!a2^3vn(e20s_j96t~&)q=NEK`|IG=ywQ^$u`PK
z%2-#8?TgSyn5cvZNW;oaf^ZdYFX$l49R}4^ZXy7euik7n>|L-)Jx>H3jz$EWH=IZy
zIJ7vU@Fx!YyvoT#0&7U!&`vhXVns%M3AbU4I(Z3b3|Y_r?K
zSFsVI^g%k!@Xd-_-bx_$ANVlD(W87aEie)iZ94Q_T2>f`!}Z6!S3FFiY*z>Nb?diX
zz}8{tIt&cct6}&rl+bssRR4W&YhY~`n{{Be4vgDYVEiS3t+{EaVN(s4OYs8=f3r~y
zn&u6C$`~3m5j6P{(Ap0ob!c)uG|nPu`X!(}wDm%z92z*o1zhVNG5U7{lbT>~KB9^j^vQyrgO
zhc8yIcnNfyPCGSprr{s>RdGq5e~8q}UjkqKlv2mj@CO~+n~nKm|KBY^SNt+EE5RDE
zE%}MC+j{X?N7gVX8B>G@)CDs063E(bE$gPLoGV8&yw~bL
zed5zuQ@YCAyobprR!E37B#2DHCqxoHO^|RC_w}bH4-3Q4q
zaF7UdFuQ3Eyh{j!!oEbpp&I|8MP(@720*9ZUX8lD7X=(|b^xL8`P&%g+haS##nog7
z&N*=dc2IEx#~FCsB6cvp4$IPa`5-T9D=zvqV1*ZQ2Tw`Sl&}VmQNRi!h6w{?Tvt0p
z#US+#2^Whopv7AAKC96fe;6CjZFh)?=XStQDgHF;;vw}v$Re}#HlR`_^+xC^S0jUU
zEBzYimM?)G)tDJL3Sla41Zl%ZiMSEjYM=$ml&d!)7qX7Ff}_%T4P;}y@M#;6LV-`$kPY~B4LJal
zGp&#jN2T)`$iSzoD;9?HIu4Pb57&hgXvIlgI^q4jt~!A}pXhefZl~J29koF3XRN+}
z!q&Q%lk)>De|Aq(JC&{rSmm~1g6yn}Q>^qAR=Ni><+ftnx*)yM$jTMUtq%W*Z20BR
zRDBa{=~T?hXMu?(fb0VL;Vpb4H#kh;>@vFVgiUg}uLiv_Tjp>0_H@+C=+!iPKP&qA
z5O>t~<7~L^mm`Za-v0n?hxVijWo~41baG{3Z3<;>mp=3ZWKu&yGBz|pLohHxLPa+>
zG&nLsLNzitGeR^)HA6u+LP0(tJVQb6vy#%?l8deT2TZ=7sbaai?9eHK6r?4
zKtWOQEeNGnR`j%tiLo&;+WQCCNk?K#tn7@&66`FDl?CwqHJ|+Eo-_B}$r&@Vjm#|G
zg4=Kh%KVcc0!|DQ>Ja(VF%S_D6niI?S=g;^+U~Wg_qjFk4GKh07)&aZXE?x
zIYMM_)T=mK)OR^*Y74nWp*(9nw_wS2J
z)t)V^DdjwSKdDsX+3zl;S`Sk%lPx#

delta 20880
zcmV(yKS;)J9;_P@{9-PIJSAvIpd2o^c9MRr%$tE;}M7EkUIp8Vy8Up)VG
zd-L+G5=p@dD{XRnmk6D+OvuEVj7epZ+imi5`i_fi^~>#FhjlOCifpuqvm(cPo_~l$
zn4H4%lEX}LeGL`t9Dal^~wc78#Q1z1&W
z`lju|mvXad+HHNj-+#hV5nk5E@NCX|H+SK=tu}3i?aFTLqypC$ylcZ`9#{v})mkYv
z-rg9;Z`H92olF$Dc0-k{1)lOV^?xWB7v_=~X_zvk$iG$91K5H5&2j%JjQSl!X}Zc0
ze)$%BkLVFCsTAW%O=<1Jf<*?TDwvVkbkSG~I%vUhk*E8zy?1&J9F@l~D#N3z_OMsA
z4O9DFTXnl;zYQM;eZC04w{DRi8)pvuU-)RL)uyzO08_NUhC4%~P3aZ`oiqth*otBpys
zI;UyY+xk!)JLs__`d(>%+7@Kn(oOSH;muv^PTL{;2&x@NKgf{TJby9J7pxwHIJmt|
z`@V+JvS2jFoU-Y$X%Bu=jN$fuTfaX!gN8T6Ek-^Dy1`wu-#0G&JyKk<>7goVz7Lsh
z0ELGW8g>iL<))U{;p3}oCj^nt!mka_SP{5?$EBqeYP14bB8%?-#<^+k!}iQ^5}K((
z%$z4RIdhpF&YYOx5PyD{4@6$iP|!@vJWi052e6sj(FO+J5O;dKtZ+^T9=XmUhf>eb
zhe1uAtf3d0r&nv?=*!>)i1hy45Ch|n7h6(Bm?j2k=1&LL(X}Ws>zwcCe*&25qr#nRMy9i$s%w3EER)Y7k?#?8D~bT>2eq1^0cXa
zQt7WnZ^f}ZzhL>NXINGOp|`l?c*aTNd?{I>E;hM)&L2P01e;j62SUB~7h7mJ1=Dz3
zacgahWHjkD9LHwqEBCPBk;TXB@G|aJjCPaAY)VbG`*t?98Iz5-yMM-JDi;W7m-=XZ
z0EEx=ySfWLM}Of9!O=&s7t^FeVNu&0=Zpdig)5bIKypxE>URyP#^qIf+|ul_-gaRWb${1YeLvvW(?g)pN9qGKxCmYU
zD1VJ01k+0*MbHDoD=a+j>Wy2w3u_;A+3`A~ciZk_*@j
zn5-b9ebXEsP99d|QP~)D5eQ*SI3bT)=Z^Ssg`?Yh+&*467-^=?%ZA9HeMC?N@;gG0
zmzrER_pr-hCy%`F+ez+5nJ>cJ8)u9K^rUhG?SG6Qpbs#b2UThTmy43Rmpd8^l44Af
zVy@q7JdKbYJ_RM^2f^sW`N)q*Cs#$=YP&yw&To5ZlmEUu3d3ld!I!0;(oTio>u1?
z@6y^bM-5UyA|sx9Pd?!S#a-?m!^~M+C0NeY8T8MfC?GU3WWl>x6m5Nj;VL_vS*8Bx
z*y<^>01#Owzlhu7Dz{6}%svOrCSateMBT+SM`b_8cPK1p1X^2|7=hKDg5o)+aDQ^?
z-zHZ*vnN7RZpL!GnQy0z9OpM}Hc_HsMveDc=-x~~-}51LYPHLZI7wf4D>D$0f(o=<
zMKS)9vsH7Vq~BH%P#hz6qH5pcl`f_W2N+%unriDYs%cLBwPMLN@3fgR^VK9(#`@eg
zp&qf$2k(?^efNpbXg?%5|J>}UUw;{K+p-ke01oOUAr{%w7P43|UFU-8Wml0>wX~?_
zkU~O13eXyQXJoN9{Gx1)(Z-!Kio+oY$3`1QD(Jzt=v?oP^k7L3mh@l)`kK-xd!n$T
z?j1nteSQ4HYHhUj5YPF|CDR;=A{r9SDk298ma%C}eW>61S%a~VAFfYmz;M)%?t79XKE!4Zk0COx93eQG97Q#F!M0jW#R@B}O$C#fB5yGkX
z7I3ACxKZ(##KR9*f4%6BKz}}h+%dXGkq>Q2Dl3X&goWM99SsHzVC-6f8MWAlRv_Mc
zWE^i-B5{CPYj^~;{~l^N85YJWM^14t=up6Q5A
za+H{Z$2b!F_W_S5wopvq#TXy|m&i%uB$767W5WWcG_%)0N8zF(vVi(};LThIIWvXT
z+5c3r?0^L{1`B-^4~(=v%LVoC&jr=zxnO?X1v$$vJfBFyc>1_sK&4QfNfF2!p8ZSddtt!qLt*$4i!dVPv4&h5l?L42-UF
zJ0<~o}X
zb;1?F7iYr
ztDC!()LNuTwn}U&l(wsz-Rg(vPgQZ>)WxBCU#>SQS5d#`Mpu6vPX}&qo3&QaFYCla
zMOzj1*7eKiSoCf6jt4gMY;OCas%Mj(+N$1Gr{dt}J8srL-FyX{Zln~sWo}OE4Xo4D
zWsU3goj{^bUdwd44G~KN1LKuk$;?8(UM0o~m91`$Fvr);TBXqg&XmCeeoAcTzo-BXVJ|%gB=jnuA!B?
z#%&fjSDtY=3nNzvJhZaJ_yaRww2XGsV=~4sxvog7?O(2!W%6`rX`De`G|jpWvqpb+Hf7Q-lR5;-mEj}
zH@q!s;{Dul5ARq-cjtQBS53{F1u8uer&Dbug-&Ua`Xmk&i11#N)%|rQWeyutX)xSr
zhS79}2F#%Wu1#j5zxDLup*kGsRa^Gwwx(WE80dsWQ8V1}fYltFpLoRJ83Q{cFzsKtheX=5T0m>_b)G6H#Au
zZ@W*CpTX~a@)|UAvKJ$AV^$j#3uVa3{20r0HY9)8ek}6LZ_Z{S0a1Yng!CeN1-5_g
z(Xl0#JskK&bO(7V95}mY3OLD-vTu6l2cZviBY{1`#=TR8&TVl5eJXeK?-pl!
z45qNK1#Q}{r#~kcK6t0#Xgv2RIDAFSUplU9(%a38bl>G?tsW}G_a`Gf07QQ(&
z&V+wrf3$gYEQ^}Dy!c8d0PNWneL+-gHgN=Y?Mw;#4vkZWVK!)xbU-S+%rLGiw~cGV
zz|(b+nFPYjQzGmBA2PdOK_w!!US@6MSok;i2;!e*mwDMi*Ou91*}=O}b`8C1sDBXr
zph%;9;q`@_P(dm7;WC3fHp{XDo6EfL29bZgBqQcEz+OlTrUrhRY?5}M-44+48WX+Q
z7d?%2J~?zQ;4yEzvL)a0hu`{qXh=-NU!IUrSm#{@fwl}bVFr=IRy%ga*>om|x>5%I
z=-qhIE|J8GoP(D
zGC>4Dw8>;fRK7HrF<#)@2ngD%XP3oj&fG7p1?C95Sy(I3@r_bO$P7gekWDM1eAt6G
zSy)tMVB^{UzG^DcaD&~%*<}et<0qIqT5d?md
ziBzhUfjD3kJK@&|Cl8P7%tq+VVswRv19{{a`8iO-ff{OdeR;wi#fm*2^gXmy&##%E
zmi`gAsp-!*ufC+l0Dox_KMU@73HN&^)+7h$)bF8{`e=U%U4DK@
zHhN^>k=}!p59gh8!$4GBL*F{igoU+&HNV|fZ>xUu_p)f~XHo>fr8dsP;c$c&u+m?P
z5gijJ#J6Mc
z1szT^CK_bhJ`Zr!hJ<|0V9z>YbkbAyp~vjw5zmb*6uSotg+FAWz}dt?we}{4)WC8(
z!Wbg%${wMq@CTgFJdREc1RG8@zP#&chSb}zp42p;7xEwK1QxsUKs)Z}-JZB*B-l{r
zI2dahYdMuvfDKbb**t$aD}28mDk^2|x$-Eq05pWJ&M@ck;;hlWb)p1>S+_Axd^qbt
zxiB!C=6hvuMS{Q)z$@x)be3L<7Qui
zG9I5P2ms5_sU<>iHnuld*ggP?z*9r_hF0X@JZtNWVT
z)1nP%>Vlg!N=RTa;Oe{Q%P%>|Dj4&yueQj`-gAcvV~s!Nk{LqO&wh(D3b*s8oU4uv
z#B%Rkt8RPlIvT}RnMTbh2K4XJMz661lIK1rwQJLD+v)i%x=i(f%Hf_=?#&IBpWujlnGSYfxPnKC7e
z)!~453otH$aATKXTmT0#a)-$Ga%rGZ*&JnGUrH`76=Hv4M%ka@BcVs?A4_}y2#`9z
z2KSMXBw*yk;9La&SH|g>#KiI9mXsQ3S3WO7lkgb_780blJK
zSEA?fZBci3O$(l3S%^awHU??J)@S@($`rCN1xV$(P-XT}Lj4dmR*MEPQRCAS8fgZN
z1pCoxxI>3CbAJ#VLdJU>?_@2m>0<~xRJubMZeo90EKYTOd-j;im0i)v(xnX+wQM|%
zaWKd^X}pQ4j@6-HI)C}YQx9cGr2F|?P|S0&W0;!b%@UMV#$r%(tk3lZVPDDkkfZD!
zVy_mqXe9%nrE-x(ZAmdF2bq%izgjH2m`smnXyXM149Wy=mevCAGpN(E4*dHCHZq}e
zHDP~4VtCD>FGyaQ$ROY(4g(JE@Wr2RUi|#R%Tncl;4ms_)DF4Y9$)B+R
zIv>w}Ag_ypnbVoeC`#f2M+GZSz2O2b*i5KCdmiLtK#Uc*seBR;F9axUDs%=uJBHYj
zsQ*QP<5axrhhvcH`i-DWC8$Fd?&pgukw|9XxZ>G>H;9|!=>5_?V&-gefu*M;tYUv~
ziC|s93NP_iH=Pz17r9fA?N~KCp7)EUiAL}-TtQA`r$=z)x}%XUF*A?}aAxfCI{ZH(
zaUvt&gl`x)A$`3NqsOox1A{tT^j8C+=?$#}h0-%QVVZ$#QgKRtKyvA$n9Q;0#98|H
zORz)=nf$g;?>9juABq>bGF;aFDhYoQH5u`JGM9Ap1Oytcbp=H!E+@O9#9DenU!s|y
z>9MGqX@7A(F#VJb{S;3fVE>0y%#CZ}9`1Tq1faPdCV+`wq#+SSB^NcC{hyO!C06z?w
zKu`RDWc-w;8@~~GqP#1?H%ZA~GL^?u)p7|7ug06P)Jv5!L>KNH4le6pENT3AS*YyR
zvO>WG&8e@BG@Fe0Q;LmKRnUJM(pGv`)?2DY@AwQCSv49wfgT->O1bia4Jvr5R5{Pe
zuWD-DSL9A>BDm#x4!$br0qR2ynpa&`LY9jf2;;!{1gt$3GoB3@lRaydLedt_OCi89
z&CDK~u3y5{L&?R6*vN2}!!rwKxx1ST#!Q7FQ8iVT4#?$+#qlBw!^3}U@eSbQ==0AB
zr=PC#g5Ls3JCnl8A(xx02S9khOJORvqb%wc&Es*J*XcMkDU}Cb0}P2C)z9KFW%Gsg
zIQYUL9;fAc8YnpZ?@mQMXJjdzlkWQf?Jv-{3_br~3qLZUZ?-fSDyJ3f!D8*D#o`4f
z?u*Bg#i9j^;6sF!)({jUKv@+iLY9%X<^c&n)&Bw~_we_g<)DDv=-=X~eJ
z(Rig*RL-cPj5W$=Wuj9iE32b2MOlx3t!%Lh)6NBzluMY7DrCK|S;b_P(<+%jHYpV4
ztje0iDtioil`|FuRWRWa_Q==~`x)@@*w9+D@YvKwP-9o8i}Iig)+uanB3K6=EPx*K
zNm~^_U$8g?91na$f5c}3UBWp{S%B115*){3LIs@P2BV^ZNGT~y*$|709-o^008N2v5`Za2`RvT1
zP(=nX0vC;av1QODEBpoHtpa|*fVqKe!GXJ>5`8A*&{6?DHS<|j;QQ1)r;FC*&h%k^6
zEtbO=ON3z>M`1KjGPF1yEqfwY;%sJ*=My{i#xI^cQR9gkUtYhtR^xMZ8sRvv)%fCa
zb*8@m9@C5Q&HrzfYJ7G6ez_QbxV~B~uU2;|lWsA-e_q~Qf4sd|-obxl;2+C(m*>x}
zKdDnq7c?{hd$z#xTdaeT3R&a4-S1!j{tv|&Qp^g73_$4WO
zof(|9b4Xo@LxCaAAhppAu+Lef-VL;xo>W1fAFgjtZq6^3unAf7>U_0Y-d@2{Hldch-amyD?;A$cN7iFzenDpD
zC)$0qw!nzA{jJ08Apj?UY&`W|cVXXP3C_-fb&Cd+!J~M-e6<=
zR-haLqE~ni!~~#T`F(iqb6{V@JQS3)6K<0|4wN)M+@?LDE~GL9=V#G+9S589EL!Wz
zpdU4Y_P9)(vB$Abn}=$g1CerKKV4LQaB9tF2kUYi)>+TSdP=&!(-z!m%`93JtAoyu
zdQ<3?dZ=caM9P0d#cL1Xp~81>ZS)9^n|JP=IUEV*{Cj7P^XsI%_ulD;$E|q}l`|73
z>T&wo%SRuCAL@W*C)k#+3by0ZtE+49IpwoMp5xwjGQAj|UEjW2-U=suHvVn=ar^;a
zpr|u$eSssJQV>A-eAEpQfS^C*;^bQHT?XCB$G0o+9RGiQdG+)7^vM(1VSIYAy1c#`
zPsV?~{*nKG`f0Vgx%M*VjMajlY~<{B(J>{O-f@{Pybd
z>iu`-&p7bbM#}u0QNE%?FMG@Gj8Zz!qxBh=6D*G8(bQ!H#3P%9P-RVxa!9E(tt?#TU^=>rlSKFvWL65*|iOG
zv~8Gqw(*Ds_x{@LOUl?bN{u(~|9f-${;fLQoMDqGm-qJFP*1w~E3waE^4J>IY6_}St#7mKwHw-KkLk1=n!zO{+#ciTYAZEZj9oQ#Js
zMBJUYQ)^4&Dru98JcqQw%4}^ro_(8wp8OlldTsrYiorQ++pmg_TG?&0=Gk^zW96Xj
zBa5zDn}UD*H~oLM&4R_8w$V4}R5+Gf$EG4iFUT@>4fOsF59`=(_xX*OuJ1v`(r)gvE6ozY6ZWgG}?CHWWYr>s0%?DL9_*9uZqbBE4
z5IW<;iRnNw+l})TaVcsf@NEpxb2hdR)3;l#)IbrihHn}a!s>*h)uL%a?=w%z*CaxU
z3=@CV2w*#vE+WTd?wnX=Y4XUFKWNJAeI%~H592795%MCfE+oNHD3b{cg0KLL6k^Yv
z4MQBxS5$FX9GaSm%?%|?@|P!7y+vvQXpj4(X8L^7A!t!K$&=-Qa-^^s+e3Ybntb~s
zN+cm=tbs%qC!RrTjLB2?N$gfe`qmgmH8y`fIqKXoC{r&iCmc7?h1v+54N>eB8Z7cN
zeB^r`p!A9K5gLOf`z+q?c*10yK%GRjVC=0#-b|L&kf1yokKxN1G&@H%#sFGTCHrQrLFb`M+bs56JrJDEdG}Y(`+n~=)_aAXUvR}
z^vEDeeB=@6wI$cBB+kOpkbbTIV%qCfCQnoLv}^76QAA9MhSe)%r{B_gr_erSRdaMz
zo;{J5F}0{CEIS>pqYOW
zYd@7B6#;{FLY()l9Hk_DjnK6cF*HginL7BuY5*RMtigArkk6P+Dox`o99lo~otqlw
z>n?3yX_3UtWFShJ(JQ0%tCa}NL{&8v*E+B3I|(X1O;dIqk9tUJC1r}z@=!-#X*|j#
zgd9_(b?aROMR|Bqtzu$NpN%G4t95_K#rb!2{@o1s2Y3Fh0BvE=e`M#Ui)XrvWz}VMfMS84d`Gjga+th!{G0#%8
ze$Sg@Z7J)#*GO=h4)2OoyqlRSKhqRGgPM
zngs~CN##0T%`D2J1YQCPLCQ{4S4p%~Vv*KH47ni>oujysmKr*Fao)VS1Q?yX@aL^0
z%6)~?kUd(twQO5s9gJFfof-q4heAilhHnrL<%gvyc3`rheJBb9Y
z);kcGzgQ!Z3JCNXUa=!(JFa!s-D`77XNz9}+gnZIgMpp5ft>+<3=lFm3NK7$ZfA68
zG9WfMH8Pi>*#Q&+I5Cr8gDQWuR(CuU{{Od0*^!J~MmBe{N7;LCvJQ8g
z#t~ouam7GfY#}HBHbjDy1E2yyKu}n}J;2r*pasEz4ZU3wSR)L;
zg+u^sA&y{YC;$ls7(#zc0Y(N&`UU_MeH|lR18xCqVFPzpS0w5`Toeopja2vmigMb9
zN&tv4A3(*(!0_U?Aq0Wt@4yGpHpKcb`d|Yu(zTQfCcYgF6}{fJ8Y6xH^CTTLKbis}in}oGv{w9VU1m>cq^RI9K
z8Up!88%HquXRfBMt|q_*3`1Zb2r$A98;Aj8+|dA=pEm3VV$b=DAOxV`jzV4Z(E7`T
z`n%0v)a8-bm05o{`}%`D|Gi^igge^jkJ*?-8ivN8e^rD4pfG311$*?xp1}}5
zZCY~LYRXCmhWwgX?IHNJkl5)U1TbEhpV1ffO2{Zly?X3&KhI}(3iSEMTd3U)?A{9#ZC_KVaP4fcQlFerD3zwe)p
zep^+myiJVJaJpKr|BeQ`z?{AROZ>m$OduD$xwMfe7qIicd@!^!
z%nM?#3&Yqs{u=Tx`>#@R2nS~ffFC3*0K6?K^vh#(aoU`*vc_Ibu#3y>!Y}$SA6Bt;
za0CR61_%Rxx**ue{EIzyJ1)2b1eFyuP325@|A&8=eg-Qc?2z^_gabfGL=*rbXg5ci>8)^4962{iF!ycO7eh?4{5VZdTf&hY$zae&hf({q&0ro!l
z9U^}M5QO~!vE7~jhC-kV2zIfx|J_3r+rs%`uD^4ztX%%cxq$A!-5_iN;tz=Bg!}_y
zNxA+`#_pq_D;R}=ft~GPQ0VWqUf3Zh81j$Rfw1fUVj|6*CuEA$M~hyuF`xD;KG$Drp&z_yCb_$`MTKd^T5;@M2)kUQc!j9H14zNUkK;
zlVL1i6qKeldB3=km3vJ+BipP%csQ*#w)O_VsApY4qQ2?qiBSJJ)lL$tMRDziEJDw`
zQBd_&lry#0Bqc)wtsxWm-`A^vnO%A2`-^_YIXrb1ra(VK_2`?A;6kv$=Mr|IC?4>!qF
zGN{OviOA?~rF@x4;;wT`=)%O5UA1u_C&gQj;qx0F-%c1N@syWoM#Ku}VHuw`NC;F<
z`ZUZZm4rek;hmfA_4Psfxy4%w+#gIL1O(z1kk3Dnneq?eOQ#wNbQ
znd(PCnUOA|Vg*sON|HC~oR&`>vMD7f6
zt(#!hk)d;^L-jf4a(D73tmK#8Kn>C`S6^P*k!CJCjlIoT)$`TY+SY$`peBMVoCk1y
zL~-r5b>2`|An6_FQD8A-DKAX|iEj}Q9BOTDC8p|rX3eW+?1P95VbGhb)p6Plt35w_
zf>#dR4ZN~2ItpM|L=dllFR%_2L-z;-!io*njOZ0@NNh%HXd|7F@Ar#Xz^x(pU3)sJZ%{L
zh7Fs{asq4AkW5zzG$ijfn_N9!3BbM>=f@)al=GG}fiex}k7E9?sv#9G+CBQm_c!*g
zjJ&wP{V_)cR?;0m7ydGwoKrO*FKEo#zb1VCpf3PpaGp?RqxF{cA+i{E>`GnQ)7G?k
zAer_^BKyP8xypaz7fI5O$pK1fu{PBG@0-*v35>^AFjne7Z53aX3SK9ERa8ioKRo`3
zO-1!ZY>(c#{lbu@B$uUiPaVO8e_NteFkN(N$GQMVSG(u3P$&$CxSTeNP$eUEaK?4~
zmiyPCOYeYA>Q7EotZ`Rzs2kFgwHM>}6i9b-mm2aDOC5h%+x7$+Clnm!=q6V8_)W9K
z@uEMUzsqQ(h?gK7G)jSczgzXTVpHOJFqhAvcRSlFdU4+GTgBo>nzkxU!cKBcK9IJ@
zIZ#zv-I2T$bl3IGkdbCXNk=gOq;aCgjaEbFv|xzefT2y9RzeQ{9TIZAx6g!=->`Oi
zP}MaqxbJ^!1c7%ET7)*&?&-3YN{Qg25bYukE&)}43eNmTR`}zyOX)d6Li&&L-r&n9
zQh5y}OL5DiYxy=+MZcJU@rh~TE0vKRg`!WQ6G6dsSM%kS+Ic1}VcrNk7H(!K`_(r}
zJNE~ClN)jEv?4!SoTgqG?#zRAj%$kqc1cF%I8J9OHHJ*_5@6g
zE6gjzwQRk2e5W&Uh3o1-{p3^$3wi$Q+?|*&_aOO+SC*{K$OZBjPf?#^+q{B>d6%SP
zTGAAgls>)b>Q>vfP$P+%VR+yd0;9}36HjBq(LW-LqxPC
zRb_u0AE(-m=yjYNmA**hotUYCs!FlvXH!?oN!mq^09XWjzh`w-LGSJw1ivR{iZ!2A
zd0yyPp;VD`&`h?pzUT8E{hbw>5|i3&^Px@DB;`umu%?^!t2{izmRT)>UKwlc0oCU+
z9}Et+YxF=SgkAOM+^{X^iUlBR^fxEF{F)U4d=Aycg-qZPaf@Jh`9UK~+U?zXW
z+VxgBMk>V>AfKtG2Mnk5YZg!nj(VsY^4UNvc7VK0s84IAY;ljsU_k;VSTzh
z8wc*Ui~d4W6%kuQ=;IqE!iyUcqd5D0!q!DJjS;q5Gjk^v_Zs^F|1O?i0Lh(YV)*w2
zI&Dh4b2172jwyGPVDh0t2w_LfT}ywfqWXlRk<+jMu%0)56vHP_0x!qdab&>NqEVAD
z0q!cRdz;`lbbE!@gMu4?-z%Mx#+H{~(w%BuC+Tjs*v-=p-E+|cI<6N#z0JTpMl`he#0u`duLbZ#?E9<
zsN1VIcUk|t%A7FkY*SY)jr~JszPXw|U)4cZM?wyah
zCQtUfsZmf#-Tx8I;DCMUy2;o*^U~g1AD2QC
zKRexf;veq%SJ!?sdf(I2;D;5=K}I9f*=~M%0f|z?HLpl4xh
zelT*C)DB$66d
zR*s8oQ2^*}C`cq5gm=qN;MQqf-eO5;PfPN@ddkF16@MkaWwp_gjrx@~FT;c~2>pg+
zNT#HisotAKBkOgO_;#!op(|zl=T~kD22Z%JXXr>%p5gu|%94NZjY>4A)$VGK%Iw0s
zB~&L}@U-#VC+MYPbuxYM)^ShV@D**19&)brfbv%Df$@xic;
zI)wy2z=171`SKjDUy#4@w3*%qcK^Mpoi(ENYgRXsiX8hVpG4VZ2+!*-gpK2r>JNlT
zru$G_<7ir)z2<)+7rPGLnX^G&Cu_Rg`}(EL$PZy+#N`2n{`KU9MR$)aBDj*=3#jDK
zon?ewW)q3N0E6x=RJpgF?2R|V-_nGoK8cov;)73*ZOI4)trNJ*{6Tm$=}(yELorbEaZH{D0tfZ
z$IRQlbV1^oE6lr0wmPBcz`csBVuu?7ci$6QJ;*vTY|ws~YvJB?uC0>PnfsYUBOBD=
z=^%O>k)Afz+EU$9>3Vo&Bw|RO?AYyMjc>GRd;kx
z@!okM)bW3aQrfvLYb^C}?JDJxo})00h}y*I6rcrYzf?;|MEJ_T^hprIQ{D6rXPjL7
z!4*6Y3T+dbcGyc=dtT&?&R7+G(S9XL)rmu8;k9n^qa^5;#l
za}~b<+9=cgzIL)n6)II#esHgA`G*Hyep<+((?&JpggviLr5OJ>`F{Q^q=KNFex5Fh
zy?%ZH*E>^T9_aRVJo@{=ES@466>$JgU`(X+q2?uVJ~WKtdr{g4`_IIod_rrpPlfiv
ze~^EPTf6#!rhatA<>5P*G;_>$JY_2Qc$dzF!s3#4i9fK;g0OIvyZW0x==L~M*#D1qtIJXV4U^Y
z+Q@BA%7iIIkDSwjYqurL`|cjQlxmLol6-%@Zj{3B=Q8|2cG&(au>-y@ZmRf=
z%+@VVEtvd%*PUagB0S-^ZCr7;=?d!kJwasR9JDJ0--mr-9cK_C+l26p;Afu6nzd6tvGBIf(#rskQ&;Fpg_)GmA
zTt$0kLgTMKJ|A~_YqU8`Yh98HTUvigbx_(Rn~ljG8IM1Gdg0MzynR#Mh
z7n;}YV^l=4tT>@}VtQh36~DIej=jw!Ldre!QfYH3(!)bX-St2cS$~(OI1;IM%SHOR~Spaee+GQO~qb4p!SNFx1^!XS`zXFaFf33P-yvB)F2y79z^R2`4HI{6{K|e@ag2Y4LTcgG7b2|`
zu6KIV9c{@{vucb>%w_HI-O0Xhs@h_lLM34OHxb8;UYj;X^o8FZ7R-rh<3%|DXoJy$tNtoO3D!Opx<
zU2YPju)Ceq)^5VvX8eEUxYuJWXF(6Xo$%TG0OyOpfKR#b3I(QeoOWXxA1WxAxTnPK
zXU2R?`KvR*NRl3H$;jK+5wDHz6&6Yz~>wCdtn=a7vHuLze)<=@=#
zohc=EJ}N%7HC+2*$<3)!;Ep=vGejc0wAUQ^v;7uU#q7(*B$88ezEa-
zO&P!9VOf_y>XY>b(j&`_iywaIUT=RGXB|FF{$|lwY&C@#P?E#WNK9VF#^JW(qNCr%aUiR1UK8$F_TU4j+onh_?
zmZ&fE`XGOGnM%HL_==9nr$x?bB2M4xK!WsCS*a$~Ac*b|XtyPHX|VfIzQ;p8xt)f)
z5xTFs(!K*%AWAxtV+q&9ueyzXm{n5_f&fd;KOkpB0|%
zsaqPW1$?djy0C+ln?oH$?VbHsE1XWoItkmh_cVW{8=Wqf8_^{l&=iMjGfsAs??c*%h
z=jMNOUffp_G2y4Z;C>j@EXJ0nb#CVGXY2I^capKCI2h3d_9;T&6ZYtD@(oQD5Jr^ij(~MSRE?c};FC=I
ziN1x)JgX2M+Tu5r$nNp6_xC+yL@w
zP73wcXYDTyK6^{C2$>3B`Ht4dgZEfM(6{WiNfh
zn9KE6)T7?6%wr$Nws0>hlz880#3R&-t>FD*8*ljfz4h*sa2@laxxv>>W4bBe?$ZpB
zY0v&oLixaFUIM)B86u^FNQqr`BJO|Jac}4w65DQc@QHAApOU^T48sRHgex=uP^QqS
z9U*i4LR3>L3gjut%;y|*qUMfyACx#KTX_ji=Kg}X_~rOJc6ksDZ)P(R@4?3kYrgu>
z(a^`wz54Ewnx{XTTXx<23eau9DB7-PYo=*@dQ}^I+wJ^RoV!MS>|qOY*cX2#BJ6IrHi6+e%!{atmH7KzIprs2C}$|?KAIG-->5WE0;HSgJ4Y+22Wc+d|!
z?7gmDjTHI#q|*(xfD`AOKKH4WX}Jjp|7Lc9Ht%VI#&G3O-5KuNk&w40aa*dvFF`l_
z>u+eutk<HY)QGiNf%$wT$j7K}^!Kq}0SKkqgVI4?RU-pHeE-M4v=m{Zba64pv_^
zDSZQ_bE~d}FI%i>&mRnOVb~Ar^Xk3;Teq-dJ
zo3r}HM(S&}pqdGW=(Ql36{V>6RPlIzR+Ztpo}phVreQa2k0=WrzX%o{pKsLgEoojU
z4?P>a2ftMCtf4dHW!Y`FE{oc`zA-wsuH(a$M+^*)Ht30lkc@*rvY0%msO;ke;M6!R
zu3>uC3Ap>^6aw)qM3aBqX*|U~geUi^IXZOtJ%-S>u1`U5XhkDxJOZ7VPO#d^9$(Z@
zoB2w@xBi4XWS&V{f>-=&ZmL
z4@y3C)KpnM-fUugnzG7dPr>M5^s#s&XLofaQ@(D{s>J5*C@eGW+2$@o;)=wxTj5t*
z=F@x&cm}%hrZ|g3T{RXzGanlT8hsTB8p$KO72RlWv0Lq=;hZ<@)VqUU@=X4ck1&AE
zNP8HGBO^6E`h$Nk;RE?qTm<;4{#V=gUpgf}37t~TufG50c`ftt;_B=7Df>NWFKKy3xDN7Sei<}|Su61{B#VgfaNK9Q5Q|YNZ
z=x&b+udcCqVBR2*O+nIod3U>HKb0$thm+@{*0#A-0cC&qS{^GMMuSF<=OKF{-dp@V^jp3nS44?g4T+xC1Lw{W5
z{pB!-zk^uZ5Nn-^VT*12dL?rY&q2t%{DRRVzfvBz4N%Bb`zCEVWSHvSt~QiFcqM*?
zY2eHxsw02FjZ>)l%J%qJ&HhAaDbqR5sk(uVFI+$jI;7VDcoB#0$-1sjqt0;uDLheb
z+aKp-kP>*0de06Me>z6FN|cJK3bawlpBydkxU6X8IjYF-=lQL@+iYoo^Epf)A!PGe
zZ2D3VwSsTRn+=5^C4z`HQnjg@P0@PHNuy=O!KwbaCMGCFGe>sv}pve+9A(<&^
zC1W!C(jourfyY1sveT%TM~DgSFD13icSLNidSqMDzgxI%`>Ukxzp+7S?fJnSkX=5`Pn5(>aT~nkSHdf-H>zBtewWpO@KQ_#J0$
zbuCgV@?1fG#!mx>@YYplILQG`fx|w}x_8gI_dNPt7F~EwP%|KJL#@VbStX9sbqg->
zXkxgqIF8_KFSl;3zy+Ft+PiOh@M$9VRU;e_(y-?4bk&!bT~LRS?g6j0qEqdu56U>@2Qrfd@?uZd1Hbn^-o(GOvJ7yPI}+WGdQSF{
zE?n7cxu$QwhC*gGYmk*q0-iPhb|24uY*Au>g_TtVusBetafuSuU1M%aw3m_@m;A(%
zw|H=W#yC;U1ldJ1l~MXW_Tl2OZhV|u<31UI^HeTL7v
zjQj!1Hx#r-b{c%jc3+5P?2=3SnZ=+2<$!M}^HmDfQE8uwbUQs&a>7d%QzvAlLk?CD
zH-_P?3t}9PVwL_6W@snZlYyNRhfK)gk5`Oou(BpEpK%=`6p8yhS
zCyH};B?M8D!&+PtWnd`dnkeT~hjC6cGXM-G)dEIHMznIU0=zH`KtSCZn8Xm1z<3w{
z&4_>)Arvf5jD-FmvC&F&7}PC1;FWO(P%2^>RA70_9ROj(N+k$)#3}3*@sz|JY-WGZ!4(GQp3Wh)*tFh#;4SwOrcD_hy_JA+{Q~<>~
zxd(DHx2I4^d!j7kyXXFvqy00?7eAF0du-mK9V$MV#TJq5X`qkoIRYjCWy1wLfk?pS
z63~v&&`-fgpe$i9QaiS1P*D89mvHF=6@LZ>d~-nIpI{z)Hj=q4c|4)Vyv3E1D#l>3
zvdj&Xg+gpNI<%I2_*R{mvhI`xcFS>#zW81mlpE{|pI$S|cY!4=;jdr&Yv+sRFdlxL
zFnjZ-@`4M=aTZ^W+e6uU39&AHT}C~Y@5G`M-u%k5oY6l^8(3JK5)2seb%_ia=YK#K
zHuQzj0yeAlIzmoRk=Wazd<-~*4Pk8pBeYR_xM8&K<~vv^(Jr*bseIWAhr|MD#&WNI
z>HFcr3|&+xh9@|W1!}cqBUpuW%Xd=E=Rw_(OoK-neI=z*>uAGyo%{KD-8Fhwk_t4NF0c4?EPta6R4ujp%f3=*Hvcl4B=(f9NMWWS;*y`^2+
zqZz%U_w<22JdGZ*0nO?E=zrg|pnW=^Lpq|Pbes)}>~0!LV_bxevT;7_(J_!Wr8D|{
z6@h9S0rDd9Jh&?zte)IVdh<`mzrw^|mr`ceV
z579y|vin()j#qOG(FNG3#t03AFs)RVlldylHcoqaH!X0Y$qZCF1=4!?WK^V2Rb{ws
zlyP+U?q>gEsf@tXCVx#pR25uf0wM8w*C?Lp8g=in7nZ_sm7z8`1lG#i5n
z7qm$Gx4o1W*<=F7+@CG&d77t1FTcBsox|LgS2+lXaJQiD`26VV;<`eeS-z90~m10$oVpj#a!|NZds?a5(ul&lp}Lc|X{6@ND>rcI1joMH7evClhA
z)@;_;7alLX;n~6;Ugr0}z!m-A{o*I@8KYfV*<@&8`M5jl-xaX^>fX~YPqTgx?I-&-
zOADGkcER}2uj2ol`v2d-(H8V>I2*4%!}@7=3=_k@e2{^q4&m8;rXe+R^Bonl33%jh
zvuuK!9|}X*1Ao4!x9sdYEHud<7dAe97-t!s=0y*lKZfAHXMw4A8&5v=AqYHhw_0H=7L&!>rlUJ9agkj2XYI7SB2Z^u)6;1!|})WbKy4Z
zZwaX=2*>dvfTM8omxVL!rXOt3ix~6_8g8m`tJC20`hV@)zYgZo+6~ewT}mt63#roR
zH5>g9Hu#C(+!u=gXA#=qcB^c*5ZJ_fEDjLEZfLNjw>khm9UZ@ae=%pvZD32zZNpzL
zF>ZgEEtjli>s90mFTnveIP?A~k3zs@@LUZk^j62k`G?EvqqpU`c=^=79Tt;~(VZi^
zKDfh6BY$HeOM-Lq=Z;3%&Q3bGkI#T*H8|6BG#U>d!}dzYOsuCugsJKfOFSZM%m6k#;3d2p1ABQKOhqYT4>E+OnR?p3ou|={3EmNMcsF
z!qwVKP;
z($Q*aTPCN(+C~UF9JHaQW)PX2aP9ht%^#&>=MclMLUixaX}nV*eE%
zx;T22k0(>~ghZPTeV3LM#v!@>nD?d+6DZr&j(y$wZC9{$7`hGv-Slc0{yR$OJ6Edz
zhHz_OZC9IhV73m7+g4!wC4jBDX@96;Qw`D(&1!5yy#%)Q+^7ak^M*cU42_uxntTap
z?TJVonp_W!vk01g324u_?5%4yX=t{>Uc4zAisf>9h?}lJ>-ZY#Mb=uphG_uQ+XLKi
z(OAdV(6MrD1Pc2S_%P+>|-h@!56wV)crbK!3L>+o_>5
zjs1b&D=z61k$U+{;HzgTbv%vzpksTpF<R(YHEF!;m@
z3BHB|kxBRxk%X@%NVtjn`hQcChlxTXu}Sy{Yq`+Kl4R5Lylr|&B*GlbZkhw{62hQx
zE|HK_6mCkD*1B+VKrw1xcT@Ey_It~t0n!0Qps5Eui
zI1D$o9kJc1c5X+^flyB?>eB*;(nQSh-hNxjmREj}_zAIq8)~R<2MUb@)fJ;gdg8^-Zwlrean;3rsWt
zL>JHxZ{ZuU!C?xy%kaJv4$0-Y8uZ3&8NcD%(_t^8A12xPY+B?4JWML|P1
zL^nf4Ha0~uGCm+YML0x7L_s$)MK?7@L`60>MMgA5K|?o0H$z1>HbpTqK3xhgOl59o
zbZ8(nI5C%>^aCq@%~Ot17{~GJcb`u6^!6wTjJy9u-kRl<9-pIQomi_=^jD@ws
z*m~RBSXfxFG8T=kof&KV0j!uw-)r-k-?^{r-1mLvI%Z~TnOQs!7hn-g_N-_mLAA{C
zzRV!3leWNS=z@0G1#PgEz6yv?2~`wYp%aQc+W9vUD|wS%+?
z8fDFyZ)<{$R9m3HvYE6EwzKpfYLOM5_zr$N`Vn@JyrosvQ}8NBi0qC2D$W-DyBsxk
z%7zBL&-eQ8a`M;%McL?IuXe+3zTv4KO0ubvS9@U(?1etq4+FBrCU4mXgD?b#pe$Qi
z^6CK?hJ!GF0>|L6?Czwu9Dz|7gQGAB+*6hGZPin?vO8r@Ic2Y3d8$$N
z>w%|Qh1pM@>J;9$d8$`fedwt{;p